Transactions of the American Institute of Mining, Metallurgical and Petroleum Engineers

Name: Transactions of the American Institute of Mining, Metallurgical and Petroleum Engineers
Author: American Institute of Mining, Metallurgical, and Petroleum Engineers

American Institute of Mining, Metallurgical, and Petroleum Engineers

Transactions of the American Institute of Mining, Metallurgical and Petroleum Engineers

American Institute of Mining, Metallurgical, and Petroleum Engineers · 1871 · Historical

Transactions of the American Institute of Mining, Metallurgical and Petroleum Engineers by American Institute of Mining, Metallurgical, and Petroleum…

Public-domain full text preserved in the Mountain Man Mining Library. Original source: archive.org.

at| : . com/

ij

The Branner Geological Library

LEILAM])''SIMFOMD 'OWIORVKJIVEKSliTTT

philadelphia: an a ooi, pbintbbs.

Preface.

The present volume, and Volume XXIII., to be issued immedi- ately, contain the papers and discussions of the Chicago Meeting of the Institute, held in August, 1893, and constituting Divisions C and D, devoted respectively to Mining and Metallurgy, of the International Engineering Congress, held in connection with the World's Columbian Exposition. No attempt has been made to separate in publication the papers of the two Divisions, or to arrange them in the order of their presentation at the meeting. They appear in these volumes in the order in which they have been made ready for final publication.

Volume XXIV. will contain additional discussions of these papers, held at subsequent meetings of the Institute.

R. W. R.

May 1, 1894.

Contents.

Paoe

Opficeks, . \iil

Publications, ix

Pboceedings.

Chicago Meeting, being part of the International Engineering Ck>ngre88, August,

1893, xiii

Papees.

Tests of Hydraulic Material, by H. Lb Chatelieb, 3

Geological Distribution of the Useful Metals in the United States, by 8. F. Em- mons (See Discussion, p. 732), 53

Mining and Mineral Statistics, by C, Le Neve Foster, 95

Segregation and its Consequences in Ingots of Steel and Iron, by Alexandre

PouBCEL (See Discussion, Physics of Steel," vol. xxiii.), . . . .105

Note on Experiments on the Specific Gravity of Gold Contained in Gk)ld-Silver

Alloys, by Henby Louis (See Discussion, p. 724), 117

The Detection and Measurement of Fire-Damp in Mines, by G. Chesneau (See

Discussion, p. 725), 120

The Le- and Zinc-Deposits of the Mississippi Valley, by Walter P. Jenney

(See Discussion, p. 621), 171

General and Special Observations Concerning Ore-Dressing, by O. Bilhabz (See

Discussion, p. 699), 225

On a Remarkable Deposit of Wolfram-Ore in the United States, by Db. Adolf

Gublt 236

Microscopic Metallography, by P. Osmond (See Discussion, " Physics of Steel,"

vol. xxiii.), 243

The Bessemer Process as Conducted in Sweden, by Pbof. Richabd Akebman

(See Discussion, p. 661), 265

The Origin of the Gold-Bearing Quartz of the Bendigo Beefs, Australia, by T. A.

RiCKABD (See Discussion, p. 738), 29

Summary of American Improvements and Inventions in Ore-Crushing and Con- centration, and in the Metallurgy of Copper, Lead, Gold, Silver, Nickel, Alu- minum, Zinc, Mercury, Antimony, and Tin, by James Douglas (See Discus- sion, p. 647), 321

The Open-Hearth Process, by H. H. Campbell (See Discussion, p. 679), . . 345

The Bertha Zinc-Mines at Bertha, Va., by William H. Case (See Discussion,

p. 696), 511

Blowing-Engines, by Julian Kennedy (See Discussion, p. 709), . . .537

An Improved Hanging Compass, by Guy B, Johnson, 543

Microstructure of Steel, by Albebt Sauveub (See Discussion, "Physics of

Steel," vol. xxiii.), 546

The Mineral Deposits of Southwest Wisconsin, by William P. Blake, . . 558

The Separation of Blende from Pyrites : A New Metallurgical Industry, by Wil- liam P. Blake (See Discussion, p. 723), 569

Vi Contents.

Page

Improved Slag-Pots, by H. A. Keller (See Discossion, p. 675), 574 A FurDace with Automatic Stoker, Trayelling Grate, and Variable Blast, In- tended Especially for Burning Small Anthracite Goals, by Ecklet B. Coxe, 581 The Hydrogen-Oil Safety-Lamp, for Lighting and for Accurate and Delicate Detection and Measurement of Inflammable Gas and Vapor in the Air, by Prof. Frank Clowes (See Discussion, p. 725), 606

Discussions.

Discussion of paper of Dr. Jenney (See p. 171), 621

Discussion of paper of Mr. Douglas (See p. 321), 647

Discussion of paper of Prof. Akerman (See p. 265), 661

Discussion of paper of Mr. Keller (See p. 574), 675

Discussion of paper of Mr. Campbell (See p. 345), 679

Discussion of paper of Mr. Case (See p. 511), 696

Discussion of paper of Oberbergrath Bilharz (See p. 225), 699

Discussion o( paper of Mr. Kennedy (See p. 537) 709

Discussion of paper of Prof. Blake (See p. 569), 723

Discussion of paper of Mr. Louis (See p. 117), 724

Discussion of papers of Prof. Chesneau (See p. 120) and of Prof. Clowes (See

p. 606), 725

Discussion of paper of Mr. Emmons (See p. 53), 732

Discussion of paper of Mr. Rickard (See p. 289), 738

Officers.

For year ending February, i8g4.*

HENRY M. HOWE, Boston, Mass.

THOMAS M. DROWN, Boston, Mass.

DAVID T. DAY, Washington, D. C.

JOHN STANTON, New York City.

(Term expires February, 1894.)

A. J. BOWIE, jR,f San Francisco, Cal.

ROBERT G. LECKIE, Londonderry, N. S.

E. G. 8PILSBURY, New York aty.

(Term expires February, 1895.)

JAMES C. BAYLES, New York aty.

J. A. PORTER, Durango, CoL

J. C. SMOCK, Trenton, N. J.

(Term expires February, 1894.)

H. L. HOLLIS, Chicago, HI.

GEORGE W. GOETZ, MUwaukee, Wis.

CHARLES KIRCHHOFF, New York City.

(Term expires February, 1895.)

H. H. CAMPBELL, Steelton, Pa.

W. L. SHEAFER, Pottoville, Pa.

A. M. SH(X)K, Tracy City, Tenn.

(Term expires February, 1896.)

Secretarg. Sreastirer.

R. W. Raymond, Theodore D. Rand,

13 Burling Slip, New York City. Philadelphia, Pa.

The following officers were ''elected at the Annual Meeting, February, 1894 : Pnddenty John Fritz, Bethlehem, Pa.: Viu-Presidenii (to serve two years), J. F. Holloway, Cuyahoga Falls, Ohio, J. C Piatt, Waterford, N. Y., E. V. lyinvilliers, Philadelphia, Pa. ; Managers (to serve three years), T. A. Rickard, Denver, Col., H. O. Hofman, Boston, Mass, John A. Church, New York City ; TrtoKwrer Theodore D. Rand, Philadelphia, Pa. ; Seerelaryf Rossiter W. Raymond, New York City.

Publications.

The publications of the Institute comprise :

1. The minutes of the Proceedings of each Meeting, published in pamphlet form.

2. Such of the papers presented or read by title at each Meeting as are furnished by the authors and approved by the Council for full publication. (In nearly aU cases in which papers, the titles of which appear in the Proceedings, are not subsequently published, they have been withdrawn by the authors.) These papers are pub- lished separately in pamphlet form, and are marked " Subject to Revision." A small supply, beyond the edition distributed, is re- tained to meet subsequent demand. There are no copies on hand of papers read before 1880. The stock is nearly complete from

3. Annual volumes of Transacticms containing the list of officers, rules, etc.; the Proceedings; and the papers, revised for final pviUca- tion, (In this revision after the preliminary publication, authors are permitted to use the largest liberty ; and the changes and addi- tions made in papers are sometimes important It should be borne in mind, by those who study or quote a paper in the preliminary edition, that they may not have in that form the ultimate and de- liberate expression of the author's views. It should be added, how- ever, that in the majority of cases there is no essential change, the correction of typographical errors and additions of later informa- tion being the usual alterations.)

4. Special editions of separate papers, for which there is demand. These are fully revised, and are usually issued in pamphlet covers.

5. Books. Under this head the only publications are an Index to Vols. I. to .XV. inclusive ; an Index to Vols. XVI. to XX. inclusive ; a Glossary of Mining and Metallurgical Terms ; and a Memorial of Alexander Lyman Holley.

All the foregoing publications are sent free to members and asso- ciates not in arrears at the time of publication. They are also for sale*at the office of the Secretary, or are sent to purchasers by mail or express, charges paid, on receipt of the price by the Secretary, as follows :

Publications.

Classes 1 and 2, above mentioned, at the following rates :

No. OP Pages.

Single Copies.

10 Copies.

20 Copies.

8 orlesH

So 06

8 to 12 IndiisiTe

12 to 16 "

16 to 20 "

20 to 24 "

Over 24 "

The following papers contained in Vols. XXII. and XXIII. are for sale at special rates : " Genesis of Ore-Deposits," by Prof. Franz Posepny, price $1 ; " Microscopic Metallography," by F. Osmond, price 30 cents ; " Micro-Structure of Ingot-Iron in Cast Ingots," by Prof. A. Martens, price 50 cents ; " Micro-Structure of Steel," by Albert Sauveur, price 30 cents ; " The Open-Hearth Process," by H. H. Campbell, price $1 ; " Iron Alloys with Special Reference to Manganese Steel," by R. A. Hadfield, price 30 cents.

Class 3. This class now comprises twenty -three volumes of Trans- actionsy for sale as follows : Vols. I. to IV., inclusive, at $2 each ; Vols. V. to VIII., inclusive, at $3 each ; all other volumes at $5 each, except Vol. X., of which the supply on hand is smallest, and the price of which is $10. These prices are for. paper covers. Half- morocco binding, $1 extra per volume.

Class 4. This class now includes " Steel Rails " (Papers by Messrs. Sandberg, Dudley and HoUey, and discussions at two meetings in 1881, from Vol. IX. of the Transactions), price $1 ; " Technical Edu- cation" (Papers and discussions at the XVIIth [Philadelphia] meeting, in 1876 — mostly not in the Transactions), price 50 cents ; " List of Members, Rules, etc.," price 25 cents.

Class 5. Index to Vols. I. to XV., inclusive, of the Transactions, price, in stiflf cloth covers, $1, in half-morocco, $2. Index to Vols. XVI. to XX., inclusive, price in paper covers, $1. Indexes I.-XV. and XVI.-XX., bound together in one volume, cloth, $2.50. " Me- morial of Alexander Lyman Holley," in cloth, with frontispiece- portrait, price 81. "Glossary of Mining and Metallurgical Term," by R. W. Raymond (from Vol. IX. of the Transactions) in cloth, price 50 cents ; " The Genesis of Ore-Deposits " (Paper of Prof. Franz Posepny, with discussions), price in cloth, $2.

Authors' Editions.

Extra copies, when ordered before the types have been dis- tributed, are furnished to authors under Rule VIL, at the following rates:

Publications.

No. OP Pages.

50 Ck)PIE8.

100 Copies.

250 Copies.

Each addi- tional lOOcopies above 250.

4 or less

$150

4 to 8 inclusive

8 to 12 "

12 to 16 "

16 to 20 "

20 to 24 "

24 to 28 "

28 to 32 "

Covers (Including print- ing on first page of the same), extra.

When a paper coDtains one or more separate plates or " folders/' these will be charged in reprinting as follows : One page or one fold, the same as four pages in the above table ; each additional fold, the same as four additional pages. These prices are for plates on the ordinary paper, used in the edition issued subject to revision." If special bank-note paper is desired, such as is used in the Volumes of Transaetioru, the price for the plates will be doubled.

All communications and remittances should be addressed to R. W. Raymond, Secretary, P. O'. Box 223, New York City.

Proceedings Of The Sixty-Fifth Meeting, Being

Part Of The International Engineering

Congress, Chicago, August, 1893.

General Committee of the World's Congress Auxiliary ON Engineering Congresses.

E. L. Corthell, Chairman; O. Ghanute, First Vtee-Chairman ; D. J. Whittemore, Second Viee-Chairman ; Max E. Schmidt, Secretary; C. L. Strobel, John W. Cloud Thomas Appleton, Ira O. Baker, H. L. HoUis, F. W. Grogan, William Forsyth Robert W. Hunt, Joseph Hirst, F. A. Ross, William Sooy Smith, W. W. Curtis W.L. Marshall, Allan Strale.

AnoetaUd Engineering HeadquarterslO Van Buren Street, Chicago.

Special Columbian Committee of the Institute.

John Birkinbine, Chairman; David T. Day, Secretary.

Engineering Headquarters m the Columbian Exposition — Mines and MiAng Build- ing.

Sessions were held in the Memorial Art Palace, Michigan Ave- nue and Adams Street.

The Congress comprised the following Divisions :

Division A. — Civil Engineering, in charge of the American So- ciety of Civil Engineers.

Division B. — Mechanical Engineering, in charge of the American Society of Mechanical Engineers.

Division C, — Mining Engineering, in charge of the American In- stitute of Mining Engineers.

Division D. — Metalluiical Engineering, in charge of the Ameri- can Institute of Mining Engineers.

Division E, — Engineering Education, in charge of a Special Com- mittee, Prof. I. O. Baker, University of Illinois, Champaign, 111., Chairman.

Division F. — Military Engineering, in charge of Major Clifton Comly, U. S. A., Grovernor's Island, N. Y.

Division O. — Marine and Naval Engineering, in charge of Com- modore George W. Melville, Chief Engineer, U. S. N., Washing- ton, D. C.

The opening session was held on Monday, July 31st. Mr. C. C.

Xiv Proceedings Of The Chicago Meeting.

Bonney, President of the World's Congress Auxiliary of the World's Columbian Exposition, and Mr. O. Chanute, yice-Chairman of the General Committee on Engineering Congresses, made brief addresses of welcome, to which responses were made by Sir Benjamin Baker, Vice-President of the Institution of Civil Engineers, England; Baron de Rochemont, Inspecteur des Ponts et Chausses, France ; C. O. Gleim, representative of the Associated Engineering Societies of Grermany; Hugo Tesola, Chief Engineer of State Railways, Austria ; Professor Alfred Nyberg, Institution of Ways and Com- munication, St Petersburg, Russia ; Cavaliere Celso Capacci, Royal Italian Commissioner; William Metcalf, President of the American Society of Civil Engineers; Eckley B. Coxe, President of the American Society of Mechanical Engineers ; Henry M. Howe, Presi- dent of the American Institute of Mining Engineers ; Professor Ira O. Baker, University of Illinois; Lieut. Henry L. Harris, U. S. A., representing Major Comly, Chairman of Division F; and Commo- dore Geoie W. Melville, U. S. N.

Divisions C and D then held a joint session. President H. M. Howe in the chair, at which the following papers were presented :

*The Growth of American Mining Schools, and their Relation to the Mining Industry, by Prof. S. B, Christy, Berkeley, Cal.

Mining and Mineral Statistics, by C. Le Neve Foster, Llandudno, Wales.

♦Genesis of Ore-Deposits, by Franz Posepny, Vienna, Austria.

Geological Distribution of the Useful Metals in the United States, by S. F. Emmons, Washington, D. C.

Lead- and Zinc-Deposits of the Mississippi Valley, by W. P. Jenney, Deadwood, S. Dak.

Mineral Deposits of Southwest Wisconsin, by W. P. Blake, Shullsburg, Wis.

Origin of the Gold-bearing Quartz of the Bendigo Reefe, Aus- tralia, by T. A. Rickard, Denver, Col.

A Remarkable Deposit of Ores of Wolfram in the United States, by Dr. Adolph Gurlt, Bonn, Grermany.

Division C subsequently held separate sessions, as follows : Vice- President E. G. Spilsbury presiding, and Prof. W. H. Pettee, Ann Arbor, Mich., acting as Secretary.

Printed in yol. zziii.

Ic

Proceedings Of The Chicago Meeting. Xv

Wednesday Morning,* August 2d.

Papers presented :

The Detection and Measurement of Fire-Damp in Mines, by Prof. G. Chesneau, ficole des Mines, Paris, France.

The Hydrogen-Oil Safety-Lamp, by Prof. Frank Clowes, Not- tingham, England.

fExperimental Investigations on the ''Loss of Head" of Air- Currents in Underground Workings, by D. Murgue, Besses, France.

Thursday Morning, August 3d.

Papers presented :

The Bertha Zinc-Mines, by W. H. Case, Bertha, Va.

General and Special Observations Concerning Ore-Dressing, by O. Bilharz, Berlin, Germany.

An Improved Hanging Compass, by Guy R. Johnson, Long- dale, Va.

Friday Morning, August 4th.

Papers presented :

A Furnace with Automatic Stoker, Travelling Grate and Vari- able Blast, Intended Especially for Burning Small Anthracite Coals, by Eckley B. Coxe, Drifton, Pa.

Tests of Hydraulic Materials, by Professor H. Le Chatelier, Paris, France.

fElectricity in Mining, by Francis O. Blackwell, Lynn, Mass.

The Division was then adjourned, to meet with Division D as a general meeting of the Institute for the election of members.

Division D held separate sessions as follows. President H. M. Howe, presiding, and W. B. Kunhardt acting as Secretary :

Tuesday Morning, August Ist.

Papers presented :

Summary of American Improvements and Inventions in Ore-

The session of Division C was omitted on Tuesday morning, in order that the members might attend the session of Division B, in which the report of the Standing Committee of the American Society of Mechanical Engineers on " Tests of Ma- terials '' was presented.

t Printed in vol. zxiii.

Xvi Proceedings Op The Ohicaqo Meeting,

Crushing and Concentration, and in the Metallurgy of Copper, Lead, Grold, Silver, Nickel, Aluminum, Zinc, Mercury, Antimony and Tin, by James Douglas, New York City.

Note on Experiments on the Specific Gravity of Gold Contained in Gold-Silver Alloys, by Henry Louis, Singapore, Straits Settlements.

*The Limitations of the Stamp-Mill, by T. A. Rickard, Denver, Col.

Improved Slag-Pots, by H. A. Keller, Butte, Montana.

The Separation of Blende from Pyrites ; a New Metallurgical In- dustry, by W. P. Blake, Shullsburg, Wis.

Wednesday Morning, August 2d.

Papers presented :

Microscopic Metallography, by F. Osmond, Paris, France.

♦Microstructure of Ingot-Iron in Cast Ingots, by Professor A. Martens, Berlin, Grermany.

Segregation and its Consequences in Ingots of Steel and Iron, by Alexandre Pourcel, Paris, France.

Microstructure of Steel, by Albert Sauveur, South Chicago, 111.

♦Further Observations on the Relations between the Chemical Constitution and the Physical Character of Steel, by William R. Webster, Philadelphia, Pa.

*Heat-Treatment of Steel, by H. M. Howe, Boston, Mass.

Thursday Morning, August 3d.

Papers presented :

The Bessemer Process, as conducted in Sweden, by Prof. Richard Akerman, Stockholm, Sweden.

The Open-Hearth Process, by H. H. Campbell, Steelton, Pa.

*Iron-Alloys, with Special Reference to Manganese Steel, by R. A. Hadfield, Sheffield, Eng.

♦Consumption of Fuel in the Taylor Gas-Producer Plants at the Aspen and Marsac Mills Compared, by C. A. Stetefeldt, San Fran- cisco, Cal.

Friday Morning, August 4th.

Papers presented :

♦Review of American Blast-Fumace Practice, by E. C. Potter, Chicago, 111.

New Process for the Production of Pig-iron, Refined Iron,

♦ Printed in vol. xxiii.

PB0CEEDINQ8 OF THB CHICAGO MEETING. Zyu

Ingot-Metal and Weld-Metal, by Alexander Sattmann and Anton Homatschy Donawitz, Austria.

♦Sulphur in Cast-Iron by W. J. Keep Detroit, Mich.

Blowing-Engines, by Julian Kennedy, Pittsburgh, Pa.

♦Recent Advances in Pyrometry, by Prof. W. C. Roberts- Austen, London, Eng.

On Saturday morning, August 6th, was held the concluding gen- eral session of the International Congress, Mr. O. Chanute presid- ing, at which reports were presented by the representatives of the different Divisions, and numerous brief addresses were made by dis- tinguished foreign engineers.

Membebs and Associates Elected. The following persons, duly recommended by the Council, were elected at the joint session of Divisions C and D.

Members.

O. 6. Amsden, Needles, Cal.

Bichard G. Anderson, Deadwood, So. Dakota.

Francis H. Backhouse, Sydney, N. S. W.

Frederick W. Bacom, Marysville, Mont,

Edwin Ball. Palmer, Mich.

G. H. Bamhart, Mammoth, Arizona.

G.F. Becker, Newport, B. I.

J. F. Beny, Johannesbarg, Transvaal.

Willard D. Bigelow, Washington, D. C.

George D. Blood, Park aty, Utah.

Charles E. Bowron, Tracy City, Tenn.

Arthur R Call, . Amador City, Cal.

Prof. Edward D. Campbell, , Ann Arbor, Mich.

W. J. Collins, Melbourne, Australia.

James M. Colwell, . . . ' . . Kittanning, Pa.

John Daw, Jr., Brooklands, Acton W. England.

De Witt C. Doney, South Chicago, III.

C. H. DooUttle, Chicago, ni.

Max Duchanoy, Paris, France,

Wm. a Edwards, Calico, Cal.

Lester H. Ely, Drifton, Pa,

George W. Eustice, Puerto Berrio, Colombia.

Louis Faragher, Cape of Qood Hope, So. Africa.

Charles Ferry, Troy, N.Y.

E. J. Gilbert, Ely, Minn.

William F. C. Hasson San Francisco, Cal.

J. D. Hawkins, Aspen, Colo.

August Heinbeck, . ... . Gunderson, Mont.

A J. Higgin, Adelaide, So. Australia.

William L. Honnold, Houghton, Mich.

E. J. Horwood, Broken Hill, N. S. W.

John Howell, Broken HUl, N. S. W.

Printed in vol. zziii.

Xviii Proceedings Of The Chicago Meeting.

George P. Hyde, Joliet, 111.

John A. Jardine, Hartranft, Tenn.

Frederick D. Jones, Toangstown, Ohio.

Beiji Kanda, Sado Mines, Japan.

O. W. Kennedy, Scottdale, P*l

Henry M. Kurtz, Clearfield, Pa.

T. Benton Leiter, Sheridan, Mont. -

Ernest Lidgey, Ballarat, Australia.

C. H. McCullougb, South Chicago, 111.

G. W. W. McKinnon, Broken Hill, N. S. W.

J. F. Martin, Gawler, So. Australia.

Charles L. Miller, Chicago, 111.

Balph Nichols, Hailey, Idaho.

F. J. Odling, Melbourne, Australia.

F. H. Oliphant Oil City, Pa.

Edward Payne, Ballarat, Australia.

J. F. Pearson, Broken Hill, N. S. W.

Willard B. Bising Berkeley, Cal.

William B. Sands, Jr., Sparrow's Point, Md.

D. A Scheidel, San Francisco, Cal.

R Shepherd, Broken Hill, N. S. W.

F. Stubbs, Sheffield, England.

€}eorge D. Swift, Iron wood, Mich.

W. F. A. Thomae, Zeehan. Tasmania.

James J. Tonkin, Linares, Spain.

Dean S. TumbuU, Sheet Harbor, N. S.

E. W. Walter, Silverton, Colo.

Utley Wedge, Cleveland, Ohio*

R H. Williams, Toungstown, Ohio.

H. J. Williams, London, England.

Luke Williams, Hobart, Tasmania.

Harry Wilson, Broken Hill, N. S. W.

Associates.

E. T. Bolles, Denver, Colo.

Elmer Turner, Cleveland, Ohio.

Associates made Members.

L. K. Davis, Carl Eilers.

Excursions and Entertainments.

In the general excursions and entertainments connected with this meeting the members of all the Divisions of the International Con- gress participated. They comprised a visit to the Exposition on Monday afternoon, a trip on the whaleback steamer on Tuesday afternoon, a visit to the " two-mile crib " in Lake Michigan ; special receptions at the Associated Engineeering Headquarters on Monday and Saturday evenings; a reception given on Friday evening by Hon. F. J. V. Skiff, chief, and the exhibitors of the Department,

Pbooeedings Of The Chicaqo Meeting.

X\X

in the Mines and Mining Building at the Exposition, and a visit on Monday, August 7th, to the Chicago Drainage Canal. As will be seen, the sessions were held in the forenoon exclusively, leav- ing every afternoon and evening free for social intercourse. The headquarters in the Mining Building,* as well as those at 10 Van Buren Street, were extensively patronized by members and guests.

Registry.

The attendance at the Congress in its various divisions, is esti- mated to have been about 1000. Between 700 and 800 persons reg- istered during the week at headquarters ; many others, previously roistered, were doubtless still in the city ; and many residents of Chicago had not registered at all. The following is an incomplete list of members and guests of the Institute who were present :

A. v. Abbott. W. S. Ackerman. Bobert AUison. P. ArgaU. F. E Bachman.

D. Baker.

H. P. BeUinger. T. Bergendal. John BirkiDbine. W. P. Blake. A. BoDzano. Samuel Brady. W. H. Bradley. A. E. Brown.

E. D. Campbell. H. H. Campbell. W. H. Case.

H. M. Chance. J. P. Channing. S. B. Christy.

F. W. Clark. T. C. Clarke. F. L. aerc.

W. B. Cogswell. H. A. Cohen. F. Collingwood. J. M. ColweU.

E. S. Cook.

F. K. Copeland. E. B. Coxe.

A. C. danningham. Gteorge M. Davidson. F. W. Davis. David T. Day.

J. E. Denton. H. P. Dickinson. E. v. dInviUiers. W. H. Dodge. T. M. Drown.

E. L. Dufonrcq. W. F. Durfee. D. Eagan. Henry Englemann.

B. F. Fackenthal, Jr.

A. W. Fiero.

C. H. Foote. J. E. Forcyce.

C. Le Neve Foster. P. Frazer. H. C. Freeman. J. Fritz.

F. Glein.

G. W. Ooetz, C. W. Goodale. F. W. Gordon. H. B. Griffiths.

B. A. Hadfleld. F. A. Halsey. A. Hardt.

H. K. Hartzell.

These headquarters, comprising sitting-room, reading-room, library, etc, were placed by Hon. F. J. v. Skiff, Chief of the Department of Mines and Mining, at the disposal of the Institute, and tendered by the latter, with his consent, to the Associ- ated Engineering Societies and their gaeets.

Proceedings Of The Chicaoo Meeting.

E. C. Hegeler. J. W. Hegeler. G. C. Henning. H. B. Herr.

G. C. Hewitt. H. D. Hibbard. R. W. Hildreth. H. L. Hollis. L. Holmboe. H. M. Howe. H. W. Hughes. A. E. Hunt.

C. W. Hunt.

F. R. Hutton. J. C. I'Anson. J. T. B. Ives.

D. S. Jacobas.

G. R. Johnson. T. J. Jones. W. J. Keep. J. a Kellogg. W. Kent.

W. B. KunharcU. G. F. Kunz. O. Lamm. P. Larsson. L. G. Lanroaa. J. £. Lawton. G. M. Lehman. H. R. Leonard. J. F. Lewis. N. Lilienberg. C. Macdonald. Thomas Macfarland. i R. C. McKinney. A. Martens. E.D. Meier. J. W. Meier. F. J. H. Merrill. W, H. Merritt. George Merryweather. C. L. Miller. C. Morgan. W. H. Morris. H. S. Munroe.

F. H. Newell. A. R. Nickels. P. E. Nostrand.

E. E. Olcott.

G. A. Packard,

K W. Parker. R. A. Parker. H. w! Parkhurst C. O. Parsons. G. S. Patterson. W. E. Patterson. C. P. Perin. W. H. Pettee. C. Pettigrew. O. Pfordte. H. K. Porter. E. C. Potter. R. W. Raymond. Ellen Richards. R. H. Richards. R. Rickard. T. A. Rickard. C. S. Robinson. T. W. Robinson. R. P. Rothwell. A. Saaveur. H. S. Smith.

C. H. Smyth, Jr. E. J. Spencer. K G. Spilsbury. A. A. Stevenson. A. Swazey.

W. J. Taylor. G. H. Thomson. R. H. Thurston.

D, Torrey.

D. Townsend.

E. K R. Tratman. G. A. Trube.

J. 8. Unger.

W. H. Van Arsdale.

H. Veeder.

H. A. Vezin.

L. Von Rosenberg.

J. R. Wagner.

H. Wedding.

J. D. Weeks.

S. T. Wellman.

M. White.

W. G. Wilkins.

H. V. Winchell.

Arthur Winslow.

T. W. Yardley.

P. Yeatman.

E. Zellenkoff.

Papers.

roL. ixii.- 1 Pigiji

Tests Of Hydraulic Materials.

Bt H. Le Chatelier, Professor At The £C0Le De8 Mines, Paris, Prance,*

(Chicago Meeting, being part of the International Engineering Congress, August, 1893.)

I. Introduction.

During the last twenty years, the study of the methods of test- ing hydraulic products has made great progress. Richly endowed laboratories have been devoted to this study — in France, the labora- tories of the Service des Fonts et Chausses at Paris and at Boulogne- sur-Mer, that of the Military Engineers, and that of the Boulonnais , Cement Works; in Switzerland, the laboratory of Zurich; in Ger- many, that of Charlottenburg near Berlin ; and in Russia, that of St. Petersburg.

Numerous conferences and congresses, such as the conferences of Munich, Dresden, and Berlin, and the congress upon the materials of construction at the Exposition of 1889, in Paris, have discussed the results obtained. But, while recognizing the unquestionable services rendered by all such endeavors, and the good intentions which have co-operated in this work, we cannot deny that the ob- ject aimed at has been sometimes overshot. The growing multi- plicity and complexity of the proposed tests rendering them imprac- ticable outside of laboratories of investigation, necessarily leads to a more restricted use of such tests, instead of the increased use which might be desirable. Here, as everywhere, the best things involve inconvenience ; the attention continuously fixed upon the tests has led to the study of them for their own sake, and too often their essential purpose has been lost sight of, namely, the purpose of permitting the engineer to estimate by certain and rapid processes the product which he is about to employ. In a spirit of reaction against the unconscious and inevitable tendencies of investigating laboratories, the present paper has been prepared, in order to enlarge the use of methods of testing by opposing their abuse.

The definition of the most convenient methods for determining the qualities of hydraulic cements and lime, is a very delicate ques-

Translated bj the Secretary.

4 Tests Of Hydbaulic Matebials.

tion. Ift* .approaching such an inquiry, it is necessary to remember that Alftfeting of hydraulic products is an entirely different process frooa'- of metals. The latter possess at the moment of the test the qualities which they will preserve practically unchanged, pro- . viUed they are not altered by subsequent manufacture, and are suit- **ihly preserved against oxidation. Moreover, the principal qualities valuable in use (tenacity, elasticity and malleability) are perfectly defined, and lend themselves to direct measurement. They are qual- ities of an exclusively mechanical order.

On the other hand, the hydraulic products are subject to perpet- ual development from the day when they are produced to the day of their complete disintegration, which sometimes takes place after a few years. It may be said that their essential, almost their only peculiar quality is that of resisting as long as possible the external agencies which attack them in every way, such as freezing, drying, running water, and sea-water. This quality of resistance depends, moreover, in equal measure upon certain physical or chemical prop- erties, often very little known on the one hand, and on the other hand, upon the greatly variable conditions to which they are sub- jected in use. It is always difficult to separate with precision the two components of the complex result ; and yet an exact knowledge of the part that each plays in the qualities of the material employed is indispensable to a choice of rational methods of testing. The qualities of mechanical resistance or adherence, while certainly in- teresting to determine, have but a secondary importance in the vast majority of practical applications of cement.

The choice of methods must, therefore, be preceded by prelimi- nary studies of the three following factors :

(1) The exterior qualities which co-operate in the deterioration of hydraulic qualities (rain, drying, sea-water, etc.); (2) the intrinsic properties which render them more or less able to resist such causes (chemical composition, degree of calcination, fineness of grinding, etc.) ; finally (3), the conditions of application which play a part in the more or less rapid alteration of the products (the water of tem- pering, sand, compactness, etc.).

These experimental inquiries can only be pursued with advantage in fully furnished laboratories, under the direction of engineers capa- ble of conducting such investigations with the indispensable scien- tific spirit. They differ entirely in object and in method from tests properly so-called. The latter have a purely practical purpose, that of recognizing by a simple and rapid process the presence, in a given

Tests Of Hydraulic Materials. 5

comraercial product, of the qualities which, previously, laboratory researches have shown to be useful, and the absence of the properties similarly shown to be harmful. They can be carried out by simple manipulations, apart from all scientific knowledge, and requiring only fidelity and some manual skill. Such tests are indispensable to manufacturers for the conduct of their business, and to consumers forjudging of the products delivered to them.

Unfortunately, the scientific study of the properties of hydraulic products is still but little advanced. A small number of their qual- ities are defined with precision ; for a still smaller number, conven- ient processes of measurement have been found. In the lack of these indispensable means, we are reduced, in order to make the best of the actual situation, to guide ourselves provisionally in the choice of testing methods, either by theoretical considerations, which are some- times open to question, or by the vague experience of practitioners. Under these circumstances, the value of some testing-methods neces- sarily remains undetermined ; it is prudent not to repose in these tests too great a confidence, and it is, on that account, useless to con- duct them with a minute accuracy which they do not deserve.

II. Chemical Composition of Hydraulic Products.

The theoretical study of the chemical reactions which combine in the hardening of hydraulic products, is the surest foundation of all researches upon methods of testing. But, in order to understand the interaction of the different chemical phenomena, it is indispensable to bear in mind the laws that govern the phenomena of solution, since these play a leading part in all the transformations of cements.

1. OenercU Laws of Solution.

Many substances have the property, when placed in contact, of mixing intimately with one another in indefinite proportions and giving a liquid homogeneous mass. This is called solution.

Although the proportions of materials in solution are variable, yet there is generally a superior limit which cannot be passed. Thus, when a large mass of chloride of sodium is placed at ordinary tem- perature in contact with a .given amount of water, only a certain quantity of the salt will be dissolved, however prolonged its contact with the solution already formed. The composition of the solution thus obtained is independent of the quantity of salt present. Such a solution is called saturated. The degree of saturation varies, more- over, with different circumstances, such as temperature. Taus,

Te8T3 Of Hydraulic Materials.

nitrate of potassium is infinitely more soluble hot than oold ; the hydrate of calcium, on the contrary, is but half as soluble at 100® C. as at ordinary temperature, and the maximum solubility of gyp- sura is shown at about 35° C.

But among the circumstances which influence solubility there is one which possesses a leading importance for the study of cements. I refer to the state of hydration of the salt under consideration, and, to speak more generally, the various physical or chemical conditions to which it is subjected. Gypsum or hydrated sulphate of calcium has not the same solubility as the unhydrated sulphate of calcium. The carbonate of calcium already formed will not give with water a solution of the same concentration as is obtained from carbonic acid and lime simultaneously introduced into water. The same is true for the allotropic conditions of the same substance. The red iodide of mercury and the yellow iodide have not the same solubility.

These differences in the solubility of the same salt in different states were recognized long ago by the French chemist Loewel,* but the fact did not at once receive the attention which its importance deserved. The following diagram shows the curves of solubility ob- tained by this savant for sulphate of sodium :

SolubUlty of Sodlmn Sulphate of dlfteint degrees of h jdratton.

Temperature In degrees Centigrade.

100*

It will thus be seen that at 10° C. the solubility of the three forms of sodium sulphate, stated in parts of the anhydrous salt contained in 100 parts of water, is :

NaO,SO„ 5S

NaO,80,.7H5,0 32

NaO. SOa, lOHjO, 10

Annales de Chimie, et de Physique, 3d series, vol. xlix., 1857, p2.

Tests Op Hydbaui.Ic Materials. 7

The ordinary sulphate, having ten molecules of water, is less solu- ble (in terms of the anhydrous salt dissolved) than that which con- tains seven molecules of watr, which is, in turn, less soluble than the anhydrous salt. The same fact was proved by Marignac for the sulphate of calcium. Plaster of Paris or calcined (dehydrated) gypsum is at ordinary temperature five times as soluble as natural gypsum or sulphate of calcium with two molecules of water.

But, it should be added, that of these solutions of differing satu- ration-points, it 18 only the least concentrated which is stable, and which there is occasion to observe. The production of the more concentrated solutions is generally a passing phenomenon.' In cer- tain cases, however, these stronger solutions may be maintained by special contrivances for a longer or shorter time as so-called mper- scUurcUed solutions, which are in fact only saturated, but saturated for a particular state of the salt, which state is not stable in the presence of water under the given conditions of temperature. It is easily proved in the ordinary preparation of supersaturated solutions of the sulphate of sodium that these liquids are in reality saturated witl the anhydrous sulphate. In fact, they always contain, if too much water has not been added in preparing them, crystals of anhydrous sulphate, which remain indefinitely in contact with the liquid with- out dissolving in it, thus giving the character required by the defi- nition of a saturated solution. But this solution is not stable. The simple introduction of a small quantity of the hydrated sulphate with ten molecules of water will provoke a crystallization en masse from the solution, reducing the latter to the lower concentration correspond- ing with the saturation of the new hydrate formed, which is the only one stable at ordinary temperature.

In all cases the least soluble of the different varieties of the same solid substance is the most stable under the given conditions. For the sulphate of sodium below 33 C, the anhydrous form is most stable and also least soluble. At 33 C. exactly, the anhydrous salt and the salt with ten molecules of water are equally stable,, and their co* efficients of solubility are the same. The curves of solubility for these two forms intersect, as will be seen in Fig. 1, at 33 C.

These properties of solutions give immediately the key to the mechanism of hardening of materials capable of 'setting " in con- tact with water.

2. Hardening of HydraiUic Maiei'iaU.

The property of hardening in contact with water, which character- izes hydraulic materials, belongs to a large number of differenXchemie

8 Te8T8 Of Hydraulic Matebials.

cal compounds, some of which can be stadied more easily than limes or cements. The salphate of sodiam, previously fused and finely pul- verized sets rapidly when tempered with a little water behaving like plaster only still more quickly, by reason of its greater solubility. The mechanism of this setting is easy to trace.* The anhydrous salt dissolves in contact with the water, the concentration increases rap- idly, passes the point of saturation for the salt with ten molecules of water, and tends to reach that of the anhydrous salt But such a solution is not stable in contact with the small quantities of hydrated salt which form on the surface of anhydrous salt exposed for a oertain time to air. The supersaturated solution in process of formation soon ciommences to crystallize, giving the salt with lOHjO. Mean- while, the anhydrous salt not in the presence of its completely satu- rated solution goes on dissolving. These two inverse phenomena of crjTStallization and solution take place each at its own rate of speed, whence there results for the liquid a certain state of mean supersatu- ration, which does not completely disappear until there is no longer ny anhydrous salt remaining to be hydrated. The final set is thus the result of a transitory solution, rendered possible by the difference in solubility of the different states of the salt. In the case of sul- phate of sodium, by reason of its great solubility, this transitory solu- tion can be very clearly proved. It is sufficient, instead of taking the salt as fine powder, to break it into fragments of pea-size, and place these in the upper part of a vase filled with a saturated solution of the hydrated salt. The supersaturated solution which forms is able to flow down through the laier interstices between the frag- ments and to fall to the bottom of the vase, where the IOH2O salt crystallizes out in a compact mass. The transfer of the salt corresponding to its hardening is the certain proof of a transitory solution.

With less soluble salts, like the sulphate of calcium, the phe- nomena are the same, though less easy to observe. Calcined plaster gives with water a supersaturated solution containing 10 grammes of sulphate per liter (1 per cent.), as against 2.36 grammes (0.24 per cent.) in the saturated solution of the hydrate.

The aluminates of calcium occurring in cements, although scarcely soluble when once hydrated, give, when anhydrous, in contact with water, solutions of relatively high supersaturation. By agitating pulverized aluminate of calcium for five minutes with an excess of

H. Le Chatelier, Recherches ezprimentales sur la constitution des mortlers hjdrauliqaes," Annales des iftnes, Mai-Juin, 1887, page 20.

Tests Of Hydraulic Matebials. 9

water, 0.6 gramme of this salt may be dissolved per liter (0.06 per oent.); but it soon precipitates hydrated crystals, and finally there is DO measarable quantity of aluminate left in the liquid.

The same process should take place with the silicate of calcium, which is the essential hydraulic element in all limes and cements. But the solubility of this salt is so low, and the determination of small amounts of silica is so delicate a matter, that it has not yet been possible to demonstrate experimentally the supersaturation of water with the anhydrous silicate. This has, however, been done for an analogous salt, the silicate of barium, which is somewhat more soluble, and which likewise sets in contact with water.

The hardening of hydraulic materials thus seems to be in all cases the result of the crystallization of hydrated compounds which have passed through a momentary state of solution. But the solid- ity ot such a crystalline mass may vary within considerable limits according to the form, dimensions and mode of segregation of the crjTStals thus produced. In general the cohesion or hardness of an isolated crystal is much greater than its adherence to adjacent crys- tals. The more the crystal-surfaces in contact are developed, the greater the total adherence and the strength of the mass. Crystals in long plates or in interlocked fibers should give a considerable strength. Now it happens precisely that all crystals deposited from supersaturated solutions present these characters, and do so in more marked degree, the higher the degree of supersaturation of the mother-solution. The crystallization of supersaturated solutions of sulphate, acetate or hyposulphite of sodium gives filiform crystals sev- eral centimeters long and less than 0.1 mm. (0.004 in.) in diameter. The same may be recognized in the setting of plaster, ahiminates of calcium, etc. But this elongation, aud the resultant solidity, may vary to a considerable extent with the degree of supersaturation, and this in turn depends upon multiple conditions — the fineness of the anhydrous salt, the number of centers of crystallization, etc. It is for this reason that hydraulic products very nearly alike often give in use, whether in actual construction or under experimental tests of strength, results so difierent.

3. Swelling by Slaoking.

In most cases the action of water on anhydrous salts is a cause of hardening; but this is not always so. We know that lime, in be- coming hydrated, cracks, disintegrates; in a word, is slaked or 'slacks.'' In order that a hardening shall take place, the anhydrous

10 Te8Ts Of Hydraulic Materials.

salt must be able to subsist for a certain period in contact with water simply dissolving in the latter, but not combining with it. This is impossible for certain substances which have a strong affinity for water, such as lime, baryta, boric acid — anhydrides which combine directly with water, forming hydrates prior to any solution. In these cases, which are very few, there is a disintegration of the sub- stance and a simultaneous swelling with enormous force. For in- stance, it is well known that a fragment of quicklime enclosed in a brick may, in slacking, burst the brick.

Substances which set in contact with liquid water may swell when hydrated in contact with the vapor of water. This is the case with sulphate of sodium, plaster and aluminate of calcium. But in this case the hydration is often extremely slow, and may even go no further than the exterior surface. This phenomenon of swelling is one of the most active causes of the disintegration of cements ; and its examination should be one of the most important objects of the methods of testing.

4. Chemical Compounds in Cements.

The compounds usually encountered in cements and hydraulic limes are lime and its combinations with silica, alumina, oxide of iron and sulphuric acid.

Lime. — Anhydrous lime always slacks; that is, disintegrates and swells during hydration.

The time required for slacking lime varies greatly with its degree of compactness. The porous lime obtained by burning pure lime- Gitone at low temperature slacks instantly in contact with water, while the compact lime produced by calcining nitrate, or, at high temperature, somewhat argillaceous carbonate of calcium, takes sev- eral days.

Slacking is more rapid and swelling greater, the higher the tem- perature. This is a fact of capital importance, which is utilized in the treatment of hydraulic limes not conveniently slacked when cold, and also in the testing of cement to detect the presence of free lime.

Swelling is greater from vapor of water than from liquid water.

Finally, slacking is considerably hastened by the addition to the water of a small quantity (2 to 6 per cent.) of calcium chloride, or of salts like magnesium chloride, which yield with the lime of the cement chloride of calcium. Thus, a lime which takes 48 hours to slack in pure water does so in half an hour if ground with a 3 per

Tests Op Hydraulic Materiau9. 11

cent, solution of calcium chloride. These facts, discovered by M. Candlot*, have received several ioterestiDg applications.

The presence in hydraulic limes and cements of anhydrous lime free, that is to say, not combined with acids or water, is, as has been observed above, one of the most important causes of the destruction of these products. When the slacking of the lime is so slow as not to be accomplished until after the setting, it causes by its swelling, if it is in notable quantity, cracks in the whole mass and a conse- quent disintegration of the mortar. If it is present in small quan- tity, it produces an increased porosity, facilitating the destructive action of exterior agents. The addition to a good Portland cement of 1 per cent, of compact lime (from nitrate) is enough to cause dis- tinct cracking.

Free magnesia gives rise to analogous but less important swell- ing.

Silicatea of Calcium. — There are three silicates of calcium, viz. :

(1) CaO, SiO„ or wollastonite ; this does not occur normally in cements. It is produced only against the walls of the furnace and at the expense of their silica especially when the lining, as in some works, is made of sandstone. It possesses no hydraulic property, and is an inert material.

(2) 2CaO, SiO, ; this silicate has the singular property of spon- taneous decrepitation to powder upon cooling, a consequence of alio- tropic change of condition. The phenomenon resembles that which is produced under the same conditions of cooling with oxide of lead, sulphate of potassium, and especially with the double sulphate of copper and potassium. This spontaneous crumbling of the dicalcic silicate, very frequent in basic Uast-furnace slags, also occurs fre- quently in Portland cements. It is the more marked the weaker the proportion of lime and the higher the temperature of burning.

This silicate possesses no hydraulic properties. It does not harden in contact with water. But it is rapidly attacked by dissolved car- bonic acid, with the formation of crystalline carbonate of calcium, and it may thus contribute in some degree to the final hardening of mortars.

(3) 3CaO, SiO,; this is the only really hydraulic silicate; it is par excellence the active element of hydraulic limes and cements. In Portland cement, of which it constitutes the greater part, it occurs

Role da chlonire de calcium et du salfate de chaaz sur la prise et le darcisse- ment des roortiere. BuUetin de la SociSU cTeneouragemerU pour Plndxutrie nationaU, Juillet, 1890. Paris. 8ie de la Socit, 44 rue de Rennes.

12 TiESTS OP HYDRAULIC MATERIALS.

in pseudo-cubio crystals. It is produced by the reaction of silica and lime in the presence of fusible combinations formed by iron and alumina. When overheated, it appears to be decomposed into di- calcic silicate and free lime, thus losing its hydraulic properties. In contact with water it sets, dividing so as to give hydrated monocal- cic silicate crystallizing in microscopic needles and calcium hydrate crystallizing in large hexagonal plates which are visible to the naked eye in all Portland cements.

3CaO, SiO, + Aq CaO, SiO 2.6H,0 + 2 (CaO, H,0).

The hydrated silicate in the presence of an excess of distilled water decomposes until the moment when the solution contains 0.052 gramme of CaO per liter.

The supersatnration which precedes crystallization is difficult to recognize for the monocalcic silicate. It is very clearly proved, on the other hand, for the simultaneously formed hydrate of lime.

This silicate is but slightly sensitive to the action of water vapor. This enables it to pass unharmed through the period of slacking in the manufacture of hydraulic limes. It is upon the proportion of this silicate that the quality of a hydraulic product principally depends.

Aluminates of Calcium, — There exist different aluminates of cal- cium, all of which set very rapidly in contact with water. The most important, the tricalcic aluminate, is simply hydrated in contact with water like plaster, producing highly supersaturated solutions.

3CaO, Al A + Aq 3CaO, Al A, I2H3O.

This relatively unstable hydrate loses its water and effloresces in warm and dry atmospheres; it may thus become a cause of destruc- tion in mortars used in air. In contact with distilled water it de- composes until the solution contains 0.22 gramme of CaO per liter. This salt, according to the researches of M. Candlot, combines with sulphate of calcium to form a double salt which crystallizes with a very large quantity of water

3CaO, AIP3, 2.5(CaO, SO,), 60H,O. This compound appears to play an important r6le in the destruc-

♦ Candlot — Loe. eit, {BuUelin de la SociiU cCencouragemem) and Candlots CimetUa el ChavXf No. 250, Librairie Baudry (1S91).

Tests Op Hydraulic Materials. 13

tion of mortars in sea-water. It is formed at the expense of the sulphuric aoid in the sulphate of magnesium. This is a point of capital importance which deserves to be studied more completely, and which will some day undoubtedly be utilized in the testing of cements destined for marine works.

The existence of hydrated aluminates containing less than three molecules of water has not yet been established, but that an alumi- nate of the composition, CaO, AlO,, Aq, exists may be rarded as highly probable. Such a compound cannot, however, be formed under normal conditions in cements which always contain an excess of lime.

Calcium chloride combines with the intermediate aluminate, 2CaO A]fi but the combination is destroyed in the presence of an excess of lime, the tricalcic aluminate being formed in its stead. When cements are tempered with sea-water, the combination with calcium chloride occurs during the setting, which it retards, but it disa|)- pears as soon as the hydration of the calcium silicate has liberated sufficient lime.

Ferriies of Calcium. — These compounds swell like free lime under the first action of water, and then give birth to a white hydrated tricalcic ferrite, which latter is decomposed by the carbonic acid of the air with the production of brown sesquioxide of iron. These compounds do not exist in well-burned Portland cements, which never assume in the air the characteristic burned color mentioned.

SUico-Alumino Ferriies of Caloium. — There is produced in the Portland cements a fusible silico-aluminate identical with that which forms the essential element of crystalline blast-furnace slag, in which the sesquioxide of iron partially replaces the alumina :

3CaO, AI3O3, 2SiO,.

This substance is completely inert under the action of water ; it does not appear even to be attacked in the long run by carbonic acid. Its only useful function is to serve as a fiux to favor during the burn- ing the combination of silica and lime.

Thb silico-aluminate crystallizes in Portland cement by reason of the slow cooling, but may, on the other hand, retain a vitreous texture when cooled with sufficient suddenness. This is the case, for in- stance, when blast-furnace slags are precipitated while still liquid into cold water. The properties of this compound then become entirely different. It is attackable by weak acids and at the same time by alkalies. It combines particularly with hydrated lime in setting,

14 Tb8T8 Op Hydraulic Materials.

and gives rise to silicates and aluminates of lime identical with those which are formed by entirely different reactions during the setting of Portland cement. It is upon this property that the manufacture of slag-cements which assumes daily greater importance, is based.

III. — Classification op Hydraulic Products.

One of the facts most clearly shown by the daily experience of constructors is that hydraulic products of similar average composi- tion vary greatly according to the methods of their manufacture, in the qualities of resistance to disintegration which they show in prac- tice. The artificial cements, called Portland cements, enjoy an un- disputed reputation for use in the air during frosty weather or for marine work. It does not follow, as one might be tempted to be- lieve, that products like the slag-cements, the chemical composition and the resistance to mechanical tests of which are similar, must possess the same qualities. They are perhaps better, perhaps worse. In fact, it is impossible in the absence of suitable methods of testing, to know what would be their quality except by a sufficiently pro- longed trial in actual practice. A rigorous classification based upon the methods of manufacture is thus rendered provisionally indis- pensable as an additional condition of tests for these products. This classification involves inevitably four essential grand divisions, which will be divided in turn into a number of catories, more and more extended as the industry of hydraulic products advances.

1. Portland Cements.

Artificial so-called Portland cements are obtained by mixing in given proportions argillaceous and calcareous materials, which are then burned up to the temperature of scorification (semi-fusion). They require after burning no other addition than that of substances accidentally less burned coming from the same manufacture and a small quantity of water spontaneously borrowed, as a general rule, from the atmosphere.

Examined in thin plates under the microscope, they are formed of tricalcic silicate in crystals, with very feeble double refraction, embedded in a crystalline ground-mass, without individual forms, of silico-alumino-ferrites of lime. These are the two essential ele- ments of these cements. If the lime is in excess, aluminate of lime

Prost — Note stir la fabrication et lea propri6t des cimento de laitiers. — AnnaU de$ MineSf 8e srie, tome zvi., pp. 15S to 20S.

Tb8Tb Of Hydraulic Matebiau3. 15

is first formed ; then for a still greater excess, ferrite of lime, and finally free lime. If, on the other hand, the lime is deficient in quantity, a dicalcic silicate is formed, recognizable by the spontaneous crumbling of the burnt pieces of cement. When the mixture is im- perfect or the burning insufficient, the reactions remain incomplete; and although the average composition may be suitable, there is a simultaneous production of free lime and aluminate of calcium with dicalcic silicate. In a Portland cement of normal composition the proportion of lime, according to the chemical formulas of the com- pound, should be greater than that determined by the following formula :

Sio, — aia-fa"" '

in which CaO, SiO,, AliO,, FjO, represent not the equivalent weights, but the number of equivalents of these substances present ; that is to say, the quotient of the weights of the substances divided by their equivalent weights. This proportion of lime must never, on the other hand, reach the relation indicated by the following formula :

which corresponds to the exclusive formation of aluminate of cal- cium. It is necessary, by reason of the inevitable imperfection of the mixture, to keep always well below this limit, beyond which there will remain uneombined lime.

In the use of this formula, magnesia should be added to the lime and sulphuric acid to the denominator after dividing its number of equivalents by 3.

Notwithstanding the care bestowed upon the burning of Portland cements, it is very seldom that they do not contain a small quantity of free lime. It is indispensable that this should be slacked or hydrated before use. This hydration takes place at the expense of the water which is deposited upon the surface of the pieces during cooling in the furnace, or else of water which is furnished by special sprinkling. The reaction is favored by the elevation of temperature produced in crushing and a long storage in silos. The water which thus reacts has remained in the cement not in the condition of simple absorption, but in chemical combination in hydrated silicates and aluminates, which are less stable than the hydrate of lime, possess

16 Tests Op Hydraulic Materials.

a stronger tendency to dissociation, and are capable, in consequence, of gradually decomposing and slacking the quick-lime.

The composition of commercial Portland cement of good quality is usually within the following limits :

Commercial Products.

Per Cent.

Silica, 21 to 24

Alumina, . . 6 " 8

Oxide of iron, 2"4

Lime, 60 " 65

Magnesia, 0.5 2

Sulphuric acid 0.6 " 1.6

Water and carbonic acid, 1 " 3

The ashes of the fuel, which, for the most part, pass into the cement, play an important part in the variation of the proportion of silica. For the same cements of good quality the characteristic quotients (Formulas 1 and 2, p. 13) of the numbers of equivalents of different substances present vary within the following limits :

CaO Quotient : — : — tti ft-p varies from 2.5 to 2.7.

2. Hydraulio Limes.

Hydraulic limes are obtained by the burning of natural siliceous and aluminous limestones and of the reduction of the mass to pow- der, after burning, by the action of water, which slacks the remain- ing free lime. The only addition is that of a certain quantity of grappiers — t.e., fragments of lime which have resisted slacking and which are crushed in mills.

The principle of the manufacture is to leave in the burned pro- duct a suflBcient quantity of free lime to reduce the whole mass to powder by its slacking. This result is obtained by starting with limestones containing more lime than is required to produce, with the alumina and silica, the tricalcic limestone salts, and by burning at an elevated temperature; or limestones less rich in lime are burned at a lower temperature to produce incomplete reactions.

It is, on the other hand, important that the hydrated lime which remains inert during the hardening shall be as small in proportion as possible, hence the slacking must take place under the conditions of the most energetic swelling ; i.e., at a high temperature, in which

Tests Of Hydraulic Materials. 17

case a smaller proportion of quicklime will effect a complete disin- tegration. In practice, by utilizing the heat of hydration alone, the temperature of the mass may be raised to 200 C, provided the heaps of lime treated are large enough to preclude too great a loss of heat by external radiation. This is a delicate operation to con- duct, because, at the high temperature named, the water added to the mass, and not yet combined, is completely volatilised, nd the lime will not find sufficient water to slack Jt unless regular additions of wetted lime are made to supply the vapor of water in the neces- sary quantity to the hotter parts of the heaps.

The presence of free lime resulting from incomplete slacking is the most serious 9nd frequent defect encountered in hydraulic limes and to be detected by testing-processes. The second defect, likewise frequent, is the result of imperfect burning, producing too little of the active silicate of calcium.

The best hydraulic limestones contain very little alumina. The alnminates of calcium play the part of inert bodiee ; they become hydrated during the slacking, and therefore cannot contribute to the hardening. Moreover, they contribute little or nothing to the slacks ing process. The silicate of calcium, on the contrary, upon which the vapor of water has practically no effect, passes through the period of slacking without alteration, retaining its hydraulic properties.

By theory, a good hydraulic lime (supposing it ta carry neither alumina nor iron) should contain approximately four equivalents of lime for one of silica. Three equivalents of lime should combine with the silica, while the fourth remains free for slacking. But this combination of the lime and silica, especially in the absence of nota- ble proportions of alumina, is always incomplete, and leaves more than the necessary proportion of free lime, experience shows diat the best hydraulic limes do not contain more than three eqjaivalents of lime to one of silica. Their composition, in widely-separated French districts (Ardfeche, Eure, Indre-et-Loire), varies little from, the following :

Ptroent.

Silica, 22.0

Alumina, 2.0

Oxide of iron, 2.0

Lime, 62.0

Magnesia, 1.5

Snlpbaric acid, 0.5

Water, ♦ 10.0

Vol. Xxii.— 2

18 Tests Of Hydraulic Materials.

In limes of inferior quality the tenor of silica may decline to half the above figures, while that of water, and sometimes of alamina may be doubled. The percentage of lime is always about the same, varying, at most, from 65 to 65.

3. Naturcd Cements.

Natural cements are made by burning limestones, less rich in lime than the hydraulic limestones proper. They may be divided into three classes: {a) quick-setting cements (Wassy, Roman) ; (6) slow* setting cements; and (c) opjEner-cements.*

(a) The -eeUmg cements are obtained by burning at very low temperature, just enough to decarbonate very argillaceous lime- stones. They are characterised by a very rapid setting, followed by a final hardening, which is very slow — much slower than that of Portland cements. Their nnxkrate burning produces incomplete chemical reactions, forming aluminate of calcium, which is the cause of their rapid setting, and probably also silico-aluminates of cal- cium (like pozzuolana) which react gradually upon the free lime contained in these products. The presence of this free lime is not injurious, because, by virtue of its porosity and the low heat at which it was burned, it is hydrated immediately upon contact with water, before setting has taken place. Hence it cannot operate to disin tegrate a mass not yet hardened. These cements generally contain 6 to 10 per cent, of sulphate of calcium, the presence of which seems to be indispensable to retard the setting of the aluminate of calcium, which would otherwise be too rapid for practical use.

Analysis.

The composition of natural quick-setting cements is generally within the following limits :

PerceDt

Silica, . . . . , 22 to 24

Alumina, . , . . , 7 to 10

Oxide of iron, . . . . , . . , .4to6

Magnesia, . . lto3

Lime, 45 to 55

Sulphuric }icid, .. . 2 to 4

Water and carbonic acid, 2 to 6

They differ from Portland cements in having a larger proportion

There ia equivalent English term for gmppier. The word signifies anj kind of workshop-Befuse, and -has been used French manufacturers of hydraulic

Tests Op Hydraulic Materials. 19

of sulphuric acid (which appears to be one of their essential elements) and considerably less lime.

(6) The slow-setting natural cements, burned at high temperature, resemble Portland cements; but natural limestones seldom possess the homogeneity of artificial mixtures, and it is difficult to avoid in these cements the presence of a considerable quantity of free lime, the slacking of which is consequently a more important matter than in the Portland cements, in which the amount of free lime must be a minimum. This slacking of the slow-setting natural cements is a delicate operation, which cannot be performed by sprinkling, like the slacking of ordinary lime, and is too much lefl to chance. Studied with greater care, this process could doubtless be made to produce from the cheaper natural cements results as uniform as are reached with artificial mixtures.

The composition of the cements of this class varies from that of the quick-setting cements to that of genuine Portland.

(c) The grappier-cements are obtained by grinding the pieces which have escaped in the disintegration of hydraulic lines. These grappiers are a mixture of four distinct materials, two of which (unburned limestones and slags formed in contact with the siliceous walls of the furnace) are wholly inert, while the other two (unslacked lime and true slow-setting cement) are eminently hydraulic. It is necessary that the last-named should predominate in the grappiers if the grinding is to give a useful product. Cement-forming grap- piers are regularly obtained only in burning at high-temperature very slightly aluminous limestones, containing about three parts of calcium carbonate to one of silica.

These rop/ner-cements are much more likely to contain free lime than the natural slow-setting cements obtained by burning much more highly aluminous limestones. Since they are made by grinding a mixture of grains of cement with variable amounts of inert materials, they may be highly irregular in composition. The grains of cement have nearly the theoretical composition of the tri- calcic silicate, 3CaO, SiOa. The following table gives approximately the composition of the best grappier-cements, that of the pure cement- grains and that of the pure tricalcic silicate.

lime-cements to denote the refuse of their industry. When this refuse, or waste, came to be utilized in the manufacture of a new kind of cement, the product was called oppter-cement.'' A well-known brand of it in the United States is said to be Lafarge's Portland Cement, which is imported and sold here more particu- larly for facing or pointing the stone walls of fine buildings, as it is said that it

does not stain like many other oement8.~B. W. R. r\r\n]o

Ic

20 Tests Of Hydbauuc Materials.

Cement. Grains. Silicate.

Silica, 26.5 26.0 26.5

Alumina, 2.5 8.5

Iron, 1.5 1.0

Lime, 63.0 66.0 73.5

Magnesia, LO 1.0

Sulphuric acid, . . 0.5 0.5

Water and carbonic acid, . .5.0 1.0

100.0 99.0 100.0

4. Mixed Cemeni.

Mixed cetneBts are made by mixings afler calcination various hy- draalic and non-hydraulio substances. By far the most important class is that of slag-eementa obtained by mixing granulated blast- furnace slag with slacked lime, hydrauHc or otherwise. The slags, retained in a vitreous condition, play the part of true poziuolana, and give a more rapid setting the richer they are in alumina and lime. An excess of lime however, by hindering the vitrification by quenching, diminishes their pozauolanio property. They should have preferably about the composition of the formula 3CaO 2Si02, which gives in cooling very hard, compact,, crystalline slag. A very fine pulverization of the slag is indispensable to secure the reaction of the lime.

Good slag-cements will vary little from the composition t

SiOi AlA Fe,Oi CaO MgO SOj Ufi Percent. 24 14 1 51 2 1 7

The category of mixed cements comprises also many products of secondary quality and importance in which Portland cement forms the essential element, while the accessory constituent, added to re- duce the price, may be hammer-scale,, slags, hydraulic of fat lime, or chalk. These cements are never sold under names indicating their true character. They are, in fact, merely adulterations of Port- land cement.

It is very important to the consumer to know exactly to what class the product delivered to him belongs. Very seldom can he rely, in this respect, upon the declaration of the manufacturer, who always has good reasons for including his wares among Portland cements, which command the best price. Among the many methods of control which might be employed to assure agree- ment between the name claimed for the product and the true con- ditions of its manufacture, the most efiective is the establishment

Tests Of Hydraulic Materials. 21

of a permanent control at the manufactory. But this could be Buccessfulij done by govermental inspection only. Most consumers must resort to experimental tests bearing upon certain characteristics of cements, whether related or not to their hydraulic properties. In this direction, cAtntco/ analysis, the determinojtion ofabsdiUe density, the examination under a magnifying-glass of the refuse of sifting, etc., give useful indications. But it is impossible to define precisely the tests of this kind to be employed. The question to be settled exists independent of any rule. Has there been fraud or not in the description of the material sold? The number of possible frauds, and consequently, of the means for their detection, is unlimited. To define and limit the required tests would be equivalent to saying that a fraud which escapes them is not a fraud. The consumer has an indefeasible right to inform himself concerning the nature of the wares he purchases by any convenient means, even outside of ex- periment, provided the means afford proof.

IV. Agents of the Disintegration of Hydraulic Products.

After recalling the views held at the present day concerning the constitution of hydraulic products, it is equally important for the choice of suitable tests, to consider the more active agencies of their disintegration. It will be remembered, at the outset, that there are two causes of destruction inherent in the hydraulic materials them- selves, and, independent of exterior media (water, air, etc.), except so far as the latter may modify the intensity and rapidity of their action.

Interior Agendes, — The first is a deficiency of the active hydraulic constituents, without which there can be no induration. Yicat was the first to connect the induration of limes and cements with the hy- dration of certain chemical compounds formed in the burning. It had been supposed, previously, that the active cause of hardening was to be found in the sand used in mixing mortar. This confusion of ideas resulted from comparing indiscriminately the hardening in air or under water of mortars containing pozzuolanic or ordinary sands. The intervention of such active constituents justifies testing methods, particularly tests of rupture. A deficiency of hydraulic constituents is more or less injurious, according to the conditions of employment. Apart from the mechanical stresses to which mortars are subjected, we might be content, in most cases, with limes or ce- ments very poor in active elements, particularly for construction in

22 Tebto Of Hydraulic Materials.

air, where desiocation and carbonation are active agencies of hard- ening. It is to resist the attacks of physical and chemical agencies of disintegration that the employment of products as rich as possible is useful .

The second internal cause of destruction is the presence of free lime or magnesia, the slacking of which, after setting has begun, leads to swelling and even cracking of the mortar. Swelling, even when unaccompanied by cracking, is very dangerous, because the resulting increase of porosity renders the mortar much more attack- able by frost and by the magnesium salts of sea-water. Free lime, in proportions too small to produce swelling, may cause, in slack- ing, equally injurious internal tensions. But this is a liiooted point. There are no experiments proving the reality of this danger, which has been indicated thus far by theoretical prevision only.

Exterior Agencies, — In relation to the action of exterior agencies, three catories may be distinguished corresponding to the three media in which hydraulic mortars may be employed, viz., fresh water or damp earth, air, and sea-water.

In standing fresh water, in damp soil, and in currents sufficiently slow to exert no mechanical wear, special causes of alteration are not thus far known. There might be fear of continued solution of the lime proceeding from the decomposition of silicates and alumi- nates of calcium described above. This, indeed, takes place with dis- tilled water, constantly renewed. After an extremely long period, there will remain only a gelatinous mass of silica and of hydrated alumina. But in natural waters, always containing bicarbonate of calcium and free carbonic acid, such solution is impossible. On the contrary, there is formed carbonate of lime, the crystallization of which only augments induration.

It is not impossible that certain aluminous elements become altered in contact with water charged with sulphate of calcium, especially if under pressure tending to make it filter through the mortar. But heretofore alterations of this nature have been proved only in ma- sonry exposed to the air on at least one face; hence, it is not certain that they would take place in masonry completely immersed.

In constructions in the air, the conditions favorable to the preser- vation of mortars are manifold. But they have been studied here- tofore in a very superficial fashion only. Variations of temperature and of hygrometric condition are, in every instance, the two most active agencies of disintration.

An excessively low, like an excessively high, temperature may be

Tjb8Tb Of Hydbauuo Materials. 23

injarious. At temperatares below zero centigrade, the freezing of the water which impregnates the mortar acts as in porous stones. The increase of volume assumed by the water in solidifying, pro- duces an expansion which tends to break the cement, splitting it in plates parallel to the isothermal surfaces, that is to say, to the free surface of the masonry. As in frost-split stones, this disintegration is the more easy, the smaller the mechanical resistance of the mortar, and the larger the aggregate volume, and the smaller the individual dimensions of the interstices. When the interstices are sufficiently large, ice may flow through them by reason of its plasticity under a pressure too low to produce rupture of the mortar. It is for this reason that mortars made with coarse sand are least affected by a frost — the interstices are less numerous but larger.

At high temperature certain solid hydrates may effloresce upon losing their water and be reduced to dust (as is the case with crystal- lized carbonate of sodium), thus causing a disintegration of the mortar. This may be expected to be the case with aluminates, and particularly with the alumino-sulphate of calcium, but oonclu- sive experiments are lacking. In all cases this dehydration, if it takes place, should be favored by dryness of the atmosphere. We may hypothetical ly attach to this cause the well-known &ct that certain cements after having remained months under water and ac- quired a very great hardness, crack, and even decrepitate into a sandy mass when exposed to dry air. The excessive dryness of the air has another inoonvenient result. It opposes the suitable harden- ing of certain mortars by hindering the condensation of water upon them in sufficient quantity for this result. This influence of atmos- pheric dryness is very different upon different hydraulic products. In our temperate climates it is but slightly noticed for Portland cements, which harden very well in the air. It is much more marked in the pozzuolana mortars such as slag-cements. In these latter products, the substances which mutually react, being separate, cannot be brought into contact except by the intermediation of a sufficient quantity of liquid water. The addition of hygrometric substances, such as results from tempering with sea-water, generally favors hardening in the air by augmenting the thickness of the layer of water condensed by capillarity on the surface of the grains and especially in the narrow spaces which surround their points of con- tact. There are still other and less known causes of the alteration of mortars in the air. For instance, the carbonation of lime takes place in the neighborhood of zero and gives a hydrated carbpinate of

24 Ts8T8 Of Hydraulio Materials.

calcium, CaO, (X),, SHjO, which when the temperature rises above C, undergoes destruction, giving a pulp without consistency com- posed of water and anhydrous carbonate of calcium.

The filtration through masonry of water carrying sulphate ot calcium has in many instances occasioned the disintegration of mor- tars upon the face of the masonry exposed to the air; and this has led to the supposition that evaporation has played a part in the result. It appears, moreover, probable that cements carrying alumi- nate of calcium are the most alterable. These facts of disintegration are still little understood ; but their existence has been proved with absolute certainty in the masonry of sewers and fortifications.

In sea-water the causes of disintegration are still more energetic. The hydrates of the calcium salts, to the formation of which indura- tion has led, are decomposed by the salts of magnesium ; soluble chloride of calcium is formed which is carried away ; and at the same time there is a deposit without consistency of hydrate of magnesia and sulphate of calcium, as well as of silica and alumina set free by the decomposition of the calcareous salts. The carbonic acid of sea- waters, or that which is disengaged by marine vegetation alone, limits the action of this cause of disintegration by forming crystalline car- bonate of calcium, which is not attacked by magnesia salts.

The second and still more dangerous action of sea-water, unques- tionably connected in some degree with the foregoing, manifests itself in the swelling and splitting of the mortar. The sulphate of calcium formed at the expense of the sulphate of magnesia, or pre-existing in the cement, is here the most active agent. This mode of disin- tegration is especially to be feared with very aluminous products; it is apparently a consequence of the formation of the sulpho-aluminate of calcium of M. Candlot. In all cases, this action of sea-water is the more rapid, the more active its renewal, not only at the contact but especially in the interior of the mortar. The tides produce dif- erences of pressure which, causing water to pass alternately in and out of the joints, considerably promote the ruin of masonry, which therefore lasts much longer in tideless seas, like the Mediterranean, than in the ocean.

Finally, the insufficient mechanical resistance of the mortars them- selves should be mentioned as a cause, though a much less frequent one, of their destruction. This may show itself in special construc- tions, such as bridge-piers, where the crushing of the mortar upon itself in the joints will cause a deformation of the structure, or in retaining-walls, where lack of adherence of the mortar to the stones will permit the rupture of the masonry above the foundation

Tests Of Hydraulic Materials. 25

V. — Methods op Testing.

Among the many methods proposed, only a few are really valu- able and at the same time simple enough for industrial use. These only will be mentioned here, all those being passed by, the efficiency of which is doubtful or the study of which has not gone far enough to warrant definite conclusions. The practicable methods may be classified according to their respective objects, that is, the qualities for which they are respectively intended to test a hydraulic material.

There is no direct method of determining the quantity of active elements in a cement. We must be content with indirect methods, the most satisfactory of which is based upon the measure of resistance to rupture. This resistance does, in fact, increase with the quantity of active elements, but not in strict proportion. Many conditions conspire to determine the resistance in each particular case. Deter- minations of the fineness of grinding are also employed for this pur- pose, it having been observed that the grain of cement exceeding a certain size is no longer hydrated to the center, and behaves like inert sand. Chemical analysis may also give indications in this respect, since the absence or too small proportion of silica involves a similar deficiency of hydraulic compounds. Finally, it has been long believed that the density may give indications as to the degree of the burning upon which depends the state of combination of the elements present.

To estimate the presence of the expansives properly so-called, t.6., of free lime and magnesia, recourse is had to the swelling caused by their hydration. This is best determined by direct measurement of the increase of volume, for which, however, simple observation of the cracks which may be occasioned by the swelling, is often sub- stituted. These tests for invariability of volume are facilitated by the action of heat or of solutions of calcium chloride, whieh greatly augment the intensity of the swelling.

The above are the only two qualities of cement for which there exist at present methods of testing, approximately satisfactory, though not perfect. There is no sure means of comparing the resistance of different cements from the standpoint of their use in air or in sea- water. It does not seem impossible, however, to arrive some day at satisfactory methods for this purpose.

The different tests of hydraulic materials may be classified again, according to the nature of the phenomena which they directly deter- mine, and this is the order in which they will now be discussed, as follows : chemical methods (chemical analysis), physical methods (fine-

./

26 Tb3T8 Of Hydraulic Materials.

ness, density, invariability of volume), mechanieal meihoda (rapidity of setting, resistauoe to rupture). The test for rapidity of setting, included in this list, throws no light upon the resistance of hydraulic materials to disintegration, but determines a quality of considerable importance from the standpoint of the facility and the proper manner of use.

In studying these methods, we shall consider the object and principle, the practical execution and the degree of precision of each.

The same test may often be made with very different apparatus and arrangements, each manner of execution being preferable for the object immediately in view. The simpler way will be chosen for current work, the more precise (and especially more independent of the interference of the operator) ought to be reserved for contested cases, in which it will have the advantage of being endorsed by tra- ditional usage. If different ways of making a test are to be recog- nized, it must be understood, of course, that their results ought to be in substantial harmony, and that in case of disagreement the most precise method should always be deemed the normal and legally de- cisive test

A. Chemical Tests.

Analysis.

The hydraulic properties certainly pertain to substances of per- fectly defined chemical composition, but these are mixed with one another and with inert materials in variable proportions. The re- sultant product has no definite composition, and we can do no more than assign superior limits to the proportions of certain substances contained in it Moreover, chemical analysis does not recognize the state of combination of the different elements present, which varies always with the homogeneity of the mixtures burned and with the duration and temperature of burning. Finally, chemical analysis is a delicate operation, only to be suitably performed by educated and practiced operators. The erroneous results so often reached by it are much worse than no results at all.

Complete chemical analysis, therefore, cannot be reckoned among normal tests; its place is in the scientific laboratory, where the general properties of cements are studied.

Without making a complete analysis of a cement or a lime, we may nevertheless propose to determine separately certain elements, knowledge of which would be particularly interesting. The sub-

Testb Of Hydraulic Materials. 27

stances of particular importance for various reasons are magnesia, sulphuric acid, water, and carbonic acid.

Magnesia, — Accidents occasioned by certain magnesian elements have led to the supposition that the presence of magnesia in hydraulic materials is highly injurious. In reality, this is not the case. Even in marine work the magnesian cements may undoubtedly give good service. The accidents above alluded to, and the similar results obtained in laboratory experiments, have been due to the employ- ment of badly-proportioned cements, containing free uncombined magnesia and too small a quantity of clay. Corresponding mixtures containing lime instead of magnesia would have caused still more serious accidents, yet it would not be ooncluded that there must be no lime in cement. There is then no special reason for specially ascertaining the proportion of magnesia. Moreover, its determina- tion is long and delicate, and involves a complete analysis.

Sulphuric Add. — The presence of sulphuric acid has no incon- venient effect upon work in fresh water or damp soil, but is very injurious to marine work. But the exact determination of sulphuric acid is a long and delicate analytical operation, which can be suc- cessfully carried out only in a chemical laboratory. It should there- fore be reserved as an exceptional test for materials destined to marine use.

Waier and Carbonie Acid.— The determination of water and of carbonic acid may give interesting indications of defective burning or slacking, and the degree of deterioration of hydraulic materials. But from this standpoint such determinations are accessory to the tests for resistance, which directly reveal the diminution in active ingredients resulting from the fixation of too much water or carbonic acid in the combination with the lime. The chemical determina- tions, therefore, should not be included among tests proper, but should be considered as processes of investigation capable of throwing light upon the causes of defects otherwise discovered.

In conclusion, neither complete nor partial chemical analysis of the constituents of hydraulic materials can be ranked among normal tests. But chemical analysis may render real service in controlling the classification of a product concerning which there is reason to doubt the declaration of the manufacturer. Thus, a slag cement can be distinguished from a Portland by its tenor in alumina and water; certain natural cements by their contents of sulphuric acid ; the a/>pter-cements by their carbonic acid and water, eta

28 Tebts Of Hydbauuc Materials.

B. Physical Tests, 1. Fineness of Grinding.

It is interesting for several reasons to determine the fineness of hydraulic powders. Up to a certain limit, increase of fineness im- proves the quality of limes and cements. Larger grains only con- tribute partially to the hardening; they become covered with a layer of hydrate, which protects them against the further action of water; and thus they play to some extent the part of inert sand. It is often conceded (but this still remains to be completely verified) that only the grains which will pass through a sieve of 900 meshes per square centimeter (about 5700 per square inch or 77 meshes per linear inch) are wholly utilized. But for inspectors' tests there should be a maximum fixed for permissible excess of size above this. As to the finer portions, it is not known whether it is inaportant that the pro- portion of fine dust should be more or less considerable. Hence, no requirement in that respect can be formulated. It may be added, that with the usual processes of manufacture the relation among the proportions of grains of different sizes is nearly constant.

Apart from this first consideration, large grains of overbumed lime are much more injurious than small ones of equal aggregate weight. Their slow slacking lasts much too long, and is the more dangerous the farther hardening has progressed. On the other hand, the concentration of the developed strains upon a smaller number of points causes an inequality of these strains which favors the forma- tion of cracks. This fact of daily experience is an additional reason for fixing a superior limit to the size of grains.

The proportion of grains exceeding a given size is determined by screening. The sieve may be agitated by hand or by machine. The result of the operation always involves more or less uncertainty, because the meshes not being all of exactly the same size, and the grains not having the same dimensions in every direction, a pro- longed shaking at last causes a certain quantity of grains to pass which were at first rejected. The importance of this uncertainty is proved by the two tables of experiments given below, one furnished by M. Durand-Claye and the other by M. Candlotf These ex- periments were made upon 100 grammes of two different Portland cements, with sieves of 5000 meshes per square centimeter (about

Sur let EitaU de Tamisage, Report of M. Durand-Claye to the Commission on the Unification of Testing-Methods.

t Ciments ei Chaux Hydrauliques by M. Candlot, p. S8.

Tests Op Hydraulic Materials.

180 to the linear inch). M. Durand-Claye employed a sieve 1 decimeter (3.94 inches) in diameter, mechanically operated on the Tetmayer system and making 200 revolutions per minute; M. Candlot used a 3-decimeter (11.82-inch) sieve, agitated by hand. There were left on the sieve in each case the following percentages at successive stages of the screening :

Mechanical Sieve.

Per Cent.

After 500 I'evolutiona

. 41.2

1000 "

. 39.4

" 1500 "

. 3S.6

" 2('00 "

. 38.0

" 2600 "

. 37.6

Hand'Sieve.

Per Cent.

After 5 minates,

. 29.6

"10 "

. 29.1

. . . 28.4

. 28.0

" 40 " . .

. 27.7

The difference caused by stopping after 1000 revolutions with the machine or 5 minutes' shaking by hand, as compared with a very prolonged operation does not exceed 6 per cent., which is amply sufficient accuracy. Successive operations with the machine give results agreeing within 1 per cent. With the 900-mesh sieve screening will be still more rapid.

Canclvmon. — The normal test for fineness of grinding should be made with a sieve of 900 meshes per square centimeter (No. 80 of the French gauge, equivalent to 77 meshes per linear inch), made of wire 0.15 mm. (0.006 inch) in diameter, which gives 0.18 mm. (0.007 inch) for the diameter of the openings. The sieve should have a diameter of 1 decimeter (3.94 inches), and should be operated by machine, the screening being performed upon 100 grammes (3.5 ounces), and to be stopped after 1000 revolutions. For current tests it might be sufficient to screen by hand for 10 minutes with a sieve 3 decimeters (11.82 inches) in diameter. The results of the two methods are practically identical. The normal value of the residue upon the 900-me8h sieve varies for different hydraulic products about 5 per cent.

2. DengUy. Great importance has been long attached to the determination of

30 Tests Of Hydraulic Materiaus.

the ''apparent density" of hydraulic products, this name being given to the weight of a Hter not rammed. The greater this density, the higher should have been the heat of calcination, and consequently, the better should be the quality, by reason of the greater completeness of the chemical reactions. But M. Candlot has proved that this relation between the density and the degree of calcination does not necessarily exist. The ''apparent density'' de- pends almost exclusively upon the fineness of grinding; it may likewise depend, to a certain extent, upon the chemical composition, principally the percentages of water and carbonic acid. If the less thoroughly burned cements are generally lighter, this is chiefly because they are lees hard and usually more finely ground, and being more porous, they absorb more easily water and carbonic acid. But it is possible, by avoiding too fine a grinding and the absorption of water, to preserve a high density, even for cements insufficiently burned.

Conclusion (1). — The test for apparent density has no value, and ought not to be included among normal tests.

It has been proposed recently to seek by tests of true density that which the tests of apparent density cannot give. It has been proved, for instance, that the specific gravity of well-burned Port- land cement is in the neighborhood of 3.15, while for poorly-burned cement it will fall to 3. This difference is measurable, though small but it depends, in reality, not directly on the degree of burn- ing, but only on a greater percentage of water in the cement. An extra 2 per cent, of water will make the above difference. Now, it is possible to avoid this absorption of water, and thus to preserve in cements imperfectly burned a true density almost equal to that of well-burned cements. I will cite, for example, two natural cements coming from the unequal burning of the same rock.

Cement.

Quick-settiDg, . Slow-ettiDg,

The quick cement, which is the less thoroughly burned, has already the regular density of Portland cements.

Conclusion (2). — The determination of the absolute density is no more serviceable than that of apparent density as a basis for a normal test of quality. The determination of density, like the chemical analysis, may be used with profit to chk the classification of hy-

Time of

Water.

Specific

Seuing.

Per Cent.

Gravity.

5 minutes.

1 hour.

Tests Of Hydraulic Materials. 31

draalic products. If the indications it affords are less complete than those of analysis they have, on the other hand, the advantage of being much more rapidly and simply obtained.

3. Tests for Invariability of Volume.

The destmction of cements and hydraulic limes almost always manifests itself by fissures which, extending progressively in all directions, ultimately lead to complete disintegration. These cracks are accompanied sometimes by swelling, sometimes, on the contrary, by contraction, but in all cases by a change in the apparent volume of the mass.

The tests for invariability of volume have for their object the recognition of these changes of volume. If there existed such tests satisfactory for all possible cases, they would be almost enough, taken alone, to determine completely all the qualities of cements, since the deterioration of these products is almost always accompanied by changes of volume.

There is to-day but a single mode of alteration which, in the present state of our knowledge, we can attempt to foresee by means of a test for invariability of volume, namely, that which results from the presence of free lime or magnesia. The effect of these substances varies with the proportions in which they are present un- combined. If this is very considerable, the setting is quickly fol- lowed by fissuring in all directions, which disintrates the mortar, and reduces it to a sandy mass. A smaller proportion produces a number of separate fissures which do not effect complete disintra- tion, but take away all the strength of the mortar, which then be- comes rapidly destroyed by exterior agencies. If the proportion be still smaller (between 1 and 3 per cent, of the total weight of the cement), it produces no visible cracks, but merely a rular swelling. The interstices between the grains of cement are enlarged, and only microscopic and discontinuous fissures are formed. The strength of the mortar is not extrenely reduced ; and in fresh water (for in- stance, under the conditions of the customary preservation of bri- quettes inteeded for mechanical tests) it remains indefinitely, after the first swelling, without further alteration. But it behaves very differently in air or is sea-water, where, as numerous experiences have proved, the increased porosity, necessarily resulting from the swelling enormously augments the alterability of the cement by the action of frost or magnesium salts. Finally, if the proportion of free lime is still smaller, it causes neither cracks nor swelling; but

32 Tests Op Hydraulic Materials.

the slacking of this lime roust Devertheless necessarily develop in- ternal tensions. Theoretically, these tensions must be injurious; but no proof on this question has been furnished as yet from prac- tice, and notwithstanding the conclusions which may seem to be theoretically justified, it is therefore still open to discussion.

Tests for invariability of volume are usually made with thin cakes preserved in fresh water, air or sea-water, in which the pro- duction of fissures, the rising of scales, or the occurrence of surface- pitting is observed. But such tests have no practical value. In fresh water, only exceptionally bad materials show visible effects; in air and sea-water, a larger proportion exhibit traces of alteration ; but only after so many months as to deprive these indications of practical usefulness. Moreover, the production of scales and cracks is very capricious, depending especially upon the distribution of the coarser grains at greater or smaller distance from the free surface ; and it reveals nothing as to the simple effect of swelling, which is, of itself, very dangerous.

The only rational method of testing for invariability of volume is to determine directly the swelling, by measuring the linear expan- sion of a mass of cement. The idea most naturally suggested is that which was first developed in such tests by MM. Durand-Claye and Debray,"*" namely, the making of measurements upon long and very thin rods. But the manipulation of these rods is delicate ; and the preparation for the test is very tedious. A much more simple and yet sufficiently precise measurement of the expansion can be made by letting the cement harden in cylindrical moulds of a di- ameter equal to their height (for example, 30 mm. or 1.2 inches), con- structed of metal, 0.5 mm. (0.02 inch) thick, slit along a genera- trix, and provided on each side of the slit with two long needles (150 mm. or 6.0 inches, for example) which serve to magnify any widening of the slit. This widening is equal to the enlargement, not of the diameter but of the circumference of the cylinder of cement. Very slow-setting cements, or limes, the water of which would evaporate or drain away in air, it is indispensable to im- merse as soon as moulded, without waiting for them to set. The immersion in water of a porous mass, filled with air, may some- times, by reason of capillary phenomena, give rise to a certain ex- pansion, and even to more or less disintegration, if the hardness be

Note of MM. Durand-Claje and Debray, prenented to the Chmmisaion Unu ficaiion dti MHhodes (VEstai,

Tests Op Hydraulic Materials. 33

insufficient. During the moulding and until setting has taken place, the mould should be kept firm by means of a suitable holder, which is removed after setting and before the measurements are bun.

The departure of the needle-pointers may be measured without hesitation to about 1 mm. (0.04 inch). The amount and espe- cially the rapidity of expansion vary considerably with the tempera- ture, and to a certain extent with the proportion of the water used in tempering.

Candimon. — The normal test of invariability of volume, to de- termine the presence of free lime or magnesia, should be made upon, the regular mixture or paste, as determined by the test for rapidity of setting. This paste should be moulded in a cylinder of the height and diameter of 20 mm. (1.2 inches) formed of metal 0.5 mm. (0.02 inch) thick, slit upon a generatrix and provided with two needles 150 mm. (6.0 inches) long, welded to the two sides of the slit. The mould, kept firm by a holder and preased between two plane surfaces applied to the two ends, should be totally immersed in water as soon as it has been filled, and should be kept so immersed until setting is complete.

For products of good quality, the distance between the points of the needles does not attain to 1 mm. (0.04 inch) in 28 days from the time of the end of setting. This test for invariability of volume, when made cold, has but little interest, since it detects only excep- tionally bad products. Mediocre cements endure this method with impunity. For practical tests, it may be sought to accelerate and exaggerate the expansion due to the slacking of lime. For this purpose it is sufficient to elevate the temperature. If the briquettes, as soon as they have set under water are plunged into boiling water. the swelling takes place in a few hours, instead of many days, and is, besides, considerably exaggerated. The products which when cold merely crack, disintegrate when hot; those which swell cold split hot; and those which contain too little lime to swell cold do so very distinctly when treated hot.

The following are the results of experiments upon a good Port- land cement, mixed with different proportions of burnt lime or mag- nesia, coming from the calcination of nitrates or the burning of impure dolomite. These products had been calcined at the tem- perature of the fusion of wrought-iron, and then passed through a sieve of 10,000 meshes per square centimeter (64,500 per square inch). The figures in the columns indicate the separation of the needle-pointers in millimeters.

Vol. Xxii.— 3

TESTS OF HYDRAULtC MATERIALS.

Portland Cement, to which has been added—

CaO. Ipr.ct

CaO 2 "

CaO 3 "

CaO 6

MgO... 6 "

MgO 10 "

Burnt do]omite,5 "

Test in cold water. Days.

Disi

ntegr

DiiiBto gntid. ated.

Test in boiling water.

Min. I Hours.

Disi

Disi ntegr

Days.

ntegr'atcd.

ated.

The following table gives the results obtained with commercial products of doubtful quality. That the swelling is in all cases due to the presence of free lime, is proved by the fact that it can be made to disappear almost entirely by heating the cement for twenty- four hours before using, with a small percentage of water:

Material.

Cold water.

Hot water.

Days.

Min.

Hours.

Days.

Unbarned Portland..

Do. heated 24 h. with2pr.ct.of water

Another unbumed Portland

Disi

ntegr

ated.

Disi

ntegr

ated.

Do. heated 24 h. with4pr.ct.ofwater

Another unbumed Portland

Do. heated 24 h. with 2 pr. ct. of water

Natural qiiick-aett'g cement

Do. heated 24 h. with2pr.ct.ofwater

Orappier cement

grited.

DiiiBle'grsted.

Conoltision. — It is possible by the test of invariability of volume in boiling water to detect the presence of free lime with much greater

/

Digitized by'

Tests Of Htdbaulic Materials. 36

certainty than by the similar cold test. This is done by placing the briquette in cold water, which is gradually heated to the boiling point and there maintained for six consecutive hours.

C. Mechanical Tests. 1. Teats for Rapidity of Setting.

The setting comprises the initial period of hardening, during which the mortar has not yet lost its semi-fluid or plastic state, in becoming solid. To determine completely the rapidity of setting, requires a series of observations of hardness at convenient intervals of time. It is usual to make only two such determinations, as amply sufficient for the needs of practice. The periods are observed, at the ends of which the hardness of the mass has attained two special, arbitrary values, corresponding, the one to the change from the semi- fluid to the plastic state, and the other to the change from plasticity to solidity. The first is called the beginning, and the second the end of the setting.

The hardness of a pasty mass is measured by a process due to Yicat, and consisting in the measurement of the effort required to force into a given quantity of the mass a solid body of definite form and size, such as a rod of steel, or, more simply, the thumb. For pastes of pure cement, it is said that setting has commenced when a round rod of polished steel 1 mm. (0.04 inch) in diameter, weighted with 300 grammes (10.6 ounces avoirdupois) does not completely traverse the mass confined in a vessel, to the depth of 40 mnu (1 .S inches). It is said that setting has terminated when the same rod, gently ap. plied to the upper surface of the mass does not sink visibly into it.

The rapidity of setting varies fidr the different hydraulic materials from one minute to ten days, or in the proportion of 1 : 15,000. It is considerably influenced by numerous conditions outside of the nature of the material itself. The principal of these are the propor- tion of tempering-water, the nature of this water, the temperature, and the conditions of preservation.

(a) Proportion of Tempering-water. — The rapidity of setting de- creases as the quantity of water increases ; but the variation in very different in different cements, as the following figures will show :

Ixerease caleaUted Dnnition of setting, on original Quantity. Mor. Minutes. Per cent

Cement. Portland A,

Temperine-water. Percent

/ 24

PortkDd B,

f 25 ' i 35

Quick-eettingWay,{ JJ

36 TiBTS OF HYDBAULIC MATERIAUB.

For cement B, each 10 per cent, of increase in the quantity of water causes the rapidity of the setting to vary as 1 to 3 ; while, on the contrary, for cement A, the relation would only be as 1 to 1.6. It may be assumed as an average for cements that the rapidity of set- ting will be affected as 1 to 2 by an increase of 10 per cent, of water in tempering. For hydraulic lime, the influence of the proportion of water is much smaller. The following are the results obtained with different limes :

Proportion of water, . . 50 per cent., 60 per cent., 70 per cent. Teil lime, duration of setting, 1 day, 2 dajB, 3 days.

Bondj lime, duration of setting, 3 days, 5 days, 7 days.

When, instead of pure cements, we take sandy mortars which always carry for a given weight of cement much more water than can 1)0 incorporated in pastes of pure oement, the setting is much slower than that observed in the tests (which are not usually made on mortars). Moreover, the different products vary greatly in this respect. For a mortar formed of one part by weight of Portland cement to three parts of sand, M. Candlot found the duration of set- ting to be from 2 to 20 times as great as that of the paste of so-called normal composition.

(6) Composition of the Tempering-water. — The influence of salts dissolved in the tempering-water varies greatly, according to the researches of M. Candlot, with the nature of the salts, the nature of the hydraulic product affected, and its degree of stateness. Sea- water often retards setting considerably. This is due to the action of its magnesium salt and not to the chloride of sodium, solutions of which appear to have no effect upon the duration of setting. Salts of calcium behave, as to retarding action, like those of magnesium. The retardation is less pronounced, the more stale the cement.

The following are examples of the comparative action of fresh and sea-water upon different products :

Cement A, . Cement D, . -j

(c) Temperature. — The rapidity of setting increases rapidly with the temperature, as the following figures of M. Candlot show, but the rate of increase varies among the different products.

Tempering- water. Per cent

Duration of setting. Fresh water. Sea-water.

Fresh,

20 minutes. 5 hours.

Stale,

16 hours. 16 hours.

Fresh,

14 minutes. 20 minutes.

Stale,

3h.,50m. 5h., 45 m

Tests Of Hydraulic Materials. 37

Cement E, Cement F,

Tempemtare.

Duration of seUIng.

Deg.C.

Hours.

Minutes.

f

( 25

( 7

( 30

(d) Oondiiiona of Preservation. — Setting is much slower when the paste has been preserved under water than when this has been done in the air, even though the air has been kept saturated with moisture so as to prevent all evaporation. There is produced under the influ- ence of mere weight an outward dripping of the water which per- vades the mass, and this efiect is the more marked the slower the setting. If a paste of cement, or especially of lime, which has just finished setting in air, be immersed, it will be observed that the ab- sorption of water causes a marked re-soflening. The setting in air, then, is not a true setting; it is due in large measure to the mechani- cal elimination'of a portion of the water. The following are aver- age comparative results for hydraulic materials of different classes :

Duration of setting.

Material.

In air.

In water.

Quick-eetting cement, .

7 minutes.

15 minutes.

Portland

2 hours.

3 hours.

Slag . .

. 6 "

8 "

H}'draalic lime,

. 36 "

5 days.

In order to render tests of setting concordant and comparable, it is thus indispensable to keep as constant as possible these difierent elements which affect its rapidity.

For tempering, ordinary drinking-water, such as is found in all countries, should be employed ; that is to say, water which contains no other substance dissolved in noteworthy quantity than the bicar- bonate of calcium. The temperature should be kept between 15 and 18° C. throughout the tests. The paste should be immersed in water as soon as tempered, and left there until setting has ended. These three indispensable conditions are simple and easy.

(e) Normal Paste. — By far the most delicate point is the determi- nation of the proper proportion of water to be used in tempering. From the practical standpoint it would be interesting to perform the test upon a true mortar, like those employed in construction. But it has been admitted hitherto that there are no practicable means of following precisely in the test the hardening of sandy mortars.

38 Tests Of Hydraulic Matebiaus.

The tests are always made — a fact to be regretted — upon pastes of pure ceroent. As will be shown below, this usage should be abandoned.

The pastes to be tested are tempered as closely as possible to uniform consistency, called the normal consistency, for the definition of which there are three equivalent methods. One of these, amply sufficient for all cases, is based upon the condition that the paste shall not drip from a trowel, bat a minute addition of water shall make it do so. The other methods, which can be considered as normal because they are independent of the judgment of the operator, utilize the measurement of the foree required to cause a round rod or disk, 1 centimeter (0.4 inch) in diameter, to penetrate a given quantity ot the paste. As M. Tetmayer proposes, the condition may be that such a rod, weighted with 300 grammes (10.6 ounces), shall stop sinking 6 mm. (0.24 inch) from the bottom of a vase 40 mm. (1.60 inches) deep filled with the paste. From the strictly exj>eri- mental standpoint, a more satisfactory definition would be obtained by adopting the requirement (substantially equivalent to the fore- going) that a circular disk, 1 centimeter in diameter, carried by a slender rod, and loaded to weigh 50 grammes (770 grains), should stop half-way in a vase like the above. It is, in fact, at this point of half-depth that a given load produces the maximum variations in sinking, and permits therefore the greatest precision of observa- tion. Testing a disk instead of a cylinder avoids the interference of lateral friction, which varies with the degree of oxidation of the metal of the cylinder.

These three methods are more than sufficiently precise for con- trolling the consistency of the paste, but do not indicate with great accuracy the corresponding proportion of water, for the consistency varies considerably with the more or less prolonged mixing. There is an uncertainty of about 10 per cent., which, in some cases, might have a considerable influence upon the corresponding rapidity of setting. To make the mixture as susceptible as possible of com- parison, all its conditions should be prescribed. Among the possible conventional rules, the following may be recommended : Mix with a spoon, in a metallic dish or casserole, for five minutes; use a con- stant quantity of 2(X) grammes (7 oz.) of water, to which succes- sive additions of cement are made, until the instant when the de- sired consistency has been attained.

(/) Methods of MedsuremerU. — In measuring the force required to cause a metallic rod of given section to penetrate into a paste, it

Tbst8 Op Hydraulic Materials. 39

is found that this force does not by any means increase directly with the depth attained. . The increase is at first rapid to a short distance below the free surface; then tlie force remains practically constant for medium depths; and finally it again increases rapidly at a short distance from the bottom. As a result of this law of variation in the force of penetration the maximum precision of test is secured from observations at mid-depth. The commencement and the end of setting, according to Vicat's definition, will thus correspond to weights of 50 and 3000 grammes (1.8 and 106 oz.) for penetration to mid-depth in a vase 40 mm. (1.6 in.) deep.

The direct observation of the penetration or non-penetration of a rod with the constant weight of 300 grammes (10.6 oz ) is less precise, because of the nearness of its two extreme limits. The variation in sinking is extremely slow, and the uncertainty as to the duration of setting is at least one-fifth for the beginning and half as much for the end of the process.

The end of setting may be simply tested by pressing with the thumb. When under a strong effort of the arm the thumb does not produce a depression which can be detected by lightly passing a finger over ic, the setting may be deemed complete. This test is much less uncertain than one might be tempted to believe. In a given operation, the uncertainty as to the duration of setting does not amount to one-fifth. Between two different tests the variation may be a little gter, because the thumb may not be pressed with the same force at different times ; but the difference in observed time is not half as great as that in the force applied, because, during this phase of hardening, hardness increases in a much greater pro- portion than the lapse of time.

For the test with the thumb, as well as for the observation of the end of setting by Vicat's process, it is indispensable that the upper surface should be perfectly smooth — neither diluted with water, nor carbonated. This is secured by filling the mould with paste to overflowing, and smoothing it off with a glass plate, which is left upon it during the first period of immersion.

If it be desired to estimate the degree of precision attending this determination of the rapidity of setting, the consideration of the var- ious causes of uncertainty above stated forces the conclusion that it is impossible in practice to reckon upon a smaller uncertainty than one-half, especially with some relatively quick -setting materials which are especially sensitive to the influence of variations in the quantity of tempering-water and in temperature.

40 Te8Tb Of Hydraulic Materials.

Condvsion. — The normal test folr setting sbonld be provisionally made with a pure paste of normal consistency, defined by the re- quirement that a plane disk, 10 mm. (0.4 in.) in diameter, weighted to 60 grammes (1.8 oz,), shall penetrate to mid-depth in a maos 40 mm. (1.6 in.) deep. For the beginning and the end of setting, the moments should be observed when a round rod, 1 sq. mm. (0.0015 sq. in.) in section, weighted to 50 grammes (1.8 oz.) and 3000 grammes (106 oz.) respectively, penetrate to mid-depth in a mass 40 mm. (1.6 in.) deep.

The paste should be immersed in water as soon as moulded, and should be tested with the rod at intervals of time between one-fifth and one-tenth of the total probable duration of setting. In deter- mining the beginning or the end of setting, the mean of two con- secutive observations should be taken, one of which showed a depth of sinking above and the other below mid-depth.

It is useless to '' prick '' the mass more frequently in the hope of attaining greater precision, because, by reason of the inevitable irreg- ularity of the mass at different points this procedure would give . consecutive results contradicting each other.

In the practical test for setting, tlie normal consistency of the paste is fixed by the condition that it shall be at the point of ceasing to adhere to the trowel. Setting is considered to have ended at the moment when the thumb, pressed with sufficient force (say about 10 to 15 kg. or 22 to 33 lbs.) ceases to produce a depression distin- guishable by lightly ])assing the finger over the spot pressed.

{g) Teat for the Setting of Mortars. — If the test for setting has not been universally practiced upon mortars, the cause lies in prejudices against which we cannot too strongly protest A degree of precision much greater than it can in any way achieve having been attributed to the test for setting when applied to pure pastes, it has been fancied that this imagined precision would be impaired by the presence of irregularly distributed grains of sand. In fact, the test of setting is at least as accurate for mortars as for pure pastes. It is to be hoped, therefore, that this inexcusable error will be abandoned in the near future.

It is necessary, as preliminary, to fix for mortars the three indis- pensable definitions of the normal consistency, the beginning and the end of setting, and the principles to be followed in establishing these requirements are easily perceived.

It is indispensable to employ soft mortars, analogous to those used in actual construction, without using, however, so much water

TESTS OF HYDRAULIC MATEBIAIjB. 41

that it will too rapidly separate from the mortar when left undis- turbed. In mortars, even more than in pure pastes, the addition of a very little water suflBces to transform them from the sandy state into a liquid slime. During this transition many properties vary, with great rapidity, and may be utilized in defining the normal consist- ency desired ; the exudence of the water under the pressure of the trowel ; the resistance offered to tempering ; the mobility of the material under the influence of a greater or smaller angle of incli- nation, etc. As a normal procedure, 300 grammes (10.6 oz.) may be prescribed as the weight required to make a disk or round rod 1 centimeter (0.4 in.) in diameter penetrate half-way throuph a mass of mortar 40 centimeters deep. This will give, for the mortar made in the proportion 1 : 2 with normal sand (which will be defined under the head of tests for rupture) a very satisfactory plastic con- sistency. For leaner mortars, it is necessary to prescribe a greater weight ; for instance, 600 grammes (17.6 oz.) for a 1 : 5 mortar.

To determine the moment of setting, a rod must be used, the di- ameter of which is a high multiple of that of the grains of sand. For the normal sand described below (about 1 mm. or 0.04 inch in diameter), a rod-diameter of 1 centimeter (0.4 in.) is quite suitable. It must not be attempted to make such a rod penetrate the mortar to mid-depth, because the latter would become disintegrated to a con- siderable distance, rendering continued tests impossible; and also, because much too great a force would be required. It must suffice to observe the moment when the rod, placed gently upon the surface, ceases to produce an appreciable depression. But this observation is much easier than for pure pastes, by reason of the large diameter of the rod. Instead of judging by the eye, which too easily con- founds a depression with a simple smoothing, the touch should be em- ployed. By gently rubbing the finger over the surface it is possible to recognize with great precision the moment when depression is no longer produced.

The rod should be so weighted that the commencement of setting shall correspond with the moment when the mortar can no longer be handled with a trowel, and the end of setting, as shown by the rod, shall coincide with the end as observed by the thumb-test, which is equally applicable to mortar and to pure paste. For a rod of 1 centimeter (0.4 in.) diameter, the weight at the commencement of setting should apparently be about 400 grammes (0.88 lbs.); at the end about 10 kilos (22 lbs.).

Wben such rules have once been established, the normal test for

42 Te8Tb Of Hydraulic Materials.

setting will be made not upon a pure paste, but upon a definite mortar — for instance the 1 : 2 mortar made with normal sand.

2. Test for Resistance to Rupture.

The value of tests for mechanical resistance does not lie, as might be supposed, in the fact that the mortars are to be subjected to heavy stresses in the conditions of actual use. Such stresses, per haps, never equal 10 kilos per square centimeter (142 pounds per sq. " in.), while in the tests they may go to several hundred kilos. In this respect the conditions of use are very diflTerent from those of metals.

The principal reason for tests of rupture for cements is that there ought to be a certain proportionality between their strength and the* quantity of active hydraulic material which they contain. Now, their resistance to deterioration by external agents is evidently the greater the larger the amount of their active constituents, and henoe the greater their strength. But in order that this proportionality may exist, all of the many other circumstances which could influ- ence mechanical resistance must be the same in any two cases com* pared. It is impossible to fulfil this condition completely, and thei must always be some reserve in the interpretation of tests of rupture. If the proportionality were rigorous, the classification of such differ- ent products would be the same, whether tests were made on pure pastes or on mortars, by compression or by tension, in fresh water or aea water, etc. But this is by no means the case. These differ- ent classifications give only analogous, not identical, results. It is only within the limits of precision of their agreement with each other that the results obtained can be reckoned upon. When it is desired, for example, to classify Portland cements by tension- or compression-tests upon pure pastes, or 1 : 3 mortars, it is found neces- sary, if the results are to be identical, to modify some of the figures to the extent of 50 per cent. The reasons for these discrepancies are many. Experimental errors of measurement, as will be shown below, count for more than 20 per cent. Another cause, likewise very important, is the unequal compactness of the briquettes sub- jected to rupture. A finer cement will require more water in tem- pering, but the difference will not be the same for the mortar and the pure paste. Thus it is, that in comparing Portland and lag- cements, the former are favored by tests of the pure paste and the latter by tests of mortar.

TEarrs of hydraulic materials. 43

Numerons oonditions apart from the natare of the cement ioflaence its resistaoce to rupture.

The form and size of the briquettes have a cousiderable impor- tance which was long ignored. It is difficult to establish a rela- tion between the resistances of briquettes of different sizes, even though geometrically similar. Thus, in tension-tests, there is no proportion between the force and the section of rupture. If we would find any such proportionality, it must rather be sought be- tween the force and the perimeter. This fact has been incontestably established by the experiments of M. Durand-Claye, who has given at the same time a plausible explanation. It is necessary, then, to adopt for tests of rupture a uniform and perfectly defined type of briquettes. There is no use in trying to connect the figure of the rupturing stress with the unit of surface of the section. It must suffice to take this figure just as the test gives it.

The composition of the briquettes, that is, the amount and nature of the water and sand employed, and above all, the compactness of the mortar and its richness in cement, have great influence upon the rupture. It is easy to regulate uniformly the nature of the water and sand employed, but it is almost impossible to secure uniform compactness.

The nature of the water, as fresh or sea-water, has little influence upon the resistance under brief stresses, which are the only ones to be utilized in practical tests. It is only ailer a year, and for pastes of pure cement, that this influence becomes considerable. Neverthe- less, it is preferable always to use fresh water.

The nature of the sand, that is, its chemical composition, and still more, the form and size of its grains, have a great influence on its resistance. It is necessary to use in tests a normal sand which can always be reproduced in identical character. In France use is made of the quartzites of the mountain of Roule, near Cherbourg, ground in ordinary mills to a size intermediate between those which pass respectively sieves of 64 and 144 meshes per square centimeter (413 and 929 meshes per sq. in). If breaking by hand with a ham- mer, or crushing with a jaw-crusher, be substituted for grinding in a mill, there will be obtained, instead of rounded and polished grains, angular splintery fragments, producing very different results in the tests. Id many respects, the use of calcareous sand, made by grind- ing marble or compact limestone, seems preferable. By reason of the cleavages of the carbonate of calcium, it yields a sand much more independent of the process of grinding. That its substitution for

44 Tests Of Hydbaulic Materials.

the quartzose normal sand employed in France would not otherwise cause appreciable change in resistances to rupture, is shown by the comparative investigations in this line of MM. Alexandre and Feret.

The compactness of the mortar and its richness in cement are eminently variable and difBcult to define, yet enormously influential upon the resistance to rupture. These variations are probably the chief source of the serious uncertainties attending rupture-tests. According to the experiments of M. Feret, the resistance to rupture of different mortars, composed of a given sand and cement, depends principally upon the relation in each mortar of the quantity of ce- ment to the volume of the interstices occupied by water and air. Variations in the proportions of cement, sand and water, unequal ramming of the briquettes, and the variable volume of the impris- oned air come into play only so far as they modify the characteristic relations in question.

It appears, however, upon discussion of M. Feret's experiments, that the characteristic relation to be considered is not as he suggests,

Cement Cement

or

Volume (water + air) Volume (mortar, sand, cement)?

but rather

Cement Cement

or

Volume (cement + water + air) Volume (mortar — sand)

which seems, a priori, more reasonable.

The resistance to tension will vary in proportion to this charac- teristic relation ; and to crushing, in proportion to its square.

Fig. 2 shows the curve constructed according to the experiments made by M. Feret upon the same cement with two different natural sands, one rather coarse and the other fine. The cement had been passed through a sieve of 5000 meshes per square cm. (32,500 meshes per square inch) to remove the larger grains, which could not have had time to become completely hydrated during the two months which were allowed for hardening.

The variations of observations from the theoretical line are not greater than the experimental errors involved in such observations.

If these results should be verified for all cases, it would follow that one or another composition of mortar might be indifferently used for the rupture-tests, provided that the compactness of the mortar, that is to say, the volume of cement and of sand contained in a given quantity of it, were determined at the same time.

Tests Of Hydraulic Materials.

(a) MeOioda of Bupfure. — Among the principal possible methods of rupture only three have been employed hitherto in tests of cement — bending, tension and compression.

Bending, which is the simplest in experimental application, was employed by Vicat, who, however, under the influence of erroneous theoretical notions, discarded it in favor of tension, to which latter

Volume of cement Volume (mortar - sand;

method was attributed the advantage of giving resistances propor- tioned to the section of rupture.

Bending-tests deserve, by reason of their simplicity, to be recom- mended in connection with works of secondary importance. They do not secure great precision, because the force of rupture thus ap- plied, is largely affected by the condition of the superficial layers, which are most exposed to accidental causes of alteration. It is indispensable here, as in all rupture-tests, to apply the increment of

46 T£8Tb Of Hydraulic Materials.

stress {la vme en charge) at regular prescribed rates, avoiding espe- cially every sudden increase which will produce the anticipated rup- ture. The desired result is very easily obtained by suspending from the middle of a briquette, supported at the two ends, a bucket of known weight, into which water or shot is allowed to flow at a prescribed rate.

The method of tension is at present most widely used, but the preference for it is not well founded. Here, as in rupture by bend- ing, only the surface of the briquette acts in a really useful way, and its inevitable irregularities and alterations so greatly affect the pre- cision of the results that they can in no case be trusted nearer than about 20 per cent. This preponderant influence of the superficial parts was first shown by the fact that the resistance of briquettes of different sizes increases, not with the section, but, on the contrary, with the perimeter. Finally, M. Durand-Claye has shown that the interior of a briquette may be removed without notably diminishing its resistance to rupture by tension, and has given a complete theo- retical explanation of the phenomena which seemed at first sight paradoxical. On the other hand, tension-tests do not possess, like bending-tests, the advantage of great simplicity. They require, like those of compression, expensive machines and delicate manipulation.

The method of rupture by crushing, which seems at first the most rational, was abandoned because it gave very discordant figures. This was because, in the hydraulic presses originally used for this test, it was impossible to distribute the pressure uniformly over the surfaces. The test-blocks were crushed first at one edge, and became disintegrated by degrees. This objection can be removed, however, by very simple devices, and the crushing-test can be made to-day to give results of at least twice the precision of those of other rupture-tests.

A second hindrance to the adoption of the crushing-test was the widespread prejudice that it required the treatment of very large blocks, and, &s a necessary consequence, the use of powerful, costly and cumbrous machinery. But it can be made with cylinders of 25 mm. (1 inch) diameter and the same height, which dimensions will be suiBcient, even for mortars, provided the sand employed is not coarser than the normal " sand, defined above. The crushing- stress need not exceed 5000 kilos (1 1,000 pounds), which can easily be exerted by hand, with the aid of a vice actuated by gear. The test- cubes, though perhaps a little difficult to mould, have, on the other hand, the advantage that their resistance to crushing can be directly

Tests Op Hydraulic Materials. 47

compared with that of bricks or stones. Cabes of 7 centioielers (2.8 inches) are found convenient.

The grt advantage of crushing-tests is that they measure the resistance of the total mass of the briquettes not merely that of the sapertictal portions, which are exposed to various causes of altera- tion. Hence the results should be more uniform and more inde- pendent of accidental circumstances than those of bending or tension- tests.

The sourcSs of error involved in rupture-tests of any kind are numerous and important. The accidental errors of experiment pro- ceeding from the machinery may easily amount, under the most favorable conditions, to 10 per cent, of the observed value. Besides the uncertainties inherent in the rupture itself, there are those due to the mixing of the paste, which is never the same for two opera- tions, and to the manner of preservation, which differs in different laboratories. The temperature, the conditions of the removal of the water, and the access of carbonic acid from the atmosphere are not identical. An estimate of 10 per cent, as the value of these errors is certainly below the truth. It is evident, therefore, that in different, scries of tests made upon the same material a closer agreement than within 20 per cent, can never be counted U|)on. In some cases the uncertainty may be much greater, the results, for instance, upon pastes of pure cement kept in sea-water for a suiBciently long time, varying as 1 to 2. In that case the briquettes acquire a peculiar fragility which leads to enormously discordant figures of rupture. With the slightly-kneaded and heavily-pounded mortars which are at present frequently employed in testing, the uncertainty may easily amount to 60 per cent., because the inevitable irregularities of pounding forcibly produce variations in compactness. From no pounding to the maximum pounding the resultant resistance varies as 1 : 3. Operators working together and trying to pound alike cannot pro- duce results agreeing more closely than within 20 per cent It is comprehensible, therefore, that the work of separate operators may vary by 50 per cent. For this re&son the use in such tests of dry pastes, which need to be pounded, should be positively forbidden.

Conclusion. — The normal rupture-test should be made by the compression of cubic brkjuettes of 7 centimeters (2.8 inches), com- posed of a mortar of normal sand, containing, by weight, 1 part of cement to 2 of sand, tempered with water to the normal plasticity, as defined above under Test for the Setting of Mortars. The bri- quettes, tempered with fresh water, should be preserved in the same

Ic

48 Tests Of Hydraulic Materials.

water, protected as far as possible from carbonic acid, and at a tem- perature kept between 15® and 18® C. The test should be made after 28 days. Pressure should be uniformly applied, with an in- crease of 6 kilos per minute.

The normal sand should be either a quartzose sand of grain be- tween 64 and 144 meshes per square centimeter (413 and 929 meshes per sq. in.), prepared by grinding in a mill, or a calcareous sand of the same fineness, obtained by grinding compact, hard limestone.

The practical test will be made by bending, upon pastes of pure cement of normal consistency (t., barely not dripping from the trowel) moulded into bars 120 mm. (4.8 in.) long by 10 mm. (0.4 in.) square, and preserved in water at about 16® C. They will be broken at the end of 28 days by placing them upon two knife-edges 1 decimeter (4 in.) apart, and applying at the middle a weight increasing regularly 2 kilos (4.4 lbs.) per minute.

(6) Tests of Rupture in Hot Water. — These tests are made upon briquettes which have been preserved in hot water, after the termi- nation of setting. The reason for using hot water is, that the chemi- cal reactions of hardening are considerably accelerated by it, so that they may be nearly or quit€ completed within 28 days, which is not the case in cold water for very slow-hardening products, like hy- draulic limes. For finely ground Portland cements, the compara- tively rapid hardening in cold water is mostly complete after 28 days, and the proportion of active constituents may be measured by the mechanical resistance, determined at that date. But a similar procedure with limes would be liable to grave errors. The firat rank would be accorded to those limes which harden more rapidly at first because of a slightly greater fineness or an incomplete slack- ing, but which may contain a much smaller total of active constitu- ents. The hot-test, in which the briquettes are immersed in boiling water immediately after setting, and so kept for 28 days, or even for 7 days only, gives the same classification as cold tests of much greater duration. It will be understood, however, that the final re- sistances obtained are not the same. But the experiments made hitherto with the use of hot water have not been numerous enough to justify the adoption of this method for better or worse as a normal test. It is simply a question which deserves to be studied.

For the tests of rupture, as well as for those of rapidity in setting, one should keep in miifd not to depart very far from actual usages. It would seem rational, however, to make an exception in the fol- lowing two particulars :

Tests Of Hydraulic Materials. 49

1. In all mortar-tests, which are made for the purpose of compar- ing different cement?, it is customary to mix the cements with a defi- nite weight of sand, such as 1 : 2 or 1 : 3, and to vary the quantity of water till the same consistency is secured for the different mixtures. A far preferable practice in all cases would be to mix the cements with the same weight of water, say 50 per cent., and to vary the addi- tion of the sand in order to obtain the uniform consistency. By this method the same weight of cement would be obtained in the same space or void left by the sand, and comparisons of different cements would be rendered thereby far more satisfactory.

2. Id making mortar-tests it is current practice to use either a soft paste of the consistency employed in actual working, or, according to the Grerman method, a very dry paste which has to be strongly beaten in the moulds. Neither of these consistencies is suitable for testing purposes — very dry pastes yield resultsthat vary greatly with the intensity of the beating, this being dependent on the endurance of the workmen ; soft pastes, on the other hand, remain filled with air-bubbles and allow the escape of some of the water, thus marring the uniformity of their composition. The best results are obtainable with a paste of intermediate consistency— one that is neither dry nor yet soft, but moderately firm, and which will allow the water to flow out upon very gentle tapping that does not fatigue the workmen.

Such are the only tests which can be recommended with entire confidence and which should be required in specifications. They are four, determining the fineness of grinding, the invariability of volume the rapidity of setting and the resistance to compression.

Besides these tests, which may be considered as definitely estab- lished in principle, and only open to future modification as to details, there are others, the objects of which are equally interesting, but which have not been sufficiently studied to permit their efficacy to be determined or the best method of their application to be pre- cisely defined.

The indications which will here be given should be considered as interesting suggestions to be followed, but not as established results. We have to do with only two classes of such tests, namely, tests for the mechanical properties of mortars, and tests for the detection of alnminate of calcium.

3. Tests for the Mechanical Properties of Mortars.

It is pretty generally agreed that the tests of rupture by tension or compression described above, measure with precision the meaa-

Vol. Xxii.— 4 Ic

50 Tests Of Hydraulic Materials.

ical properties of mortars. They do indeed measure certain me- chanical properties, but not those which are utilized in construction. When masonry is broken by tensile stress the mortar, in most cases, does not break, but is detached from the ashlar and stones — it is its adherence which is concerned. When a mortar is crushed in the joints of an arch, it is not broken by shearing, as in compression- tests for rupture. It crushes itself upon itself; the grains of sand sliding by one another, to reduce the volume of the original interstices among them.

The measurement of adherence and that of resistance to crushing, properly so-called, may be the object of special tests; and apparently few and very simple investigations would suffice to determine the degree of usefulness of these new tests, and to settle their experi- mental methods upon undisputed foundations, if they are to be admitted.

Adherence depends upon certain conditions, of which the chief are the nature of the cement, the chemical composition of the stones and the physical state of their surface. To separate the effect due to the cement, it is necessary to operate with stones, the chemical character of which is perfectly known and is analogous to that of the sand employed in construction, for instance, quartz and marble, of which the surface of contact has been treated with emery of a prescribed fineness. Numerous forms of experiment, based on the principle of the tensional testing-machine have been proposed for this purpose. They are all designed to make a heterogeneous briquette, one-half of which is formed by the stone to be joined and the other by the mortar to be tested. M. Candlot has attempted to avoid the incon- venient necessity of having a large number of " stones of adherence " by employing as " block of adherence " artificial stones, made with a mortar of one part cement to two parts of sand obtained by grind- ing the quartzite or marble upon which the adherence is to be studied; tlie results being the same as upon blocks of massive stone. If this should be verified as a fact, it would considerably facilitate tests for adherence.

But in any event it would be necessary, before taking pains to define a normal adherence-test, to make sure that it presents any utility; that it will teach something more than the tests already ad- mitted. It is indeed certain that cements of different nature, such as Portland and slag-cement, do not possess the same property of adhesion, and that this property is the more developed, the riclier the cement in active elements ; but it is not demonstrated, and even

Tests Of Hydraulic Materiau3. 61

remains very doubtful, that cements of the same class, Portland for example, which will prove themselves equivalent under the ordinary rupture-tests, may not exhibit as to adhesion variations considerably larger than the errors of observation.

A systematic study of this question will probably lead to the con- clusion that adhesion-tests, while interesting as indicating the most favorable conditions of employment for mortars destined to certain exceptional uses, or in aiding the selection of preferable mortars for such uses, present no interest of any kind as tests of inspection for the acceptance of material.

The crushing of mortars may be studied in the laboratory under conditions similar to those of practice. It is sufficient to enclose the mortar in a rigid mould, open on the top only. When a steel piston of given area, for instance 5 square centimeters (0.77 square inch) is pressed upon the free surface of the mortar, a given force effects a penetration, the depth of which increases in proportion to the increased force — at least, so long as the depth does not exceed 2 or 3 mm. (0.08 or 0.12 in.). Either the force which effects the initial penetration, or still better, the increase of force necessary for a given increase of penetration, may be measured.

Such examples would be certainly interesting in studying the most suitable composition of mortars for certain particular cases, or to determine the period, at the end of which the centering of an arch might be taken away ; but it does not appear, a ptiori, that there will be found reason to include these tests among those of in- spection ; doubtless they will teach no more than the ordinary test for rupture.

The same conclusion may be expressed concerning tests of suscep- tibility to frost This is another property which is connected with the mechanical qualities of mortars. Those which are rich in active constituents and which endure well the tests of rupture will evidently be best suited to resist the expanding-stress of ice; a separate test seems unnecessary. But if such tests have no interest for the in- spector, they are, on the contrary, very important to the student of mortars, to determine the most favorable nature and proportion of sand for constructions in air. In stones, the frost-test has a value, because these possess at the time of delivery a definite porosity, whereas the porosity of mortars, on the contrary, results exclusively from the conditions of use and not from the quality of the product delivered.

52 Tests Of Hydraulic Materiau9.

4. Tests Relating to the Aluminate of Calcium.

A deeper study of the properties of the aluminates of calcium will certainly lead to the discovery of methods of testing permitting the estimation beforehand of the resisting qualities of a cement to the action of air and of sea- water. The experiments made at the present day upon the stability of cements in air and sea-water cannot serve as inspector's tests, because they do not give results until after many months often more than a year. But it is well-nigh certain that the unequal resistance of cements employed under these conditions is related to their greater or smaller tenor in aluminates of calcium; and it seems practicable to recognize quickly the presence of these aluminates.

Practice with constructions in air has shown that certain cements, — slag cements, and some quick-setting cements — tempered to pure paste, end after one or two years' exposure to atmospheric extremes by losing every kind of resistance. Sometimes they decrepitate into small fragments ; more frequently only fine fissures of contraction are produced. Exclusively siliceous cements do not exhibit such ac- cidents ; pure aluminates of calcium, on the contrary, do so in a much more marked degree, cracking in all directions aftr a few days' ex- posure to the sun.

Similar observations have been made by M. Candlot in the use of cements in sea-water, the predominant cause of their disintegra- tion being the action of the sulphate of magnesium upon the alumin- ates of calcium.

If this influence of the aluminates of calcium should be conclusively established, the following characteristics might be tested to detect the presence of these compounds :

1. We might seek to utilize the alteration of the hydrated alumin- ates in a dry vacuum. Briquettes of pure aluminate left three months under such conditions have cracked and lost all resistance.

2. We might utilize the action of a temperature of 100 C, which being prolonged for a few minutes only, suffices to remove from some products an important part of their resistance, even in the absence of any swelling, which might be caused by the presence of expansive constituents.

3. Finally, we might avail ourselves of the properties discovered by M. Candlot in dilute solutions of chloride of calcium, retarding considerably the setting of the aluminates of calcium, otherwise extremely rapid in fresh water. The hydrate of calcium appears to behave in the same manner. ,

oeologioal distribution of the useful metals. 53

Conclusions.

In the present state of knowledge and practice the following tests are to be recommended :

1. Fineness of grinding — acxrding to the residue from a sieve of 900 meshes per square centimeter (5850 per square inch).

2. Resistance to crushing— determined upon blocks of mortar composed of 1 part of cement to 2 of sand, tempered to plasticity ; the sand having a uniform grain of about 1 mm. (0.04 inch) diameter and the blocks being cubes of 70 mm. (2.8 inches) or cylinders 25 mm. (1 inch) in height and diameter.

3. Invariability of volume in boiling water.

4. Rapidity of setting — the mortar being composed of 1 cement to 2 sand, and tempered to normal consistency.

5. A fifth test, for the detection of aluminates, should be the object of investigations conducted in the hope of becoming able to foretell the stability of cements in air or sea-water.

GEOLOGICAL DISTBIBUTION OF TEE USEFUL METALS llf THE UNITED STATES.

BY S. F. EMMONS, WASHINGTON, D. C. (Chicago Meeting, being pan of the International Engineering Congress, August, 1893.)

The first paper which appears in the published Transactions of our Institute is that read by our respected Secretary at its first meeting in Wilkes-Barre in May, 1871. It is entitled "The Geo- graphical Distribution of Mining Districts in the United States," and presents a brief but masterly review of what was known of the distribution of our deposits of useful minerals, particularly the metals, not only from a geographical but from a geological stand-point.

At the request of Dr. Raymond I agreed, somewhat hastily, per- haps, to write for this occasion a brief sketch of the geological dis- tribution of the deposits of the useful metals in this country, in the light of the increased knowledge of the present day. In the time given no personal investigation was possible, and as it was therefore out of the question to attempt to make anything that could be con- sidered an original contribution to the history of our ore-deposits, I have been obliged to limit myself to an examination of such pub-

54 GEOLOGICAL niSTRIBDTION OF THE USEFUL METALS.

lished data within my reach as bore upon this subject, and could be consulted in the brief time I have been able to give to it. Had the geological investigations undertaken by the Tenth Census been con- tinued systematically by the United States Geological Survey or by the Eleventh Census, it might have been possible to make a fairly complete review of the subject. As it is, the principal result of my examination has been to show how very unequal and in many directions extremely meagre are the data of any kind that are avail- able, and to demonstrate the great need that exists for a systematic investigation of this important subject by some scientific organiza- tion, for its field is too vast to be covered by any single individual, and will be of little permanent value unless carried out on some uniform plan by which the relative accuracy of its results may be assurecl.

The utmost that T can hope, therefore, for the very imperfect, and from a statistical standpoint possibly somewhat inaccurate, review here presented, is that it may oflTer a suggestion to other workers in the field of lines of investigation that may be profitably pursued in the future.

In the twenty-two years that have elapsed since Dr. Raymond's paper was written, many important contributions have been made to our knowledge of the geological structure of the continent, but a great part of these contributions, especially in late years, have been rather in the line of modifications and reversals of preconceived theories, than in the firm establishment of new ones. We seem now to have removed most of the unstable stones from the foundation of our geological knowledge, and to be nearly ready to build up a perma- nent structure in the immensely enlarged field that progress in vari- ous lines has opened to us. In like manner the special study of ore-deposits and of their relations to geological structure which had hitherto been rather neglected by field geologists, has in the last decade received more attention, though perhaps not as much as it deserves ; many false conceptions have been cleared away, and im- portant progress has been made toward a more rational method of correlating their phenomena.

It would occupy too much space to give a complete list of the various papers and authors consulted in this examination, and it must suffice to say that they have been found for the most part in publications from the following sources : Tenth and Eleventh Census; Director of the U. S. Mint; various United States and State Geological Surveys ; American Institute of Mining Engineers; American Journal of Science; American QeologxH; Colorado Scientific Society; Engineering and Mining Journal; Zeitaehrifi JUr praktitehe Oedogie, etc.

Geological Diotbibution Of The Useful Metals. 65

In the realm of eruptive or igneous rocks, the great change that has come about has l)een the gradual abandonment of the theory that the mineralogical or structural character of the rock is a criterion of its age. It is no longer a necessary conclusion, for example, that because a rock is a trachyte, rhyolite or basalt, it is of Tertiary or later age. Well defined rocks of types formerly classed as Tertiary have been found to be as old as Cambrian, and the petrographical character of a rock is now admitted to be dependent on other causes besides geological age. It still holds good that most of the so-called volcanic rocks are of Tertiary or recent eruption, but many crystal- line rocks, actually granitoid in structure, are also of Tertiary age and it is now necessary for the geologist to determine the age of the various eruptives of each district by their relations to sedimentary rocks of known age.

From the internal structure of the eruptive rock, whether more or less completely crystalline, one can judge whether it has consolidated at considerable depths and under the pressure of great weight of superincumbent rock-masses, hence very slowly, or at or near the surface and with comparative rapidity. In many cases this furnishes a further aid in the determination of the relative age of different varie- ties of eruptive rock occurring in a given region.

As regards the origin of eruptive rocks and the determination of the natural order of succession of the many types distinguished by their different chemical and mineralogical composition, most of the theories hitherto held are gradually being discarded or merged into what may be called the theory of differentiation of igneous magmas, which is now being worked out by the more advanced petrologists in this country and in Europe, and which promises to throw impor- tant light upon the origin of ore-deposits also. It proceeds from what is known as Soret's principle that in a cooling solution of a salt, the salt will concentrate in the parts of the solution which cool first, and reasons that in a molten rock-magma a similar sepa- ration or differentiation of substances may take place. For in- stance, it has long been observed that in eruptive dikes of moderate dimensions, those portions of the dike adjoining the walls, which, when the matter forming the dike was injected, may be supposed to have been relatively cold have a finer-grained texture than the in- terior of the dike, and in certain cases there is a concentration of the more basic minerals composing the general mass in the outer zone or in different parts of the rock mass. It is assumed that on a larger scale the different varieties of eruptive rock which belong to one

56 Qeolooical Distribution Of The Uhbful Metals.

general period of eruption in a given district and are, so to speak, consanguineous, proceed from one general molten magma in the depths of the earth ; and that in this magma a chemical and mineral- ogical differentiation takes place by virtue of which each successive eruption of igneons rocks differs in character from the one which has preceded it, according to laws not yet fully made out, but which, according to the preponderance of chemically acid or basic material, undr varying conditions, produce in the erupted rock a correspond- ing preponderance of acid or basic minerals.

On the other hand, considerable advance has been made in the classification of the crystalline rocks, which were formerly all grouped indefinitely as Archaean. More detailed and systematic field- studies in the areas occupied by typical series of crystalline rocks have shown that there are several series that can be distinguished as originally sediments made up of debris of older series, with a greater or lens proportion of eruptive material, in which there is evidence of the former existence of organic life, and which are older than the oldest known Cambrian beds, and younger than Archaean, the latter term being limited to non-clastic rocks, in which there is no evidence of life. Petrological investigation, in the light of the most ad- vanced studies in this branch of geology, has shown the enormous capabilities of metamorphism, in that a crystalline and more or less schistose product may result from the alteration of either sedimentary or eruptive rocks, the original form of which may be entirely unde- terminable if such rock cannot be traced continuously in the field to some less altered condition in which sufficient traces of its original character can be found to admit of its satisfactory determination. As a result of these investigations, so much discredit has been thrown upon the classification and subdivisions of Eastern crystalline rocks by Hunt and his school, which were based on petrological distinc- tions now shown to be unessential and local in their character, that, until the areas covered by them have been systematically and care- fully studied, the relative age of different parts of the series must remain a matter of doubt. In a few cases, fossil evidence has been found in the Appalachian areas to show that certain crystalline beds are altered sediments of Cambrian or later age. In others, remains of organic life have been found which are older than any known Cambrian forms. In most cases, however, it can only be deter- mined on stratigraphical grounds or lithological evidence that the rocks in question are older than any known Cambrian, and younger than the fundamental complex of non-clastic crystallines for which

Geological Distribution Of The Useful Metals. 67

the term Archsean is still retained. To these rock-series the general designation Algonkiaii has been given.

The Algonkian, as thus defined, necessarily includes a great many rock-series in different parts oF the continent, which, in the absence of palffiontological evidence, cannot be correlated in age, and whose relative succession must be determined in each geological province separately and by itself. They have been thus far systematically studied only in the Lake Superior region, where the new classifica- tion was first proposed by Irving and Van Hise. Here they con- sist of an aggregate thickness of over 60,000 feet of rocks in which three general subdivisions, separated by great unconformities or time-breaks, have thus far been recognized, the Keweenawan or copper-bearing series, which consists of sandstones, conglomerates, lavas and tufis, being the upper, and resting unconformably upon the two great iron-bearing series, the Upper and Lower Huronian, which include all the at present economically important iron de- posits of the region. Two, and possibly three, series of Algonkian rocks, each of great thickness, and some showing a large develop- ment of eruptive rocks, have been recognized in the Rocky moun- tains, but for the Appalachians, where geological study is rendered more difficult by the intense complications of structure, great meta- morphism, and deep covering of weathered material and soil, it can only be said as yet that certain rocks hitherto called Archaean are certainly either altered Palaeozoic or Algonkian, while it remains for further study to determine of the greater part whether they belong to either of these systems or may properly be classed as Arohaeau.

Iron.

Iron is not only economically the most important of the metallic products of our country, but it was among the first of the metals to be concentrated into workable deposits ; for the greatest bulk of our ores, probably not less than two-thirds of the total product, has been obtained from the most ancient geological formations.

Iron in the Older Crystalline Rocks,

Up to the close of the last decade nearly 100,000,000 tons* of iron-ore had been derived from deposits in comparatively limited areas of crystalline rocks, which were formerly all classed as Ar-

The long ton is used in this paper, as more nearly corresponding to the metric ton than the short ton of 2000 lbs.

58 Geological Distribution Op The Useful Metals.

chseao. Under the new classification most of these ore-deposits are known to occur in undoubted Algonkian rocks, and of the balance the age of the enclosing rocks whether Algonkian or ArchseaUy is as yet undetermined, so that it may be considered doubtful whether concentration of iron-ores had so far proceeded in pre-Algonkian time as to produce an economically valuable ore-deposit.

In the Lake Superior region, which leads the list (of producers), with a total product of 57,000,000 tons, and whose ores are mostly hematites, with some admixture of magnetites and limonites, the iron-producing horizons belong either to the Upper or Lower Hu- roniau, the former of which is separated from the latter by a dis- tinct unconformity, and in whose beds are found fragments of iron- ore presumably derived from the latter; whence it may be inferred that iron-ore deposits had been formed in this region during Algon- kian time. A small amount of limonites has been produced from Cretaceous strata of Wisconsin and Minnesota, along the borders of the Palseozoic continent, which are of mo:e geologic than economic importance. At the western end of Lake Superior, in connection with remarkable new finds of great bodies of iron-ore in the Mesabi range, occur a series of eruptive gabbros which contain a great deal of titaniferous iron, not, however, as far as known, sufficiently con- centrated to form valuable ore-deposits. The age of this series and its relations to the various sedimentary series are not yet definitely determined ; but although as yet of no economical importance, its geological bearing has considerable significance and suggestiveness.

The Adironddok or Lake Champlain region and the HighlandB of New Jersey are the next largest producing districts, the latter having yielded a little more, the former a little less than 17,000,000 tons.

In the Adirondacks is a central core of eruptive gabbro (norite of Hunt) rich in titaniferous iron surrounded by beddeil rocks of Algonkian age, in which occur most of the producing mines. The enclosing rocks consist of foliated gneisses, quartz! tes and coarsely crystalline limestones carrying graphite. The ores are mingled magnetites and hematites. Whether the ore-deposits of the interior region are to be considered of Archssan or Algonkian age has not yet been determined.

The ores of the New Jersey Highlands, which are almost exclu- sively magnetites, occur in rocks lithologically similar to those car- rying the main producing ore-deposits of the Adirondacks, in that they consist of foliated gneisses, quartzites and limestones, and which will probably I)e found to be Algonkian.

Geological Distribution Of The Useful Metals. 69

In Soviheastem Missouri the region around Pilot Knob, Iron and Shepherd Mountains, is the next limited area of pre-Cambrian rocks which has been an important factor in the iron industry of the country, its total product having been something over 5,000,000 tons. Here the ores were originally segregated in irregular vein- like masses in porphyries, which are associated with clastic rocks, and which are undoubtedly of pre-Cambrian age. Workable ore- deposits have been formed from the detritus of these veins in adjoin- ing pre-Silurian valleys which have since been covered by Silurian limestones; hence the formation of the ore-deposits dates certainly from the pre-Silurian and probably from pre-Cambrian time. The enclosing rocks have not yet been satisfactorily correlated with any particular series of the Lake Superior Algonkian, but there seems little doubt that they are properly to be considered as of Algonkian age. '

Of geological interest is the occurrence of magnetites associated with basic igneous rocks at Magnet Cove, in Arkansas.

Appalachian Areas. — In the large areas of crystalline rocks which lie to the eastward of known Paleeozoic rocks, and which stretch from Maine to Greorgia many deposits of magnetites and hematites have been found, and some worked to a limited extent. Those of North Carolina are at the present the only ones of considerable economic importance. Less is therefore known of their true geo- logical relations, but the little that is known leads to the reasonable supposition that many, and possibly the greater number, are of Al- gonkian rather than Archsean age.

Next in importance are the Virginia ores obtained east of the Blue Ridge, which are magnetites, hematites and limonites occurring in metamorphic schists and sandstones. It is probable that many of these deposits will prove to be in altered Palaeozoic beds, and it is quite impossible at present to say what part of their product is to be credited to Algonkian or older rocks. Many of the scattered deposits of magnetic ores of New England which have been some- what intermittently worked may prove to be Algonkian or older, but in this area, where most of the modern work on crystalline rocks in the east has been done, so much has been definitely proved to be altered Palaeozoic that geologists hesitate about expressing an opinion on the age of what remains to be studied. Of considerable geological suggestiveness is the fact observed in some localities, and which further study may discover to be yet more common, that the titaniferous magnetites in crystalline schists are associated with an-

80 Geological Dibtbibution Of The Useful Metaij3.

cient basic eruptives, gabbros and allied rocks in which magnesia or lime predominates over alamina in the silicates.

Western Arexu. — Except in Colorado, the iron industry west of tlie Mississippi river has not yet assumed sufficient commercial impor- tance to have brought about a thorough investigation of the resources in iron-ores. Many occurrences of what are apparently important deposits are known, but our knowledge of their geological relations or of their commercial value as compared with those of the east is very imperfect. Magnetic ores do occur in the crystalline rocks, but have not thus far proved of economic importance and so far as known no ores have yet been found in the few rock series that have been definitely recognized as of Algonkian age, except in Llano county, Texas. In the Rocky Mountains the ores of the three best known localities, viz. : On Horse Creek in Wyoming, at Caribou, Boulder county, and on 'Grape Creek, Custer county, in Colorado, are all titaniferous. It is significant that at the former locality is the only important body of norite or ancient magnesian-silicate rock thus far discovered in this region. At the other two localities it is only known in a general way that basic eruptives are found in the region. The Algonkian series thus far recognized in the west, with the possible exception of the Grand Caflon series, have not shown evidence of such great development of basic eruptives as those in the iron-bearing regions of the east. In the light of our present geo- logical knowledge it would appear that the west is not likely to de- velop any iron-bearing districts in the older rocks comparable in commercial importance with the eastern areas mentioned above. It must be remarked, however, that so much of this rion is as yet unknown geologically that future developments may very possibly furnish grounds for a modification or reversal of this judgment.

Iron in the Palceozoic Rooks.

The rocks of the Cambrian and Silurian formations east of the Mississippi river have yielded the greatest amount of iron*ore next to the Algonkian rocks. Within this geological range the Lower Silurian has been the most productive. Next to that comes the Clinton division of the Upper Silurian, while the Cambrian beds are supposed to have yielded less than the other divisions, but it is to be borne in mind that more accurate geological studies are con- stantly increasing the range of this horizon and may be expected to increase also the iron-ore product that belongs properly to it. A single mine, the Cornwall ore-banks in Pennsylvania, has produced

Geological Distribution Of The Useful Metals. 61

nearly eleven million tons of magnetite. No figures are avail- able whioh will give the total product of the many mines at these horizons scattered throughout the length of the Appalachians, mostly along the line of the great limestone valley. An estimate based on the pig-iron production would show that it has been between 16 and 20 millions of tons in the last decade. The ores occur in lime- stones and schists, and are generally limonites, supposed to have been hydrated and concentrated at or near the outcrop. The Corn- wall ores are magnetic, and occur in the LfOwer Silurian limestones, near the contact with Mesozoic sandstones, associated with a body or irregular dike of trap. Magnetic ores are found at many other points along the Mesozoic contact, but in less considerable bodies, and often without the trap. The Cornwall ore-bodies, and probably many others, evidently result from the oxidation of sulphides. Spathic ores occur at this horizon in New York in considerable bodies. Specular and limonite ores are also derived from the Cambro- Silurian limestones of southeastern Missouri.

The Devonian beds have yielded comparatively little irop, the principal production from this horizon coming from limestones and argillaceous sandstones in Pennsylvania, and from calcareous sand- stones at the Oriskany horizon in West Virginia.

From the Carboniferous limestones on the other hand a very con- siderable amount of carbonate and brown hematite ores has been derived, mainly from the regions west of the Appalachians. Some black band ores also come from these horizons. No statistics are available to show the relative amount of ore derived from this source, but in the year 1890 its tonnage was about one-seventh of that derived from Silurian beds.

West of the Mississippi river most of the important iron deposits now known can be assigned with more or less certainty to Palaeozoic limestones ; and in many cases it can be proved, in others assumed as probable, that they result from the alteration of pyritous ores, formed by i*eplacement of the limestone. Many of these ore-bodies are mingled magnetites and hematites. Others are limonites. No spathic ores have yet been developed as far as known. It is of in- terest to note that in Colorado the maguetic ores are generally asso- ciated with considerable eruptions of diorite.

Of the age of the ores found in PalsBOzoic rocks it can only be said with certainty that they have been formed later than the beds in which they were deposited, for the hypothesis that they were formed simultaneously with the enclosing beds by precipitation from sea-

62 Geological Distribution Op The Useful Metals.

waters is now abandoned by most g(H)logi8ts, and it is generally con- ceded that they result from the concentration by underground or surface-waters, as the case may be, of material originally dissemi- nated through the surrounding rocks. If, as there seems to be some reason for I>elieving, the limonites of the Lower Silurian in the Appalachians were formed by the leaching and concentration near the surface of the outcrops of the limestones, they were probably formed since the Appalachian uplift at the close of the Carboniferous.

Iron in the Meaozoic Rocks.

Although stratified rocks of Mesozoio age are often characterized by an abundance of disseminated iron minerals, the beds of these horizons have thus far developed no important workable deposits of iron-ore. In the east, limonites and clay ironstones have been mined to a limited extent in the Trias of North Carolina and in the Cre- taceous of Maryland. Workable deposits of limonite, as already mentioned, are found in the lower Cretaceous of Wisconsin and Min- nesota, which were originally sulphides. Clay ironstones also occur in the Laramie or coal-bearing strata of the up|)er Cretaceous in the Rocky Mountain region, but do not seem likely to prove of more than local economic importance. Iron occurs also in California, in upturned Mesozoic rocks associated with dioritic eruptives. The limited development of important iron deposits in rocks of this ho- rizon would seem to be due rather to the absence of considerable limestone beds than to their recent age. In Mexico, where there are considerable developments of limestones in the lower Cretaceous (which is wanting in the Rocky Mountain region), large deposits of hematite are found in these limestones, associated with diorites, which are necessarily Mesozoic, or later, in age. If all the diorite eruptions of Colorado prove to be Mesozoic, or later, as is already determined of the greater portion of these rocks, even the iron-ores in Palceozoic limestones which are associated with diorite, will prob- ably be found to have been formed in, or since, Mesozoic time.

Iron in Tertiary and Recent Deposits.

Although the concentration of ores in older rocks has, undoubt- edly, continued in many cases through Tertiary and recent times, no de|>osits of more than local economic importance are known in sedi- mentary rocks formed during these times. Bog-ores, formed in Quaternary or recent times by the leachiug and redeposition of older deposits, are common in all parts of the country where these older

Geological Distribution Of The Useful Metals. 63

deposits exist. In the States of Oregon and Washington they are often covered by, or included in, recent basalt flows. Basic eruptives of recent age may contain concentrations of iron minerals by differ- entiation, but, as yet, none have been proved to be of economic im- portance.

Oenesii of Iron Deposits.

It has always been a matter of wonder to the geologist, as well as to the layman, how such enormous concentrations of metallic min- erals as occur in the great iron mines could be brought about, and whence their materials could have been derived. In the light of the more exact studies of modern times, the easy reference of such knotty questions to the '' unknown source in depth/' is no longer available, especially since, in the case of the Lake Superior deposits, the last stronghold of the little band of geologists who still maintained the eruptive origin of iron-ores, the careful and systematic researches of Irving and Van Hise have demonstrated that the supposed eruptive bodies of iron oxide have been deposited from aqueous solutions as replacements of carbonates, and that the eruptive contact phenomena result from the fact that the enclosing rocks, instead of the iron-ores themselves, are of igneous origin.

That iron minerals, such as pyrite, magnetite, and ilmenite are frequent and almost universal constituents of eruptive rocks, is well known, but they occur as original constituents of the rock, that have formed within its mass more oi* less contemporaneously with the other mineral constituents, and not as later injections into an already consolidated rock mass ; whereas, critical studies of existing ore-deposits have so universally proved them to be of distinctly later origin than the enclosing rocks, that the burden of proof lies upon those who would maintain a contemporaneous origin for any partic- ular deposit. The truly scientific method in the study of such ques- tions, at the present day, is the reverse of that which was followed in the early days of geology, when, after the observation of a few isolated facts, some great geological mind was led to a general theory, and humbler followers were only too apt to do mild violence to na- ture in order to make her facts conform to it. It accumulates, year after year, a multitude of facts of patient observation supported by studies with the microscoi)e and in the laboratory, avoiding general theories, and only making such deductions in regard to local con- ditions as are supported by the overwhelming evidence of facts.

Although we are yet far from having a sufficient accumulation of facts bearing upon the origin of iron-ores to justify the putting forth

64 Gbological Distribuiton Op The Useful Metals.

of any geneml theory, it may be allowable in the present case to indicate the lines of research to which the facts that have lately been accumulated seem to point as promising the most remunerative results.

A great deal of light has been thrown upon the manner of forma- tion of iron-ore deposits by the researches of Irving and Van Hise in the Lake Superior region, and by the discussion of replacement of limestones by iron-ores in general by J. P. Kimball. By both the process of formation of workable iron-ore deposits is rarded as a concentration by the agency of percolating waters, such concentra- tion being influenced by physical or structural conditions, and local- ized, it may be, by a pre-existing nucleus of iron-bearing minerals as original constituents of the rock. The deposition is considered to be in very large degree a metasomatic replacement of the rock mate- rial, and only to a very limited extent a deposition in pre-existing open cavities.

For the Marquette region it is found that though the LfOwer Huronian carried iron originally, the concentration into workable de- ]K)sits, of both this and the Upper Huronian series, was brought about subsequent to folding and erosion, and that hence the age of the deposits as such is Upper Huronian or later. Evidence is also found that the deposition was a secondary concentration from waters percolating downwards along the paths of great water channels until stopped by some impervious base. The original condition of the ore is regarded as probably iron carbonate, though it is ad- mitted that this may have been a replacement of calcium carbonate.

In the Palseozoic limestones and shales of the Appalachians, the iron-bearing solutions appear in most cases to have been also down- ward-going currents, or water sinking from the surface under the in- fluence of gravity rather than hot ascending solutions. The original mineral was the carbonate or the sulphide of iron (pyrrhotite or py- rite), and instances are adduced where the limestones carry in their mass over 2 per cent, of iron carbonate, and in other cases pyrite is known to occur in about the same proportion. Whether these minerals were chemically or mechanically deposited with the limestones or were introduced subsequently remains to be determined, but it appears im- probable that deposits were formed simultaneously with the enclos- ing rocks by chemical precipitation from sea waters, of sufficient size to constitute workable deposits.

If it be admitted, then, that our workable deposits of iron-ore are mainly concentrations of iron minerals already disseminated in sedi-

Geological Bistbibution Of The Useful Metals. 6&

mentary beds, and that these coDoentrations have occurred in different forms and places according to varying local structural or chemical conditions, it still remains to be determined what was the original sonrce of the iron in different regions, and why the concentrations are so much greater in one place than in another.

A line of investigation that seems to promise interesting results is suggested by recent researches by Swedish geologists on the forma-) tion of concentrations of titaniferous ore by the so-called differentia- tion process in basic eruptive magmas. In Sweden and Norway according to them, actually workable deposits of titaniferous iron have been formed by differentiation within the eruptive magmas of> labradorite, hypersthene, or olivine rocks. Van Hise had already sug- gested for the titaniferous magnetites of the eruptive gabbro of Lake Superior that in the crystallization of these rocks, before the magma had solidified, magnetite, which is one of the early minerals to sepa-' rate, had slowly settled to the base of the mass by virtue of its supe- rior specific gravity. But it is still questionable whether in this dif- ferentiation process gravity is a controlling influence, since in most observed cases it is evident that some other force must have influ- enced the concentration. Metallic concentrations in eruptive rocks have been observed before, the most remarkable of which is the liody of metallic iron at Ovifak in Greenland. Although in these cases the ores may properly be said to be of eruptive origin, it may still be doubted whether their concentration as workable deposits is not due, in a measure at least, to secondary action, as has been observed in the case of the Lake Superior gabbros.

While, therefore, one may not necessarily expect to find economic cally valuable deposits in such rocks, the question naturally suggests itself whether the occurrence of large areas of older basic eruptives,: which in some parts contain a relatively large proportion of iron- bearing minerals, may not fairly l>e considered to be an indication that neighboring sedimentary beds may contain large concentrations of iron-ores, which have been derived from them. Where such basic eruptions are older than the beds, this derivation would be mainly mechanical, the ores being sediments resulting from the abra- sion of the eruptives, more or less concentrated according to varying- conditions of sedimentation. Where the eruptives are younger and have broken through or overflowed the sedimentary beds, the deri- vation would be mainly chemical, through leaching out and redepo- sition by the agency of percolating waters.

While there seems to be some genetic connection between the

VOL. XXn.— 6

66 Geological Distribution Of The Useful Metals.

greatest concentrations of iron-ore and considerable developments of ancient basic emptives, important deposits also occur where no such relation can be traced. The frequent association of iron-deposits in the west with large bodies of eruptive diorite, suggests that though the very basic rocks would naturally afford the greatest amount of iron, even a relatively acid rock, like diorite, may contain concentra* tions by differentiation, which have yielded to the action of perco- lating waters and thus allowed their basic constituents to be trans- ferred to easily soluble rocks like limestone.

Since water is the principal agent in the final concentration of ore- deposits, it is important in searching for them to study the physical conditions which will favor its ready action both in taking up and in throwing down. In the northwest. Van Hise has found an im- pervious basement a general favoring condition. This assumes down- ward percolation, but within the crust of the earth the circulation of waters may be in any direction according to local conditions. It may possibly be safe to assume that iron-ores which occur mainly as oxides would have been deposited from oxidizing waters or those which come recently from the surface, and that pyritous ores would be more likely to bespeak a derivation from subterranean waters. Types of the former are furnished by the Lower Silurian ores which pass in depth into ferriferous limestone, and are apparently a concentration due to leaching by surface waters, in which other minerals have been removed in greater proportion than the iron oxide. The so-called gossan ores, occurring in the eastern or metamorphic belt of the Appalachians and which pass into pyritous ores in depth, would appear to be good types of the latter class.

Much remains yet to be done in the study of the structural rela- tions of iron-ore deposits. One of the interesting problems is fur- nished by the line of magnetic deposits occurring in Pennsylvania and southwards in limestones at the contact of Mesozoic sandstones, and frequently associated with trap-dikes. If this prove to be a line of displacement, as there seems to be reason to assume, it would af- ford a natural water-channel for the collection of iron-bearing waters from various series of iron-bearing rocks, which would preferably collect in limestones, and more readily from their broken edges. Whether the function of the trap has been to furnish heat for mag- netization of the iron oxides, as has been frequently assumed, or to interrupt the ore-bearing currents and thereby induce precipitation, may well be the subject of further investigation.

There is .reason to assume that the concentration of iron-ores in

Ic

GEOiiOOiCAL DnrrRiBaTroN op the useful metals. 67

the more southern parts of the Appalachians, especially in the ex- tremely complicated regions of the Carolinas, Tennessee, Alabama and Georgia, will be found to have more or less intimate structural relations with the many fault-zones that abound there.

A careful study of the magnetite deposits in Colorado is likely to throw some light upon the true genetic connection, if any exists, between the occurrence of this oxide and the vicinity of eruptive bodies. In the light of present knowledge it would appear that the iron-ores occur as magnetite in the vicinity of large bodies of erup- tive diorite and as limonite elsewhere at the same horizon. It has been asserted, moreover, that there is no evidence of an intermediate hematite or limonite stage in the alteration from pyrite to magnetite. As against the theory that in these cases the formation of the mag- netic oxide is due to the heat of the eruptive body, it is probable, in the opinion of the writer, that these ores have been concentrated since the eruption of the diorite.

Manganese.

Manganese in nature is so intimately associated with iron that it is rare to find an ore of this metal that does not contain the other in greater or less proportion. The origin and geological range of the two metals would therefore naturally be expected to be similar, and it will suffice for the purpose of this brief article to point out some of the contrasts in their behavior and of their mineral compounds.

Manganese is almost as widely disseminated in nature as iron, but occurs in very much smaller quantities. As an original constituent of crystalline or eruptive rocks, it generally occurs as one of the. bases of compound silicates.

In derived or sedimentary rocks it occurs almost invariably in one of its many oxides, though the carbonate is found, to a limited ex- tent, in some deposits. The fact that the salts of iron and manga- nese are isomorphous would suggest that the original form of depo- sition of the two minerals would be similar, but workable deposits of manganese in this country have not yet been explored beyond the range of the oxidizing influence of surface-waters, so that it is not yet demonstrated that the oxide in nature is generally derived from the carbonate of manganese in calcareous rocks. The rare occur- rence of the sulphide might be accounted for by the great instability of the known mineral compounds of manganese and sulphur.

The geological distribution of ores which are worked for their con- tents in manganese alone is extremely limited. In the decade 1880-

68 Geological Distribution Of The Useful Metals.

1890 the total product for the United States is given at 192,816 tons, of which 180,144, or 93 per cent., were derived from the Cambro- Silurian horizons of Arkansas, Virginia, and Greorgia. The Appa- lachian deposits are generally assigned to the Cambrian sandstones ; those of Arkansas occur in what is supposed to be the upper part of the Lower Silurian. Both are described as residual-claj deposits; that is, thej result from the decay, leaching, and concentration by surface waters of rocks containing manganese originally disseminated through them either as carbonate or oxide. In Arkansas these original deposits were limestones ; in Virginia and Greorgia they are supposed to have been shales in the Cambrian sandstones at their contact with overlying Silurian limestones, and are more or less in- timately associated with the lower series of Silurian iron-ores.

In Colorado, manganiferous iron-ores, of economic importance in making spileisen, are derived from the Lower Carboniferous lime- stone of Leadville, and result from the surface oxidation of pyritous silver- and lead-ores. It is rather remarkable, however, that no manganese-bearing minerals of any importance have yet been dis- covered in the unaltered sulphides. Deposits of manganese asso- ciated with iron-ores, which may be of economic importance, occur in limestones in Colorado which are probably of similar horizon. In California, manganese-ores have been derived from highly-altered Cretaceous rocks.

In the earlier, or pre-Cambrian rocks, manganese forms locally an economically-important constituent of iron-ores ; the most notable occurrence is in the Gogebic district of the Lake Superior region, where the Colby mine produces some pure manganese-ore. Of geo- logical interest are local occurrences of manganese-bearing minerals in the crystalline areas of the Appalachians, of which the iron-zinc ores of New Jersey may be mentioned as of some economic impor- tance as producers of manganese. The central Texas area is most suggestive in that siliceous oxides and silicates of manganese are in- timately associated with manganiferous varieties of garnet in foliated gneisses and quartzites of probable Algonkiau age. Of the ancient basic eruptives it can only be said, as yet, that manganese occurs at times as an important constituent (2 per cent, or over), but generally in subordinate quantities and in far smaller proportion than iron. It is also a frequent constituent of later igneous rocks, and in the veins of Butte, which are fractures in a rather basic granite, it has been secreted in notable quantities as silicate and carbonate in associa- tion with sulphides of the other metals.

Geological Distribution Op The Useful Metals. 69

In Tertiary and recent formations manganese occurs in analogous positions to bog-iron in the form of wad. Manganese concretions form also a notable portion of the material found on the ocean floor at abyssal depths.

Genesis. — The same lines of genesis suggest themselves for manga- nese as for iron, but owing to its much smaller percentage as a con- stituent of original rocks, it will be less easy to detect the probable localities favorable for secondary concentration in workable bodies.

In this secondary concentration the study of structural conditions which would produce natural water-channels is equally important. It has already been observed that the manganese-ores of the southern Appalachians occur along a great faulted zone, and the frequent men- tion of breccia conditions in other deposits suggest that further study may show that the concentrations of this mineral, as well as of iron, have been along such lines of displacement more frequently than has been hitherto realised.

In the relative chemical behavior of the salts of manganese and iron in terrestrial economy there are certain unexplained contrasts which would appear to offer remunerative results to those who would occupy themselves with its study.

Nickel.

Workable deposits of nickel-ore are of extremely rare occurrence in this country and, geologically, its minerals are not of very fre- quent occurrence. The total production of the metal in the last de- cade is given as less thau 1000 tons (2,049,676 pounds).

But one mine, that at Lancaster Gkip, Pa., has been profitably worked for the metal on a large scale. It has produced 4,000,000 pounds, and at one time yielded one-sixth of the world's product. Besides this, only a few deposits of possible economic value are yet known within the United States. With regard to all of these, the published geological information is extremely meagre, and presents little that is of final value.

Most of these deposits in the East occur in crystalline rocks, which may prove to be Algonkian or altered Palaeozoic. The Gkip mine in Pennsylvania produced a sulphide-ore associated with pyrrhotite in a hornblende rock (which may prove to be an altered eruptive) enclosed in mica schist. There is a suggestive occurrence, in the belt in which this mine is situated, of chrome-ores with magnetite, associated with altered magnesian silicates ; a somewhat similar series of occurrences seems to run through New England, but the

70 Geological Distribution Of The Useful Metals.

data are too indefinite to suggest a necessary geological correspond- ence.

Near Webster, in North Carolina, an area of olivine rock more or less altered to serpentine, contains a considerable concentration of nickel and chrome minerals, which are supposed to have been an original constituent of the (eruptive?) rock. The nickel here occurs as generally in such associations, in the form of a nickeliferous mag- nesian silicate. The ores developed near Biddies, Oregon, are of similar mineralogical character, and are associated with serpentine. Both occurrences present remarkable points of resemblance with the famous deposits of New Caledonia. The ores of Churchill county, Nevada, are also silicates, and are apparently connected with an altered basic eruptive.

In distinctly sedimentary rocks the principal occurrence is at Mine La Motte in southeastern Missouri where it occurs in veins in Cambrian limestones associated with other metallic sulphides. The mineral millerite is also found at various points in the mineralised limestones of the Mississippi valley.

It may be well to mention here the nickel-deposits of Sudbury, Canada, which are the most important producers of the metal on the continent at the present day, and hence better known geologically. Here the metal occurs as a sulphide, nickeliferous pyrrhotite, asso- ciated with chalcopyrite. An interesting point is the association of the deposits with a greenstone or trap (diabase?) containing magnetic and titaniferous iron, copper and nickel. The region is described as much faulted and dislocated, and traversed by dikes. The ore-bodies usually occur at the contact of the schists and greenstone, and char- acteristically carry breccia fragments of the country rock of which the ore fills the interstices. No mention is made of the occurrence of serpentine.

In spite of the fact that the geologist, from whose description the above facts are taken, considers the ore to have been formed not from solution but by secretion from a fused magma, they seem to me to afford convincing evidence that it was formed by concentration through percolating waters of the material originally disseminated through the rock in water channels formed by fault-planes or zones of displacement.

Genesis of Nickel-Deposita.

The frequent connection in nature of nickel and magnetic pyrites, and of nickel with native iron and magnesia, in meteorites and at

Geological Di8Tbibution Of The Useful Metalb. 71

Ovifak in Greenland, is suggestive of an intimate connection be- tween the two metals in fused magmas. The frequent occurrence of its silicate ores in connection with serpentine and associated with chrome and magnetic iron has often been remarked bj geologists and chemists as pointing to a genetic connection between these minerals and magnesian silicate rocks. It is to be noted, however, that both the silicate of nickel and serpentine are secondary products. Ser- pentine is known to result from the metamorphism of many rocks, both eruptive and sedimentary, most commonly from basic magne- sian silicate rocks in the first case, and from calcareous sedimentary rocks. The silicates of nickel may well be assumed to have resulted from the secondary alteration of sulphides, if the assumption is cor- rect, that in those cases where it so occurs in association with magnetic pyrites, the neighboring basic eruptives have not yet reached the extreme of serpentinous alteration. As with iron, therefore, certain portions of basic eruptive magmas may be supposed to have been relatively rich in nickel-bearing minerals, and by secondary concen- tration these may have beeu transferred to the water-channels of adjoining rocks. The greater this alteration of the rock the greater concentration of nickel-ore, as a general rule, would one expect to find.

Tin.

Tin seems to be the rarest of the useful minerak in this country. The known deposits of its ores of supposed economic importance occur in crystalline rock series associated with granites. The en- closing rocks in the Black Hills have been determined to be of Al- gonkian age. The age of those of Virginia and California has not been determined, but may be provisorily assigned to the same gen- eral series. Its mineralogical occurrence is also, so far as known, confined to the areas of crystalline rocks. Its ores here, as in the older countries, have a mineral and rock association peculiar to themselves, consisting in the former respect of combinations with borax, fluorine, tungsten, niobium, tantalum and other rare metals, and in rocks, of granites and crystalline schists carrying lithia micas. In the Black Hills cassiterite is associated with notable quantities of colnmbite and tantalite, minerals so closely resembling it in their physical qualities, that they have been frequently confounded.

Tin occurs almost invariably as oxide, the sulphide being only of mineralogical importance; and it is generally in crystalline form, and more or less free from the ordinary metals. These character- istics and associations are so universal that it probably is little

72 GEOLOGICAL DISTBIBUTION OF THE USEFUL METALfl.

looked for in other associations. Yet it appears that in Bolivia, where tin-ores form an important part of the mineral product it occurs in andesitic or trachytic rocks of Cretaoeons or Tertiary age, and in association with sulphides of silver, copper, lead, zinc and iron, and without the usual accompaniment of tourmaline, topaz, fluorspar or apatite. The sulphide is, moreover, of not uncommon occurrence, though it does not appear to have been necessarily de- posited in this form, since the oxide is sometimes enclosed by sul- phides of the other metals.

Although tin-deposits in the crystalline rocks of the east have not yet proved of economic importance, and no considerable product has yet been derived from them in the west, it is as yet too early to des- pair of finding paying deposits of tin in this country, since the ex- tensive regions of andesitic eruptions are as likely to prove tin-bear- ing in the northern as in the southern hemisphere. Tin minerals are the extreme opposite of those of copper, in that they are ex- tremely difficult to detect by the naked eye.

Owing to its high specific gravity and hardness tin oxide is, like gold, often found in river gravels, and the infrequency of tin-placers in this country is in so far an unfavorable sign, but here again its resemblance to magnetic oxide of iron, and the want of familiarity with its physical characteristics would account for its escaping the notice of the average prospector.

Coppeb.

The total production of metallic copper in the United States from 1845 to 1890 is estimated at 1,056,436 tons. Of this the Lake Su- perior region has contributed 604,829 tons or 57 per cent. Within the last decade Montana has become the next important producer, yielding 36 per cent, to Lake Superior's 45 per cent, of a total pro- duction of 731,889 tons, while for the year 1890 it produced 43 per cent, as against Lake Superior's 38 per cent, of a total produc- tion of 115,668 tons. Next in importance is Arizona, which yielded 15 per cent, of the total production of the decade. These three regions together are now furnishing 95 to 96 per cent, of the total product of the country; the other 4 or 5 per cent, of the decade's production, being largely derived from ores sold in the open market, cannot be accurately sregated. Their ores come from the follow- ing States or groups of States which are given in the order of their relative ihiportance: Utah and Colorado, New Mexico, California, the New England States, Wyoming, the Southern States, Nevada

Geological Distribution Op The Useful Metals. 73

Idaho, the Middle States ; over three-fourths of the total coming from the four first named.

In the Lake Superior region a single mine, the Calumet and Hecla, produces 50 to 60 per cent, of the total product.

In considering the geological distribution of workable deposits of copper, it is to be remarked that the older rock series have produced the greater amount of the metal, and, while the age of the enclosing rocks is not necessarily a measure of the age of the ore-deposits en- closed, since the process of deposition is a long continued one and may be going on at the present day, it will be seen that there is reason to assume that the greater concentration in these rocks has a probable dependence on the length of time during which processes of rock alteration and mineral concentration may have acted.

Copper in the Algonkian Rocks.

Lake Superior District. — The copper of this region occurs almost exclusively in the native state. It is obtained from upturned rocks of Algonkian age, known as the Keweenawan series, which were . deposited unconformably on the iron-bearing Huronian series, and are in turn overlapped by nearly horizontal Cambrian sandstones. The Keweenawan series consists of eruptive sheets and of sandstones and conglomerates largely made up of a detrital eruptive material ; the eruptives are mostly basic rocks which were poured out as lavas upon the surface, but there are some eruptions of the acid type. The lavas and sedimentary beds are interstratified, the former pre- dominating near the base, the latter at the top of the series, which is estimated to reach a maximum thickness of 46,000 feet. These rocks have experienced intense and long continued metamorphic action or alteration, which has produced amygdaloidal structure and in certain parts has resulted in an entire change in the mineral com- position of the rock. The copper occurs in regions of intense altera- tion, in small veins traversing the beds at right angles to strike and dip, in the amygdules and in the interstices of the conglomerates. The ore of the Calumet and Hecla and adjoining mines is obtained from a thin bed of conglomerate in the lower part of the series. It has been demonstrated that the copper had been deposited from wet solution as a stage in the process of rock alteration, and as a pseudo- morphous replacement of certain minerals in the rocks. It is assumed that these deposits are a concentration of copper-salts once minutely disseminated throughout the entire great series of rocks, and it is probable that the original form of the mineral in these rocks was

74 Oeolooioal Distbibution Of The Useful Hetau.

sulphkle. Minute specks of copper minerals are found in the suc- ceeding sandstones, but as yet no concentrations sufficient to form workable deposits. The deposits are remarkable for the compara- tive absence of other metallic minerals except silver, which also occurs, to a limited extent, in the native state.

Montana, — The Montana ores are almost exclusively derived from the mining district of Butte. Here they occur in a series of strong fissure-veins in a basic granite, which in places is almost a diorite in composition. A later eruption of rhyolite cutting through the granite appears to have some relation to the ore-deposition. A considerable body of porphyrite occurs in the granite a few miles to the east of the mines. The geological relations of this granite, are not yet deter- mined, and its age is, therefore, as yet unknown, but it is probably pre-Cambrian. The copper-bearing ores are mainly chalcocite and chalcopyrite. The veins are a system of strong more or less parallel fractures which have afforded channels for the copper-bearing solu- tions which have deposited the ores in the interstices and as a replace- ment of the minerals composing the granite. In most of the veins but little copper was found within 200 or 300 feet of the surface, the values being mainly in silver ; silver continues with the cop|>er in depth but in smaller proportion. No native copper has been found, except slight traces in the cropping of veins where the upper part had been eroded down to the copper-bearing level.

Colorado. — Frequent copper stains are often found on the rocks of a newly discovered area of Algonkian schists which extends along the east side of the Arkansas Valley near Salida and southward into the Sangre de Cristo range. This series of banded or stratified rocks is largely made up of very much altered eruptives. At one point actinolitic schists are so strongly impregnated with carbonates and oxides of copper that they have been exploited for their copper values.

Copper minerals are found at various points in deposits in the older crystalline rocks of the Rocky Mountain system associated with other metals, but the copper values are obtained for the most part only as a by-product.

Arizona. — Copper-ores occur in the central part of the Territory in the crystalline schists, which may correspond to the Algonkian series exposed in the Grand Cafion 100 miles north of Prescott, and in which observers have traced a general lithological resemblance to the Keweenawan series of Lake Superior. The principal mine thus far has been the Verde, just north of Prescott, where the ores are

GEOLOGICAL DISTBIBUnON OF THE USEFUL METALS. 76

massive carbonates near the surface which pass into oxysulphides in depth. The ores contain gold and silver irregularly distributed throughout them.

Appalachians. — From Vermont to Georgia large masses of pyrite or pyrrhotite occur in the crystalline schists, which all carry more or less copper, and at times enough to be profitably worked for this metal. At different periods mines have been actively worked in these deposits in Maine, New Hampshire, Vermont, North Carolina, Tennessee, Georgia and Alabama. The enclosing rocks have been generally classed as Archsean (Huronian), and the deposits in some cases considered lenticular interstratified deposits. Some observers consider them fissure-veins formed on fracture-planes. There is a general succession in the southern States, where products of decom- position have not been removed by glacial denudation, of limonite at the surface, succeeded by copper oxides or carbonates, all passing into unaltered pyritous ores in depth. Of geological interest are the deposits of native copper in the South Mountain of Maryland and Pennsylvania, which occur in rocks recently determined to be largely eruptives and of Algonkian age, and which present points of resem- blance with the Keweenawan of Lake Superior.

Copper in the Paheozoio Rocks,

The Palaeozoic rocks in the East, with the exception of the Cam- brian limestones in southeastern Missouri and the lower horizons of the upper Mississippi lead region, do not seem to have developed any workable bodies of copper-ore. Some of the deposits in the Appalachians may be found to occur in altered Palozoio rooks, but even should this be the case, it can scarcely be hoped to increase their econouic value.

In the West, on the other hand, the Palaeozoic limestones have been the most important producers of copper next after the granites of Butte. The deposits, as far as known, are always associated more or less intimately with eruptive rocks. The main occurrences of oopper-ore are widely scattered throughout the Cordilleran region, and generally associated with other metals in such proportion that the copper is extracted only as a by-product. With regard to these deposits, and even of those where copper forms the principal value, geological data are as yet extremely meagre.

Arizona, — The greater part of the product of Arizona comes from the three districts of Clifton, Bisbee and Globe, in the southern and eastern part of the Territory ; other mines occur around this region

76 Geolooioal Digptribution Op The Useful Metals.

in Arizona and adjoining portions of New Mexico in which similar geological conditions appear to exist The ore occurs mainly in the Lower Carboniferous limestones and, to a less extent, in adjoining bodies of eruptive rocks, described as felsites, diorites and granites. The ores thus far produced have been mainly more or less oxidized, a certain amount of native copper having been found at the outcrops. They carry relatively little silver. As a rule, the ores occurring in the limestones are the most deeply oxidized, those in the eruptives passing from carbonates to sulphides at comparatively shallow depths, while in the limestones oxides and oxysulphides are found at depths of several hundred feet. The ore follows natural water-channels, either contact-planes of eruptive and sedimentary rocks or fractures in the latter, passing also on to stratification-planes in the limestone. Its age can only be said with certainty to be later than the eruptive bodies, which probably do not date from earlier than Mesozoic time.

Colorado. — In Colorado the principal copper values have been derived from mixtures of chalcopyrite and pyrite in the Silurian limestone of Leadville. The region has been the scene of intense eruptive action in Mesozoic time, the igneous rocks, generally of the acid type, being intruded in vast numbers of sheets and dikes be- tween and through the sedimentary beds. The ores of the region are mostly argentiferous sulphides of lead, zinc and iron, replacing the Lower Carboniferous limestones, which were carried by solutions following contact- and stratification-planes, and fracture-planes cut- ting across the bedding. Where the fracture-planes have been par- ticularly strong, in passing downward into the more or less siliceous limestones of the Silurian, lead and zinc have partially given place to copper in the pyritous deposits. In other portions of this region, however, copper-ores are found concentrated in the Carboniferous limestones also.

Utah. — In Utah the principal product comes from the mines of the Tintic district, where the ores occur as the replacement of Palseo- zoic limestones, the values being about equal in gold, silver and copper.

Copper in (he Mesozoio Rooks.

The importance of the Trias in Europe as a producer of copper- ores has led to the expectation of finding them also at this horizon in America, and some ore is said to have been extracted from these beds in Connecticut and New Jersey as far back as colonial times, but they have not been worked of late years. They are generally associated with the bodies of trap or diabase which are so frequent

Oeoloqical Distribution Op The Useful Metalb. 77

in this region. Disseminated oopper-ores are known to occur in the Triassic of Texas, New Mexico and Colorado, but so far tliey have not proved to be in sufficient quantities or sufficiently concentrated to be of much economic value. Copper minerals also occur in asso- ciation with other ore in the Cretaceous of the Rocky Mountain region generally in regions where there has been eruptive activity. In California considerable copper has been obtained from pyritous ores which occur exclusively at the contact of diabase and the up- turned Cretaceous slates along the foot-hills of the Sierra Nevada.

Copper in Tertiary and Recent Rooks.

No workable deposits of copper are known to the writer in sedi- mentary rocks deposited in Tertiary or recent times though certain occurrences in Arizona which seem, from the published descriptions, to result from the surface leaching of deposits in older rocks, seem to present a certain analogy with recent deposits of bog-iron and wad.

In igneous rocks which have been erupted in Cretaceous or post- Cretaceous time ores occur locally which yield values in copper in association with other metals, notably in the San Juan region of Colorado. The water-channels in which these ores have bn de- posited were probably formed during the post-Cretaoeous movement, and the ores are hence of Tertiary age.

Oenem of Copper-Deposits.

The observed facts with rrd to copper-deposits seem to point to eruptive rocks as the original source of the metal, and to indicate that its original form in deep-seated concentrations or deposits was that of sulphide. There seems to be less ground for supposing it to have been generally disseminated in marine sediments than in the case of some of the other metals, though very strong arguments have been advanced by geologists of great ability in favor of the theory of its chemical precipitation from the waters of the Triassic ocean by the agency of decaying organic remains. Its concentration, either in sediments generally or in ore-deposits, seems to have been by chemical rather than by mechanical processes. The assumption that certain portions of eruptive magmas are exceptionally rich in this and associated metals furnishes a good working basis for explaining its concentration in most well-known ore-deposits. Its chemical behavior, especially in deposition, presents some peculiarities not always easy to explain. In its ready assumption of the metallic

78 Geological Distribution Of The Useful Metals.

state, as in certain other actions, it resembles gold and tilver. Pom- pelly explains the absence of the baser metals in the Lake Superior deposits on the assumption that the copper has been reduced from its salts by protoxide of iron which would not have acted on the salts of the baser metals, which would have been carried farther on ; and that once reduced to a metallic state, the copper was in a condition of greatest permanence in presence of the usual reagents. To ac- count for the unusual amount of metal in this rion, there is an ex- traordinary amount of eruptive material to draw from, and unusually intense and long-continued action of metamorphic or alterative pro- cesses to produce the concentration. On Keweenaw Point, traces of sulphur are found in the melaphyr ; and in two mines, copper occurs as sulphide associated with other metals. Other exposures of this same series of rocks, in Wisconsin and (innesota, where no ore- deposits have yet been found, are said to carry small quantities of metallic copper, associated with sulphides of iron and copper.

It is worthy of remark, that the native silver which occurs with the copper in this region is never alloyed, but separates from it by rolling.

The unusual richness in copper of the ores along the limits of the zone of oxidation in veins at Butte and in the Appalachians is readily explainable by the leaching down of this metal and the removal of the less permanent salts of the baser metals. It is more difficult to account for the frequent sudden appearance and disappearance of copper at diflTerent parts of an unaltered deposit of mixed ores, as at Leadville and other places. There does seem to be a more frequent association of deposits of gold and copper with relatively acid rocks as of iron, chrome and nickel, with basic magnesian rocks, but ex- ceptions are so freiuent and data so incomplete, that it is questionable whether this association can properly be assumed to have a genetic cause. The pebbles in the copper-bearing conglomerate of Lake Superior, for instance, are said to be mostly of acid eruptives ; but this may result from the superior hardness of these rocks over the altered basic rocks. In Leadville, the ores carrying copper in the most important deposits are in a more siliceous limestone than in those which contain no copper, but other copper-deposits in the same re- gion are in the more pure magnesian limestone, and both here and in Arieona more copper is found in the relatively basic limestones than in the adjoining acid eruptives.

The presence of copper in nature is so easily detected on account of the bright colors of its surface alteration products, that it may be

GEOLOGICAL DISTRIBUllOK OF THE USEFUL METAUS. 79

assamed to have been so thoroughly prospected that no important sources of the ore remain undiscovered. It seems probable how- ever, that the belt of pyritous ores with limonite caps which stretches through the crystalline zone of the Appalachians, and contains gen- erally small percentages of copper, may yet prove a source of this metal of commercial importance in connection with other products of these great deposits.

Lead akd Zinc.

Lead and zinc are so closely associated in nature that their geo- logical distribution may well be considered under one head, with mere allusion to the points of difference as they occur. The original condition of both metals in ore-deposits is the sulphide. The com- mercially valuable oxidation products are, in the case of lead, the carbonate and sulphate, with a very limited amount of phosphate, arsenate and oxide; for zinc, the carbonate and silicate, and in one region the oxide. Lead sulphide contains more or less silver, and is frequently mined primarily for its silver values. Zinc sulphide generally contains more or less iron chemically combined, and is sometimes argentiferous.

The total product of the country in metallic lead, from 1825 to 1890, is given as 2,521,028 tons. For that of zinc in the same period no figures have been found, but the product of the last decade is 402,053 tons against 1,325,755 tons of lead, or less than Up to 1870 the greater part of the product of both metals was derived from the Mississippi valley deposits. Since that time the develop- ment of argentiferous lead-ores in the West has gradually reduced the proportion of lead coming from this source, and in the last de- cade it has been only about 11 per cent, of the total product. Zinc in the silver-bearing ores has, until very recently, not been utilized, and its main source still continues to be the Mississippi valley lime- stones. The States of Wisconsin, Missouri and Kansas furnished 83 per cent, of the total metallic product for the decade. It is difiS- cult to arrive at the relative proportion in which the two metals exist in nature, as their relative product, even if accurately known, would not furnish a fair criterion, for the reason that, owing to the peculiar conditions attending the reduction of the ores to the metallic state, the zinc is, in large measure, lost or rejected in treating mixed ores and the lead as generally saved. Chamberlin estimates that the upper Mississippi region has produced, up to 1877, 400,000,000 pounds of zinc against 320,000,000 pounds of lead.

80 Oeolooical Distribution Op The Useful Mbtaub.

Lead and Zinc in the Older Crystalline Rocks.

Throughout the crystalline belt of the Appalachians, sulphides of lead and zinc occur in veins in association with pyritons ores, but, so far as known, not in sufficient concentration to constitute workable deposits. They are also known to occur in similar manner in the Algonkian series of Lake Superior. In either case, owing to their want of commercial value, so little is known of their general distri- bution that it is not possible to outline the regions of greatest preva- lence, and it can only be said that they are probably quite unequally distributed.

The franklinite deposits of northern New Jersey are the only de- posits in the older rocks which have been an important source of zinc commercially. They occur in a marbleized limestone traversed by dikes of granite, which is probably of Algonkian age, and upon which the Lower Silurian limestone rests unconformably. Although implicated with, it is likely to prove distinct from, the foliated gneisses in which the magnetite deposits of this rion are found. The mineralogieal occurrence is unique, franklinite being composed of protoxides and sesquioxides of zinc, iron, and manganese, with which are associated the anhydrous silicate and red oxide of zinc, and some magnetite. The locality is near the western limit of the areas of exposed crystalline 8chi.ts.

In the western areas of granites and crystalline schists, sulphides of lead and zinc are common associates of the ores of other metals, especially in silver-bearing deposits, and are frequently accompanied by igneous rocks in dike form. It is seldom, however, that they are sufficiently concentrated to be worked for these metals alone. In such deposits the oxidized forms, especially of lead-ores, are in very limited amount. One notable exception is in the Terrible mine in Colorado, which is a large body of remarkably pure carbonate of lead which replaces a porphyry dike in granite.

Of the age of all these deposits it can only be said in most cases that they are concentrations that cannot have commenced before Algonkian time, and may properly belong to a very much later geological period.

Lead and Zinc in the Palceozoic Bocks.

It is from the limestones of Palseozoic age that the bulk of the

lead 2ind zinc product of the country has been derived. In these

rocks not only are the concentrations of their minerals greater than

in the crystalline rocks, but they are more distinctly segregated from

OEOIiOQICAL DISTRIBUTION OP THE USEFUL METALS. 8L

the sulphides of other metals and from each other, sometimes form* log considerable accumulations of the sulphides of either metal alone or with the omnipresent pyrite. Alteration of the sulphides to the various oxidized products has also proceeded to a greater distance from the surface in the limestones than in other rocks.

In the Appalachians, although galena is observed in minute con- centrations disseminated through the limestones of various horizons, the workable deposits of either metal seem thus far to be mainly confined to the Lower Silurian. Near Bethlehem, Pa., are the largest producing zinc-deposits at this horizon, the ores being largely carbonates and hydrous silicates, passing into sulphides in depth and comparatively free from lead. They occur in the same general line of strike, and but little more than fifty miles southwest from the franklinite deposits of New Jersey. Similar ores have been found in these limestones from point to point along the same line in New York, Pennsylvania, Maryland, Virginia, and Tennessee. In southern Virginia the sulphides of the two metals occur in workable quantity in a .region where iron-ores abound at the same general horizon but generally lower down in the limestone; the zinc-blende is sometimes segregated by itself and sometimes associated with galena. The occurrences are just northwest of a highly mineralized area of the older crystalline rocks.

TTie Uisaiasippi valley deposits, that have furnished such large amounts of these metals from comparatively undisturbed regions where there is no evidence of eruptive action, present exceptional conditions which have given rise to much speculation as to their, origin and manner of formation. In the upper Mississippi region,, which includes parts of the States of Illinois, Wisconsin, and lowa. the ores occur exclusively in the Silurian limestones, as the Carbe* niferous limestones have been entirely eroded away. Two main.sys- terns of fractures cross these limestones, which have served as chan- nels for the metal-bearing solutions, and from which they have spread out into 4hin shaly beds of the limestone. One series runs north and south ; a later series east and west, the latter carrying most of the ore. Lead predominates in the upper portions of the horizons, zinc in the lower. The fractures show evidence of displacement both vertically and horizontally. In the southwestern Missouri region, which includes also adjoining portions of Kansas and Arkansas, the ores occur in the Sub-Carboniferous in cross-fissures, in breociated zones, and along the bods, and extend to a limited extent into the overlying Coal Measure shales. There are two systems of fault-

Vol. Xxii.— 6

82 Geological Distribution Op The Ubeful Metaij3.

fissures. Lead occurs in the upper part of the deposits ; zinc is roost abundant in the lower, and now forms 90 per cent, of the produot of the region. The Lower Silurian (magnesian) limestones carry galena and zinc-blende in fractured regions of central and southeastern Missouri the latter being in subordinate amount. In southeastern Missouri a system of strong fault-fissures with parallel cross-frac- tures in the Cambrian limestone, which carry galena and pyritous ores with some nickel and cobalt, are furnishing more than half the lead product of the State, but no zinc. Here, as elsewhere, the ore has spread from the fissures to a certain extent outward into the strata, parallel with the bedding. This region lies directly east of the older granites and iron-bearing porphyries of the Ozark uplift, and the system of fractures throughout the State was probably formed by the movement which produced that uplifts Of more geo- logical than economic interest are the fault-fissures in the Sub-Car- boniferous limestones of southern Illinois and western Kentucky, which carry galena in a gangue of fluorspar, and similar fissures in the Lower Silurian limestone of central Kentucky, where galena occurs in a gangue of barite.

In the West, the Palaeozoic limestones have also furnished the greater part of the lead that has been derived from that region, but the zinc sulphide that occurs with it, in no inconsiderable amount, has thus far not been utilized industrially, and it is therefore impos- sible to form even an approximate idea of the proportion of the two metals in ore-deposits. A limited segregation of the two metals is observable in very large deposits, but both are mixed with pyrite, and economic conditions are such that ores containing a considera- ble percentage of zinc are either not mined or rejected. The mining of either metal would, in most cases, not pay were it not for the silver contents of the ore. The large concentrations of ores are almost invariably in immediate or proximate vicinity to considera- ble bodies of eruptive rock. They are found in the various moun- tain uplifts from the Rocky Mountains to the foot of the Sierra Ne- vada, and small amounts are obtained from California and Dakota. The greatest single producers have been the Carboniferous limestones of Leadville and Aspen in Colorado, and the Cambro-Silurian lime- stones of Eureka, Nevada. It has been observed, that while in the unaltered deposits of Leadville zinc sulphide is present in nearly equal proportions with lead aud iron, in the oxidized ore zinc is almost entirely wanting.

Although the deposits spread out into the limestones, of which

Qeoixx3Ical Distribution Of The Useful Metals. 83

they are metasomatic replacemenbt, the ore-bearing solutions have generally reached them along fracture- or oontact-planes, which in- dicate a Mesozoic, or even a later, age for the formation of the de- posits. This is also said to be true for the Mississippi Valley de- posits; but for those of the Appalachians no suflScieut study of their structural features has been made to justify the haearding of an opinion.

Lead and Zinc in Mesozoic and Tertiary Rocks

No deposits of lead or cine of economic importance are known in sedimentary beds of Mesozoic or Tertiary age though these metals are known to occur as mineralogical associates of silver-ores in sedi- mentary rocks of Mesozoic age, and in eruptive rooks of both Meso- zoic and Tertiary age.

Genesis of Lead- and Zinc-Deposits*

For the original source of lead and zinc there seems no valid ieason why we should not look to the massive eruptive rocks, as in the case of other metals. It is true, that their mineral combinations do not form prominently visible constituents of these rocks, as do the iron-bearing minerals, nor have concentrations of them yet been dis- covered which could be considered to be the result of differentiation in an eruptive magma. As tlieir deposits are found in nature, they are essentially precipitations from aqueous solutions, and their fav- orite habitat appears to be sedimentary limestones. Moreover, for the very extensive and important deposits of the Mississippi valley there are no known eruptive rocks within reaeh from which their metals could have been derived, and tlie opinion of most of the geologists who have made careful study of these deposits is that the metals in them were originally deposited with the limestones in a disseminated form, and that the present deposits are merely concen- trations of these finely disseminated minerals by downward perco- lating waters. On the other hand chemical analysis has detected their presence in appreciable amounts in some eruptive rocks not directly connected with ore-deposits, which is sufficient proof that portions of eruptive magmas may contain them as original constit- uents. If it is admitted that they were deposited with the Missis- sippi valley limestones, whether chemically or mechanically, they must have been derived from some earlier rock masses, and may well have resulted, either in £rst or second instance orven farther back, from the disintegration or decom|K>sition of older eruptive masses. The latest student of the Mississippi valley deposits Jenney). le

84 GEOLOGICAL DISTRIBUTION OF THE TJSEFnL METAU9.

whose most detailed studies were made in the southeastern Missoun region, finds the fissures in the limestones to be fault-fissures, and argues that they are probably deep-seated, and that the minerals have probably been brought up through these fissures from some deep- seated source in crystalline or eruptive rocks below. The fact that, in the upper Mississippi region, blende, which is at the lowest hori- zon, is generally of earlier deposition than galena, might be consid- ered an argument in favor of this hypothesis, though it is explain- able otherwise. On the other hand, their general association with barite in Silurian limestones, and the fact that fluorspar is found with lead only in Sub-Carboniferous limestone, is in so far an argu- ment of derivation from the limestones themselves.

In the West, the frequent association of their deposits with erup- tive rocks is most striking, and it seems likely that more systematic studies in the Appalachians may discover a probable association of areas of concentfation of their minerals with eruptives, from which they might indirectly have been derived. A most fruitful field of investigation lies open here, and one that is comparatively untouched for no general truths can be derived from the study of a single de- posit or group of deposits, and, as yet, the work either of individuals or organizations, in this country, has scarcely gone beyond this stage. It would also be interesting to determine how far the segration of the minerals of the two metals was due to differentiation in the original magma, and how far to a process of gradual selection in successive concentrations. In composite sulphide-deposits of great extent, like those of Leadville, which may be assumed to be the first concentration after that in the original magma, there appears to have been a certain amount of selective segregation by which certain por- tions contain a larger proportion of one or of the other metal, com- parable to the imperfect separation of the first part of an ore-dress- ing process. In the mineral economy of nature there is a generally observed tendency for like to seek like, as far as freedom of move- ment admits. It suggests itself, therefore, that those deposits which contain one metal, to the practical exclusion of the other, may be the result of a succession of such selective concentrations, and hence more removed from the original source than the more mixed deposits.

Quicksilver.* Quicksilver is said to be at least three times as abundant in nature as silver, but its occurrence in deposits of workable value is com-

Data with regard to this metal are from G. F. Becker's monograph on the sub- ject, 1888, and from private information communicated by hinUg bvGoOQlc

Oeolooical Distribution Of The Useful Metals. 85

parativelv rare, the bulk of the world's product having come from Almaden in Spain, Idria in Austria Huancavelica in Peru, the province of Kwei-Chau in China, and California. The production of the last has been nearly one-sixth of the total recorded production of the world. In 1880, California, the only State which has pro- duced important quantities of this metal, yielded nearly one-half of the total product, but in 1889 the California product had diminished to less than one-quarter of a smaller total.

The workable deposits of quicksilver in California have been found thus far, with some curious but economically insignificant ex- ceptions, in the coast ranges. The ore occurs in complicated systems of fissures and fractured zones forming deposits which Becker has called chambered veins, chiefly in highly metamorphosed rocks be- lieved to be of early Cretaceous or late Jurassic age, and which are lately composed of granitic detritus. The deposits are usually re- lated to volcanic rocks, rhyolite, andesite and basalt. At New Almaden a long rhyolite dike occurs at a short distance from and nearly parallel to the general ore-bearing fissure system. The for- mation of deposits from heated solutions appears to Ije actually in progress at Sulphur Bank and Steamboat Springs and the greater part of the deposits seem to be later than the close of the Miocene.

The principal ore is the sulphide (cinnabar), with whicli are found luetacinnabarite and native quicksilver. Tiemannite occurs and has been worked to a slight extent in Utah. Pyrite or marcasite is in- variably associated with cinnabar; and quartz, opal, calcite and dolomite are the usual gangue minerals, and bituminous minerals are not infrequent. Sulphide of nickel in small quantities is frequently found in the ores, and gold, copper and silver arc locally associated with them. Though the ore is found mostly in metamorphosed rocks, it occurs also in granite, andesite, basalt and Tertiary sand- stones.

Genesis of Quicksilver-Deposits.

The original source of the quicksilver and the associated metals is believed to have been in or below the deep-seated granites. The deposits are regarded as having been precipitated from heated solu- tions containing sodium sulphide, rising through fissure-systems, by relief of pressure and contact with surface-waters. The quicksilver- minerals have been deposited in interstices between rock-fragments iRod in masses of porous texture, particularly sandstones, but noth- ing like actual molecular substitutions or pseudomorphosis has been

Geological Distribution Of The Useful Metal9.

observed either in California or Spain.* In the Bavarian palati nate, however, cinnabar has been found to play the part of a fossil- izing mineral and has therefore replaced organic matter.

The only recent important development of quicksilver-ores in California is at the Mirabel mine, formerly known as the Bradford ; in the Mayacmas belt, Lake county. This mine yielded, in 1892, 3245 flasks. The production at the older mines, and particularly at the New Almaden, has fallen to a very low point.

Gold and Silver.

In manner of occurrence and chemical behavior gold and silver present more points of resemblance than of difference in nature, and may therefore more conveniently be considered together. In the native state all gold contains more or less silver, so intimately asso- ciated that it may be considered an alloy, though owing to the far smaller proportion of the former metal in nature, it cannot be de- termined that native silver always contains gold. From an indus- trial point of view, however, these ores are to be considered distinctly, most gold ores, which carry the metal in the native state almost ex- clusively, being comparatively free from silver values; on the other hand, while most silver-bearing ores contain this metal in various combinations with other metals and carry an appreciable portion of their values in gold, a considerable number are practically free from it.

The product of the two metals for the United States is estimated for the last century and the last twenty years respectively as follows :

Gold-Product

Period.

Troy onnces.

Kilograms.

coinage value. -"of

1792 to 1890

90,543,090 34,926,9?1

15,816,060 1,089,683

DoUare. 1371.706,769 33

721.920,000, 84

1870 to 1890

Silver-Product.

1792 to 1890

771.719,116 692,617,321

24.001,291 20,277,266

997.755,645 895,355,645

1870 to 1890

Becker draws a distraction between molecular sabstitntion and the deposition of ores in porosities which are due to precedent chemical action.

Geological Distribution Of The Useful Metals.

Of the total gold-produot of the country about 60 per cent, has been derived from the western slopes of the Sierra Nevada in Cali- fornia. Of the total silver-product about 50 per cent, can be traced with approximate accuracy to a number of single mines or raining districts, the relative production of which is given in round numbers in the following table:

Period.

Kilofframs of

KllogmmB of Gold.

Total

CoioAge value

of product

! Comstock Lode, Nevada,

4.820,000

214,000

$342,866,664

Eureka District, "

1,130,000

8,400

73,842,959

Leadyille DiBt, Colo.,...

2,700.000

6,000

116,236,178

j Aspen Dist., Colo.,

600,000

24,201,976*

Batte Dist, Montana,

1,650,000

7,400

69,637,824

Granite Mt. Mine, Mont

14,807,366

Ontorio Mine, Utoh,

660,000

27,688,892*

Horn Silver Mine, Utah,

250,000

10,303,730*

If we attempt to form an idea of the relative abundance in nature of the two metals by comparing their relative product we get vary- ing proportions for different periods, dependent on the conditions of the mining industry, and upon which it must be confessed a certain amount of doubt is cast by the uncertainty of statistics. Still the comparison is interesting as showing certain definite limits to the ratio.

Jiatio of Oold and Silver in the Product of the:

World .'From 1493 to 1890...

World

United States... United States

1792 to 1890... 1870 to 1890... 1792 to 1890... 1870 to 1890

1 to 19. 1 to 13i. 1 to 16. ItoSJ. ltol9.

Oold and Silver in the Older Crystalline Rocks. In the Appalachians both gold and silver are found in veins in the crystalline areas of the Appalachian system at various points from Maine to Georgia. Silver, being generally in too small pro-

Gold not segregated.

88 Geological Dxstribution Of The Useful Jietau.

portion to be of economic importance, has been, for the most part, neglected. Gold, on the other hand, especially in the southern por- tion, where the glacial sheet has not planed off the decomposed sur- face of the country, is often concentrated, either in gossans or in placers, in paying quantities, and has been mined intermittently since the very earliest days of our history. The enclosing rocks are gen- erally schistose or foliated, and may prove to belong, for the most part, to some of the Algonkian series. The deposits have in many cases been considered to be quartz-lenses parallel with the bedding of the rock, but from my own observations I am inclined to think that more careful study will prove the apparent bedding to be the result of compression, and that the deposits are on actual fracture- planes. The association is invariably with quartz and pyrite or its decomposition products, but the great bodies of pyrite of the pyri- tous belt seem to carry proportionately less gold than smaller veins at some little distance from them. In many cases the association with eruptive dikes is known to have had an important influence in the concentration of the ores. The number of such cases will un- doubtedly be much increased when more careful studies are made of the region. The great depth to which the rocks in the southern region are altered beyond recognition makes such studies more diffi- cult than in the West or North.

In the Northwest both gold and silver are known to occur, and have been mined to a limited extent in veins in the Huronian jrocks of the Lake Superior region, where also the association with Eruptive rocks is noticeable. Silver seems, from present evidence, to be in larger proportion than in the Appalachians. Here alsd they are not, apparently, directly associateil with the great masses of iron ore.

In the West the crystalline areas of the Black Hills, Wyoming and Colorado, and the granites of Montana, Idaho and Nevada, have been the principal producers of the precious metals among the older rocks. The Black Hills rocks have been determined to be of Al- gonkian age; those of Wyoming and Colorado have thus far been considered Archsean. In either case the deposits, which are on fracture-planes, are associated with eruptive rocks. In the Black Hills they are proved to be of pre-Cambrian age by the existence of their detritus in Cambrian sandstones. In Colorado the deposits, which are later than the dikes, whose age has not yet been deter- mined, may also be pre-Cambrian. The ores of the Black Hills are almost exclusively gold-bearing pyrites. Those of Colorado are in

Geological Distribution Of The Useful Metals. 89

places mainly gold-bearing pyrites in others mixtures of sulphides of silver, lead and zinc; in others, again, tellurides of gold and silver.

The age of the granites of the norihem Rocky Mountains is not yet definitely known, but, until proved to the contrary, may be as- sumed by analogy to be pre-Cambrian. In the Butte and Granite Mountain districts the values are mainly in silver associated with sulphides of copper, lead, zinc and iron in varying amounts. In other districts they are mainly in auriferous pyrites, and, again, in others in sulphides and arsenides of silver. The granite is itself eruptive and often cut by later eruptives. The deposits are in sys- tems of fracture-planes, which sometimes extend into adjoining rocks and are, necessarily, later than the granite.

Oold and Silver in the Palceozoic Rocks.

In sedimentary beds of the various Palaeozoic formations deposits carrying values in silver are more common than those carrying gold in appreciable amount. Silver, like lead, seems to find a preferable habitat in limestones; gold occurs rather in siliceous rocks.

In the Eastf as far as known, no MPrkable deposits of either metal occur in Palseozoic rocks, though some of the gold-bearing schists may yet prove to be of Cambrian age, if the assumption of this age for the gold-bearing rocks of Nova Scotia is well founded.

The small values in silver of the lead-ores of the Mississippi valley are hardly to be taken account of.

In the mountainous regions of the West, on the other hand, the Palaeozoic sedimentaries are important producers of silver and carry a certain amount of gold also. Their deposits, moreover, are almost invariably iif the vicinity of, if not in immediate juxtaposition with, bodies of eruptive rock. The most important deposits of silver- bearing sulphurets in the limestones of these horizons have been found :

1. At Leadville, Aspen and a number of smaller districts of Colorado, encircling the Sawatch uplift, mostly in Carboniferous limestones, whose gold values have been insignificant.

2. In the Silurian and Cambrian limestones of Eureka, Nevada, a third part of whose values has been in gold.

3. Various mining districts in New Mexico, Arizona, Utah, Mon- tana, Idaho and Nevada, which have yielded considerable amounts of silver from Palaeozoic limestones, in which the lead-product is generally of economic importance, though in many cases, especially

90 Oeolooical Distribution Of The Useful Metau.

in the two laRt-named States the silver minerals are suflSciently free from lead or zinc to be amalgamated.

In siliceous Palseozoic rocks the Ontario and a few smaller mines in the Wahsatch Mountains, and some in Bingham, in the Oquirrh Mountains of Utah, whose ores occur in Carboniferous quartzites, are the most important silver-producers. They also contain some lead, but no important values in gold. In Colorado, on the other hand, the underlying Cambrian sandstones are often gold-bearing where the limestone horizons are rich in silver. In one case it has been stated that the gold has probably been leached from the over- lying argentiferous sulphides. In California a limited number of the deposits of the gold-belt occur in siliceous rocks to which a Car- boniferous or earlier age has been provisionally assigned. They are, however, contemporaneous with and analogous to the bulk of the ores of this belt which occur in Mesozoic rocks.

Although these deposits in many cases have a considerable extent parallel to the bedding, even when they do not also occur on cross- fractures, such fractures are generally to be found from which the ore-bearing solutions have spread out into the l>eds. Their age may therefore be determined as lat than the fracturing, which may be assigned to some well-known or(raphic movement. Such move- ments took place near the close of the Jurassic and again at the close, of the Cretaceous, the latter being the more important and wide- spread and, moreover, more generally accompanied by eruptive activity. To one or the other of these periods, therefore, mut the commencement of ore concentration at these horizons be assigned.

Gold and Silver in the Meaozoio Rocks.

In the Ead there are no known deposits of economic value either in gold or silver in Mesozoic sedimentary rocks.

In the Wedf on the other band, rocks of Mesozoic age have yielded the greatest part of our gold product, and in certain localities, a notable amount of silver.

The auriferous slate of California, which is the common name given to the belt of rocks along the west flanks of the Sierra Nevada from which has proceeded the greater part of the gold produced in the State, includes in the light of the most recent investigations, both Mesozoic and Palseozoic rocks. The greater part of the gold-bearing veins occur in the former, which are considered of Jurassic and early Cretaceous age, and in the intrusive bodies of dialMse and diorite which abound in them. AH these rocks are steeply upturned and

Geological Distribution Of The Useful Metau9. 91

are unconformably overlaid by nearly horizontal beds of late Creta- ceous age which contain detrital gold but no gold veins. Hence the age of the deposits may be assumed as post-Jurassic and probably Cretaceous. Most of the veins carry gold and a little pyrite with no other metallic minerals in a gangue of white quartz. The veins in diabase contain a good deal of galena, blende and pyrite. The majority of the veins are meridional and show a certain pafallelism with contacts of eruptive masses ; in certain parts, however, cast and west veins are the more productive.

In western Nevada high-grade silver-ores occur in Triassic rocks, and in Utah silver is found in paying quantity in coarse white sand- stones supposed to be also of Triassic age which, exceptionally, are not associated with eruptive rocks. In Colorado both gold- and silver-bearing veins occur in Cretaceotis beds, closely associated in either case with peculiar types of eruptive rocks. At Leadville, Colorado, gold-bearing veins occur in eruptive rocks, and in them as a somewhat unusual occurrence, native gold is sometimes asso- ciated with galena.

Oold and Silver in the Tertiary Rooks.

In Tertiary sedimentary rocks no workable deposits of either gold or silver have so far as known been yet developed in this country.

Of the great variety of eruptive rocks in which valuable deposits of either metal occur in the West, the greater part, if granites are left out of consideration, are of Tertiary age, and of those that may have been erupted a little earlier most of the concentration into ore- deposits has taken place during Tertiary time.

The most important of these deposits, the great Comstock lode, which has produced more gold and silver than any lode or single district in the country, is a fault-fissure cutting across a variety of eruptives of dioritic and andesitic types assumed to be of Tertiary age. About 40 per cent of the bullion obtained from Comstock ores is in gold.

The greater part of the silver product of California comes from the Tertiary rhyolites of San Bernardino county, which belong ge- ologically, like the Comstock, to the Great Basin province. Many of the deposits of rich silver minerals in granites through western Nevada, eastern Oron and Idaho are so associated with eruptions of rhyolite that they may be assumed to be of Tertiary age. In all these deposits there is a very small proportion of the base metals — lead and zinc. The Horn Silver mine of Utah is on the contact of

92 Geological Distribution Of The Useful Metals.

limestone and a recent eruptive, and man/ adjoining deposits are entirely within recent eraptive rocks. In Colorado many of the important silver veins of the San Juan region, and those of the new Creede district are in recent eruptive rocks, presumably of Tertiary age. The same is true of the gold-deposits of the Summit district and of the new Cripple Creek mines. The recent gold developments in WasJington are in an eruptive region, though little is as yet known of the actual relations of the deposits.

Gold and Silver in Detrital DeposUs,

Gold is very largely .derived from detrital or placer deposits, which are the mechanical concentration of material resulting from the abrasion of older gold-bearing rocks. Such gold carries a small amount of silver, but in less proportion, as a rule, than the gold of adjoining vein-deposits. Such deposits are, for the most part, of recent formation, but placers are known to be of much earlier forma- tion. Among old placers may be mentioned those of the Cambrian sandstones of the Black Hills of Dakota, those of the late Cretaceous sandstones in the foothills of the Sierra Nevada and, still more im- portant, the auriferous gravels of old river-beds of California and Idaho, supposed to be of Pliocene age. Many of the gold-bearing gravels of the Rocky Mountain system in Colorado and Montana belong probably to the glacial period, and are older than the gravels of modern streams.

Genesis of Gold- and Silver- Deposits.

The frequent association of deposits of gold and silver with emp tive rocks, the world over, has long been remarked. Chemical in- vestigation of many eruptive rocks has detected their presence under such conditions as leave little doubt that they were original con- stituents of these rocks. Recently a Grerman geologist kas reported the discovery of gold in a late eruptive rock in Chili, which could be actually seen, by the aid of the microscope, to be an original con- stituent of the rock. There seems very good reason to assume, therefore, at any rate as a working hypothesis, that the original or ultimate source of these metals has been the eruptive rocks.

With regard to their subsequent dissemination in sedimentary beds, whether by mechanical or chemical agencies, there appears to be less satisfactory evidence, as there are few known concentrations which can with much probability be assumed to have derived their material exclusively from sedimentary rocks. The concentration of

Qeoloqical Distribution Of The Useful Metals. 93

the metals in workable ore-deposits has evidently been by the agency of aqueous solutions; detrital deposits are only the mechanical re- arrangement of such concentrations, though some maintain that these have been enriched by precipitation from solutions.

Aside, then, from the study of the structural relations which would afford favorable conditions for the concentration of metal* bearing solutions and the precipitation of their contained salts in workable ore-bodies, which is of common interest and importance with regard to deposits of all the metals, a most fruitful field of re- search, and one which promises results of economic as well as scien- tific importance, is afforded by the study of the chemical and mineral- ogical affinities of these two metals, and their probable behavior under the conditions which may have existed where deep-seated de- posits were formed. Much obscurity still exists as to the actual chemical condition of ailver in galena and of gold in pyrite. The suggestion has recently been made that combinations of these metalf, as alloys or otherwise, with small amounts of tellurium, bismuth, etc., are much more common in nature than has hitherto been sus- pected, and may be the reason of the unexplainable difficulties found in amalgamating certain ores, and further investigation on this line may produce important results. There are some features, also, with regard to the behavior of these metals under the action of atmos- pheric waters, and their consequent concentration along the zone of alteration of sulphide-deposits which are not entirely clear, and demand more systematic and careful investigation.

Conclusion.

What is at present known about the distribution of ore-deposits west of the 100th meridian does not seem to call for any serious modification of the statement as to their general distribution made by Clarence King in 1870. It is probable that if the subject were carefully worked up in detail it would be found that the meridional zones laid out by Mr. King contain, as Raymond has suggested, a greater variety of minerals than he was at that time aware of.

In the eastern half of the continent it is evident, from the facts given above, that certain geographical areas are peculiar in contain- ing great concentrations of certain varieties of minerals, but it seems hardly necessary to recapitulate the peculiarities of these areas, since it is the geological rather than the geographical distribution that is of practical importance. The former must have a genetic bearing; the latter can only have such bearing through geological causes.

94 Geological Distribution Op The Useful Metals.

UnfortuDately, our knowledge of the geological relations of the ore- deposits of our country is as yet too incomplete to afford material for any exhaustive generaliasations on the geological relations of the useful metals as a whole, or the underlying genetic causes of such relations. The fissure systems, or the natural water-channels which have admitted of the concentration of the metals into workable de- posits, have, as pointed out by King, Becker and others, certain definite relations with the great orographic movements, and these relations admit of our forming an idea of the relative age of the de- posits. They do not, however, afford any reason why certain min- erals are more prevalent in one district and certain others in another ; nor do they necessarily afford any clue to the original source of the metals. A certain amount of systematic geological work has already been done by our Geological Surveys towards the solution of these important problems, which are of practical, well as scientific, im- portance, but a vast amount remains yet to be done, and many large fields are still practically untouched.

The suggestion offered above as a working hypothesis seems to be one worthy of consideration by the workers in this field. If the metallic minerals do concentrate in eruptive magmas within the crust of the earth in accordance with some law not yet clearly known, but which results in what is called differentiation, by virtue of which certain areas of igneous rocks, formed by successive extrusions of material of differing chemical composition which have cooled at or near the surface, are found to be unusually rich in minerals contain- ing a given metal or class of metals, a basis is afforded to account for the unusual abundance of deposits of these metals in a given area. In the case of the older eruptive rocks, the accumulation of mineral combinations of the metals into workable deposits may be the result of many processes of concentration, both mechanical and chemical. The concentration of material derived from younger eruptive rocks, on the other hand, would be more direct, and mainly chemical, by the sole action of percolating waters. In either case, did investiga- tion prove certain areas of eruptive rock were unusually rich in mineral combinations containing a given metal, it would afibrd reasonable ground for looking for valuable deposits of that metal in the vicinity, especially if the geological conditions of rock-alteration or metamorphism and dynamic movements are such as to favor con- centration.

If sedimentary beds carry disseminated minerals, or concentration of such disseminated minerals into ore-deposits they might have

Miking And Mineral Statistics. 95

been derived ultimately from the abrasion of bodies of igneous rocks rich in their minerals by differentiation. How close a proximity would constitute a vicinity would vary widely under varying geo- logical conditions. It is quite uncertain how far percolating waters carrying minute amounts of metallic minerals in solution might travel through underground water- passages without depositing their load, but the possible distance is evidently very considerable ; and the argument sometimes advanced against the lateral secretion theory, that proof can be found in certain cases that the mineral of a vein could not have been derived from the immediate wall rock, is no valid argument against this theory in its broader acceptation, which admits the secretion from neighboring bodies of rock not necessarily in immediate proximity but possibly at considerable distance and not visible at the surface.

For the derivation of sediments, the possible distance of the source of materials has still wider limits; but analogy from the conditions under which sediments are deposited in present oceans would bring it within a hundred miles as a probable limit; here, also, it may readily happen that the eruptive body from which the metallic min- erals were derived is not visible on the surface.

Mining And Mines Al Statistics.

BY C. LK NEVE FOSTER, D.Sa, F.R.8., LTJkNDUDNO, NORTH WALES. (Chicago Meeting, being part of the International Engineering Congress, August, 1898.)

The object of this paper is to offer a few suggestions for improv- ing the mining and mineral statistics presented by the governments of various nations on both sides of the Atlantic At the present time one may fairly complain of two evils — lack of uniformity in some cases and want of completeness in others. The consequence is that the statistics are either less plain than they might be, or are useless for comparisons, or fail to present a faithful picture of the htAte of the mining industries of a country.

The principal subjects upon which information is desirable are: the amount and value of the mineral products raised and the number of the mining population. With these data it is possible to form an estimate of the importance of the mining industries of any given

96 Mino And Mineral Statistics,

land; but raany countries also publish lists of mining accidents and compare the fatalities with the number of persons employed, in order to deduce a death-rate which may be treated as an index of one of the risks pertaining to the trade* If these death-rates are collated from year to year, it is possible to gauge the increase in the security of the workman's calling, which is being effected either by legisla- tive enactments or new inventions, or both these causes. The chance of an accident is not the only danger incurred by a workman ; there may be a rapid shortening of the lives of all the members of a trade due to unhealthy surroundings, far worse for them as a whole than the occasional casualties incidental to their crafl.

We may therefore say that some of the most important data are:

1. Quantity of mineral raised.

2. Value of the minerals.

3. Number of persons employed in producing the minerals.

4. Number of fatal accidents among these persons.

5. Death-rates from accidents.

6. Duration of life of the persons employed in and about mines. In addition to all tliis, many countries will like to know the

revenue they obtain from mining property belonging to the State, while from a commercial point of view it is not unimportant to be able to ascertain the financial results of the industry; that is to say, how far mining is profitable to those who risk their capital in seek- ing the hidden treasures of the earth's crust. However, for the purpose of this paper, I will confine myself to the subjects enumerated under my six headings.

1. Quantity of Mineral Raised.

Various points for discussion crop up even in so simple a matter as a record of the quantity of mineral raised in a country, viz: a. Standard of weight. 6. Definition of a mineral.

c. Advisability of having one complete general table.

d. Arrangement of the minerals.

c. State of elaboration of the minerals.

This is not the place for entering into any battle concerning weights and measures. I accept the unfortunate fact that various standards are in use, and I would simply propose that each country in addition to recording the total quantity of each particular mineral according to its own customary weights, should also add the amount reckoned according to the metric system.

Mining And Mineral Statistics. 97'

Mere reduction to a aniform standard is an easy matter compared ' with the next pointy i.e., the definition of a mineral. Some coun- tries, such as France, include peat as a mineral ; in the United King- dom it is omitted. It would be easy to contend that kauri gum, the semi- fossil resin of a New Zealand pine, has not been buried long enough to put it on a par with amber ; and yet it is always included among the mineral products of the colony. Even when a line has been drawn across the borderland between the vegetable and mineral kingdoms, other difficulties await the statistician yearning for uni- formity. Are or are not mineral waters to be included ? No objeo- tion is ever made to the products of brine springs, and, if so, why should sodium chloride he favored to the exclusion of other salts? In the statistics of the United States credit is taken, and very rightly,! for the natural gas which is being obtained in such large quantities;* bat in the figures given for Prussia not a word is said about car- bonic acid ; although the trade in the liquid acid, compressed froni' the natural gas, is far more important as regards value than that of* some of the Bubstances mentioned. No doubt the reply to this objec- tion is that carbonic acid is not a mineral belonging to the state ; but' this explanation scarcely satisfies the inquirer in search of informal . tion upon the mineral resources of the German empire.

The want of one complete general table showing at a glanoe* the total mineral production of the country is a common evil.* Owing to the statutes which regulate mines and mining property in some countries, certain kinds of mineral are treated differently to* others in the official tables. Take France for instance. We find a tablet entitled : '' Resum6 g6n6ral de la statistique des mines et des autres exploitations minerales." The unwary foreigner would very likely jump at the conclusion that this table was a summary of the total mineral product of France; but he would be wrong. It does not include such minerals as gypsum, phosphate of lime, slate, stone, eta, the workings for which are "quarries" by the French law and not mines. Here we at once have a fundamental difference between the United Kingdom and France. In the former it is the nature of the excavation, according as it is above ground or below ground, which determines for legal purposes whether any given working is; a mine or not ; in the latter the question is decided by the nature of

ZeilKkrififuT das Berg-y Hutten- und Salinen- Wemm m iVcuMwcAm Siaaie, vol. xl., Berlin, 1892, sUtistical part, pp. 20 and 21.

t SUiH$Uque de Pinduttrie inmra/ ei de9 appaartii$ d vaptw m France d en Algirie fHmr Patmie 1890, Paris 1891, p. 41.

Vol, Xxii.— 7

98 Mining And Mineral Statisti03.

the miDeral and not by the method of working it. It is evident that raining engineers and students, anxious to learn something about France, would be saved time and trouble if the official volume contained upon one page a complete table of all the minerals raised, no matter how they were obtained or how they were classed accord- ing to the law of the country. So much careful attention and labor are expended upon the excellent French statistics, that the extra trouble of preparing such a summary is not worthy of consideration. Precisely in a similar manner the mineral statistics of the Kingdom of Italy might be misread by a person who was not on his guard. A general table is given headed '' Products of the Mines/' and as it includes salt from brine wells, petroleum, mineral waters, and boric acid from steam-pufis, one might fairly suppose that the picture of the mineral wealth of Italy Was complete, especially as no hint is given that other minerals are produced in the country. Some of these, which are considered as the product of quarries," are men- tioned in a subsequent table (p. cxviii.). Further, supposing that the reader has m&stered the fact tliat he has to consult two tables, it may not strike him that salt is obtained from sea- water on the coast of the mainland and its adjacent islands. The quantity so obtained is stated in a third table (p. cii.). No doubt these mat- ters are plain enough to the government mining engineers; but they are not necessarily apparent at first sight to the Italian unacquainted with all the resources of his country and certainly not to the foreigner.

Even the minutely precise statistics of the Prussian Governmentf are not so complete as one would wish. Not a word is said about slate and other stone quarries, though some are worked underground as true mines, and I have already alluded to the absence of any ref- erence to carbonic acid.

. I do not wish it to be supposed, while criticizing foreigners, that I consider my own kith and kin to be perfect. For instance, in spite of there being page after page of statistics compiled with the greatest care for the colony of Victoria, I do not find any one table giving a summary of all the minerals raised. Two tables have to he consulted and all the values added up before a total can be ar- rived at.

Rivista del BcrviMio minario nd 1890, Florenoei 1892, p. xcii. t Op.eiL

X Annual Report of the Secretary of Minujw the Yoar 1891, Melbourne, 1892, pp. 101 aud 109.

Mining And Mineral 8Tati8Tigb. 09

Great diversity of arrangement is found in the tabular statements of the amounts of mineral raised. Sometimes the most valuable mineral is chosen to head the list; in other cases the order in which the minerals are placed seems to depend entirely upon fancy. To my mind there is much to be said in favor of the alphabetical system adopted in the United Kingdom. It commends itself by its simplicity, as well as by the fact that it is known to every- body.

I feel loth to find fault with a volume so full of useful informa- tion and so carefully indexed as The Mineral Reourcea of United SiaJteSj calendar years 1889 and 1890, lately issued by the Director of the Geological Survey; but I cannot .help remarking that the General Summary" (pp. 6 and 7) would have been improved if the various minerals had been arranged alphabetically. Indeed, the various articles in the book itself might just as well have been printed in a similar order.

The difference in value caused by "dressing," or some other method of treatment, is so great that the statistics of two countries are useless for the purpose of comparison unless the same stages of elaboration are taken in both cases. A few examples will illustrate my meaning. In the statistics of the United States just referred to, we find the value of the metah added to the value of the coal and other non-metallic minerals; in Great Britain we beep the mining products proi)er separate from those obtained by smelting. The effect of such a fundamental difference of practice is far too great to be ignored, and must always be taken into account in making com- parisons. The value of the iron-ore produced in the United King- dom in the year 1891 was £3,365,860, and it was calculated that the iron produced from this ore was worth £11,886,819, a difference of more than eight millions; and if we had adopted the American sys- tem, the total value of our mineral products would have appeared nearly nine millions sterling higher than the figure at which it is put in the official publication.

In the Italian stitistics, strange to say, we have quicksilver reck- oned in the metallic state and all the other metals taken as ores.

Though it is comparatively easy to make a separation between the products of the miner and those of the smelter, that is to say, between ores and metals, differences of opinion may arise con- cerning the stage of elaboration at which the value of a metallic or a non-metallic mineral should be taken. Some tin mines, for in- stance, do not dress their ore at all, but sell it to the owners oftamp-

100 MINIKG AND MINERAL STATfSTICS.

ing mills, who produce a smeltable product by procesBes of crushing and washing, which, as a rule, are carried on by the mining company itlf ; therefore, some persons might argue that the weight and value given should be those of the undressed and not of the dressed ore. In most cases I think mining engineers would agree that the practice of reckoning the ore in the dressed state is the more useful.

Some agreement between nations concerning the non-metallic minerals is necessary if comparisons as to value are to be made. One country ay reckon its clay and brick-earth in the raw state, while another will give their value when converted into drain-pipes, tiles, and bricks ; again, one country may take sandstone as it oomea from the quarry, and another wait until a similar stone has been made into grindstones. While Italy takes credit for clean sulphur ready for the market, Spain* gives the value of the raw sulphur-bearing rock as It comes up from underground under its minerales.'' The sul- phur subsequently extracted from it by liquation is then placed under the' productos metal drgicos '' with metals obtained by smelt- ing, and also with coke and briquettes. The term 'productos metaldrgicos '' is a dangerous pitfall. In the '' General Summary for 1888 "t we find, for instance:

Produccin.

MlneraloB. Productos metaldrgicosw

Toneladas. Toneladas.

Ziuc. 74,353 26,173

At first sight the reader might suppose that the 26,173 tons rep- resented spelter obtained from the 74,353 tons of ore, but, on refer- ring to a foot-note in small print, it appears that only 5117 tons are metallic zinc and the rest calcined calamine. It is of the utmost importance that titles should be so worded as to leave no doubt in the mind of the reader as to what is really meant. The framer of a table oflen fails to remember that his readers may not possess all his knowledge of the subject, and are therefore liable to fall into errors wiiich he may consider as hardly possible.

2. Value op the Mineraus. Some difference of opinion may exist concerning the place at which the value of a mineral should be calculated, whether at the mine, at

Comisidn eaUtva de EstadUtica Afinera, Datoa estadUtieoa eorrespondieiUei al afSitcondmieo de 1887-88, yalos aflOB n<UuraU$ de 1887 y 1888, Madrid, 1890. Plate opposite p. xiii. t Ibidem, p. 353.

MINING AND MINERAL STATIStigd. . 101

the smelting works, at a railway station, or at a port* of shipment. Probably uniformity will be best secured by taking yljae at the place of production before any freight has been paid. IiEould be convenient also that any international tables should statel(bre total values according to the gold standard of some important country. -

3. Number of Persons Employed. ;-

' A statement of the persons employed in and about the mines and-

open works should show how many work below ground and how

many above ground, and should distinguish males from females thus:

Total number of persons em- ployed in and about the mines, brine-wells, mineral springs, open works, etc.

A. Persons employed below ground, -j

Females.*

{Males. Females.

In some countries it may also be advisable to have a classiBcation according to race, and separate the white men from their brown, black or yellow brethren.

It is not easy to lay down very strict limits for Class B ; if the persons employed in dressing ores are included, it only right to treat in a like manner those who are engaged in washing coal apd making it into patent fuel. The manufacture of coke, when carried on by a mining company, is the preparation of its mineral for the market, and if this is once admitted as a mining process. Class B must contain the persons who deal with the by-products which are sometimes collected from the escaping gases. However, the numbers belonging to Class A, which is of far more importance than B, can always be given with suiScient accuracy for all practical purposes.

4. Number of Deaths from Accidents.

In Great Britain we treat as a fatal accident any casualty which causes the death of the person injured within a year and a day of the date of its occurrence. Some definition is necessary, for I have known persons argue that an accident was not fatal, because the victim was not killed on the spot and lingered days or weeks before succumbing to his injuries.

Of course, where persons employed are classified according tp

It must not be supposed that I am advocating the retention of female labor. below ground ; bat as long as women are allowed to work underground in some oonntries, the fact should be recorded.

102 . ipieiNG AND MIKEBAL STATISTICS.

race, it wilj be well to carry on the same distinction in the case of accident;.. For the purposes of what may be called the international tables Da'minate subdivision of accidents according to causes need be attiented ; quite enough will be done if the accidents happening belo.w ground are distinguished from those happening at the surface.

6. Death-Rates from Accidents.

As death-rates from disease are usually calculated per 1000 of the

population, it seems most convenient to adopt the same method in case of accident** especially as it has already found favor in many countries. It is very necessary to calculate the death-rate of the underground workers separately and to avoid mixing up the miners with the surface hands, as is commonly done. Unless attention is paid to this point, no comparison between the relative degreefl of danger of different kinds of mines can be attempted. The surface workers have a small risk compared with the underground men and if their proportion is large, the average death-rate of the total num- ber employed is lowered very considerably and the miner's occupa- tion appears less risky than it really is.

An example must be given in order to make this clear. In Great Britain the working of mines is regulated by two Acts of Parlia- ment, viz., the Coal-Mines Regulation Act and the Metalliferous Mines Regulation Act. The former governs mines of coal, dratified ironstone, fire-clay and shale; the latter applies to all other mines, and therefore includes underground workings for ores and stones of different descriptions.

The following table, which makes a comparison between the death- rates of the two statutory classes of mines of the United Kingdom, has I)een compiled from the official figures which are published annually.

If the death-rate is calculated from the total number of persons employed above and below ground (columns 4 and 8), the mines under the Coal Mines Act appear to be more dangerous than those under the Metalliferous Mines Act in both periods, but when the error due to the comparative immunity from accidents of the surfiM workers is eliminated (columns 2 and 6), it turns out that during the last eight years the mines under the former Act have proved less fatal to the underground workers than those under the latter. In other words, a given class of mines may appear more dangerous or less dangerous than another, according as the surface hands are ex- cluded from or included in the calculations.

Mining And Minebal Statistics.

nines of the United Kingdom of Oreat Britain and Ireland, together with the Me of Man.

Under the Coal-Mines Regulation Act

Under the Metalliferous-Mines Regula- tion Act.

Av. annual death-rati

fh)m accidents per lOCX

persons employed.

the No. of the surface w'rk'rs beam to the to- tal No. of per- sons employ*d above and be- low ground.

Ay. annual death-ratt

from accidenu per 1000

persons employed. .

Proporl the NO. surface bears tc Ul No. sous en above i lowgrc

Below ! Above ground ground

A bore

and

below.

Below ground

Above ground

AbOTO

and

below.

Ten vrs.

1874 to

Bgktjem 1884 to

Per cent. . 197

Per cent.

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

In the Prussian statistics we find a carefully-prepared table ia which the death-rates are given for each year from 1867 to 1891 for four classes of mines, vilk., workings for coal, brown coal, ores and sundry other minerals. These figures enable the progress on the path of safety to be traced in the case of any one of the four classes, but they do not afford the means of comparing the risk of the un- derground men of one class with that of the underground men of another. It is precisely this underground risk which requires to be calculated and known.

6. Duration op Life.

As pointed out in my introduction, a knowledge of the effect of a trade upon the duration of life is no less important than the know- ledge of its liability to accidents. Certain occupations may expose their workmen to little risk of accident but to great chance of dis- ease. The vital statistics of a calling are consequently just as im- portant as the figures which deal solely with accidents. The mor- tality of males engaged in different occupations has been carefully studied in this country by Dr. Ogle, Superintendent of Statistics at the General Roister OiSce, London, and his Report f may fairly be taken as an excellent model for investigations of this kind. Many of his conclusions are of the utmost importance. He finds, for '

♦ Op.eiL,. 29.

t SuppUmerU to the Forty-Fifik Annwd Report of tkt Begistrar-Oeneral of Birtki, Deatki and Marriagei in England, London, 1885.

101 Mining And Mineral Statistics.

instance, that the death-rate of coal-miners is low; in other words, that, in spite of their liability to accidents, they live deci- jdedly longer than the average male. On the other hand, figures j which he has checked with the greatest care tell us that the Cornish (miner has a very unhealthy occupation, almost the worst of the hun- jdred callings which are selected for comparison. These facts natu- Irally suggest the advisability of similar researches into the healthi- iness or unhealihiness of mines in other countries. I I need scarcely remark that I have no wish to dogmatize while making these proposals; on the contrary, I am most anxious to 'learn the opinions of others, and I have great hopes that most per- sons will agree with the main object in view, viz., the establishment 'of some uniform system of mining statistics* At the same time, it j should be recollected that I am not proposing that any nation should I alter the form of any statistical tables which experience has led it to adopt; I merely suggest that, in addition, it should furnish annu- (ally a brief set of tables drawn up in accordance with the opinions of a representative Mining Congress at Chicago. These tables, after 'having been published with the official reports of each country, oould then be collated by some central body, such as the American .Institute of Mining Engineers, and published in the form of a small pamphlet as The Mineral Statistics of the World, References to the original sources of information might be given, so that persons re- quiring more details would know where to find them. One more suggestion : Would it not be possible to take advantage of the Chi- cago Mining Congress for collecting a short account of the nature and scope of the official mining departments of all nations which have one? I think many persons would be glad to have this infor- mation, especially if they could learn precisely what official docu- ments are issued annually.

In conclusion, I trust that my perhaps too candid opinions will not be taken in ill part by the compilers of the statistics which I have ventured to criticize. I wish it to be understood that I fully appreciate the skill with which their arduous labors have been ac- complished ; I merely suggest tneans by which their talents and experience may be rendered still more beneficial to mining engineers and to the world at large.

8Egb£Qation And It8 Consequekce8. 105

8Eqbeqati0N And Its Consequences In Ingots Of Steel And Iron,

BY ALEXANDRE POURCEL, PARIS, FRANCE.* (Chicago Meeting, being part of the IntemationAl Engineering CongreaBt Augustt 1893.)

The phenomena of liquation in steel or iron ingots of all sizes, but naturally to fsyesXest extent in the heaviest ingots, have been noticed ever since the commencement on a large scale of the Besse- mer and open-hearth manufactures ; but they have been studied with care during the last few years only. The English and the Ameri- cans were the first to publish the results of investigation upon phe- nomena of this class, to which they have given the name of aegre- gaiian.

Historical,

Passing by that which has l)een written concerning the segrega- tion observed in gray pig-iron, of which instances were given by Karsten, we shall confine ourselves here to that which is known of this subject with exclusive reference to ingot steel and iron,

Mr. H. M. Howe, of Boston, in his Mdallurgy of Steel, the most complete and well-arranged treatise upon steel, gives a riaumi of observations, with the results of numerous analyses, bearing upon the phenomena of segregation.

According to Tchernoff (with whom we discussed this question at length in 1878, in connection with the cast-steel plates of large di- mensions designed for the Italian fleet, which we showed him at Terre-Noire), Kalakoutsky called attention in 1866 to the lack of homogeneity in Bessemer ingots, especially when cast in sand.

Our own attention had been called to the phenomena of liquation in 1868 by a somewhat surprising fact which appeared in almost every one of a number of steel-rail ingots coming from the works of Messrs. Piern? and Emile Martin at Sireuil, to be rolled at Terre- Noire. In the heating or in the first groove of the rolls, a slice of from 6 to 10 centimeters (2.4 to 3.9 inches) separated itself from the head of each ingot. As a general consequence, the section of

Translated bj the Secretary. The metric measaree have been retained, bat

106 8Egbeoation And Its Consequences.

rail coming from the upper part of the ingot was weak, while the section from the bottom of the ingot endured admirably the tests required by the railway.

These observations led us to adopt at Terre-Noire in 1870 a special treatment for Bessemer ingots intended to be forged into cannon. The metal was poured into a cast-iron ingot-mould of hexagonal section and truncated pyramidal form and very thick walls sur- mounted by a mould of heated sand.

Thus, besides the weight of the metal required for the cannon- ingot, there' was poured an additional weight lialf as great in an en- velope previously brought to red heat in order to preserve its fluidity as long as possible after the solidiflcation of the mass contained in the ingot-mould.

This arrangement was always used at Terre-Noire in casting ingots of all weights for shaft-forgings.

It is evident that this does not absolutely avoid the partial liqua- tions which taice place in the center of the ingot, principally in the zones nearest to the part cast in sand, and serving as massdotte (sinking-head); but this is an inconvenience without consequence for cannon, and of little importance for forge-shafts; since in the first case the non-homogeneous material disappears in the boring of the piece ; and, in the second case, finds itself in the region of the neu- tral fibers of the solid. The accident which occurred to the boilers of the Lividia gave rise to very lively discussion in England upon the causes of the lack of homogeneity in steel boiler-plates. The specifications at that time adopted for boiler-plates could seldom be met by one plate in ten, so that as a general rule it was arranged to take the tensile test-pieces from the upper part of the plate and the tests for quenching from the lower part.'*'

It has been attempted to attribute to blow-holes, as principal cause, the defects observed in forged or rolled pieces, such as rails or plates of hard or soft steel, but principally soft steel.

English equivalents have been added in parenthesis, the following being taken as a sufficiently accurate basis of calculation :

1 centimeter 0.3937 inches.

1 meter 3.28 feet.

1 kilogramme 2.2046 pounds avoirdupois.

1 square millimeter 0.00155 square inches.

1 kilo per sq. mm. 1422.3 pounds per square inch. — R. W. R.

The terms upper and lower part, top and bottom, head and foot, as applied in this paper to plates, mean the parts of a plate coming respectively from the top and bottom of the ingot

8Egbegation And 1T8 Consequences. 107

But at the meeting of the Iron and Steel Institute, held at Lon- don in the spring of 1881, Mr. Stubbs, of Manchester, put his finger upon the real evil, in mentioning the heterogeneous results of analy- ses made upon different parts of an ingot of large section. He gave figures. A test-piece taken about 60 centimeters (23.6 inches) from the head of the ingot of 2.3 m. (7.5 feet), gave by analysis very dif- ferent results from those of a test-piece taken 75 centimeters (29.6 inches) from the bottom.

Mn.

Si.

P.

Top,. .

. 0.92

Bottom, .

. 0.37

0.49S

Mr. Windsor Richards, in the discussion of Mr. Stubbs's commu- nication, said that he had often observed in test-pieces taken from different points of one plate, variations of 0.05 per cent, of carbon. But Mr. Snelus questions the results of Mr. Stubbs ; and when at the ffillowing meeting in September, 1881, he communicated his in- vestigation upon ingots poured in sand, which confirmed the analy- ses given by Mr. Stubbs, he insisted particularly upon the fact that the marked segregations of which he had been speaking were always insignificant in flat ingots for plates, cast in cat-iron moulds. This is sometimes true but not always. The examples which we give below do not in fact confirm the assertion of Mr. Snelus.

Order op Segregation op the Principal Elements op

Steel.

There is no absolute rule according to which the different metalloids and metals entering into the composition of industrial steels are liquated, but the average of a large number of operations has shown that the order of segregation is somewhat as follows : Carbon and phosphorus, sulphur, silicon and manganese. Copper liquates quickly when it is present in notable proportions. Homogeneity may be given to cupreous steels by adding a minute proportion of aluminum. It is equally difficult to obtain homogeneous masses of chrome- or wolfram-steel, but these alloys are beyond the province of the pres- ent paper. We may, however, mention as highly homogeneous the nickel-bearing steels, particularly those in which this metal is present in large proportion ; the reason being, undoubtedly, that it accelerates solidification.

108 segregation and itb consequencbb.

At What Moment does Segregation Take Pucb?

Segregation operates during the congelation of the cast piece, and exhibits itself in the parts which solidify last. Mr. Howe gives examples which prove the homogeneity of the metal in the ladle. In fact, as we have observed many limes for soft steel, the metal first emerging from the ladle does not sensibly differ from that which leaves it last. In tapping directly from Martin furnaces into ingot- moulds, there are often sensible variations between the metal of the first ingots and that of the last. Mr. H. H. Campbell* gives results of his tests upon samples taken during the first and the last cast, and for a dozen consecutive heats. Under tensile tests of cylindrical bolts, 8 inches long, the variations in resistance among these twenty- four samples reached 2960 pounds per square inch ; the elongation varying from 21 to 25 per cent. The variation in carbon did not exceed 0.045. Finally, as Mr. Howe concludes, even for carbon and silicon one cannot attribute the segregation observed in hard or soft steel ingots to an imperfect mixture of the final additions.

Moreover, while segregation is specially pronounced in an ingot in its central portion, and around the space of the piping, differences in chemical composition bearing principally upon the proportions of carbon and phosphorus in samples taken from different points of a horizontal section,! can also be demonstrated. It appears, then, that homogeneity is a quality almost impossible to realize in a block of steel.

Consequences op Segregation.

The mother-metal contained in the apparatus where it was formed, whether the Bessemer converter or the open-hearth furnace, is prac- tically homogeneous. In the ladle it can be obtained perfectly ho- mogeneous. But once solidified, in the form of the ingot or casting, the block of metal presents portions softer and .less impregnated with foreign elements than the mother-metal, from which, on the contrary, other parts differ entirely and do not even recall it.

In Stahl und Eiaen, Aug, 1891 , p. 643, analyses are reported which were made upon test-pieces taken from different parts of a Bessemer steel roll of seven tons, cast at La Louvidre. Its truncated conical form, reposing upon the larger base, caused the later solidification of the lower portion, the mass having been cast from the bottom. The mother metal had the following composition :

♦ T/an.,xiv., 359.

t Colonel Maitland, Proceedinga of the Insiitutum of Civil Engineen, London, vol. Ixxxix., 1887.

Segregation And Its Consequences. 109

P.

Mn.

The sample taken from the apper trunnion, which solidified first, gave by analysis :

C. 81. Mn.

0.215 0.338 0.910

That is to say, less carbon, but the same quantities of manganese and of silicon, the small differences in which may be attributed to errors of analysis.

The metal of the lower trunnion was harder than the mother- metal. It contained :

C. 81. Mn.

0.314 0.280 0.980

As for the metal last solidified, which formed the walls of the cavity of the piping, its composition no longer resembled that of the mother-metal, as the following figures show :

Si.

S.

P.

Mn.

On the interior of this recess, or pocket of the piping, there was formed a sort of cake, 50 millimeters thick, with a smooth surface, the chemical composition of which was still further removed from that of the mother-metal, being.

P.

Mn.

This is one of the most curious examples whicl% could have been given of the phenomena of liquation in a metal comparatively soft and pure.

In blocks of smaller weight and limited dimensions, subjected to the influence of solidification as rapid as casting within thick walls will permit, liquation may still be observed very distinctly. One may judge from the following analyses made from an ingot of Mar- tin steel, weighing about 450 kilos (992 pounds), and having a height of 1.1 meters (3.6 feet), and a section of 260 by 260 mm. (10.24 inches square).

1. Upper section :

r. S. P. Mn.

Border, 0.330 0.040 0.033 0.420

Center, 0.530 0.077 0.067 0.430

110 SfiGBEQATION AND ITB 00NBEQU£NC£8.

2. Lower section :

a 8. p. Mn.

Border, 0.280 0.029 0.016 0.890

Center, 0.290 0.030 0.038 0.390

3. Middle section :

c. ,8. p. Mn.

Border, 0.320 0.026 0.026 0.400

Center, 0.320 0.048 0.048 0.400

In the portions in which the mother-metal has solidified almost instantly in contact with tlie mould, there is preserved a practically homogeneous composition ; carbon, for example, varying only from 0.280 to 0.330. And one may also consider the block as suflBciently homogeneous in the lower half of its mass.

Segregation is less marked in ingots of extra soft metal cast in cast-iron moulds of considerable thickness. It is, however, still im- portant, and explains the difference often shown by the results of tests on pieces taken from different portions of a plate. Two sam- ples, taken from the sound part of a flat ingot, one on the outside and the other in the center, 20 centimeters (7.9 inches) from the upper

edge, gave :

c. 8. p. Mn.

Center,. . . . . 0.14 0.053 0.072 0.576

Exterior, 0.11 0.036 0.027 0.610

Manganese is the element most uniformly disseminated in hard or soft steel.

Mr. Herbert Eccles, at the meeting of the Iron and Steel Institute held in London in 1888,* reported some interesting experiments made upon bare cut from plates of soft steel, with the view of inves- tigating the cause of defects which they presented. A bar showing silky fracture had tensile strength of 26 tons per square inch and 25 per cent, elongation in 10 inches. Subjected for several weeks to the action of dilute hydrochloric acid, it was attacked with tolerable regu- larity without showing any special corrosion upon the medium sec- tion of fracture. Another bar taken from the same cast gave, under tensile test, a fracture with brilliant points, and the tensile strength was only 23J tons, with an elongation of 17 per cent. Subjected to the dilute acid, the granular part of the fracture was rapidly attacked, and after a certain time the sample appeared as if composed of two layere separated by an empty space.

Journal of the Iron and Sted In$tUuUf 1888, No. 1, p. 70.

P.

Md.

8EGBEGATION AND ITS CONSEQUENCES. Ill

Analysis gave different resalts for the central granular part and the silky part which covered it

0. 8. p. Mn.

Oninular party 0.160 0.078 0.112 0.570 Silky part, 0.115 0.030 0.038 0.576

Another bar of silky fractnre, which had given 26 tons per square inch of tensile strength with 31 per cent, of elongation in 10 inches, was but slightly more readily affected on the inside than on the out- side by the dilute acid. However, analysis revealed some difference between the metal of the exterior and that of the central portion.

Central portion, . Exterior portion,

It happens sometimes that plates of ingot-iron split along the edges. The texture is laminated. Such plates are the most defec- tive. A sample several square inches in size, taken from the shear- ing of such a sheet and submitted to the action of acidulated water separated into two layers, the part between being almost wholly eaten away. The following are the analyses of the central and exterior portions :

C. 8. p. Mn.

Central portion, 0.24 0.166 0.127 0.614 Exterior portion, 0.16 0.054 0.060 0.648

The analyses made upon plates of the famous boilers of the Lividia give considerable variations for samples taken from the same sheet, as the following percentages show:

Carbon 0.096 and 0.200 ; phosphorus 0.039 and 0.095 ; sili- con 0.036 and 0.177 ; manganese 0.331 and 0.371.

On the whole, it seems impossible to prevent the occurrence of segregation in cast-steel, hard or sodb, in large or in small mass. Even in tool-steel melted in the crucible and cast in small ingots of from 30 to 40 kilos (66 to 88 pounds), it is rare to find a bar from 2 to 3 meters (6.6 to 9.8 feet) long having perfect homogeneity.

Microscopic examination reveals differences between two samples taken near each other from the same hammered or rolled piece. We have thus to do with an inevitable evil, the effects of which, however, we may ameliorate in large part by localizing it.

Conclusions. The arrangement inaugurated at Terre-Noire, in 1870, has been aniversally adopted to diminish the most pronounced effects of seg-

112 8Eorbgation And Itb Oonsequencbb.

regation in obtaining the largest steel ingots for plates and for heavy artillery. Success has not been obtained for products of such impor- tance without many failures, and even to-day we are far from being content with the result accomplished, especially for armor-plates. Recourse has been had to hardening, and on good grounds ; but hardening cannot render uniform the resistance to shock of a block which has not homogeneous composition. We know the metal we must not use, but do we know the metal the chemical composition of which responds exactly to the requirements ?

In all cases the metal which forms the final armor-plate difiers from the mother-metal prepared in the furnace, and the probleii> thus set us, What is the mother-metal which ought to give a final product of certain composition ? is evidently not easy to solve so long as the solution depends upon many variables. One and the same mother-metal may furnish cast pieces of different composition.

For cannon of large caliber, if we reject, in addition to the part cast in sand and called the (sinking-head), one-third of the upper part of the ingot, we can obtain a tube practically homogene- ous in composition, because the central part is naturally removed by the boring of the tube. With extra-soft steels, destined for ship- or boiler-plates, the solution for practically perfect homogeneity lies in the obtaining of a metal more closely deserving its name of extra-soft metal. We must recognize the error which has been committed in large constructive industries, whether private or governmental, in requiring of a metal called extra-soft, and slightly or not at all sen- sible to annealing, tensile strength amounting to 42 or 48 kilo- grammes per square millimeter of section (68,770 pounds per square inch).

It is certainly right to require for boiler-plate a metal practically unaffected by hardening. In that case it is by elongation and by striction ("necking") — in which all the pure iron products are de- ficient— that we should define the mechanical properties of the metal, leaving tensile strength aside.*

The manganese steels have no striction ; neither have those which contain a high proportion of nickel. I refer to steels respectively carrying more than 10 per cent, of manganese or 20 to 25 per cent, of nickel. The cement or non-hardening carbon exists only in feeble proportion in these alloys, in which the iron, by a sim- ple quenching in oil, appears to be preserved almost wholly in condition /? An alloy of 25 per cent, of nickel with 0.80 of carbon, after quenching in oil, gave under tensile test, 80 kilos per sq. mm. (113,760 pounds per square inch) tensile strength, and 60 per cent, elongation in 10 centimeters (3.9 inches).

Segregation And Its Consequences. 113

We can sinoerely declare that in a long iDdostrial career, the ex- perience of which has a certain practical value (since we inaugu- rated in 1867 at Terre-Noire the manufacture of extra-soft steel with ferronianganese containing 80 per cent, of manganese) we have never been able to realize or to see others realize the desideratum of a homogeneous plate which successfully endured the harden- ing test with the tensile strength of 42 kilos (59,736 pounds per square inch) heretofore required for boiler-metal. The lengthwise sample, cut from the bottom of the plate and satisfying a rigorous quenching-testy rarely gave a maximum of 40 kilos (56,892 pounds). The lengthwise sample from the top of the plate was mediocre, and often absolutely bad, under the hardening test. And as to cross- wise samples, while the bottom one would sometimes bend double, with a metal giving more than 40 kilos tensile strength, the top one was always defective.

We can even cite an instance (though an exceptional one) in which a boiler-plate 22 mm. thick (0.87 inch) made from the lower part of the ingot, showed an extreme lack of homogeneity :

P.

Mn.

Average sampley

Samples from the middle

portion, the thickness be-

ing reduced by planing to

10 mm. (0.39 inch),

014O

Another example, furnished by a plate 30 mm. (1.18 inche thick,, exhibits, on the contrary, a satisfactory homogeneity m the foot of the plate, while the head presents a very heterogeneous composition. Lengthwise and crosswise samples from the head gave by analysis:.

c. 8. p. Mn.

Exterior, lengthwise, . . 0.24 0.025 0.050- 0.)60

Exterior, crosswise, . . 0.24 0.017 0.052 0.150

Interior, lengthwise, . . 0.32 0.061 0.100* 0.170

Interior, crosswise, . . 0.40 0.070 0.088 0-140

The corresponding samples from the bottom of the plate give- practically uniform results :

P.

Mn.

Exterior, lengthwise,

. 0.250

Exterior, crosswise, .

. 0.250

Interior, lengthwise.

. 0.250

Interior, crosswise, .

. 0.260

Vol. Xxii.— 8

Digitized by V

114 Segregation And Its Consequences.

Under tensile test, for a length of 100 mm. (3.9 inches), the strength and elongation of these specimens corresponded with their chemical composition.

Specimens from above :

Lengthwise. CrosBwiBe.

Tensile strength, kilos per sq. mm., . 47 and 46 45 and 47.7

Tensile strength, Ibe. per sq. in., . . 66,S48 and 65,426 64,003 and 67,844 Elongalion, per cent, 27 and 82 13 and 21.5

Specimens from below :

Lengthwise. Crosswiae.

Tensile strength, kilos per sq. mm., . 41.7 and 42 42 and 41.5

Tensile strength, lbs. per sq. in., . . 59,310 and 59,636 59,736 and 59 026

Elongation, per cent, 32.5 and 33 33 and 33.5

In spite of the comparative purity of the metal, the pieces taken for hardening from the bottom of the plate did not sufficiently meet the required conditions.

To what distance from the foot of the plate does this state of homo- geneity extend? Can we be assured of securing it with practical certainty by rejecting one-third or one-half of the upper part of the ingot? This necessarily depends upon the rapidity with which the metal solidifies in the mould. Flat ingot-moulds with thick walls have long been used to obtain a rapid solidification, which, however, must always be a function of the temperature of casting and the transverse section of the ingot, varying according to its weight.

Aluminum Steel.

In our opinion, the injurious consequences of segregation must be suppressed by reducing, as far as possible, the elements subject to liquation.

Upon the basic or neutral open-hearth, and starting with an ini- tial bath of approximately pure materials, it is easy to obtain a metal containing not more than 0.1 per cent, of carbon ; 0.02 of phos- phorus, and traces of sulphur, with 0.10 of manganese. By adding 0.1 per cent, of aluminum, the metal can be cast quietly, and with- out altering its composition. Consequently, if from an ingot so cast and destined for boiler-plate, one-fourth to one-third of the upper part (in which the carbon and phosphorus may reach respectively 0.12 and 0.03, for exampfe,) be cut off, the remainder will be a block of approximately perfect homogeneity.

Operating in this way, we have obtained in a large establishment

8|Greoation And Its Consequences. 115

in the northeast of England* ingots from 2 to 3 tons, 18 inches square, which being first treated by the hydraulic press and subse- quently rolled into billets, were utih'zed almost without waste in the manufacture of wire for telegraphic cables.

We think that the specifications prepared by the late engineer, M. Cornut, for the Association des proprietaires (Tappareils d vapeur of the north of France, express the conditions most suitable for boiler-plate.

The elongation lengthwise of the annealed plate ought never to be less than 30 per cent, in 20 centimeters (7.89 inches); as regards tensile strength, 40 kilos per square mm. (56,892 per square inch) as a maximum seems to us too high.

For ship-plates, whatever may be the importance of having a much stronger metal in order to diminish thickness and weight, it IS our opinion that too much is sacrificed to this consideration to the neglect of (1) the more easy and certain manipulation of a more malleable metal, and (2) the action of sea-water, which may be a fifth or a fourth more rapid upon a metal with 45 kilos (64,003 pounds) tensile strength than upon a softer and more homogeneous metal with only 38 kilos (54,047 pounds) tensile strength.

In the construction of bridges, our preference for the use of an extra-soft metal runs counter to the general desire of having for this purpose a metal of high elastic limit. But, nevertheless, it has not been wished hitherto to secure this precious mechanical quality in bridge- metal by increasing the hardness beyond a certain very mode- rate limit. Why not use a new alloy? Chrome-steel has already been tested ; and when to a pure metal, like boiler-metal, 0.2 to 0.4 per cent, of chromium has been added, homogeneous blocks have been obtained, and the limit of elasticity has been raised notably — up to two-thirds of the breaking-strain — without sensibly altering the elongation.

In the direction of alloys there may be found various advantage- ous solutions of the problems involved in the manufacture of metals destined for civil constructions.

As a final conclusion of this summary survey, we would call at- tention to the fact that tensile tests and mechanical tests in general may determine a priori the intrinsic qualities of a mass of fluid metal, but not those of a solid metallic block, whether before or after work has been done upon it.

The Port Clarence Steel Works, owned by my friend, Sir Lowthian Bell, Bart. The figures refer to 1890.

Segregation And Its Cx)Nsequences.

Appendix I. — Physical Tests.

Ordinary soft Martin steel plate, 14 mm.

Boiler plate, Martin steel. 17 mm.

(0.56 in.) thick.

(0.68 in.) thick.

o

I-- 1 lid

Remarks on frac- ture, etc.

&p\

s .

.1 1 i

J ;

J6.0'

Granular; large

1 flaw in centre. !

1 large white lines.

n.o

44. 9' Ditto, ditto.

1P,fl 48.5 Ditto, ditto.

46. Oi Finely granular; 1 flaw at one com'r

3 2l.fi i43.0 Ditto, ditto.

4 '..,.. 22.0 143.5 Ditto, ditto.

is.ol

46.6. Ditto, serious flaw

20,0 44.8 Ditto, ditto.

d

on one edge.

10.0L**.>47.3 Irregular: balf-

It. 3 46.5

Ditto, ditto.

grain, half-flber.

13.0; 47.2

Ditto; serious flaw

n

center hard.

on one corner.

15.0 44.5 Irregular; two

'20.0

Normal ; central

1 i strong white lines.

fi

part appears very

aa.O '42.5 Fine white lines.

Hard. "'

„ 17,5 '45.5 Piped.

44.2, Beau tiftil necking; 1 some blow-holes.

?

124.5 42.6 Ditto.

lii.b 4S.2i Many white lines

21.0 47.2iPiped; some hard

1 at center.

1 1

spots in center.

40.7 Large flaw incen-

44.7 Normal; fine

1 ter ; no necking.

white line.

IS.ol 43.9 Normal. 1

11 ;i9.5' '44.0

Piped; large

white lines.

I23,& 41 .6 Fine white lines. 1

i<l

'

23.0. '41.7 Piped ; some white

S'

12 14.5! 43.3

CkMtrsely granu-

points. 23.5 |i,... 42.3 White line on

1 one edge. 'iO.O Normal; fine

o.

lar; serious flaw on

one edge.

"i ?

13

45.2 NormiU. .

17 25.0

44.0 Ditto.

1 1 impure lines.

15 121.5

42.8 Ditto.

s

"-

18 I19.5 „ 41.4 Many Impure lines.

16

44.5 Piped; much

19 ,13.0 :46.9 Hachure of

i necking.

1 1 white lines.

Ditto; some

15.o! 43.7 Normal ; some

1 white lines.

1 impure lines.

18.5 44.0

Ditto; gran ul'r spot;

27.0' 40.1 Ditto, ditto.

25.0 39.4 Piped.

I several large

u

white lines.

|18.0, 40.4 Normal.

1

♦19

17.5 44.6*

Normal; some

!31.0 38.7 Piped.

11

1 1

1 white lines. J

25 23.5' 38.6 Ditto.

Si

20.0 44.9 iMuch neckinjf ;

26 21.0

39.2 Piped ; some

1 1

1 some blow-holes.

1 1 white lines..

'22.0'

42.5 Normal ; do.

40.7, Much necking;

some blow-holes; 'slight flaw at cent'r

r

Av.jl7.7 22.0 42.8'42.3|

23

n.5 41 .5 Piped; some

1 1 blow-holea.

?4

W.O 40.8!Much necking;

1 1 some blow-holes.

:t

1 1 kilo per sq. mm. — 1422.8 lbs. per sq. In.

41.2 1 Much necking:

1 fracture without

flaw.

21,0

Normal; large white line atcent'r. J

Av.

20.9 44.1

1 !

Planed.

Specific Gravity Op Gold In Gold-Silver Alloys. 117

Appendix II. — Chemical Analyses.

Ordinarv soft Martin steel-plate, 14 mm. (0.56 In.) thick.

No. Part.

Si.

8. P. 1 Mn.

H

Os

Mean 0.260, 0.066 |O.M2,0.065,0.160

Interior. 0.2901 0.019 0.083 0.1f9 0.150

f Interior. 0.270' 0.140 i0.088 0.078 0.160 t Exterior ,0.240! 0.028 0.020 0.060 0.150

Mean 0.270 0.019 ,0.038 0.093 0.170

Interior. 0.270 0.028 ;0.040 0.078 0.150

; 0.046 0.077 0.150

i Exterior 0.260 0.037 ;0.022 0.065 0.150 '/Slagl' I. I

f Interior. 0.260

t Exterior 0.230| 0.093

0 10.028/

r Interior. 0.270 1>0.043,

1 Exterior ,0.220 0.065 0.022 0.062 0.150

0.026 0.078 0.160 0.028 0.060 0.150

I ! I

0.035 0.057 0.160

[0.080 0.076 0.160 0.027 0.068 0.160

15 Mean 0.240 0.064

, f Interior. 0.2ro! 0.075 ' ' Slajf

f Interior. 0.260 \0.(M6 Exterior 0.'230 f Slag

Mean ,0.260

I I

0.054 0.073 0.1ft 0.042 0.060 o.ia

0.0!44J

0.170 0.048 0.085 0.150

f Interior. 0.200 0.028 0.030 0.067 0 160

1 Exterior 0.240 f Slag) 0.018 0.058 0.160

t0.06of,

Mean 0.220: 0.046 |0.032 0.075 0.160

I I I !

Mean 0.250 0.075 0.0S8 0.070 0.150

f Interior. 0.22o' 0.093 0.018 0 063 0 150 t, Exterior 0.2001 0.046 ,0.019 0.057 0.150

0.028 0.074 0.150 022 0.068 0.160

Interior. 0.230' 0.037 ,

f Slag) 0.

to.oesf I

I i i Exterior 0.210 0.130 0.015 0.049 0.150

I Mean 0.243 0.064 0.0310.070,0.155

Exterior ,0.230

Interior. '0.250 0.074 0.020 0.060 0.150

2

i.g.

also

mi

O o

2

Boiler-plate Martin steel, 17 mm. (0.68 in.) thick.

No.

Part. C. Si.

S.

Interior. 0.: Exterior 0.1

Mu.

Mean 0.250 0.032 0.043 0.080 0.100

320 0.096 0.070 0.077 lO.lOO 180 0.056 ,0.030 i0.049 0.100

Interior. 0, 'Exterior 0

028 0.023 0, 210 0.028 0.

048 ;0.065 022 0.057

Mean 0.280 0.028 0.058 ,0.073

Interior. 0, Exterior 0,

Interior. 0. Exterior 0.

Interior. 0, Exterior 0,

250 0.065 0.030 0.067 O.lOOi

,220 0.032 0.020 0.063 0.100

I I '

280 0.032 0.010 0.078 0.110

210 0.046 0.026 0.065 0.090

240 0.046 0, 210 0.046 0.

042 0.091 iO.lOO. 028 0 060 ;0.090

Mean 0.240 0.028 0.028 0.065 iO.lOO

Interior. 0, Exterior 0

Interior. 0, ExtHnrO,

ii.ir.O f ijjrO

.250 0.028 0. .200 0.046 0.

230tr'ce 0 200 0.019 0

,240 0.019 0 .220 0.016 0

088 0.078 |0.10(

020 0.049 10.090

024 '0.070 0.110

022 0.047 0.100

028 0.117 0.110

026 0.057 'O.Ioo

19 Mtan 0.250 0.028 0.040 0.068 0.090)

25|

I"

Inierinr. 0, jEsK'HorO,

Mean 'o.

iMmn.,.- 0.

Interior. 0 Exterior 0

InirriorK 0.

E>.[ijrior 0

240 0.(6 0 2000.076 0

230 0.042 0,

220 0.028 0

.200 0.042 0 .210 0.023 ;0

210 0.070 0 180 0.046 ,0

031 10.077 lO.llOl 028 1O.O68 O.lOOi

028 0.062 0.090 030 ,0.057 lO.lOO

018 0.044 0.100

020 ;o.054 0.090!

028 0.065 iO.llOl ,082 ,0.052 0.100

Ml-u n 10.229 0.0403 0.0305,0.0652 0.099

JSOTE ON EXPERIMENTS ON THE SPECIFIC OBAVITT OF GOLD CONTAINED IN GOLD-SILVER ALLOTS.

BY HENRY LOUIS, SINGAPORE, STRAITS SETTLEMENTS. (Chicago Meeting, being part of the International Engineering Congress, August, 1893.)

Four alloys of gold and silver were prepared, containing the metals in the following proportions by weight : 1. Gold:Silver:: 1:2.62.

118 SPECIFIC GRAVITY OF GOLD IN GOLD-blLVER ALLOYS.

2. Gold: Silver:: 1:3.15.

3. Gold: Silver:: 1:4.11.

4. Gold: Silver:: 1:5.17.

These alloys were rolled, with repeated annealing, into thin strips, rather thinner than is usual for assay-cornets. The strips were thoroughly annealed, cut into pieces and dropped into hot parting- acid. They were boiled twice with No. 1 parting-acid for one hour and ten minutes altogether, and then with No. 2 parting-acid for thirty minutes, washed thoroughly with boiling distilled water, and left under water under the exhausted receiver of an air-pump for thirty hours.

Their specific gravities were then taken with the following results, the temperature being 15° C.

No. of Ex- periment.

WeiRbt of

Gold

in air.

"GrainsT"

Weight of

Gold in water.

Weight of

equal bulk

ol water.

Specific Gravity.

Grains.

The quantities operated on were very small, and the balance and weights by no means first-rate, so that the results are not likely to be very accurate. The first three results are sufficiently close to- gether, but the very high figure obtained in the fourth experiment must, in all likelihood, be due to some error. Hence, it will be safest to average the results of the first three experiments, and, pending a more complete and accurate investigation, which I hope to make of this subject, to take the specific gravity of the gold resi- due left on dissolving the silver-gold alloys at about 20.3 at 15° C. Apparently, the varying proportions of the two metals in the alloys do not affect the result.

Gold thus left, on dissolving out the silver, is highly spongy, and, on annealing, it undergoes a very evident shrinkage. The result now obtained proves that, on annealing, the molecule of gold does not contract, but, on the contrary, expands, the shortage of bulk being due to a diminution of the fthysical interspaces between the particles of gold, while the interatomic distances must increase, the diminution being, of course, greatly in excess of the increase.

Specific Gravity Of Gold In Gold-Silver Alloys. 119

With regard to the figures here given, it may be noted that Rose obtained, for the specific gravity of precipitated gold, results vary- ing between 19.49 and 20.72. His figures, taken together with mine, seem to point clearly to the existence of a heavy allotropic modification of gold. It is of course possible, that the brown amor- phous gold obtained, either by precipitation or by the removal of the silver from a silver-gold alloy, may be a mixture of ordinary and allotropic heavy gold, in varying proportions ; there can, how- ever, be little doubt of the allotropism and of the further fact that gold exists in the alloy in this allotropic form.

A further partial and indirect justification for this latter state- ment may be found in the experiments of Matthiessen on the specific gravity of alloys {Phil. IVans., 1860, p. 177), in which he shows that the observed specific gravities of gold-silver alloys exceed the calculated ones (calculated upon the specific gravity of ordinary gold) in the ratio, approximately, of 1 : 0.997. This would appear to suggest that the gold in these alloys is in a heavier state than ordinary normal gold, but any conclusions based on these observa- tions are liable to modification, seeing that it is not yet known in what condition the silver in these same alloys exists.

It is possible that these results may afford a clue to the explana- tion of the widely different specific gravities found by different ob- servers of native gold specimens (containing silver), of approximately the same composition, but from different localities.

In order to complete our knowledge of this subject it will be nec- essary to study the character of the gold by dissolving out the alloy- ing metal from a series of alloys of gold with different metals; and I hope to be able to undertake this investigation before long.

The above experiments were conducted, by the kind permission of Professor W. Chandler Roberts-Austen, in the research-laboratory of the metallurgical department of the Science Schools, South Ken- sington.

120 The Detection And Measurement Of Fire-Damp.

The Detection And Mea8Ubembnt Of Fibedamp

In Mines.

BT G. CHESNEAU, PARIS, FRANCE, ENGINEER OF THE CORPS 07 MINES, PROFESSOR AT THE NATIONAL SCHOOL OF MINES IN PARIS, AND SECRETARY OF THE FRENCH FIRE-DAMP COMMISSION.*

(Chicago Meeting, being part of the International Engineering Congress, August, 1893.)

Introduction.

Two great discoveries of this century have diminished the dangers of fiery coal-mines, — the safety-lamp, conceived in 1815 by Sir Humphrey Davy and successively improved by many engineers, such as Clanny, Marsaut, Mueseler, Fumat, etc., and safety-explosives, the appearance of which is comparatively recent, but which, thanks to the persevering labors of several technical commissions (especially the French commission on explosives), and of the manufacturers and experimental users of these powders, appear likely to play, in the protection of the miner, a part as important as that of the safety- lamp itself.

But while it is true that by the Use of a safety-lamp of approved form, and the exclusive employment, with all prescribed precautions, of these new explosives, complete security should be assured, even in highly fiery atmospheres, experience has shown that the imprudence of workmen, or the bad condition of a lamp, may nullify these pro- tective means, so that it ought to be the constant care of the engi- neers of fiery mines to dilute the fire-damp by good ventilation so far as to maintain always and in all parts of a mine a proportion of this gas below the lower limit of explosiveness belonging to a mix- ture of air and methane. Good ventilation still remains the best preventive of explosions from fire-damp, and since, even with a rational and well-controlled mine-ventilation, the volumes of air traversing galleries and rooms are liable to vary within wide limits by reason of open doors, falls of rock, or other accidental causes, it follows that but a few thousandths of fire-damp should be led in the normal atmosphere of the mine-workings, so that a considerable

Translated by the Secretary.

The Detection And Measurement Of Fire-Damp. 121

dimiDutioti (amounting, for example, to one-third or one-fourth) in the volume of circulating air may not for a moment raise the pro- portion of fire-damp to 6 per cent., which makes the mixture explo- sive. A normal proportion of 1 per cent, in the workings should therefore be deemed very high, and 2 per cent, ought to be permitted in very exceptional cases only.

Mine-inspectors should therefore be constantly advised of the per- centage of fire-damp in all air-ways and rooms in order to be able to distribute rationally among the different passages the air furnished by the fan, to reinforce the current as needed in chambers where the percentage of fire-damp is increasing, and to stop work in those places where this measure is not efficacious. It is only such daily control of the amount of fire-damp in chambers and passages that will permit a good ventilation, constantly adjusted to the varying quantities of gas liberated, either in old workings or from freshly- broken coal.

But while the illuminants and explosives employed in mines, as well as the means of ventilation, have been the subjects of great im- provements, the detection and measurement of fire-damp are still, in many (even important) mines, in a very rudimentary condition. It is often deemed sufficient to make inspection for the presence of the gas with oil-lamps, which do not indicate with practical certainty less than 2.5 per cent, of it, — a proportion much greater than can be safely tolerated in the workings. In some cases more or less frequent tests are made upon samples of the mine-atmosphere, taken in flasks, and subsequently analyzed above ground in the laboratory. If daily tests are thus made from the same points in the general re- turn air-ways of a mine, this procedure may give valuable indica- tions of the total liberation of fire-damp. But it furnishes no infor- mation as to the relative condition of all the fiery parts, and some chambers may contain a dangerous abnormal percentage, not revealed by any alarming increase shown in chemical analyses of the air of a general return-current, supplied by a large number of chambers. This evil is the more serious because, by reason of its smaller den- sity (specific gravity, 0.558), the fire-damp tends to accumulate along the roof of the workings in which it is liberated, and mixes but slowly with the air-current, so that, as may frequently be observed, a distance of a few centimeters only may separate an explosive layer from one of relatively pure air. A sample of air taken from the latter may thus give a very erroneous idea as to the safety of a chamber.

122 The Detection And Measurement Of Fire-Damp.

Such laboratory-determinations, therefore, do not take the place of a detective apparatus which can be carried to all parts of a mine, and will show immediately the composition of the surrounding at- mosphere, furnishing without delay to the engineer the information which would be given by hundreds of analyses, occupying several days. Nevertheless, the laboratory-determinations are very valuable in checking from time to time the reports of the portable fire-damp indicator and guarding against errors of observation.

More precisely, the handling of a fiery mine as rards the detec- tion and measurement of fire-damp required for proi)er ventilation, should comprise, on the one hand, the use underground of indicators which can be observed not only by the engineers but by their sub- ordinate inspectors, and, on the other hand, accurate laboratory-ap- paratus, open to no systematic error, and available to check from time to time by a relatively small number of analyses, the observa- tions made with the indicator.

During the recent labors of the French Commission on fire-damp upon different types of lamps and the determination of fire-damp, I have been led to study the conditions in practice of different indica- tors used in various mines, and, while determining the causes of error or danger which they involve, to design an indicator which is easily portable in the mine and possesses also greater accuracy and safety than the forms hitherto known. At the same time, Prof. Le Chatelier, of the National School of Mines, a member of the same commission, has introduced important improvements and simplifica- tions in laboratory apparatus for the determination of fire-damp, and these improvements have been tested by sufficiently prolonged prac- tice to warrant the consideration of the procedure above prescribed as a well-established rule.

Before describing these new means and methods, I deem it well to sketch briefly what is known concerning the detection and meas- urement of fire-damp, and the experiments which led me to my de- sign of an indicator. This rapid survey may be of use to engineers seeking to improve the apparatus described by relieving them from the need of fruitless researches.

I shall describe successively, in the two following chapters, the determinations of the laboratory and the portable indicators used in the mine.

The Detection And Measurement Of Fire-Damp. 123

Chapter I.

PRACTICAL METHODS FOR THE DETERMINATION OF FIRE-DAMP IN THE LABORATORY.

To say nothing of the general methods of organic analysis, prop- erly so-called, which may be applied to fire-damp, but are too com- plicated for a rapid daily operation, the proportion of methane or formene contained in air may be practically determined by two dif- ferent processes, the combustion of the fire-damp itself, or the obser- vation of the limits of inflammability of the combustible gases.

1. Determination of Fire-Damp by Combustion.

In a measured volume of fiery air, ascertained by any suitable method to contain less than the quantity of fire-damp which its oxygen would suffice to burn completely*, the combustion of the fire-damp is eflTeoted by means of an incandescent metallic spiral. The methane, even when highly diluted, burns according to the fol- lowing formula :

CH, -f 2O2 COj + 2UJ0. Volumes 2 4 2 4

If the vapor of water remains in the gaseous state there will be no change in the volume, since one molecule of methane and two of oxygen, that is, three molecules of the mixture, yield after combus- tion three molecules of the burnt products, one of carbonic acid and two of water. But at ordinary temperatures, this water is con- densed, and the combustion of each molecule of methane causes a diminution of two molecules in the total gaseous volume. The combustion of 1 per cent, of fire-damp diminishes therefore by 2 per cent, either the volume or the pressure of the mixture, according as the operation is performed at constant pressure or at constant volume. If the carbonic acid produced should be absorbed, the diminution would be triple instead of double.

M. Coquillion (who first pointed out the possibility of utilizing for quantitative determination the combustion of fire-damp in con- tact with a spiral of palladium brought to bright-red heat by an electric current) constructed in 1887 an apparatus based on this prin- ciple, which permitted the proportion of fire-damp to be determined

For a sample taken in the mine, it is easy to prove with a lamp that this limit has not been passed. #

124 The Detection And Measurement Op Fire-Damp.

by the diminution in volume of the air measured at constant pres- sure before and after combustion.

The indications fiven by this apparatus were somewhat uncertain, for several reasons.* The gas being measured over water, variable amounts of it were dissolved therein. There were no refrigerating envelopes, such as are indispensable to maintain a constant tempera- ture in the gas of which the change of volume is to be measured. Finally, the arrangement of the burner did not secure a sufficiently prolonged contact between the glowing spiral and the gases, and combustion was always incomplete.

The studies of Dr. Schondorf, director of the laboratory of re- search of the Prussian Fire-Damp Commission, showed that it is possible, with a spiral of incandescent platinum, as well as with one of palladium, to obtain complete combustion in a mixture of air and methane, however small the proportion of the latter.

M. L. Poussigue, director of the collieries of Ronchamp, France, has constructed a modified Coquillion apparatus, giving much more precise results than the original. His apparatusf consists of a grad- uated measurer connected with a burner by a tube with a stop-cock. In the burner is a spiral of platinum which can be rendered incan- descent by an electric accumulator. Finally, this burner is itself connected with a vessel containing a concentrated solution of caustic potash for the absorption of the carbonic acid. The measurer, which may contain about 200 cub. cent. (12.2 cubic inches) of the air to be analyzed, is surrounded with a jacket cooled by water-circulation, which serves to maintain the temperature of the gas at a constant point, as shown by a thermometer ; and the pressure is observed by means of a mercury manometer, open to the air. The whole appa- ratus is of glass.

The air being introduced into the measurer, its volume and pres- sure are noted. It is then caused to pass through the burner and the solution of caustic potash (which absorbs the carbonic acid of the air) and conducted back to the measurer, where its volume and pres- sure are again noted ; after which it is passed a second time through the burner, the platinum spiral having now been brought to incan- descence. The carbonic acid produced by the combustion of the methane is absorbed by the potash, and the gas is once more returned to the measurer, where volume and pressure are finally observed.

Annales des Mines, Paris, 1892, 9e srie, tome ii., p. 469, "Sur le Dosage du Grisou " par H. Le Chatelier, Ingnienr en Chef des Mines, t Bulletin de la SiMHde V Industrie Minirale, 1892, tome vi., p. 249.

The Detection And Measurement Op Fire-Damp. 125

All the observed volumes are reduced by calculation to the initial pressure, and the diminution of calculated volume, after the first absorption, is equal to three times the volume of methane contained in the air tested.

This apparatus is open to the objection that the gases remain a long time in contact with the potash solution which may absorb from the air or return to it a certain amount of gas.

Moreover, the measurer being pretty large, it takes a compara- tively long time to secure the complete combustion of the methane.

In the apparatus constructed by M. Le Chatelier,* special care

Fig. I.

Pointcd Scrcw-Tap

has been taken to remove, or reduce to the smallest possible impor- tance, all the sources of error inherent in the inexperience of the operator, and particularly to avoid all joints through which the gas could escape. All rubber-tubing has been discarded ; only a single stop-cock, required for the introduction of the sample, has been re- tained; and this cock is a pointed screw-tap (Fig. 1), not a plain cock, and therefore gives almost absolute guaranty of tightness.

The Chatelier apparatus, as shown in Fig 2, consists of a glass cylinder. A, 20 mm. (0.8 inch) inside diameter, capped with an iron top, in which is the cock, R, for the introduction of the gaseous mix- ture to be analyzed, and the subsequent escape of the same, at the end of the test. The bottom of the glass cylinder is closed with mercury, the level of which can be varied at will through the move- ment of a body of mercury contained in a flask, F, connected by a rubber tube with the lower part of the glass cylinder, the bottom of which is drawn out to fit this connection. The cock, R, being open, gas may be made to enter or leave the cylinder. A, by simply lower- ing or raising the flask, F,and the cylinder thus constitutes a closed

AnnaUt de$ Mines. srie, tome ii., p. 469, " Sur le Dosage du Grisou,*' par H. Le Chatelier, from which the description here given is taken almost literally.

126 The Detection And Measurement Op Fire-Damp.

receptacle, serving at once as a measurer and as a combustion-cham- ber.

The mercury-reservoir communicates also with an open tube, T, constituting a manometer under atmospheric pressure, the height of the mercury in which measures the pressure of the confined gas.

Scale KotfaUftise. Le Chatelier's CJombastion-Apparatus for Determining Fire-Damp.

The combustion of the mixture is secured by the incandescence of a platinum spiral, 8, connected with two platinum wires, i.t'y which pass through the iron cap, in insulating sheaths, and to the conduc- tors of an electric battery. One of these platinum wires terminates below, very near the spiral S, with a slender point. Before and after the combustion, the mercury is levelled in the cylinder to this point, so that observations are made always at constant volume, and only the height of the mercury in the manometric tube, T, is noted. The volume of air in the measurer is about 22 cu. cent (1.34 cu. in.).

The cylinder. A, is wholly immersed in a glass reservoir, B, filled with water, the mass of which regulates the temperature, reducing its variations and permitting its easy measurement. The water in

The Detection And Measurement Of Fire-Damp. 127

this reservoir, moreover completes the tightness of the joint of the iron cap, which is screwed on a ring terminating A at the top, so that it can be removed when the interior of A is to be cleaned.

In making a test, the cock, R, being open, the mercury is first levelled to the point above described, and the division, h, at which the mercury stands in the manometer, and which will be the zero of the manometer-scale, is noted. The air is then expelled from A by elevating the mercury-reservoir (which is placed upon the support, D) ; after which the cock, R (by means of a rubber tube so short and small that the volume of its contents may be disregarded in comparison with that of the gas introduced into A), is connected with the receptacle holding the air to be analyzed. The mercury- flask is then lowered, so as to draw in the sample of air, which is at the same time expelled from its former receptacle by the introduc- tion of water (preferably salt-water, which absorbs less gas). The operation is now reversed, and the air in A is forced back again into the receptacle containing the air to be analyzed. This is done for the purpose of mixing with the whole mass of air in the receptacle the small amount of pure air which was in the short rubber tube, and thus reducing to a minimum which can be disregarded, its influ- ence upon the test. The measurer. A, is now filled anew ; the mer- cury being levelled to the neighborhood of the point previously de- scribed, but without attempting an exact adjustment. Five minutes are allowed for the establishment of an equilibrium of temperature, which is read to within one tenth of a degree C by means of a thermometer immersed in the water of B. The pressure of the gas- eons mass is then observed, for which purpose the mercury must be exactly levelled to the point. This is done before a window or a light which strongly illuminates the surface of the mercury. At the mo- ment of contact, the height, A, of the mercury in the manometer is noted ; and the pressure, P, of the gaseous mass will be (H being atmospheric pressure, and h the former manometer-reading, as above explained) :

P H + A' — A.

The platinum spiral is then made incandescent by an electric car- rent. The arrangement and the temperature of this wire have great influence on the rapidity of combustion. A wire coiled in a spiral is much more efiective than a straight one of the same length, be- cause the heating efiect at a given instant is exerted upon a smaller

128 The Detection And Measurement Of Fibe-Damp.

qiiaDtilj of gas, and necessarily raises it to a much higher tempera- ture. The spiral should be in the lower part of the combustion- chamber, so that the circulation produced by the heating may bring all parts of the gas into contact with it. Combustion is evidently the more rapid, the hotter the spiral. Its temperature should be be- tween 1300° and 1600° C. Outside of these limits, either combus- tion will be too slow or there will be danger of fusing the spiral. Some practice of the eye is necessary to the proper regulation of the temperature. A beginner will always fuse the spiral a few times. For approximate regulation, the two following indications may serve: (1) The spiral should be sufficiently luminous to seem to emit rays, but (2) not so brilliant as to blind the eye to the separate coils of which it is composed. Good results are obtained by making the spiral of six turns, 3 mm. (0.12 in.) in diameter, 1 mm. (0.04 in.) apart, and of wire 0.3 mm. (0.012 in ) in diameter. If a wire be used made of platinum-copper alloy, containing 3 per cent, of cop- per (which has the advantage of being highly refractory), a suitable incandescence of the spiral may be efiected with a source of electric- ity capable of giving a maximum current of 6 amperes, with an electromotive force of 12 volts.

If a battery be used, it is necessary to interpose a rheostat of vari- able resistance, which will permit the control and the progressive increase of current-intensity. More convenient is a hand-dynamo, with which the current can be easily controlled by more or less rapid revolution of the crank. The machine is, moreover, always readv to work ; and the tedious installation of batteries and accumulators is dispensed with.

Fifteen seconds of incandescence of the spiral suffices to excite complete combustion in the gas ; but it is well to repeat incandes- cence twice, after the interval of some seconds, because the expan- sion of the gas by heat causes the descent below the spiral of a por- tion which may thus escape combustion, whereas the momentary cooling causes all the gas to pass again above the spiral. This reas- cension of the gas may be aided by elevating the mercury reservoir ; but care must be taken at the moment of the passage of the current, not to permit both conductors to touch the mercury, which would divert the whole current and thus render the incandescence of the spiral impossible.

Combustion being complete, it is necessary to wait 10 minutes for the re-establishment of an equilibrium of temperature. This is an absolutely indispensable precaution ; and it is necessary to wait even

The Detection And Measurement Op Fire-Damp. 129

longer, if the incandescence of the spiral has lasted more than thirty seconds.

As at the beginning, the temperature, and the height, A", of the mercury in the manometer, are noted.

The proportion of 6re-damp is calculated from the diminution, A' -A", of the pressure of the gaseous mass at constant volume, cor- rected for the difference in temperature. This is the reverse of the methods of Coquillion and Poussigue, by which the percentage of fire-damp is deduced from the change in volume of the gaseous mass, either kept at constant pressure, or reduced to constant pres- sure by calculation.

With Le Chatelier's apparatus the proportion of fire-damp is cal- culated from the temperatures and manometric readings as follows :

The pressure P, the volume F, the absolute temperature t + 273° C, and the number of molecules iV, of a gaseous mass, have the known relation :

Pv=Rnt . . . . (1)

in which £ is a constant, depending upon the units of measurement adopted. After combustion, the volume remaining constant, this relation may be expressed by :

P'V=Rnt (2).

Whence, subtracting (2) from (1) and dividing the remainder by (1) we have:

or, substituting for P and P' these values :

P=ir -I- A'-A

P' ir-|-A'' — A

and introducing the proportion x of fire-damp, which is, in parts

It 18 necessary to avoid prolonging the operations of one analysis beyond one boar. By the oxidizing action of the humid air upon the iron cap of A, oxygen is absorbed, and the yolume of the air in the measnrer is consequently reduced. But. this effect is appreciable only after some hours.

VOL. XXn.— 9

130 The Detection And Measurement Of Fire-Damp.

by volume, equal to one-half the change of volume divided by the initial volume, or:

we have

' <H+h'—h~t + 273 V + 273

1 / h' — h'' i — t' , H-273

But in practice the tAnperatures t and V do not differ more than

i -f- 273 one degree, and the term - may be rarded as unity without

involving in the determination of the proportion of 6re-damp an error of more than of the result, which is quite insignificant. We may, therefore, take for the calculation of x, the formula :

' — i\H + h'—h t+273'

In rftost cases, moreover, the second term of the parenthesis may also be neglected. With this apparatus, the manipulation of which is not difficult, numerous determinations have proved that the pro- portion of fire-damp can be measured to one-thousandth.* It is subject to no special source of error, since the air in the measurer remains in contact with mercury, which has, under the conditions of the test, no physical or chemical action upon it. Besides, the appa- ratus requires for analysis but a very small volume of air, and hence permits the collection of mine-samples in small flasks, 100 cubic cen- timeters (6.1 cubic inches), for instance, being enough for two or three analyses :

2. Determination of Fire-Damp by the Limits of Combuatibilily.

The very ingenious principle of this method — due to an Ameri- can, Mr. Shaw, the engineer of a fiery colliery in Ohio — is the fact that the limit of inflammability of a gas — that is, the smallest pro- portion of it which, present in air, will give a combustible mixture — is a rigorously definite quantity, capable of being determined with great precision.

Care must be taken, whenever the mercury is removed for the purpose of pnri- firing it or of cleaning the glass cylinder, to make afterwards one or two '' blank tests, 80 as to get rid of the air which may adhere between the glass and the mer- cury, and may, as the latter is depressed in the measurer by the expansion of the gas heated from the platinum spiral, mix with the gas of the sample to be tested.

The Detection And Measubement Of Fire-Damp. 131

The experiments undertaken by M. Le Chatelier upon mixtures of air with illuminating gas or with methane, in successively in- creased proportions, for the purpose of verifying the proposition of Mr. Shaw and of determining the degree of precision of the test for fire-damp based upon it, gave the following results

a. Mixtures of air and illuminating-gas. (Given in percentages of the latter.)

No. of Ezpeiiment.

luflammable.

Non-inflammable.

. 8.45 8.25 8.05

7.95 7.8

7.96 7.7

8.1 8

. 8.6 8.4 8.2

8.1 8

. 8.4 8.2 8.15

8.0 7.95

6. Mixtures of air and methane (prepared by the action of soda- lime upon the fused acetate of soda). Given in percentages of pure methane.

Inflammable.

Non-inflammable.

8.2 6.2 6.2

6.1 5.9 5.9 5.6

7.05 6.1

6.0 5.9 5.6

No. of Experiment

1 .

These figures give as the limit of inflammability when the initial temperature of the mixture is between 10° and 20° C, 8.1 per cent, for illuminating gas and 6.1 per cent, for methane. Mr. Shaw had found 8 per cent, for the former and 6 per cent, for the latter. The agreement shows that the limit of inflammability may be determined at least to within xi(nr of the total volume of air and combustible gas.

In operating upon a mixture of air with two different combustible gases, it is observed that the mixture becomes inflammable when

n and n' being the volumes of the respective gases mixed with suflS- cient air to make 100 volumes of the mixture, and iVand being the respective limits of inflammability of the two gases tested separ- ately.

Annates dfs Minea, 1891, srie, tome xix., p. 388: ''Note sur le Dosage da Orison par les LuniteB d'lDflammabilit/' par H. Le Chatelier, Ingoieur en Chef des MineA,

132 The Detection Anb Measurement Op Fibe-Damp.

To determine the proportion of fire-damp contained in a given iample of air, it is therefore sufficient to add either illuminating-gas or methane until the mixture becomes inflammable.

The volume x of fire-damp contained in 100 volumes of the sample from the mine is then given by the following formula, which is easily deduced from the preceding one :

in which n is the volume of combustible gas added to the mine-sam- ple in order to make 100 volumes of a mixture at the limit of inflam- niability ; and N is the volume of the same gas which would have to be added to pure air for the same result.

Mr. Shaw based upon this principle an apparatus automatically indicating, above-ground, the proportion of fire-damp in the air-cur- rent traversing a system of pipes at a given point in fiery under- grouni workings. This very ingenious apparatus* is much too complicated, and therefore too costly, for general use in mines. Moreover, it does not lend itself easily to the analysis of samples taken underground.

M. Le Chatelier has constructed on the same principle an appa- ratus which is, on the contrary, very simple, and which, with the aid of a gauge and any combustible gas, permits the rapid and pre- cise determination of the proportions of fire-damp contained in the air.f

In order to effect in rigorously definite proportions the mixture of combustible gas with the sample of air to be tested, M. Le Chate- lier employs a glass gauge, 35 mm. (1.4 inches) in diameter and 250 mm. (9.8 inches) long, contracted at its lower end to 20 mm. (0.8 inch), so that it can be closed with the thumb (see Fig. 3). This gauge is prolonged at the upper end in a smaller tube, 10 mm. (0.4 inches) in diameter and 250 mm. (9.8 inches) long. The volume of the gaseous mixture is limited by a line drawn at 50 mm. (2 inches) above the lower opening ; the upper tube is graduated to thousandths of this volume.

To make a test, the gauge is filled with water, and set in a pan of

♦ ArmaleB da Mine$, srie, tome xix., p. 379: Avertieseur de Qrisoa Thomafl 8haw/' par M. Bajrard.

t AimaUt de$ Mines, srie, tome xiz., p. 388 : " Note snr le Dosage du Griaoa pu lee Limitee d' Inflammability'' par H. Le Chatelier; and Griaoa" par H. LeCliatelier, p. 111.

The Detection And Measurement Of Fire-Damp. 133

water. The combustible gas (illuminatiDg-gas, for iDstanoe) is inlro- daced through a capillary tube, permitting the delivery of very small bubbles, of a volume less than one-tbousaodth of that of the gauge, that is, less than 0.2 cu. cent, for the dimensions given above. It is equally indispensable, to secure a regular delivery,

Fig. 3.

Scale, K of slse In practice. Le Chatelier's Qaage for Testing Fiery Gases.

bubble by bubble, .that the volume of the tube between the stop- cock and the ori6ce of delivery should be as small as possible. The volume of combustible gas is reduced by calculation to that which it would occupy under atmospheric pressure, the height h of the water in the gauge above the level of the water in the pan being measured, and the observed volume of gas being multiplied by the

rr L

factor — --- , in which H is the atmospheric pressure expressed ia the height of a water-column.

134 The Detection And Measurement Of Fire-Damp.

If the water in the pan is sufficiently deep, the volume of the gas at atmospheric pressure can be directly determined by lowering the gauge until the water stands inside and out at the same level.

The gauge is now filled to its lowest mark by the introduction of the air of the sample to be analyzed. It is then taken in the hand by the lower end, the orifice being closed with the thumb, and is reversed and so held until the water remaining in it has completely replaced the gas contained in the upper prolongation. It is then violently shaken for some moments, the water in the larger portion being rapidly moved to and fro, to effect a mixture of the air with the combustible gas. Finally, it is elevated in vertical position, until the operator is ready to ignite it, when it is quickly turned upside down, the thumb is removed from the orifice, and a burning match or small gas-flame is introduced without delay.

If the mixture is inflammable, a pale blue flame descends almost to the bottom of the gauge ; if not, nothing is seen. Combustion, in fact, extends some distance beyond the flame which has kindled it; but the fugitive aureole which surrounds the latter for an instant is most frequently masked by the light.

The limit of error in determining the percentage of the gas by this method does not reach one-thousandth of the total volume, provided the above precautions are strictly observed. They are in- dispensable to prevent incomplete combustion, near the limit of inflammability.

When illuminating-gas is the combustible gas adde<I, there is a peculiar source of irrularity requiring special precautions. The vapors of benzine in such gas dissolve in water, varying thus the combustible efficiency and hence the limit of inflammability. To avoid such irregularities, it suffices to operate with a large quantity of water, and to change it from time to time, so as to keep its solvent action always uniform. With the use of methane made from soda- lime and acetate of soda, or with tire-damp collected in an inspirator, this cause of error need not be feared.

Whatever be the combustible gas employed, it is necessary to verify from time to time its limit of inflammability, N. It is to be observed that a mixture of combustible gas with air can be as easily added for this test as a pure gas, provided the mixture is combusti- ble, and that the volume of it required, to bring the sample to the limit of inflammability, is not greater than can be contained in the narrowed and graduated extension of the gauge.

Since each operation requires but one or two minutes, and four or

0.21 to 0.24

0.40 to 0.42

THE DETECnON AND MEASUREMENT OP FIRE-DAMP. 135

five tests at most would determine within one-thousandth part the limit of inflammability, even though the nature of the sample were not at all known beforehand, it is evident that this method is at once rapid and accurate, although it employs only the simplest means.

The following is a comparative statement of tests made by M. Le Chatelier, with this method and the method of incandescence, upon five samples of ordinary percentages.*

No. of Sample. Percentage of flre-damp determined by

Inflammability. Incandescence.

4, 5,

This agreement shows that by either of the two methods the fire- damp can be determined to within less than 0.1 per cent., and their simplicity (especially that of the method of the limit of inflamma- bility) recommends them as means of laboratory-determination for current mining work, which is seldom provided with carefully equipped laboratories and special analysts, and therefore needs appa- tus of the least complexity, requiring but brief training for their proper use.

Chapter II.

Portable Fire-Damp Indicators For Underground Usb.

Attempts have been made to modify into portable forms the in- struments based upon the combustion of fire-damp in contact with a metallic spiral. Thus, M. Coquillion has constructed a portable ap- paratus on his system. But it requires that the operator shall carry, besides the apparatus itself, a strong (and therefore large and heavy) electric battery, the handling of which would be difficult in work- ings upon thin or highly-inclined coal-seams. Moreover, each analysis, to be precise, requires a considerable time, and there is no advantage in performing underground the tests which can be more conveniently made above ground upon samples collected for the pur- pose. It therefore does not appear likely that an apparatus of this class will be adopted in daily practice.

1. Indicators Baaed on the Physical Properties of Ftre-Damp. It has been proposed to utilize, for the detection of fire-damp, dif-

♦ Annates des Mines, srie, tome ii., p. 477 ; Sur le Dosage du Grisou," par H. Le Chatelier.

136 The Detection And Measurement Of Fire-Damp.

ferent physical properties of the gas, such as the phenomeDa of dif- fusion through a porous body or the variations in density of a mixture of air and methane. When a porous body, like a plate of biscuit porcelain, separates two spaces, filled, the one with air and the other with a mixture of air and fire-damp, the pressure, having been at first the same on both sides, will increase in the former space until it attains a maximum, from which it will then decline to its original point. The increase of pressure at the moment of its maximum is nearly proportional to the proportion of fire-damp in the mixture filling the second space.

M. Ansell constructed, about 1868, an apparatus upon this prin- ciple, consisting of a metallic barometer, closed below with a plate of biscuit*

This plate, protected ordinarily by a metallic cap, is exposed at the moment of observation and the rise of pressure is noted. Un- fortunately, this ingenious instrument is not very sensitive. It shows a rise of pressure of only 0.2 millimeter (0.008 inch) of mercury for 1 per cent, of fire-damp present. Its indications are, moreover, liable to be vitiated by the presence either of carbonic acid (the influence of which is opposite to that of fire-damp) or of the vapor of water (which acts in the same direction as fire-damp). Moreover, it is only possible to make one observation with this in- strument before returning it to an atmosphere of pure air. It is thus for good reasons that M. AnselPs apparatus has not been adopted in mining practice.

The devices for utilizing the diminution of the density of air caused by the mixture of fire-damp with it are not capable of giving any better results. Mr. Forbesf has proposed to utilize this prop- erty through the increased rapidity of the vibrations of sound in a medium lighter than air. Numerous other inventors have projKJsed, and are still continually proposing, to measure the proportion of fire- damp with a sort of hydrostatic balance, the beam of which is equi- librated in pure air by a small weight at one end, and a large recep- tacle with thin walls, filled with pure air, at the other. In a fiery atmosphere, lighter than pure air, the equilibrium is disturbed, and the receptacle filled with pure air will descend in proportion to the amount of fire-damp present.

These indications, like those of the Ansell apparatus, are vitiated

Annates des Mines 1881, Rrie, t xix., p. 186 tt seq. ; "Sur les Procs propres k Dceler la Preseoce du Grisou,'' par MM. Mallard et Le Chatelier. t Trans, of N, of E, Insi, of M. E,, 1880, vol. xxix., p. 171.

The Detection And Measurement Op Fire-Damp. 137

by the presence of carbonic acid (specific gravity 1.529), which is heavier than air, and by the presence of watery vapor, the density of which (0.622) differs little from that of methane (0.558). It may therefore be regarded as hopeless to seek in this direction practically valuable indicators of fire-damp.

We will pass from this enumeration of devices, possessing only theoretical interest, to consider those which are susceptible of prac- tical use.

It may be definitely said that, for the purpose of continuous measurements in all parts of a mine, it has been found possible to utilize, up to the present time, only the three following pro|)erties of fire-damp :

1. The heating, by the combustion of the fire-damp contained in the air, of a platinum wire already brought to a certain temperature by means of an electric current. According to the proportion of the gas and the initial temperature of the platinum in pure air, the metal is carried from incipient red to bright red heat by the slow combus- tion of the gas in contact with it, and this heating may serve to measure the percentage of gas in the surrounding atmosphere.

2. The flames of oil-lamps may be employed to indicate the pres- ence of fire-damp by the elongation which they exhibit in air im- poverished in oxygen.

3. Finally, observations may be made on the aureoles produced in contact with the flames by the combustion of fire-damp in the hot zones nearest to them, the height of an aureole depending upon the size and temperature of the flame as well as the proportion of fire- damp, and its visibility depending upon the relative brightness of flame and aureole.

Upon this third property are based the most widely employed in- dicators, as well as the grisoumeter, which I designed a few months ago, and which seems to satisfy the requirements of precision and safety. Before describing the instruments of this class, however, I should give some account of the indicators, based upon the first and second properties named above, which are capable of useful ser- vice.

2. Indicaiors Baaed upon the Heating in Contact with an' Atmosphere

Containing Fire-Dampy of a Platinum Wire, Traversed

by an Electric Current.

The tendency to electric lighting in mines lends great interest to this class of indicators. Although the problem of a portable elec-

138 The Detection And Measurement Op Fire-Damp.

trie lamp has not yet been practically solved,* it is to be expected that this will be achieved in the near future. A frequent cause of explosions, due to the lighting of mines, will be removed, and, as a consequence of the new method of illumination, it will be necessary to provide the miner with an indicator of 6re-damp, utilizing the electricity of his lamp. The property of a platinum wire, when traversed by an electric current, of becoming strongly heated in an atmosphere containing fire-damp, seems to be well adapted to com- plete the coming system of colliery-illumination.

Two sources of error involved in this method should be pointed out. It appears to be beyond doubt that the heating of the platinum wire is proportioned to the percentage of fire-damp ; but in order that observations may be mutually comparable, it is necessary that the initial temperature of the wire in pure air shall be rigorously constant, and hence that the electric current which produces that temperature shall be constant, — requirement difficult, but appa- rently not impossible, to realize.

Curve of Temperature of Plaiinum Wire, Traversed by an Electric Current in Atmospheres Containing Fire-Damp.

The same cannot be said of the following difficulty. If the pro- portion of fire-damp be gradually increased from 0 to 100 per cent., the temperature of the platinum wire will first rise gradually until the percentage of fire-damp (9 to 10 \)er cent.) corresponding to the most rapid combustion is reached, after which the temperature will decline, arriving, when the percentage of fire-damp is 100, at the point from which it started. If the percentage of fire-damp be rep-

Recent experimentfl with the ''Stella" lamp at the mines of Anzin, in France, did not give satisfactory results. The lamps deteriorated in U8e with considerable rapid it J.

THE DETECTION AND MEiLSUREMfiNT OF FIRE-DAMP. 139

resented by abscisrse'aod the corresponding temperatures of the wire by ordinates, the phenomenon will be represented by a curve start- ing at the origin, reaching its maximum at about 10 per cent, of fire-damp, and returning to the horizontal axis at 100 per cent.

It follows that to one and the same temperature t of the wire, two very different percentages B and C will correspond, and particularly that any apparatus whatever, based on this principle, whether the brightness, the heat or the varying conductivity of the platinum be observed, will give precisely the same results in pure fire-damp as in pure air. With oil-lamps, which are extinguished by explosive mixtures, misunderstanding of the indications of such an apparatus need not be feared, but when the lamp itself is an electric one, there will be no way to decide which of the two possible percentages is indicated by the instrument.

With these reservations we may proceed to state how the property in question has been utilized hitherto.

Mr. Liveing, an English engineer, has proposed to utilize the dif- ference in brightness of two platinum wires brought to dull red heat by the same current, and placed, the one in pure air and the other in the air to be tested.

His apparatus, which was studied and described by MM. Mallard and Le Chatelier,'*' is composed of two platinum wires, one of which is placed in a closed glass vessel containing pure air, and the other in an envelope mainly of metallic gauze. A current can be passed through both from a hand-dynamo. The gauze envelope is exposed to the surrounding air which is to be tested, and a portion of it is of glass, to permit the luminosity of the wire it contains to be clearly transmitted.

Between the two wires is a movable partition, with two faces, in- clined at 45, painted white, and illuminated respectively by the light from the two wires. If the air is pure, the brightness of the two wires will be the same, and the partition when placed midway between them will be equally illuminated on both faces. If the air contains fire-damp, the wire exposed to it will be the brighter and the equal illumination of the faces will require the removal of the partition further from that wire. A scale may thus be established by making tests upon mixtures of known percentages. The apparatus is, unfortunately, not accurate, the possible error being 0.3 to 0.5 per cent., or even more, and the manipulation in mine-chambers is somewhat inconvenient.

AvnaUa de Miiies, 1883, 86rie> tome iii., p. 31.

140 THfi DETECTION AND IfEABUBEMENT OF FIRE-DAMP.

Mr. Murday, another English engineer, has 'recently published a description of a thermo-electric indicator," which, according to the inventors statement, will indicate the presence of 0.1 per cent, of g. The principle is the same as that of the Livcing apparatus; but instead of the dilTerence in luminosity of the two wires, their difference in elongation due to temperature is measured. This is effected by means of levers and a needle, which form a sort of dif- ferential thermometer. The apparatus is rendered bulky and diffi- cult to transport by the batteries required for its operation. More- over, it contains delicate parts which will scarcely endure the very dusty atmosphere of many collieries.t

It does not appear to be suited for use except at a fixed station conveniently arranged for it.

3. Indicators Based on the Elongation of Flames in an Atmos- phere Impoverished in Oxygen,

The flame of an oil-lamp, adjusted in pure air at the height suita- able for good light, elongates and may become smoky in an atmos- phere containing 6re-damp. This elongation, which was studied in 1880 by MM. Mallard and Le Chatelier,t is due to the fact that the zone surrounding the flame consists, not of pure air, but of a mixture of air and fire-damp in combustion. The proportion of free oxygen in this zone is thus greatly diminished, and the volume of the flame must be correspondingly increased, if it is to receive upon its sur- face the quantity of oxygen necessary to burn the combustible pro- ducts furnished by the wick. Moreover, the action of the fire-damp is not merely due to the volume which it occupies, thus diminishing by so much the volume of the air available for supporting combus- tion. It acts also by absorbing for its own combustion double its own volume of oxygen, and, consequently, a volume of air many times larger. While 3 to 4 per cent, of carbonic acid in the air produces upon the flame almost no appreciable effect, a few tenths per cent, of fire-damp suffice to elongate the flame. With a Davy lamp, the flame of which had been adjusted in pure air at a length of 27 mm. (1.08 in.), MM. Mallard and Le Chatelier observed the following elongations :

Colliery Ouarrfuin, 1892, vol. Ixiv., p. 789, "Murday'a Thermo-Electric Fire- Damp Detector."

t According to information received from English engineeni, this apparatus has not jet been tested underground, and its practical value cannot be decisivelj es- timated. I AnnaleB des Mines loc. eitt 1881.

The Detection And Measurement Op Fire-Damp. 141

flre-Damp, Elongation of

Per cent. flame.

0.33 . . Difficalt to measure for lack of a fixed starting point

0.75 lto2 millimeters.

1.50 3 to 4

4.00 20 "

It appears from these figures that a miner, by simple inspection of the flame of his uil-lamp (other safety-lamps give analogous in- dications) could estimate the fire-damp from about one-third of one per cent, upwards, with considerable precision. Unfortunately, it is seldom possible to utilize the indications of this kind furnished by oil-lamps for the following reasons :

In the first place, the elongation, at least when it is small, is diffi- cult to observe in the almost always agitated air of a mine gangway. Besides (and this is the most serious hindrance), in passing gradually from pure air to air containing 3 to 4 per cent of fire-damp, and after having travelled a more or less extended road through the mine-workings, the miner will indeed observe an elongation of the lamp-flame ; but it will be difficult to affirm that this elongation re- sults from the presence of fire-damp, and is not due to other causes, such as a higher temperature, a deoxygenation of the air, etc., or even simply to bad adjustment of the wick. Elongations, extending even to filage, are in fact observed at any moment, in atmospheres free from fire damp. This method is, therefore, only a diflTerential one, permitting us only to observe, at a given place in the workings, the differences of composition which *may be exhibited at the same instant by two distinct parts of a gangway or breast. It is applica- ble, for instance, to the examination of the recesses of the mine-roof for fire-damp; and, up to a certain point, it may serve to compare the greater or smaller quantity of fire-damp in the recesses with the emouut in the air of the gangways ; but the percentage of fire-damp cannot be even roughly approximated by any deduction from the observed flame-elongation.

It is possible, however, as I have proved, to obtain indications of great accuracy in absolute value, and perfectly comparable with each other, by substituting alcohol for oil, and limiting by a special ar- rangement the admission of air upon the flame. I have succeeded in making upon this principle an alcohol lamp of great sensitiveness. It has, however, the serious drawback that it is much influenced by variations of temperature. It is, therefore, not suited for detecting and measuring fire-damp in chambers, when the temperature of the

142 The Detection And Measurement Op Fire-Damp.

air generally increases as it traverses the workings. On the other hand, it is perfectly adapted for testing the percentage of fire-damp in a general air-return where the temperature remains constant for several weeks, and where it is sufficient to place at a fixed station the apparatus, adjusted in pure air at the same temperature as that

Fig. 5.

Am. Bank Notk 00N.T>

Scale, full sice. Chcsneau's Fire-Damp Detector. (First Form).

of the return. The flame of alcohol is much more brilliant in an atmosphere impoverished in oxygen than in pure air. It has about the luster of a candle-flame, and could probably be employed for continuous photographic record. While I do not consider this ap- paratus as practical for a tour of inspection proper through under-

The Detection And Measubement Of Fire-Damp. 143

groand workings, its special qualities are such as to lead me to give here a description of it.

This apparatus consists of an alcohol-lamp entirely similar to my grisou metric lamp of the type described at the conclusion of this paper (p, 151). It differs from that type simply (as is shown in Fig. 5) in having:

1. A smaller diameter of wick (7 mm. 0.28 inch instead of 9 mm. 0.36 inch).

2. Openings at the base of the shield (not shown in Fig. 5) ; and finally,

3. The addition, around the wick-holder, of an envelope or shade of thin copper of truncated conical form, pierced at its base with a crown of circular holes regulating the quantity of air which feeds the flame.

The flame of the alcohol (ethylic alcohol, of 90° by the Gay-Lus- sac alcoholometer) is adjusted in pure air, after 15 minutes burning, in such a way that its tip (very brilliant and well-defined, like that of an oil-lamp flame) is level with the upper edge of the shade. Under these conditions, with the dimensions adopted for the height of the wick-holder and of the truncated conical shade respectively, and the number and size of the orifices in the latter, the following elongations of the flame above the shade will l>e observed:

Fire-damp, per cent., 1 2 3

Elongation, mm 12 45 80

Elongation, inches, 0.48 1.8 3.2

Up to 3 per cent, of fire-damp, the tips of the flames are extremely clear; above that point they begin to be confused. The lamp goes out in quiescent mixtures of fire-damp at the lower limit of inflam- mability. In explosive mixtures of air and illuminating-gas it does not always go out, but its gauze docs not become sensibly red.

The variation of level of the alcohol in the reservoir has a notable influence upon the length of the flame.

The influence of temperature is such that an elevation of a dozen degrees produces in pure air the same effect as about 2 per cent, of fire-damp. Evidently it would not be impossible to regulate the

One lamp has been constructed and tried with perfect success in an underground loor made October, 1891, at shaft No. 3 of the Cie, de Lens in gangways of constant temperature. I only recognized at a later period the perturbing influence of vari- ations of temperature, which led me to abandon the apparatus. No description of it has been published heretofore.

144 The Detectiox And Measurement Op Fire-Damp.

height of the wick by means of a lateral screw and graduated dial, so as to correspond with the regulation of the flame in pure air for each temperature, and the apparatus would then be adapted for de- terminations in all parts of a mine, fiut this would require the continual reading of a thermometer, and would not present the sim- plicity required of instruments designed for mine use. This circum- stance is regrettable, in view of the unequalled clearness of the alco- hol flames given by this apparatus, fiut that advantage is singularly reduced by the increase of clearness which I have obtained, by means of a special alcohol, in the aureoles caused by alcohol-flames in at- mospheres containing fire-damp, as will be explained in the following section.

4. Indications Founded upon Flame Aureoles in Atmospheres Con- taining Fire-Damp,

Air containing from 6 to 17 per cent, of fire-damp is inflammable ; that is to say, when it is tranquil the combustion excited at one point will extend progressively, with more or less rapidity, to the whole mass.

If the proportion be less than 6 per cent., a flame or sufficiently hot body introduced into the mixture causes combustion only in a larger or smaller zone surrounding the hot body. The gas still burns, but only with the aid of external heat.

The quantity of gas burned, or, what is the same thing, the ex- tent of the zone in which combustion is effected, is thus greater, the more heat is emitted in the unit of time by the hot body. If, for instance, an oiMamp be introduced into an atmosphere containing 6 per cent, of fire-damp, the white flame of the lamp will be seen sur- mounted with a bluish flame which has received the name oi aureole. This aureole, at the proportion of 5 per cent., will have a height of 2 to 3 centimeters (0.8 to 1.2 inches) at most if the flame of the wick be reduced to a minimum. It might be more than 10 centimeters (3.9 inches) long if the flame were of ordinary dimensions, but in that case the brilliancy of the latter would render almost invisible the extremely pale light of the aureole.

It is, in fact, by observing these aureoles, after having lowered the wick until the flame shows no longer any brilliant point, that miners have endeavored, since the time of Davy, to estimate rapidly the proportion of fire-damp in the air of mines.

MM. Mallard and Le Chatelier, who were the first to study with care the aureoles given by the Davy and the Mueseler lamp in con-

The Detection And Measurement Of Fire-Damp. 145

tioooas slow oarrents of air mixed with methane, concladed from their experiments* that, ad indicators of fire-damp, these lamps do not begin to give tolerably clear indications below 2 to 3 per cent. It is necessary with these small proportions of methane to commence, before the operation, by lowering the wick slowly, so as to cause the luminous part of the flame to disappear. For 3 per cent, the conical aureole thus obtained is 6 to 8 mm. (0.24 to 0.32 inch) long, and attains, for 6 per cent., in the Davy lamp the length of 30 mm. (1.2 inches).

At and above 6 per cent., these lamps can no longer serve as in- dicators— the Davy lamp becomes filled with flame, and the Mueseler goes out.

Practically, oil-lamps cannot be relied upon to show with certainty the presence of less than 3 per cent of fire-damp — percentage which should never be expected in the ventilated portions of fiery mines. The security attribute to inspection with oil-lamps is there fore wholly illusory.

By furnishing the Mueseler lamp with a fixed shade, 8 mm. (0:32 inch) high, below which the flame is caused to disappear at the mo- ment of observing the fire-damp, MM. Mallard and Le Chatelier suc- ceeded in rendering the aureoles visible, even for small percentages ; the flame still having sufficient length (8 mm.) to produce much longer aureoles than those obtained when the wick itself is lowered until th" flame is no longer luminous. With this arrangement it is pos- sible to estimate the fire-damp, from 0.5 per cent, upward, with great facility, and from 1 per cent, upward with considerable precision. A trained observer can, with this lamp, give the percentage of fire-damp to within about 0.6 per cent.

The use of this apparatus has not become general, as much oa account of the diminution of the light of the lamp by the added- 8hade as by reason of the indistinctness of the aureoles for low per- centages.

In the course of their studies upon flame aureoles, MM. Mallard and Le Chatelier were the firstf to point out the use which could be made of the much larger aureoles given by flames of alcohol and of hydrogen ; but they considered that, in practice, the use of alcohol would be dangerous and that of hydrogen difficult, and confined themselves to recommending the use of the aureoles from oil-lamps, with the additional introduction of the shade for the Mueseler lamp.

Awnalea da Mines, 1S81, srie, tome ziz., p. }B6etaeq. t AimaUi de$ Mines, 1881, srie, t. xiz., pp. 205, 206.

VOL. xxu.— 10 . , ,

146 The Detection And Measurement Of Fibe-Damp.

The idea of the alcohol lamp was taken ap by Herr Pieler, an Austrian engineer, whose apparatus, now widely used in Germany, Austria, and the north of France, is merely an ordinary Davy lamp supplied with alcohol instead of oil, and furnished with a shade sur-

Fig. 6.

Fig. 7.

iirt™

ipoo o o o o o ooo<

Movaslc Cuirass.

Scale, h of slxe In practlos. Pieler Lamp.

rounding the flame which just reaches the edge in pure air, and gives in fire-damp perceptible aureoles from 0.25 per cent upwards.

The recent experiments of the Austrian and the French Commis- sions' on Fire-Damp have shown that, in this form, the Pieler lamp

I Annales des Mineg, 1892, sie, tome i., pp. 47 and 239.

The Detection And Measurement Of Fire-Damp. 147

is unsafe. An explosive current of 4 to 6 meters (13 to 16 feet) per second is sufficient to propagate ignition to the outside, even through a double gauze. Hence, it should only be used when pro- vided either with a perfectly close shield, which is removed at the moment of observation, as recommended by the Austrian commis- sion (see Figs. 6 and 7), or else, with a complete permanent 'shield containing a window which is opened for the observation. The latter is the form employed for the last eight years in the Anzin colliery.

According to the Austrian commission, the comparative heights of the aureoles given in fiery '' atmospheres by the Mueseler, Wolf (benzine), and Pieler lamps, are as given in the Table on p. 148.

The Austrian commission admits, that for ordinary percentages, below 3 per cent., of fire-damp, the determinations of the Pieler lamp are correct to within about 0.25 per cent., and that, from 3 to 6 pr cent., the error may amount to 1 per cent, in consequence of the heating of the lamp and the distillation of alcohol vapors. It will be seen, further on, that the precision practically attained is inferior to these figures.

Before describing my own researches for a mine-indicator, safer and more accurate than the Pieler lamp, I should mention the recent attempts to transform, at will, an ordinary oil safety-lamp into an indicator using alcohol or hydrogen, so that the same lamp may be employed for lighting and for the detection of fire-damp.

In England, Mr. Frank Clowes' has invented a method of trans- forming, at will, an oil- into a hydrogen-lamp by means of a small tube which traverses the lamp-reservoir, and may be put in commu- nication with a steel reservoir containing hydrogen. The latter is

3 inches in diameter and 8 inches long, weighs 4 pounds, and may contain, under compression, the amount of hydrogen equivalent to

4 cubic feet at atmospheric pressure. When it is desired to test for fire-damp, the cock of the reservoir (serving also to regulate the flow of gas) is 0|)ened, the hydrogen ignites in the flame of the lamp, which is then extinguished by turning down the wick, and the flow of hydrogen is so regulated that its flame will be, in pure air, 10 mm. (0.4 inch) long. The aureoles have the following heights for corr spouding percentages of fire-damp :

Fire-damp, per cent, . . . 0.25 0.50 1.00 2.00 3.00 Length of aureole, mm., . . 17. IS. 22. 31. 52. Length of anreole, inches, . . 0.68 0.72 0.88 1.24 2.08

♦ Proc. of Roy, Soc, 1892, vol. 51.

148 THE DETECTION AMD UEASUREUEltT OF FIBE-DAMP.

'S'

I,

m

s

2 o 2

Pointed cone. tt

tt

Trunc. cone, slightly spread'g in up'r part.

tt

Cylinder, spreading widely at the top.

Flames turned over along the gauze.

f

si

Beaches top

of the gauze.

tt

tt tt

Si

Z a

d

o

O

Pointed cone.

tt

Right cone, spread'g toward top of gauze. The aureole forms a bell, open above, In projecting flame tow'rd base along the gauze.

M

► tt

!l

Os

Reaches top of the gauze.

t

[

'5

tt

Poinied cone. tt

tt

Cone, very much elongated.

it

Cylinder.

Inverted truncated cone. The flame fills the whole chimney and reaches the end of it in the glass cyl- inder, forming a hell. Total extinction.

M S

.

ss

§

n

16J

Point of aureole- cone disappears in the chimney.

Top of the aureole stays in chimney.

u

durep-ajU[ JO adinuaojdci

The Detection And Measurement Of Fire-Damp. 149

At higher percentages the aureole extends above the glass cylin- der of the lamp, where it cannot be seen. After the oliservation is finished, the oil-wick is relit from the hydrogen flame and the cock of the reservoir is then shut.

M. Legrand, an engineer at the AnEin mines, has constructed, on the principle of this apparatus a compound lamp for oil and alcohol, having two reservoirs and two distinct wick-holders, and operated at will, exactly as is the Clowes lamp, either for lighter for the detection of fire-damp.

These instruments, though exti*emely interesting, do not appear to be very accurate, because they are not regulated, after each ob- servation, in an atmosphere of assured purity.' But they may be very good for proving the absence of fire-damp.

Although the Pieler lamp, as the experiments of the French and Austrian commissions have shown, is rendered much less dangerous by the addition of a complete shield, its gauze glows none the less strongly in explosive mixtures, and is liable even to melt in a lively current of sufficient rapidity. The heating causes abundant distilla- tion of alcohol, and even after removal of the gas the alcohol-vapors continue to be liberated, their combustion keeps the gauze red-hot, and it is impossible to extinguish the lamp except by plunging it into water.

I have found that the Pieler lamp, with gauze and shield of the type adopted at Anzin, when placed in the apparatus described in the Annates des Mines (1892, t. i., p. 48) in a very slow current of air and formene, likewise glowed strongly when the proportion of gas exceeded 4 or 5 per cent., and, when carried back into pure air, it remained filled to its whole height with a large, brilliant flame, due to the combustion of alcohol vapors, which required often more than a quarter of an hour to shorten down to the level of the shade, while the gauze remained red-hot for several minutes. Even when the observation-window and the air-openings in the shield are closed the lamp is not only not extinguished in explosive mixtures but the heating of the gauze and of the whole lamp is not sensibly dimin- ished.

There was, therefore, good reason for the recent recommendation of the Austrian Fire-Damp Commission that, in very fiery mines

Moreover, if desiring to inspect the gangways with the Clowes apparatus for several minutes without interruption, the diminution of the pressure of hydrogen in the reserroir gradually lowers the flame and the aureoles are no longer com ptrable.

150 The Detbgtiok Axd Measurement Of Fire-Damp.

with strong air-currents, the Pieler lamp should be provided with a shield, and that it should not be employed except when a preliminary test, always to be made with the Mneseler or the Wolf lamp, had failed to indicate fire-damp. With double gauze and complete shield, as in the lamp of M. Dinoire, used at the Lens collieries, the above defects are still present, though in diminished dree. The Dinoire lamp, passed suddenly from pure air into an explosive mixture, fre- quently goes out ; but if it has been at all heated by a 1 or 2 per cent, mixture, it will continue to bum in the explosive mixture at rest, and when the air-orifices are open its interior gauze will become red-hot. If these are closed, as well as the observation window, the combustion of the explosive mixture takes place only in the upper part of the gauze cylinder, and the lamp becomes thereby much less highly heated.

M. Dinoire's double-gauze shielded lamp is much safer than the other types, but at the cost of distinctiveness in the aureoles, which, by reason of the double gauze, cannot be clearly observed for proportions below 1 per cent

Moreover, all these various types have this defect : that when the lamp is heated in a mixture of high gas-percentage, the alcohol-flame takes a long time to resume its standard height in pure air, and for more than thirty minutes the aureoles are no longer comparable with those observed at the first lighting of the lamps.

It may, therefore, be said that all the indicators now in use em- ploying alcohol leave much to be desired as to both accuracy and safety, apart from other causes which may influence determinations of gas-percentages, such as the nature of the alcohol, composition of the air, temperature, etc., to which I will return further on. But it still remains true that for small percentages of fire-damp, not shown by ordinary lamps, alcohol by virtue of the constancy of its flame for several bourn, fumbhes the clearest and most definite aureoles, and I have therefore conducted my own investigations in the class of alcohol-lamps, with the view of devising a type which, with sen- sitiveness at least equal to that of the Pieler lamp, should have the advantage of not heating so much, and, at all events, of never be- coming red-hot in explosive mixtures at rest or in motion.

To avoid the disadvantages of the Pieler lamp, I have endeavored to transform into alcohol-lamps several types known as highly safe.

The Marsaut lamp, fed with alcohol, behaves in explosive mix- tures very nearly as it does when fed with oil, and is consequently much less dangerous than the Pieler. But by reason of the method

The Detection And Measurement Op Fire-Damp. 161

of circalating the fresh air and the combustion products, the aureoles due to fire-damp are rather vacillating, and the presence of the gas causes a permanent heating of the glass (or of the metal shade which may be substituted for it), which modifies the adjustment of the lamp and renders it unsuitable to determinations for comparison.*

To obtain a uniform adjustment, which will render the alcohol- burning indicator not merely a qualitative but a truly quantitative instrument, I have found as the necessary condition that the fresh- air supply shall never be mixed with the products of combustion, and, moreover, that it shall circulate around and constantly cool the outside of the alcohol-reservoirs. I was thus led to adopt an arrange- ment for supplying the air at the bottom, as, for instance, in the Fumat lamp.

Starting from this point, I have, after many attempts, adopted the following type, the first trials of which have been so satisfactory that its principal features may be regarded as settled.

(a) Desoription of the Fire-Damp Indicator of O. CAeneau.f — This consists essentially (see Fig. 8) of a brass reservoir for alcohol, A ; a circular crown made of double-wire gauze, B, serving for the admission of fresh air; a hollow cylinder, C, made of sheet metal, surrounding the wick-holder and serving as a shade ; an iron-wire gauze, D, with 196 meshes to 1 square centimeter (1264 to the square inch), and 140 mm. (5.6 in.) high, the collar of which, resting on the hollow sheet-metal cylinder, completes the shade; and, lastly, of a sheet-iron shield, E, provided with a window for observation, thor- oughly closed by a sheet of mica, as thin as possible to permit ac- curate observations. The distinguishing characteristic of this shield is that it is fitted at its base with an annular diaphragm, G, which rests on the collar of the gauze, and thus completely closes the bottom of the shield, so that the outer air can never reach the gauze directly. Be- tween this diaphragm and the collar of the gauze there is interposed a washer of asbestos cardboard, o, o, the object of which is to lessen any heating in the lower portion of the lamp, when the latter is plunged into mixtures containing a large proportion of gas, and when

ForaBimilar reason the Wolf benzine-Ump, which gives very clear aureoles for iniDate proportions of fire-damp, presents, as I have proved, highly variant in dications, according as the adjustment of the flame (secured by lowering the wick nntil the Inminoos portion disappears) is performed in pure air or in air containing traoes of fire-damp.

t ArmnUs dea Mines, 1S92, srie, t. ii., p. 203 aeq. " Note sur un Nouvel Indicatear de Orison, par G. Chesnean, Ingnieur des Mines.

152 The Detection And Measurement Of Fire-Damp.

Scale. H of size In practloe.

THE DETECnON AND MEASUBEMENT OP FIRE-DAMP. 153

the shield has a tendency to become heated. The upper portion of this shield is provided with apertures, 6, 6, protected by a fixed shtfde, d, which prevents the air-currents from impinging directly on the gauze at too great a speed. A movable shade, H, formed by a cylinder of thin sheet-copper, pierced with a window, protects (when necessary) the bottom of the observation window from sharp currents of fresh air, and thus prevents a mist from being formed inside the mica through its being cooled on the outside. It suffices, when making an observation of the aureole, to bring the window of the movable shade opposite that made of mica, when in a calm at- mosphere the mist will disappear of itself a few minutes after lighting, and only form again when the lamp is taken into a current of cold air.*

A scale is painted in white on each side of the window— one side giving tenths of 1 per cent, of fire-damp ; the other, centimeters.

A fixed shade, I, pierced with holes which can be closed when nectissary with a movable part operated by a button, protects the gauze crown, B, against draughts.

In the interior of the reservoir. A, there is a certain quantity (6 grammes — 92 grains) of wadding, the purpose of which will be ex- plained further on. It is not intended, as in the Pieler lamp, to hinder the alcohol from running out when the lamp is upset I have found, in fact, that where a Pieler lamp, tightly stuffed with cotton, has been filled with alcohol, and then drained of all that will run out, the flame will only give for a few minutes perceptible aureoles in gaseous mixtures. At the end of a brief period, the flame shortens, the wick chars, and no aureole is observed even from 1 to 2 per cent of fire-damp. The reason is, that the alcohol, strongly held by the cotton in the reservoir, rises in the wick but very slowly. In- dee<1, the flame gives aureoles which can be compared only when the reservoir contains alcohol which can run out freely if the lamp be overturned. With the quantity of cotton used in the new indicator, about 120 cu. cent. (7.32 cubic inches) of alcohol could run out from

The formation of this mist in dusty mines rapidly soils the plate of mica, which be removed when it can no longer be made transparent by rubbing with a Tery fine linen cloth. It is best to change the mica frequently, the expense being insignificant by reason of its small size and thickness. It need only be strong enough to permit its insertion into the proper place, which is done by protecting it on both sides during the operation with stout paper, the paper being withdrawn when the mica is in place. Even very thin mica will stand the jarring of a mine- intpectioD tour ; moreover, it is well protected for half its height by the movable fhiidfi

164 The Detection And Measurement Of Fire-Damp.

the completely filled reservoir. This is equal to the cx)n8umption of four hours at least. This partial suppression of cotton in the new indicator presents, however, no danger, because the lamp, if turned to a horizontal position, goes out before any alcohol can drip from the wick-holder. Such is not the case with the Pieler lamp, in which the alcohol, if not retained by the packing, spreads, burning, over the gauze, if the lamp is much inclined. The alcohol is introduced into the reservoir by an orifice closed by the screw-plug, L; and tightness of the joint is obtained by a lead or leather washer inter- posed between the plug and its seat. The weight of the lamp, filled with alcohol, is 1460 grammes (3 pounds, 3 ounces).

(b) Properties of the New Cheaneau Indicator. — As the entrance of the air and the exit of the burnt gases are absolutely separated in the new indicator, there can be no mingling between them as in the ordinary Pieler lamp, in which the air charged with gas may come to burn at any height in the gauze, and heat it strongly. Introduced into mixtures of air and marsh-gas with an increasing proportion of the latter, at rest, the new indicator shows aureoles which only reach the top of the gauze at about 3 per cent., with the height of flame and alcohol adopted, which will be dealt with further on. For proportions between 3 and 6.5 per cent, the alcohol flame proper becomes elongated, the aureole widens and becomes cylindrical, but is more and more depressed at the top; the lamp can no longer give out the increasing volume of the products of combustion, and the gauze remains completely dark. A little below 5.75 per cent the alcohol-flame is drawn out to near the top of the gauze, always with- out making it turn appreciably red, the quantity of air drawn in through the gauze crown, B, being insufficient to cause the complete combustion of the gas and of the alcohol vapors disengaged through the heating of the lamp.

From 5.75 per cent, (a proportion lower than the limit of inflam- mability of the mixture of air and marsh-gas, which is 6 per cent.), all flame disappears from the gauze, and the mixture of air and gas burns with a very pale flame only in the crown, but without heating the reservoir to any considerable extent. Everything is extinguished after a few seconds, and very rapidly if the orifices of the shade, I, be closed.

This extinguishment, which occurs with certainty, is to be ex- plained, I think, by what is known of the limit of inflammability, which, being 6 per cent, for marsh-gas mixed with air at ordinary temperature, b lowered by increase of temperature, so that in the inte-

The Detection And Measurement Op Fire-Damp. 155

nor of the lamp the heated mixtare becomes inflammable at about 6.76 per cent. When this limit is reached, the ignition produced at the lamp-flame extends downwards to the current of gas circulating through the lamp. The velocity of the circulation is very low, the orifices of entry and exit being small, and the products of combus- tion, moreover, augmenting its volume more and more as the per- centage of fire-damp is raised, by reason of the greater heat liberated by combustion. Thus at 6.75 per cent, ignition is propagated to the gauze crown. But the lamp-flame is quickly extinguished, being environed only by the products of the combustion of the inflamma- ble mixture; the temperature of the interior of the lamp is sud- denly lowered ; the mixture is thus rapidly cooled to below its limit of inflammability and is speedily extinguished at the gauze-crown.

While in a Pieler lamp, which has remained for one or two min- utes in a mixture containing a high percentage of gas, even at rest, the heating causes a considerable elongation of the alcohol-flame, which persists fifteen to thirty minutes after return to pure air, the flame of the new indicator, under the same conditions, resumes very rapidly in pure air its original height, no matter what may have been the percentage of fire-damp to which it was ex- posed.

Plunged into explosive mixtures of air, with a maximum of illu- minating-gas moving very rapidly (the orifices of the shade, I, being closed), the indicator has given the following results

1. The velocity of the current being 6.36 m. (20.8 feet) before reaching the lamp, and 11m. (36 feet) at the lamp, the gauze remains dark, and there is no apparent combustion there. The explosive mix- tare bums in the crown, but develops there a temperature of only 260° to 300° C. at most (commencement of fusion of tin-solder). Aftr five minutes, during which the lamp has been violently shaken in all directions, the gas is cut off; the lamp relights itself, the alco- hol, vaporized by the heat developed from the combustion of gas in the crown, burns with a reddish dim flame in the gauze cylinder, which remains dark, and this flame rapidly diminishes, resuming the height of initial adjustment in thirty seconds.

Three similar experiments gave the same results.

The apparatus employed consists of a wire-gauze sheath, traversed by a current of air from a fan, and mixed with illuminating-as from a meter. The lamp is sus- pended in the middle of the sheath, and observed through a mica window. The sec- tion of the sheath being partially obstructed by the lamp, the velocity of the current is greater in this position of the lamp than before or after.

156 The Detection And Measurement Of Fire-Damp.

2. With the same current of air and a progressive admission of gas, the series of phenomena observed was the same as in mixtures at rest. At about 8 per cent, (the limit of inflammability of illuminat- ing-gas with air) the alcohol-flame proper, which had been elongated too near the top of the gauze, but was growing dimmer and dimmer, suddenly disappeared from the gauze, and the mixture burned only in the gauze-crown. The proportion of gas at which the lamp seems to be most heated (the gauze, however, still remaining dark) is about 6 per cent.

3. With a current-velocity diminished to 2.75 m. (9 feet) before reaching the lamp and 4.75 m. (15.6 feet) at the lamp, the flame goes out at 8 per cent, of illuminating-gas, as in mixtures at rest

As to safety, therefore, the new indicator has all the properties of ordinary safety-lamps of approved construction.

The same experiments repeated with the orifices of the shade, I, open have given the same results.

Experiments made in rapid currents have shown in addition that the lamp may be placed in currents of 8 to 10 m. (26 to 32 feet) velocity without being extinguished, provided it has been lighted for fifteen or twenty minutes before being exposed, and no attempt is made to shield it with the hands (or any other means tending to hinder the draught), which will cause it to go out.

c. Nature of the Alcohol Employed, — I have observed that the length of the aureole varies considerably, according to the nature and the degree of volatility of the alcohol. Thus, with the same adjust- ment of the flame, an ordinary Pieler lamp, for example, gives, with different alcohols of varying drees of concentration, as shown by the centesimal alcoometer, the following aureole- heights in a 1.4 per cent, mixture:

AlcoOmeter-reading. Height of Aureole.

Alcohol. Deg. Mm. In.

Ethylic, 88 40 1.6

" 95 80 3.2

Methylic, 95 55 2.2

" 99 75 3.0

Engineers using the Pieler lamp have not given sufficient atten- tion to this peculiarity, which, together with variations in the adjust- ment of the flame by different observers, is a frequent cause of error in determinations of fire-damp with that lamp. A very skilful per- son, who has been able to observe in an apparatus supplied with mix- tures of air and marsh-gas in known proportions the behavior of the

The Detection And Measurement Op Pibe Damp. 157

Pieler lamp which he uses regularly, provided with its own alcohol of uniform quality, can estimate the percentage in a chamber or gang- way, where thg temperature is not high, to within 0.25 per cent, from 0.5 to 2.5 per cent. But with the causes of error above named, de- terminations do not come within 1 per cent, of accuracy, as I have had occasion to prove in several French collieries. In one mine the determination with the Pieler lamp gave 1 per cent., and the labora- tory-analysis 2 per cent. ; in another the lamp indicated 3 per cent, when there was only 1 5 per cent., and in a third the Pieler lamp, being too tightly packed with cotton, showed no fire-damp at all in a gangway containing 1 per cent.

The choice of the alcohol to be used, and the mode of adjusting the lamp in pure air, should therefore be so defined as to eliminate these sources of error.

I have become convinced that in order to obtain aureoles capable of comparison it is necessary (and also sufficient) to employ always an alcohol of the same chemical nature, and marking at the same tem- perature the same degree upon the Gay-Lussac centesimal aleo- dmeter.

I have found that the alcohol which gives the best results in the new indicator is methylic alcohol or wood-spirit, of 92.5° Qay-Lussac at 15° C. (59° F.), or 93° Gay-Lubsac at 20° C. (68° F.). Practi- cally, indeed, the same aureole- heights for the same gas-percentages are obtained with mixtures of wood-spirit of different descriptions and various densities, some pure, others impure and dilute, pro- vided the mixture gives in the alcooroeter exactly the degree above mentioned.

Having had occasion to prove that the estimates of aureole-heights in the Pieler lamp by different observers vary greatly by reason of the extreme pallor of the aureoles and the diffused form of their tips, and, moreover, that the low luminosity of the alcohol-flame rendera its adjustment in pure air very uncertain, I sought to ascertain whether the addition of metallic salts to the alcohol would not, by coloring the aureoles, make them more clearly discernible. I have tried the majority of the metallic salts capable of coloring flames, such as the salts of sodium, lithium, thallium, barium, strontium, copper, boric acid, etc. All the above metals vividly color the alcohol-flame and likewise augment the brightness of the aureoles; but with all the salts tested (except cupric chloride), the alcohol-flame is much more brilliant than the aureoles, and the contrast does not augment, but for some of them even diminishes the distinctness of the latter.

158 THE DETECrriON AND MEASUREMENT OP FIRE-DAMP.

With cupric chloride, on the contrary, which is very soluble in alco- hol the difference in brightness between flame and aureole is not very great, and hence the clearness of the aureoles is considerably in- creased by the addition of a small quantity of this salt, which gives the flame a green tinge, especially at the edges, while it imparts to the aureole a fine greenish-blue color. As the cupric chloride has a tendency to pass into the state of subchloride (insoluble in alcohol) when aiming into contact with the brass of the wick- holder, it is ad- visable to add to the alcohol a small quantity of hydrochloric acid, which maintains the subchloride in a state of solution. The cotton enclosed in the reservoir has the property (and hence its employment) of absorbing and retaining any insoluble subchloride which it thus prevents from encrusting the wick. Without the cotton the alcohol becomes, after an hour, heavily charged with a milky turbidity, and the aureoles become much less clear.

As it becomes impregnated with the subchloride of copper, the cotton gradually loses its absorbent power. This condition is indi* cated when alcohol which has remained some time (an hour, for example) in the reservoir comes out turbid. The wadding must then be drawn with a small hook and replaced with a new one, the wickholder having been screwed down to the bottom, to keep its place for it in the middle of the wadding. This operation requires but a few minutes.

In consequence of the deposition of subchloride on the wick, the latter must be changed before each tour of mine-inspection. At the end of each tour, the wickholder should be removed and the reser- voir emptied of its remaining alcohol, to avoid the corrosion of the metal by the cupreous alcohol. The alcohol removed can be used again. If it is slightly turbid, it can be clarified by putting some wadding in the bottle into which it is poured.

The coloration of the aureoles is the greater, the greater the quan- tity of chloride ; but too large a proportion causes a voluminous de- posit of salt on the wick, which modifies the flame-adjustment, ren- dering it also somewhat uncertain, and similarly affecting the observations of the aureoles, by coloring too highly the luminous glow at the tips of both flame and aureole. On the whole, a satis- factory result is obtained by adding to one liter (0.22 gallon) of methylic alcohol, of the proper degree, 2 cu. cent. (0.122 cubic indies) of a saturated solution of crystallized cupric chloride in con- centrated hydrochloric acid (or about 30 drops of this solution per Ut6t The cupreous alcohol, thus prepared from normal alcohol of

-J 7.-— , T.-n..

'i

-s

X" led by

160 The Detection And Aieasureiient Of Fire-Damp.

92.5° Gray-Lussac, will mark, after the addition of the chloride, 92° Gay-Lu8sac at 15° C. The presence of this metallic salt ren- ders the aureoles perceptible at 0.1 to 0.2 per cent, of fire-damp.

From 0.5 per cent, upward, as is shown in Fig. 9, which gives the appearance of the aureoles for increasing percentages of marsh- gas, it is clearly seen that the aureole is composed of a blue cone, slightly tinged with green, the point of which is surrounded and surmounted by a whitish glimmer, which forms a sort of hood on the blue cone, and the intensity of which diminishes rapidly toward the top.

With alcohols of various kinds and without cupric chloride, these two parts of the aureole may be observed, but much less clearly. With poor alcohols, the cone tends to lose itself in the glimmer. This is one of the chief sources of error in the estimation of aure- ole-heights with Pieler lamps and ordinary alcohol by observers who have not had opportunity to see the form and tint of the aureoles given by the alcohol they are using, in an apparatus where it could be tested upon mixtures of known proportions.

d, A djuatmeni of the Flame and Aspect of the A ureoles. — The lamp, filled with alcohol up to the level, xx. (Fig. 8), of the filling-orifice, is allowed to burn for twenty to thirty minutes, with the wick high, the alcohol-flame extending at least 5 mm. (0.2 inch) above the flange of the gauze, aud the orifices in the shade, I, being closed. The object of this precaution is to cause the lamp to assume as quickly as possible its definite rimc of heating. The observer, standing in the dark, then opens the orifices of I, to let the fresh air reach the wick freely, and slowly lowers the wick by means of the regulating- screw, M, to the exact point at which all the greenish-blue light which terminates the luminous part of the flame has disappeared below the level, yy, of the flange of the gauze, which is 37 mm. (1.5 inches) above the tube zz. There still remains above the flange a pale light, about 20 mm. (0.8 inch) high, sometimes traversed by small red points, especially when the lamp is first lighted ; but this is easily distinguished from the greenish-blue light which borders the illuminating flame proper of the alcohol (Fig. 10). Thanks to this greenish-blue light, given by the copper chloride, the adjustment by different observers will vary but little, much less in any case than in the use of ordinary alcohol, the point of the flame of which it is not easy to distinguish with precision.

For the proper performance of this adjustment, which is an essential matter, the flange of (he gauze should be so held, at the height of the eye, that both sides

The Detection And Measubement Of Fire-Damp. 161

The flame being thus adjusted, it will be found that when the indicator is placed in mixtures of air with formene in i.creasing percentage, the cone and the glimmer are both augmented in height, that of the latter being about double that of the former. When there are traces of fire-damp in the air, the pale glimmer, which in

Fig. 10.

*aLe with audi id potnli

gitiuiiMtr 'Qrcoo border

pure air shows above the flange, becomes blue for a little over half its height ; but up to 0.5 percent, it is difficult to distinguish the blue cone from the glimmer. From 0.5 to 2.5 per cent, the cone is very distinct from the glimmer; and its poiut is the more easily located, since there is a sort of throttling of the glimmer at that level.*

of the flange may be in the same horizontal plane. When the adjustment by means of the screw M is believed to be correct the lamp is slowly lowered, and the eye, looking below the flange, perceives the yellow flame with its green border then the greenish-blue exterior portion, and finally the glimmer. It is then easily seen whether the greenish-blue point is approximately at the level of the flange, and its adjustment can be corrected if necessary. The lamp is then slowly raised again, to see whether the greenish-blue light disappears when the plane of the flange passes the eye ; and verification of the adjustment is completed by finding that in thb position the least raising of the wick causes the greenish-blue flame to appear above the flange.

If the cone-point is indistinct, its position can, nevertheless, be sufficiently well determined by imagining the prolongation of the oone-sides (which are always clear towards the bottom) and assuming their intersection to be the point desired.

To obtain distinct aureoles it is necessary that the wick shall not be too tight in the tube of the holder. For the adopted diameter of 9 mm. (0.36 inch) for this tobe, the wick, when flattened, should have a total width of 15 mm. (0 6 inch)

Vol. Xxii.— 11 ,,

162 The Detection And Measurement Of Fire-Damp.

At 2 per cent., the flame proper of the alcohol, brilliant yellow, at the bottom, and passing gradually to green above, begins to rise above the flange of the gauze, as a somewhat indefinite, blunted cone. Hence from 2 per cent, upwards, the aureole is composed of three part, the yellow, alcohol flame at the bottom, the blue cone in the middle, and the glimmer at the top. As the percentage in- creases these three parts increase in height.

Above 2.5 per cent, the glimmer occupies the whole top of the gauze cylinder, and the tip of the cone becomes a little less distinct. At 3 ier cent., this tip reaches the top of the gauze, the whole inte- rior of which is pervaded by the glimmer, while the alcohol flame proper, growing more and more brilliant, shows already a height of 25 ram, (1 in.) of yellow, and greenish-yellow, light.

Above 3 per cent., the blue cone expands more and more, tending to a cylindrical form, and blending progressively with the glimmer. Moveover, cone and glimmer grow gradually darker; so that at about 4 per cent, there can only be seen the alcohol flame proper, yellow at the base, with a much-rounded green top, bordered with blue. This flame at 4 per cent, is 46 to 50 mm. (1.8 to 2 in.) high ; then rises to about 80 mm. (3.2 in.) at 5.5 per cent., and mounts up to the top of the gauze for a percentage between 5.5 and 5.75, be- yond which all flame disappears from the gauze cylinder, while the mixture burns in the crown, with a very pale blue flame, which is entirely extinguished after a few seconds.

I have found that, up to 2.5 per cent., the height of the blue cone is but little affected even by considerable variations in temperature or in the amount of carbonic acid in the air; but this is not the case with the glimmer, which grows much larger and also dimmer, with increased temperature, and diminishes with the increase of car- bonic acid.

Thus, in raising the temi)erature of the gaseous mixture from 17° to 37 C. (62.6° to 98.6° F.), I have observed an elongation of only about 10 per cent, in the height of the blue cone; and the addition of 2 to 3 per cent, of carbonic acid to a mixture containing 1.5 per cent, of fire-damp did not at all affect the height of the blue cone.*

which will make it, when impregDated with alcohol, fit conveniently. The wick should be ctit off flunh with the tube at botii ends ; and the tube should be screwed down to the bottom before filling the reservoir with alcohol.

The gauze parts should be frequently brushed to clear the meshes from obstructing rust or coal-dust ; otherwise, the draught of the lamp will be impaired, and the aure- oles will be dim.

The influence of temperature is the greater, the more concentrated the aloohol.

THE DETECrriON AND MEASUREMENT OF FIRE-DAMP. 163

Moreover, since the tip of the blue cone is much easier to locate than the top of the total visible glimmer, I have adopted the rule of taking as the characteristic point in the height of an aureole for a given percentage of fire-damp, the tip of the blue cone (See A, Fig. 9, in the aureole of 1.6 per cent.). For percentages below 0.6, the blue cone not being easily distinguished from the glimmer, observa- tion of the total height of the visible glimmer will still give a very accurate result. But whatever be the percentage, the observer, in case of doubt as to the height of the blue cone, can bring near the indicator an ordinary lamp, the light of which will, by contrast, render the glimmer invisible, leaving clearly visible (above 0.25 per cent.) the blue cone, the tip of which can then be located (the light being removed a little further off). It is possible, thanks to the ad- dition of copper chloride, to see the aureoles at some distance, and to read the indicator in a chamber containing lamps; whereas with ordinary alcohol complete darkness and close proximity of the ob- server are required. This is an appreciable advantage in under- ground inspections.'*'

For proportions above 3 per cent., it is necessary to make use of the height of the alcohol flame proper, the exact determination of which is somewhat difficult; but since from this percentage upwards, the fire-damp is shown by ordinary oil lamps and its precise deter- mination has no value in practice, I have not sought to increase the accuracy of the indicator for these higher percentages, since that could only be done at the sacrifice of some precision in determining proportions below 2.6 per cent.

Since, in practice, the high percentages are generally encountered only in pockets, and, most frequently, surrounded by air containing but little fire-damp, there is a risk, on testing such recesses for gas, of passing very quickly from a low to a high percentage of gas, causing the aureole to change suddenly from a blue cone several cen- timeters high, to an alcohol flame proper of about the same height. But it is impossible to confound the latter with the true blue cone, first by reason of its form and tint, and secondly by reason of the comparatively high temperature developed in the upper part of the shield by the products of combustion in highly fiery mixtures.

It is on this account that I have adopted as normal an alcohol of mean concentra- tion, which gives satisfactory aureoles and is not much influenced by the raising of temperature.

For mixtures of air and illuminating gas, the aureoles are much less clear and the blue cones less colored than for natural or artifical mixtures of air and for- mene.

Ic

164 The Detection And Measurement Of Fire-Damp.

To determine the heights of the blue cone, the glimmer and the alcohol flames corresponding to each percentage of gas, I have used the apparatus constructed at the £coh des Mines for mixtures of air and marsh-gas, in which the air and the marsh-gas are measured separately in a meter before being mixed. The marsh-gas, prepared by the reaction of lime impregnated with soda on fused acetate of soda, and containing a certain proportion of air owing to that con- tained in the producing apparatus, was previously analyzed by the method of the limits of inflammability, and the proportions expressed in quantities of pure fire-damp. To determine the quantity of air that should be passed into the apparatus, so as to realize the condi- tions of air-supply in practice, I placed the lamp in a cylinder of glass, such as encloses it in actual use, causing the marsh-gas to reach the lamp, and allowing the outer air to enter at the bottom of the cylinder. The height of the aureole indicates approximately the quantity of air induced by the natural draught of the lamp. I thus found that, for a cylinder of 15 cm. (6 in.) diameter, the (fraught was not less than 60 liter (2.1 cubic feet) per minute; and it is under these conditions that all the determinations were made. Up to 2 per cent, of marsh gas there is, however, but little difference in the aureole heights for air-currents comprised between 6 and 80 liters (0.21 to 2.82 cubic feet) per minute, while for larger percentages the aureoles are much lower, with a too slight air supply.

A number of determinations were made with several lamps of the type shown in Fig. 8. The results, expressed graphically, show that the indicator is capable of giving very correct determinations; for the points representing the height of the blue cone observed for each proportion always group themselves at very slight distances from the curve representing the mean. Fig. 11 shows the results obtained with one lamp, the height of the shade of which is 37 mm. (1.5 in.), while the adjustment of the flame was in the first determination rather high, and in the second rather low, so as to show the limit of errors

.possible in practice. The two full-line curves, representing the mean of the results furnished by each of the determinations, almost

r coincide up to 2 per centf It will be seen, moreover, from deter-

Annalea dn Mines, 1892, srie, tome i., p. 47.

t For proportions below 3 per cent, (corresponding in the appafatoe nsed to a supply of marsh -gas of less than 0.2 liter, 12.2 cu. in., per minute), the action of the marsh-gas meter is somewhat irregular. This may explain, without the assump- tion of error in observing the height of the blue cone, abnormal results like that of high flame determination, No. 2, in Fig. II.

The Detection And Measurement Of Fire-Damp. 165

ininatioDS 12, 13, 14, 15, made with high flame, hours after lighting the lamp, and thus some two hours after Nob. 1 to U, that the adjustment is well maintained. It is safe to reckon in practice upon getting through a period of three hours determinations capable of mutual comparison.

iir

nt

oo

A

Jri

Ol

C-M

cia

LUl

C(

Nc

1)

/

11(1

tAO

cue

roe

or

to

K*0

Ami

r-

wo

♦ A

Uiu

th

[Mb

tiM

Itk

M

k

/

1k

mm

oft

ofni

ObM

rva

ton

/

(-

A

8*.

m

Kf

J"

itt

a

h

ig

.'

miS

t

—

A

—

t

&

—

z

—

%X

Pwewuc* Ar*-duBp.

The following table gives the mean of the results of the two series of tests above described, and may be considered as applicable to a correct flame adjustment, as I have verified it by several determina- tions made with the utmost practicable precision of adjustment

I have ascertained by numerous experimentfl, that notwithstand- ing the heating of the lamp when the proportions of gas exceed per cent., the height of the blue cones, on passing immediately after- wards to slight proportions of gas, is not appreciably affected.

166 THE DETECTION AND MEA8UBEMENT OF FIBE-DAlfP.

Table of Indioatima.

ii

&§ d

Height blue coi above shade

Height

above

shade.

Height

theflait

itself

ab. shad

Corresp' proporti

of fl re-dam

Rkmakk8.

(Height of Shade, S7 mm. — 1.48 in.)

Mm.

Per cent

20 to 25

Very pale reddish or yellowish glimmer. Ck>ne aifficult to distinguish from glimmer.

Do. do.

Cone tolerably well defined.

Cone well defined.

Cone very well defined.

Do. do.

Do, do.

Do. do.

top of gun.

Do. do.

Do.

Do. do.

Do.

Cone very well defined. Flame not very

distinct.

Do.

Do. do. Do.

Do.

Do. do. Do.

Do.

Do. do. Do.

Do.

Do. do. Do.

Do.

Do. do. Do.

Do.

Point of blue cone indistinct. Do.

Do.

Do. do. Do.

Do.

Do. do. Do.

Do.

Do. do. Do.

Flame very wide at base, bordered at top

with a blue fringe. Total extinction.

When the proportion exceeds 2.75 per cent, the bhie cone becomes pale aAer a few minutes when the temperature of the lamp is high, and its height slightly increases (by about -j), to again remain fixed. It is, therefore, advisable to make the observation as quickly as pos- sible when dealing with these high proportions. For the same reason it is well to make the observations with the orifices open in the screen, I, so that the fresh air may impinge directly upon the gauze crown. I have also found, that even with the orifices closed, the height of the blue cone remains constant for more than a quarter of au hour, even with proportions so high as per cent. ; in fact, it is only at a higher proportion that the lamp becomes sensibly heated, without failing, however, to return to zero after remaining for a few seconds in fresh air. Moreover, to obtain an accurate and regular flame-adjustment during an underground inspection, it is better, as a general rule, to leave the orifices open, provided there be no currents of air strong enough to extinguish the flame.

e. Experiments with the Cheaneau IndiccUor Underground. — A

The Detection And Measurement Op Fire-Damp. 167

number of indicators of this type have been put in use at several collieries since July, 1892, on the recommendation of the French Fire-Damp Commission,* and I have personally made experiments in diflerent mines, to test both the safety and the precision of the apparatus. The trials thus made up to the present time (January, ' 1893), have included 15 separate instruments of this pattern, and the results have confirmed in all respects those of my laboratory ex- periments.

As to its safety, the lamp has always extinguished itself at the end of a few seconds in explosive mixtures, without becoming more highly heated than a Mueseler lamp under the same conditions. In no case has the gauze been observed to attain, in explosive mixtures, incipient red heat

With regard to the distinctness of the aureoles, the tests have been generally very satisfactory. In some mines, where the proportion of fire-damp was always below 1 per cent., and where a little diffi- culty was experienced in locating the tip of the blue cone it was found that the alcohol employed was much below the normal alco5metric degree.

In my various mine-inspections I have always found that the lamp instantly announces the smallest changes of percentage in the sur- rounding air; and when it is suddenly transferred from an atmos- phere containing fire-damp into pure air, the return to zero is imme- diate.

I have thus been enabled to observe repeatedly that in the region of the roof of fiery chambers the passage from almost pure air to the explosive mixtures held in the recesses of the roof or between the timbers occupies less than 0.1 meter (4 inches). Hence, before ex- ploring the roof of such a chamber the observer should assure him- self, by means of an ordinary lamp, that an explosive mixture is not present, in order that he may avoid the extinguishment of the indi- cator, which often occurs so quickly as to give no time for returning the lamp into pure air at the instant of the discovery of the explo- sive mixture. It has been observed, also, that the mixture of air and fire-damp is rarely homogeneous in a working-chamber, or even in an air- way, where the percentage often increases from wall to roof.

These lamps are constnicted by M. Anguste Decamps, lampist of the Lens Co., at Li6viii, Pas-de-Calais. They are used at the collieries of the following compa" nies: Anzin, Lens,Livin, Blanzin, Houilldrtfsde St. Etienne, Ronchamp, Beftsdgea, and bj the mine- inspectors of the districts of Pas-de-Calais, Loire, Saone et Loire and ATeyron.

168 The Detection And Mea6Ubement Of Fire-Damp.

It 18 only at the end of a sufficiently long passage, the walls of which give out no gas, that the proportion of fire-damp at all points of the cross-section is the same.

As regards the percentages announced by the indicator, the results obtained have equalled in precision those of the laboratory. I will give some details of the experiments on this point because of the im- portance of the degree of accuracy to be calculated upon as a guide in the management of mine- ventilation.

In three mine-inspections I have personally taken samples at points where the percentages reported by my indicator were fairly constant over a considerable distance ; and these samples I have subsequently analyzed in my laboratory at the £jcok des Mines with the combus- tion-apparatus of M. Le Chatelier. The following is a comparative statement of the results carried to tenths of one per ceut:

Comparative ReauUs of the Chesneau Indicator and of Chemical

Analysis.

Place and i Time.

Flre-ilauip per indi- cator. Per cent.

Fire-damp per an- alysis. Per cent.

Difference of indi- cator. Per cent.

P

Anzin Co. Bmr ,

No. 2. July 7, '92.

0.4 ,

— O.l

The lamp was adjusted andergroand (546 m. 1780 ft. level) at a lighting- statioQ, in perfectly still air.

Lens Co. Shaft No. 7 Oct. 23/92

+ 0.3 -f 0.2 -hO.3

The lamp was adjusted m a current of fi-e$h air, coming directly from the down- cast air-way.

Lifevin Co. 8hafi.No.l Oct. 26, '92

+ 0.1

— 0.1

The lamp was adjusted at the mrfaee, in a dark place; the samples were taken between 400 and 500 m. (1304 and 1630 ft.) below surface.

This table shows, that in the tests made at the Lens shaft No. 7, where the lamp was adjusted in a current of fresh air, the percent- ages which it indicated were a little too high ; but, on the contrary, in the two other series of tests, the adjustment having been made in a still atmosphere, the differences between lamp-reading and analysis do not exceed 0.1 i)er cent Moreover, the fact shown at LiSvin shaft No. 1, that an adjustment made above ground gave accurate percentages at a depth of 500 m. (1630 feet), permits the inference that the adjustment may be made equally well in or out of the mine.

The Detection And Measurement Of Fire-Damp. 169

This enables us to secure in all cases the necessary conditions for ac- curate indications, which is, that the flame m\Mt altoaya be adju8ted in pure and perfectly calm ah* and in a dark place.

At the Pelissier shaft and the purap-shafl of the S<)ci6t6 des Houillires de St. £tienne, where I exhibited underground to the engineers of the company on June 25, 1892, the operation of my in- dicator, experiments were made (after a few preliminary trials) in the following manner, in order to check by chemical analysis with an improved Coquilliou apparatus the indications of my lamp, ob- served always by the same mine inspector.

Six or seven comparative tests were made for each tour through the mine. After three or four hours, the inspector became able to determine the fire-damp very accurately with my lamp. I give, by way of example, out of forty-three tests, three series taken respec- tively at the beginning,'*' middle, and end :

Practical Testa of the Oiesneau Indicator.

Fire-damp, per Coquillion.

Di (Terence

No of Test.

per iudlcator.

of indicator.

Per cent.

-f 0.2

-hO.4

-f 0.7

-f 0.1

-h 0.2

— 0.1

— 0.2

+ 0.1

— 0.1

-h0.1

0.31 :

-f 0.1-

+ 0.1

— 0.1

After these experiments, the concordance of results was deemed to be so well established that constant checking by analysis might be dispensed with, and the management was content to make, at con- siderable intervals only, comparisons this kind, which showed

J,\, the first day when the apparatus was used by the inspector.

170 The Detection And Measurement Op Fire-Damp.

that the indicator-reports of fire-damp never varied more than 0.1 per cent, from the results of analysis.

I will give in addition the following comparative statement, in which the indicator-observations were made by the engineers of the mines, after a few days' apprenticeship, and the analyses of the cor- responding samples were executed at the £oole des Mines with the Le Chatelier combustion-apparatus:

Further Comparative Testa of the Indicator,

Place.

Anziii Co. Rennrd

pit, No. 2.

0.7*

Lievin Co. Shaft No. 1

E

O.l

-f- 0.02 -h 0.06 -h 0.10

Remarks.

The adjtistroeot was a little too low, for according to the engineer the total lumin- osity in the first test was bat 15 mm. (0.6 in.) high ; whereas it should be at least 20 mm. (0.8 in.) in pure air.

Observations made with alcohol of 94° Gay-Lussac

It follows from all these comparative trials, which actually exceed two hundred, that the mean difiPeiience of the percentages given by the new indicator, from those furnished by analysis, does not exceed 0.1 per cent., when the adjustment is made according to the direc- tions given above. The accuracy of this apparatus, after a little practice in its use, is therefore substantially the same as that of the laboratory apparatus.

It thus appears to satisfy, in precision and safety, the reasonable requirements for a portable fire-damp indicator for use in mines, and, with the addition of either of the two forms of analytical labora- tory apparatus '(described on pp. 125 and 132), to check from time to time the indications it gives, it constitutes a simple plant for secur- ing the object stated at the beginning of this paper, namely : To follow with precision the variatioTia in the percentage of fire-damp at all points in the mine, so as to be able to regxdaie a4)cordingly the ven- tilation and exploitation in every part,

Indicaons noted as indistinct.

LEAD- AND ZINODEPOfilTS OP THE MISSISSIPPI VALLEY. 171

TEE LEAD- AND ZINaDEPOSITS OF THE MISSISSIPPI

Valley,

BY WALTER P. JENNEY, E.M., PH.D., DEADWOOD, SOUTH DAKOTA. (Chicago Meeting, being part of the International Engineering Ck>ngreflB, Aogust, 1893.)

Introduction.

An investigation, condacted by the author, was bun in Sep- tember, 1889, by the United States Geological Survey, having for its object the study of th questions bearing upon the occurrence and manner of formation of the deposits of lead- and zinc-ores in Mis- souri and the adjoining States. The field-work was completed in Deceml>er, 1891. During the first season the State Geological Sur- vey of Mi.<souri furnished an assistant, and in other ways co-oper- ated in the research within the boundaries of that State. In order that the ore-deposits of the southwest might be compared with those of other sections of the Mississippi valley, an examination was made of the lead- and zinc-mining area of Wisednsin-Iowa, and of the argentiferous lead-mines of the region extending westward from Hot Springs, Arkansas, to Indian Territory. In the discussion of the general geology, and in comparisons between the ore-deposits of the Mississippi valley and those of the Rocky Mountains, the writer has drawn largely from his personal experience, having been engaged for many years in professional work as a geologist and mining engineer in that region. The Director of the U. S. Geological Survey and the State Geologist of Missouri have kindly consented to the publi- cation of the present paper.

While giving the more important economic results, an endeavor has been made to set forth the unity of plan which appears in the formation of the deposits of lead and zinc of the Mississippi basin, and to 0how their relation to the argentiferous ores of the Rocky Mountains — that the ores of gold, silver, and mercury, and also of copper, antimony, zinc, and lead, have a universal common origin irrespective of the geological formations in which they occur.

172 lead- and zinc-dep06it8 op the mississippi valley.

Location, Topography, and Structure of the Lead-

AND Zinc-Mining Regions op the

Mississippi Valley.

The level surface of the country drained by the Mississippi river, lying between the Alleghaniesand the plains along the eastern slope of the Rocky Mountains, is broken by a number of remarkable areas of uplifl. At the north, stretching from the Great Lakes southerly, covering the State of Wisconsin and contiguous sections of Iowa and Illinois, occurs a vast promontory of land, described by Chamberlin under the title of the Wisconsin Island.*

In Ohio, Indiana, and central Kentucky is situated the Cincin- nati anticlinal, made known by the labors of Newberry and later in- vestigators. The elevated region of southern Missouri and north- western Arkansas, reaching from the confluence of the Missouri and Mississippi rivers southwesterly to Indian Territory, has been named by Broadhead the Ozark upliftf

The most southerly of these great areas 6f upheaval, designated by Branner as the Ouachita uplift, forms an easterly and westerly range, extending through central Arkansas and Indian Territory from the vicinity of Little Rock, Arkansas, nearly to the pan-handle of Texas.t

The deposits of lead- and zinc-ores of the Mississippi valley are associated with certain of these uplifted areas. The lead-region of the upper Mississippi ijj located in the southern part of Wisconsin and adjoining portions of the States of Iowa and Illinois, in the southwestern section of the Wisconsin Island. The lead-mines of southeastern Missouri, and the deposits of lead and zinc of the south- western part of the State and of the extension of that field into Kansas, Indian Territory, and northwestern Arkansas, are all in- cluded within the boundaries of the Ozark uplift. Quartz veins, carrying argentiferous galena associated with zinc-blende, occur at intervals in the Ouachita uplift from the region about Hot Springs and Little Rock, Arkansas, westward into Indian Territory.

The Cincinnati anticlinal is exceptional, and from some cause ap- pears to be barren of deposits of lead- and zinc-ores. In Henry county, Kentucky, in the extension of this uplift, small deposits of

Oeohgy of WiscoMin, vol. i , T. C. Chamberlin.

t The Geological History of the Oxark Uplift," by G. C. Broadhead, Am. QtoUh- gift, Jan., 1889. X JUpariM of the Arkanms Oeol, Survey, J. C. Branner.

Lead- And Zixc-Deposits Of The Mississippi Valley. 173

lead-ore ooeur. Galena is also found in Hmited quantities, associated with fluorspar, in southern Illinois, in the vicinity of Rosiclare. These deposits have lately been investigated by S. F. Emmons, of the United States Geological Survey, and are made the subject of a paper in our Transaotiona.* This district extends across the Ohio river into the contiguous counties in Kentucky, and appears to form an outlier of the Ozark uplift.

The Ozark Uplift — In outline rudely quadrilateral, this elevated rion covers the southern half of Missouri and nearly all of that portion of Arkansas lying north of the Arkansas and directly west of the Black river, extending for a short distance over the contiguous corners of Indian Territory and Kansas, an aggregate area exceeding 50,000 square miles. The axis is approximately a line prolonged from the Mississippi river, at a point equidistant from St. Louis and Cairo, to Grand River, at the junction with Spavinaw creek, in the northeastern part of Indian Territory. The course of this axis is south 65 west; in this direction the ex- treme length of the Ozark uplift; is nearly 320 miles, the breadth varying from 200 to 250 miles.

The structure is that of an elevated plateau, bounded by mono- clinal folds; in the central areti, the sedimental formations are nearly horizontal, dipping in the marginal belt, at gentle angles, away from the uplift, until the inclination of the strata is lost in the surround- ing level country. In past geologic times, the Ozark area has, at several periods, formed an island in the ocean that extended over this portion of the Mississippi basin ; now, it is circumscribed by rivers, conforming in their course to the palsdo-littoral zone. On the northeast flank, the Missouri river flows in the monoclinal fold from Boonville, Missouri, to its junction with the Mississippi river; in the interval between St. Louis and Cairo, the Mississippi follows the general direction of this' flanking fold, a small portion of the uplift lying on the east bank of the river in the State of Illinois. From the source, in southeastern Missouri, to its junction with White river in Arkansas, Black river conforms to the southeast marginal belt. On the south, the Ozark uplift is bounded by the valley of fhe Arkansas river, and on the extreme west by the Grand or Neosho river in In- dian Territory and southeastern Kansas. f

The elevation of the Ozark uplift is from 400 to 2000 feet above

Flnorspar DepositA of Southern Illinois," Trans, xxi ,31. t *'The Geological History of the Ozark UpHH," by G. C. Broadhead, Am. Oeohgud, Janaary, 1889.

174 Lead- And Z1Nc-Deposit3 Op The Mississippi Valley.

the sea; the oentral p1a(%aa, comprising the greater area of the upheaval, having an elevation of 900 to 1600 feet. No complete topographical survey of the region has been made ; it is not prob- able, however, that any point attains an elevation exceeding 2100 feet. On many of the published majKS of Missouri, the Ozark moun- tains are somewhat indefinitely marked, traversing the southern por- tion of the State. It is, perhaps, best that the designation of any portion of this elevated region as a mountain-range should be dis- continued, for even the most hilly parts bear little resemblance to mountains, and would attract slight notice but for the contrast presented by the prevailing level surface of the Mississippi valley. The surface of southwest Missouri is gently rolling, with numer- ous high prairies and flat-topped divides. In the southern part of the State, and in northern Arkansas, the Oeark area is more hilly and broken. This topography has resulted from denudation by the present system of streams, modified somewhat by the char- acter of the geological formations covering the surface. All the streams that drain the region are still corrading their beds, and, in the southwest, they have but little modified the plateau-character of the upheaval.

The sections of the Ozark area drained by the tributaries of the Arkansas and White rivers, are generally well timbered ; the hills are covered with oak of good size, and the bottom-lands support a heavy growth of hickory, sycamore, and gum. In northern Ar- kansas, tracts of yellow pine {pimca minis) occur on the tops of the higher divides. The belt of high prairies extending from Spring- field, Missouri, southwest to Galena, Kansas, is in most part desti- tute of trees, or, at best, supports a stunted growth of oak.

The mining-fields of the Ozark uplift, designated by Whitney as the lead-region of the Lower Mississippi,* comprise the lead- district of southeast Missouri and the lead- and zinc-fields of the southwest. The southeastern Missouri lead-field is of limited ex- tent, embracing portions of the counties of JefiTerson, Washington, St. Francis, St. Genevieve, and Madison, covering an irregular area in the nortlieastern part of the Ozark upheaval. The lead- and zinc-mining district of the southwest covers that section of the country drained by the tributaries of the Missouri and Mississippi rivers where the States of Kansas, Missouri, and Arkansas, and the Indian Territory, corner nearly together.

Repori on the Upper Miasistippl Lead-Region, J. D. Whituejr, 1S62.

Lead- Akd Zinc-Dep0Sit8 Op The Mississippi Valley. 175

In Kansas, only the southeastern part* of Cherokee county, the extreme southeastern corner of the State, belongs to this mineral area. In Missouri are included the counties of Jasper, Newton, McDonald, Barry, Lawrence, Dade, Stone, Taney, Christian, Doug- las, Greene, Webster, Wright, Polk, Dallas, Camden, and Morgan, constituting the southwestern section of the State. In Arkansas, this mineral region covers the northwestern tier of counties, embrac- ing Benton, Carroll, Madison, Boone, Newton, Marion, Searcy, Baxter, and Stone counties, with some outlying deposits in Law- rence county. The field also extends into the northeast part of In- dian Territory, covering, in the Cherokee nation, the area included between the Grand or Neosho river and the west boundary of Mis- souri and Arkansas. The total area of the southwest mining region exceeds 20,000 square miles; the most productive section comprises Lawrence, Jasper and Newton counties in Missouri, and Cherokee county in Kansas.

The Ouachita Uplift, — This closely-associated upheaval has been investigated only in the eastern part, within the boundaries of the State of Arkansas ; its great extension in Indian Territory is com- paratively unknown and unprospected. Commencing in the vicinity of Little Rock, Arkansas, it stretches, with a generally westerly course, through the central part of the State, and across Indian Ter- ritory, terminating before reaching the pan-handle of Texas. The extreme length of the range is nearly 450 miles ; it forms a com- paratively naiTOw belt, averaging, in the Arkansas section, a breadth of from 40 to 60 miles.

In structure, the Ouachita uplift differs from all the other up- heavals in the area drained by the Mississippi river, and resembles in miniature a strongly-folded mountain-range. The Wisconsin, Cincinnati and Ozark uplifts are more in the nature of elevated plateaus of the Uinta type of Powell. The Ouachita uplift is not composed of a single flat-topped arch, but of a number of parallel anticlinal ridges and synclinal troughs sharply folded, the folds corresponding in direction with the general course of the range. This folding of the strata is so great throughout central-western Arkansas that the sedimentary usually occur set on edge or dipping in steep angles, in strong contrast with the horizontal posi- tion of the strata in the Ozark uplift. This difference in the struc- ture of the Ozark and Ouachita uplifts becomes the more notable from the narrow interval of separation ; a narrow trough filled with the Coal-Measures and now occupied by the valley of the Arkansas,

176 Lead- And Zinc-Deposits Op The Mississippi Valley.

haviug a width of from 30 to 50 miles, divides these two great up- lifts. Contrasted with the Ozark area, the Ouachita uplifl is char- acterized by a much more hilly and mountainous topography and a greater elevation, the higher peaks attaining an altitude of 2500 to 2850 feet above the sea.

Central Arkansas is well timbered, especially in the western part, in the section adjoining Indian Territory. The ridges are covered with a heavy growth of oak and yellow pine. Along the river- DOttoms, white oak, hickory, black walnut, gum aud sycamore at- tain a very large size.

Deposits of silver-, lead- and zinc-ores are found grouped in small mining-districts irregularly scattered through the extent of the Ouachita uplift in Arkansas. In the continuation of this range in Indian Territory, ores of silver, lead, zinc and gold are known to occur at a number of localities, but owing to the reservation of vast tracts for the use of various Indian *ribes and to the general un- settled condition of the country, the mineral wealth remains unde- veloped.

The Wisconsin Uplift, — Of the great area included in the Wiscon- sin Island, it is necessary to consider here only the lead-region situ- ated in the southwestern part, comprising the northeast portion of Iowa, bordering on the Mississippi river, the extreme northwestern corner of Illinois, and a large section in southwest Wisconsin. The region of the productive mines covers a tract nearly circular in out- line, and from 60 to 80 miles in diameter; its total area within the three States is 3000 square miles.

In the southwest section of the Wisconsin uplift, the bedding of the strata is nearly horizontal, with a slight prevailing dip to the southwest. The lead-region lies from 580 to 1300 feet above the sea, the higher mounds reaching an altitude of 1000 to 1700 feet. The surface is that of a gently undulating plain, which results from the wide-spread erosion-carving of the horizontally-bedded rocks. Prominent features in the topography of the mining- region are the conical and flat-topped mounds that rise above the general surface to an altitude of 100 to 400 feet, as remnants lefl by the denudation of the sedimentary formations which once extended in an unbroken sheet over the area. Extensive tracts in the lead-region are treeless, though some timber of value is found in the valleys and on the bot- tom-lands of the streams.

The central and northern sections of the Wisconsin uplift, beyond the limits of the mining-region, are more elevated and exceedingly

Lead- And Zino-Dep06It8 Op The Mississippi Valley. 177

rough and broken in sarfaoe, comparable to the mountainous distxieta of the Ouachita and Ozark upheavals.

General Geology op the Areas op Uplift op the Mississippi Valley.

The lead- and zinc-mining regions are closely related in their local geology to that of the uplifts in which they occur. Viewed in its broader and more general relations, the geology is that of the Mis- sissippi valley, modified, it is true, by peculiar conditions that have locally supervened, but forming a part of the great geologic domain which extends from the Alleghanies to the Rocky Mountains. The Ozark and Wisconsin uplifts date back to the Archaean age, when they formed a portion of the earliest land of North America, con- temporary with the Labrador Continent, the Allegheny Mountains and the Black Hills of Dakota, and other outlying islands and spurs of the Rocky Mountain chain. A more recent origin has been at- tributed to the Ouachita uplift and the Cincinnati anticlinal : the oldeHt strata exposed to view are of Lower Silurian age. Possibly, these upheavals may have begun with disturbances reaching back into Archaean time, it being a general law, that disturbed areas have been subjected to recurrent periods of dynamic action from the most remote geologic eras.

The Ozark Uplift. — The more important events in the geologicali history of the Ozark area, having a bearing upon the occurrence of the ore-depoeits, may be briefly sketched. The Archaean rocks- were covered by the advancing ocean in the later Cambrian age ;; during the long interval between the Cambrian and the close of the Subcarboniferous period, the area of the Ozark uplift was an island! or a group of islands, possibly completely submerged at times beneath, the sea. With the opening of the Coal- period, the Ouachita, Ozark, and Wisconsin islands were united by the elevation of the oceauf- floor and formed a part of the continental area extending, fnom northern Canada to the Gulf States, and from the eastern slope of the Alleghanies to Kansas and Indian Territory. Encircling marshes of the Coal-period surrounded the Ozark area, the southeast shore alone bordering on the Gulf of Mexico.

From the Carboniferous period to the present day, this continental area has been above the sea, so that Mesozoic and Cenozoic times are not represented in the Ozark uplift by any formations of marine ori- gin. The deposits of the Glacial period and the Drift appeau to have

VOL. xxiL— 12 C"nirn]o

? IvL

178 Lead- And Zinc-Deiosits Of The Mississippi Vallet.

reached their southern limit near the Missouri river and not to have extended over the Ozark uplift.

Throughout the great central area of the Ozark uplift, the Palaeo- zoic strata are nearly horizontal, having been elevated without any considerable disturbance in the bedding. Along the marginal zone, the sedimentary strata form a gentle monoclinal fold, and dip radi- ally away from the central elevated plateau. The movement of elevation has been accompanied, especially in southwestern Mis- souri and southeastern Kansas, by local faulting of the beds in this marginal area; the more pronounced faults formed at that time being either parallel to the axis, or tangential to the flanks of the uplifl. Many examples of this form of upheaval occur in the plateau-regions of the Rocky Mountains, characterized by a central elevated plateau, or flat-topped arch, bordered by monoclinal folds along the flanks; the marginal folds becoming faults which dislocated the beds, where the simple flexure of the strata was inade* quate to compensate the extent of the vertical movement. The effect of the Ozark uplift is seen in the disturbance of the rocks through- out eastern Kansas, where the formations have a general northwest dip away from the upraised area. In the northern counties of Arkansas, a general southerly and southwesterly dip of the strata is observed, approximately normal to the marginal belt of the Ozark uplift in that section.

Different geological formations are productive of lead and zinc in the several districts in the Ozark area. In southeastern Missouri the great lead-mines of Bonne Terre and Mine la Motte occur in Cambrian limestone. In the same regions the Lower Magnesian limestones of Oalciferous age have contained deposits of lead- and zinc-ores which have yielded heavily in the past. The zinc-mines of Northern Arkansas, in the southern part of the Ozark uplift, also occur in the Lower Magnesian limestone. The upper beds of the Subcarboniferous formation, designated as the Cherokee limestone and Seneca chert, carry the large deposits of lead- and zinc-ores in the southwest. In the area covered by the Subcarboniferous rocks, are situated the mining-camps of Aurora, Webb City, Granby and Joi>- lin in southwest Missouri, and Galena in Cherokee county, Kansas.

7%e Ouachita TJplijL — Before discussing the closely-contiguous

Ouachita uplift, it should be noted that many errors occur in the

older published statements of the geology of the central Arkansas

on. Owen regarded the granitic outbursts in the vicinity of Lit-

lock and Magnet Cove, Arkansas, as of Archeaan age, forming

Lead- And Zinc-Deposits Of The Mississippi Valley. 179

outliers of the granitic rion of soutbeastero Missouri. Dr. J. C. Brenner has shown that these rocks are of eruptive origin, certainly later than the close of the Carboniferous, and possibly even as recent in intrusion as post-Cretaceous or early Tertiary. Dikes of igneous rock (peridotite) are described by Dr. Brenner in Pike county, Arkansas, on the southern flank of the Ouachita uplift, intruded through beds of Cretaceous age.*

The Archsean grenites and porphyries of southeastern Missouri extend within forty-five miles of the Arkansas-Missouri boundary, and an out-lying grenitic mass appears on Spavinaw Creek in Indian Territory, about thirty-five miles from the Arkansas line; other and similar outliers of the Archaean rock are exposed at various points in Indian Territory and Texas. No occurrence of rocks of ArchsBan age has been found anywhere within the State of Arkansas.

The marked contrast of the complex structure of the Ouachita uplift with the simple formations of the otlier areas of upheaval in the Mississippi valley has been noted. This narrow belt of rough and hilly country, stretching from central Arkansas westerly to Texas, bears many points of resemblance in structure to some of the greater mountain-renges. A series of sharply compressed anticlinal folds, heavily denuded, expose the strata upturned on edge in par- allel ridges, having a general east and west trend. The plication of the beds involves the whole series of Palseoeoic rocks, many thousand feet thick. No less distinct are the geological formations of this belt from those of other sections of the Mississippi valley; the rocks of this highly disturbed region are greatly metamorphosed, and the sedi- ments from which certain of the be<ls have been formed appear to have differed in character and chemical composition at the time of their deposition. The prevailing formations of the Ouachita uplift are metamorphosed shales and quartzites, with novaculites — peculiar flint-like rocks occurring in massive strata and composed of nearly pure silica. These novaculitee, though much older, are closely re- lated in composition, and probably in origin, to the great ohert-beds of the Subcarboniferous formations of the Ozark region.

The age of the geological formations included in the Ouachita uplift has been but imperfectly determined. Strata of limestone and shale, carrying fossils of Trenton age, have been identified near the middle of the exposed series of rocks ; and a few species of grapto- lites recognized as Calciferous forms have been found in the slates as-

Annual Report Oeol, Sur, Afkansaa, 1890, vol. ii., p. 377.

180 Lead- And Zinc-Deposits Op The Hissis8Ippi Valley.

sociated with the novacah'tes. Along the northern flank of this np- lift the rocks of the Coal-Measures are aptumed by the movement effecting the elevation ; the southern littoral belt is formed by Creta- ceous limestones resting unoomformably on the denuded and folded strata of the Palseozoic rocks. Evidences of uncomformity were ob- served by the writer between the Trenton formation and the under- lying novaculites and also between these novaculites and still older quartzites and schists; so that it is not improbable that the exposed section may reach back to the earliest Silurian, possibly to the Cam- brian, though no palsBontological evidence has been obtained from these older rocks whereby their age could be determined.

It is difficult to correlate the formations of the Ouachita and Ozark uplifts, although the interval of separation is so small. This correlation becomes the more difficult, not only from the metamorph- ism of the rocks but from the general absence of fossils in the Pal- ozoic strata of the Ouachita area. The Archsean granites of the Ozark region are not represented in the Ouachita uplift ; the older quartzites and shales at Hot Springs, Arkansas, may possibly be of Cambrian age, but bear little resemblance to the limestones and sand- stones of that horizon occurring in southeast Missouri. The great magnesian limestones, regarded as Calciferous in age, which extend over so large an area in southern Missouri and northern Arkansas, with a thickness of 2000 to 3000 feet, have not been recognized in the Ouachita uplift, although the exposures of these rocks in the northern part of the State, included in the Ozark area, are separated from this uplifl simply by the valley of the Arkansas river with its narrow belt of Coal-Measures. It is not until the Trenton lime- stones are reached that any correlation can be made ; limestones of this age having been recognized by Prof. H. S. Williams in tde vicinity of Batesville, Arkansas, and at other points near the Mis- sissippi river in the littoral belt of the Ozark uplift. No rocks of Devonian age have been found in this section ; and the great Subcar- boniferous formations of Missouri, embracing the limestones of the Chouteau, Burlington, St. Louis and Chester epochs, aggregating more than 600 feet in thickness, if represented at all in the Ouachita uplift, occur as coarse sandstones nearly barren of fossils.

Fissure-veins having a quartz gangue and carrying argentiferous lead-ores associated with ores of zinc, and in some instances of anti- mony, occur at intervals in the Ouachita uplift, in the vicinity of Little Rock and westward to Indian Territory. These veins traverse clay-slates and quartzites, probably of Calciferous or Trenton age.

Ic

Lead- And Zinodep08It8 Op The Mississippi Valley. 181

The Wucowsin Uplift, — More closely, apparently is the geology of the Wisconsin Island relatefl to that of the Ozark uplift. Of simi- lar structure and formation each of these insular areas began in later Cambrian time with a nucleus of Archsean rocks, gradually extending in area by the deposition of broad littoral belts of Palaeo- zoic sediments ; the greatest growth taking place in westerly and southerly directions.'*' The geological formations of the Wisconsin uplift are capable of a more or less perfect correlation with those of the Ozark region ; when compared, however, with respect to the oc- currence of the ores of lead and zinc, it is seen that the special geological formations that are ore-bearing in a marked degree in either of these insular elevations are barren in the other, or carry but small deposits.

The magnesiau limestones (Calciferous ?), in the Wisconsin-Iowa rion are known to carry lead-ores in a few localities ; while in Missouri and Arkansas, over a vast area of limestones of this age, numerous deposits both of lead and zinc occur. The principal ore- bearing formation of Wisconsin-Iowa, the Trenton limestone, covers a large section in the southwest part of the Wisconsin Island, but occurs in the Ozark uplift only in comparatively limited areas along the east and southeast marginal belt, and is not known to carry ores of lead and zinc. The Subcarboniferous limestones which form the productive ore-horizon in the mines of southwest Missouri do not occur anywhere in the elevated region of Wiscon- sin. Limestones of this age cover an extensive area in central Illi- nois, extending into northeastern Missouri and southeastern Iowa, in the broad interval between the Ozark and Wisconsin elevations, but they appear to be everywhere barren.

A long promontory of Archsean rocks stretches south westward from the Canadian territory, north of Lake Superior, nearly across the state of Minnesota; and in Silurian time it became united to the Wisconsin uplift. This region forms a connection between the geologic province of the Great Lakes and that of the Mississippi valley. Nearly three hundred miles from the lead-region, on the opposite shore of the Wisconsin Island, occur the mines of native copper of the Upper Peninsula of Michigan with the associated in- trusive rocks.

A large section of the Wisconsin upHfl, including the lead-region, has never been covered by the ice of the Glacial period or by the de-

Geology of Wiseonsinf vol. i.

182 Lead- And Zinc-Deposits Of The Mississippi Valley.

posits of the Drift, though entirely surrounded by glaciers and by the great sheet of Drift spreading over the Upper Mississippi val- ley. This driftless area has been investigated by Chamberlin and Salisbury, who, besides other causes, topographic and climatic, ascribe the exemption of this track from all glactation to the diversion of the glaciers by highlands on the north.*

The region covered by the Cincinnati anticlinal, or uplift, in Ohio, Indiana and Kentucky was not visited, as the deposits of lead- and zinc-ores associated with this upheaval are small and are not worked at the present time. The structure of the Cincinnati axis is that of a broad and extremely flat-topped arch, the elevation being accompanied by comparatively little disturbance of the strata. This uplift is probably more recent in formation than the other elevated areas in the Mississippi valley; the oldest rocks exposed are of Trenton age, the region having become dry land at the close of the Juower Silurian. t Although, from their character and chemical composition, the rocks that cover this area would be looked upon as favorable ore-horizons, no considerable deposits of lead or zinc have been found anywhere within the limits of the elevation. The scattered occurences of lead-ores in the southern extension of the Cincinnati anticlinal into Kentucky are small and usually unworkable. The Drift in Ohio and Indiana reached its extreme southern limits near the Ohio river, covering all the Cincinnati uplift, with the exception of the area in the State of Kentucky.

Tlie Relation of these Areas of Uplifts to tJie Rocky Mountains, — There is a marked parallelism between the early geological history of the Ozark and Wisconsin uplifts and that of the Black Hills of Dakota, the hilly region west of Fort Laramie, in Wyoming, and other outliers of the eastern slope of the Rocky Mountains, in that the Cambrian formation rests upon the denuded edges of the up- turned metamorphic Archaean strata, and is overlaid by the Calcife- rous rocks. The Upi)er Silurian and Devonian rocks are either of local occurrence or are altogether wanting, and the Subcarboniferous limestones are deposited unconformably upon the older Palaeozoic beds.

A similar section was described by the writer as occurring in the Organ Mountains in western Texas, where, in an ascending order,

♦ "The Driftless Area of the Upper Mississippi,'' by T. C. Chamberlin and R. D. Salisbury.— 5'ix/A Ann, Report U. S, OeoL Survey, 1884-6, p. 205.

t The Trenton Limestone as a Source of Petroleum and Natural Gas in Ohio and Indiana," by Edward Orton.Eighth Ann. Report U. S, OeoL Surv., p. 674.

Leaiv And Zinodep08It8 Op The Mississippi Valley. 183

the exposed strata are ArchaBan granite and Cambrian quartzite, suc- ceeded by limestone of Calciferous and Trenton age, overlain un- ooDformably with massive limestones of the Carboniferous.* At the opening of the Carboniferous the parallelism fails ; for the Rocky Mount&in region remained an open ocean, depositing limestones of great thickness, while the Mississippi basin was elevated above the sea, and deposits of coal were formed in it where local conditions were favorable. The greater portion of the Rocky Mountain range did not emerge from the ocean until the close of the Cretaceous, when it became a part of the continental area.

The general resemblance in formation and in the succession of the asBociated sedimentary rocks in these elevated areas, from the Black Hillfl, along the line of outlying islands and spurs of the eastern border of the Rocky Mountain system, to the Mexican boundary, and from the Great Lakes through the Wisconsin, Ozark and Ouachita uplifts to central Texas, is evidence of a uniformity of origin, a cor- respondence in the dynamic movements which have taken place, and a relation that has been maintained with regard to the general prevalence of the ocean and to the deposition of the sedimentary strata over this extensive region, from Archsean to Carboniferous time.

Dynamic Geology of the Areas op Uplift, Considered

WITH Reference to the Formation of the

Ore-Deposits.

The location of the deposits of lead and zinc, the origin of the mineral-depositing solutions, the means by which these solutions have been introduced into the strata, and the formation and occur- rence of the ores all appear, upon examination, to be dependent upon the dynamic disturbances which have taken place in the past geologic history of these elevated sections of the Mississippi valley. The deposition of the ores of lead and zinc in the Ozark area and Wis- consin uplift has not been accompanied by igneous disturbances or by intrusions of igneous rocks within the mining-areas. This is the more remarkable, as the deposits of nearly all the mining-regions of the globe conform to the law announced by Humboldt, That the deposits of the precious metals and of leady zinc and merctiry, are usua/ly associaied with intmsions of igneous rocks.'' The igneous

" Notes on the Geology of Western TexM,"— W. P. Jenney, Am. Jour. Sei, BeT.f vol. viii., p. 25.

184 Lead- And Zinc-Deposits Op The Mississippi Valley.

rocks of the Archaean area of southeastern Missouri, included within the Ozark uplift, are far older than the earliest sedimentary deposits carrying lead- and zinc-ores. The eruptive rocks of the northern shore of the Wisconsin Island, in Upper Michigan, are all regarded by Van Hise as pre-Cambrian,* and therefore cannot have influenced the formation of the deposits of lead and zinc in the Palseozoic rocks of the southern part of the uplift. The presence of igneous rocks and of numerous evidences of intense igneous action in the Ouachita uplift and the relation of that upheaval to the Ozark area have been noted.

The more pronounced dynamic disturbances appear to have taken place at two distinct periods : the earlier, accompanying the conti- nental elevation which involved all this portion of the Mississippi basin, occurring at the close of the Subcarboniferous period ; and the later one in post-Cretaceous and early Tertiary time, culminating in the elevation of the Rocky Mountains, and characterized by widely extended outbursts of eruptive rocks and by igneous actions that Eeero to have been the direct cause of the Assuring and faulting of the strata and the formation of the ore-deposits, f

The result of this investigation of the deposits of lead and zinc in the Mississippi valley has made it possible to announce the general law that all workable deposits of ore occur in direct association with faulting fissures traversing the strata and with zones or beds of crushed and brecciated rock, produced by movements of disturbance. The undis- turbed rocks are everywhere barren of ore.

While it is true that the ore-deposits are thus associated with areas of disturbance and fissures faulting the strata, so that it may be said that no ore-deposit occurs without a crevice or fissure in the rocks through which the ore-depositing solutions were introduced, it by no means follows that all fissures, even though they fault the rocks, are connected with the ore-deposits. In the barren sections of the mining- districts many disturbed areas occur where no action of ore-deposition appears to have taken place, and this is equally true of mining- regions in other parts of the world.

For the occurrence of ore-deposits, it is requisite, not only that the strata should be disturbed and faulted, but that the fissures should ))enetrate to and form open channels connecting with the zone of sup- ply of the ore-forming solutions, which may be located at a consider-

Joum. of GeoLy vol. i., 1893.

t For the discussion of this aubject see the section on The Time at which the Formation of the Ore- Deposits Occurred," post, p. 217.

Lead- And Zinc-Deposits Op The Mississippi Valley. 185

able depth in the earth ; also that the pressure should be sufficient to force the mineralizing solutions to the surface; that the solutions should contain metallic substances in adequate quantity, and that the physical and chemical conditions should be such as to permit ore- deposition. Through the absence of any of these conditions dis- tricts otherwise favorable for ore may remain unmineralized.

IThe Ore-Bearing Fissures. — There are many evidences which can- not be set forth in detail, that the fissures associated with, the ore- bodies have furnished the channels through which the mineral- depositing solutions were introduced. AmongjHich evidences are: the mineralizing action is often seen to have been diminished as the distance from the fissures increased ; hard, impervious bars of rock have acted as dams to the solutions, causing certain areas of otherwise favorable rock to remain unimpregnated and barren ; the minerals also are frequently observed to have been deposited first near the walls and in the outer or lower }>arts of the ore-bodies, while the minerals of later formation occupy the central and upper parts of the deposits. The master-system of fissures is frequently traversed by a belt of parallel cross- fissures; and at such crossings or interseclions, the richest and largest deposits of ore are frequently located, a result due to the cross-fissures aiding to keep open the channels through which the mineral-depositing solutions entered, and more completely shattering and brecciating the beds, thus afibrding free circulation for the solutions.

No general law is apparent with respect to the course of the ore- bearing fissures in the Southwest, though a tendency is noted in a number of districts for the fissures to form in systems that are either parallel to the axis of the Ozark uplift (N. 60° to 80° E.) or rudely parallel to the marginal belt in that special section. In the vicinity of Joplin, Missouri, the course of the more prominent fissures is usually included between N. 30° W. and N. 30° E. At Aurora, Missouri, the belts of fissures have courses N. 20° E. to N. 30° E., and E. and W. The course of the fissures seems to have had but little influence in the mines of the Southwest upon the character or size of the ore-deposits. Locally the belts of ore-bearing crevices having a given course maybe the more productive; but in other districts the rule will not be found to hold good. In the Wiscon- sin-Iowa mining region the productive crevices have a general E. and W. course: the N. and S. systems are unmineralized, or carry ore-bodies of small size only.

The fissures through which the ore-depositing solutions appear

188 Lead- And Zixc-Dep08It8 Op The Mississippi Valley.

to have been introduced are usually nearly vertical, rarely dipping at angles of less than 60 degrees; and they vary from a narrow fault- ing-seam less than an inch wide (the smooth and often polished rock-surfaces of the faulting-plane being, in places, in close contact) to prominent crevices measuring 2 to 5 feet between the walls, filled with angular breccia formed by the attrition of the cheeks of the fissure moving upon one another. The brecciated material filling the fissure is often observed to have been more or less altered and mineralized by the waters traversing the crevice; and the included rock-fragments are, in some instances, silicified. Longitudinally, the fissures have an extent which, in most localities, cannot be deter- mined ; the underground workings of the mines seldom permit the inspection of more than a few hundred feet along the course of the main belt of fissures. At the lead-mines of Bonne Terre, in south- eastern Missouri, the stopes of the mine follow a strong belt of par- allel fissures for 3000 feet ; and at Mine La Motte, the master fault- ing-fissure is opened at points along its strike for a still greater dis- tance.

In some localities, the fissuring of the strata has been accompanied by only a slight displacement of the rocks, notwithstanding which, the associated ore-deposits are large. The greatest vertical displace- ment of the strata, by faulting-movements, in the southwest region, occurs in Cherokee county, Kansas, in the area covered by the Cher- okee limestone, where the fissures that traverse the ore-deposits are observed to cause a vertical displacement of the strata of a few inches up to 3 or 4 feet, and the combined faulting-movements, due to cer- tain belts of parallel fissures, give aggregate vertical throws of 40 to 100 feet. In that locality, broad reefs or belts of chert, impressed by the movements of faulting, with a vertical, tabular, sheeted struc- ture, course through the region, oft;en traced by outcrops for a dis- tance of one or two miles. These chert-belts have been brecciated and recemented by silicification into a very hard rock, which resists denudation, so that, by the erosion of the surrounding beds, they have been left projecting above the surface, resembling, to some ex- tent, the outcrops of mineral-bearing quartz-veins in other mining countries — a resemblance which is increased by the occurrence of bunches of oxidized lead- and zinc-ores in these belts of brecciated chert. At intervals, along their course, the chert-outcrops overlie and cap extensive deposits of zinc-ore, occurring in zones of brec- ciation in the Cherokee beds beneath ; and they should be regarded as the most favorable surface-indications for the occurrence of ore.

Lead- And Zinc- Deposits Of The Mississippi Valley. 187

The evidence of faulting and fissuring of the strata in other camps in the Southwest 18 usually strongly marked. Though the vertical movements of the beds are often slight, the crushing and displace- ment of the strata, and the grooved, striated, and polished surfaces of the rocks, show that faulting has occurred.

In the zinc-mines of Marion county, Arkansas, the fissures trav- erse all the geological formations from the Third Magnesian lime- stone to the Subcarboniferous, through a vertical range of 500 to 600 feet ; the vertical displacement of the strata by the ore-bearing fisjnres is from 5 to 25 feet.

There are evidences that the larger and more prominent fissures have a great extension in depth, and penetrate the Archaean floor on which the sedimentary formations rest. At Mine La Motte, the courses of the crevices in the granitic ridges are rudely parallel to the master-fissures in the underground workings of the mine. The Cambrian limestone and sandstones at this locality are probably no- where over 400 feet thick ; the vertical displacement of these beds by the master-fissure and its branches aggregate not less than 100 feet — a displacement in such massive strata which it is difficult to con- ceive to have taken place except as caused by a faulting movement so profound that the fissure must of necessity penetrate deep into the underlying Archeean. In general, throughout southern Missouri and northern Arkansas, the fissures associated with the ore-deposits appear to be best defined in depth, whereas in the surface-formations they appear to be split up into numerous branching crevices and fracture-planes.

In conclusion, it may be said of the fissures which occur in direct association with the deposits of lead- and zinc-ores in the Ozark and Wisconsin uplifts, that they are not the result of local causes, and are not confined to a narrow vertical range, or to rocks of a similar lith- ological character, but, on the contrary, that these fissures are the result of forces connected with wide-spread dynamic disturbances, aflecting the North American continent, and that the fissures are faulting-planes of indefinite vertical extent, traversing all the geo- logical formations from the crystalline rocks to the Coal-Measures.

Occurrence op Lead- and Zinc-Ores in the Ozark

Uplift.

Ore-Bearing Formations. — Some geological formations appear to be everywhere barren of ore; others occasionally carry small de- posits, workable where the conditions are exceptionally favorable,

188 L£Ad- And Zing-Deposits Of The Mississippi Valley.

but in each mining-region certain strata are ore bearing in a degree exceeding all other formations combined. The lead-ore now mined in the Ozark uplift is mostly from a single formation in south- eastern Mia<ouray the lead-bearing Cambrian limestones of Bonne Terre and Mine La Motte; while nearly all the zinc-ore of the South-west is yielded by the Subcarboniferous Cherokee limestone. Deposits of zinc- and lead-ores occur, indeed, in magnesian lime- stones of the Lower Silurian in central and southeastern Missouri and northern Arkansas, and formerly yielded the larger proportion of the product of that region ; but most of the mines in the Lower Silurian limestones have been worked out, and at present the yield is small compared with that of the other producing horizons. Local deposits of the ores of lead and zinc are found in the Devonian limestone and in beds of tlie Chouteau and Burlington epochs of the Subcarboniferous; but the relative yield from these formations is also small. The shales of the outlying coal-basins of Carbonif- erous age carry workable deposits of galena in Jasper county, Mis- souri. In Moniteau county, Missouri, the coal of these basins is intersected by thin seams of galena and zinc-blende, filling joints in the mass of the coal ; in mining the coal these thin seams are separated and saved, thus affording an occasional shipment of ore.

The relative importance of the different geological formations is shown by their production. The total product of lead-ore of the Ozark area, for the year ending December 1, 1889, was 54,500 tons. Of this, 30,000 tons, or 55 per cent., was produced from the mines in the Cambrian limestones. Of a total production of 122,500 tons of zinc-ore for the same calendar year, 119,000 tons, or 97 per cent., was derived from the deposits in the Cherokee limestone, and in ad- dition, 13,500 tons of lead-ore, or 21 per cent, of the total produc- tion, was extracted from the same formation.

Ore- Horizons. — In those geological formations in which workable bodies of ore occur, certain beds appear to have been peculiarly favor- able to the deposition of the ore. The strata immediately above and below the ore-bearing beds are usually barren, or carry only small deposits. This characteristic of productive strata separated by barrea beds is found, in many districts, to be uniform over large areas. The favorable strata are designated as ore- horizons and may be of any thickness, limited only by the occurrence of beds of uniform char- acter and composition. They are from 3 to 15 feet thick in the zinc- mines of northern Arkansas, and from 40 to 80 feet thick at Webb City and Joplin in Jasper county, Missouri ; and they extend through-

Lead- And Zinc-Deposits Op The Mississippi Valley. 189

out all the beds of the lead-bearing limestone at Bonne Terre in the southeastern part of the State, attaining there a thickness of not less than 220 feet.

In relative order of importance the productive strata are, first, limestones, including both pure lime carbonates and the dolomites, but especially cavern-forming limestones; second, thin-l)edded strata of chert or highly siliceous shales; third, pervious beds of clay-shale or calcareous shale. Sandstones and massive or impermeable strata are commonly barren. The presence of organic matter in the rock appears to have exerted a favorable influence on the deposition of the ore. That the ore has been deposited in these horizons and not in the intervening beds appears to be due to the structure and chemical composition of the productive stratum and to physical conditions that have supervened,-— causes that are discunsed in detail in the chapters that follow on the occurrence of the ores in the several geo- logical formations.

OccuiTence of the Ore-Deposits. — All the deposits of lead- and zinc- ores in the Ozark uplift belong to the great class of fi?8ure-fed im- pregnations, and may be designated as " runs/' a term by which this form of deposits is known to the miners. A run may be described as an irregular ore-body, formed at the intersection of an ore-horizon with a vertical fissure. The roof and floor of the ore-deposit are formed by the massive, unmineralized beds which bound the favor- able stratum above and below. Laterally the ore-body extends on either side of the plane of the fissure for an irregular distance, which de|)ends upon the extent to which the rock was brecciated and frac- tured, allowing the mineral-depositing solutions to penetrate its mass. Longitudinally the ore-deposit or run stretches the length of the sec- tion of the fissure which remained open during the formation of the ore. Thus, in its simplest form, a run is limited vertically and later- ally, but longitudinally it usually has a considerable extension. It is an ore-body which extends beyond the walls of the fissure through which the mineral-depositing solutions were introduced. Where a system of closely adjacent parallel fissures intersects an ore-horizon, the resulting runs are often connected laterally by ore, so as to form an irregular compound run. Such compound runs most frequently occur where the ore-horizon has been intersected by difierent systems of fissures and cross-fissures, resulting in a more complete and ex- tensive brecciation of the rocks, and thereby allowing the ore-de- positing solutions that are introduced through difierent fissures to

190 Lead- And Zixc-Dep08It8 Op The Mississippi Vallet.

intermingle. The extensive mines of Joplin and Webb City, Mia- soari, and Gklena, Kansas, belong to this class of oorapoand rans.

The dimensions of simple runs vary greatly, but they are com- monly 10 to 50 feet wide, 5 to 30 feet high, and from 100 to 300 feet long, these dimensions being exceeded only in exceptional oases. At Webb City there are compound runs which extend with an irru- lar course 300 to 600 feet longitudinally, the width varying from 75 to 150 feet, and the height of the ore-body from 40 to 60 feet.

In many miningistricts of the Ozark area, runs of ore occur in the stratified formations at different geological horizons, and the same fissure or system of fissures forms runs, situated onealiove another at the several intersections with the different ore-horieons.

A run resembles a fissure- vein, in that the ore is deposited from solutions which ascend through a fissure from an unknown source in depth ; but in fissure-veins of the simplest type the ore is in- cluded between the walls of the fissures, and has an extension in depth that is more or less continuous and unbroken, while in runs the ore is confined to the favorable strata and is deposited outside of the walls of the fissures; the fissures are usually barren, except where they traverse the ore-horizons. The occurrence of mineral de- posits in the form of runs is not confined to the lead- and zinc-mines of the Mississippi valley ; ore-deposits of this form exist in many localities in the mining regions of the Rocky Mountains, in situations where the sedimentary rocks are nearly horizontal in bedding and have been intersected by vertical fissures, introducing the ore-form- ing solutions.* The relation of runs to other forms of mineral occurrence is shown in the annexed general classification of ore-de- posits, based upon the manner of formation and origin of the minerals. For convenience of study, the great class of deposits formed by ascension is further subdivided with respect to the char- acter of the enclosing strata.

Classification of Ore-Deposits. Division with Respeet to the Forces Giving Rise to the Deposits.

L DepoeitR formed by the action of II. Deposits formed by the action of

mechanical forces. Examples: Placers chemical forces.

and auriferous conglomerates. (This class is further subdivided as

follows:)

This subject is more fully discussed in the chapter on "The Lead and Zinc Min- ing Regions of the Mississippi Valley Compared/ poBt, p. 212.

laEAD- AND ZtNO-DEPOSriB OP THE MISSISSIPPI VALLEY. 191

Division with Respect to the Origin of the MineralSy tchether Derived from the Enclosing or Adjacent Strata or from an Exotic Source.

UK IrregDlar depoeits formed by ]at- IP. Deposits in which the origin of

nl secretion or sregation from the the ores is in mineral-solutions ascend-

enclosing or adjacent strata. Examples : ing through fissures from some unknown

Deposits of manganese and bog iron- source in depth.

ores; certain carbonate ores of iron of (This class is further subdirided as

the Coal-Measures, etc. ibliows :)

Division with Respect to the Structure or Character of the Enclosing

Strata.

IP. — a. Deposits confined to the walls 11'.— 6. Irregular deposits extending

of the fissures through which the min- beyond the walls of the fissures and im- eral-formiog solutions were introduced. pregnating the country-rock. Exam- Example: Fissure- veins. pies: Mineralized lodes or zones; im- pregnated beds or runs, and deposits filling pre-existing caverns.

Deposits op the Lead- and Zinc-Ores in the Cherokee Limestone op the Subcarboniperous.

Stratigraphical Oeology, — The Cherokee limestone covers an area of more than 4000 square miles in the southwest part of the Ozark uplift;, extending from the vicinity of Springfield, in Greene county, Missouri, to Grand river, in the northwest portion of Indian Terri- tory. In Missouri this limestone is the prevailing surface-formation in the counties of Greene, Lawrence, Jasper, and Newton ; it covers the southeast part of Cherokee county, Kansas, the northeast portion of Indian Territory, and a limited area of northwestern Arkansas.

When not denuded, this formation has a thickness of 185 to 220 feet. In the area covered by the Cherokee limestone the strata are nearly horizontal ; gentle synclinal and anticlinal folds give a slight undulation to the surface. Along the course of the streams the re- gion is cut by shallow ravines, while the divides are broad, high prairies. Fossils are abundant in this formation, which is probably the representative of the Warsaw or St. Louis epoch of the Subcar- boniferous. The Cherokee formation is made up of beds of lime- stone irregularly interstratified with layers of chert A marked variation is noticed in the relative proportion of the chert and lime- stone; in localities frequently but a short distance apart, the Chero- kee is seen to change from a massive limestone with only occasional nodules or included layers of chert, to a stratum made up of alter- nate thin beds of chert and limestone, in which the chert exists in larger proportion.

192 Lead- And Ziko-Dep08It8 Of The Mississippi Taljley.

Analyses of the Cherokee limestone show that it is uniformly a remarkably pure carbonate of lime, averaging from 98 to 99 per cent, of carbonate of lime with traces of magnesia, alumina, oxide of iron, and insoluble matter. Organic matter and bitumen are present in appreciable quantities, and from the solution of the rock by the chemical action of surface-waters the bitumen is liberated and col- lects as a thick pitch in the cavities. Owing to the purity of this limestone and its coarsely-crystalline character, it is readily attacked by surface-waters, subterranean erosion forming numerous caves and sink-holes where this limestone outcrops near the surface. The chert of the Cherokee formation is white, bluish-white, or buff in color, breaking with a smooth, conchoidal fracture. By analysis it contains 98 to 99 per cent, of silica, with traces of alumina and oxide of iron. About 3 or 4 per cent, of the silica exists in the hydrous condition, and is soluble in a solution of potassium hydrate. The chert of the Cherokee is cleanly separated from the interstratified limestone; though so intimately interbedded with the latter, the siliceous nodules and layers occur pure and distinct in the mass of the calcareous rock.

Areal Location of the Ore-Deposit, — The Cherokee limestone is not everywhere ore- bearing, but, following the general law of mia- eral occurrence, the ore-depositn have been formed only in local areas of disturbance. Between and surrounding such areas of mineraliza- tion extend broad barren tracts of undisturbed strata. The larger number of the productive mining-districts are grouped near the mar- ginal belt of the Ozark uplift, in such a situation that the movements of elevation that have taken place and the consequent disturbance of the strata have been more profound and extensive than in the cen- tral plateau. Exceptions may be noted of mineral districts situated in the central area, where the rocks are horizontal, although much disturbed and cross-fissured. A marked instance of such occurrence is in the mining-camp of Aurora, in Lawrence county, Missouri, and the belt of deposits traceable from this district east to the vicinity of the city of Springfield.

It is probably futile to attempt to reduce to a general law, or to define in belts, the location of the ore-districts in this region. From the complex nature of the movements of disturbance that must have taken place, which alone appear to have influenced the special mineralization of certain areas, such attempts at broad generaliza- tion roust be imperfect and liable to error.

Occurrence of the Ore. — Broadly considered, the ore-deposits are

Lead- And Z1Nc-Dep061T8 Of The Mississippi Valley. 193

mineralized zones of brecciated chert, more or less intermingled with limestone and the products of its alteration and decomposition. These mineralized zones or compound runs show, in many districts in the southwest, a marked tendency to form in the areas of intersection of different fissure and cross-fissure systems, and to be favorably influ- enced by the local occurrence in the Cherokee formation of thick strata made up of thin layers of iuterstratified chert and lime- stone.

As the natural result of the manner in which the ore-bodies in the Cherokee have been formed, and of the peculiar character of the enclosing rocks, they are very irregular in shape. The ore-deposits are limited only by the boundaries of the brecciated zone, the extent of the mineralization in the pervious and shattered beds, and the thickness of the favorable stratum or geological horizon. Many of the compound runs are of remarkable size, especially in the vicinity of Webb City and Joplin, Missouri, and Galena, Kansas, where stopes occur 75 to 150 feet wide, 40 to 80 feet high, and 200 to 400 feet long, from which all the extracted material has been milled: The smaller runs are 15 to 50 feet in width, 5 to 30 feet in height, and continuous in ore for a longitudinal distance of 100 to 500 feet. In one instance, near Joplin, Missouri, a run of ore was traced by connected workings for a length of nearly 1000 feet on the course of the fissure.

The fissures that traverse the ore-bodies and form the channels through which the mineralizing solutions have entered show, in some deposits, particularly in the vicinity of Galena, Kansas, strongly marked polished and striated wall-faces ; but commonly the course of the fissures is indicated by the greater disturbance of the adjacent strata, the tabular-sheeted structure of the breccia forming the cheeks of the faulting-planes, and the presence of vugs and o|)en channels in the ore-bodies.

The ore-deposits of the southwest do not invariably occur in the form of runs. In a few mines in Jasper and Newton counties, Mis- souri, nearly vertical fissures traverse the Cherokee limestone without disturbing the stratification or producing any brecciation except that of the rock included between the cheeks of the fissure, so that in the deposition of the ore the solutions could not escape from the fissures and penetrate the wall-rocks, but deposited the ore in the breccia in- cluded between the walls, forming ore-bodies which are vertical tabular sheets, simulating in structure and occurrence the ore-shoots of fissure-veins. A fissure of this character near Joplin, Missouri,

194 Lead- And Zinc-Deposits Op The Mississippi Valley.

has a dip of 45 to 55, and carries an ore-body stoped to a depth of 60 feet. The floor of the stope, for a length of 200 feet, is con- tinuously in ore with a width between the walls of 4 to 12 feet The ore does not penetrate the horizontal beds of limestone, chert, and shale, but is confined to the fissure, the walls of which are rough and irregularly eaten out -by the ore-depositing solutions. Future exploration alone will show how such ore-bodies, resembling the shoots of ore in fissure-veins, will behave in depth when followed into the underlying formations.

The Gangue, — The gangue of the lead- and zinc-ores is made up of the rocks of the Cherokee formation and the products of their alteration, chert usually predominating. It occurs either as shattered beds, the strata retaining the original horizontal position and bed- ding; or, if the movements of disturbance and settling have been great, the rock is converted into a breccia of angular fragments mingled in the most confused manner, in which all stratification is destroyed. In many districts masses of drab-colored, coarsely crys- talline dolomite surround the ore-bodies or occur in them, in which case they are more or less mineralized. Occasionally crystalline pink or while dolomite forms the cement of the breccia; more commonly a peculiar dark-colored silicified rock, resembling in appearance a fine-grained quartzite, fills the interspaces in the brecciated zones, and is frequently impregnated with ores. This gangue-rock, named by the writer cherokite, occurs only in connection with the ore-de- posits in the Cherokee limestone in the southwest. It is the prevail- ing gangue of the zinc-mines at Galena, Kansas, and is associated with the ores of lead and zinc in the Joplin, Grauby, and Webb City districts in Missouri.

Unaltered limestone forms the gangue of the ore in a few mines. Clay-shales and calcareous shales are the gangue of some mines in the upper l>eds of the Cherokee. Tallow-clay is abundant in the chert-breccia near the surface at Granby and Aurora, Missouri.

Dolomite results from alteration of the Cherokee limestone, and is connected directly with the formation of the ore-deposits. Broad tracts covered by the Cherokee formation stretch for miles in the barren intervals between the mining-districts in the southwest, the calcareous beds showing no evidence of alteration until an area of disturbance and faulting is reached, where the limestone is often converted into dolomite in the belts along the sides of the fissures. The effect of this dolomitization is to change the fine grained struc- ture of the limestone to that of coarsely crystalline dolomite, and to

Lead- Akd Zinc-Deposits Op The Mississippi Valley. 196

obliterate all fossils and evidences of organic origin ; and were it not for the horizontal bands of chert in the mass of the dolomite and the occasional preservation of bedding-planes, there would be nothing to indicate that the rock had once been stratified. The area adjacent to the fissures over which the limestone is altered to dolomite is oflen of much greater extent than that of the oreeiosits, and large masses of dolomite surround and separate the ore- bodies from the unaltered limestone-strata.

This alteration of limestone to dolomite has not taken place to the same extent in all the mining-camps of the region. In some localities all forms of dolomite are absent, either owing to the more or less complete removal of the limestone by the action of Surface- waters prior to the dolomitization, thuH leaving little material for the solutions that deposited magnesia to act upon, or by reason of the ab- sence of magnesia in the solutions which formed the ore. The dolo- mite is a soft to medium hard, dark gray rock, with a coarsely crystal- line structure, made up of white crystals of dolomite, one to three sixteenths of an inch long, embedded in dark drab-colored cement. The following analyses made by Mr. L. G. Eakins, U. 8. Geol. Survey, from samples taken near Joplin, Missouri, show a difference in composition, in great part due to the variable amount of the in- soluble residue which formed the cement of the crystalline mass.

Analyses of Dolomite,

No. 1, No. 2.

Lime, 21.46 2872

Magnesia, 14.79 17.36

Alumina and sesquioxide of iron, . . . 1.32 1.03

Insoluble residue, 29.77 11.66

Carbonic acid by calculation, . . . . 33.18 41.55

100.47 100.32

Cherokite forms the cement of the chert-breccia in the ore-bodies. It commonly occurs as a brown or drab-colored rock, hard, dense and tough, breaking with difficulty under the blow of a hammer with a coarse, rough fracture. Analyses and microscopic examina- tions of samples of cherokite show that this rock has been formed by the silicification of the residual sandy mud resulting from the dissolution of the Cherokee limestone by waters charged with car- bonic acid. The formation of the cherokite seems to have closed the ore-depositing period, this silicification of the gangue having taken place subsequent to the formation of the ore. Of the follow-

196 Lead- And Zinc-Deposits Of The Mississippi Valley.

ing analyses of cherokite, Nos. 1 and 2 were made by L. G. Eakins, U. S. Geol. Survey, and No. 3 by the St. Louis Sampling and Testing Works. Sample No. 1 came from Joplin, Missouri, and Nos. 2 and 3 from Gralena, Kansas.

Analyses of CherokUe.

No. 1. No. 2. No. 3.

Silica, 95.77 97.83 94.72

Alumina and oxide of iron, . 1.84 1.89 4.00

Lime, 0.54 0.1 1 1.18

Magnesia, 0.24 0.09 trace.

Water 1.17 0.77

99.56 100.19 99.90

Under the microscope, thin sections of cherokite show a highly crystals'ne structure, with imperfect interlocking crystals of quartz as the predominant mineral, and spots of dark bitumen, or carbon, and scattered oolitic grains, probably derived from the limestone. Cherokite is frequently impregnated with blende and galena ; inter- mixed with brecciated chert it forms the prevailing ore in many mines in Jasper county, Missouri, and Cherokee county, Kansas. On account of its extreme hardness and specific gravity it cannot, without difficulty, be cleanly separated from the blende.

Occasionally masses of the soft residual mud, resulting from the subterranean erosion of the limestone, that from some cause have escaped silicification, are found in the ore deposits, and not uncom- monly carry very perfect disseminated crystals of blende and galena. By the oxidation and removal in solution of the minerals included in the mass of the cherokite, the rock is converted into a skeleton. Where the rock was originally densely mineralized, the resulting skeleton is light and cellular, and the cavities show the imprints of the faces of the crystals which once filled them. So perfect are some of these moulds in the rock that casts can be taken in wax reproduc- ing the form of the crystals. The minerals thus decomposed and removed were in all cases minerals of primary origin, blende and galena, and rarely pyrite ; calcite, barite and other minerals of recent deposition have not been observed included in cherokite.

Disposition of the Ore. — The original form in which the ores of lead and zinc were deposited in the Cherokee formation was that of the sulphides, galena and zinc- blende. These ores appear to have been deposited in two ways, by replacement and by crystallization. By replacement, a chemical interchange takes place; a particle of

L£Ai>- And Zinc-Deposits Of The Mississippi Valley. 197

rock is dissolved and removed by the ore-depositing solutions, while at the same time a particle of ore is formed in its place. Bj crys- tallization, the galena and blende fill the interstices in the breccia and line cavities in the mass of the ore. The chert-fragments of the ore-bodies are very rarely penetrated by blende and galena, ow- ing to the impervious and insoluble character of that rock. By this raetasomatic replacement, ore has been deposited in the mass of the limestone- fragments in the breccia and in all the calcareous material, whether limestone or dolomite, throughout the ore-body. The small size and crowded growth of the crystals of blende which make up the great mass of the ore-bodies in certain nines, is evidence of a crystallization from concentrated solutions ; as the solutions became weaker, the blende and galena of later deposition formed large crys- tals.

5ni6 Ores. — Only in exceptional cases are the ores of the South- west suflBciently free from gangue to allow them to be shipped with- out dressing. In most mines all the ore that is extracted requires to be crushed and concentrated by washing to clean the mineral from the gangue and to separate the galena from the blende. The ores when thus concentrated are remarkably pure and command a high price. This is due in part to the general absence of iron pyrites in the ores of the region, and in part to the ease with which the ore is cleanly separated from the gangue.

The yield of the ore varies: two to five tons of the richest ore produce one ton of concentrates; while in the case of the poorer ores, 10 to 20 tons are required to yield one ton of concentrates.

Vertical Distribution of the Ores in the Deposits, — Where blende and galena occur in the same ore-body, the blende usually predomi- nates in depth, often to the total exclusion of the lead-ore in the lower portions of the deposits, while galena is commonly found in formations near the surface. There is no sharp line of demarkation between the blende and galena in the deposits, and in the zone of transition the two minerals are intermingled. The blende in the lower and outer parts of the ore-body is usually free from galena and in the central and upper portions gradually gives place to the lead-ore.

In certain localities in the Southwest, the ore-deposits are not con- fined to a single geological formation, but, following the fissures, channels and openings in the strata, extend vertically from one hori- zon to another. Deposits of this character occur near Webb City, Missouri, where the Cherokee limestone is overlain locally by shales

198 Lead- And Zinc-Deposits Of The Mississippi Vall.Et.

of the Coal-Measures, and show a more strongly defined separation of the ores of the two metals; the Cherokee carrying deposits of blende with scarcely a trace of galena, while the galena is concen- trated in the superficial formation. This peculiar vertical distribu- tion of the lead and zinc has been noted not only in the deposits of the Southwest, but also in the mines of the Upper Mississippi re- gion and of certain mining districts of Europe.*

The productiou of galena in the mines is not as great as formerly, owing to the fact that the deposits near the surface are naostly worked out, and as the explorations are pushed in depth the relative output of blende constantly increases. At present date the produc- tion of zinc-ores is from 5 to 20 times that of galena in the different district8, averaging for the entire Southwest about 10 tons of zinc- blende to one ton of galena.

Alteration of the Ores, — In most of the mines in the Southwest very little alteration of the ores has taken place since their deposi- tion. At the mining-camps of Granby and Aurora in southwest Missouri, however, an extensive alteration of the ores has occurred ; the blende is converted into calamine or rarely into smithsonite, and the galena into cerussite. But the production x)f these oxidized ores is rapidly decreasing; the great proportion of the output of the mines of the Southwest is now galena and blende.

The Minerala ConstUufing the Ore-Bodies.

The minerals in the ore-bodies of the Cherokee limestone have l)een mostly formed by primary deposition; where the conditions were locally favorable, a secondary deposition has taken place from oxidation and the action of the surface-waters on the primary deposits.

The minerals of primary deposition in the Cherokee limestone are:

Sulphides. — Blende and galena, constituting the principal ores. Cadmiferous blende, locally abundant in some districts. Existing as subordinate minerals and probably resulting chiefly from second- ary deposition : pyrite and marcasite, chalcopyrite (rare).

Sulphates, — Barite locally occurs in a few mines in the Southwest, but is usually of secondary formation.

Carbonates. — Dolomite, formed by the alteration of the Cherokee limestone.

♦ Metallic Wealth of the U. S.;" and " Lead-Mines of the Upper Miw.," by J. D. Whitney; Geology of Wis., vol. iv., T. C. Chamberlin.

Lead- And Zixc-Dep08Its Op The Mississippi Valley. 199

Silicates. — Cherokite, resulting from the silicification of the resid- ual sediment formed by the decomposition of the Cherokee lime- stone. Quartz occurs rarely in the deposits of the Cherokee.

The minerals of secondary deposition are : ' Sulphides. — Pyrite and marcasite, formed from the iron derived from the wall-rocks. Blende, galena, chalcopyrite and greenockite, produced by alteration from the primary ores in the zone of oxida- tion in the upper portions of the ore-bodies, and re-formed as sulphides by the reducing action of organic matter in the deeper levels.

Sulphates. — Anglesite, from alteration of galena. Barite, occur- ring locally, and probably derived from the gangue of the ore. Gypsum, a rare mineral in the Southwest, though sulphate of lime is present in all the mine-waters. Soluble in water and derived from the oxidation of the ores : sulphate of zinc and several different sul- phates of iron, occurring as efflorescences in the old workings of the mines. Sulphate of cadmium is present in the waters of many mines. Sulphates of magnesia, lime and the alkalies are also pres- ent in mine-waters, and result from the action of metallic sulphates upon the gangue and country-rock.

Carbonates, — Resulting from the oxidation of blende and galena: smithsonite, hydrozincite, aurichalcite (rare), cerussite and cadmife- rons smithsonite (rare). Derived from the gangue or from the wall- rock : dolomite and calcite.

Silicaies. — From the alteration of blende : calamine and zincife- rous tallow-clay. From decomposition of chert and limestone by surface-waters : tallow-clay.

Phosphates and Chlorides. — Pyromorphite (rare), a product of the alteration of galena.

Organic. — Bitumen set free in the decomposition of the rocks by surface-waters.

The Order of Deposition of the Minerals.

The minerals forming the ore-bodies in the Southwest appear to have been deposited in a uniform order, which seems to be constant in all the mines examined where the facilities for observation were such that the paragenesis of the different minerals could be deter- mined. This order of formation is:

Minerals of Primary Deposition. — 1. Crystalline dolomite, fre- quently forming the wall-rock of the deposits.

2. Crystalline blende, becoming more perfectly crystallized as the

200 LEAD- AND ZINO-DUPOSITS OP THE MrSSI&SIPPI VALLEY.

deposition proceeded, and often intermixed with galena in the ores of later deposition.

3. Crystalline and crystallized galena.

4. Pyrite, occurring in relatively small proportion to the other minerals.

. 5. Cherokite and crystallized quartz (rare).

Minerals of Secondary Deposition. — From the nature of their depo- sition, the paragenesis of the minerals of secondary formation is greatly varied, and the same mineral may occur moi*e than once in the series. While thus conforming to no absolute order, the sec- ondary minerals may be arranged in groups according to the more commonly prevailing sequence of deposition, the series being thus continued :

; 6. Crystallized white and rose-colored dolomite, lining cavities in the ore-body and filling the interstices in the breccia.

7. Crystallized blende, usually of garnet or ruby-red color, often in small brilliant, translucent crystals. Crystallized galena, the commonest forms being the cube and cube-octahedron ; the plain octahedron is of rarer occurrence.

8. Crystallized pyrite, marcasite, chalcopyrite, calcite and barite. Amorphous tallow-clay.

9. Anglesite, cerussite, calamine, smithsonite and greenockite, re- sulting from the alteration of the ores.

It may be noted that the primarily deposited ores are almost wholly composed of the simple sulphides of zinc and lead and of the gangue-minerals dolomite and cherokite. Calcite appears to be in -all instances of secondary and late formation. Pyrite, marcasite and barite are probably of both primary and secondary deposition, but seldom occur other than in relatively subordinate importance.

The Cause of the Concentration and Deposition of the Ores in the Cherokee Limestone.

For purposes of discussion the causes that have induced the ore- formation may be divided into three classes:

Structural. — In the Cherokee formation thin- bedded strata, espe- cially where the limestone is interstratified with chert, are observed to be peculiarly favorable to ore-occurrence. Such thin and brit- tle stmta of alternating limestone and chert, included between more massive and unyielding beds, were shattered by the slightest movement of disturbance or faulting, and subsequently, in the depo- sition of the ore, these fractured and brecciated strata afforded free

Lead- And Zixo-Deposits Op The Mississippi Vai.Iy. 201

circulation to the mineralizing solutions, the effects being augmented by the greater number of bedding-planes. The coarsely crystalline structure of the limestone and its porosity and permeability also increased the action of carbonated waters and further aided in open- ing channels for subterranean circulation. The position of this geo- logical formation, occurring near the surface over a great area, has likewise been an important factor, from the action of sub-aerial waters and from the increased extent of the brecciation due to fissuring and movements of disturbance; it being a well-recognized law that the fracturing and crushing of a stratum of rock by dynamic movements takes place the more readily and to a greater extent, the less it is weighted by superincumbent strata.

Chemical — The extreme purity of the limestone of the Cherokee renders it readily soluble in waters containing carbonic acid. Not only are the limestone strata thus dissolved and removed by the action of sub-aerial waters, but in the subsequent deposition of the ores, the limestone fragments that remain in these breoeiated beds are metasomatically replaced by the ores. The limestone itself acts as a chemical precipitant and neutralizes any free acid or acid salts present in the ore-forming solutions, and the organic matter and bitumen originally contained in the calcareous rocks, but set free by their dissolution, constitute powerful reducing agents in the ore- deposition.

PhysioaL — Physical conditions antecedent to the formation of the ore, have influenced the localization of the deposits and greatly in- creased the extent of the brecciated zones. Prior to the formation of the mineral deposits, the Cherokee limestone, where it occurs near the surface, was subjected to the prolonged action of surface- waters ; in the districts where the beds were faulted and disturbed, an exten- sive subterranean erosion of the calcareous strata has taken place. This removal of the limestone interbedded with the chert in the shattered and Assured belts caused a settling of the formation and of the superincumbent rocks, resulting in the more complete breccia- tion of the chert and in the formation, ultimately, of broad zones, consisting almost entirely of angular fragments of chert intermixed in the most confused manner, and mingled with residual clay of the eroded limestone.

The action of these combined physical and chemical forces began with the elevation of the Ozark uplift above the ocean at the close of the Subcarboniferous period, and continued with attendant fault- ing and disturbance of the strata for a vast period of time, during

202 Lead- And Zinc-Depositb Op The Mississippi Valley.

which the Ozark area remained dry land — an interval when the cli- matic and atmospheric conditions are believed to have been, at least during the Coal-period, singularly favorable for the production of surface-waters charged with carl)onio acid and the organic acids resulting from the decay of vegetation. At a much later age, when other dynamic disturbances produced a more profound and widely extended Assuring of the formations and inaugurated the period of ore-deposition, these zones of brecciated rock directly connected with the fissures and containing the chemical elements requisite to effect the reduction and precipitation of the metals, afforded free escape and circulation for the ore-depositing solutions and formed a matrix admirably fitted to receive the minerals.

Ore-Deposits in the Magnesian Limestones op the Lower

Silurian.

Stratigraphical Oeology. — The rocks of the magnesian limestone series cover nearly four-fifths of the area of the Ozark uplift, reach- ing from the Missouri river south to the valley of the Arkansas, and from the Mississippi west to the headwaters of White river. This formation aggregates 1200 to 2000 feet in thickness and is made up of three or four distinct limestones, separated by strata of soft sandstone. The position of the beds is nearly horizontal, except where they are locally disturbed by dynamic movement. The lime- stones are composed of beds of very variable composition, shaly layers frequently alternating with thick strata of compact limestone. By reason of the irregular hardness of the different beds and the greater elevation of the area covered by the magnesian limestones, an extensive erosion has taken place. The topography is marked- by long narrow divides, with steep and rocky slopes and high cliffs alon the streams, constituting altogether the roughest and most mountainous region of the Ozark upheaval. Numerous caverns, some of great extent, exist in this section.

Occwrence of the Ores. — Lead- and zinc-ores are found at widely separated localities in the area covered by the magnesian limestone scries in central, southern and southeastern Missouri and in the con- tinuation of this formation into northern Arkansas. In general the ore-deposits are small, though in the early settlement of the country, large and very productive bodies of lead-ore were found in these limestones in southeastern Missouri and were worked for many years — but now most of them are exhausted.

The ore-deposits are found at certain favorable horizons in the

Lead- Akd Zikc-Dep08Its Ok The Mississippi Valley. 203

limestones of this series, in runs formed at the intersections of the horizons by faulting fissures. The intervening strata are barren. The ore-horizons are commonly shaly or thin-bedded strata; where the limestones occur in thick beds they are less frequently ore-bearing, and the sandstones of the series are everywhere barren. In northern Arkansas it is noticeable that thin-bedded, highly siliceous strata form the horizons, probably from the facility with which such brittle beds, inclosed between rigid strata, are shattered by the faulting movement and form a permeable matrix for the ores. Generally it is observed that in the limestones of this formation the pervious character of the bed, rather than the chemical composition of the rock, has influenced the deposition of the ore. The solnbility of dolomite in surface-waters has to some extent influenced the deposi- tion, through the enlargement, by this agency, of the crevices and openings in the limestone.

In the deposits in the magnesian limestones, the small extent of the brecciation which has accompanied the faulting has induced a concentration of the ore-deposition within narrow limits, with the result that the ores of this formation are usually rich and free from gangue. The ores are in most part pure and capable of being band-sorted for shipment ; only the poorer ores are crushed and concentrated.

The more notable mineral occurrences in this formation are the prevalence of smithsonite in place of calamine, as the mineral re- sulting from the alteration of blende, and the deposition of crystal- lized barite and quartz in the gangue of the mines in certain locali- ties. In southeastern Missouri barite frequently occurs as the gangue of galena, and the outcrop of the ore-bearing formation is charac- terized by a peculiar chert filled with nests of crystallized quartz, known by the miners as mineral blossom." Crystallized pink dolomite, resembling that of the Cherokee limestone and possibly of similar origin, is occasionally found in the ore-deposits and fis- sures in the magnesian limestones of northern Arkansas.

Depositb op Lead-Ore in the Cambrian Limestones.

Stratigraphical Geology, — In southeast Missouri, in the lead- mining region of Bonne Terre, Doe Run and Mine la Motte, a dark brownish-black magnesian limestone forms the ore-horizon. The age of the formation may be Upper Cambrian (Potsdam). This limestone rests uncomformably on the higher points of the Archeean beJ-nck at Mine la Motte and at Doe Run. In the deeper chan-

204 Lead- And Zinc-Deposits Op The Mississippi Valley.

Dels between the ridges of granite and porphyry the lead -bearing limestone overlies a soft sandstone. The thickness of the lead- bearing formation is 220 feet at Bonne Terre; at Mine la Motte it is locally much reduced by denudation, attaining 160 feet as a maximum.

The strata of the Cambrian are liorizontal, or dip locally at gentle angles from some axis of disturbance. In the mineralized areas the beds have been subjected to extensive faulting. This dynamic ac- tion, measured solely by the vertical displacement of the strata, is most intense at Mine la Motte at the extreme southern end of the district, and decreases to the northwest, being smaller at the Doe Run Mine and smallest at Bonne Terre. At Mine la Motte a large fault- ing-fi?sure bounds the workings of the mine on the west, and is con- nected with the master-system of fissures traversing the ore-body. The strata at this mine have been faulted by these fissures prior to the deposition of the ore ; the vertical displacement of the beds is from 5 to 10 feet in the stopes, and the aggregate throw by this main faulting-fissure and its branches exceeds 100 feet. At Bonne Terre no considerable displacement of the strata has occurred ; the work- ings of the mine are traversed by belts of parallel fissures with walls striated horizontally, showing that a lateral movement has taken place.

Occurrence of the Ore. — The galena occurs disseminated through the beds of limestone in crystalline grains from to -inch in diameter, and also deposited in flat irregular thin sheets in the shaly partings between the layers of the rock. The position of the ore in the limestone is determined by the intersection of a system of nearly parallel vertical master-fissures, traversing the strata in a narrow belt, by one or more belts of vertical cross-fissures. The masses of galena occur around such places of multiple and complex intersec- tion, where different systems of faulting-planes, intersecting on diverse courses, and crossed by smaller diagonal fractures, have breccia ted in situ a large section of ground.

In the subsequent mineralization of the beds, the ore-depositing solutions, rising from below through the system of master-fissures, followed paths of easiest flow where the intersection of the cross- fissures aided in keeping the channel open, and on reaching the hori- zon of the black limestone, penetrated the fractured beds. The form and extent of the ore-body has been determined by the distance to which the solutions penetrated the fissured limestone from the chan- nels by which they were introduced. The richest ore is usually

Lead- And Zino-Deposits Of The Mississippi Valley. 206

found ahont the intersection of the fissures, and as the ore-body is followed laterally it becomes poor as soon as the outer boundaries of the area of disturbance and Assuring are reached. The mineral- forming solutions deposited only galena and small quantities of iron pyrites carrying nickel and cobalt. Not a trace of zinc-blende occurs in these deposits in the Cambrian, though in the raagnesian lime- stones blende is found in considerable quantity in many mines of southeast Missouri. The ore-deposits may be described as thick beds of limestone irregularly impregnated with galena ; but on ex- amination they are seen to be compound runs of unusual size, formed much in the same manner as the mines of the Southwest. The ore- deposits are of great size. At Bonne Terre a stope in the mine fol- lows a belt of fissures for nearly 3000 feet ; the width of the work- ings is 100 to 200 feet, and the height 25 to 60 feet. At this mine all of the extracted ore goes to the mill and is crushed and concen- trated ; the average yield as it is taken from the ground is about 8 per cent, of galena, equivalent to 6 per cent, of lead.

Cavse of the Accumviaiion of the Deposits of Oalena in the Cam- brian Limestone. — At Bonne Terre the mineralized beds occur at a depth of 50 to 225 feet ; at Mine la Motte they outcrop at the pres- ent surface of denudation. No subterranean erosion of the limestone has taken place; even the more prominent fissures have been but little enlarged by the circulation of surface-waters. The ore-bearing strata at Bonne Terre are overlain by beils of magnesian limestone of close texture, containing but little organic matter. These upper beds, though traversed by the faulting-fissures are but little fractured and are unmineralized. The underlying sofl sandstones are also barren.

The causes of the deposition are, apparently : the presence of or- ganic matter in relatively large proportion in the ore-bearing beds; the existence of numerous shaly partings between the layers of lime- stone; the pervious structure of the rock, which favors the penetra- tion of the mineralizing solutions and the deposition of the ore by replacement; and the extensive fracturing of the strata by faulting movements, due to the brittle character of the formation, in which hard layers alternate with thin, soft, shaly partings.

Production. — From Mine la Motte, Doe Run and Bonne Terre the aggregate production for the calendar year 1888 was 20,750 tons of pig-lead. In 1889, the production was increased to 21,320 tons, forming nearly 60 per cent, of the total production of non-argentifer- ous lead in the United States for that year, or 1 1 per cent, of the yield of lead in the United States from all sources.

206 lead- and zinodep0sit8 of the mississippi valley.

Deposits of Argentiferous Lead- and Zikc-Orbs in the Ouachita Uplift.

Slratiffraphical Geology. — The ore-bearing rocks of soathwestern Arkansas are metamorphic day-shales and quartzites, probably of Calciferous or Trenton age, extending through the central ridges of the Ouachita uplift from Little Rock, Arkansas, into Indian Terri- tory. Prior to the deposition of the ores, these rocks were strongly folded, metamorphosed and subjected to heavy denudation. In many localities in the area of the upheaval, dikes of igneous rocks are intruded into the older formations, and in the eastern section, in the vicinity of Little Rock and Magnet Cove, Arkansas, extensive extravasations of eleeolite-syenite occur.

Dynamic Disturbances. — The Ouachita uplift, like the Ozark area, was elevated above the ocean at the end of the Subcarbonifer- ous period. The last great upheaval took place at the close of the Carboniferous; strata of the Coal-Measures being folded by the move- ment. The intrusion of the igneous rocks in the Ouachita area has been determined by Dr. J. C. Branner to have taken place in post- Cretaceous time.*

It is probable that the formation of the veins of quartz carrying silver, lead and antimony was due to these igneous disturbaoces; but the evidence so far obtained resi)ecting the age of the deposition of the ores is of a negative character.

Occurrence of the Ores. — Deposits of silver-bearing lead- and zinc-ores are found at intervals scattered irregularly throughout the extent of the Ouachita uplift. The more prominent mining districts are the Kellogg mines near Little Rock, Silver City in Montgomery county, and Antimony City and Silver Hill, in Sevier county, Ar- kansas.

The ores occur in fissure-veins which have general northeast and southwest varying to east and west courses, parallel to the axis of the uplift. The fissure-character of these deposits is strongly marked ; the veins dip at steep angles, from 40 to 90 degrees, and traverse the country in straight courses, in some instances indicated on the surface by outcrops of quartz. The fissures cut the bedding of the inclosing rocks on both strike and dip ; the walls are slick- ensided and striated by the faulting-movement, and the crevices are more or less filled with material ground from the walls. Fre- quently the vein-material exhibits a tabular, sheeted structure, par-

Eeporte of the Arkanea$ Oeoiogieal Swve}/.

Lead- And Zinc-Deposits Of The Mississippi Valley. 207

allel to the cheeks of the fissure. The majority of the veins are narrow, usually averaging one or two feet in width, and seldom ex- ceeding four or five feet. The ore-deposits form shoots confined to the fissures, and do not penetrate or impregnate the wall-rocks.

The prevailing ores are crystalline zinc-blende, carrying from a trace to a few ounces of silver per ton ; crystalline galena with from 4 to 60 ounces of silver per ton, and iron pyrites and chalcopyrite containing traces of silver. In several mines argentiferous tetrahe- drite occurs associated with galena and blende, together with a min- eral resembling polybasite. These silver-minerals are very rich, assaying from 100 to 1200 ounces of silver per ton.

In Sevier county the veins carry deposits of antimony-ore (stib- nite) ; in the same district lead- and zinc-ores are closely associated with the lodes of antimony ; and in one occurrence, a vein which has produced stibnite near the surface, has changed at the depth of 90 feet to zinc-blende and galena low in silver, but with little antimony. These antimony-ores are usually poor in silver. The vein-rock is either opaque white crystalline quartz with a banded structure parallel to the walls, or a breccia of angular fragments of the wall- rock, cemented by quartz. The minerals are disseminated through the mass of the quartz ; the fragments of wall-rock in the gangue are unmineralized. Crystalline calcite and chalybite occur as sub- ordinate minerals in the gangue of certain mines. Mispickel is found in small quantity in the antimony-lodes in Sevier county. The paragenesis of the minerals is obscure; the deposition has prob- ably occunied from hot, alkaline solutions carrying silica and the metals; the minerals appear to have been simultaneously formed by cr}*stallization in a magma of soil, gelatinous silica, which subse- quently was converted into the crystalline quartz of the gangue. Since the deposition, very little alteration of the ores has taken place, the zone of oxidation extending but a few feet below the sur- face.

In occurrence, these argentiferous zinc- and lead-veins closely re- semble the silver-bearing veins of the Rocky Mountain region, to which they are probably allied in origin, and from which they cannot in any way be differentiated in classification.

The greater proportion of the ores require mechanical concentra- tion, though in certain mines the ores are sufficiently rich and free from gangue to be shipped after hand-sorting. Owing to various causes all of the mines of this region were closed at the time of my examination (March, 1891). In a number of localities the outlook

208 Lead- And Zfnc-Deposits Of The Mississippi Valley.

is such as to warrant further and more extended prospecting in depth.

Deposits of Lead- and ZfncOres in the Wisconsin Uplift.

Straiigraphicai Geology. — The formations of the Wisconsin-Iowa mining region lie almost horizontal, with a slight dip in a general southwest direction. In an ascending order occur the Potsdam sandfitone, the Ix>wer Magnesian limestone, the St. Peter's sandstone, the Trenton and Galena limestones, the Maquoketa shale and the Niagara limestone. Over the greater portion of the mining-region the Galena limestone is exposed at the surface, the Niagara limestone and the Maquoketa shale having been removed by denudation. Owing to this denudation and the prevailing southwest dip of the formations, the Trenton limestone occurs near the surface in the ex- treme eastern and northern sections of the mining-area, and the Galena limestone that remains uneroded increases in thickness to the south and west until near Dubuque, Iowa, the upper beds outcrop, thinly covered by tiie Maquoketa shale, and'finally disappear beneath the later formations.

Dynamic Disturbances, — The Wisconsin Island, together with all of the surrounding land, was elevated in the great continental up- heaval at the close of the Subcarboniferous. The mining-rion, like that of the Ozark uplift, has been continuously above water from the Carboniferous to the|)resent time, nor was it covered in the Glacial period by the great iee-sheets or by the deposits of the Drift.

Two systems of fissures occur in the lead- and zinc-rion ; one with courses substantially north and south is characterized generally by small dimensions ; the other, an east-and-west system, includes the great ore-bearing crevices. These fissures are faulting-planes traver- sing all the different geological formations ; as a rule they are nearly vertical, in a few instances dipping at angles of 35 to 45 degrees. They appear to have been formed at distinct periods ; the north-and- south system is the older and has been faulted by the east-and-west master- fissures; the direction of the faulting movement was from west to east in a horizontal thrust, with very little vertical displacement of the strata. In a mine at Dubuque, Iowa, where the horizontal throw was determined by measuring the displacement of the north-and- south fissures by tliose of the east-and-west system, it was found to be 30 to 40 feet, while the vertical component of the movement was only 3 or 4 inches. The north-and -south fissures are usually less than a foot wide; the east-and-west crevices are much stronger, and the

Lead- Akd Zinc-Deposits Of The Mississippi Valley. 209

evidences of the faalting movement produced by them are more pro- nonnoed. In the vicinity of Dubuque, Iowa, certain of the east-and- west fissures are from 5 to 20 feet wide between the* walls, and have been traced by workings at intervals along their course for from 2000 10 6000 feet.

After the formation of the fissures a long period must have elapsed, during which the crevices and openings in the limestone strata were greatly enlarged by erosion, due to the circulation of sub-aerial waters, prior to the introduction of the ore.

Occurrence of the Ores. — The great ore-bearing rocks are the Trenton and Galena limestones, aggrating 300 feet in thickness. Ore has been found in small quantities in both the overlying and underlying geological formations; but the only horizon other than the Trenton that promises to carry workable deposits is the Lower Magnesian limestone.

Conforming to the general law of mineral occurrence, the ore- deposits are grouped in fissured and disturbed areas, constituting the mining districts; and where faulting fissures are absent, the ore- horizons, over broad tracts, are barren.

Ore-Horizons, — Certain horizons, locally called openings, in the Trenton and Gralena limestones, are found to be ore-bearing to a much greater degree than the intervening beds. Such horizons are often persistent over considerable areas, and are, commonly, thin- bedded, shaly, and |)orous strata. Less frequently' layers of massive limestone form the ore-bearing beds, where the mck, from its physi- cal structure or chemical composition, is readily attacked by carbon- ated waters. Nearly all the ores mined are extracted from the east and west crevices; the north-and-south fissures commonly carry thin sheets of ore, and are but little worked. Where the two systems of fissures cross in an ore-horizon, the ore-deposit in the eastand-west crevice is ofen enriched and of greater size near the intersection.

The form of the ore-deposits is that of runs, which occur where the favorable horizons are traversed by faulting fissures. The course of the runs conforms both to the strike of the fissures and to the nearly horizontal position of the strata. These runs, while subject to great variation, are of two general forms : First, where thin- bedded rocks constitute the ore-horizon, the ores are oflen deposited, by replacement, impregnating the layers of the rock on one or both sides of the fissures, and for a considerable distance beyond the walls. Second, in horizons in the massive beds of the limestone, and in the less permeable shaly strata, the ores have been deposited by crys-

210 Lead- And Zinc-Deposits Op The Mississippi Valley.

tallizatioD and form more or less regular sheets, included witbiu the eroded walls of the crevices. When ore-bodies of this character are of considerable longitudinal extent, and are continuous for some vertical distance in the fissures, they resemble somewhat the ore- shoots of fissure-veins.

The gangue of the ore in the deposits within the crevices is com- monly composed of fragments of the wall-rock, mingled with the clay resulting from their decomposition. In the imprnated beds, the gangue consists of the disturbed and slightly altered rock which forms the ore-horizon. The ores from the impregnated beds usually have to be crushed and concentrated, while most of the ores depos- ited in the open crevices require hand-sorting only to prepare them for sale.

In the vertical distribution of the. ores in the deposits, galena is more abundant near the surface, while blende prevails almost exclu- sively in depth. In the northeast section of the mining region, blende occurs in large proportion in the Trenton and in the lower beds of the Galena limestone, even under conditions where these formations are exposed at the surface.

Minerals Associated with the Ores. — Iron pyrites and marcasite are very abundant, and appear to have been of primary deposition, thus indicating the presence of a large amount of iron in the ore- forming solutions — a view confirmed to some extent by the large percentage of iron contained in the blende. Chalcopyrite also is abundant in the northeast section of the mining-region, and, in some localities, it is mined in limited quantities. Smithsonite usually re- sults from the alteration of blende ; calamine, from the absence of any form of soluble silica in the ore-deposits, is of rare occurrence. Although the wall-rocksf are highly magnesian, dolomite rarely oc- curs in the gangue, though calcite is quite common. Barite is also comparatively rare, though more prevalent in the deposits in the Trenton limestone.

Order of Deposition of Minerals. — The minerals constituting the ore-bodies usually have been deposited in the following order:

1. Crystalline pyrite, forming a thin band separating the ore from the wall-rock.

2. Crystalline blende.

3. Crystalline and crystallized galena.

4. Crystalline and occasionally crystallized pyrite and marcasite.

5. Crystalline and crystallized calcite.

6. Crystalline barite.

Lead- And Zinc-Deposits Of The Mississippi Valley. 211

Cause of the AocumiUation of Ores in tlie Trenton. — The existence of the ore-deposits appears to be due :

1st. To the exposure of the Galena limestone for a great length of time, antecedent to the introduction of the ores to atmospheric agen- cieSy and to subterranean erosion. This action of the oxidizing sur- face-waters not only enlarged the then existing water-channels, open- ings, and crevices, but softened and decomposed the wall-rock and the permeable strata adjacent to the fissures, thus preparing a favorable gangue for the ores in the subsequent mineralization.

2d. To the great solubility of both the Trenton and Gralena lime- stones in waters containing carbonic acid. The Galena limestone and certain layers of the Trenton are highly magnesian, approaching a pure dolomite in composition. From the greater solubility of mag- nesian than of lime carbonate in carbonated waters, these dolomites are less resistant to subterranean erosion than the limestones. Sur- face-waters also more rapidly erode the Galena limestone on account of its soft, coarsely crystalline structure and the irregular occurrence of hard and soft spots in the rock.

3d. To the presence of interstratified soft and pervious shaly beds, affording free circulation for subterranean waters.

4th. To organic matter contained in the rock, and most abundantly in the shaly beds. This, by its oxidation, has, to some extent at least, furnished a source of carbonic acid to the circulating waters, and in the deposition of the ores has acted as a precipitant of the metals.

The Occurrence of Lead- Ores in the Lower Magneaian Limestone. — Small deposits of lead-ore were mined in early days from the Lower Magnesias limestones, and patch-diggings,'' or deposits of flt-ore resulting from the erosion of this formation, were found in a number of localities on the northern and western borders of the mining region. The only mine at present worked in the Lower Magnesian is situated abottt 6 miles northwest of Lansing, Iowa, in the extreme northwest border of the lead-region. At this locality a strong vertical faulting fissure, having a course nearly north and south, traverses the lower beds of this formation. The crevice averages from 6 to 10 feet in width, and is filled with tough red clay, which forms the gangue of the ore. The galena occurs in a vertical sheet, 3 to 4 inches thick, embedded in the clay. The ore is very pure, assaying over 70 per cent, of lead and 2.6 ounces of silver per ton. About 25 tons of clean ore have been shipped from this mine. This occurrence of lead in the Lower Magnesian limestone is of interest as showing

212 Lead- And Zinc-Deposits Of The Mississippi Valley.

the great vertical range of the ores in the geological formation of this region.

The Lead- and Zinc-Mininq Regions of the Mississippi Valley Compared.

In this discussion it is necessary at the outset to separate the Ouachita uplift with its Appalachian structure, its highly disturbed and metamorphosed rocks, its igneous intrusions and silver-bearing fissure- veins, from the other areas of upheaval in the Mississippi valley, and to place it with the silver-mining regions of the Rocky Mountains, to which this elevated range in southwestern Arkansas has so many points of resemblance. The Ouachita uplift cannot be classed with the lead- and zinc-producing regions proper, as silver is the more valuable and important coitituent of the ores.

Unity of Plan in the Occurrence of the Ore-DepoelU. — The lead- and zinc-regions of the upper and lower Mississippi have the follow- ing points in common :

1. The mining regions are located in great insular areas, elevated above the general level surface of the Mississippi valley.

2. These uplifts have a nucleus of Archaean rocks, surrounded by encircling belts of the Palaeozoic formations, and have been centers of recurring dynamic action from Archaean to Tertiary time.

3. The mineral rions have been continuously above the ocean since the opening of the Carboniferous period, and are non-glaciated and driftless.

4. The ore-deposits are confined to limited areas, distributed without apparent regularity over the uplifted regions, but showing some pendency to roup near the marginal belts rather than in the central plateaus.

5. The mining-districts are local areas of disturbance and faulting of the strata.

6. The deposits of lead- and zinc-ores are not confined to any spe- cial geological formation, but occur to a greater or less extent in all the limestones, shales and chert-beds from the Cambrian to the Coal- Measures. Neither is any specific horizon found to be ore-bearing in all sections, but each mining region has its own peculiar produc- tive formation. The ore-bearing horizons are predetermined by the position and structure of the formations, the permeability of the beds, the solubility of the calcareous rocks in carbonated waters and the abundance of organic matter in the strata.

7. The absence of igneous rocks in the sedimentary formations

Lead- And Zinc-Deposits Op The Mississippi Valley. 213

and of all evidence of igneous action in connection with the deposi- tion of the oreSy other than such as may have taken place deep in the earth's crust

8. The ore-deposits occur in runs or imprnated beds and are invariably associate<l with faulting fissures. The location of the ore- bodies is in many instances determined by the foci formed by the intersection of a system of master-fissures by belts of cross-fissures and diagonal faulting-planes.

9. The minerals and ores have been deposited in a definite order of succession; and this paragenesis, at least for the minerals of primary deposition, appears to be uniform in each mining region, though somewhat varied, for the minerals constituting the series, in the dif- ferent sections of the Ozark and Wisconsin uplifts.

10. Simple minerals, few in number, constitute the ores; the zinc and lead were originally deposited as blende and galena. Quartz is rarely present in the gangue; and all minerals of gold, silver, anti- mony and arsenic are absent from the ore-bodies. Galena commonly occurs near the surface, giving way in depth to blende; this vertical distribution is not invariable nor yet peculiar to these regions, for it is also reported of the deposits of these ores in the stratified rocks of Europe.

Prominent Features of the Occurrence of the Aiyentiferom Lead- and Rnc-Ores in the Ouachita Uplift, — These are:

1. The evidences of intense dynamic action, not only in the fold- ing and metamorphism of the rocks, but in the later disturbances in post Carboniferous and post-Cretaceous times.

2. The occurrence of extensive igneous intrusions that were prob- ably connected with the formation of the ore-deposits.

3. The ore-deposits occur in districts distributed irregularly through the central axial belt of the uplift.

4. The ores are deposited in fissure-veins and do not extend into the wall-rocks, but are included within the crevices. The form of the deposits is that of shoots, or sheets of ore bounded by the walls of the fissure and having an indefinite extension longitudinally and in depth.

5. The deposition of the ores does not seem to have been in any way influenced by the character or chemical composition of the wall- rocks.

6. The minerals constituting the ores and forming the gangue have probably been simultaneously deposited from heated alkaline solutions carrying silica and the sulphides of the metals.

214 Lead- And Zinc-Deposits Op The Mississippi Valley.

7. The complex character of the minerals associated in the fis- sures, including argentiferous galena, blende, pyrite, chalcopyrite, stibnite, mispickel, argentiferous tetrahedrite, polybasite (?), chalyb- ite, ankerite (?), calcite and quartz.

CaiLse of the Peculiar AasocioHon of the Minerals in the De- posite. — The lead- and zinc-ores of the Ozark and Wisconsin uplifts, when compared with those of the Ouachita uplift and of the Rocky mountains, differ radically from the complex minerals of the silver* bearing fissure-veins, in the more simple occurrence of the few mine- rals that form the ore and in the absence of all minerals of arsenic, antimony and silver as well as of crystalline quartz. It is probable that this peculiar occurrence of the minerals is due to the formation of the ore-deposits under special physical and chemical conditions; that the mines of the West and of the Ouachita uplift have resulted from a deposition from solutions of high temperature and pressure, arising from the proximity of intrusions of igneous rocks and of widespread igneous action, while in the Ozark and Wisconsin areas no igneous intrusions have occurred and the minerals have been deposited by solutions of relatively lower temperature.

The prevailing uniform order of superposition of the minerals in the deposits in the Cherokee limestones, dolomite coming first and being followed in order by blende, galena and cherokite, may be due to a deposition effected by successive upflows of solutions charged with magnesia, zinc, lead and silica — upflows that were regional in their extent and that resulted from conditions existing in the source of supply in the crystalline rocks. In the ore-bodies of the Cherokee the formation of dolomite and the deposition of the ores of lead and zinc graduate insensibly the one into the other in the order of time ; no sharp line of demarkation is observed to separate these minerals in the deposits.

The dolomitization was most intense in the early stages of depo- sition; great masses of the Cherokee limestone were then altered to dolomite, the action continuing with diminishing energy through the ore-forming period. Aftr the culmination of this dolomitiza- tion, blende appears to have been formed ; the deiX)sition of the blende gradually giving way to that of the galena, so that in the zone of transition the ores of the two metals occur intermingled in the ore-bodies.

The epoch of the silicification of the gangue closed the ore- forming period and seems to have taken place after the primary for- mation of the ores had entirely ceased.

Lead- Akd Zikc-Depobits Of The Mississippi Valley. 215

The occunrenoe of lead near the surface and of zinc in depth, seems to be the result of causes which are to a great extent inde- pendent of the geological age or nature of the rocks in which the deposits occur, and is apparently influenced by the position of the forniation with respect to the surface. It is possible that the depo- sition of the galena may be due to the cooling and diminished pres- sure of the mineral-solutions on reaching the surface, and to the mingling of the ore-forming solutions with surface-waters carrying organic matters, both in suspension and in solution, and also nitro- genous compounds, ammonia, carbonic acid, oxidized alkaline salts and free oxygen. The peculiar manner in which cinnabar is now forming at Sulphur Bank, California, by the action of ammonia in the ascending mineral solutions has been investigated by Becker.'*' The finding of weighable amounts of sulphate of ammonia in crys- tals of barite from the lead- and zinc-mines of central Missouri by Dr. Charles Luedeking of St. Louis, is an indication that salts of ammonia may have been present influencing the deposition of the minerals, under conditions not hitherto suspected.f

The tendency of galena to deposit in rocks containing bitumen and organic matter, has undoubtedly, to some extent, influenced this concentration of the galena in certain strata ; but the constancy of the occurrence of lead-ores in formations near the surface, the world over, shows that the cause of this vertical distribution of the ores must be sought in other directions than in the chemical composition, or lithological peculiarities of the formations carrying the deposits.

Origin of the Ores. — The evidence obtained in this investigation indicates that the ores and the associated minerals have all been deposited from aqueous solutions, probably of moderate or normal temperature and pressure, and that trhe fissures connected with the ore-bodies have formed the channels through which the mineralizing waters were introduced. It is also evident that the lead and zinc were not derived from the geological formations in which the depa<its occur, or from the overlying or underlying sedimentary strata, but that the source of the metals was exotic and was probably deep- seated in the primitive rocks.

Form of Occurrence of the Ore-Deposits of the Upper and Lower Mississippi, — The deposits of lead- and zinc-ores of the Ozark and Wisconsin uplids do not occur invariably in the form of runs. In

Report im the QmeksUver Deposits cf the Pacific Slope.

t Paper read before St. Louis Academy of Rcience, April, 1890.

216 Lead- And Zinc-Dep06Its Op The Mississippi Valley.

a few instances noted in Jasper and Newton counties Missouri, fis- sures traverse the Cherokee limestone without dbturbing the strati- fication, so that in the deposition of the ores, the solutions unable to escape from the fissures and penetrate the wall-rock, deposited the ores in the breccia included between the walls, forming ore-bodies which are vertical, tabular sheets of ore, closely resembling in struc- ture the ore-shoots of fissure-veins. Near Dubuque, Iowa, lead- ores occur in vertical sheets, often of large size, deposited within the eroded walls of the crevices ; smaller bodies of galena, in like man- ner included within the fissures, are found in many localities in the Wisconsin region. These occurrences indicate the close relationship of runs to the fissure- veins, and show that the form of the ore- bodies has resulted from the composition and structural condition of the wall-rock.* Where the ascending solutions could escape from the fissures and penetrate the beds, owing to the permeability of the rocks or to their fracture or brecciation, runs have been formed ; and where the wall-rocks have been impervious, and therefore have con- fined the solutions within the fissures, the ore-deposits are in shoots.

The fact that in runs the ore is confined in vertical extension to the thickness of the ore-horizon cannot be regarded as a specific dif- ference, for in fissure- veins the ore-shoots are often strongly influ- enced in occurrence and in location by the character of the wall- rocks, and many instances are seen in the mining-regions of both America and Europe where fissure-veins become barren in passing from one geological horizon to another, and are productive only where traversing certain strata.

OccufTence of Runs in the West. — The occurrence of ore-deposits in favorable horizons and in the form of runs is not confined to the Ozark and Wisconsin areas, but has been observed by the writer in many mining-districts in the West, in localities where the sedimen- tary rocks are substantially horizontal, and are traversed by vertical faulting fissures affording the channels for the introduction of the ores. The occurrence of runs in the Potsdam sandstone of the Black Hills, South Dakota, may be note<l, where they are formed by the intersection of certain calcareous or shaly beds, which constitute the favorable horizons, by vertical faulting fissures. In this mining- region the runs carry argentiferous galena in the Bear Butte district, while in the neighboring district of Bald Mountain, over a broad area and in the same Potsdam formation, extensive and richly-im- pregnated beds or runs of refractory gold-ore are mined. Large com- pound runs of argentiferous lead-ore are found in the Carboniferous

Lead- And Z1Nodep08It8 Op The Mississippi Valley. 217

limestone at Horseshoe Gulch, near Leadville, Colorado, and the re- cently discovered mines of silver-bearing galena near GU)od Springs, Lincoln county, Nevada, similarly occur in runs or impregnated beds in limestone of Carboniferous age.

The ore-deposits of the Ozark and Wisconsin uplifts, with all the multiform local variations, may be classed under two generic forms : deposits filling enlarged sections of the crevices; and impregnations extending outside of the fissures which introduced the mineralizing solutions. These forms are all included under the designation of runs, adopted from the miners of the Southwest

In conclusion, no good reason appears for placing these ore-de- posits in a separate class from fissure- veins because of distinctions based upon the form and position of the ore-bodies relative to fissures. On the contrary, these runs, considered in connection with the fissures through which they were formed, belong to the great class of ore- deposits in which the origin of the ore is in mineral-bearing solu- tions, ascending through the fissures from some source of unknown depth in the crust of the earth, of which class fissure-veins are the simplest typical form.

Time at which the Formation of the Ore-Deposits Occurred. — All the deposits of zinc- and lead-ores of the Mississippi valley appear to have been formed at the same period of time, without respect to the age of the geological formations in which they occur. Though the time at which the deposition of the ores took place cannot be definitely fixed, we are not without evidence bearing upon the ques- tion. The ore-forming period was certainly subsequent to the Car- boniferous, as the ore-bodies in the Cherokee limestone extend up into the overlying shales of the Goal-Measures. Further, a great length of time was expended in effecting the subterranean erosion of the limestones that antedated the ore-forming period ; a work accom- plished by the slow solvent action of percolating surface-waters, and one that could commence only when the region became dry land at the opening of the Carboniferous.

Boulders of float-galena derived from the ore-deposits have been found in the Drift in Iowa'*' and in Pliocene gravels along the Mis- sissippi river near the mouth of the Ohio;t water- worn fragments of chert carrying galena, occur in the Orange sands near Helena, Arkansas, regarded as pre-Pleistocene in age.| Fossil bones and teeth of the mastodon, megalonyx, and an extinct species of peccary

♦ White. t McCJee. t Branner.

218 Lead- And Zinodeposits Op The Mibsib8Ippi Valley,

have been discovered in the upper parts of lead-bearing crevices in Wisconsin.** These and other evidences obtained in the course of this investigation make it probable that the primary de|X)sition of the ores had terminated before the later Tertiary. Since the c]os*e of the period of deposition, the ores near the surface have been oxidized by atmospheric agencies to a depth locally as great as 100 feet. CaveSy many of which are of great extent, have been formed in the limestone by the action of surface water, and the larger streams draining the mining-regions have deepened their beds from 50 to 100 feet.

Between these phores of time, the epoch of the Coal -Measures and the Pliocene, the formation of the ore-deposits in all probability occurred. No ores are now being deposited in the strata, and con- ditions other than those at present existing roust have prevailed in the ore-forming period. That the deposition was the result of dynamic action deeply located in the crust of the earth, follows from the connection of the ore-deposits with faulting fissures, and the acceptance of a source of the metals in depth.

The only epoch in Mesozoic or Cenozoic time known to have been characterized by other than local dynamic action in the Mississippi region, is that which begins with the continental disturbances at the close of the Cretaceous and continues to the later Tertiary, — period distinguished by intense dynamic action which extended its effects over the western half of the North American continent, and was signalized in many widely separated regions by great outbursts of eruptive rocks.

In the Ouachita uplifts the eruptive intrusions are postCreta- ceous.f The outbursts of trap and basalt in Indian Territory and Texas are of the same age.| The Ouachita and Ozark uplifts are closely contiguous, the narrow valley of the Arkansas river mark- ing the separation ; and while they differ radically in formation and structure, these insular areas are intimately related with respect to the dynamic events that have taken place in their geologic history. The extensive igneous disturbances of the Ouachita uplift were presumably connected with the formation of the veins of argentifer- ous lead- and zinc-ores; analogy would imply that the ores of lead and zinc of the Ozark region were probably deposited at the same period.

The epoch of the formation of the argentiferous ores of the Rocky

♦ Whitnej. f Branner. t Hill.

Lead- Akd Zinc-Deposits Of The Mississippi Valley. 219

Mountains was inaugurated by the disturbances at the terraination of the Cretaceous period and continued through the Tertiary.* The silver- and lead-deposits of ihe Black Hills of Dakota were deter- mined by the writer to have been formed at this same period and to have resulted from igneous action connected with the extravasation of the eruptive rocks. A parallel has been drawn between the early geologic history of these areas of uplifl in the Mississippi valley and that of the outlying elevations along the eastern slope of the Rocky Mountains. A similar uniformity can be traced, in these far sepa- rated mining regions with respect to the dynamic disturbances that have taken place, and the formation of the mineral deposits.

In conclusion, all evidences point to the deposition of the lead- and zinc-ores of the Mississippi valley as the result of disturbances which, while local in their action, were connected with the wide- spread disturbances, accompanied by intrusions of igneous rock and the formation of mineral-deposits, that in post-Cretaceous and Ter- tiary time extended over the Rocky Mountain system from Mexico to British America.

Discussion op the Theories of Ore-Deposition.

The absence of all igneous rocks and all evidences of igneous action in the ore-bearing formations of the southern Missouri and the Wisconsin-Iowa mining-regions, renders it unnecessary to con- sider any theory in explanation of the deposition of the lead- and zinc-ores, based upon igneous injection or sublimation of the metals. The discussion may therefore be confined to the theory of ascension, or deposition of the minerals from solutions rising from unknown sources in depth, and of lateral secretion, or the deposition resulting from the concentration, in the crevices, openings and favorable beds, of the minute particles of the metals originally disseminated through- out the inclosing strata. In this discussion, the term lateral secre- tion is used in a somewhat restricted sense, to mean that the ores have been derived from minute disseminated particles of lead, zinc, iron and copper, primarily deposited in the inclosing sedimentary rocks, or in the beds overlying or underlying the ore-horizons; and that by a process of segregation through the medium of the subter- ranean circulation of atmospheric waters, they have been leached from the strata and concentrated in the crevices and fissures in which they are now found.

Clarence King.

220 Lead- And Zinc-Deposits Op The Mississippi Valley.

This theory of lateral secretion was advanced by Whitney in The Metallic Wealth of the United States, published in 1854, and subse- quently in more detail, in his Report on the Upper Mississippi lead-region. Whitney expressed the opinion, from his investiga- tions of the Wisconsin -Iowa lead-rion, that the crevices now filled with ores of lead and zinc, originated through the action of local causes, or offerees limited in operation to a comparatively narrow vertical range, and to rocks of similar lithological character ; and that the origin or source of the metals was in the Trenton and Gra- lena limestones. He regarded it as improbable that the ores re- sulted through deposition from solutions ascending from the deep- seated Archsaan rocks, and in summing up the discussion of the origin of the ore-deposits, he says : '' In view of all these facts we consider it as a matter settled beyond all possibility of doubt that the lead-deposits of the Northwest must have been introduced into the fissures from above, and by precipitation from a solution."* The results of the present investigation have led to a different con- clusion, not only as to the origin of the metals, but also as to the manner of formation of the ore-deposits. This conclusion is sup- ported by many minor details in the occurrence of the ore-deposits and the associated fissures that cannot be set forth in this brief paper.

The strongest argument in favor of the mode of deposition from solutions rising from unknown sources in the crust of the earth, is that the ore-bodies are associated with faulting-fissures of indefinite extension in depth, and that all the evidences of the occurrence of the ores and minerals in the deposits point to these fissures as the channels through which the mineralizing solutions were introduced. The localization of the ore-deposits is difficult of explanation by any theory of lateral secretion. The Cherokee limestone extends over an area of 4000 square miles in the Southwest, yet it is everywhere barren except in the mining-districts dotted over the rejfion. The magnesian limestones cover one-third of the area of the State of Missouri, but carry ores in districts of limited extent distributed without regularity, while great tracts of these limestones are un- mineralized. These Silurian limestones are exposed over large sec- tions of the Mississippi valley, but are ore-bearing in only a few localities. If the minerals and ores in these formations were de- posited by lateral secretion, it is not easy to understand why the ores do not occur more generally distributed. The true explana-

Heport of a Geological Survey of the Upper Miumippi Lead Region by J. D. Whitney, 1862, p. 39S.

Lead- And Zin0-Dep08It8 Of The Mississippi Valley. 221

tion 18) that these local areas are mineralized because they are centers of disturbance and faulting of the strata, and the surrounding terri- tory is barren because of the absence of any dynamic action that has been capable of forming fissures extending to the deep, whereby the ore-depositing solutions could gain access to the strata.

The minerals that form the ores and the gangue are not such as would be predicated, were the metals derived from the inclosing strata. Had the lead and zinc been segregated from the limestones in which the ore-deposits are found, it would be natural to expect that the prevailing gangue of the ores would be calcite; but this mineral occurs-in relatively small proportion in the ore-bodies and is always of secondary deposition, overlying the primary ores. Dolomite is a very rare mineral in the ore-deposits in the Galena limestone in Wisconsin, although that formation is highly magne- sian in composition. On the other hand, vast masses of dolomite form the gangue of the ore-bodies in the Cherokee limestone in the Southwest; yet analysis shows only traces of magnesia in the lime- stone itself.

Neither will any theory of lateral segregation explain the para- genesis of the minerals in the Cherokee limestone; or how it hap- pens, when the conditions are apparently identical in the same hori- zon, that one or more of the primary elements, dolomite, blende, galena or cherokite, may be absent in an ore-body, though abundant in the mines in the vicinity ; or the occurrence of deposits of blende ' free from galena, and of corresponding mines of galena without even traces of zinc-ore, and of ore-bodies formed of these minerals inter- mingled, all occurring under apparently identical conditions and in rocks of the same lithological character. If the deposition of these minerals is assumed to have been effected by successive upflows through the fissures from some source in depth, of solutions of dif- ferent chemical composition, it is easy to understand the uniform order of formation throughout the region ; while the absence in certain mines of one or more of these primary minerals, may be explained by variations in the mineral contents of the solutions in different localities, or by the temporary suspension of the mineralization from some cause.

The examination of the Cherokee limestone, which is the princi- pal ore-producing formation of the Mississippi valley, affords no evidence that the ore-deposits have been derived from that horizon, or from the sedimentary rocks situated above or below it in the geo- logical column. The traces of lead and zinc detected by an analysis

222 Lead- And Zinodep06It8 Of The Mississippi Valley.

of the unaltered Cherokee limestone are entirely inadequate (were it possible to conceive of a perfect concentration of the disseminated metals in the rocks from miles on every side of the mining-districts) to form the lai ore-bodies that are grouped together.

It is reasonable to suppose that the solutions resulting from lateral secretion would be weak and would deposit their metals slowly, but this does not correspond with the occurrence of the ores in a number of the mines in the southwest, where the small size and crowded growth of the crystals indicate a rapid deposition from concentrated solutions. Lateral secretion, if it takes place at all, is in all prob* ability confined to the zone of oxidation ; and by the general to- pography of the mining regions, this zone is limited to a depth of 100 feet from the surface, and averages less than 60 feet The minute traces of lead and zinc disseminated through the sedimentary rocks probably exist as sulphides, and in this form they are in- soluble and permanent in ordinary subterranean waters, unless they become oxidized and decomposed — an action that is confined to the zone of oxidation near the surface.

The perfect faces and sharp edges of crystals of blende and galena, lining the water-channels in the lower parts of the ore-bodies, show that, below the zone of oxidation, no solution or decomposition of the metallic sulphides takes place. Even where the lead and zinc be- come oxidized and pass into solution in the circulating waters, con- tact with the organic matter contained in the rocks in the presence of alkaline sulphates, which occur in all mineral waters, immediately reduces and precipitates the metals as sulphides.

PbO,C02 + CaO,SO, + H,0 + 2C - PbS + CaO,CO, + 2COj + H,0.

It is this protective action of the organic matter disseminated through the strata, that has limited the zone of oxidation to so shal- low a depth in the mining-regions; for until all the carbon contained in the rocks is first consumed by oxidation, no decomposition of the minerals can occur or any sregatiou of the metals take place. There is nothing to warrant the belief that the normal underground circulation of waters will effect the extracon and segregation of the ores disseminated through the rocks below the zone of oxidation ; and while it is probable that by some process of lateral secretion, the galena and blende which are occasionally found in small quanti- ties, crystallized in the interior of geodes, or filling the cavities in fossil shells, are derived from adjacent strata, no action of this

USAD- AND ZlifC-DEPCSITS OF THE MISSISSIPPI VALLEY. 223

nature is adequate to form the immense ore-deposits of the Missis- sippi valley.

The deposition of the ores has long ceased ; no ores of lead or zinc are now being formed, other than by secondary deposition resulting from the decomposition of the primary ores, and it has been shown that the ore-deposits were formed probably in post-Cretaceous and Tertiary time, certainly later than the Coal-period. To account for a deposition of the ores by lateral secretion at a definite period of time, and not in the ages prior or subsequent to the Tertiary, although the country has been continuously dry land from the beginning of the Carboniferous to the present, it would be necessary to assume that the conditions under which the segregation of the metals took place were not only substantially diflferent from those now subsist- ing, but were different from those occurring in the interval between the termination of the Subearboniferous and the close of the Creta- ceous. Nothing, however, appears to warrant the assumption that;, in the ore-forming period, the prevailing conditions of climate were such as to give the circulating sub-aerial waters a different chemical composition, or increased solvent power, or a temperature or pres- sure greater than that of the subterranean waters of the present time.

Conclusion. — The burden of proof is upon the advocates of the theory of lateral secretion ; for the mineral-deposits of the Appa- lachian and the Rocky Mountains, of the Lake Superior region and of southwestern Arkansas, that surround on all sides the lead- and zinc-mining regions of the Mississippi valley, are all formed by ascension. The peculiar occurrence of the ore-deposits in the form of runs, is characteristic as well of certain mining-districts in the West as of the Ozark and Wisconsin uplifts ; of mines of lead, silver and gold that are unquestionably formed by ascension and not by lateral secretion.

The only theory which, in all observed instances, will account for the occurrence of the deposits of lead- and zinc-ores and the associ- ated minerals in the Upper and Lower Mississippi region is that of ascension, the source of the metals existing deep in the primitive rocks. With the discovery that the ore-bearing crevices are fault- ing-planes of indefinite vertical extension, the classification of the deposits of the Mississippi valley as the fillings of 'gash-veins," or crevices formed by the contraction or shrinkage of the rocks, and confined to a narrow vertical range within the geological horizon, mubt be abandoned.

224 lead- and zinc-deposits of the mississippi valley.

Rules for Prospecting.

As one result of this investigatioDy the following rules maj be laid down for the guidance of the miner :

I. — The old rule "to /oHowj the ore," holds good in this as in other mining-rions, but cases arise where other guides must be sought ; as where the ore-bmly is worked out and search is made for other deposits in the vicinity; and in the prospecting of virgin ground and the exploiting of areas but little developed.

II. — In all underground prospecting the general rule may be given, to foUow the mare prominent vertical fissures in the search for ore; for these have been the channels through which the solutions have entered the rocks and formed the ore-bodies, and along the course of which, in favorable ground, the deposits of ore occur.

III. — In prospecting new ground, attention should be given to the indication of the course of the fissures and cross-fissures; the tDork should be concentrated upon the areas of crossing or intersection of the different belts of fissures ; for experience has shown that the largest ore-bodies are situated at such crossings of different fissure- systems. On the surface the course of the fissures may be traced in some localities by the direction of low bluffs or breaks, or by sags or lines of depression in the even contour of the topography ; also by the strike of outcrops of silicified rock, more or less mineralized and stained with iron. When carefully searched, such outcrops often afford traces of the oxidized minerals resulting from the weathering of galena and blende. Evidences of the disturbances of the rocks should be carefully observed ; such as, beds dipping locally at steep angles, or in a direction different from that of the prevailing inclina- tion of the strata in the region ; and the occurrence of belts of folded, crushed or brecciated rocks. The character of the vegeta- tion upon the surface is usually an uncertain guide in prospecting; in rare instances, the course of an ore- bearing crevice is marked by a narrow belt which isdestitute of vegetation on account of the pois- onous action of the salts produced by the decomposition of the min- erals. Underground, the course of the fissures, not only in the mine but in all workings of the vicinity, should be carefully sur- veyed and platted. The probable intersection of the different fis- sures may be determined by prolonging on the map their surveyed courses both underground and on the surface.

IV. — An advisory rule may be given, never to sink a shaft without having put dovm a drill-hole, in order to ascertain the character of

OBSEBYATIOKS COKCERKIKQ ORE-DBEaSINO. 225

the nnerlying formations, lest time and money be wasted from striking hard and massive strata, or areas of barren rock. In the future, prospecting drills, such as are commonly used in sinking artesian wells, will be more generally employed in the search for ore. The diamond-drill is not adapted for this work in prospecting in the Cherokee formation, on account of the loose and open struc- ture of the ground and because the hard chert cuts out the diamonds. In the Cambrian limestone the massive and uniform structure of the beds and the absence of chert are favorable for the successful em- ployment of the diamond-drill.

Okseeal And Special Obsebvationq Concerning 0Be-Dbe88Ing,

BT O. BILHARZ, ROTAL SAXON OBERBERORATH ON THE RETIRED LIST, BERLIN, GERMANY.*

(Chicago Meeting, being a part of the International Engineering Congress. August, 189S.)

Design and Operation of Dressing-Works.

The rules laid down by Peter von Rittinger in his classic Lehr- buck der Aufbereiiungskunde (Berlin, 1867, p. 513) concerning the designing of dressing- works still possess authority for such works and their operation. Nevertheless, some modifications and additions suggested by practice may be of value at the present time.

Rittinger says :

" 1. Choose the apparatus and machines which can he, under the existing con- ditions, easily constructed and kept in running order/'

This rule contemplates the construction of apparatus (mostly of wood) on the spot, with the command of good shops and especially of good workmen in wood. But in many cases, particularly for the sake of a prompt supply, the construction of apparatus is now lefl to special manufactories, which pay more attention in the choice of materials to durability and strength than to easy manipulation and erection. Nevertheless, here, as in many other instances, simplicity of construction and solidity of execution are to be recommended. Especially important is the convenient and cheap replaceability of

Translated by the Secretarj.

226 Observations Ooncerning Ore-Dressing.

those parts which are exposed to more or less inevitable injury. It is even advisable to keep duplicates of such parts on hand.

" 2. The more valuable the products to be obtained, the more perfect should be the arrangements adopted."

The rule is to be unconditionally endorsed. Yet it may be pointed out that frequently arrangements must be made for the treatment of a material of small intriuRic value, which, however, promises by reason of the great mass of its natural deposits, to sup- port a profitable business after certain difficulties of manipulation have been overcome. In such a case also, and perhaps even in a still greater degree than where the final product is of greater value, the highest perfection of the works may be important.

The general aim in mechanical preparation is, whatever may be the ultimate value of the product to be extracted from the crude material treated, to secure the highest results with the simplest and yet most perfect means and arrangements, so far as these desiderata can be practically combined.

''3. If it is a question of the most perfect arrangements practicable, and if other considerations do not narrow the choice of apparatus and machines, the following will be used :

( 1 ) For finely disseminaled mineral :

a. For crushing : Stay-batteries* and syphon-discharge or overflow-batteriesf onlj when the crushing is to be very fine.

6. For sorting: Syphon classifiers or spitzlutten with pointed boxes or spitzkasten for the slimes.

c For separation : Continuous percussion-tables.

d. For saving free gold : Gold-mills.

(2) Far coarsely disseminated mineral:

a. For crushing : Stay-batteries with coarser screens, and rolls only when the ores are very coarsely disseminated. 6. For sizing : Drum-sieves.

e. For separating: Continuous plunger-jigs or Setzpumpen.*

I think that " the most perfect arrangements practicable " should be sought in every case; and would extend the conception of such

♦ Stay battery {Slausatz). A stamp-battery with screen covered by a water-tight plate with a discharge opening below. By this arrangement the water-level in the mortar is raised, and the discharge through the screen becomes continuous, instead of intermittent and during the splash only. The result is an economy of water, and greater uniformity in crushing.

t Syphon-discharge or overflow-battery (Sehubersatt). Stamp-battery without screens, discharging through a slot and a rising channel, the latter having the efieet of a hydraulic classifier in determining the size of the escaping grains.

Observations Ooncernikg Ore-Dressing. 227

perfection to the planning of the works at the beginning in such a way as to permit at any time, without interruption of the existing operations, such additions and extensions as might be needed.

la. As to very fine crushing {todtpochen), I am persuaded that it should be avoided on general principles, in every case; since even in a very finely disseminated ore a grade of granulation may be effected which permits enrichment by jigging, with less loss than is inevitable in the treatment of slimes.

The liberation of even the finely disseminated minerals by crush- ing should rather proceed by successive steps, and this can be se- cured by coarse stamping {Roschpochen) since the mixture of grains from the finest sieve can be returned if desired to the stamps.

16. For classification, the syphon-classifier {Spitdviie) may be pro- nounced the most perfect apparatus, but with the improvement upon Rittinger's which is shown in the Bilharz spitzlutte, in which the discharge-opening constitutes a slit of the entire breadth of the channel, so that every partial variation of speed in the discharge is avoided. The Bilharz apparatus has also the noteworthy advan- tage, that it consists of a number of separate, similar elements, in- creasing successively in breadth, each element being conveniently controlled by itself as regards pressure of clear water and speed of discharge. It is thus easily adaptable to any demand. Increased capacity, for instance, can be secured without interrupting the opera- tion or changing the arrangement of the plant, by the introduction of a larger number of elements of the proper breadth. For a tem- porary reduction of capacity, it is simply necessary to shut off some elements.

No objections have been encountered in practice to the use of spitzkasten, in the form proposed by Rittinger, for slimes and fine sands.

Ic. In separation, the continuous operation of tables is at the present day, when (in America, Germany and even Hungary) high wages play so important a part, an indispensable feature of every dressing-works using such machines. It has been secured advan- tageously in the so-called Stein tables, or continuous percussion belt- tables, introduced by the writer at the large central dressing-works of the Hiramelfahrt mine at Freiberg, where it was soon proved that this table is preferable to any other for eflSciency, durability, economy of power and cheapness of installation.

iSuch tables are now manufactured in excellent style and up to 1.25 to 1.40 meters in width at the Gruson, Buckau, Magdeburg. ,

228 Observations Ooncebniko Ore-Dbbbsing.

Id. lustead of the cylindrical gold-mill contemplated by Rittin- ger, with its revolving mullers, not based upon the principle of con- tinuous operation, apparatus of this class is to be recommended in which charging, amalgamating and discharging go on continu- ously— especially in machines which bring the gold-bearing material into the most intimate contact with the. mercury, by centrifugal action.

2a. What is now understood under coarsely disseminated mineral ought not under any circumstances to go direqly to the stamps. It will always be advisable to begin the reduction with crushing-rolls. I repeat that liberation of coarsely disseminated mineral must take place by successive steps, in order to leave a minimum share of the work to the wet treatment proper, upon the tables.

26. ClassiGcation according to size of grain has been recently pronounced superfluous by Hoppe, and the English practice of jig- ging without any such preliminary sizing has received general acceptance.

For my part, I concede that where the substances to be separated differ widely in specific gravity, a satisfactory result can always be achieved without careful sizing. But in materials possessing very slight differences of specific gravity, as for instance, blende and py- rites, a very accurate classification by grain is necessary to a success- ful separation.

2c. Jigs with valved pistons under the sieve, called sefzpumpen' by Rittinger, have fallen entirely out of use, by reason of the great trouble of maintaining the pistons and valves. The continuous sand-jigs of the Harz, or the Bilharz StaucAiefr-element-jigs, answer the purpose better and more completely.

4. The capacity of machines and apparatus most be rigliUj proportioned to the quantity of the material to be treated.''

Since the character of a machine is essentially affected by the kind of material treated (t.6., the form and grain-size of its several constituents), I think this consideration, rather than that of quantity, should be the controlling one.

'*5. If tlie site of the dressing- works b not predetermined, let it be so chosen that the transportation of materiab to be treated in the works shall be as simple as possible.''

Above all, repeated handling or dumping should be avoided, especially when the material might be thereby crushed in a maoDer

Observations Concerniko Ore-Dressing. 229

injurious to the utilization of tlie products to be won from it. On the other hand, hoists or lifts (oden simple water-hoists) available in the supply-system of raw material may often decide the location and design of the works, to which in any case they cannot fail to be appropriate and advantageous.

For the rest, the best utilization of suitable differences of level for the site of the works is to be studied, and the full advantage to be secured from thin is in most cases so important that it should have mor weight than the simplicity of the transportation-lines. Some complexity in the latter might be accepted if necessary to secure the greater benefit.

"6. The several apparatus and machines should be relatively so arranged that the middlings of each can be carried forward in the shortest and simplest way to the next following manipulation. Especially ishonld care be taken not to let the middlings descend unnecessarily, that they may not have to be hoisted again."

This rule should be observed in its full scope. Intermediate ele- vators are to be avoided as far as practicable, yet cannot be com- pletely dispensed with, even in works arranged in stories on the large scale. For coarse stuff (battery sands) bucket-and-chain ele- vators are to be recommended ; for fine sands and coarser stamp slimes double raff-wheels; for fine slimes, centrifugal pumps.

" 7. The simplest and cheapest motors should be used to drive the several ma- chines. In this class belong, as a rule, the motors driven by water-power, and only where water-power is not easily available should steam-power be employed.

' S. Too many machines should not be operated by one and the same motor; since, otherwise, operations might be too seriously interrupted by the stoppage of this one motor, and also because it is difficult to regulate the running of the indi- vidual machines.*'

In this rule Rittinger evidently had in mind only the running by water-motors, which require no special attendance, and when attached to small departments severally, can be easily stopped or started with- out affecting the general operation. But steam-engines are now almost exclusively to be recommended in order that the available clear water (in most cases (mly too scanty in supply) may be saved for the concentrators; and the employment of many small engines in a works would not be advisable, because, on the one hand, the cost of attendance would be greater, and, on the other hand, a large engine, in good condition and running at full capacity, is more eco- nomical of fuel than several smaller ones taken together.

The position of the motor should be such that the apparatus de- manding the greatest amount of power should be nearest to it.

230 OBSEBVATIONS COKCBBNIXG OBE-DRESSINa.

That all the principal apparatus of a conceDtrating-works should be dependent upon one central motor would not be disadvantageous, inasmuch as the whole operation is strictly continuous, and hence a stoppage of single apparatus or departments is not practicable.

But this does not prevent the appropriate use of small motors for supplementary work outside of the main operations ; for instance, the often desirable working at night of accumulations of old mill- products in works which are running on single shift mainly. In this way only certain departments may be run as required.

Continuous day-and-night running is, however, always preferable to the simple single-shift day-system, and where this continuous system is in use, the side-operations above alluded to are provided for with great advantage in small departments, totally independent of the main works.

9. By means of suitably disposed windows and a proper position of the working- room, provision should be made for adequate interior lighting by day. At night the end must be sought by a liberal lamp-illumination."

Rittinger aimed this rule at a defect of all the old dressing- works, from which the older establishments of Freiberg and Hungary are even yet not wholly free.

In modern designs special weight must be laid on the securing of as large lighting-surfaces as possible. E;$pecially to be recommended is the use of overhead lighting (even in snowy regions, notwithstand- ing certain inconveniences which there attend it). When side-illumi- nation is employed, apparatus which requires constant attention and regulation should be so placed that the workman, standing with his back to the light, will have the full effect of it upon the apparatus to be watched.

The use of electric lights (incandescent over single machines and arc-lights over hand-sorting tables, yards Bnd dumps) is highly advisable and removes a great objection to the night-work, which would be otherwise less worthy to be recommended.

Electric lighting excludes danger from fire, possesses greater ca- pacity for distribution, greater convenience, and in almost all oases is, for the same intensity of light, lower in cost.

To these nine rules of Rittinger*s I would add a tenth, as follows:

10. The arrangement of a dressing-works must be clear, comprehensive, logical, and especially calculated to insure continuity of operation for the whole range of the business.

Rittinger distinguishes four groups of dressing-works, according to the greater or less simplicity of the processes employed, namely :

Observations Ooncerning Ore-Dressing. 231

" Group L — The materials to be concentrated require no crush- ing." (Here gold-sands and placer-materials are spepiallj meant.) "Group II. — A single crushing suffices to liberate the mineral." " Group III. — The crushing must be done twice." Group IV. — The crushing must be preceded by a sizing." This classification is neither complete nor practical. In which of these groups could we rank the concentration of calami ne-earth, which, although one of the processes involving the greatest number of manipulations, requires no crushing, but, on the other hand, a very careful sizing? On the other hand, it is not impossible that even gold-bearing sands might need to be pulverized. Again, there are few cases in which a single crushing is sufficient; on the con- trary, in most, if not in all cases, a successively advancing or, in other words, a repeated reduction in size should take place.

Works for mechanical preparation may well be divided primarily into :

I. Works treating coal and bituminous substances.

II. Those treating metallic 8ul)stances.

And the latter may be subdivided according as they perform (a) a simple, or (6) a multiplex separation — that is, the separation of several constituents.

I. Works for the Preparation op Coal and Bituminous

Substances.

These must, usually, operate upon large quantities, and the follow- ing are control'ng conditions: (1) Convenient delivery of crude material. (2) Maximum simplicity and rapidity of its treatment. (3) Maximum facility for disposing of both the valuable product and the refuse.

Local conditions are primarily decisive for the general arrange- ment. If thre be a navigable river, or an available branch rail- way, in the vicinity of the point of production of the raw material, it is always advisable to locate the works immediately by the river or railroad. If a bluff be available, it is well to build the works in terraces ; on level grounds, a main hoist for the crude material should be provided, and the buildings should be arranged in stories, so that in either case the crude material will need to be elevated but once, and the separation of the valuable from the worthless portion can go on continuously in such a way that each will fall into the bins or. chambers designed for it, from which it can be directly loaded and removed.

232 Observations Concerning Ore-Dressing.

The connection between the shafl-house, adit-mouth or open work- ing, and the dressing-works, is made by means of either surfaoe-trams (operated with horsce or with cables), or aerial railways. Electrical motors can be employed, especially when water-power is at hand.

Under all circumstances, the crude material should be conveyed from the point of production to the dumping arrangement at the works in the same vehicle, without intermediate dumping and re- loading. Arriving at the dumping-platform, it should be passed over a grating or a movable screen to separate at once the lumps from the smalls.

a. Coal with, abundant lumps does not, as a rule, require (for the treatment of the smalls) an extensive plant, unless the coal is largely intercalated with slate ; in which case the lumps must be coarsely broken in order to be sized together with the smalls, then jigged, and finally also huddled {geschldmmert). Especial weight is laid, in the preparation of coal upon the production of separate sizes to meet a commercial demand. Hence the large perforated sheet-iron cylin- ders or separaling-drums have become one of the essential features of coal preparation.

The separation according to size is followed by the jigging of the different sizes, down to fine sand and slime. In view of the large difference in specific gravity between coal and slate, sizing would be of comparatively small importance where the commercial demand for it was not decisive.

A recently introduced apparatus for the production of small sizes down to the finest coal-slime (suitable for the manufacture of bri- quettes or of coke), is the Bilharz round piston-jig, with a bed {Orau- penbeit) of feldspar chips. This machine works continuously, and has a large capacity.

Essentially, therefore, an establishment of this class will consist of :

1. Coarse bar-screens for lumps.

2. Sizing apparatus, drum-sieves, Karlik pendulum sieves.

3. Coarse and fine jigs.

4. Classifiers for fines and slimes ; launders, spitzlutten, and spits- kasten.

5. Circular slime-jigs.

In which series, between 1 and 2, jaw-crushers may be introduced when the coal is tough and so very slaty as to prevent its use with- out breaking and cleaning.

6. Brown coal, bituminous slates, and earthy ozocerite require mostly, in the first place a preliminary washing.

Observations Concerning Ore-Dressing. 233

II. Works for the Preparation op Metallic Substances.

a. Simple Separations, — Nearest to the arrangements of a coal- washing plant, are those of the so-called calamine-washing works, the function of which is to separate and save the grains of metallic ore which are imbedded in clay. Here the washing away of the worthless matrix plays the principal part This is rapidly and effectively done by the so-called Crickboom washers (iauerromm/n), large sheet-iron cylinders mounted upon friction-rollers and provided with exterior cogged bands by which they are revolved in one direc- tion, while blades carried upon an independent axis passing through the cylinder and resting upon outside bearings, are revolved swiftly in the opposite direction, so that the material lifted in the interior of the cylinder as it revolves is caught in falling by the blades and chopped up. Tough clay is quickly reduced in this apparatus to fine particles, and these are dissolved to a thin " broth," out of which the valuable grains of mineral can be easily separated. For this pur- pose the diluted mass is conducted into separating-drums, from which the different sizes of grains pass to coarse and fine jigs.

Since the clayey materials washed away all contain fine particles of the metallic substance in the form of sands and slimes, the thin pulp escaping from the wash-drums {Trommeln) is conveyed into a system of syphon classifiers {Spitdutten) or hydraulic classifiers {Slrom gerinne), in which the floating particles are separated according to their rate of settling.

The separated ingredients which have thus settled in the differ- ent departments of the syphon-system, and, resisting the upward cur- rent of water in the syphon, have been discharged from it, are then concentrated upon Harz current-jigs for coarse material, or ui>on the Bilharz plunger-jig or the circular Stauchsiebatromaetzmaschine, for fine material, and finally, upon Stein tables with Bilharz closure (water- bed) for the finest sands and slimes.

The tailings from the tail-race are settled, to clear the water, in reservoirs which are periodically cleaned out.

b. Complex Separaliona, — 1. A very similar process is employed upon ores of manifold composition, that is, comprising several asso- ciated minerals, likewise imbedded in clayey gangue. Such are found in the Rhenish coal-regions, near Aachen, where the clayey matrix is the product of a decomposition of the country-rock. Only, in this case, the first washing is immediately followed with a crush- ing of the cleansed ore by rolls — in which operation a progressive treatment by successive stages is sought.

234 Observations Concerninq Ore-Dressing.

The steps of this process are as follows :

(1). Washing in Criekboom drums.

(2). Hand-sorting into pure lump-ore, mixed ore for the rolls, and country-rock. (N.B. Where both pyrites and blendes are present, special care should be given to the separation of blendous from py- ritic ores for the rolls.)

(3). Successive crushings, each connected with its own jigging.

(4). Sizing and jigging of coai'se stuff.

(5). Sorting of fine sands and slimes in spitzlutten or hydraulic separators.

(6). Fine jigging.

(7). Slime treatment (HerdarheU\ which includes washing on tables and belt-machines.

2. The preparation of ores of complex composition, extracted from the veins in the Palaeozoic rocks, is conducted without preliminary washing in a manner otherwise similar to the foregoing. Such ores, since they carry but little clayey materials, are dumped as they come from the mine, over grates or screens, in order to separate lumps from smalls. The former, mostly rough, angular pieces, are broken in rock-breakers, and a portion (of about the size of road-metal) goes to the hand-sorting tables, while the smaller pieces, together with the smalls from the mine, are delivered to the so-called rock-jig {Bergesdzmaachine) for direct removal of the worthless pieces.

Ores of this kind are generally composed of galena (usually argen- tiferous), zinc-blende, arsenical and iron pyrites. Often gray copper and its almost invariable associate, copper pyrites, are also present. A preliminary hand-sorting is required, and care should be taken, besides setting apart the rich ores not requiring further crushing, to divide the concentrating-ores into lead-ores proper on the one hand, and ores on the other hand, in which galena is not predominant — the latter being subdivided, if practicable, acconling as blende, iron pyrites, or arsenical pyrites predominate, in order that each class may be treated by itself.

The further conduct of the process is similar to that already sketched, with this addition — that, notwithstanding the hand-sorting above mentioned (which may be a coarse sorting, as distinguished from the elaborate fine sorting still practiced at Freiberg upon dry ores — Durrerze\ the jigging gives middle products; namely, mate- rial rich in galena, which goes back to the rolls, and material poor in galena, which is coarsely stamped for further liberation of the metallic particles. In any case the last crushing with rolls is fol-

Obsebyations Cx>Ncerninq Ore-Dbessinq. 235

lowed by stamping or granulation in ball-mills, with which is again connected the classification of fine sands and slimes in a current, the fine jigging and the slime treatment as above specified.

For all these materials, however, precisely as for the ores of the previous class, it is important to give to the stamps as small a part of the whole operation as possible.

3. Material containing gray copper is carefully removed in the hand-sorting, and separately treated — mostly dry.

The same is true of the rich silver ores, called Durrerze in Frei- berg. In that district, however, these ores occur in particular veins only, and can be kept separate after mining.

Concerning the degree of treatment to be given to such ores, the local conditions of the point of production and its distance from the smelting-works ara controlling considerations.

4. Gold-bearing materials contain their value either as free gold, or partly mechanically and partly chemically (?) attached to other metallic constituents.

(a). 0r€8 Containing Free Gold Only. — Reduction to fine sand and flour is almost always inevitably necessary by reason of the finely- divided condition of the gold in these ores. The process is usually performed in stamp-mills, but the Gruson ball-mill may also be recommended. The crushing may be either dry or wet. With dry crushing, pneumatic separation has given good results, furnishing at once the concentrated product desired.

With wet crushing, the escaping pulp passes first over the amal- gamating-plates, thence to the amalgamators (Lazlo's or others), and is then further concentrated, according to its value, in diminishing degree, in Bilharz spitzlutten or Rittinger spitzkasten. A final treat- ment upon the Stein-Bilharz continuous percussion belt-table may advantageously follow.

(6). Ores Containing Combined Gold. — The treatment of such ores is directed to the separation of the ingredients which carry or con- tain the gold. The most frequent of these is arsenical pyrites; less frequently, rich silver-ores are the gold-bearing constituents; and, rarest of all, galena.

In all cases the crushing must be performed with gre&t care and by successive stages, and all tailings must be carefully saved, since they may, in many instances, be still further treated by a chemical process, such as the MacArthur-Forrest.

Successive crushing and accompanying concentration by jigging is to be recommended for ores of this class, and the inevitable fine-

236 Bemakkablb Deposit Of Wolfbam-Ore Ik The V. 8.

jigging and 8lime-treatment is to be performed with the apparatus above specified.

The products of the treatment by coarse and fine jigging of aurif- erous arsenical pyrites can be best utilized by shipment to smelting- works, where they are roasted and smelted with lead-ores. The same can be done with the but half-concentrated slimes from the huddles or tables whenever the gold in them is combined with silver or lead.

Oy A BEMABKABLB DEPOSIT OF WOLFBAMOBE IN THE UNITED STATES,

BY DR. ADOLF GURLT, BONN, QVEiUXSY. (Chicago Meeting, being part of Uie IntemaUonal Engineering Congren, August, 1898.)

It has long been known that minute quantities of foreign sub- stances, when alloyed with steel, are capable of materially altering its physical properties. Thus, half a century ago, Faraday and Stodart, in England, discovered in the famous Indian steel, known as wootz," small quantities of aluminum, from 0.0728 to 0.693 per cent, and attributed its extraordinarily good quality to the presence of such minute quantities of that metal. Since aluminum is now produced by electrolytic means on a large scale and at low cost, its addition to cast-steel has been frequently recommended of late. Chromium was alloyed with steel fifty years ago by the celebrated French metallurgist, Berthier, of Paris, and has recently been suc- cessfully applied to the same purpose by Holtzer, of Terre-Noire. In 1860 titanium was alloyed with crucible steel by Robert Mushet, of Coleford, the alloy containing up to 0.50 per cent, of titanium, which was extracted from Norwegian ilmenite. The titanium-steel was much vaunted by him for its excellent qualities. Rhodium, a very rare metal of the platinum series, was maintained by Berthier to improve steel materially in strength and hardness. Manganese has always been held by Sheffield steel-melters to provide steel with excellent qualities, and this influence was quite recently investigated by H. M. Howe and L. Campredon. Nickel was tried as an addi- tion by Berthier, not to speak of gold, silver and platinum ; and nickel-steel has recently been brought into prominence by J. Riley and others, who recommend it for its great strength and toughness, particularly as material for armor-plates.

Remarkable Deposit Op Wolfram-Qre In The U. 8. 237

Wolfram-steel was made as early as 1855 bj Dr. F. Koeller at Reichramming, in Austria, and a few years later was imitated by Mushet who introduced it by the name of" Mushet's special steel" to the general public. Prof. Heeren, of Hanover, as well as Prof. L. Gruner, of Paris, investigated its remarkable properties, and found that when forged red-hot and cooled slowly it possesses an ex- traordinary degree of hardness, which, however, gives way to soft- ness when plunged red-hot in cold water, quite contrary to most other species of steel. As the degree of hardness can thus be ru- lated at will, it appears that wolfram-steel, when prepared with pure metal, has quite as good a chance in future as the best kinds of chro- mium- and uickel-steelsy and other advantages besides.

The influence of small additions of foreign metals, when alloyed to good steel, was recently demonstrated in an exhaustive way by Prof. Roberts-Austen, of London, whose well-known researches need only be mentioned here.

Among the wolfram-ores the mineral now called wolframite is the most frequent. It was known for centuries to German and Cornish tin-miners as an obnoxious mineral, though they had no notion of its true character. They had found by experience that when smelted with tin-ore in the furnace, it impeded the reduction of the tin and &cilitated its scorification, so they thought it ate up the tin as the wolf eats the sheep. For this reason the Germans named it " wolfart" or " wolfert" or " wolfrig," from which the special min- eralogical name wolfram or wolframite was derived. In Cornwall the miners applied the term " call " or " mock lead " to it because its great weight led them to suspect that it contained lead. But the Swedish chemist Scheele proved, in 1781, that this mineral, as well as another which he named tungsten, t.e., heavy stone, contained a specific mineral acid now called tungstic acid, and that wolframite is essentially a tungstate of iron and manganese, while tungsten, now known by the name of scheelite, is a tungstate of lime. That this mineral acid contains the metal wdlfram was discovered in 1786 by the brothers Jos and Fausto de Luyart. These two tungstate com- pounds are the most frequent wolfram-minerals, while two others, yellow oxide of wolfram or wolfram-ocher and scheelitine (stolzite) or tungstate of lead, are of rare occurrence.

These minerals were employed in 1848 by the English chemist, Robert Oxland, for the preparation of tungstate of soda, to be used as a mordant in dyeing cloth and, as proposed by Versmann and Lyon Playfair, for the impregnation of vegetable tissues, linen and

238 Remarkable Deposit Op Wolpram-Ore In The U. 8.

cotton, to render them non-inflammable and almost fire-proof. The same compound, when free from impurities, such as tin, copper, arsenic, bismuth, and iron, is the basis of the manufacture of wolfram-metal and other preparations of tungstic acid and the oxides of barium, copper, chromium, or of blue oxide of wolfram, or the tungstate of wolfram* oxide and soda. This last preparation has a shining bronse luster, and like most of the other tungsten com- pounds, which are characterieed by yellow, green, blue, pink, and gold colors, it is used in the manufacture of stained papers, etc. This brief sketch suflSces to indicate the great variety of industrial u.es to which wolfram-minerals are applicable when found in any considerable quantity.

In the autumn of 1887 I examined a highly interesting deposit of wolfram-ore in the United States, of which I propose to give a short description, with a view to recommending a more extensive utilization of the deposit. In my investigation I was greatly assisted by Mr. E. S. Olmstead, who lives near Stepney Depot, a hamlet on the Housatonic railroad, and whose good services and knowledge of the country were very useful to me.

Locality, — The deposit is situated near Long Hill station on the Housatonic railroad, in Trumbull parish, Fairfield county. Conn., about 8 miles from the city of Bridgeport. East of the Pequannok river, which empties into Long Island Sound at Bridgeport, the country is formed of rolling hills, of which South Hill attains a height of about 250 feet above the valley. The mining property covers this ground to the extent of about 66 acres, and is now owned by Mrs. A. E. Hubbard, of Port Chester, Westchester county, N. Y.

Previous Mining. — The occurrence of wolfram-ores in this lo- cality was known more than fifty years ago, and is mentioned by Shepard and Percival, who made an early official survey of the State.*

Shepard deals with the mineralogy and Percival with the geology of the State; and although the scientific views of both explorers are now somewhat antiquated, yet their observations of facts remain valuable. On the authority of these explorers, some statements about the occurrence of wolfram minerals have been widely spread through American and European scientific literature, frequently,

♦ E. U. Shepard, M.D., Hepori on the QtotogicdL Survey of OonnectiiMt, New Haven, 1837. James G. Percival, Report on the Gcoiogical Survey of Cbnneeticut, Kew Haven, 1842.

Remabkable Deposit Op Wolfram-Ore In The U. 8. 239

however, with slight inaccuracies as to the locality,, which is some- times placed in Monroe and sometimes in Trumbull parish. As a fact, the deposit lies in the latter; and the error may be explained by the circumstance that the owner at that time, a certain Charles Lane, had been mining in both parishes, — in Monroe for lead and bismuth, which occurred in association with magnetic iron pyrites on a quartz- lode traversing gneiss, and in Trumbull at the above-mentioned place. This mining enterprise ceased when the gold-discoveries in California began to attract the metal-miners of the Eastern States, but the Trumbull mine was started again at a later date by Thomas R. Hubbard, and worked till about 1874, with very indifferent re- sults, for copper, lead and silver, under the superintendence of Capt. James Arthur. Although wolfram-minerals were then known or supposed to be connected with a very flat vein, they attracted no at- tention except as mineralogical specimens. The next proprietor, William L. Hubbard, utilized only the limestone on the property, burning it in a kiln for agricultural and building purposes. Several lodes which intersect the property have been worked for quartz, fluorspar, and feldspar, the products being sold in New York for the manufacture of porcelain.

Oeology, — Hubbard's mine is in a district composed chiefly of a younger metamorphic amphibole-neiss, of a dark blackish color, containing quartz, feldspar, mica, and amphibole, alternating with l>eds of mica-gneiss, in which mica predominates. The gneiss en- closes a bed of crystalline limestone, 35 to 45 feet thick, exposed over an area of about 25 acres, and only covered by gneiss in South Hill near the southern boundary of the property, where the copper-, lead- and silver-mining had been carrial on. The limestone is highly crystalline, in fact a marble, and is intersected by several true-flssure veins, in the neighborhood of which it contains foreign minerals, such as mica, pyroxene, analcite, and amphibole. The lodes or veins are in no sense confined to the limestone, but traverse the underlying as well as the overlying gneiss. They may be enumer- ated as follows :

(1) Quartz Lode. — A lode of white, milky quartz over 6 feet wide, called the " Champion lode," is found in the western part of the property. It bears northwest and southeast with an almost vertical dip, and may be traced for more than half a mile. It is generally composed of admirably pure quartz, but sometimes contains small quantities of magnetic pyrites, copper- ore, and galena near the walls. As, according to Shepard, a similar quartz-lode in the old mine near

240 Remarkable Deposit Op Wolfram-Ore In The U. 8.

Monroe contained in addition bismuth, tin-stone, arsenical pyrites, and blende, it would not be surprising if, upon proper investigation, tin-ore should also be found in the Champion lode of Hubbard*s mine, though at the time of my visit I was unable to detect any traces of it, possibly because the pits on South Hill, sunk about 1873, were inaccessible.

(2) Topaz Lode. — This remarkable lode intersects both the gneiss and limestone in the east part of the mine, bearing almost parallel with the Champion lode and dipping east. The vein is about 8 feet wide, and the highly interesting feature about it is that its center consists of white, vitreous, and compact topaz, which is often found crystallized in druses or cavities, the crystals being 3 to 7 inches long and 6 to 6 inches in diameter. The large crystals are seldom transparent, while the smaller pellucid ones are usually white, with a tinge of green or yellow. Forchhammer gives for their composition :

Per cent.

Silica, 35.39

Alumina, 55.96

Fluorine, 17.35

The crystallized topaz sometimes contains other minerals such as pale grn beryls. The middle band of topaz is bordered on either side by capel or a zone of violet fluorspar, quartz, blende, and margarodite or white, pearly mica ; the latter is of rather rare occur- rence, and, according to the analysis of Smith and Brush, contains:

Per cent.

Silica, 45.70

Magnesia, 1.15

Water 4.90

Alumina, S3.76

Soda, . 2.85

Iron oxide, 3.11

Potash, 7.44

Chlorine, trace.

Fluorine, trace.

The topaz vein appears to affect the side-rock to some degree. Thus, in the proximity of the vein, the amphibole-gneiss occasionally carries calcite and yellowish epidote, and the limestone carries analcite, py- roxene, coccolite, amphibole, euclase, mica, and sphene or titanite. It is reported that this vein has occasionally yielded also wolframite, which, however, I failed to detect.

Bemabkable Deposit Of Wolfbam-Obe In The V. 8. 241

(3) Feldspar Lode. — The topaz vein and its enolofling rocks are cat by a powerful lode, which is composed almost entirely of fine, granular, white al bite- feldspar, sometimes containing beryl and blende in small quantities. The vein is 7 to 8 feet thick, and has a northwest and southeast strike. It does not seem to disturb the to- paz vein perceptibly, nor does it exhibit well-defined walls; the vein-filling gradually merges into the country-rock without any dis- tinct line of separation.

(4) Wolfram Ore-Bed. — The wolfram-ore, consisting of wolfram- ite, scheelite, and wolfram-ocher, occurs upon a so-called contact- deposit, 3 to 6 feet thick, which is embedded between the crystalline limestone and the lower gneiss, and is conformable with their gen- eral dip and strike. It is evident that this contact-deposit is inti- mately connected with the above described lodes which traverse it, and that to them, as feeders, it owes a great portion of its minerals, which may have been deposited from solutions or thermal waters that rose along the fissure-lodes and found lateral vents along the contact of the gneiss and limestone. The close genetic connection of the lodes and the ore-bed seems to be confirmed by the fact that the topaz-vein occasionally yields wolfram minerals. When I visited the mine I found on the western outcrop of the ore-bed, about mid- way between the topaz and quartz lodes, a few shallow pits and a short adit in the gneiss. The' pits had been sunk vertically through the ore-bed into the gneiss and then had been abandoned, instead of following the bed, which dips east 20 to 25 degrees. Nevertheless, a not inconsiderable quantity of wolfram-ore, chiefly scheelite and wolframite, had been taken out. I put a couple of laborers at work in these small excavations and also sunk costeaning pits at other points along the outcrop, always following the slope of the bed, and though I had only a few days at my disposal, some rather significant results were obtained.

In its principal mass the ore-bed consists of vitreous, translucent quartz, of an entirely different character from that of the Champion and topaz- veins. It usually forms a compact mass, containing cavities or druses studded with quartz crystals that are frequently covered by a thin film of yellow wolfrara-oeher. The quartz penetrates through crevices and fissures into the limestone and the amphibole gneiss, both of which near the contact must, therefore, be taken as part of the ore- bed. The quartz also contains iron pyrites, epidote, calcite, mica, and the wolfram-minerals, scheelite and wolframite. The latter occur embedded not only in the quartz but in the adjoin-

? IvL

242 Remarkable Deposit Op Wolfram-Ore In The U. 8.

ing metamorphosed beds of the oountry-rock, as well-shaped crystals or solid lumps and strings. The crystals are numerous and often of considerable size. It is remarkable that the wolframite crystals never show the peculiar crystallization of this mineral, but always that of scheelite. They are really pseudomorphs, and indicate that the original wolfram-mineral was scheelite or tungstate of lime, which was subsequently altered into tungstate of iron and manga- nese. The crystals are sometimes only partially converted, showing both minerals in the same individual. Both the crystals and lumps are usually loosely embedded in the matrix and easily detached.

As the ore-bed is traceable all along the outcrop between the lime- stone and gneiss, it is probably continuous in the whole basin of the gneiss, which is filled by the limestone over a surface area of about 26 acres. It is to be supposed that the wolfram-minerals are not uniformly distributed through the whole bed, but that rich and poor spots will be found. Though it is hazardous to estimate the value of an ore-deposit before it is actually worked, considering the high specific gravity of the wolfram minerals, which is nearly that of cast- iron, I think it safe to assume that this bed may contain 2 to 3 per cent., by weight, of wolfram ; taking the average weight of the ore at 2 tons per cubic yard, one yard would yield 90 to 130 pounds of wolfram — an estimate which will have to be verified by actual work.

Working of the Ore-Bed, — The practical exploitation seems by no means difficult, as an adit can be driven from the valley, close to the bridge across the Pequannok river. The adit can be extended all along the ore-bed, and will drain it for a vertical height of 160 to 170 feet. Inclined shafts from the adit to the surface will open the bed along its natural slope. The extracted ore can be hand-picked and suitably concentrated by wet dressing, for which the proximity of the river offers a favorable opportunity.

Conclusion, — In considering the paragenesis of the minerals on this property, one is forcibly impressed with the fact that they are all essentially tin-minerals; t.€., such as are usually associated with tin-ore, and it would, therefore, seem no remote probability to find tin also in this locality. It would certainly be desirable to keep a sharp lookout for it while working the deposit for wolfram.

Microscx)Pic Metallography, 243

Microscopic Metalloqbafet.

BT P. OSMOND, PARIS, FRANCE.* (Chicago Meeting, being part of the International Engineering Congress, August, 1893.)

When a metal (whether a simple substance, an alloy, or a com- pound) presents, in each of the smallest parts to which it can be re- duced by mechanical division, a constant chemical composition, it is defined as chemically homogeneous. But chemical homogeneity by no means necessitates mechanical homogeneity. There are no per- fectly amorphous metals. Crystalline forces, on the one band, and tensions or compressions, on the other band, determined by ine- qualities of temperature in the mass during heating or cooling, give rise in every metal to the formation of geometric elements of struc- ture. These elements may, moreover, assume perfect crystalline forms, with their planes of cleavage ; or embryonic crystalline forms, segregated in the midst of a paste of confused organization ; or pseudo-crystalline forms of expansion and contraction, like those produced in the drying of gelatinous precipitates or (most frequently) these different forms associated. In either case the adhesions between two adjoining structural elements, respectively homogeneous, may be very different from the interior cohesion of each of these elements. The mechanical properties of this complex aggregate cannot be sim- ple, and, when it is submitted to strains sufficient to produce defor- mation, the/esults may be very different according to the type of structure or the manner in which it is affected by the strains. It may be fairly said that a metal does not present a single resistance to tension, or a single resistance to compression, etc., but in fact, at least theoretically, as many resistances to tension, compression, etc., as it may possess coexistent features of structure; only, in any single test, it is the weakest of the resistances involved which makes itself evident.

For further explanation, let us imagine a cylinder A B (Fig. 1), formed of two similar halves A and B, placed together upon their common base x y. Such a system evidently presents no resistance to axial tension, while it would resist axial compression like a single

Translated by the Secretary. I

Microscopic Metallography.

block. This is an extreme case, yet in practice, though a metal is -not generally traversed by complete interruptions of continuity, it is always intersected by one or more networks of surfaces of unequal resistance, and in the different applications of the metal it is not always the same network which is concerned.

The above is true, with still stronger emphasis, of those metals which are not chemically homogeneous, and the mass of which is consequently formed of different constituents, each possessing its own peculiar properties and certain relations of adhesion with its neighbors. The mechanical properties of the aggregate are therefore a resultant which may be very complex.*

In a word, metallic bodies are analogous to rocks, and as the study of rocks has given birth to the special science of petrography, the exact knowledge of metals calls for the creation of a corresponding science of metallography. This science has already given rise to important separate researches, but it has not yet arrived at complete autonomy. It has not acquired, perhaps, either the full conscious- ness of its future or the full possession of its field. It is the natural development and scientific transformation of the uncertain art of the interpretation of fractures. Its first task should be to collate the ma- terials of observation and to determine the nature and proportions of the diverse constituents composing metals in different conditions. It should then inquire by systematic experiment how these constitu- ents are modified in dimensions, local or general distribution, chemi- cal relations and mutual aggregation, under the influence of the three independent variables : temperature, time, and pressure. Finally, it will have to determine the technological properties coi> responding to an ascertained structure.

By supposing one of the constituents to be gaseous, we may logically reduce the tudy of blow-holes, pores, and fissures to a chapter of metallography.

Microscopic Metallography. 245

The results thus obtained will teach consumers what structure tbej ought to require of a metal in order that it may satisfy known demands, and producers will be taught by what processes they can secure such a structure.

This is a vast programme and far from having been actually re- alizedy yet the study of the structure of metals has already accumu- lated a number of documents far too great to be summarized wholly in this paper. We shall limit ourselves, therefore, to that part of metallography which is based upon the employment of the micro- scope, and may thus be considered as a distinct and well-defined specialty.

Historical.

Petrographers are accustomed to examine translucent specimens in thin sections, and their methods, as based upon the phenomena of polarization, are not applicable to opaque bodies. Notwithstanding some profound analogies, microscopic metallography has not been developed from petrography. It is more closely connected as a nat- ural extension with the study of meteoric irons, and, as has oflen happened in the history of the sciences, it appears to have had sev- eral independent origins. The publications of Dr. Sorby go back to 1864, and those of Prof. Martens to 1878. But in spite of this difference in date, the labors of the latter present all the characters of complete originality. While Dr. Sorby devoted himself to the development of a complete method of examining sections of opaque bodies under the highest magnifying powers, and the application of this method to different products of the metallurgy of iron, Prof. Martens studied at first, though without neglecting the examination of sections, the general laws of fractures, fissures, blow-holes and crystallization in metals. Prof. Wedding, of Berlin, has also made numerous investigations on the structure of iron and steel. Those of Messrs. F. Lynwood Garrison and Dudley are well known in America. In France, M. Barba introduced in 1880 the use of the microscope in the works of Creusot, and gave the first impulse to the labors of Messrs. Osmond and Werth, which have been pur- sued since that time by methods perfected under the influence of Dr. Sorby.

Microscopic metallography is thus cultivated to-day in the prin- cipal metallurgical countries. Starting from scientific laboratories, it is extending more and more into industrial laboratories, where it will probably become an indispensable auxiliary to chemical analysis and physical tests.

246 microscopic metallography.

Technical.

As we have seen above, the microscope may be employed for ex- amining either fractures or prepared sections. Prof. Martens justly observes that the two methods are complementary. Nevertheless, since fractures are, according to the expression of Dr. Sorby, sur- faces of weakness, and moreover, do not lend themselves to the use of any but very feeble magnifying powers, the employment of sec- tions alone is the general practice; and we shall con6ne ourselves to this, referring for the rest to the important memoirs of Prof. Martens (Nob. 5, 7 and 20.)*

A plate of metal being given, to be studied under the microscope, the first o})eration to which it must be submitted is that of polish- ing. The recent discussion in Germany between Prof. Martens and Prof. Wedding (Nos. 31, 37, 38 and 39) has fully shown the im- portance of this preliminary operation, demonstrating how an im- perfect preliminary grinding may result in erroneous interpreta- tions. It is, however, difficult to formulate absolute rules ; for the arrangement of the polishers (if mechanical means are employed), the choice of polishing substances and of the substances used as supports may be varied according to convenience within wide limits. Be- ginners would do well to consult and follow the directions of Dr. Sorby (No. 21) or those of Prof. Martens (No. 35). Dr. Sorby em- ploys at the beginning emery papers mounted upon plate-glass, using consecutively finer and finer grades, down to the finest fur- nished in commerce. He continues the work with "fine-grained water-of-Ayr stone;" then with the finest crocus; and ends with the best washed rouge, with which he sprinkles a wet cloth held upon a perfectly plane wooden backing. When the final emery paper is of good quality, the use of the rouge may follow immediately without any intervening step.

Prof. Martens substitutes for emery the stone called Oebtein, largely employed by Swiss watch-makei*s, the powder of which he washes with the greatest care, so as to separate it into different classes of fineness, each of which shall be perfectly uniform.

When the substance upon which the polishing with rouge is fin- ished, is somewhat soft, such as cloth, leather or parchment, and the substance polished is composed of constituents of different Iiardness,

The figures given in this place, and throughout the remainder of this paper, in parenthesis following the name of an author, refer to the publications, a list of which is appended to tlie paper.

Hicr0600Pic Metallography. 247

the surface of the latter becomes lightly engraved the harder con- stituents appearing in relief.

This preliminary indication is always of great value, and some- times suffices of itself (as for instance in the case of cement-steel), to reveal the whole structure of the metal. It is therefore indispensa- ble, when one is studying a specimen for the first time, to polish it under these conditions and then to examine it with the microscope. If, however, for the purpose of studying the most delicate details, a perfectly plane surface is required, the polishing should be finished upon an unyielding support.. Prof. Martens employs for this pur- pose mixtures of colophony with wax or pitch. Theoretically, the polishing cannot be too perfect; but it is not easy in practice to arrive at perfection. Happily, this is not necessary as a general rule. If the emery papers or other hard materials have made some scratches deeper than others, which cannot be removed without ex- cessive labor, these striations are recognized at the first glance; and if they are few, they do not seriously hinder observations. It is poible, therefore, to obtain in less than a quarter of an hour a very satisfactory preparation, even without the use of any other instru- ment than the hand. Of course time may be gained, when many plates are to be polished, by using a small machine. Mr. R. Fuess, of Steglitz near Berlin, constructs one which is convenient for this purpose, and which is used in the laboratory of Cliarlottenburg.

In most cases, the structure of the metal is not shown by polish- ing only, and must be made to appear by physical or chemical processes which produce different effects upon its different constitu- ents. These processes vary with the nature of the metals concerned. We shall have occasion to describe them further on, in connection with particular cases.

Whatever be the methods of preparation, the metals are opaque, and their illumination under the microscope presents special difficul- ties which are not encountered in the study of translucent sections. Natural illumination will serve only for very low powers, which are usually insufficient for this work. Hence the microscope must be provided with special accessories.

For oblique illumination we have the parabolic mirror of Sorby (Fig. 2) and the mirror of Lieberkuhn (Fig. 3), both of which may be mounted upon the objective. For perpendicular illumination re- course is had to Beck's vertical illuminator (Fig. 4) ; this is a small transparent mirror m, placed in the axis of the microscope, which receives the light by the slit a b and reflects it upon the objective.

Kicru600Pic Hetallographt.

The lenses of the objective concentrate the light apon the object and the latter is seen through the transparent small mirror. Prof. Martens

employs for similar conditions a small prism (Fig. 6), which was de- vised by M. Nachet, a constructor of Paris, for the labors of the international Commission on the meter. With these two latter de-

Microscopic Metallographt. 249

vices the lUumiDation is almost as good for high powers as for low ones; one may go to 1000 diameters and beyond.

For the lower powers, up to about 300 diameters, the two systems of illumination, the oblique and perpendicular, are equally practi- cable ; but it is not a matter of indifference whether one or the other be employed. In general, the perpendicular illumination shows the structure best, oblique illumination clearly exhibiting only the planes which are suitably situated with reference to the incident light, and often giving, when taken by itself, very misleading indications. It IS in this way that the small polyhedra of the soft steel (Fig. 1 a, Plate I.) seem in oblique light to unite (Fig. 1 b) and simulate a ribboned structure which has nothing in common with the facts. On the contrary, oblique illumination alone shows the watered (rooir) effects which may be brought out by rotating the object under the objective and which it is important to know (compare 5 b with 6 a, Plate II., and 2 b with 2 a, Plate III., etc.). Indeed the two systems pf illumination complement one another, and when necessary, correct one another, within the limits of the magnifying power which permits the concurrent use of both.

It remains to fix by photography the indications of the micro- scope. The microphotography of minute objects has made great progress in the last few years. The first specimens of Dr. Sorby (No. 21) were magnified 9 diameters; those of Messrs. Osmond and Werth (No. 14), 18 diameters, vf O smond (No. 29), in 1890, reached 300 diameters; and Prof. Martens (No. 32) can now go up to 800. The equipment of the experimental laboratory at Chariot- teuburg, in which this remarkable result was made possible, is cer- tainly the most beautiful and complete in existence. It has been already described (No. 32) in detail. The only objection that can be urged against it is its expensiveness. This criticism has no force from a scientific standpoint; but, since it might deter industrial es- tablishments from the employment of the method concerned, we deem it worth while to observe that much can be accomplished with less costly instruments. All the photographs from which the illus- trations of this paper have been prepared were made with a small Nachet camera, costing 90 francs; and the source of illumination was a simple colza-oil moderateur lamp. With this crude apparatus, powers of 300 diameters may be employed ; and the time of exposure (which can be reduced one-half by the use of a good petroleum lamp) did not exceed an average of 20 minutes, being 40 minutes in the most unfavorable cases. If one be working alone, and have

250 Microscopic Metallography.

not to make a large iiuml)er of photographs, one can develop a slate while its successor is under exposure; and in such a case there will be really no loss of time.

Results — Steels.

A, Preparation of the Sections. — Apart from the preliminary pol- ishing, the processes employed to reveal the structure of a section of steel are :

1. Coloration by heat, which applied long ago to meteoric irons, and has been used in this connection by Prof. Martens (No. 6) and Prof. Wedding (No. 13). As is well known, the polished sur- face of the steel assumes, under the action of heat in an oxidizing atmosphere, a series of colors, dependent upon temperature, time and the composition of the metal. All other things being equal, the different constituents cannot assume the same at the same in- stant, and hence they present in characteristic colors a picture of the structure.

2. Attach by Chemical Reagents. — Of these, the most frequently used is nitric acid, the action of which is, in general, the more rapid, the harder and more impure the metal. Dr. Sorby (No. 21) uses it highly diluted, which renders its effects feeble. A satisfac- tory mixture is that of one volume of 36° Beaum6 acid with four of water. In this the metal is left for from a few seconds to one minute, according to its composition. If one is not sure on this point beforehand, the operation may be conducted by successive immersions of ten seconds or even of five seconds each; the plate being examined with the microscope after each immersion. In this way, it is soon possible to arrest at the desired point the action of the attacking acid. Prof. Martens recommends chlorhydric acid diluted with alcohol or ether. Other reagents may likewise be em- ployed, but, so far as appears, without advantage, save in some special cases.

The attacked plates are immediately washed in abundant water; then in alcohol ; wiped upon blotting-paper and dried. A jet of compressed air, if at command, furnishes the best and most rapid means of drying without any oxidation. It is sometimes necessary to dry the damp plate with a piece of soft linen, for example, when the structure is veiled by a black film of carbon due to hardening.

J5. The Cdnstiiuents of Steel. — In the carbon-steels (with which alone we are now concerned) three constituents are unanimously recognized. These are described by Dr. Sorby (No. 21) as (1) free

H1Ck0600Pic Metallogbapht. 251

iron ; (2) the nacreous or pearly constituent; (3) iron combined with carbon. On the other hand, Prof. Wedding (No. 13) distinguishes: (1) crystalline iron ; (2) homogeneous iron, softer, and (3) homo- geneous iron harder, than the crystalline. Finally, Mr. H. M, Howe, who in his admirable Metallurgy of Steely has devoted to metallography some excellent pages, proposes to give to these three constituents the names Ferrite, Pearlyte and Cementite. The use of this triple nomenclature may easily have created certain confu- sion in the minds of readers, leading them to suppose that the con- stituents recognized by Prof. Wedding were not the same as those of Dr. Sorby. But this is not the case. An attentive study of the original memoirs, aided by practical experience of the matter in question justifies the assertion of the identity of the things under the diversity of the names. By writing in line the three names pro- posed for each constituent, we obtain the following table :

Sorby. Wedding. Howe.

Free iron Homogeneous iron, softer than the crystalline Ferrite.

Pearly constituent Crystal 1 i ne i ron Pearly te.

Iron combined / Homogeneous iron harder than the crys- 1 Cementite with carbon. / t talline.

The names of Prof. Wedding have the disadvantage of giving for each constituent a definition which, as we shall see later, is open to controversy,. On the other hand, the names proposed by Mr. Howe seem to us felicitous ; they are brief; they do not prejudge (as it would be premature to do) the nature of the several constituents ; and, by their mineralogical terminations, they establish a just rela- tion between metallography and petrography. For these reasons, they will be adopted below.

Ferrite is completely or comparatively pure malleable iron. Ac- cording to Dr. Sorby, it crystallizes at a high temperature, and recrystallizes at a lower temperature. In mixture, its forms may be modified by neighboring constituents.

Pearlyte, so-called because it assumes a nacreous aspect under certain conditions of illumination, resolves itself under high magnify- ing-powers into alternate hard and soft laminae. According to Dr. Sorby (No. 16) who discovered and defined it, the soft layers are about twice as thick as the hard ones, and while variable in that respect, average about 0.6

/I 0.001 millimeter.

MICROSCOPIC MErALLOGRAPHY.

iron.

The soft laminae seem to be iron, and the hard ones carburet o& The latter has been isolated, though in a more or less altered condition, by Dr. Muller, Sir Frederick Abel, and Messrs. Osmond and Werth. Its probable chemical formula is Fefi, Fig. 6, taken from Dr. Sorby, shows this curious structure of pearlyte. The two photographs given by Prof. Martens (Figs. 10 and 11 of No. 32), presenting enlargement to 500 and 800 diameters respectively, may be compared with it. Nevertheless, this structure-type does not seem to us to be universal. Our personal researches have led us to

Rfl. 6.

believe there is room to distinguish between granular and lamellar pearlyte. The former consisting of small grains of iron, irregularly disposed, and surrounded by iron carburet, constitutes the dominant form in steels forged up to dark red heat, and passes gradually into the lamellar pearlyte in proportion as the structure has been con- stituted at a higher temperature. Lamellar pearlyte ap|>ears to be characterized, under oblique illumination and moderate enlargement, by a watered (moire) appearance, the absence of which indicates the granular pearlyte. M. Werth and the writer (No. 14) have applied to the latter the term simple cellular to designate its structure. But all intermediate grades may be observed also.

Micr0600Pic Metallography. 253

Cementite is a hard, carburetted substancey occurring in cement- steels, and probably also in the hardest cast-steels. Dr. Sorby (No. 21) classes it, at least as to species and without attributing to it a fixed composition, with the characteristic constituent of white pig- iron. Its forms in blister-steel, and also its hardness, permit its easy recognition, although its color is nearly that of iron (Plate IV.) ; but its chemical nature is imperfectly known. We can say only that it is rich in carbon, but not whether it constitutes a definite carbide, identical or not identical with that of the thin hard laminae of pearlyte, or whether it contains carbon dissolved in an allotropic modification of iron.

C Relations Between the Structure and the Composition of Steds. — Just as quartz, feldspar and mica, for instance, form by their manifold associations a large number of different rocks, so ferrite, pearlyte and cementite give rise to all the varieties of steel, in such a manner that steels of different hardness, or the different conditions of the same steel, vary from one another only in the proportion and distribution of these elements.

Let us examine first a series of steels of different hardness forged in circular bars of 12 millimeters, reheated to a uniform temperature of 750 C. and cooled slowly under the same conditions.

The extra sofl steel (0.14 per cent, of carbon) is an agglomeration of small polyhedra of ferrite having an average diameter of 30 ft and the more or less imperfect form of pentagonal dodecahe<lra (Fig. 1 a, Plate I.).* This form is not strictly crystalline; but it is that which contiguous polyhedra assume by the reason of their mutual limita- tions in simultaneous development. Each polyhedron appears to be a crystalline individual the interior of which is more or less clearly lamellary ; the existence of the lamellae being indicated (1) by the variable lustre of the different polyhedra according to the incidence of light upon them ; and (2) by the visible traces of laminary edges which appear upon the section under high powers (Fig. 2 c, Plate I.). Ferrite thus appears to have, theoretically at least, the same struc- ture as pearlyte, and differs from the latter only by the absence of interposed carburet. But in fact the lamellae in question are more or less plicated and interwelded. As for the iron carburet, naturally

To reprodnce the author's photographic plates for the pages of the Tia-Moetions, the dimensions of Plates L and II. have been reduced one-third, and of Plates III. and IV. one-fourth. The magnifying powers used by the author have been cor- respondingly reduced in the text and on the plates to the number of diameters actually shown.— R. W. R.

254 Microscopic Metallography.

rare in steel of this dass, it is segregated on the exterior of the poly- hedra among which it constitutes a discontinuous film of variable thickness. At points where this film becomes thicker it sometimes passes into pearlyte. Traces of carburet may also remain as inclu- sions between the lamellae of the ferrite.

The steel of medium hardness (0.45 per cent, of carbon) is a mix- ture of almost granular pearlyte with ferrite in dislocated grains which are entangled with each other like vermiculated work in architecture (Figs. 2 a and 2 c, Plate II.). Upon photographs taken in perpendicular illumination the ferrite is light and the pearlyte dark ; even at 300 diameters the latter is not resolved into its ele- ment. Each irregular grain of ferrite is probably formed by the juxtaposition of several relatively regular.

Hard steel (1.24 percent of carbon) of which we give photo- graphs (Figs. 1 a, 1 b, 1 .c, Plate III.) does not appear to us to con- tain cementite. The proportion of ferrite in it is necessarily very much reduced and forms nothing more than a fine stippling, assum- ing sometimes, under an enlargement of 30 to 45 diameters the appearance of a net-work (Fig. 1 a). The not very clear watered appearance which is perceived under oblique illuminations (Fig. 1 b, Plate III.) indicates a general structure in the condition of laminary pearlyte not fully developed. This structure appears fairly well under enlargement of 200 diameters (Fig. 1 c, Plate III.); the alternate light and dark soft and hard lamellse of the pearlyte are plicated and irregular; one cannot see here, as in the diagram of Dr. Sorby, separate islands limited by the constant orientation of the lamellae.

With the forged steels which we have just described, it is conve- nient to compare blister-steel (Plate IV.). This is a mixture of pearlyte and cementite. The cementite either forms straight lines intersecting variable angles or long-drawn curves describing a laie, roughly polygonal network. Figs. 2 a and 2 b, Plate IV., show, under an enlargement of 60 diameters, the details of these two vari- eties of cementite. Both resolve themselves into juxtaposed lamellae of the average thickness of 6 The pearlyte is granular in places, but more frequently laraellary.

D, Relations between Structure and Different Conditions of the same Steel, — We have seen how the structure of steels taken from the same condition varies with their contents of carbon. The dif- ferences are not less marked for the same steel in different states.

' the word state we mean the total mechanical properties of the

Micr0600Pic Metallography. 256

metal, as they have been determined during the manufacture under the influence of the three independent variables — temperature, time and pressure.

a. Infiuence of Temperature. — The three steels which we have compared with each other and described in one of their states have been heated at variable temperatures and slowly cooled under uni- form conditions from the maximum temperature.

Extra Soft Steel — In proportion as the temperature of annealing is elevated, the polyhedra of ferrite increase in size, retaining at first their regular form. The average diameter of the polygons upon the section is about 32 fi after heating to 750° C, 39 fi after heating to 920°, and 48 fi after heating to 1013°. Above 1000° the poly- hedra, while continuing to increase in size, become more and more irregular and tend to elongate themselves. At about 1300° C. the structure is completely modified ; the ferrite has arranged itself into groups of large lamellse, having a parallel direction in the same neighborhood. Figs. 1 and 4, Plate 1, show the principal stages of these transformations.

Steels of Medium Hardness. — (Figs. 1-4, Plate II.) If the tem- perature of annealing (750° C.) has not passed the point a of Cher- noff, the finest and best structure is obtained, as we have described it above (Figs. 2 a and 2 c, Plate II.). The temperature being raised, the pearlyte becomes more and more lamellary, and its irreg- ular aggregations tend to form more and more large and rular polyhedra. Figs. 1 a and 1 c, Plate 2, give the structure as it will be after heating at about 850° C. The polyhedra of pearlyte are not yet much more definite than in Figs. 2 a and 2 c ; they become more so iu Figs. 3 a and 3 c after heating to 1115° (D., and they are perfectly formed after heating to 1330° C. (Figs, 4 a and 4 b.) At the same time the ferrite, original groupings of which were very similar to those of the pearlyte, tends to pass to the surface of the polyhedra, while thrusting into the interior of the latter long paral- lel ramifications (Figs. 3 c and 4 a). In the burnt steel (Fig. 4 a) the ferrite forms a continuous network enveloping the large grains of pearlyte, the total presenting a good type of what M. Werth and ourselves have called the compound cells of the second order. The large grains are themselves subdivided into smaller elements (com- pound cells of the first order), each of which is characterized by the orientations of the lamellse and by its varying lustre when photo- graphed under oblique light (Fig. 4 b). These subdivisions are not definitely limited, and, under high powers, may be seen to penetrate

256 Microscopic Metallography.

Hard Steel — (Plate III.) The action of heat upon hard steel exhibits itself in a watered appearance growing more and more large and bold, of which Figs. 1 b and 2 b, Plate III., show the extreme variations. The structure of the burnt steel after heating to 1330° C. is seen in Fig. 2 a; the polyhedra exhibit, under per- pendicular illumination, a fine white piping of ferrite.

6. Influence of Time, — Methodical experiments have not been made upon the influence of the duration of annealing. It may be admitted, provisionally, that a prolongation of this period is equiva- lent to a certain elevation of temperature during a shorter period ; but this is a question which needs to be further elucidated and con- cerning which current notions are perhaps not well justified.

The influence of the rate of cooling is somewhat better known. In a general way the hastening of refrigeration hinders the develop- ment of the structure which would be produced during a slower cooling. The result is a finer grain, though we cannot say always a fine grain, because it is necessary to take into account the tempera- ture which had been reached in heating. On the other hand, a quick hardening impedes considerably the formation of the iron car- buret, which constitutes the hard lamellae of the pearlyte.

Figs. 5 a and 5 b (Plate II.) show the structure of steel of medium hardness quenched in water. In perpendicular illumination and enlargement of 30 diameters this structure appears almost amor- phous, but under oblique light it is easy to perceive a very fine watered (moire) appearance, which indicates that the very small grains of the hardened steel still possess a lamellary structure, and this indication is confirmed by the employment of the highest mag- nifying powers. If the quenching was done at white heat the lamellfe are much longer and the slide recalls the picture of pearlyte given by Dr. Sorby. But it is diiScult to say whether the lamellse in question are composed of two constituents of equal hardness, as is the case in the pearlyte of annealed steel, or of a single constitu- ent traversed by the planes of cleavage. Tlie latter opinion appears the more probable, but the aspect of the slide might be the same in either case.

The microscopic study of the structure of hardened steel presents, in fact, great difficulties and requires great caution. Dr. Sorby did not succeed in making out this structure with certainty, and the works of Mr. Behrens (No. 30), which describe in this instance a soft continuous network surrounding small hard grains, seem to us to require confirmation.

Microscopic Metallography.

c. Influence of Pressure. — At elevated temperatures the pressure appears to have no other effect than that of hindering the develop- ment of structure. Dr. Sorby has shown that the grains of iron, considered as forming each a crystalline individual, are not drawn out in rolling, and consequently have been constituted after the metal has passed the rolls. Provided that the rolls or the ham- mer knead it in a suflBciently energetic way, it probably possesses at the end of the operation an amorphous structure, which changes afterwards in accordance with its actual temperature and the rate of its cooling. It thus follows that these two factors exert upon the quality of the finished product a controlling influence which is too frequently ignored even at the present <lay.

If the rolling or forging ends at a relatively low temperature, it impresses special characters upon the structure. Photographs Nos. 53 and 54, given by Prof. Martens in his memoir on the structure of rail-heads (No. 37), show the ferrite to be arranged in part in the direction of the rolling. Our Fig. 7 represents a schematic structure of a small round bar of extra soft steel 12 milli-

ng. 7.

meters in diameter, cut perpendicularly to its axis and treated with dilute sulphuric acid. It exhibits a sort of trefoil, formed of parallel bands, alternately brilliant and dull, the latter being depressed as compared with the former. Fig. 5 a, Plate I., gives, under an en- largement of 30 diameters, a point of the striated region. The sample here photographed has been quenched, but quenching has served only to bring out more clearly a pre-existing structure. Since the metal is but slightly carburized (0.14 per cent, carbon), we can

Ic

258 MICROSCOPIC MErALLOGRAPHY.

no longer attribute the dark bands to the presence of hardening carbon, and this hypothesis would be negatived also by examination of plates not quenched. It is probable, therefore, that the light bands are the more compact and the dark the more porous, and hence the more deeply attacked by the acid, the result being that they appear obscure under perpendicular illumination. Ou the other hand, since the general features of the figure manifestly reveal its origin, we are led to conceive that rolling and forging act like vibratory forces and determine undulations which produce ventral and nodal surfaces in the metal which they deform. This is an- other important subject of study.

If pressure is exerted at low temperature so as to produce a flow of metal and a permanent deformation, the structure will preserve traces of this process which can be easily discovered with the micro- scope. The results actually depend on the relative hardness of the constituents. The softer portions flow first, escaping in a direction perpendicular to that of the pressure, while the harder portions tend to break. Such is the origin of fiber and of lamination.

E. Accidental Defects. — We designate thus the blow-holes, pores, fissures, hard cores, and inclusions of slag, which may be encountered in steels. As a general rule, defects of this class pertain rather to maerostructure than to microstructure, and the study of them does not require the delicate means of investigation above described. A summary polishing, followed by a superficial treatment with dilute sulphuric acid, usually renders them visible to the naked eye, if they are not already so upon the surface of fracture. It is, moreover, a good practice to employ this method in addition to the microscopic one. The two modes of testing complement each other, and do not involve double pains. But it is not our intention to develop this standpoint, which is aside from the purpose of the present paper.

To return to microscopic defects, the most commonly encountered of these is the presence of minute cavities or pores, apparently due to the imperfect interlocking of unsound particles of the ingot. The study of this defect has been well performed by Professor Martens (No. 37). Such pores must not be confounded with the accidents of polishing, which disappear under perfect manipulation. When they are grou|>ed in considerable numbers in one region, this porous re- gion appears dull and dark after treatment with nitric acid, first, because it is more profoundly corroded than the more compact re- gions ; and, secondly, because the drying of these cavities, being slow and imperfect, occasions there a commencement of oxidation.

Microscopic Mktallography. 259

Another question, pertaining in part to metallography and in part to analytical chemistry, is that of the unequal distribution of foreign bodies in a mass of steel. It is known that carbon, sulphur, and phosphorus are concentrated in that portion of the ingot which solidifies last, and around the cavity called "the pipe' (retasHurt). These irregularities are indicated under the microscope by the rela- tion of the surfaces occupied by ferrite and pearlyte respectively.

F. Qmclnauma. — We shall not attempt to give the theoretical interpretations of the facts which we have outlined. The chemical or molecular changes to which steel is subject at certain critical tem- peratures, render the problem exceedingly complex, and it is difficult to offer, even hypothetically, a satisfactory solution.

The work of description is itself incomplete. Certain questions, such as the effect of the duration of heating, have not yet been at- tacked ; others, such as the mode of action of forging and rolling, open a vast field of research ; the influence of other substances than carbon is known in a few special cases only.*

Notwithstanding all these gaps, the metallography of steel can render from this time forward services of practical value. Two facts seem to us to have been ascertained :

1. When difierent steels are compared in the same state, that is, after having been treated in all respects in the same manner, their microstructure is, to a certain extent, characteristic of their hard- ness.

2. When samples of the same steel are compared in different states, the microstructure is an indication of the manipulations which the steel has undergone, and, notably, of the temperature from which, and the rate at which, it has cooled.

The first of these facts is not very interesting. It is clear that the aspect of a section is far from defining with the same precision as chemical analysis the potential properties of a steel.

The second fact, on the contrary, leads to an important conse- quence. Here micrography complements analysis. While the chemi- cal composition declares what a steel could or should have been, the microscope shows us what it really is. Hence, it may be concluded :

(1) That the metal examined has or has not been manufactured by the best methoils — a means for the improvement of such methods;

(2) that the metal will or will not successfully endure the strains to which it is intended to be subsequently exposed.

For mangaDese-steelfl, the notes of Mr. Tetflkichi Mukmi (No. 41) mmy be ooDSiilted.

260 Microscopic Metallography.

Unquestionably, in the present stage of our knowledge, such de- ductions can be nothing more than personal estimates, lacking pre- cision and certitude. Micrography cannot replace the old methods of testing, until it shall have established numerical relations between the mechanical properties of steels and the varied aspects of their structure. In the meantime, it can only qualify, somewhat vaguely, the metal examined, as good, passable, mediocre, or bad. But it has an immediate value, nevertheless, in that it throws more light than is derived from tensile tests, upon the causes of the actual condition of a steel. It enables us to '' reconstruct '' the thermal treatment to which the steel has been subject, and to modify this treatment if practicable. In this way, Mr. Dudley (No. 33) has been able to draw from his microscopic observations conclusions both just and useful concerning the manufacture of rails.

It may be fairly said, that a steel derives all its mechanical prop- erties from its chemical composition, its molecular condition, and its structure. This statement is not exactly novel ; it was put forth about 1878 by the engineers of the Terre Noire works, who, in their turn, had predecessors holding the same view. Yet the proposition has remained in dispute, and has received but scanty recognition. Thus it has been sought to determine the relations between the varia- bles; for instance, between the chemical composition and the wear of rails; or, even between the amount of rolling and the resistance to shock ; relations which unquestionably exist, but which are most frequently masked in industrial products by irregularities of struc- ture. Now that the essential influence of structure has been made plain for all time to come, we may hope that this new independent variable will be introduced into such investigations, or eliminated beforehand, so that, to the greatest benefit of both producers and consumers, the relations may be ascertained which connect the quali- ties of steel with their chemical, physical, and mechanical oo-effi- cients as determined by the tests of manufacturers and inspectors.

Wrodght-Irons.

Wrought-irons have been less studied than steel, and their micro- scopic examination, though more simple, is also less practically im- portant. They are less sensitive to the action of heat, and hence more easily manipulated. In like manner they can be more easily classed according to the appearance of fractures. We have nothing to add to the excellent description of their structure given by Dr. Sorby.

mick06c?0pic metallography. 261

Cast-Irons.

Besides the three constituents of steel, cast-irons contain graphite, an essential element of gray cast-iron, and also, according to Dr. Sorby, small ruby and dark crystals, and a residual compound. This makes in all six elements of structure, and although it is not neces- sary to consider the complications which, in the case of steel, result from the multiplied conditions of manipulation, the study of cast- irons appears none the less to be a very arduous undertaking. Prof. Martens and Dr. Sorby (Nos. 5, 7, and 21) have given highly inter- esting descriptions of a certain number of specimens. But the en- deavor to connect the indications of the microscope with the mechan- ical properties of the material has been merely tentative. We may cite, however, the paper of Mr. F. Lynwood Garrison (No. 15), read in February, 1886, before the American Institute of Mining Engi- neers (Trans., xiv., 913). Mr. Gkirrison thinks that good car-wheel irons are characterized by the uniform distribution of graphite in strong, well-defined, regular plates, while this substance, in pig-iron of poor quality, is irregularly distributed in isolated and relatively large patches.

The method of coloration by heating is to be particularly recom- mended in the examination of certain cast-irons. Prof. Martens has thus obtained, notably for spiegeleisen, colored images of remark- able distinctness and beauty.

Alloys of Copper.

The following observations are taken from a recent note by Mr. Guillemin (No. 40) :

The plate is attacked either with dilute cold nitric acid or with sulphuric acid (diluted to one-tenth, and under the influence of a weak electric current of 2 volts and 0.1 ampere).

Micrographic examination permits the direct classification of the usual alloys in a few categories. Thus the bronzes and brasses are distinguished as :

Tin-bronzes.

Phosphor-bronzes.

Brasses containing less than 37 per cent, of zinc.

Muntz metal and analogous alloys, containing more than 37 per cent of zinc.

Aluminum-bronze.

Alaminum-brasses.

262 Micboscopic Metallography.

Delta metal.

Roma bronze, etc.

In the white "anti-friction " alloys, containing tin, antimony, and copper, it is easy to recognize the presence and even, with a little practice, to make a close estimate of the proportion of lead.

By examining the ingots of rose-copper composing different casts from the same smelting of ore, those which have been perfectly re- fined can be distinguished, and the rest can be classified according to the degree in which they have been refined.

It is known that the mechanical qualities of brasses and bronzes are profoundly modified by the addition of small quantities of alumi- num and of phosphorus. The presence of these two elements can be detected with certainty by micrographic examination. Thus the tracings assume invariably the form of veins in marble when the alloy contains aluminum, even in proportions so minute as to be de- tected with difficulty by chemical analysis. In like manner phos- phorus produces in tin-bronzes an absolutely characteristic image, recalling a fern-leaf. This reaction, however, may be masked by the presence of zinc, in amount exceeding 4 per cent.

Finally, for a known alloy, the microgram indicates also the con- ditions of casting, as well as the nature of the mechanical work sub- sequently performed upon the specimen. Thus the image shows whether bronze has been cast too hot or too cold ; whether it has been stamped or rolled; and, in the latter case, the direction of the rolling is distinctly marked.

In a word, micrograph ical analysis enables us to determine Quickly and summarily the nature of a bronze or an industrial alloy by the simple inspection of a polished and etched surface, and to recognize whether the alloy has been simply cast or has been rolled or stamped.

Bibliography.

1. H. C. Sorby.— "On a New Method of Illustrating the Struct- ure of Various Kinds of Steel by Nature Printing" (Sheffield Lit- erary and Philosophical Society, February, 1864).

2. H. C. Sorby. — "On the Microscopical Structure of Meteorites and Meteoric Iron " (Proceedings Royal Society, vol. xiii., p. 333, and British Ass. Report, 1865, part i., p. 139).

3. H. C. Sorby. — "On Microscopical Photographs of Various Kinds of Iron and Steel" (British Ass. Report, 1864, part ii., p. 189).

4. H. C. Sorby. — " On the Microscopical Structure of Iron and

Micr06Copic Metaixography. 263

Steel " (Dr. Lionel Beale's " How to Work with the Microscope/' 4thecl., 1868, pp. 181-183).

5. A. Martens. — "Ueberdie Mikroskopische Untersuchung des Eisens" (Zeits. des Ver. Deuts. Ing., vol. xxii., pp. 11, 205, and 481, Jan., May and Nov., 1878; and vol. xxiv., p. 397, Aug., 1880).

6. H. C. Sorby.— "Lecture delivered in Sheffield's Firth College," read Oct. 20, 1882 ("The Engineer," vol. liv., p. 308, Oct. 27, 1882).

7. A. Martens. — "Ueber die Mikroskopische Untersuchung des Eisens " (Verhandl. des Ver. zur Beforderung des Gewerbfleisses, Sitzungslierichte, 1882, p. 233).

8. J. C. Bayles. — " Microscopic Analysis of the Structures of Iron and Steel" (Trans. Am. Inst. Min. Eng., vol. xi., p. 261, 1883).

9. O. DoUiak.— "Beitrage zur Mikroskopie der Metalle" (Mit- theil. iiber Gegenstande des Artillerie- und Geniewesens, 1883, Heft 9, p. 467).

10. A. Martens. — " Erlauterungen einer in der kon. Bergakade- mie zu Berlin befindlichen Sammlung von 120 Schliffen zur Dar- stellung des Mikroskopischen Gefuges verschiedener Eisen und Stahlsorten," Berlin, 1884.

11. F. Osmond et J. Werth.—" Structure Cellulaire de TAcier Fondu " (Comptes rendus de I'Acad. des Sciences, vol. c, p. 450, Febr. 16, 1885).

12. F. Lynwood Garrison. — "The Microscopic Structure of Iron and Steel" (Trans. Am. Inst. Min. Eng., vol. xiv., p. 64, 1885).

13. H. Wedding. — " The Properties of Malleable Iron, deduced from its Microscopic Structure " (Journal of the Iron and Steel Inst., 1885, p. 187).

14. F. Osmond et J. Werth. — "Th6orie Cellulaire des Propri6t& de I'Acier" (Ann. des Mines, 8th series, vol. viii., p. 5, July to Aug., 1885).

15. F. Lynwood (Jarrison. — "The Microscopic Structure of Car-Wheel Iron" (Trans. Am. Inst. Min. Eng., vol. xiv., p. 913, 1886).

16. H. C. Sorby. — " On the Application of very High Powers to the Study of the Microscopical Structure of Steel " (Journal of the Iron and Steel Inst., 1886, p. 140).

17. H, Wedding. — "Die Mikrostructur Verbrannten, Eisens" (Stahl und Eisen, vol. vi., p. 633, Oct., 1886).

264 Microscopic Metallography.

18. H. Wedding.— "Die Mikrostructur des Eisens' (Stahl und Eisen, vol. vii., p. 82, Febr., 1887).

19. F. Lynwood Garrison. — " Microscopic Structure of Steel Rails" (Trans. Am. Inst. Miu. Eng., vol. xv., p. 761, Febr., 1887).

20. A. Martens. — " Ueber das Kleingefuge des schmiedbaren Eisens, besonders des Stahls" (Stahl und Eisen, vol. vii., p. 235, Apr., 1887).

21. H. G. Sorby. — Microscopical Structure of Iron and Steel" (Journal of the Iron and Steel Inst., 1887, p. 255).

22. H. Schild. — " Die Neuesten Forschungen auf dem Gebiete der Mikroskopischen Untersuchung von Stahl und Eisen " (Stahl und Eisen, vol. viii., p. 90, Febr., 1888).

23. H. Wedding. — " Znsammenhang zwischen der Chemischeu Zusamroensetzung und dem Kleingefuge einerseits und der Leit- ungsgute des Telegraphendrahtes anderseitn" (Mittheil. aus den kdn. technischen Versuchsanstalten, 1888, Erganzungsheft i., p. 6).

24. H. Wedding. — " Ueber Fortschritte in der Lichtabbildung des Kleingefuges von Eisen und uber die Herstellung von Sehliffen" (Stahl und Eisen, vol. ix., p. 263, Apr., 1889).

25. A. Martens. — "Ueber die Mikroskopische Untersuchung des Kleingefuges von Eisen" (Stahl und Eisen, vol. ix., p. 393, May, 1889).

26. F. Osmond. — " Le Fer et TAcier." (Luraire filectrique, vol. XXXV., p. 2d5, Febr. 8, 1890).

27. H. Wedding, — " Das Kleingefuge des Eisens. Mikros- kopische Originalphotographien mit Erlauterungen" (Berlin, 1891).

28. Sir Fr. Abel. — " Presidential Address " (Journ. of the Iron and Steel Inst., 181, part i., p. 18).

29. F. Osmond. — " Note on the Microstructure of Steel " (Jour- nal of the Iron and Steel Inst, 1891, part i., p. 100).

30. H. Behrens. — "Sur la Structure Microscopique et sur la Trempe de I'Acier et de la Fonte" (Recueil des travaux chim. des Pays Has, vol. x., p. 261, 1891).

31. H. Wedding. — " Das Gefuge der Schienenkopfe " (Stahl und Eisen, vol. xi., p. 879, Nov., 1891).

32. A. Martens. — Die Mikrophotographische Ausrustung der kon. Mechanisch-Technischen Versuchsanstalten " (Mittheil. aus den kon. Technischen Versuchsanstalten, 1891, Heft 6, p. 278).

33. P. H, Dudley.—" Microscopic Structure of Steel " (Journal of the New York Microscopical Society, Oct., 1891).

Plate III.

Hard Steel.

Plate IV.

, "ti nini l!/i alww \Ue rletntk of I lie rem entile.

Blister Steel.

The Be6Seher Process As Oonducted In Sweden. 265

34. A. Martens. — " Ueber Einige in derMechanisch-Technischen Versuchsanstalt ausgefiihrte Mikroskopische Eisenuntersuchungen" (Mittheil. aus den kon. Technischen Versuch&anstalten, vol. x., p. 57, 1892, Heft 2).

35. A. Martens. — '*Die Mikroskopische Untersuchung der Met- alle'' (Glaser's Aunalen, vol. xxx., p. 201, 1892).

36. H. Behrens. — " Revue G6n6rale des Sciences Pures et Appli- que" (vol. lii.. p. 343, May 15, 1892).

37. A. Martens. — "Das Gefuge der Schienenkopfe" (Stahl und Eisen, vol. xii., p. 406, May 1, 1892).

38. H. Wedding.—" Das Gefuge der Schienenkopfe " (Stahl und Eisen, vol. xii., p. 478, May 15, 1892).

39. A. Martens. — " Das Gefuge der Schienenkopfe " (Stahl und Eisen, vol. xii., p. 530, June 1, 1892).

40. G. Guillemin. — "Analyse Micrographique des Alliages" (Comptes rendus de I'Academie des Sciences, vol. cxv., p. 232, July 25. 1892).

41. Tetskichi Mukai. — "Studien iiber Chemisch-Analytische und Mikroskopische Untersuchung des Manganstahls" (Freiberg, 1892).

Tee Be88Emeb Process As Conducted In Sweden

BT PROF. RICHARD AKERMAN, STOCKHOLM, SWEDEN.* (Chicago Meeting, being part of the International Engineering Congreta, August. 1893.)

At the International Sessions of the Iron and Steel Institute of Great Britain, the American Institute of Mining Engineers and the Verein Deutscher Eisenhuttenleute, held in Allegheny City, Pa., in October, 1890, Sir James Kitson, then President of the Iron and Steel Institute, presented an account by Sir Henry Bessemer of the origin of the process which bears his name.f The paper graphi- cally told how, after the first successes immediately following the first reports given at Cheltenham of the new process, there rapidly succeeded such obstacles and difficulties in its practical application, that the whole matter was nearly abandoned in 1858. Suddenly, however, the situation changed, so that at length it became possible

Translated by Philip W. Moen and Emanuel Trotz, Worcester, Mass. t Trans, xix., 810.

266 The B£68Em£R Process As Ookductted Ik Sweden.

to produce by the Bessemer process, with comparative uniformity, a completely satisfactory product. But Sir Henry Bessemer did not touch with a single word upon the question, what caused the change. He merely said :

" Happily for me the end [of his mental anxiety] was nigh, and in a few more months I had fully succeeded in producinfj steel worth from £50 to £60 per ton from charcoal pig-iron which had cost me only £7 per ton, the conversion of the crude iron into steel being effected by simply forcing minute streams of cold atmos- pheric air upward through it for a space of fifteen minutes."

It is no more than just to mention here, that the fortunate turn in affairs was brought about, substantially at least, by a Swede, Consul G. F. Goransson, who as early as 1857 had begun his ex- periments with the Bessemer process at Edske blast-furnace. Until the middle of 1858, he, like Sir Henry himself, had succeeded only exceptionally in turning out a good product while following the advice of the inventor to lay the greatest weight on having a high pressure of blast. By departing from that advice, and securing instead, by means of a large tuyere-area, an abundant supply of blast, Goransson was able, beginning with the 18th of July, 1858, so to shorten the time necessary for the process and thereby increase the heat of the blow, that an improved product was obtained ; and from that day forth the success of the Bessemer process was first assured. Before this, at least at Edske, the blows had generally run too cold, so that the product originally became so thick that two classes of serious trouble occurred. In the first place, the slag could not properly separate from the metal, but remained partly in nodules in the ingots. Because the steel was insufficiently fluid, the ingots were heterogeneous, being harder in some parts than in others. Moreover, they were unduly filled with blow-holes, which were too often oxidized, so that they could not be effaced by weld- ing. In the second place, great waste was occasioned by the exces- sive quantity of steel, even as much as one*half or more, which solidified in the ladle and had to he treated as scrap.

Concerning the result, obtained solely by the placing of larger and more numerous tuyeres in the converter-bottom, Mr. Gh)ran8- son himself says

the 18th of July, 1858, occurred the first blow in the remodeled farnace, and the results were completely changed. The blow was warm and regular, the

Nagra minnesblad om BeasemermetodeTis genonoramU vid Edtke vMMugn, 1858, och deas utverkring vid Sandvikens bruk instUlf 1883, p. 170.

The Bessemer Process As Conducted Is Sweden. 267

steel flowed so readily and was so liquid, that all the slag rose through it to the top, and the ingots obtained were free from slag. Continaed efforts in the same direction showed that we were now on the right path and that the problem was solved."

And in another place :

" I sent 15 tons of ingots to a firm in Sheffield and 15 tons to Messrs. Henry Bes- semer & Co.'s works at the same place, and went over to England in September to have them tried. I fonnd then that Mr. Bessemer had not succeeded, but had to grannlate in water the steel he got from the converter and afterwards remelt it in crucibles. As such a process could not give any profit, the friends who assisted him were losing all hope of success, but they all came down to Sheffield to see my ingots tried before they finally gave up this business. The Sheffield firm, thinking it their interest not to forward the Bessemer method, got the whole lot burnt at the wash-welding, but the 15 tons hammered and tilted at Messrs. Henry Bessemer & Co.'s works turned out to full satisfaction alter having been tried for knives, scis- sors, razors, other tools and plates.

The result of these trials has led to the further improvement and to the present extension of the Bessemer process.''

One condition which materially aided Mr. Goransson in making regularly a good Bessemer product was the freedom of his ma- terial from phosphorus and sulphur. The Swedish experiments first clearly showed that the Bessemer process, far from removing phosphorus in a higher degree than the older processes, as its sponsor had hoped at the outset, had, on the contrary, greater need of materials low in phosphorus than any of the older methods ; and after the trials at Edske had been crowned with the success men- tioned, Swedish pig-iron only was used for two years at Sir Henry Bessemer & Co.'s works in ShefiQeld, in making tool-steel.f

The relative freedom of the Swedish pig-iron from sulphur was of further aid, in making the Bessemer steel produced from it less liable to red-shortness than that prepared from other pig-irons. The latter appeared to need so urgently an addition of manganese for the removal of red-shortness, as to induce the general belief that the Bessemer process first came to be practicable through the additions of manganese introduced by B. Mushet. The exaggerated character of this opinion may be readily inferred from the fact that, at some of the Swedish Bessemer works, no ad- dition whatever of either spiegel or ferro-manganese is made for the best steels ; whil6 at other Swedish works, where such an ad-

History of the Manvfaeture of Iron in aU Ages by James M. Swank, p. 405. t See Sir Henry Bessemer's above-quoted description in Trans A. I. M. xix 827.

268 The Bessemer Pbocb88 As Conducted In Sweden.

dition is employed, it is confined to an insignificant qaantitj com- pared with what is used in the majority of Bessemer works in other countries. But this circumstance is by no means entirely due to the small tenor of sulphur in the Swedish pig-irons. As will be shown later on, it stands quite as much in connection with the fact that such Swedish pig-irons as require little or no addition of man- ganese at the close of the process, carry proportions of silicon and manganese suited to the purpose for which they are to be used.

Mr. Goransson's plan included the taking of pig-iron, without remelting, direct from the blast-furnace to the converter. This practice was followed at Edske from the beginning of 1857, and has been continued ever since at all the other Swedish Bessemer works. The experiments made at two works, about 1870, in using remelted pig-iron, speedily showed that the product derived from such material was both costly and inferior to that made from molten pig-iron taken direct from the blast-furnace. It was not, therefore, at Terre Noire in France, as has often been stated, but in Sweden that the taking of pig-iron direct into the Bessemer converter without re- melting was first practiced. Moreover, it was done, in pursuance of the Swedish example, in Carinthia and Styria, before there was a Bessemer plant at Terre Noire.

The plan, adopted in Sweden from the beginning, of taking the pig-iron direct from the blast-furnace, was accompanied with diffi- culties, more particularly at the outset, because it was essential that the blast-furnaces working on Bessemer pig should steadily produce a pig-iron of suitable composition. Although it is now generally known in our Bessemer works that the art of getting a Bessemer plant, arranged in the Swedish manner, to work satisfactorily con- sists principally in a proper management of the blast-furnace to that end, it was a long time before not only the importance of this was seen, but the best way to accomplish it was learned.

The silicon in our forge-irons was generally between 0.2 and 0.4 percent.; but such pig-iron was unfit for the Bessemer converter, as it would make the blow too cold. On the other hand, as long as the blowing-engines lacked the necessary power, it became evident that pig-iron with more than 1 per cent, silicon caused an equally cold blow on account of the unreasonable prolongation of the pro- cess. After 1870, when the Swedish Bessemer works had at length secured blowing-engines of requisite strength, it was found best for the quality of the product to run the blast-furnaces so that the Bes- %mer pig should contain from 0.9 to 1 per cent, of silicon.

The Bessemer Process As Conducted In Sweden. 269

While in other countries pig-iron containing from 2 to 3 per cent, of silicon has been chiefly used, most Swedish Bessemer works have never desired more than 1 per cent, of silicon, but they have wanted at the same time 1.5, or, better still, from 2 to 4 percent, of manga- nese. This, together with about 4.5 per cent, of carbon, and the greatest possible freedom from phosphorus, sulphur, copper and ar- senic, is the chemical composition desired for Swe<iish Bessemer pig- iron ; but it is essential besides, that when tapped from the blast- furnace it should be thin and " hot;'' otherwise, with so little silicon, especially when the proportion of manganese is not particularly great, the blow in the Bessemer converter will be " cold," or rather, not sufficiently hot.

With charcoal for fuel, and with as thorough and careful a roast- ing as is ordinarily given to the ore in Sweden, the charges of the Swedish blast-furnaces contain, as a rule, but a trifling amount of sulphur; and there is, therefore, unlike the practice in coke blast- furnaces, no necessity for any special effort to remove sulphur in the blast-furnace. In other words, the condition, so unusual in other countries, obtains in Sweden, that it is generally unnecessary to keep the temperature in the blast-furnace any higher, on account of sul- phur, than is needful to effect reduction and a regular working of the stack.

In this connection I embrace the opportunity (although it has nothing to do with the Swedish Bessemer practice) of strongly em- phasizing the fact that white pig-iron in Sweden is just as good, and often fully as free from sulphur, as the gray, although foreign con- sumers generally believe, judging from their own pig-irons, that only the gray pig is suitably free from sulphur, and therefore present an absurd requirement in asking for a gray pig-iron, with not more than 0.2 per cent, of silicon.

The cheapest way of producing, under the Swedish conditions named above, a pig-iron containing 0.9 to 1 per cent, silicon, is to so arrange the charge that, with alumina considered as a base, the slag would be about a 2.5-silicate. With a charge in which the metallic iron does not exceed 50 per cent, there would then be re- quired only a slight excess of heat to obtain a pig-iron with the silicon-ratio in question, and it could be produced with but little larger consumption of charcoal than would be necessary, with the same charge and temperature of blast, merely to maintain regular running. But such a pig-iron, when run from the blast-furnace, would be oold," and would not give a sufficiently warm blow in

270 TH£ BESSEMER PBOCCSS AS OONBUCfTED IN SWEDEN.

the converter ; and it therefore becomes necessary for Bessemer uses, even with that kind of charge, to keep the blast-furnace running with an excess of heat. As a still further consequence, the pig-iron has a tendency to contain more nearly 2 than 1 per cent, of silicon, and perhaps even more; and this higher silicon, in its turn, con- tributes to a warmer blow in the converter.

There are indeed sonde Swedish Bessemer works which, to gain a cheaper product in producing iron or soft steel for ordinary purposes, use a pig-iron made with an acid blast-furnace charge, and hence low in manganese, and containing usually about 2 per cent, of silicon. This acts in the Bessemer converter about the same as the Bessemer pig-irons generally used in other countries, in that it yields a slug- gish slag, and requires, after the close of the blow, an addition of manganese which is lare in comparison with the ordinary practice in Sweden. Such conditions constitute, in this country, not the rule but the exception ; and it is not that kind of Bessemer process, but that which is typical of Sweden, which alone will be the subject of this paper, and which alone is under consideration in what follows.

All ingot-metal is divided in Sweden into twoclasses: steel, high in carbon, and capable of hardening ; and iron, low in carbon, and incapable of hardening. Since, as this paper shows, Swedish Bes- semer metal contains relatively small quantities of such impurities as manganese and silicon, it is the percentage of carbon alone which fixes the limit between steel and iron. This limit is said to be at 0.40 per cent, of carbon — all ingot-metal with higher carbon being called steel, and all with lower carbon, iron.

For the Bessemer pig ordinarily made in Sweden, a basic, not an acjd, charge is used, with a slag which (alumina being reckoned as a base) usually lies between a 1.5- and a 1.2-silicate. The reason for this is, in the first place, that the more basic the charge is kept, the greater e;ccess of heat must there be (especially as our charges are low in alumina) in the production of pig-iron, if it is still to contain the desired amount of silicon (0.9 to 1 per cent.); and so much the hotter will it be when it comes to the converter. This is all the more necessary because, on the one hand, the pig-iron is so low in silicon. On the other hand, on account of the long interval between the blows, and also because the charges with us are so small (mostly between three and four tons), the converter is generally by no means as warm when the pig-iron is turned in, as is most often the case in other countries. Another reason is that the supply of highly man- ganiferous ores is smaller than that of those poorer in manganese, and,

THE BESSEMER PBOCESS AS OONDUCrTED IN SWEDEN. 271

as 16 well known more of the manganese of a charge enters into the pig-iron when, other things being equal, the charge is kept more basic. Moreover, most of our more manganiferous ores are rich in lime and not seldom contain pyrites; and since it is more difficult to completely remove sulphur by calcining from calcareous than from sih'ceous ores, our more manganiferous charges often stand in need of a greater basicity and, at the same time, a hotter run in the blast- furnace, on account of the sulphur. Finally, it seems probable that a more basic charge co-operates in obtaining the higher percentage of carbon (4.5 per cent.) which is desired in the pig-iron.

With the small production of our blast-furnaces there can be, of course, no thought of equalizing the composition of the pig-iron by some arrangement more or less similar to Jones's pig-iron mixer ; and hence it becomes all the more necessary to endeavor by the most careful furnace*management to maintain that uniformity in running which is indispensable to good results in the use of direct-metal in the converter. Great uniformity in the run- ning of the blast-furnace is thus the end most earnestly sought at the Swedish blast-furnaces; and this circumstance, in connection with low wages and the limited market for products of the highest class (such as are chiefly considered in this paper), constitutes the principal reason for the comparatively slight importance attached in Sweden to large output. It is, however, at the Bessemer blast- furnaces, not enough to keep a good and uniform run with the same kind of charge. Some Bessemer works change their blast- furnace charges to produce steel of different grades of hardness; experience having shown that a charge most suitable for a certain purpose is not necessarily equally well adapted to another.

This question stands in closest connection with Mr. C. A. Cas- persson's plausible explanation concerning the influence which the heat during the Bessemer process has on the presence of blow-holes in the ingots. By close observations, made during many years to ascertain the relation l)etween the composition of the pig-iron and the conditions of the blow, with special regard to the temperature, and also by breaking ingots from each observed blow, to determine the appearance and position of the blow- holes, Mr. Caspersson was at last able to predict with certainty, from the known composition of the pig-iron and the nature of the blow, the character of the ingots as regards blow- holes.*

See JemkonioreU AnnaleTf 1882, p. 295. In translation, Stahl und Eisen, 1883, p. 71 ; and in summary, TKe Journal of the Iron and Sud Inatitute, 1883, No. 1, p. 480.

272 The Bbbsebieb Process As Conducted In Sweden.

In this way Mr. Caspersson found that extremely cold blows, the temperature of which is designated as No. 0, give ingots which are full of blow-holes from the surface to the middle, and consequently unserviceable. Good ingots are not obtained until a so much higher temperature (No. 1) is reached, that the blow- holes are limited to a belt not less than a couple of centimeters (f inch) from the surface of the ingots. Blow-holes in this place do much less harm than those which, with however small an opening, reach the surface of the ingots. The former do not oxidize as the latter do, and are therefore more easily welded. Moreover, defects in the surface of ingot-metal generally cause greater harm than similar defects farther in.

The temperature now in question, or temperature No. 1, is the most common in making Swedish Bessemer steel, and for this rea<n ingots with a belt of blow-holes, which have no opening to the sur- face, are those most common here. Since the announcement of Mr. Caspersson researches, we have in Sweden, more than in other countries, made it a study to avoid the next higher range of tempera- ture (No. 2), as it occasions blow-holes situated so near the surface that they come in contact with the air, and in consequence become oxidized. These surface blow-holes are, to be sure, very often smaller than those situated somewhat deeper, and caused, as men- tioned, by the next lower temperature; and ingots of Bessemer- metal at the higher temperature now in question contain, therefore, generally fewer blow-holes than any of those previously described. But, since, as has been said, even the finest surface blow-holes are far more injurious than considerably larger blow-holes lying a little farther in, the former should always be carefully avoided.

A somewhat higher temperature (No. 3) gives ingots compara- tively solid, both at the surface and towards the center, and is there- fore desirable, particularly for steel high in carbon, where the belt of blow-holes is still more difficult to weld than in ingot-iron lower in carbon. At a still higher heat (No. 4), the ingots certainly con- tinue to remain comparatively free from blow-holes, or rather become more and more so, but begin simultaneously to suffer from " piping," which grows more troublesome as the heat is increased.

The changes in the state of the blow-holes in ingots, to which

reference has been made, and which are caused by differences in tem-

— ture, hold good, however, only for ingots of the same hardness.

metal with a higher percentage of carbon, as is well known, the

ing-point becomes lower; and as the temperatures above dis-

The Bessemer Process As Conducted Ix Sweden. 273

cussed really indicate different amounts of surplus heat above the melting-point, it is natural that the amount of heat which with soft iron, for instance, scarcely suflBces for temperature No. 1 and gives ingots with blow-holes, should give, for harder iron and soft steel, temperature No. 2 and ingots with surface blow-holes, and for hard steel, temperature No. 3 and comparatively sound ingots. The tem- peratures described above are, in other words, by no means absolute, but relative, varying with the desired hardness of the product, because the absolute temperature corresponding to a given temperature- number increases as the product becomes softer.

As has been mentioned already, relatively little or no manganese IS added at the end of the blow at most Swedish Bessemer works. Recarburizing is accordingly insignificant, especially as ferro-man- ganese has latterly been used here much more than spiegeleisen, and the blows in Sweden are therefore oftenest made more or less "direct," that is, they end as soon as the carbon has been reduced to the amount wanted in the product, or very little below. This method, peculiar to Sweden, is preferred to the one sometimes formerly practiced here, namely, recarburizing with pig-iron, which was added in a molten state, together with iron rich in manganese. The reason for this preference is, that Bessemer works having proper control of the blast- furnaces succeed, in the manner first mentioned, in getting a better and more even product than by recarburizing. One essential con- dition, however, is, that the amount of silicon in the pig-iron must not materially exceed 1 per cent., for in that case one might ea<ily (especially in making hard steel, where the blow has to be ended earlier) run the risk of a large percentage of silicon ; whereas, as the conditions with us now generally are silicon is mostly below 0.1 and very often not above 0.05 per cent. In those exceptional cases where pig-iron with about 2 per cent, of silicon is used, recarbur- izing with pig-iron is also generally employed.

It is the above-mentioned conditions that furnish the principal reason for keeping the blast-furnace charge, at typical Swedish. Bessemer works, the more basic, the hotter the Bessemer operation, is desired to l>e. But this does not prevent hard steel from being richer in silicon when produced with temperature No. 3 than with. No. 1; for even if no addition is made at the end of the process,, the percentage of silicon in the final product is, as is known, not dependent exclusively upon that in the pig-iron, but also upon the temperature of the blow. For no material does the tendency to* oxidize increase more with a rise of temperature than for carbon,.

VOL. XXII.— 18 r\r\n]o

? IvL

274 The Bessemer Process As Conducted Ix Swbdex.

and if the blow is stopped when the amount of carbon is reduced to only 1 per cent., the percentage of silicon in the steel may be 0.15 per cent, if the grade of heat during the blow has been about No. 3, while it amounts to but 0.04 or 0.06 per cent, in a steel of the same grade of hardness, if produced from a pig-iron almost alike in silicon, but made from a K'ss basic charge, and hence carrying so much less surplus heat, that the temperature during the blow remained about No. 1.

As a rule, Bessemer ingot-iron and soft steel containing less than 0,6 per cent, of carbon cannot be produced advantageously except with temperature No. 1, and ingots of these grades of hardness have consequently a belt of blow-holes; for it is only exceptionally that the heat can be so raised, even for soft steel, as to reach No. 3, and the intermediate temperature No. 2, with its injurious surface blow- holes should, of course, always be avoided. On the other hand, hard steel is produced with a temperature which corresponds to No. 1 as well as to No. 3, and temperature No. 3 is for very hard steel with 1 to 1.4 per cent, of carbon the same as No. 1 for iron with 0.2 per cent, of carbon. Likewise for steel with 0.6 to 1 percent, of carbon there can be used both the temperature No. 3, giving sound ingots (temperature No. 3 for such steel being the same as No. 2 for soft steels, with its undesirable surface blow-holes), and also temperature No. 1 with its belt of blow-holes ; but the latter temperature is, on the other hand, entirely insufficient for producing ingot-iron, and causes even for soft ingot-steel bad blow-holes all through.

As ingots produced at temperature No. 3 become comparatively sound, many might be led to expect that all grades of hardness from 0.6 per cent, carbon up should be made at temperatures correspond- ing with No. 3. This is, however, by no means the case; on the contrary, it is more usual for harder steel, as well as for ingot-iron, to employ temperature No. 1. The reason for this is, that No. 3, as compared with No. 1, leaves in the product not only primarily a larger percentage of silicon, but also a larger percentage of manga- nese. As the amount of these materials naturally becomes still greater in Bessemer products produced at temperature No. 4, and ingots thus prepared have moreover the disadvantage of being " piped," it is a matter of course that temperature No. 4 h used still less than No. 3. An effort to reach it is made, however, when a perfectly sound ingot is of particular importance, though in such a case the upper part of the ingot, with its " pipe,*' has to be sacrificed.

In the conditions thus stated must be sought the reason why, at

THE BfiSSEMER PBOCBBS AS a>NDUCrED IN SWEDEN. 275

the same works, the blast-furnace charges are sometimes changed somewhat for making different grades of steel. When the greatest importance is attached to the soundness of the product, a more basic charge is used for steel with 0.6 per cent, of carbon and over, than for softer products, which, if blown from the same pig-iron, would have surface blow-holes. When the steel is intended for purposes very sensitive to red -shortness, a pig-iron produced from an extra- basic charge, rich in manganese, is used, even for the softest iron, having lees than 0.15 per cent, of carbon.

Some examples of the percentage-composition of blast-furnace slags from Swedish Bessemer works (small proportions of alkalies being omitted) will be found in Table I.

The percentage of silicon in the pig-iron is, as already said, gen- erally from 0.8 to 1.1, and that of manganese varies from only 0.6 up to 6, but it is only at one or two works that it it exceeds 4 per cent.

Table I. — Slag- Analyses from Swedish Bessemer Blast- Furnaces,

A1,0,.

CaO.

MgO.

MnO.

FeO.

CaS.

'ate grade lina reek- 1 as A base

DomDarfvet..

5.S1

? 1.19

Hagfors

0.43 1.20

Westanfore...

? 1.30

Nykroppa

?

o

LanghjttaD...

?

Sandviken

?

Bangbro

A consequence of the low percentage of silicon in the pig-iron is that the boil, or violent ebullition of the carbon, generally begins from one and a half to three minutes after the blow has begun, and the ordinary time for the entire blow is not more than seven to ten minutes, not counting the time used in taking samples. The area of the tuyere-holes, which in Sweden are very large as com- pared with the charge of pig-iron, contributes in large measure to this result. This area generally amounts to from 30 to 35 square centi- meters (4.66 to 5.43 square inches) per ton of pig-iron; excep- tionally, it may on the one hand reach 50 and on the other

276 THE BESSEMER PROCESS AS OOXDUCrED IN SWEDEN.

hand 15 square centimeters (7.75 and 2.33 square inches) per ton. That the proportion just now mentioned is so large, depends, how- ever, by no means on an absolutely large total tuyere-area, but simply upon the small charges, which in general do not amount to more than 3000 to 3500 kilograms (6614 to 7716 pounds) of pig- iron. The converters are proportionately small, their diameters being about 1.5 to 1.6 meters (4.92 to 5.25 feet), but at the bottom only 1.2 to 1.3 meters (3.94 to 4.26 feet), while the interior height from the bottom to the mouth usually varies between 2 and 2.5 meters (6.56 and 8.2 feet.)

For more than twenty years, rotating converters exclusively have been used in Sweden, and for the better preservation of heat they always have the mouth on the side. For the same reason the mouth of the converter is always very narrow, being often only 0.2 to 0.25 meter (8 to 10 inches), but occasionally 0.3 meter (12 inches) in diameter. The total area of the tuyeres amounts generally to 80 or 120 square centimeters (12.4 and 18.6 square inches), and is most often divided up into from 70 to 200 holes, for the most part 9 to 10 millimeters (0.35 to 0.39 inch) but now and then 6 to 15 milli- meters (0.25 to 0.59 inch) in diameter. The pressure of the air is most frequently between 400 and 1000 millimeters (16 and 39 inches) of quicksilver (equal to 7.8 and 19.6 pounds per square inch) and the blowing-engines are generally of from 600 to 900 horse- power.

The procedure during the blow depends not only upon the initial temperature of the pig-iron and upon its chemical composition, with which the generation of the heat varies, but also upon the degree of fluidity of the pig ; because when the pig-iron is viscous, the pro- cess may be so delayed, through increased resistance to the blast, par- ticularly if the blowing-engines are not especially strong, that the boil does not begin until much later than would be the case with a fluid, but in other respects similar, pig-iron. Such viscosity is here often the result of a large percentage of silicon ; and to the reasons previously given for a basic blast-furnace charge, which gives with the same temperature a smaller amount of silicon, may therefore be added this, that not only the pig-iron, but also the Bessemer pro- duct made from it, becomes more liquid, a result which is still fur- ther promoted by the amount of manganese in the pig-iron.

With the exception of silicon, there is no material which has such an influence upon the operation of the Bessemer process as manga- nese. To investigate its effects, researches were made as early as

The Bes8Emeb Process As Conducted In Sweden. 277

Table II. — Analysea of Variotia Manganiferous Bessemer Pig-Irons, and of the ResvUing Baths and Slags.

Name of

Character of

Time of blow;

after which

samples were

taken.

Iron contained.

Composition of Bessemer Slags.

work. sMunpie.

Si. Mn.

FeO.

MnO.

MgO.

CaO.

AltO, SiO,.

Silicate grade.

Mir.lpAn.

46 fin

Ungbytt*n

Bessemer bath.

2m. 45s.

2m. 16s.

34.72' 18.95

.24' 2.60

.78, 48.76

"

4m. 80b.

.03' .12

21.08| 15.48

.80 8.25

.98 69.82

"

5m. 80s.

.72' 48.48

2.56! 61.00

Pig-iron.

1 Os

StndTlken.

Bessemer bath.

2m. 30s.

3m.

.Os .11

Trace

1.84' 53.44

j

u

6m.

.03 .09

17.04, 22.80

1.94 57.80

u

5m. 45s.

.02| .07

18.48 22.23

1.58! 56.76

5.83I 47.16

Pig-iron.

.88 1.16

1

.77 2.14

Nykroppa..

Bessemer bath. Im. dOs.

2m. 30s.

.10 .15

13.50 29.76

.23 .42

2.28| 53.26

5m. 806.

.05 .16

9.54| 23.70

8.90, 62.84

M

6m. 80s.

.04 .08

2.14 44.52

Plg-iron.

1.02 1.83

.U

" 1 10.88 27.22

3.86 50.20

BtQgbro

Bessemer bath. 2m. 458.

3m.

.03 .22

14.20| 26.81

2.86 55.26

j

M

4m. 458.

.03 .12

18.52 31.01

2.70, 47.20

5m. 45s.

.03 .09 1.06 5.12

81.19, 25.43

2.24 40.50

Piv.lmn.

.70 11.16

18.37 19.10

4.36 46.72

i.

Westanfors

Bessemer bath. 2m. 80s.

4m. 15s.

.43 3.26

4.20 46.38

3.08 46.87

u

8m. 35s.

6.24' 52.26

2.49 39.07

M

9m. 20s.

2.94 37.63

1875 and 1879, under my direction, by Dr. A. Tamm* and Mr. C. G. DahleruSyt at five works, which at that time were using pig-iron with very different percentages of manganese. Some of the results obtained are summarized in Table II.

The opinion, that combustion during the Bessemer process is per-

JemhmioreU ArmcUer 1878, p. 444, and 1880, p. 462. t Ibid, 1881, p. 314.

278 THE BESSEMER PROCESS AS CX)NDUCrrED IN SWEDEN.

formed exclusively by the blast direct, and consequently without the aid of the slag formed by it, has been all too prevalent. Should any one still retain this erroneous notion, the slag-series from L5ng- hyttan in particular, and those from Sandviken and Nykroppa be- sides, may powerfully aid in bringing about a sounder view. That the percentage of oxidized iron in the slag during the boiling-period can be diminished as much as from 34.7 to 21.1, points out in an indisputable manner, that the oxidation of the carbon in this case must have been performed in an essential degree by the reduction of oxidized iron from the slag. In this refining, the slag has in reality assisted the free oxygen in still greater measure than would gener- ally be noticed at first sight; for besides the fact that the amount of silica in the slag, considered absolutely, has only slightly increased during the interval (so that in place of 48.8 silica to 34.7 ferrous oxide, there is found, two and one quarter minutes afterwards, 59.8 silica to 21.1 ferrous oxide, or a proportion equivalent to 48.8 silica to 1 7.2 per cent, ferrous oxide), it is evident that a material part of the combustion of carbon which has been effected by free oxygen did not in point of fact occur directly, but was brought about by the in- tervention of slag. For during the boil, as well as at the beginning and end of the process, iron must have been oxidized and subse- quently reduced again by carbon, although this action escapes notice, and indeed would do so even if more iron were thus reduced.

That the proportion of oxidized iron does not always diminish during the "boil," but, as in the slag-series from Bangbro and Westanfors, may, on the contrary, increase, affords no proof that the slag has been inactive in refining, for the condition alluded to is evidently the result of there having been more iron oxidized during the boil than was reduced out of the slag by the cafrbon. Without a doubt, even in this case, refining has been assisted partially by the slag, although it cannot be said in this instance, as it could in the other, to have taken place in part at the expense of slag previously formed.

An interchange must take place in the Bessemer, as in all other refining processes, between oxidizing or oxygen-yielding materials, on the one hand, and reducing or oxygen-taking materials on the other. How far, at any particular moment, more or less iron is oxidized than is simultaneously reduced again by the other elements of the pig-iron depends as much on the chemical composition of the metal- and slag-bath as on the temperature and internal mixture. At any moment during the process there must be a spontaneous

The Bessemer Process As Conducted In Sweden. 279

tendency towards the state of equilibrmniy necessitated by the con- ditions set forth ; and a result of this is that the content of oxidized iron in the slag increases at the beginning, until it thus becomes so oxidizing that as a consequence iron is reduced as fast as it can be oxidized by the oxygen of the blast.

This occurs earlier, or with a lower content of oxidized iron in the slag, the hotter the blow ; because the tendency of carbon to take up oxygen increases far more rapidly than that of iron, when the temperature rises above yellow. A given result of this is that the warmer the Bessemer blow the smaller will be the loss of iron by combustion, and the more acid, and, as a consequence, the thicker and tougher will be the slag.

Another condition which aids considerably towards the establish- ment of the state of equilibrium, even when the amount of oxidized iron in the slag is comparatively low, is a larger proportion of man- ganese in the pig-iron ; for manganese, besides contributing to the higher temperature, with the resultant lower proportion of iron in the slag, also makes the slag more liquid, so that it mingles more intimately with the mass of iron, and, in consequence, operates more in oxidizing, albeit the content of iron is lower than would be the case Vith a less liquid slag. It is also an important fact that the oxide of manganese takes to a certain degree the place of that of iron, though not to such an extent that the combined amount of MnO and FeO in the slag must not be considerably greater than if the amount of manganese were less. From this it follows that the blow progresses more slowly, and that the waste increases with the amount of manganese in the pig, though in less than equal ratio.

These conditions bring it to pass that of two pig-irons with the same amount of silicon, but different amounts of manganese, that which contains the most manganese will give the greater quantity of more liquid and less oxidizing slag, which during the boil does not take part in the combustion of carbon as actively as does a smaller quantity of slag which is richer in iron ; and the result is, therefore, that when the blow does not progress so rapidly that the rise of temperature during the boil compensates for the falling off in the combustion of carbon due to the decrease in the content of car- bon, a Bessemer slag, produced from a pig-iron richer in mangan- ese, can, even during the boil, be richer in oxidized iron, in that more iron is oxidized by the oxygen of the blast than can be sim- ultaneously reduced out by carbon. This is, e.g,y the case with the slag- series obtained at Bangbro and Westanfors, but there are cases

280 The Bessemer Process As Conducted In Sweden.

where, on the contrary, the content of iron, even in Bessemer slag rich in manganese, diminishes somewhat, though comparatively little, during the boil.

When the silicon in the pig-iron is not higher than in the pig- irons given in Table II. (maximum 1.14 per cent.), almost all the silicon is necessarily removed before the boil,'' unless the blow be very hot, either because the pig-iron contains more manganese than usual for the Bessemer process, or for some other cause, as, for in- stance, an initial high temperature of the pig-iron ; but in the last- named case the removal of the silicon is delayed by the temperature, and the reason for it is the already oft-repeated fact that when the temperature rises al>ove the melting-point of pig-iron, the tendency of carbon to be oxidised increases more rapidly than that of silicon. With manganese the case is somewhat similar, so long as there is not over two per cent, of it in the pig ; for the incom- parably greater part of the manganese is then removed before the boil. But with higher contents of manganese (4 to 6 per cent.), as shown by the Westanfors series, the oxidation of the manganese is more evenly distributed throughout the entire blow. According to information kindly given by Mr. J. A.. Brinell, the following are the ordinary results at Westanfors, where the blow is always direct, and there is, therefore, no subsequent addition of manganese:

In steel and iron produced from pig-iron with 4 per cent manganese and 1 per cent, of silicon, the carbons in the product are accompanied by the percentages of manganese and silicon shown under them.

Percentages,

Carbon, . 1.3 1.1 0.9 0.7 0.6 0.3 0.2 0.15

Manganese, 0.6 0.55 0.5 0.4 0.3 0.2 0.15 0.12

Silicon, . 0.06 0.05 0.045 0.045 0.04 0.03 0.02 0.015

In Bessemer steel from pig-iron with 5 to 6 per cent, of manganese and 1 p&r cent of silicon, the corresponding Bgures are as follows :

Carbon,

Manganese,

Silicon,

Before we had learned to understand the workings of an un- usually " hot " Bessemer blow and to regulate properly the heat in question, such exceptions could occur, even in this country, as that a product was obtained quite unexpectedly with 0.5 carbon, 1.10 manganese and 0.5 silicon. Such anomalies no longer occur if the

THE BESSEMER PROCESS AS CX>NDnCrED IN SWEDEN. 281

blast-furnaoe is suitably under control ; and for this reason, as well as for those already given, it may be asserted with full confidence that the technical success of the Swedish Bessemer process depends above all on the running of the blast-furnace.

With so small an amount of silicon in the Bessemer pig-iron, and so large an aggregate of tuyere-area per ton of pig as is generally the case in Sweden, the refining or slag-forming period must be very short, most frequently one and one-half to three minutes, but at times only one minute, and, on the other hand, from four to five minutes for pig-irons with more manganese. As a whole, the blow generally lasts between five and ten minutes ; but for pig-irons with high manganese, that time may extend to 15 minutes and sometimes even to 20 minutes. Since blows are made at ordinary Swedish Bessemer works approximately direct to the desired carbon, it must be evident that the duration of the blow will increase with the degree of soft- ness desired in the product.

As there is but little silicon in the pig-iron, and hence the tem- perature in the converter is relatively low, the slag is only slightly acid, and it may even be somewhat basic. Hence it is considerably more fluid than the converter-slags usual in other countries.*

More especially is this the case with the slags derived from the more manganiferous pig-irons, which, when the charge is poured, are so liquid and so white that one not accustomed to them often cannot tell when the iron ceases and the slag begins to run out of the ladle.

Inasmuch as with the less manganiferous pig-irons, the mangan- ese practically passes entirely into the slag, it follows, as appears from Table II., that about 0.8 per cent, of manganese in the pig- iron will suffice to render the protoxide of manganese the prevail- ing base of the slag, until such time as the carbon of the bath is so lowered that the latter consists of soft steel ; and since, more- over, the Swedish Bessemer slag generally remains about a bisilicate, it is easy to understand that its principal mass when cold is most frequently crystalline, on account of the rhodonite which has crys- tallized. But the color is generally not a handsome red, except in

The direct effect of a relatively low temperature would, of course, be to thicken the slag, not to thin it Prof. Akerman probably means here that the low tempera- ture hinders the slag from attacking the lining of the vessel and from taking up silica therefrom. Thus the slag remains relatively basic A relatively basic slag is more flaid than an extremely acid one would be, even at a higher temperature. — Tbkslators.

282 The Bessemer Process As Oonducted In Sweden.

the slags of hard steel, because the higher the percentage of fer rous oxide runs above 10 per cent, in the slag, the more does the clear pink color of rhodonite change to a darker brown ; and the longer the blow is continued after the boil, the richer, of course, will the slag become in oxidized iron.

To the fluidity of the slag is due, in my opinion, the extraor- dinarily small tendency to red-shortness in the Swedish Bessemer products. For the more liquid the slag, the more completely does it absorb the oxidized iron ; while a viscid slag, because it cannot mix intimately with the mass of molten iron, leaves iron oxide in it ; and thence results red -shortness.

The amount of manganese left in the iron itself by the more manganiferous pig-irons, certainly contributes still further to the same end, in that the oxidation of the iron is hindered by the metal- lic manganese present, which reduces again iron already oxidized ; but if the effect of manganese on red-shortness were confined to this, the manganese present in the pig-iron at the beginning, at least when not above two percent., could not be very effective in the man- ner now in question ; because, if less than two per cent., it would be, as the Table shows, so far removed during the earlier stages of the blow, that the small remnant in soft iron could not possibly accom- plish the result described. Experience, however, proves that red- shortness, even in soft iron, is prevented in some measure by an even smaller amount of manganese in the pig-iron ; and this circum- stance confirms the opinion that even the manganese which has been slagged must have this effect. The manner in which this occurs certainly must be, that the fluidity of the slag, increasing with the contained manganese, assists in washing away the oxidized iron, which otherwise would remain in the mass of iron and render it red-short.

The slight need in ordinary Swedish Bessemer works of a subse- quent addition of ferro- manganese (for generally only 0.2 to 0.6 per cent, is added), depends meanwhile by no means exclusively on the lower amount of oxidized iron in the product, but also upon the comparative freedom of the latter from sulphur, phosphorus and silicon. Since it is nearly free from these elements, there is naturally no need of adding manganese to counteract the injurious influence which their presence would occasion.

In what precedes, only the advantages of manganese have been presented ; but it has its disadvantages also. It not only materially increases the cutting-action of the slag on the lining of the converter,

The Bessemer Process As Conducted In Sweden,

80 that the durability of the tuyeres, bottoms and walls of Bessemer converters is rapidly diminished with the increase of manganese in the bath ; but it also increases the waste, in&smuch as protoxide of manganese is by no means as effective in refining as ferrous oxide, while the slag, for this reason, as has already been shown, requires more of manganous and ferrous oxide together than of ferrous oxide alone.

Moreover, the waste depends partly, as said, on the temperature, to which it stands in inverse proportion, and partly on the degree of hardness of the product. Naturally, the waste will be greater the softer the iron produced ; because, to make the product softer, we must blow longer, and more iron is to be oxidized in a given time the '

CASPCRSSON'S Convcrtcr-Ladlc

the less the contents of carbon in the bath of iron. It is, therefore, rash to give any exact numerical measure of the waste ; but it would keep mostly between 10 and 10.5 per cent., though it may fall, on the one hand, to only 9, while, on the other hand, in the case of pig-irons with higher manganese, it may rise to 12, and now and then even to 12.5 per cent.

Notwithstanding the small amount of silicon in the Swedish pig- iron, the waste at our Bessemer works is not much smaller than is common elsewhere. This again stands in conjunction with the colder run in the converter which generally prevails here and, in its turn, brings about the noteworthy difference, already referred to, in the character of the , in that our Bessemer slag, which are

284 The Bessemer Process As Cx)Nduct£D Ix Sweden.

comparatively slightly acid and from that to somewhat basic, are very liquid in comparison with the decidedly acid and sluggish slags which are the common ones in other countries. The temperature be- ing wrongly estimated from the fluidity of the slag, causes many to imagine contrary to the fact, that Bessemer blows in Sweden are hotter than in other countries.

Besides the actual waste given above, there was, down to 1880, a loss of 2 or more per cent, due to the formation of ladle-skulls; but this trouble was luckily obviated by Caspersson's converter-ladle,* so-called, which, since the date mentioned, has been used with the greatest success at the majority of the Swedish Bessemer works. This arrangement, illustrated in Figs. 1 and 2, consists of a narrow ladle, which is furnished low down on the side with a lateral opening. A, that fits the mouth of the converter, D. Immediately after the blow, the ladle is brought up to the converter-mouth and made fast by means of wedges, C, in the lugs, B (see Fig. 1). The converter is then turned down still farther, so that its position is as indicated in Fig. 2, and a part of the bath of metal runs into the converter- ladle, while most of the metal remains in the converter. It is precisely on this circumstance, coupled with the smaller depth of bath occasioned thereby, that the advantages of the converter- ladle depend. A ladle, being never as hot as a converter at the end of a blow, operates, therefore, always to chill ; but as the converter-ladle is quite small, its power to chill the metal-bath, the incomparably larger part of which remains in the converter, after it has been turned down, is far less than that of an ordinary ladle, which receives all of the molten mass.

The result is that the liquid Bessemer metal can be allowed to re- main in peace and quiet longer, to permit the gases to escape, and without fear of partial freezing before being tapped into the moulds, than before the advent of the converter-ladle ; at least, longer than was possible with a product only as moderately hot as the Swedish Bessemer metal generally is, especially since the quantity handled in Sweden at each blow, as has been shown, is comparatively very small, and consequently the cooling effect of the ladle is proportionally great. By the aid of the converter-ladle, ladle-skulls, formerly so common here, can be prevented. But to this great advantage may be added another, viz., that the product is somewhat freer from blow- holes, partly because with the converter-ladle a longer time is given

JemkontoreU Annalerf 1880, p. 471, and The Journal of the Iron and Steel Insiituief 1880, ii., p. 699.

Tue Bb88Emeb Process As Conducted In Sweden. 285

Casjpssoh's Straincr-Funnkl

286 The Bessemer Process As Conducted In Sweden.

for the escape of the gases and for the more complete separation of the slag by rising to the surface, and partly on account of the raore moderate speed of the stream of metal incident to the smaller depth of bath.

Since the converter-ladle, during teeming, is fast and immovably fixed to the converter, it is evident that the former cannot, like an ordinary crane-ladle, be carried over the moulds, but that the moulds roust be brought under the converter, either on cars or on a turn- table.

As has been mentioned already, the formation of blow-holes in in- gots stands in the most intimate connection with the temperature that prevails during the blow. If the temperature has been too high, as when temperature No. 2 instead of No. 1 has been employed, and there is reason to fear surface blow-holes in the ingots, this serious detriment can be somewhat lessened only by allowing the product so to cool before being tapped that it then has the same heat as that which ordinarily prevails with temperature No. 1. If this is to be attained only by letting the metal stand longer in the ladle before being tapped into the moulds, there is danger with us of ladle-skulls, and sometimes even of the partial closing up of the nozzle together with the spurting stream occasioned thereby ; and this, in turn, gives ugly and faulty ingots. The dangers of these disadvantages is cer- tainly very decidedly diminished by the converter-ladle; but even with its assistance, one cannot always make use with safety of as warm a blow as would be desirable for the complete prevention of red-shortness. To meet this diflSculty, Mr. Caspersson has comple- mented his convorter-ladle with a so-called strainer-funnel, illustrated in Figs. 3 and 4.

The plate-iron funnel, 6, furnished with the handle, a, and lined with refractory material, has the easily-withdrawable chamotte-bot- tom, c, which is provide<l with many larger or smaller holes, the number of which is larger in the same degree as their diameter is smaller. When teeming takes place, there is first laid upou the mould, d, a cast-iron disk, e, with an opening in the middle,/, cor- responding in size to the changeable funnel-bottom, and a little Hide- hole, which affords a chance for the gases separated during teem- ing to escape, and also for the man who directs the tapping to judge when the moulds have l)een suitably filled. The preheated funnel is placed over the central opening, /, and the metal, tapped out through the converter-ladle into the strainer-funnel, is divided by means of the holes in the bottom into more or less numerous

The Bessemer Process As Conducted In Sweden. 287

streams of smaller or larger size. This division of the metal occa- sions, during its passing down into the mould, both a more complete removal of gases and greater cooling than takes place in the ordinary mode of tapping. The finer the strainer-holes, the more effective does its work become. There should be, therefore, at least two dif- ferent kinds of strainer-bottoms at hand, one with larger and the other with smaller holes, and the choice between them can be made according to the temperature of the blow. After an unusually hot blow, a bottom with smaller holes is taken ; and if the tempera- ture during the blow has been just hot enough, no strainer at all is used.

It might be supposed that the division of the metal into many fine streams would be likely to be dangerous by reason of the risk of partial oxidation with accompanying red-shortness ; but experi- ence shows that this is prevented by the gases which escape from the metal and speedily fill up the mould covered by cap, e, thus hin- dering the contact of the streams with the air.

Although too high a temperature produced during the blow can be suitably lowered by means of the converter-ladle and the strainer-fun- nel, without fear of making scrap, the higher proportion of silicon and manganese in the product incident to an over-hot blow cannot be thus remedied. This can be regulated to a certain degree by the practice employed in other countries, of adding cold steel scrap during the blow; but the method of Mr. Caspersson in this con- nection is still more effective, namely, to charge finely-crushed rich ore during the blow; for, besides the chilling which the addition of cold ore occasions directly, it causes an indirect cooling through the con- siderable absorption of heat involved in the reduction of the iron-ore.

By a proper use of these aids the blower, if he can judge the temperature accurately enough, can attain surprising uniformity, even in Bessemer products. This is particularly necessary in a country where most of the Bessemer steel made is used for pur- poses for which crucible steel is usually required elsewhere.

The Swedish plan of not recarburizing to any considerable extent, but of blowing approximately direct, must of course be accompa- nied by difiSculties in deciding when the blow should end, which are greater than those attending the usual course of continuing the blow till the metal is fully decarburized and afterwards, by means of dif- ferent additions, recarburizing to the desired hardness. No skill is needed to end the blow, if it is to be prolonged till the drop of the flame becomes conspicuously evident. Quite different does the prob-

288 THE BEaSEMER PROCESS AS CONDUCTED IN SWEDEN.

lem become, when the blow for the production of steel mast be in- terrupted while the carbon-flame is still so copious that changes in its size can scarcely be detected.

In these earlier stages, no great assistance can be had from the spectroscope, because it is only when carbon is so reduced that one can take note of the diminishing of the flame, that the changes in the spectra first become sufficiently apparent to be a sure guide by which to judge of the degree of hardness. Hence, although the spectra of the Bessemer flame were scientifically examined in Swe- den as early as the beginning of the decade commencing in 1860, the spectroscope has never come into general use here for this purpose.

That by which the Swedish Bessemer blower most guides himself during the blow, in producing steel of the higher carbons is, in reality, the character of the s|)arks; and many a Bessemer blower has acquired a marvellous skill in this direction, when the great difficulties are taken into account. Nevertheless, with a view to prevent faulty blows, it has become more common, in the course of time, for the blower to turn down the converter immediately before he considers the right point reached ; to convince himself by a hasty hammer-test whether the degree of refining has been judged correctly or not, and then to continue the blow for a few seconds. In producing ingot-iron, on the contrary, no such tests are re- quired ; and one judges, as has been said, by the flame.

As the Swedish Bessemer metal contains, in comparison with that made elsewhere, relatively little of any elements other than iron and carbon, it is natural that in this country the degree of hardness should have been determined by the contents of carbon, more espe- cially since the colorimetric method of carbon-determination, worked out by Prof. Eggertz, furnished so easy a means for a quick and generally satisfactory determination of the carbon in the Bessemer product. But besides the colorimetric determination of carbon, which is never omitted in Sweden, the product of every blow is subjected to a hammer-test, to determine not only the degree of hardness, but also how far the product is free from red -shortness.

When the metal is intended for purjwses of construction (which IS not as common in Sweden as it is elsewhere, since a very consid- erable part of the Swedish Bessemer product is used for tools and the like), it is, of course, subjected to the usual mechanical tests.

In this connection, and in such a paper as this, it should be specifically mentioned that the first scientific explanation of the in- fluence of carbon on the qualities of iron was made, as far as I

IThia [kAper was printed without the final revision of the author, who desires the following

'cUona: P. 209, 3d paragraph shonld read, "white pig-iron in Sweden is oommonly aa

-m sulphur/' etc.; P. 278, line 28, should read, "and indeed must do so, sa still more

M thus reduced ;" P. .278, third line fh>m bottom, " as much should read " as weU." j

Origin Of Qold-Brabinq Quartz Of Bendigo Reefs. 289

know, by the direction of Mr. G. F. Goranssotiy at Sandviken, and was carried out by Messrs. L. E. Soman and O. Kollberg. The latter published a report in Jernhmtords Anncder for 1865, p. 307, not only of the increase in absolute tensile strength as carbon in- creased, until it somewhat exceeded 1 per cent., hut also of the de- crease in specific gravity as the carbon rose. This paper by Koll- berg is all the more worthy of notice, as it gives (p. 314) not only the earliest analyses of Bessemer products, both metals and corres- ponding slag, from different periods of the same blow, but also an explanation of the course of the Bessemer process based on these investigations. This explanation is not only the first which - shows warrant for being called really scientific, but is, moreover, much more true than many an essay which has appeared even in later times concerning the peculiarities of the process in question. Though unknown to the majority of even scientifically educated metallurgists, it is a really fundamental and fully classical work, and should never be passed over by any one who wishes to give a true picture of the history of the Bessemer process. It is indeed none too early to accord a suitable recognition of his merits to the memory of him who has long since passed hence.

THE ORIGIN OF THE QOLD-BEABINQ QUARTZ OF THE BENDIQO REEFS, AUSTRALIA,

BY T. A. RICKARD, DENVER, COLORADO. (Chicaco Meeting, being part of the International Engineering Congreas, August, 1893.)

The lode-formation of the Bendigo gold-field was described in a former paper.* It presents a striking identity of arrangement with the general geological structure of the region, which is one of com- parative simplicity. The alternating beds of slate and sandstone which constitute the prevailing country-rock have been extremely contorted ; yet notwithstanding their highly developed cleavage it is possible to discern in their structure the evidences of original sedi- mentation. Along the crests of the waves which are the axes of anticlinal folds occur bodies of gold-bearing quartz, of great econom- ical importance, which the miners have appropriately termed "sad-

♦ "The Bendigo Gold-Field," Trans, xx.,463.

Vol.Xxh.— 19

290 Origin Of Gold-Bearing Quartz Of Bbndigo Reefs.

dies." They are more extensive in strike than in dip, are found to occur in a recognized succession, and have been followed by a very complete system of mine-workings reaching from the surface to more than half a mile below.

I. Tbe Rock-Formations. . At the commencement of our inquiry into the origin of this gold- bearing quartz, it will be necessary to consider the relative ages of

Fig. I.

l:iii;';3voLCAWTo

III' :':'[:] QWANiTE

Qeoloqical Sketch Map

Op A Portion Op

Victoria.

the different formations of the region. Numerous graptolites have made it easy to label the slates and sandstones in which the reefs occur as Lower Silurian. Any further subdivision, however, of this great thickness of rocks has been rendered almost impossible by reason of the striking lack of variety in the fossil remains, the gen- eral confinement of their occurrence to the slates; the marked

Okigin Op Gold-Bearing Quartz Of Bendigo Reefs. 291

scarcity of conglomerates and breccias; and finally, the notable similarity of textnre and composition which characterizes the succes- sive beds of slate and sandstone.

The enormous thickness of this series has been referred to. Dr. A. R. C. Selwyn, at one time head of the geological survey of Vic- toria,* estimated it at 36,000 feet. The same authority is in accord with the present government geologist, R. A. F. Murray, in consid- ering that a thickness of 7600 feet of superincumbent rock at one time covered the beds which now form the surface.

Fio. 2.

lyv-yjQRANiTE

Geological Map

Bl<2 HILL, BENDlQO.

CAlEt ie CHAIH6 %'

Kssh Pos-Lpliocene

The. geological map of the colony here reproduced (Fig. 1) indi- cates that the Silurian rocks of the BendigoJ gold-field are overlain to the north by Tertiary shales, while south they abut against the granite mass of Mount Alexander, which, in a horseshoe form, sepa- rates this mining district from that of Castlemaine.

The contact between the two older formations is best seen in cer- tain road-cuttings at Big Hill, 7 miles from the town. The accom-

Now director of the Geological Survey of Canada.

t Taken from Geology and Physical Oeography of Victoria by R. A. F. Murray.

I The old name " Sandhurst " appears on the map in place of " Bendigo."

292 ORIGIN OP GOLD-BEARINQ QUARTZ OP BfiyDIGO REEPB.

panying map'*' (Fig. 2) shows, on a scale larger than that of the general map, the principal features of the surface geology of this part of the district. It will be noticed that the crest of the ridge is formed, not by the granite but by the Silurian rocks where hard- ened by their contact with the granite. The railway passes through the ridge; but the tunnel indicated on the map is rendered useless for the purpose of obtaining a geological section by reason of the brickwork which effectually hides the rock-structure. A short dis- tance to the west, however, the main wagon-road, in crossing the hill, affords several interesting sections.

The northern edge of the cutting shows the commencement of the physical and chemical changes produced in the Silurian rocks by contact-metamorphism.f The slate graduates into a crystalline schist and the sandstone becomes a quartzite. More mica has been developed. At this point the structure of the country shows no marked disturbance; but in approaching the highest point of the road, attention is drawn to a distortion of the bedding by the occur- rence of a large black lava dike, which evidently follows a line of fracture.

Further southward the rocks begin to exhibit more marked altera- tion, being hardened, somewhat bleached, and very much jointed. At the crest of the ridge the granite itself is first noticed. (See Fig. 3.) On the west the cutting gives a partial syncline, the members of which show joints. Underneath, and upon a level with the road itself, the first intrusions of granite are to be seen. Small veins of it extend halfway up the height of the cutting. The rocks forming the immediate foreground of the sketch are cut up by joints and cross-joints. The bedding is indicated by the contour of the em- bankment which forms the side of the road.

On the southern slope of the ridge the cleavage of the slates is very marked, and the sandstones are traversed by several systems of fracture, while both rocks have become hardened and more brittle. The granite again appears in the form of small veins penetrating the overlying rocks. (See Fig. 4.) The vein of granite, G, intrudes among a series of thin beds which have been so altered that it is difficult to recognize which of them were originally slate. Two

systems of joints, C, C, are readily noted, more particularly in the

A reproduction of a small part of a general map of the gold-field, published hj the Department of Mines, Melbourne.

t Also termed hj Daubr, " the metamorphism of juxtaposition/* page 133, JEhtdes Synthdiques de Oiologie ExphimeniaU.

Origin Of Qold-Bearing Quartz Of Bendioo Reefs. 293

Is

mi

s

29it ORIGIN OF GOLD-BEABINQ QUARTZ OF BENDIQO BEEFS.

sandstones, A, A. A little further on, the gradually diminishing sides of the cutting gave the section reproduced in Fig. 5, where G, H und K are veins of granite (2, 2 J and 15 inches wide respectively), which have noticeably affected the sandstone bed, A. The slates, B, B, are also baked, and their ordinary cleavage (60® to 65®) is largely obliterated by cross-fracturing.

The wagon road does not show the actual contact of the two forma- tions; but the railway, in approaching the tunnel from the south, gave me the sketch seen in Fig. 6. Here the main mass of the granite throws out small branches which intrude between the bed- ding-planes of the slates and sandstones, penetrating them for a con- siderable distance. Some of these veins, C, C, are not more than a couple of inches thick. The darker lines, B, B, indicate segregrations of ironstone.

Approaching the tunnel-entrance a crossing of two granite veins is observed (Fig. 7).

To the facts above given, the following observations may be added. At 500 yards from the contact the effects produced are only faintly observable in the occurrence of joints recognized as not habitual ; at 300 yards the fracturing has become very marked and the rocks have lost their usual color. Even in the immediate vicinity of the contact the bedding of the Silurian rocks has not been notably disturbed. The changes produced in the slates and sandstones consist in their being somewhat bleached, considerably hardened, and very much fractured.

The penetration of the sedimentary rocks for such a distance from the contact by such small veins of granite, without however causing any disturbance of the bedding, affords a study of peculiar interest. To explain the behavior of the granite, 'it must be allowed to have been in a mobile condition. Mobility suggests fluidity, and fluidity in rocks is generally held to be the result of a molten state. An acid rock could not, however, by reason of heat alone, be in a plastic, much less a fluid state, except at an extremely high temperature. That temperature has been approximately determined as being pro- bably more than 2500® but less than 3000® F. Mallet's experi- ments on siliceous slags proved that their melting point was about 3000® F. Though we find that the general effects of heat upon the slates and sandstones are indeed observable, yet the actual surfaces in contact with the granite are not particularly affected; and the

Also those of Sir Lowthian Bell.

ORIGIN OP GOLD-BEARING QUARTZ OP BENDIGO REEPd. 295

evidence as a whole strongly indicates that no such tem[)erature as that just mentioned could have existed at this point.

The volcanic phenomena of to-day furnish a clue to the problem here suggested. The ejection of lava from volcanic vents is accom- panied by enormous volumes, not of smoke but of steam. The movement of a lava-flow down the mountain-side is marked by those clouds of watery vapor whose after condensation precipitates the heavy rains which are often more injurious to man than the desolation caused by the lava itself. The subsequent examination of the cold lava will show it to have a vesicular structure due to the myriad bubbles of superheated steam which it contained at the time of ejection. In this way, by the aid of the steam which forms a large part of their bulk, acid lavas are often as mobile and apr parently as fluid as those which have a basic composition. More- over, Daubree has proved that siliceous rocks which require a tem- perature of 2500° to 3000° F. before they undergo true fusion will yet become liquid at 800° F. if in the presence of superheated steam.

It is suggested, therefore, that the penetration of these slates and sandstones for some distance from the contact by small veins of granite was due to a condition of suspended solidification caused by the presence in the igneous rock of imprisoned steam, taken up by the intrusive mass of granite when at lower depths.'*' The granite, it is true, does not now exhibit a vesicular structure such as would or- dinarily characterize a rock whose mass has been interpenetrated by steam ; but it must be remembered that it cooled under the pressure of a great thickness of overlying formations and was extrudeii at so early a period in geological time that any such structure would long ago have been obliterated by those infinitely slow physical and chemical changes which are forever bringing about the " decay and repair " of rocks.

To the petrographer the granite presents no particular feature of interest It is of the normal type, consisting of mica, quartz and orthoclastic feldspar.

It was extruded at a period previous to the Devonian, but later than the Lower Silurian. The following facts warrant this state- ment. In the sections afforded by the road-cutting and railway- embankments at Big Hill, we have seen that it intrudes between and

8inoe tlie above was written I have read Rev. Osmond Fisher's Physics of Earth* 8 Crust in which the author argues in fa?or of the existence beneath the earth's crust of a liquid substratum of rock holding water-vapor in solution.

296 OniQIN OF GOLD-BEARING QUARTZ OF BENDIGO REEFS.

aoross the bedding-planes of the slates and sandstones. It must necessarily, therefore, be of later origin. Further, while the great series of the Lower Silurian rocks has been much folded and con- torted, yet this has not altogether obliterated the original lines of their sedimentation, and we find that when the strike of the slates and sandstones makes an acute angle with the locally irregular line of the contact, their dip abuts against the granite. Viewed as a whole, however, the general line of the contact is at right angles to the strike, and, therefore, to the axes of the folds of the Silurian rocks. This fact proves not only that the granite was extruded at a period following the deposition of the Silurian sediments, but also subsequently to that folding which is one of their most noteworthy features. The stratigraphical position of the granite is further in- dicated by the fact that elsewhere in the Colony its eroded edges are overlain by rocks known to be Devonian, which are neither pene- trated nor altered at the surface of contact. Fig. 8 represents a section at Mt. Hump Creek.

Fig. 8.

A. SILURIAN B. QRANtTE 0. DEVONIAN

Having determined the relative ages of the slates and sandstones in which the quartz-reefs occur, and of the granite with which the former come in contact, we have one other formation to consider. I refer to the lava dikes, incidental references to which have been fre- quently made in my previous papers on this subject. They form a marked and instructive feature of the Bendigo mine- workings. A few additional sketches will serve to illustrate further the characteristic behavior of these " lava-streaks,*' as the miners call them. This is well shown in the cross-section of the deep levels of the 180 mine. (Fig. 9.) The course of the dike, whose width averages about 9 inches, can be readily followed as it cuts through the successive sad- dle-formations. In Fig. 10 a dike (in another mine) is seen to tra- verse a sandstone, cutting through several small quartz-veins, until

From Geology and Physical Oeography of Vicloria, by R. A. F. Murray.

Origin Of Gold-Bearing Quartz Op Bendigo Reefs. 297

it reaches a bedding-plane, which it then follows. In Fig. 11 an- other lava-streak, following the structural lines of the country-rock, becomes " pinched out" in the plane of the section along which the sketch is taken. Doubtless it found an easier passage elsewhere. lu

>a4

Fig. 12 a larger dike, 14 inches wide, exhibits centers of decompo- sition marked by concentric formations of white zeolites. There is a dark band of less altered lava following the center of the dike. In Fig. 13 a lava-streak separates into two small branches before finally dying out.

298 Obiqin Op Gold-Bearing Quartz Of Bendiqo Reefs.

It requires but little observation underground to determine the recent origin of these dikes, which evidently followed a passage along those lines in the country-rock offering the least resistance.

The lava is of Tertiary age, either Pliocene or post-Pliocene. It is identical in lithological character with the basalt, successive sheets of which, in a manner much resembling that to be observed in Cali- fornia, overlie the Miocene and Early Pliocene gravel of the " deep leads" of the alluvial mines of Victoria. The gold-field of Ben- digo was probably also, at one time, covered by the lava extruded from the dikes ; and while all vestiges of such an outflow have since been removed by erosion, yet by a comparison of its structure, com- position and behavior it is easy to see the similarity existing between the rock which forms the " lava-streaks " of the deep workings of the mines and that of the basaltic plateaux overlying the rich allu- vium of the neighboring districts of Clunes and Heathcote, and the more distant ones of Ballarat and Ararat.

It is not easy to determine accurately such a rock as the Bendigo lava. Igneous rocks in the vicinity of ore-deposits are almost in- variably much altered, and therefore very diflScult of examination under the microscope. In the case of a basic rock, such as the one under consideration, these difficulties are very much increased. Upon examining a fair sample of the lava, taken from 2000 feet below the surface. Prof. Judd, F.R.S., at the Royal School of Mines, London, decided that, while now much alterel, it was originally basaltic in character and contained free crystals of olivine of con- siderable size. The augite and magnetite are abundant and well preserved, but the alumino-alkaline constituent of the rock has almost disappeared. Leucite, nephelite, or even melilite may have been present. The Germans would call such a rock "melaphyr," a term which Dana,* however, highly disapproves.

In the mines the lava is generally far advanced in decomposition. The olivine becomes converted into serpentine. The vesicles, often arranged so as to indicate that the mass has flowed, are filled with zeolites. At the surface, or within the domain of surface-waters, this alteration has of course progressed even further, and only white soapy clays remain to contrast with the black, compact, and homo- geneous rock of the deep levels.

One feature of the behavior of these dikes has peculiar interest, namely, their penetration through such an enormous thickness of overlying rock. In the 180 mine I have traced from the surface to

Manual of Mineralogy and Peirographyf p. 485.

Origin Of Gold-Bearing Quartz Of Bendigo Reefs. 299

a depth of over 2600 feet a dike only 9 inches wide. It is fair to assume that the dike must have found its way for a distance many times exceeding that for which it can be actually observed, for allow- ance must be made for the portions eroded since the Tertiary period, and also for that much greater unknown depth from which the lava came.

The question arises, did the lava effect its passage slowly or was it shot upward through its entire length and height in one instanta- neous operation ? The latter explanation requires the formation of a continuous fissure, itself unwarranted by the facts in this c&se, and the conception of its instantaneous filling with lava along its entire extent further presupposes that the material of the dikes remained molten during the brief time of such a performance. Again, the force which was able to shoot the lava upward through the tortuous fractures, penetrating many thousands of feet of overlyinvr rock, would also cause it, when it reached the surface, to be ejected with great violence and to a great height into the air.

Such an explanation is in accord with the catastrophic theories of the past, but it is opposed by the modern study of volcanic action, and does not harmonize with the facts as observed in the mines of Bendigo.

The accurate observation of volcanic phenomena by Scrope and Judd'*' has shown that the molten material issuing from the vent of a volcano usually wells up slowly. The violent paroxysmic out- bursts which do occasionally take place are essentially surface-occur- rences, and are due to a sudden relief from pressure obtained by the escape of accumulated superheated steam.

The course of the dikes and the effect of the lava upon the rock- surfaces in contact must be now consideredv Nothing in their mode of occurrence is more remarkable than thei/* tortuous and very irreg- ular passage through the overlying rocksfj They do not fill a clean- cut, continuous fissure; they rarely preserve a straight line for any great distance, but follow bedding-plane, joint and cross-fracture, as each in turn presents itself. Where their line of passage takes them across the quartz-reefs, the crossing is generally effected near the apex of the saddle. The conclusion arrived at is, therefore, that the |>enetration of the basalt through the slates and sandstones was not the work of a few seconds, during which it was shot upward instan- taneously through its entire height and length, but rather that it

As described hj the former in his book entitled VoloanoSf and hy the latter in his Voloanoes.

300 Origin Op Gold-Bearing Quartz Of Bendigo Reefs.

required a long time, and took place during a period when the Silu- rian rocks were subject to a tangential strain which caused their fracturing and thereby, little by little, oflTered a passage to the basalt, which \mng at the time under pressure, was seeking its way upward.

The effects produced upon the Silurian rocks, through which at Bendigo the basalt penetrates, and upon the granite and gravel, upon which in the neighboring districts it lies, are in each instance so slight as to be barely observable. The feeble changes produced upon the rock-surfaces in contact suggest themselves as due rather to the action of water than to the effects of heat.

It is, indeed, true that rocks having the composition of these ba- salts are fluid at a temperature low in comparison with that required to melt an acid lava; but observation of the behavior of the dikes, and of the effects produced upon the beds through which they pass, impresses one with the conviction that the temperature of the mate- rial filling them did not reach that which even the more basic of basalts require for their fusion. The conclusion forces itself upon us that the basalt of the " lava-streaks'' did not owe its mobility to a molten condition arising from intense heat. The government geologist, Mr. R. A. F. Murray, states the belief that the dikes were "in great measure the product of hydro-thermal action."* This is my view also, which, more definitely expressed, is as follows: The material of the dikes, when forcing its passage through the Silurian rocks, was more in the condition of a boiling mud than what we ordinarily imagine as the state of a liquid basalt; its mobility was due not so much to the fact that it was basic in chemical composition, and there- fore more readily fused, but because of the superheated steam of which it was full. Such imprisoned steam was superheated because originating at a great depth. It has been found that within the very restricted limits of human observation heat increases with depth. As we sink through the earth's crust the average increment of heat is at the rate of 1® F. per 47 J feet of descentf Water cannot re- main a liquid, in spite of increasing pressure, at a temperature above that of its critical point. This was first determined by Caigniard de la Tour to be 773° F. More recently, however, Battel li proved J

♦ Op. cU., p. 137.

t Sir William Thompson (Lord Kelvin) estimated it at 1" per 61 feet Prof. Prestwich collected 530 observations made in 248 localities, and selecting those onljT in which the necessary precautions had been taken, found the average to be as above stated, 1° per 47 feet.

X At Turin, in 1890. Slr.iuss gives the critical point as C. and A. Najedin as 358° C. This information I owe to Prof. Hallock and to Mr. Carl Bams.

ORIGIN Of GOLD-BEARING QUARTZ OF BENDIGO REEFS. 801

that it was 364° C, equivalent to 687® F. Such a temperature, taking the mean surface temperature to be 50° F., would be reached at a depth of about 30,000 feet. In areas under disturbance and in volcanic regions local conditions would tend to raise the temperature and consequently to diminish the depth required to reach the horizon where water can no longer remain liquid. A depth of less than 25,000 feet might then suflSce. In the gold-field of Bendigo we have to deal with a series of rocks over 35,000 feet thick. The basalt of the dikes has come up through that thickness, and consequently must have had its origin at a depth greater than that required to reach the temperature of the critical point of water. Such water as it contained must have existed, therefore, as water-vapor, which, as it approaches the surface, we may call superheated steam. This steam was in a state of compression, seeking to be relieved of its load, and striving to go where only that load could be lightened, that is, up- ward. The expansive force of such imprisoned steam rendered it a powerful agent in assisting the basalt to force its slow way through the crevices and fractures of the overlying rocks until relief was obtained at the surface. The lava would have lost some of its heat and moisture by contact with the rock-faces along which it passed, but having a temperature much above the lowest tempera- ture of hydro-thermal fusion, it could lose much heat without solidi- fication, and was therefore enableil to arrive at the surface where it probably welled forth water, steam and mud, overspreading the older rocks and the later gravel with one of those sheets of basalt which form so marked a feature of the surface geology of the Colony of Victoria.

The Silurian rocks form the Ultima Thule of all research into the geological history of the region. Of the rock- masses by the disintegration of which they were formed, no vestige remains. Their uniformity of structure and composition, their thin bedding and general regularity, their wide extent and great thickness, all prove that they were the sediments deposited during enormous periods of time from a comparatively shallow ooean. The frequent ripple- marks now observed at a depth of many hundred feet below the surface tell of estuarine seas and shallow reaches of water. The absence of any large bodies of conglomerate and breccia indicates a continuity of uniform conditions.

If the members of this great series of beds were deposited in water of a generally uniform depth then it follows that the bed of

302 Obigin Of Gold-Beasikg Quartz Of Bendjgo Beefs.

the Silurian ocean mast have undergone a subsidence which kept pace with the rate of sedimentation. The subsidence occurred along a certain line of weakness in the earth's crust, which later became the longer axis of a trough-like depression. The gradual deepening of that depression corresponded with the slow rate at which it was being filled.

As each layer of material became covered by another it began to undergo a change; as the superincumbent mass became thicker and heavier, the thin slime and the fine sand were consolidated by pres- sure and commenced to take to themselves the form and structure of the rocks which we term slate and sandstone.

When the area of depression had been loaded with a thickness of seven miles of sediment, subsidence ceased and elevation began. The existence of so great an accumulation of material determined the choice of this part of the earth's crust as the place of an eleva- tory movement The lowermost layers of sediment were subject to an enormous pressure whose effect was to cause condensation fol- lowed by contraction. This was the local cause which started the action of the tangential strains due to the general compression of the earth's crust. This compressive strain brought about the fold- ing of the Silurian sediments, it caused them to be gradually ele- vated and compelled the sea-waters covering them to recede slowly. Their upper portions became dry land, and erosion at once com- menced.

The forces still at work continued to exert a lateral compression upon the sediments which now had become rock-masses. The effects are seen in their highly developed cleavage and in the flexures and undulations which are to-day their most marked feature. Their cor- rugation made them stronger but less pliant.

Though bent and folded, the Silurian beds were not at this time broken by extensive fissures. This was prevented by their great thickness and by their comparative flexibility. That flexibility was due in the first place to the water which they still held over from the period of their sedimentation, and secondly to their fine grain and thin bedding.

The neighboring rock-masses, forming the rim of the original area of depression, were, however, unprotected by any such depth of overlying deposits ; they were older and more rigid ; and they became consequently broken by fractui'es and fissures through which volcanic agencies found a vent. It was then that the pi u tonic gran-

Obiqin Op Gold-Bearing Quartz Op Bendigo Reep8. 303

ite, perhaps, absorbing into its mass the lowermost portions of the Silurian beds, welled up toward the surface, giving out its heat as it advanced. The granite as it was extruded must have exerted a certain pressure against the slales and sandstones. It may have been the agent which caused those minor transverse undulations whose axes are at right angles to the main ore-bearing anticlinal axes of the Bendigo gold-field.

A period of comparative quiet followed. The Silurian rocks un- derwent denudation ; their upper portions were disintegrated ; from their dibria other formations were built up, to be in turn eroded and washed into the waters of the ocean. While the upper earth was undergoing change, the under world was also the scene of silent chemical and physical processes, removing matter here and laying it down there, destroying and upbuilding, ever shifting their centers of activity but never at rest. In the latter part of the Tertiary period the basaltic lava found its passage through the fractures traversing the Silurian rocks and covered the golden alluvium lying ui)on their eroded edges.

II. The Auriferous Quartz.

The origin of the deposits of gold-bearing quartz, belongs to no particular period. The agencies which brought the ore of the reefs to its present i>osition began to operate when first the sediments of the Silurian seas were laid down, and have continued until now.

In the reefs of Bendigo the two most important substances are, the metal, gold, and the matrix quartz. The other mineral con- stituents found in association with these two are relatively unim- portant. Their origin is a matter of much interest but it can be discussed apart.

At Mt. Tarrangower there is an inlier of sandstone which suggests this. For the following particulars I am indebted to Prof. Ulrich, of New Zealand : " The 8ilarian inlier lies about 16 chains away from the Granite-Silurian boundary, form- ing a small hillock about one acre in extent. It consists of a nearly black, dense, metamorphic sandstone, and its boundary with the granite can, in several places ronnd its circumference, plainly be seen, but there is no change in the rock other than that feldspar particles make their appearance near the boundary. However, in a water-course which cuts across the contact the dark Silurian rock can in one place be seen to run down into the granite ; and there it becomes gradually quite light-colored, % strongly feldspathic and also micaceous ; in fact, so altered that any geologist seeing a hand-specimen of it would no doubt call it a fine-grained granite.''

304 Origin Of Gold-Bearing Quartz Of Bendigo Reefs.

1. The Quartz.

The Silurian sediments contained a large proportion of sand, and the resulting rocks are very siliceous. Silica occurs in the rocks partly in a free state, as quartz, but a proportionally larger amount is found combined as a silicate. Free quartz is soluble in heated waters. The quartz combined in the more complex form of silicates has been demonstrated by Daubree and others to be readily dis- solved in hot water, more particularly superheated steam.

The formation of the quartz lodes has taken place in two stages, namely, leaching followed by precipitation. In both operations water was a necessary factor. It is the all-powerful disintegrating and transporting agent of the land-surface, where its chemical ac- tivity is intensified by the presence of dissolved carbonic acid; it is also, though in a different way, no less efficient underground where by reason of the prevalence of a higher temperature it becomes more energetic both physically as a force and chemically as a reagent ; and finally it is increasingly powerful at great depths where even the tremendous pressure of superincumbent rock-masses will not prevent it from being transformed into steam and developing a chemical activity to which the mineral constituents of the rocks can offer but feeble resistance. It was the agent which, from a state of general distribution, collected the quartz and brought it into the fractures and fissures, crevices and cavities where the miner now finds it.

All rocks, though compressed and dried as they appear, contain moisture.* All rocks are pervious to waterf to a varying extent by reason of capillary action, and all rock-masses are permeable because of the fractures and joints which traverse them. It has been shown by Daubree, Bischof and others, that only a very small quantity of water such as we find the rocks actually to contain is required to produce the most pronounced changes iu their chemical constitu- tion, particularly when aided by pressure, accompanied by high temperature.

The heat, in most instances accompanied by pressure, required to

make the water in the rocks intensely active, was afforded at various

times and for long periods. At the earliest stage of their history

Sey were subjected to high temperature. From the moment that

The French call it eau de constitution. They also have a term eau de carrihre, ich is the equivalent of our " quarry-water."

The amount of water in the rocks and their porosity has been measured by irrj Hunt, Prestwich, Delesse, and others. See Chemieal and Oeological E98ay$ y the first and Ocoloffy, Chemical, Physical and Stratigraphicalf by the second.

Origin Of Oold-Bbarxno Quartz Op Bendigo Reefs. 305

one layer of sedimeDt was deposited on the ocean bed and became covered by another it began to acquire a more elevated temperature in consequence of its increased distance from the surface of radiation.* As the thickness of the sediments kept pace with the sinking of the bottom of the depression in which they were laid, the successive beds acquired a temperature proportioned to their depth. The lowermost members attained a distance of over 35,000 feet from the surface. Though the increment of temperature be not constant with increasing depth yet the heat which obtained at a horizon seven miles from daylight must have been extremely high.

The great thickness to which they attained caused an enormous pressure to be exerted by the upper upon the lower members of the series. Pressure does not develop heat unless motion also occurs. When, however, at the close of the period of their deposition certain elevatory movements took place, their great thickness mut have helped to develop an energy which became converted into an exceed- ing heat This period of elevation and corrugation was long; and during its prevalence the slates and sandstones must have been sub- ject to the full play of those slow chemical and physical forces which, though less striking than violent volcanic outbursts and sufiden earth-movements, are yet the most powerful of the agencies which modify the rocks.

Afterwards, when the granite was extruded, further heat was difiiised through the adjoining slates and sandstones. In the long interval which separated the time of the extrusion of the granite from that which marked the ejection of the basalt, the Silurian rocks were subjected to slow movements of elevation and depression which, while not so energetic as those which took place at the beginning of their life-history, were yet sufficient to develop a rise of temperature. In later times, the period during which the lava penetrated the older ' rocks must have been marked by an increase of their heat. Though the extrusion of both granite and basalt have left more striking evi- dence of their occurrence than the slower movements referred to, which took place at other periods, yet the heat they afforded was not

An obseivatioD doe to Babbage, and quoted bj Daubr.

t See Rev. Osmond Fishera ofiKt Earth's Onutf 1889, chap. i.

t Le Conte, ElfmentsofOeology, p. 93, says, " the lower portion of sediments 10,000 feet thick wonld be raised to a temperature of about 260®, and of 40,000 feet thick to that of 860®." In the instance under discussion the eroded portions of the Silu- rian sediments are estimated to have been 750D feet thick, and their total original thickness to have been 35,000 feet.

VOI,.XXII.-20 Digmedby

306 Origin Of Gold-Bearing Quartz Op Bendigo Reefs.

SO widely diffused, and the effects produced were comparatively l<)Cal. The heat given out by both granite and basalt was however sub- sequently widely distributed.

This distribution was effected through their contact with under- ground waters which, becoming converted into steam, transferred some of their heat to other circulating waters, and these in turn in- jected their heat and steam through the pores and crevices of the slates and sandstones.

The hot water and superheated steam thus produced during long continuing periods were the restless agents in leaching the rocks, in dissolving out the silica from its state of combination and in after- wards transferring it to underground currents which bore it away until changed conditions compelled them to deliver it up by preci- pitation as quartz.

Thus, in this and other ways, the quartz became separated out through the pores of the rocks by a kind of sweating proce88,t to be segregated along the joints' and fractures traversing them.

The experiments of Daubrfee proved the action of superheated steam in dissolving the silicates and the subsequent precipitation, upon the lowering of the temperature, of crystalline quartz of a character similar to that found in association with the gold and other metals of ore-deposits.

The changed conditions compelling precipitation were very vari- ous in kind and due partly to physical, partly to direct chemical causes. In traversing the minute underground passages of the rooks, the hot solutions would meet with portions of loose or crushed rock, giving larger space, increased surface, and diminished pressure. This would favor precipitation. The workings of mines oflen afford illustrations of ore-deposits which owed their existence to such con- ditions. Similarly hot waters meeting and mingling with colder currents or passing between comparatively cool rock-surfaces would have their solvent power diminished, leading directly to the depo- sition of the material in solution. The hot springs of the present day in many parts of the world give us familiar examples of the precipitation of silica resulting from the lowering of the tempera-

One feels inclined to speak of "superheated " water, but there is no such thing. Superheated water, without reference to pressure, is steam. Geikie, though his writing is generally a model of accurate expression, slips into the use of this term on page 284, Text book of OtoUx,

t Or, as Daubr6e puts it : Le quartz a M foumi aux veines par uoe sorte d'exsudation de la roche encaissante."

Obigin Op Gold-Bbarino Quartz Of Bendioo Reefs 307

ture of the issuing waters. A third cause may be quoted. Tho le- ducing agency of organic and other matter is able to precipitate silica from its state of solution as a silicate. Of this, silicified wood is a familiar illustration.

The seas, the rivers, the thermal springs, and the underground waters of to-day carry notable quantities of silica in solution. Foreh- ammer found sea-water to contain silica, and in certain samples he determined the quantity to be as much as 3 parts in 1,000,009 parts of water. A cubic mile of the ocean would, at this rate, contain 13,500 tons of silica. Deville showed that the river Loire, at Or- leans, contained in 100,000 parts 13.46 of solid matter, of which 30 per cent, was silica. The geysers of the Yellowstone Park, those of California, Iceland, and the north island of New Zealand, all de- posit silica from the waters ejected by them. Steamboat Springs, in Nevada, and Sulphur Bank, in California, may also be instanced. The occurrence in New Mexico and Arizona of extensive areas which have undergone submergence, and whose forests have become pet- rified by the action of percolating siliceous solutions, affords a striking instance of the transference of silica by underground waters and its precipitation under favorable conditions. That the waters of mines contain silica in solution has been proved by thesilicification" of drift-wood, beautiful specimens of which have been found in tho auriferous gravels of the " deep leads" of, California and Australia. Such silicified wood has also been shown to be gold-bearing.

The concentration of quartz in one place rather than in another, its deposition between certain beds and along certain fractures rather than elsewhere, are the results more of simple physical conditions than of complicated chemical reactions, and are due to causes to be discussed in considering the structural geology of this district,

2. I%e Gold.

The waters of the ocean contain gold. In 1851 Malaguti and Durocher determined the occurrence of silver, but did not extend their inquiries into the question of the presence of gold in sea- water. This fact was first accurately determined by Sonstadt in 1872.t His experiments were not quantitative, but he stated, in

This, as poioted out to me by Prof. Le Gonte, is a double process, consistiog of the filling of the iDterstices by the precipitation of silica and the actual replacement of the woody fiber itself.

t *'0n the Presence of Gold in Sea- Water,'' £. Sonstadt, Chemical Nau, Oct 4, 1872, ., p. 169.

308 Origin Of Gold-Bearing Quartz Of Bbndigo Rebfb.

parenthesis, that the amount was certainly less than 1 grain in the ton."* More recently, however, Miinster found an average of 5 milligrammes per ton.f In endeavoring to arrive at an approximate estimate it must be remembered that local conditions, such as the temperature of the water, will affect the amount in solution. Son- stadt's researches were made with water obtained near Ramsey, in the Isle of Man, while Miinster got his from the Kristiania ford. In each case the sea-water was that of a northern latitude. In warmer regions it is probable that precipitation, due to the presence of putrescent organic matter, may diminish the amount of gold held in solution. Ijet us, however, take 5 milligrammes (equiva- lent to of a grain) as an approximation. This, though in itself a minute quantity, will be found to represent an enormous total amount of gold in the waters of the ocean. From the results ob- tained from the careful soundings carried out by the Challenger and similar scientific expeditions, it has been computed that the ocean has an average depth of 2500 fathoms, and that it contains 400 million cubic miles of water.§ This is equivalent to about 1 ,837,030,272,000 million tons, which, upon the basis of 5 milligrammes per ton, would represent 10,250 million tons of gold. By way of contrast, it may be added that, according to Soetbeer, Leech, and others, the gold-production of the world, from the binning of 149 to the end of 1892, a period of exactly four centuries, has amounted to only 5020 tons. The present output is equal to about 200 tons per annum.ll

The gold in sea- water is kept in solution as an iodide. 1[ The amount of free iodine present in the ocean is very minute but a large proportion of that element occurs combined as an iodate of calcium.ft From the results of a series of six experiments, Sonstadt found that a cubic mile of sea-water contains about 17,000 tons of

Sonstadt is often incorrectly quoted as having shown that sea-water contains 1 grain of gold per ton.

t The amount of silver was determined to be 20 mg. per ton. See " On The Pos- sibility of Extracting the Precious Metals from Sea-Water," Jouraa/ ij the Society of Chemical Indwiry, April 30, 1892, xi., p. 351 ; abstract from Norsk Tekniak Tidnkrift, vol. 10, No. 1. t See note in the Chemical News, Oct. 4, 1872, xxvi., p. 161.

Geikie's Text-book of Geology, 2d ed., p. 33.

n The colonies of Australasia, since 1851, have produced 2830 tons, of which fendigo contributed 375 tons.

T The Chemical News, October 4, 1872, xxvi , p. 159. rbid., April 26, 1872, xxv., p. 196. t Ibid,, April 26, 1872, p. 197.

Origin Of Oold-Bearing Quartz Of Bendigo Reefs. 309

iodate of calcium or 11,072 tons of iodine.* This represents the occurrence in the entire ocean of no less than 4,428,800 million tons of iodine.

The iodine which maintains the gold in sohition is obtained from the iodate of calcium. Gold is soluble in extremely dilute solutions of iodine, which under ordinary conditions are in turn readily reduced by organic matter. That the gold in the sea is not precipi- tated is due to the presence of the iodate of calcium in which it is not soluble but which, being readily decomposed by putreaeible or- ganic matter, liberates the iodine required to keep the gold in solu- tion.

There is reason to believe that the sea-waters of to-day contain much less iodine than those of former geological periods.f That there is so little free iodine in the ocean is due to causes parallel to those which bring about the noteworthy 'absence of carbonate of lime. Marine animals abstract the latter while marine plants absorb the former. How great is the work done in this way, is evidenced by the dimensions of the coral reefs and by the extent of the forami- niferous and other marine limestones.

The abstraction of iodine is no less striking. Sea-weeds, and more particularly those which grow at great depths, are the chief source of the iodine of commerce. When, after a storm, such sea- weeds are. cast upon the shores of Great Britain,§ France and Swe- den, they are collected and burnt, and from their fused ashes, termed kelp,'' the iodine is subsequently extracted by a simple chemical process. From 13,000 kilos of kelp about 100 kilos of sodium car- bonate and 15 kilos of iodine are obtained.

That iodine is not now so plentiful in the sea as during farmer geological periods has been suggested by chemical investigations into the composition of rocks. Certain sedimentary formations contain notable quantities of it. It has been found in some aluminous shales in Sweden, and also in certain varieties of coal and ixirf. The saline waters of several springs contain large amounts of it. Even rain-water has been known to give a recognizable iodine reactioa

♦ Ibid., May 24, 1872, p. 242.

t Thomas Sterry Hunt, Chemical and Oeologieal Eemytt p. 142. t It 18 also recovered from the nitrate of ('hili.

Partlcalarly the western bles of Scotland, and the west coasts of Wales and Cornwall. II The Pri$ieiple8 of Cheinietry, D. Mendeljeflf, vol. i., p. 490. f Fownes's Cftfmisfry, p. 185.

310 Origin Of Oold-Bearing Quartz Of Bendigo Reefs.

when tested, such iodine having been obtained by the agency of winds which have been blowing over certain areas of the sea where it was being liberated by the action of organic matter upon the iodate of calcium.*

Let us now return to Bendi|2. The conditions which prevailed during the time when the Silurian seas washed the earth's surface, were no less favorable to the solution of gold than those which ob- tain to-day. It is indeed probable that the waters of the palsBOzoie ocean contained more gold than those of the present geological period. When the sediments sunk to the bottom they carried with them, entangled amid the silt and sand, a large proportion of sea- water, most of which was subsequently yielded up as they were pressed down by the weight of the overlying deposits. The sea- water as it was rejected by the solidifying strata, left behind it a re- siduum which contained in a highly concentrated form the original constituents of the Silurian ocean. The sediments which fell to the ocean floor also contained portions and fragments of vegetable life which iu their decay served to decompose the iodate of calcium con- tained in the residual sea-water and so set free the iodine. This is a very active element and would at once form a fresh combination. The excess of iodine thus obtained may have served as a solvent for any particles of metallic gold which by mechanical means accompa- nied the sands laid down beneath the sea and which thus became added to that already derived from original chemical solution in the waters of the ocean.

The great body of slates and sandstones, not less than seven miles in thickness, was thus slightly but distinctly gold-bearing. It is so still ; the amount of the precious metal has remained practically the same and the changes which have occurred have merely brought about its less uniform distribution.

Ad the sediments were further consolidated, the vegetation which they originally containeil became disintegrated into its elements, and the iodine which it had abstracted from the sea-water when alive was now yielded up. The iodine thus derived became another fac- tor in the solution of the gold which was disseminated through the solidifying strata.

From a state of eve.i dissemination through great rock-masses to its concentration along certain lines of fracture which we call reefs, the gold arrived through the agency of the water still retained by

The Chemical News, May 17, 1872, xxv., p. 231.

Ic

Obigin Of Gold-Beabikg Quabtz Of Bendigo Reefs. 311

those rock- masses. The geological history of the district has shown that from various caases the heat required to make those underground waters intensely active was at hand for long periods. It is probable that the time of the extrusion of the granite was in this respect one of unwonted chemical activity. So also was the later period, when the lava dikes penetrated the overlying strata.

As an iodide, the gold readily circulated through the medium of the waters occurring in the rocks. We know gold to be readily soluble, even in extremely dilute iodide solutions. It would be wrong, however, to suppose that this or any other one chemical agency was the universal solvent for the precious metal. There is reason to suppose that in different formations and unaer diven>e con- ditions gold is soluble in varying combinations, all having a common want of stability. It has been suggested by Le Conte and others, and some recent discoveries* in the mines have tended to confirm the supposition, that the gold of certain lodes has )>een deposited from the solutions of the persulphate of iron. Alkaline sulphides have long been mentioned in the text-books as the chief agents in the dis- solving of the metallic sulphides, and very pretty chemical formulsB have been evolved to explain the resulting complicated reactions.f Occasionally gold is found free from association with metallic sul- phides; then Bischof's experiments are quoted, and its origin is put down to a silicate of gold soluble in alkaline waters. Again, gold readily combines with free chlorine, forming a salt easily soluble in water. The chloride of gold has, however, not been considered as likely to exist in nature because of the rarity of the occurrence of free chlorine and because of the instability of the compound which it forms with gold. The latter quality is not peculiar, but charac- teristic of all the salts of gold as we know them. Free chlorine does exist in nature, and though not frequently detected, this is not to be wondered at, seeing that its occurrence is usually associated with that of hot water or steam. Volcanoes, especially during times

Depoeits in the Quartzite Formation of Battle Mountain/' F. Gniterman, Proeeeding$ the Cohrodo Scientific Societjf, vol. iii., part iii.

t The supposition that gold is insoluble in alkaline sulphides requires modifica- tion. At the June, 1893, meeting of the Colorado Scientific Society, Mr. L. Q. Eakins reported the results of recent experiments, which show that fine gold, digested for four to five days in a cold, strong yellow solution of ammonium sul- phide, has a solubility equal to 2 per cent Weak solutions, with seven days' diges- tion, gave confirmatory results. Sodium sulphide also was found to be a solvent, but to a very slight dree only. Mr. Eakins is still prosecuting this series of in- teresting experiments.

312 Origin Op Gold-Bearing Quartz Of Bendigo Reefs.

of temporary quiescence, emit large quantities of hydrochloric acid,* and the rims of the volcanic vents, then called "fumaroles/' show incrustations of chloride of iron, chloride of ammonium, etc. The comparatively large amounts of such salts as exist in sea-water which were found at yesuvius,t was one of the facts used in support of the theory that the steam of volcanic eruptions came from the ocean. That these and other compounds, including hydrochloric acid, owe their origin to free chlorine, and were formed in the presence of hot water and steam, is most probable. Therefore it would seem well to add the chloride to the other probable combinations in which gold may circulate ynderground.

In the great laboratory of nature, the chemical changes which take place in the long periods of time with which we have to deal must be of an infinite variety, and therefore no number of chemical formu> IsB can represent the reactions which have followed each other from the moment when the gold was liberated by the disintegration of some pre-Silurian rock to this later day, when the miner finds it en- closed in the quartz of the reefs. Therefore, while it is suggested that as an iodide it may have travelled through the waterways of the Silurian rocks, it is not meant that in its long joumeyings before it arrived at the place where it is now found, it did not pass through many changes of chemical combination.

The precipitation of gold from solution can be brought about in many ways. . More than twenty-five years ago certain members of the Greological Survey of Victoria carried out a series of experiments which yet remain our main source of information on this subject. In 1864, an accidental discovery by Richard Daintree led to the de- termination by him and his colleague. Prof. George H. F. Ulrich, of certain conditions under which the precious metal is deposited from solution. It was demonstrated subsequently by a number of care-

♦ Judd'8 Voleano€, p. 213.

t Palmieri's VesuviuSf p. 121.

I The following interesiiog account of what has become an historical event was given to me by my distinguished friend, Prof. Ulrich, now of the University of Otago, New Zealand, and with his permission, I reproduce it here:

I was engaged in the laboratory analyzing a seolitic mineral, whilst Daintree himself was busy with some photographs. Suddenly he made ejaculations of aston- ishment, and on my asking the reason, he showed me a small, common medicine- bottle, which contained (more than half full) a water-clear fluid in which floated a part of the cork. The cork remaining in the neck of the bottle was acid-eaten. At the bottom of the bottle there was a large speck, or rather little nugget, of bright gold. He explained that he had made a concentrated solution of chloride of gold for the purpose of toning his photographic prints. He had placed in it a small

Origin Of Oold-Bbarikg Quartz Of Bendioo Reefs. 313

ful experiments made bv Wilkinson,' that organic matter, in any of the ordinary forms in which it is found to occur in the alluvial drifts, is readily capable of precipitating the gold of even the weakest solutions, and of depositing it as a thin metallic coating upon other particles of gold or upon any of the sulphide minerals, such as py rite, galena, or blende, with which its occurrence is usually associated.

Several years later, in 1871, Daintree was in Loudon and at Dr. Perc/s laboratory, at the Royal School of Mines, commenced another series of experiments. In different bottles he placed solu- tions of chloride of gold in strong solutions of chloride of sodium, containing a piece of pure rock-salt, and to each he added a crystal of one of the several metallic sulphides commonly found in gold-ores. Four years later, and before any results had been obtained, he died. The chemical reactions which he had set to work were, however, pro- ceeding slowly but surely. I now quote Mr. Richard Pearce, who was an eye-witness of both the binning and the consummation of these experiments.

" When Daintree died, one of the bottles, namely, that containing a gold solution and a crystal of common pyrite, was removed to Dr. Percy's laboratory in Glouces- ter Crescent, and there, in 1886, the work which Daintree had begun came to frui- tion. On the smooth surface of the crystal of pyrite there had been deposited a cluster of crystals of gold.''

It had taken fifteen years to obtain the result, and the man who started the investigation did not live to see his work fulfilled ; but what, it may well be asked, is the brief time of even a generation when compared to those vast SBons during which Nature, in her greater laboratory, is carrying out operations similar to these?

It is interesting to be able to add as a sequel to the above that in the ore of the St. Ijouis mine, in Gilpiu county, Colorado, crystals of pyrite have been found which, while in themselves non-aurifer- ous, were yet dotted over with clusters of crystalline gold remarka- bly similar to those obtained by Daintree's experimentf

The persulphate of iron, which is also a vehicle for the removal of gold from one place to another, is reduced by organic matter,

speck of gold. ,' And now see (turning the bottle upside down), the speck has grown to such a size that it won't go through the neck of the bottle.' We both agreed that the organic matter of the cork had been the cause of the decomposition of the solu- tion and the growth of the gold."

The late Mr. Charles Wilkinson, Government Geologist of New South Wales, at that time junior assistant in the Geological Survey of Victoria.

t My authority is again Mr. Richard Pearce.

314 Origin Of Gold-Bearino Quartz Of Bendioo Reefs.

with the formation of gold-bearing pyrite. Under like conditions silicate of gold solutions would precipitate their gold. Again, me- tallic sulphates or alkaline sulphide waters containing metallic sul- phides in solution would, on meeting alkaline carbonates in the presence of organic matter, be compelled, by reduction and neutrali- zation respectively, to form a deposit of metallic sulphides.* Finally, iodide solutions are readily reduced by organic matter or by ferrous sulphate, while the presence of both together would lead to the formation of an auriferous pyrite.

We need now to recall the fact that the original sediments con- tained imbedded with them numerous remains of the organic life of the ancient seas, many of whose forms were subsequently preserved in the rocks as the fossils which now enable us te determine the stratigraphical position of the slates and sandstones. Other rem- nants of the more minute organisms such as abounded in the waters of the ocean, though they left behind them no recognizable forms, yet were doubtless mixed among the silt and mud which fell to the bottom of the Silurian seas. Such organic matter in process of time became decomposed into its constituent elements, those which were soluble being removed by underground waters while the insoluble remained to undergo further decomposition. In this case, as often happens in the laboratory of Nature, the solvent and the precipitant formed part of the same substance, whose disintegration liberated the one from the other, both, however, to meet again and to react upon each other at a later period. Thus the iodine, which is a sol- vent for gold, was set free while the carbon, which is a precipitant, remained in the residue and became converted into the graphitic material which darkens the rock encasing the quartz lodes.

This graphitic material was the precipitant for the gold solutions. Its occurrence in certain lodes has been pointed out by othersf but its very frequent association with gold quartz has not been fairly recognized. In the main auriferous belt of California, passing through the counties of Amador, Calaveras, Tuolumne and Mari- posa, an encasing formation of black slate is known by experience to be a favorable indication, while a gray or greenish-gray country- rock is considered less encouraging. At Amador City, in the county of the same name, I have seen men coming up from underground with faces and hands all sooty black by reason of contact with the

Le Conte*8 EUmenU of Oeology, p 245.

t As, for instance, by Sandberger, in connection with the veins of the Enge- birge.

Origin Of Gold-Bearing Quartz Of Bendigo Reefs. 315

graphitic slate which there and elsewhere in that rion encloses the quartz veins. In New Zealand, more partionlarly in the province of Otago, the gold-bearing lodes are similarly characterized by a black selvage, the somber tint of which is due to carbonaceous ma- terial. In the mining districts of the continent of Australia, as well as in the gold-fields of the island of Tasmania, the slates and metaraorphic schists, which so generally form the prevailing country- rock, are dark, and the clay which lines the walls of the reefs is black and graphitic. Such is the case also in the mines of Bendigo. It is not, therefore, necessary to go far to find an -agent capable of precipitating the gold from its state of solution in the underground waters of the rocks.*

That gold does occur in solution in the underground waters of to- day has been shown by the evaporation of large quantities of ordi- nary mine-water and the finding of gold in the residuum,t also by its occurrence in the incrustation of boilers using mine-water. It has been proved by the examination of old mine-timbers left in abandoned workings and the discovery in the decayed and often silicified wood of crystals of pyrite which were distinctly gold- bearing.§ Furthermore, it may be added that in the gravel of the deep leads, at Ballarat, for example, there has been found driftwood yielding assays of from a few pennyweights to several ounces of gold per ton. II

The occurrence of the metallic sulphides, such as arsenical and ordinary iron pyrites, galena, and blende, the minerals most com- monly found at Bendigo in association with the gold, brings up a large field of conjecture. SandbergerTf showed that iron, zinc, lead, copper and other metals occurring in lodes can also be found in cer- tain silicates common to the crystalline rocks, such as olivine, augite, hornblende and mica. Ijis deductions have not been, however, en- tirely accepted. At Bendigo the lode-formation, though evidencing

How far electro-chemical and electro-magnetic forces may have aided theee re- actions cannot be estimated. That their took some part is highly probable.

t First proved by Daintree.

X Also by Daintree, at the mines at Maryborough.

i Notes on the Physical Geography Qedogy and Mineralogy of Vietoria by A. B. C. Selwyn and G. H. F. Ulrich. Melbonme, 1866.

y Investigated by the above-mentioned, and also by J. Cosmo Newberry, analyst to the Geol. Survey of Victoria.

f Engineering and Mining Journal, March 22 and 29, 1884, xxxvii., pp. 218 and 232. Translation of first chapter of Untersuchungen iiber Erzgange,*' by Fridoiin Sandberger. Wiesbaden, 1882.

316 Origin Op Gold-Bbarino Quartz Of Bendigo Reef3.

metamorphic action, is yet sedimentary. It may, perhaps, be sag- gested that the granite contributed the metallic sulphides associated with the gold and quartz, but there is no evidence to support such an explanation. While the frequency of the occurrence of ore- deposits in association with eruptive rocks is a fact now widely re- cognizedy yet it has given rise to generalizations not altogether war- ranted. In this particular district the contact is not a place of ore- deposition, and the main series of producing mines is seven miles distant. There are, it is true, certain anticlinal formations of quartz not far from the granite; but they appear to have no relation to the contact, and they are not economically of any importance. The geological evidence of the region has suggested that the granite was a factor in the process of ore-deposition, but it does not indicate that this rock was the source of the gold or of the associated sul- phides.

In this connection I would hazard the remark that the near neighborhood of igneous rocks is favorable to the occurrence of ore- bodies, not always or necessarily because such rocks were the origin of the precious metals which were leached out from them, but often because the extrusion of such eruptive rocks afforded the heat and steam which gave an intensified chemical activity to percolating solutions.

The Silurian sediments were obtained from the erosion of pre- existing rocks, whose disintegrated particles probably contained the material required to form the sulphide minerals of the reefs. Whether the metals were dissolved in the waters of the palsBOzoic seas and were subsequently chemically precipitated, or whether they were deposited by mechanical agency among the silt and sand which fell to the ocean bed, is a question not now to be determined.

In the reefs of the Bendigo gold-field t|ie metallic sulphides are present in an unusually small proportion. Iron pyrites is much the most abundant. A great deal of the richest ore of the deepest workings of the mines is a clean white quartz, almost entirely free from any accessory minerals. Such quartz frequently contains the gold in an extremely coarse form, in pieces weighing many penny- weights.* The presence of pyrite is not necessarily an evidence of the poverty or richness of the quartz. Rich ore-bodies often con- tain a good deal of it, just as poor ones sometimes do. Other min-

At the Lazarus Mine, at a depth of over 2000 feet, I saw pieces of gold ex- ieding an ounce in weight in a large reef of clean, white, splintery quartz.

Origin Op Oold-Bearinq Quartz Op Bendioo Reep8. 317

ing districts afford a similar experience. We must recognize that the sulphates of iron act very differently with an increase or de- crease of the oxygen they carry. Ferrous sulphate, FeO, SOj, is a precipitant* for gold solutions, while ferric sulphate, FeO,, 3 SO.f a solvent. May not this fact help to explain the irregularity of the phenomena attending the presence of pyrite in gold-ores?

The limitations to the length of this contribution must prevent the discussion of theories explanatory of the process by which the gold and quartz were actually brought to the place where we now find them. Modern teaching and experience has advanced that theory of ore-deposition which has been, not at all happily, called lateral secretion." It is to the effect that the origin of the pre- cious metals as found in lodes is to be ascribed to the immediately encasing rock, out of which they have been leached by solutions, which afterwards by endosmotic flow, penetrated the walls of the fissure and there deposited the gold and silver. Such an explanation is, I respectfully submit, more in harmony with the teachings of a professor's laboratory than with the testimony of underground ob- servation. No theory so narrow as that framed on a phenomenon as specific as endosmosis can live in the air of the mines where the modes of occurrence of the ore have an infinitude of variety which even the most general of explanations can hardly hope to cover. As our stock of ascertained facts slowly accumulates, as each dis- tant mining district sends in its quota of recorded observations, we shall, I believe, find that " lateral secretion in its narrowest mean- ing is rarely tenable, that is, that while the material of ore-deposits may have been and probably was derived from the leaching of the rocks, it did not necessarily or often come only from those which are immediately adjacent to the walls of the lodes. In the case of the Bendigo saddles " we are led to believe that the gold and the quartz were derived from the mass of the surrounding formation rather than from that small portion only which immediately adjoins the reefs. In the process of ore-deposition and lode -formation this gold-field was a part of a greater area, in which similar phenomena of segregation! took place, an area which included nearly all of the

It is the precipitant employed in the chlorination-works of California and else- where. Charcoal is also nsed for the same purpose.

t The formula is not rigid. The so-called seAqui-sulphaten of iron are of vari- able composition.

t This word " segregation " — the separating out of material and its regathering together elsewhere — best covers the process of lode-formation in this particular region.

318 ORIGIN OF QOLD-BEARIXG QUARTZ OF BENDIGO RE£Fd.

Fio. 14.

Drawing of Bendigo made in 1851.

Origin Op Gold-Bearing Quartz Of Bendigo Beefs. 319

Domerous and productive gold-mining districts of the colony of Victoria. Bendigo was especially favored because of the very pe- culiar structure of its beds of slate and sandstone.

It may be objected to the general explanation which has been offered with regard to the origin of the gold and the quartz that if it be accepted as true, all sedimentary rock formations should be equally capable of profitable exploitation. The answer is obvious, namely, that to the miner the mere dissemination of gold in rocks is a fact economically of no importance, since it is only by its concen- tration in certain quantities and in certain forms of ore-deposit that it can repay him for the toil and expense of its extraction. In the Silurian rocks of the Bendigo district conditions obtained and agencies were at work which were particularly effectual in collecting the gold from its wide and even dissemination to its concentration in the reefs.

The richness of the gold-field is owing primarily to two causes, the structure of the country-rock whose extreme and very regular folding gave rise to unusual facilities for the percolation of under- ground waters, and the occurrence in the slates and sandstones of a precipitant able to compel the de|)osition of the gold.

It is a striking fact that while the anticlines necessarily alternate with synclines, the latter are not the places of ore-deposition. " In- verted saddles," as the miners call them, do indeed occasionally occur; but they are not economically of any importance. The ex- planation is a simple one. By the mechanical principle of the arch the anticlinal structure tends to preserve a passage for mineral solu- tions, while on the contrary the basin or trough, formed by the synclinal arrangement of the beds, tends by the action of gravity to become closed. It is not probable that the apex of the anticlines was marked by an open way ; but weare ju8tifie<l in supposing that along the anticlinal axes there were portions of rock more loose and more permeable than the country surrounding them. These we may liken to arched canals through whose long passages mineral solu- tions have circulated from pre-Devonian times till now, bearing with them that golden freight which by reason of the reducing action of the carbonaceous matter lining their walls they were com- pelled little by little to lay down.

Thus we arrive at a stage when from wide and uniform dissemina- tion through enormous rock-masses the gold and the quartz have become concentrated along certain lines and in certain localities, but nature is never at rest; they are no sooner laid down than they be-

320 Origin Op Oold-Bearino Quartz Of Bbndioo Rebfb.

come again wanderers throagh the andergroand waterways. The caases operating to remove them from one place and to concentrate them in another are forever at work. We speak of secondary deposi- tion in cases where we think we clearly recognize the removal of metallic ore from one place to another; bat as a matter of fact all the ore-deposits of the mines are concentrations, and in their nature secondary, from the period when their constituent parts formed a portion, relatively large or infinitely small, of the first sediments laid down npon the floor of the ocean to that time long afterward when the pick of the miner disturbs that which for ages has been going through a process of continual change and evolution.

Thus the silica which as fine sand fell to the bed of the sea, in process of time united with other elements and became a part of a complex mineral which we call a silicate. That silicate, by the rever- sal of the reactions which had brought it into existence, became subse- quently disintrated, the silica was set free, and as quartz later on became the matrix enclosing the gold. From that state of admix- ture it was separated by the contrivances of man or by the less noisy and more powerful agencies of heat and cold, wind and rain to be swept into the running stream which carried it into the waters of the ocean in whose silent depths it was destined again to 'sow the dust of continents to be."

Similarly the Silurian slates and sandstones may have been them- selves derived from the erosion and disintegration of the granites, other portions of which afterwards intruded among them. Or, again, there is reason to believe that the granite may have been formed by the extreme metamorphism of the lowermost members of the Silurian series. Upbuilding and disintegration in the mineral creation like life and death in the organic world, are but correlative parts of one continuous process. They are different aspects of that indestructibility of matter and transmutation of energy which are taught no less by the formation of a quartz reef than by the unfold- ing of a flower.

At that point in the long sequence which marks the present time, the eye lingers on rolling woodland and winding road, grassy meadows and fleecy flocks, glancing from the busy activity of the railway to the peaceful quietness of the farm, to be finally arrested as it catches the gleam of the Bendigo mines, white islands in the dark blue sea of the lovely Australian bush, that vast forest of EvxicdyplxiB whose leafy waves have replaced the watery wastes of palsBozoic times.

'ig. 14 is a reproduction of a drawing of the Bendigo field as it

Amebicak Improvements And Inventions. 321

appeared in 1861. The upright mass of quartz in the foreground is the " west leg of a saddle- formation, while just beyond the two pools of water (now ornamental lakes in the gronbds of Mr. Greorge LanselPs residence) there is seen an actual saddle outcropping at the surface.

Summaby Of Amebican Impbovements And Inven

Tions In Obe-Cbushino And Concentbation,

And In The Metallubqt Of Coppeb, Lead,

Gold, 8Ilveb, Nickel, Aluminum, Zinc,

Mebcuby, Antimony And Tin,

BY JAMES DOUGLAS, NEW YORK CITY. (Chicago Meeting, being part of the International Engineering Congress, August, 1898.)

Introduction.

American metallurgical inventions have not always been abso- lute metallurgical improvements, if accurate work be the standard of comparison; but when we review the new methods and machinery which have been generally accepted, we find that some have become essential to the metallurgist the world over, while all are so admir- ably a<lapted to the peculiar requirements of our local conditions that they have entirely supplanted foreign processes and appliances. The necessity of diminishing to the utmost the employment of labor (owing to its great cost), and the inexorable requirements of our modem joint-stock corporate system, which demands that large quantities be worked up in order to secure large returns, regardless betimes of waste, have been the impelling motives to invention, and indicate the special directions in which originality has been exer- cised. As might be anticipated, the West ha& been the most fen tile field of inventions by miners and metal lurgistoy by reason of its great and varied mineral wealth, locked up in ores sometimes lean and often of intricate composition, which must be profitably handled, at a great distance from fuel and from building-materials, by workmen who are paid wages very much higher than those re- ceived by their fellow- work men of like craft in Europe, and in some localities twice as high as those paid for like services even> in the eastern States.

Moreover, paradoxical as it may seem, the technical inexperi- ence of many of those engaged in mining and metallurgy has helped rather than hindered the progress of invention ; for while there may have been ignorance, there has been radical fueedoin from

VOL. XXII.— 21 . ... ., vjOOgle

322 American Improvements And Inventions.

prejudice and from prepossession for time-honored or, in western estimation, time-worn methods ; and, therefore, thongh innumerable utterly impracticable plans have been proposed, tested,' and rejected, others, which would have been scanned very suspiciously by well- trained metallurgists, have been adopted and not found wanting.

In the very brief survey of so wide a subject, which of necessity will be made, all reference to the metalluiy of iron and steel will be omitted, and little more than a list of the more important specifi- cally American improvements will be given under the following heads :

I. — Machinery for Coarse Crushing and Fine Grinding. 11. — Machinery for Mechanical Concentration. III. — Furnaces for the Calcination of Ore. IV. — Modifications of Cupola Furnaces and Cupola Smelting.

V. — Modifications of Reverberatory Smel ting-Practice. VI. — New and Modified Metallurgical Methods for the Treatment of Ores of (1) Copper; (2) Lead; (3) Gold and Silver; (4) Nickel; (5) Aluminum; (6) Zinc; (7) Mercury; (8) Antimony; and (9) Tin.

I. — Crushing Machinery.

In no department of metallurgy have more decisive modifications been made than in appliances for the coarse and fine crushing of ore; and this country may claim credit for well-nigh every important modern improvement in machinery for these purposes.

The Blake, Comet, and Grates crushers are American inventions, while the Ball stamp, using the direct impact of a stamp-head, driven by steam, and the California revolving drop-stamps, in their present perfected form, are products of America.

The Blake crusher was patented in 1858 by Eli Whitney Blake, of New Haven. It is constructed on the principle of the nut- cracker, a movable jaw breaking the ore against a fixed jaw. In the Blake, as now constructed, the movable jaw is at its maximum dis- tance from the fixed jaw, where it hangs from two pivots. Motion is communicated from an eccentric to the free end of the movable jaw by a pitman or vertical connecting-rod and toggles.

JE)iirerent mechanical contrivances have been employed to produce this motion. In the original Blake, a toggle-joint was used, the lower end of the movable jaw resting on a fixed point, and motion being communicated to its upper free end.

To insure a smaller and more uniform size of particles in the crushed material, the Dodge crusher still communicates motion to

American Impbovements And Inventions. 323

the apper extremity of the movable jaw, which oscillates on fixed pivots projecting from its lower extremity.

In the Foster crusher, the movable jaw has a grinding motion ; and other crushers differ from the Blake proper in the means em- ployed to move the jaw, and in devices for reducing the strain on the frame of the machine by throwing it on the foundation. The Robinson-Rea Co., of Pittsburgh, manufactures the Blake Challenge crusher, one commendable feature of which is a frame of structural iron. But whatever their differences, all have in common a fixed and a movable jaw.

This is true of another type of breakers, the Comet and Grates, whose movable jaw is a cone, — either suspended, as in the Comet, made by Fraser & Chalmers, or supported on a step, as in the Gates, — which oscillates with a gyrating motion in a heavy cup- shaped mortar open at top and bottom. The movable jaw, there- fore, in this type of breaker, is circular, and crushes all around its periphery, whereby not only is the crushing surface increased, but the strain upon the jaw is equalized.

The next important improvement in crushing-machinery was the perfecting of the steam-stamp, which had been proposed in Europe by Jordan in imitation of the Nasmyth hammer, but was first practi- cally built by W. Ball. Ball is said to have erected his first mill at the Copper Falls mine on Lake Superior in 1855, and there to have driven his stamp by steam employing the rudiments of the gearing still in use, but discharging by only one gate from, the mortar. At the Pewabic mill, built in 1860, Ball, it is said, added an additional gate. As now constructed, the mortar discharges through four gates, and the machine has been so improved in every detail, that one stamp- head, moved by a 20-inch cylinder, with a 30- inch stroke, will crush 280 tons a day coarse for concentration, or 160 tons fine for amalgamation. In those last erected the founda- tion, instead of being elastic, is as rigid as metal will make it, with the result of a gain in yield.

The so-called California stamp originated also, it is said, in the Lake regions, the first battery with revolving stems, which is its es- sential distinction having been put up early in the sixties by Gates, of Chicago, in the Quincy mill. The mortar of the old mill wiis of the original English style, the newer type of mortar being of California introduction.* The revolving stamp, however, has not proved ap-

For this ioformation I am indebted to Mr. J. F. Blandy, whoee contribution to ▼ol. iL (18T3), p. 208 of the Trans, A. I. M, K, is the first of the series of papers

324 American Improvements And Inventions.

plicable to the crushing of Lake ore. In the Central mill, the only large mill that has not introduced the Ball stamp, the up-and- down-drop stamp is still used.

For western mills there are two patterns of stamps now built, designated as the California and the Colorado. They differ in the height of the stamp, the height and speed of the drop and the level of the discharge. The California stamp-head and stem weigh usually from 700 to 800 pounds and drop ninety times a minute. The pulp is discharged from both sides through screens set low in the mortar. In the Colorado mill, the stamp weighs usually from 400 to 500 pounds. The drop is slow, thirty to forty times per minute, and the discharge is set high, so as to permit large amalga- mating plates to be secured to the inner walls of the mortar. Each pattern has its advocates, but the Colorado is less generally used than the California. Where concentration follows amalgamation it is objected taas producing more slimes;: and coarse crushing cannot be effected by it.'*' Of the many mills which have sought the favor of the public, as a substitute for the stamp,, the Huntingdon, a well constructed roller-mill, in which amalgamation may be effected, and the Sturtevant mill fw fine dry crushing, have alone met with gen- eral acceptance.

Though no cardinal change has been made in the construction of rolls, those made by Grates, of Chicago which combine under one housing rolls, screen, and elevator, and the splendidly built steel rolls of the Krom pattern, are good examples of American design and workmanship. The tendency in this country, now that open- hearth steel is used for the tires, is toward high speeds, running up to 170 revolutions a minute..

II. — Concentrating-Ma€hinery.

American metallurgists have devised few important improvements in machinery* for the mechanical concentration of ore, if we except the contrivances for concentrating gold by so-called hydraulicking. This method depends on the powei of water, in large volumes and under high pressure, to break down and disintegrate even firmly

in the TranaacUons on t]ie Ball stamp. In 1856 Carr suggested an air-spring or cush- ion in working the steam-hammer, the same device now employed in the Hushand stamp built by the Hayle Company, Cornwall, and resembling the atmospheric stamp used in the old Phcenix mill, Lake Superior.

Raymond on The Relation Between the Speed and Effectiveness of Stamps," Trans. A. I. M. E., i., 40 ; Rogers. Mines and Mills of Gilpin County, Colo- rado," Ibid,, xi., 29. 1

AlfEBICAN IMPROVEMENTS AND INVENTIONS. 325

cemented banks of gravel and sand, and concentrate the minate quantities of gold which the gravel contains. The machinery con- sists of pipes of sufficient size and strength to convey water to a nozzle so constructed as to be directed by man-power against the bank which is being attacked, and of a system of flumes, laid with a proper dip and floored with riffles, through which the rush of water carries the smaller rocks, gravel and sand, while the gold set- tles and is arrested and amalgamated by the mercury in the riffles. This method was first applied in a primitive fashion in Michigan City, Cal., in 1855, but now, as Bowie says, a 40-inch wrought- iron pipe has been substituted for canvas hose and a stove-pipe, and an inch-stream replaced by a river of water discharged through a 9-inch nozzle under 4C0 feet pressure." As a result gold is recovered from very favorably situated gravel-beds at a cost of less than three cents per cubic yard ; and until prohibitory legislation forbade the discharge of the debris into the present river-beds, $12,000,000 of the precious metals were thus extracted annually from the old river- beds of California alone.* Ten centaper cubic yard (1 J to If tons) is a profitable yield from well situated gravel-banks. Extreme in- stances of what can be accomplished are given as follows by J. Hays Hammond in his article on The Auriferous Gravels of Cali- fornia/' page 113, NinUi Annual Report of the California Stale Mineraloffid :

''The old washings of surface-ground near Malakoff, Nevada county, from 1870 to 1874, are estimated at about 3,250,000 cubic yards, the yield of which was about cents per cubic yard. The Bloomfield company from November 29, 1876, to October 13, 1877, washed 1,591,730 cubic yards of top-gravel which yielded cents per cubic yard.*'

This is the roost rapid and wholesale system of concentration ever devised. The steam-shovel is applied by the Bucyrus Company to the rapid handling of auriferous gravels in localities unsuited to the hydraulic method.

The mechanical concentration of crushed ore has been carried out with greatest financial economy in the Lake rions — I do not say with the greatest economy of mineral. There the ores of native copper are so lean, that in most instances treatment, to be profitable, must be automatic, and the quantity of ore handled very large.

Barchard's Report an Production of Gold and Silver in the United States, 18S3, p. 716; Blake*8 Geological Beconnaisaanee in Califomia, p. 265, 1858; Bancroft's Hit tory tf OaUfomia, vol. xvii., p. 645.

326 Amebican Improvements And Inventions.

Hence large volumes of water are used, sufficient to carry the crushed material through the grating of the Ball stamps to the jigs, and to trausport the final and intermediate products of the concen- trators to their destinations without handling.

The Ball stamp, referred to elsewhere, is used in all the Lake mills, and has replaced rolls in the larger Montana concentrating-works, not only on account of the small space it occupies, in proportion to the work it does, but by reason of the rapidity of repairs and of re- placement of wearing parts.

In the Lake mills, and wherever Ball stamps are used, the stamp discharges the ore to hydraulic separators ; trommels and all other mechanical classifiers having been abandoned. Thence the ore is carried by the powerful current of water to jigs and buddies. On the Lake, the CoUom jig is largely used. In it the pitons of adjacent compartments receive a sharp stroke from a rocker and are returned to position by the recoil of a spring. But the Hartz jig, with its eccentric motion and multiple compartments, is more popular. The Evans table, which is the accepted slime concentrator in the Lake mills, is a modification of the German round revolving buddle.

I extract the following items from the summary of results re- ported by the Atlantic Mining Company of Lake Superior for 1892 :

Rock stamped, 300,900 tons.

Product of mineral, 5,02S,560 pounds.

Product of refined copper, 3,703,875 pounds.

Yield of rock treated, 12:31 pounds per ton or, 0.615 per cent.

Gross value of product per ton of rock treated, . . $1.4645

Cost per ton of minings selecting and treating,

and all surface-expenses including taxes, 8398

Cost per ton of transportation to the mill, 0333

Cost per ton of stamping and separatiug, 2)09

Cost per ton of working expenses at mine, 1.1240 Cost per ton of smelting and marketing product,

including N. Y. office expenses, 1767

Cost per ton of running expenses, 1.3007

Total expenditure per ton of rock treated, 1.3351

Mr. Stanton, whof excellent management has made the Atlantic mine and mill the standard of comparison on all questions of economy, gives me the following additional particulars as to his mill- practice : " With regard to our stamp-mill practice, the following statement 'ers what we do at the Atlantic mill, viz.: 6 heads steam-stamps, inch diameter of cylinders, 26-iuoh stroke, carrying 100 pounds

Ic

American Impbovements And Ikventi0N8. 327

steam-pressnre. Capacity soraetbinp raore than 1000 tons (2000 ponnds) in 24 hours. The force employed is: 1 superintendent, 10 stamp-feeders, 2 head-runners, 2 mill-runners, 4 firemen, 2 ma- chinists, 1 blacksmith, 2 carpenters, 6 laborers and spare hands, 1 sweeper and lamp-cleaner, 3 wash-bosses, 18 wash-boys, 1 cooper re- pairing and heading mineral-barrels— 52 hands.

'' The necessity for two carpenters arises from the large amount of repairs on a wooden flume, miles in length, which conveys the wash-water to the mill and on the long launders carrying off the tailings.

''The cost of the entire treatment of the ore in the mill, including repairs, taxes and insurance, is about 26 cents per ton of ore stamped. You are no doubt Familiar with the method of separation employed in our mills, which is so often commented upon ; but it may not be amiss to say that the wash-water carries the fine stuff from under the heads to the V-separators, the slimes passing to the slime-tables for treatment and the sands to the jigs. We employ 100 jigs (200 sieves) and 16 slime-tables, all of which are automatic in their action. The various processes require from 36 to 40 tons of water for each ton of rock handled in the mill. In the Atlantic mill we save only about 13 pounds of copper for each ton of rock stamped, equal to 0.66 per cent., our concentrates averaging about 73 per cent, of fine copper.''

The Frue vanner, whose revolving-belt receives a sharp lateral shock, and its rivals, the Embrey, Triumph, and other endless-belt machines, are improvements, both in construction and in accuracy of motion, over the old Brunton belt, which was the type that Capt. Frue followed in his original vanner, used in 1874 on Silver Islet in Lake Superior. The suitability of these automatic belt-machines to the concentration of the auriferous pyrites which escapes, in the un- classified slimes, from the gold-amalgamation mills of the West, where two vanners usually receive the tailings and water from each battery of 6 stamps, has led to their incorporation as an almost in- variable appendage to every Western gold-mill. The American metallurgist in this department has aimed at the improvement and adaptation to his special requiremeuts of old machines rather than to the creation of new inventions.

The extreme aridity of the Rocky Mountain zone has created a demand for dry concentrators, and has stimulated invention in that direction. Several machines, notably those of Krom and Paddock, have attained a considerable degree of success in concentrating

328 American Improvemgvts And Inventions.

larger particles by agitating a layer of ore moving over a perforated bed by intermittent jets of air produced by bellows; but the result has generally proved so much less perfect than that attained by wet eonoentrationy and the maintenance of the machine in repair so much more costly, that the system, whatever support it may obtain from theory, has not made headway where water is available.

III. — Calcining-Furnaces.

The necessity for reducing the cost of labor has led in the West to the invention or adoption of several types of mechanical calciners for fines and concentrates, in which form the ore is generally delivered to the metallurgist for treatment In several districts, complex silver- ores are crushed dry and roasted with* salt preliminary to amalga- mation or leaching, and most of the sulphuretted ores of copper are crushed and concentrated before being roasted. Three types of mechanical calciners are widely used :

1. Shafl-furnaces with one long drop.

2. Cylinder- furnaces, with lifters and repeated short drops : divided into (a) those with intermittent and (6) those with continuous discharge.

3. Stationary hearth furnaces, with mechanical rabbles.

1. The first shadrfurnace was that proposed in 1865 by Whelp- ley & Storer. In it four fire-places surrounded the head of the shaft, and the ore fell into a focus of flame, which was drawn down the shaft by an exhaust -fan that served at the same time as a spray- wheel. The complexity and cost of the furnace and its associated machinery prevented its adoption ; and the questionable practice of employing the highest heat at first in roasting ore created a preju- dice against it in the minds of metallurgists. Nevertheless, its ca- pacity was very large and the calcination done by it fairly good. The dense shower of incandescent ore which filled the stack was more brilliant than even the blast from a Bessemer converter.

The Stetefeldt furnace is the only shaft-calciner which has been extensively used, shafl-furnaces with interrupted fall, like the Gfer- stenhofer and the shelf-burners, so widely used for pyritic fines, being too slow to meet with general aooeptauoe. The large Stete- feldt furnace,' such as that of the Marsac mill, Utah, is 50 feet high and 6 by 6 feet at base, furnished with fire-place at base of stack, and auxiliary fire-place at the entrance to the dust-chamber, and provided with a most perfect automatic distributing feed.* When

and MaUe- Roasting in Utah," by R. N. Terhune, Trans, A. I. M. E, xvi., 21.

American Improvements And Inyentionb. 329

employed for chloridizing silver-ore, dry-cru8hed through a No. 30 screen, the capacity of sach a farnaoe reaches the high figure of 60 tons a day. The ore being thoroughly ignited, and partially chloridized during the fall of 20 feet, is very perfectly chloridized while at rest on the hearth of the furnace. Experiments in oxidizing finely- ground ore and matte have seemed to show that 50 per cent, of the sulphur present was oxidized during the fall; but as the oxidation does not afterwards proceed actively in the quiescent mass, the final results are too imperfect to entitle the Stetefeldt in its present form to rank as a succeasful oxidizing-fumace.

2. Of the second class, mechanical calciners of the revolving- cylinder type, two are prominent : the White-Howell, with continu- ous discharge, and the Bruckner, with intermittent discharge. The first combines, with the main features of the simple Oxland revolv- ing cylinder, and its four lifters built into and projecting beyond the brick lining, some of the advantages of a shaftfurnace, for the fire-box is so placed that the ore, falling in a continuous stream from the cylinder, passes through the flame as it travels to the cylinder. As a chloridizing-roaster for silver ores, it has been the closest competitor of the Stetefeldt; but it has not received, as an oxidizing-fumace, the favor bestowed on the Bruckner, of which no less than one hundred and thirty-six are in operation in the Anaconda works alone. There are two sizes of the Brfiek- ner furnace. The cylinder of the larger is 8 ft. 6 in. in diameter and 18 ft. 6 in. long. There are contracted orifices, one in front, to receive a movable fire-box, and the other in the rear, leading to the dust-chamber. From 9 to 13 tons of concentrates, inserted in one charge, are roasted every 24 hours, down to from 7 to 12 per cent, of sulphur, with a consumption, in Montana, of 1 ton of Rock Spring (Wyoming) lignite. The movable fire-place, attached to the cylin- der till the charge is ignited, is then removed, and is only re-attached to complete the roast.

At the Germania works, Utah, the Bruckner, by dint of careful management, is made to roast lead-ores, the cylinder being revolved once in forty minutes. The same furnace is used as a chloridizing- roaster, and, as modified by Hofiman, is provided with flue and fire-place at both ends to be used alternately. To accelerate the oxi- dation, the Clark oxidizing- and desulphurizing-apparatus is some- times attached. This is a water-jacketed perforated pipe, passing from end to end of the cylinder, through which jets of air are forced upon the surface of the burning ore, increasing notably the

330 American Improvements And Inventions.

duty of the furnace. I use a continuous-discharge cylinder-calciner, provided with a heavy central tile flue, which is supported by four slotted tile partitions. The fire-place gases, when fuel is required, are confined to the central flue, and heat the annular space between the flue and lining, which is thus converted into a muffle. The admission of air is regulated by a register. In roasting highly sul- phuretted ores or concentrates, the mass of brickwork, once heated, so retains and equalizes the heat that no fire is needed. A 60-inch cylinder will roast 6 to 12 tons a day of mill-tailings or pyrites fines to between 2 per cent, and 3 per cent, of sulphur. As the compart- ments are confined, an ore that sinters easily is either unsuitable or must be roasted in small quantities.

3. The two rake-furnaces which have met with most general acceptance are the Spence, and the O'Hara as improved by Brown. The former takes its name from Peter Spence, of Birmingham, by whom it was patented in England. The American form pos- sesses, however, many mechanical improvements not embodied in the original Spence patents. There are more than half a hundred now used in the eastern States for burning pyrites smalls in acid- works, where their capacity is about 3J tons a day ; but in the Par- rott Works, Butte, they roast 7 tons a day down to the percentage of sulphur required by the smelters.

The furnaces are built with four (and in some instances five) hearths, provided with a fire-place below the lowest hearth, when a dead roast is required, as in preparing gold concentrates for the Plattner process in the Treadwell mill, Alaska. The rakes are at- tached to a rigid frame, which is operated by an automatically-revers- ing engine; and the adjustments are so nicely made that the ore is admitted at the same rate that it is discharged, and falls from hearth to hearth in equal quantities, at whatever speed is desired. Undue wear and tear of the older Spence furnace are said to have been over- come, and the cooling of the rake provided for.*

The O'Hara furnace, in which the ore is moved for wan! by rakes attached to an endless chain, is an American invention. It has also undergone notable improvements. As now manufactured by Fraser & Chalmers, under the name of the Brown-Allen improved O'Hara, it is provided with two hearths, and the rakes are attached to car- riers which extend through slots in the side-walls into chambers in which the moving parts play, isolated from the roasting ores. The —

See Peters's Modem ArMiiean Methods of Oopper-SmeUing, 3d edit., p. 135.

Ic

American Impbovementb And Inventions. 331

furnace will roast down to 7 per cent, of sulphur as much as 40 tons daily; but despite the provisions for protecting the mechanical parts from the action of the roasting ore, the wear and tear is high. To lessen this defect, Mr. B. Pearce, of Argo, has recently constructed a circular rake-furnace, whose two hollow arms, moved by simple ma- chinery situated in the open space within the annular bed, are cooled by air> which, passing through them, is forced by the gentle pressure of a fan against the rakes and on the surface of the agitated roasting ore. The furnace may be accepted as the most perfect specimen of the rake-type. As used at Argo, it is heated by fuel, and is, there- fore, unsuited for acid-manufacture.

IV. Cupola-Furnaces and Cupola-Smelting.

The most radical American departure from precedent in cupola- smelting practice has been the introduction of water-jacketed fur- naces. The great cost of refractory bricks in the West forced smelters in remote situations to substitute other structural material for the construction of cupolas. Sandstones were used where available, and where the smelting-charge permitted, but metal furnaces natu- rally sufirgested themselves. The Longfellow copper mine, in Arizona, was opened more than twenty years ago, when the nearest railroad terminus was about seven hundred miles distant. As every fire- brick cost one dollar, their use for smelting a basic ore, such as the Longfellow then was, was out of the question. A furnace was built of plates of coarse copper, cast on the spot, which were sprinkled with water. These, in time, were replaced by copper boxes, open at top, through which cold water flowed ; it being found possible to make sound hollow castings of the copper, alloyed with iron, which was the product of the furnace itself. These clumsy sections, out of which an effective water-jacket was built, gave place to boiler-plate furnaces in 1882. By that time the present type of water-jacket had come into general use for smelting both copper- and lead-ores. The water-jacket furnace was a natural growth out of the water- tuyere and the water-back, but to whom is due the credit of its in- troduction is a moot question.

The copper-smelters generally prefer to make their furnaces of boiler-plate, the water circulating between an outer shell and an inner shell, in contact with the smelting-charge. Lead-smelters, on the contrary, prefer to build their furnaces of cast-iron sections, any one of which can be removed and replaced without much loss of time or serious disturbance. The width of the water-space

332 American Improvements And Inventions.

varies from 3 inches to as mach as 9 ioches where the water is very calcareous, and where, therefore, a deposit collects rapidly. Some furnaces are inverted cones, others are oblong, and those used in lead -smelting almost always have a bosh. As the jacket merely replaces the brickwork of the ordinary furnace, every smelter designs his furnace of the proportions and shape he finds best adapted to the ore he has to treat All large works using wrought- iron and steel jackets employ their own mechanics not only to patch defective spots, but to replace worn-out inner shells. The use of water-jacketed furnaces in preference to brick is, however, not universal. One of our largest genenil smelting-works, the Orford, employs a very large cupola of the Raschette type, built of brick or of brick and water-pipe coils, the contention being that more heat is lost by convection in the one furnace than by radiation in the other, and that the danger of explosion is an element that should weigh in the selection. The prejudice against water-jackets (especially in Europe) arises often from faulty construction, and also from the practice of counter-sinking all rivet-heads on the exposed surface of the inner shell. When high boiler-plate jackets are used, unless projecting points are lefl to hold up the shell, which immediately forms on the cool surface of the iron, and the rim which collects below the feed- door, the shell slips down before the tuyere and persistently deranges the working of the furnace.

Outer wells are used in all copper-furnaces not running metal. HerreshoiT employs a water-jacketed well which is admirably suited for collecting the matte and discharging the corrosive slag when smelt- ing pyrites residues. The most novel appliances for automatic dis- charge are Arents's so-called syphon-tap for the lead-furnace, and the Orford well for copper furnaces. The latter consists of a recep- tacle for slag and matte, divided by a longitudinal partition into two unequal compartments, the partition being provided with a ver- tical slot extending some inches upwards from the bottom. Into the large compartment slag and matte flow from the furnace, and from it slag flows continuously, while the matte only runs through the slot in the bottom of the partition and flows through a spout about two inches below the level of the slag-spout of the first compartment. When the flow of matte is rapid, this automatic contrivance works to admiration.'*'

Two blast-producing machines are distinctly American, the Root

For details, see Peters, 3d ed., chap. ix.

Amebican Improvements And Inventions. 333

and Baker blowers, both designed to produce a Femi-positive blast bj means of revolving vanes, which almost touch, and, therefore, force forward the entangled air. These occupy an intermediate point between the fan and the blowing-engine, giving as much pres- sure as the fan without injecting as much air or involving as rapid 8i)eed.

V. Reverberatory Furnaces and Practice.

The reverberatory practice of this country shows no such radical departure from precedent as is exhibited in the use of the water- jacket cupola. Some of the copper- refineries of the Lakes, in the days when large masses of native copper were frequently found in the mines, were built with removable roofs. At Anaconda and other new works in Montana and elsewhere, cylinder-calciners, placed above the reverberatories, discharge their red-hot ore into hoppers suspended above the reverberatory-roofs. At the Argo works, Mr. Pearce has altered the shape of his reverberatory hearth, increased the size from 10 by 15 ft. to 14 by 24 ft., and has thereby increased the capacity from 16 to 28 tons per day and de- creased notably, if not proportionately, the coal-consumption.* He also taps, as well as skims, his slag. Not content with past achieve- ments, Mr. Pearce is now building a reverberatory with an expected capacity of 35 tons.

At the new works of the Boston and Montana Company, at Great Falls, Montana, tilting-reverberatories, suggested by the tilt- ing open-hearth furnaces of the Pennsylvania Steel Ck)mpany, have been built, and are said to work successfully ; but full particulars as to construction and operation have not been published.

Gras has not been as successfully applied to copper-furnaces in this country as in some of the European works.

VI. Treatment op Ores. 1. Copper.

The particulars in which methods of copper-smelting have been modified through modifications of the plants, have been already described.

Of our two large Eastern smelting-establishments, the Baltimore Company follows, in the main, the English methods of reverberatory- smelting, and treats its argentiferous product partly by electrolysis and partly by conversion into blue-stone.

Peter8 Modem American Methods of Copper-Smeiiingf p. 324.

334 American Impbovemexts And Inventions.

The Orford combines capola-practioe with reverberatory-concen- tration and is contemplating the erection of an electrolytic plant The Lake establishments melt and refine their concentrated ores at one operation, in large furnaces modified in design to fulfil their special requirements.

At the Arizona carbonate-mines all the copper is made bj one fusion in water-jacketed cupolas; the matte, which is produced simul- taneously in greater or smaller quantity, being generally kiln- or heap- roasted and added to the ore-charge. The pyritic ores of the Verde mines are lieap-roasted, matted and concentrated in converters. In Montana, the ores of Butte, which receive almost without exception preliminary concentration, are in most works roasted in one of the previously described mechanical calciners and matted in reverbera- tory furnaces; the ore being in some works passed hot from the calciner to the reverberatory. The Parrott Company, under the pro- gressive guidance of Mr. Franklin Farrell, was the first iu this country to Bessemerize matte, and undoubtedly the first anywhere to make as much as 14,000,000 pounds of copper a year by the pneumatic process. Now all the large Montana works are adopting the same speedy method of matte-concentration, and raising 50 to 60 per cent, matte up to 97 per cent, black copper at one blow, in the deep converter with elevated tuyeres ; though, I understand, the bottom- tuyeres have been used as successfully with a very strong blast. The Bessemer pigs of the Parrott works are electrolyzed at the Bridge- port works of the same company ; and steps are being taken by all the other copper companies of the Butte district, to separate the pre- cious metals and refine the copper by electro-deposition.

The only large concern the sole aim of which is the treatment, apart from the mining, of argentiferous and auriferous copper, is the Boston and Colorado Smelting Company, of Argo, Col. The method employed for the extraction of the silver is the Ziervogel ; that for the extraction of the gold from the copper-bottoms is maintained a secret.*

Apart from the argentiferous mattes of Butte and Arizona, and those made by the Boston and Colorado Co. and its affiliated com- pany in Butte, and separated at Argo, most of the precious-metal copper-matte issues as a by-product from the lead-furnace. The silver of the San Juan, Durango and other districts of southern Colorado is

"Progress of Metallurgical Science in the West," by R. Pearce, Tran$. A. I. M, E; zviii.y 55.

AMERICAN IMPROVEMElirre AND INVENTIONS. 835

more or less associated with oopper-glanee, and copper is becoming more abundant in the Leadville ores, and thereby so increasing the supply of oopper-Iead matte as to bring Colorado into greater promi- nence as a copper-producer. Argentiferous copper is also imported in considerable quantities from the Sierra Mojada of Mexico. The copper-lead mattes are treated by electrolysis, at their own works, by several of the lead-smelting companies.

The Pueblo Smelting Company has, after extended experiments, carried into practice a very interesting method, patented by Mr. John Crook. The steps of the process as I saw it at first practiced are as follows:

1. The ground matte is immersed repeatedly in a bath of lead, maintained at a temperature below the fusing-point of the matte, which dissolves a very large percentage of the precious metals.

2. The matte is skimmed from the lead bath, and submitted to so- called scorification on the hearth of a reverberatory. The excess of lead flows off, and the volatile constituents of the matte are so thor- oughly evolved that, from a very arsenical antimonial matte, excel- lent copper is made, and the copper and iron are reduced to a sponge.

3. This sponge is fused with silica, and copper results. Particu- lars are not published as to the consumption of lead and exact ex- traction of gold and silver; but the method embodies a number of novelties, of the greatest interest to chemists and metallurgists.'*'

The Kansas City Smelting and Refining Company is preparing to extract the copper from lead matte, by a method which I worked out in association with the late T. Sterry Hunt. It depends on the action of sulphurous acid on copper sulphate, in solution with a suf- ficiency of chlorine to produce cuprous chloride. This pure copper- salt, insoluble except in a strong chloride solution, readily separates on the injection of sulphurous acid gas from a roasting-furnace. If the gas be not diluted by too much air, acid is simultaneously generated in excess of the equivalent quantity, due to the unavoidable, and in this case, desirable presence of some free oxygen in the injected gases. The sub-chloride can be reduced by iron to cement-copper, and the chlorine recovered as chloride of iron, or to suboxide by lime, and the chlorine recovered as chloride of calcium. The precious metals remain with the iron oxide in the matte- residues.

For further particulars as to present practice, see Hofman's Mttatlwrgy of Lead, D.267.

336 American Impbovements And Inventions.

There has been, however, but little patronage extended to wet copper-methods, mainly because we do not possess, within accessi- ble reach of the chemical centers, any large bodies of cupriferous pyrites, whose residues, after the extraction of sulphur and copper, would possess value as an iron-ore. The treatment of the low- grade oxidized ores of the southwest by lixiviation is awaiting realization.

In the past, various attempts have been made to employ old and new wet methods, but none have proved commercially successful, nor have any survived till to-day.

Bessemerizing, as a method of concentrating copper-matte, is rap- idly displacing the older and slower processes; but with us the machinery does not differ notably from that in vogue abroad. Here, as in Europe, attempts are being made, and with more suc- cess than there, to utilize the heat generated by the oxidation of sulphur and iron, and thus smelt pyrites without fuel, and also to Bessemerize and smelt at one operation, as was essayed by HoUway, in 1879, in cupolas of special construction. Whatever improve- ments have been made in the electrolysis of copper are kept secret

2. Lead,

American practice in lead-smelting is distinguished rather by the quantity of work got out of a plant of a given size than by any marked departures from European methods.

The larger lead-furnaces, invariably water-jacketed, and driven with a higher blast than of old, smelt in some cases more than 100 tons of charge per furnace, and yet, through skillful metallurgical handling, and close attention to slag-composition, do work which will compare in perfection with the best in Europe.

Our prominent lead-smelters have brought more accurate scien- tific training and acumen to bear on their professional labors, than perhaps any other class of the fraternity. Yet, withal, their science has not dwarfed their inventive faculties or made them slayes to the teachings of the schools.

Reverberatory-smelting in the English type of furnace, is little, if at all, practiced in this country. In Mississippi and Missouri, a modified Carinthian hearth, known as the air-furnace, is credited with producing about 10 per cent, of the lead from that rion, while over 60 per cent of the pure lead-ores of the Mississippi Valley are reduced on a hearth, which differs from the typical Scotch hearth

Amebican Improvements And Inventions. 337

in being water-jacketed, and therefore capable of continuous run- ning.*

At the Lone Elm smelting-works the Lewis and Bartlett process of collecting the lead-fumes in bags has been adopted, and the heavy losses of this system of smelting have been reduced. The furnaces which produce 174,626 tons of our total of 206,626 tons of leadf are large cupolas, growing ever larger, originally designed on the Ras- chette model, then verging towards the Piltz, and finally reverting to the Raschette type. They are now so modified in essential details by each smelter that no single furnace can be taken as the exact pat- tern of the whole. A standard size is now 42 by 120 inches* All are built with a bosh and all are jacketed within the zone of fusion. Most are provided with an Arents syphon-tap, and a dis- charge built into the crucible-wall, communicating with the bottom' of the lead-well; in all, the superstructure is supported on pillars to* facilitate the removal of defective sections. In most, coke and char- coal mixed is the fuel, but in one instance at least uncoked lignite,, mixed with very poor coke made from the same lignite, is used very successfully. Most are fed by hand ; some through thimbles. In- most cases only the dust and fume that will collect in long flues is saved ; in other cases, sheet-iron chambers, with large cooling-surfaces,, are used, and in more than one, the Bartlett method of filtering the fume through woollen bags, as in collecting zinc-white, is applied Preliminary roasting is done in some works in the O'Hara and Bruck- ner mechanical furnaces. Mr. Matheson, of the Pueblo Smelt- ing and Refining Ck)mpany, has, it is believed, at last perfectel a tapping-jacket which, attached to the ordinary water-jacketed fur- nace, permits a continuous flow of slag, and the maintenance of the liquid charge at a constant level below the tuyeres. While the arrangement of the newer works is essentially American (excellent examples of which are Eilers's works at Pueblo, Col., and Great Falls, Montana), the works being laid out with a view to the utmost economy of fabor, and the utilization of the utmost efficiency of the machinery employed, the only points to which attention may be directed, as distinct American inventions and suggestions, are the- water-jacketed walls, the Arents syphontap, the Matheson slag- tap, and the Bartlett method of collecting fumes. No notable de- parture from accepted foreign practices or appliances is generally

Hofman*8 Metallurgy of Lead, p. 89.

t Rothweirs Mineral Industry, vol. i., p. 307.

Vol. Xxii.— 22

338 American Improvements And Inventions.

made in the departments of refining and parting, if we except the use of Moebius's electrolytic process for treating dor6 bars.

3. Oold and Silver.

In the field of gold and silver metallurgy more originality has been displayed and exercised by American workers than in, perhaps, any other.

Under the head of concentration I have referred to the system of hydraulicking, which was the product of California. The gold- mill, with its series of automatic operations, is another offspring of California ingenuity. Manual labor is almost entirely replaced by ocular labor; for superintendence, and not work, is the function of the mill-hands. The ore dumped into the breakers falls into large bins, whence it slides into automatic feeders, which supply the stamps with regulated quantities. The free gold is extracted partly by liquid mercury in the mortars and by copper plates at- tached to their sides, and partly on an apron of amalgamated plates, over which the crushed pulp flows as it issues from the battery- screen. Frue or other automatic vanners receive the tailings, sepa- rate the sulphurets, and discharge the waste. When the power is supplied by water, and the stream is divided to Pelton wheels, coupled directly to the separate groups or even pieces of machinery, the absence of intermediate running-gear increases not only the sense but the reality of automat icity, and makes a skilfully arranged and thoroughly equipped California gold-mill one of the triumphs of modern mechanical metallurgy.

The treatment of the iron sulphurets which, in deep mines, invari- ably accompany the free gold, and sometimes lock up all the precious metals, has been confined, practically speaking, to two methods.

(1.) Smelting with either lead or copper, or alone with the produc- tion of an iron matte. The general experience has been that an iron matte is a less perfect absorbent than either a copper matte or lead.'" In these departments we can point to no notable improvements or departures from European practice, unless it be in a nearer ap- proach to success in pyritic smelting.

(2.) The chlori nation of gold-ores, as proposed and practically car- ried out by Plattner, including the roasting of the ores, the impr- nation of the moistened mass in leach-tanks by the chlorine gas, ad- mitted through a false bottom, the extraction of the soluble gold by

" Progress of Metallargicsl Science in the West,'' R Pearoe, Trans. zviii., 55.

American Improvements And Inventions. 339

water, and of any silver which may have been ehloridized by hypo- sulphite of soda, is the method used at Grass Valley and other points. But the slowness of the operation in stationary vats led Mears to substitute revolving barrels, whereby not only the chlorinatiou, when gas was injected, could be accelerated, but the ingredients for gener- ating chlorine could be agitated with the ore itself.'*' A difficulty, encountered in the treatment of some ores, of leaching the mass of pulp when discharged from the barrel, has been met by constructing lead-lined iron barrels provided with a filter, using the barrel when revolving to effect the solution, and when at rest as a filtering vat, and accelerating the filtration by pressure. Such barrels are con- structed of a capacity of 8 tons per charge.

The cyanide method has been applied in some cases with admitted success, but it cannot claim American parentage.

A Western silver-mill has no prototype in the older European prac- tice. It was a Nevada outgrowth of the California gold-mill, employ- ing the same crushing-machinery, similarly combined, but substituting amalgamating-pans for battery- and apron-amalgamation (on account of the greater resistance to amalgamation of silver than gold) and settling-pans, in which the mercury is collected from the pulp. As usually designed and workid, the ore passes from the battery into settling-tanks, whence it is shovelled, in charges of suitable size, to the pans. In them it is saljected for hours to mercury under the grinding action of iron mullers revolving in close proximity to iron dies secured to the bottom of the pans. Live steam is usually in- jected, but the Boss pans are jacketed. Chemicals {salt and sulphate of copper) are generally added by rule of thumb, or no rule at all. They undoubtedly play the same part as in the patio, the Krohnke and the Freiberg barrel processes, yielding sub-chloride of copper, which aids in the decomposition of the silver sulphides ; but it is difficult to distribute fairly the effects on the amalgamation among friction produced by the grinding of iron on iron, heatnd the reac- tions due to the chemicals.

In the Boss system the palp flows continuously through a series of pans and settlers, the reactions being, it is claimed, accelerated by the thinner layers under treatment in each pan.

The Nevada mill and the Washoe process do quick work, but not clean work.

"The Chlorination of Gold-Bearing Sulphides;' hj E. G. Spilsbary, TnxM zvi., 359.

340 American Improvements And Inventions.

Where the silver-ore, to use a Western term, is so " base as to resist amalgamation, it is crushed dry, salt-roasted, and may then be amalgamated in pans and settlers. Barrel-amalgamation is not resorted to in any of the large mills.

But in this case the Patera process is sometimes used. It was first extensively applied by Ottokar Hofmann (to whom is also due the system of rapid trough-lixiviation, IVana,, xvi., 662), in Par- ral, Mexico, in 1868, and since then has been widely employed; oftener in small establishments, or by miners who have extemporized a plant out of old barrels, than in large works. The only impor- tant modification in the process itself is that made by Mr. Russell, of the Marsac mill, Utah, which consists essentially in the addition of sulphate of copper to the solution of hypo. Mr. Stetefeldt has brought his engineering skill to the aid of Mr. Russell's chemistry ; and the resulting plant, as designed and running in the Marsac mill, Utah, contains economical devices for the discharge of the leach- tanks, for accelerating precipitation. and for the collection and reduc- tion of the silver sulphide.

4. Nickel

Until the ore and matte from SudUury,. Ont, were imported, the only nickel comniercially produced in this country, apart from small quantities which came as a by-product from the lead mines of Missouri, was extracted by Mr. J. Wharton from the Lancas- ter Gap mine. Pa., and converted into metal and nickel salts at his works in Camden, N. J.,, by methods which were kept pro- foundly secret. Now that nickel-steel has created an increased de- mand fir the metal or its oxide, the methods of reduction have been extensively investigated. One result has been the patent- ing by Messrs. J. L. and R. M. Thompson, and the practical application on a large scale by the Orford Copper Company, of a method for.eeparating nickel from copper-nickel ores and mattes by smelting the mixed sulphide of iron, copper and nickel with an alkaline sulphide, or with acid salt-cake. On tapping the charge, sulphide of nickel, with a small proportion of the copper and iron sulphide, settles to the bottom of the pot, while the sulphide of iron and copi>er, with a very small proportion of sulphide of nickel and a large proportion of the salts of the alkali used, form a very distinct, easily separated, upper layer. By retreating the bot- toms, a still further elimination of foreign metals is effected, until ultimately pure nickel sulphide is obtained. The consumption of

American Improvements And Inventions. 341

alkali is reduced by recovering it from the tops. The company, by this method, puts on the market large quantities of nickel oxide made by roasting the sulphurets, and contaminated with only a tri- fling amount of iron oxide*.

The oxidized nickel ores of the West have not yet entered the market extensively.

6. Aluminum.

The successful agent of recent investigation in the metallurgy of aluminum has been electricity ; and this is an agent which we have not been backward in trying to harness to our metallurgical vehicle and turn to our use. This country may claim unquestionable priority of invention for one electrical method of treating aluminous com- pounds, and at least independent simultaneity for another.

The Cowles Bros., of Cleveland, Ohio, patented in 1885 a process and the necessary appliances for producing aluminum by the de- composition of corundum in the electric arc, and for forming alu- minum alloys by decomposing the corundum in the presence of granules of oopper 'and other metals. They place the alumina, whether corundum or prepared oxide, with charcoal, and the metal with which the aluminum is to be alloyed, between the electrodes ot a dynamo in a carbon-lined and covered trough, A current of about 8 volts and 3000 amperes is turned on ; the alumina is decomposed, and as aluminum is formed, it combines with the fused copper, or other metal, and the alloy collects in globules or masses. The oxides of calcium, magnesium, manganese, silicon, barium, and the alka- lies, are decomposed in the Cowles furnace with the production of these elements and metallurgical alloys, which it would be difiQcult to make otherwise.

The only other process under which aluminum and its alloys are made in commercial quantities in this country, is that pat- ented by C. M. Hall, and operated by the Pittsburgh Reducing Company, which " reduces the metal from its oxide, alumina, by elec- trolysis, this alumina being held in solution by a molten fluoride bath, which itself is not decomposed by the electric current. The latter is conveyed to the melted solution by means of carbon cylin- ders placed in the bath for positive electrodes, a carbon-lined metal pot forming the negative electrode.

" The oxygen of the alumina goes off at the positive electrode as

Nickel patents issued to J. L. Thompson and R. M. Thompson for improve- ments in the separation of sulphide of nickel, No. 489,576, January 10, 1893; No. 489,676, January 10, 1893 ; No. 489,574, January 10, 1893.

342 American Impbovehents And Inventions.

carbonic acid, wearing away the carbon at the rate of nearly a pound of carbon to the pound of aluminum produced. The reduced metal settles to the bottom of the pot, from which it is easily tapped or ladled off, practically free from the electrolyte. Remelting entirely purifies it."*

There are points of resemblance between the Hall process and the Bernard process used at Creil (France), and more remote simi- larity between the Cowles system and that in operation at the large works of Miilhausen, on the Rhine.

6. Zine.

Despite the wide diffusion of zinc, and the large size of some of the zinc-deposits in this country, no important improvement has been made in its metallurgy since the invention by Wetherill, in 1855, of a method for making zinc-white by burning with forced draft, on the perforated grate-bars of a specially-constructed furnace, a layer of anthracite pea- or nut-coal upon which is spread a bed of naturally or artificially oxidized zinc-ore, mixed with about one-third its weight of anthracite slack. The zinc is simultaneously volatilized and oxidized, and the white zinc collected in muslin bags. At the New Jersey Zinc Company's works, at Newark, franklinite ore from its mines at Mine Hill, N. J., is thus treated for the extraction of most of its zinc. The remaining iron and manganese are run down into spiegeleisen, and the zinc-fumes which escape from the blast-furnace are treated for spelter.

There is a profitable field for metallurgical research and invention in devising an economical method of recovering the zinc from the mixed ores of lead, zinc, and copper, of the Rocky Mountains, where they are, at present, far from being a source of profit, the most trou- blesome and wasteful element with which the furnace-man has to deal.f The Cowles brothers were experimenting in this direction, and with this aim, with their electrolytic furnace, when their atten- tion was deflected from zinc to aluminum.

7. Mercury. The mining and metallurgy of quicksilver is almost wholly con- fined to California. Of the total of 101,336 flasks made in the country, only about 2300 flasks are reported in the last census year as coming from Oregon and Utah.

''Properties of Aluminum/' by Hunt, Langley and Hale, Tran, A. L M, E., xviii., 628. t See "Treaunent of Zinc-Lead-Sulphides,*' Both well's Jftiu Ind., p. 316.

American Improvemento Akd Inventions. 343

The Uerror and anza-fiirnaces used in California do not differ in principle or in their main features of construction from the dis- tillation-furnaces for coarse and fine ores, of Europe ; but in their accessories ingenuity and money have been expended with a view to minimizing labor, while in compactne&s of plan and neatness of construction, the California works contrast favorably with the older European establishments.

In 1876 the necessity of saving labor led to the construction and adoption at New Almaden of the Huttener and Scott vertical shelf- furnace, for roasting cinnabar fiues. The Livermore inclined shelf- furnace is a later invention.'*'

8. Antimony.

At present, antimony is not made in this country in commercial quantities, though in Utah, Nevada, Montana, Colorado, and Ar- kansas, there are deposits of stibnite of considerable extent. In 1881 a very good grade of antimony was made from the oxidized ores of Sonora, Mexico, at Oakland, California. The reduction was effected in a low 36-inch diameter cupola, and the refining was done in crucibles ; but the enterprise was short lived. The same was true of the operation of the Stayton Mining Company, which some years earlier treated antimony ores from California in reverber- beratories; also of the works run by Faber at Battle Mountain, Nevada, on stibnite from Humboldt county, Nevada; of the more extensive establishment in Salt Lake City for treating the antimony ores of southern Utah, and of a Philadelphia company, which made for a time in Philadelphia a good article out of the concentrated ores of Arkansas, by the English crucible method.

9. Tin.

Though tin exists in Arkansas, is widely diffused from Maine to Alabama in the Appalachian range, and has been exploited in Dakota and California, the mining of it so far has nowhere been suf- ficiently remunerative to create a smelting industry. The concen- trating-mills constructed in Dakota and California include sub- stantially the machinery and automatic methods employed in the mechanical treatment of other ores, and therefore differ widely from the dressing-floors of a Cornish tin-works.

♦ a B. Christ/s Quicksilver Reduction at New Almaden."— IVana. A. L M. R, xiii., 547.

344 american improvements and inventions.

Conclusion.

From this hasty review of AmericaD departures from European precedents, it will be seen that though our metallurgists cannot claim the honor of any great metallurgical discoveries, they have brought to bear on the practice of their profession acute ingenuity in altering machinery and methods, no as to adapt them to their altered surroundings. The high price of labor, coincident in some instances with very lean ore, has necessitated the adoption, wherever possible, of automatic machinery arranged, so to speak, consecutively, which does its work very cheaply, if not very perfectly. The cost of refrac- tory building material in the West has led to the replacement, in the construction of furnaces, of brick by water-jacketed metal. The distance from a market has excluded our metallurgists from taking advantages of certain economies, which materially assist their foreign rivals, notably in the saving and utilization of by-products. The cost of fuel, and sometimes of fluxes, makes it at times more profit- able to lose than to save.

In the Lake Superior mills, for instance, it would not pay to reduce the quantity treated, and increase hand labor, in order to save a few hundredths of one per cent, of copper. There are places where, limestone being almost more costly than lead, it would be unprofitable to make a clean slag. At the carbonate-mines of Ari- zona, the high copper contents of slags could be reduced only by making, instead of metal, a matte, the reduction of which would cost far more than the value of the copper sacrificed. How far it is right to waste nature's resources in order to realize immediate profit, is a matter for the economist and the legislator to decide. The met- allurgist of America, while fully conscious of the imperfections of his practice, having nevertheless set before him the task of developing with profit the mineral resources of the country, has fulfilled it by dint of enterprise and originality, wastefully it may have been some- times, but yet successfully. Where, however, through improved means of transportation, fuel and fluxes have become accessible, he has shown, as in the case of lead-smelting in Colorado and Utah, that whereas in the early days he reluctantly adapted himself to circum- stances, and in order to make lead at all had to make foul slag, to-day he can make his profit by doing work so excellent from both the theoretical and practical points of view, that it may safely challenge comparison with the best performances of Europe.

The Open-Hearth Process. 345

The Opbnbeabth Pb0Ces8.

BY H. H. CAMPBELL, 8TEELT0N, PA. (Chicago Meeting, being part of the Interaational Engineering Congress, August, 1893.)

Introductory.

The foIlowiDg paper deals almost exclusively with the results of practice at the works of the Pennsylvania Steel Company at Steelton, Pa. From the records of the furnaces at this plant, both acid and basic, large and small, are drawn the data for the various demon- strations, and on a long and intimate experience with their opera- tions are founded many theories and generalizations. It may seem that universal laws should not be built upon a basis so restricted ; but in reality the condition is the strongest guarantee of validity. Each table investigates some certain variable, and by virtue of the common origin of the factors, we have good ground for assuming the corresponding agents to be equal.

The examples may or may not be instances of good practice, and may be open to criticism if regarded solely as exponents of metallurgical success. But it must be remembered that the lists are not made up of selected specimens, nor are they chosen to show su- perior results ; they are gathered to investigate, by strict scientific methods, the formulae of the operation. In such work no untoward results should be eliminated, save those which bear upon their face the stamp of experimental error. If the worst members be omitted, then the best must also be omitted. Only by using every available case can an honest attempt be made to deduce a true average. It is recog- nized, however, that with the most careful selection it is difficult, if not impossible, to collect data representing universal conditions. In each table the records tell of a special practice, and all conclusions drawn from them are true only within the limits between which the equations have been integrated.

Not assuming the prerogative of speaking for American melters in all the opinions herein expressed, nor acting as a camera in por- traying an ephemeral state of national metallurgical development,

346 The Open-Heabth Process.

this paper aims only to formulate the laws and principles on which the open-hearth furnace is working out its history.

The following table of contents may facilitate the use, as it will show the scope, of the paper :

Chapter L — The Economic Position of the Open-Hearth

Furnace.

Section 1. — Comparison of the Furnace and the Converter.

Section 2. — The Basic Furnace.

Section 3. — Combination of the Open Hearth and Converter.

Chapter II. — The Regenerative Furnace and its Machinery.

Section 4. — Construction of the Furnace.

Section 5. — Method of Operation.

Section 6. — Dynamic Equation of the Grases.

Section 7.— The Hearth.

Section 8.— The Valves.

Section 9. — The Ladle-Crane.

Chapter III. — Fuel.

Section 10. — Quality of the Gas Required.

Section 11. — Bituminous coal: The Siemens Producer.

Section 12. — The Action in the Producer.

Section 13. — The Composition of Siemens Gas.

Section 14. — The Thermal Equation of the Producer.

Section 15. — The Use of Anthracite Coal.

Section 16. — Water-Gas.

Section 17. — Natural Gas.

Section 18. — Petroleum.

Chapter IV. — Regulation op the Temperature.

Section lO.The Law of Thermal Increments. Section 20. — Estimation of the Relative Temperature.

Chapter V. — The Thermal Equation op the Furnace. Section 21. — Summary of the Thermal Factors.

The Open-Heabth Process. 347

Chapter VI. — The Acid Process.

Section 22.— The Hearth.

Section 23.— The Charge.

Section 24.— The Method of Charging.

Section 25. — Conditions of Oxidation during the Period of Fusion.

(a) The kind of gas.

(6) The nature of the flame.

(c) The construction of the ports.

{d) The method of charging and arranging of stock.

The time of exposure.

(/) The kind of scrap.

(g) The kind of pig-iron. Section 26. — Value of the Elements for the Absorption of Oxygen. Section 27. — The General Chemical Law.

Section 28. — Quantitative Investigation injo the Chemical History. Section 29. — General Review of the Quantitative Results. Section 30. — History of the Ore- Additions.

Section 31. — Quantitative Investigations into the General History of the Ore-Additions.

(a) Composition of the slag.

(6) Composition of the metal.

(o) Reduction of the FeO. Section 32. — Quantitative Investigations into the History of the Ore at different Periods.

(a) Composition of the slag.

(6) Composition of the metal.

(c) Reduction of the FeO. Section 33. — History of the Metal.

Chapter VII. — The Basic Process.

Section 34. — Pseudo-Basic Practice.

Section 35. — Material for Basic Hearths,

Section 36. — The Passive Joint,

Section 37.— The Charge.

Section 38. — The Basic Additions.

Section 39. — Method of Charging the Basic Additions.

Section 40. — Conditions of Oxidation during the Melting- Period.

Sectioii 41. — The Protective Power of Phosphorus.

Section 42. — The Greneral Chemical Law.

Section 43. — Quantitative Investigations into the General History.

348 The Opex-Hearth Process.

Section 44, — History of the Slag, with Special Reference to FeO. Section 45. — The Elimination of Phosphorus.

(a) Dephosphorization during melting.

(6) The final dephosphorization. Section 46.— The Elimination of Sulphur.

(a) By the action of metallic manganese.

(6) under ordinary conditions when the initial sulphur- content is below 0.10 per cent.

(c) under special conditions when the initial sulphur-

content is above 0.10 per cent

(d) By manganiferous ore.

(e) Conclusions.

Chapter VIII. — Preferential Relations,

Section 47. — Sih'con and its Oxide. Section 48. — Manganese and its Oxides. Section 49. — Carbon and its Oxides. Section 60. — Phosphorus and its Oxide.

Chapter IX. — Recarburization.

Section 51. — The Time when the Recarburizer is Added.

Section 52. — Function of the Recarburizer.

Section 53. — Recarburization in the Furnace.

Section 54. — Recarburization in the Ladle.

Section 65. — Certain Factors affecting the Loss of Manganese.

Section 56. — The Addition of Silicon.

Section 67. — Loss of Manganese on the Basic Hearth.

Section 58. — Rephosphorization.

Chapter X. — Conditions op Successful Practice.

Section 69. — Personal Equations. Section 60. — The Determining Variables.

Chapter XI. — The Use and Loss of Material.

Part I.The Acid Process.

Section 6L — The Conditions which Limit Comparisons. Section 62.— Theoretical Losses.

(a) From sand and oxidizable elements.

(6) From iron held in chemical combination in the slag.

The Open-Heakth Process. 349

(c) From shot in the slag.

(d) From the ore used.

Section 63. — The Conditions of Melting.

(a) Time of Exposure.

(6) kind of Gas. Section 64. — Summary of Results.

Section 65. — Results of Acid Practice with Pig and Scrap. Section 66. — Results of Acid Practice with Pig and Ore.

Fart II. — The Basic Process,

Section 67. — Theoretical Losses.

(a) From sand and oxidizable elements.

(6) From iron held in chemical combination in the slag.

(c) From shot in the slag.

(d) From the ore used.

(e) From the lime used. Section 68. — The Conditions of Melting. Section 69. — Summary of Results.

Section 70. — Comparison of Different Methods. Section 71. — Results of Basic Practice.

Chapter I. — The Economic Position of the Open- Hearth Furnace.

Definition. — The open-hearth furnace, as discussed in this paper, is a renerative gas-furnace, in which pig-iron, wrought-iron, steel, or similar iron products are exposed to the direct action of the flame and converted into steel.

Sec. 1. — Comparison of the Furnace and the Converter.

As an outlet for the crop-ends and miscellaneous scrap from the steel-mills of the country, the open-hearth furnace fills a position of general economic importance. More than this, it has been found peculiarly fitted for producing certain grades of structural and other steels, and furnaces have been operated where scrap was scarce, and where it was necessary to use pig-iron alone as the regular charge. Thus the process has been raised from a subsidiary to an independent position.

In the field of structural work, as in other branches, the Bessemer converter is always a colossal rival. In the United States, Bessemer

350 The Open-He A Bth Process.

linings are almost universally acid, and therefore in the following detailed comparison of the two processes it will be unnecessary to consider the metallurgical position of the basic converter or the applicability of its products to the stringent demands of modern engineering.

(a) In making steel with the phosphorus from .06 to .10 per cent., the converter can turn out a product at less cost than the open hearth. For many purposes it answers well enough. In compar- ing it with open-hearth metal of exactly the same contents of carbon, silicon, manganese, phosphorus, sulphur, and copper, no marked difference can be detected by the ordinary methods of testing; but many engineers, after long experience with both metals under vari- ous trials and stresses, think that the Bessemer metal suffers a larger number of inexplicable breakages. Experiments on punched test- pieces indicate that this arises from its greater liability to crystallize under shearing-stresses.

(6) To make Bessemer steel with low phosphorus requires very careful selection of the raw material. If cupolas are used, the pig- iron will absorb both sulphur and phosphorus from the coke, and the percentage of these elements will be increased by the waste in the converter. The pig-iron must therefore' contain much less than the maximum of impurity allowed in the steel. It must also contain the requisite amount of silicon for successful blowing, and the scrap used in the vessel should be as pure as the iron. These conditions are not impossible or impracticable, but they enhance the cost of the product considerably and prevent the employment of pure Bessemer steel in competition with common open-hearth metal.

(c) No solid fuel is used in the open-hearth furnace, and therefore phosphorus cannot be absorbed from it by the metal, while sulphur will not be unless the coal be unusually sulphurous. There is also more leeway in the selection of stock. In the acid-lined furnace, low- or high-silicon iron can be used, as well as washed pig, char- coal-blooms, or basic scrap, while the smaller amount of oxidizable elements reduces the increase in the percentage of phosphorus and sulphur due to waste. In the basic furnaces the ability to remove both these objectionable impurities renders the question of stock a matter of secondary importance.

(rf) The conditions of Bessemer practice preclude a system of test- ing during the operation. The vessel may be turned down for a time, but a skull is likely to form on the lining where the metal lies, and this skull, if dissolved by a subsequent hotter charge, may

The Open-He A Bth Process. 351

appreciably affect the composition after a test has been taken. More- over, the chilling of the charge during the testing is a matter of vital importance in the practical working of the mill. The fact must also be noted that the economy of the Bessemer process depends upon continuous and rapid running; any system of interrupted work is bound to raise the cost of the product.

{e) The open hearth admits a system of testing without any corre- lated disadvantages. Such testing is necessarily crude, since the bath is continually changing under the action of the flame, and the results must he obtained quickly. The tests must also be taken before the addition of the recarburizer, and hence will be only relative. After the addition of the recarburizer, the rapid oxidization of the man- ganese renders delays out of the question, but in spite of the limited opportunities, the tests that can be made are of great value in con* trolling and regulating the composition of the charge.

(/) Uniformity and homogeneity, therefore, are two of the most important factors in the comparison of the merits of the Bessemer and the open-hearth product; but unfortunately no conclusive testi- mony can be given to the skeptical or even the careful mind, although ex parte arguments are easily constructed. No one conversant with the facts doubts that Bessemer heats can be made which are as homo- geneous throughout as any open-hearth charge that was ever melted. No one doubts that in good practice the proportion of Bessemer heats which are not homogeneous is a small percentage of the whole number. But the question is not as to the homogeneity of ninety- nine heats ; it is about the quality of the hundredth. And it is not one single test of this hundredth charge that is required, but a large number of tests, taken from all parts of the cast. One piece of steel differing radically from the rest, wipes away all favorable arguments drawn from any number of other tests indicating homo- geneity.

The terms homogeneity and uniformity are ofln confounded. The first should mean the likeness of all parts of the same heat, and the second, the likeness of one heat to another. It may be said that the second term includes the first, but this would imply a stretch of meaning, as by common acceptation it may be said that certain products are iniformly imular. What is required in steel products is a combination of uniformity and homogeneity ; and obviously, no series of records however long, can demonstrate the existence of this necessary combination, unless all heats have been thoroughly tested and have been found homogeneous.

The Opek-Heabth Process.

Table 1. — Tests Showing Homogeneity op Open-Hearth

Metal.

Heat 10,699. Acid open hearth. Test-bare, %" rolled rounds.

Elastic

ntimate

Limit

Strength.

lbs. per sq. in.

36,900

53,510

36,460

64,790

36,000

66,160

65,690

36,090

55,830

56,830

36,740

66,370

36,360

66,090

36,460

57,610

36,125

56,900

37,680

56,600

36,900

67,610

37,220

57,420

37,130

57,280

36,000

57,050

67,190

67,440

36,460

57,670

37,166

57,680

36,640

57,350

Av. 36,610

66,638

Clong. in

Sin. per cent.

31.2")

Reduct.

of area.

per cent,

Heat 10,910. Acid open hearth. Teebare, 1%'' rolled rounds.

per cent, per cent.

,. , lbs. per lbs, per sq. in. gg,

31,140 31,790 31,540 31,250 31,080 31,160 31,250 31,040 32,050 31,660 31,700 32,550 32,570 33,330 33,.-)80

62,760 62,750 52,000 52,320 62,320 62,830 53,160 63,840 52,480 62,680 62,960 63,050 63,860

Av. 31,809

52,863

5aio

Heat 11,018. Acid open hearth. Test-bare, %" rolled rounds.

Heat 1,820. Basic open hearth. Test-bare, %" rolled rounds.

Elastic

Limit.

Lbs. per sq. in.

Ultimate Strength. Lbs. per sq. inch.

36,700 37,150 36,060 36,420 86,060 35,780 36,700 35,780 36,700 35,700 37,020 37,4 0 37,260 37,480

58,400 57,840 56,880 66,940 56,700 57,180 66,800 57,440 56,800 57,440 66,900 57,180 67,320 56,780 67,420

EK>ng. in

8 in. Per cent.

31.0U

Reduct. of area. Per cent

Elastic

Limit.

Lbs. per sq. in

31,530 33,650 31,600 33,340 32,760 33,260 32,935 33,270 32,900 31,920 32,185 33,880

s7A5"gt!El2"S*! Reduct. T \iA ntr ® ™- of area. inh. Percent.. Percent.

48,340 47,380 48,450 48,230 49,176 48,560 47,730 48,786 48,640 49,440 47,836 48,060 48,360 48,400

Av. 36,634 57,t:01 30.36 63.94 A v. 32,745 48,384 34.75 72.49

35.U0

THE OPEN-HEARTH PROCB&fS. 353

In the case of low-carbon steels the open-hearth possesses slightly better facilities for securing homogeneity and uniformity than the Bessemer. For high-carbon steels the conditions of manufacture make the hearth far superior. The metal in the converter is always blown until nearly all the carbon is eliminated since the stopping of the blow at any definite intermediate point has proved impracti- cable. All the carbon-content of the steel, therefore, must be added in the recarburizer, and absolutely perfect homogeneity can only be secured by absolutely perfect mixing. In the open hearth, on the other hand, high-carbon steels are made by interrupting the process at the desired stage. It is plain that no mixing is required, so far as carbon is concerned, since about the same quantity of recarburizer will be used for a given manganese, whether high or low steel is being made.

The manganese in the recarburizer distributes itself very evenly under proper conditions. In TVarw., xiv., page 358, some experi- ments by the writer are recorded, proving this point. Confirmatory evidence may be obtained by rolling the charge into small bars and testing a large number of pieces. Some variations in physical prop- erties will be caused by differences in finishing-temperatures, and some by the fact that the test-bars, not being machined, are not exactly round, and hence their area may not be accurately calcu- lated; but the data will sufSciently indicate that the variations are not of practical importance. On page 352 are the records of several- heats thus tested. The first three were made in a twenty-five-toa acid, and the last in a fifteen-ton basic furnace.

The foregoing comparison of the open hearth and the coirverter may not be a convincing argument against Bessemer metal. It may justify engineers in using the cheaper article in many structures, but it will also sustain the more cautious members of the profession who refuse to incur a known or a probable risk. In adhering to the safest course, engineers are continually calling for a metal with* lower phosphorus. The limit has been .lO per cent.; it is now .08 ; soon it will be .06 ; it should be .04. Every step in this direc- tion means more work for the open hearth, and it nieans also the development of the basic furnace.

Sec. 2. — The Basic Fwmaoe.

The stock required in acid work for a pure product is, and always must be, more expensive than for the common grades. The basic- lined furnace renders possible the elimination of tlie impurities.

Vol. Xxii.— 23

354 The Open-Heabth Pbocb88.

Like every new device, it has Its censors. One of the commonest arguments against it is that good steel cannot be made from bad stock. This statement must be carefully qualified in speaking of the acid process, since the finest open-hearth- and crucible-steels may be, and usually are, made from washed pig-iron, puddled iron or charcoal-blooms, all produced by the application of a basic process to a more or less phosphoric iron. So far as the assertion is applied to the basic open-hearth process, it is unquestionably wrong ; the first piece of good steel made from phosphoric iron relegated this doctrine to the realm of ignorant prejudice. Some engineers prefer not to employ basic metal rather than incur the least risk of unsatisfactory material from a new process. Against such reasoning little can be said, since it has already been admitted that Bessemer and open- hearth metals of the same composition may be unlike; but it is well to consider that a bad heat of steel made by the basic process may be due to ignorance and inexperience in new methods, while such an excuse is out of place when applied to the Bessemer practice, which has been developed for a generation.

Sec. 3. — Combination of the Open Hearth and the Converter,

As above observed, open-hearth furnaces sometimes have been run entirely on pig-iron. This practice results in a greatly-diminished output. Attempts have been made to hasten the operation by charg- ing melted pig direct from the blast-furnace. One disadvantage of this is the uncertainty concerning the composition of the charge. Other troubles arise from the long intermissions when only one or two open-hearth furnaces are operated. Much labor is avoided by this practice, but otherwise there is little gain. When solid pig-iron is charged, the conditions are favorable for the oxidation of the metalloids during melting, and in this period much of that work is done. In a liquid charge none of this preliminary oxidation has occurred, and the time necessary for its completion in the furnace cancels a part of the expected gain.

The foregoing considerations have led to some experiments on the use of a Bessemer converter as an auxiliary. Theoretically, nothing is more attractive than a plant with a blast-furnace to produce the pig-iron, a converter to desiliconize and partially decarburize it and an open-hearth to finish it into steel. Unfortunately, practical diffi- culties limit the applicability of this plan. A few of them may be summarized as follows:

.(a) When the blast-furnace is making sulphurous iron the entire

The Open-Hearth Process. 355

8teel*plant has to be stopped. This may be obviated by having several blast-furnaces and a mixer. The use of cupolas allows the selection of the pig-iron but entails considerable additional expense. (5) The converter, to be worked economically, must be run con- tinuously in order to pay for the large investment in hydraulic machinery, blowing- engines, bottom-making- and drying-establish- ment and the other incidentals of a Bessemer-plant.

(c) If the converter is small, several heats will be necessary to make one open-hearth charge of reasonable size. A furnace should have a capacity of at least 15 tons, while a converter of that size is part of an expensive plant. If three or four converter- heats are necessary to make one open-hearth charge, and if only one vessel is used, it is evident that more than an hour will be required to blow the necessary metal and allow time for pouring and recharging. This fact, and the trouble caused by changing bottoms during the preparation of a furnace-charge, make a second vessel very desirable.

(d) The waste is augmented by this duplex system. In the ordi- nary methods of open-hearth work there is a reduction of the iron contained in the ore, and this helps to make good the loss in weight due to the burning of silicon and carbon. When these elements are oxidized in the converter they represent a total loss. Furthermore a considerable amount of metal is projected from the vessel-mouth in the form of sparks, and more shot is contained in the viscous con- verter-slag than in open-hearth cinder. Ordinary Bessemer practice shows a waste of 10 per cent. It is doubtful if this would be ma- terially lessened by stopping the blow (for the duplex system) when the metal contained 1 per cent, of carbon, for most of the material ejected from the mouth of the vessel is lost during the combustion of the first ix)rtions of carbon. The loss in shot, also, would be fully as great, since the slag would be even more viscous, and the only gain would be the actual carbon-content. If no carbon be removed in the converter, the duration of the heat in the open-hearth furnace will be nearly as long, and the work nearly as costly, as though the metal had not been blown at all. On the other hand, if carbon be entirely eliminated, there will be difficulty in producing a slag in the hearth, and it is impossible to prepare the charge properly unless a suitable cinder covers the metal. In basic practice the metal must be allowed to remain in the furnace long enough for the removal of phosphorus. With a moderate percentage of this impurity present, there should be at least 1 per cent, of carbon also, so as to afford sufficient time for the production and action of a good slag.

356 The Open-Hearth Process.

(e) It is quite possible to eliminate all the silicon in the converter ; but with the use of direct-metal from the blast-furnace, and a demand for a hot product in order to avoid skulls, and with the stopping of the blow before all the carbon is burned, the silicon would often exceed one-half of one per cent, in the blown metal. This in itself would not be disastrous, but it would trespass seriously upon the clear ground of theory and would reduce the value of the work done in the converter.

(/) In combination with a basic open-hearth furnace, the desili- conizing of a phosphoric iron in an acid converter is very ad- vantageous, as the basic hearth is thus relieved of the element which produces most of the scorification, and is the cause of most of the lime additions. The installation of the duplex process in such a case would do away with the demand for low-silicon pig-iron, and so im- prove the chances for the elimination of sulphur in the blast-furnace.

The conditions above enumerated show that the combination of the converter and the open-hearth furnace is perfectly feasible. Only on a large scale, however, can it prove a commercial success, and even on this basis it is more attractive to the theorist than to the practical metallurgist.

Chapter II. — The Regenerative Furnace and its Machinery.

Sec. 4. — Construction of the Furnace.

The process about to be discussed is conducted in a regenerative furnace, so constructed that no solid fuel, except perhaps soot, comes in contact with the metal upon the hearth, and the heat of the waste- products of combustion is used to raise the temperature of the air and gas before their entry into the combustion-chamber. Possible variations in construction are innumerable, but experience has de- termined within certain limits the conditions of successful practice. The furnace proper, or working-room, is constructed of the most refractory silica bricks. The walls are made at least 13 inches thick to prevent loss of heat. The roof is usually made only 9 inches thick, that radiation may keep its temperature below the point of soflening. This danger is much overrated ; with a proper construction in other respects, the roof may be made of any desired thickness. The use of silica bricks involves a system of strong binders to resist their enormous expansive force during the heating

The Open-Hearth Process. 357

of the furnace. Expansion cannot be prevented, but it can and must be controlled, so as to keep the roof and walls in proper shape. In vertical interior walls, where binding-rods are impracti- cable, space must l)e left; for the bricks to push up on heating. The whole structure should be bound together in the strongest possible manner; for a crack once formed is immediately filled, or partly filled, with dirt or pieces of brick from near the line of rupture, and the next alternation of strain uses this filling &s a fulcrum and the crack extends or a new one starts. All theoretical lines of force are disregarded, and deformation seems an inborn necessity which even- tually compels the entire rebuilding of the structure. This event may be postponed for a long time by a strong and rigid system of binders, and also by making all arches with as small a radius as

The furnace proper should measure at least 15 feet in the clear between the ports, to give room for the complete combustion of the gas. In large furnaces a length of 40 feet is not excessive. The foundation should be entirely independent of the chamber-arches, and the hearth should rest on steel beams, so placed that there is a free circulation of air around and under them, and that any metal breaking out can run off without causing damage or delay. These rules are not always rigorously observed by constructing engineers, but every departure from them is a mistake.

The proper size or capacity of the furnace is entirely a question of convenience and ex|)ediency. There is an impression among some persons, not well informed, that a large furnace does not make as good steel as a small one. This is an error. So far as the working of the furnace or the quality of the steel is concerned, there is no reason why five hundred tons shouM not be treated at once. The objection to an abnormally large furnace is entirely in the handling of the product. Even with a heat of one hundred tons the size of the casting-crane becomes rather formidable. This, however, is not a vital consideration. A more important matter is the cooling of the charge during the process of casting. If very small ingots are to be poured by the usual methods, the time required will be so great that the steel must l)e very hot at the beginning, which is bad, or very cold at the end, which is worse.

The main argument against large furnaces arises in the rolling- mill. When ingots are to be worked before they get cold, — and this practice is bound to become almost universal, — it is evident that a flood of one hundred tons of ingots at one time, occurring once in

The Open-Hearth Pboceb8.

LongitudiDal Section through Center of Furnace. E, E, Air Chambers ; F, F, Gas-Chambers ; H, Gas-Port ; I, Air-Port ; K, Furnace- Hearth; Lf Flues to Valves; M, M, Binding-Rods.

Fio. 2.

Cross-Section of Furnace through Line A B C D, of Fig. 1. A, Furnace-Hearth ; B, B, Gas- Ports; C, Air-Port; D, Furnace-Runner; E, Air- Chamber and Ports Leading to Hearth; F, Flue to Reversing- Valve ; H, Air-Re- versing-Valve; I, Gas-Reversing-Valve ; K, Gas- Box and Regulating- Valve ; L, Flue to Stack ; M, Stack-Damper; N, Binding-Rods.

The Open-Hearth Process. 359

ten hours, is not as desirable as lots of twenty tons coming every two liours. For this reason it is better to have five 20-ton furnaces than one of 100 tons.

Figs. 1, 2, 3 and 4 show different types of furnaces. Figs. 1 and 2 represent a common form. ]t is not an example of the best con- struction, but rather of the way in which the problem of foundation is often only partly solved. The sketch exhibits the system of re- generation by vertical chambers and butterfly-valves.

Figs. 3 and 4 show an entirely different type. The chambers are horizontal, the working-floor thus being brought down to the general level and the handling of material made easier. The furnace is ellip- tical and has a riveted steel-plate jacket. This construction of fur- nace and chambers was devised by Mr. H. W. Lash, of the Carbon Steel Company, Pittsburgh, Pa. In vertical regenerators the dy- namic action tends to uniformity of tem|)erature : the hot descend- ing gases naturally seek the coldest spots, and the cold ascending gases find the hottest places; in each condition there is a ten- dency to equalize the heat throughout a given cross-section with a consequent maximum efficiency. In horizontal passages there is no such equalizing process; on the contrary, the gases tend to travel through the upper part of the chamber to the exclusion of the lower part. The only advantage of this construction lies, therefore, in the accessibility of the charging-floor.

In the drawing only one port is shown, and the brick work form- ing it is enclosed in an iron frame-work, so built that the whole cage can be changed when the arches are worn out. This is an improve- ment introduced by Mr. C. E. Stafford, now with Shoenberger & Co., Pittsburgh, Pa. The furnace itself is placed upon rockers, and the tap-hole is above the slag-line. By tilting with a hydraulic cylinder, the whole, or any part of the charge, is poured out when desired. The end-joint is made by the abutting of two flat, water- cooled cast-iron rings, one being fastened to the cage and the other to the furnace. For these innovations the author is responsible, as well as for the valve-system which is given in detail in Figs. 5 and 6. This construction is adapted for either acid or basic work. The first two furnaces built after this pattern were started in 1890 with a basic hearth, and in nearly three years, up to the time of writing, have given no trouble whatever.

Another form, devised by the author and used successfully for sev- eral years, is made in the shape of a barrel, with a port at each end ; it siis on wheels and is capable of complete rotation around its longi-

The Open-Hearth Process.

A

O

- — B

f-

e

"Is

Go

£i

g-i if

fa O

S

It

ii Si

:|DigitizedbyGOQe

The Open-Hearth Pb()C£88.

tudinal axis, the end-joints l>eing fitted with flat water-cooled rings. It is first lined with a course of clay-brick and then, after being brought to a proper heat, sand is thrown in and set over its entire circumference in the same manner as hearths are constructed. When

Fig. 5.

Vertical Section through Gas-Reversing- Valves. C, Suck ; D, Main Gas-Tube; E, E, Branch Gas-Tubes, Showing Valves; F, F, Gas-Chambers ; H, H, Gas-Chamber- Flues to Re versing- Valves ; I, I, Stack Re- versing-Valves for Gas ; K, Flue from Reversing- Valves to Stack ; L, Stack-Damper for Gas; M, Valve-Reversing-Gear ; N, N, Water-Cooled Valve-Seats; P, P, Air- Chambers.

any part of the roof or sides is worn away, the furnace is put in sucli a position that the thin portion can be built to a proper thickness with sand. This work can be done during a twelve-hour shift at the end of the week, without taking ofl the gas or cooling the furnace, and without any labor except that of the melter and helper. The

The Open-Heabth Process.

building of the entire circle, so that it will not flake off, requires care and time; but the quality of the sand that can be used is far superior to any silica brick ; and since the lining can be made two or three feet thick, the loss by radiation is materially diminished. This construction allows the use of a high tapping-hole, but it is not adapted to basic work, as there seems to be no basic material that can be set by heat so as to overhang.

Fio. 6.

Horizontal Section on Line A, B, Fig. 5. A, Air-Inlet ; B, B, Air-Chambers ; C, Stack ; D, Air Re versing- Valve ; E, E, Gas- Inlets; F, F, Gas-Chambers; H, Stack-Damper for Air; I, Stack-Reversing- Valve for Gas ; E, Flue from Reversing-Valve to Stack ; L, Stack-Damper for Gas ; N, N, Water-Cooled Valve-Seals.

The Pemot Furnace. — Another form of construction, which was heralded as a triumph at its advent, is the Pernot furnace. The hearth is placed upon rollers, and may be revolved in a plane slightly inclined from the horizontal. At the time of its introduction, one of the main advantages claimed for it was the rapid working which would result from the revolution of the hearth. Being set at a slight angle, rotation brings all parts of the bottom to the surface, pushes up into the flame all unmelted lumps of stock, and causes a thorough agitation of the metal afler fusion. Experience soon showed that the

364 The Open-Heabth Progeb8.

hastening of work was dearly paid for in a waste of iron and a soori- fication of the bottom, so that in furnaces of this type, now in ope- ration in this country, the hearth is not revolved while the charge is melting, except to bring unftised areas beneath the flame. It is ro- tated to facilitate the work on the tap-hole, as a partial revolution raises the position of the outlet so that slag and steel drain away from it ; it is also moved during charging, thereby reducing considerably the labor involved, and it is kept in motion after the addition of the recarburizer, to facilitate thorough mixing. The value of this last point is not to be disrarded, but it is apt to be overrated by advo- cates of this type of construction. ,

Against this furnace must be urged the difficulties attending the construction of the ports, the care of the ever-o|)en joint, and the impossibility, with any existing form, of repairing the hearth in the interval between charges without removing the bottom from its posi- tion beneath the roof. From this last difficulty great danger arises of the steel burning through the bottom, and this is especially to be feared on account of the machinery which is placed beneath the hearth. Although furnaces of this class have been run successfully for years, and are in operation to-day, yet the author ventures to express the opinion that their application to open-hearth work is a thing of the past.

Sec. 5. — Method of Operation,

Through suitable valves the gas enters a renerative chamber built of brick, with thick walls, and filled with brickwork so laid that a large amount of heating-surface is exposed, while at the same time free passage for'the gas is assured. The air enters a similar chamber. As the bricks in both chambers have been previously heated, the gas and air, in passing through them, are raised very nearly to the temperature of the surfaces with which they have been in contact. The air and gas meet just before reaching the melting* chamber, and produce a flame which travels over the hearth. The resulting products of combustion pass to the stack through chambers similar to those through which the gas and air entered. After the brickwork in the first set of chambers has been partially cooled by the incoming gases, the currents are reversed by means of suitable valves, and the gas and air enter the furnace by way of the chambers, which have been heated by the products of combustion. By the repeated reversal of the currents, the chambers are kept at a tem-

The Open-Hearth Process. 365

peratnre of about 1200° C, and this, therefore, is the temperature of thte gas and air entering the combustion-chamber.'*'

The working of the furnace depends very much upon the arrange- ment of the ports through which the gases come and go. The gas should enter below the air, because, being the lighter, mixture is thereby facilitated, and also because this arrangement does not expose the metal on the hearth to a stratum of air and cause excessive oxi- dation. The point where the two gases meet should be about 6 feet from the metal ; if much less than this, combustion can hardly begin before it is checked by contact with -the cold stock ; if much more, and if the burning mixture is conducted between confining walls, the brickwork will be rapidly melted. Some diflSculties are over- come by making the roof extremely high and keeping the flame clear of the metal, trusting to radiation for heat. All other things being equal, the oxidation in a furnace of this design will be reduced to a minimum, and for some purposes this is a most important con- sideration. In certain cases it has been found that the time.of the operation is lengthened by the high roof. This would naturally be expected, and where it is not the case, a comparison may possibly have been made with a bad previous practice, in which the gases rolled aimlessly into the working-room and into direct contact with the charge.

Sec. 6. — Dynamic Equation of the Gases.

Both gas and air should enter the combustion-chamber under a positive pressure, forcing them into contact with each other, and throwing the resultant flame across the furnace in such a way that the draught of the stack on the outgoing end can pull it down through the ports without its impinging against the roof. A prev- alent idea among furnace-men is that the draught of the stack pulls the gases into the furnace. This h entirely wrong. Some furnaces may have been built in that way, but they are beneath considera- tion. A slight outward pressure at the door-openings is essential to good work, and such a pressure cannot possibly be caused by a pulling action of the stack. When the vertical distance from the port to the flue where the gas or air enters, is 15 feet or more, the force of the upward push due to the hot brickwork will create suflS- cient pressure; but with furnaces that are surrounded by a working- floor on the {neral level, the ascending power of the currents will

For a farther discDBslon of this point see chapter iv., page 3S7.

Ic

The Open- Hearth Process.

be very weak anless the chambers are sank to anasnal depths. There may not be much difficulty with the gats, since this is often under a slight pressure when it enters the valves ; but it will be found advan- tageous to force the air with a blower. The difficulty in making the gases enter the furnace will be more serious toward the end of a run or campaign, since the spaces between the checker-bricks become more and more filled with a deposit brought over by the products of combustion. This deposit is composed of the dust of material thrown into the furnace — sand, dirt, rust, scale, lime, ore, etc., and also of the fine globules of metal projected from the bath by the violence of the carbon-reaction, which particles, being immediately converted into oxide of iron, are swept along toward the stack. The deposit, therefore, consists largely of very fine iron oxide.

The respective areas of the gas- and air-ports must be proportioned to the kind of gas used, as the composition of the gas will deter- mine the required volume. The proper amount of air and the vol- ume of the products of combustion will not vary directly with the quality of the gas. The following is the result of calculations on certain gases, taking as a basis the volume of each required to pro- duce one thousand calories when perfectly consumed. The first is a fair sample of soft-coal producer-gas; the second is pure CH repre- senting natural gas; the third is an equal mixture of CO and H, representing perfect water-gas.

(1) Prod'oeer-Oas :

Calorific valae of 1 cu.ra. of the gas 1170 calories. Hence 1000 calories are developed by 0.855 cu.m. of gas, which requires for com- bustion 0.869 cu.m. of air ;

/ CO,, 0.267 cu.m. producing 11,0, 0.183 cu.m.

In, 1.202 cu.m.

Total products of combustion, 1.602 cu.m.

Calorific value of 1 cu.m. 8480 calories. Hence 1000 calories are developed bj 0.118 cu.m. of gas, which requires for combustion 1.126 cu.m. of air ;

rCO„ 0.118 cu.m. producing H,0, 0.236 cu.m.

i N, 0.890 cu.uL

Total products of combustion, 1.244

Co, 22.81

[)ercent.

H, 8.5

K

Ch, 2.4

H

C,H,. 0.4

u

C0„ 6.2

u

O, 0.4

It

N, 60.3

it

Total

(2) Natural Oas :

Ch„ 100 1

aer cent.

The Open-Heakth Process. 367

(3) Water Oas:

CO, 50 per cent Calorific value of 1 cn.m. of the gas 2830

H, 50 calories. Hence 1000 calories are developed

by 0.353 cu.ro. of gas, which requires for

combustion 0.842 cu.m. of air; rCO,. 0.177 cn.m. producing H,0, 0.177 cn.ra. In, 0.665 cu.m.

Total products of combustion, 1.019 cu.m. Summary,

For development of 1000 calories. Air required. Combustion-prod 'a. cu.m. cu.m.

Producergas 0.869 T.602

Natural gas, 1.126 1.244

Water-gas, 0.842 1.019

At first sight, the volume of the products of combustion would appear to be a matter of no practical importance; but it must be re- membered that all these products escape through the same ports and passages that admit the gas and air, and also that the entering gases are in compression, while the outgoing products are under tension, and that being also heated to a very much higher tem- perature, their volume is considerably expanded. Since the work- ing of the furnace will be determined by the amount of fuel that can be burned, it follows that the facilities provided for the escape of the burnt gases will be oue of the main factors in its success.

Sec. 7.— The Hearth.

The part of the furnace containing the hearth should be built of steel plates with tight riveted joints. Every bottom, when broken up after long use, shows that melted metal has penetrated through cracks and found its way to the enclosing shell. If such wandering steel comes in contact with a cool plate, it spreads in a thin film until it fills the vacant space and then chills; but if in the area covered (which may possibly be several square feet) there is an unclosed rivet-hole, a continuous stream of steel may be started from the bath, that will enlarge its channel with startling rapidity and in a few minutes empty the whole heat upon the ground. For the same reason, the shell is carefully lined with layers of bricks with joints broken, and on this lining the true bottom is built.

In acid work this bottom is made by spreading sand in successive layers and hardening each layer by exposure to a full working-tera-

368 The Open-Heakth Pboce88.

perature. In basic work, the bottom is sometimes built up in this manner of magnetite or dolomite, but generally it is rammed or laid in while the furnace is cold.

In the ordinary stationary furnaces, the tap-hole is of necessity at the lowest point of the bottom, and must be closed by refractory material so mixed and set that it will not break open during the melt, yet soft enough to allow a bar to be driven through it to tap the charge, and solid enough to resist erosion while the metal is run- ning out. In basic work est>ecially, this combination is not easily attained. In the construction shown in Figs. 3 and 4, the tap-hole is placed above the slag-line, and the charge is poured by tilting the furnace op its rockers. The bath of metal, when melted, should be from 15 to 24 inches in depth. If it is shallower, the oxidation is excessive; if deeper, the time of melting and working is prolonged.*

Sec. S.—The Valves.

The admission of gas and air to the chambers is regulated by some simple form of throttle-valve. In addition, re versing- valves are necessary to direct the course of the currents. The itinerary of the journey made by the gas and air separately is as follows: through the regulating- valve and through the reversing-apparatus into a regenerative chamber, and thence into the furnace; then jointly, as products of combustion, from the furnace into the second set of regenerative chambers, and through the reversing-apparatus into the stack. For the complete control of this system, the ordinary four- way butterfly-valve (shown in Fig. 2) is theoretically a perfect mechanical contrivance. Its simplicity, its neatness, the ease with which it is manipulated and the small space which it occupies, have led to its general adoption and to an equally general unwillingness to recognize its radical and irremediable defects. From the nature of the case, it is exposed on one side to the incoming gases and on the other to the escaping products of combustion. The waste-gases should not be hot in ideal work, but, unfortunately, circumstances will occur in practice whereby for a short time they pass to the stack at a temperature of redness. The inevitable result is a warping of the valve or its enclosing box. The same result may be effected by a leakage of gas, if soot or tar should clog the seat unequally, or if air should leak in and burn the gas in the valve-box. When once the valve is warped, the destructive action is cumulative, since gas

See Sec. 22, page 397.

THE oVeN-HEARTH PROCESS. 369

leaks and burns continually at the opening. It is a most objection- able feature that this leak does not show itself openly in any way ; but more vital is the fact that there is no way of remedying the dif- ficulty without complete replacement of the valve or box. Adjust- ment is impossible when a fixed pivot and two metal seats are in question. Water-cooling has been tried with some success, but it should be applied to both the box and valve. The better way is to construct an entirely different type of valve, shaped to withstand unequal heating and provided with means by which the injured parts may be easily replaced. Many such arrangements have been devised. Fig. 5 shows a form designed by the author; it is quite possible that in some other comer of the world it has been used liefore.

In addition to the valves which determine the influx of the gases, dampers are provided to regulate the flow of the waste-products. Not only must the total pulling-power of the stack be controlled, but there should be some arrangement by which the gas- and air- chambers can be governed separately. This is necessary because the amount of waste-products passing through any particular chamber determines the temperature of that chamber, which in turn deter- mines the temperature of the gas it delivers to the furnace. During the run of a furnace it often becomes necessary to alter the relative amounts delivered to the various outgoing passages; but as tight seals are unnecessary, no complicated apparatus is required.

Sec. 9.— The Ladle- Crane.

For handling the metal and slag after tapping, various systems are in use, and no particular one can be declared unconditionally the best. The weight to be handled, the sizes of ingots to be made, the number of furnaces, and the available room will all be factors in the problem. To the mechanical engineer the travelling-crane seems the true solution, but the metallurgist condemns it, as the multiplicity of parts makes breakages too frequent. In the ma- chine-shop such accidents may not be serious, but in an open- hearth plant they may be disastrous. If a charge cannot be tapped' when it is ready, the metal oxidizes, scorification of the bottom ensues, and the whole heat may break out and be lost. Relief may be oB-* tained by adding pig-iron, but it is only temporary, and if the metal is allowed to chill it may be given up as lost, so far as a successful cast is concerned. In such a case most furnace-men would lihe pour it into a bottomless pit. In case of a breakage of the support- ing chains with a charge in the ladle, the possibilities are appalling;

VOL. XXn.-24

370 The Open-Hearth Process.

25 tons of steel flowing over tracks and moulds, around ladles and engines, fastening them inseparably together and chilling into an unbreakable mass — all this is beneath consideration when the danger to human life is counted.

Confronted with these conditions, the metallurgist has used the hydraulic crane, and each year has made it safer and stronger. The cylinder should be large enough for twice the load it will ever be called upon to lifl, and the superstructure must be able to withstand the utmost pressure of the cylinder. Struts should be used as far as possible instead of tie-rods, and all the parts should be made of rolled-steel shapes. In such a crane the only place of weakness is the connecting water-pipe, and the breakage of this involves merely a slow descent of the load.

The introduction of an electric traveller, with a hydraulic lift and pump carried on the bridge, offers for the first time a combination of the above-described elements of safety, with the evident advantages of the travelling-crane class.

Chapter III.— Fuel..

Sec. 10. — QuaMy of the Gas Required, It has been stated that the fuel used in the furnace must be a gas. The quality, within reasonable limits, is not of vital importance, provided the quantity be sufficient. Ordinary producer-gas contains over 60 per cent, of non-combustible material, and yet is all that can be desired so far as thermal power is concerned. Certain substances, such as sulphurous acid and steam, are objectionable; but this arises rather from their chemical action upon the metal than from any in- terference with calorific development. With coal of ordinary quality, sulphur causes no trouble, but when it is present in large amounts it is absorbed by the steel. The injurious effect of free steam is a matter for debate. With gases containing much moisture the oxida- tion of the metal will be unquestionably greater than with a dry, carbonic oxide flame; but it is difficult to see why it will be greater when the moist gas is burned than when an equal amount of water is produced by burning a quantity of hydrogen. If steam is so baneful as is sometimes alleged, the use of water-gas or any highly hydrogenated fuel must be a mistake, but it is quite possible that the bad reputation of steam is due to the gases with which it is usually associated, and that the effects of its presence in producer-gas could be caused also by a corresponding increase of the hydrogen-content.

The Open-Hearth Process. 371

In manj places, increased oxidation by the flame during melting is considered advantageous, as it allows the use of a larger percentage or pig-iron. While the charge remains high in carbon, and oxida- tion is the one thing desired, it would seem immaterial whether the result be obtained by the use of ore or by the flame, so far as the effect upon the metal is concerned. During the last stages of the operation, however, the conditions are changed : a perfectly neutral flame is then desired, so that the metal shall be in an environment similar to that of steel in a crucible during the attainment of a dead melt." In practice, such a flame cannot be obtained for any length of time, but the conditions are more nearly approached in burning carbonic oxide than in burning hydrogen. By proper attention, the difficulties attending the use of an inferior fuel can be largely over- come, and the question is narrowed to a consideration of relative cost. From the standpoint of general use, bituminous coal ranks first; natural gas rules in a favored but limited district; petroleum is available over a wider area, while anthracite, coke, lignite, peat, wood, and even sawdust, have been the reliance of successful fur- naces.

Sec. 11. — Bituminoua Coal: The Siemens Producer.

Soft coal can be converted into gas for metallurgical purposes by burning it in a thick fire, the products being carried through tubes to the furnaces without the intervention of a receiver. The forms of apparatus devised for this work are many, difiering only in detail. Fig. 7 shows the Wellman producer, the design of one of our most practical engineers. It is a modified form of the old Siemens type, with a steam-blast attachment, and fulfils the following requirements, which every good producer should meet:

(1) Each fire is independent, and leakage from one to another cannot occur. This renders it possible to repair any particular pro- ducer at any time.

(2) It is constructed in the strongest possible manner. A circu- lar form is the easiest to in shape, and a continuous enclosing envelope of rolled steel canncfc be improved upon.

(3) The coal is dumped in the center of the fire, and poking is re- duced to a minimum.

(4) The stepping-back of the brickwork from the grate-bars par- tially overcomes the tendency of the air to creep up the walls.

(5) Ample provision is made for putting in a false grate by push- ing bars through the fire above the ashes in such a manner that they act as a grate while the clinkers below are removedpigjjg byGoOQlc

The Open-Hearth Process.

(6) During this operation the fire must be separated from the main receiving-tube, else the gas will back down through the fire, on the workmen. The isolated position of each producer renders it easy to provide a separating-damper for each fire.

(7) By means of the steam-jets, the relative amounts of air and steam can be regulated to obtain the best results.

Fio. 7.

Vertical Section through Center of Wellman Gaa-Producer. Ash-Zone; B, Combustion-Zone ; C, Distillation-Zone; D, Steam-Blower ; E, BearinfjT.Bar ; F, Grate-Bar; H, Hopper; I, Stoking-Hole ; K,Neck to Gas- Main ; L, Door for Cleaning Neck ; M, Damper ; N, Gas-Main.

(8) The gas is taken from the side of the cylinder, the top thus being left clear, so that every part of the fire can be reached with a poker.

Sec. 12. — The Action in the Producer.

In a producer in operation we may consider three different zones :

(1) The ash-zone.

(2) The combustion-zone.

(3) The distillation-zone.

The first is theoretically of little importance. Its sole function is to heat the air and steam before their contact with the burning carbon. In the actual conduct of the process, however, the existence this zone must be borne in mind. The fire is cleaned at least

The Open-Hearth Process. 373

twice a week, but during the intervals the ashes accumulate to such a thickness that the fuel-bed must be deepened to allow room for the zone of combustion ; otherwise, abnormal amounts of carbonic acid and steam will pass through it undecomposed.

In the zone of combustion the air and steam meet the carbon and the following reactions occur :

C + 20 COj CO, + C 2CO H,0 + C <X) + 2H

The zone of distillation comprises the upper part of the fire. The addition of fresh fuel tends to lower its temperature, but the heat of the combustion in the region beneath distils the volatile components.

The zones of combustion and distillation are not as distinct in practice as the above division would indicate. The complete dis- tillation of coal is attained only after long exposure to a high tem- perature, owing to its tendency to coke into large masses which are not easily broken. (Jonsequently, lumps of coke frequently pass through the fire and drop from the grate while still retaining some volatile constituents, and in the zone of combustion there is not only a maFS of burning coke, but also considerable coal that has been only partially carbonized. Thus it happens, that while a certain proportion of the gases of distillation escape from the top of the fire without being subjected to intense heat, the remainder is exposed to the high temperatures and chemical forces of the zone of combus- tion. A series of complicated processes of dissociation and syn- thesis is the result, in which the length of time and dree of heat are important factors. Under otherwise equal conditions, a hot and deep fire will best promote the desired reactions, but even with a bed of incandescent coke ten feet in depth, the escaping gases will contain CX),. In practice, the depth of the fire cannot be much over six feet. If deeper, it is found that, no matter how thor- oughly the upper surface may be stirred, the lower part of the fire is not thoroughly broken up by the poking, and the zone of combus- tion becomes honeycombed with large cavities, affording passage for undecomposed air and steam. This condition is most marked along the walls, and the intense beat produced on the interior sur- faces of these coke chimneys causes clinkers to fuse to the brick- work. Practice therefore demands that the thickness of the fire be limited, that steam be used to avoid extreme temperatures, and that the mass be kept thoroughly broken by frequent poking.

374 The Open-Heabth Pboce8S.

Sec. 13. — The Composition of Siemens Oas. Under the above coDditions, the producer-gas will have about the following composition :

Per cent- by vol.

Co,. . 5.2

C,H, 0.4

Co, .22.8

H, 8.5

Ch„ 2.4

The presence of O is due to the fact above mentioned, that con- siderable time is required to complete some of the reactions. The gas, as a whole, may be regarded as a quantity of combustible mate- rial, mixed with a solution of oxygen in nitrogen. When the amount of oxygen is reduced to less than one-half of one per cent., the solution is so weak that chemical action ceases.

The presence of CO, may be ascribed to similar causes. Steam is more inert, and requires a much higher temperature for its dissocia- tion. C2H4 is derived from the zone of distillation ; at high tem- peratures it readily resolves itself into C and CH, its elemental components. CH is likewise dissociated by heat, but the action is slow. The H is derived from the breaking up of some of the hydro- carbons, CH4, C2H4, etc., and from the decomposition of steam. The CX) is the result of the reaction CO, + C 2C0. Besides the nitro- gen of the air used in the partial combustion, steam is present in quantities varying with the moisture of the coal and the condition of the fire. This steam never appears in ordinary analyses, since it is condensed in taking the sample. It is possible to estimate it by conden-sation in a separate tube, but precautions should be taken first to filter out the tar, which, otherwise, would collect in the condenser. This filtration, in turn, must be so conducted that water is not de- posited with the tar. Sulphur is present, either as sulphurous acid or sulphuretted hydrogen, or, possibly, it may be distilled uncom- bined. Certain coals produce also large quantities of tar, which are deposited in the conducting tubes. Part of this product, as well as some rich hydrocarbons, are decomposed by the heat of the fire, forming large quantities of soot. When the fires are very hot, the amount of soot is three times as great as when they are cool, and it may reach five cubic feet for every ton of coal consumed. This necessitates expensive and troublesome delays for cleaning the tubes.

The -Open-He A Eth Process. 375

In Sec. 15 it is claimed by Blauvelt, and the argument has also been elaborated by Taylor {Trans., xviii.,* p. 868 ei aeq.), that the un- stable hydrocarbons contribute very essentially to the calorific value of the gas. Both maintain that chemical analysis, as usually made, does not represent the true equation, since these rich and powerful con- stituents are condensed in the receiving-tube, and do not appear in the result. There are some good grounds for tliis belief, as the writers show, yet there is also reason for supposing that the influence of these condensable components has been overrated. Thus, Taylor says that to obtain satisfactory results producer-gas should not be cooled below SOO" F. (149° C), whereas it may be stated as a fact, that a five- ton furnace has been run with gas passing through four hundred feet of exposed iron pipe, four feet in diameter, and that it was cooled to 40® C. where it entered the flues. The fuel-rate was not obtainable under these conditions, but there was nothing to indicate that an excess of gas was used, and the furnace turned out four heats in every twenty-four hours from cold stock.

In Sec. 14 f certain records of fuel-consumption are given for a case where the gas was cooled to about 150° to 200° C. Under these conditions the following experiments were made to determine the value of the tarry constituents.

Experiment 1. — A sample of 170 liters (0.170 cu.m.) of gas was taken from the tube as it entered the valves at the furnace. To assure a fair average, the sample was drawn off very slowly, daring more than twelve hours. The tarry matter was condensed in cool tubes and weighed after the water had been evaporated by a current of dry air.

Gaa after Per cent,

extraction of tar. by volume.

C,H4, 0.2

O, 0.6

Cx), 22.8

H, 5.8

Ch4, 3.2

N (b/ difference), 63.4

Calorific value of the cooled gas per cubic meter, 1139 calories. " of 0.170 cu.m.. 194

Tar 2.10 grammes 0.0021 kilo.

♦ Infra, p. 380. f fra, p. 378.

/

The Open-Hearth Process.

Assuming the compoeition of the tar to be C, 92.6 per cent., II, 7.4 per cent, (see Experiment 2), we have :

Calories. Carbon, 0.001945 kilo at 8080 calories per kilo . . 15.716

Hydrogen, 0.000155 kilo at 29,633 calories per kilo . 4.593

Calorific power of the tar, 20.309

Calorific power of gas (as above), 194.000

Total calorific power of gas and tar, . . . . 214.309 HeaenergJr in the tar 9.5 per cent.

Experiment 2. — A sample of 170 liters of gas was taken in the same way and place as in the first experiment, bnt the tarry con- stituents were collected by filtering the gas through a tube filled with asbestos. After drying the residue in an air current, a combustiou was made, and the resulting CO, and Hfi were weighed. The short tube through which the gas had been conveyed from the box to the filter in taking the sample, was put bodily into the main combustion- tube, so that all tar and soot which belonged with the gas-sample should appear in the result.

Gas after Per cent,

extraction of tar. by Tolome.

Co, 22.0

H, 6.4

Ch4, 3.6

N (by difference), 62.0

Calorific value of the cooled gas per cubic meter, 1193 calories. " of 0.170 cu.m., . . . . 203 "

Products of combustion of the tar :

Grmmmet. Per cent.

Carbon, . . . 2.6870 92.6

Hydrogen, 0.2142 7.4

2.9012 loao The total calorific value will be :

Calories. Carbon 0.0026870 kilo at 8080 calories per kilo . . 21.711

Hydrogen 0.0002142 kilo at 29,633 calories per kiIo= . 6.347

OOorific power of the tar 28.058

Calorific power of the gas, as above, 203.000

Total calorific power of gas and tar, . . . . 231.058 Heat -energy in the tar 12.1 per cent.

The Open-Hearth Process. 377

Thus it appears that bj the scrubbing-process the gas was robbed of only one-eighth of its calorific power. It is true that this does not represent all the material which was suspended in the gas as it left the producer, but it represents what is more pertinent to this inquiry, namely, the suspended matter in the gas as it actually entered the valves of furnaces in regular and economical operation. When it is shown that only one-eighth of the calorific power of such a gas is due to sus|)ended hydrocarbons, it would seem incredible that complete scrubbing would give a worthless fuel, and experiments indicating the contrary may have been controlled by extraneous conditions.

The deposition of tar and soot in the tube before reaching the valves has nothing to do with this argument. It will influence, however, the consumption of fuel per ton of product. It would be possible to collect the gas from the top of the fire and determine the total hydrocarbons by filtration ; great care would be required in obtaining a reliable sample, for the frequent coaling and poking are continually altering the equation of any one fire, and it is impossible to preserve a representative condition even for an hour. In the above experiments, where the gas was collected at the valves, it came from twenty fires, the products of which became thoroughly mixed in their passage through the tube, thus giving a fairly regular product, while the long period over which the sampling extended insured a true average.

The amounts of tar and soot deposited in the conducting-pipe may be ascertained by the quantities removed from time to time in cleaning the tube. No records of the weight of the soot are at hand ; but during a run of eight months, and a consumption of 8000 tons of coal, the tar collected amounted to 44,000 pounds, or 5.5 pounds per ton of coal, which is 2.5 kilos per 1000 kilos of coal. The calorific value of the tar, computed as in the second experiment is:

Calories.

0.0021 kilo of tar, 20.309

1 " " 9,670.

2.5 " 24,175.

The thermal value of the coal given in my former paper {JVana,,

xix., p. 182) was 8198 calories per kilo; hence 1000 kilos

8,198,000 calories, and the calorific energy in the condensed tar is

24,175

three- tenths of 1 per cent, of the total.

would

8,198,000

Even if all this tar could be kept in the gas, it would not be a

The Open-Hearth Process.

clear gain for when rich hydrocarbons come in contact with hot checker-bricks, they are decomposed and carbon is deposited. This deposit of carbon, if small in amount, is burnetl on the next reversal of the valves, but its energy is lost through the stack.

The soot, which probably does not exceed the tar in calorific value, is still more troublesome in choking the valves and regenerators, so that nothing would be gained by preventing its deposition in the conducting-tube.

Sec. 14. — The Thermal Equation of the Producer.

In the gasification of bituminous coal there is a loss of thermal energy from several causes. The value of each factor, as well as the sum of all the factors, will vary with every kind of coal and every detail of manipulation. The investigation of a certain par- ticular practice gave the following result:*

Total heat-value of the coal, 8198 calories per kilo.

Per Cent, of Calories. Total Heat Generated. Potential heat lost in ash or carbon, . . . 4S3 5.9

Latent heat of volatilization, 600 7 3

Decomposition and heating of steam, . . . 356 4.3

Radiation and conduction, 355 4.3

Sensible heat of gas, 1085 13.2

Total loss of heat, 2879 35.0

All these losses except the last would occur in a system of direct heating. The sensible heat of the gas is regarded as a total loss, since a rise in temperature at the entrance-flue of the furnace means a similar and equal rise in temperature for the products of combus- tion escaping into the stack. It is therefore important so to adjust the calorific work of the producer that the heat developed is util- ized in the heart of the fire and the escaping gases are kept as cool as possible. This is usually accomplished by blowing steam through the fire and using all surplus heat for its dissociation. It must be remembered that the cooling of the upper part of the fire by steam from the grate implies cooling of the zone of combustion to the same degree, so that the utilization of the sensible heat of the upper sur- face of the fuel involves the presence of an increased amount of un- decomposed steam in the gases.

See chapter v., page 395, and also, for details of the calculation. Trans., xix., p. 136 et seq.

The Open-Hearth Process. 379

Some engineers advocate — with plausible and, at first sight, con- clusive reasons — placing the producer near the furnace, under the impression that thereby they save the sensible heat of the gas. It is true that when the gas is hot, less heating of the gas-chambers is re- quired, and hence less checker-work will suffice; hut this is a small matter, having no bearing on the fuel-economy. Whatever Is gained by hot gas at the incoming end is lost on reversal in the outgoing products of combustion. Moreover, a special system of valves must necessarily be constructed to handle hot gases : ordinary valves soon warp and leak, and water-cooling is not to be thought of, for this involves chilling the gas, which is manifestly opposed to the intent of the practice in question.

From these facts and the data presented in Sec. 13 it appears :

That soot is objectionable in the regenerators.

That some of the tar which reaches the chambers is not utilized.

That even when the gas is thoroughly cooled, the deposition of tar is trifling (representing in one case only three-tenths of 1 per cent.. of the total calorific power).

That the tar deposited by condensation can be utilized as fuel elsewhere.

That cool gas is very desirable for the preservation of valves and dampers.

That hot gas does not tend to economize energy, since the loss of heat in the escaping products of combustion offsets the apparent gain.

It remains for the advocates of hot gas to show that the consump- tion of coal is less when the producer is near the furnace, than when it is far away. In the two months just elapsed the records of two 25-ton acid furnaces, where the gas travels 250 feet before reaching the valves, show a production of 6260 tons of steel, and a consump- tion of 827 pounds of coal per 2240 pounds of product. This is by no means a minimum. Two other consecutive months can be named showing 810 pounds and one month only 783 pounds. These are not results of specially selected periods. They are working figures, and there is no shortage to be allowed for. In making comparisons it should be known that, owing to the construction of the furnace, two-thirds of the products of combustion escape at red heat, and that the coal contains from 10 to 20 per cent, of ash. Furthermore the tar which condenses in the gas-tube figures as part of the fuel- consumption, when, in fact, its energy is utilized by burning it in other furnaces.

380 The Open-Hearth Pbocebs.

Sec. 16. — The Manufacture of Producer- Oaa from Afiihraciley and

a Comparison of AfUhracUe and Bituminous Gases,*

By W. H. Blauvelt.

AVhile anthracite is the ideal fuel for domestic purposes and steam- making, to which the clear atmosphere of our eastern cities is witness, yet for metallurgical purposes it has never been as much favored as bituminous coal. This has been particularly the case in operations requiring gaseous fuel, where the attempt to use anthracite promptly developed difficulties in both the manufacture and use of the gas, which do not appear where sofl coal is used.

When producers were run by natural draft, it was necessary to use the larger and more expensive sizes of anthracite in order to permit the draft to pass through the fuel-bed with sufficient rapidity. At- tempts to use the small sizes, " buckwheat " and " rice," with a forced draft, have met with but partial success in most types of producers. One reason is, that when the small coal is incandescent, it will run through the smallest opening in a most persistent stream, occasioning a large loss through the grate when the producer is cleaned, and also injuring the quality of the gas for a considerable time. The grate, with its attendant losses of fuel and excessive expenditures of labor and time in cleaning, has been dispensed with in recent inventions, so that for several years there has been no difficulty in making from the smaller sizes of anthracite a good quality of gas, very low in carbon- dioxide and with but trifling loss of carbon in the ash. But the problem of successfully using the culm, or waste, from the breakers, by gasification in producers, seems to be as far as ever from practical solution. When smaller sizes than buckwheat are used, every attempt to force the rate of combustion to a reasonable speed results in de- flecting the steam- and air-blast to the walls, thus creating an annular zone of intense combustion, with an unburnt center, and producing only a small quantity of gas containing a high proportion of carbon- dioxide. (For a full discussion, see Trans.y vol. xx., p. 625.)

Assuming a gas of good quality, made from the cheaper sizes of anthracite, with an expenditure of labor no greater than is required for bituminous coal, let us consider the qualities of this gas and its

Tlie use of anthracite coal for the manufacture of gas has been the subject of much experiment. In the following memoir the subject is discussed by Mr. W. H. Blauvelt, who, for many years, was connected with the practical operation of the Taylor producer, the most successful apparatus in this field of work. It may be needless to say that some statements and theories of Mr. Blauvelt are not in accord with statements and theories advanced in the foregoing sections.

The Open-Hearth Process. 381

availahility for various purposes, especially in comparison with gas from bituminous coal. The first notable difference between the two gases is seen in their appearance while burning. Anthracite-gas burns with a bluish, almost colorless, flame, whereas bituminous-gas burns with a luminous flame, radiating considerable heat.

Analyses of the two gases do not give any indication of this dif- ference, as may be seen by comparing the percentages of hydrocar- bons in the following analyses, which are of representative gases from anthracite and bituminous coals respectively, made according to the best practice :

Per cent, by volume. Anthracite gas. Bituminous gas.

Co 27.0 27 0

H. 12.0 12.0

CH4 and CjH, 1.2 2.5

Co 2.5 2.5

N, 57.3 56.0

100.0 100.0

Theoretically, a gas made from a bituminous coal containing 30 to 35 per cent, of volatile matter should contain about 12 to 14 per cent of hydrocarbons, or say six times the amount shown in the above analysis. The deficiency in the analysis is explained by the fact (now l)etter understood than formerly) that when soft coal is gasi- fied, only a small portion of the hydrocarbon is converted into fixed gas, the remainder being merely volatilized and carried as vapor to the furnace. This vapor is condensed in the apparatus for samp- linjr, and therefore does not reappear in the analysis, — a circumstance which illustrates the importance of maintaining the temperature of bituminous gas during its passage from the producer to the furnace at a point which will prevent the condensation of the hydro- carl)on vapors, in order that the full value of these vapors may be secured.

In discussing the relative values of hard- and sofl-coal gases, we must therefore consider the actual composition of the latter rather than its apparent composition as shown by analysis. The nearest approach we can make to the true composition of gas from sofl coal is probably the analysis of the theoretical gasification of bituminous cord in Mr. W. J. Taylors paper on "The Energy and Utilization of Fuel " (Trans.y vol. xviii., p. 870). He assumes a soft coal of the following composition :

382 The Ope.V-Hearth Process.

Per cent

Hydrocarbons, 32

Fixed carbon, 65

Water and ash, 13

The theoretical gasification yields a gas containing :

Per cent, by vol

Co, 27.8

Co,, 2.7

CH,andC,H 13.2

H. 8.3

N 47.8

This gas has about 58 calories (230 B. T. U.) per cubic foot, while the same volume of the best quality of gas from anthracite (for ex- ample, one with the above analysis) has only about 34 calories, — difference of over 60 per cent, in favor of the former. But the coals from which these gases were made show a thermal differ- ence of only 7 per cent., containing respectively 7yOO and 7400 calories per kilo. These figures explain the essentially different actions of the two gases at different temperatures, and show why bi- tuminous-gas is so much superior for heating at high temperatures, while in low-temperature work its superiority does not appear ; in fact, at low temperatures the values of the two gases are nearly equal when the gas from a given amount of coal is taken as a basis of compari- son. To illustrate, a ton of iron can be heated in a furnace of good construction with the gas from, say, 300 pounds of bituminous coal, whereas, using anthracite, the same work can hardly be done with less than 600 pounds of coal. But in work like boiler-firing, where only a low temperature is reached, the amount of work done by a pound of bituminous coal is but slightly greater than that done by the same amount of anthracite. Reference to any lists of boiler-tests, where both coals were used, will confirm this statement. U{)on con- sidering each class of work, the reason for this difference becomes appa- rent. The shell of a boiler cannot be heated beyond a point which is far below the flame-temperature of the gas, and the work done is simply the transference of the heat-units from the gas to the water in the boiler. In boiler-firing it is, therefore, no advantage to bitumi- nous-gas that the heat-units are more concentrated than in anthracite- gas. In estimating the work of the fuel the only question is, how many heat-units does it contain, that are capable of absorption by

The Open-He A Bth Process. 383

the water, assiiining the boiler to be so set as to absorb all available heat from the gases.

In a high-temperatare furnace the material to be heated is usually charged cold, and the object is to raise it as rapidly &s possible to a high temperature. Under these circumstances the difference in in- tensity of 60 per cent, in favor of the bituminous gas is a great ad- vantage. A comparatively small portion of the generated heat is absorbed by the furnace-charge, and the gases leave the furnace with- out being materially cooled. It is apparent, therefore, that intensity is desirable rather than the mere delivery of a certain number of heat-units to the furnace per minute, and that the inferior intensity of anthracite-gas can be offset by its consumption in greater quantity only to a point where the necessarily increased size of the furnace causes the heat to be radiated as fast as it is supplied.

Theoretically, the maximum temperature of the furnace should be that of the flame-temperature, provided that all radiation could be prevented. Assuming the gas and air for combustion to enter at a temperature of 1000* C.,* the anthracite-gas flame-temperature would be approximately 2850 C, and that of the bituminous-gas, 3190 C. Although these temperatures are much higher than can be reached in practice, yet they may be assumed to show approxi- mately the difference of temperature obtainable with the two gases, since with properly-constructed furnaces, the loss by radiation would not differ materially with this difference in temperature.

The difference in flame-temperature is not sufficient to account for the great superiority of bituminous-gas, and we must, therefore, look to the fact that the gas itself, being more concentrated, burns more quickly, thus generating a higher heat before it can be dissipated by radiation, and also, to what is probably a more potent factor, the luminous quality of the gas, which causes the heat to be given out to the charge by radiation as well as by conduction. The precise amount by which the radiating power of the gas increases its heating power has not yet been determined, but in view of the above comparison, we must apparently hold radiation responsible for much of the su- periority of bituminous- over anthracite-gas. Until within a few years little attention was paid to the value of heating by radiation, but the importance of this factor is now acknowledged, and the ac- cepted type of high-temperature furnace is one with a high roof and

This assumption b based on the writer's observation of the color of the checker- work in the regenerators of steel-furnaces, and on u comparison of the respective temperatures of the furnace- hearth and stack.

384 The Open-Heabth Procb8S.

ample room for the development of the flame to the highest tempera- ture. This flame is relied upon to heat the charge by its radiant power, no attempt being made to force it down upon the charge, as with the old low-roofed furnace. Luminosity and increased heating- power may be given to anthracite-gas by the addition of petroleum- vapor. A small stream of crude oil may be dropped into the pro- ducer or, preferably, into the rear part of the combustion-chamber of the furnace. When petroleum is volatilized very rapidly, it is largely decomposed into solid carbon (soot) and hydrogen. It is therefore essential that the petroleum be introduced at a point where the temperature is low, or the desired enriching effect will not be obtained.

To summarize briefly, anthracite-gas differs from bituminous-gas in its lack of hydrocarbons, which makes it a third weaker and almost non-luminous. The duty obtainable from the two gases is practi- cally the same at low temperatures when equal amounts of coal serve as the basis of comparison, but at higher temperatures the bituminous gas immediately shows its superiority; at still higher temperatures, a point is reached where the furnace cannot be made any hotter by the combustion of any quantity of anthracite-gas, and the more intense fuel must be used for any further increase of tem- perature. Theoretically, the maximum temperature of the anthra- cite-gas furnace should be the flame-temperature of the gas,1)ut radiation from the furnace lowers it very much — probably to a point between the temperature for melting steel and that for heating iron. This maximum point may be increased by the addition of petroleum, but only to a limited extent, except by the extravagant use of oil. The bituminous-gas owes its superior value to its hydrocarbons, which give it intensity and also luminosity, qualities which impart to the gas largely-increased heating power, although the degree of increase has not been accurately measured.

Sec. 16.— Water- Oas* It has been shown that by the continuous use of steam in a pro- ducer, the extreme in one direction gives a hot fire with a great loss

Water-gas has not yet been generally nsed in open-hearth work, bnt a con- Hideration of gaseous fuels would be incomplete without mention of it. The de- velopment of the procera in this country is due to Mr. Bnrdett Loomia, of Hart- ford, Conn., and to his engineer, Mr. llawley Pettibone. The following discussion has been reviewed and revised by them, and therefore the statements may be ac- cepted as authoritative oonceming the present development of the Loomis water- gas process.

The Opek-Heabth Process. 385

in sensible heat; in the other direction it results in retarded com- bustion in the producer and an excess of steam in the gas.

To avoid this dilemma and furnish a fixed, clean gas .fit for all work where fuel is used under a strong and uniform pressure capa- ble of easy carriage for long distances, the intermittent system has been devised. The temperature of the fire is first raised by a dry air-blast, and then steam alone is passed through it. During the air-period, the gaseous products are like those of an ordinary pro- ducer and may be burned in a regenerative furnace. During the steam -period, the product consists of a mixture of nearly equal vol- umes of CO and H. This gas is usually passed into a separate holder. The cooling effect of the steam is such that it can be used only about one-half of the time, the other half being required for recovery of the fire under the air-blast.

It is found advantageous to exhaust the air downward through the fire, thus preventing caking or fusing of bituminous coal, and converting the tar and volatile matter in the fresh coal into a fixed gas. This dis()osition enables the operator, while the top door is open for the in-draught of air, to examine the fire and feed coal where it is needed, so that the fires do not require poking. The upper part of the fire is hottest, and since the steam is admitted in the reverse direction, the longer retention of an area of intense heat for decomposition of steam is assiired, and the brickwork in the bottom of the generator, heated by the sensible heat in the producer- gas, superheats the steam that passes into the fuel-bed.

The gases pass from the generators through boilers, where the sensible heat is utilized in generating suificient steam to run the exhauster and for decomposition. As the gases are thereby cooled, no water-cooled valves are required. The exhauster delivers the gases through holders under a strong and uniform pressure, making it practicable to use small pipes, and as the gases are fixed they can be carried any distance. Small holders are used, since two or more generators furnish a steady supply of both gases, and their propor- tion can be varied at the will of the operator.

The rich gas made by this processes a useful and valuable agent where intense temperatures are required and the work is intermit- tent. Under these conditions a regenerative furnace is unsuitable, for if it be kept at a maximum temperature, it requires nearly as much fuel when empty as when it is charged ; and if allowed to cool, it requires a long time to bring it up to the proper heat.

Water-gas furnishes means by which, an intense temperature may VOL. XXII.— 25 T

386 The Open-Hearth Process.

be quickly prodaced without any preliroinary heatiug of the fuel. This special advantage has no bearing upon open-hearth work, for there regeneration is indicated as a matter of economy whether the gas be rich or poor. Wherever the fuel is to be preheated there is no advantage in having a gas which does not need it.

For regenerative furnaces the question of cost is the most impor- tant consideration. Up to the present time there has been no pre- tence of manufacturing a given quantity of heat-units by the inter- mittent system as cheaply as by a continuous process. But it is now claimed that the improvements in apparatus, in the simplification of machinery and the reduced labor of attendance, will allow this sys- tem to compete with the old processes. Whether this claim will be borne out by trial upon a large scale remains to be seen.

Sec. 17. — Natural Gas.

In Western Pennsylvania the occurrence of natural gas solved for a time all questions of fuel for heating-purposes. This gas contains from 2 to 12 per cent, of nitrogen, with less than one-half of 1 per cent, of carbonic acid, the remainder being paraffines containing, by ultimate analysis, about 70 per cent, of carbon and 23 per cent, of hydrogen. By ordinary methods of analysis, natural gas gives from 67 to 93 per cent, of marsh-gas, the remainder, where the marsh-gas is low, being principally hydrogen.

Attempts to pass this gas through the ordinary regenerative chambers were unsuccessful, owing to the deposition of soot upon the brickwork from the dissociation of the CH. The purity and power of the gas, however, render preheating superfluous, and as it can be carried to the ports in a small iron pipe, the construction of the furnace is much simplified, and all the products of combustion pass through one opening into the air-chambers.

Sec. 18. — Petroleum.

The existence of a large deposit of petroleum of such an aromatic character that it is not applicable to domestic use, has been a con- stant incentive to employ it in the place of the solid fuels. For many years it has been burned in all kinds of heating-furnaces by atomizing a jet of the oil with steam and supplying air for combus- tion. To use it in a regenerative furnace, the additional problem of its conversion into gas has to be solved. So far as the writer knows, the most successful method for this purpose is the Archer process. A jet of oil is atomized by superheated steam, and the re-

The Open-Hearth Process. 387

suiting raixtare is carried through an arrangement of pipes and per- forated discs which are heated to a temperature just below redness, the whole being constructed so as to mix the vapors as much as possible. The product is not a gas, though often called so; it is a vapor, and must be delivered through covered pipes into a hot chamber. It can be burned in a muffle, but the flame labors under the work. Before combustion can be completed, the oil-vapor must be converted into gas, and the absorption of heat for this work, as well as for the heat- ing and decomposition of the steam, keeps down the temperature.

For open-hearth work, the mixture must be inje(*ted into a part of the chamber where the temperature is sufficiently high to prevent condensation, and where it will pass through enough checker-work to insure gasification. The waste-heat of the furnace is thus used to perform the preliminary expansion, and the furnace is supplied with a fixed and rich gas. If oil were injected alone, the spaces between the checker-bricks would soon become clogged with soot from the decomi)osition of rich hydrocarbons. The steam which is used in the atomizer serves the double purpose of an injector and a preventive of this deposit, the reaction being C + HjO CO 2 H. In the practical operation of an Archer producer, or in any ordinary atomizer, the amount of steam required for the purely physical work exceeds the amount required for this chemical action. It seems reasonable to suppose (and many inventors have acted on the sup- position) that under the combined action of steam and heat there should be an atomic interchange by which all the aqueous vapor would be decomposed by the rich hydrocarbons. Unfortunately such is not the case under ordinary conditions. The passage of the mixture through 20 feet of small flues with rough brick surfaces, at yellow heat, may leave in it as much as 20 per cent, of free steam. That this is not due to an abnormal amount at the injector, is proven by the presence at the ports of smoky hydrocarbon com- pounds. These can be destroyed by admitting air into the chambers and converting the carbon into carbonic oxide, but the consumption of fuel will be increased.

Chapter IV. — Regulation op the Temperature,

Sec. 19. — The Law of Thermal Increments.

In Ktarting a furnace it is not safe to attempt the immediate intro- duction of producer-gas, for when cold the gas can scarcely be burned

388 Th£ Open-Hearth Process.

with air at the ordinary temperature, and serious explosions are likely to result. It is necessary, therefore, to heat the furnace proper and the chambers by a wood fire, or otherwise, until they show signs of redness. When finally the gas and air are admitted, precautions being taken to avoid explosions, they cool the chambers through which they enter, and heat the furnaoe-roora and the checkers on the outgoing end. At the expiration of an hour, the cooling of. the incoming end will be indicated by the appearance of the flame while the outgoing end will be appreciably hotter. Upon the reversal of the currents, the opposite action will obtain, and each successive alternation of the currents will be attended by a rise in the tempera- ture of the entering gases, the combustion-chamber, and the outlet- chambers. The gas and air, as they come to the valves, are unaf- fected by this rise, and since they determine the temperature of the checker- bricks near the valves, it is obvious that whatever may be the temperature attained in the furnace, the products of combustion will pass on their way to the stack over brick surfaces which will be at a certain constant temperature, so that with ample regenerative capacity the loss of heat in the escaping gases will depend upon the temperature of the entering gas and air.

As the temperature of the furnace rises, the loss by radiation in- creases, and if the walls which surround the combustion-room and the chambers are thin, the loss at high temperatures may equal the gain from the reneration and thus reach a maximum. In a good furnace, however, the integration of the thermal increments extends between the limits of temperature at which the gas ignites and at which combustion ceases. This latter point, where gas and air refuse to unite, is often mentioned in theoretical treatises, but it never dis- turbs the practical metallurgist, for long before combustion ceases the most refractory walls and roofs will have melted. The diflSculty is not to obtain sufficient heat, but to avoid an excess, with the conse- quent destruction of the furnace. To guard against this evil the oonsumptioa of gas is so regulated, after the desired temperature has been attained, that the radiation will equal the thermal incre- ment.

Sec. 2Di — Estimation of the Relative Temperature.

It is difficult if not impossible, with any existing appliance, to determine the dree of heat attained in melting steel ; but for all practical purposes it suffices to know the relative temperature, and this can be determined by the eye with wonderful accuracy. In

The Opek-Heabth Process. 389

furnace-work, blue glass is used to cut off the fierce rays of light and heat, but even so, from a scientific point of view, it is the eye alone that gauges the temperature. There is a feature in Bessemer practice which enables us to estimate the usual limit of error in the ocular test; it is customary to put a weighed quantity of cold pig- iron or scrap into the converter with the charge iu order to reduce the temperature of the blown metal, and experience with many thou- sands and tens of thousands of blows shows that the difference caused by a variation of 100 pounds either way in such an addition can be seen by the naked eye. A rough estimate of the alteration in tem- perature can be made by the following system of average :

Conyeiter-charge. Galories.

6820 kilos of pig-iron at 1600° C. conUio . . . 10,912,000 45 kiloB of scrap at 0° C. contain 0

6865 kiloe at 1590° C. contain 10,912,000

Refrigerating effect 10° C.

When the scrap is added after the completion of the blow, this method is not entirely worthless. It neglects the variations in specific heat for different temperatures as well as the latent heat of fusion ; but since the heat-absorbing power is much less at low tem- peratures, the error caused by ignoring this variation will tend to counterbalance the omission of the heat of liquefaction. An error of 100 per cent., therefore, is almost impossible and for present pur- poses this is sufficiently accurate since the result would be just as surprising if it were 20° instead of 10°.

A more careful determination may be made by taking into ac- count the specific heats and thermal values of the factors involved. In the following calculation the constants, with the exception of the final temperature, are taken from a paper by Prof. Akermnu in the Journal of the Iron and Steel Indiiutey 1 872, ii., p. 1 10. As the author has elsewhere pointed out {Trans., xix., p. 180) the values for the specific heat given by Prof Akerman may be accepted, for though other writers differ from him in the values which they give for dif- ferent temperatures, the totals between extreme limits agree suffi- ciently well. The methods used by Akerman are correct and the results accord with the data ; but in the condensation of the calcu- lations as given in the Journal the formulae are confused and mis- leading. The proper method is indicated below: the initial tem-

390 The Open-Heabth Process.

perature is retained at 1400 C, but that of the products is assumed at 1600'' C, as the bath is undoubtedly heated by 200"" C.

The heat available for heating the bath is that produced by the combustion of silicon, manganese and carbon mintu the heat ab- sorbed in raising the products of that oombusticHi to the tempera- ture of the bath. The amount thus absorbed will be the total sen- sible heat in the products minus the heat in the initial metal due to the heated state of the above elements. Thus the heat lost in the escaping carbonic oxide is not the product of the total specific heat of the gas multiplied by the temperature at which it passes away, but it is this total minus the specific beat of the carbon multiplied by the initial temperature of the metal.

The initial temperature of the bath is taken as the starting-point, notwithstanding that at the time when the carbon burns the bath is hotter than at the beginning of the operation; for such an increment must be produced by the combustion of some of the metalloids, and all the heat of that combustion is assumed to be imparted to the metal.

Combustion of SUicon. — The specific heat of silica is unknown, but the net absorption of heat is assumed to be the energy required to heat the oxygen and nitrogen of the air consumed in the com- bustion.

ReACtloiM. 10 kilos Si + 11.429 kilos O develop 78,300 calories. CMoriea.

Absorbed by oxygen, 11.429 X .218 X 1600 3,986

" nitrogen, 11.429 X 38.262 X .244 X 1600 . . . 14,937

Total heat-absorption 18,923

Net heat-evolution (78,300 — 18,923) 69,377

per kilo of Si 6,938

Combustion of Carbon. —

Specific heat of C, 0.241.

Specific heat of CO. 0.2479.

10 kilos C -h 13.33 kilos O 23.33 kilos 00, which develop 24,730 cal's.

Calories. Absorbed by oxygen, (23.33 X .2479) 1 600 — ( 10 X .241) 1400 . . 6,880

" nitrogen, 13.33 X 44.627 X .244 X 1600 . . . 17,422

Total heat-absorption, 23,302

Net heat-evolution (24,730 — 23,302) 1,428

perkiloofC= 143

The Open-Hearth Process. 391

Combudion of Iron. —

Specific heat of pit-iron (melted) 0.11.

FeO 0.17. 10 kilos Fe + 2.857 kilos O 12.857 kilos FeO, which develop 12,013 cal's.

CalorieH. Ab)rbedbyoxygen, (12.857 X. 17) 1600— (10 X. 11) 1400= . . 1967

" nitrogen, 2.857 X 9.569 X .244 X 1600 . . . 3736

Total heat-absorption, 5693

Net heat-evolution (12,013 — 5693)= 6320

" per kilo of Fe= 632

CombiLstion of Manganese. — The thermal value of manganese is as- sumed to be the same as that of iron. The constants employed are :

Specific heat of pig-iron between C. and 1200® C. . . 0.16

Latent heat of fusion of pig-iron 46.

Temperature of melted steel. 1600** C.

Specific heat of melted steel (same as elementary Fe) . .0.16

Weight of bath. 15,000 lbs. 6820 kilos.

Weight of cold metal added, 100 lbs. . . . . 45 " Percentage composition of the addition. Si 1.50, Mn 0.30, G 3.75.

The calories developed by the combustion of the oxidizable com- ponents of the metal added are:

Calories.

8i-0.675 kilos develop 4001

Mn— 0.135 " 88

C— 1.688 " " 822

Total heat developed, 4411

The calories absorbed by the metal added are :

Calories. In heating from O*' to 1200** C, 45 X (1200 X 0.16) . . 8640 Latent heat of fusion, 45 X 46 2070

Total heat absorbed, 14,490

Thus the addition of cold pig-iron exerts the following effects:

Calories.

Absorption of 14,490

Development of 4,41 1

Net absorption of 10,079

392 The Opek-Hearth Process.

This araoiiDt taken from 6820 kilos of melted steel reduces its tem- perature 9"" C.

If steel scrap be used for the addition, containing Si 0.10, Mn 1.00 and C 0.40 per cent., only 595 calories will be produced, and the net absorption (assuming the thermal values for pig-iron and steel to be the same) will be 13,895 calories, causing a reduction in temperature of J 3® C.

A comparison of these two calculations shows that they corres- pond perfectly with the results of practice, since 1000 pounds of steel scrap are equivalent in refrigerating effect to 1300 pounds of pig-iron.

There are many assumptions in this calculation, yet practice proves the general accuracy of the results. One assumption which may seem to be unsubstantiated is the final temperature of 1600 C. for the steel. The theoretical increment may be provisionally calcu- lated as follows : 100 kilos of metal ; initial temperature, 1400 C. :

Calories. Combustion of 1.50 per cent Si 1.5 kilos 5938 cals. . 8907

Total heat-production, 9408

Specific heat of steel 0.16

Heat absorbed by 100 kilos for each C. . . . . 16

Theoretical increment -— - 688® C, giving a final tempera- ture of 1988C.

This does not allow for the heat carried away by an excess of air, by decomposition of steam, and by fusion of lining. Moreover, the calculation is provisional, for with a final temperature difiering from 1600° C, the thermal values of the metalloids change very appreci- ably. If 2000° C. be taken as the final temperature, the loss in the products of combustion is greater, and the results are changed as follows: Silicon develops in burning 5465 calories per kilo; iron, 451 ; while carbon becomes a negative factor, absorbing 524 calories per kilo. These values make it necessary to recalculate the theoreti- cal final temperature by some system of successive approximation, a refinement however, which seems superfluous, since the factors are not known with sufficient exactness to warrant the labor. It will he safe to assume that a change of 400° from the basis of the first calculation embraces all possible errors in temperature determi- nation. The following calculation shows theefiectof the metal-addi- tion on this new basiS; namely, that the final temperature is 2000° C.

The Opex-Hearth Process. 393

(1) Absorption of heat by 45 kilos of pig-iron :

Calories. Calories. In heating from 1200° to 2000° C, 7560 Latent heat of fusion 2070

Total absorption, 18,270

Production of heat in burning 0.675 kilos of Si, . 3689 Production of heat in burning 0.135 kiloe of Mn, . 61

Less amount absorbed in burning 1.688 kilos of C, . 884

Total production, 2.866

Net absorption, 15,404

Cooling effect on 6820 kilos of steel 14" C.

(2) Absorption of beat by 45 kilos of steel scrap :

Calories. Calories. Total absorption as above, 18,270

Production of heat in burning 0.045 kilos of Si, . 246

roduction of heat in burning 0.450 kilos of Mn, . 203

Less amount absorbed in burning 0.180 kilos of C, . 94

Total production 355

Net absorption, 17,915

Cooling effect on 6820 kilos of steel 17° C.

These figures show that the probable error in the work is, pro- portionally, very large. Similar radical alterations in results might be made by assuming different values for the specific heats. But, although the ratio of error to determination is very great, yet the ab- solute nature of the result is not affected, since the refrigerating effect is shown to be very small under the most unfavorable hypotheses. By the rough system of average first employed, and on any basis of theoretical calculation, the addition of one hundred pounds of scrap to a bath of seven tons will cause the temperature to fall only ten or twenty degrees Centigrade.

As previously observed, this variation is perceptible to the naked eye. An excess or deficit of five hundred pounds in the addition of scrap makes the charge either so hot that it causes trouble, or, so cold that it skulls badly in the ladle. A variation of two hundred pounds from the true amount, which would show a variation of about thirty degrees, is the greatest error allowable in good practice.

394 The Open-Hearth Procebs.

Incredible as it may seem, within these narrow limits lies the sac- cess of the blower and the melter.

The estimation of temperature by the eye may be unscientific, but it is eminently practical. Probably, no pyrometer applicable to rough work can ever be constructed which will determine these high temperatures within the narrow limits required. Moreover, no in- strument or method can be satisfactory which gauges only the tem- perature of the furnace-room. It must reach the metal itself. It often happens, in fact it occurs in every melt, that the furnace- gases, and the roof and walls, are at the maximum allowable tem- perature, while the bath itself is far below the required degree of heat. It also happens, occasionally, through some slight interrup- tion in the fuel-supply or through inattention, that the combustion- chamber is temporarily cooler than the metal.

The condition of the metal itself must be known, and in the very nature of the case, this is a difficult subject for scientific experimenta- tion. When the slag-covering is thin, and the boil lively, the tem- perature of the metal may be accurately judged by the appearance of the surface of the slag. When the bath is quiet, other means must be resorted to. Some operators stir the metal with an iron rod and note the rate at which the rod melts and the appearance of the end when it is withdrawn. The objection to this test is, that the slag which adheres to the rod as it is thrust into the bath sometimes protects it from melting, even in the hottest steel. The character of the cinder, and the precise manner in which the rod is put in, will affect the result, as will also the duration and rapidity of the stirring. The vigor of the stirring will be regulated by the liveliness of the boiling which is caused by the ebullition of gas. The variability of these factors interferes seriously with the usefulness of this mode of testing the bath. The best way is to take out a sample of the metal in a small test-ladle. Such a test is required for the determination of carbon by fracture or otherwise, and its character and temperature can be noted while pouring it into the mould.

As for the temperature of the furnace-room itself, the continually changing conditions will always render any form of pyrometer im- practicable and difficult of application.

In the Journal of the Iron and Steel Institute, vol. ii., 1886, p. 953, a method of temperature-determination is described, consisting of noting the rate at which a bar of iron is heated when exposed to the flame. If the melter has experience and good judgment, he can tell by the eye alone, better than in any other way, the temperature

The Open-Hearth Process. 395

of the bath, side-walls, roof, and flame, but if he lacks the proper qualifications no test-bar, certainly none used as the writer in the Journal describes, can compensate for his deficiencies.

Chapter V. — The Thermal Equation of the Furnace Sec. 21. — Summary of the Thermal Factora.f

Calories.

Heat generated from coal during one charge (Sees. 46 and 84, iAid.), 2,377,420

Heat generated from the combustion of the oxidizable ele- ments in the metal (Sec 85, ibid.), 143,000

Total heat generated, 2,520,420

Per cent, of total calorific power Calories. generated.

Loss due to carbon in producer-ashes (Sees. 12

to 25 inch, ibicL), . . . . . . 140,650 6.6

Losses in producer other than in ash, includ- ing distillation, dissociation, radiation, and

sensible heat of gases (Sees. 12 to 25 incl.,

ibid.), 694,840 27.6

Loss dae to the sensible heat of the products

of combustion escaping into the stack (Sees.

29 and 30, ibid,), 246.710 9.8

Losses due to combustible gases escaping into

the stack, and from an excess of air in the

products of combustion (Sees. 32 to 42 incl.,

ibid.), 69,390 2.8

Absorption of heat by the metal for heating,

melting, and superheating (Sees. 82 and 83,

ibid,), 290.000 11.5

Radiation bj difference (Sec 88, ibid,), . . 1,078,830 42.7

2,520,420 100.0

If we neglect the loss of heat in the producer, and take as a basis the potential heat coming to the furnace in the gas, we shall have :

In a paper on the " Physical and Chemical Equations of the Open-Hearth Pro- cen," IVont., xix., p. 128, the writer has discussed at length the factors of the thermal equation. Only a summary is therefore presented here, accompanied by numbered references to the sections of the earlier paper for fuller information. A 25-ton acid furnace with vertical chambers, working with gas from bituminous coal, formed the sabject of investigation.

t See Sec 87, ibid.

396 The Open-Heabth Process.

Calories. Per cent

Lobs from Bensible heat in waste-gases, . . 246,710 14.6

Loss from imperfect combustion, . . . 69,390 4.1

Absorption in melting and heating, . . . 290,000 17.2

Badiation by difference, 1,078,830 64.1

1,684,930 100.0

Practical men (among whom all metallurgists wish to be num- bered) will not accept a conclusion which is founded solely on theo- retical calculations, and which apparently conflicts with general, though perhaps unformulated, tradition or experience. That only one-ninth of the total energy of the coal is used in the actual melt- ing of the stock appears at first sight contrary to reason, nor does it seem much more acceptable to say that one-sixth of the potential power of the gas is thus utilized. But corroborative evidence is easily obtained. In Sec. 88 {he. cit.) the fuel-consumption for main- taining an empty furnace at a temperature considerably below work- ing-heat is given, the figures showing that nearly as much fuel was used as if the furnace had been in active operation. This does not prove that each item which contributes to the 88 per cent, of loss is correct, nor does it validate the determination of the loss due to radiation, for only the absence of all other methods can justify the resort to estimation by difference; yet it furnishes convincing evidence that the figures for the absorption of heat by the metal are substantially correct.

Since every furnace, whether fast or slow, must work at about the same intensity of heat, and since for any given temperature the radi- ation and other losses will be in exact proportion to the time in- volved, it is evident that the fuel-consumption per ton of steel will be inversely proportional to the product; in other words, all other things being equal, a furnace with large ports and sufficient gas, making four heats a day, will have only one-half the fuel-ratio of a furnace of the same general type making two heats. Strict accuracy aside, experience confirms this general statement.

Chapter VI. — The Acid Process.

The acid process, as here discussed, consists in melting a charge of pig-iron or a mixture of pig-iron and low-carbon metal upon a sand- hearth, and converting it into steel by the action of the flame, with or without the aid of ore, and by the addition of proper recarburizers.

The Open-Hearth Pboce98. 397

Sec. 22.— The Hearth.

In open-hearth work, as usually conducted in this country, the hearth is made of sand, so set or sintered that it will resist the me- chanical and chemical actions of the melted metal and slag. The " setting is performed by raising the temperature until the material hardens into a solid mass. Sometimes a natural sand is used, and sometimes a mixture of two or more sands. When two kinds of sand are employed, the relative proportions can be altered at will and, in shop-parlance, the mixture may be made " hard " or "soft." One of the sands used in a certain mixture, with which the writer is familiar, is a pure quartz, running 99.5 per cent, silica; it is entirely unaflTected at working temperatures ; the other sand contains 96.9 per cent, silica and 2.2 per cent, alumina, and it softens at a point somewhat below the full working-heat. A higher temperature only softens it still more, and at no point is there any hardening. A mixture of these two sands, containing 97.5 per cent, silica, becomes a solid mass when it is exposed to the proper heat. It is easy to understand how the more fusible sand becomes plastic under the heat, and, acting as a slag upon the pure quartz, absorbs the grains of silica, incorporating them into itself and becoming more and more viscous; but it is difficult to understand how this can con- tinue until the bottom is so hard that it will resound when struck with an iron bar even after heating for only fifteen minutes, and in the face of a continually-rising temperature, which, we should expect, would keep up the viscosity. Upon this esoteric action rests the practical success of the melting- operation, for the ca])acity to resist erosion depends quite as much upon the physical condition of the hearth as upon its chemical composition.

Up to within a few years no generic name was required to define the operation conducted upon the sand-hearth, but with the advent of the basic lining the contradistinctive title of acid process has become necessary. It is somewhat of a misnomer, since sand is not an acid, but an anhydride ; moreover, the slag in the operation is not acid, but basic enough to attack the hearth. Nevertheless, the name is appro- priate and worthy of retention. The bottom fulfils no necessary chemical function. It is true that its wear supplies silica to the slag ; a slag rich in bases will cut the hearth badly, while one that is highly siliceous will have little effect upon it. The hearth thus serves as a balance-wheel to keep the percentage of silica in the slag within cer- tain limits, but no fixed formula for this regulating action can be drawn; the most siliceous slag that can be carried in practice b basic

398 The Open-Hearth Process.

enough to appreciably attack the bottom if sufficient time and beat are allowed.

This relation between the slag and the hearth has a certain influ- ence on the chemical history of tlie process, but since, in good work, the charge and the operation are regulated to obtain the least scori- fication, the changes that occur in the composition of the bath may be ascribed solely to the interactions of the stocky the flame, and the ore.*

Sec. 2S.—The Charge.

For economic reasons, the charge varies at different furnaces, sometimes being all pig-iron and sometimes only one-fifth pig and the rest scrap. Since there is no elimination of phosphorus or sulphur, the amount of these elements in the stock must always be limited to the amount allowable in the finished product. What this amount will be depends upon the use for which the steel is intended. The content of silicon, manganese, and carbon in the charge is not limited by such narrow bounds. These elements are oxidized during the process, and their presence in greater or lesser amount alters the working of the charge rather than the composition of the product.

In the manufacture of soft steel it is the aim of most melters, where the cost of scrap is not prohibitory, so to regulate the proportion of pig-iron that the melted bath shall be free from silicon and manga- nese and shall contain from three-quarters to one per cent, of carbon. During the elimination of carbon the metal is in a state of continual ebullition, and its temperature and condition, as well as the charac- ter of the slag, may be completely controlled in preparation for recarburisation and casting.

If too small an amount of pig-iron is used in making up the charge, the molten bath will contain neither silicon, manganese, nor carbon, and will I)e viscous and pasty. Such a mass will be oxi- dized rapidly by the flame, and the oxide of iron will scorify the bottom. The only remedy lies in the addition of pig-iron ; as a rule the extra amount required will be more than would have sufficed at first to make a proper charge. This addition in itself may be of small importance, but the loss of time, the oxidation of the metal, and the cutting of the hearth will be of more serious consequence. If, on the other hand, too much pig is added, it is easy to lower the carbon-content by the action of the flame or, even more rapidly, by

For further discussion of the iDfluence of scorification on the metallargicil history of the bath, see Sees. 25 d, 27, and 29 iira, pp. 400, 404, and 409.

The Open-Hearth Process. 390

the addition of ore; hence it is safer to err bj supplying an excess rather than a deficiency of carbon.

Sec. 24. — The Method of Charging.

In* some places it is customary to charge the stock in installments. A party usually including all the pig-iron, is melted and superheated, and then the scrap, which has been preheated in another furnace, is dissolved in the bath. This practice arises from several consi<lera- tions :

(1) When preheating of the stock is practiced it becomes almost a necessity to charge in separate lots, as otherwise the capacity of the preheating-furnaces would have to equal that of the melting- hearth, and several preheaters would be required for the proper disposition of the stock.

(2) In hot summer weather it would be a destructive tax on human endurance to fill a 25-ton furnace at one time with preheated stock.

(3) The successive additions of scrap constitute a process of dilu- tion, which may be arrested at any time when the desired composi- tion is attained.

(4) It is sometimes supposed that the oxidation is less when pre- heated scrap is dropped into the melted bath beneath a blanket of slag than when all the metal is piled in immediate exposure to the flame. If, however, the oxidation in the preheater be included, as it should be, this supposition is doubtless an error.

(5) Preheating is advisable whenever, for any reason, it becomes necessary to make several separate and partial charges. The addi- tion of cold stock to a melted bath chills the liquid and checks all action, thus retarding the melt and producing superficial oxidation attended with cutting of the bottom.

(6) Intermittent charging may be advisable when the regenerative chambers of the furnace are so small that the addition of an entire charge of cold stock robs them of their heat and necessitates a long period of recuperation before melting can begin. This condition can obtain only under a radically bad construction, and entire re- building would be the proper remedy.

Thus the system of intermittent charging results directly or indi- rectly from inferior design, and preheating is its necessary concomi- tant. In a good furnace the whole charge can be put in cold at one time, without any injury from sudden cooling of the roof or trouble from chilling of the checkers, or oxidation of the metal, and with great economy in fuel and labor. ,

400 The Opex-Hearth Process.

Sec. 25. — The Conditions of Oxidation during the Period of Fusion.

The quantity of pig-iron to be used for a given result will depend upon many conditions, among which are the following: (a) The kind of gas. (6) The nature of the flame.

(c) The construction of the ports.

(d) The method of charging and arranging the stock, (c) The time of exposure.

(/) The kind of scrap. (g) The kind of pig-iron.*

(a) The Kind of Gas. — In Sec. et seq. the influence of certain components of the burning gas in producing oxidation of exposed metal was discussed. During the period of fusion these influences attain their greatest potency, for the direct contact of bare metal and hot gases, without the intervention of slag, favors chemical activity.

(6) The Nature of the Flame, — It is the custom in many places to carry a very smoky flame, especially during melting. Under such conditions the tendency to oxidation will obviously be less than with a short, sharp flame. The practice can easily be carried to an extreme, for a large excess of gas will not only involve a loss of potential heat by the escape of unburned gases to the stack, but will also delay the melting, and thereby tend to increase the oxidation.

(c) The Construction of the Ports. — That the gas should enter the furnace below the air has been mentioned in Sec. 5, one object being to avoid direct contact between the air and the stock. In the same section the advantages of heating by radiation were discussed, the flame not touching the metal. Just as these important features can vitally affect the oxidation, so, in proportion, smaller variations in design will also exert their influence. With a wrong construction of ports, or a neglect to keep a proper system in good repair, the flame will not be uniform throughout its breadth ; in some places vertical strata of air will impinge upon the metal, while parallel to them clouds of gas will pass through the furnace unconsumed.

(d) The Method of Charging and Arranging the Stock, — In de- scribing the system of intermittent charging (Sec. 24) it was said that while the loss by oxidation in the furnace possibly was reduced, it was probable that by including the waste in the preheater the total loss would be greater than where all the stock was charged at one time. In this latter method of charging, the arrangement of the

The Open-Hearth Process. 401

Stock has an important bearing upon the amount of oxide formed. In ordinary practice the scrap is first charged and the pig-iron is spread upon it. Under the action of the flame the pig will melt first, being the more fusible and the more exposed. It will trickle over the hot viscous scrap and a process of carburization will be started which will produce a compound intermediate between steel and pig-iron. If the iron could be spread evenly over the bath, it might be said that at no time would a piece of low steel be exposed to the flame, but always a surface of carburized metal, and thus each atom of iron would be protected by the contiguity of an atom of silicon or carbon for which the oxygen has a greater affinity.

Practically it is impossible to attain perfect protection. The flama wanders through the loosely laid stock and oxidizes the steel before any melted pig-iron can touch it ; when only a small propor- tion of pig is used there will be places where the scrap is entirely un- covered, and there large amounts of iron-oxide will be produced. If this cinder forms a pool on the viscous surface of the charge, it will meet, sooner or later, with high-carbon metal, and an interchange will occur with reduction of iron, the result being the same as if mixture had taken place at an earlier stage. But if the fused oxide comes in contact with the hearth, scorification will ensue with formation of sili- cate of iron, and though at a later period this scoria may be mixed with high-carbon metal, the harm cannot be completely remedied. A portion of the iron may, perhaps, be reduced, and a higher silicate be formed, but silica once having entered the equation as a factor, is there to stay, and will permanently hold a greater or less amount of iron-oxide. Hence to avoid a waste of iron, the pig-iron should be so distributed as to receive the direct action of the flame and particularly to guard the line of contact between the surface of the metal and the hearth, which is known to furnace-men as the "slag-line."

(e) The Time of Exposure. — The maximum oxidation of iron in the case of unprotected low-carbon scrap, will occur at a tem|3era- ture just below fusion. In a slow-working furnace, wh*cre, owing to insufficient gas or excessive radiation, the thermal increments at high temperatures are small, the longer exposure will occasion increased Iu88, while with rapid melting, other factors being equal, the waste will be reduced to a minimum. With the use of a large proportion of pig-iron, and the attendant diminution in scrap, the chances of oxidizing iron are lessened and the difference between slow and fast working may not show itself in the waste.

VOL. XXII.- 26 C"r\rn]o

? Ic

402 The Opex-Ueabth Pbocbss.

(/) The Kind of Scrap. — Inasmuch as oxidation can take place only on the surface of the metal, the use of finely divided scrap will augment the loss. Such material also is harder to melt, owing haps to its non-continuity and poor conductive power, and its longer ez|)osure occasions increased waste. These objections do not furnish reasons for going to the other extreme, and using scrap in large masses which also are slow to heat and slow in melting. The nature as well as the size of the pieces will affect the result, for if the scrap be high in carbon, the iron will be shielded in great measure, while if the foreign elements be exceptionally low, the iron will be not only entirely unprotected, but so infusible that the time of exposure must be considerably extendeil,

(g) The Kind of Fig-iron, — Since we rely upon the metalloids in (he pig to preserve the iron from oxidation, the value of any indi- vidual lot of pig will depend upon the amount of such foreign ele- ments present. Sulphur and phosphorus, as already observed, are not elimmated in the acid process, and therefore play no part in this action. The protecting agents are silicon, manganese, carbon, and rarely titanium. The relative influence of each will be discussed in the next section.

Sec. 26. — Value of the Elements for the Absorption of Oxygen.

The protective capacity of the elements depends in great measure upon their power to absorb oxygen, and may therefore easily be ascertained from their chemical quantivalent relations:

Si +20=SiO, Mn + O =MnO

28 + 32 60 55 + 16 71

C +0=C0 Ti +20 TiO,

12 + 16 28 182 + 32 214

Fe + O FeO

56 +16 =72

From which we find :

Units of oxygen.

1 unit of silicon combines with 1.143

1 unit of manganese combines with 0.291

1 unit of carbon combines with 1.333

1 unit of titanium combines with 0.176

1 unit of iron combines with 0.286

The Open-Heabth Pbocesb. 403

Or stated id another way :

1 nnit of ozjgen combines with 1 unit of oxygen combines with 1 unit of oxygen combines with 1 nnit of oxygen combines with 1 unit of oxygen combines with

0.875 unit of silicon. 3.44 units of manganese. 0.750 unit of carbon. 5.69 units of titanium. 8.50 units of iron.

From these figures the protective capacity of any ordinary iron compound may be calculated. Taking for example two pig-irons which are average samples of large piles, and common rail steel scrap, we have :

Weigbt, pounds Oxygen absorbed Percent, in 1000 pounds in 1000 pounds

of metaL A— Pig-iron: Silicon, . . , 0.84 8.4 Carbon,. . . 3.65 36.5

of metal

Total oxygen absorbed by 1000 pounds of metal,

. 58.3 p<]

B-Pig-iron: Silicon, . . . 4.24 42.4 Carbon,. . . 2.86 28.6

Totol oxygen absorbed by 1000 pounds of metal,

. 86.6 pd

C—BaU-steel: Silicon,. . . 0.07 0.7 Manganese, . . 0.85 8.5 Carbon, . . 0.40 4.0

Total oxygen absorbed by 1000 pounds of metal, . 8.6 pds.

In regular work on a charge of 60,000 pounds, it is found that the amount of pig-iron of class A required to produce a bath containing 1 per cent of carbon when melted, is about 20,000 pounds. The mixture, exclusive of iron, will absorb oxygen as follows :

Pounds of oxygen absorbed. 20,000 ponndsof pigiron A at 58.30 per thousand . . 1166 40,000 pounds of rail-steel Cat 8.6 O per thousand . . 344

Total oxygen absorbed, 1510

From this we may calculate the following comparison for pig- iron B:

Equation B + C 60,

i.€., the total weight of the pig-iron and scrap is to be 60,000 pounds. Equation 2. 86.6 B + 8.6 C 1510,

404 The Open-Hearth Process.

assuming that the total absorbing capacity of the charge is to be the same as that of the first mixture, or 1510 pounds.

Equation 3. 8.6 B + 8.6 C 516,

obtained from equation 1 by multiplying by 8.6.

Equation 4. 78 B 994,

by subtracting equation (3) from equation (2), hence,

1000 B 12,750 pounds, 1000 C 47,250 pounds,

or, 12,750 pounds of the high-silicon, low-carbon pig-iron will replace the 20,000 pounds of the low-silicon pig. Practice bears out this calculation. The difference is still more marked when the high-sili- con pig is richer, and the low-silicon pig poorer, in carbon.

During the melting operation a certain amount of iron is always oxidized, and this quantity, varying with the conditions of furnace- work, must be considered in any accurate calculation.* In the fore- going figures this action upon the iron has been neglected, as the equations would have been more complicated, and the only object was to show the general law. In practice every competent metallurgist can estimate closely the required amount of any given pig-iron, without resorting to algebraic computations of protective quanti- valences.

Sec. 27. — The General Chemical Law.

The record of an open-hearth charge is essentially one of oxida- tion. It is true that certain oxides and other compounds are con- tinually undergoing reduction, but the sum of all the reactions in any period will be on the side of oxidation. The mutual reactions of the elements and compounds of the charge upon one another must balance, while the effect of any flame that has active calorific power, is always oxidizing.

As a general rule it may be said, that the most easily oxidizable elements are burned first. This seems so self-evident that the guarded form of the statement may seem to involve the negation of an axiom. The limitation applies most forcibly to the period of fusion, for, as

For a further discussion of the oxidation of iron, see Sees. 27 and 31 e.

THE OPEN-HEABTH PBOCfiBS. 405

previously observed (Sec. 25 d), the reactions during the melting of the stock are largely determined by physical conditions. The pres- ence of considerable amounts of silicon in a rapidly oxidizing bath, when the metal is extremely hot, presents an apparent exception, which is due to a change, at different temperatures, in the relative aflSnity of oxygen for carbon and silicon. There is also a real excep- tion, which is based upon practical conditions. At first sight it would appear that all the operations of heating the bath and burning the silicon, manganese and carbon, could proceed without the interven- tion of any medium between the metal and the flame. It is true that the oxidation of the silicon and manganese will produce a slag, but when these elements are burned, there seems to be no reason why their products may not be removed, the carbon burned, and the metal superheated to any desired point. This, however, is not possible in practice. It is found necessary to have a sufficient covering of slag on the metal, otherwise the iron will be oxidized by the flame, causing serious cutting of the hearth, and the metal will cool even under the best of flames.

What will constitute a sufficient quantity of slag depends on the surface to be covered and on other conditions of practi(je, but that such a determining factor exists should always be remembered with a view to furnishing in the components of the charge ample material for a proper blanket of slag.

In ordinary practice a certain aVnount of oxide of iron will be formed, and will be present at all stages of the operation, and the metallic iron involved, whatever may be its exact amount, must be regarded as a separate agent, acting with the metalloids in the work of protecting the great mass of metal which will not be attacked until the silicon, manganese, and carbon have been burned. After the complete fusion of the charge, and the formation of au adequate quantity of .slag of proper composition, there is no further strong tendency toward an increase in the amount of slag. There will, however, always be some slight cutting of the hearth and a conse- quent addition of silica; the ore will also contribute a small amount of the same material, and since there is an inherent tendency in the slag to attain a given composition, there will be a slight increase in its volume so long as the heat is held in the furnace.

During the period of melting the charge, the flame burns the oxi- dizable elements of the bath by transmitting the oxygen through the slag. This action may occur in some measure by the oxidation of fer- rous to ferric silicate in the slag, followed immediately by its reduc-

406 The Open-Hearth Pbocbbb.

tion by the bath to the ferrous condition; or the oxidation may be produced by the burning of the fine globules of iron which are con- tinually being projected through the slag by the ebullition of carbonic oxide, and to the transference of their oxygen to the underlying metal. The combustion of silicon and manganese by this process is very slow, as their oxidation is not attended with any commotion of the metal, whereas the burning of carbon keeps the bath in constant action. The work may be hastened by the addition of ore in suc- cessive doses, or by blowing air into the bath through a tuyere. This latter practice has been tried with some degree of success, but the scorification of the hearth, the cutting of the roof and walls, and the loss of iron through the ports have prevented its extended adoption. A large quantity of air is required to hasten the operation appre- ciably. Two thousand pounds of pure ore (FcjOj) contain six hun- dred pounds of oxygen. To furnish the same quantity of atmos- pheric oxygen, 2586 pounds or 35,920 cubic feet of air would be required. After the silica in the slag has once been satisfied, there is no further urgent demand for bases, and any ore added thereafter may be regarded as free iron oxide, which will react upon the car- bon or other reducing elements of the bath with the formation of metallic iron.

Sec. 28. — QuarUUative InveatigatioTis into the Chemical History*

Exact calculations upon the open-hearth bath are almost impos- sible. A certain amount of slag (much more than many furnace- men suppose) will be left in the furnace after the tapping of every heat. This residue will be picked up by the next charge, which in turn leaves one to its successor. Therefore it is impossible to ascer- tain the exact weight of slag for any one charge, and incorrect to assume that it is formed solely from the regular and known compo- nents of that charge. It is practicable, however, to approximate very closely to accuracy by taking a series of consecutive charges made under the same general conditions, melted with the same kind of gas by the same men, and showing none of those accidental variations which occasionally arise, and by weighing and analyzing all mate- rials and determining the scorification of the bottom by the amount

Id this and the following sections the results of certain quantitative investiga- tions of practical working are given. These results* accompanied with explana- tions of the modes of calculation, were first published by the writer in Dransactiont, vol. xix., p. 160 et seq., in the paper already referred to.

The Open-Hearth Process.

of sand required to repair it. With sufficient data obtained in this way, the weight of the slag may be calculated as follows for any period of the melt:

The final slag, after tapping, is weighed cold. By subtracting from this weight the MnO produced by the addition of the recar- burizer, and the sand derived from the tap-hole and ladle-linings, the amount of slag which was in the furnace before tapping may be computed. Given the analysis of the slag at that time, it is easy to calculate the weight of its various constituents, among which will be the manganese ; if the ore contained no appreciable quan- tity of this element, the amount which, in one form or another, was present throughout the operation will be known ; and since the percentage of manganese in the slag and in the metal can be deter- mined by analysis, and the weight of the metal can be calculated for any stage of the work, all the requisite data are at hand for a deter- mination of the weight of the slag at any given time. With this determination as a basis, the quantitative estimation of the factors is a matter of simple arithmetic. In the following results of investi- gations of two kinds of practice. Group I. represents nineteen con- secutive heats melted with soft-coal producer-gas, and Group II. six consecutive heats made with Archer oil-vapor.. The figures are averages of each series :

Group!. Group II. Pounds. Pounds.

J 1 ) Pig-iron nwd : Si 1 .72, C 3.50 per cent 11 ,700 20.700

(2) Steel scrap used : Rail, Si 0.07, C 0.40 per cent., . . 39,680 29,600

Boiler, Si 0.02, C 0.13 per cent., . . . 6,970 7,200

(3) Ore used : 81 per cent FeO, 9 per cent, free O., . . 1,020 860

Percent Percent.

(4) Metal when melted : Si, . .02 .05

" " Mn, 09 .06

" " C, 64 .64

(6) Metal before tapping : Si, 02 .01

Mn, 04 .02

C, 13 .12

(6) Slag just after melting :SiO, 60.24 49.46

MnO, 21.67 1316

" " FeO, 23.91 33.27

(7) Slag before tapping : SiO 49.40 49.36

" " MnO, 1H.60 11.30

" FeO, 29.79 34.11

Pounds. Pounds.

(8) Slag nnder ladle, 4050 5670

(9) Slag before tapping (estimated), 3900 5520

(10) MnO in slag before tapping [From (7) and (9)], . . 644 624

(11) MnO formed after melting [From (4) and (5)], . . 86 28

(12) MnO in slag after melting [(10) — (11)], ... 608 696 ,

40S

The Open-H Earth Process.

(13) MnO in slag after melting [See (6)], .

(14) Slag after melting [(12) (13)], .

(15) Constituents oxidized during the melting-

ing assumed to be 55,003 pounds:

Per cent. 21.G7

Per cent.

Pounds. Poands. . 2810 ' 4530

period ; weight of metal after melt-

Group I.

Pounds. O req'd.

Si, 219 250

Mn [see (12i], 473 135

Fe [from (6) and (14)], . . . 62:< 149

C, 279 372

Pounds of oxygen required, . . . 906 (16) Constituents oxidized during the oreing :

Group II. Pounds. O req'd.

Si.

Mn,

Fe,

Group I. Pounds. O req'd.

Pounds of oxygen required.

Group II. Poundfl. O req'd. .. 22 25

Summary of Above Results,

Group I. Pounds.

(17) Oxygen supplied to the bath during melting [See (15)], 900

(18) Oxygen supplied to the bath during oreing [See (16)], 303

(19) Total oxygen supplied to the bath [(17) + (18)], . 1214

Per cent.

(20) Oxygen supplied during melting [(17) -7- (19)], . . 74.6

(21) Oxygen supplied during oreing [(18) -4- (19)], . . 25.4

Pounds

(22) FeO in slag after melting [From (6) and (14)], . . 672

(23) FeO in ore added [See (3)], 826

(24) FeO which would have been pre.ent in final slag if no

reduction had taken place [(22) -h (23)], . 1498

(25) FeO in final slag [See (7) and (9)] 1162

(26) FeO re<luced during oreing [(24) — (25)]j ... 336

(27) Fe in the FeO reduced [See ('20)], 261

(28) Fe in ore added [See (3)], 643

Per cent.

(29) Ore reduced [(26) H- (23)] 41.

Pounds.

(30) Metallic iron in final slag [See (25)]. 904

Group 11.

Pounds.

Per cent.

Pounds.

Per cent.

Pounds. 146)

Carbon is here calculated to CO, while in my former disctission the produc- tion of COa was predicated. In that case the effect of oxidation upon the products of combustion was being cmsidered, and since CO is burned to CO as soon as it rises into the flame, the former premise was valid. The present case, however, deals with the protective action of the carbon, and as this element leaves the surface of the bath as soon as its lower oxide has been formed, it is evident that the calcu- lation must no rest on a new basis.

The Ope.V-Heakth Process. 409

Sec. 29. — General Review of the Quanliiatwe Remits.

. Comparing the results of the oil-flame (Sec. 28, group II.), with those of producer-gas (group I.), we find that with the hydrogenated fuel containing steam, the stronger oxidizing action necessitated the use of a greater quantity of pig-iron, but yet less ore was required, for although the carbon was higher when the bath was melted, the greater power of the oil-flarae was felt as well after fusion as during the first part of the operation. The burning of the greater quantity of silicon in the pig, and the increased scorification of the hearth due to the more oxidizing flame, produced a larger amount of silica, and since this component regulated itself to almost e;cactly the same per- centage in the two slags (as observed in See. 27), it took from the bath the bases that were required to satisfy it. The percentage of MnOand FeO together is nearly the same in the two slags, being 46.29 per cent, for group I. and 45.41 per cent, for group II. This evidence that the increased amount of silica grasped a proportionate quantity of bases necessarily implies a corresponding increase in the quantity of slag, as the record, indeed, shows, there being 5520 pounds of slag for the oil- as against 3900 pounds for the coal- group.

The quantity of manganese present was limited by the composi- tion of the original charge. The total content of oxide of manganese in the slag before tapping (see item 10) was 644 pounds for group I. and 624 pounds for group II., indicating 499 pounds and 483 pounds respectively of metallic manganese in the original charges (neglecting the small and unavailable percentage held by the metal throughout the operation). When the demand came from the silica for still more bases, the only means of meeting it was by oxidizing the iron of the bath. The metallic iron in the final slag is 904 pounds in group I. and 14C5 pounds in group II. (see item 30). Thus the presence of an excess of silica is a cumulative evil, since it calls not only for a proportional waste of iron, but for an ad<litional quantity to supply the deficit of manganese.

Sec. 30. — History of the Ore-Additions.

By referring to See. 28, item 29. it will be noticed that the per- centage of ore reduced was nearly the same in both groups. This is an accidental circumstance rather than an illustration of a general law. The argument has already been elaborated that the slag, upon reach- ing its proper condition, has no further strong tendency to perma-

410 THE OPEy-HEARTH PROCESS.

nently absorb more basef, and that any oxide of iron then added is free to react upon the carbon of the bath with the formation of metallic iron. It therefore follows, if the same quantity of ore he added to successive heats, and all other governing conditions remain constant, that the amount as well as the percentage of ore reduced by each heat will be the same; but if the quantity of ore varies, the calculation of percentages is useless. Each pound of ore carries a certain quantity of silica; the erosion of the hearth is continually supplying more, and some oxide of iron will be consumed in main- taining the proper composition of the slag. The remainder will be chemically reduced so long as the bath contains silicon, manganese, or carbon in any considerable amount. It is evident, therefore, that not only the quantity reduced, but the percentage also, rises with the amount of ore added.

Sec. 31. — Quantitative Investigations into the Oeneral History of the

Ore-Additions.

In order to equate the action of the ore, the following investiga- tions were made on charges which differed widely in the amount of ore they contained. In some cases where only single heats were ex- amined at one time, care was taken to select such as should be typi- cal of regular practice, in order that the difficulty, alluded to in Sec 28, of obtaining accurate data for the slag, might be in a measure overcome by the fact that the slag which remained in the furnace after each heat was equal in quantity and kind, as nearly as could be estimated, to that which the heat had taken up from the preceding charge. By thus selecting several representative heats for each class of charges, the average of each group will fairly exhibit the law of its kind. Each group embraces from two to five members. The results obtained from the two groups treated in Si?c. 28 are also given for comparison :

Group I., no ore used.

Group II., 100 to 500 pounds of ore used.

Group III., 600 to 1000 pounds of ore used.

Group IV., 1100 to 1500 pounds of ore used.

Group v., 1600 to 2500 pounds of ore used.

Composition of the Slag, — (1), after melting ; (2), before tapping.

The varying percentages of FeO in the final slags may seem to conflict with the allege<I tendency of the slag to attain a regular composition (see Sec. 27). A comparison with the last column of the table will show, however, that a high percentage of iron arises

The Open-Hearth Process.

Table 2. — Variatioks in the Composition op the Slao.

MnO.

FeO.

FeO MnO.

GroQp I

Ii

20.H8

(1)

" Iii

" Iv

V

19 heats with pro- ducer-gas

(See section 28.)

6 heats with oil- gas

(See section 28.)

from a deficiency of manganese. The amount of manganese is limited by the character of the charge, and all further basic supplies must neces.sarily be drawn from the iron (compare Sec. 29). The results of this last column show a remarkable agreement and regularity, which argues strongly for the validity of the work.

CompasUion of the Metal. — (1), after melting; (2), before tapping.

Table 3. — Variations in the Composition of the Metal.

Si.

Mn.

GrouD I..'.

Hi

V

19 heats with producer-gas

(Section 28.) 6 heats with oil-gas

(Section 28.)

The results of table 3 will be discussed in Sec. 33 ; they are in- troduced here, however, because the manganese data are required for the quantitative determination of the slag.

The carbon of groups I., II., and III. apparently follows no regular law. It would be naturally expected that group I., with .36 percent, of carbon when melted, would require more ore than group 11. with only .18 per cent. That it did not, may be explained by supposing that the samples of the first group were taken too soon after melting, and that by the time the metal was sufficiently hot the carbon had been eliminated, while in the second group the tests

The Opex-Hearth Pbocess.

may have been made afler the bath had been properly superheated and when it was ready to receive the ore. Every furnaoe-man will understand that it was impossible to establish any definite point which should be designated as "after melting;" and further, that the figures in the carbon column, but no others, depend wholly on such a determination.'*'

Reduction of the FeO. — The following is a comparison of the groups with regard to the amount of FeO reduced during oreing, the method of calculation being the same as in Sec. 28.

(a) Pounds of FeO added in the ore.

(6) Pounds of FeO reduced during oreing. .

(c) Percentage of the FeO addition reduced.

(d) Pounds of FeO permanently added to the slag by the ore-addi- tions (a-5).

Table 4. — Showing the Reduction op FeO.

Group I

"'2*7

11

Iii

" V

19 heats with producer-gas (Section 28j

6 Jieats with oil-gas (Section 28)

It will be noticed that the items in column (0) of table 4 corrobo- rate the statement made in Sec. 30, that the percentage of ferrous oxide which is reduced rises with the quantity that is adde<l. In group I., where no ore was used, it appears by column (6) that seven pounds of oxide were reduced from the slag during the superheating period, which in this group corresponds to the period of oreing. Although this small amount is far within the limit of error of the determination, yet such a result is quite probable. As already pointed out (Sec. 25 d), the reactions during melting are determined by local environment, and just at the moment when fusion is accom- plished the slag as a whole may be regarded as an aggregation of small |K)ol8 of cinder, each being the result of special conditions. The average may contain a greater or a less amount of oxide of iron than is required to secure a free-working slag.

If high-silicon pig-iron has been used in greater quantity than

Further remarks on table 3 will be found in Sec 33, page 416.

The Open-Hearth Process. 413

8ailice8 for protection, the slag will contain an excess of nilica, which will demand bases either from the bath or from ore-additions; but if the amount of pig-iron has been small, so that, as in the present case, the bath contains no excess of carbon, and if at the same time the content of silicon be low, there may be more than sufficient oxide of iron in the slag to satisfy the silica, in which case the sur- plus will be reduced by the carbon of the bath.

In group II. it is possible that an excess of ore may have heen used, — a more likely occurrence when the melted charge is low in carbon than when the bath has been in good working order and has been run down under complete control. The fact, already noticed, that 223 pounds of FeO were used when the carbon was only .18 per cent., would indicate that too much ore may have been added, and this view is somewhat strengthened by the analysis of the slag, which shows lower silica than any other group.

In groups III., IV., and V. sufficient ore was used to mask the irregularity whih arises from the readjustment of the slag imme- diately after melting, and to show that the action was under complete control. It is true that the figures in column (d) are not alike for the three groups, nor do we find, as theory would lead us to expect, that they increase slightly with the greater amounts of ore, yet the practical metallurgist, who appreciates the possible errors and diffi- culties of such investigations, will interpret the results liberally, and find in them a verification of the statement (Sec. 27) that the slag has no tendency to permanently hold an indefinite quantity of bases, but that when ore is added it may be looked upon as free oxide which can work out its own history with the reducing elements of the bath.

Sec. 32. — Quantitative Investigations into the History of the Ore at Different Periods.

The action of the ore may be better understood by examining the results at different stages of the operation. For this purpose four heats are selected, in each of which 1500 pounds of ore were used, and separate calculations are made for each. The average may be supposed to represent true conditions. It would be preferable to have consecutive charges, but it is a mere coincidence when exactly the same quantity of ore is used in successive heats ; hence, the al)ove plan is more practicable. The amount of slag varies widely in the individual heats, but this does not signify any error. The silicon in the charge, the sand clinging to the surface of the pig-iron, and the

The Op£K>Heabth Pboce88,

erosion of the bottom, determine the quantity of the slag. The fact that the figures show a regular progression in each case is indicative of correctness and of the constancy of the relations.

For the following analyses samples were taken at four different times, namely :

(1 ) After melting and before any ore was added.

(2) After the addition of 500 pounds of ore.

(3) After the addition of 1000 pounds of ore.

(4) After the addition of 1500 pounds of ore or just before tapping.

Table 5. — CoMPosmoN op the Slao at Different Periods.

Number of Heat

Pounds of

CoMstltuente,

After addition of ore, as

ore added.

shown in Ist column.

Average

None.

SiO,, per cent

62.4?

i.

u

a u

None.

MnO, per cent

u

U M

M M

None.

FeO, per cent

M U

30r68

U M

2H96

None.

FeO and MnO, per ct

Table 6. — Weight of the Slag, in Pounds.

Heat No.

Weight of steel Upped.

Pounds of Ore added.

None.

61,600 61,900 51,200 54,500

Average

52,300

Incremen eacli inte

lbs. 2151

bs. 282 lbs.

THE OPEN-HEABTH PROCiSS.

Table 7.— Composition of the Metal at Different Pebiodb.

Heat No.

Silicon, per cent

Manganese, per cent.

AAer adding Ore, as below.

After adding Ore, as below.

None.

5

Ibe.

None.

50U lbs.

lbs.

763)

undet.

M

undet.

.01 !

.00 ,

trace

The above results will be discussed in Sec. 33. Reduction of the FeO. — The weight of the FtO iu the slag is as follows :

Table 8. — Iron Oxide in the Sijg at Different Periods.

Heat No.

Poun After

ds of FeO in the Slag.

adding Ore, as below.

None.

lbs.

43'i

Av...

From these figures the amount of oxide of iron added to the slag, during each of the intervals, may be deduced by subtracting the average of each column from the average of the next one, thus :

Average, as above, 709, 803, 856, 991.

First

Second

Third

Interval.

Inter'aU

S62

(1 ) Increment of FeO, pound.*, .

(2) Amount FeO added in ore, pounds,

(3) Amount FeO reduced (2—1), pounds,

This shows, that during the addition of the first 500 pounds of ore the amount of iron oxide reduced was, practically, the same as during the addition of the second 500 pounds; but, in the third period, there was a slight decrease in the amount reduced and an increase in

The Open-Health Process.

the amount imparted to the slag. A reference to the percentage of silica in the slags, and to the total content of the oxides of manganese and iron, shows that this increase was not accompanied by any in- creased basicity of the slag, although this may easily occur in careless or ignorant practice after the carbon and other reducing elements are nearly eliminated.

Sec. Z,— History of the Melal,

With regard to the metal, the results of the foregoing investiga- tion may be tabulated as follows :

Table 9.— Showing the Effects op Ore-Additions on the Composition of the Metal.

Foreign Elements oftho MetaL

Before UppiDg,

laagu- ( After melting,

I 1 eent ( Before tappiig,

Before tippiig,

2 i

o V

,05

u

unde

term

V o

trace

ined

The calculations on the two groups of charges in Sec. 28 show, that in the case of the soft-coal producer-gas heats, 74.6 per cent, of the total work of oxidation was done during the melting, as against 78.8 per cent, in the heats with oil-fuel. These percentages are purely empirical and special, depending primarily on the total con- tent of oxidizable material. It may be assumed that the amount of oxidation during the melting period is constant, so that the per- centage will depend upon the excess of oxidizing agents in the charge. The circumstances which modify this general law have al- ready been discussed (Sec. 25). If the proportion of pig-iron is reduced to the lowest poasible point, so. that no ore is required, the oxidation during fusion will approach 100 per cent, of the total,

The Open-Hearth Process. 417

while if the charge oontains no scrapi the greater part of the oxi- dation must be done after the fusion is completed.

The composition of the metal, just after fusion, will be dependent in great measure upon the excess of oxidizable elements in the orig- inal charge. In the two groups referred to (Sec. 28), the composi- tion of the stock as it went into the furnace was about as follows :

Stock for group I. : Silicon 0.40, manganese 0.90, carbon 0.95.

Stock for group II. : Silicon 0.70, manganese 0.80, carbon 1.45.

All the charges referred to in the foregoing pages, with the excep- tion of group II., in Sec. 28, were made with sofl-coal producer- gas, and hence will be fairly represented by group I. In this prac- tice the melted metal is found by the above summary, table 9, to show the following composition :

Metal of group I., after melting: Silicon from .01 to .13 per cent.; manganese, from .02 to .19 per cent.

Metal of group I., before tapping: Silicon from .01 to .06 per cent. ; manganese, from a trace to .10 per cent.

The tendency, therefore, is toward the almost complete elimination of silicon and manganese during the melting period, the carbon being protected by their greater aflBnity for oxygen. When the quantity of these two elements in the original charge exceeds the amount re- quired to absorb the oxygen that has to be taken up during melting, the surplus will be found in the metal; almost all of it will be burned by the ore, however, before the carbon is affected.

In the above table, heat No. 7635 exhibits high silicon and man- ganese, but the greater part of the excess was eliminated in the firnt stage of the oreing (compare Sec. 32). A further examination of the same table shows that in those groups for which less than 1500* pounds of ore was required, the highest silicon afler melting, and previous to adding the ore, was .05 per cent, and the highest mani- ganese .09 percent. Before tapping, the silicon had been reducedito- .03 as a maximum, and the manganese to .04.

The amount of ore required for a charge will not follow closely the- content of carbon. The flame is constantly at work, and ore is addedt when the melter thinks it advisable, rather than when it is absolutely, necessary. If the charge is hot it dissolves the ore rapidly, leaving to the flame little chance to do its share of the oxidation. If the charge is cold, only a small amount of oi*e will be added, and the oxygen will be derived from the gases. Thus any attempt to make an arbitrary equation of the action must fail. It may be broadly said, that if the bath contains 1 per cent, of carbon, 1500 pounds of VOL. xxn. — 27 T

418 The Open-He Abth Process.

ore may be used in bringing it down to .08 per cent. The first 500 pounds will reduce it to about .80 carbon, the second to .40, and the third will finish the work. If silicon and manganese should be as low during the interval between the first and second ore-additions as at a later time, the burning of carbon might be the same then as later, but either the presence of these protectors, or the less favorable physical condition of the slag which accompanies high carbons, re- tards the action at the start. When large quantities of high-silicon or high-manganese pig-iron are used, the first additions of ore are consumed by the unburned excess of these elements, and hundreds or even thousands of pounds of ore may be added before the car- bon is affected.

If the temperature of the metal is very high, the last traces of silicon are not oxidized. It is well known that with the Bessemer process steel can be made containing 1 per cent, of silicon, if blown sufficiently hot, since the relative affinity of oxygen for silicon and carbon is a function of the temperature. In the open hearth the combustion of silicon is very slow when no ore is used, and under the non-conducting slag the metal has no chance to attain the extreme temperature that can be reached in the converter. On the other hand, where ore is used, the repeated additions tend to keep the bath cool and facie'tate the complete elimination of the silicon. The open-hearth, therefore, cannot rival the converter in producing high-silicon metal by non-combustion, but under suitable conditions the amount carried along in the metal may be quite appreciable, and by holding the bath at a very high temperature, with a siliceous slag, there will even be a reduction of the silica of the hearth according to the equation :

SiQa + 20 Si + 20O.

This variation in the affinity of silicon and carbon for oxygen is only one of a large number of similar chemical phenomena, some of which will be more ftilly treated in Chapter VIII. (page 462), under the head of preferential relations.

Thus far no account has been given of phosphorus and sulphur. Numerous determinations and universal experience prove that the content of phosphorus in the steel will be determined by the initial content of the charge. Every particle of it that occurs in the pig- iron, scrap and ore will appear in the finislied metal.

The percentage of sulphur in the metal cannot be predicted with quite so much precision. Traces of this element may be burned

The Open-Hearth Process. 419

during melting and may pass away as sulphurous anhydride, though the proportion thus eliminated is small. On the other hand, there is a tendency to absorb sulphur from the gases. If the coal is reasonably free of this impurity, the increment from this source is unimportant, but with impure fuel a long exposure to the flame may result in raising the sulphur-content from .05 to .25 per cent. It seems reasonable to suppose that the absorbent power will be stronger if the metal be pure. Thus, Stead says that metal with .01 per cent, sulphur gives steel with .025 to .040 per cent. This is a greater increment than will be obtained from good fuel and ore and stock containing .04 to .06 per cent, sulphur. It has been plausibly argued that the percentage of lime in the coal will de- termine in great measure the deleterious power of the sulphur ; for should lime be present in sufficient quantity, most of the sulphur, it is sa'd, will be retained in the ash at the producer, and thus play no part in the history of the bath.

That the sulphur of the ore must be an important factor seems very evident. Willis f claims that 30 per cent, of the sulphur of th ore is alisorbe<l by the metal, but in doing so he erroneously assumes that the whole increase of the sulphur-content is derived from the ore-addition, regardless of other contributory influences. The proportion that is absorbed will depend largely upon accident. If the sulphur in a certain lump of ore is contained in a piece of pyrite, and that lump floats in the bath in such a way that the pyrites is exposed to the flame, it is probable that the sulphur will escape as sulphurous anhydride; but if the pyrite be submerged in the metal, it is quite probable that the sulphur will be absorbed. The quantitative history cannot, therefore, be accurately equated.

Chapter VII. — The Basic Process.

The basic process, as herein discussed, consists in melting a charge of pig-iron, or a mixture of pig-iron and low-carbon metal, upon a hearth of dolomite, lime, magnesite or other basic or neutral material, and converting it into steel in the presence of a stable basic slag, by the action of the flame, with or without the use of ore, and by the addition of the proper recarburizers.

On the Elimination of Sulphur from Iron/' by J. £. Stead. Journal cf the Iron and SUel InstUuUf vol. ii., 1892, pp. 223 et 8eq.

t Reactions in the Open-Hearth Process," by A. Willis. Journal oj the Iron and Steel InetituU, vol. i., 1880, p. 91.

420 The Open-Heabth Pbocb38.

Sec. Si.— Pseudo-Basic PraoUoe. The experiment has been tried in acid practice of adding lime shortly before tapping, the object being to make the slag somewhat basic, and thereby remove a part of the phosphorus and sulphur. As the period of exposure is brief, the damage to the bottom is not disastrous, and such practice has been in a measure successful, with- out, however, meriting the name of a process. The results both for the steel and the bottom roust be uncertain, and the practice can be viewed only as a temporizing expedient. For rular and positive work it is obvious that, within practical limits, conditions of equi- librium must be present, and since in the basic as in the acid process the slag is the active agent, it follovrs that for basic working the bottom must be of such a character that it will either withstand the scorifying action, or that its erosion will not seriously affect the composition of the slag.

Sec. 35. — Material for Basio Hearths, To meet these requirements, the hearth must be either basic or neutral — it is immaterial which. This may be a surprising state- ment, but the fact cannot be too strongly insisted on that, in the basic as well as in the acid pnxess, the hearth is only a passive agent. It is simply the retort in which chemical work is carried on. Among the various substances which have been used to make a proper bottom, carbon, bauxite, chromite, magnesite, lime, and dolomite may be mentioned.

Of these, carbon is the only one which may properly be called neutral, though the term is oilen incorrectly applied to bauxite, mag- nesite, and chromite. Oxide of magnesium and oxide of aluminum are as true bases as any substances in nature or in art. The fact that in a sintered condition they resist the attacks of both silica and basic oxides has no bearing on the issue. Absolutely pure CaO and absolutely pure SiO, do not readily unite when in contact, as both melt only at the highest temperatures, and rapid chemical union is not possible without fusion. Yet it would hardly do to call these substances neutral to each other. In the same way the vitrifi- cation of magnesian limestone converts it into a state in which it is acted upon very slowly by water and resists the action of scoriae. It would be a mistake to suppose that the vitrification had in the least altered its chemical composition, — it cannot be said even to have assumed an allotropic form. The phenomenon is one which every chemist will recognize simply as a change in solubility induced by an

The Open-Hearth Process. 421

alteration of physical structure. As no line can be drawn between lime and dolomite, so none can be drawn between dolomite and mag- nesite. All are bases pure and simple, and it would be a matter of much regret if, with regard to them, an improper term should become established in metallurgical nomenclature.

Chromite occupies a somewhat different position. In one sense it is a salt, — satisfaction of chromous acid by iron oxide ; in another it is a collection of bases, for chromous anhydride is an anomalous body, having as its fundamental element one which frequently acts as a base with ordinary acids. When the mineral does yield to scori- fication, the only way in which it can act as an acid is by dissolving unaltered. Alteration into a chromate condition is inconceivable under exposure to the reducing action of the bath ; on the other hand, decomposition with attendant deoxidation is quite possible, for part of the eroded material may be reduced to metallic chromium and iron, while the remainder enters the slag as the basic factor of a complex silicate. Thus, on the whole, since it is only the dissolved portion of any lining which can affect the chemical equation, and since chromite can never take an aggressive acid part, and for some reasons may be regarded as basic, it would seem hardly correct to style it neutral.

Carbon, — Attempts have been made to use this material for certain portions of the lining, but although not seriously affected by the ac- tion of slag, it is so readily absorbed by the metal that its use is out of the question. It is mentioned here, not on account of any merit of its own, but because the suggestion to so use it is a proposition of perennial growth which may not be omitted from the catalogue.

Bauxite, — This mineral, when pure, withstands intense heat and resists the action of basic slags. Its great shrinkage upon heating, however, causes it to crack and renders it unsuited for use in a hearth. Thorough calcining might, perhaps, remove this difficulty, but it would also destroy the plasticity of the bauxite, which for some purposes is its most attractive feature.

Chromite. — Chromite of iron is one of the most infusible of sub- stances and one of the best to resist both acid and basic influences. Owing to these very characteristics it is difficult to fuse or sinter it into a solid and impermeable mass, and in some places, where bot- toms have been made of it, disastrous erosion has resulted from the unsatisfactory mechanical structure.

Magnesite. — Without any doubt magnesite is the best material for the basic hearth. When properly pul in, by setting thin layers in

422 The Op£N-H Earth Process.

the same manner as sand-bottoms are built, the cutting by the slag is very slight, but the rarity of the mineral, the expensive manipu- lation required to properly calcine it, the long freightage, and in this country a heavy import-duty, render its cost almost prohibitory. Magnesite bricks have been nsed for the construction of furnace- roofs and walls, — in some places with reported success, — but the author is ready to record that his attempts to use them on overhang- ing work have been total and expensive failures. The bricks were purchased through responsible agents, and were supposed to be of the best, but they shrunk to such a degree that every joint opened wide and every brick cracked, and the material itself softened under the intense heat. Each brick became a sort of separate stalactite, and with the breaking and dropping of the exposed portion, a new part was exposed to pass through the same experience. In the case of the hearth, the temperature is not so high, and the softening is a matter of less consequence. There, if necessary, the mineral may be used without thorough calcining, but as the shrinkage and distilla- tion have to occur somewhere, it would seem unprofitable to use the open-hearth furnace for a kiln. A bottom made of unburned magne- site must be very porous, and this implies a large absorption of slag and metal and a very unreliable hearth.

Lime. — The cheapest and most abundant basic material is burned lime. Theoretically, it is perfectly suited for making a hearth. Practically, it is open to objection, for it slakes so readily on expo- sure to the air that it cannot be kept in stock. If the bottom is built of partially slaked or imperfectly burned lime, the dissociation of the water and gas, when the furnace is heated, causes a disintra- tion of the mass, and it is then easily eroded by the metal.

DolomiU, — The difficulties encountered in the use of lime from quick slaking, have been obviated by the substitution of magnesian limestone or dolomite. This material, when thoroughly burned, ground, and well mixed with tar, will keep for several weeks in a dry place, an advantage which fully compensates for the slight addi- tional expense. The bottom can be built by ramming successive layers of loose stuff, or it can be made of unburned bricks which have been pressed into convenient shape. The only advantages of the bricks lie in the facility of handling, and the certainty that every part of the hearth will be compact.

Sec. 36. — The Passive Joint. The poor results from using magnesite-brick on overhead-work

The Open-Hearth Process. 423

have already been set forth. Unsatisfactory results were h'kewise obtained in making and using bricks of well-burned dolomite. With both these materials, the shrinkage is dependent upon the tem- perature and the time of exposure. A piece which has been bume<l and shrunk at a certain temperature will contract still further if heated to a higher point, or if exposed to the same temperature for a longer time. The same is probably true of bauxite and lime. Possibly chromite can be made into bricks which will be capable of withstanding the strains, the heat, and the variations in tempera- ture and shape, to which furnace-roofs are subject. Conclusive ex- periments on this point are expensive, and there seems to be no vital need of determining the question, as silica roofs and walls can be used with good results, when the silica work is separated from the basic hearth by a passive joint. Bauxite, chromite, mjtgnesite, and carbon have been tried for the separating medium, and of these chromite is doubtless the best, but success depends in this case, as in so many other details of practical work, quite as much upon the method of application as on the material. It is a safe assumption that in an open-hearth furnace any two snbstances will slag together if they are in contact under pressure and exposed to the direct flame. Hence the design should be such that when the joint is subjected to pressure it will be protected from the intense heat. (Compare Sec. 35, which also explains why " passive " is preferred to "neutral" in designating and characterizing the joint).

Sec. ST.— The Charge.

The remarks in Sec. 23 upon the proper quantity of pig-iron in the charge for acid work, and upon the disadvantages of a defi- ciency, are applicable in great part to basic practice; but the propor- tion of pig-iron that may be used advantageously is somewhat higher with a basic hearth, since in this case it is less objectionable to mix ore with the metal in the original charge, and by this means the work of oxidation during the melting may be hastened ; but the amount of ore which can thus be used is limited, as the ebullition of carbonic oxide gas causes serious frothing of the bath.*

In the kind of metal that may be used, the newer process admits a most important expansion of the specifications. In acid work sulphur and phosphorus are not eliminated, and therefore, as previously observed, the content in the stock cannot exceed the allowable con- tent in the steel. In basic work, both these impurities may be reduced, and the range of available material is therefore widely

424 The Open-Heabth Process.

extended. This gain, however, is not secured without compensation. Each pound of phosphorus or sulphur involves the consumption of basic material for its chemical satisfaction, while by the change in the character of the lining, the silicon and sand of the pig-iron have be- come enemies, requiring bases for their capture. These basic addi- tions are a source of expense beyond their first cost ; extra labor is required for the handling and charging; extra fuel is needed for the melting; the wear of the furnace is greater on account of the dust which impinges on the acid brickwork, and the checkers are clogged more rapidly. Besides the duration of a heat is longer, as the presence of much ore and lime retards melting, and as a result the wear of the bottom is greater, and the output of the furnace less.

The increased cost due to these factors cannot be accurately stated. No attempt to construct a universal cost-sheet can be successful. The varying prices of iron, scrap, ore, lime, dolomite, bricks and fuel would be sufficient to make the problem impossible, even if the value and efficiency of labor could be disregarded. The machinery of production must also be considered, by which onerous labor may be lightened or dispensed with, and last but not least, the question of management, by which metallurgical disadvantages may be over- come, or fortuitous conditions swept away.

Sec. SS.—The Bade Additions.

In the acid process there is generally sufficient silicon in the pig- iron and in the sand clinging to it, to furnish the silica required for a proper quantity of suitable acid slag. On the basic hearth, bases and not acids are required, and it is evident that the only way in which the bath could supply them would be by the oxidation of iron. But since it is impracticable to maintain an extremely ferruginous slag in contact with highly carburized metal, owing to the continuous reducing action, it follows that the scorification of the bottom can be avoided only by the presence of irreducible bases. In some places magnesite has been used for this purpose, but common limestone, either burned or unburned, answers the purpose so well, that it will be unnecessary to consider anything else. So far as basicity is con- cerned it is immaterial whether CaO is added alone (burned lime), or in combination with CO2 (limestone); but the carbonic anhydride evolved from the stone affects the operation in more than one way.

(1) It acts as an oxidizing agent on the charge, completing the following reactions :

The Open-Hearth Process. 425

Si + 2COa SiOj + 2CO

Mn + CO, MnO + CO

C + Co, 2 Co

Fe + CO, FeO + CO.

These relations allow the use of an increased proportion of pig- iron, which is asaally an advantage.

(2) The volatilization of the CO, necessarily absorbs heat, thereby cooling the interior of the mass of metal where the flame has no chance to supply heat. The melting is thus delayed considerably.

(3) The passage of the gas through the semi-viscous mass and its reaction on the carbon of the metal cause foaming, and this creates a danger that the metal may run out of the doors or down the ports, or that the slag may come in contact with the silica side-walls. On the other hand, if burned lime is used, there is no absorption of heat from dissociation, and no production of gas, and, therefore, it becomes practicable to mix ore with the charge. This acts upon the elements as follows :

3Si + 2Fe,03 3810, + 4Fe 3Mn + FejO, 3MnO + 2Fe 3C + Fe,03 3CO 2Fe.

The reactions represented by these equations are not completed im- mediately. Where varying proportions of FeO, FcjO, and CO, are present, there are numerous intermediate steps, but the ultimate re- sults are the same. The formation of CO occasions frothing, but for a given effect upon the carbon of the bath less gas is produced with ore than with limestone, as the following comparison shows :

(Limestone) CO, +C 2CO. (Ore) Fe,0, 3C 3CO 2Fe.

In the first equation, representing the action of carbonic anhydride from heated limestone, two volumes of CO are formed for every atom of carbon taken fi*om the stock, while in the second only one volume is produced. In the thermal relations there is just as vital a difference between the two equations. In the first, the formation of CO from CO, is a reducing process which absorbs heat; in the second, the oxidation of three molecules of C to CO liberates more energy than is absorbed by the decomposition of Fe,03. Hence lime-

426 The Open-Hearth Process.

stone acts as a refrigerating agent, owing both to the distillation of its carbonic anhydride and to the action of the gas upon the metal- loids, while ore acts as a calorific agent during its dissociation. The relative oxidizing powers can easily be computed :

CaCO, CaO + CO + 0 100 =56 +28 +16

1 fi Available oxygen 160 pounds per thousand of CaCO,.

Fe20s 2Fe + aO

160 =112+48

Available oxygen --7 300 pounds per thousand of FcjOj.

Thus the oxidizing action of ore is nearly twice that of limestone ; the amount of gas produced per unit of carbon is just half as much ; the effect is heating rather than cooling, and metallic iron is obtained as a by-product. It is frequently possible to use ore and limestone together, but as there is a limit to the allowable amount of frothing, it is evident that the stone must replace a certain quantity of ore. The actual weights of the reagents which can be used will depend upon the proportion and composition of the pig-iron.

The superiority of the ore over limestone is partly obscured by the fact that the greater frothing of the metal, when the stone is used, causes increased oxidation, and this, with the attendant advantage (or disadvantage) of a longer exposure during melting, produces a some- what higher efficiency for the stone than the above calculation would indicate. Any comparison which ignores these influences is neces- sarily defective.

Against all the strong arguments in favor of using lime and ore instead of raw stone, there is to be set the fact that the burned ma- terial costs considerably more per unit of basic content. The final balance-sheet of profit and loss of the two systems upon which the adoption of one will depend, will be largely affected by local pecu- niary considerations.

Sec. 39. — Method of Charging the Basic Additions.

In the foregoing section it has been assumed that at least a por- tion of the basic material is charged with the metal in order to pro- duce a basic slag as soon as possible. The arrangement of the stock

The Open-Heartii Pboce8S. 427

should be such that during melting the slag-Iine will be protected from the silica formed from the pig-iron. The lower portion, or bottom, of the hearth must also be protected, since sand and drops of siliceous slag are apt to fall through the stock and cut into the lining. It would seem &s though these conditions could be met by placing the basic material where it would cover the vulnerable parts; yet it is found to be bad practice to put large masses of lime or stone in one place in contact with the hearth, as both these substances are almost infusible, and stick to the bottom, occasioning trouble by filling the melting-room and depriving the charge of its proper basic supply. A compromise works best, some of the lime being put near the slag-line and the rest mixed with the stock in such a way that it does not rest on the bottom and yet is in position to catch the silica from the pig. A part of the lime may be reserved and added when necessary, but there is no particular advantage in this practice. The portion of the ore which is put in with the charge should be buried in the metal, so that there will be an opportunity for the chemical reactions. The remainder is thrown in when needed, as in the acid process. The general arrangement of the pig and scrap should be made in accordance with the principles explained in Sec. 25.

Sec. 40. — The Conditions of Oxidation During the MeUing-Period.

The various oxidizing forces which are at work during the melt- ing-period of the acid process (Sec. 25) likewise prevail on the basic hearth, though with several additional and modifying relations. In the former case the production of a large quantity of iron-oxide in- volve<l a waste of metal by causing scorification of the bottom, and the silica so taken up retained a certain proportion of iron-oxide in the slag throughout the oreing period. In basic work this does not occur. The excess of iron-oxide cannot find free silica, and there- fore continues in the slag, unsatisfied, awaiting future reduction. Hence, if ore is mixed with the metal the waste is not augmented, as it is very likely to be on an acid hearth.

In the history of the metalloids, a vital difierenoe is introduced. In contact with basic slag, the sulphur and phosphorus are removed, and they thus become factors in the protection of the metallic iron.

Sec. 41. — The Protective Potter of Phosphorus.

In Sec. 26 the capacity of the metalloids to absorb oxygen is shown. In basic work the following equation of phosphorus must also be considered :

428 The Opek-Heabth Process.

62 + 80 142,

One unit of phosphorus combines with 1.290 units of oxygen.

One unit of oxygen combines with 0.775 unit of phosphorus.

With these figures we maya>mpute the change in protective value which a certain amount of phosphorus exerts in a pig-iron such as that given in Sec. 26.

Pig-iron A (see Sec. 26); oxygen absorbed per 1000 pounds of pig, 58.3 pounds.

In 1000 pounds of metal 1 per cent, or 10 pounds of phosphorus absorb 12.9 pounds of oxygen.

Hence, if the carbon and silicon are not changed, we have :

Ox7gen>ab0orbed

per 1000 poonds

of metal.

Pounds.

Pig-iron as in Sec 26, 5S.3

Thesaroe with 1.00 per cent phosphorus, 71.2

The same with 2.00 per cent phosphorus, S4.1

The same witli 3.00 per cent phosphorus, 97.0

By pursuing the same course of calculation as detailed in Sec. 26, we obtain the following mixtures, each having a total oxygen absorb- ing capacity of 1510 pounds for a charge of 60,000 pounds.

Pounds. Pig .

Pig iron A (see Sec. 26) 20,000 40,000

Pig-iron A + 1.00 per cent P., 15,880 44,120

Pig-iron A + 2.00 per cent P 13,170 46,830

Pig-iron A + 3.00 per cent P., 11,240 48,760

The effect of sulphur does not appear in the calculation for reasons to be explained later (Sec. 46).

Sec. 42. — ITie General Chemical Law.

That oxidation is the normal condition of open-hearth work has already been referred to (Sec. 27), and also, that subject to certain conditions imposed by the formation of ferrous oxide, the most easily oxidizable elements are burned first. In basic work the formation of iron oxide is not so potent a factor in the general history of the bath. The silica derived by oxidation from the silicon of the pig combines with the basic additions, leaving apparently nothing that demands the combustion of iron. The incorporation of iron oxide

The Open-H Earth Process. 429

into an acid slag may be only an instanoe, however, of the general law that all forces tend to work along lines of least resistance, which being interpreted for this case means that a slag will seek to com- bine with anything that promotes fusibility.

If given the opportunity, a siliceous slag absorbs either bases or silica, but preferably bases, and particularly those which impart the greatest fluidity. If this preference should seem to conflict with the behavior of the slag in the acid process, wherein, as may be remem- bered, the slag upon attaining a certain degree of basicity, no longer exhibits a strong tendency to change, the reducing agency of the carbon in the bath should be considered, which there alone checks the further absorption of iron oxide.

Whatever the reason, basic open-hearth slag is always found to contain a certain quantity of ferrous oxide. With an increase in the percentage of lime, the proportion of iron decreases. Inasmuch as the substitution of lime for iron produces a more viscous slag, this would seem to invalidate the theory just advanced, but the decrease in iron oxide here is due to another influence which limits the basic- ity. The more bases that are added, the less necessity is there for an additional amount, since the weight of silica necessarily remains constant; as the need for iron oxide in the slag diminishes or dis- appears, the reducing action of the metalloids comes into play, and the slag bins to be robbed of iron, its most fusible base.

The presence of manganese oxide in the slag modifies, without completely changing the relations just described. By furnishing an additional base and imparting greater fluidity, it tends to render the presence of iron oxide less necessary. This matter of fluidity is of great practical importance: the slag must run from the furnace freely with the metal, else the hearth will soon be filled; furthermore the slag must be so basic that it will not cut the hearth. These two conditions, fluidity and basicity, determine the nature and amount of the basic additions. Roughly speaking, CaO and MgO together cannot exceed 55 per cent, in the final slag without producing a very viscous cinder. The same result will follow from a reduction in silica below 10 per cent, although the presence of unusual amounts

In Sec. 45, Pari JL certain of the data will show that this theory of the auto- matic increase of fluidity is not entirely gratuitous, but represents certain active conditions, profoundly modified perhaps by limitations which will be discussed in Chapter Vlil. (page 462).

430 The Open-Hearth Process.

of the oxides of iron, manganese, and phosphorus alter the relations by introducing liquefying components.

Given the proper additions to fulfil the above conditions, the slag when melted will be somewhat viscous, owing to the prevailing low temperature. It will contain about all the silica and much of the manganese of the charge and all the lime which has not stuck to the hearth. Its content of iron oxide will be determined by the conditions of oxidation during melting (see Sec. 40) and the disposi- tion of the ore in the stock. More truly even than in the acid pro- cess, the slag at. this time may be regarded as an aggration of pools of different cinders, each made under different conditions (see Sees. 25 and 31); but as it gradually becomes completely fluid the various separate relations are readjusted to a proj)er average. The silica finds bases in the lime, manganese and iron, or takes them from the hearth ; the manganese burns to MnO and unites with the silica; the phosphorus changes to PjO and combines with the lime. The general relations of the iron are the same as on the acid hearth, and here, as there, it holds true, that any addition of ore may be regarded as so much free ferric oxide, which will react upon the re- ducing elements of the bath and form metallic iron (compare Sec. 27).

The practice of melting with an insufficient amount of lime and tapping the slag soon after fusion is sometimes followed, for the pur- pose of avoiding the necessity of thoroughly satisfying the silica so removed. This plan is far more seductive on paper than in the shop, as the removal of any considerable portion of the slag requires great care and trouble. There is also a strong probability of losing steel in the operation and of scorifying the hearth so badly that the extra time consumed in repairs will more than offset any advantage.

The opposite of this practice is to form a very calcareous slag during the melting and to flush off with it a goodly proportion of phosphorus. But this plan is also objectionable on account of the viscous nature of such a cinder. All things c(msidered, the best practice with an ordinary stationary furnace (save, perhaps, when very bad iron is worked and the quantity of slag is excessive) appears to consist in making a good slag at the beginning of the melt and holding all of it in the furnace to the end. With the use of a tilting- furnace (see Sec. 4) the slag may be partly removed by decantation, or almost entirely disposed of by pouring out the entire charge and returning the metal to the furnace.

The Open-Hearth Process.

Sec. 43. — Qtmniitative Investigations into the General History.' The same method of quantitative investigation which has been described for the acid process (Sec. 28) is applicable to basic work. An element of difficulty, however, is introduced, from the fact, already noted, that lime is apt to stick to the hearth. Thus the slags cal- culated for the first part of the heat do not show the proper quantity of bases. Moreover, the viscous character of the cinder militates somewhat against the accuracy of the sample, as the more liquid portions will be most apt to be taken up by the test-ladle. The residue of slag in the furnace afler tapping is also much greater than in acid practice, and the wearing and filling of the hearth vary between wider limits.

In the table below, giving a summary of 17 basic heats, there is an item of 1500 pounds of dolomite. This was used in repairing the hearth between the charges. Before the series was begun, the hearth had l>ecome partially filled with lime and slag, and during the run this deposit was eroded by an amount estimated at 4200 pounds. In the calculation the material thus added to the heats is assumed to consist of one-half CaO and one-half normal slag. The error in this assumption will not be large enough to seriously affect the result.

Table 10. — Data of Seventeen Basic Charges; Five-ton

Furnace; Dolomite Hearth; No Ore with Charge;

Melted witH Oil-gas.

Material used .

Weight. Pounds.

Composition, per cent.

Mn.

p.

8iOs. CaO.

MgO.

FeO.

Bessemer pig...

Basic pig

Scrap

102,000

],('00

2,740

10,200

1,500

[4,200]

0.58 ! .082 1.03 1 3.089 1.11 .094 0.60 0.02

70 -ft

[5.00]

78.1*

Pit scrap

Ore (FeA) ... Lime

Dolomite

Wear of bot'm Ferro

Spieirel

6.100'

. ... 1804

Wear of roof in 17 heats estimated to be 860 lbs. SiO,. Wear of ladle-linings, 1.200 lt)s., of which the SiOs will be about 750 lbs. Sand on pig-iron estimated to be 400 lbs. Note.— Figures enclosed in brackets [ ] are estimates.

The following data were first published in the anther's paper, Ti'anaf xix., p. 128, previously cited, and to it the reader is referred for additional iaformation.

The Open-Heabth Process.

Table 11. — Weights aitd Analyses of Slaos and Metals OF THE Seventeen Basic Charges op Table 10.

MlTAL.

Slao.

Composition, per ct

Pounds.

Compositioii, percent

Poundfl.

Mn.

P.

SiOfr MnO. ' CaO. ' MgO. FeU. F/h.

A B D

19*2,'750 197,720

19.21 11.12 42.16 6.64' 13.8 5.149 16.37 10 36 42.78 7.87 16.29 4.848 15.08 9.01 42.161 8.45' 20.34 3 850 15.75 14.1 39.05j 10.40. 16.65 2.961

In estimating weight of metal in test C, allowance is made for the addition of the recarburizer, less its oxidized manganese.

A after complete fusion of the metal ; B after beginning of the boil ; C bath ready lor the recarburizer ; D products on leaving the steel-ladle.

In the paper above mentioned, from which these tables are trans- ferred, the detailed calculations are given by which the various com- plicating conditions, due to the wear of the hearth and the roof, are equated and allowed for. The summary of the results is as follows:

Table 12. — Slags at Different Periods of the Seventeen Basic Charges op Tables 10 and 11.

Slag.

Teat

Total Pounds.

SiO,.

A B D

17,470 21,390 2H,264 30,780

3,356 3,502 3,959

4,848

Constituents of the Slag, in Pounds.

MnO.

CaO.

MgO.

FeO.

1,943 2,216 2,365 4,343

7,365

9,151

11,069

12,020

1,160 1,683 2,218 3,201

2,390 3,484 6,340 5,125

1,037 ;

1,011 1

Note. — 1965 pounds of ore added between A and B. 775 " " " BandC.

The deficit of CaO in the slags A and B illustrates the effect produced by much of the lime-addition sticking to the bottom during the first part of the operation.

Calculating the oxidation performed during the melting and after fusion y we have:

Oxygen required.

Pounds.

. 6484 ,

Oxidised

during melting.

Pounds.

Silicon, 808

Manganese, 1224

Phosphorus, 361

Carbon, 3170

Iron(2390 — 87)x 1791

Total oxygen absorbed during melting,

The Open-Hearth Process. 433

Oxidised Oxygen

after ftislon. required.

Pounds. Ponnds.

Silicon, 97 111

Manganese, 212 62

Phosphorus, 64 83

Carbon, 1138 1517

Iron(6340 — 2390 — 174) . . . 2169 617

Total oxygen absorbed afler fusion, 2390

Note. — The subtraction of 87 pounds and 174 pounds in the abore parenthesis is to allow for the 87 pounds of FeO which enters the slag during each period from the scorification of old slag on the hearth. (See original paper, loe. eit,).

Summary.

Pounds. Percent Oxygen absorbed during melting, 6484 73 " " after fusion, 2390 27

8874 100

All the factorR, except the last, are deduced from the alterations in the composition of the metal. All the iron oxide is assumed to be formed by oxidation, which is not strictly correct, since ore is added after fusion, in which the iron exists as FesO,. This point will be considered more fully in Sec. 44.

In my former paper* data for three different .sets of charges were given. The third of those sets was selected for discussion here, as calculation showed that, for that set, the content of phosphorus in the products agreed fairly well with the amount charged in* the stock. In the case of the other two sets there was an* unaccounted loss of phosphorus, and, at the time, it was explained that the results were announced subject to future experiments which should deter- mine whether part of the phosphorus might have been- volatilized, or whether the record simply bore witness to an error in conducting the tests. It is, therefore, with much personal satisfaction, and a sense of scientific vindication, that I can now refer to Professor Wedding's statement f that in basic Bessemer practice there is a volatilization of from thirty to forty per cent, of the total phosphorus when the charge is very hot, while there is no loss in the case of cold heats. This phenomenon should not, hoW'dvery.be made a basis for reviving visionary and impracticable schemes to purify the metal by- heat alone. The history of the acid process- demonstrates that tem-

Loe.eit.

t See Progress of German Practice in Metallurgj,'* 2Vans.,.xiT., p. 367. VOL. XXII.— 28

The Op£N-H£Arth Process.

perature unaided is powerless to do the workand thata preliminary oxidation and basic combination of the phosphorus is necessary.

Sec. 44. — History of the Slagy with Special Reference to FeO.

In the heats which we have been considering no ore was added with the charge. The FeO formed during the melting-period by the oxidizing effect of the flame amounted to 2390 — 87 2303 pounds. (It has been explained that 87 pounds of FeO entered the slag during each period from the scorification of old slag on the hearth.) After fusion, the quantity of FeO added to the slag from all sources was 6340 — 2390 2960 pounds. Of this amount the slag from the hearth furnished 174 pounds. The 2740 pounds of ore furnished 2422 pounds of FeO, carrying 1884 pounds of Fe. The ore also contained 269 pounds of free O. This free O served its purpose in burning the alloyed elements, but it would seem, at first, as though the Fe had l>een entirely wasted, since more FeO entered the slag during the period of fusion than was contained in the ore. This, as we shall presently see, would be a wrong inference.

The analyses of slags A and C are here repeated for convenience :

Table 13. — Composition op Slag after Complete Fusion (A), AND after Beginning of the Boil (B).

Slag.

Constituents, Per Cent.

Test. Pounds.

MnO.

CaO.

MgO.

FeO.

A

17,470 26,264

42J6

It may be argued, that if slag A can contain 19.21 per cent, of SiOg and still hold over five per cent, of PjO, it should be unneces- sary to dilute the acids and reduce the PjO to less than four per cent. A furnace, however, must be run on practical as well as theoretical lines, and two broad rules may be said to govern the operation :

First. — If the charge be low in phosphorus, so that, when melted, the amount to be removed is small, the dilution of the slag by the incorporation of undissolved lime and ore will be such that the per- centage of P3O2 in the final slag will be less than in the cinder im- mediately after fusion.

The Open-Hearth Process. 435

Second. — If the charge be high in phosphoras, so that, when melted, the amount to be removed is large, it will be found necessary to add bases to complete the work, and the dilution of the slag from these additions, ami from the solution of undissolved portions of the original charge, will lower the percentage of PjO in the final slag.

If with the low-phosphorus charge the addition of basic material be reduced to such a point that none remains for the slag to pick op during the last part of the operation, the bottom will be scorified. On the other hand, with high phosphorus, if an attempt be made to keep down the lime, a long lime will be required for the removal of the phosphorus, and the lengthened exposure will be accompanied by a cutting of the hearth and waste of iron to form the required base. These are not a priori deductions, nor yet rigid laws suscep- tible of proof, for it is quite possible to create exceptions by unusual circumstances, but, as expressions of prevailing rules of practice, they represent actual experience.

For reasons which will appear later, a slag becomes more stable as its content of phosphoric anhydride diminishes, and since forces work along lines of least resistance, we should expect to note a ten- dency, however slight, toward an absorption of bases by the slag for the dilution of this component. Other and more powerful forces are also at work. SiO, and FeO are both called for by the slag, since both promote fluidity. The supply of SiO, is limited, but the amount of FeO is determined only by the reducing power of the bath. If the supply of CaO is ample, oxide of iron is held so feebly by the slag that the carbon of the bath reduces it ; but if the supply of CaO is low, the slag tends to hold the bases more tenaciously. With high silicon and carbon the reducing action of the bath may be very potent, whereas with low carbon reduction does not readily occur. Since FeO imparts fluidity, no attempt is made in practice to drive all the iron out of the cinder. Between the operations of the melter, therefore, and the internal relations of the components, tlie composition of the resultant slag varies between quite narrow limits. In a series of twenty-seven heats, of 15 tons each, made from poor stock, a portion of the slag was poured off soon after melting. During the last part of each heat an ad(iition of lime was necessary, on account of the absorption of silica and phosphoric acid. The addition was regulated entirely by the appear- ance of the cinder and the metal. The results were as follows, the analysis being made on an average sample of all the slags :

The Open-Hearth Process.

Table 14. — Average SLAChANALYsrs op Twenty-seven

Heats.

Peb Csirr.

810,.

CaO.

FeO.

Alter melting Before tapping...

Four-fifths of the lime was added with the chai and the re- mainder after melting, while about 400 pounds of ore was used to decarburize each heat. In spite of the incorporation of this basic materia] in the slag during the interval between the two stages at which the samples were taken, it will be seen that by careful super- vision, and through the action of the internal chemical forces, a remarkably uniform composition was maintained.

Sec. 45. — The Elimination of Phoaphortis.

. By means of such investigations as those in Sec. 43, the general rule of good practice may be discovered, without disclosing, how- ever, the relative influence of all the special conditions involved. This additional information cannot be easily gathered, yet the study of a variety of metals and slags may at least show that certain reac- tions can take place under more than one condition, even though it does not define the limits of their occurrence.

Part L — Dephosphorizaiion during Mdting, — In 1886 F. W. Harbord endeavored to show, from an investigation of two charges, that the amount of dephosphorization during melting would depend upon the percentage and kind of oxide of iron in the slag during that period. His thesis cannot be considered proven, for not with- out a thorough knowledge of all the attendant conditions, nor without carefully weighing the influences due to arrangement of stock, char- acter of flame, and the many minor details of the operation, can even the most superficial comparison of any two charges be made. From the internal evidence of the paper referred to it would seem that the writer did not possess this necessary information. More than this, it is a physical impossibility to take a true sample of the slag during the melting, as each square foot of its surface is covered by

''Some Preliminary Experiments on the Removal of Metalloids in the Basic Siemens Furnace."— Jtmmo/ of tht Iron and Steel Intiiiute 1886, No. 2, p. 700.

The Open-Hearth Proce89.

a different cinder, and the progress of the reactions in each is con- tinoally changing the relations at every spot. Considerable action between the metal and the basic additions takes place below the surface of the slag, and it is almost impossible to thoroughly investi- gate this part of the operation.

It is merely a truism to say that when all other conditions are equal and favorable, the removal of phosphorus will be hastened by an increase in oxidation. One of the most potent agencies in this removal is the ore charged with the metal. Its effect upon the phosphorus will be largely determined by the quantity of carbon that has to be burned. With a large proportion of scrap in the heaty a little ore exerts great influence on the phosphorus, while in a charge consisting chiefly of pig-iron the oxide must be present in greater quantity.

Whatever may be the temporary and special action in any partic- ular portion of the slag during melting, sufficient lime should always be present to hold any TJO that may be formed, else the reducing action of the bath-carbon will soon force the phosphorus to return to the metal. The presence of lime is also necessary to prevent scorification, as already explained (Sec. 88).

The advantages of a certain percentage of carbon in the bath ailer melting, and the necessity of having a calcareous slag set limits to the oxidizing forces which should be brought into operation during fusion, but it is impossible to fix rules which will apply to all cases. The following data, tables 16 to 18, will not establish any gen-

Table 15. — Phosphorus Elimination — Initial Phosphorus IN the Charge, over 2 per cent.

Charge All Pig-iron.

Ore added

with

charge.

Pounds.

it-

PhoBph'r'8 after melt-

iDg. Per cent

Ph08ph*r*8 removed. Per cent.

Igi

Carbon in metal aft'r

melting.

Per cent.

Slag, after melting. Per cent.

FeO.

PjO..

CaO.

4,500 4,500 5,400

2,29

undt'd.

Ave

nige...

eral law but rather show the scope and sphere of dephosphorization. Each table deals with heats made, from metal of approximately similar initial phosphorus-content. The separate heats are arranged

The Open-Hearth Pk0Ces8.

acoordiDg to the percentage of phosphorus-elimination ; in tables 17 and 18 the list is divided into two equal groups, so as to com- pare the average results of the heats which show the greater dephos-

Table 16. — Phosphorus Elimination — Initial PnosPHtiRus IN THE Charge, between 1 and 2 per cent.

Charge f Pig-iron ; Steel scrap.

1'

a

Ore added

witn

charge.

Pounds.

Initial PbofiphVi Per cent

Phosph'r's

alter melting. Per cent.

X 6 k

Carbon in metal aft'r melting. , Per cent

Slag, after melting. Per cent

S

SiO,.

PeO.

None. None. None. None.

und'td. ti

Average...

phorization and those which show the less. Owing to its lack of homogeneousness during melting, no account is taken of the slag

Table 17. — Phosphorus Elimination — Initial Phosphorus IN the Charge, 0.55 per cent.

Ore added

with

charge.

Pounds.

Phosph'r's

after melting. Per cent

Phosph'r's removed. Percent !

1 Carbon in P metal aft'r § melting. 1 Percent

Slag, after melting. Per cent.

SiOs.

FeO.

CaO.

undet'd.

andet'd.

:l

It

unHet'd.

undet'd.

Are

rage

9

undet'd.

undet'd

8

|10

andet*d.

undel'd.

(C

Ave

rage

Group A shows a comparatively large elimination of phosphorus, t Group B shows a smaller elimination of phosphorus than group A.

THE OPEN-HEABTH PnOCESS.

Until after the fusion is oompleted, but it is assumed that the data for the amount of ore originally added, and for the percentage of carbon after melting will convey some idea of the oxidizing condi- tions during melting, while the composition of the slag after melting, though not actually representing the forces previously at work, will indicate the nature of the environment in which the effects were produced.

Table 18. — Phosphorus Elimination — Initial Phobphobus IN the Charge, 0.19 Per Cent.

Charge Pig-iron. Steel scrap.

Number.

mB ft-

Carbon in

metal after

melting.

Percent

Slag after melting. Per cent.

FeO.

3

4

s

None None None None None None None None None

H.09

Average

19.fi8

n

13 O 16

None None None None None None None None None

. .084

Average

Qronp A shows a comparatively large elimination of phosphorus, t Oroap B shows a smaller elimination of phosphorus than group A.

Part IL — The Final Dephosphorization. — The point of interest to the average furnace-man is the general record of the removal of phosphorus. In the following tables, 1 9 to 22, only those of the preceding heats are selected in which no slag was removed during the operation. The analyses of the final slags and metals do not re- veal all the circumstances which controlled the operation, but they clearly indicate the basic conditions under which the various forces

The Open-Hearth Pbocbb8.

were at work. The consumption of lime is not given, as it is a vari- able quantity, depending upon the amounts of silicon and silica in

Table 19. — Compositiox op Basic Slag — Initial Pho6-

PH0RU8 IN THE ChaRGR, 3 PeR CeNT.

pi

Slag before Tapping, per cent

reo.

PiO,.

MnO.

CaO. MgO.

nndeterm.

undeterm.

Average

die charge. The actual weights are arbitrary and local, but the composition of the slag is governed bj definite laws.

Table 20. — Compostiion op Basic Slag — Initial Phoh- PHORU8 IN the Charge, 1M Per Cent.

B

o

lis

Slag before Tapping, per cent

i:

FeO.

PtO*.

MnO.

CaO.

MgO.

&6

13.32 .

Average

3.60 1

2.57 !

8.45 1

Pq 13

O 16

.0P9

Average

The Op£N.Hearth Process.

4n

The results are classiBed, as before, in tables of similar initial phosphorus-content, each table (except No. 19) being divided into groups, so as to show the composition of the slag corresponding to varyipg degrees of phosphorus-elimination as indicated by the phos- phorus in the ingot-metal (6r8t column). Ifhe lines of division are so drawn as to make the number of members in each group as nearly equal as possible without separating charges that show exactly the same elimination. In each group the individual heats are ar- ranged in the order of efficiency of the dephosphorization, i.e., the lowest phosphorus-content in the ingot metal (first column) heads the list.

Table 21. — Composition op Basic Slag — Initial Phos- phorus IN THE Charge, 0.19 Per Cent.

Group A.

Group B.

Group C.

Group D.

Is

Slag before

h

Slag before

h

Slag before

Tapping, per cent.

Tapping, per cent

£

Tapping, per cexii

Tapping, percent.

:

SlOfr 1 FeO.

SiO,.

FeO.

£-

FeO.

,028

M

24.6?

W

25.8S

22.2/

15.27,

Av.

Ion

21.02,

Av.

By referring to tables 19 to 22, it will l>e noticed that, as a rule, a low content of SiOj accompanies good dephosphorization,yet there are numerous exceptions in which high silica occurs with high elimination, and in each group slags will be found containing more SiO, than the average SiO, of the succeeding group.

The Opex-Heartu Process.

Table 22. — Composition of Basic Slag -Initial Phosphorus

IN THE CHARttB, 0.10 PER CENT.

Group A.

Group C.

Ji

' Oroltp I>,

: 1

sSSlff before Tftppliiff.

°S

ji

Blsffbefbre

SlMg before ' t: 1 3

Tapping, 1 JS lit

BtBgbc(bi

Tapplttg. 1

Tftpplii.

Pt cent, 1

0 J3 .1 per cem.

per penL

ti

mi

£5

Hio,.

reo.

£2

46 ,012

FeO.

P9 .016

BiO,.

FeO.

J3 2

146:.022

Si FW).

.one

L. 22.27'

7.

Ik. 50

20.fil

15.05 Ift.ri

nOOf? ' l-35| 18.00

47 .012

14,47

100 .Olrt

Ik. 27

1U,00 H7|.022

10. 7, 15.7fi

S

.WW 1 16,3,V I9.3tt

4S ,U12

101 .Olft

15.47 Hk'.022

24.73' *.91

.Oue 1 17.TA 17.7J1 4& .Ol

I7,fft

15..S3

iy,yo 149' .oa

21 .48' 12.73

,W 1 .012

24. M

li.7i(

103 .016

24.90 15(>: .O-ifl

22.31! 12.23

e

Am 21.S7i 2S,li

51 .012

21.fi7

n.i;*

H.ao 151 .0-.3

16,63

.nos U-U\ 34. S2

3ffl ;,012

14. M

Wr .01§

]fy.&2

20.31: 1.5J;J>J4

22.4*

A

.008 ; 21.'J0 15.21

fa ,012

*iM.,W

rj.w

106 .oir>

14.36, I5.1i. 044

21 .7H

.(Km I If). hi 26.3(i

M mi

16. 7§

21.1b

107 .mij

16. 1.54 Ai'I*

22.70, 12.64

.Ot ; 18.63' t2l.W9

56 ,,t>12

lOH .0!fi

!S).*6 1.S.V.024

17.85: i-.fti

u

.two i 1,20 1.M

5ft '.ft2

l!*,2a

W 017

12,91 ! I5ti .0'.d4

'Ja,10 11.44

.of(ft nj.aji as.'ifi

57 1.012 16. aa

110 .017

ll,5&

23.12 10.9ft

n

.IKB* 1 lUlKli l.W

!.012 17.2S

2r..l7

111 .017

l.yi ,0i6

21, ua l& 36

ax iri.4i 2a.ai

.ot2i ai.flf*

17. Od

112 .017

19,21

l59,GrjiJ

20.1*5' 15 .4

U

0W ii.3; Ifi.'il

M

.012; 18.90

I8.72!

113 .017

1fl0i,(*27

20.45 ri,V}

la

ft!

.012' 14. la

21 M

U4 .017

24.no

.UIO 11.15 2fKl7

,0131 16. i;

ll' .017

17,33

20,72.' 162 .027

2I.It1 10,45

.niO H.4ft; lii.72

Am 16. ai

'JflM

no .017

24.18 1 lR3u027

.010 >.17! J5,4fi

.0131 U,67

22,27

117 .017

12.rt2

/4.27 1 1041 .tt

19.40 13,fl

au

.010 ! rj.45' 17.30

6,5

.013; 'ib/£-

Uk .017

21, m

i5.ia: 2i,tt

-Oki 'Ham)' Usa

m

Ma, 21 .8H

XA.ri

119 ,Hv7

M.Of; lfi6'.D20

ti.n- H.m

£

Am IK. 21 17.BI

.om 14. fa

n.

121* ,on

22,49

H.O'J

.010 21.fl-2' ii.HO

,013 ±;.n

Ik. 91

121 Am

m..029

.010 1 ir>.27 23.54

f1

.OliJ 19,20

vn .018

31 .72,

169 .Osm

16,78, 21.1*

.010 16.70 lJ*,3f>

,014 '2;L\rS

9.yo

12a .018

16,70

18.'J7i

nu.ftio

ISi.SfJl IJ.l*

aa

.010 i IfiJO lyj8

.OH 22. lV

12Jiy

;24 .018

20, M

171 .Qmo

2i.ra>; 11. lu

.010 1 14.51. 22. Ih

.OH lX:ja

21 .W

lij .018

172 .032

I5,:iii !n.64

.010 1 14. 7o, I.IR

7S

.014 2K65

14.4H1

126' .018 ,

13,10

74.451 8,91

2tt

.aiO , 20.HW' lJi.<W

.014, 22.72

127' .018

17, i5

174 .0 4

21.43 9.45

no

.010 1 Mm 22,46

.Oh 'Zi.I&

14, M

ah57

16.3fl;

175 .flJW

24.0fti lo.go

.010 1 VxM MM

-fi

.OH 23.ft7

129 .Oij*

W.4&

17,00

17 .OlM

21 .M

3-J

,010 Ifl.HO, \ti.m

77 .014 1S.5& Is.M

laol .01i>

16,12

22..'4'J

1T7 .(7

9.6al

sa

'S .011 :*2.45 14.0(1

20.5rj

m.o*o

23.;in

71> ,J>14 17.N5 ,;v.00

VSl

It, 13

m/M

10,45

17.40 17,Ki

.OH LS.fl 1J*,45,

.oil*

m .012

2Lf,B

aa

.011 14.33, IH.iJM

M .Oh 111.75:

.osy

17,00

IHjWii

181 .045

'21.96

.on 1 nM\ lo.m

.014 16. Jm 21 .S2

v& .oia

lO.Ofl

182 .04t'*

2rt.5l

.oil , 17. 2&! Li>.:x

K.H m4 14.70 2U31

1H ,046

ail.M

0.7J

!I9

,011 1 16,00. 27J.Nj

.oh; i7.wr, n.7j

.aJU ; lll.t4

14,00

.011 21.10| 14.8'

35 ,,0I5i VZM K..2

J ijy* Am : 17.15

17.fi6.

Ik.5,648

.Otl 1T,6| J2.74i

86 .015 32. 16.4?!

IKfi' .Ot50

3fi.l8

>oii 11. aw i.7j'

87 ,01B I5,S2 19, eo

140 .021

14,27,, lK7',a-W

.011 14.3d n.v

m Mb ni.9u 1*0. rio

Hi; Arjl

.011 14.10' 20. 9i;

,.016 ai.oEt i:t.;*(v

Hil ,01 I 17.7:>

28.091 18a .062

.011 1 l.Hi ]$.71

90 , .Ol.-i lfi.i5 24 ,iW

14.' Am 22.15

14,18 Iso

s.:r7

ill 1.015 2AA ll,*l

Hll -O-il , 13,77

is.tio

..ih.iH...'

t(2 ,.015 v'ts.m, 13/jr?

,Qii 1 20.04

l&.T-S

.WHliJ

M ,.015 2;*. 57 15,f)0 ,.015 21.6-Sl l&.OO

—

t

96 !.0t6' 1ft. GO 35,73 9(7 I.OUi 2.1. a-J! H,7i 0H ,015| n.47 26. oy

"Z'Z

.010 1 17.04 20, B3

' 1 1

Ay.,. 014 Ift.ea 1.09

Av.

.018 j Ifl.iS

17.8Ti

2l.n

ta.TS

THE OPEN-n EARTH PROCESS.

Table 23. — Summary op Tables 19 to 22.

W

O

A. B.

A. B. D.

It

Is

Slao before TAPPiifo— Per cent

5|l

SiOs.

FeO.

P.0,.

MnO.

CaO.

MgO.

und'd.

15.65' 21.01

The followiog is a list (table 24) of the individual heats of each group which show respectively the highest SiO, and highest FeO in the slags :

Table 24. — Basic Slaos Containing a Maximum of SiO, and

FeO.

Heat in table 20 showing mazimam SiO...

Heat in table 20 showing maximum FeO

Heat in table 21 showing maximum SiO, Heat in table 21 showing maximum FeO...,

Heat in table 22 showing maximum 8iO,

Heat in table 22 showing maximum FeO

8iO,. FeO. CaO. FeO. CaO. 8iO,. FeO. SiO,. FeO. SiO,. FeO. SiO,. FeO.

Omovr

Osoor OKOvr Qkovv B. a I D.

26.0121.63 24.60l29.15 12.64| 8.27 14.7415.16 31.7223.90

Low SiO,, it will be observed, accompanies high FeO and vice versa. The same phenomenon appears in the preceding summary, table 23. In order to ascertain its scope, the results have been ana- lyzed as follows :

The Open-Heabth Process.

Table 25. — Classification op Basic Slags to Show the Relation Between SiO, and FeO.

a

Limits of

h

ConstitaentB of the Slag ; per cent.

s

in Slag, percent

SiOa.

FeO.

P.0,.

MnO.

CaO.

MgO.

8iO,+ PeO.

below 10

above 10

8 to 12

13 to 14

t(

15 to 16

i(

K

18 to 19

U

20 to 22

23 to 27

10 to 13

*014

it

fi

in.

tt

(1

11 a5

25 to 29

These divisions are purely arbitrary, the only object being to obtain as many of them as possible, and yet retain a sufficient number of separate heats in each division to give a reliable average. It will be seen that with only one or two breaks in the line of averages, a rela- tion is clearly established between the content of SiO, and of FeO in the slag. This relation is shown in the column headed SiOj + FeO. With a given phosphorus-content in the charge, the percentage of SiOj and FeO combined adjusts itself to a constant quantity. This does not mean that the original phosphorus-content is the only con- trolling influence. Without doubt the percentage of phosphoric acid in the slag is the determining factor. This percentage is not given in the foregoing tables but it will be noted that the tables 20, 21, and 22 deal with charges of different initial phosphorus; and it may be taken for granted that under the same working-conditions, higher phosphorus in the charge means higher phosphoric acid in the slag. Where phosphorus is high, this combined percentage of SiO, and FeO runs about 27.5 per cent. ; with medium phosphorus

The Open-Hearth Pboce88, 445

it is abont 35 percent., and with low phosphorus about 36 to 37 per cent. Variations in the details of practice will alter this figure, but under the same general conditions the attainment of a constant content of FeO + SiO, in the slag seems assured.

It has already been argued (Sec. 44) that very little FeO can be permanently introduced into the slag while the bath is high in car- bon unless certain chemical requirements demand it. In the absence of such demands, the only way in which an addition of FeO could at any time reduce the percentage of SiOj, whether the carbon l)e . high or low, would be by dilution. This would necessarily also de- crease the proportion of CaO, MgO and of all other constituents. In other words, a relative decrease in SiO, from 20 per cent to 10 per cent., by the process of dilution, would require an addition of FeO equal to the entire weight of the slag. Since no such addition actually takes place, it follows, when all other things are equal, that if high FeO accompanies low SiO,, it is because certain chemical forces call for it; and when it is shown that the sum of FeO + SiO, is a definite constant, one of the components must evidently fulfil some function of the other. As FeO is basic and SiO, acid, the function cannot possibly be directly related to the basicity of the slag. The author believes that the phenomenon is to be attributed to the universal law by which forces act along lines of least resistance, and in obedience to which a tendency is manifested in this case to attain a certain dree of fluidity. Both SiO, and FeO promote this con- dition. If lime be added in sufficient quantity to dilute the silica of the charge to a low percentage, the slag will hold enough FeO to make it fluid. If, on the contrary, the proportion of lime be small, the content of SiO, will be high, and no increment of FeO will be necessary (see Sec. 44).

Thus far the investigations show that low SiO, accompanies good dephosphorization, and that high FeO accompanies low SiO, — hence, also, that high FeO accompanies good dephosphorization. It re- mains to be discovered whether an increase in FeO has any connec- tion with the amount of phosphorus removed. In order to prose- cute this inquiry, the divisions of tables 21 and 22, given in table 25, have been further analyzed by subdividing each into two sections, one representing a greater phosphorus-elimination and the other a less. Thus, in each division, two groups, having the same silica-content, will be compared with regard to their different actions upon phosphorus.

The Open-Heabth Process.

Table 26. — Classification of Basic Slags Acoobding to Phosphobus Elimination.

Subdivision A.

Subdivision B.

gj

Greater Elimination.

Le8 Elimination.

Q

Phoe.

in Metal, perct.

Phoa.

in Metal, perct.

Slag.

SiOs*

FeO.

Sio,.

FeO.

SiOTT FeO.

"

u

(i

Vii.*

Ssm

It

.Oil

It

18,30

(1

or the nineteen comparisons in table 26, seventeen show a higher content of FeO in the slags which accompany the lower phosphorus in the ingot. The average excess of FeO in the seventeen cases is 1.88 per cent. The average difference in phosphorus is .012 per cent. It should be remembered that the excess of FeO does not denote increased basicity, for the whole investigation is made upon slags of the same silica- content, and of the two sets of heats in the table, each has an approximately constant initial percentage of phos- phorus. It therefore seems certain that under otherwise similar con- ditions, lower SiOj means better dephosphorization and that with a given percentage of SiO,, FeO is an important agent. Whether the iron oxide is valuable for its oxidizing power, or for its enhance- ment of the fluidity, remains undecided.

Sec. 46. — Tke Elimination of Sulphur.

The basic process demands low silicon in the pig-iron, and with this low silicon comes oftentimes high sulphur. But this does not always follow, and it is easy to show that there is no necessary con-

Cannot be divided with similar silica-content in the sabdivisions.

THE OPEN-HEARTH FJEtOCESS. 447

nection between the two, since a calcareons cinder in the blast-fur- nace tends to keep both silica and sulphur away from the iron, while a deficit of lime leaves the sih'ca and sulphur unbound, and open to the reducing action of the gases. In accordance with these simple preferential relations, iron may be made under proper condi- tions with low silicon and low sulphur, while under bad conditions it may have a high percentage of each. The following are samples of large lots of pig-iron with which the writer has had the good and the bad fortune to deal.

81. P. 8.

0.46 2.3 trace.

2.75 0.5 0.54

But whatever special samples may indicate, it is true that the low-silicon pig-iron of the market contains more sulphur than high- silicon pig, for the reason that the reactions in the blast-furnace de- pend fully as much upon temperature as upon the chemical equa- tion. Hence it becomes desirable that methods should be introduced to eliminate sulphur upon the basic hearth. The results of usual practice as given by various investigators are contradictory. Wed- ding'*' states that there is little or no elimination. Saniterf testifies to the same experience and recommends the use of calcium chloride, which his experiments indicate as an efficient agent. Harbord had previously given opposite evidence and recently he has been cor- roborated by Stead§ who cites Hardisty's results and summarizes the sulphur equation of the basic open-hearth process as follows:

(1) " The manganese added in the metal, in passing out, may carry some sulphur with it.

(2) " The manganese reduced from the slag during dephosphoriza- tion effects an elimination of sulphur.

(3) '' The calcareous slag in contact with the upper surface of the bath containing carbon, may absorb sulphur.

(4) '' Some of the manganese added in the ferro does undoubtedly leave the bath again, carrying with it a small quantity of sulphur."

These carefully guarded statements may be accepted without qualification, as may also the more positive one which follows :

"Progress of German Practice in Metallurgy."— IVarw., xix., 1891, p. 380.

t " A New Process for the Purification of Iron and Steel from Sulphur." — Joum. I, and S, Irut,, 1892, No. 2, p. 216 et teq.

X *'On the Removal of Metalloids in the Basic Siemens Furnace."— /&uf., 1886, No. 2. p. 700 et wq,

2 "On the Elimination of Sulphur from Iron."— /Wrf., 1892, No. 2, p. 266.

The Open-Hearth Process.

'The changes during elimination of the sulphur in the basic pro- cess, however, have not jet been thoroughly explained, and the sub- ject is worthy of more attention/*

On certain of the points involved the following results may shed some light. They are records of work done by the Pennsylvania Steel Company.

The factors investigated are:

The elimination of sulphur by the action of metallic manganese.

The elimination of sulphur under ordinary conditions when the initial content is below 0.10 per cent.

The elimination of sulphur under special conditions when the initial content is above 0.10 per cent.

(a) By the Action of Metallic Manganese. — The following are records (table 27) of a few charges in which sulphur was deter- mined just before adding the recarburizer and also in the steel ingot. These charges were made from desiliconized and partially decar- burized Bessemer metal which was poure<l into the furnace in a melted state. The analysis of the initial metal is also given, for if any desulphnrization occurred during the decarburizing period, we should naturally expect the action to continue during the solution of the recarburizer and should aseril)e to that action a part of the change in the sulphur-content. The carbon in the original charge is added so as to indicate the extent of this previous period.

Table 27. — Effect op Recarburization on Sulphur.

Bath.

No. of Charge.

Initial Metal.

Before Recarb.

Slag

Steel i After Tapping. j Ingot. '

s.

810,.

reO.

CaO.

MgO. 1

2,15

undet.

The charges are arranged according to the degree of desulphnriza- tion during recarburization. Charge 203 is placed above 202 be- cause unlike its successor it was not desulphurized during decarburi-

The Open-Hearth Process.

zation and hence the diminution of S, due to the ferro, seems more genuine. Charge 202 in turn precedes No. 198 because in it the period of decarburization was obviously longer, and, with the same elimination of sulphur, a less energetic action is indicated. Charge 199 is placed last, for although the initial sulphur was not deter- mined, it is quite certain to have been much above .02 per cent, and probably was .05 to .07 per cent.

In heat number 104, sulphur is decreased by .02 per cent. ; in four heats the decrease during recarburization is .01 ; while in two heats, Nos. 199 and 201, there is an increase during recarburization of .01. In both cases of increase, however, there was a previous de- sulphurization during decarburization to an extent equalled by only one member of the first group. It might be argued that the higher percentage of iron oxide in the slags of the second group, Nos. 199 and 201, gave rise to unstable conditions. If this were true and resulphurization were the result, we should expect rephosphoriza- tion also to occur. In the following table the relations of sulphur and phosphorus in the forgoing charges are shown for the particu- lar period under investigation. The oxidation of manganese during recarburization is also given since that might possibly have a bear- ing on the elimination.

Table 28. — Relation of Sulphur and Phosphorus TO Recarburization.

Charge Number.

Mn

Burned. Perct.

Sulphub.

Phosphorus.

Before Recarb.

After Recarb.

Incre8e+ Decrease

Before Recarb.

Atter Recarb.

Iiicrea8e+ Decreaw—

—.02 —.01 —.01 —.01

-f.Ol -f.Ol

—.001 -f.002 +.009

+.004 +.009

These figures show no fixed relation between the desulphurizatioa and either the rephosphorization or the combustion of manganese.

There were also investigated sixteen heats made from high-sul- phur stock charged cold. The results are slightly complicated hy the circumstance that desulphurization was going on at the time of adding the ferro. The rate of elimination at that period, however

The Open-H£Abth Process.

was very slow, and in the few minutes during which the recarburizer was acting, the reduction in sulphur-content from that cause would be trifling.

Table 29. — Removal op Sulphur from High-Sulphur

Charges.

Co

sal

—.048 —.044 —.038 —.030 —.028 —.021 —.020 —.016 —.011 —.008 —.004 —.000 +.011 +.017 +.017 +.055

Slao

After Tapping.

FeO.

HnO.

CaO.

ondet.

undet.

u

undet.

u

It

i(

undet.

The records of rephosphorization and of manganese burned dur- ing recarburization are here given for comparison :

Table 30. — Relation of Sulphur and Phosphorus in High-Sulphur Charges to Recarburization.

Charge Number.

Mn

Buraed.

Per ct.

Sulphur.

Pho8Phobu8.

Before After Recarb. Recarb.

Increased Decrease—

Before Recarb.

After Recarb.

Increase + Decrease—

1M4

.29*

.31*

V34' .26'

—.048 —.044 —.038 —.030 —.028 —.021 —.020 —.016 —.011 —.008 —.004 —.000 -f.OU -f.017 -f.017 +.055

.Oil

—.010 -f.OOl 1 -f.006 ' -f.003 1 -f.006 -f.004 -f.008 —.007 -f.OOl -f.007 —.001 —.006 -f.003 —.005 —.016 -f.007

The Opex-Hearth Process.

These charges are arranged as in table 27, in the order of dimin- ishing desulphurization. No law of variation can be interpreted either from the concurrent rephosphorization, or from the oxidation of manganese during recarburization or the composition of the slag with respect to sulphur, silica, iron oxide, manganese oxide or lime. The records of charge No. 1(>08 may be wrong, for possibly the test which showed .04 per cent, of sulphur in the metal before recarburiz- ing did not fairly represent the whole heat, but only a particular locus in the bath.

With two charges the experiment of adding 20-per cent, spiegel was tried, and of allowing the bath to stand for a time so as to afford opportunity for the partial combustion of the manganese. The results were as follows :

Table 31. — Effect of Spiegel on Sulphur.

Test.

Crargi 966.

Chaeqe 1686.

Mn.

S.

L Before adding 20-per cent spiegel 2. Before addinir recarburiser.

3. After addinar recarburizer

Amoont of Mn in the spiegel

0.18 p.ct. ofch'ge.

0.10 "

n. o.t. nf rh*aA.

Increase of Mn in the bath (2-1)

It will be seen that the results are irregular. Heat 966 with a combustion of 0.10 per cent, of Mn shows a desulphurization of only .001, while heat 1686 with only 0.01 per cent, of Mn burned shows a reduction of .01 in sulphur-content. This preliminary washing had no apparent influence on the action of the recarburizer. In both heats a reduction of sulphur occurred during recarburization, but it seems to have had no relation to the work done by the first addition of manganese.

Concliunons. — Though the forgoing .data do not offer very solid ground for generalizations, they seem to warrant the following statements:

(1) There is no certainty that desulphurization will take place during recarburization, but there is a probability of a reduction of about .01 per cent, in the sulphur-content.

The Open-Heabth Pb0Ces8.

(2) The presence of a sulphurous slag due to a high initial content of sulphur does not necessarily interfere seriously with the action, but under certain conditions renders possible a considerable resul- phurization.

(3) The conditions favoring desulphurization during recarburiza- tion do not seem to depend ui>on either the amount of manganese burned, or any particular composition of the slag but on some factor not usually recorded.

(6). Under Ordinary Conditions When the Initial Sulphur-Content is Below 0.10 per cent. — In the investigation of sulpliur, as of other elements, an obstacle in the way of conclusive results is the difficulty of knowing the exact composition of the metal charged. While it may be determined by careful sampling, yet strict accuracy is often a matter of doubt. To avoid this source of error, a series of tests were made in which the initial charge consisted of partially blown Bessemer metal, of which it was practicable to get a true sample. The records of those tests have been given in tables 27 and 28, but they are here repeated in order to classify them according to the desulphurization effected.

Table 32. — Desulphurization of Low-Sulphub Stock.

FiBT Sbribs. Six Head. AnAlysM.

Charge

Initial Metal.

Steel IngoL

After Tapplag.

Number.

P.

S.

P.

810).

FeO.

MnO

CaO.

0J8

.06 .020

NoTE.~Noiirmaganiferoi]s ore was used for decarburisiiig.

The following charges were made from cold stock, and table 33 gives the composition of the metal after fusion. This leaves untold the story of the sulphur during the melting, but it is safe to assume that if sulphur is removed during decarburization it will likewise be

ttacked during the earlier period — an assertion which the records iu

[Vision (c) of this section will substantiate.

The Open-Hearth Process.

Table 33. — Dbbulphurization op Low-Sulphur Stock.

Second Sibixs.

Fifteen Heats. Analyses.

Charge

Metal alter Fusion,

Steel Ingot.

Slagr After Tapping.

Number.

P.

SiO,.

FeO.

NoTi.~Non*manganiferou8 ore "was used for deearburizing.

With such discordant results it would be quite possible to make different combinations of the figures tell different stories. The most rational mode of comparison is by dividing them into groups accord- ing to the sulphur-elimination, by which we obtain :

Sulphur- Number of heats Average of slag,

elimination. in group. SiOi. FeO.

.05, 1 13.43 30.45

.04, 3 18.88 19.71

.03, 2 16.63 19.72

.02, 3 17.49 22.48

.01, 3 19.54 18.67

.00, 2 17.33 24.87

Gain of sulphur, 1 15.66 34.36

They may also be combined as follows :

Group I. — First six heats showing sulphur-eliminations of .05, .04y and .03 per cent, inclusive.

Group II. — Second six heats showing sulphur-eliminations of .02 and .01 per cent, inclusive.

Group III. — Remaining three heats showing no sulphur elimina- tion.

Number of Average of slag,

heats in group. SiOs. FeO.

Group I., 6 17.22 21.51

Group IL, 6 18.51 20.57

Group III., 3 16.77 28.03

454 The Open-Heabth Process.

It is evident that neither the silica nor the iron oxide is the de- termining factor. The oxide of manganese in the slag is not given, but it was probably quite constant in all the heats, and not over 3 or 4 per cent.

Other determinations were made without the corresponding slag analyses, resulting as follows :

Third series ; seventeen beats. Heats.

Elimination of .03 per cent, sulphur 3

.01 " " 3

Gain of sulphur, 2

Summing up the results of this third series and those of tables 32 and 33, we have :

Heats.

Elimination of .05 per cent, sulphur, 1

.04 " " 4

Gain of sulphur, 3

It should be remembered that in the above records the first sample- test is taken after fusion, and even when no elimination is shown it is quite possible that that there has been a previous desulphuriza- tion. Thus, in the third series, which is not given in detail, one set of heats shows no elimination, yet the steel carries only .02 per cent, of sulphur, while another set gives a gain of .01 sulphur with .04 in the ingot. In both these cases it is almost certain that the metal was purified during melting.

Conclmion. — Under ordinary conditions, with an initial sulphur- content below O.IO per cent., sulphur is usually reduced by about .02 per cent. ; but oftentimes no elimination occurs.

(c) Under Special Conditions when the Initial Sulphur-Content is above 0.10 per cent. — The following are the records of the same sixteen heats given in table 29 which were made from high-sulphur stock with manganiferous ore. When the chaise was melted, and a good slag had been formed, part of the cinder was decanted by tilting the furnace, after which lime was added in such quantity as was deemed necessary to give a proper slag.

the open-hearth process. 455

Table 34. — Desulphurjzation op High-Sulphur Stock.

Sulphur.

Charge Number.

In

In Bath after

In Steel

In

In Bath after

In Steel

Chnrge.

MelUng.

Ingot.

Charge.

Melting.

Ingot.

.086 ;

.089 ;

.062 .

Further data concerning these heats are given in the following tables, 35 to 40, arranged to show the relation of the sulphur-elimi- nation to the composition of the slag and to the basic additions.

Table 35. — Arranged According to the Elimination op Sulphur during Melting.

"

it

Slag, after melting.

FeO.

CaO.

MnO.

n.

undet'd.

K

a

U

undet'd.

und.

undet'd.

und.

u

undet'd.

u

und.

undet'd.

undefd.

u

The slags and metals of table 36 are the products as they leave the steel-ladle. In some respects thej do not represent all the actual working conditions ; the sulphur in the metal, for example, is in- fluenced by the action of the recarburizer and by exposure to the air during pouring, wjiile the composition of the slag will be affected by the silica derived from the scorification of the ladle-linings. The

The Open-Heahth Process.

Table 36. — Arranged Acxx)rding to the CJontent of Sul- phur IN THE Steel.

2tS

S5

Slag, after tapping.

sa

Ib

SiO,.

FeO.

CaO.

MnO.

undet'd.

u

undet'd.

und.

undefd

iindet'd

undetM.

.2

undet'd.

u

und.

ti

undet'd.

undetM.

true history of the sulphur will be found in the slag and metal before adding ferro. Though samples were not taken for all the charges at that particular stage in the process, yet in tables 37 and 38 such

Table 37. — Arranged According to the Absolute Elim- ination OF Sulphur Between the Time op Pouring

OFF THE Slag and just before RECARBURIZrNG.

Sulphnr in metal

after melting.

Sulphur in metal

before recarbur-

izing.

Eliminat'n of sulphur

between melting &

tapping.

Sulphur in slag after melting.

Slag, Just before recarburizing.

S.

SiO,.

FeO.

CaO.

MnO.

undet.

M

undet.

nndet.

1

u

und'd.

u

l(

analyses as were made are given, and the heats are variously ar- ranged so as to reveal, if possible, any relation between desulphuri- zation and the composition of the slag. With the same end in

The Open-Heabth Pr0Cbb8.

Table 38. — Arranged According to the Percentage of eumination of sulphur between the time op pouring OFF Slag and just before Recarburizing.

el

Sulphur iu metal

after melting.

d

hi

in

p.

mi

£'3

Slag, juRt before recarburiiing.

FeO.

CaO.

MnO.

undet'd.

.'3

undet'd.

undel'd

und.

u

It

vieWy the slags are arranged in tables 39 and 40 according to their sulphur-content.

Table 39. — Slags after Melting, Arranged According to THEIR Sulphur-Content.

Slag.

after MelUng.

Charge number.

810,.

FcO.

CaO.

MnO.

undet'd.

undetM.

u

K

undet'd.

uudetM.

u

1630

undet'd.

tt

In table 40, for the first time, there appears to be a connection between the factors ; low silica is associated with high sulphur in the slag. The content of the metalloid in the slag, however, seems to bear no relation to its content either in the metal before recar- burization or in the steel.

The Opek-Hearth Process.

Table 40. — Slags before Adding the Rbcarbitrizbr, Ar- ranged ACXJORDING TO THEIR SuLPHUR-CoNTENT.

u 9

Sulphur in inetAl before re- carburiz'g.

y

Slag, before adding the Recarburizer.

SiOfr

FeO.

CaO.

MnO.

undetV.

u

.Uk)

undetV.

undet'r.

it

undet'r.

The only inferences which the foregoing work seems to admit are the following :

Qyiiclusiona. — (1) Under proper conditions an initial content of from 0.20 to 0.30 per cent, sulphur is reduced considerably during the melting-period, and is further reduced during the oreingso that the steel will contain less than 0.10 per cent.

(2) The law of sulphur-reduction is not to be found in the equa- tion of the slag as ordinarily presented but in some other metallur- gical reaction.

(3) A decrease in the percentage of silica in a slag increases its power of absorbing sulphur where other factors remain unaltered.

(d) By Manganiferous Ore. — The effect which is produced upon the sulphur-content by an addition of manganese in the form of Spiegel or ferro has already been considered (see table 27 et eg.). Manganese, however, may also be added in the form of ore, from which it can be reduced by the silicon, carbon and phosphorus of the bath. In the al)sence of any direct comparison between the value of a given quantity of manganese reduced from the ore and a similar quautity added in the form of an alloy with pig.or scrap, it may not be unnatural to suppose that the greater efficacy possibly lies with the ore, since the reaction which attends its use sets the element free in a nascent state.

In open-hearth work it is not customary to sample each charge accurately before it goes into the furnace. In large steel-works wh*ere domestic scrap is used, the percentage of phosphorus is usu- ally known, since it is determined exactly by the content in the ore and fuel. But the percentage of sulphur changes with every cast from

The Open-Heabth Process.

the blast-furnaces. It is only by careful and expensive sampling that accurate determinations of a mixed charge of pig and scrap can be made. In the case of the heats investigated in (c), the initial sul- phur-content is not given with strict accuracy, but where we are con- sidering a reduction in content from .20 to .08 per cent, an error of .02 per cent, in the initial sulphur is not of vital importance. When, however, we start with sulphur at about .06, such an error is inad- missible. Scientifically speaking, the mere fact of having low sul- phur in a steel does not indicate an elimination unless higher sulphur has been shown in the particular charge from which the steel was made. Practically, however, if one mode of working yields uni- formly lower sulphur in the product than another, and if the metal- charge has been the regular run of domestic iron and scrap, in which the sulphur is known within narrow limits, and if the results cover a sufficient period to obliterate the effects of accidents and coinci- dences, it is not only reasonable, but logical, to assume that the difference in practice is the cause of the difference in results. Rea- soning thus from such premises, the efficacy of manganiferous ore in the elimination of sulphur has been deduced, and the conclusion is corroborated by the experience of so many metallurgists that it may be accepted with confidence. In the following tables the total amounts of ore used on the high-sulphur heats are given, together with the amounts of manganese expressed in percentages of the metal charged.

Tables 41 and 42.-

-Desulfurization during Meiting and Oreing.

Table 41.

Table 42.

Reduction

in sulphur

during

melting.

Ore added with the metal. Pounds.

Mang'nese in ore, ex- pressed in the charge '

Reduction

in sulphur

during

oreing.

Ore added during oreiiifif. Pounds. Mang'nese in ore ex- pressed in per cent of the charge.

InOO

+.021

1:

igitized by

G

ioogle

460 The Open-Hearth Pbocebs.

It will be noticed that the elimination of sulphur is least in cbai 1662, table 41, where no manganiferous ore was used, and greatest in charge 1546, table 42, where the highest percentage of manganese was added. There is no rular law, but the charges in the first half of the list in table 42 show an average addition of 0.52 per cent, of Mn, and those in the last half an addition of 0.31 per cent Doubtless the action occurs through the feduction of metallic manganese by the silicon and carbon, and the formation and separa- tion of sulphide of manganese."' It is probable that on exposure to the flame and to the oxidizing action of the slag, this sulphide is decomposed, the sulphur burning and passing away in the waste- gases. Some of the sulphur may combine with the lime either directly, thus :

S + CaO + C CaS + CX).

Or through the mediation of the manganese reaction :

MuS + CaO CaS -h MnO.

The manganese thus carried into the slag may possibly be reduced like ore by the carbon of the bath, and again unite with sulphur and repeat the cycle. The supposition that this occurs to any important extent is, however, rather gratuitous.

The oxidation of a part of the sulphur is indicated, in the above experiments, by the fact that the cinder does not contain as much sulphur as existed in the stock. The decanted slags and final cin- ders of three heats were weighed and analyzed. Though the results for the two periods of each heat cannot separately be computed be- cause the amount of slag left in the furnace after decantation is an unknown quantity, yet the total sulphur in the two slags may be found and compared with the amount eliminated from the metal. The following is the calculation :

Metal charged :

Pounds Per cent Pounds

Heat. Charged. Sulphur. Sulphur.

1606 35,000 0.28 98.0

1608 35,000 0.28 98.0

1611 35,000 0.20 70.0

Total sulphur, three heats, 266.0

See Massenez ''On the Elimination of Sulphur from Pig- Iron/' Journal of \hn L and S. InsL, 1891, No. 2, p. 76.

The Open-Hearth Pb0Ce8S.

Steel obtained:

Ponnda

Per cent.

Pounds

Heat.

Steel.

Sulphur.

32,000

Total sulphur, three heats, 79.7 Summary for metal:

Pounds.

Three heats, sulphur charged, 266.

" sulphur in steel, 79.7

" sulphur

eliminated

f

Sulphur in slags:

Pounds

Per cent.

Pounds

First slag.

Heat

Slag.

Sulphur.

(Decanted), .

u

Second slag, .

a

u

Total sulphur in both slags, three heats.

Summary for slags:

Pounds.

Three heats, total sulphur eliminated, 186.3

sulphur in slags, 118.4

sulphur unaccounted for, 67.9

36.4 per cent.

These three heats show much higher sulphur in the final slags than most of the other charges, and fully as much as the others in the slags sampled after melting (compare table 35). They are therefore as unfavorable specimens as could be selected to show a volatilization of sulphur. Other experiments, with less complete but scarcely less conclusive data, might be cited to confirm the evidence.

Obviously, the conditions most favorable to volatilization may be intermittent or continuous; if at any time transient causes should favor the conversion of the sulphur into gas, it would matter little what the composition of a later cinder might be. Thus, for exam- ple, the arrangement of the stock on the hearth and sundry minor details of operation might dispose of a considerable proportion of

462 THE OPEN-HEARTH PBOCfiBS.

the sulphur without leaving any record iu the chemical equation of its passage or even of its presence.

(e) Condtuiions. — Although we have not been able to obtain the complete equation of sulphur, jet the partial records afford a basis for the following propositions :

(1) Under certain conditions, not perfectly understood, the content of sulphur is not reduced, although the environment may be suffi- ciently basic for the removal of phosphorus (see division 6).

(2) Under certain other nearly similar conditions, sulphur may be eliminated in the presence of a simple calcareous slag (see division 6).

(3) When the content of sulphur is below 0.10 per cent, its rate of removal is very slow (see division 6).

(4) The addition of metallic manganese tends to eliminate sulphur, but the amount removed is small and variable (see divi- sion a).

(5) The addition of manganiferous ore tends to eliminate sulphur (see division d).

(6) Under certain conditions regularly obtainable in practice, high sulphur (0.25 to 0.30 per cent.) may be reduced below 0.09 percent without the intervention of any agents except lime and manganiferous ore (see divisions c and d),

(7) The quantitative results indicate that in basic converter- and open-hearth practice, phosphorus can be vaporized under certain con- ditions, and in both of them sulphur is partially volatilized after its passage into slag (see division d and Sec. 43).

Chapter VIII. — Preferential Relations.

In the foregoing pages it has more than once been shown that the relative intensity of the several metallurgical factors has been a most potent determinant of the result. Some of the principal relations which have thus far entered into the problems of the acid and basic practice are the following.

Sec. 47. — Silicon and its Oxide,

In acid practice, oxygen unites at low temperatures with silicon in preference to either manganese or carbon, but at high tempera- tures the relations are reversed ; consequently, at high temperatures, oxygen is transferred from silica to carbon with the production of silicon (see Sec. 33). With a slag rich in bases there is need for all the silica, and a stronger power than carbon alone is required for its

THE OPEN-nEARTH PROCESS. 463

reduction. Under these conditions carbon will fail to complete the work, and even the relative affinity for oxygen at a given tempera- ture is changed. Thus in banic practice the demands of the slag prevail over minor preferences, and silicon is seldom found in the metal in large amounts. In an exceptional case coming under the notice of the writer, the metal contained .09 per cent, silicon and 1.29 per cent, carbon after melting, but the slag was unusually sili- ceous, analyzing :

SiO, 30.63; FeO 13.41; CaO 39.17; MgO 7.15.

When the carbon had been reduced to 0.41, the silicon was .01, and the slag ran :

Si02== 23.00; FeO= 15.30; CaO 43.35; MgO 7.22.

Sec. 48. — Manganese and iia Oxides.

On the acid hearth there is need for all the available bases, and hence the oxidation of manganese is completed early in the opera- tion, its affinity for oxygen under these conditions being interme- diate between that of silicon and carbon. In basic work there is less call for the product of its combustion, and the bath, therefore, usually retains manganese in an appreciable percentage to the end. The amount so held varies widely. If the bath is very high in carbon when melted, so that there is a chance for much oxidation during oreing, the content of manganese at the end of the operation will usu- ally be small. But if the proportions of pig and scrap are so regulated that the charge, when melted, contains only about 0.50 to 0.75 per cent, carbon, and if the content of manganese in the original charge was about 0.60 per cent., then 0.20 to 0.30 per cent, manganese may be left in the dearburized metal. With a rular practice, a fairly constant percentage may be counted as being retained, in making the calculation for the recarburizer.

When manganiferous ore is added to a charge, the manganese may be reduced even in the presence of active oxidizing forces, which are manifested by the concurrent combustion of carbon. The following series of tests upon such a chaise demonstrate this point.

As manganese and iron are closely allied, it is necessary to keep in mind that the chemical relations of the manganese may be some- what modified by a physico-chemical tendency to form an alloy, and, furthermore, that in case of both silica and manganese oxide, the theory advanced in Sec. 45, Part Il.y concerning the natural effirt

Thb Open-Hearth Process.

Table 43. — Reduction op Makoanbbe.

Metal.

8Laq.

Mn.

SiO,.

FeO.

Ca'K

MnO.

l.?0

to realize a maximum fluidity may affect their respective preferential relations. In any event, the only oxide which can permanently ex- ist in contact with the metalloids of the bath is the simple MnO. The other combinations, while perhaps transiently present, will be robbed of their surplus oxygen by silicon and carbon.

Sec. 49. — Carbon and its Oxides.

The carbon of the bath is removed in the form of carbonic oxide. It may happen that an atom of carbon in contact with a lump of ore is converted into CO,, but the remaining carbon of the metal will immediately reduce it to CO. As above observed, the forma- tion of this oxide at a low temperature is dependent upon the prior claims of silicon and manganese for oxygen ; but at high temperatures the carbon assumes control, stops the combustion of silicon and even decomposes its oxide; it adjusts the proportion of iron in the slag in accordance with the basicity, and under certain circumstances will reduce metallic manganese from the cinder.

Sec. 50. — Phosphorus and Us Oxide.

In the acid process no appreciable elimination of phosphorus oc curs. All the bases in the slag are needed by the silica. Theoreti- cally it would be possible to liberate part of the irou oxide for com- bination with phosphoric anhydride, leaving the remainder of the bases to form a somewhat lower silicate. But when the percentage of silica is high, its power of holding the iron oxide is so strong that the phosphoric anhydride can obtain no foothold in the slag; if formed, it is immediately reduced by carbon, and the phosphorus is drawn back into the metal. When the slag is low in silica, there is a chance for the weaker phosphorus to enter it. From a chemi- cal standpoint it might combine with the oxide of iron, or of man- ganese, or of calcium. In practice, it enters into combination with

The Open-Hearth Pbocess. 465

the lime, which is superficially explaineI by saying that phosphate of lime is more stable than the phosphates of iron and manganese. The definition of stability, however, is not so easily given. We may suppose that the violent molecular changes going on in the open-hearth bath during the combustion of carbon effect a con- tinual rearrangement of the elements, and that the particular pro portion of the components at any loons of the bath will determine the momentary chemical reactions at that point. Everything will be in a transition-state. Phosphate of iron must be rarded composed of phosphoric anhydride and iron oxide; if the latter be set free for an instant in the molecular transmutations just men- tioned, it is liable to be reduced by carbon, thus leaving the phos- phoric oxide exposed to a reducing action, to which it immediately yields. But in the case of phosphate of lime, which is a combina- tion of phosphoric anhydride with oxide of calcium, the latter will be unaffected by any ordinary reducing action, and therefore the phosphoric oxide will not be left, even for a single moment, without a partner.

Thus the history of phosphorus, which metallurgists of the last generation could not write, appears to be the record of a simple pre- ferential relation. The precise determining conditions are not so* easily equated. The critical percentage of silica, and its absorbent power at that percentage, vary with each particular combination of the remaining elements, with the intensity of the reducing condi- tions, and with the duration of the exposure. In slags taken sood after melting, and before the carbon-reaction has broken up all un- stable compounds, we may find such results as these which are copied from records of slags made in regular work.

FeO.

WOi + PjC

but, for a final slag, the following are fairer examples of the possf- bilities, these slags having likewise been made in rular practice t VOL. xxn.— 30

466 The Open-Heabth Pboce8S.

FeO.

8I0, + P,(V

In basic practice, it does not always happen that a slag is sat- urated with phosphoric anhydride. Lime is frequently added to satisfy the silica and prevent cutting of the hearth ; it is simply a coincidence if exactly the maximum allowable quantity of phospho- rus be present. Each additional per cent, of oxide of phosphorus in the slag decreases its capacity for more, since, in this respect, phos- phoric anhydride acts in the same way as silica. The above column which shows the sum of SiO, and P2O5 indicates that the total acid content of the slag is the measure of its power to absorb phosphorus.

In addition to the varying absorbing-capacity of the slag, the com- position of the metal exerts an important influence on the transfer of phosphorus. When the carbon is high, the reducing action is powerful, and thus there is a tendency to postpone the phos- ])horus oxidation until the end of the operation. It is probable, however, that the affinity of )>hosphorus for iron decreases as the phosphorus-content increases, so that when phosphorus is present in large proportions it may be oxidized, even in the presence of a strong carbon-reaction. With the combination of high carbon and low phosphorus, the percentage of elimination is very uncertain. The work may be hastened by a thick, limey slag, but such a cinder oc- casions trouble by filling the hearth.

As a corollary of the proposition that affinity decreases as the pro- portion of the phosphorus increases, may be mentioned the enhanced attraction of minute quantities to the iron. This phenomenon exists in the case of all the metalloids with which the steel-maker has to deal. Silicon drops with comparative readiness to 0.02 or 0.01 per cent. ; manganese to about the same point ; carbon to 0.05 per cent, or thereabout; but the remnants then hold on with the vigor of despair. This strong persistence of traces, which holds good not only in iron metallurgy, but throughout the domain of chemistry, is merely the last member in the series of preferential relations.

The Open-Hearth Process. 467

Chapter IX. — Recarburization. Sec. 51. — The Time when the Eecarburizer is Added,

The conditions of the metal and slag can readily be determined by taking samples from the furnace with a small test-ladle. Thus the changes in composition can be noted, and the process may be interrupted whenever it has reached the desired point. The samples are usually cast in a test-mould, so as to give an ingot with a cross- section about one inch square. The ingot is chilled in water and then broken, and the carbon is estimated from the appearance of the frac- ture.

The reliability of such a determination depends upon the con- stancy of the conditions of casting and chilling and the ezpertness of the judge. Roughly speaking, the content may be ascertained within 10 per cent, of the true amount, though with long practice on one kind of steel still greater accuracy may be attained. With variations in the content of phosphorus, its great influence on the crystallization must be allowed for. In high-carbon steels silicon is frequently, and manganese occasionally, present. Both exert con- siderable influence upon the structure. When the stock varies, the only safe way is to pour a small quantity of the metal upon an iron plate, cool it as quickly as possible without hardening the piece, and determine carbon by color. This may be done in about ten minutes. Silicon and manganese may also be determined, but this is usually unnecessary. A knowledge of the character of the slag is necessary on account of its important bearing on the waste of manganese. It is well, however, not to speculate upon a certain effect for a certain change in the slag, but rather to alter the cinder to a constant char- acter. Each melter may have his preference for a particular kind of slag, but all should know how to produce a given condition by the use of sand, or ore, or heat.

Sec. 52. — Function of the Recarburizer.

In the manufacture of rail-steel in the converter, the term re- carburizer'' is not inapplicable, but in ordinary open-hearth prac- tice the word is certainly misleading. As previously observed, high- carbon metal is made by interrupting the process when the bath contains the desired carbon percentage. The function of the final

468 The Open-Hearth Pbocess.

addition, therefore, is to furnish manganese, or silicon, or both, while carbon, under ordinary circumstances, is looked upon as a necessary but objectionable companion of these two elements. Remanganizer and resiliconizer would be truer terms, but custom has not coined them. On the other hand it has broadened the meaning of the term 'recarburizer," until it now signifies all those additions which are incorporated into the metal at the end of the operation to make it suitable for the purposes for which the steel is intended.

The exact chemical equation resulting from the addition of the recarburizer cannot be given. The bath contains a certain substance, either gaseous or liquid, which will burn manganese or other easily ozidizable material. A heat which is low in carbon usually lies dead in the furnace, but if a cold steel rod be thrust into the metal, large volumes of gas will be given off. It would seem possible that this gas might be the oxidizing agent, but two facts oppose the hy- pothesis :

MraL — If the bath be stirred with several such rods, the gas may be in great measure removed ; the chilling of the metal in contact with the rods reduces its solvent power and the evolution of gas at that spot furnishes a starting-point for further action in the same way that a glass rod affords a chance of escape for steam in a beaker of superheated water. But it is found that the removal of the gas in this way does not decrease the waste of manganese to any great extent. In some cases it may diminish the intensity of the boiling action, and thus by decreasing the exposure may occasion less oxida- tion, but it can hardly be considered that the absence of the gas itself is the cause of the saving.

Second. — The analysis of the expelled gas does not indicate that it can be the active agent. The experiment of collecting a sample of it is attended with some difficulty, but may be performed by thrust- ing a long piece of 2-inch pipe into the bath and keeping it in mo- tion, the outer end projecting about 10 feet beyond the furnace. Splashes of melted metal and gas will be immediately expelled from the end of the tube. After this first explosion, a piece of glass tubing, provided with rubber connections at each end, should be attached to the outer end of the pipe. The gases will be chilled by their passage through the long pipe (which can be kept cool by water if necessary), and after a sufficient quantity has passed through it and through the glass tube to wash out the previously contained air, the contents of the tube can be imprisoned by means

THE OPEN-HEARTH PROCaESS. 469

of claraps. A sample of gas thus collected had the following com- position :

00], 8.8 per cent, by volume.

O, 1.2 "" "

N, 4.4 by difference.

The presence of CO, may be due to imperfect removal of the air from the pipe. If it be regarded as an original constituent its oxi- dizing power may be considered cancelled by the reducing effect of the hydrogen. Practically, therefore, the gas seems to consist almost entirely of carbonic oxide, which could have no oxidizing effect upon metallic manganese. Hence it becomes reasonable to suppose, that one of the oxides of iron is present in greater or less quantity in the bath, and (hat the more easily oxidizable element exchanges places with the iron in the combination. No arbitrary law of the extent of such transference can be given. The condition of the bath and the nature of the slag will determine what amounts of manganese and silicon will be burned during the recarburization.

Two methods for recarburizing have been investigated — first, in the furnace ; and second, in the ladle.

The first method would seem the surest one to avoid a ])08sible hard spot in the steel ; but, as a number of experiments will pres- ently show, uniformity of product is more readily attained by add- ing the ferro in the ladle. This second method has been objected to on account of a supposed difiSculty in obtaining thorough mixing. In poor practice, trouble will arise from this source, but it can be avoided. One writer has compared the action to the dissolving of lumps of salt in water — a comparison which is manifestly wrong. The history of the recarburizer is not solely one of solution. There is a strong chemical affinity between the manganese and the oxygen which it is designed to seize. The union produces an agitation which in turn increases the vigor of the combustion, thus making the energy of the reaction cumulative. Only when the amount of manganese is so great that the metal is killed and all action stops, will there be any concentration of manganese in one spot, unless the charge be too cold to melt the addition. Furthermore, the lumps of recarburizer do not remain intact like lumps of salt. If pieces of ferroman- ganese be placed inside a furnace-door they fly to pieces from inter- nal strains. Similarly, when thrown into a bath of steel they will

The Open-H£Abth Pb0Ce8S.

Table 44. — Comparison op Losses of Manganese when the Carbon is Constant and the Addition Variable.

Division I.— Carbon below .16.

Mn added.

Per cent of total

charge.

to

Over 1.29

.a

Average Percentages.

Carbon in Steel.

Mn added.

Mn in SteeL

Loss of Manganese.

Per cent of charge.

Per cent, of amount added.

42,3

Division II.— Carbon .16 to .25.

to .80 " 1.19 " 1.49 " 1.59

Division III.— Carbon .26 to .35.

Below 1.10

1.10 to 1.19

1.20 1.29

1.30 1.39

1.40 " 1.49

1.50 1.60

Division IV.— Carbon .36 to .45.

to

Division V.— Carbon .46 to .55.

Below 1.00

1.00 to 1.09

1.10 " 1.19

1.20 " 1.29

1.40 1.49

Over 1.49

The Op£N-Hearth Process.

Table 44. — Continued.

D1VI8IOK VI.— Carbon .66 to .66.

Average Percentages.

Loss of Manganese.

Mn added

o

O

Per cent, of total charge.

Carbon

Mn

Per cent.

Per cent, of amount added.

S

in Steel.

added.

in Steel.

of charge.

Below 1.00

1.00 to 1.09

l.Ol

1.10 " 1.19

1.30 1.39

Over 1.39

Division VIL— Carbon .66 to .76.

Below .90

.90 to .99

1.00 **1.09

1.10 " 1.19

1.20 'M.29

Division VIII.— Carbon .76 to .85.

Below 1.00

1.00 to 1.09

break up and hasten to the utmost the work of solution and reduc- tion. In Sec. 1 results have been given which show the homogeneity of steels made by adding the ferro in the ladle, and other sources of information on this subject are there referred to.

Sec. 53. — BecarburizcUion in the Furnace (25-1071 Add).

Table 44 shows that the loss of manganese by recarburizing in the furnace increases with the amount added. Not only does the actual weight which is burned increase, but the percentage also — ue., if twice as much manganese be added, more than twice as much will be burned. Thus, each increment of manganese in the finished metal calls for a greater sacrifice in the furnace. This fact may be ascribed to the concurrent action of at least two causes :

First. With the greater addition of ferro there is a chance for a more thorough mixing and removal of the oxygen in the bath, which explains an increase in the amount of ferro burned, though it can hardly account for the larger percentage of loss.

The Open-Heabth Pbocess.

Second. With the greater addition, a much longer time is neces- sarily required for melting and mixing. During this period, the recarburizer floats about on the surface of tlie bath, and is exposed to the oxidizing action of the flame and slag, so that the manganese given to the bath by the portions of ferro which melt first is partly burned before the last of the ferro has been dissolved.

Corroborative evidence of the importance of this second influence is obtained from a comparison of the loss when recarburizing with a lower grade of ferro. Thirteen heats, between .35 and .45 carbon, were made with the addition of 1.50 per cent, of Mn in the form of 80-per cent, ferro. Ten other heats were made at about the same time, in the same furnaces, by the same men, and of the same stock, with the addition of the same total Mn in the form of a 45-per cent, ferro.

The results were as follows :

Table 45. — Comparative Losses op Manganese with Low- Grade and High-Grade Ferro.

Ferro used.

No. of Heats.

Manganese.

Loss of Manganese.

Added, per ct. of charge.

In Steel, per cent.

Pr. ct. of charge.

Pr.ctof amoant added

80-per cent Mn. 45 " "

fn considering the increased loss with the lower grade of ferro, it must be remembered that together with the manganese greater quantities of silicon and carbon are introduced. In table 44, as in others, many exceptions occur in the several groups, but these ex- ceptions may often be accounted for by the small number of heats which they represent. The general trend of the resuRs may easily be seen without illustrating them with a graphic diagram.

Table 46 shows that the loss of manganese decreases as the content of carbon in the bath increases. This would naturally be expected, since the amount of oxygen present must be less in the presence of greater reducing agents.

The Open-Hearth Process.

Table 46 — Rbcarburization in the Furnace (25-Ton Acid).

Comparison op Losses of Manganese when the Ferro-

Additions are Constant and the Carbon Variable.

Manganese.

I/Mw of Manganese.

No. of

Carbon.

Heats.

Added.

In Steel,

Perct.

Per cent of

per ct of Charge.

perct.

of Charge.

Amount Added.

Group I.

Mn added. .60 to .69 '

Group II. Mn added, .70 to.79 i

Group III.

Mn added,. 80 to .89

Group IV.

Mnadded,.90to.99

Group V.

♦ .38

Mn added, 1.00 to 1.09...

►

Group VL

Mnadded,1.10tol.l9...

Group VII.

Mn added, 1.20 to 1.29...

Group vm.

Mn added, 1.30 to 1.39..

,86

Group IX.

Mn added, L40 to 1.49...

Group X.

Mn added, 1.50 to L60

.64""'

iedbyJDO

The Open-Heabth Process.

Section 54. — Recarburization in the Lddle {25'Ton Acid Heat).

Table 47. — Comparison op IjOsses of Manganese, when the Carbon is Constant, and the Ferro-Addition Variable.

Division L— Carbon below .16.

p.

o

Mn. added.

Per cent, of total

charge,

.30 to .40 " .60 .80 " 1.00 " 1.10 "

Average Percentages.

Loss of !

No. of heats.

Carbon

Mn

Mn In

Pr. cent

of charge.

in Steel.

added.

Steel.

Per cent, of amount added.

Division IL —Carbon .16 to .25.

.40 to

.60'*

.60"

1.00 "

L30*

Division III.— Carbon .26 to .35.

.80 to .89

.90 " .99

1.00 1.09

1.10 " 1.19

L20 1.29

.12 ,

Ml

Division IV.— Carbon .36 to. 45.

.60 to

.80 "

.90 "

1.00

1.20 "

M9

Lio

Loo

Division V. — Carbon .46 to .55.

60 to 70 80 90 " 00 " 10 "

Loo

Lio

The Open-Hearth Process.

Table 47. — Continued.

Division VI.— Carbon .66 to .65.

Md added.

Per cent, of total

charge.

.60 to .69 .70 .79 .80 " .89 .90 .99 1.00 " 1.09

No. of heats.

Average Percentages.

Carbon In Steel.

Mn added.

Mnin Steel.

Loss of Manganese.

Pr.ct

of charge.

Per cent, of amount added.

Division VIL— Carbon .66 to .75.

.50 to .69

Loo

Lio

Loo

S

Lio

Division VIIL— Carbon .76 to .85.

.70 to .80 "

Lio

Table 47 proves that the loss of Md in the ladle increases with the amount added, thus indicating that in the case of low manganese the washing of the metal is not complete. But no increase in the per- centage of loss is shown, from which it is to be inferred that the existence of such a relation when the recarburizer is added in the furnace must be due to some effect of the flame or slag rather than to the reactions in the metal itself. (See remarks on Table 44 in Sec. 53.)

Table 48, like table 47, shows a decreased loss in the case of the higher carbons. As in the other tables, some exceptions will be ex plained by the small number of heats represented in the group.

The Open-Hearth Process.

Table 48. — Recarurization in the Ladle (25-ton Actd Heat). Comparison of Losses of Manganese when the Additions are Constant and the Carbon Variable.

Manganese.

Loss of Manganese.

No. of Heats.

Carbon.

Added, perct.of

In Steel,

Per ct. of

Per cent, of

charge.

perct.

Charge.

Amount Added.

Group I.

Mn added, .40 to .49 '

GRorp II.

Md added, .50 to.59

Group III.

Mn added, .60 to .69

Group IV.

Mn added, .70 to .79

Group V.

Mn added, .80 to .89

Group VI.

Mn added, .90 to. 99

,71

Group VII.

2L6 .

Mn added, 1.00 to 1.09..

.4H

Group VIII.

Mn added, 1.10 to 1.19.-

Group IX.

Mn added, 1.20

The Open-Heabth Process.

Sec. 55. — Certain Fadon Affecting the Lose of Manganese.

The tables id the preceding section are made up entirely from the records of a 25-ton acid furnace working on soft-coal Siemens pro- ducer-gas, with a charge of about one-quarter pig-iron and three- quarters steel scrap. Further results for a 5-ton acid furnace, run- ning on oil-gasy with the same kind of a charge, are also available. The bath in this case is much shallower, the average depth being between 7 and 8 inches, and a much larger surface is exposed in pro-

Table 49. — Rbcarburization in Ladle. Comparative Records. Acid Hearth.

Division I.— Carbon below .16.

5-TOJi FuBNACE. Shallow Bath.

25-TON FUBNACS. Normal Bath.

Loss of Manganese.

Av. percentages.

Ixwof Manganese.

o

Garb.

MaDganese.

Per ct.

of Charge.

Per ct. of Amount Added.

Garb.

in Steel.

Manganese.

Perct

of Charge.

Per ct of Amount Added.

jg ,Stl.

114 .14

Added.

In Steel.

Added.

In Steel.

Division If.— Carbon .16 to .25.

14; .17 8' .18 13) .19 38 .20

Division III.—

Carbon .26

to .35.

Division IV.—

m .36 to .55.

Division V.— Carbon .66 to .75.

.85 .70

The Opek-Heabth Process.

portion to the weight of metal. The oxidizing conditions are there- fore much greater and presumably the waste of manganese would be increased.

In the foregoing table 49, are given, in parallel columns, the items in the history of the small furnace for which sufficient data are available, and the corresponding items from table 48 for the large furnace.

The same small furnace was next operated with a bath of about 50 per cent, greater depth than before, and under these conditions the loss was as follows :

Table 60. — Recarbukization in Laj>le. Comparative Records. Acid Hearth.

5-Toic Furnace. Deeper Bath.

' 25-TON Furnace. Normal Bath.

Av. percentag<.

Low of Manganese.

o

Garb.

in Steel.

MaDganere.

Perct.

of Charge.

Per ct. of Amount Added.

Carb.

in Steel.

Manganese.

Perct.

Per ct of Amount Added.

Added.

In Steel.

Added.

Steel.

.35 ' .08 .44 ; .13 .44 1 .10

It will be seen that with the shallow bath the loss of manganese in the small furnace far exceeded that which obtained in the large furnace, while with the deep bath in the small furnace the loss was about the same as in the large one. Oil-gas was used in the small furnace for the heats run with the shallow bath, and also for part of the time with the deeper bath. A comparison of the losses with oil- gas and with coal-gas for the group of 180 heats, reported in table 50, results as follows :

Table 61. — Comparative Losses op Manganese with Oil- Gas AND Siemens Gas.

Acid Hearth.

Carbon in Steel.

Per cent of Manganese.

Added. In Steel.

Oil-gap, 68 heats

Coal-gas, 112 heats...

The Open-Hearth Phocess. 479

The exact coincidence is, of course accidental, but the figures in- dicate that with the deeper bath the nature of the result was not affected hy the character of the flame. It does not follow that the same is true of the shallow bath as the superior oxidizing power of the oil may always keep a little extra oxygen in the bath. Certain it is that the above average does not agree with some other results of practice, for in a change which was made for a short time from coal- gas to oil on a deep-bath 25-ton furnace, it was found necessary to increase the addition of manganese.

Sec. 56,— The AddUion of SUiccn.

In the manufacture of ordinary steel there is much diversity of practice regarding the use of silicon. In both low- and high-carbon steels, it is regularly used by some melters to make the metal lie quiet in the moulds. By others such additions are deemed super- fluous, since proper furnace-manipulation may lessen the necessity for their use. Still other operators melt the charge under such con- ditions that the bath contains from 0.10 to 0.20 percent, silicon, and there is no call for a further addition. In the manufacture of steel castings, silicon is quite generally used as part of the recarburizer, and when so added, broadly speaking, none is oxidized. Special circumstances, such as an over-oxidized bath, a slag very rich in iron oxide, or too long an exposure of the metal to the flame after recarburization, may produce an elimination of silicon, but usually the steel contains all the silicon that has been added, whether it was introduced as a rich ferro-silicon or in small quantity as an adjunct of the manganese, and whether added before or after the rest of the recarburizer. That the bath often contains silicon at the end of the operation has already been noted ; it may be unnecessary to add that this content must be estimated and duly allowed for if there are any indications of the presence of an appreciable percentage.

Sec. 67. — Loss of Manganese on the Basic Hearth.

In the foregoing consideration of acid practice, no account has been taken of the manganese which may be in the decarburized metal, for, in the practice pursued, the amount will always be below .03 per cent. Only by carrying a very high percentage of manganese in the original charge, and by melting under such conditions that no boiling or oreing is necessary, can the bath contain any great amount of this element. On the basic hearth, however, the presence

The Op£K-H£Abth Pbocb38.

of a considerable percenti of manganese before recarburisition is not at all anusual (see Sec. 48), the exact amount depending upon the initial content and the conditions under which the bath has been worked. Obviously, the history of the recarburizing reaction cannot be written without a knowledge of the amount of manganese present in each case. As this detailed information is wanting for the charges hereinafter investigated (since, in practice, such determinations are not made when the stock and the mode of working are regular), the following tables give only approximate results. In table 52 a

Table 52. — Recarburizatiox in Ladle (5-Ton Basic Heat).

Shallow Bath.

1 Dekp Bath.

Manganese added.

Per cent, of charge.

o

Average percentages.

1

, 1

Average percentages.

Carbon

InSteeL

Manganese.

Carbon in Steel.

Manganese.

Added. , In Steel.

Added.

In Steel.

.20 to .29 .30 " .39 .40 .49 M " .59 .60 .69 .70 .79 .80 " .89

(52

16

"i'3

!38

division made between heats made with a shallow and those with a deeper bath, in the same manner as described in the last section. Whatever influence the depth of the bath exerts here is obscured

Table 53. — Recarburization in Ladle (18-Ton Basic Heat); AND Comparison with Acid Practice.

18-ton Basic Furnace.

25-ton Acid Furnace.

Manganese

added. Per cent, of

Average percentages.

Manganese.

charge.

Carbon in Steel.

o d

Carbon In Steel.

Added.

In steel.

%

Added.

InSteeL

.40 to .49

.50 " .69

.60 " .69

.70 .79

.80 " .89

The Open-Hearth Process. 481

by variations in the composition of the decarburized metal. The fact, that when little manganese was used, the content in the steel exceeded that of the addition, merely illustrates cases where the de- carbarized bath was known to contain manganese, and on that ac- count less of the recarburizer was used. In practice with an 18-ton furnace, the charge usually contained much lower initial manganese than is shown above, and the decarburized metal was nearly free from it. The results for these conditions are given in table 53, the parallel right-hand side showing the comparable data in acid prac- tice, .the figures being taken from table 47, Sec. 54.

Sec. 58. — Bephosphorization.

The passage of phosphorus from the slag to the steel does not occur in acid open-hearth work, but, on a basic hearth, it sometimes becomes a factor of no small importance. 1 n the basic Bessemer, con- siderable trouble has arisen from this phenomenon, since the addition of melted spiegel to the overblown converter-metal causes violent reactions, accompanied by the production of large volumes of car- bonic oxide, which acts as a reducing agent. In the open hearth, it is customary to use the recarburizer in solid form, and the action. is not violent. Even when melted spiegel is used, there is no strong ebullition, and, under ordinary conditions, the rephosphorization is unimportant. No determination of the exact equation for the re- action can be made, however, on account of the effects concurrently produced by the dissolving of the recarburizer and the play of the flame upon the metal. If, in a given case, dephosphorization had been taking place prior to the addition of the recarburizer, the re* duction in phosphorus-content, due to that normal action, might have been greater than the subsequent increnient caused by the re- action of the recarburizer. The resultant of the two forces may be a considerable diminution in the phosphorus-content, especially if the dissolving of the ferro requires any length of time. In con- sidering the extent of the rephosphorization, it is evident that the content of in the slag must have an important bearing on the result. The following tables, 64 to 59, are compiled from a large number of heats, and are classified according to the percentage of PjOft in the slag. In each table the charges are grouped to exhibit similar degrees of rephosphorization, and for every group the average analysis of the slag is given.

Many of the variations in each table are within the limits of error, and are not explained by the composition of the slag. The. system

VOL. XXII.— 31 r\r\n]o

? IvL

The Op£N-H£Abth Pbocebs.

Table 54. — Action op Phosphorus During Becarburizatiok.

Slags oooUioiog lew than 5 0 per oeoL PsQ|.

LimfU of Change In Phoipbonis-content.

Slag.

Increase of P between .004 and .009 per cent incl. ;

(average of 15 beau)

Increase of P between .002 and .003 per cent. incl. ;

(average of 14 heats)

Increase of P betwen .000 and .001 per cent, incl.;

(average of 14 heats) —

Decrease of P between .001 and .016 per cent. incl. ; (average of 17 heats) -

FeO.

CaO.

ondet

of averages applied id these tables gives no information concerning the individual heats. Thus, an average of 16 per cent SiO, may

Table 55. — Action op Phosphorus Dubing Reoabburizatiox.

Slags containing 5.0 to 10.0 per cent. Tfi.

IJmita of Change in PbosphoniBKK>ntent.

Slag.

Rephosphorixation.

Increase of P between .006 and .014 pr. ct. inclusive: (average of 8 heats...

FeO.

Pa.

CaO.

Dephosphorization.

Decrease of P between .003 and .016 pr. ct. inclusive: (averaireof 9 heats)

result from the combination of six heats of 12 percent, and six heats of 18 per cent 9 or from eleven heats of 14 per cent and one heat of 26 per cent.

Table 66. — Action of Phosphorus During Recarburization.

Slags containing 10.0 to 15.0 per cent. PfO.

Limits of Change In Phosphoras-coiitent

Slag.

Repbosphorlzation.

Increase of P between .005 and .019 pr. ct. inclusive: (averaire of 10 heats)

SiOs.

FeO.

Pa.

CaO.

Dephosphorization.

Decrease of P between .004 and .106 pr. ct inclusive: (averaire of 9 heats)

The Open-Hearth Pbocb38.

Table 57. — AcrrioN of Phosphorus During Recarburization.

Slags containiDg 15.0 to 20.0 per cent, PsO.

Limits of Change in PhosphorosHSODtent

Slag.

Rephosphorixation.

FeO.

PfO,.

CaO.

Increase of P between .013 and .017 pr. ct. induflive f average of 5 heatsV

ondet. nndet.

Dephoflphorixation.

Decrease of P between .007 and .017 pr. ct. inclusive : (avenure of 5 heats)

To show more clearly the nature of the forgoing results, those heats which produced slags with the greatest content of SiO, and of FeO are arranged in tables 58 and 59 :

Table 58. — Action of Phosphorus During Recarburization.

Slags below 5.0 per cent. P,0|.

Change In Phosphorus,

Slag.

SiOs.

FeO.

CaO.

Increase of P, per ct., .009 " " .001

Decrease " .002

Increase " " .003

" " .002

Decrease - .001

nndet.

w

M M

Table 59. — Action op Phosphorus During Recarburization.

Slags over 5.0 per cent. PsO.

Change In Phosphonis.

Slag.

SiO,.

23.95*

FeO.

CaO.

Increase of P, per ct, .046 " " " .003

Decrease .002

" " .016

Increase " " .008

Decrease " " .045

nndet.

NoTi.— This slag is not included in any of the foregoing tables.

484 The Open-Hearth Process.

From these data we may draw the following conclusions :

(1) With slags containing under 5 per cent. P3O5, and not over 20 per cent. SiO,, the rephosphorization need not exceed 0.01 nor average over zero per cent.

(2) With slags containing from 5 to 10 per cent. PjO,, and not over 19 per cent. SiO; the rephosphorization need not exceed 0.015 nor average over 0.005 per cent.

(3) With slags containing from 10 to 15 per cent. P2O5, and not over 17 per cent. SiO,, the rephosphorization need not exceed 0.02 nor average over 0.005 per cent

(4) With slags containing from 15 to 20 per cent. P2O5, and not over 12 per cent. SiO,, the rephosphorizatiou need not exceed 0.02 nor average over 0.01 per cent.

Chapter X. — Conditions of SucJcessful Practice.

Sec. 59. — Persomd Equations,

The operation of melting and decarburizing a charge of metal can proceed under various conditions. There may be either a large excess, or a bare sufficiency, of silicon, manganese and carbon ; the flame may be sharp or soft ; melting may be slow or fast ; decarburization may be eflected by additions of ore in small or large doses, or by the flame alone, with or without dilution by scrap ; the recarburizer may be added in either the furnace or the ladle. All these modifications are possible, if the melter thoroughly understands his own method. Each successful furnaceman has certain peculiarities, finding, as he does by long experience, that some particular detail is essential or beneficial in his particular practice. Into this class of local, tran- sient, and personal equations fall the various ideas concerning the pre-eminent importance of a certain casting-temperature, of a certain duration for the boil, and of certain other specific conditions which should exist when the recarburizer is added. In this connection it is well to bear in mind that bad results in the working of steel are not always due to bad furnace-work, but sometimes to faults in the rolling-mill. Thus, of two mills that were supplied with the same kind of material at the same time, the deliveries extending over many months and amounting to many thousands of tons, one rejected four times as many plates as the other on account of imper- fections. Thus, apparently, the good results of a particular system may be due not to any virtue of the furnace-process, but to fortuitous

The Open-Hearth Process. 486

conditions in the rolling-mill, or to the fact that a peculiarity in one department is perfectly adapted to some peculiarity in the other. Such an adaptation is usually not accidental, but arises from the selec- tion of means to ends. It is the fruit of much study and trial ; but it does not follow that a particular detail, however advantageous as part of a special system, can be heralded as a metallurgical sine qua non.

Sec 60.— The Ddermining Vartahlea*

The two agents at the command of the melter are heat and oxygen, and the differences in their use and control, according to Mr. Hib- bard, give all the variations of which the process is capable under a given set of conditions. But these given conditions vary widely — the proportion of pig charged and the nature of the bottom are the principal factors, while the proportion of stock charged at first, the amount of manganese in the metal, the rate of melting, the kind of flame and the amount of waste figure as the minor ones. All these variables have been discussed in the foregoing pages, but Mr. Hibbard has ingeniously arranged them in the form of a table. His classifica- tion of hearths, however, as acid, neutral and basic, is open to criti- cism ; for though it might be proper to distinguish between bottoms made of dolomite, which are universally styled basic, and those of niagnesite which are often, but erroneously, called neutral, yet to consider such a distinction as fundamental appears to be a serious error.

The argument against the popular misconception of this subject has already been presented (Sees. 35 and 36), but it may not be amiss to repeat, that the nature of the bottom has very little direct influ- ence on the result — it determines whether acid or basic slag can be carried, and that is the limit of its legitimate function. It is true that the character of the slag determines the chemical history of the bath, and hence it might be said that, indirectly, the nature of the hearth had determined the reactions. But, in Mr. Hibbard's table, the nature of the slag is duly taken into account, so that by the sub- stitution of the nearer and the truer cause, the original determinant is deprived of its pre-eminence. It should never be forgotten that the highest function of the hearth is to remain passive. If the

The best presentation of the subject of furnace-manipulations which has come to the writer's knowledge, is to be found in a paper by Mr. H. D. Hibbard, pub- lished in The Iron Age July 2, 1891. In the present section this essay will be con- sidered at length.

486 The Open-Hearth Pbocesb.

attempt were made to run basic slags on a sand bottom, or highly siliceous slag on a basic bottom, then the hearth might be considered a living agent ; but, in practice, the slags are always adjusted to the bottom, and the latter is relegated to a very subordinate position as a final agent. From this point of view the term '' neutral " should be eliminated from the table. (See Sees. 35 and 36.)

Mr. Hibbard gives four grand working-divisions as follows :

Low temperature and low oxidation ;

Low temperature and high oxidation ;

High temperature and low oxidation ;

High temperature and high oxidation.

It is on these general divisions that the minor refinements of prac- tice are superposed. The author expresses the belief that eventually the basic process will be given up to the manufacture of soft steel, and the acid process to that of hard steel. The basic is considered the preferable process for the low-<rbon metal —

Since the manganese is being taken up constantly from the slag, which man- ganese is at work removing oxygen and silica from the metal, a most desirable state of affairs having a direct bearing in improving the quality of soft steel. Ore can be used freely enough to induce the degree of ' boil desired without oxidizing the metal.''

The continuous travel of manganese back and forth through slag and metal (which has been referred to in Sec. 46 d), must be regarded as a matter of theory, so far, at least, as its occurrence to any con- siderable extent is considered. As for the use of ore without thereby oxidizing the metal, the reader is referred to Sec. 27, where the theory is advanced that the oxidation of the metal is not a function of the ore, or of any protecting manganese, but is primarily controlled by the demands of the cinder. The last sentence of the quotation might be construed to imply that a lively boil in the acid prooew injuri- ously oxidizes metal. That such, however, is not the author's mean- ing appears from this subsequent paragraph :

In the manufacture of soft steel* the writer, in common with most melters, con- siders a fairly lively boil beneficial to the quality. What the direct cause is we do not know. Something is probably worked out, and if so, that something is proba- bly gaseous, and, further, an oxidiiable gas. These suppositions are supported by the fact that oxidation and its effect, boiling, removes almost completely those ele- ments (silicon and manganese) which are known to have great solvent effect on the gases, allowing the free escape of the latter. The writer has, with no good reason, long suspected hydrogen to play a part here."

Mr. Hibbard has withdrawn this last observation on learning the

The Open-Hbarth Process. 487

analysis of the bath-gases as given in Sec. 52. Doub* must also be expressed as to the correctness of the supposition that " the some- thing worked out " is " probably an oxidizable gas." Possibly a gas, but probably oxide of iron, albeit oxide of iron is eliminated by the action of carbon in the form of carbonic oxide, which is an oxidiza- ble gas, yet this is merely a coincidence of terms and not of chemical reaction. Further remarks on the boiling are given as follows :

" Giyen carbon in the bath, and with silicon and manganese largely remoyed, the rate of boiling depends on the proportion of oxide of iron in the slag nn- combined with silica, together with the degree of heat on the metal. High tem- perature retards boiling while low favors it."

The universal tnith of this last statement must be questioned. It is true, that at high temperatures there is a tendency in acid prac- tice to reduce silicon from the hearth (see Sec. 33), and thereby check the action by deadening the bath ; but this is not sufiBcient ground for a generalization on the effect of temperature for all linings and all conditions as embodied in the thesis that high temperature retards boiling.

"The melter tries to get the degree of boil before tapping, which experience has tanght him to be the best for the grade he is making, and herein lies the control of oxygen, and to some degree of temperature, by the average workman. For soft steel, dead action is bad, providing C is present, while for steel castings or hard steel lively or wild action is injurious."

Concerning this last proposition it should be said that, while it may hold true for some particular practice, it is not a general and arbitrary law. Possibly, the same criticism might also be made on the following paragraph, for it is feasible to completely alter the status of the metal and slag just before the introduction of the recar- burizer,80 that antecedent conditions cannot be regarded as arbitrary determinants :

" A heat of soft plate-steel, made with a minimum of boil, will have a higher elastic limit and tensile strength, with less elongation and reduction of area, for a given chemical composition, than one made with a lively boil but in which the oxi- dation is kept under control. Steel oxidized still further, and given a wild boil, but kept within workable limits as regards red-shortness, has still less elastic limit and tensile strength, but less elongation as well, though such steel shows at times remark- ably high reduction of area, the test-piece drawing in greatly at one point when palled. The finishing temperature at the rolls also has a great influence in deter- mining these physical properties, but is beyond the limits of our subject. Makers who work with more strongly oxidizing conditions must, for a plate of given tensile strength, aim for higher carbon than those who use oxygen more sparingly."

488 The Open-Heabth Process.

Concerning the slag, Mr. Hibbard says :

"The amount of the slag depends on the amonnt of dirt in the charge, the rate of melting, percentage of waste, quantity and quality of ore used, and waste from the bottom. When it is large it acts as a fly-wheel on boiling, making it slow to start and slow to stop, and in that way may mislead or cause trouble. Further, act- ing as a blanket, it will require the furnace to be kept hotter to maintain the metal at the proper temperature. It is usually accompanied by extensive cutting of the bottom, and is generally not a cheerful thing to see, especially in the basic pro- cess.

" The color of the slag indicates the amount of free oxide of iron (probably Fe804) contained, and therefore the degree of oxidation to which the metal is being subjected, the black having the most and the yellow, perhaps, none.

The texture of the slag indicates this more fully. Earthy and crystalline slags are only black, while vitreous slag may be either black, olive, or yellow. The tex- ture varies much with the size of the test taken. The thinner the test, the more likely it is to be vitreous. This extreme is met in slag tests taken on a small peel or spatula touched to the slag in the furnace and quickly withdrawn. The writer usually pours a cake on the charging-floor plates, but in this case the thickness of the cake must always be considered. The slags of the basic process are always black and earthy, or should be for satisfactory work.''

On the interesting and important subject of ca.sting-condition8, the following is quoted :

" The action of the gases during teeming has much to do in determining the mar- ketable fitness of the product In grades of soft steel to be rolled, the gases should evolve freely during teeming and setting. In hard steels and steel castings they should be kept in solution. The middle ground between these extremes is full of danger, and unsound steel is likely to result.

"At many works especial difficulty is met in making some one grade of carbon, all others, higher and lower, being easily produced. Their practice, particularly in reference to the degrees of oxidation maintained, but casting-temperature as well, finds a carbon at which the gases neither escape freely nor are all kept in solution, and the steel as cast is unsound. Under thb head comes for consideration ' piping' and rising ' steel. We find in soft sted one kind and in hard another kind of pipe. In soft, very low-carbon metal, with rather pronounced oxidation and absence of appreciable amounts of silicon or aluminum, the gases are so freely evolved in mi- nute bubbles that they may form half the volume of a mould apparently full of steel at the end of teeming. Later, when this gas has escaped, the metal sinks down and leaves a hollow shell or pipe of metal which has chilled next to the mould. Basic steel is prone to do this.

" Hard steel which pipes, does so because of the shrinkage of the metal during setting and cooling. The gases are kept in solution by ingredients having that effect, and the loss of volume of the steel in setting for the whole ingot or casting is concentrated in the pipe.

" Piping or incipient piping steel is to be desired. In soft steel a slight tendency to pipe is aimed at. This steel will rise a little just before the last of it has solidi- fied if allowed to do so, but it can be prevented and good ingots obtained.

"Rising steel is that which increases in bulk after teeming, the result of the for- mation of bubbles in its interior which do not rise and escape. The casting-tern-

The Open-Hearth Process. 489

perature of soch steel is too high, and this brings us to the consideration of the effect of temperatnre on steel. The trouble, when this is too high, is well known, but not the reason why. It seems to be chiefly from the effect on the gases It is not easy to see how heat per can work injury, but by increasing the solvent power of iron for gases to the point that they are eyolved so slowly in the cooling steel that they cannot escape, being entangled, as it were, in the plastic steel, but yet not held completely in solution, high heat can and does exert a positive influence in making bad steel."

A point which has not been considered is the different structure which may obtain with different temperatures. The greater part of the crystallization-record will be made while the steel is fluid, and, therefore, the extent of such phenomena will depend on the excess in the temperature of the molten metal over the point of solidifica- tion. In Sec. 20 it was pointed out how close the tem))erature limits of practice are. Thus if the critical temperature be 1600° C, and one charge is tapped at 1620° and another at 1660°, the period dur- ing which the metal will remain fused in the moulds will be about three times as long for the hotter as for the cooler charge. This alone would make many differences in structure possible, and might even cause the creation of malignant alloys — all of which is hypo- thetical, but not impossible.

Furthermore, the greater contraction must be considered, which takes place when the metal cools from very high temperatures. If this contraction were uniform, the damage resulting from it would be less; but, whatever the casting-temperature, the steel chills in contact with the mould and forms an envelope which must neces- sarily follow the interior in all its changes. The hotter the metal the thinner this shell is and the more it has to contract, thus doubling the evil. From one or all of these causes, and possibly from others not enumerated, arises the bad quality of hot steel. Con- tinuing this subject, Mr. Hibbard says :

*'To make steel requiring the free solution of gas in the mould, the temperature must be as low as may be without giving a skull in the ladle. With oxidation rather high the limit of permissible temperature is raised, but red-shortness lies in wait to head off the melter who would use this agent freely to avoid close watch of the heat of his bath.

" High-carbon piping-steel should not be too hot, as this increases the pipe. For steel castings, high temperature is needed to insure filling all parts of the mould. The pipe must be guarded against by use of large risers or sink-heads.

Having the steel hotter than the casting-temperature earlier in the heat has but little influence, except indirectly, by its effect on the slag. An abnormally high temperature melts from the roof, walls, or bottom, or from all of them, a quantity of siliceous material in the acid process, which adds to the slag, changes its quality, makes it more vitreous, and so affects the work. In the basic process, the siliceous

490 The Opek-Hearth Process.

material melted from the roof and sides is sometimes enough to destroy the dephos* phorizing quality of the slag. The normal coarse of the temperature of a heat is to rise steadily until it reaches the proper degree for casting at the end.

" The weight of the charges has a bearing on the resulting quality. When the weight is below 10,000 pounds, trouble is found in having the metal hot enough to cast properly without melting: the furnace. The difficulty of this increases about inversely as the square of the decrease in weight. Below 3000 pounds a ladle can- not be used satisfactorily, but the metal must be run directly from the furnace to the mould. On the other hand, large heats cast into small ingots engender bad results somewhere. If the casting-temperature is right at the beginning of teeming, it will be too low at the end ; and if right at the end, too hot at the beginning.

The time of making a heat may or may not affect the quality. Where abnor- mally long, due to bad furnace-work, which may come from bad gas, badly-designed furnace, bad draft, or ignorance, the quality is pretty apt to be injured. A melter can hardly have had enough experience with this class of heats to know exactly what to do to keep the quality up to his best, though it may pass inspection.'*

Chapter XI. — The Use and Loss op Material.

The object of this chapter is to show how the details of practice affect the cost of operation through the waste of stock and the use of ore and lime. In order to make a complete comparison of two processes, — " pig and ore " and pig and scrap," for example, — it would be necessary to consider the effect upon the cost-sheets of the diminished product under the first practice, and the increase per ton of product in labor, fuel, and repairs. These considerations are not embraced in the scope of pure metallurgy, but there are others which fall properly within its domain.

Part I. — The Acid Process.

Sec. 61. — The Conditions which Limit Comparisons.

The factors of the acid process which will be considered are:

(1) The loss in weight due to the presence of sand and to the oxi- dation of certain constituents of the charge.

(2) The loss of metallic iron held in chemical combination by the slag.

(3) The loss of metallic iron held mechanically, as shot, by the slag.

(4) The amount of ore required.

(5) The conditions of melting.

In most published data the percentage of loss is given in definite figures, as though it were a fixed exponent of practice. Thus, in H. M. Howe's admirable treatise on The Metallurgy of Steel, we find

The Open-Heakth Process. 491

it stated that at Landore, with '' pig and ore/' the loss was 22 per cent. ; at South Boston (no mention being made of the kind of ma- terial used) it was 16.63 per cent. ; at Creusot 6 and at Terre-Noire only 5 per cent. The inference that the practice at South Boston was very bad, at Creusot much better, and at Terre-Noire still better, might not seem unnatural, yet it would not be warranted. It is evident that in calculating the loss at fiandore, the ore was included in the material used ; the heavy loss at South Boston would indicate that the same method of determination was followed there; whether this is also true of Creusot and Terre-Noire, or whether the ore was omitted in one and not in the other, is a point on which no informa- tion is given.

This omission may be explained by the current custom of includ- ing the ore in the column of material charged. Assuming this to have been the case in each of the above records, other data for a proper comparison still remain lacking. It is quite possible that in the South Boston practice a large proportion of pig-iron was used. This, in itself, would occasion an increased loss, owing to the greater amount of silicon and carbon to be burned. Besides, more ore would have to be used and, since hematite, even when absolutely pure, can contain only 70 per cent of metallic iron, it is obvious that the loss would be doubly increased. Thus, a loss of 16.63 per cent, may represent the very best of work with much pig-iron, or the very worst practice with a good mixture of pig and scrap. In the same sense the work at Creusot may have been far superior to that of Terre-Noire. In short, the figures signify little without additional data concerning the composition of the charge.

In calculating percentages of loss, as given below, the ore is not included, for it appears unreasonable to enter its 35 or 40 per cent, of silica and oxygen on the debit side of the account. Similarly, it might be argued that allowance should be made for the silicon and carbon of the pig-iron, and, strictly speaking, the claim should be conceded, for the real criterion must be the waste of metallic iron ; but as the silicon and carbon reduce iron from the ore, they make some return for their waste. Besides, the percentage of impurities in pig-iron is much less than in ore, and the error is not so important. Furthermore, pig-iron is used to greater or less extent in every practice, while ore may not be used at all. The basis for the calcu- lations is purely theoretical, and it must be acknowledged that the results can never be fully realized in practice. Experience proves that the amount by which actual results fall short of theoretical ones

492 The Open-Hearth Process.

increases with the number of operations which enter into the process. Undecisive as the theoretical losses therefore are, yet they constitute a basis of comparison between different systems which is not without value. In this limited sense the following calculations should be understood.

Sec. 62. — Theordieal Losses.

a. From Sand and Oxidizahle Elements, — The sand, silicon, man- ganese, and carbon of the pig are a total loss so far as product is concerned. The sum of the respective percentages of each will rep- resent the waste from this cause.

6. From Iron held in Cliemioal Combination in the Slag. — The preceding section tells only a part of the story. The sand which adheres to the pig-iron forms slag ; so also does the silica arising from the oxidation of the silicon. As the quantity of slag is deter- mined by the silica (Sec. 27), each pound of silicon creates a certain amount of slag, which contains certain amounts of iron, chemically and mechanically combined. Thus, each pound of silica induces a certain waste of iron, the exact amount depending upon the compo- sition of the cinder. The proportions of SiOj and FeO in the cinder are determined largely by the quantity of manganese in the charge (Sec. 29) ; the slag made with coal-gas fuel, given in Sec. 28, contained SiO, 49.40 and FeO 29.79 per cent. ; with no manganese present, the slag might be composed entirely of these two con- stituents.

The following summary will show the law governing three repre- sentative conditions. It will be noticed that Slag A closely approxi- mates to the slag just referred to in Sec. 28.

Slag A : SiOj, 60; FeO, 30; Fe 23.33 per cent. One pound of silica carries 0.47 pound of iron. One pound of silicon carries X 0.47 1.01 pounds of iron.

Slag B : SiOj, 50 ; FeO, 40, Fe 31.11 per cent. One pound of silica carries HH 0.62 pounds of iron. One pound of silicon carries f f X 0.62 1.33 pounds of iron.

Slag C : SiO,, 50 ; FeO, 50, Fe 38.89 per cent. One pound of silica carries f f f J 0.78 pounds of iron. One pound of silicon carries ff X 0.78 1.67 pounds of iron.

c. From Shot in the Slag, — A certain amount of metal is always mechanically held, as shot, in the slag. In the case of a viscous cin- der the amount is quite large, but in ordinary acid open-hearth slag, the percentage will not run over two hundredths of one per

The Open-Hearth Process. 493

cent, of the total elag. Thus, each poand of silica will carry away in the form of shot only 0.0004 pounds of iron, or an amount too small to consider (49.40 is the percentage of SiO, in the slag given in division 6 of this Section).

d. From the Ore Used. — As most of the ore undergoes reduction, it is not an item of working-expense only but becomes a source of iron. Though the cost of fuel, labor, etc., of a large open-hearth furnace renders the reduction of this iron expensive, yet, since the ore is needed as an oxidizing agent, its correlated reduction is a con- sideration of importance. Theoretically, its use should have little eflJsct in increasing the amount of slag, but practically there is a continuous scorification of the bottom during its addition (see Sec. 63). The silica of the ore will act as a slag-maker (see division 6 of this section), and as the ordinarily available ore contains about three per cent, of silica, every hundred pounds of ore will carry from 1.41 to 2.84 pounds of iron into the slag from the action of its silica. On a sixty-thousand pound charge the loss from one hundred pounds of ore would thus be from .0023 to .0039 per cent., and with five thousand pounds of ore it would be from one-tenth to two-tenths of one per cent, of the charge. Even if the ore carried six per cent, of silica, the loss from this cause would not be a matter of vital im- portance.

Sec. 63. — The Conditions of Melting.

The amount of silica derived from the hearth will obviously be an important factor, yet it cannot be accurately equated, since the scorification is affected by many circumstances (Sees. 25, 27, et cU,), If it be assumed that the charge is arranged in the furnace to the best advantage in all oases, and melted rapidly with a full, clear flame, the important points for consideration will be the time of ex- posure afler melting, and the kind of gas used.

a. Time of Exposure. — When a large proportion of the stock consists of pig-iron the metal, after melting, is high in carbon, and sometimes high in silicon also, and the operation is lengthened. This involves a certain amount of scorification. On the other hand, the melting itself is rapid, and the cutting during the period of fusion is very slight

In a series of fifteen consecutive sixty-thousand-pound charges of two-thirds scrap and one-third pig, the wear of the hearth was twelve hundred pounds per charge, as determined by the quantity of sand required to repair the slag-line. This gives, for the loss of iron

The Open-Hearth Process.

when FeO 30 per cent, of the scorified cinder (Slag A, Sec. 62), 1200 X 0.47 664 pounds of iron 0.94 per cent, of the charge. With heats of pig-iron alone, it is safe to assume that the wear of the bottom would have been 1800 pounds per charge, and the FeO 40 per cent. (Slag B, Sec. 62), giving for the loss of iron in the scorified cinder 1800 X 0.62=. 1116 pounds of iron 1.86 per cent of the charge.

6. Kind of Oas. — In Sec. 28 the data for two groups of charges are given — one melted with coal-gan, and the other with oil-gas con- taining steam. The following calculations are based upon those heats:

Table 60.— Comparative Effects of Coal- and Oil-Gas ON THE Waste of SiO,.

Theoretical Waste of SiOt

Charges Melted with

Coal-gas.

OU-Oas.

(1) Total SiO, ID slag before tapping, pounds

(2) Toul SiO, from combustion of St

and from L4 per cent, of sand adhering to pig iron, pounds

(3) SiCfrom bottom (1-2)

Ratio

It will be seen that the theoretical calculation for the coal-gas series, which gives 1293 pounds of wear, agrees closely with the ac- tual figures, 1200 pounds, given in division (a) of this section. If the scorification with coal-gas involves a loss of 0.94 per cent of the metal, as calculated in (a), then, according to the above ratio, the loss with oil-gas containing steam will be 1.27 X 0.94 1.19 per cent. In addition, the stronger oxidizing conditions, by themselves, pro- duce a slug with a higher content of iron oxide. The slag produced with coal-gas contained 29.79 per cent. FeO, and with oil-gas 34.11 per cent., or a ratio of 1.00 to 1.15. The increased loss under oil- gas, due to the greater scorification and to the more basic slag com- bined, will, therefore, be 1.16 times as much as the amount due to the scorification alone, and the true ratio will be 1.00 to (1.16 X 1*27), or 1.00 to 1.46. Thus, if the loss with coal-gas be 0.94 per cent (as above), with oil-gas it will be 1.37 per cent.

Per cent.

Per cent

The Opex-Hearth Process. 495

Sec. 64. — Summary of BesuUs.*

The percentage of metal lost in acid work may be summed up as follows :

(a) The percentage of the weight of the charge lost by the re- moval of foreign elements is equal to the sum of the percentages of those elements in the charge.

(6) The percentage of iron lost in the slag, due to the sand amounts to the product of the percentage of sand multiplied by :

8I0> FeO.

P

0.47 when the slag contains

(c) The percentage of iron lost in the slag, due to the oxidation of silicon amounts to the product of the percentage of silicon mul- tiplied by :

SiO,. FeO.

Per cent Per cent

1.01 when the slag contains 60.0 30.0

1.33 " 60.0 40.0

1.67 " " 60.0 60.0

(d) The loss of iron in shot is inconsiderable.

(e) The loss due to silica in the ore does not exceed two-tenths of one per cent.

(/) The percentage-loss of metal, due to scorification of the hearth, may be taken as 0.94 per cent., with a charge of pig and scrap ; and 1.86 per cent., with a charge of all pig-iron.

(ff) With oil-gas containing steam, the loss of iron is 1.46 times as much as with coal-gas.

The foregoing summary does not include the relations of the recarburizer, or of the ore-additions to the loss. The loss of man- ganese should be nearly the same for each kind of practice ; in the following pages it is assumed at 0.20 per cent, of the charge. The effect of the weight of the recarburizer upon the percentage value of any factor is neglected. An item of waste amounting to 1 800 pounds on a heat of 60,000 pounds would show a loss of jY(/A= P® cent. If care were taken to include the recarburizer, the corrected percentage would be 2.98 per cent., a difference of only 0.02 per cent In the following comparative calculations therefore, this correction is omitted. The quantitative effect of the ore must be

All the statements of this section are subject to the assamptions and limita- tions ennmerated io the beginning of this chapter.

496 The Open-Hbabth Process.

computed for each separate practice ; knowing the amount used and its composition, it is to compute what the bath should receive from the process of reduction.

Sec. 65. — Results of Acid Practice With Pig and Scrap.

The figures for any one charge are worthless in determining the loss in a furnace. Mistakes in weighing are frequently made through ignorance or carelessness, and cumulative errors are not at all im- probable. Besides it may happen that either more or less than the average amount of scrap and skulk is made in the particular heat selected, and an accurate inventory of such material is not always practicable. The conditions of melting, the disposition of the stock and the scorification of the bottom may all be abnormal, and may make the result far from representative. Particularly to be avoided is any deduction from the results obtained on a new bottom. No matter how well the new sand may be hardened, it acts as a sponge to absorb slag and metal, and the first charge is a source of great trouble to the inexperienced. In the following quantitative data, care has been taken to eliminate as far as possible the above sources of error, so that the results shall indicate the history of regular practice.

Type I. — Pig and Scrap.

The figures are averages per heat of fifteen consecutive charges of 10-carbon steelmade by the Pennsylvania Steel Company, of Steel- ton, Pennsylvania.

Per cent, (est.) r Sand, 1.50

Pig-iron, 11,947 pounds U

V C, 3.50

Per cent ( r Si, 0.02

J 10,600 pounds, . . . . Mn, 0 35

34,893 pounds, Mn, 1.00

.0,0.40 Recarburizer, 300 pounds, . . . ; . Mn, 80.0

r Fe, 63.0 Ore, 1000 pounds, j SiO 6 0

Sand (repairs) 1170 pounds, SiO 98.0

r SiO„ 49.54

Slag, 4133 pounds, ] MnO, 17.09

(FeO, 28.72

The Open-Hearth Process. 497

Fuel, Siemens's prodacer-gas.

Pig-iron, scrap and recarboriser; poundff, . . . 57,800

Product, steel ; pounds, 55,307

Loss, pounds, 2493

Loss, per cent., 4.3

Calculating the theoretical loss by the formulsB summarized in Sea 64 (the same letters referring to the same items), we have :

Percent. Percent. Sand, 0.31

r Sa( I 8i,

(a) Average composition of charge, not in- .

cludiog there. . . . ]?;"' Z

I 2.78

(ft) 0.31x0.47= 0.15

(c) 0.59 X 1.01 0.60

(/) 0.94

Total, 4.47

The loss of manganese in recarburization on these heats would be about two-tenths of one per cent, of the total weight of the charge. The ore contained 630 pounds of iron or 1 .09 per cent of the total charire. The complete theoretical statement will therefore be:

Per cent.

Losses as above, 4.47

Loss of Mn, 0.20

Gross loss, 4.67

Less iron added from ore, 1.09

Net loss, 3.58

The actual weights thus show a loss of 4.3 per cent., while theory calls for only 3.58 per cent., a difference of 0.7 per cent. Part of this deficit is carried down the ports; part is lost during tapping, in sparks and minute splashes ; the remainder is within the probable error of the determinations. No allowance has to be made in the computation for the carbon and manganese in the product, owing to the method of equating the loss from the recarburizer, which is the agent from which the steel derives almost all its alloyed compo- nents.

Sec. 66. — Remdts of Acid PracHoe wUh Pig and Ore. Type II. — Pig and Ore.

The figures are averages per heat of ten consecutive charges of VOL. XXII.— 32

498 The Open-Hearth Process.

.10-carbon steel made by the Carbon Steel CompaDj, Pittsbargh, Pa.:*

Per cent (eit.) f Sand, 1.40

Pig-lroD, 60,000 ponnds, J Si, 1.60

Iq 350

Percent.

Ore, 13,600 pound.. {gw/IS

RecarbnrizeE, 400 pounds, Mn, 80.00

Fuel, Natural gas.

Pig-iron and recarburizer ; pounds, 60,400

Product, steel ; pounds, 60,000

Loss, pounds, 400

Loss, per cent, 0.7

Calculating the theoretical loss by the formulae summarized in Sec, 64, the same letters referring to the same items, we have :

Percent Percent

{a) Average composition of charge, not in- f i aa

(b) 1.40X0.62, -.0.87

(c) 1.60X1.83, 2.13

Total,. -. 1.46X4.86 7.10t

(e) The low percentage of silica in the ore, counterbalanc- ing the large quantity of ore added, 0.20 Loss of manganese in recarburizing, 0.20

Totalloeses, 14.00

MetaHic iron in ore 18 pounds, 15.20

Theoretical net gaiUf 1.20

The actual weights show a loss of 0.7 per cent., while theory calls for a gain of 1.2 per cent,, a difference of 1.9 per cent. As mentioned in the consideration -of the pig-and-scrap process, a part of this deficit is carried down the ports in fine spray. The amount

The author is indebted to Mr. H. W. Lash, General Manager of the Carbon Steel Company, for these results, which form a*ohapter in a history of most excel- lent practice.

t This multiplication of 6, c, and /by 1.46 is to allow for the increased scorification with natural gas, which is here supposed to act in the same manner as oil-gas.

The Open-Hearth Process. 499

lost increases with the quantity of ore used, and henoe should be more, in the present case, than in Type I. There will also be the same loss from sparks and splashes, and the probable errors will play their part in the history.

Part II.— The Basic Process. Sec. 67,— Theoretical Losses.

a. From Sand and Oxidizable Elements. — The sand, silicon, man- ganese, phosphorus, and the carbon are a total loss so far as product is concerned. The sum of the respective percentages of each will represent the waste due to their presence. When much sulphur is eliminated it may be added to the list, but cases are rare in which this element need be taken into account. .

6. From Iron held in Oiemioal Combination in the Slag. — In the consideration of acid work it was found that the quantity of slag could be easily predicated from a knowledge of the slag-creating powers of silica and silicon, since these were the determining factors for the given practice. In basic work, as already pointed out (Sees. 42, 44 et al.), the relations are more complicated. If pure stock is used, silica will be the governing £Eu;tor, the only requisite being that the silicic acid shall not much exceed 20 per oent. so that the hearth shall not be scorified. If impure stock is used, slag must be produced in such quantity that the phosphoric acid shall not con- stitute more than a certain percentage, which varies with the content of silica.

The following are calculations on typical hypothetical slags; it w assumed that a certain proportion of the slag will consist of MnO and MgO, with smaller amounts of AlO, etc.

Slag A: SiO, 30.0; 5X) or P 2.18 ; FeO 10.0 or Fe 7.78 ; CaO 40.0 per cent.

Such a slag might be produced just after melting a charge of semi-phosphoritic stock with a slight deficit of lime:

1 pound of silica is associated with . . . pounds Fe.

1 pound of silicon is associated with . -f ,

1 pound of phosphorus is associated with . 18.3 '' CaO.

Slag B: SiO, 20.0; PA 10.0 or P 4.37 ; FeO 10.0 or Fe 7.78 ; CaO 45.0 per cent,

The reader is referned to Sections 46, 46 And 58 ior examples of similar slags obtained in practice.

600 THE OPEN-HEARTfl PROCESS.

This also is an intermediate slag from the melting of impure stock with a slight deficit of lime :

1 pound of silica is associated with . . pounds Fe. 12.25 " CaO.

4.82 " CaO. 1 pound of phosphorus is associated with 10 3 CaO.

Slag C: SiO, 14.0 ; PA 15.0 or P 6.55 ; FeO 10.0 or Fe 7.78 ; CaO 45,0 per cent.

An intermediate slag, from impure stock melted with sufficient lime :

1 pound of silica is associated with . .

13.21 " CaO.

fl.20 Fe. 1 pound of silicon is associated with . . I A 88 " CaO

1 pound of phosphorus is associated with 6.87 '' CaO.

SlagD: SiO,= 20.0; PA 3.0 or P 1.31; FeO 15.0 or Fe — 11.67 ; CaO 45.0 per cent. A final slag, from moderately pure stock :

J .,. . . J . r 0.58 pounds Fe.

1 pound of silica is associated with . 1 2 25 " CaO

fl.24 " Fe. 1 pound of silicon is associated with . . 1 4 82 CaO

1 pound of phosphorus is associated with . 34.4 CaO.

Slag E: SiO, 15.0 ; PA 5.0 or P 2.18; FeO 18.0 or Fe 14.0; CaO 45.0 per cent.

A final slag, from semi-phosphoritic stock :

/ 0.93 1 1 pound of silica is associated with . . I OO

Fe.

r 0.93 pounds Fe.

/ 1-99 " 1 pound of silicon is associated with . . 1 6 43 CaO

1 pound of phosphorus is associated with . 20.6 CaO.

Slag F : SiO 12.0 ; PA 10-0 or P 4.37 ; FeO 20.0 or Fe 15.55 ; CaO 45.0 per cent.

A final slag, from phosphoritic stock. The percentage of iron is higher than necessary, but is not in excess of what will obtain if the operation is not carefully rulated :

, . . j 1.30 pounds Fe.

1 pound of silica is associated with . . 3 75 " CaO

r 2 79 " Fe. 1 pound of silicon is associated with . . 8 04 " CaO.

1 pound of phosphorus is associated with . 10.3 CaO.

The Open-Hearth Procebs. 601

Slag G : SiO, 12.0 ; PA 13.0 or P 5.68 ; FeO 13.0 or Fe 10.11 ; CaO 45.0 per cent.

A 6na1 slag id good practice, from phosphoritic stock :

1 pound of Bilica is associated with . ( q 75 CaO

. ... riO '' Fe.

I pound of silicon is associated with . . 1 8 04 " CaO

1 pound of phosphorus is associated with . 7.92 " CaO.

c. From Shot in the Stag.— In acid work the loss due to metal mechauically held in the slag was found to be very small. In basic practice the cinder is more viscous, and under ordinary conditions contains about six per cent, of shot. If the slag amounts to six per cent, of the charge, the loss from this cause will, therefore, be about one-third of one per cent. Should the slag amount to eighteen per cent., the loss from shot would be about one per cent of the charge.

d. From the Ore Used, — The silica added with the ore will act as shown in division 6 of this section. Theoretically the oxide of iron should have no effect upon the basic hearth, but, as a matter of fact, a continual scorification is in progress during the period of oreing; this is due not to any direct action of the ore-addition, but to the longer exposure of the metal to the liquid slag. The silica con- tained in the ore will act as a slag-maker in accordance with the formnlffi in division 6.

e. From the Lime Used. — From an examination of the slags in division 6 of this section and from the remarks made concerning them, it is evident that the proper addition of lime will depend upon the amount of silica to be satisfied and the extent to which the latter must be diluted for the elimination of the phosphorus.

The percentage of silica in the lime itself will largely determine the amount of lime required. It becomes, in fact, an item of con- siderable importance when a slag of very low silica-content is de- sired. In ordinary burned lime the silica and calcium oxide do not constitute over eighty per cent of the total. The remainder is mostly carbonic acid, which was not exlled in calcining, and water, which was absorbed subsequent thereto. The following analjrses represent two very common kinds of lime :

SiOt. CaO.

Per cent Per cent.

(1), 2.5 77.6

(2) 6.0 74.0

502 THE OPEN-HEABTH PROCfiSS.

The relative value of these limes in absorbing silica is deduced l)elow :

Slag A (above) : SiOj 30 per cent. ; CaO 40 per cent., whence the weight of CaO 1.33 X SiO,: Lime No. 1 ; 2.6 per cent. 8iO, :

Pweent.

Total CaO (as above), 77.5

CaO necessary to satisfy its own silica I.3S x 2.5, . . 3 3

CaO available for foreign silica 74.2

Lime No. 2 ; 6 per cent. SiO, :

percent

Total CaO (as above), 74.0

CaO necessary to satisfy its own silica 1.33 x 6.0 . . 8.0

CaO available for foreign sUica, 66.0

The proportional value of the two limes as absorbents of foreign silica is as 6G : 74 or as 1 to 1.12.

Slag G (above) : SiO, 12.0 per cent. ; CaO 45.0 per cent., whence the weight of CaO 3.75 X SiO Lime No. 1 ; 2.5 per cent. SiO :

Total CaO (as above), 77.5

CaO necessary to satisfy its own silica 3.75 X 2.5, . . 9.4

CaO available for foreign silica, 68.1

Lime No. 2 ; 6.0 per cent. SiO, :

Total CaO (as above), 74.0

CaO necessary to satisfy its own silica 3.75 X 6.0, . . 22.5

CaO available for foreign siKca, 51.5

The proportional value of the two limes as absorbents of foreign silica is as 51.5 : 68.1 or as I.O : 1.32.

This is not the entire history. These ratios do not express the whole relative values; in slag G the combined iron amounts to 10 per cent., and the shot to 6 per cent, additional, making 16 per cent, in all ; an increase of 32 per cent, in the volume of the cinder means an increase of 32 per cent, in the loss.

With pure stock the slag may not exceed six per cent, of the charge if the purer lime be used. This will give a loss of .06 X 16 0.96 or nearly one per cent, of chemically or mechanically en-

The Open-Hearth Process. 603

trained iron. An increase of 32 per cent, in volume of the slag would in this case mean an increase in loss of 0.32 X 0.96 0.31 or one-third of one per cent, of the charge. With phosphoric or sulphurous stock, the slag may amount to eighteen per cent of the weight of metal, occasioning a loss of .18 X 16 2.88 per cent, of iron. An increase of 32 per cent, in the loss in this case would amount to 0.32 X 2.88 0.9 per cent, of the entire charge.

If a charge contains little silicon and much phosphorus, it may be advisable to use an impure lime in order to create sufiBcient slag to hold the phosphoric acid. It seldom happens in practice, how- ever, that the accidental silica is insufficient for this purpose; the use of a particularly siliceous lime, therefore, can be justified only by necessity, or by a considerable difference in the first cost.

The foregoing figures show that the efficiency of any lime in ab- sorbing silica depends upon the composition of the slag which it is destined to make. The complete calculation may be summarized and illustrated as follows : Slag A: SiO,, 30.0 per cent. ; CaO, 40.0 per cent. ; weight of CaO 1.33 X SiO,.

Lime No. 1. Per cent

Total CaO, 77.6

8iO 2.5 per cent.

CaO necessary to satisfy its own silica 1.33 X 2.5, . . 3 3

CaO available for foreign silica, 742

Lime necessary to satisfy 1 pound of foreign silica <-i —

1.79 pounds.

In this manner the following table has been worked out for lime with 2.6 per cent. SiO, :

SlOt

Slag. percent. .

A, 30.0

B 20.0

C, 14.0

D, 20.0

E, 15.0

F, li.O

G, 12.0

Sec. 68.— The CondiHona of MeUing.

The scorification of the basic hearth depends primarily upon sil- ica ; the character of the flame is a matter of secondary importance

CaO er cent.

Pounds of lime required for 1 pound of foreign silica.

The Open-Hearth Proce88.

(8ec8. 39, 40 and 42). It is evident, however, that when all things are equal, the scorification is proportional to the time of exposure.

Sec. 69. — Summary of Results,*

The percentage of loss in basic work may be recapitulated as follows:

(a) The percentage of the weight of the charge, lost by the re- moval of foreign elements, is equal to the sum of the percentages of those elements in the charge.

(6) The percentage of loss due to each per cent, or unit of silica, -either in the ore, or lime, or attached to the pig-iron, and

(o) The percentage of loss, due to each per cent, or unit of silicon, are shown respectively in the following table :

Table 61. — Loss of Iron dub to Silicon and Sij-.ica.

Theoretical Slag.

Composition of Slag, per cent.

Per cent, of

Iron LoAt for each

per cent, of

A

B D E F G

FeO.

CaO.

(/) The loss due to shot in the slag amounts to 6 per cent, of the weight of the slag.

(g) The efficiency of the lime is governed by conditions of the slag, set forth in Sec. 67 e.

{h) The scorification is proportional to the time of exposure.

(h) The loss due to oxidation of manganese in recarburization may be taken at 0.20 per cent, of the charge.f

Sec. 70. — Comparison of Different Methods.

The following five representative methods or types of conducting the open-hearth process will now be compared :

All the statements are subject to the assumptions and limitations enumerated in the beginning of this chapter.

t See remarks at end of Sec. 64 concerning the valuation of the recarburiser and the ore.

The Open-Aearth Process.

Metal Charged Solid, Type L — Pig and scrap; low phosphorus and low silicon. Type II. — Pig and scrap ; mediam phosphorus and low silicon. Type III. — All pig-iron ; low phosphorus and low silicon. Type rV. — All pig-iron; high phosphorus and low silicon. Type V. — All pig-iron ; high phosphorus and high silicon. In this comparison certain constant assumptions are made viz. :

(1) The pig-iron is cast in chills and car- ries no sand.

Per cent

Si, 0.05

(2) The scrap averages . .

Mn,0.60 P, 0.10 C, 0.30

(3 1 The lime contains* . . . . -

r SiO 2.5 I CaO,77.5

(4) Low-silicon iron contains .

r Si, 0.70 I C, 3.70

(5) High-silicon iron contains . .

f Si, 2.00

I C, 3 00

Si, 0.70

Type 1.-15,000 pounds of pig-iron, .

P, 0.10 C. 3.70 Si, ao5

26,000 pounds of scrap, ...

Mn, 0.60

P, 0.10

L C, 0.30

1000 ponnds of ore (est), . .

f SiO,. 3.0 I Fe, 60.0

pounds recarburizer.

SiO„ 20.0

Final slag assumed to be similor to

CaO, 45.0

slag D (Sec .67),

I FeO, 15.0

Losses (see Sec. 69).

SiO„ 0.13

Per cent.

(a) Ayerages in the charge, including ore and lime,

Si, 0.29

Mn, 0 38

P. 0.10

Ic, 1.58

(6) .13 X 0.58.

, ,

(c) .29 X 1.24,

(/) 1503 X. 06 90 pounds, . .

(k)

Total losses,

. 3.35

Less 1000 pounds of ore 600 poun<

JsFe, .

. 1.50

Net loss,

1 s.*;

See Section 67

The Op£N-H£Arth Proceeb.

Pounds S10|. The charge contains 0.29 per cent, or 116 pounds Si,

oxidizable to 249

The ore contains 3 per cent. SiO . . - . . 30

Total SiOt,

279 X 3.13 873 pounds of lime required.

Tf.i .io 873 X 0.775 Total slag, —

Typb II.— Charge 40,000 pounds of naetal.

! 1503 pounds.

Per cent. ( Si, 0.70 i 13,333 pounds of pig-iron, averaging "J P, 0.85

i C, 3.70 Si, 0.05 Mn, 0.60 P, 0.10 C, 0.30

26,667 pounds of scrap, averaging

1000 pounds ore (est.), pounds recarburizer.

Final slag, assumed to be similar to slag Q Losses (see Sec 69).

(a) Averages in the charge, including ore and lime,

(6) 0.17 X 0.84, .

(c) .27X1.80, .

(/ ) 2477 X 0.06 1 49 pounds,

(A)

( SiO„ 3.0 1 Fe, 60.0

SiO„ 12.0

CaO, 45.0

FeO, 13.0

L P,0 13.0

' SiO. 0.17

Si, 0.27

Mn, 0.40

P, 0.35

C, 1.43

Total losses. Less 1000 pounds of ore 600 pounds Fe,

Percent.

Net loss, 2.32

Pounds 8iO The charge contains 0.27 per cent, or 108 pounds Si,

oxidizable to 231

The ore contains 3 per cent SiO, 30

Total SiO,,

261 X 5.51 1438 pounds lime required.

Total slag 0-775 3477 pounds.! 45

In the reduction of the amount of pig-iron from that used in Type I., allow- ance is made for the superior oxygen- absorbing power of the high-phosphorus pig. t The phosphorus in the pig-iron in Type II. is assumed at 0.85 per cent The

The Open-Hearth Process.

Typb in — 10,000 pounds of pig iron, .

6000 pounds of ore (est.)f .

pounds of recarburizer.

Final slag assumed to be similar to slag D (Sec. 67), . . . . Losses (see Sec. 69).

(a) Averages in the charge, including ore and lime,

Per cent SiO„ 3.0 Fe, 60.0

rero

( Si, 0.:

I c. s.:

SiOy 20.0 CaO, 46.0 FeO, 15.0

Per cent

Si,

p,

(6) .60 X .58,

(c) .70 X 1.24,

(/) 4204X0.08 252 pounds, 0.63

(k) 0.20

Total losses, Liess 6000 pounds of ore -

3600 pounds of iron,

Net gain, 1.85

Pounds SiO. The charge contains 0.70 per cent or 280 pounds

Si, ozidizable to 600

The ore contains 3 per cent SiO,, . . .180

Total SiO„

780 X 3.13 2441 pounds of lime required.

Total slag — — 4204 pounds.

Percent

Type IV.— 40,000 pounds of pig-iron, .

8,000 pounds of ore (est.), .

pounds of recarburizer.

Final slag assumed to be similar to slag G (Sec. 67),

r Si, 0.70

P, 1.15

I C. 3.70

f SiO 8.0

Fe, 60.0

SiO„ 12.0 CaO, 45.0 FeO, 13.0

phosphoric acid produced by the charge will be 321 pounds. By assumption, the final slag can carry only 13.0 per cent of Pfi 322 pounds. Hence, if the pig-iron contain more phosphorus, either a greater volume of slag must be produced, or, it most carry a higher percentage of the imparity. This is a case where the use of an impure lime would not be disadvantageous. It is true, that the loss of iron would be increased (as shown in Sec. 67 e), but the practice would be justified by the ne- cessity for removing the phosphorus.

The Openhbarth Procb88.

Umen (see Sec 69).

(a) Avenges in the charge, iocladiDg ore and lime,

Per cent Per eent

8iO 0.89

8i, 0.70

P, 1.15

I C, 3.70

(6) .89 X .84, .75

(c) .70 X 1.80 1.26

(/) 7970X0.06 478 pounds, 1.20

{k) 0.20

Total losses, 9.85

Less 8000 pounds of ore 4800 pounds Fe, .1 2.00

Netgain, 2.15

Pounds 810t. The charge contains 0.70 per cent or 280

pounds 8i, ozidizable to 600

The ore contains 3 per cent SiO 240

Total, SiO,

840 X 5.51 4628 pounds of lime required, 4628 X 0.775

Total slag

7970 pounds.

Type V. — 40,000 pounds of pig-iron,

10,000 pounds of ore (est), .

pounds of recarburizer.

Final slag assumed to be similar to slag G (Sec. 67),

Losses (see Sec. 69).

(a) Averages in the charge, including ore and lime,

Percent r Si, 2.0 P, 1.15 I C, 3.0 f SiO8.0 i Fe, 60.0

{SiO„12.0 CaO,45.0 FeO,13.0

SiO„1.45 Si, 2.0 P, 1.15 [ C, 3.0

Percent

(6) 1.45 X. 84=, 1.22

(c) 2.00x1.80=, 3.60

(/) 19,112X0.06 1147 pounds, 2.87

(k) 0.20

Total losses 15.49

Lem 10,000 pounds of ore 6000 pounds Fe, . 15.00

Net loss,

The Open-Heabth Process.

Pounds SiOf The charge contains 2 per cent or 800 pounds Si,

ozidisableto 1714

The ore contains 3 per cent. SiO, 300

Total SiO„

2014 X 5.51 11,097 pounds lime required.

Total slag ttt =19,112 pounds.

The following is a summary of the results obtained from the foregoing calculations. The amounts of ore are estimates of the proper quantities. The weight of recarburizer is not given ; the loss of metallic manganese is practically constant whatever the kind of the addition; this loss is anticipated in the table. In each case the weight of the recarburizer should be included in the material charged :

Table 62.— Comparison op Different Types of Open- Hearth Working.

Type.

Pounds of charge.

Per cent of Iron.

Pounds of lime.

Pounds

Pig-iron.

Scrap.

Recarb.

Ore.

Loss.

Gain.

of slag.

40,000 40,000

26,667

1,000 1,000 6,000 8,000

2,441

11,097

1,503 4,204 19,112

Sec, 71. — ResvJls ofBasio Practice. The remarks in Sec. 65 upon the possibilities of error apply equally to basic practice. It is to be regretted that no results are available of a long series of charges made from the same stock and with the same basic additions. The following record however, represents the continuous work of one furnace in the production of over twenty-four hundred tons of steel. The figures for the pig- iron give the average composition of the total amount, the calcula- tion being made from the quantity of each kind that was used, and its composition. The same is true of the figures for the scrap, re- carburizer and ore.

Per cent. r Si, 1.30

Pig-iron, 3,992,500 pounds averaging,

Mn, 0 25 P, 0.20 C, 3.60

The Open-Hearth Process.

Scrap, 1,781,900 pounds averaging,

Recarbnrizer, 104,860 ponnds averaging,

Total metal charged, 5,878,760 pounds averag- ing

Ore, 437,300 pounds averaging. Lime* 413,400 pounds averaging, .

Product, 5,565,000 pounds averaging,

Final slag assumed to be similar to Slag E (Sec. 67),

Calculation of slag.

Per cent. Si, 006 Mn, 0.60 C, 0.25 P, 0.12

Si, 6.40 Mn, 32.00 P, 0.15

Si, 1.00 Mn, 0.92 P, 0.18 C, 2.59

f Si03.25 I Fe, 62.00 fSiO 2.50 I CaO, 77.50 Si, 0.01 Mn, 0.40 C, 0.15 P, 0.02

SiO,, 15.0

CaO, 45.0 iFeO, 18.0

CaO in lime,

CaO in dolomite used in repairing the hearth, .

Total CaO added,

If CaO 45 per cent, of slag, the weight of slag Combined iron in the slag 14 per cent or 2.17 per cent of the charge.

8hot .06X 911,970=

or 0.93 per cent of the charge. Losses (see Sec. 69).

Pounds.

320,386

90,000

410,386 911,970 127,676

54,718

Theoretical.

(a) averages in the chai*ge, including ore and limei . . .

Per cent SiO 0.42 Si, 1.00 Mn, 0.92 P, 0.18 C, 2.69

Per cent

(h) and (c) as above,

(/) as above, .

Gross loss,

Less 437,300 pounds ore 227,396 pounds iron added,

Net loss,

The Bertha Zinomineb. 511

Actual I068 as hj above records : Poandi.

ToUl metal charged, 5,878,760

Prodact, 5,56'),000 pounds steel at 09.42 per cent, iron 5,532,728

Loss, . . . ' 846,037

or 5.9 per cent of the charge.

Thus the'actual weights show a loss of 5.9 per cent., while theory calls for 4.54 per cent., a difference of 1.36 per cent. The loss from sparks and splashes, and from material carried down the ports, and errors in assumptions and determinations, will account for the dis- crepancy.

Tee Bebtha Zinomines At Bebtha, Ya.

BT WILLIAM H. CASE, BERTHA, VA. (Chicago Meeting, being part of the International Engineering Congress, August, 189S.)

The Bertha zinc-mines of the Bertha Zinc and Mineral Company are in that part of the State of Virginia popularly known as Southwest Virginia, and are included in that division of the "Great Valley" named in recent years the " New River-Cripple Creek Mineral-Region of Virginia." The mines lie wholly in the southeast part of Wythe county, two miles from Pulaski county on the east, and three miles from Carroll county on the south. The only other zinc-mines that are opened in the region are those of Manning & Squier, which ad- join the Bertha zinc-mines on the northeast, and the mines of the Wythe Lead and Zinc Company, at Austinville, eight miles to the southwest. A little farther southwest of Austinville, near Ivanhoe, a new company is preparing to mine zinc-ores. The mining prop- erties named are included within a distance of ten miles along a direct line, bearing north 56° east, on the south side of New river, which here flows northeastward in a tortuous course. The greatest dis- tance of any of the mine-openings from the river is a little less than two miles.

It is not the magnitude of the operations in zinc-mining in this particular region, as compared with that of other well-known dis- tricts, that merits especial notice. Aside from the exceptional purity of the spelter made directly from the Bertha ores, some peculiar features of the zinc-ore deposits, as shown at the Bertha mines, and

Fig. I.

The Bertha Zinc-Mines. 613

some interesting mining-problems that have attended their ezploita- tion, have suggested the preparation of this paper.

1. Oeologiodl Feaiures.

The mining-property of the Bertha Zinc and Mineral Company at Bertha, comprises 742 acres, in fee, the larger portion of which is considered to be zinc-ore bearing. The tract has a northwest frontage of one and one-half miles along New river. The North Carolina Extension of the Norfolk and Western railroad skirts this boundary along the river front. Topographically, the tract is an elevated plateau, with a rolling surface and rounded hills, its gen- eral elevation being about 300 feet above New river, and 2200 feet above sea-level.

Geologically, the tract is wholly underlain by the Lower Mag- nesian limestone, No. II., of the State Geological Survey. The limestone rests directly on the Potsdam slates, and is everywhere covered with a varying thickness of heavy, red and brown clay, ex- cept where it forms bold and picturesque blufis along New river. The principal geological features of the district are shown in Figs. 1 and 2. Huronian and Laurentian areas succeed the Potsdam to the southeast, towards the Blue Ridge.

By reference to Figs. 3 and 4, it will be seen that the mining oper- ations are upon the outcrop of strata of zinc-bearing limestone, a little more than a mile from the river. The dip of the limestone is between 6° and 7° toward the river, and the zinc-bearing strata pass under the river at a depth of several hundred feet. Strictly speak- ing, the limestone does not outcrop to daylight. As shown on the section, the outcrop is covered with heavy clay, varying in depth from a few feet to more than one hundred feet. This clay probably results from decomposition of the magnesian limestone.

The zinc-ores thus far mined at Bertha are the silicate and car- bonate of zinc (calamine and smithsonite), the larger proportion being silicate. Their occurrence, in segregated bodies, and their free- dom from lead and iron, are unique. Reference to the enlarged sec- tion. Fig. 6, will facilitate a clear understanding of their situation. As will be seen, the outcropping limestone has weathered with ex- treme irregularity, so that, if the overlying clay and ores were en- tirely removed, there would be presented a veritable wilderness of limestone pinnacles, cones, columns, and domes, of varying heights, locally called " chimneys,'' some of which have an altitude of nearly one hundred feet above the bottom of the cavities at their base, but VOL. xxn.— 33

The Bertha Zinc-Mixes.

o o

z o o

&

o

sua amiivM miJi

The Bertha Zinc-Mines. 515

they are generally of less height. Rarely, the limestone is cav- ernous, with small caves, and narrow openings many feet in length, sometimes overarched with limestone, and, in other cases, having the appearance of clefts or crevices in the rocks, which, however, do not continue in depth, but invariably terminate below in solid limestone near the level of the base of the chimneys. When they are uncov- ered, the limestone chimneys show clearly the same parallel planes of stratification as the bedded limestone on which they rest, and of which they form a part. Fig. 4a, from a photograph of one of the open-cut mines, shows some of the so-called limestone chimneys that have been uncovered. A number of chimneys were blasted down and removed in the process of mining. The lower part of the chim- neys and the cavities at their bases are not shown, being hidden by clay and limestone filled in after the ore had been removed.

When the zinc-ores (silicate and carbonate) are found, they invari- ably rest against the sides of the limestone chimneys, and of the coni- cal, hemispherical, and trough-like cavities at their bases; at times, they cover also the upper surface of the limestone, sometimes en- tirely enveloping the chimneys, particularly the lower ones, but, more often, covering them only partially. Frequently, however, a chimney will be almost or entirely barren of ore, as also will be the cavities and depressions at its base, the occurrence of the ore being quite irregular. The small caverns, clefts and crevices, above referred to, usually carry ore, wholly or in part. Where they are- covered or overarched with limestone, the ores often rest upon the floor, leaving an open space above. In the open crevices, the clay rests upon the ore, but, as before mentioned, the occurrence of these- caves and crevices is quite exceptional.

The ore upon the limestone varies in thickness from a few inches, to many feet. The greatest observed thickness, normal to the sur- face of the linaestone, has been 40 feet between the chimneys.. Nor- mal to the sides of the chimneys, a thickness of between 5. and 10 feet occurs frequently, but the average thickness is less than 5>feet. The section. Fig. 6, shows the usual occurrence of the ore..

The ore-bodies consist of hard and soft ore, and are won. simply with pick and shovel. The hard ore occurs throughout the mass of the ore-body in all sizes, from small grains up to masses of sev- eral tons' weight. Occasionally, block-holing and j)Owder. are re- sorted to, to reduce the large masses to a size convenient to handle, but, commonly, wedging and sledging answer. These boulders,, small and large, ofln have a concretionary structure, like ''pot-

The Bertha Zinc-Minb9.

ore/' with beautiful incrustations and crystallizations of zinc-ore in vugs. An irregular, shelly, or honeycomb structure is common, with cavities and cells filled with gray, yellow and brown powdery zinc-ore having the api>earance of wood- ashes and pulverized yellow ochre. When the formation is not compact and small cavities oo-

f. ScEtllirjcr PuurU

I. Incilnfri flaic i CvTlif* S

V ScliwJ f[n

. .Opfn Cm yUaimf

TTamTrmcki

rig. 4

ur in it, the ore facing these cavities is covered with minute bril- liant crystals, and is often drusy in appearance. Slab-ore is common where cavities are larger. Rarely, the hard zinc-ore clings to the limestone in sheet-form, and is won by gadding. Frequently, this solid sheet of ore is found several inches away from the limestone,

The Bebtha Zikc-Mines.

Si

B

u

o

The Bertha Zinc-Mines. 619

and the space between it and the limestone is filled with loose, powdery silicate and carbonate, while it is not unusual to find a thin sheet of sandy material next to the limestone, evidently a product of the decomposing siliceous magnesian-Iimestone chimney. Beautiful milk-white incrustations of silicate of zinc, like a thick enamel, with surfaces covered with mammillary and delicate stalactitic forms, occasionally occur between series of sheet-like layers of solid zinc- ore, while often the spaces of this sheeted structure are filled with pockets of fine, light, rich ores in honeycomb-form.*

The various and wonderful forms and appearances of these ores cannot be farther dwelt on here, but they are worthy of separate examination and description. I have briefly alluded to them and to their mode of occurrence, simply to establish or invite the sug- gestion of a theory of their origin, which, at Bertha since the open- ing of the mines, has been an interesting theme of speculation among those who have examined the deposits. The ore-l)ody is entirely distinct between the overlying clay and the underlying limestone. The material in which the grains, pebbles, boulders, and other forms of the zinc-ores are directly contained, forming collectively the ore- body, is of a clay-like nature; it is both hard and soft, and is dubbed by the miuers "hard buckfat," and "soft buckfat." Strictly, this is a lean zinc-ore of great purity, too low in zinc to be profitably used in tha present practice of smelting ; it is separated from the richer ores by roughly hand-sorting the larger masses in mining, and by subsequent washing and jiggiug. It is now stored in dams be- low the dressing-works for future treatment for " blue powder/' or oxide, previous to its conversion into spelter. The soft buckfat dis- solves in water by a violent washing operation ; the hard buckfat is of a finegrained, brittle structure, like common chalk, and does not dissolve, but, being of lower specific gravity than the hard ores, it is largely separated from them in crushing and jigging.

The clay that invariably exists as an immediate covering, or blan- ket, upon the ore- bodies, is extremely compact and unctuous. The line or plane of demarcation between this clay and the ore-body is not always visible, but usually there has been a movement of the yielding clay upon the firmer ore-body, resulting in a distinct parting or cleavage-plane, and when this is not readily seen, the miner's pick or shovel determines the distinction as unerringly by "feel" or sound, singly or together, as it would the difference be- tween clay and sand or gravel, when thrust into them. Outside of this peculiar clay is the red or brown clay forming the general sur-

The Bertha Zinc-Mines.

face of the country. This apper clay is the gaDgue, or home, of brown Hmonite-ore, which it carries to a greater or less extent at the Bertha mines, and very abundantly elsewhere in the New River-Cripple Creek mineral-region. The section A-B, Figs. 7 and 8, shows the relation of the series, and the forms assumed by its members. They are approximately concentric. Slickensides in the clay, as smooth and polished as in the walls of hard-ore bodies, are frequent near the chimneys, and where cohesion has been overcome they are noticeable in the ore-bodies, and are very common in the clay near barren chimneys. They here indicate a slow settlement and movement of the clay mans, as the underlying limestone gradually decomposes. It is noticeable, that where ore-bodies are the largest, the chimney

Rg. 5.

6000 7000

ENLARGED SECTION 8HOWINO RELATION OF ZINC-ORE TO THE LiME8TONE6 AND CLAY.

which they cover, or to which they are attached, is very seamy and porous, and presents the appearance of having had soluble constit* uents leached from it. Almost invariably the barren chimneys pre- sent smooth, hard surfaces, like cut stone. The exceptions are, per- haps, too many to establish a rule, but the difference is marked.

As a rule, the limestone in the chimneys, and immediately under- lying the ore, is barren of ores of any kind. This has been proven by the breaking up and removing of many chimneys, by many drifts through the bases of chimneys where drift-mining was carried on, by breaking into the limestone at the bottom of crevices, cavities, and depressions, and by general tests with hammer and drill.

The Bertha Zinc-Mines. 521

In a very few places zinc-blende has been noticed in the sides of chimneys or of depressions, covered with the silico-carbonates evi- dently too meager in blende to weather rapidly, yet andeiing an ap- parent slow change. So far as explored, such bodies have not extended far in depth, nor has their lateral continuance been observed. They appear to be remnants of bolated bodies, occurring in certain hori- zons of the stratification.

Apparently (to the writer at least) the zinc-ores mined at Bertha have been derived by slow weathering and leaching from zinc- blende contained in very irregular bodies or deposits, through a depth of a hundred feet or more of strata of the magnesian limestone, the blende being formed in situ contemporaneously with the lime- stone, by some such process of nature as is indicated by Dr. Sterry Hunt, in his Chemical and Oeologioal Essays, second edition, chap, xii., " The Origin of Metalliferous Deposits."

Why the silico-carbonates are found only contiguous to the lime- stone, has never been satisfactorily explained, nor why, as a rule, the chimneys and the immediately contiguous limestone, should be barren of blende. It has been suggested that the barren rock im- mediately underlying the silico-carbonates as mined, retards weather- ing, and that underneath, at moderate depths the unchanged blende- ores will be found. Explorations to test this ground have been started. The theory of the contemporaneous formation of sulphurets with the limestone, would suggest that blende-ores are to be sought for not only under the silico-carbonates, but also in the downward dip of the strata which carry them as a weathered outcrop. Ex- plorations with the diamond-drill to test for blende on the dip, out- side of the area of silico-carbonates have been started, but at the time of writing have proved only that the inference is true as to the iron-ores, the sulphurets of which, in bands and bunches in the solid limestone, have been penetrated at or near their estimated horizon in the stratification, while the zinc-blende ores have not yet been penetrated."*"

The mining of zinc-ores has thus far been confined to a surface- belt or strip 1500 feet wide extending across the tract in the

Since writing the above, a test-abaft in limestone that was being sunk just out- side the boundary of the silico-carbonates on the dip, to look for a continuance of a mall streak of blende that had been discovered nnder 4 feet of silico-carbonates in the top of a chimney, entered a body or a stratum of sine-blende at a depth of 57 feet below the surface, and continued in it 8 feet and was then still in blende. This seems to confirm the idea that the blende-ores should be sought for on the dip of the strata which carry the silico-carbonates at their outcrop.

The Bertha Zinomines.

z CO

a. o

?

7i O

The Bertha Zcnc-Mines. 623

direction of the strike of the limestone, south 35° to 40° west. Outside of this belt, to the southeast, on the Bertha property, as also on the Manning and Squier property to the east and northeast, a second belt of similar ores is known to exist, but has not yet been explored sufficiently to determine its extent. It is parallel with the first strip, and at a slightly lower geological horizon, unless, as may be the case, the strata have been faulted in a line nearly parallel to the strike, and this second belt occurs in a down>throw from the first belt. As the rock here is everywhere covered with clay, ob- servation has not yet been sufficient to determine whether or not such a fault exists.

2. Methods of Mining,

The method of mining first pursued at Bertha was the usual one practiced in the district in mining limonite-ores in clay, known as " open-cut" mining; with the difference that in digging iron-ore, the dirt or clay is only occasionally stripped or separated from the ore- l)earing zones and put in spoil-banks as ''deads;'' generally it is dug with the ore and goes with it to the washer ; whereas in open-cut mining for zinc, the enclosing clay is first removed as "deads" to spoil-banks, and the zinc ore-bodies are then dug and sent to the washer, the process becoming one of " stripping." While this pro- cess was confined to the low ground, and to the slopes of the hills where the clay-covering was comparatively light, the ore was quite easily won, the chief difficulties occurring only in excavating the deep cavities between the chimneys, below the working- or hauling- level ; in such places the material was at first cast up in stages, some- times from depths of from 30 to 40 feet, and later it was hoisted out by steam-derricks. The occasional filling of the pits with rainwater, mud and sliding clay, was a drawback to this system of mining, as were also the interruptions to a steady output by stormy or inclement weather, and the occasional intervention of barren areas of ground that had to be excavated before ore was again reached, and the blasting and removal of many chimneys. As the depth of the strip- ping was increased by advancing into the slopes of the hills, the proportion of " deads " to ore was also largely increased, and the costs as well as the difficulties assumed serious proportions. Fig. 7a is a view of the working-level under this system.

To obviate these difficulties, but mainly to secure a steady and suffi- cient output for the furnaces, as well as to lessen the costs, the under- ground system of extraction was started in the fall of 1889. The hills

The Bebtha Zinominb3.

t-iy. 7. SECTION THROUGH A-B SHOWING OCCURIICNCC OF ZiNC ORC IN RELATION TO THE LlMCSTONC AMD ClAT THE POSITION or SHArTS AND METHOD OF HOISTINO ORE.

m

A

U

:n\

Ejcpcanation:

Plan

Howino

System Of Shafts And Drifts.

Bertha Zinc Mines.

Seal* of Feet

0 10

; 14 OTE: I>rift hown in SoIiaT Llie* indlemte Che Generally Upper Working '"

' .. wDotud " " ' P J? Lower '

The Bebtha Zik0-Mine8.

£

s

"3

The Bertha Zikc-Mines.

a o

a

The Bebtha Zinc-Mines. 529

were entered by timbered drifts in clay from the previous outside working-level, and the tram-tracks were continued in them from the outside. When the limestone chimneys were encountered, the drifts followed their contour, where practicable, and where it was not practicable to do so and preserve the necessary alignment for tracks, the limestone was cut into as much as necessary, and the chimney was encircled with a timbered barrow-drift to win the ore. A second timbered drift was run upon the one below, and thus, one after another, drifts were carried around and over the chimney, all the drifts above the lower one holding the " deads." The ore was milled through chutes, or passage-ways, to cars or barrows in the bottom drift. Where the thickness of the ore-bodies exceeded the width of the drift, additional drifts were carried outside the first, until the limiting clay was reached and all the ore extracted. The ore below the working-level was extracted by sinking and timbering

SECTION THUOUGH C-D OF FIQ, fl,

t'CTHOO or TIMBZRjrG ORIFTtt

a small shaft, similar to a winze, near the chimney till the bottom of the cavity or depression was reached ; timbered drifts were then carried around the cavity, similar to the upper timbering, and the ore was hoisted by windlass through the shaft or winze.

Later, the set "-timbering around and over chimneys was modi- fied, especially where the ore was less than three feet thick ; in such cases a small timbered raise was made to the top of the chimney, and the ore was removed from the top first, thence the excavation was carHed down the sides in sections, the clay being temporarily supported by head-pieces of slabs or lagging held by a small timber strut, in a species of mining by " withdrawing." The main drifts were strongly timbered, in order to keep them open as long as needed, while all other drifts were allowed to collapse, which they did some weeks or months after the ore had been exhausted, by the gradual but persistent " squeeze " of the tenacious clay. The chief difficulties VOL. XXII.— 34 T

The Bertha Zinomines.

Fig. 10.

Fig, 11.

THE BERTHA ZrNC-MINEB. 631

enoouDtered in this system of drift-raining without machinery were the eventual long tramming, as the work progressed, necessitating considerable re-timbering of main drifts, and the occasional poor ventilation of the shafts or winzes. As the work could be conducted steadily in all kinds of weather, and as costs were favorable, the re- sults were encouraging.

In the meanwhile, the sinking of some test-shafts from the sur- face to the limestone, on a distant portion of the tract, suggested their utilization for exploiting the ore. By a studied arrangement, they might be sunk to the lowest depressions in the underlying lime- stone, and thus serve to overcome the principal diGBculties that had been encountered in drift-mining.

These test-shafts are common in the ore-bearing clays of the South. They are circular, and, as sunk at Bertha, have an average diameter of feet, which dimension bears a close relation to the horizontal extent of a miners spinal column when in position for digging in one of these shafts. With a pick and shovel fitted with special short handles, and a bucket to suit, a miner, with his anatomy disposed in the shaft so that his legs form the vertical, and his spine the horizontal limb of a right-angled triangle, can, with his heli)er oper- ating a primitive windlass at top, sink a shaft feet in diameter, at Bertha, to depths of 100 feet and more for 22 cents per linear foot. The application of these shaft:s to mining Bertha zinc-ores seemed perfectly feasible by using a removable, cheap lining to preserve the form of the shaft, and by directing the shaft-location by means of underground exploring-drifts and simple surveys, so that they should be sunk in each one of the principal basins or cavities that occurred between the limestone chimneys. This stem, named from the form and diminutive appearance of the linings the pipe-shaft" system, was approved by the managing directors, Messrs. R. T. Wilson & Co., and was introduced in the summer of 1890, and since that time has successfully furnished all the ore required to fully supply ten Welsh-Belgian furnaces of 140 retorts each, working at Pulaski exclusively on Bertha ore.

The iron shaft-linings devised and used are shown in detail in Fig. 10, and require but little explanation. Where the clay stands intact, as is usual till the shaft reaches limestone, and its correct location is thus proved or established, the segments for each section are lowered one by one and bolted together in place, or, if the clay is inclined to flake and fall, the sections may be completed at the top and lowered as required, thus protecting the miner from falls of clay, and follow-

The Bertha Zino-Minbs.

ing the excavation as it progresses. As the first cost of these linings is somewhat great, being $4 per linear foot at Bertha, although they can be used many times, and the cost chargeable to each shaft is thus reduced, only enough lining for three shafts was at first used, and this lining has been continued in satisfactory use, especially in weak ground and at temporary shafts.

On a(K;ount of cheap lumber and for convenience, it was found very satisfactory to make linings of 2-inch oak plank, the circular shafl being squared to 3-foot sides, and lined with a set made from a

Fig, 12,

DCTAIL OF DRIFT TIMBERING ScaTe of St: .bHd==JMHiHMb

t

14-foot plank by cutting the same into four equal lengths, and fram- ing the ends by simply halving or boxing, as shown in Fig. 11. The lininpj of the shafts by contract, including the squaring and hoisting, costs 25 cents per foot, and the plank costs 42 cents per foot of shaft. Hence, the total cost of sinking and lining amounts to 22 -f 26 + 42 89 cents per linear foot of shaft. The lining- planks, like the iron linings, are reclaimed when the shaft is aban- doned, and many planks are used a second and even a third time, and the broken ones are utilized in underground timbering.

In the pipe-shaft system of mining. Figs. 7 and 9, the main drifts

The Bertha Zinc-Mines.

for exploiting the ore are of much smaller section, and less expen- sive, than the main drifts in drift-mining where tram-cars are used. The posts are spaced but feet apart on the sill and feet apart on the cap. The general level of the exploring drifts is the general level of the rims of depressions between chimneys. As the drifls

Iron Bucket

follow the limestone laterally they are very crooked. Special, deep mine-barrows, made of boards, are used in these drifts, the extreme wheeling-limit rarely exceeding 100 feet. No limestone is cut in this system of mining. The greatest depth thus far required for a

The Bertha Zinomines.

shail has been 126 feet. The average depth of working-shafts is about 90 feet. The system of timbering, aside from that of the main drifts, is about the same as in drifl-mining, and is shown in Figs. 12 and 14. The ventilation is good, exoept occonally near the top of those chimneys which have no ladder-shafts extending from

the top to the surface ; but on account of the nearness of shafts this is easily regulated by a little bratticing. With incandescent light- ing, that is now being installed, the ventilation will be excellent.

In the shaft-system, portable light hoists and boilers are used, the hoist operating two drums, and all being mounted on a heavy timber

/

BERTHA ZtNOHIKES. 635

frame for moving from place to place. A single brakeman attends the two drums and the boiler, and a single lander attends each shaft. Steel buckets, Fig. 13, 30 inches in diameter and 40 inches deep, are used, holding 1300 to 1600 pounds of ore. A simple tripod of heavy poles, 36 feet long, or a bolted frame of squared timbers in a portable form, is used for the pit-head frame, as shown in Figs. 7 and 76. The buckets are inverted by a gallows-rope or chain, which carries a hook at one end to engage with a ring in the bottom of the bucket, and dump directly into tram-cars. The system is intended to be easily portable and elastic and adapted to unevenly settling ground. As fast as the workings progress, surveys are made with a hanging-compass and are mapped. These surveys show just what ground has been exhausted, and where to locate shafts on the sur- face. The extension of the timbered workings from one shaft to the point where they meet the timbered workings of neighboring shafts, proves that all the limestone has been uncovered between the shafts, and with competent supervision it proves also that the ore in that section has all been extracted.

The ore from the mines is conveyed in tram-cars by steam-loco- motives to the brow of the plateau above the river, and dumped from trestles into storage-bins, as shown in Fig. 6. Underneath the bins runs a small plank trough or flume, 18 inches deep and 12 inches wide, with a semicircular cast-iron lining at the bottom. The flume under the bins is covered with short pieces of plank to retain the ore. At the upper end of the flume, water is introduced from a tank, to which it is pumped from the river, and the ore is allowed to enter the flume by taking up, one after another, the short pieces of plank that cover the flume, it being admitted as fast as the rapid current of water will carry it along the flume to the dressing- works below. By this plan the ores are well washed and freed from ad- hering clay when they arrive at the dressing- works. The usual form of gravity-plane, shown in Fig. 6, was formerly used to convey cars of ore from the top of the hill to the dressing- works, but at present it is used only for taking timber and merchandise up to the mines, and cars of heavy lump-ore down the plane to the dressing- works. The limonite ores which are occasionally mined at Bertha are similarly transferred in a second flume to a separate washing- and screening-plant at the dressing- works.

3. Shipment and Product. As this paper is not intended to describe the treatment of the

Ic

536 THE BEBTHA ZrNC-MINES.

Bertha zinc-ores, I will only say, in a general way, that the ores are dressed by washing and jigging at the dressing-mill, connected with the mines, to as near the grade of a 40-per cent, concentrate as is practicable without too grejtt loss, and the product is shipped to ten furnaces owned and operated by the company at Pulaski. Here it meets the semi-anthracite coal of the Altoona coal-mines, which is used in reducing the ores directly to spelter. The relation of the zinc-mines, coal-mines, and smelting-works, all owned and operated by the company, is shown in Fig. 1.

The following is the average analysis of spelter made by a direct smelting of the ore in the Welsh-Belgian furnaces :

Percent.

MeUlIic zinc, 99.90

Lead and iron, 0.10

Total, 100.00

The Bertha mines were discovered in 1866 by Mr. David S. Forney, a pioneer in the active development of this mineral-region, not from any visible outcrop or float, but in the amateur pursuit of mineral- ogical and geological investigations, suggested by the favorable ap- pearance of the region to which he came from Pennsylvania to pur- sue his profeasion of landscape-artist.

The mines are referred to in a report for gratuitous distribution by the Norfolk and Western Railroad Company, prepared by An- drew S. McCreath and E. V. d'Invilliers, entitled the New River- Cripple Creek Mineral-Region of Virginia,'* published at Harris- burg, Pa., 18S7, from the excellent map accompanying which I have taken many of the topographical and geological features shown in Fig. 1. References to the mines may be found in the Transaclions (viii.,341; x.. Ill; xii., 31 ; xviii., 632), under the head of Bertha Zinc-Mines" and "Falling Cliff Zinc-Mine." The Falling Cliff mine and other adjoining tracts have been merged in the Bertha Zinc and Mineral Company.

Blowing-Engines. 537

Blo Winqenqine8.

BY JUUAN KENNEDY, PirTSBUROH, PA.

(Chicago Meeting, being part of the International Engineering Congress, August, 1893.)

The different types of blowing-engines in use are so numerous that it would not be practicable to consider them all in this paper. I shall therefore only take up briefly a few well known types.

The style of blowing-engine most largely used in this country is the vertical engine with air-cylinder above, cross-head between steam- and air-cylinders, and two fly-wheels, each having a wrist-pin in its hub, or in one arm. This kind of engine can be built cheaply, takes up little room and is very accessible. Its disadvantages are that the cross-head is liable to break, and that putting the wrist-pins in the wheels tends to set up vibrations in them. It is likely that on the whole this type of engine will continue to be built quite ex- tensively.

The same general arrangement has also been used to some extent, but not widely, in horizontal engines. The Bethlehem Iron Company has some very fine blowing-engines of this type, except that they are compound, having the one steam-cylinder replaced by two, side by side. These engines are noticeable not only for the very excellent workmanship on them, but also on account of the weight of the pistons being carried by steam-pressure applied in chambers in the lower side, the steam being supplied to these chambers through the hollow piston-rods. This arrangement, I believe, has always worked well.

Another type of engine which has given good satisfaction is the vertical double-engine, having air-cylinders above, steam-cylinderB below them, and cranks at bottom, the engines being coupled to cranks on the ends of the shaft, placed at right angles to each other, the shaft; carrying the fly-wheel at its center. Engines of this type, except that they are arranged horizontally, are also used to a con- siderable extent. These engines avoid the disadvantage of having wrist-pins in the fly-wheels, and also dispense with the long cross- head with its attendant disadvantages. They give a very uniform pressure of blast and are very convenient for starting. The vertical engines of this type, as compared with the horizontal, take up less

Blowinu-Exoine8.

Blowing-Engines. 639

room, and the wear on cylinders, due to carrying the weight of pis- tons, is avoided. On the other hand, the machine is very high, and there is considerable vibration. The horisontal machine avoids this, iR very accessible and is cheaper to construct. With proper atten- tion, there seems to l)e no serious trouble with wearing of cylinders, so that wherever ground-room is ample, the horizontal double- coupled engine seems to be very suitable. The principal objection urged against this style of engine is that in the event of a break a large machine is disabled, whereas, if two single engines are used, one can keep the works going whife the other is being repaired. This is doubtless correct ; but I think that too much weight is often given to this consideration. With machines strongly proportioned and carefully built, there should be very few stops ; on the other hand, in the case of Bessemer engines, which are starting and stopping at short intervals, the fact that one attendant can handle the double machine is worthy of consideration.

The double-coupled type of engine is also particularly adapted to compounding. As most blowing-engines run under a comparatively constant load, and as the increasing use of water-tube boilers in iron- and steel-works renders it easy to maintain high steam-pressures, I have no doubt that before long compound blowing-engines will be adopted in a large majority of the new plants built.

In looking over the different kinds of blowing-engines, we cannot fail to be impressed with the fact that, in nearly every case, the air- valves are the weak point in the machine. In the great majority of cases the maximum s|)eed of the engine is about half what it could be, if the air-valves could work fast enough. To remedy this fault, several plans have been resorted to. In some cases fairly good re- sults have been obtained by making the valves very light, giving them but little lift, and arranging thera so that they shall seat by gravity. In some cases valves of this kind are so constructed that the air in entering the cylinder is compelled to pass through a large number of very small openings. This is a very objectionable arrange ment, not only on account of the increased amount of friction, but because the air in passing over the metal grids in thin streams will absorb quite a considerable amount of heat from the heads, which, in the case of engines working against high blast-pressures, are made very hot by the heat of compression.

This heating of the incoming air expands it and proportionally reduces the weight of air entering the cylinder at each stroke. I have observed this in the case of an engine, which was so constructed

Blowing-Engines.

k

a

S

P

Blowing-Engines. 54 1

as to cause the air to travel about three inches over the hot metal in thin films about in. thick. Alongside of it was another engine of the same size and make, except that valves were used which allowed the air to pass over about one inch of metal, the openings being of such size that each stream of air was 2 in. in thickncRs. Careful and repeated tests of these engines, when both were in good order, showed that, while the indicator-diagrams were practically the same, the one with the large valves would burn about 10 per cent, more ceke in the furnace: a result which could only be explained on the supposition that, in the ease of the engine with small air-openings, the incoming air, in parsing through the small and tortuous passages in the heads, was heated about 25 degrees C. more than in the case of the other engine. It is plain, therefore, that a blowing-engine should have air-valves which will not only give ample area of inlet-pas- sage, but give this in a small number of good-sized openings.

Figs. 1 and 2 are the plan and elevation, and Fig. 3 is the dia- gram of air-valves and valve-gear, of a compound horizontal blowing- engine, now being constructed by the well-known builders, the E. P. All is Company, for the Ohio Steel Company. The engine is a Reynolds-Corliss cross-compound; steam-cylinders 40 in. x 78 in.; air-cylinders 60 in. ; stroke 60 in., with reheater in intermediate re- ceiver, and is provided with an independent condenser. In general design this engine, as will be seen at a glance, is very similar to the large quadruple-expansion engine by the same builders to be seen at the Exposition. The air-cylinders are so arranged as to draw the air through pipes, which project above the roof of the building, and to discharge it below the cylinders.

The inlet-valve is a plain rotary valve held to its seat by the blast-pressure, which is admitted to the back of the valve by a |)ort Irom the discharge-chamber, and is driven from a wrist-plate. The outlet-valve, as will be noticed, is a triple-ported valve, which is closed at the proper time by the wrist-plate.

The connection between wrist-plate and valve is made by a tele- scopic extensible rod, which pushes the valve shut, but [>ermits the wrist-plate to reverse its motion without pulling the valve open. To the valve-lever is attached a vacuum-pot, which tends to pull the valve open. When the valve has been closed it is gripped by the receiver-pressure acting on the back, holding it against the seat, and remains stationary during the return-stroke of the piston, and also while the piston advances towards it again, until it has compressed the air in the cylinder to nearly the same pressure as in the re-

Blowing-Engines.

ceiver, at which time the pressure on the back of the valve becomes so nearly balanced that the vacuum-pot can move the valve, which is then quickly thrown open. The telescopic connecting-rod is so constructed that a small dash-pot is formed at the bottom of the

tube, to avoid shock should the plunger strike the bottom while the valve is opening or when the closing-motion begins. It will be ob- served that no trip- or releasing-gear of any kind is used with these valves, the holding and releasing being done by friction, controlled in the simplest possible manner by the air-pressure in receiver and

AN IHPROYED HANGINGMX)MPAaB. 543

cylinder. The ootlet-valves are also held against their seats by long flat springs bearing in the center on the back of the valve and at ends on blocks set in pockets at end of the valve. It will be seen from the drawings that these blocks have a clearance of J in. at the bottom, so that if for any cause the valve should be pre- vented from opening at the proper time it will be only forced back from the seat, the opening of in. being sufficient to allow the engine to run at full speed with wrist-plate and vacuum-pot disconnected from outlet-valves. This valve-gear is extremely sim- ple, and practical tests have shown it to work admirably. Thi* engine is intended to run at a speed of 60 turns per minute if nec- essary.

In conclusion, the tendency in designing blowing-engines seem to be in the following directions :

1st. Compounding.

2d. Obtaining valve-gear which will give liberal openings at both inlet and outlet, and which can be operated at a fairly rapid speed.

The latter advantage can probably be best secured by the use of metal valves operated as far as possible positively, which will also* do away with the vexation due to the use of leather,, gum and other short-lived materials.

AN IMPBOVED HAHmNQ COMPASS.

BY GUY R. JOHNSON, LONGDALE, VA. (Chicago Meeting, being part of the International EngineeriBg COBgreas AngtBt, 1893.)

In working brown iron-ore mines on the system employed atf Longdale, namely, stoping from the top down, the usual procedure- is to drive a succession of upraises from the lowest adit to the- highest, or to the top of the deposit, as the case may be. These up- raises are usually driven 120 feet apart, along the strike of the vein.. As the work goes on, they are connected by levels, 1© feet apart (vertically), and then serve as chutes.

A transit is useless in surveying these chntes, and the writer has* for several years made use of a hanging compass. The method of using this instrument, described in a former paper (Trans.y xx., 105), embracing as it does the use of a cord and several strips of wood,

An Impboved Hanging Oompass.

o

u

An Improved Hanginq 00Mpas8. 646

18 well enough in the chutes proper where the total distance to be surveyed is comparatively small but when extended to the levels between the chutes becomes very slow and cumbrous. It was to obviate this objection that the instrument here described was de- vised.

My first idea was to carry another compass of the ordinary type, but the additional weighty and inconvenience of carriage through the steep and narrow chutes caused this plan to be abandoned. After some correspondence with the instrument-makers to ascertain the practicability of the change, the instrument herewith illustrated in Figs. 1, 2 and 3, was ordered.

In these figures, A is the compass swung in gimbals, as in the usual form of the instrument; B and B' are two small levels, sunk into the bottom of the compass-box, one on the N.-S., and the other on the E.-W. line. With these the instrument can be levelled per- fectly. Outside levels would have interfered with the gimbals.

The folding sights are of the usual pattern. In surveying in the chutes these are never raised, unless the line of sight is near the magnetic meridian.

D is a plate into which the cap E of the Jacob's staff screws. This departure from the usual form of Jacob's staff-heads, in which the instrument turns on a spindle, the prolongation of the ball of the ball-and-socket joint, was occasioned by the necessity of having an easily portable instrument. The socket for the usual spindle would have made the carrying-case much too bulky.

The compass having been made fast to the head, revolves on the center F; the head £ turning with the instrument.

In surveying the levels, the screws G and G' are loosened, and the compass is taken out of the gimbals. It is then used with the Jacob's staff, as in the ordinary form of te instrument

With the arrangement above described, which entails the carriage of only one extra piece (the Jacob's staff), the levels can be sur- veyed in about one-half the time formerly occupied, and the porta- bility of the instrument is not affected. Also better speed can be made in the chutes, as the slight extra weight, especially that of the plate D, makes the compass a great deal steadier on the cord.

For further details of the conditions and method of surveying at the Longdale mines, I refer to my paper in the Transactions AlreeAy cited, to which this is merely supplementary. VOL. XXII.— 36

546 Microstructure Op Steel.

Micro Structure Of Steel,

BT ALBERT 8AUVEUR, SOUTH CHICAGO, ILL. (Chicago Meeting, being part of the International Engineering CongresB, August, 1893.)

I. Crystallization of Steel.

The following propositions and corollaries are intended to present, as concisely as possible, some of the evidences gathered while study- ing the microstructure of steel.

Each proposition is accompanied by photo-micrographs showing the stractare of illustrative samples. But it must be understood that the special instances thus chosen for illustration are by no means the only ones on the examination of which the propositions are based. On the contrary, they have been taken almost at ran- dom among a great many, in some instances more than a hundred, similar cases.

Propobition I.

A slow and undisturbed cooling from a temperature x or higher produces crystallization.

Plate I., Plate II., Fig. 1, and Plate V., Figs. 1, 2, 3 and 4 show the crystalline structure of several pieces of steel. Plate I. reveals this structure in the center of a slowly-cooled Bessemer ingot, 1.5 feet from the top. Plate II., Fig. 1, is from a small 2-in. square test- ingot of the same heat, slowly cooled. Plate V., Figs. 1, 2, 3 and 4 show the structure of different parts of steel rails. The original photographs were all magnified 70 diameters, and have been reduced in engraving to 63, or to 55 diameters, as shown in the plates.

Proposition II.

Undisturbed cooling from an initicd temperature lower than x is not accompanied by orystaUization,

Plate III., Fig. 5, shows the structure of a piece of rail-steel forged from a white heat until it had fallen to dull-red. The struc- ture is almost amorphous. Practically no crystallization has taken place during the subsequent cooling.

Corollary. — Pieces of steel finished ai a temperature lower than 7 do not take a crystalline structure.

MICROSTRUCrUBB OP STEEL. 647

Proposition III.

The temperature x varies with the chemical composition of the steel. Each impurity {at least, carbon and phosphorus) lowers x, although in widely different degrees.

Plate II., Figs. 1 and 2, show the structure of two small test- ingots of the same size and shape cooled down from the same initial temperature. They have practically the same composition except as regards carbon, the former being a rail steel with 0.40 per cent. C, the latter a soft steel with 0.11 per cent. C.

Plate Il.y Figs. 3 and 4, exhibit the structure of two pieces of steel similarly treated and varying in composition only by their per- centages of phosphorus, which are respectively 0.145 and 0.063 per cent. It will be noticed that the high- phosphorus and high-carbon steels have respectively a much larger grain than the low-phosphorus and low-carbon. The crystallization of the former has kept on (the grain has continued to increase) long after it had ceased in the latter.

Corollary I. — Carbon and phosphorus and probably ail impu- rities, at least when present up to a certain amount, increase the size of the grain. Carbon, by increasing the amount of carbide of iron (which is the hard constituent of steel) confers hardness and strength on the metal ; by increasing the size of the grain it causes decrease of ductility and eventually brittleness. This, of course, can be di- minished to a great extent by proper heat-treatment

Corollary II. — The purer the steel, the higher the finishing-tem- perature can be wiihovi causing coarse crystallization.

Properly conducted experiments might teach us what this tem- perature X is for various chemical compositions ; and the knowl- edge would be extremely valuable. It would tell us at what temperature a steel of a given chemical composition should be fin- ished, in order to avoid an objectionable crystalline structure.

Although this is certainly a complex problem it is not, we believe, an insoluble one. In the latter part of this paper we have outlined a method of working which, so far as it has been tested, has given gratifying results, and which may furnish the means of getting the desired information.

Proposition IV. The higher the initial temperature from which the steel is allowed to cool undisturbed, the larger the grain for a given composition. Compare Plate I. and Plate II., Figs. 1 and 2. Plate V., Figs. 1

648 Micr06Tbugtube Op Steel.

and 2, show the structure in the center of the head of two rails fin- ished respectively hot and cold. They have the same section and were rolled from the same heat.

The influence of the initial temperature on the size of the grain is here well illustrated.

Proposition V.

The slower the cooling , the larger the grain for a given composition.

Compare Plate I. and Plate II., Fig. 1. The initial temperature was the same. Both ingots are from the same Bessemer heat. The smaller grain of the small test-ingot is due to its more rapid cool- ing.

Corollary (to Propositions IV. and V.). — Finished pieces of steel will have a coarser grain in those places which have been finished hot- test, and where subseqtient cooling has been more gradual.

Proposition VI.

The size of the grain is independent of the amount of work the metal has received.

If two test-ingots taken from the same heat, one say 4 inches and the other 2 inches square, are forged down to say one inch square, the bar reduced from the larger ingot will show the smaller grain (Plate III., Figs. 1 and 2). At first it might be inferred that the smaller grain is due, partly at least, to the greater amount of work received by the larger ingot, but we believe it is caused only by the lower temperature at which that bar was necessarily finished ; for if instead of being left to cool slowly after forging, from their respec- tive temperatures, both bars are similarly heated, say to a bright red, and then slowly cooled, they will take the same structure (Plate III., Figs. 3 and 4.)

This, at all events, is an argument in favor of the non-existence of a " special effect of work.'* Work as such seems to have no efiect on the grain of the steel ; and it is hard to understand how a treat- ment which does not alter the structure of a metal can influence its physical properties.

II. Microstructure op Steel Rails.

A polished and etched section of a steel rail when examined under the microscope does not by any means reveal in all its parts the same structure. This heterogeneousness of structure is due to the different temperatures at which the various parts of the rail

Microstructure Of Steel. 649

leave the finishiDg-roIlB and to the aneqnal rate of their subsequent cooling.

The propositions formulated in the first part of this paper should enable us to foretell such variations in the structure of a rail.

The outside of the rail, leaving the rolls coldest and cooling quick- est, will offer less chance for crystallization, and should therefore show a smaller grain than the inside, which being the hottest and cooling down more slowly will favor the crystallization of the metal, the size of the crystals or grains reaching its maximum in the center of the head. It might also be inferred that the smallest grain will be found at the extremities of the flange, since this is always the coldest-finished region. Proceeding from the outside to the inside, we shall find the tendency towards coarse crystallization gradually increasing.

The same propositions tell us, moreover, that for a given chemical composition, the higher the finishing temperature of the rail and the slower the cooling, the more pronounced will be the crystallization. It follows also that the heavier the rail the more crystalline will be its structure; for it will generally be finished hotter than a lighter one.

Plate rV. shows the structure of twenty-one different portions of the transverse section of a rail. Facing the plate is an outline of this section, with corresponding numbers indicating the spots where the photographs were taken. The enlargement is 33 diam- eters. The rail was finished fairly hot. Its weight is 80 pounds to the yard. As we proceed inward towards the center of the head we notice how the structure, at first very close, gradually becomes coarsely crystalline.

The photographs of Plate V. show these different structures per- haps more plainly. They were taken at the spots marked 1, 2 and 3 in the accompanying outline of the section.

It is a necessary consequence of such variations of structure that the physical properties of the different parts of the rails should vary accordingly. As might be expected, the crystalline structure of the center of the head will show less power of elongation and of reduc- tion of area, than the closer structure of the side of the head and especially of the extremities of the flange. Test-bars were cut from three different locations and submitted to tensile strain. The results agreed remarkably well with the microscopical evidences.

Thus it appears that, so far as physical properties are concerned, a steel rail is a somewhat heterogeneous body. This is the more

Miceosthuctcjre

Of 8Teel.

appuent .n raxls finished uodaly hot In colder-finished ntils the

andtil rTu""!'" of the head, the side of the head and the end of the flange rpectlvely. If we compare them witll the oorrpondrng photographs of the hot-finished nul (Pkte V, rS' i:i K ! in each case their grain is smaller.

finlZtn? 'r '''' 2 and 3, of both hot- fad cold-

Pki/sic Teas of Hoi. and QM-rmished Bail,.

Test-bftre, cut from :

Size of grain Inl q. mm. magni nified 100 diam.

Center of head...

Side of bead I 2I8

End of flange 52

S'

Tennile strength. Honffatten. I Keductlon. Lb8, persq. inch Pr.ct. length.! pSwot

86 99,900l 100,900 15 75 100,000. 103.900; 19 35 102,900 105,60o' 22J

?.

Am

U3

Ca

cd

S

s

20

32}

I fo&" T of elongation and of .Eduction.

deflLVon rhictTh structure of a rail that the

sbow ng hao those from the side of the head, the web or theflanee-

igh a finishing temperature ujr

.other consequence of this variation in the structure is that

MICR06TRUCTURE OF feTEEL. 551

when rails are in use, some portions, more yielding, will not bear their fair share of the stress, but will throw it partly upon the other more resisting portions, thus reducing the power of endurance of the rail as a whole.

It seems, then, that the ideal rail should have as homogeneous a structure and as small a grain as possible. Theoretically we should get nearer to this result by finishing cold (not so cold, however, as to produce the brittleness due to cold-rolling) and cooling quickly ; or by re-heating the rail to a temperature from which on subsequent cooling the metal does not crystallize.

III. Some Experiments on the Relation Between the

Physical Properties op Steel and its

Microstructure.

When a piece of steel has failed to meet expectation, whether under preliminary test or in actual use, we generally turn to chemi- cal analysis for an explanation of the failure. The investigations may occasionally include the cutting and testing of some test-bars, and the etching of a section of the deficient piece, but it rarely amounts to more, at least as far as rails are concerned.

Seldom, if ever, do we stop to inquire whether the heat-treatment may not have been defective. But this is not a factor that should be disregarded in view of its all-important effect on the structure and consequently on the properties of the metal. No doubt our neglect of it is chiefly due to the lack of means at our disposal for ascertaining with sufficient accuracy what the past heat-treatment of a finished piece has been. Yet heat-treatment leaves its unmistak- able impress on the structure of the metal. Jointly with the chemi- cal composition it regulates the grain. If we could find to what extent each of these two variables contributes to the formation of the grain, then the final result being known and one of the factors (the composition), the other could easily be inferred. It is here that the microscope seems destined to become an extremely important in- strument for the steel-manufacturer, since it furnishes us the means of estimating that final result, the size of the grain. In other words, from the size of the grain of the steel under consideration and its chemical composition, we could infer what its heat-treatment has been, and this would indeed be a very valuable piece of information and might explain many failures which chemical analysis had left unaccounted for.

In the following pages we use the names given by H. M. Howe to

552 IflOROSTRUCrUBE OF STEEL.

the different constituents of steel described by Dr. Sorby. It will, perhaps, be well to briefly recall their meaning. As far as the micro- scope has been able to ascertain, all steels not quenched seem to be made up of three constituents, namely: ferrite, cementite, and pearly te. Ferrite is iron free from carbon. Cementite is iron with cement carbon, probably a carbide answering to the formula FejC; it is the hard constituent of steel. Pearlyte, when examined under high power (at least 300 diameters), shows itself to be a mixture made up of very thin plates alternately of ferrite and cementite in the proportion of two-thirds ferrite to one-third cementite. It has a pearly appearance, whence its name. It seems as if during slow cooling the ferrite and cementite combine as far as possible, leaving an excess of ferrite or of cementite, as the case may be.

We have seen that all steels crystallize when allowed to cool un- disturbed from a sufficiently high temperature. This crystalline structure is composed, roughly speaking, of grains of pearlyte sur- rounded either by ferrite (in the case of soft steel or steel of medium hardness) or by cementite (in the case of harder steel). A section cut at any angle through the metal will exhibit what we may call a net-work structure, the pearlyte forming the meshes, the ferrite or cementite the net-work itself.

In the case of very soft steel the grain is very small and the proportion of ferrite is considerable, but we may still compare the structure of a section cut through such a steel to a net-work with small meshes and an exaggerated net.

Now, the physical properties of a sound piece of steel depend :

1. On the proportion of pearlyte and ferrite, or pearlyte and cementite, present in the metal (and this is governed solely by the percentage of carbon) ; and

2. On the size of the grain, which, in turn, is a result of the chemical composition and heat- treatment.

In the following experiments we have endeavored to ascertain the relation between the size of the grain and the properties of the metal.

We started with the intention of considering only these two vari- ables, namely, size of grain and physical properties, hoping to be able, by the plotting of a sufficient number of cases, to get some curve from which we could foretell with fair accuracy the properties of a steel, the size of the grain of which was known. We soon found, however, that it was necessary to take into consideration a third variable, namely, the percentage of carbon, on which depends

Section showing the Microetructure of a Slowly Cooled Bessemer Ingot, Magnified

63 Diameters.

Plate II.

Fig. 1. — Untreated Bessemer rail-steel. Structure of small test- ingot, 2 inches square, slowly cooled. Carbon, .40 per cent.

Fig. 2. — Untreated soft Bessemer steel. Structure of test-ingot, 2 inches square, slowly cooled. Carbon, .11 per cent.

Fig. 3. — Bessemer steel. Hammered bar. Phosphorus, .145 per cent.

Fig. 4. — Bessemer steel. Hammered bar. Phosphorus, .063 per cent.

Plate II.

r/0. /

no. 2,

rio.3

Sections showing the Microatriicture of Bessemer Steel, Magnified 5f5 Diameters.

Plate III.

Fig, 1. — Bessemer rail-stee]. Small test-ingot 2 inches square, forged down to a bar 1 inch square, and slowly cooled.

Fia. 2. — Bessemer rail-steel. Test-ingot, 4 inches square, forged down to a bar 1 inch square, and slowly cooled.

Fig. 3. — Steel of Fig. 1 reheated and slowly cooled.

Fig. 4. — Steel of Fig. 2 reheated and slowly cooled.

Fig. 5. — Bessemer steel. Bar hammered from a white to a dull- red heat

Plate III.

F7Cl/

nQ.s

rtQ.3

r/c. Ik

Pig B

Sections showing the Microstnicture of Bessemer Steel, Magnified 55 Diameters.

Plate IV.

Structures of different parts of the same rail-section. The pho- tographs were taken at the places marked with the corresponding numbers in the accompanying rail-profile.

SectioDs showing the Microstructure of Different Parts of the Same Bessemer Steel Rail-Section, Magnified 33 Diameters.

Plate V.

Figs. 1, and 5 are taken from a hot finished rail, respectively in the center of the head, the side of the head and the flange, at spots numbered 1, 2, and 3 in the accompanying rail-profile.

Figs. 2, 4, and 6 are similar photographs taken from a a>ld, finished rail.

Plate V.

FlO.t

F1Q.2

Fig. 3

no.y-

fJG.S

F/ae

Hicrobtbucture Of Steel. 653

the relative arooaDts of pearlyte and ferrite (or eementite) present This will be easily understood, for, although two steels may have the same grain if they contain respectively say 60 and 10 per cent, of ferrite, their properties must be very different. The grains of the former will be farther apart, the intetpace being filled with soft ferrite, while only a narrow net of ferrite will separate the grains of the latter.

Consequently, a constant relation between the size of the grain and the quality of the metal can exist only, if at all, among steels with the same, or nearly the same, carbon-content. Hence the neces- sity to classify our results into groups corresponding to each percent- age of carbon, or at least to each variation of 0>06 or 0.10 in that percentage. For instance, all the rail steels containing from 0.30 to 0.40 per cent, carbon will form one group, while the soft steels, with from 0.10 to 0.16 per cent, carbon will form another.

But, before going any farther, it will be well briefly to de- scribe here our method of proceeding to ascertain the size of the grains.

Our outfit for that purpose consists of a microscope and its acces- sories, a camera lucida (Abbe's improved design), and a planimeter. The sample whose grain it is proposed to measure is carefully pol- ished and etched in the usual way. The microscope is kept in a vertical position. In the middle of our table, by the side of the microscope, is a round board, carrying a sheet of white paper and the planimeter, and capable of being raised or lowered at will by means of a screw-motion. We are thus enabled to keep a constant distance between the eye-piece and the paper, securing for specimens of different thicknesses a uniform enlargement. The specimen and paper being properly illuminated, it will be found comparatively easy, after a little practice, to follow accurately the outline of any grain with the steel point of the planimeter, thus ascertaining its area. Or, if found more convenient, the outline of the grain may be first traced with a pencil on a sheet of paper, and the area after- wards measured with the planimeter.

In many cases the lines of demarcation of some grains appear, in certain places, very faint, or their course doubtful ; but we can generally complete them with the pencil, without introducing much error in the figuring of the size of the average grain.

What we call here grains, are, of course, the meshes of our net- work. These are the sections through the grains of the steel ; and for the sake of comparison, the average area of such sections is just

554 MICEOSTRUOrUBE OF STEEL.

as good as the average surface or the average volume of the grain itself.

Such enlargement is used as is found desirable to enlarge the grains to a convenient size. Of course, in dealing with small grains, such as those of very soft steel, or of harder steel finished very cold, a much higher magnifying power is needed. In some instances as high a power as the one furnished by a -inch objective will be found necessary to enlarge the grains sufficiently. In plotting our results we have reduced all our areas to a uniform enlargement of 45 diameters.

With teel of average hardness, the amount of ferrite is very small ; the grains are in close contact and the work of measuring can gen- erally be much shortened by nlecting altogether the area occupied

Fig. 1.

Grains of Moderately Hard Rail-Steel, Magnified 45 Diameter?.

by the net-work itself. All that is necessary then is merely to follow with the planimeter the outline of the space including all the full grains visible in the field of the microscope, to count them out and to figure the average area.

As an illustration we have drawn, in Fig. 1, the grains of a steel rail, magnified 45 diameters. All the visible full grains are in- cluded inside the heavy line. The area limited by that line is found to cover 2250 sq. mm., which, divided among 45 grains, gives 50 sq. mm. for the average, or 247 sq. mm., when magnified 100 diameters.

HICBOBTRUarUBE OF STEEL. 655

In dealing with soft steel, this method is no longer possible. It is then necessary to measure individually a sufficient number of grains. Fig. 2 shows the grains of a soft Bessemer steel containing 0.11 per cent, carbon, magnified 225 diameters.

By measuring each of the 42 individual grains, we find a total area of 334 sq. mm., which gives an average size of 8.1 sq.mm., or if reduced to an enlargement of 100 diameters, 1.6 sq. mm.

Most of our test- bars were cut f inch square and 20 inches long. A thin section was cut at one end of each test, polished, etched and its grain measured. The tests were then subjected to tensile strains and the maximum load in pounds per square inch and the elonga- tion and reduction per cent, were recorded.

At this early stage of our experiments, we can only publish the results we have obtained for rail-steel. These are shown in the accompanying diagram. Fig. 3. Even here we must stop our curves when the grain reaches 226 sq. mm. since for larger grain the num- ber of cases plotted has been too small to determine the curve.

Fio. 2.

Grains of Soft Bessemer Steel, Magnified 225 Diameters.

The lines of reduction and elongation follow remarkably well the changes in the size of the grain ; both falling rapidly as the grain increases, and the reduction diminishing more rapidly than the elongation.

- The size of grain does not by any means so much affect the tensile strength. The fall of that line is very gentle. The grain increas- ing from 35 to 221 sq. mm., t.e., six-fold, the corresponding de- crease of tensile strength is only about 10,000 pounds.

Will the tensile strength curve continue to fall with further in- crease of the grain? If we consider the only two cases plotted and

Hicb08Tbuctube Of Steel.

TEN8ll£ tTRENOTH, 186. tfU SQUAHC.INOH. t t It Z f t t 2 8 S t S S I I i S

I I I I I I I I I I I I I S I f 1 I I

s s

I i i

I s s

Peh Cent Of Elongation And Rcouction.

]

17 i i- L

- -rt y'HT! - - "H

f

eo* L itltS it i it

ISr i 7 ± - :

J? / / fl

5 - J L i -L J- w - -

o 5 / y o

WPJil it '.L it

O O " / / /

-? isi -4 tfli - i."' i - -

He ' u : :

s

B

1

A L L ' ' " '

It'

4,

s (

f . .

K 1—

B J ' J '

1 , "

B

% -

K 1 , . ,

t' ' 1 1 1 1 1 ' 1 1 1 1 ' 1 1 ' 1 " ' 1 ' ' 1 ' 1 ' 1 1 1 1 1 1 ' 1 ' 1 " i 1 ' 1 1 n ' "

laCBOfiTBUCrURE OF STEEL. 657

corresponding to a grain of 338 and 376 sq. mm. respectively we notice a decided increase of tensile strength ; but of course many more steels with such grains will have to be tested before we can venture to show any conclusion.

These results show that there is a constant relation between the size of the grains and the properties of the metal, and what that re- lation is. They constitute, however, only the first step towards a very extensive series of experiments necessary to accomplish our purpose. We must next endeavor to determine the extent to which the grain is affected by different finishing-temperatures and by each unit of the various impurities, carbon, silicon, sulphur, manganese and phosphorus.

If the plotting of a sufficient number of cases should reveal a con- stant relation between these variables and the size of the grain, it will be seen at a glance what*extremely useful results would follow. We should be enabled to detect improper heat-treatment. Tables could be constructed from our curves, which would give the size of the grain corresponding to a given chemical composition and a given heat-treatment, and consequently the physical properties of a steel produced under these conditions.

Such tables would help us to solve the two very important ques- tions : What heat-treatment will best suit a given chemical compo- sition, so as to get a certain desirable structure ? and what chemical composition will best suit a certain heat-treatment, desirable for mechanical reasons, economy, etc., so as to get the desired structure?

Already in the present state of our knowledge, we get much useful information from microscopic examination of the steel, especially when it includes the measurement of the grain.

Such a method throws much light on the way a certain steel is affected by a certain heat-treatment. It shows, within limits, what maximum finishing-temperature a steel of a given composition will stand, without assuming an objectionable crystalline structure. But above all it tells us a great deal about the past heat-treatment of the metal. If the piece of steel under examination shows a size of grain which is not justified by its chemical composition, it is safe to infer that the heat-treatment has been defective ; the metal was allowed to cool undisturbed from too high a temperature, thus inducing coarse crystallization.

I here desire to express my thanks to Mr. W. C. Post of the Illinois Steel Company, for his conscientious assistance and his careful prepa- ration of the specimens.

558 The Mineral Deposits Of Southwest Wisconsin.

The Minebal Deposits Of Southwest Wisconsin.

BY WILLIAM P. BLAKE, NEW HAVEN, CONN., AND SHULLSBURO, WIS. (Chicago Meeting, being part of the International Engineering Congren, Angnst, 1898.)

The numerous and copious reports of geological surveys made in the lead and zinc rion of Wisconsin leave, perhaps, but little room for any original work, or for descriptive details of the nature and origin of the deposits not already given in the exhaustive memoirs of Percival, Whitney, and Chamberlin and their associates. How- ever, since the completion of the last survey, in 1879, much more attention has been given than before to the exploration of the ores of zinc, especially of late years, since the region has been better opened up by railways and the zinc industry of the country has assumed large proportions, giving a constantly increasing demand for zinc-ores.

A residence of more than a year in the region, and the active direction during that time of mining operations over a considerable area, have familiarized me with the forms and peculiarities of many of the deposits; and it appears probable that some observations upon them, especially from the mining and commercial standpoint, may interest the members of the Institute. A few notes on the structure of the deposits from the mineralogical and chemical standpoints have already been presented by me to the Wisconsin Academy of Arts and Sciences.'*'

The presence of lead-ore in the soil at many points along the Mis- sissippi river was well known to the aborigines, and early attracted the attention of the frontier traders, who purchased from the Indians the ore and even lead smelted out by the squaws. The demand for the lead-ore soon increased, and it became one of the early and potent fiictors determining the settlement of southwestern Wisconsin, and the development of its mining and agricultural resources. The mines of that section, together with those of Missouri and the Mis- sissippi valley, may be said to have been the cradle of mining in the western United States. The deposits of ore being at or near the

December Meeting at Madison, 1892, "Notes on the Stnicture of the Ore-De- posits of Southwest Wisconsin."

The Mineral Deposits Op Southwest Wisconsin. 559

sorface and being Dumerous and widely distributed, afforded to poor men an opportunity to mine on their own account with little or no capital. The "diggings/' which have often been termed "poor men's mines/' soon attracted a large population. Laborers and miners were drawn thither from Cornwall and other mining centers and those not bred to the use of the pick, gad, and windlass soon gained an experience in the use of tools and methods most useful to them in the new fields of the great west and in the mountain ranges sloping to the Pacific.

We are also largely indebted to the lead- and zinc-deposits of Wisconsin for the early institution of mineral and geological surveys by the general government, and also for the state surveys, by which the progress of geological work in America was greatly promoted. No pure love of science for its own sake, however, moved and stimu- lated either national or state legislatures to vote money for such surveys. It was in every case the hope of gain and the expectation of promoting the revenues of the state that led to the organization of geological surveys in those days, and that really lie in most in- stances at the root of the sources of state appropriations at the present day.

The chief centers of shipment of ores are Shullsburg and Benton, both within the region drained to the Mississippi by the Shullsburg branch and the Fievre river debouching at Galena. A series of mines extends from below Benton to and beyond Shullsburg. Some of the most prominent are the Yon Dusko, Buncombe Hill, Ben- nett Brothers, Diamond Joe, Byrne's, Coltman's, Blende, Ida Blende, Sallie Waters, Raisbeck's, Leary and Coulthard, and Cuba City; the Zinc Carbonate Co.'s mines (with mill), and the mines of the Wisconsin Lead and Zinc Co., including the Monte Christo, Helena (with mill), Blaine and Logan, McCarty, Galena Level, Little Giant (with mill), Oakland Level, Wagner or McFeeley, Bonanza, Stop-line, Hempstead or Old Elevator diggings, and others. There are numerous shafts, tunnels, and drainage-adits, some of great length and draining extensive areas. The shafts sel- dom exceed 80 feet in depth. Most of the ores are hoisted by hand- windlass or trammed out of the tunnels, but open-cut mining for "dry-bone" (earthy smithsonite) has been successfully practiced in some places.

Area and Geology.

The lead- and zinc ore region of Wisconsin extends through por- tions of Grant Lafayette and Iowa counties, and has an area of

560 The Mineral Deposits Of Southwest Wiscx>Nsin.

aboat 1776 square miles. It lies just north of the Illinois State- line and south of the Wisconsin river, extending in a general north- easterly direction from the Mississippi. This is also the general direction of the groups of deposits and of the main lines of ore- occurrence.

The geology is simple. The ore-deposits are confined to nearly horizontal strata of dolomite and limestone of Lower Silurian age, which lie between the Potsdam sandstone and the equivalents of the Cincinnati or Hudson River shales. These strata enumerated from below upwards, include the Lower Magnesian limestone, the St. Peter's sandstone, the Trenton limestone, and the Gralena limestone,* or dolomite, this last having a vertical thickness of from 126 to 200 feet, and being the galena-bearing rock. It rests upon the beds of the Trenton limestone, a compact blue limestone generally known in the lead- and zinc-region as 'glass-rock,'' from its brittleness and oonchoidal fracture. The two formations are separated in the neigh- borhood of Shullsburg and New Diggings, in Lafayette county, by thin layers of a brown bituminous shale called "oil-rock," upon which, in that vicinity, the zinc-deposits spread out and end, though in other districts they are known to extend lower into the blue lime- stone.

The oil-rock, so far as yet known, attains its greatest development in the mines in the vicinity of Shullsburgf and along the Shullsburg branch, notably at the Dry-Bone diggings and at the Gralena Level, where it is overlaid by a strong sheet of sphalerite and barite with crystallized galenite.

Forms of Ore-Deposits.

The ores are found in vein-like sheets in vertical and inclined crevices and in cavern-like enlargements along the course of such crevices known as "openings." Again, the ores form flats or sheets extending for some distance laterally between the strata, but, in my

The CliflT limestone of Owen (1S40), known also as the Upper Maesiao lime- stone, but including other formations as high in the geologic series as the Helder- berg limestones of the Devonian. The area of the Galena limestone, which maj be taken as limiting the occurrence of lead ore and also sinc>ore, is given by Strong (OeoL WU., ii., p. 684) at 1613 square miles.

t Notably in some mines of the Hempstead ; at the Butler shaft on the Little Giant tract, and at several other places on the property of the Wisconsin Lead and Zinc Co. At the Galena Level it underlies an extensive sheet of jack (blende), and is interstratified therewith. The oil-rock, when dry, bums with a bright flame and much smoke, and it yields upon distillation a thick, Tiscid petroleum.

The Hinebal Deposits Op Southwest Wisconsin. 561

experience generally resting upon the comparatively impervious Strata of the brown shales, the latter being often partly decomposed and forming beds of blue clay.

Observation confirms in general the theoretic conclusions of Whitney, Chamberlin, and others regarding the genesis and distri- bution of the ores. The ores everywhere present the same general sequence of lead-ore in the upper portions of the diggings, with carbonate of zinc gradually passing into sulphide of zinc below, which last is associated with pyrite of the mareasite variety.

In the lead-region the ore-deposits are seldom referred to as "veins" or " lodes," although so far as they extend vertically in the formations they have what we generally term vein-structure to a remarkable degree. They are called " crevices " or " openings," and sometimes "sheets," " pitches," or " runs."

Although there is locally great diversity in the ascertained or sup- posed direction of the ore-bearing crevices, varying from northwest and southeast to northeast and southwest, or from north and south to east and west, the major axis or direction of the principal groups of deposits appears to be northeasterly and southwesterly, approxi- mately coincident with the general geologic structure of the State, as shown by the trend of the uplifts further north, by the direction of the Baraboo axis, and by the direction of the mound-forma- tions— the remnants of the shales which have, for the most part,, been swept away.

Direction of Crevices.

So far as my experience extends (it has been chiefly in the region beyond and below Shullsburg to ,New Diggings), the crevices with the northeast trend are the most common and persistent, and are crossed at intervals by other crevices, of less extent, at right angles, giving what are called " 10 o'clocks," or having a direction nearly N. 20° to W. This general parallelism of the ore-bearing groups with the structure of the region is significant of some rela- tionship or dependence which has not hitherto been fully investi- gated. Percival gave great attention in detail to the direction of the crevices, and generally refers to them as " norths and souths " and " easts and wests," while recognizing the general bearing as quartering or northeasterly and southwesterly — his east and west ranges being in fact the same as those approximating a northeasterly and southwesterly course. He says : " The term east and west is applied to the leading ranges, although they may deviate even 46°

? IvL

562 THE MINERAL DEPOSITS OP SOUTHWEST WISOOlfSIlf/

from a due east and west coarse."'*' He also pointed out the fact that the traces of order and connection in the surface arrangement appear no less remarkable than in the vertical arrangement, and that the ore-bearing ranges in their bearing (direction) and grouping have been governed by some general laws, and have not been merely local accidents.f

Horizon op Ore-Deposits.

The question whether the ores of lead and zinc are confined merely to the horizon of the Galena limestone and to the upper portions of the Trenton or extend much lower, even into the Lower Magnesian limestone, has been much discussed by the successive geologists of Wisconsin since the days of Percival, who favored the view of the deep-seated source of the ores and their presence in the lower forma- tions. This view was emphatically discouraged by Whitney and by Chamberlin, whose conclusions have been generally accepted, and no effort has been made to test the question practically by a deep shafl, as has been several times proposed.

Chfep Deposits op Blende.

The bulk of the blende comes from bedded or horizontally dis- tributed ore and not from vertical or inclined crevices. The de- iposits show less irregularity of formation than those of galenite, being for the most part below the general water-level, and not affected by oxidation and the attendant solution of the rocky walls and the for- mation of irregular caverns. But the beds are not without cavernous -spaces or " openings," such as are ordinarily called " vugs." We may make two broad or comprehensive 'divisions or groupings of the de- posits into 1. Irregular and brecciated deposits; and 2. Rular sheets or beds.

In the irregular and brecciated we may include most of the dry- bone derived from the oxidation of the blende in place, which dry- bone continues to occupy the same places and passes downwards into unchanged blende. Sometimes the original bedding of the rocks is bat little changed, and there is no disturbance, but in other places there is great confusion and breaking up of the bedding with the accumulation of Irregular masses of rock, surrounded and invested with a coating of ore, by which the rocks are united into one mass.

♦ Report of 1856, p. 76.

t Percival's Beport of 1856 (second report publinhed after his death), p. 63.

the mineral deposits op southwest wisconsin. 663

Breociated Deposits.

In all of the deposits of blende and pyrite, not only in those formed in the midst of a breccia but in the regularly formed horizontal beds, there are places where the ores have been broken across, with the layers disjointed, so as to show faulting of these layers on a small scale, but such breaks have been recemented and reunited as strongly as before. It is evident that there has been movement, probably in most cases by the falling down of masses of ore by their own weight, when perhaps the rocks below became eaten away or the soft clays became washed out, thus removing the support.

In the case of the breociated deposits, where no distinctly regular flat sheets are seen, the original bedding of the rocks themselves is obliterated or destroyed by the falling-in. It is not always easy to decide whether the broken condition of the beds is the cause or the effect of the percolation of mineral solutions, but I incline to the view that, in most cases, it is the result of the flow of the solutions of acid salts of the metals which have eaten away the rocks below with- out fully replacing them in bulk, and that being so undermined the rocks have fallen in from time to time, and in the confused masses of fragments the zinc and iron sulphates have found the most favor- able conditions for deposition or replacement.

The Varieties op Ore.

There are now four different kinds of ore shipped from the Wis- consin mines, namely :

tiemical compound.

Mineral species.

Local name.

Lead sulphide.

Galenite.

Mineral.

Zinc carbonate.

Smithsonite.

Bone.

Zinc sulphide.

Sphalerite or blende.

Jack.

Iron sulphide.

Pyrite or iron pyrites.

Sulphur.

Of these, the zinc-ores largely preponderate, and the production is increasing with the constantly growing demands of the spelter and the white zinc-oxide industries.

The lead-ore is not now so much sought as formerly, and most of the old deposits are regarded as exhausted, although now and then new discoveries are made. The production may be rrded, how- ever, as chiefly incidental to the exploitation of the zinc-deposits. As these are followed, new deposits of lead -ore are sometimes uncov- ered, and in some of the zino-deposits a little galenite is intermingled with the carbonate and with the blende.

564 The Mineral Deposits Op Southwest Wisconsin.

The bulk of the shipments of lead-ore is in the form of coarse lump-galenite in large sheets ("sheet mineral"), or masses ("chunk mineral "), or in crystalline masses as broken out from the deposits and culled by hand. When in large cubical crystals the ore is known as "block-ore" or "cog-wheel mineral," or "dice-mineral," if in smaller cubes. The name " mineral " for all forms of lead-ore is in general use. When the galenite is the product of separation by jig- ging from either dry-bone or jack, it is of course broken up into small pieces and finer, and is then somewhat contaminated with resi- dues of pyrite and of jack, or possibly barytes, by which its percent- age is lowered. It is sold by the " thousand " (thousand pounds), and the price varies with the market value of pig-lead and the im- mediate needs of the smelters. In 1892 it ranged from $20 to $25 per thousand pounds, or from $40 to $50 per ton. It has special value to smelters by reason of its purity, being without arsenic, anti- mony, or silver, and thus furnishing a clean lead for corroding.

The total shipments of zinc- and lead-ores from Benton, the prin- cipal ore-shipping railway station in the southwestern part of the lead- and zinc-region, amounted in 1892 to 13,800,000 pounds, of which the lead-ore was 800,000 pounds, or only about 6 per cent, of the whole. This may be taken as representative of the present ratio of production by weight, though in the case of some new discovery of lead-ore the shipments for a season may be greatly increased.

The lead-ores are sent largely to the works of the Pennsylvania Lead Co. at Pittsburgh, and some go to Aurora, near Chicago- Formerly, the galenite was smelted at the local furnaces; but since the opening up of the country by railways all this has changed ; local smelting has been given up and the furnaces are disman- tled and in ruins.

Zinc-Ores — Smithsonite.

Smithsonite, or " bone," as mined and sent to market, is usually in three grades or sizes, the result of the methods of sorting and cleaning :

1. The large masses or picked bone, culled by hand as mined, and comparatively free from rock or other substances.

2. The washed or jigged bone, in smaller fragments, cleaned as far as possible from iron oxide, blende, and rock.

3. The finer portions or "smittems," more or less contaminated with heavy ochery olay limonite and ferruginous rock, which can- not be removed.

The Mineral Deposits Op Southwest Wisconsin. 565

Of these sizes the first, being the cleanest and carrying the highest percentage of oxide, commands the highest price. The value of the second grade varies greatly according to the presence of more or less limonite, which is not easily separated. The third is the lowest grade and the least desirable, as it is too fine to be cleaned by cull- ing. In selling these ores at the close of the season's work, it is customary to sell the whole together at an average price per ton for all grades, since it would be difficult to dispose of the smittems alone.

Dry-bone is used for the manufacture of white zinc-oxide for paints, and is not used for the production of spelter.

Although by opening up new deposits of this ore the production may be temporarily increased for some years, it is easy to foresee that, with the large and ever increasing demand, the supply will diminish and the value will advance. Missouri supplies more 'jack'' than bone," and the Wisconsin deposits are at present the most available for the latter product. Some ore has been obtained from New Mexico, but the deposits are too far away to permit all the grades of ore to be marketed. Only the very highest grades bear the great cost of transportation. There are also deposits in Arkansas, but these are also difficult of access and are now idle. The smithsonite or "bone'' is sent chiefly to the works of the Mineral Point Zinc Oxide Works, at Mineral Point, in Iowa county, Wisconsin, on the line of the Chicago, Milwaukee and St. Paul railway, and some now goes to the newly established works of the Lanyon Zinc Oxide and Paint Co. at Waukegan, on the shore of Lake Michigan, about 30 miles north of Chicago. A few shipments are made to the St. Louis works of Page & Grouse, and a new establishment at Dubuque, Iowa, is projected. The value of the best grades during the year 1892 ranged from $18 to $24 per ton, delivered on the rail. The quality of even the best grades varies at different localities. At some, the bone is light-colored ; at others, it is more or less discolored by iron oxide or ocher; manganese oxide is rare. Thick, solid crusts of compact and vitreous ore, like some obtained in England, are seldom seen. Zinc silicate is but rarely found. Mineralogically, the specimens are not beautiful, but they sometimes have the form of very interesting casts or replacements of crystals of calcite.

Although the attention of the miners had been directed, before the year 1855, to the commercial value of the dry-bone (smithsonite), which they were then throwing away as useless, it was Percival, the poet and geologist, who first effected a practical utilization of this

566 THE MINERAL DEPOSITS OP SOUTHWiST WISCONSIN.

ore by sending two barrels of it to the works of the New Jersey Zinc Co., at Newark, N. J., where its value was practically demon- strated.*

Blende — " Black-Jack."

The blende of Wisconsin is very different in its appearance from the ruby-red variety of the Missouri mines. It is, instead, gener- ally of a dark color, and hence appropriately called "black-jack;" but this color is seemingly due not so much to the presence of com- bined iron as to organic matter. This is partly destroyed by heating, and the blende then becomes lighter in color. The proportion of zinc in the clean blende compares well with that of any other locality, ranging from 60 to 64 per cent, or more. Some of the de- posits yield a considerable proportion of the light honey-yellow or amber-colored variety, generally known by the miners as resin- jack." I have not yet seen any colorless and perfectly transparent blende in the Wisconsin ores, or in those of any locality in the Mis- sissippi valley.

The jack, like the carbonate ore or bone, is prepared for market in three principal grades or conditions:

1 . The massive, hand-sorted, culled jack, cleaned as far as possible from all rock, pyrite, bone, and lead.

2. The medium-sized fragments or "sieved jack," picked oat from the mixed and broken stuff of the mine and cleaned by washing.

3. The jigged ore and "smittems."

4. To the above we may now add the dressed-jack from the roasting process.

The first commands the best price and is most desired by the smelters, though often carrying a considerable amount of pyrite in thin scales, which cannot be removed by the cul ling-hammer. The second grade contains a still greater admixture of pyrite. If pyrite is abundant in the original mass it cannot be profitably removed by hand, and the mixed ore is laid aside as unmerchantable, though sometimes sales are made to smelters at a greatly reduced price. The same observations apply to the jigged jack. If comparatively free

♦ Percival, in his first Beport, 1855, wrote, p. 97: "The ores of zinc, although Terj abundant in manj instances, particularly in the flat and pitching sheets and in the lower openings, have never jet been turned to any account There can be no doubt that they must be hereafter sources of profit when we consider the large and increasing demand for zinc, both in its metallic form and as an oxide (zino- paint)."

The Mineral Dep06It8 Of Southwest Wisconsin. 667

of pyrite, it oommands a good price, but it is generally impossible to obtain any large amount not thus contaminated. In the culling and sorting, employment is given to many boys during the summer months, the clean lumps of jack, freshly cleansed from the dirt of the mine by free washing, being picked out by hand at so much per box or per hundredweight.

The chief market for the Wisconsin blende is at the spelter works of Wenona, Peru, and La Salle, Illinois, the route of delivery or transportation being by the Illinois Central railroad branch south- wards from Warren.

Occurrence op Barite.

Barite (heavyspar) is an occasional associate of the deposits of blende, and usually occupies a medial position in the veins, being formed upon the horizontal sheets of jack, and generally in heaviest layers upon the lower side of the openings. It is rarely crystalline, but forms in rounded, snow-white mamillary masses, contrasting sharply with the dark-brown or black of the jack. Any gale- nite present is usually associated with, or planted in the midst of this barite.

We may easily account for the formation of sheets of barite in the veins upon the hypothesis that carbonate of baryta, a soluble salt, flows downwards from the surrounding rocks into the fissures and beds, and there meeting with sulphate of lime in solution, formed by the decomiiosition of the walls during the deposition of zinc- blende, is changed into the insoluble sulphate, calcite being formed at the same time. In fact, we find numerous crystals of calcite planted upon the surfaces of barite.

Barite is objectionable, commercially, in connection with either the blende or galena. It is so heavy that it is separated with difficulty from galenite in the jigs, and it ciinnot be separated from the jack ; consequently, in the treatment of mill-stufi containing heavyspar a mixtnre of jack, heavyspar, and pyrites is obtained ; and although by the new process now employed the pyrite can be removed, the barite remains. Although innocuous metallurgically, its weight and bulk increase the cost of handling, freighting, and smelting, and make the blende-concentrate nearly valueleas. In some of the de- posits the barite may be nearly all separated by careful hand-culling, and, when clean, can be sold at a price which pays a little more than the cost of breaking and saving it.

563 the mineral deposits op southwest wisconsin.

Pyrites— Marcasite.

The pyrites, generally known to the miners as "sulphur," is . chiefly marcasite. It occurs next to the walls of the crevices, or coating the masses of dolomite, and is tightly attached to them, while the blende is superimposed. As mined, the two minerals are broken and thrown down together, and the coarser parts are then separated by hand. The layers of marcasite vary from one-sixteenth of an inch to several inches in thickness, and when thick enough to detach from the sheets of blende, the marcasite is saved for sale. At the Helena mine large quantities have been saved and sold at $3 per ton. It carries a little iron oxide on the surface, a little adhering rock, some calcite, and a trace of jack. Analysis of the piles in bulk shows 46 per cent, of sulphur. It is a very free-burning pyrite, and does not contain arsenic or antimony. It is a desirable article for the produc- tion of sulphurous acid, and is used also by smelters to add to charges which require more iron sulphide.

In the operation of band-sorting and culling the blende from the rock and pyrites (much of this work being done with the hammer), considerable quantities of "smalls" or chips are made from which the blende cannot be separated by hand. There are also large deposits of mixed blende and pyrites where the two minerals are so mingled that they cannot be separated by culling. In fact, the deposits of zinc-blende in heavy sheets free of pyrites or sufficiently so to be made available as a commercial product by hand-sorting are rare, and even in such mines there are portions which are avoided or lefl standing because separation cannot be effected even in a mill. For in milling and concentrating zinc-ores as they come from such deposits, while it is feasible to separate thoroughly any rock or flint, and also to take out the much heavier galena, a concentrate is obtained which carries not only tho blende but the pyrites which was present. Such concen- trates are not fit to use for the production of spelter and are not marketable. The treatment of such concentrates for the separation of the pyrite and the production of a clean high-grade merchantable zinc-blende is the subject of another paper.

The Separation Of Blende From Pyrites. 569

The 8Epabati0N Of Blende From Pyrites: A New Metalluboical 1Ndu8Tbt.

BT WILLIAM P. BLAKE, NEW HAVEN, CONN., AND 8HULL8BURQ, WIS. (Chicago Meeting, being part of the International Engineering Congress, Angust, 189S.)

In another paper I have described in general the close association of zinc-blende and iron pyrites in the ore-deposits of southwest Wis- consin. These two minerals generally occur together at the same horizon a sheet of one being overlain by a sheet of the other; and also the pyrites is disseminated through the blende to such a degree that the latter is rendered comparatively worthless for the produc- tion of spelter. Ordinarily the pyrites, which is mostly the trimetric marcasite variety; occurs in about equal quantity with the blende and is separated by hand-culling and chipping; but there are large de- posits where the ores are so closely mixed that they can neither be separated by culling nor by any mechanical method. The specific gravity of the two minerals is so nearly the same that attempts often made to effect even a partial separation by jigging have proved un- successful.

The smelters of zinc-ores in the Mississippi valley will not take ores which carry more than 7 per cent, of iron ; and it is therefore necessary to cull the massive "jack with great care and at consid- erable expense, in order to bring it up to the market-standard of purity.

The results of trials to effect a separation of blende and pyrites in other regions have been equally unsatisfactory. The literature of this branch of metallurgy abounds in examples.

In Wisconsin efforts have been made in the past to " burn " the mixed ores in heaps, in order to soften or destroy the pyrites, .and facilitate its separation from the massive blende.

The same method has also been resorted to when galenite is mixed through the substance of the blende, for the purpose of disintegrating the ore, so that it could be separated by jigging. The custom was to pile the ore and the fuel together. As might have been expected, the results were far from satisfactory. As is usual in heap-roast- ing, some portions of the ore were overheated and decomposed; zinc was lost by volatilization ; the galenite in some portions was

570 The Separation Op Blende From Pyritb8.

partially reduced ; and the pyrites was unequally burned. The method was at once wasteful and ineffective.

In consequence of the impossibility of effecting a separation of these substances, many mines throughout the Wisconsin zinc-region have been but partially worked or closed altogether, while in others, the extraction of ore has been confined to such portions of the deposit as would yield massive blende in a condition to be hand-sorted or culled from the pyrite.

It was believed by some that, with good machine-jigs and a care- ful classification of the ore into the proper sizes, a marketable con- centrate of nearly clean blende could be obtained. Two large mills were erected by the Wisconsin Lead and Zinc Co., and hundreds of tons of concentrates of different sizes were turned out ; but the per- centage of pyrite remained so high that these concentrates could not be sold. The galenite was easily secured, but the blende and pyrite were inseparable, and neither was in a marketable form.

At the request of the above company, I undertook the design and construction of a special form of furnace for the purpose of roasting the mixed concentrates, preparatory to their separation by jigging. The conditions of the problem were :

1. Automatic feeding and delivery.

2. Oxidation of the pyrite without acting upon the blende, and without matting.

3. Avoidance of a heat sufficient to affect the lead-ore.

These conditions were satisfied by the form of furnace described by me in a former paper {Trans., xxi., 943). The pyrite is oxidized to such a degree as best suits the conditions of separation. The blende is not decomposed, scarcely losing its splendent luster, and the galenite is hardly tarnished. The grains or fragments of the concentrate are delivered from the furnace without being in the least dree adherent, and are as granular and separate as when fed in. Beyond a slight breaking up of the fragments of blende by decrepi- tation, there is no change of form in the concentrate, except as to the pyrites, which suffers a change of bulk accompanying the loss of its sulphur. The pieces of marcasite swell up, expand, crack open and exfoliate in a peculiar manner, and, being more spongy and bulky than before roasting, and the blende fragments being made smaller by decrepitation, the two substances are in a condition favorable to a separation, which is then effected by jigging in the usual way. The products are clean commercial or marketable blende, or " dressed- jack '' and tailings of red sesqui-oxide of iron, or, rather, a mixture

The Separation Op Blende Fbom Pyrites. 571

of the oxides, from which an excellent raetallic paint can be made. The small residue of zinc-ore contained in it, and consisting partly of ''chats'' and partly of any zinc carbonate originally present, add to its value as a pigment. Any lead-ore present is removed in the first compartment of the jigs ; the furnace-operation having had the effect of splitting off the blende and leaving the galenite free. This is another and very important advantage of the roasting process ; for a blende containing over 2 per cent, of lead is unfit for the pro- duction of the best spelter, and ordinary "leaded-jack*' will carry more than this quantity with it. But the roasting frees the lead-ore and it can not only be taken from the blende but is by itself a more valuable product than the latter.

I am thus able by this process to treat successfully any ore com- posed of zinc-blende, pyrites and galenite. The ore (which may also contain a large amount of rock and gangue), is first crushed, sized and concentrated, giving a concentrate containing, say, 25 per cent, of blende, 25 per cent, of pyrite (marcasite chiefly), 5 to 10 per cent, of galenite, and the remainder dolomite and flint. This is sent to the furnace, dried and roasted, and is then re-jigged, giving a marketable blende of high grade, assaying in the best samples over 62 per cent, of zinc, less than 3 per cent, of iron and less than 1 per cent, of lead (often less than 0.5 per cent, depending upon the nature of the original ore). This is a much cleaner and purer blende for smelters' use than can usually be got by hand-culling the best grades of massive blende-ore.

The key to successful separation is found in the complete and even roasting. Every particle of the bisulphide of iron must be decom- posed. Even a remnant or kernel of unchanged pyrites will cause the fragment to remain with the blende. The outer coating of oxi- dized iron may be broken away and go off in the tailings; but the inner portion, containing the unchanged pyrite, will not go over. Any neglect or carelessness at the furnace is speedily detected at the jigs, and imperfectly roasted ore has to be sent back.

By this roasting process a large amount of sulphurous acid is pro- duced, which at present is allowed to escape into the atmosphere, but could be utilized for the manufacture of sulphuric acid. When the furnace is in full operation, from 3 to 5 tons of pyrite are burned in twenty-four hours, this amount being, roughly stated, the amount of pyrite in the concentrates put through the furnace.

The process of making clean blende for sale to the spelter-works now consists of the following operations :

572 The Separation Op Blende From Pyrites.

1. CrushiDg, sizing and jigging the crude ore, giving a mixed concentrate of blende and pyrite; most of the galenite being taken out in the operation. The product is known as raw concentrates."

2. The raw concentrates are spread upon the drying-floor at the top of the furnace and are dried and warmed before being fed into the hopper, from which the concentrates descend automatically into the roasting-chamber.

3. The roasting operation; the concentrates being turned and moved automatically, and being dischaiedat the side of the furnace without manual labor.

4. The cooling in heaps, the result being " roasted concentrates/' which are conveyed to the mill ; and,

5. Jigging and cleaning the roasted concentrates, giving clean blende of 60 per cent, and upwards, and any lead-ore which was contained in the concentrates, and the tailings for making paint

It is found practicable to take the poorest refuse ore, which may contain only 10 per cent, of blende or less, the remainder being pyrites and rock, and make a very low-grade concentrate in the usual way, and then by the proper roasting and re-jigging recover the greater portion of the blende in a marketable condition. It is of course less difiBcult to secure a high-grade product from raw con- centrates carrying a large percentage of blende. It is at best, even afler the roasting, a delicate operation to hold and secure the blende without loss, and at the same time to throw off the large percentage of roasted pyrite and the residues of rock. The latter are still con- siderable; for, in preparing the concentrates from the original ore for the roasting process it is the best practice not to make a perfectly clean concentrate at the first jigging. It is better to allow some of the rock to go in with the product, as the final loss is then much smaller. This rock is thrown off in the re-jigging which follows the roasting.

The raw concentrates being already sized and having been kept separate during the roasting, there is, theoretically, no necessity for passing them a second time through screens ; but it is found that, owing to the decrepitation of a part of the blende, it is desirable to screen out the finest before jigging ; and this is done as the roasted concentrates are fed to the jigs.

The best results are obtained, especially with the fine sizes, upon bottom-discharge jigs.

For particulars of the control of the roasting in the automatic furnace, I refer to my paper in the Transactions already cited.

The Separation Op Blende Fbom Pyrites. 573

I find that it is not necessary to remove the last atom of sulphur, and that by proper manipulation the oxidized pyrites may be left in a condition to be attracted ind removed by magnets. A special in- stallation for this purpose has been made on a small scale and oper- ates well ; but inasmuch as the jigging is so satisfactory, magnetic separation has not been employed to any considerable extent. It is, however, interesting to note that the fully oxidized pyrites which is least affected by the magnets is the most easily removed by the jig- ging, while the fragments which are most responsive to the attrac- tion of magnets are those more difficult to remove on the jigs. The two methods might thus work together to advantage, but the low value of the blende may not justify a second handling. The mag- netic method may, however, be used to advantage upon the product of the last compartment of the jigs, where the lighter fragments of blende and the heavier fragments of roasted pyrite fall together.

Experiments have been made by Mr. J. W. Meier, upon the mag- netic separation of iron from blende, the pyrites being first roasted in a muffle so as to convert it as much as possible into the magnetic oxide. His results were so encouraging as to lead him to believe that the process could be made a commercial success. Mixtures carrying 33 to 35 per cent, of zinc and over JO per cent, of iron were easily converted into concentrates with 50 to 54 per cent, of zinc and only about 4 per cent, of iron. The chief difficulty was to get all of the pyrite into the magnetic condition. Fragments fully roasted and changed to. the sesqui-oxide could not be removed by the magnet. It seemed necessary after the removal of the sul- phur to introduce some carbonaceous material or vapor so as par- tially to reduce the higher oxides to the magnetic oxide.

Electro-magnetic separation of ferruginous impurities from zinc- ore has been successfully practiced in Sardinia at the Monteponi mines, where in dressing calamine, an intermediate product contain- ing limonite and ocher, corresponding to the " rust'' of the Wiscon- sin miners, is obtained. This is first roasted with a portion of finely ground coal so as to effect the reduction of the hydrous ferric oxide to the magnetic oxide. At the same time the portions of dolomite remaining as impurities are rendered caustic, and can be slaked in water and removed, while the residue is passed through a magnetic separator, by which the iron is removed, giving a final product of commercially clean calamine — zinc silicate* — averaging 42 to 45 per cent of zinc.

The apparatus was set at work in Maj, 1890, and dtiring the first twelve months

574 Improved 8Lag-Pot8.

The paper of Mr. Ellis Clark, Jr., on " Ore-DressiDg and Smelt- ing at Pribram, Bohemia " (Tans,, ix., 420), contains (p.451), a description of a magnetic separation of rotfsted siderite from zinc-ore.

It is probable that the magnetic method may be sacoessfally ap- plied upon mixtures of smithsonite ('bone") and iron oxide (the rust '' of the Wisconsin miners) ailer a preliminary roasting. Bone is not easily saved by jigging after roasting, in which process it loses its carbonic acid, and becomes very friable.

The method of roasting and re-jigging above described has been in successful operation at the Helena mill, three miles west of Shulls- burg, for nearly two years, and hundreds of tons of high-grade zinc- blende have been produced and sold to the zinc-smelters from ores that could not be culled. The process utilizes ores which were before thrown aside as useless, and also makes a market for the '' chippings " of the culling-floors, which have accumulated for years at the chief mines. Many large deposits of mixed ores are thus made available and the wealth of the region is correspondingly increased.

In the operations of concentrating I have been greatly aided by the industry and skill of Mr. John F. Perkins, mill superintendent.

The work is now upon an established systematic basis. Ores from any source are taken and milled in quantity, and the results heretofore unsalable are now of the highest market value. The practical miners and ore-dressers of the zinc region concede the success and value of the process.

Improved 8Laq-P0Ts,

BY H. A. KELLER, BUITE, MONTANA. (Chicago Meeting, being part of the International Engineering Congress, August, 1893.)

Amoko the important implements of most of our Western lead and copper smelting-works is the slag-cart or buggy, commonly called slag-pot. A large number of such slag-pots being in service, dura- bility, as well as lightness of running, is an important requisite, since repairs, though trifling in each single instance, may amount to a con- siderable item of cost in the course of a year. They should therefore

seldom necessary and inexpensive when necessary. Light ron-

at 5000 tons of crude ores containing 20.17 per cent of zinc were treated, pro- ing 1630 tons of merchantable zinc-ore of 41.67 per cent., since carried to 45 per Min, and ScL JVeM., December 17, 1892, from Ofsterr, ZdUeh. fir Berg, ic

Ihpboyed Slag-Pots.

ning isy in many respects, also a considerable advantage, for a light- running cart frequently carries more slag and necessitates lees exertion in moving than a poorly constructed one. Thus fewer trips result, with consequent saving in both repairs and labor. The importance of the latter item is specially appreciated by the metallurgist who has' had to deal with the problems incidental to large and constantly grow- ing slag-dumps.

To aid in overcoming the great expenses due to large dumps, three devices are in use at the Parrott works, at Butte, Montana, namely, casting part of the liquid slag into brick ; granulating another por- tion of it by water, which subsequently carries the fine slag oiT, and lastly, the using of well-constructed buggies over an iron track to dis- pose of the remainder.

Fig. I.

Eiler's Slag-pot

The slag, which must be siliceous, is bricked by casting it into moulds. These moulds were originally designed by Mr. J. E. Gray- lord, and were fully described and illustrated by Prof. Thos. Egles- ton.* These bricks, 6X6X12 inches in size, are admirably adapted for building-foundations, and find for that purpose a ready market. A good workman casts about 500 in a ten-hour shifl.

In the West, slag has been granulated and disposed of by water for a number of years at Anaconda. This method was likewise adopted at the Parrott works in March, 1892, though a limited sup- ply of water, and lack of fall besides, made the problem somewhat more complicated. Early in 1891, the writer made provisions for

Manufacture of slag-bricks in Montana, School of Mine$ Quarterly, Tol. xii., p. 1S9.

576 Improved Slaq-Potb.

such an arrangement in a lead-smelting works. The fine slag thus obtained may be utilized for a number of purposes. . It is the object of this paper without going into details on brick- ing or granulating, to describe and illustrate slag-carts to be pushed by hand, which have been designed after an extended and varied ex- perience.

In the accompanying illustrations, Figs. 2 to 5 inclusive represent the cart now in use at the Parrott works, and Figs. 6 to 9 a cart simi- lar to the one introduced by the writer at the Philadelphia works at Pueblo, Colorado.* Parts of these pots have been in use for a num- ber of years, while other parts are of more recent date.

The cast-iron track shown in the drawings is laid into that part of the slag-dump which by constant usage is apt to become specially rough and uneven. A rough dump, besides adding to the work of the slag- wheeler, greatly increases the necessary repairs. Further away from the furnaces the dump is levelled by " slag squares or slabs of slag formed by pouring, liich are constantly kept up to its edge. These are best made 2 feet by 4 feet and from 8 to 12 inches deep. After a mould of these dimensions has been formed by means of rails or cast-iron plates and cold slag, it is partially filled with large pieces of cold slag, which are then cemented together with liquid slag poured simultaneously from several pots.

The slag-cart consists of three parts, the bowl or pot proper, the wheels, and the handle or foot.

There are two styles of bowls now in general use. These are represented in the accompanying Figures, as No. I and No. 2. On account of its straight sloping sides, the pointed bowl. No. 2, allows the matte to settle more readily. It is therefore preferred when but little matte is suspended in the liquid slag, which matte is to be saved in shell and bottom. To avoid unnecessary disturbance, this bowl is provided with a 1 J-inch hole, through which the liquid slag is tap[)ed by a |-inch bar, usually of hexagonal steel. Such a tap-hole was first used in this country by Mr. W. B. Devereux, at Aspen, Colorado.f Its location varies, of course, with circumstances. After much experimenting, the writer, for instance, when employed at the Philadelphia works, determined to locate it as shown in Fig. 8.

The bowl No. 1, with rounded sides, has the advantage of greater capacity than one of similar dimensions with straight sides. Such

Hofman's Metallurgy of Leady p. 203.

t Se Kerrs MUaUhuHmkundt, 1881, p. 100.

Improved Slag-Pots.

a bow] is therefore preferable when it is intended to dump out the entire cone. A more sliallow bowl (shown in Fig. 1), introduced by

Mr. A. Eilers and universally used in early Ticadville smelting- practice, is gradually disappearing with the increased size of slag-

VOL. XXII.— 37 nirn]o

? IvL

578 Improved Slag-P0T8.

dumps, since it does not permit as great capacity as those shown in Figs. 2 to 9.

Bowl No. 1 requires the axles to be fastened to it with set screws, while the straight sides of No. 2 leave room for securing the axles with wedges. In the latter case, each axle is provided with a square stub J-inch larger than the diameter of the axle. The axles of the third form of bowl (Fig. 1), are carried by one continuous piece of square iron, bent to conform with its shape. This is fastened to the solidly cast side-lugs by stirrup-clamps and further held in place by passing between two guiding-lugs at bottom of bowl.

The lug shown at the rim of each howl is a great protection to the spot where dumping causes most wear. It was first suggested by Mr. C. T. Limberg of Leadville, and is now universally used.

The splash-guards prevent the liquid slag from spilling upou the hubs. They were, I believe, first introduced at the Grant works, Denver, Col.

The false bottom for pot No. 1, shown in dotted lines in Figs. 2 and 4, has been in use for several years at the Parrott works. It consists of a cast-iron dic held in place by a countersunk {-inch holt If the original bottom is very badly damaged, a washer may have to be used besides. Though such an arrangement does not give satisfaction with a heavy flush, it behaves admirably with slag or matte running slowly. With matte, it has the additional advantage of producing a flattened cone which is more easily broken than a tapering one. New bowls are used for slag at the Parrott, and last about eighteen months. After that, being provided with tliese false bottoms, they last almost as long on matte. Another contrivance for using a bowl after its point is worn out has been described by Mr. R. H. Terhune.*

By the device shown in No. 1 , the bowl is made reversible, the cart being at the same time steadied by fastening the handle with two straps instead of one.

The style of wheel representedf consists of a cast-iron hub, with wroiight4ron spokes and tire. It is mounted upon a steel roller- bearing. The hub is tapped to receive the end of the spoke which for that purpose is threaded. For further tightening, each spoke is .provided with a jam-nut. After the spokes have l)een carefully ad- justed, the wrought-iron tire is shrunk upon their outer ends and is 8ul)sequently fastened to them by means of countersunk rivets. This lire being the part of the wheel subjected to most wear, must be

Tran$,f zy., 92. f Patented by Cole, Gay lord and Keller.

IMPROVED SliAQ-POTS.

made saffioientlj heavy and strong, withoat being clumsy. The ad- vantages of a wrought over a cast tire, are evident, particularly when, as in this case, the former is well fastened and easily repaired.

The anti-friction rollers require no oiling, or at least but little. A double set of these rollers is put in loosely around each axle. Thus arranged, they are prevented from wearing so as to cause apprecia-|e

580 Improved Slag-Pots.

ble friction i.e., by running upon one another. By using f -inch rollers, an 1-inch axle takes two sets of nine; and for each -inch increase of axle-diaraeter an additional roller is required.

The foot or handle is practically the same for all forms of slag- cart. The essential point is that it shall be of suflBcient length, and its crosspiece wide enough for convenient pushing. It is fully illus- trated in the drawings and needs no further comment.

All rivets and bolts in either No. 1 or No. 2 are f-iuch in diame- ter. The average weight of No. 1 is 663 pounds. Many matte-cones taken from such pots with false bottoms gave :

631 pounds each for 55 per cent, copper.

603J pounds each for 52 per cent, copper.

597 pounds each for 47 per cent, copper.

574J pounds each for 44 per cent, copper.

A large number of slag-cones taken from pots with original bot- toms gave an average of 496 pounds. Their composition was :

Per Cent.

SiO„ 3S

Fe(J, 46.5

Al,()„ 10.5

CaO 3

Total, 98

The copper contents are from 0.3 to 0.4 per cent., present either as CujO or as suspended noatte.

It may be of interest to mention here that a large number of cop- per-matte analyses, from reverberalory- and blast-fumaoes, com- prising the different grades and extending .over several years, gave a constant tenor of from 21 to 23 per cent, sulphur. Many of these analyses showed the presence of magnetic iron, sometinaes in con- siderable quantity. Copper-mattes would accordingly correspond to either of these formulae

(Cu). + (Fe), or

(Cu),+ (FeS),+ (FeA).. No 1 slag- pot easts, as a rule, six full slag-bricks. A great many bricks, being weighed, gave an average weight of 54 pounds each. The slag given above weighs therefore 216 pounds per cubic foot Taking the same quantity of water at 62J pounds, gives a speciBc gravity of 3.47. These seem fully as accurate as similar determina- tions made in many laboratories of western smelting works, most of which are necessarily crude in this line of research.

The writer is indebted to Dr. Edward Keller for these formulie.

:ogle

FURNACE FOR BURNING SMALL ANTHRACITE COALS. fSl

A Fubnace With Automatic Stoker, Travelling

Qhate, And Variable Blast, Intended

Especially For Burning Small

Anthracite Coals,

By Ecklky B. Coxe, Drifton, Pa.

(Chicago Meeting, being part of the International £i:igiueering Congress, August, 189S.)

Having been appointed, on February 19th, 1890, a member of the Commission created by the Legislature of Pennsylvania for the pur|)ose of investigating the " Waste of Coal Mining, with the View to* the Utilizing of the Waste" (waste in this case referring specially to the small sizes of anthracite heretofore thrown away), my atten- tion was naturally directed to the most economical method of burn* ing these small sizes.*

Although I had been familiar with the burning of anthracite for many years, I found when I came to study the matter carefully that I was far from thoroughly understanding it, particularly with reference to the burning of the smaller sizes, and that the liter- ature of the subject, which was not very extensive, had more of a tendency to confuse than to enlighten me. I began therefore to in- vestigate the subject so as to arrive if possible at some definite prin- ciples applicable to the question. With the aid of my assistant, Mr. John R. Wagner, I Ijegan at Driftou a series of experiments upon the burning of small anthracites.

We first erected a small furnace with a grate 3 feet long and 2 feet wide, with which we experimented about two months. From the information thus acquired, we designed and built another furnace of about the same size, which gave us facilities for extending our investigations still further. Afler experimenting several months with the second furnace, we built a third one, which was much larger, having a grate area of about 68 square feet, which was ex- actly the size of the grate used in the boiler-plants of Coxe Bros. &

Commonwealth of Pennsylvania. — Report of Commission appointed to inves- tigate the Waste of Coal Mining, with the View to the Utilizing of the Waste, May, 1893.

582 Furnace For Burniko Small Anthracite Coals.

Co. We experimented with this several months, and then erected an improved furnace, a description of which is the object of this paper, and placed it in one of our regular boiler-plants.

I quote from the report of the Coal Waste Commission some of the results of our experiments :

A number of experiments were made in the testing laboratory of Coxe Bros. & Co., by Mr. John R. Wagner, in burning small coals, from which the following conclusions were arrived at:

A series of careful experiments were made with a forc draught, obtained in one case by a fan and in the other by a steam jet, which showed :

Fird. — That the ashes produced by a steam jet were never as low in carbon as those produced by the fan ; that is, an appreciably larger per cent of the carbon was utilized by the fan-blast. This appears to be due to the fact that when the carbon in the ash over the grate is reduced to a certain point the steam dampens it somewhat, and it ceases to burn sooner than it does when dry air only is blown through it.

"/Second. — That with the fan-blast the rate of combustion per square foot per hour is greater than with the steam jet.

" Third. — It was found that where a bed of coal was ignited and burned out, the percentage of carbon in the ash is much less than where coal is successively added to the burning mass. In practice it is not generally possible to allow the bed to bum out sufficiently before adding the cold, unignited coal ; the result is a damping down of the 6re, which causes the ash to cease burning sooner than it would do if there were no reduction of temperature and checking of the draught due to the adding of the coal.

" Fourth. — There seems to be no doubt that the introduction of steam into the ash-pit decreases very materially the tendency of the coal to clinker on the grate in comparison with the fan-blast or natu- ral draught. It also changes the color, volume, and character of the flame, and, owing to producer action, increases the distance that the flame extends beyond the bridge-wall. In many cases it is not prac- tical, or at least it is very difficult, to 6 re the smaller siies of coal without the steam jet on account of the clinkering. This efiect of steam on clinkering is probably due to the fact that the steam, to a certain extent, moistens the ash close to the grate and prevents the ash from reaching there as high a temperature as it would with dry air. It is also probable that the decomposition of the steam into carbonic oxide and hydrogen, which takes place to a certain exteot.

Furnace Fob Burnino Small Anthracite Coals. 683

584 Furnace For Burning Small Anthracite Coals.

and which, of course, is accompanied by a reduction of temper- ature, tends to prevent clinkering. The decomposition of the steam, accompanied by the formation of carbonic oxide and hy- drogen, will probably account for the difference in the flame re- ferred to.

" Fifth. — A careful study of the burning of culm, that is, the burning of small coals with more or less dust in them, in these and other experiments, seemed to show that in almost all cases it is ac- companied by a very high percentage of carbon in the ash, which analysis showed, in some cases, reached 58 per cent. Unless special precautions are taken to prevent it, a large portion of the fine coal runs down through the grate. When the culm gets red hot it acts almost like dry sand and works its way into the ash-pit, thus in- creasing largely the percentage of carbon. Where coal has to be transported any distance, the value of tl>e culm at the mines being very small, it is probable, from the investigations made, that it would be cheaper to remove the dust and transport only the larger coal.

Sixth. — It has been found that the percentage of iron pyrites, which occurs to a greater or less extent in all coals, increases very rapidly with the smallness of the coal. This is due to the fact thai the iron pyrites occur generally in thin layers or in incrastatioos od the coal. These thin layers are broken off and pulverized in the preparation and handling of the coal, and are therefore found to a much greater extent in the very small coal. It is, of eourse, well known that the presence of iron pyrites in fuel is very undesirable as it generates sulphurous acid and has a tendency to destroy the grates or other iron work around tl>e boilers, l)esides, in many eases, increasing the tendency to clinker.

Seventh. — That while the fan-blast produces the best ash and gives a more perfect and greater rate of combustion, yet in many cases it is more advantageous to use the steam-blower on account of the clinkering, which may cause very serious trouble. In certain localities, particularly in cities, the noise of the steam-blower is some- times a disadvantage.

Eighth. — While, it is not positively demonstrated, it is thought that the question of mixing small coals from different veins or dif- ' ferent localities is a matter of importance. It would appear that sometimes two coals, each of which, when buri>ed separately, give reasonably satisfactory results, when mixed together, clinker and give trouble, probably because the ash of the combined coals forms a much more fusible silicate than either of the &shes separately.

Furnace For Burning Small Anthracite Coals. 585

tized by

686 Furnace Fob Burning Small Anthbacite Coals.

" Ninth. — It would seem that the combustion of the small anthra- cite is more perfect when the coal remains undisturbed, or as nearly as possible in the condition in which it was put in the fire, instead of being turned over so (hat the partially consumed and the uncon- sumed coal are mixed together/'

Our experiments were not sufficiently extended and exhaustive to justify us in asserting that all these conclusions are absolutely true, but only that they seem to us probable.

Another point referred to in the same report, and which our further experience seems to confirm, is the fact that the temperature developed by the burning of the smaller coals decreases with the size of the coal ; this naturally involves a larger heating surface in the boiler in order to develop the same number of horse-powers, — that is to say, if you are burning pea coal and obtaining one horse- power for every nine square feet of heating surface, you would probably require from 20 to 25 per cent, more heating surface if you are using No. 3 Buckwheat ; although you may be evaporating practically the same amount of water per pound of combustible.

It is also stated in the paper that it is possible that the best results in burning these small coals may be obtained by using a blower under the grate and a suction apparatus in the stack. This statement should be modified, as the following is probably a more correct statement of the case : Where the pahsage of the gases through the boiler- furtlUce to the stack is free and unimpeded, and the stack reasonably high, it may he necessary to check the draught by a damper near the outlet ; while should the furnace and boiler be so constructed that the gases travel a long distance and are more or less seriously impeded in their flow to the stack, which is not very high, it may be necessary to put some suction apparatus in the stack. In other words, there is a certain speed for every boiler-plant which the gases should have in passing through, in order to obtain the most econ- omidal results, and some device should be adopted to maintain it

Having determined, in a general way, what seemed to me the proper conditions for burning small anthracite economically, I started to design a furnace which would, as far as possible, fulfil the required conditions which were :

1st To ignite the coal and burn it up without mixing it with fresh fuel ; that is, that fresh fuel would not be added to the already partially consuiiied coal.

2d. To have the furnace so arranged that the combustion would be continuous and uniform ; that is to say, that when the furnace

FURNACE FOR BURNIKO SMALL AKTHRAaTE OOAUB. 587

u n

£

r

u

Id

o

588 Furnace For Burxinq Small Anthracite Coals.

was in use the condition of the fire would be practically the same at any hour of any day of any week of the year.

3d. To make the work of firing as easy as possible, so that a minimum numlier of firemen would be employed, and that the whole operation of the furnace would be controlled by an intellint man, who would have more use for his brains than for his muscles ; the idea being that in a large and complete plant the coal would be brought from the source of supply by elevators or drags, and fed to the furnace without hand labor, and that the ashes would be car- ried to or dumped into a pocket, where they could be easily loaded into cars in the same way. No pokers, slice bars, or other similar tools should be needed.

The illustrations which will be referred to in the description of the process are as follows :

Plate I. — Diagram illustrating the process and furnace for burn- ing the small sizes of anthracite coal.

Plate II. — A reduction of the working drawing from which the iron work of the automatic stoker furnace for the Stirling boilers at No. 3 Colliery, Oneida, Schuylkill county, Pa., was built.

Plate III. — Detail drawing showing the construction of the grate- bar and water-back which forms the side wall of the furnace, the air seal by which the air is prevented from passing between the movable grate-bar and the fixed side of the furnace, and the construction of the chain which carries the grate-bars, also the method of securing the upper part of the grate to the lower.

Plate IV. — Side and transverse elevation of the Stirling boiler plant at No. 3 Colliery, Oneida, Pa., showing the manner in which the grate is placed under the boilers, also the arrangement for moving the grate and supplying air to the furnace.

Plates v., VI. and VII. — Photographs of one of the Oneida automatic grates. Plate V. shows the skeleton of the frame before the sheet-iron work and grate-bars are put in. Plates VI. and VII. are two views of the iron work of the furnace completed ready for shipment.

Plates VIII. and IX. — Views of the boilers at the No. 6 Slope at Eckley, Pa. Plate VIII. was made by removing the front of the boiler house, and shows the running mechanism of the grate and the blast-pipe. Plate IX. is a view of the same boilers taken from inside of the boiler house, and shows in the background the fan connected with the blast-pipe and the next set of boilers not yet finished, for which another stoker is being built.

Ftrnace For Burning Small Anthracite Cx)Aij9. 589

690 Furnace For Burning Small Anthracite Coals.

The diagram in Plate I. which I shall use for the general descrip- tion of the grate is not an exact representation of the furnace as built, it being intended more especially to explain the principle of its action.

The furnace consists essentially of a travelling-grate moving from the right toward the left. The coal which is brought to the hopper 20 by a drag, spout, or any other convenient method feeds down by gravity over the fire-brick 14 onto the travelling-grate. The coal is carried slowly at the rate of from 3J to 6 feet per hour toward the other end. In the beginning of the operation, the coal on the right-hand side of the furnace is ignited, the other part being covered with ashes or partially consumed coal. After the furnace is heated, the fire-brick 14 which we call the "ignition brick," be- comes hot, and the coal passing down under the regulating gate 21, l)ecomes gradually heated, and by the time it reaches the foot of the ignition brick is incandescent. In some cases the coal becomes hot enough to ignite soon after it passes the regulating gate 21. Under the grate there are a number of chambers made of sheet-iron which are closed on all sides except on top. The blast from the fan which is used to furnish the air is blown into the large air chamber which is the second one from the right. These air chambers are open on top, but the partitions are covered by plates 27, 28, 29 and 30. These plates are of such width that no matter what may be the position of the grate-bars 18, there is always one resting upon this plate, so that the air cannot pass from one chamber to another except by leakage along the bar. The result of this arrangement is that if we are blowing into the large air chamber with a pressure say of 1 inch water-gauge, the pressure in the next air chamber to the lefl would be about inch, the next to that i inch, and the next to thai J inch. Of course these figures are not strictly correct, and are used nierely for the purpose of illustrating, as I am now describing only the general principle of the apparatus. The pressure in the air chamber to the right would be say inch. The result of this state of afiairs is that the coal when it arrives on the grate is subjected to a pressure of blast sufficient to ignite it, but not too strong to impede ignition.

In order to regulate exactly the pressure of the air in each of the compartments the partitions are provided with registers, by the sim- ple opening and closing of which the pressure in the air chambers can be varied to suit the conditions.

As the thoroughly ignited coal passes slowly over the second coio-

Furnace For Burkino Small Anthracite Coals. 591

592 Furnace For Burning Small Anthracite C0Au3.

partment (where the air pressure is a maximum ), it bums briskly ; then slowly passing over the 3d compartment where the air pres- sure is less and better suited to the combustion of the thinner layer of partly consumed coal, the bed continues to diminish in carbon, and to be subjected to less blast, until, finally, the hot ashes are cooled off (before being dumped) by a very gentle current of air, which is heated and mingles with the carbonic oxide produced in the zone of intense combustion B and converts it into carlK)nic acid ; the object being to subject the coal as soon as it arrives on the grate to a pressure of blast which is the prepe one to ignite it ; then to burn it with a blast as strong as will produce good combustion, and as the carbon is eliminated and the thickness of the bed becomes smaller to diminish the blast to correspond to these conditions. The mass of coal remains all the time in practically the same position and condition in which it was placed on the grate, except so far as altered by the combustion. It is evident that there would be a tendency of the air to pass out between the brick rest 13 and the top of the grate bars 19, which have no coal on them, and if no provision was made to prevent it the air would pass under the air chamber along the line of travel of the grate and enter the furnace through the ash exit at 17, thus forcing a large excess of air into the s|)ace under the boiler and causing a loss in two ways: First, in the power necessary to furnish the air, and, second, in the heat carried off by the surplus of air going out th(* stack. This is avoided by having the returning line of grate pass into a water pan 36. By means of the partition 39, which passes down below the surface of the water, a water seal is obtained which absolutely cuts off all connection between the front and back ends of the lower portion of the furnace along the line of travel of the grate. The ash-pit, which is practically the part to the left of the plate 39 is closed by a door out of which the ashes are taken and the front end of the boiler is closed by a sheet-iron casing, which passes down into the water in the water-pan, thereby preventing the air from passing out between the brick ret 13 and the grate bars into the free air. There is space enough between the extreme right hand end of the water- pan and the vertical wall of the casing to allow any ashes or dirt that may accumulate in the water-pan to be taken out very easily. This is ver clearly shown in Plate VI., where the opening between the bottom of the water-[)an and the vertical casing is distinctly shown.

From this brief description the continuous action of the furnace can be easily understood. The coal passing continuously down from the

FURNACE FOR BURNINQ SMALIi ANIHRACITE COALS. 593

Vol. Xxii. —

594 furnace fob burning small ANTnuAcrrE goals.

Ignition brick is ignited gradually, l)umed out, and the ashes ae carried off or dumped hy the grate bars as they descend, as can l>e easily seen on Plate VII.

The coal burns out from the bottom, that is, the first thin layer of complete ash forms on the bottom and gradually becomes thicker until it roaches to the top. At first, the ash is very hot, but the gentle current of air passing through it gradually cook it oU) and when it is dumped into the ash-pit it is not very hot. The ahtidtii portion beginning in C and extending into D n'[>resents the ),nitlual formation of the ash, and the part to the left of that shows the anU practically cooled or cooling.

A certain portion of air from which the oxygen is not removed passes through and cools the ash, but in the first sections of ttie l>ed of fuel near A, a certain amount of carbonic oxide is formed, due to the fact that the amount of air blown through i; not sufGt-ient to properly consume all the carbon and the incandescent carbon de- composes the carbonic acid, forming carbonic oxide very much as in gas producers. This carbonic oxide is burned in the furnace by the air which has passed through the ash. Our experiments have shown us that if we allow the gases to pass through the furnace with a velocity that will permit the carbonic oxide to burn completely before reaching the parts of the furnace too cool for the combustion to take place, weget a better result, and in one of our plants we have found an increase in efficiency and economy by putting a damper in the stack and checking the flow of gases. Of course, there is a velocity for each furnace above or below which you have less economy and less efficiency, pro- vided you are burning a certain number of (>ounds of coal per hour.

Having thus briefly described the process, I will now give some details as to the construction of the grate and the method of placing it under the boilers.

One of the first difficulties we encountered in our experiments with the travelling grate, was the fact, that if we had a fire-brick side- wall there would be a tendency to form clinker along it. This clinker would retard the coal that should be carried forward, and have a tendency to break up the fire near the walls and allow the air to escape, giving considerable trouble. This has been avoided by making the sides of a hollow cast-iron bar (called the water- back), No. 15, Plate III. This bar is horiisoutal on the bottom, but the upper part rises at the rate of inch to the foot toward the front end, which is also the hottest end. The water is fed in at the back end, and flows out at the top at the front.

Our experiments with the Stirling boilers show us-that ifj we

Furnace For Burnikq Small Athracit£ 00Al9. 596

696 Furnace For Burninq Small Anthracite Coals.

pass the Feed-water which is necessary to supply the boilers tlirough this water-backy on each side, the water leaves the water-hark at a temperature of 110° to 120°. This goes directly to the feed-pump, and the heat is all utilized. We also found that the coal had a ten- dency to burn a little more rapidly along the water-backs, so that the layer of partially consumed coal became thinner there more quickly than in the center, thus allowing a too free passage to the air at that point. This has been avoided in two ways. First, by having no holes in the grate-bar at that point; second by making the water-back narrower at the top than at the l>ottomj which gives a larger quantity of coal to be consumed ahm the water-back, so that, if anything, the tendency is to have tlie layer of ashes there a little thicker than in the center. It is also imiortant that there should be i>ractically a tight joint between the end of the bar and the side along which it slides. This is aoainiptishd {as shown on Plate III.) by having a casting, 5, a portion of which, forming an in- clined plane, makes the fixed side. The joint is made by round bars of iron cut in sections about 1 foot long. These bars of iron rest on the inclined plane and roll against the end of the bar, 18. If one bar, 18, protrudes more than another, it simply pushes this iron bar, 43, back; if it recedes, the iron bar follows it down. If the bars, 43, were all cut off square at their ends, the moving grate-bars, if not exactly of the same length, might catch upon them if one pro- jected a little beyond the other ; the bars 43 are, therefore, rounded off at the end, as is shown at 43, Plate II. Since we have adopted this plan we have had no trouble whatever with the leakage of air. The grate is formed of two parts; the lower, 18, which is X-shaped, consisting of the vertical rib and the horizontal plate. The hori- zontal plate is ])erforated with a number of conical-shaped holes, wider at the bottom than at the top, as is shown in the drawing. At each end is a lug, which fits into the chain, 11, also shown in Plate III. There are two holes cast in the bar, and two holes drilled in the alternate or long links of the chain, and by means of two bolts each end of the bar is fastened to the chains. The upper part of the bar, 19, consists of square plates, which are about inches square. The holes in these are wider at the top than at the bottom. They are simply placed upon 18, being separated inch from it by three little stO|>s 56; this makes an air-space about of an inch high between the plates. The holes are so arranged that the lower ones are immediately under the center of the solid parts of 19. In this way it is impossible for the coal, no matter how fine, to roll through,

Ic

Furxace For Burnuq Small Antbbaoite Ooai. 697

698 Furnace Fob Burning Small Anthracite Coals.

as the natural slope of the coal will not reach the opentngs in 18. In order to hold them in their places, tiro diach-pitis 54, of soft iron are cast into 19. The plates 19 are simply placed in position over 18, andy with a couple of strokes of the hammer, the soft iron clinch- pins are bent as shown, thus holdine the up|)er part of the bar (irmly in its place, and allowing it to ii remove<l easily when neces- sary. It will be observed, that 19 projects a little over 18, on the left-hand side, and that 18 projects beyond 19 on the right-hand side, so that, when two complete bars are together, tbey overlap and close the joints so that no coal can fall through. By constructing the grate in this way, I he only parts exposed to the hot fire are the small square plates, 19, on top. The main or carrying-bar, 18, is pretty well protected from the intense heat, does not warp or twist, and shows, so far, no sign of giving out; this is very important. The expansion is also taken care of.

The construction of the chain 11 is easily understood from Plate I II. The two chains pass over 3 pairs of sprocket-wheels 44 in Plate If. In order to prevent any sagging or friction, these chains run on a set of rollers 12, which are carried by two roller-bearers 7 on top, and the two roller-bearers 6 on the l)ottom. The way in which the chains travel is very distinctly seen in Plate 2, Fig. II. The object of the lower sprocket-wheels at the front end is to bring the chain into the water-|)an and form the water-seal.

The grate is driven by worm-wheel gearing 31, 32, 33, 34 and* 36, Plate II., the velocity of the shaft 9 being about 1 revolution per hour, and that of the grate from to 6 feet lineal. The speed of the grate is very slow, and cannot easily be detected by the eye.

Plate II. is a photographic reduction of the working-drawing from which the grate at Oneida No. 3 was constructed.

Plate IV. shows the plant at Oneida No. 3. It consists of two 150 horse power Stirling boilers of the ordinary type to which this grate has been applied. In this case the firebrick arch 60 covers almost the whole of the grate, and the gases from the entire grate mingle at the outlet. The result of having this fire-brick arch is to keep up an intense heat over the grate, giving a chance for most of the carbonic oxide to unite with the oxygen of the free air before the gases become cold by contact with the heating surface of the boiler. It appears probable that it will be an advantage to remove the heatin|r surface of the boiler from the combustion-chamber, so that the gases will not come in contact with the cooler iron surface until the carbonk) oxide has been entirely burned and a thorough mingling of all the

IirBNAOB VOB BUBNINa SMALL AKTHBAOITK COALS. 599

S

600 Furnacb For Busking Shall Anthracite C3Als.

gases has taken place. In this cae the p*ant, which will consist eventually of several batteries of boilers is so arranged that a drag will carry the coal into a coal-hopper in front of each boiler, and that the ashes will drop into an ash-pit 50 in each battery, from which they will be loaded into a car when the pit is full by simply opening the gate at the end of the pit and scraping them out.

This drawing also shows the method by which the engine drives the fan 59 and the shaft upon which the cone-pulleys 49 are situated. These cone-pulleys drive the cone-pulleys 48 by which the worm- gearing is actuated. They enable us to change the speed of the grate without changing that of the fan, and to change the speed of the fan without changing that of the grate, as the relation between these two speeds varies with the character and size of the coal.

The main shaft of the engine by means of worm-gearing drives the drag that is to carry the coal into the coal-hopper.

The method by which the air is carried from the fan into the middle compartment of each grate is also shown on this plate. In this case the coal is fed in front and the ashes taken out at the back of the boilers.

The Plates VIII. and IX. show the fire- front of an improved set of cylinder-boilers with mud-drums etc., to which the stoker has been applied. The worm-gearing and air-pipe and also the fire-front are here shown.

In Plate IX. the fan is shown in the background, and, in front, the casing covering the end of the grate where the ashes are dumped. There is a car in the tunnel below, into which the ashes are drawn. The large fan fieds into the large air-pipe, from which the small pipe on Plate VIII. branches. There are two other similar nests of boilers adjoining, which are being supplied with similar grates. The large air-pipe isntended to supply the two remaining nests of boilers as well as the one already in operation.

There will be found on pages 604-606 a description of the Plates, in which is given a list of all the parts shown on each plate, each part being designated by a number, which is the same on alL

We have been running successfully the oldest plant about eight montlm. We have made many improvements, principally in the line of simplificationnd elimination of unnecessary parts.

Since we erected the Stirling boilers, some six weeks ago, we have been making experiments with them, using different sizes of coal. A Ublt of the results is hereto appended. We do not claim that theae rults are complete and absolutely accurate. They areoorrect

/

FURNACB FOB BITRNINO SMAIili AlffTHRAOTTB OOAM. 601

as far as tbey coaM be under the cironmstanoes. We have Dot as yet arranged to analyze our stack-gases or determine to our own sat- isfaction the moisture in our steam. We are engaged in this at present but we do not wish to give the results until we have veri- fied tbem by repeated experiments and checked up the calorimeter. The moisture is about 2 per cent. We have we think, established one fact, and that is that the size of the coal does not materially affect the number of pounds of water evaporated per pound of com- bustible. It does affect the number of pounds of water evaporated per square foot of heating-surface. As I said before, the tempera- ture at which the smaller coals burn is not as great as that developed by the larger coal, and therefore one square foot of heating-surface will not absorb as much heat when you use small coal as when you Use large ; but the economy (that is, pounds of water evaporated per pound of coal) appears to be about the same in all cases. Of course, tb commercial value at present of No. 3 Buckwheat is very much less than that of pea-coal.

We append herewith a table showing the size of mesh through and over which our pea-coal, and Nos. 1, 2, and 3 buckwheats are made.

Sise of Coal.

Pea coa]

No. 1 Buckwheat No. 2 No. 3

Over a Round Hole.

inch diameter

Through a Round Hole.

Ji inch diameter.

This paper is not as full and complete as we would wish to make it, but the time at our disposal since we got our boilers in shape has not allowed us to make as full and complete a series of experiments as we would wish ; but we think the results already obtained are of sufficient interest to justify us in presenting the paper to the atten- tion of the Institute.

We give here the record of the tests made with the plant at Oneida No. 3 and with that at No. 6 Eckley.

In the course of these tests it has been shown to our satis- faction that the best results would probably be obtained by extend* ing tlie air-chambers to as near the dumping-end of the grate as possible, and regulating by the registers the pressure (which may be very alight) in the last aip-chafflber, sa that a small amount of air

602 Furnace Fob Bubnikg Small Anthbacitb Coals.

may pass throogh the asb as near to the dump as possible. Tbe amount of carbon io the aah can, we think, be diminished materially by attention to this point.

In the new plants now under oonstruetion we are extending the air-chambers further towards the dump than we did in those with which the experiments were made. See Plate 11., where there is room for two more air-chambers.

Dimensiona and Proportions. (For tests 1, 2, 3, and A,)

Type of boiler, Stirling Water Tube.

Number, ; . Two.

Tubes, 155 S-inch tnbes to each boiler.

Square feet of heating surface, . . . . 1725 square feet each. Hone-power (by builder's rating, 30 lbs. of water from 100 F. to 70 lbs. pres- sure), . . . . 150.

Type of grate, . . Ck>xe Travelling Grate and Mechanical Stoking Fumaoe.

Size of grate tf feet wide by 9 feet 2 inches long.

Grate surface, 55 square feet.

Ratio of heating to grate surfirae, 31.4 to 1.

Kind of blast, Fan blower.

Dimensions and Proportions. (For test 5.)

Type of boiler, GylinderBoiler of Improved Setting.

Drums. 5 1 main, 34 inches diam. by 36 feet long.

2 mud drums, 34 inches diam. by 20 feet 4 inches.

Short connections, 4 10 inches diam., 4 14 inches dia.

Gne cast-iron water tube boiler, in flue, . . . .2} inches inside diam.

Heating surface :

Three main shells, 575 square feet.

Two mud drums, 393 "

Eight connecting tubes, 21

Gne cast-iron water tube boiler, 873 "

1862 "

Type of grate, Goze Travelling Grate.

Size, 7 feet 6 inches wide by 9 feet 2 inchea long.

Grate surface, 68.75

Ratio of heating surface to grate surface, 27.1 to 1

Manner of Conducting Teds.

This' type of grate is admirably adapted to the purpose of Boiler Tests, as there is no need of starting fire with wood, or cleaning fire at starting or stopping of test ; as the fire can be maintained in

Fubnacb Fob Bubnino Small Anthracite Coals. 603

Si

c4c eoQOf

s

g

Is

to6r-i eo 1-J t>I eo c4 CO 54 1-; iij-i ;o c4 i-J -jco ' c4 lo c4 oi i-h

'1

s

,

:ij

ss

S5 g2S:|v

"'1

g

1

604 Furnacb Fob Bubninq Shall Anthracitb Coals.

exactly the same condition throughout the run of a whole week or month.

The hourly records of ooal fired and water evaporated show that as close results can be obtained with this grate in a six or eight hour test as in a twenty-four hour test on hand-fired and hand- cleaned grates.

The tests were started at 8 a.m. ; the first of the coal to be tested was delivered into clean feed-hoppers at 4 a.m., to allow the fireman to get the right air-pressure and speed of grate for a given horse- power, and to have as much as possible of the grate covered with fire and yet as little carbon as possible carried 9ver into the ash-pit. The boilers were run continuously throughout the week, but during the tests the steam consumption at the colliery was only from 60 to 60 per cent, of that generated, the remainder being blown ofi at some distance from the boilers.

Hourly and half-hourly observations were made.

The water was weighed in a barrel placed on a platform-scale, and was fed from the weighing-barrel into an iron tank, 34 inches in diameter, by 12 feet long, set on end into the ground, and projecting 26 inches above floor-level. From this tank the feed-pump was supplied. The feed-pump was run continuously. At the time of starting the test the level of the water in the boilers was arranged so as to produce a flutter at a certain one of the gauge-cocks, and, at the same time, a mark was made on the gauge-glass. This level was readily kept, and at the time of observations the level was checked by both the gauge-cocks and gauge-glasses. At the same time the level in the feed-tank was brought to a certain height indicated by a straight-edge laid across the top of the tank. This straight-edge had a nail pointing down to the water surface. Between observa- tions no heed was taken to the level of water in the feed-tank.

The coal was weighed in a nail-k, each keg being levelled off with a straight-edge. A sample of coal was taken by taking a handful from every keg before levelling it.

Moisture determinations were made by spreading the coal out on pieces of sheet-iron and drying it in the sun or over the boilers. Several kfulsof coal levelled off were also weighed after being air- dried, from which the weight of dry coal fired was calculated.

Plate I. — List op Parts.

A. — Ignition period over air-spaoe of moderate pressure.

B. — Full combustion period over air-space of maximum pressure.

FURNACE FOB BUBNINO 8MALL ANTflBACITB 00AL8. 605

C. — Last stages of combustion ov air-space of decreasing pres- sure.

D. — Period during which last trace of carbon is consumed and during which heat is extracted from ash.

Numbered Paris. — 9, top sprocket-wheel shafts; 10, lower sprocket-wheel shaft ; 11, endless carrying chain ; 13, igni ting-brick rest; 14, igniting-brick ; 17, exit or rear wall-supporting plate; 18, grate-bar bearer; 19, grate-bars; 20, feed hopper; 21, feed regula- ting gate; 22, air pan ; 23, 24, 25 and 26, Ist, 2d, 3d and 4th par- titions respectively in air pan ; 27, 28, 29 and 30, 4th, 3d, 2d and 1st air cut-off plates respectively; 36, water pan; 39, plate, which dipping into water, separates the space outside and below the com- bustion chamber into the two spaces A and B, Plate II., and forms the air seal ; 41, bridge-wall ; 44, sprocket wheels for chain : 45, ignit- ing arch.

Plates II and III. — List op Parts.

Numbered Parts., — 1, sole plate for main posts ; 2, posts at dump- ing end ; 3, posts at feed end ; 4, center posts ; 5, bars forming tie for posts, rest for water-back and support for round iron bars 3, resulting in air seal at side walls or ends of grate bars; 6, bottom roller bearers ; 7, top roller bearers ; 8, pedestals ; 9, top sprocket- wheel shafts; 10, lower sprocket-wheel shaft; 11, endless carrying chain ; 12, rollers, for carrying chain and grate bars; 13, igniting- brick rest ; 14, igniting-brick ; 15, water-backs forming sides of fur- nace; 16, special fire-brick over ash exit; 17, exit or rear wall- supporting plate; 18, grate-bar bearers; 19, grate-bars; 20, feed hopper; 21, feed rulating gate; 22, air pan; 23, 24, 25 and 26, 1st, 2d, 3d and 4th partitions respectively in air pan ; 27, 28, 29 and 30, 4th, 3d, 2d and 1st air cut-off plates respectively ; 31, worm on horizontal shaft ; 32, worm-wheel on vertical shaft ; 33, worm- wheel shaft (vertical) ; 34, worm on vertical shaft ; 35, worm-wheel on main sprocket-wheel shaft ; 36, water pan ; 37, door in casing to remove grate-bars ; 38, end of casing at feed end ; 39, plate which dipping into water, seimrates the space outside and below the com- bustion chamber into the two spaces A and B and forms the air seal ; 40, cleaning doors in air pan ; 41, bridge-wall ; 42, skew-back on post at feed end ; 43, round bars bearing against ends of bar- bearers and preventing leakage of air; 44, sprocket wheels for chain ; 45, igniting arch ; 46, air pipe; 52, cover separating ash pit from gas chamber back of bridge-wall ; 61, end of water pan ; 62, blank plate forming air cut-off at ash end.

606 The Htdbogen-Oil 8Apbty-Lahp.

Plate III.

bars formiog tie for posts rest for water-back and support for round iron bars 43 resulting in air seals at side walls or ends of grate-bars; 11 endless carrying chain; 15 water-backs forming sides of furnace; 18, grate-bar bearers; 19, grate-bars; 43, round bars bearing against ends of bar bearers and preventing leakage of air; 54, clinch pins to secure short grate-bars to bearer; 55, stops; 56, overlap of grate-bars ; 57, vertically inclined wall of water- back ; 58, seat for air seal.

Plate IV.

32, worm-wheel on vertical shaft ; 33, worm-wheel shaft (verti- cal); 34, worm on vertical shaft; 37, door in casing to remove grate bars ; 38, end of casing at feed end ; 44, sprocket wheels for chain ; 46, air pipe ; 47, ash-pit doors ; 48, cone pulley on horizontal worm shaft; 49, cone pulley on driving shaft; 50, ash-pit ; 51, brick floor of gas chamber; 52, cover separating ash-pit from gas chamber back of bridge- wall; 53, ash car; 59, fan blower; 60, combustion arch.

TH£ HYDBOQENOIL 8AFETY LAMP, FOB HQHTING AND FOB ACCUBATE AND DELICATE DETECTION AND MEASUBEMENT OF INFLAMMABLE 0A8 AND VAPOB IN THE AIB.

By Prof. Frank Clowes, Universiit Colleqs, Nottinqham, England.

(Chicago Meetings being part of the International Engineering Congress, August, 1893.)

I. Introductory Remarks.

This lamp has been devised to bum oil from a flat wick in the usual way for lighting-purposes ; and also to burn a hydrogen-flame of standard size instead of the oil-flame, when delicate and accurate gas- testing is to be carried out. The change from the oil-flame to the hydrogen-flame, and vice versa, can be made with great rapidity and without opening the lamp or running any risk in the presence of gas.

The oil-flame serves for illumination ; and when the wick is drawn down by the '' pricker," so as to abolish the light, the pale blue reduced oil-flame serves to detect fire-damp or gaV' in any propor-

'

The Hti>Sogenoil Safety-Lamp. 607

tion between 3 and 6 per cent., and to measure such proportions with fair aocaracy.

The hydrogen-flame, set to standard size, detects gas when present in proportions varying from 0.2 to 3 per cent, and measures such proportions with precision.

The presence of gas is detected by the appearance of the pale "flame-cap;'* its proportion is estimated partly by the character of the cap, but mainly by its height. In order to render the cap more easily seen, a vertical strip of the interior of the lamp-glass, about an inch in breadth, is smoked by a wax-taper. This is arranged to form a background against which the cap is viewed, and serves to throw up the cap and to prevent its obliteration by cross reflections from the smooth glass-surface.

The height of the cap may be estimated with (nsiderable accuracy even without measurement. A vertical wire scale, however, may be arranged on the wick-holder, so as to stand in front of the flame and the cap ; the cross-wires on this scale, when seen against the flame or cap, give actual measurements. Each cross-wire is ar- ranged to pass in front of the cap at a height of 0.2 inch from its top; since the wire is invisible at the top of the cap owing to the feeble lighting-power at that level, the illumination of a scale by the flame and cap is impossible ; the light is wholly insufficient for the purpose. The method pursued, and the apparatus employed for originally measuring the cap-heights are described below. It should be specially noted that the hydrogen test-flame was in each experi- ment set to standard height in the presence of the gas which was being estimated. It is accordingly Unnecessary to obtain gasfree air in order to obtain the standard flame when testing any atmos- phere. The flame is set to its standard height in the air which is being tested, the cap-height is at once read off, and can be translated into percentage of gas by reference to the scale.

A careful series of comparative tests has fully confirmed the opinions expressed by Pieler and by Mallard and Le Chatelier, that no flame equals that of hydrogen in delicacy for detecting gas in the air. The writer has succeeded in introducing this flame into an ordinary safety-lamp, so that it can be used alternately with the oil- flame ; and considerable experience of its use both in the laboratory and in the coal-mine has proved the apparatus to be of a thoroughly practical nature. The writer has also carefully mapped out the in- dications of the hydrogen- and of the oil-flame by means of his specially devised test-chamber. The mapping of the hydrogen-flame

THE HTDBOaEN-OU SAFETY-LiLlCP.

ipdications is entirely new einoe the impassibility of iDtrodueiDg the hydrogen-flame into a safety-lamp had previously rendered such measurements useless.

Careful measurements have also been made of caps produced over the hydrogen-flame and the reduced oil-flame in known percentages of coal-gas, of water-gas, and of petroleum vapor. The hydrogen- test proves to be of extreme delicacy and accuracy when applied to the detection and estimation of all these gases and vapors, and in the case of petroleum-vapor is already applied towards the examina- tion of the atmosphere in petroleum-stores, and in the tanks of petroleum cargoteamers.

II. Description of the Hydrogen-Oil Safety-lamp, and of THE Method op Using It.

The lamp to which the hydrogen-test is attached is by preference a modification of the Gray lamp, which takes its air-feed from near the top, and delivers the air around and beneath the flame. Figs. 1, 2 and 3, show the lamp and attachment.

The hydrogen is contained in a small steel cylinder, easily carried in the pocket. This cylinder can be instantaneously attached to the

Fig. 1.

Portable Hydrogen-Oil Safety-lamp.

lamp, and on opening a valve it discharges its gas through a narrow copper tube which traverses the oil-reservoir. The end of this tube is brought up parallel with the wick>tubeand dose to and just above the wick. Accordingly, when the gas is turned on, it is at onoe

Ic

The Hydrogen-Oil Safety-Lamp.

kindled by the flame of the lamp. The lamp is prepared and used

Fio. 2.

Portable Hydrogen-Oil Safety-lamp Hjdrogen-CyliiMler Detached.

in the ordinary way for lighting-purposes; a mixture of xx)l7a <h1 with petroleum (water-white) oil in equal measures being used.

Fig. 3.

Portable HydrogenOil Safety-lamp, witli Attached Hydrogen-Cylinder (Diagrammatic Section). VOL. xxil.— 39

The Hydrogen-Oil Safety-Lamp.

When employed for gas- testing, an opening on one of the side- pillars is exposed by sliding up a sleeve. This hole is closed by the thumb. The lamp is then gradually raised towards the roof, and the behavior of the flame is watched. Should the gas be in sufiBcient quantity to threaten the extinction of the lamp-flame, the. thumb is withdrawn, and the supply of air, being taken from a lower level, prevents the flame from being lost.

Fig. 4. Standard Flame.

Percentage of Methane (Fire-damp). Actual Heights of Cape over Hydrogen-Flames.

If the percentage of the gas is to be measured the wick is drawn down by the pricker until the flame just loses its bright tip, and if a cap is seen, its height serves to measure with some approach to accu- racy the proportion of gas, according to a scale given below.

If no cap appears over the reduced oil-flame, the absence of gas is not proved, since less than 3 per cent, is not indicated by this flame.

The pocket hydrogen-cylinder is then attached to the lamp; the cylinder serving as a handle is grasped in the left hand, while the hydrogen gas is slowly turned on by means of a key applied to the

The Hydrogen-Oil Safety-Lamp.

cylinder-valve by the right hand passed round behind the lamp. A tongue of flame shoots up from the bright flame as the hydrogen enters; the wick is then drawn down until the oil-flame is ex- tinguishedy and holding the lamp with the hydrogen-flame on a level with the eye, the flame is set by means of the cylinder-valve to 10 millimeters (0.4 inch) by viewing it behind the standard wire scale. The height of the cap, if any, is then noted, and measures

Fig. 5.

2 3 4 6 6

Percentage of Methane (Fire-damp). Actual Heights of Caps over Colza-Petroleum Flame. The pale blue flame, about 3 mm. in height, is shown.

The partly luminous flame, giving maximum caps, is not shown, but the cap heights over it are dotted in.

the percentage of gas, according to a scale given below. If no cap is seen, the gas is less than 0.2 per cent in amount.

To bring back the oil-flame, it is simply necessary to push up the wick, which is at once kindled on touching the hydrogen-flame. The- hydrogen gas may then be shut off, and the cylinder detached and replaced in the pocket until it is again required.

When using the lamp in the mine for the detection and measure* mentof gas, the standard hydrogen-flame is thus made to supple- ment the reduced oil-flame, and the two flames carry the indications

The Hydrogen-Oil 81Fety-Lamp.

from 0.2 up to 6 percent, of gas.* But by varying the height of the hydrogen-flame used, this flame may easily be made to indicate and to measure gas through the whole range, as is shown in a subse- quent diagram.

The ordinary safety-lamp is thus made to serve for lighting, for rough gas-testing in the usual way, and for delicate testing in the return air-ways of the mine. The advantage of the delicate testing is that it enables a general opinion to be formed quickly as to the

Fio. 6.

Percentage of Methane (Fire-damp).

Cap-Heights with Hvdrogen-Oil Lamp. (Actual Size.)

Ht 15 mm. hydrogen flame. H 10 mm. hydrogen flame. H, 5 mm. hydrogen flame. Oi Maximum oil-flame. O Pale blue oil-flame.

condition of the ventilation of the mine by the examination of the returns.

The pocket hydrogen-cylinders are rapidly charged by attaching them by means of a connecter to a large store-cylinder of hydrogen under pressure. If such a store-cylinder is not kept at the mine the

The Hydrogen-Oil Safety-Lamp.

small cylinders travel easily by post to some center, where they may receive their charge. A full charge serves for about eighty separate tests, each of which occupies about thirty seconds; each lamp is fur- nished with two pocket-oylindersi so that one may be used while the other is being recharged.

Fio. 7.

Percentage of Gas.

Heighu of Caps over Hydrogen-Flame in Coal-Gas and in Water-Gag, The dotted curves are for water-gas. Hj 15 mm. hydrogen flame. H Standard 10 mm. hydrogen flame. H, 5 mm. hydrogen flame. O Oil-flame giving maximum cap.

X. B. — The tail or streak above the cap was not measured in for 5 and 6 per cent, of water-gas.

III. Advantages op the Hydrogen-Flame.

The advantages gained by the use of the hydrogen-flame for gas- testing are in the main the following :

1. The flame is non-luminous, whatever its dimensions may be; it therefore does not interfere with the perception of the cap. The reduced oil-flame is never non-luminous; and the alcohol-flame gives as compared with the hydrogen-flame, an amount of light which seriously interferes with the perception of the cap.

The Hydrogen-Oil Safety-Lamp.

2. The flame can always be adjusted at once to standard height, and maintained at that height sufficiently long for the completion oi the test. Other testing-flames are constantly varying in dimensions, and most of them cannot be set to standard size at all with any cer- tainty. Thus a Colza petroleum-flame exposed in air containing a low percentage of gas, when twice adjusted, gave caps of 8 and of 20 mm. The reduced oil-flame oflen fell so quickly that cap- readings with low percentages of gas could not be taken at all.

3. The caps produced over the hydrogen-flame re larger than those produced by any flame of corresponding size.

4. The size of the hydrogen-flame can, therefore, be so far redaced as to enable it to be used in an ordinary safety-lamp. The size of the flame may further be suitably varied in size, so as to increase or decrease the height of the cap. This serves either to increase the delicacy of the test or to extend its range.

5. The hydrogen-flame shows no trace of mantle or cap in air free from gas. The Colza petroleum and the benzoline flames show pale mantles in gas-free air, which may be easily mistaken for a low per- centage of gas. The Pieler alcohol-flame also shows a pale, tall, thread-like mantle in gas-free air, which might be mistaken for a low percentage of gas.

6. The standard hydrogen-flame burns vigorously, and is of fair size; therefore it cannot be extinguished by accident. The reduced flames ordinarily used in testing burn feebly and are readily lost

7. Hydrogen is supplied practically pure, Oil and alcohol are apt to vary much in composition, and, therefore, to give flames whose indications vary with the sample of liquid which is being burnt.

Cap-HeiglUs in Millimeters in Fire-Damp (Methane).

Percentage

of Methane.

Hydrogen-flame.

Colza-petroleum flat flame.

Standard 10 mm.

15 mm. in thegafl.

5 mm. in the gas.

Small blue 3 mm.

Flame part- ly lumin'us 6 mm. 1

Enters top

of lamp.

Enters top of lamp.

Enteretop .

The Hydrogen-Oil Safety-Lamp.

IV. Cap-Measurements Obtained for Hydrogen- and Oil-Flames.

The accurate measurement of the caps appearing in air which con- tains fire-damp over the flames used in the hydrogen-oil safety-lamp, has yielded the results tabulated on the preceding page. The flames and caps are also represented drawn to actual size in the diagrams, Figs. 4, 5 and 6.

Multiply millimeters by 0.04 to convert them to inches.

CoalnOae and Water-Gas. — The following were the indications yielded by the hydrogen-oil safety-lamp when testing with it fur coal-gas and water-gas in the air :

Cap-Heights in Millimeters due to Coal-Oas.

Percentage of Ccval-Gas.

Gap-heif ht 1

n mm. over bydrogen-flame.

Cap-h't. in mm.

over oil-flame

reduced until

the cap 18 at

maximum.

Standard 10 mm. flame.

Flame raised Flame reduced to 15 mm. to 5 mm In the Gas. in the Gas.

3.0 .

Enters top.

65.0 (over).

Cap- Heights in MiUimeters due to Water- Gas.

Percentaire of Water-Gas.

Hydrogen-flame.

Colza-petrol'm

flame, maximum size.

Standard 10 mm. flame.

Flame raised to 15 mm. in the gas.

Flame reduced

to 5 mm. in

the gas.

Nil.

t4

M

ao

36*

40*

50.6*

Nil.

t4

s

8 !

In these measurements the tail-like prolongation of the conical cap wi|s not included. IC

The Hydbogen-Oil Safety-Lamp.

If these heights are required in raches, they should be maltiplied by 0.04.

Cap-Heights in Petroleum- Vapor. — In order to test the delicacy of the hydrogen-flame as au indicator of petroleum-vapor air satu- rated with benzoline-vapor was mixed in known proportiou with fresh air and the safety-lamp was introduced iuto the mixtare. The results are tabulated here, and prove that the hydrogen-flame detects of the amount of benzoline-vapor which naakes air explo- sive, and -j. of the amount which b inflammable in air.

Proportion of

benzolized air

to Mr for

mixture.

Behavior of the Mixture with ft naked flame.

Height of the Cap over the hvdrogeu safety-

ture.

1 :4

Violently explosive.

1:5 1:6 1:7

[Burns rapidly, and would probalily be explosive if fired in large quan- tity.

1:8

Bums round a flame only.

1:9

Non-inflammable.

1:23

"

52 mm.

a

r43mm.

4 mm. with Ash- worth's benzo- [ line-flame.

1:72

M

31 mm.

1 :144

M

32 mm.

V. Special Apparatus Used in Measuring Cap-Heights.

In order to enable the cap-heights to be measured in the labora- tory a SQEiall test-chamber was constructed in wood, of precisely 100 liters capacity. It had a glass window in front for examining the lamp-flame, an opening below for the introduction of the safety- lamp, and an opening above for renewing the atmosphere. The upper and lower openings were closed air-tight by water-seals. When opened simultaneously, they renewed the atmosphere of the chamber in two minutes. When the air inside the chamber as to be charged with a certain percentage of gas, the requisite volu me of the gas was introduced from a small gas-holder, and was mixed with the air by means of a large, light flapper, worked by a bnl®

The Hydrogen-Oil Safety-Lamp.

from oQtside. The mixture was effected in a few seconds by this

Fig. 8.

Test-Chamber.

means. A gas lighter than air was fed into the top of the chamber, a gas heavier than air was introduced near the floor of the chamber.

Fig. 9.

/ f

Test-Chamber (Section, Front).

The wooden chamber was blackened on the interior, and was made gas-tight by brushing over the inside and outside with melted par-

THE HYDROGEN-OIL SAPFfY-LAMP.

affin-wax. In a long series of experiments the test-chamber per- formed its duty admirably and rapidly.

The Figs. 8, 9 and 10 serve to render the construction of the test-chamber evident.

The cap-height was measured by pressing an ordinary flat paral- lel-ruler against the window of the chamber, and adjusting it until the cap was just included between the rules ; the intervening space

FiQ. 10.

Test-Chamber (Section, Hide).

was then marked on a piece of paper pressed beneath the rule, and the distance was read off on a millimeter scale, and corrected for parallax.

By specially-devised apparatus it was proved that the cap-height was independent of the movement of air around the lamp, even when the velocity of the air far exceeded that of the ventilation-cur- rent in the mine. The ordinary amount of coal-dust in the air of the mine was also without disturbing effect on the test.

. Discussions;

[Note. — The following discussions of papers contained in this volume are printed without regard to order of succession. It has been impracticable to secure from the various parties, corrected reports of their contributions in time to effect any particular desired arrangement in this respect.

The papers of Messrs. Pourcel, Osmond and Sauveur, were con- sidered at Chicago together with others of cognate character ; and the whole discussion will be found in vol. xxiii., under the head, "Physics of Steel."]

Tbe Lead- And Zinc-Deposits Of The Mississippi

Valley,

Discussion of the Paper of Dr. Jenney. (See page 171.)

William P. Blake, New Haven, Conn., and Shullsburg, Wis. In a memoir upon the progress of geological investigations and surveys of the State of Wisconsin, particularly of the lead- and zinc-region, which I presented to the Wisconsin Academy of Arts and Letters in December, 1892, I reviewed the published opinions of leading geologists in that field, particularly with regard to the origin of the ores and the existence of faults and breaks in the strata, and the relation of such disturbances to the ore-deposits.

It was shown in that paper that Percival, the geologist and poet, had recognized as early as the year 1856, or before, the existence of dynamic disturbances in the form of faults at many places in the lead-region and the connection of such faults with the localization and origin of the ores.

Percival also advocated the then commonly accepted theory of the deep-seated origin and upward flow of mineralizing solutions, by which the lead- and zinc-ores were deposited.

It was also pointed out 4hat, after the death of Percival, Prof. Whitney and Prof Chamberlin in succession differed in ioio from Percival and failed to find or to recognize any faulting or disloca- tions of the strata. This brought the subject down to the close of the last geological survey of Wisconsin in 1879.

The views of Prof. Whitney and my own conclusions, based on observations during more than a year were stated as follows :

" It should be noted that Whitney did not disease the cause of the linear dis- tribution of the mineral -bearing crevices, their origin and relation to the lines of uplift which had so greatly impressed Percival, nor did he find any faults in the strata which Percival had so specifically noted and described ; Whitney's statements relative to these subjects being : ' There is no evidence in the lead-region of the de- posits of ore or the crevices being situated over or near faults or dislocations of the surface, or of being in any way connected with subterranean or deep-seated move-

These remarks of Prof. Blake were issued as a separate pamphlet, under the title, "The Existence of Faults and Dislocations in the Lead and Zinc Regions of the Mississippi Valley, with Observations upon the Genesis of the Ores.''

622 Lead- And Zinc-Deposits Of The Mississippi Valley.

menta of the crust of the earth such as would allow of the metalliferous solutions having access from below.'*

This is remarkable, inasmuch as dislocations do exis(, as Percival stated, and that they do certainly appear to have some relation to the mineral deposits — a rela- tion which needs investigation/'f

It was my purpose to follow ap this subject and to present later in detail some of the phenomena of faulting, which I have since found are common to both the Wisconsin and the Missouri zinc- and lead- regions.J

Being therefore in accord with Dr. Jenney in this one point of the existence of faults in both regions, I am much pleased to find his confirmation of my conclusions as shown in his brilliant paper before us ; but while I recognize the existence of an occult relationship of these disturbances to the distribution of the mineral deposits, I am not yet prepared to accept his theory of the mineralization of the rocks by solutions ascending from deep-seated sources, for the evi- dence so far is insufficient to warrant this solution of the question, especially as there is also much evidence in favor of the view of local lateral secretion, and the descent of solutions from higher to lower horizons.

The General Relation of Mineral Deposits. — Dr. Jenney's paper brings before us many problems of earth-structure and the phe- nomena of mineral veins. In the introductory portion he seeks to show a direct relationship of the mineral deposits of the Missis- sippi valley to those of the Rocky Mountain system, carrying the precious metals, and to establish a theory of their universal common origin. Now, while this may be true of all the elements in the very broadest sense, the generalization overrides and subordinates the very strongly- marked difference between deposits of lead- and zinc-ores of the Mi-ssissippi valley, free from the precious and most other metals, and the ores of the crystalline rocks and other forma- tions of the Rocky mountains where the precious metals occur in close association with those of lead and zinc.

Tlie Correlation of Uplifts. — In support of the theory of com- munity of origin, attention is first directed to the series of" uplifts,"

J. D. Whitney, Geology of Wiscondn, vol. i., 1862, p, 393.

t Memoir above cited, presented December, 1892; published September, 1893.

X The existence of dislocations of the strata is shown at the Oswego tract, Jop- lin, Mo., where the coal-measures, and even coal, are found together with lead* and Einc-ores in juxtaposition with older rocks, and in general northeast and southwest courses.

Lead- And Zinc-Deposits Of The Mississippi Valley. 623

so-called, in which the deposits occur, namely the Wisconsin, the Ozark and the Ouachita uplifts, while the Cincinnati axis is not overlooked.

At first, this correlation of the uplifts is somewhat dazzling and distracting ; but when we find from Dr. Jenney's clear description (based upon Prof. Branner's) of the Ouachita uplift, that it differs radically from either of the other uplifts, being a well-defined axis of pronounced plication with an east-and-west trend, and accom panied by mineral veins of the usual type, with argentiferous and auriferous lead- and zinc-ores, and on the other hand that the Ozark uplift and the Wisconsin uplift are but gentle swells of the strata, still retaining practically their horizontal position, while the ores in both are similar— clean zinc- and lead-ores,* without silver or gold-, we easily recognize the well-known parallelism between the forma- tions of the Ozark and Wisconsin regions, but fail to find their unity with, or relationship to, the Ouachita.

In these remarks, therefore, I prefer to leave out of view the dis- cussion of the phenomena of the Ouachita region and of the deep lead-mines of Missouri, and confine myself chiefly to a discussion of the origin of the lead- and zinc-deposits of Wisconsin and of the Ozark region, between which direct comparisons based upon simi- larity of formation and of ores can be made.

In the Missouri lead- and zinc-region, and also in the Wisconsin region there is practically no " uplift," in the usual acceptation of the word. Chamberlin characterizes the so-called Wisconsin uplift as a " very gentle upward bending of the strata attended by compen- sating depressions on either hand."* He also compares it to the Cincinnati axis formed at the end of the epoch of the Hudson River or Cincinnati shale, and the comparison is well summarized by the sentence : mountain folding at the east, a less forcible arching in western Ohio, and a still more gentle flexure in Wisconsin."! But whatever significance may attach to the fact of even a trace of flex- ure of the formations, the structure must be referred to the Appala- chian system, rather than to that of the Rocky Mountains, the gen- eral axis of structure being northeasterly rather than northwesterly, while in the Ouachita the structure is well-defined and trends west- erly, belonging to the Rocky Mountain system, and having nothing in common with either the Missouri or the Wisconsin lead- and zinc- region.

Wisconsin Geological Reports, vol. i., 1883, p. 174-75. f P' 175.

624 Lead- And Zinc-Deposits Of The Mississippi Valley.

But Dr. Jenney himself directs attention to the "marked con- trast of the complex structure of the Ouachita uplift with the sim- ple formations of the other areas of upheaval in the Mississippi valley" (page 179) thus nullifying to a great degree his claim of unity and correlation in the uplift and their attendant phenomena.

The field of comparison, then, may be narrowed down to the two notable zinc- and lead-regions characterized structurally by nearly horizontal strata.

We shall not, however, lose sight of the fact that the traces of structure in the Wisconsin region, while partaking of the direction of the great Appalachian folding, are at the same time parallel to the most ancient shores of Laurentian and Huronian time, and to .the long northeasterly trend of the Keweenawan of Lake Superior. The structure is parallel to the western shore of Lake Michigan, to the depression marked by Green Bay and Lake Winnebago, and, finally, to what is called the Baraboo axis, extending northeasterly from Baraboo, a short distance north of the chief lead- and zinc- region. The original foundation rocks would thus seem to have given rise to, and to have determined, the structure and its direc- tion.*

The Baraboo Axis of Wisconsin. — The Baraboo axis appears to terminate, or to be turned westward, north of the Wisconsin river, but if it were prolonged southwesterly in its general direction from the northeast, it would pass directly through the center of the lead- and zinc-region. This did not escape the observation of Percival, who wrote:

The theories of the origin of folded and plicated strata, though involved in the subject of the paper, need not be discussed, but it is well to add the views of Prof. Chamberlin, Qed. TTis., L, 1883, p. :

Gathering these observations together, it appears that the axes of the folds have a general trend from north of east to south of west, and with this the strikes of the series in the adjacent regions of Michigan, Canada and Minnesota generally agree.

'' According to the law of folding and uplieaval already indicated, the disturbing force is to be sought along lines at right angles to these axes, either to the east of south or to the west of north. In the latter direction, however, we encounter the great Laurentian belt that stretches northwestward to the Arctic sea, from which an active force could hardly be expected because of its previous solidification and the fact that it was undergoing denudation. In the opposite direction was an extensive sea, whose bottom constituted a broad flat arch. On this, near the land, there had accumulated heavy sediments, causing it to sag, and hence to assume an attitude

unfavorable for resistance The settling of the arch, to accommodate itself

to the shrinking earth, presumably caused it to exert a powerful lateral pressure, and the sagged portion yielded and was compressed and crumpled."

Lead- And Zinc-Deposits Of The Mississippi Valley. 625

I have examined at different points an extensive range of sienitic rocks not laid down in former maps and reports traversing the country from the south side of Fox river through Marquette and Waushara counties, and apparently, from its ar- rangement, having an important bearing on the phenomena of the lead district as well as on the general arrangement of the secondary strata south and east. The extensive ranges of gray quartzite in the Baraboo country and east of Portland have also an apparent connection with the same."

And while writing these pages a paper by Prof. Van Hisef comes to handy showing the existence of a vertical-shear zone in these Baraboo beds; a clearly defined plane of faulting, trending, seem- ingly by his description, from east to west.

Here, then, is additional evidence of the value of the observations by Percival, who may be said to have fallen at his post in the front of the battle for knowledge of the origin of the lead- and iiinc- deposits of Wisconsin and in defence of his belief in the existence of

Faulis and Dislocaiiona in the Lead-RegUm. — He wrote in 1856 :

"The opinion expressed in my former report that the mineral was derived from beneath, is strengthened not only by the general results of my observations in the diggings, but by the appearance of disturbance in the strata, particularly along the line of the great body of mineral traversing the middle of the district, and by the relation in the bearing of that body to the extensive ranges of primary and metamorphic rocks towards the northeast, indicating that the mineral may have arisen from a mass of such rocks beneath the secondary strata/'

He had already in the report referred to, shown that there was a great degree of order in the surface arrangement or direction of the mineral deposits, that they were disposed in extensive series of par- allel groups, norths and souths and easts and wests, or rather north* easts and southwests. He made a map, on which he indicated the direction and grouping of the chief deposits by parallel lines and pointed out that a " systematic order presents itself pervading the whole district which indicates that the mineral deposits are not casual but regularly arranged. This may be regarded as an impor- tant confirmation of the facts already stated in relation to the arrangement of the mineral in veins ; " and again, p. 91, " there is a degree of orderly arrangement in the succession of the diggings such as to indicate that they are not merely casual deposits but parts of a connected whole;" and on page 100 he writes:

J. O. Percival, Second Report, 1856, p. 12.

t C. R. Van Hise, "Some Dynamic Phenomena Shown by the Baraboo Quartzite Ranges of Central Wisconsin." — The JoumcU of Otology t No. 4, p. 347, May- June, 1S93. Published at end of July, 1S93. X Setxmd Annual Report, 1856, p. 63.

VOL. Mil.— 40 ,

626 Lead- And Zixodep06It8 Of The Mississippi Valley.

"The traces of order and connection in -the surface arrangement appearno less remarkable than in the vertical arrangement* What I have here given is only a small part of what might have been stated ; but I trust it will suffice to show that the ranges in their bearing and in their grouping from the smallest to the most ex- tended combinations have been governed by some general laws and have not been merely local accidents."

Percival had before noted the existence of faults and disturbances of the strata in the lead-region and enumerates no less than five well defined examples chiefly along the Fievre river and Sbullsburg Branch and the West Pecatonica near Mineral Point; also, on the East Pecatonica, and on Grant river.

Had Percival lived there is no reason to doubt that he would have elaborated and worked out this subject. His posthumous report gives evidence of this in repeate<l references and in the additional facts given, especially in his descriptions of points of depression as well as of elevation, wherein he pointed out that the mounds occupy centers of depression and further describes a line of fracture along the Pecatonica with an abrupt elevation of the strata on the south.

A synclinal depression, accompanied by a brecciated condition of the strata, was noted by Dr. J. P. Kimball, Prof. Whitney's assist- ant, at the Peaslee diggings. It is described as a synclinal axis along the ridge with the strata dipping to the center at an angle of 4°.*

Theories of Ore-Deposition. — Dr. Jenney in his " Discussion of the Theories of Ore- Deposition," cites the theory of lateral secretion ad- vanced by Prof. Whitney, and says that the results of his investiga- tions 'M)ave led to a different conclusion, not only as to the origin of the metals, but also as to the manner of formation of the ore-de- posit " (p. 220). He regards the fact that the ore-deposits are associ- ated with faulting-fissures of indefinite depth as the strongest evi- dence in favor of the view of the ascension of the mineralizing solutions from below, and says that the 'Mocalization of the ore-de- posits is difficult of explanation by any theory of lateral secretion" (p. 220), and that all the deposits are formed by ascension (p. 223).

All the phenomena with which I have become familiar in the past two years of active work in the Wisconsin region point to a downward flow of the metal-depositing solutions rather than to any upward flow. The form of the deposits indicates this. Although we may have a nearly vertical fiasure or "crevice" in the upper parts of the Galena dolomite more or less filled with carbonate of zinc and crystals of galenite, the ore generally " takes a pitch " and forms

J. D. VVhitnev, Geol. of WUconsin, i., 1862, p. 299.

Lead- And Zinc-Deposits Of The Mississippi Valley. 627

on both sides of a middle rock or mass of the strata not cleft by the fissure and the fissure is lost ; it does not "go down," and by follow- ing the ore downwards irregularly we usually find it ending in a flat sheet spreading out like a foot upon an impervious layer of clay or clay shale. A sheet of blue clay is significant of ore resting upon it at some point. Again, in caverns containing dry- bone, if the sheet of bone upon the floor of the opening is examined before it is broken up, the lines of drip of solutions from a network of crevices above can be clearly traced. This is evidence of, at least, a secondary flow and ore-deposition, by percolation from above downwards. . And the evidence is complete that all or nearly all of the zinc carbonate was once in the condition of blende, and that it has been in the state of solution probably as sulphate, and has been redeposited by con- tact with the calcareous beds as carbonate, with the concurrent for- mation of sulphate and carbonate of lime and sulphate of magnesia.

Jndigenoua Origin or Deposition. — The original impregnation of the rocks with the ores is another question. Much may be said in favor of each of the various theories. The evidence as it stands to-day, to me, is in favor of the view of, primarily, a general dis- semination of the ores in the mass of the strata, from which during decomposition they are drawn or leached and finally concentrated in the fissures or crevices and caverns in the vicinity, most of which openings are probably formed by solution of circulating waters charged with carbonic acid while assisted by the solutions of sul- phates of zinc and iron, which last passing below the permanent water level and thus escaping further oxidation, are reduced to sul- phides and deposited as such.

The close adhesion in all cases of these sulphides to the adjoining dolomite, without any clay parting, and the frequent envelopment of a mass of dolomite by the zinc sulphide go to show that the chemical composition of the wall-rocks has been a factor in the deposition.

So also the lateral extension of the ores between the strata for a considerable distance, evidently following the planes of most rapid absorption, shows the flow of the solutions and their action upon the rocks. This lateral flow of zinc-depositing solutions is shown also upon a small scale by the angular skeleton forms exhibited by masses of " dry-bone " (zinc carbonate, smithsonite), where it has penetrated the thin cleavage planes of crystals of calcite, the mass of which has since disappeared leaving only rhombohedral cavities. Pseudo- morphs in smithsonite after calcite are also common in the deposits of "bone," and smithsonite often forms in crusts upon crystals of

g'°''-

628 LEAD- AND ZINC-DEPOSITS OF THE MIfeSISSIPPI VALLEY.

Ores in Solid Dolomite, — In driving through the hard bars of dolomitic limestone, supposed to be utterly barren, and in which there was no trace of decomposition or change by atmospheric agencies, I have found crystals of galenite imbedded in the massive unchanged dolomite, to all appearance a part of the original mass, or at least formed when it was formed, but accompanied by a small mass of calcite, both filling a former cavity. If, now, we suppose a rock like this to be shattered and traversed by crevices through which air and water could penetrate, we can easily imagine the gradual solution of such masses of lead sulphide and also of any disseminated masses of zinc sulphide or of iron sulphide and the gradual flow into the larger crevices of such solutions with the attendant re-formation of sul- phides, to be again oxidized inta sulphates and carbonates, giving us as a result an accumulation of ores in crevices and workable deposits. The evidence is strongly in favor of the view of the long-continued decomposition, downward flow and re-composition of not only the ores of zinc but of lead and of the pyrite from the upper formations to the lower, as the general water-level of the region subsided and as the upper formations by long-continued exposure through geologic ages were gradually decomposed in place. By such a process the present zinc deposits would seem to have accumulated and to represent the originally diffused ores in many formations, possibly as high in the geologic scale as those of Missouri,, or the Lower Carboniferous. This is, however, improbable owing to the dense and impervious na- ture of the intervenijig Hudson rivershales.

Prof. N. H. Winchell* has suggested that the sulphide ores have been derived from formations formerly overlying the region and since swept away. He thinks that the cretaceous formations once covered the lead-region, that the ores were primarily deposited from the ocean of that periorl, and that the sulphides gradually found their way downwards through the strata to their present horizon.

Locdlization of Deposits by Faulting, — With, regard to the apparent localization of the deposits by faulting, which Dr. Jenney finds it difficult to explain uj>on any other theory than that mineralizing solutions rose from below through the faulting-planes, I may be permitted a little tentative theorizing also.

The fracturing of the formations by upheaval or subsidence, with

the resulting faults, may have produced the needed crevices along

certain lines, and thus have determined the direction of the deposits,

but without any direct contribution of metalliferous solutions upon

Iron Ores of Minnesota, p. 153.

Ic

Lead- And Zinc-Deposits Op The Mississippi Valley. 629

the plane of the fault itself; or, possibly, the faulting-plane may have been the source of a flow of water by which the adjoining rocks have been leached, or by which marine organic life may have been destroyed in great quantities along the lines of fault, or in their immediate vicinity. This alone, if possible, would have been suffi- cient to cause a deposition of ores contiguous to the faults by the products of decomposition ; or we may suppose that noxious gases ascended through such breaks and, by the destruction of organic life along certain lines, or by direct precipitation of the metals, laid the foundation in the forming sediments of future ore-deposits.

The establishment of a line of faulting, or of break, through the beds around the borders of the Hurouian and Laurentian rocks early in the Silurian era, say at the close of the deposition of the Trenton limestone, would pre-determine other and later movements, and as we may suppose that the formations have not since been stationary, the same line of faulting may have been repeatedly opened and would extend upwards through all the later deposits, and what- ever influence such breaks or planes of movement exerted upon the localization of ore-deposits would have continued through all the ages wherein like favorable conditions for ore-deposits existed.

Influence of Organic Remains — Taking the phenomena as we find them and following the indications they offer, we shall conclude that the decomposition of organic remains, the death and decay of myriads of mollusks and other forms of life on the ocean floor de- termined the deposition of metallic sulphides from the sea-water by the influence of the escaping sulphuretted gases. This is substan- tially the theory propounded by Prof. Whitney in his Metallic Wealth of the United States, and elaborated somewhat in his later contributions to the geology of the lead-region.* We must also take into consideration the evident sudden and cataclysmal incur- sion of hydrocarbons with an attendant precipitation of shale, a fouling of the before-clear waters of the ocean with petroleum and mud, and the sudden destruction of the ocean-life of the period. If we can accept either of these views, a relationship between the faults and disturbances of the beds and the deposition of the ores is estab- lished and explained.

Petroleum Shales, — The Oil Rock. — It is not impossible, also, that the dislocations of the strata permitted the escape of sulphuretted and hydrocarbon gases from below, as already suggested, which act

GeoL Wisconsin, i., 1862, p. 404.

630 Lead- And Zinc-Dep08It8 Of The Mississippi Valley.

ing upon the sea-water caused the metals to be precipitated along or near to the outflove in the midst of the sediments, while also gases by destroying animal and vegetable life along the outflow added to the accumulation of decaying organic matter and promoted the precipitation of the sulphides of the metals.

Reference has already been made to the existence in the zinc-re- gion of a carbonaceous shale known as the " oil-rock." It is well named, for it is an oil-bearing rock, a true petroleum-shale from which oil may be distilled. It appears to have suddenly formed while the seas were teeming with life. We find it now as thin brown-colored partings, not thicker than paper or card-board, in the midst of the upper layers of the Trenton limestone, increasing in number and in thickness until it forms a shaly bed, some inches io thickness, a part of which is dark brown in color and is highly charged with petroleum.

The oil-rock, when dry, has a light snuflF- or chocolate-brown color and an earthy lustre, but when wet it is several shades darker and in the mine looks like old decayed wood. It has no ligneous structure, however, and no indication of organic origin and no strong odor, but, when dry, it is combustible. Fragments can be lighted in the flame of a candle and will burn with a luminous flame and give off considerable black smoke, which has a strong bituminous odor. Exposure to the weather for years does not destroy this property.

The transition from the compact, homogeneous, blue limestone without structure, with conchoidal fracture and few fossils, to the oil-rock, is sudden and is marked by a darker colored layer, lami- nated and irregular for a quarter to half an inch in thickness, fol- lowed by a layer of shells from one to three inches thick. These fossils are thickly crowded together so as to make up the bulk of the rock. They are best seen in cross- fractures of the rocks, which show innumerable edges of shells.

This record of life and of death is followed sharply by the oil rock, in which a few fragmentary fossils occur, but are rare. The shells, however, in the oil-rock are well-preserved and are embedded in the midst of the rock.

The presence of 8|>ecks and grains of galenite in the midst of the homogeneous limestone, but not far from the organic exuviie, is not to be overlooked ; in fact the presence of both blende and galenite in considerable quantities diffused in crystals in such shale beds, and in an accompanying blue clay, is extremely significant, and the deposit has the appearance of being indigenous.

Lead- And Zinc-Deposits Op The Mississippi Vaixey. 631

In all the principal mines about Shullsburg and westward along the Shullsburg Branchy and including the celebrated Raisbech mine, further north, the oil-rock is at the base of the ore-deposits. It ap|)ears to limit the deposition in depth ; accompanied by blue clay it becomes an impervious stratum, and is the bed-rock of the ores, below which it is considered useless to mine. This is the general rule; but exceptions are reported to exist.

This oil-rock also marks the change from the limestone of the Trenton to the Galena dolomite above.

The origin of this petroleum-shale, and its relation to the life of the period, are extremely interesting subjects.

In the Missouri region it is probable that the carbonaceous shales of the Coal-Measures have held a relation to the ore-depositions similar to that held by the oil-rock in Wisconsin.

It may be urged by some that the death of the organisms and their decomposition was the cause or the source of the hydrocarbons ; if so we have yet to explain the cause of the sudden and sweeping de- struction of marine life. Was it due to the sudden incursion from below, through long lines of fracture of the formations, of poisonous gases or solutions inimical to life, or could it possibly be referred to the sudden incursion of fresh water along such faulting- planes ; such a flow as might have arisen from the porous beds of the Pots- dam sandstones? Or, shall we turn to the phenomena of the great change in the chemical composition of the sea-water, the advent of magnesian solutions, or conditions which rendered the deposition of the Galena dolomites possible?

But whatever the causes may have been, we know the result. There was a sudden formation of a thin brown shale, highly charged with petroleum ; and we find that this petroleum-shale, this horizon of hydrocarbons, is to-day the chief lower horizon of deposition of the lead- and zinc-ores. Certainly, if this shale did not influence or determine the original primary accumulation of the ores, it appears to have exerted a very important influence upon the secondary or later deposition, from solutions percolating downwards.

We look in vain in the Shullsburg Branch region for evidence of the upward flow of mineral solutions. The drainage throughout the Galena dolomite is downwards to a general water-level, which in that region is at the top of the Trenton. The effects of atmospheric in- fluences are everywhere apparent in this Galena dolomite formation above the water-level. Yellow ochers abound. The rocks are decomposed and softened, shattered and broken, affording free

632 Lead- And Zinc-Deposits Of The Mississippi Valley.

ingress and circulation for surface waters through the dolomitic beds, and downward along clefts and crevices where the rocks are more indurated and impermeable. All the phenomena indicate a down- ward flow and percolation through crevices to the impervious layers of shale or clay at the top of the Trenton limestone. It is true that even if clefts of the beds exist below the oil-rock it would be difficult to see them in the bottom of the workings, but when we are able to lift out sheets of zinc-ore spread on an apparently imper- meable bed, it would be strange if at some point we did not find a prolongation of ore downwards into the lower rocks, if the ore had its source below.

Flint Beds and Deposits of Ore, — Having observed the prevalence of flints in the vicinity of lead-deposits, I am led to rard them, or the conditions of their origin, as having played an important part in the deposition of the ores, chiefly probably as determining the extraction or precipitation of the ores from sea-water, much in the same way that the decomposing organisms at a lower horizon seem to have done. In many of the flint beds I have found large quan- tities of limonite in octahedral crystals, evidently the result of the alteration by oxidation of pre-existent sulphide of iron — iron pyrites.

Brecciated Deposits. — With regard to brecciated deposits, reference has already been made to one in the synclinal of the Peaslee ridge along the Shullsburg Branch ; others are not infrequent. Notable examples are found in the Little Giant cut of the Wisconsin Lead and Zinc Company — the old Dry Bone diggings — and in the hill beyond at the McFeeley level, where I removed some thousands of tons of rock and ore last winter. In some parts of this mine the dolomitic* beds were much broken up, and the fragments were re- cemented by deposits of blende and pyrite, these minerals investing and covering the masses with a coating from half an inch to 2 or 3 inches in thickness. The appearances all point to an undermining of the beds by solution, followed by a falling-in of the unsupported blocks, many of which remain, but are slightly displaced and wedged in, like stones in an arch.

If this brecciation had been caused by movement along a faulting- plane, we should expect to see crushing eflects such as are familiar to us in the deep-seated veins of greatly disturbed regions, accom- panies] by a wall on one or both sides, while in this brecciated zone the transition is gradual from the broken-up parts to the well-formed beds in place. In some parts of the mine the beds pit<h sharply downward; and beyond they have their normal horizontal position.

Tip A Tv- And Zinc-Deposits Of The Mississippi Valley. 633

The general direction is northeast and southwest, and this I have found to be the general trend of the most extensive ore-deposits of the Shollsburg region, with "crossings'* of crevices, trending N. 20°-30' W. and S. 20°-30° E., also ore-bearing. I am, at present, unable to say positively that such brecciated deposits are not u))on faulting planes, but there is no direct evidence of any fissure or plane of movement passing through these brecciated zones, so far as yet observed.

Similar phenomena of inclined and broken-up beds of Gralena dolomite are visible in the open cut of the Cuba City Mine, where the beds pitch downward at an angle of 30° to 40°, but the lowest part of this inclined series has not been reached.

The Blue lAmeaUme or " Glass RockJ — In Dr. Jenney's first mention of the age of the principal ore-bearing formation of Wisconsin-Iowa as the Trenton limestone (p. 181), it would have been better to indicate specially the Ghilena dolomite which, though by the books a member of the Trenton group, is by nature so strongly marked and differentiated by composition, by appearance, by a change of fauna and by mineral contents, that in the lead- region we cease to regard it as a part of the Trenton. It is the old Cliff' Limestone of Owen and the home of the lead- and zinc ores which rarely penetrate into the compact underlying blue limestones, locally known as the "quarry-rock," the "glass-rock," and the " blue limestone." This last we in the lead region regard as the true Trenton. It is sharply separated from the Galena dolomite by a thin parting of petroleum-shale known to the miners as "oil- rock," and in. my experience this layer becomes the floor — the lower floor — of the zinc-deposits.

I am aware that Strong, in his chapter upon lead- and zinc-ores,* states that the workable deposits of lead-ore in the lead region occur in the Trenton division of the Lower Silurian formation ; but he pro- ceeds to explain that the Galena limestone is especially productive of lead-ore.

Whitney has shown, in an admirable way, by means of a diagram,t the vertical range or distribution of the ore-deposits in the Galena dolomite, from which it would appear that the depth to which the mineral impregnation extends in that formation depends largely upon the amount of surface denudation by which its thickness is reduced. While the vertical distance traversed by the solutions is approximately the same in the thicker parts of the Galena dolomite,

Qtol, WiMtmsin, i., 1883, 637, 638. f Wisconsin, i., 1862, pliriL

634 Lead- And Zinc-Deposits Op The Mississippi Valley.

they reach downward about half way to the Trenton h'mestone, but when the greater part of the Gralena dolomite has been denuded they extend lower down to, and even penetrate, the blue limestone.

'Jlie Lead Region not Glaciated, — It is a significant fact that the lead-producing region has not been glaciated, while the glaciers extended over the adjoining regions. This evidence is in favor of the view that the deposits have been formed by leaching downward, the unglaciated area having been long exposed to atmospheric agencies and decomposition without removal, while the rocks were decomposed in situ. Similar conditions prevailed in Miasouri, where, according to Dr. Jenney, the land has been above water from the beginning of the Carboniferous period to the present time.

Conclusion, — The subject of faults and dislocations of the strata in the lead-region appears to have been laid aside with the death of Percival. It received very little attention from Prof. Whitney, or from the Chamberlin Survey. It isof 8()ecial importance in view of the relations of the movements to the mineral deposits. The faults should be carefully studied and mapped and the position of the mineral deposits shown in connection with them. The work will be difficult owing to the heavy deposits of clay covering up and ob- literating the evidences of movement, but some of the more im- portant breaks may be located and followed along the banks of the streams, and some may be seen in the mines.

In connection with a survey to determine the position of such dislocations, it would be well to investigate the possible occurrence of lead- and zinc-ores in the Lower Magnesian limestones along the lines of disruption. This can be done by drilling at a com- paratively small cost, and, although it is not probable that paying deposits will be found, a practical demonstration would be not only very satisfactory to the mining men of the lead and zinc rion but it would be of great interest to science.

Arthur Winslow, Jefferson City, Mo. : All who have studied the lead- and zinc-deposits of the Mississippi Valley have probably felt that the explanations of their origin heretofore suggested, nota- bly by Whitney and Chamberlin, are not altogether satisfying ; that at least some additions may become necessary and perhaps some im- portant changes must be made. So radical a change, or reversion to the old ideas of Owen and Percival, as Dr. Jenney proposes, how- ever, few are probably prepared for.

Concerning crevices or fissures and their association with the ore- deposits, I am inclined to think that, as may be inferred from my re-

Lead- And Zinc-Deposits Op The Mississippi Valley. 635

marks on Mr. Emmons's paper, a little too much significance has been attached to them by Dr. Jenney. I have not been able to recognize the presence of any strong system of fissures or faults in the most productive portions of southwestern Missouri, and, where such exist in the southeast, in the larger ore-bodies, they can hardly be re- garded, in my opinion, as veins or channels of ore-supply from great depths. Indeed, one of the most pronounced faults in this region, which traverses the country about two miles north of Mine La Motte, with an apparent throw of 300 feet, has been found to be en- tirely unaccompanied by ore, though the adjacent ground has been prospected with the diamond-drill. Moreover, in those mines which work to the very contact with the underlying granite, I do not know of a single instance where ore-bearing fissures extend down into that rock. Apart from this, however, the association of ore and. crevices does not demand by any means a deep-seated source for the ores. Such crevices naturally act both as channels controlling distribution and as receptacles for accumulation, whatever the source of the ores; hence a disturbed and creviced region, which is in other respects adapted to the reception of ores, will be their most natural habitat. Therefore the explanation of the localization of the deposits based upon such conditions is equally applicable under any of the common theories of ore- derivation. The same may be said, it seems to me, concerning the observe<l paragenesis of the minerals and the growth of crystals. If we accept the broader idea of lateral secretion, which does not demand that a mineriN shall be derived from the very rock to which it is attached, but recognizes abundant flow along crevices and through porous strata and a consequent free transfer of solutions from place to place, I think all the phenomena find at least an equally ready explanation. That the metallic contents of the country-rock are insufficient to have supplied the ore-bodies remains yet to be proved, according to the best of my knowledge.

The attempt to find support for the deep-seated derivation of the ores through analogy in stratigraphy and geologic history with re- gions of the far West does not seem to me altogether successful. The last pronounced regional disturbance of both the Ouachita and Ozark uplifts was immediately after the Coal-Measure period. In Arkansas this was accompanied by great flexing of the strata. I find no evidence in the Ozark uplift of any intense disturbance of post-Cretaceous date, or of the presence, even at great depths, of flows of such igneous rocks as accompanied the uplift of the Rocky Mountains. I have already expressed myself to the effect that the

636 Lead- And Zinc-Deposits Op The Mississippi Valley.

Missouri ores cannot be properly considered to occur in the Coal- Measures of the State. Had such a profound fissuring taken place in post-Cretaceous times as Dr. Jenney requires, we should expect to find it extending into the body of the Coal-Measures, accompanied by the ores. At least faulting or other exhibition of disturbance would be found, which phenomena do not characterize these rocks.

Over and above these considerations affecting the quality of the support for the theory, there still remain the positive obstacles to be dis|K)sed of. The almost entire absence of the precious metals in the Missouri ores is a fact which further weakens the force of any anal- ogy which may exist between their conditions of deposition and those of the Rocky Mountain ores. How are the objections raised by Whitney and Chamberlin to be met — such as the facts that faults are practically absent from the region ; that there is little ore in the underlying Lower Magnesian beds and none in the Potsdam and St. Peter's sandstone ; that no deep and continuous crevices like true fissures are found ; that no hydrostatic cause is assigned for the ascension of solutions from great depths? How could the ores be carried across such thick, pervious and water-soaked strata as those of the St. Peter's and Potsdam sandstone without showing their presence in those rocks?

I recognize that it is easy to criticize, but that it is another matter to supply and support a satisfactory substitute. I confess that I am not prepared to offer such a substitute at present ; yet what I have said arises, by no means, from a mereHesire to criticize. I recognize the value of Dr. Jenney's contribution as advocating one hypothesis. The study that I have given these ores during the past few years leads me to feel that some statement of fact and some expression of opinion is called for from me on the present occasion, even though I am not prepared to offer matured conclusions. We are all in search of the true theory, and questions or criticisms which tend to make obscure points clear or to remove false ideas are in the interest of the common cause.

Frank L. Nason, New Brunswick, N. J. : The question whether Dr. Jenney's theory, or that of Whitney and Chamberlin, is true, is one of vital importance. According to Dr. Jenney's theory these metalliferous deposits should extend to practically limitless depths ; while, on the other hand, if Prof. Chamberlin's theory is true, lead- and zinc-mining in the Mississippi valley will be confined to comparatively shallow diggings.

Lead- And Zinc-Dep08It8 Of The Mississippi Valley. 637

During ray two years' work in the Ozark mountains as assistant geologist in charge of iron-ores, I visited nearly every county south of the Missouri river. In my studies of the iron-ores many facts were noted which bear directly on this question, and thus may be of interest here.

The specular ores of the St. Fran9ois mountains* it will be hardly necessary to mention, since they are confined to the Archaean area of southeast Missouri. Suffice it to say, in passing, that, with a few exceptions, the ore-deposits are of vein formation, and prolmbly are veins filled by lateral infiltration from the surrounding porphyry.

The specular ores of the sandstone region occupy the plateau of the Ozark uplift, and since they occur in the same formation as the zinc and lead of southeast Missouri, it is very likely that the origin of one is intimately connected with that of the other. Dr. Schmidt, in his report on the " Iron-Ores of Missouri," published in 1872-3, places these specular ores in the Roubidoux sandstone (formerly called second sandstone). The work of the Winslow survey has established the fact that while many of these deposits do occur in lens-like masses in the saudstone, the ores taking the place of lime- stone, which has been removed by solution, the larger deposits have been formed in caves under the sandtone and in the Gasconade limestone (formerly called second Magnesian limestone).

A brief description of one typical locality will suffice for all.

The Cherry Valley deposit, as at present known, is about 400 feet long, 300 feet wide and 100 feet deep. The ore has been com- pletely extracted laterally, exposing the side-walls. At the north end of the mine the terminal wall has not been found, but the ground is too lean to be worked. Southwardly, the breast of ore appears to be as good as ever, and all the present workings are in this direction. Below 100 feetf though no bottom rock has been reached, the ore is very sulphurous, is mixed with clay and chert, and is thus valueless.

Following southward along the strike of the ore-body, we come to a large ravine running nearly at right angles to the depression in which the ore-body occurs.

The summits of the ridges which reach above the surface of the mine are covered with heavy beds of horizontal sandstone. At the surface of the mine, 150 feet below these summits, sandstone again appears, but instead of being horizontal, it occurs in huge broken masses, dipping steeply towards the ore-body. On the sur-

This name has been given to the porphyry hills of S. £. Missouri by Arthur Winslow, State Geologist. ,

638 Lead- And Zinc-Deposits Op The Mississippi Valley.

rounding slopes of the enclosing ridges for a distance of 80 feet above the ore-body are outcropping beds of horizontally-bedded limestone. This limestone, also horizontal, appears in the ravine before-mentioned below the lowest working in the mine.

The ravine drains into Huzza creek, and this, later, into the Meramec river.

These facts point conclusively to the theory that the Cherry Val- ley deposit lies in a limestone ravine. The horizontal sandstone on the summits of the ridges, and the same sandstone dipping towards the ore-body 150 feet below, lead to the conclusion that the ravine in which the ore-body lies was etched from the limestone under the sandstone, which, being left unsupported, gradually fell in.

In the iron-ores Mr. J. D. Robertson, assistant on the Greological Survey, found remains of crinoid stems preserved in specular ore, thus indicating that the deposition of these ores was not completed until after the age of the Lower Carboniferous limestones.

Taking into consideration the numerous large springs, so common in the Ozark region, it is an easy step to conclude that the ravine in which the Cherry Valley bank is situated was once an under- ground feeder to a spring which has long since been converted, through the falling in of its roof, into an open stream.

Again, the springs of the Ozark region are not deep seated, as is proved by their roiling during heavy rains, as well as by the fact that in many cases, they are known to be but the reappearance of "sinking creeks." It is assumed that the Cherry Valley depression was caused by similar agencies, and the source of the ore is to be referred to the leaching of rocks lying above and adjacent to the deposit.

The Cherry Valley bank is typical (except, possibly, in shape) of all specular ore-deposits of this region. It is an irregular parallele- piped, while or-bodies such as that of Simmons Mountain are shaped like an irregular inverted cone. Whatever the shape of the ore-bodies they are, without exception, so far as noted, near large drainage ravines.

Limonite iron-ores, also, are found abundantly in the Ozarks; . but while the specular ore-bodies occupy the plateau part of the re- gion, the limonites are characteristically distributed farther down the slopes of the uplifl. As regards their origin, they may be di- vided into four classes: (1) those which are derived by hydration directly from the specular ore-bodies, and which, except that they are limonites instead of specular ores, have the same nature and

Lead- And Zinc-Deposits Op The Mississippi Valley. 639

source; (2) liraonites derived from leaching of the specular ore- deposits and re-deposited at a lower level ; (3) those derived from the leaching of the rocks; and (4) those derived from the original deposits of iron sulphides.

For our purpose only the fourth class need be considered. Through- out the country, in Howell, Texas, Carter, Shannon, Reynolds and Ripley counties and the counties along the Osage river, large masses of limonite are often found with cubic crystals, pointing to their derivation from pyrite by desulphurization and hydration. In the same region crevices in the limestones are formed from a few inches to three feet in width, whose walls are covered with limonite pseudo- morph after pyrite. Long stalactites of limonite are also found so abundantly that limonite is more frequently referred to as pipe- ore '' by the country people. These pipes " are all limonite pseu- domorph after pyrite.

At Cedar Bay, in Wayne county, one bank has been worked for limonite ore, and is to-day partially exposed. The ore-body lies in clay, which is stratified above ; and this clay is filled with broken chert. The ore is composed mainly of pipe-ore, and the top of the ore-masses, as proved by their position in the bank, is a chert breccia, cemented with iron. From this top the pipes of ore depend, like the limestone stalactites, in an ordinary cave. In many instances there are pipes or stalactites of ore running at all angles to those which are pendent from the roof of the cave, showing that before deposition had ceased the mass had become partially detached and stalactites had formed at a different angle. This deposit, like the deposits of specular ore, is situated on the side and near the head of a ravine.

If any doubt existed as to the stalactitic or cave origin of these ores it would be dissipated by a sti-iking deposit on the farm of Hon. J. B. Old near Thomasville, in Oregon county. This deposit is situated in a ravine which opens into the bottom-land of Eleven Point river. Here a comparatively undisturbed deposit, 100 feet wide, and visible toward the slope of the hill for the same distance, has stalactites of ore still standing in a vertical position. The pipes are cemented to- gether, but a blow from a hammer separates them, when the pe- culiar concentric structure of the stalactite is seen, as well as the nearly open tube so common in stalactites. This is an undoubted cave-deposit, where the less resisting limestone has been etched away, leaving the iron filling standing in its original position.

These typical deposits cannot possibly be explained by supposing

640 Lead- And Zinc-Deposits Of The Mississippi Valley.

the chalybeate waters to have come from below. Stalactites are irre- futable evidence of downward-raoviDg waters leaching through the more or less porous roofs of caves.

It may be said that these ores, though at present of the immediate origin ascribed to them, may have come originally from deep-seated springs. If such springs ever existed in the Ozarks, they have now disappeared, but there are innumerable springs such as are capable of producing the results described, and it is more rational to suppose that they have been active in the past rather than that deep-seated springs were, and are not. Moreover, the leached condition of the sandstone and the great denudation of the limestone by solution give strong corroborative evidence.

It may be observed in passing that the specular ore-deposits, at least, have present in them appreciable quantities of lead and zinc.

Lead and zinc are also found distributed throughout the Ozarks. They have been, and still are, dug to a considerable extent. There are, however, no such large deposits as those of the Bonne Terre and St. Joe and the southwest region.

Bearing in mind that the iron is found more extensively lower down the slopes of the uplift than it is on the plateau-like crest, it is hardly unreasonable to assume that the limonite deposits are being constantly increased at the expense of the ore-bodies on the crest or plateau.

If this explanation is admitted for the iron-ores, why not for the lead and zinc ?

During the writer's work on the iron-ores his attention was many times called to deposits of lead and zinc in the iron-ore region. These occurrences, as noted, seem to support the theory of the trans- portation of mineral by down ward -moving spring- waters.

In the Palmer lead-mine, in Washington county, the lead-ore occurs, in thickness from a half-inch upwards, in a horizontal bedding- plane between two beds of Magnesian limestone. The bedding-plane is horizontal. According to my informant, Dr. Metcalf, of Steele- ville, the ore-body widens out by filling the irregular lens-shaped cavities, such as would be formed between limestone beds by solvent water.

Near Fourche Courtois creek, in Washington county, consid- erable quantities of lead-ore are dug from the clay overlying the Magnesian limestone. In stripping the clay irregular seams in the limestone are often found, some with minute stringers of lead and occasional large bunches.

L£Ad- And Zinc-Deposits Of The Mississippi Valley. 641

Near the same place, on Huzza creek, is an almost perpendicular bluff of limestone, showing in its face a series of caves, from one foot in height upward, which have l)een etched out between two strata of limestone. They are exceedingly irregular in shape, opening and pinching into large and small cavities. Stalactites of lime were once numerous, but they have all been removed for a long distance in the hill. The series of caves has been more or less completely filled with crystallized calcite, with stalactites and stalagmites. Mingled with this are bunches of barite. During the War this place was occasionally worked for the galena, which was disseminated through the secondary filling of the caves in such a way as to prove conclusively that it had crystallized from a solution, and had not been etche<l from the limestone and fallen into the forming calcite.

In Taney county, Mo., a vein of calcite and barite has been traced for nearly seven miles. At one place a rude shaft, 60 feet deep, has been sunk on it. The debris shows crystalline grains of lead- and zinc-ore, en<!losed in the gangue. The vein itself is tight and dry, but numerous large springs are now flowing in the immediate vicinity.

In Camden county zinc is dug from Coal-Measure slates. The shaft is about 80 feet deep. The slates lie steeply dipping towards each other in a synclinal trough. The Magnesian rocks on either side of the trough are perfectly horizontal. The conclusion is inevi- table that the Coal-Measure slates have been undermined and have fallen into a narrow cave, as have the sandstones at Cherry Valley.

Many instances of the above nature could be cited, all tending to show that the movement of the water of the Ozark region, without a known exception, has been in downward springs from meteoric sources.

There is absolutely no known deep-seated thermal spring in the Ozarks of Missouri at present; nor is there, feo far as I have ob- served, any traces of the past existence of such springs. So far as solution is concerned, there seems to be evidence that when the ga- lena occurs, as it frequently does, in pockets of various sizes in the comparatively solid limestone, the bunches are often loose in the (avity and the galena is coated with oxide. This seems to indicate that through the limestone water is leaching which is capable of at- tacking these ores. If they go into solution they must, under fav- orable circumstances, be reprecipitated at a lower level.

The fact that the limonite iron-ores are so often found in this re- gion in the form of stalactites seems to prove conclusively that these

Vol. Xxii. — 41

r

642 Lead- And Zinc-Deposits Of The Mississippi Valley.

deposits at least were formed by waters leaching downwards through the rocks of the Ozarks. This conclusion is strengthened by similar forms in limestone caves, when, however, the stalactites and stalag- mites are carbonate of lime instead of oxide of iron. The leached appearance of the rocks, the abundant presence of springs, some- times of great size, the unmistakable evidence of these underground waters in great numbers in times past — all of these facts point con- clusively in one direction ; while the absence of deep-seated thermal springs at present, with no certain traces of their past existence, makes our conclusion all but certain.

There is another strong point bearing directly on our problem. In the Oeohgy of Wiscmsin, vol. iv., pp. 389-90, Prof. T. C. Chamberlin mentions the frequent occurrence of stalactites of iron pyrites, and very appropriately asks : " Could the forms above in- dicated, e.g. J the stalactites, have been formed by either gases, vapors or ascending hot waters ? "

Stalactites are by no means confined to the lead regions of Wis- consin. The writer has, in his own cabinet, specimens of stalactites of marcasite and galena from Galena, III. In the vicinity of that town and of Dubuque, la., there are caves wholly or partly filled with stalactites of carbonate of lime.

The crevices or veins in which are formed the lead- and zinc-ores in the Wisconsin lead-area are similar to those found in Missouri in the iron region of the Ozarks, as well as in the lead and zinc regions.

Dr. Jenney (communication to the Secretary): Faulting and disturbance of the strata are the controlling factors influencing the localization of the lead- and zinc-ores throughout the Mississippi valley. This is true as well for all metalliferous formations in all the mining regions of the globe, so that the law may be stated as universal that oil workable deposits of ore occur in direct association with faulting figures traversing the strata j and with zones or beds of amnheti mid brcceiated rock, produced by movements of disturbance. The undifthirhed rooks are everywhere barren of ore,

Although no ore-deposit has been found to occur except in direct

'xmnection with faulting and disturbance of the rocks, we should not

ook for all fissures, or all zones of brecciation, to be commercially

irt*-lnaringj any more than we expect all trees to bear fruit; some

h)in natural (pauses will inevitably be barren.

Prof Blake, from personal observations in the lead- and zinc- [tiittiiig regions of Wisconsin and Missouri, notes the intimate rela- lt .4£-£..(jfg fisaures and disturbance of the rocks with the occur-

LEAD- AND ZINC-DEPOSnS OF THE MISSISSIPPI VALLEY. 643

rence of the deposits of ore. This is further confirmed by S. F. Emmons in the paper presented at this meeting on the " Geological Distribution of the Useful Metals in the United States."

The real question under discussion seems to be, whether the lead and zinc in these deposits were derived by some process of lateral secretion from the overlying or contiguous rocks, or came from an unknown source, possibly deep seated in the earth's crust. In the one case, the agency aflFecting the local concentration of the ores in the deposits is thought to have been the lateral and downward circu- lation of atmospheric waters; in the other, the deposition is regarde<l as having taken place from solutions, probably not highly heated, though above normal temperature, forced upward through the fault- ing fissures traversing the strata.

If formed by the downward circulation of sub-aerial waters, the deposits of lead- and zinc-ore would be closely related in origin and manner of formation to the limonite iron-ores and the deposits of manganese; while, if the formation were due to the ascension of mineral solutions, the natural classification would be with the fissure- veins and the mines of gold- and silver-bearing lead- and zinc-ores. This question of origin has not only great scientific interest, but also an important bearing on the economic relations of these deposits. With a deep-seated origin, there is encouragement in certain districts to prospect in search for ore in the underlying sedimentary forma- tions, where the rocks are favorable in structure and composition to the occurrence of mineral, and under conditions where the expense of mining, due to water and increased depth, does not preclude profitable working. A derivation of the metals from the superficial sedimentary strata, on the other hand, holds out but little hope that workable deposits will anywhere be found in depth when the depo- sits near the surface are exhausted. It should not be understood, however, that the theory of origin in depth implies an indefinite ex- tension of the ore-bodies downward following the course of the fis- sures. From the very nature of the deposits, formed solely at the intersection by the fissures of the favorable horizons, the intervening strata being barren, workable deposits of ore are only to be sought in formations, where the rocks are from structure, lithologieal char- acter and chemical composition, and from the disturbance and brec- ciation they have undergone, rendered favorable to the pre(Mpita- tion and deposition of the minerals from the ore-forming solutions.

There are many considerations that cannot be here stated at length supporting the view of a derivation of the minerals in depths. One

644 Lead- And Zinc-Deposits Of The Mississippi Valley.

of these is the inadequa'y of the inclosing or even of the neighbor- ing rocks, as a possible source of supply, from the contained minute traces of the disseminated metals, to form even by a concentration extended radially for a distance of miles, the immense ore-bodies thickly clustered in the mining districts of the southwest. Moreover, it is diflScult.to explain by lateral secretion the formation of the vast masses of dolomite associated with the ore-<leposits, and resulting from the alteration of the Cherokee limestone — a limestone which contains but traces of magnesia, and could not have afforded the magnesia requisite to affect this dolomitization. Again, the minerals associated with the ores are not such as would be predicated from the chemical composition of the adjacent rocks.

In the upper Mississippi lead-region, late mining explorations are affording evidence of the occurrence of workable ore-deposits in the Lower Magnesian limestone, the ore-bearing character of which was denied by Whitney and Chamberlin. Near Lansing, Iowa, a lead- bearing fissure is mined in the Lower Magnesian limestone, and has produced up to this time 125 tons of clean galena. It is a strong crevice, 4 to 6 feet in width, with walls striated by the faulting movement, and is traced on its course for over 1200 feet. This oc- currence must be regarded as important evidence of the vertical range of the ore-deposits in the geological formations of the Wiscon- bin uplift.

As to the time at which the deposition of the ore occurred, careful examination has failed to show any evidenc of the formation of lead- and zinc-ores taking place in the mines at the present day, other than such deposition as is the result of oxidation of the pre- viously existing ores and their reformation and crystallization, due to the secondary action of oxidizing surface-waters, percolating downward through the upper portions of the ore-bodies. The con- clusion appears inevitable that the deposition of the ores has long since ceased, and that it must have occurred at a definite period, when the subterranean circulation of waters in the strata, and the condi- tions influencing and affecting the formation of the ore-bodies, were radically different from those now subsisting. This period can be generally fixed as later than the Coal-Measures and prior to the close of the Tertiary. The inference is that the deposition took place at the only period known to be characterizeil by igneous disturbances adequate to produce the fissuring and disturbance of the strata and the attendant formation of the ore-deposits, to wit, in post-Cretaceous and early Tertiary time.

LEAD- AND ZINC-DEPOblTS OF THE MISSISSIPPI VALLEY. 645

The widespread igneous disturbances of the Ouachita uplift, ac- companied by extensive extravasations of the eruptive rocks, have been determined by Branner to be post-Cretaceous in age. It is impossible to conceive of such intense igneous action taking place in the Ouachita uplift and not profoundly affecting the closely con- tiguous Ozark uplift — the narrow strip of Coal-Measures filling the valley of the Arkansas river forming the interval of separation between these two great upheavals.

It is not probable that the Coal-Measures of northern Missouri and eastern Kansas are disturbed or faulted to any greater extent than are the other Palaeozoic formations surrounding the Ozark up- lift. The more pronounced faulting and brecciation of the sedi- mentary strata, which alone are accorapanieil by the occurrence of lead- and zinc-ores of economic value, are confined to that upheaval. Within the limits of tiie uplift, rocks of Coal-Measure age occur locally, filling basins and depressions eroded in the older Palaeozoic beds, and are frequently observed to lie faulted and disturbed in bedding. In a number of mining districts in the southwest, these superficial formations carry workable dpcits of lead and zinc. In this connection may be citfed theleposits of galena filling fissures and brecciated zones in Coa'PMeasure shales at Webb City and Oronogo, Jasper county, Mo.— ore-formations which are connected by the fissures traversing the beds with large bcxlies of blende in the underlying Subcarboniferous chert and limestone. In the vicinity of Versailles, Morgan county, Mo., galena and blende fill seams in the coal of local outlying basins, and similar deposits of blende in coal occur near Sedalia, Mo.

Not far from the mining camp of Belleville, Jasper county, Mo., the shales of a small outlying basin are crowded with beautifully preserved impressions of ferns and coal-plants. The bedding of the shales is disturbed and faulted, apparently by the same system or belt of fissures which carries large bodies of blende in the Subcar- boniferous rocks near by. Galena, and occasionally blende, occur in crystals sparingly distributed through these plant-bearing shales ; and the growth of the crystals of galena has distorted and crowded to one side otherwise perfect impressions of ferns ; an evidence of the subsequent deposition of the minerals. From a collection of plants made at this locality, the age was determined by David White of the National Museum, Washington, D. C, to be probably that of the later Coal -Measures.

The question, whether the iron-ores are of similar formation to the lead- and zinc-deposits, is far too extensive to admit ofxiore

txiore T /CjQOgk

646 Lead- And Zinc-Dep08It3 Op The Mississippi Vallev.

than limited discussion at this time. It is worthy of note that the iron-ores occur in basip-like deposits reserablinor the outlying coal- pockets of the region and in connection with disturbances of the strata. The occurrence of small amounts of lead and zinc in these deposits of iron-ores need not be any more an indication of the origin of the iron, than the similar occurrence of blende and galena in the coal of the outlying basins, is an evidence of the origin of the coal.

Respecting the formation of iron-ores, certain ores originally de- posited as limonite are undoubtedly formed by the circulation of surface-waters ; an action that may be seen taking place in swamps and peat- bogs at the present day ; but it by no means follows, that the iron-ores resulting from the alteration and oxidation of pyrite have all lx?en de|)osite<l in this way. The great deposits of pyrite of the world are formed by ascension.

Stalactites cannot, be accepte<l as conclusive proof of a primary deposition of the ores by deseension. All evidences tend to prove that the lead and zinc were originally deposited in the form of sul- phides, as galena and blende ; and that, subsequent to the deposi- tion, the primary ores were subjected, to a greater or less extent, to oxidation and alteration by the action of atmospheric waters. This change can be seen in active operation in the mines at the present day.

Stalactites of blende, galena, marcasite and not infrequently of calamine, smithsonite and cerussite, occur in the vugs, cavities and open channels in the ore-bodies; but they appear, in all observed instances, to have been formed by this secondary action of surface- waters on the primary ores, and therefore, they may be of compara- tively recent origin. As the last footsteps passing over a road are the most prominent, and obscure all that have gone before, so, in like manner, within the ore-bodies, the most strongly marked evi- dences of water-circulation are those resulting from the recent down- ward circulation of surface-waters. Yet indications of the direction of flow of the earlier solutions depositing the ore can still be de- tected.

From the very nature of their mode of formation, stalactites can but be regarded as affording evidence of uncertain value respect- ing the manner in which the deposition of the ores primarily oc- curred, or the direction of flow of the solutions producing it.

The comparison and correlation of the deposit of argentiferous lead and zinc of the Rocky mountains with these formations of the Mississippi valley must be led for consideration in auotlier paper.

Summary Op American Improvements And Inventions, 647

SUMMARY OF AMERICAN IMPROVEMENTS AND INVEN- TIONS IN 0RE-0RU8HING AND CONCENTRATION AND IN THE METALLURGY OF COPPER, LEAD, GOLD, SILVER, NICKEL, ALUMINUM, ZINC, MERCURY, ANTIMONY, AND TIN

Discussion of the paper of Mr. Douglas.* (See p. 321.)

Prof. H. S. Munroe, New York City : In Ins reference to cop- per-dressing at Lake Superior, p. 325, Mr. Douglas says that " the . . concentration . . . has been carried out with greatest financial economy, . . I do not say with the greatest economy of min- eral." And again, on p. 344, referring to the same, " it would not pay to reduce the quantity treated, and increase hand-labor, in order to save a few hundredths of one per cent, of copper."

Mr. Douglas probably has in mind the fact that the copper lost is all included copper, locked up in grains of sand. To save this, the rock must be crushed finer. If the mesh of the stamp-screens be reduced, the stamps will crush less; and thus we shall have to reduce the quantity treated in the mill per day, while the labor re- mains the same per day, or is increased per ton, as he says.

Submitting to the loss of copper, however, is not the only solution of the difficulty. A large quantity of rich sand is now collected by the jigs, and it would be easy to save so much of this material that the esca])ing sands would be quite poor.

What is wanted at Lake Superior, and at many dressing- works in this country, is an efficient crushing-machine to treat the sands thus collected. Abroad, we have the Heberle machine and the Schranz mill for this purpose, the latter being the more successful. The Heberle grinder was tried at the Calumet and Hecla mill for crush- ing these rich sands, with what success I do not know.

I should like to ask Mr. Douglas if we have no American crush- ing-machine that will reduce jig-sands to a fine state of division

Note by the Secretary.— In Mr. Douglas's paper, on p. 321, line 15 from the bottom, betimes" should be '*at tiroes ;'' on p. 824, line 6, the first word, height," should be " weight" ; on p. 328, footnote, R. N. Terhune should be R. H. Terhune; on p. 329, line 6 from the bplloin, Hoffman should be Hofmann ; and on p. 337 the name of Mr. E. P. Mathewson, of Pueblo, Glo., is incorrectly given as Matheson.

648 Summaky Of American Improvements And Inventions.

economically, and without the production of an undue amount of slime. There are a number of machines on exhibition in the Mining building, some of which are reported to be successful. The omission of all reference to them by Mr. Douglas is significant.

I should also like to ask whether the high-speed rolls mentioned by Mr. Douglas do or do not produce a larger proportion of slime than the same rolls running at a lower speed. In the crushing of coal with toothed-rolls it has been found that there is a 8i)eed which gives the best results with a given quality of coal, while higher as well as lower speeds increase the amount of dust. Have any ex- periments, as to this point, been made with these high-speed rolls?

I must call attention to a curious slip of the pen in Mr. Stanton's description of the Atlantic mill. He speaks of the V-separator. In fact the V-separator, or double-V-trough (a modified Spiizgerinne), which was in general use 15 or 16 years ago in the Lake Superior

The Evans Classifier (not drawn to scale).

mills, is but rarely met with to-day. The Richards-Coggin separ- ator is used in its place in many mills. At the Atlantic mill and some others, the Evans separator is employed. This latter is as great an improvement over the Richards-Coggin separator as that was over the V-separator formerly used ; and both have a very great ad- vantage over the Bilharz-Rittinger SpUzluUe and other forms of sand- classifier used abroad, in avoiding excessive head of water at the discharge-orifice, without the necessity of using rising discharge- pipes or small discharge-spigots. An excessive head brings too much water on the jigs with the sands for good work, and thus increases the losses ; while the rising-tube is apt to clog, and makes it necessary to increase the force of men attending the jigs. This feature is especially important when steam-stamps are used, as broken screen- plates are common, and an accident of this sort will quickly clog a large number of spigots. By pulling out plug and spigot together, the coarse fragments are discharged from the separ-

Summary Of American Improvements And Inventions. 649

ator, and at the same time the spigot may be quickly cleaned, and then the plug put back — the whole operation consuming a frac- tion of a minute only.

The Evans classifier, as used in the Atlantic mill for a number of years, is shown in the accompanying sketch (not drawn to scale). It consists of a rectangular trough, with four circular openings in the bottom, under which are four boxes which collect the different grades of sand. Four pipes supply the necessary wash-water to these boxes. The openings in the bottom of the trough are covered by flat disks, fastened to sleeves which slide on the water-pipes, and to which they may be clamped at any desired height by means of set-screws, as shown. The trough is inclined somewhat more than is indicated in the sketch. The inclination can be further modified at any desired point by altering the height of the wooden partitions which subdivide the trough. The trough fills with sand to the level of these partitions. This sand, by the jar of the stamps, very soon becomes solid and compact, except where it is kept stirred by the ascending wash-water; at these points funnel-shaped cavities are formed, which act as SpUzkasten. The separation takes place in the thin stream of water that flows over this hard surface of sand ; the separation is therefore analogous to that which takes place on tables, and combines sizing with sorting to a greater degree than is the case with any other form of classifier. This separation of the jig-sands on an inclined surface, and in a thin stream of water, is a distinct and important advance in the art of dressing. The Evans classifier, in its practical operation, leaves little to be desired. It permits of most varied adjustment, and will produce almost any desired kind of separation. Any inclination can be given to the trough, and this inclination can be modified in any desired part of the trough by al- tering the height of the partitions. By regulating the inclination and the amount of feed-water and wash-water, we can control abso- lutely the separation which takes place on the inclined surface in each and every section. By increasing the length of the trough, the character of the separation can be still further modified.

In the four spitzkasten formed in the sand by the rising current, a sorting-action, according to equal-falling ratio, undoubtedly takes place. This sorting-action is also under perfect control by raising or lowering the disks, and by admitting more or less wash-water to the boxes below. The sorting-action can be confined to a separa- tion of slime from the sands ; or it may be pushed so far as to effect a classification of the sands themselves. It is also possible to regu-

650 Summary Op American Improvements And Inventions.

late the amount of sands coming from each box, and distribute them in any desired proiwrtion among the several jigs — feature of very great importance.

The discharge-spigots are shown, for convenience in the sketch, at the side of the small boxes. They are really in front. These spigots are of the usual form employed in the Lake Superior mills, and con- sist of wooden plugs with a short bit of gas-pipe inserted, J-inch to |-inch in diameter.

Prof. R. H. Richards, Boston, Mass. : I have little to say in the way of criticism of this very comprehensive paper. Mr. Douglas has well weighed the relative importance of inventions and improve- ments that have been made in this country. I think, however, of certain additions that should be made to it ; and these I will now proceed to indicate.

Mr. Douglas says on page 324, American metallurgists have devised fewimportant improvements in machinery for the mechanical concentration of ore, if we except hydraulicking.'*

I think if we consider the hydraulic separator, the Collom jig, the Frue vanner and their effects upon concentration in this country, as well as the spiked and fluted rolls for anthracite, the gyrating screens and the automatic slate- pickers, we shall agree that the above statement hardly does us justice.

The hydraulic separator was introduced at some time previous to 1866 at Lake Superior; the exact date I have not in my possession. I am not aware of its having been used in England before it was used here. If I am not mistaken, up to that time the so-called English system of jigging consisted in feeding an unclassified product on an intermittent or plunger-jig and obtaining therefrom a skimming or ragging from the sieve and a hutchwork from beneath the sieve; practically the whole of the fine slimes going into the tails. This method I have seen in operation in Scotland as late as the year 1876.

Now the scheme of feeding a series of plunger-jigs with a series of classified products, each one suited to its adjustments of mesh, sieve-plunger movement, depth of sieve, thickness of bed and quan- tity of water, and at the same time eliminating the fine slimes more or less perfectly for treatment upon a slime-table, all this by means of a hydraulic separator, is, I think, purely American.

Up to 1881 the hydraulic separator of Lake Superior was very imperfect ; all the mills using the same pattern. It did not give a positive action. If the hydraulic water was slightly increased it did not react at once ; and when the reaction came, it was likely to be

Summary Of American Improvements And Inventions. 651

too Violent. Consequently this separator was difficult to adjust and it always allowed some slirae on the coarser jigs.

In 1881 a new positive separator was invented. The effect of adjustment on this was instantly felt at the spigot, an<l remained constant. This was introduced in several mills ; and other mills immediately adopted the idea, though using forms of their own. The Butte mills are all furnished with separators of the positive class. The introduction of the hydraulic separator has brought about the abolition of trommels for sizes below -inch mesh in all mills, and below -inch diameter round holes for all the native-copper mills of Lake Superior, as well as two large Montana mills.

So far for the invention : now for investigation. The discussion of the jigging of water-sorted products for a two-mineral separation has been well done by Prof Munroe, of New York, showing that jigging can really separate particles of greater disparity in size than Rittinger's formula allows. This discussion has been taken up by Hoppe and others abroad, and the ideas are permeating the advanced European mills. I expect shortly to lay before the Institute an in- strument by which this law of jigjjing mixed sizes can be studied by each mill-man experimentally on his own ore, with all its crooks and turns, its shots, pins, scales, flakes, leaves, bui*8, branches, and included grains, and then he will no longer have to de|>end upon pure mathe- matics to tell him what ought to happen but does not. He can determine instead what does happen.

The Collom jig, where there is plenty of water and only a two- mineral separation, is finding increased favor. The various belt- vanners are now standard nearly everywhere in this country, and are being introduced in foreign countries. The spiked roll for breaking anthracite has been standard for many years ; the new fluted roll is gaining ground through its cheapness and efficiency.

The gyrating screen will be haile<l with thanks as soon as it is appreciated how smoothly it runs, how free from blinding, and how compact in space. The anthracite automatic slate-pickers, which separate the flat pieces from the round, simplifying the hand-pick- ing which follows, are deserving of mention.

Mr. Douglas says the Ball stamp has replaced rolls in the large Montana concentrating-works. It is true that two large mills have put in the steam-stamp, and the third has built a double mill, half with rolls half with steam-stamps, in order to settle the question which system is preferable. But the mill-men using rolls and Hartz jigs and vaimers are not yet convinced that their mills are not equal

652 Summary Of American Improvements And Inventions.

or superior to the steam-stamp separator, Collom jig, and slime-table mills. They are backing up their opinions with careful assays of carefully taken tailing-samples. The great difficulty with thesteam- stamp is its slime. This stamp ignores entirely the crushing and vanning by stages.

In comparing the Collom and Hartz jigs, I think it should be said that the Collom jig works best with much water, with a hydraulic separator, and upon a two-mineral separation, while the Hartz jig works with less water, with two-, three-, or four-mineral serration, with close sizing or with hydraulic separation, and with a deep sieve for very coarse sizes. But, on the other hand, with all these ad- vantages, its adjustments are not nearly so handily made. The Hartz is therefore a much more universal jig than the Collom, while the Collom is much handier than the Hartz.

Stephen R. Krom, New York City : Mr. Douglas says: Though no cardinal change has been made in the construction of rolls, those made by Gates, of Chicago, which combine, under one housing, rolls, screen and elevator, and the splendidly-built steel rolls of the Krom pattern, are good examples for American design and work- manship.'

If we compare the rolls of fifteen years ago with the most perfect in use today, I think it will be seen that the improvements may represent cardinal changes.

We may take, for example, the steel tires and the method of hold- ing them on the shafts, uniting in one casting the bed, statiooH pillow-blocks, and a portion of the housing ; the introduction of swinging pillow-blocks in lieu of the troublesome sliding pillt>" blocks ; the housing to enclose the rolls and prevent the escape ot dust; the substitution of the band fly-wheel pulleys, for gearioft and the self-contained battery of springs of great power.

The old style of rolls contains none of the above features. Tb cardinal changes, taken in connection with the proper proportion of the parts and good workmanship, have made rolls fiuocess rivals of the stamp-battery. I am unable to understand from wh*'' mechanical standpoint Mr. Douglas endorses the construction \irbic" combines, under one housing, rolls, screens, and elevator. Certain*? it 'cannot be justified on the ground of either efficiency, eoonofly' or simplicity. The first principle in the construction of such chines is simplicity. All combinations which will increase corup'' cation should be avoided. The housing which encases the rollsshou'" be so constructed for removal that the rolls can be easily uncovered,

Summary Of American Improvements And Inventions. 653

enabling the operator to observe at will the condition of the rolls, and to remove any pieces of steel which may have become imbedded in the faces.

Rolls require, also, an even feed across their faces ; and such a feed cannot be had if the rolls are fed by the side-stream, as in the GUtes machine.

Moreover, the screening-capacity of such a machine (as is also the case with the stamp-battery) is not only too small but is inefficient, and the screen is liable to rapid destruction or damage. A proper arrange- ment for screening should provide a coarse screen, to keep all coarse particles of ore from the fine screen. This is not practicable in the Grates machine.

Besides all the objections to such a combination of rolls, screen, and elevator combined in one machine, the cost is considerably more than that of a pair of the best rolls of the same diameter, an eleva- tor and sufficient screening-capacity, arranged to do the work effici- ently, and independent of each other.

Mr. Douglas says the tendency in this country, now that open- hearth steel is used for tires, is toward high speed, running up to 170 revolutions. I began using open-hearth steel tires as early as 1872, and have never used anything else. At first I made my rolls with gearing, and ran them at the speed of 60 revolutions, considered at that time to be "ridiculously high" (25 revolutions being then about the average). In 1882, to adapt rolls to do the work of the stamp- battery, I substituted band-wheels for gearing and raised the speed to 100 revolutions. It is the use, therefore, of fly-wheel pul- leys for driving rolls, which allows us to run the rolls at a higher speed.

Mr. Douglas alludes to the fact that the Ball steam-stamp has replaced rolls in the large concentrating-works in Montana. This result was, in my opinion, entirely due to the fact that rolls of proper construction were not used in those works.

In referring to the Krom and Paddock pneumatic jigs, Mr. Douglas says : " The result has generally proved so much less perfect than that attained by wet concentration, and the maintenance of the machine in repair so much more costly, that the system has not made headway." So far as these remarks refer to the Krom jig, they are not correct. The work done by the Krom jig is more per- fect, within the limits of the capacity of the machine (that is to say, from an 8-mesh screen to 120 meshes per linear inch), and the cost for repairs is trifling.

664 Summary Op American Improvements And Inventions.

T. A. RiCKARD, Denver, Colo. : Mr. DouglJis's paper will be of great value for future reference ; but the ground which it covers is so extensive that it was impossible not to make some omissions and some small errors.

The description of the two types of stamp mills (those of Colorado and those of California), is open to criticism. The California "stamp- head and stem" are said to weigh from 700 to 800 pounds. This is not accurate; the total weight varies from 750 to 1000 pounds. The most usual weight at the present time is 850 pounds, and it is dis- tributed as follows: stem, 335 pounds; stamp- head, 226 pounds; tappet, 130 pounds; shoe, IGO pounds.

The pulp is said to be discharged from " both sides." Does the author mean the back and the front of the mortar ? Double-dis- charge mortars have long been abandoned in the gold-mining re- gions of California and Colorado, though they are still retained in silver-mills.

The Colorado stamp weighs 600 to 600 pounds. Stamps weigh- ing 80 little as 400 pounds are not now used. A 550-pound stamp would have its weight distributed as follows: stem, 220 pounds; head, 185 pounds; tappet, 60 pounds; and shoe, 85 pounds.

The typical Colorado (Gilpin county) mill is not characterized "T so rapid a drop as 40 per minute. The minimum is 26 to 28, maximum, 30 to 32.

In the matter of concentrating-machines of the vanner class, Douglas mentions those most commonly in use. He, very prop- erly, does not express any preference for any one type ; yet I that there is one fact to which attention may be called. The rience of the mill-men of thechief gold-milling centers of Cali"*' uia (Amador and Nevada counties), gives unanimous testimony the effect that the Frue vanner requires less attention, and giv trouble, than any of the other machines which have been construct upon the principle of the endless rubber belt.

In the matter of concentration after amalgamation in an ordin' gold stamp-mill, the American metallurgist has no great reason feel proud. There has been a marked tendency to adopt a my-leader policy. That two Frue vanner.** should treat the discharged by 6 stamps, is a rule slavishly followed, whether stamps crush 8 or 15 tons per day, and whether the ore crusb contains 0.5 or 6 per cent, of the material to be concentrated.

I would also draw attention to a mistake which others havedou*' less remarked long ago. The passage of all the pulp direct to to

Summary Of American Improvements And Inventions. 855

vanner is surely wrong in principle. Some effort at sizing should precede concentration. Such sizing does not seem a costly or difficult operation. As it is now, the vanner is called upon to concentrate particles of pulp ranging from 30-mesh to impalpable slime.

The designing and building of a mill is left too much in the hands of the foundry, and too little to the man who is to we it. Again, the manager usually gives more attention to the winning of the ore than to the subsequent extraction of the values in it. There is room for much improvement in these directions.

In referring to the system of hydraulicking used in the Califor- nia gravel-mines, Mr. Douglas has omitted any reference to the " elevator." This machine is a truly American invention, and has added largely to the areas available for hydraulic mining.

Richard Pearce, Argo, Colo, (communication to the Secre- tary) : Mr. Douglas refers to the increased sizes of the hearths of the furnaces built at Argo, and T give herewith some results of figures I have recently obtained on the comparative area of hearths of furnaces which have been in operation at Argo, covering a period of fifteen years, together with the relation of the area of the hearth to that of the fire-box :

Year.

Area of Hearth.

Area of Fire-box.

Proportion of Hearth to Fire-box.

108 square feet.

4.800 : 1

143 "

6.351 : 1

202 "

8.161 : 1

: 1891

9 192 : 1

, 1893

12.000 : 1

It will i>e seen that these figures show an enormous increase in capacity. The latest furnace, which has been in operation some few weeks, and which has been christened "Columbia,'* probably exceeds in size anything before attempted in this direction. Its average capacity for fifteen days on ores of rather a refractory char- acter, has been 37.01 tons per twenty-four hours. If conditions at Argo would warrant the feeding of red-hot ore instead of the cold mixture we have at present, I am afraid to say to what extent this capacity might be further increased.

From experience obtained since the new furnace has been in opera-

656 Summary Of American Improvements And Inventions.

tion, I ana pretty well convinced that we have not yet reached the limit.

Prop. H. O. Hofman, Boston, Mass. (communication to the Secretary) : In his review Mr. Douglas has necessarily compressed a very large subject into very small compass. I should like to add a few supplementary remarks on the treatment of argentiferous lead-ores.

First as to sampling. This has been brought to great perfection within the past seven or eight years, by reason of the fact that a few largesmelters, centrally located for ores, fuels, and fluxes, have replaced many of the smaller ones which formerly existed. In the latter, hand- sampling with the ordinary implements or with improved ones, like the Brunton sampling-shovel, was sufficient ; but in the large works a more rapid and less expensive method became necessary, and machine-sampling was introduced. This consisted at first in separat- ing continuously a small part of a running stream of ore; afterwards it was found better to take the whole width of the stream at short intervals of time. Of the machines for doing this, the three repre- sentative ones are those of Brunton, Bridgman, and Constant, all excellent. To prepare the sample for the laboratory, we have the sample-grinders of the coffee-mill type, as made by our leading firms, and the quartering-apparatus (laboratory- samplers) of Bridg- man and others.

Coming to the calcination of ores, special attention must be called to the improved construction of the furnaces and to the making-up of the charges. In the stationary hearth-furnace, besides increasing its inside width to 16 feet, the roasting-hearth has been separated from the fusing-hearth by a vertical flue, and the area of the vertical section of the latter has been made smaller. The products of com- bustion passing from the fire-place over the small fusing-hearth rise in the flue and then suddenly expand, filling the entire width of the roasting-hearth. Thus it has been made possible to keep a high heat in the fusing-hearth and a low heat in the roasting-hearth, and to let the latter discharge into the former. With mechanical rabbles the roasted ore is collected in a hopper, which is emptied at intervals into the slagging-hearth. In the making-up of roasting-charges new problems had to be solved. In foreign works the ores to be slag-roasted are galena concentrates rich in lead, while the sulphide ores which our Western smelters have to treat are usually mixtures, low in lead, of pyrite, blende, galena, etc., and gangue. To slag these requires a high temperature, which is liable to cause loss of

Summary Op American Improvement'S And Inventions. 657

lead and silver by volatilization ; and, as an entirely satisfactory method of condensing the fumes from a slagging- furnace has not yet been devised, the only way of reducing the loss seems to be not to slag charges containing over 10 per cent, of lead and 100 ounces of silver to the ton — some metallurgists say even less.

In the shaft-furnace, the form and the material used are much alike at all works. While 42 by 120 inches is now a common size, some smelters think 33 by 100 inches the most favorable for doing clean and cheap work, Formerly the standard distance between the tuyeres was 36 inches. The natural tendency to increase the capacity of the furnace prompted its extension to as much as 60 inches, water-cooled tuyere-nozzles being allowed to protrude through the water-jackets, so as to make the actual distance 48 inches. The pressure from the strong blast necessary to penetrate a charge of such thickne made the heat in the furnace creep up, with the result of excessive loss in metal. This caused a reaction ; and the distance was reduced in some instances to 30 inches : at present it varies from 33 to 42 inches. Some furnace-men think 120 inches too long for a furnace, an<l do not like to exceed 100 inches.

In the management of a furnace, beside the making of a correct slag, its separation from the matte, especially with zinky ores, and the collection of flue-dust, have given much trouble, and are still doing so to some extent. Mr. Douglas praises the Mathewson tap- ping-jacket as promoting a good separation, but omits to mention that of Mr. E. F. Enrich.* which resembles it very much. Dr. Ilesf accomplishes the work in an entirely different way. He col- lects the liquid matte and slag from six blast-furnaces (42 by 120 inches) in a reverberatory furnace heated by a separate fire-place. Matte and slag are not only well separated, but their further hand- ling is greatly cheapened. It would seem as if, with a smaller num- ber of furnaces and a continuous flow, a HerreshofFor Orford well, or, if necessary, a combination of the two, would give a satisfactory separation. For the collecting of fumes, the simple and effective manner of cooling the walls of dust-flues and dust-chambers in use at the Grant and Omaha works, Denver, Colorado, should not be overlooked. I refer to the use of hollow bricks, through which air circulates. It is a simple and excellent way of surmounting one of the principal difficulties in condensing flue-dust, namely, the cooling

Berg-und Huitenmdnniache Zeilung 1892, page 187 ; United States Patent, No. 424104, March 25, 1890.

t United States Patents, Noe. 494,570 and 494,571, April 4, 1893.

Vol. Xxii.— 42 „ . , ,

658 Summary Of American Improvements And Inventions.

of the gases ; the other, the retarding of the current, has been hap- pily dealt with by Freudenberg. In the management of the fur- nace special stress 'must be laid on the fact that the making-up of the charge is now almost wholly governed by chemical principles, instead of mere experiment as in former years. For this introduc- tion of a scientific instead of an empirical method, the American lead-smelter can take much credit to himself.

The improvements made in the desilverization by the refiner are in no degree inferior to those made by the smelter, and have been along the same line. The distribution of precious metal in the base bullion has been carefully studied, and methods of sampling have been devised which are accurate and quick. Here, also, in some instances, machine-work has replaced hand-work. In the general arrangement of plant, the apparatus and its management are so planned that the base bullion, when charged into the softening-fur- nace at the top of the works, is not handled again until it is ready to be loaded at the lower end, as refined lead, into the cars. The bars are moulded, not by ladling, but by some one of the several sim- ple mechanical devices for that purpose. The capacity of thedesil- verizing-kettle, by which that of the remaining apparatus is regulated, has been increased from the original I2J tons to 30, and in a few instances to 46 and 50 tons. The kettle is discharged by that simple and beautiful invention, the Steitz syphon. The manner of working is being continually simplified. To-day, base bullion, running 300 ounces of silver and gold to the ton, is desilverized by two zinkings, if no separate gold-crust is to be made — otherwise by three. The liquation of the crusts has also been greatly improved ; and retort- ing, although invented by Parkes many years ago, only became the established method of work after it had been perfected here by Bal- bach. In cupelling, the principle of the English furnace has been adhered to; but the form, size, filling, and manner of support of the test, the apparatus for blowing, the mode of working, etc., all have been so changed and improved as to make an entirely new furnace, suited to the daily increasing demands made upon the refiner. In the record of improvements the working up of by-products must not be forgotten. This is now done as fast as they are made, thus leaving only a comparatively small amount of metal circulating in the works.

All the splendid achievements of American lead-smelting and re- fining have been made within the past 30 years and without dispar-

KothwelFs Mineral Industry, vol. i., p. 321.

Summary Op American Improvements And Inventions. 659

agement of the labors of any others who have contributed to this re- sult, I think we may well recall the names of O. H. Hahu and A. Ell- ers, among smelters, and E. F. Eurich and the late A. Steitz, among refiners, as pioneers of improvement — all of them members of the Institute.

C. A. Stetepeldt, Oakland, Cal. (communication to the Secre- tary) : Mr. Douglas's statements concerning the Stetefeldt furnace need some correction. A normal modern furnace is 48 feet high, and the effective drop of the ore is 37 feet. The auxiliary fire-place for roasting the dust is not at the entrance to the dust-chambers, but on return-flues — the dust dropping a distance of 9 feet through the flame. Between the vertical return-flue and the dust-chamber 18 a horizontal flue, with discharge-hop|)ers, at least 16 feet long. The limit of capacity for a Stetefeldt furnace has not yet been de- termined. It depends principally on the percentage of sulphur in the ore. The Marsac furnace, Utah, easily handles 70 tons of Daly ore a day, and at times receives a much larger amount, namely, when the dust collectinl in the battery dust-chambers is rapidly fed into the conveyors. At the Holden mill. Aspen, Colorado, as much as 90 tons a day are roasted, the ore containing 8 percent, of sulphur. At the Ontario mill, Mr. Russell has successfully roasted at the rate of over 100 tons of ore a day.

With base ores, especially those containing a large percentage of zinc-blende, the chlorination of the silver is finished on the cooling- floor, by leaving the hot ore banked up for 24 hours. Four Stete- feldt furnaces are now fired with great success by gas, made in Taylor gas-producers; namely, one at the Marsac mill, two at the Ontario mill, and one at the Holden mill. There are two gas-burners at the shaft, and one on the return-flue.

That the Stetefeldt furnace cannot be successfully used in its present form as an oxidizer of ore or matte containing a large per- centage of sulphur, is by no means a settled fact. The experi- ments to which Mr. Douglas refers were not reliable. Mr.Terhune himself acknowledges that they were somewhat hurried and incom- plete. Messrs. Terhune and Austin, who conducted the experiment**, were not familiar with the running of the furnace. Besides, the Marsac furnace being new, and having been fired for the first time, it was not in normal condition. Under the circumstances, the ca- pacity of the furnace was evidently overtaxed, a car-loa4 of Dixon ore, with 32.5 per cent, of sulphur, having been put through in 5J hours. Of the matte, only 3 tons were treated. Finally, Mr. Ter-

660 Summary Of American Improvements And Inventions.

hune, in giving the percentage of sulphur left in the roasted material, does not say whether this refers to sulphur combined as sulphide, or to the sulphur included in sulphate of lead. The Dixon ore con- tained 16 per cent, and the matte 13 per cent., of lead.

It is evident that the results of an experiment occupying a few hours, made in a hurry, with a new furnace, and by men not familiar with its operations, cannot be safely used as the basis for a wholesale condemnation, as is done by Mr. Douglas in his paper.

While it is rather doubtful whether the Stetefeldt furnace can be successfully U8ed for a dead-roast of ores and metallurgical products high in percentage of sulphur, it is highly probable that it can be utilized with profit in preparing ores for matte-smelting or for the blast-furnace.

The reduceil loss in silver alone would pay for finer crushing, as compared with other furnaces in which the time of roasting accom- plishes the deeired result.

Wm. p. Blake, New Haven, Conn., and Shullsburg, Wis. (communication to the Secretary) : Under the heading of Fwr- nac€8 for the Cafcinaiion of Ore, Mr. Douglas makes no refer- ence to the new form of calcining-furnace in operation for the past two years at the works of the Wisconsin Lead and Zinc Company, at Helena, 3 miles west of Shullsburg, Wisconsin, by which remark- able results have been obtained in the calcination of mixed ores of blende, galenite, and pyrite — the pyrite being roasted without roast- ing the blende, and without melting the galenite. This furnace was described by me at the Montreal meeting last February,* and a pa- per upon the process of treating such ores is presented at this meet-

Under the heading of Crushing Machinery Mr. Douglas fails to mention the multiple-jaw crusher for fine cru8hing,Jan invention of Mr. Theodore A. Blake, now in successful operation, which may be considered the most important and valuable improvement intbisline since the production of the original jaw-crusher.

Mention should also be made of the great improvements and mod- ifications of Chilean mills, as made by Frazer and Chalmers, a" others, by which their efficiency and value are greatly increased.

A New Form of Furnvce for Koahting and Oxidizing Ores,** Tram.,

t " The Seimration of Blende from Pyrites— a New Metallurgical Induatrr, Trani, xxii., 669.

t See Tram xiii., 210, and xvi., 753.

The Bessemer Process As Conducted In Sweden. 661

Under the metallurgy of zinc, it would be well to mention the util- ization of the sulphurous acid fumes from the blende for the pro- duction on a large scale of sulphuric acid, as practiced at the works of Matthiessen & Hegeler at LaSalle, Illinois ; and also the pro- duction of a white pigment by the Bartlett process,* from the mixed ores of zinc and lead, as carrfed on at Joplin, Mo., by the Picher Lead Co. — a process of great interest, which gives white amorphous lead-sulphate suitable for paint. And in this connection, reference should be made to another distinctive American metallurgical ad- vance in the treatment of argentiferous and auriferous pyritic zinc- ores, by F. D. Bartlett, at Gallon City, Colorado, by which the zinc is utilized in the form of oxide, while the precious metals are re- tained in the matte.

It is with some reluctance that I direct attention to these omis- sions, which I feel sure are the result of inadvertence on the part of my friend, Mr. Douglas, in consequence of his pressing engagements in the field.

THE BE88EMEB PROCESS AS CONDUCTED IN SWEDEN, f

o

Discussion of the Paper of Prof. Akermann. (See p. 265.)

Joseph Hartshorne, Pottstown, Pa. : I have read Professor Akerman's valuable paper with great interest. Few of the present generation of American steel metallurgists are aware of the very important part taken by Swedish engineers in the successful develop- ment of the Bessemer process, and I am afraid some of the older ones liave forgotten it. Certainly, I think, most of us would have said that Terre Noire was the first to introduce the use of metal direct from the blast-furnace. It is very well, therefore, that honor should be given to whom honor is due.

I can speak from personal experience upon one part only of Prof. Akerman's paper; namely, the effect of the temperature at which the molten metal is cast upon the quality of the finished steel. In a general way I have, of course, known for a long time that the

Trans xviii., 674.

t Note by the Secretary.— The following contributions to this discussion, whether presented at the session in Chicago or not, have been written out and re- vised by their authors since.

662 The Bessemer Pkocess As Conducted In Sweden.

colder the metal was cast, provided it ran cleanly out of the ladle, the better the ingot would be. It has only been, however, since I have been engaged in the manufacture of soft steel by the basic Bessemer process that the extreme importance of temperature in casting has been forced upon me.

As the grade of steel made becomes softer, the range of permissible temi>erature becomes smaller, and the necessity for keeping within this range becomes greater.

It is very difficult to convey an idea of temjierature when there is no means of measuring it ; but a rough scale can be formed from the amount of skull left in the ladle. At Pottstown, our heats of steel generally weigh about 24,000 pounds, and are poured either into four ingots of 6000 pounds, or six ingots of 4000 pounds each, ''he time occupied in casting is from fifteen to twenty-two minutes. We use two sizes of nozzles, one of which is 1 J inches, and the other is 2 inches, in diameter.

In making ordinary soft steel, containing carbon from .10 to .15 per cent., phosphorus .06 to .07 per cent., sulphur about .05 per cent., and manganese from .30 to .40 per cent., there is a consider- able range of temperature throughout which good results can be obtained. This grade of steel we cast through 2-inch nozzles in from fifteen to eighteen minutes. The results seem to be about the same, whether the steel, during casting, is somewhat more than hot enough to leave the ladleclear of skull, or whether it is cold enough to leave a skull of from 600 to 800 pounds. There is, however, enough variation to warrant the l)elief that a temperature which leaves about 200 to 400 pounds of skull in the ladle is the best. Of course, it is to be understood that the casting is done as slowly as possible in all instances.

In making very soft steel, however, the permissible range is much smaller. Such steel will contain by percentages .07 to .10 of carbon, .01 to .03 of phosphorus, below .04 of sulphur, and from .10 to .20 of manganese. The softest I have seen was C, .09 ; S, .031 ; P, .015; Mn, .095. For such steel the Ij-inch nozzle is the best, pro- vided it can be kept open and clean. The casting through this nozzle takes about twenty minutes. The best results are obtained when there remains from 200 to 500 pounds of skull in the ladle.

There is no trouble in making the above steel in the converter, provided the proper temperature be maintained in casting. It will roll perfectly, without the slightest trace of red-shortness. It will go through the mechanical tests to which we submit every heat of steel.

The Bessemer Process As Conducted In Sweden. 663

with perfect success. These tests I have described in my paper, read at the Reading meeting of the Institute, October, 1892 {TVans, xxi., 743), and it is not necessary to describe them again. It will be remembered that they are very searching for brittleness, red- shortness, and defects in general soundness.

On the other hand, if the steel be poured when too hot, the blooms or billets will have surface-cracks and striations, and will have a great tendency to open up and tear in rolling. The ingots, when investigated, have invariably shown the blow-holes in great number very near the surface. It has sometimes happened that an ingot which has been cast too hot has been taken from the heating- furnace and has been allowed to cool off. In such cases the surface of the ingot has been deeply pitted all over, showing that the blow- holes opened through the outer skin.

The mechanical tests of such heats show precisely the same char- acteristics. Another peculiarity is that the tests show a much more brittle steel than when the casting has been done at the proper tem- perature.

When the casting of this "dead'' soft steel has been skilfully done, the tests show an absolutely " dead " soft, homogeneous, and solid metal. The surfaces and edges are without cracks, checks, or flaws of any kind, and the fractures are perfectly solid and fibrous (so-called).

If the steel has been cast too hot, the bending-tests will show cracks at the bends; the surfaces will show hair-cracks and stria- tions ; the edges will show cracks and flaws; and the ear-test will have many of the indications of red -shortness ; while the fractures will show an open, porous, and "stringy " metal. These character- istics show themselves in the finished product, but to a less marked degree.

I have not yet been able to satisfy myself fully as to all the reasons for such diflferent results being obtained from steel of practically identical compositions, as shown by analysis. There is no doubt, however, that the surface blow-holes, described by Mr. Capersson, play the most important part in causing them. Sometimes, though very rarely, the metal contains too much oxide; but this is very easily recognized by the fractures of the tests. In such cases the excessive temperature has evidently been caused by over-blowing.

It is evident, therefore, that the temperature plays a very import- ant part in the manufacture of the softest grades of Bessemer steel, and that the greatest care must be exercised in controlling it.

664 The Bessemer Process As Conducted In Sweden'.

The same rnle applies in the casting of o|>en-hearth steel; but the permissible range of temperature is greater. This, of course, arises from the fact that less gas is contained in steel made in the furnace. It is also true, however, that when proper care is exercised in manipulation, just as good steel can be made in the converter as in the furnace. It is impoasible to distinguish our softest steels or to tell by any tests, physical or chemical, by which process they have been made. This is, I think, solely due to the care and skill exercised in regulating the temperature in casting. I must also con- fess that it is only lately that we have convinced ourselves of this truth, and have attained such skill.

Prof. Akerman has ascribed to the fluidity of the slag the small tendency to red -shortness observed in Swedish Bessemer steel. In my opinion it is to the same cause that the basic Bessemer process owes its comparative immunity from the same trouble. In all successful basic " blows,*' the slag is at least as thin as molasses on a warm day. This undoubtedly helps to wash the metal free from oxides. To this reason must probably be ascribed the fact that but little, if any, more manganese is now needed to be added in the basic than in the acid process in order to prevent red-short- ness. The natural inference from the use of the after-blow would be that more manganese would have to be added in the basic pro- cess, since more iron is oxidized. This was, indeed, generally the case in the early days of the process, as is shown by the remarks of Thomas, Holley and others. Probably, also, the dilution of the oxides by the large amount of lime added has its influence in the same direction.

The thin slag is also a factor in causing the low loss reported by Prof. Akerman, since less metal is retained in such slag in the form of shot. The basic process has the same advantage as to thin slag, but this is more than offset by the greater bulk of the slag, which is two-and-a-half to three times greater than in the basic process.

The Caspersson ladle is no doubt a valuable and even a necessary adjunct to the Swedish works, where the charges are small, com- paratively cold irons are used, and the rate of production is slow- Most of its advantages can be obtained, however, by allowing the metal to remain a few minutes in the vessel after the addition oi ferro-manganese, and by the use of shallow, dish-shaped ladles, such as those in use at Witkowitz and at the Edgar Thomson works.

The strainer may be a very good thing, but it strikes me that a man must be very sure of his temperature before he tries any ex-

The Bessemer Process As Conducted In Sweden. 665

o

periraents with it. I regret that Prof Akerman did not give the dimensions of tlie two sizes of holes.

Finally, I must again express my sincere admiration for the work of the Swedish Bessemer and blast-furnace engineers, as de- scribed in this admirable paper.

W. F. DoRFEE, West New Brighton, N. Y. : There is no doubt whatever that Consul G. F. Goransson rendered important services to Sir Henry Bessemer and the world at large by his very intelligent action relative to the practical details of manipulations in the early days of the Bessemer process — services which have not received that generous recognition, hitherto, to which their importance clearly entitles them. Prof. Akerman's paper fully confirms, with regard to this matter, information from trustworthy sources of which I have long been in possession.

This paper is of special interest to me, as some of its statements relative to practice in Sweden fully agree with conclusions at which I arrived, independently, in the course of the o|>eration8 at the experimental steel works at Wyandotte, in 1862-65.

In a paper read before the American Society of Mechanical En- gineers, in November, 1884 (Trans. Am. Soo. Mech, Eng,y vol. vi., p. 40), descriptive of the " Experimental 5teel Works at Wyan- dotte," I said :

The engine which supplied the blast for the converter was constructed from workings-drawings made by the writer. It was intended to produce a pressure of blast of 16 pounds per square inch, which was about double the pressure used at that lime for any metallurgical work, and was regarded as very heavy — in fact I was informed, at the time of commencing the plans for this engine (the winter of 18C3) that the pressure used for blowing steel in England and Sweden was but 8 pounds. I adopted the higher pressure with a view to shortening the time re- quired for a blow, in the full belief that I was taking a decided step forward in the practice of the pneumatic process, though in this I soon became satisfied that I was in error. But whatever mistakes I made in this matter of blast pressure, I had the comforting satisfaction of finding myself in most excellent company, for before my engine was finished, steel was blown in England with a blast-pressure of 25 pounds, a practice which has continued onto the present time."

Again, in the discussion of Mr. Howe's " Notes on the Bessemer Process" (TVcm/?., xix., 1170), I said, relative to the same experi- mental works :

" I was firmly persuaded, after a few blows, that the movement towards an increase of pressure was a mistaken one, and I am just as much persuaded at the present moment that the ponderous engines that have been designed to produce 20 to 25 pounds pressure, are a totally unnecessary expenditure of capital, and that equally good results can be obtained with a much more moderate pressure."

666 The Bessemer Process As Wnducted In Sweden.

These views of mine having been severely criticized, it is with no small satisfaction that I read in Prof. Akerman's paper that, until the middle of 1858, Mr. ttoransson " had succeeded only excep- tionally in turning out a good product while following the advice of the inventor to lay the greatest weight on having a high pressure of blast. By departing from that advice and securing instead by means of a large tuyere-area, an abundant supply of blad,* Go- ransson was able, beginning with the 18th of July, 1858, so to shorten the time necessary for the process and thereby increase the heal of the blow, tliot an improved product was obtained Concerning the re- sult, obtained! solely by the placing of larger and more numerous tuyeres in the converter bottom, it is refreshing to read: "The blow was warm and regular, the steel flowed so readily, and was so liquid, that all the slag rose through it to the top, and the ingots obtained were free from slag. Continued efforts in the same direc- tion showed that we were now on the right path, and that the prob- lem was solved.*'

This result was obtained by a low pressure of blast, and with metal taken from a charcoal blast-furnace, which was frequently done at Wyandotte (in fact the plant was designed with especial ref- erence to the use of metal direct from the blast- furnacef), with a result too often quite similar to the early experience of Goran&son at Edske, as described on page 266 of Prof. Akerman's pajjer.

Such results, I was fully convinced, were due in no small degree to the pressure of the blast used (14 to 16 pounds), by which much of the oxygen was forcibly driven through the metal without com- bining with the silicon and carbon therein — the power of the engine not being adequate to furnish a sufficient amount of oxygen to pass through the metal in a free state and still supply enough for promptly effecting the reactions required.

Such having been my conclusions, I proposed to obviate this difficulty by the use of a larger volume of air in a given time, at a much lower pressure; and with this in view in the design of the stationary converter which was erected at Wyandotte just before I closed my connection with the works, I increased the number of tuyeres to 13 (there were but 7 of the same size in the tilting con-

The Holies are mine.— W. F. D

t See my paper in Trans. Am. Soc. of Mech. Eng.f already cited.

X Illustrations on page 49 of the paper already referred to. In reference to tbw

controversy I said in that paper : 'it will be noticed that the upper nd

outer ends of the tuyeres are above the level of the metal in the converter, and

The Bessemer Process As Conducted In Sweden. 667

verter), each having 7 holes of an inch in diameter. The lower ends of these tuyeres were intended to be not more than 6 inches below the surface of the metal (in the tilting converter they were from 18 to 21 inches below the surface). It will therefore be evi- dent that I had made a very liberal provision for blowing with a re- duced pressure and increased volume of blast. My reasoning at that time was exactly in line with, although altogether independent of, that of the early practice in Sweden, which has been persevered in and found so advantageous ; and it is no small satisfaction to learn that the views I advanced nearly 30 years ago coincide so exactly with the practical and profitable experience in Sweden from the ear- liest days of the Bessemer process in that country ; and I regard that fact as a sufficient and decisive reply to all whose criticisms and prac- tices have been adverse to the views of blast-pressure I have persist- ently advocated.

It is to be hoped that the interest of economical production of Bes- semer steel in America will be served by the reading and publication of this most interesting and instructive paper, and that our Bessemer steel works will, in the interest and under the stimulus of an enlight- ened selfisfhness, abandon at once and forever the wasteful practice of high-pressure blowing. It would be a profitable retrospect for those who have followed that practice to calculate, in the light of the evi- dence thus obtainable, the amount of capital that has been in past years absolutely wasted on plant and fuel for excessive blast-pressures.

H. H. Campbell. Steelton, Pa. : In Table II. of Prof. Aker- man's paper there are given analyses of slags and metals showing a reduction of the iron oxide in the slag during the process of blowing,

that their lower ends are but a few inches below that level. Therefore the pressure of blast necessary to overcome the ferrostatic head at the inner ends of the tuyeres was much less than would have been required had the tuyeres been inserted in the bottom. This arrangement also permitted the blast to be stopped without fear of the metal in the converter running through the tuyeres. One of my reasons for thus placing the tuyeres was the belief that such heavy pressure as was required for bottom-blown converters, was not at all essential to the production of a good quality of steel. My reasoning was, that to expel the carbon from the metal under treatment, it was only necessary to bring in contact with it the requisite number of atoms of oxygen, which I believed (and still believe) could be better accomplished by a large volume and low pressure, than by a small volume and high pressure of blast. I have had no opportunity of practically demonstrating this belief, but will here venture the prediction that one direction in which the genius of improvement will walk in our steel works, is that which leads to a greater reduction of pressure of blast, and a corresponding diminution of the power required for its production.''

668 The Bessemer Process As Conducted In Sweden.

and from these data the conclusion is drawn "that the oxidation of the carbon must have been performed in an easential degree by the reduction of oxidized iron from the slag." There are no quantita- tive determinations given, but we can estimate from the figures at hand something concerning the weights involved. The figures given for Langhy ttan are decidedly different from those of the other works, and show a very much greater alteration in the slag; but even in this case it can be shown that the effect is not as great as the author would assume. It is elsewhere stated that the average weight of a Swedish Bessemer charge is about 7000 pounds. Whether this is true of Langhyttan, or not, is a matter of small importance, since the weights of slag and metal must always be proportionate. The composition of the pig-iron showed 0.64 per cent, of manganese ; after blowing two minutes and fifteen seconds, the metal gave 0.12 per cent. ; thus the oxidation of manganese was 0.52 per cent, of 7000 pounds, or 36.4 pounds, producing 47 pounds of MnO. The percentage of MnO in the slag at the expiration of that interval was 13.95 ; and since no manganese entered the slag except from the metal, the weight of the slag must have been 337 pounds.

That this weight of slag must be nearly correct may be shown from the following calculation on the basis of silicon: During the same interval there was oxidized 1.10 per cent, silicon, or 77 pounds, giving 165 pounds of SiOg. The percentage of SiOj in the slag was 48.76. This gives a weight of slag of 338 pounds, agreeing very closely with the calculation on the basis of manganese. The silica in this first slag comes almost entirely from the oxidation of silicon; for there would be very little scorification of the lining during the first interval, as the metal would be extremely hot, and there would be no demand for silica, the oxidation of the silicon supplying the chemical requirements. This corroboration therefore by the silica- calculation is not without value ; and the figure, 337 pounds, will be taken as the weight of this first slag. For the succeeding periods it would be possible to use manganese as the basis of calculation; it is also possible to use CaO, MgO, and AljOj. The lime, mag- nesia and alumina must have been derived from a small amount of blast-furnace slag which was poured into the converter at the begin- ning of the operation. The exact quantity may be found from the analysis and weight of the first slag, and this quantity will remain practically unaltered throughout the heat.

The calculation is as follows :

THE BESSEMER PROCESS AS CX)NDnCTED IK SWEDEN. 669

On the basis of

CaO... I MgO.., I A1,0,.

Per cent.

in Slagl.

If

Co

"

337 13.95

It will be seen that the results from CaO and MgO display remarkable uniformity. For AI2O3 there is a slight disagreement; but it must be remembered that there is a slight addition of AljO, from the scorifieation of the lining. The figures from MnO difler still more widely; but on the whole the results approximate so closely to one another that their general truth seems assured.

Assuming certain weights for the slags in question, we can check the figures by finding the weight of silica present, as follows :

Weight of slag. Pounds.

Slagl 337

Slag 2, 270

Slag 3. 380

It may seem to the American metallurgist that the weight of slag is too small ; but it must be remembered that the initial pig-iron contains only about two-thirds the percentage of silicon used in our practice, and that silica is usually the determining factor in the con- struction of the slag. Calculating from these corroborated weights we have the following history of FeO :

Sio,.

Per cent.

SiO,.

Pounds.

Slagl, Slag 2, Slag 3,

Weight

of slag.

FeO.

Pounds.

Per cent.

Pounds.

During the first interval between slag 1 and slag 2, during which time the percentage of FeO in the slag was reduced from 34.72 per cent, to 21.08 per cent, the quantitative calculation shows a reduc- tion of 60 pounds of FeO, which supplies 14 pounds of oxygen. During this interval the carbon was reduced from 4.2 to 1.1 per cent This implies a combustion of 3.1 per cent, or 217 pounds of carbon, requiring 289 pounds of oxygen for its conversion into CO.

670 The Bessemer Process As Conducted In Sweden.

Thus the oxygen from the slag was only or 5 per cent, of the total oxidation. It would seem as if there was little warrant for the conclusion that the slag in so inii)ortant an agent, especially since the slags above taken are the extreme instances in the table.

It is possible, of course, that there is a continual journey of oxygen from the blast to the slag and then to the metal, but such a supposi- tion seems gratuitous. In the open-hearth this action occurs because no oxygen can reach the metal without passing through the slag; but in the Bessemer converter the exact reverse is true, and do oxygen can reach the slag without going through the metal, except when slag and metal are churned by the blast into a foam.

The foregoing criticism is intended to show that it is time that the quantitative history of the Bessemer process should be written. By taking the weight of slag produced during a long run, with cor- responding analyses, it is possible to work out results which will be free from the suppositions embodied in the above calculations.

The observations of Prof Akerman on the relation between the temperature of the metal and its behavior during casting are valu- able as a record of Swedish practice. It is necessary to add, how- ever, that with a change in some of the minor details of manufacture a change in this relation is possible. With a given set of conditions consistent with good practice high temperature produces a certain effect. With another set of conditions, also consistent with good practice, high temperature may produce exactly the contrary effect. We have not yet reached the point where we can equate all the manifold variables and work out a universal value for one factor. Contradictory evidence comes from different works, and every one who has had long experience with the practical casting of metal made by different processes must have met with so many exceptions to any conceived relation that a formula seems unattainable.

During the present year (1893) a manager of one of our laiest American steel-works has declared that he finds it impossible to make solid Bessemer ingots from a high-silicon pig-iron, no matter how well the temperature may be regulated by steam or scrap. It is unnecessary to say that many a blower has accomplished this regu- larly. Ten years ago, on a visit to Valenciennes, I found them blowing so hot that the steel contained from 0.3 to 0.4 per cent, sili- con. 'I'his was done purposely, in order to obtain certain physical requirements which could be obtained in no other way. In Grer- many it was the practice to blow still hotter, and to leave from 0.4 to 0.6 silicon in the metal.

The Bessemer Process As Oonddcted In Sweden. 671

Notwithstanding such practice has been successfully followed in other countries without any attendant trouble, it would probably be impossible to find a single large American steel- works that would attempt to carry on such work, or that could hope for less than 20 per cent, of second-quality rails as a result.

Nothing is more firmly established in the mind of the American blower than the fact that hot metal gives bad results in the rolling- mill ; and yet every such man can recall cases where metal of this character has worked perfectly. He can also recall more instances where steel of good analysis and without any known bar sinister in its pedigree has given the worst results. I have seen a low-heat blown so hot that the metal contained 0.62 per cent, of silicon. It stood 45 minutes in the vessel before recarburization and when poured was as warm as any heat should be. But the billets ham- mered from rolled blooms of this heat were full of seams and unfit for use.

These facts indicate that there are still some things to be learned about steel-making; but they do not in the least encourage the idea that there is any mystery in the case. That word has no place in science.

In the world of iron metallurgy there is a domain which has not yet been thoroughly explored. It is possible that the microscopic investigations, the first-fruits of which are given this year to the Congress, may result in shedding light upon this unknown terri- tory.

J. L. Sebemus, Sweden : Every steel -manufacturer is well ac- quainted with the many and great difficulties caused by blow-holes in steel ingots, especially when it is desired to produce steel of the best quality. It is well-known that the number and size of the blow- holes in the ingots increase, as the amounts of phosphorus, sulphur, manganese and silicon decrease. For instance, a steel rich in silicon may give a fairly good ingot (although a number of blow-holes may be present in that case also), while a steel fit for the manufacture of tools, etc., gives an ingot, which, when broken open, presents the appearance of a sponge.

In hammering and rolling the steel, these blow-holes, of course, are contracted, their walls are pressed together, and at first glance, the bar or rail appears to be perfectly sound and homogeneous. A closer examination, however, reveals that this is not the case, but that through the rolling and stretching the walls of the blow-holes have been brought in closer proximity to each other, but without

The Bessemer Process As Conducted In Sweden. 673

any joining whatever, and thus the result of the stretching process is only that the holes have been pressed out and the enclosed gases have been compressed.

During the many years I have been engaged in the manufacture of steel, I have tried many different ways and means to prevent the forming of blow-holes, and by using silicon and aluminum I have partly succeeded. But what has been gained as to the density of the material has been lost in its purity and quality. Furthermore, by adding silicon and aluminum, it is not |x>ssible to rid the steel of the large amount of gases, absorbed by it in its fluid state. These gases are merely distributed, and the result is a large number of very small blow-holes, throughout the ingot.

Only during the last few years have I been able to overcome all difficulties, and to get rid of the gases, without adding any detri- mental substances. And I am now able to produce steel ingots, per- fectly sound and of superior quality.

This result was reached after I conceived the idea of bringing the liquid steel directly at the close of casting under the influence of centrifugal force.

As the steel gradually solidifies, the gases, which heretofore have been dissolved in it, are liberated. If no force, or only an insufficient force, is acting upon the steel, these gases remain inclosed in the ingot, forming bubbles or blow-holes ; but if, on the contrary, an adequate centrifugal force is acting on the metal, they are forced to leave their position and move towards the center of rotation.

The next step was to construct an apparatus by means of which the melted steel could be brought under the influence of centrifugal force.

The work performed by this apparatus is such as to give entire satisfaction. As far as simplicity of construction, strength, and du- rability is concerned, I believe nothing more can be expected.

Two of these rotators are now in operation in Sweden, and several are now under construction in other countries.

The machine is shown in the accompanying drawings. It con- sists of a horizontal yoke, attached firmly to a vertical shaft. The yoke itself is composed of arms set radially at right angles to each other, and to the end of each of which is fastened, by means of a pivot, a strong steel trap, movable in the vertical plane around the pivot. In this trap the moulds are placed. When the apparatus is set in rotation, the trap and the moulds, with an increasing speed, deviate from the vertical position, until finally, when full speed is

? Ic

674 The Bessemer Process As Conducted In Sweden.

attained, the moulds have taken up a radial position with the open ends turned towards the center of rotation.

When the machine is brought to rest, the moulds again take up their former vertical position.

The liquid steel is poured into the moulds above referred to from one or two ladles, so constructed as to fill four moulds at once, thus avoiding unnecessary loss of time. The fluidity of the liquid steel is an important factor in attaining the desired results.

As soon as the moulds are filled, the rotator is set in operation, and a speed of about 120 revolutions per minute is maintained until the steel in the mould is solidified.

During the whole time of rotation gases are seen to escape from the open ends of the moulds, and when they cease to do so it is known that the steel has taken up the solid state. For ingots of, for instance, about 15 inches square, the solidification will require about ten minutes; and for every ton, treated in this way, about five to six horse-power will be required.

A model, on scale, of such a machine, constructed for 15-ton charges, also 8|)ecimens of the "centrifugaled "ingots, were exhibited in the Swedish building at the Columbian Exposition.

If Bessemer or open-hearth steel is "centrifugaled," the resulting quality is the same as that of ordinary crucible-steel of the same chemical composition. This can readily be explained: The crucible cannot change the steel in any other way than by making it denser, and thus, the ingots sounder; and, if this soundness can be effected by other means, crucible- and "centrifugaled' steel of the same chemical composition must show the same physical qualities.

In teel of a high percentage of carbon, the ingots, as generally made, show an inner core clearly harder or higher in carbon than the rest of the ingot, which feature, of course, is very objectionable.

The centrifugaled " steel does not show this defect. The physi- cal and chemical conditions are the same throughout the entire ingot.

Here, however, it must be observed, that steel of high percentage of carbon shows, on the upper surface of such a " centrifugaled " ingot, a light skin, a fraction of an inch in thickness, which proves to be somewhat harder than the rest of the ingot. This fact I should explain in the following manner :

When, in an ordinary ingot, the liquid steel is solidifying, this action commences at the outer edge of the ingot and, proceeding in- wards, particles of carbon refuse, if I may use such an expression, to unite with the solidifying mass of steel, and these particles, m

Improved 8Lag-Pot8. 675

the form of carbon, or, it may be, a carbide of iron, work their way towards the liquid center of the ingot, where they are ulti- mately captured by the final solidification of the entire mass of steel. In " centrifugal ing/* on the other hand, these particles, whether carbon or carbide, on account of the centrifugal force, and their lighter specific gravity, are forced out of the liquid steel, and, as they can hardly be expected to disappear, as the gases do, they nat- urally must be somewhere in the ingot, and they are, in fact, found in the very edge of the upper end. Here, however, they do no harm, since some small part of the end of the rolledout ingot has to be sacrificed anyhow, on account of the "piping." When this small part has been cut off, every rod, bar, rail, or plate, manufac- tured from "centrifugaled " steel is perfectly homogeneous, and shows throughout exactly the same chemical qualities.

The drawing, and also the model exhibited in the Exposition, show a simple, durable, and inexpensive machine. Once built, it requires but little attendance.

Through this method the following advantages have been secured :

1. Ingots are obtained free from blow-holes and sound, without adding any detrimental substances whatever.

2. The amount of carbon in the steel is evenly distributed through- out the whole ingot.

3. The piping is reduced about 60 per cent.

4. The amount of fuel used in the heating-furnace is lessened, for the reason that, as there are no blow-holes, no welding-heat is needed ; and, for the same reason, time is saved, and also the material that would be lost through oxidation in a more intense heat. All the heat the ingot requires is only what is needed to make it pliable for the rolling-mill.

5. The steel, after being " centrifugaled," shows all the physical qualities of a crucible-steel of the same chemical composition.

IMPROVED SLAG' POTS. Discussion of the Paper of Mr. Keller. (See p. 574.)

James W. Neill, Salt Lake City, Utah : Mr. Keller observes, in his interesting paper, that his copper mattes show a constant tenor of 21 to 23 per cent, sulphur ; and continues, Many of these analy-

676 Improved Slao-Potb.

868 8howed the presence of magnetic iron, sometimes in considerable quantity. Copper-mattes would accordingly correspond to either of these formulas :

(CuS). + or, (CiS). + (FeS), + (Fe,0,)."

Judging from my own long experience with lead-copper and lead- copper-uickel mattes as also from the literature of lead-smelting, I think that both of these formulse are wrong, and that the magnetic quality of the mattes is due to the presence either of metallic iron or of the magnetic iron sulphides FeSg, FegS, etc.

Hofman, in his MetaUurgy of Lead (1893), pp. 253 and 254, says:

It oAen happens that, if the constituents of a matte are figured as sulpbides, the analysis does not show enough sulphur to combine with the metals as FeS, PbS, Cu,S, NiSy etc. The explanation given by Rammelsberg for lead sulphide, and by Monster and Schweder for iron and nickel sulphide, is that, as subeulphides do not exist, lead, iron, and nickel are held in solution by their sulphides while liquid, and separate out during solidification. Mackintosh claims the existence of a subsulphide ot nickel and iron."

From this we see that the existence of Fe is disputed ; hence, we certainly cannot attribute the magnetic qualities of these mattes to this source.

That the mattes should contain FcgO, I consider most unlikely, as the conditions of blast-furnace practice are against it. If the mattes could contain FeO, we ought certainly to expect the slags, which contain all their iron in an oxidized state, to contain also in yet larger quantities, and, therefore, to be more magnetic than mattes; but this is not the case. I have tested many lead-furnace slags for their magnetic qualities, and have found very few which are even slightly magnetic. On this point, I will again quote Hofman's Metallurgy of Lead (p. 138).

'' Last may be mentioned the magnetic property of some lead-slags to which lies first called attention. This is caused by the presence of magnetite or magnetic sul- phide of iron (FegSy). How the magnetic oxide gets into the slag is a matter for further investigation. Hahn suggests an incomplete reduction of ferric oxide ; Guyard thinks it results from the oxidization of metallic iron by lead-oxide: 4PbO+ 3Fe FeA -h 4Pb."

That the sulphides of the matte-forming metals — lead, copper, iron, and nickel — dissolve, and hold these metals in solution while liquid, is partly confirmed by the following experience.

While concentrating lead-nickel mattes at Mine La Motte, Mo.,

IMPROVED sLAO-pars. 677

in 1881 or 1882, ray attention was called to the oocurrenoe of minute metallic crystals in small cavities in the matte-cones. These crystals were as thin as a knife-blade, of metallic luster, white color, and form of crystals'zation not definable. The workmen always called them 'Mead." A test showed them to be flexible; but, when bent to spring back straight, they were attractable by the magnet ; and a qualitative test with chemicals showed reactions for nickel and co- balt. I could not collect enough for a quantitative analysis. The matte in question assayed, as nearly as I recollect, 25 to 28 percent, of lead, 12 to 16 per cent of nickel, and 6 or 7 per cent, of copper.

A still more familiar instance of the same reaction is the occur- rence of hair- or moss-copper in copjieimattes of a certain grade ; the peculiarity here being that, with increased percentages of copper in the matt£, this metallic copper does not separate out. Perhaps, in this case, the sulphide of iron is the solvent for the copper, and with the disappearance of this the copper forms a sulphide.

For a number of years past, I have tested lead-copper mattes for their magnetic qualities, and have invariably found them attract- able by the magnet. When very high in lead or copper, they are less magnetic than when high in iron. From this I should judge that Mr. Keller's mattes have also varied in their magnetic qualities with their contents in iron and copper.

As already observed in these Transactions (xx., p. 580), I have used the differenc*t in magnetic qualities of matte and slag to effect (ex- perimentally) the mechanical separation of the two, in crushed sam- ples of "slag-shells," which contained considerable amounts of matte in the form of shot, and hence also considerable silver. Some of the data so obtained may be of interest :

Aflffay valae. Ounces of silver per ton. Shell from top of pot (crushed through SO-meeh), . . . 12.1

Magnetic material, . 48.4

Non*magDetic material, 5.5

Shells from copper-matte concentration (slag very " acid "), . 5.8

Magnetic material, 36.7

Non- magnetic material, 3.2

In the first instance, the magnetic material contained 64.4 per cent., and in the second case 38 per cent, of the total contents of the slag in silver.

The construction of the improved settling-pot, illustrated here- with,* is based upon this difference in the magnetic qualities of matte

♦ U. S. Patent, No. 505,904, October 3, 1893.

Impboved Slag-Pots.

and slag ; for I have determined that this difference also exists while the materials are yet in the fluid state. The figure represents a ver- tical section through the combination-settler.

Combiiutioii Settling-Pot.

The settling-pot, A, mounted as usual, upon wheels, a a, is, when in place before the furnace, made to come in contact with the electro- magnet, B, which 18 movable upon the lever C for convenient adjust- ment. When in place, the current from the dynamo is turned on, and the molten materials are tapped from the furnace into the pot. The attraction of the magnet, working in line with the action of gravity, aids and hastens the settling of the magnetic metals or materials, and thus makes a cleaner slag. Whilst the magnetic force of the magnet might not l>e felt at the top of the settling-pot when the same is empty, vet, when it is filled with molten materials, the magnetic force will be transmitted from one globule of magnetic material to another, just as a horse>shoe magnet will attract and hold a string of needles, transmitting its magnetism from one to another to a greater distance than its own direct attraction can be felt. It is evident that with this combination-settler a more viscid slag can be used, as the settling power of the matte has been doubled. This point is, in many cases, of great value to the metallurgist The ar- rangement is simple, inexpensive, and durable; the current is of low potential, and therefore harmless, and the cost of operating and exciting the magnet is a mere trifle. That the application can be varied to suit the individual requirements of the different smelt- iug-furnaces and methods, goes without saying.

The Op£N-Heabth Process. 679

The Open-Heabth Peocess*

Discussion of the Paper of Mr. Campbell. (See p. 345.)

George W. Goetz, Milwaukee, Wis, : Mr. Campbell deserves much credit for his interesting paper. The literature of the de- velopment of the open-hearth process is distributed in many tech- nical journals, and is difficult of access to many engineers. Since the'development of the basic process, the open- hearth furnace has been raised from a subsidiary to an independent position. Certain high grades of soft steel can be produced in the basic open- hearth furnace, which cannot be produced by any other known process.

In the Uniteil States the charges for the acid process generally consist of either scrap and pig-iron or so-called wash-metal and scrap : the pig-and-ore process, as carrie<l out in England, is practiced but very little. The introduction of the basic process has made it pos- sible to use about four times as much pig-iron with scrap as can be used in the acid mixtures. The availability of a cheap pig-iron, very low in silicon and sulphur, and carrying several per cent, of manganese, is one of the reasons why the basic open-hearth process has developed so rapidly in Germany, as regards the output per fur- nace and the excellent quality of the material produced.

The acid Bessemer steel has an inherent stiffness, which is plainly shown when spring steel is made by the Bessemer and by the open- hearth processes. In order to get them of about equal stiffness, the open-hearth steel must have about 0.60 per cent, carbon ; whereas, the Bessemer steel must then have about 0.45 carbon, the remaining constituents being apparently about the same in each product.

It has proved impracticable, with American coke-pig-iron, to in- terrupt blowing at a certain point in order to make hard steel ; and experience has therefore shown that it is best to eliminate nearly all the carbon, and then to recarburize to the desired point. This practice

Note by the Secretary. — The contributions to tliis discussion, whether made orally at the meeting, or not, have all been written out since, and are here printed from the manuscripts of the authors. — R. W. R.

HfiM

680 The Op£N-Heabth Process.

is liable to give steel not quite so homogeneous as that made in the open-hearth furnace, as described in the paper read.

It must be said, however, that in Sweden, with charcoal-pig-iron very low in sulphur, and rarely above 1 per cent, in silicon, the in- terrupting of theblow is done successfully, with the production of a very high grade of hard steel. I do not think that the same prac- tice could be carried out with our coke-iron, on account of the trouble likely to arise from residual silicon and other causes, not to speak of the time lost in making tests.

I have given some attention to the combination of the open- hearth and the converter, and must agree with Mr. Campbell that the duplex system is not economical for most American circumstances. I have made a numl)er of duplex charges under the superintendency of Mr. S. T. Wellman — transferring three Bessemer heats, each of 5 tons, to make up one heat of 15 tons in the open-hearth furnace. The average time required for transfer of the three heats, getting the desired dephosphorization and carbon, and tapping, was about three hours. The heats were taken from the acid Bessemer to a magnesia- lined open-hearth furnace.

It is true that a good dephosphQrization can be attained by blowing the heats to be transferred, in such a manner that the first heat to be transferred is left hard, whereas the following heats are blown some- what soft. The reaction which ensues upon pouring one heat into the other, will then cause a rapid dephosphorization when sufficient basic additions are present. But as soon as the reaction has subsi- ded, the dephosphorization proceeds slowly.

The oxidation going on during the melting is very im|x>rtant,not only as to carbon and silicon, but also as to phosphorus ; and the above experiments showed that the dephosphorization was made &r more satisfactory in the long run by melting the entire charge in the furnace. Many consecutive duplex heats arc also very destructive of the bottom, especially when a reaction is caused by the mixture of two or more heats from the converter. The time gained by using liquid pig-iron in the open-hearth, as well as in the puddling-fai nace, is more than offset by the advantages derived from the oxidation going on during melting.

In Austria, the duplex process is carried out by a combination of the basic Bessemer and basic open-hearth processes. I understand that Mr. Kupelwieser will read a paper at the next meeting of the British Iron and Steel Institute, on the duplex process now prao-

The Open-Heabth Process.

tioed at Witkowitz, in Austria. I have had an opportunity to see the process, but I found a combination of circumstances, especially as to the availability of scrap, and the dependence upon certain grades of pig-iron, which do not exist here. Mr. Campbell gives the main reasons why the duplex process is not economical at present for Amer- ican circumstances.

As to the construction of furnaces, I should think that the tip- ping-furnaces present the advantage of being able to pour off slag and to regulate the flow of metal into the ladle. A sudden rush of a large quantity of slag and steel when tapping, which sometimes causes the ladle to run over, can thus be avoided.

mOmm

Open-Hearth Furnace and Charging-Machine.

The stationary furnaces, a sketch of which is shown in Mr. Campbell's paper, do not represent the latest constructions in use at some of our principal works. The sketch shown herewith, shows a 20-ton furnace, as designed by Mr. 8. T. Wellman, of the Well- man Iron and Steel Company. The high and straight roof, the pockets in front of the regenerative chambers, and the placing of the latter under the charging-floor, instead of under the furnace, are the special features. The mushroom-valves shown, as well as the doors, are manipulated by compressed air. This furnace has given excellent results during the last five years. A charging-maehine is also shown in the sketch.

In his chapter on furnace-gases, Mr. Campbell overlooks the im- portance of the hydro-carbons in securing a strongly luminous flame in the furnace. As to the quantity of heat, his figures are correct, but he has not considered the effect of a luminous flame on the economy of smelting, and the fact that the hydro-carbons are necessary to obtain this luminosity. A 10-ton furnace could not be

682 The Open-Hearth Process.

brought up to a steel-melting temperature by a producer-gas made out of coke, although the carbonic acid was below 2 per cent ; bat as soon as sufficient soft coal was charged upon the coke to make the flame luminous, the furnace heated up rapidly.

Natural gas fills up a checker- work with solid carbon very rapidly, but this can be prevented by admitting a little steam with it l>efore it enters the checker-work, and then all deposition of carbon is pre- vented.

To let carbon, which was deposited by the deoomposition of hydro- carbons in contact with a hot checker- work, burn out when reversing the furnace is bad practice. Whether a gas deposits carlion can be easily seen by having an observation-hole (closed by glass) for the checker-work. If carbon is deposited, the moist\ire necessary to remove it in the form of water-gas is easily determined ; thus it is consumed in the furnaces, and is not wasted to the chimney. Care must, of course, be taken that no large excess of steam is given. There are very few grades of coal, however, which do not give more moisture to the gas than is necessary.

During the last five years very much has been said and written about luminous and non-luminous flames, especially by Mr. Frederick Siemens, of Dresden. The writer had the pleasure of a three days* visit in Dresden, and fully convinced himself by argument, and by actual observation in the glass-works, of the correctness of Mr. Siemens's views on the conditions necessary for the economical com- bustion of gas, and the beneficial results of luminous flames, in com- parison with non-luminous ones on the Bunsen principle.

Before giving Mr. Siemens's views, as I obtained them from him, I wish to quote several facts mentioned in a debate on this subject before a number of British gas-engineers.

" It is an undoubted fact that the heat of the Bunsen burner is more highly concentrated than that of the illuminating flame, and it is, therefore, anivereally adopted for boiling-purposes where it is desirable to concentrate the whole heat on the vessel. For this purpose the Bunsen flame possesses another very desinble qualification, viz., that it does not deposit soot on the bottom of the vessel against which it is directed, and thereby does not obstruct the conduction of the heat by the intervention of a powerful non-conductor.

'But beyond that, the superior qualities of the Bunsen flame do not go.''

If the statement that the Bunsen burner develops a much higher temperature than an illuminating flame, is meant to be accepted as an absolute one, it may be proved to all interested that such is not the case, by an experiment suggested by Dr. O'Conor Sloane, of

The Open-Hearth Process. 683

New York, in a paper read before the American Oas-light Associa- tion in Boston (Journal of Qnslighiy vol. xxxviii., p. 878) as follows:

"Take two burners (a Bunsen and a white-flame) adjusted to consume equal quantities of gas, and introduce them into two separate tubes surrounded by a water- jacket; the same amount of water at the same temperature to be placed in each.

Let each tube have sufficient convolutions to present a large conducting-surface. Allow the flames to burn the same length of time in each tube and take the tempera- ture of the water. It will be the same in each case, proving the beat developed in both burners to have been equal.''

If these experiments were tried under an open vessel, with Bunsen and white flame respectively, the results would be very different. The Bunsen flame would boil the water more quickly than the other, because in the white-flame burner a large part of the heat developed would never strike the bottom of the pan at all, but would be radiated off in all directions.

To use Dr. Sloane's wonis, " the difference between a luminous and a non-luminous flame is one of character only. If all the heat be utilized, one is as efficient as the other."

Professor Tyndall, in Heat as a Mode of Motion (fifth edition, Longman, 1875, page 49) says :

" The Bunsen flame is much hotter than an ordinary flame, because the combus- tion is much quicker, and therefore more intense, but it is not hotter, not nearly so hot, to a body exposed to its radiation.''

Tyndall gives the proportions of radiant heat as 30 per cent, in the illuminating flame, and 12 per cent, in the Bunsen flame.

Mr. Frederick Siemens has made the following statements (per- sonally to me and in print) on the general application of heat :

In all heating-operations it is the natural tendency to try to get the highest effect with the least consumption of fuel and labor. To attain the above perfect combustion is essential ; yet perfect combus- tion does not cover all practical requirements. There are heating- operations which require particular conditions. One must consider particularly whether a large quantity of heat is required, as when firing a boiler, or whether intensity is required ; or whether both intensity and quantity are wanted, as in a steel-melting furnace.

One should not attempt to burn gas according to a fixed scheme; each case ought to be specially examined and the design so made as to meet all practical requirements. To insure perfect combustion every flame should have ample room for free development.

Practical observations have shown that a flame must be so directed

684 The Open- Hearth Process.

in a furnace that it will not com in contact with the surface of a soiid such as the walls of the furnace, or the material which is being heater! therein.

When ample space is given to a flame for its free development, not only a higher effect is attained, but also a saving in refractory materials, time, and labor. Recent constructions have shown the practical value of these facts, especially in the designs of steel- and glass-melting furnaces with high roofs.

One of the advantages of gaseous over solid fuel is that the former permits a more accurate regulation of the air for combustion.

The Bunsen burner is only of advantage when small surfaces are to be heated or when the flame is to strike a body, as shown by the incandescent gas-light.

The Bunsen burner is especially adapted to certain small opera- tions in a laboratory, in the household, or in the kitchen, but not to heat a large space and extended surfaces, as in a large furnace. In such cases a large luminous body of flame is essential for the pro- duction of radiant heat.

The flame of the Bunsen burner is non-luminous, because there is no free carbon in it, consequently it radiates but little heat, and can therefore heat by direct contact only. Although the Bunsen burner seems so simple and advantageous from a theoretical standpoint (as it is, without doubt, for certain cases), it is useless for most technical purposes, especially metallurgical operations on a large scale.

When gas is consumed in a large furnace, the gas and air must be brought together in such a manner that a flame with a high radiat- ing power is produced ; for thus the heat is best transmitted to the sides and walls of the furnace and to the material to be heated.

Furnaces fired with solid fuel' produce a more or less radiatiug flame, according to the nature of the fuel. Practice has shown that only such coal is good for direct firing in a reverberatory furnace as gives a large amount of hydro-carbon gases, for these gases increase the radiating power of the flame. The very best coke is useless for the same purpose, although the highest temperature can be produced by it when the body to be heated is placed in direct contact with it, as in crucible-steel practice.

There are no special difficulties in the production of radiant heat in the combustion-chamber of a direct-fired furnace, provided the proper fuel be selected ; but the case is not quite so simple when gas is to be used in a regenerative furnace. It is of the greatest im- portance how the gas and air are brought together in a regenerative

The Open-Hearth Process. 686

gas-famaee, accordiog to the quality of the gas and the iise of the furnace. When the mixture takes place too quickly and completely, a very, short flame is produced, which is intensely hot, but only slightly luminous, and consequently possesses very little radiating capacity. When the reverse is the case — that is, when the mixture is very incomplete, good combustion cannot take place ; and, besides a loss of heat, a number of other evils will arise, which will be men- tioned below. For good practice, and to secure at the same time a flame with high intensity and a high capacity to radiate heat, a con- dition between these two extremes must be maintained.

How gas and air must be brought together to produce radiant heat, depends upon the quality of the gas, the desired temperature, the size of the combustion-chamber of the furnace, and the special operation to be performed. In general it is best to admit the gas through a horizontal slit in the end of the furnace, over which the air is admitted through a similar slit, although somewhat larger in section. As the gas has a considerably lower specific gravity than the air, it has a tendency to rise up through the air, whereas the air sinks through the gas, and thus a sufficiently rapid mixture of the two takes place. Good results from this arrangement depend upon the temperature of the two gases and the relation of their tempera- ture to each other, as the specific gravity is changed by heat.

The light-and-heat radiating capacity of a flame depends upon the presence of free carbon particles, produced therein by a dissociation of hydro-carbons. This free carbon, although infinitely minute, is presentasasolid particle in the flame, and is heated up to the flame- temperature. Every carbon- particle radiates heat and light to all sides, and continues to do so until it is burned to carbonic acid, or until the flame is cooled to below the temperature of incandescence. The free carbon in the flame is always the last which is consumed, so that in all cases of incomplete combustion it is this carbon which remains and deposits itself as Boot in the flues, or passes out of the chimney, with the products of combustion, as black smoke. These small bodies or particles, which give to a flame the capacity to radiate heat, and which give such a high efficiency when properly utilized, cause the disagreeable smoke in our cities when the furnaces are not properly constructed or managed.

Water-gas and a mixture of water-gas and producer-gas were tried for some time abroad. Inquiring into the economy of the practice. I was told that a much larger percentage of pig-iron could

686 The Open-Hearth Process.

be used in the furnace using water-gas, which was an advantage to the respective works ; but, as the basic pi*oce86 now permits as large a percentage of pig- iron to be used as one can desire, the water-gas practice was abandoned. It must be said, however, that a number of other reasons also contributed to this abandonment.

At a number of works abroad, the temperature of a furnace is roughly estimated by the time it takes to melt a standard rod of iron, the time being measured by counting the oscillations of a large pen- dulum. The estimation of temperature by the eye, however, as men- tioned by Mr. Campbell, is quite practicable.

As to sulphur, I find that the steel will absorb more from the gases when the whole charge is melted down at once, than when suc- cessive charges are made.

As to oxidation in pre-heating, I have observed that when the hot charges consist of pieces comparatively light in proportion to their surface, the oxidation may be higher than when the whole charge is melted down at once ; whereas, when pre-heated pieces are quite heavy, the waste by oxidation is naturally less.

Inquiring as to the amount of sulphur removed in the basic open- hearth plants abroad, especially on the continent of Europe, I got figures which show sometimes no desulphurization at all, and some- times up to 60 per cent, of the initial sulphur. The figures are so unreliable, that in order to be sure not to get the sulphur above specifications, it is best, practically, to figure ongettingall the initial sulphur charged into the finished product.

It is interesting, however, considering Mr. Campbell's valuable tables, as well as the articles written by Mr. Hilgenstock (in Stahl und Eiaen) on the desulphurization of pig-iron in the mixer, that at nearly all the works where a partial desulphurization is claimed, very manganiferous mixtures are charged.

In conclusion, I wish to say that I know of no process that offers such a scope for metallurgical work as the basic open-hearth. In spite of the rapid progress it has made, the process is still in its infancy in this country. Most of our northern ores give a pig-iron admirably adapted to it, and as soon as we can control the sulphur as well as we do the phosphorus, the number of furnaces will in- crease rapidly.

I am glad that Mr. Campbell has had the satisfaction of corrob- orating Dr. Wedding's statement as to the volatilization of phos- phorus from hot charges ; and I wish to congratulate him again on his valuable contribution to open-hearth literature.

The Open-Hearth Pbocebs. 687

H. D. HiBBARD, Highbridge, JI. J. : Mr. CampbelTs remarkable treatise (for it is more than a paper) is by far the most valuable work on the open-hearth process that has ever been written, and will doubt- less be a standard work on the subjectfor a long time. One finds in it but little debatable matter, as the statements made are as a rule supported by too much evidence to be successfully assailed. About all we can do is to emphasize such points as seem especially impor- tant, though often overlooked, and also perhaps add a little here and there.

The point raised in Sec. 8, that the draft through the gas- and air-chambers should be governed separately, is very important. The need for this was first felt by the writer about 12 years ago; and 3 years ago he intended to patent a furnace in which the chief novel feature was to be an independent damper between the gas-revers- ing valve and the chimney-flue, to regulate the division of the draft — getting so far as to sketch the furnace to scale, when a pub- lished notice of a similar device in use abroad, caused a suspension of his purpose.

In Sec. 14, the author possibly underrates the value of the tar in the gas. Whether or not it is decomposed in passing through hot renerators, is a matter of conjecture. It has been volatilized once, though probably at a lower heat. If chemically " cracked/' the soot formed would be largely carried along by the gas to the hearth, or it might be changed into CO by the action of CO, and O in the gas. In either case it would reach the hearth, where its full value as a fuel would be realized.

As to the placing of the producers near the furnace, in order to save as much as paible of the sensible heat of the gas, discussed in the same section, there is one way in wliich the heat may be utilized, and that is by decreasing the size of the gas- regenerator while that for the air is made larger. This will cause the air to enter the hearth at a higher temperature. The gas being hot requires less heating, and therefore by this plan more of the waste heat may be returned to the hearth, increasing the fuel-economy by that amount. Water-cooled valves, if desirable, which they probably are not, could be used without materially cooling the gas as it rushed through them.

In the estimation of the temperature of the bath (see Sec. 20), we see, on looking into the furnace, broadly speaking, four different degrees of temperature: First, the flame, which is the hottest; second, the walls of the furnace; third, the slag on the surface of

688 The Open-Hearth Process.

the bath ; fourth the slag in the wakes of the bubbles rising from the metal itself, which is the coldest, and which indicates the degree of heat in the bath. Observations of these four give comparative temperatures ; and one' can notice if they present the usual differ- ences or not, and so judge the temperature of the metal pretty closely. Under normal conditions, the metal is the coldest thing in the hearth of the furnace, except the bottom under it. At times it will be the hottest thing in the hearth, either accidentally (as the author states) or purposely, when its temperature may be too high for casting, and the hearth may be cooled down. When this is done by shutting off the gas, as is often done with a furnace using natural gas, the walls may get very much colder than the metal ; and the wakes of the bubbles will appear white on a red background when seen through melters' blue glasses. A pyrometer seems to have no field here.

Regarding the joint in basic furnaces, between the basic bottom and the acid walls (Sec. 36), it has been pretty well demonstrated that if the joint be vertical, and therefore free from pressure, it will stand well enough for commercially successful work. The slight fluxing which will take place may be easily repaired with either chromite or magnesite.

In connection with Sec. 42, discussing the relative c(mpositions of the slags in the acid and basic processes, it should be remembered that (1) in the acid process, when ore is used to decarburize, the silica of the slag is continually combining with the oxide of iron, thus removing it from the field as an oxidizing agent or carrier of oxygen, while (2) in the basic process the amount of free oxide of iron is not decreased by such action, and therefore remains to be a carrier of oxygen to the metal.

In the acid process, ore must be added to replace that taken up by the silica, or decarburization will gradually cease. In the basic process, ore continues its work with scarcely abated vigor to the end, and much less ore will therefore do the work. Many a wild," un- manageable, over-oxidized heat of steel has been made in the basic furnace because ore was added as freely as in a similar acid furnace.

It may be said in connection with dephosphorization (Sec. 45) that a pig-iron free from sand would make a much more desirable material for use in the basic furnace than that usually obtainable for the purpose. Such a pig would favor the production of low-sulphur steel, as higher combined silicon would be permissible in the iron, which generally occurs with lower sulphur.

In Sec. 60 the author dissects a modest paper contributed by me

The Open-Hearth Process. 689

to the Iron Age. He objects to ray introduction of the kind of hearth into my list of variables, considering it unimportant. To this I would reply by asking if members would think the list complete were the kind of bottom left out of consideration. One can per- fectly agree that " the highest function of the hearth is to remain passive ;" but how often it comes down (or rather up) from its posi- tion of highest usefulness and, with a persistence worthy of a better cause, gets into the slag, where its absence would be much more desirable than its presence, and where it plays a part in the metal- lurgical operation !

Neutral bottoms include such as beauxite and ehromite, which are not very siliceous and yet which do not, by the slags they yield, act on the charge as do basic" bottoms, using the term in its usual acceptation.

In the remark, High temperature retanls boiling while low favors it," the point in my mind was the boiling due to the occlusion of gases. At higher temperatures the boiling is less, on account of the increased solvent power of the metal for gases, the escape of which produces the phenomena of boiling.

Regarding loss in melting (Sec. 63), I have found it to be pretty nearly a function of the length of time taken in melting the charge. The quicker the charge is-melted, the less the loss. The flame is essentially oxidizing, and that which melts the fastest seems to have as little oxidizing effect as any other. This loss in melting is, of course, subject to modification afterwards, during decarburization, by the action of the reducing elements; but in many, if not most, furnaces, with not over one-third of pig-iron in the charge, there is little or no reduction of metallic iron from the slag.

The furnaces from which Mr. Campbell has obtained mast of his data are very quick melters and present unusually mild oxidizing conditions.

Alphojsse Hennik, Springfield, HI. : In Section 13 of this val- uable paper, Mr. Campbell says of CjH that, "at high tempera- ture it readily resolves itself into C and CH, its elemental compo- nents. CH4 is likewise dissociated by heat, but the action is slow." These statements are open to question. According to Professor Ber- thelot, C2H4 (ethylene), as well as CH (methane), belongs to the four fundamental hydro-carbons derived directly from the distillation of the coal. By action of heat, they are, in common with CjH, and CHj, submitted to a well-defined cycle of reaction, in which the liberation of carbon does not take place.

Methane generates, by pyrogenous action methyl (CH3), ethylencjp

690 The Open-Hearth Pbocess.

(CjH), and acetylene (CgHj) with liberation of one, two, or three equivalents of hydrogen. Reciprocally, each one of these three car- bides, in the presence of hydrogen, reproduces all the others. By prolonged action of heat, CH4 (methane) will generate CjH (ethane), CjHg (propane), (butane), and the other liquid and solid par- affins. C2H4 (ethylene) will produce, by polymerization, CjHg (pro- pylane), C4H8 (butylane), and the other olefines. CjHj likewise gen- orates the vapors of the aromatic series, diacetylene, benzine, styrolene, naphthaline; and each one of these new products, being submitted to the same reactions as the original compounds, may lose part of their hydrogen. In all these transitions, the hydrogen is the active agent, uniting with hydrides already formed, or separating from them, while the carbon remains constantly combined.

Purely theoretical as these remarks may appear, they have, how- ever, a practical side which deserves attention. In a methodical, well-conducted gas-process, combined with condensing-and-scrubbing operations, dry, highly hydrogenated gas contains all the above-cited hydro-carbons, which appear only partially in ordinary analyses. CjHg, CjHg, CJJiQy which are higher members of the methane series, always present in the gas, are usually given as methane simply; and since ethane gives double, propane three times, and butane four times its own volume of GO3 during its combustion with oxygen, it is evi- dent that taking the volume of COg as representing the CH4 is mis- leading. CsHg and C4H8 are likewise given under the heading of ethylene. CH3, CgHj, and the aromatic vapors are entirely neg- lected, although the percentage of C2H2 is fully as high as that of ethylene, and may be higher by reason of its greater stability, while the aromatic vapors are, under good practice, three to four times higher.

These exceedingly mobile vapors are so thoroughly diffused during their formation in the permanent gases, that they resist condensation under very low temperature. Their stability at high heat is also very great.

The importance of the aromatic vapors in the gas will be l)etter understood when it is considered that they constitute the main luminous element of the coal-gas. In fact, a trace of benzine in heating-gas is sufficient to change the color of a large body of flame from pale blue to white. These vapors increase considerably the calorific value of the gas, and add much to the radiating power of the flame.

When unmixed with steam or air, highly hydrogenated gas, com- pletely scrubbed, will emerge on the top of the checkers, practi-

The Open-Hearth Process. 691

cally intact, and develop more heating-power in the furnaces than it is credited with.

Dr. H. Wedding, Berlin, Grermany : I beg to oflTer a few words concerning what the author- of this very interesting and valuable paper on open-hearth furnace practice has said about the nomencla- ture of iron.

We Germans, of course, have not the least objection that you Americans should call sled what we call Flusseisen, though I fear that in some years you will have lost the word "iron " altogether, because as soon as the puddling process has gone no metal of this name will exist any more for you.

We are, I think, happier; we have the name Flusseisen for all sorts of the metal iron, which, being malleable, are produced in the fluid state; and we call steel {Stahl) only those sorts of Flusseisen or Schweisseisen which practically can be hardened.

Besides this, I am surprised that Mr. Campbell, in his otherwise exhaustive paper, did not mention at all the carbonization of "steel (in your sense). We use carbonization now with good success. It will, perhaps, be remembered, that, three years ago, Mr. Thielen, of the Phoenix steel-works, explained* at the meeting in Pittsburgh, the carbonization by the Darby process, as carried on at his works. Since that time, the process has made great progress, and is now used daily in regular work. A very valuable improvement has been made by the works at Dudelingen (Luxemburg), where coal-dust, lime- powder (i.e., dry powder of hydrated oxide of calcium), and as little water as possible, are moulded in blocks and put into the casting- ladle before the steel is poured into it. The result is astonishing. The steel is nearly free from blow-holes, and has, practically, the exact content of carbon desired and cal<;nlated beforehand. The reason for this result may be found in the effect of the steam liber- ated from the hydrate of lime and working like a paddle.

The process, it is true, is used with us only for basic steel, but I cannot see any reason why it should not work as well with acid-steel methods.

H. M. Howe, Boston, Mass. : I should like to ask Mr. Campbell whether, in using his tilting furnace, he finds it expedient to pour the slag off during the heat.

One would expect a certain economy of fuel from removing the slag. Of the heat developed in the furnace, part escapes through

Tran8.y xix., 790.

692 The Open-Hearth Process.

the walls of the furnace and of the regenerators part goes up the chimney, and part passes through the slag into the metal, to raise the temperature in the latter, and to compensate for the loss of heat from the metal into and through the bottom of the furnace.

Let us take two cases in which the quantity of heat which it is necessary to pass into the metal for this purpose is constant, but in one of which we have a thick layer of slag, and in the other a thin layer of slag. Slag being a very bad conductor of heat, a thick layer of slag should conduct into the metal a very much smaller proportion of the total heat generated in the furnace than a thin layer of slag would. And in order to pass an equal quantity of heat into the metal, a greater quantity of heat would have to be devel- oped in the furnace in case of thick slag than in the case of thin slag.

The practical inconvenience of pouring off the slag has in general prevented furnace-men from doing it. Now that we have this new form, this tilting furnace, from which the slag can be readily poured by slightly inclining the furnace, the inconvenience of removing the slag is greatly diminished, and one naturally asks whether this prac- tice, so often attempted only to be abandoned* will now be found desirable.

Apropos of measuring differences of temperature with the eye, the difference between the melting point of hard and that of soft eteel seems to the eye an enormous one. Yet late observations indi- cate that there is only a difference of about 20° C. between the melting point of steel with .10 carbon and that of steel with .30 car- bon, and of about 45 between the melting point of the latter and that of steel with .90 carbon. In my own experience, while I have not watched carefully for this matter, I have sometimes been impressed with what seemed a very striking difference in temperature between adjacent bodies, though the difference certainly must have been considerably less than 10® C. In the recalescence of steel, the rise of tem|>erature is very readily detected even by an unpracticed eye ; yet in the cases which have come under my observation it does not exceed some 10° C.

Mr. Campbell : The experiments cited by Mr. Groetz seem to prove conclusively that a gaseous flame must be luminous to produce high temperatures economically. It remains to Gnd the ratio exist- ing between heat-utilization and degree of luminosity. On the one hand, it may be possible that a flame is valuable exactly in propor- tion to its luminosity ; on the other hand, it may be that only a cer- tain definite brilliancy is required, more than this being useless. The

The Open-Hearth Process. 693

determination of the law might lead to radical changes in furnace- practice by the addition to the fuel of light-giving factors, and the in- troduction of a system of photometric measurements in conjunction with the ordinary methods of gas-analysis. It must be considered, however, that although the experiments mentioned introduce a new factor into the thermal equations as written in Chapter III. of my paper, and seriously disturb the argument at some points, the fact must not be overlooked that an increase in the light-giving components of the gas does not always mean a corresponding or even an absolute in- crease in the luminous quality of the flame. On the contrary, it may even produce a diminution ; for the constituents which confer light- giving power are usually dense hydro-carbons, and a flame can util- ize only a certain proportion of these components — this proportion depending upon the temperature and other conditions. An increase in these components beyond the critical point causes a part to pass unburued through the furnace, giving a smoky flame. It is need- less to say that a flame which is extremely smoky is not at its maxi- mum luminosity ; but that, under such circumstances, the incre- ment actually decreases the luminosity, and thereby the calorific value. Thus, although it seems probable that the complete wash- ing of a gas is inadvisable, it is not proved that a gas, as it leaves the producer, is much better than the same gas which has been al- lowed to deposit some of its tar and most of its soot. The fact still stands, as related in Sec. 13, that a furnace did admirable work when the gas at the valves was cooled to the temperature of the atmosphere, and that under these conditions the flame was all that could be desired.

The plan of making peep-holes in the walls of the regenerators, to find out whether soot is deposited, does not seem to be very sat- isfactory. It is evident that, with ordinary producer-gas, any de- posit must be washed out when the currents are changed ; for if there is not as much thus washed out as there is deposited, there will be an increment on each reversal, and this must soon produce a com- plete stoppage of the passage. This constriction does not happen with some gases; but when it does, no peep-holes are necessary to tell the story. On the other hand, when the washing-out equals the deposition, it is doubtful if the process could be observed, as the amount would be too small to be noticeable.

The criticism on the absence from my paper of a drawing of a properly constructed furnace of the common type, is in some measure deserved. The omission was due to the desire to condense as much

694 The Open-Hearth Process.

as possible in matters of detail. It is possible to build a furaaoe in various ways and have it meet the requirements, and it was con- sidered sufficient to point out a common fundamental error. For this purpose there was selected a drawing which had been published within a year or two in one of the leading trade-papers as the new design of a firm of metallurgical engineers. The drawing of the tilting furnace was given, not as the only properly arranged furnace, but as embodying a feature which is new and valuable. It is else- where stated that for some reasons vertical chambers are preferable to horizontal ones. The arrangement of the parts to the best advan- tage will always remain a detail of construction, determined by local conditions and inventive ingenuity. The opinions expressed by Mr. Goetz on the duplex process are corroborative of my own, and are of interest in these days, when this method is so often mentioned as applicable to southern districts. As an instance, however, of the care with which general conclusions should be formulated, I woald say that I have performed the same experiment of pouring fully- blown Bessemer metal into an open-hearth bath of about 0.50 per cent, carbon, and there was no violent ebullition. These contradic- tory results will show that there are many factors in metallurgical history which are imperfectly understood, and a variation in which, although perhaps unnoticed, may make important changes in the na- ture of the result.

The controversy with Mr. Hibbard regarding the influence of the bottom is merely a question of words. In common with all metal- lurgists, he knows that the bottom itself exerts little influence on the metal ; but he speaks, for instance, of a basic hearth, in the brood and usual meaning of the phrase, as being the one condition which determines the procedure known as the basic practice. In that sense, by common acceptation rather than in strict scientific pro- priety, his tabulation is correct

In the Iron Age of August, 31, 1893, an article appeared, com- menting on the section of my paper relating to estimation of tem- perature. In the following issue, I made answer that the critic had somewhat misunderstood my position. Lest others also be misled, it may be well to say that many circumstances prevent the attain- ment of perfect accuracy in scrapping' Bessemer heats. In using direct metal, the variations are especially hard to forecast, even with the wonderful skill which may be acquired in the estimaticHi of analysis by fracture. In consequence of such irregularities, some heats are made that are too hot and some that are too cold. 'When

The Open-Hearth Process. 695

the iron is carefully mixed and melted one great cause of irrularity 18 lessened. Under these conditions, when thirty or forty heats are made without the delay of a minute or the changing of a bottom, as often happens in American practice, it will be found that the scrap will be very regular throughout the series, running perhaps from four hundred to seven hundred pounds per heat. In such work the practiced eye can detect the difference in temperature produced by one hundred pounds of steel scrap. This difference I have tried to calculate in degrees of temperature, and Mr. Howe has corroborated the results by some interesting facts. He also asks for some infor- mation on the loss of heat through the presence of slag, which can not be readily given. His argument is perfectly correct in portray- ing the waste of heat caused by a thick blanket of cinder between the flame and the metal; and there is not the slightest doubt that the removal of the slag will allow a much better realization of the calorific force. It would, however, be very difficult to tabulate results so as to show this effect ; for there would be many coincident and correlated phenomena tending toward the same end. Thus, the removal of the slag would greatly hasten the work of oxidation would reduce the amount of lime to be added and would facilitate the removal of phosphorus, owing to the ability to make a new slag rich in lime. These factors tend to a shortening of the operation and therefore to an economy of fuel. This would make it difficult to isolate the gain through better transfer of heat from the flame to the stock.

The remarks of Mr. Hennin might serve as a text for a discourse on organic chemistry ; but the bearing of the involved points upon the question will not justify such digression by one who does not pretend to authority in that domain of science. It must be acknowl- edged that the phrase "elemental components " is open to censure. C,H is not created by the conjunction of CH and C; for CH and CjH belong to two distinct and separate series. But CjH does entirely break up under the action of heat, as was shown in a paper which I presented to the Institute in September, 1890 {Trans., xix., p. 158).

It is true that ordinarily there is no accumulation of carbon from the passage of Siemens gas through regenerators ; but the propor- tion of C2H4 (usually about 0.4 of 1 per cent.) is so small that this action could not be expected. There is abundant evidence, however, to show that there is a heavy deposition of soot when natural-gas (CHJ is regenerated, and also that, under the same conditions, petroleum vapors give large volumes of free carbon. 1

696 The Bertha Zinc-Mines.

These facts seem to disprove the statement of Mr. Hennin that the breaking up of these hydro-carbons produces hydrogen rather than carbon ; and they may partially justify, although they do not prove, the assumption that the CjH is resolved into C and CH. The gas from a Siemens producer may be looked upon as illuminat- ing-gas, diluted by N, COj and CO. It will be found that the analyses of ordinary illuminating-gas show either very small per- centages of the hydro-carbons higher than CH or none at all. It is therefore reasonable to suppose that they are not present in Siemens gas in large amounts. This being so, it will not be necessary to take them into account when, in the analysis, the meth(x] is used of calculating the percentage of CH4 from the volume of CO, produced by combustion with oxygen. Such a correction seems to be a refine- ment beyond the scope of my paper. It is certainly beyond any- thing I had hoped to attain.

The remarks of Dr. Wedding on the nomenclature of iron refer to some entirely extemporaneous remarks of mine made at the Chicago meeting, and omitted from publication. With all due def- erence and respect for the group of metallurgists who formulated the system used in Grermany and recommended it for universal adop- tion, I have always considered this nomenclature as founded in error and impossible to practice.

The absence of any consideration on my part of methods of car burization arises from the fact that in our country the basic open- hearth produces almost exclusively low steels, below 0.20 per cent, carbon ; and it has not yet become a necessity to use the methods which have been so successfully developed by the metallurgists of other nations.

The Bertha Zinc-Mtne8, At Bertha, Va.

Discussion of the Paper of Mr. Case. (See p. 511.)

E. G. Spilsbury, Trenton, N. J. : Mr. Case's paper possesses much interest for me, because my attention was called many ago (early in the seventies) to the deposit he describes, and his ex- plorations seem to have confirmed the predictions I then made as to the form of the deposit, namely, that the ore would be found to occur, not, as was then imagined, in loose boulders, distributed

The Bertha Zinc-Mines. 697

through the clays, but, rather, in contact with the limestone. Some years ago I had charge of a zinc-mine at Landisville, Lancaster county, Pa., where the same thing was observed. The silicate- and carbonate-ores overlay the irregular points and domes of the lime- stone, just as they do at the Bertha, though the Pennsylvania deposit was by no means so extensive that of the Bertha mine. The un- derlying limestone was impregnated with blende, as it is found to be in Virginia. It seemed to be not a specially-defined vein, but one of the layers, or, probably, one of the folds of the limestone, which contained di&seminated zincblende crystals throughout. At Landis- ville we used to crush this material and treat it as is done at Bertha. Unfortunately, the deposit gradually became poorer as the mine was extended, until it would no longer pay the expense of operation.

Mr. Case's paper is instructive for all of us with regard to the method of attacking these deposits. In Pennsylvania the mines of the Bethlehem Zinc Co., in the Saucon valley, the zinc-mines near Tyrone, and the Lancaster county mines just mentioned were all worked on deposits near the surface and by the method of stripping. As is well known, the expense of this method, and of the heavy pumping of surface-water which it ultimately involved, became so large in the great Bethlehem mines as to force their abandonment. Mr. Case has not given with as much detail as I would have liked to have, the troubles encountered in the practice of that system at the Bertha mine. For many years that mine was operated by open-cut, and a largo amount of clay was removed, as will be seen on the right hand in Fig. 5 of his paper. But the expense became so great that the mine was practically about to be abandoned some five years ago, when Mr. Case took charge and devised the present method of shaft- sinking and underground exploitation, for which we are all indebted to him.

C. Le Neve Foster, Llandudno, Wales: From a geological point of view these deposits seem particularly interesting. The sec- tion given in Fig. 7 brings forcibly to my mind the phosphate-de- posits in the north of France. That figure would very well repre- sent the French deposits, if one only imagined chalk to be substituted for the limestone and phosphate for the zinc-ore. ' Fig. 7a closely resembles the appearance of the phosphate- mines when the clay has been stripjicd away and the phosphate worked out. I would like to ask how the miners locate their shafts from the surface, so as to come down into the middle or deepest point of the ore-bearing de- pressions. And I wish especially to thank the author for the small

698 The Bertha Zinc-Mines.

map included in Fig. 1, hich shows the geographical position of the Bertha minei That is a feature of particular usefulness to those of us who live on the other side of the water and arc not minately familiar with American localities; and I think it is a precedent which might be followed with advantage.

Mr. Case: The shafts are located by surveys, based upon ex- ploring-drifts made underground in advance.

W. P. Blake, Shullsburg, Wis. : The descriptions and the sec- tional view in Mr. Case's very interesting paper, showing the rela- tion of the zinc-ore deposits to the limestone, remind me of the section given by Sir Roderick Murchison, in his classic work upon Russia in Europe and the Urals,* of certain gold placer-deposits which have accumulated in deep pockets in the limestone. Similar deep pockets filled with gold gravel are found near Murphy's, in Calaveras county, Cal.,t where the limestone is so deeply corroded as to present an impassable mass of pinnacles and deep irrular cavities, all caused by the corroding effects of solutions from the formerly overlying alluvium.

The explanation of the formation of the carbonate zinc-ores in the cavities and chimneys of the cornnled limestone at Bertha is, doubtless, correct. The source of the solutions must be the pre- existing blende, which, by the rise of the strata, must have been at a higher level, and, decomposing, has formed sulphate solutions, which, percolating downward, have corroded the limestone, deposit- ing their zinc as carbonate, while the lime-sulphate has flowed away. The discovery of a blende-bearing stratum of limestone on the dip beyond confirms this theory. The blende was probably formed in the limestone contemporaneously with the rock.

It is interesting to note the very general distribution of zinc-ores in the form of blende in the older limestones of the United States and Europe. In the Appalachians the deposits appear to be most numerous and important in close connection with limestones of the age of the Trenton. This is likewise the horizon of the zinc-ores of Wisconsin.

Russia in Europe wnd the Ural MountavM i., 487. t Blake, Oeological Reconnaisance of Ckdifomia, p. 255.

Obsebvation8 Cx)Ng£Bning Ore-Drb98Ino. 699

Qenbbal And Special 0B8Bbvati0N8 Conoebnino 0Bedbe88In0.

Discussion of the Paper of Oberbergrath Bilharz. (See p. 225.)

T. A. RiCKARD, Denver, Colorado: The observations made by the author concerning the treatment of gold-bearing ores, deal with the subject only briefly and in a general way; but, coming from an authority, they invite comment.

In stating the fact that a finely-divided condition " is a necessary characteristic of ores containing free gold, the author must be refer- ring to those mining districts only with which he himself is ac- quainted. The statement would not be true of certain gold-mining rions in America and Australia. The ores of some Californian mines and the mill-stuff treated at certain of the mills of Clunes and Ballarat in Australia, and of Otago in New Zealand, contain the gold in a free but in a coarse state. Moreover,/' reduction to fine sand and flour" is not required by certain ores, even though the gold be in a finely-divided condition. In such cases the gold becomes de- tached from the quartz before the gangue has been pulverized into particles as small as those in which the gold occurs.

The dry-crushing of free-milling gold-ores is unknown in the new mining regions of the world. It is not clear whether the author has in mind'any European district in which such a process is in actual use. If he has, a description of the method and the results obtained would be very valuable. Dry-crushing is, other things being equal, twice as slow and twice as expensive as wet-crushing. In the case of free-milling gold-quartz it serves no purpose, since such ore does not carry sulphides which contain any large percentage of value, and which, by being slimed, may carry away part of that value. The dry-crushing of a free-milling gold-quartz would be considered an absurdity by an American metallurgist.

If the ore contain **free gold only," why submit it to a further treatment, such as concentration, after its passage over the amalga- mating-plate ? What does the author propose to concentrate ?

Again, why anticipate the treatment of tailings by the McArthur- Forrest, or some other process? The satisfaction felt by a mill-man in saving gold by a subsequent treatment of the tailings from his mill appears to me to be like that of the man who shoots at a rabbit.

700 Observations Concerning Ore-Dressing.

and having missed with the first barrel, gets it with the second. The one is no proof of good marksmanship, neither is the other of good milling.

Trof. R. H. Richards, Boston, Mass. : The first half of Ober- bergrath Bilharz's paper is devoted to the ten rules of Rittinger and to comments upon them. Of these I think v:e shall all approve, provided they are interpreted as was intended. To give an example of the way a different interpretation may be used, I will, for instance, refer to Rule 8 : Avoid putting too many machines on the motor." I have known of one engine driving 640 plunger-jigs, 64 circular sh'me-tables, 16 steam-stamp valves, 32 centrifugal pumps, and giving perfect satisfaction. I have also known of one engine driving only half of the above machines ; but, in addition, it was called upon occasionally to help lift heavy loads up an inclined hoist. As a re- sult, every jig in the establishment was slowed down and the work was very unsatisfactory. Again, I have known 60 stamps, driven by a separate belt to each battery of 6 stamps, the whole driven by one engine and giving perfect satisfaction. I have also known 60 stamps, mounted all on one shaft; and the mil I- manager's life was not worth living.

It is a simple matter to recognize the intent of the rules and com- ments, and having done so, to agree with them.

There are two rules, however, which I do not find in the list, and which I consider to be of importance and of almost universal appli- cation. These I will denominate for the time being Rule 11 and Rule 12.

Rule 1 1. Once middlings always middlings. By this I mean, that if a machine is working properly, a particle which goes once into the middlings will go again into the middlings if it is fed back over the same machine. It should not, therefore, be so fed back, but should be sent off to a special treatment of its own. This is espe- cially true of circular slime-tables, where, if 10 per cent, of the total feed comes out middlings, and is sent back over again, the next cycle will have 20 per cent, of stuff asking to be let into the middlings. Since, however, the attendant cannot provide for so much middlings, he puts on more water and sends nearly all of the added 10 |)er cent into the tailings, with no appreciable increase of value in the heads. The same is true of jigging, where the eventual disposition of every particle of middlings will be in the tails. The knowing ones some- times speak of this as the easiest way of getting the middlings out of sight.

Observations Cx)Ncerning Ore-Dressing. 701

Rule 12. Every machine as far as is practicable, should have its guard. I will instance first a series of six jigs in one of the great Lake Superior copper-mills. The discharge and hutch of pure cop- per of the first jig is guarded by keeping a deep layer of copper on the jig. The tailings on the other hand, of the second and third sieves are guarded by the discharges which take off included grains and allow fine copper to go through the bed. As a result, the tails are so low that they go to the lake. The first finishing-jig turns out clean hutch-work and clean automatic discharge. They are both guarded by a thick bed of copper. The*second finishing-jig guards its pure copper-hutch by a thick bed. The third finishing-jig, on the other hand, guards its tails by a thin l)ed and by running a little sand into its hutch. Such a series of guards as these insures good work and prevention from loss, even though inspection may be for a little time interrupted.

As a second instance, I will refer to a gold-mill. The usual gold- mill is run with vanning machines without any guard, on the prin- ciple that the vanner is a perfect concentrator. Now I have never vanned behind a vanner without finding a little fine ore in the tails. There is no reason why there should not be a little of this. There is no such thing as an absolute all-around concentrator. In my opinion, therefore, there should be a circular slime-table, which is safe to catch the fine ore in the presence of coarser quartz, as a guard below the vanners. One slime-table might perhaps take the tails of three vanners. The slime-table is a very inexpensive machine, and it can stand as a watch-dog during the nights and rainy days, and at other times when inspection is uncertain or even ab- sent. The experiment of trying such a machine will not cost much, and the yield of it may be considerable.

The last half of this paper deals with the classification of mills. In this regard I quite agree with Bilharz in his comment upon Rittinger's classification ; namely : that it is " neither complete nor practical.'' With so clear aqd comprehensive a conception of his subject as Rittinger had, it is not easy to understand how he should have been satisfied with it.

In regard to Bilharz's classification, perhaps it is as satisfactory as can be made. Even in this improved schedule, however, we find mills in the second great division that are almost identical with those in the first.

For the benefit of other persons, constituted like myself, who in this particular case desire a classification of mills for the single

702 Observations Concerning Ore-Dressing.

purpose of placing in juxtaposition the similarities and dissimilari' ties of the different processes, to provoke contemplation and sugges- tion, and who do not feel the need of the drag-net process of the yes-or-no principle of classification in this case, owing to the limited number of kinds of mills to be considered, I will give here a mode of tabulation which for my own purposes is more satisfactory than either of the above methods.

All ore-dressing operations begin with some mode of disintegra- tion or crushing, through the application of force. This may be done in one or in several stages* commonly in two great steps, which we can call " preliminary and " final " crushing. Disintegration is followed by concentration, which may be in one step or in more, but is commonly in two steps, which we can call " preliminary" and '' final '' washing. It happens, also, that certain middlings and by-products need further treatment ; and we can designate for these the crushing as " auxiliary " crushing, and the washing as "auxil- iary " washing.

If we arrange in vertical columns under the above headings the different mills or processes of metal- or ore-concentration, we shall have a simple means of comparing them.

The table on the following page contains a few of the principal ore-dressing combinations so tabulated.

While this paper of Oberbergrath Bilharz covers the problem of ore-dressing in his own country, it does not bring out some points which the American mill-man would like to have discussed.

Crushing by stages and washing by stages are becoming vitally important in this country. I will refer to two districts only, Mon- tana and Lake Superior. Assuming for Montana that the sieve scale falling off by J diameters is correct for a series of revolving screens preparatory to Hartzjigs, we have:

1 inch, J inch, J inch, inch, inch ; or

Now the first thing we must decide ia on which sieve the hydraulic separator shall begin. The size usually adopted is through inch.

Let us assume that the mill will be provided with three sets of rolls: No. 1 rolls. No. 2 rolls, and auxiliary rolls. There maybe also:

a. A preliminary jig, jigging on one sieve the coarse material and making only headings and tailings, sending all but its pure concen- trates directly to the No. 1 rolls.

Observations Concebnino Ore-Dressing.

I

ao'aS

Oh

a

e: o

a

p

O 9

1'

as

S

ta

B&

P

S

M

I?

IS to

is

si

B

Si

fSii

Is

e s

Sd

cd

£

Is

s&

3d

b-=

e3

§9-

n

M

§

a

r

s.

I !

I '

a

u

ftl

M

d

Os

a:

g

t

ua

;:4

;M

Pj

fii

e

s.

t

s

pq

n

g..2

ce 0

ir "1

Ei8

£5

0

Ib

¥

704 Observations Ooxcerning Ore-Dressing.

b. A set of jigs with several sieves, making only headings and tailings, the whole of the tailings going to the No. 2 rolls.

c. A set of jigs with several sieves which make only headings and tailings, sending the whole of the tailings to the auxiliary rolls.

d. A set of jigs with several sieves which make headings, mid- dlings and tailings, and send tailings. to the waste, and middlings to the auxiliary rolls.

A set of jigs which make headings, middlings and tailings, treating the stuff from the auxiliary rolls. The middlings here have to steer between two faults, too much richness in the tails on the one hand, and too much silica in the middlings for the smelter on the other.

The proper sizes to feed to the a, 6, o, d, e jigs have to be deter- mined with care and thought; the ultimate object, &s Bilharz well says, being to keep the ore as coarse as possible and avoid making unnecessary slimes.

If there is one of the above propositions more important than another, it is the feed to the d jigs. This is the first place where final waste is made, and it is therefore the limiting size of the whole mill treatment.

The main problem of Lake Superior, as I see it, is two-fold : 1. On what principle shall hand-picking be done? 2. Shall in- cluded grains be recrushed ? How shall they be caught, and how crushe<1 ?

As to hand-sorting, since every particle of silica which is in the rock must go to waste either in the mill or in the smelter, it be- comes a question of cost, where we shall send it. Let us suppose, for example, that smelting costs twenty times as much as milling (this is not an exact figure) ) then there is twenty times the argu- ment for getting rid of the quartz in the mill, provided the smelter receives enough silica for its slag.

Again, suppose for the sake of argument, that the mill-tails assay 0.7 and that the smelter's final slag assays 0.6 per cent, of copper, but on subtracting the lime, coke, ash, etc., we find that it really contains 1 per cent, per unit of rock sent into it, then we again have the ratio of 10 to 7 in favor of sending the quartz to waste in the mill rather than in the furnace.

If we look at the mine-rock of one of the conglomerate-mines, we shall clearly find in it four classes :

A. The " nigger-heads " and larger masses of solid copper, more or less clean.

Observations Concerning Ore-Dressing. 705

B. Small jagged masses, attached to clean rock.

C. Rock rich iu minutely disseminated metallic copper.

D. The remainder will be good mill-rock.

As to A, the rich lump, and D, the run of mill-rock, the ques- tion of hand-sorting is easy ; but B and C need special discussion.

B. A jagged fragment, if fed to the stamps, brings in waste rock with it that is below the assay of the tails. It then lies in the bottom of the stamp-mill where it is worn down by attrition until it is small enough to go through a f-inch hole. By this means the barren attached rock will be brought up in assay-value, and, on ac- count of the difficulty of saving slime-copper, may even go to waste richer than the average tails. Hence, B should not go directly through the mill. Should it go direct to the furnace? We have our 20 : 1 ratio of cost against that course. Such rock needs its . own machine, which stamps, breaks, thrashes, pounds or pummels it just long enough to lil)erate the most of the jagged fragments and lets the rest go to the mill.

C. These pieces, with intimately mixed copper, may make tails as high as 5 to 15 per cent, of copper without showing a sign of that metal to the eye. Nothwithstanding the higher cost of smelting, it may be clearly advisable to send this material direct to the furnace, as the only way of avoiding a ruinous waste of metal.*

J. C. I'Anson, Saltburn-by-the-Sea, Yorkshire, England : My remarks will be confined to the mechanical treatment and washing of coal and bituminous substances. In Great Britain and on the Continent of Europe, the better seams of coal are becoming in many districts more or less exhausted, and are quite inadequate to supply the increasing demand, especially of the iron and steel manufacturer. This condition, which forced the Grerman collieries, some years ago, to adopt an improved system for the cleansing of fuel, is now rapidly producing a similar result in Great Britain, and may be expected to be felt, though less immediately and universally, in the United States also.

The Liihrig system, invented by Mr. C. Luhrig, of Dresden, is the one which predominates in many parts of Germany, and has obtained a favorable footing in England, Scotland, and Wales, where eight or nine plants have been erected already. In Silesia, Saxony, Rhenish Prussia, Westphalia, Hanover, Baden, Austria-

Secretary's Note. — Further remarks upon ore-dressing, by Prof. Richards and others, will be found in the discussion of the paper of Mr. James Douglas, presented at the same meeting. VOL. XXII. — 45

706 Observations C02Toerning Ore-Dressing.

Hungary, Belgium, France, and Russian Poland there are at present more than 200 Liihrig plants; and while I do not assert that this system, as perfected by Mr. Luhrig's experience of more than twenty- five years in dealing with the most impure varieties of European coal, is the only satisfactory one in use, it can be fairly pronounced to fulfil in a high degree, at an extremely low cost, the conditions prescribed by Oberbergrath Bilharz, of maximum simplicity and rapidity of treatment and maximum facility for disposing of both the valuable product and the refuse. The process is automatic and continuous ; the quantity of water required is small and can be used over and over again indefinitely; the costs of operation and repairs are very low; settling-ponds are dispensed with ; and by-products, such as pyrites or brasses,*' can be saved as assets of value. Finally, . the " sludge" or finest coal is recovered by a new apparatus, which constitutes a special feature of the Liihrig system. It consists in a long tank of brick-work, running underneath the building, on the bottom of which the sludge is constantly settling, and from which it is constantly removed with broad scrapers, carried by endless chains. These deliver it into a hopper at the end of the tank, whence it is elevated into storage-hoppers, to be subsequently either mixed with fine coal from the washers, or used separately for firing boilers, making briquettes, etc.

A complete Liihrig plant comprises not only the washing of the coal, but its automatic and continuous treatment from the pit's mouth to the railway-car. In many cases, however, only a portion of the system, namely, the washing-plant, is required ; and this can be installed and operated without interfering with apparatus already employed for dry treatment of coal.

I will briefly describe, as an example of complete works, the new plant of Messrs. Miny & Cunninghame, at Motherwell, near Glas- gow, which is now in full operation, treating about 1500 tons of material a day — the output of three pits — which is brought to one spot by endless-rope haulage.

The contents of the tubs are discharged over vibrating screens, pierced with 2-inch round holes. All coal above this size passes from the screens on to long picking-belts formed of links (made of steel bars) with intervening spaces, through which small coal, due to breakage, falls upon a table below the return-belt, where it is col- ' ted by scrapers and delivered into the hopper, which receives the 1 paasing through the screens, he large coal on the belt is picked by hand. Any pieces of

Observations Concerning Ore-Dressing. 707

slate, etc., containing coal are thrown into shoots which carry them to a 8tone*breaker whence they are delivered to the main hopper under the screens. The refuse from the large coal is loaded into tubs and carried to the waste-heap.

The hopper under the screens has a capacity of about 100 Ions. The material received in it is elevated to the top of the washing- building, and delivered down a shoot into a large sizing-drum having several concentric shells of different meshes, which sort it into three sizes of nut and also pea and smaller sizes inch and downwards, as may be required). Each size of nut is conveyed down a separate shoot to its own washer, which consists of a wooden box divided into two compartments and having in the back-compartment a wooden plunger, actuated by an eccentric and adjusted in stroke to the size of the coal, while the front-compartment contains a sieve and two discharge-openings, one in front, near the top, through which the washed coal flows out with the water, and the other near the bottom of the space above the sieve, for the escape of the refuse. The latter is adjustable by means of a sliding-shutter.

The dirt and shale are carried by a spiral conveyor and an ele- vator to a pair of crushing-rolls, from which they again descend through a shoot to a suitable washer.

The washed nut-coals are delivered from the front of the washer over perforated drainage-shoots, to which a shaking motion is given, apd which convey them to the loading-hopj>ers, over the railway- track. While travelling down these shoots they are sprayed with a fine jet of water, which removes fine adhering particles and greatly improves the appearance of the coal.

The small coal inch and downwards in diameter), as it comes from the sizing-drums, meets the overflow water from the nut- washers and is carried into a grading-box, in which it is deposited in different compartments in successively decreasing sizes. From the bottoms of these compartments it passes to the small-coal wash- ers, which are the same in principle as the nut-washers, but have, in addition, a layer of pieces of feldspar resting on the sieve, through which the upward water-current is propelled by plungers of very short stroke. This layer slightly lifts and opens at each upward pulsation of the water, closing again with the downward current, and thus permitting the fine dirt to pass, while it prevents the fine coal from following.

The washed fine coal passes with the overflow-water down a chan- nel to a revolving drum of sheet-copper perforated with inch

708 Observations Ooxcernino Ore-Dressing.

holes, through which sludge, suspended in. the water, passes to the sludge-recovery tank beneath the building. The coal inch and upwards in size is carried by an elevator into storage-hoppers, to be subsequently sent to the coke ovens or otherwise disposed of. The elevator-buckets are perforated, so as to drain the coal before its delivery into the hoppers.

The interesting exhibit in the Mines and Mining Building at the Columbian Exposition, showing the conveyance of small coal, mixed in equal parts with water, by means of pipes and pumping-stations, to almost any distance at a small fraction of the cost of railway- carriage, suggests, in my judgment, a most useful adjunct to a coaU washing plant. By its use the owners of collieries, coke-ovens, and furnaces might, with great technical and commercial advantage, place the coke-ovens at the iron- or steel-works instead of at the collieries.

The results obtained at the Motherwell Luhrig plant, sketched above, have more than satisfied the guaranty given by the patentee, which was as follows:

CapacUy. — 1500 tons per day of ten hours for coal containing 23 per cent, of ash before washing.

Efficiency, — The washed coal between inch and inch in size was guaranteed not to carry more than 6 per cent, of ash, and the waste not to contain more than 2 per cent, of coal.

LabaF'Ck)8t, — This was guaranteed not to exceed 0.8 penny (1.6 cent) per ton of coal handled, including hand-picking, sorting, wash- ing, and loading into railway-cars.

The actual result of practice has been that the ash in the fine washed coal does not exceed per cent., the coal in the waste fidls below 1 per cent., and the labor-cost is 0.5d. or 1 cent per ton.

In one case in England the ash in coke has been reduced by this preliminary treatment of the coal from 16 to 4 percent, and, in an- other case, slack, which had been sold at 6d., is now sold, after wash- ing, at 3s. 6d. per ton, while excellent coke has been manufactured with the aid of this method from coal previously considered to be altogether unavailable for that pur[>06e.

At Kattowitz, in Upper Silesia, a plant treating 1500 tons per day of ten hours has reduced 47 per cent, of ash to 6 per cent, be- sides separating enough pyrites to pay the whole expense of treating the coal.

As to the quantity of water required, it is reported that 15 cubic feet per minute is enough for an 800-ton plant.

Blowing-Engines.

The following table, compiled from results taken almost at random, but from authoritative sources, may be relied upon as trustworthy :

Operations of Lvhig CocJ Waehing PlanU.

Works. Covutry.

SileRian Coal- and Coke- Works, Got-

tesburg, Silesia.

Pluto and Mercur, Gersdor . . Saxony.

Wolfsbank, Essen, Rhenish Prussia.

Julius Phillipp, Bochum, . . . Westphalia.

Stadt Osnabrdck, Hanovr. Prinz Sqaumburg-Lippe, Erbstol-

len, Austria.

Boussu, Belgium*

St LoDis, Anzin, France. Graf Ton Renard und Eulenbnrg,

Siedlce, Russian Poland.

Denaby Co., Mexbro', Yorkshire, . England.

Bpaclty

Un-

per ten

washed.

Washed.

hours.

Ash.

Tons.

Perec

.

5,5

Blo Wingenqines,

Discussion of the Paper of Mr. Kennedy. (See p. 587.)

F. W. Gordon, Philadelphia, Pa. : There can be no objection to the use of cross-compound blowing-engines; and when they are con- densing they are susceptible of great range of duty with slight loss in efficiency. By those who have use for the full heat of the waste gases, this type of engine is advantageous ; but many furnaces are not so located. The pig-iron is their finished product, and the gases are to be employed only for furnace-use. In this case simple engines are preferable, being less expensive and having less parts to take care of. These reasons determined the introduction of the engines, a sketch of which Ls distributed at this session. There are three of these engines to furnish blast for two furnaces, for a make of about 250 tons per day, using 76 per cent, anthracite and 85 per cent, magnetic ores ; and they are expected to blow up to twenty pounds pressure per square inch.

For blast-furnaces I prefer the single to the coupled engine, except

710 Blowing-Engines.

in compounding. The engine presented is practically the same as half of that presented by Mr. Kennedy, except in detail.

Our inlet-valve is double-ported and placed inside the cylinder, whereby it works under less constant pressure and has less motion, while the increase of clearance is but half of 1 per cent. These valves have each 452 square inches area, which, for the limited pis- ton-speed of 360 feet, is ample.

The outlet-valve openings in the head are 6 inches in diameter and there are 36 in all — 18 in each end. The valves are very light, of phosphor-bronze, have flat seats and slide on fixed stems. The valve's center of gravity is in the line of the seat, to avoid a ten- dency to cant. The end of the stem is a collar 2f inches in diame- ter, forming a piston fitted to a chamber in the valve. This acts as an air-cushion, saves the valve-seat, and prevents the noise t)f rapid closing. Loose leather collars, between which there is always air, form the cushion for the opening-movement. When compared to the usual area of air-compressor engines, these will appear excessive.

Horizontal blowing-engines have long since been out of date, yet simultaneously here are two large machines presented.

Our engine has a tail-end bearing. The piston-rod is 7J inches in diameter, and with its own weight and the weight of the blast- piston, has a deflection of inch. The slipper referred to and the full width bull-ring of the steam-piston has to carry this. The stufiGng-boxes are made to accommodate the spring of the rod. The blast-piston is turned -inch smaller than the cylinder, and the brass segmental rings only touch the cylinder.

We think that, with these plans well carried out, there can be no more objection to the horizontal blowing-engine than to the ordinary horizontal engine so extensively used ; while both possess the same advantages.

Joseph Morgan, Jr., Johnstown, Pa.: An existing example of the double horizontal coupled engine is that which the Cambria Iron Company contracted for in 1887, and finished in 1888, for its new Bessemer plant. The builders were the Southwark Foundry and Machine Co., of Philadelphia.

In design it is two horizontil steam-engines 48-inch diameter by 72-inch stroke, of Porter- Allen type, coupled at right angles to one shaft. The shaft carries a fly-wheel 22 feet in diameter, and 30 tons in weight. The point of cut-off is regulated by hand-levers moving a link-block to give the speed and blast-pressure needed. Owing to the low pressure (70 to 80 pounds) allowable upon the boilers then

Blowing-Engines. 711

in use, the engine was not compounded. Although this would have increasecl the cost of the engines, yet in building a new plant the amount saved on boilers would make the total cost about the same. I therefore quite agree with Mr. Kennedy that horizontal double compound engines will be the blowing-engines of the future.

The blowing-cylinders have 60-inch diameter, and 72-inch stroke, and are arranged tandem to steam-cylinders. In other words, the general plan is precisely that of the Youngstown engine shown by Mr. Kennedy.

The most novel features of this Cambria engine are the valves. Both inlet- and outlet-valves are metallic slide-valves of gridiron- type and fan-shape. The valves are in heads of the cylinders, and are worked by nearly positive mechanism ; and after four years' work, having blown 1,100,000 tons of steel in that time, the faces of the valves are in excellent condition and little worn. This is to be attributed to the fact that the motion of the valves takes place at times when they are subject to little or no pressure. Both inlet- and outlet-valves are so arranged as to lift slightly, in case the actuating mechanism fails, so that no excessive pressure can be produced. The valves were designed by Mr. Good of the Southwark Co.

The clearance in the air-cylinder is about 2 per cent. The high- pressure air-current is 93 i)er cent, of the steam-card. The revolu- tidhs made in blowing 28 pounds of blast on a twelve-ton con- verter, vary from 25 to 40 per minute. With improvements possible in the valve-mechanism, 60 revolutions could be easily maintained.

The apparent objection to the horizontal ty|)e of engine is its greater wear of piston- packing ; but with proper construction this is small, and is less than that of the engines ordinarily used to drive rolling-mills. All other considerations favor the horizontal type. As Mr. Kennedy says, it is more stable, more accessible, and cheaper. The steam-parts of the Cambria engine show little wear in four years; and the repair-cost of metallic blast-slide-valves is in their favor. As to loss of engine-power by break-down, we have on one occasion uncoupled one side and run the other engine alone at in- creased speed for a week ; and we have repeatedly uncoupled the blast- valve-actuating mechanism on one end of one blowing-cylinder, and run three-fourths of the double engine. To do this we have four stop-slide-valves on the nozzles of the blowing-cylinders.

James C. Brooks, Philadelphia, Pa. : Owing to the large cost of maintenance of blowing-engines for Bessemer and blast-furnace

712 Blowing-Engines.

purposes, due to the continued crystallization and breakage of the ordinary metal valves used, and the liability to accident from sach breakages, also the consequent limited speed of such machiDes, engi- neers have given much thought to the design of some device to reduce the repairs and liability to accidents, and at the same time increase the efficiency of engines of this class.

With the exception of the old-fashioned leather flap-valve, it has been almost universally the custom to use various modifications of the mushroom-valve, which are actuated in one direction by the air- pressure and in the other by a spring. To make them tight and protect them from undue jar, various forms of leather and gum cushions have been used. All, so far as I have been able to learn, require frequent renewal and expensive repairs.

To overcome the liability of breakage, these valves have at times l)een made very small. In doing this the area of the valve-openings is greatly reduced, and the consequent friction raises the temperature of the air very much. For this reason the density of air actually delivered to the cylinder is less than it would be if allowed to enter more freely. Even with the larger valves of this type the tortuous passage of the air as it enters the cylinder creates considerable fric- tion, and heats the air to a higher temperature than if it had a free passage.

In addition to this, it is impracticable to allow these valves a lift of one-quarter of their diameter, in order to get the full area, without shortening their life very much. Furthermore, these valves are very erratic in their movements, and therefore the area counted upon is never realized in their operation.

The Southwark Foundry and Machine Company of Philadelphia, with which the writer is connected, has expended a large amount of money in experiments ui>on air-valve gear, with a view to over- coming the difficulties explained above. They feel that they have succeeded in getting a free passage for the air to and from the tub, and at the same time in making the gear entirely automatic and capable of running at much higher speed than has heretofore been possible. Thus far they have .used gridiron-valves, so well known in connection with steam-engines. They have made them with as many {>ort8 as practicable, in order not to reduce the area of the valve-opening to such an extent as to have the air affected appreci- ably by the heated parts of the cylinder-head.

In an 84-inch blowing-tub they have been able to secure at each end seven laie ports for inlet-, and a like number for the outlet-

BIX)WiNG- ENGINES.

valves ; these ports being of very liberal area, giving an average of 10 per cent, of air-piston. They have, however, designs for other

O

styles of valves, upon which they are experimenting with a view to further improvements, if practicable. By having a number of ports it will be understood that the moment the valve begins to open the

/

Blowing-Engines.

area is multiplied very fast. Notwithstanding the valve is almost instantly opened the full amount, this feature is an advantage.

Fig. 1 shows the head on the blowing-tub of a double horizontal blowing-engine, made for the Bessemer plant of the Cambria Iron Company, by the same firm, in 1888. This valve, it will be noticed, is of the fan-shaped, grid iron -type. Under ordinary circumstances this would be subject to criticism, owing to the fact that there would be more wear near its periphery than at its axis, thus leading to leakage ; but the design is such that in its operation it lifts from the seat just at the moment it moves, and rests on the back-surface until the movement is completed, when it is forced back to its seat by the incoming or outgoing air, as the case may be.

The efficiency of this device is shown by the fact that after five years of constant service the wear upon the face of these valves is hardly perceptible. This valve is actuated by steam, applied to the piston of a supplemental cylinder as shown in Pig. 1. The steam is admitted and shut off from this cylinder by a pilot-valve, which in turn is moved by the accumulating pressure in the tub, acting on a differential piston and a cam, alternately.

This vertical differential cylinder is connected on the larger end to the blowing-cylinder, and on the smaller end to the receiver. The ratio of this differential piston is proportioned in such a way as to cause it to start the movement of the pilot- valves a little in ad- vance of the time at which the piston in the blowing-tub arrives at the point where the outlet- valve should open. This is done to pre- vent the building up of a pressure, in advance of the opening of the outlet- valve, greater than that in the receiver. This device has accomplished the purpose for which it is designed to the entire satisfaction of the makers and users of this engine. The inlet- valve is positive in its movement, and is actuated by a cam, as shown in the figure.

This engine, I understand, has given no more trouble to the users than the ordinary steam-engines without the air attachment used about their mill. In other words, it is not necessary for the engine to be overhauled on Sunday in order to have it ready for the week's work, as is the case with the old-style blowing-engines. The engine in which the above valves are used is a quarter-crank, with steam-cylinders 48-inch diameter, air-cylinders 60-inch diameter, both 72-inch stroke.

Fig. 2 shows a plan, and Fig. 3 an elevation, of an engine similflf to that furnished the Cambria Iron Company, with the exception

Blowino-£Ngin£S.

that it is a cross-compound. The steam-cylinders are of the well-

s'

be

a

o

s

known Porter-Allen type. This engine can be disconnected, and one side run alone.

Blowing-Ekgines.

I very much prefer Lorizontal engines, as all the parts are more

stable and accessible, and less liable to nlect, owing to thefccttha it is so easy to look after oil-cups and other parts.

/

It has been argued by many that the horizontal engine is

difficult of maintenance owing to the wear of the cylinders. In

the above case there has been no trouble whatever from this source.

Front.

Fig. 4.

Back.

Starting op. Partial Pressure in the Receiver. Fio.5.

Front.

Starting from no Pressnre. Fig. 6.

Starting from no Pressure.

The cylinders are in excellent condition, although no tail-rods or other devices, frequently thought necessary, have been used.

The above-named company has in operation horizontal tandem

'i

Blowing-Engines.

compound engines, with low-pressure cylinder, 75-inch diameter, 66- inch stroke, running at eighty revolutions as common practice, and

Fig. 8.

Fio. 9.

Fig. 10;

Fig. 11.

Indicator Cards taken from Air-cylinder of a 60'' and x Blowing-engine

at the Homestead Steel Plant, June, 1893. Cards taken bj their own expert.

Revolutions per minute, 60 ; boiler-pressure, 90 lbs. ; blast-pressure, 22 lbs.

at times higher, and after a year's use there are no signs of diflBculty with the cylinders.

720 Bix)Wing-Ekgine8.

The air-end of the blowing-engine mentioned above is entirely automatic. As an interesting exhibit I insert Figs. 4, 5, and 6. The first is a card taken when starting the engine, with partial pres- sure in the receiver. It shows that the pressure was built up auto- matically from the point at which the engine started to the maximum. The two latter show a card taken with no pressure in the receiver at the start. They illustrate the automatic building up of the pressure from the atmosphere to maximum.

In explanation of the apparent distortion of the above diagrams, I would say that the cards were not superposed, owing to the fact that the string stretched as the speed increased, and in opposite direc- tions— the indicators being one right and one left. Each of the various diagrams, when considered separately, shows the absolutely automatic action of the valve.

Fig. 7 shows an arrangement of our valve-gear on a tub, 84-inch diameter by 60-inch stroke, to work vertically, built for blast-furnace work. This gear shows the outlet-valve as opened with the pres- sure from the tub and closed by a cam. The Southwark Foundry and Machine Company has gears, which are not shown here, where the outlet-valve is operated with air from the tub in one direc- tion and from the receiver in the other. The inlet- valve in each case is positive.

Figs. 8, 9, 10, and 11 show cards taken from a Bessemer engine, operated by air both ways, now running upon the Bessemer plant at the Homestead Steel Works of the Carnegie Steel Company, Lim- ited. These cards were taken by Messrs. Daniel Ash worth and P.J. Fickinger, mechanical engineers and experts, of Pittsburgh, retained by the Carnegie Steel Company to look after the economical working of their engines and boiler plants.

I have no hesitancy in saying that enough has been done in the direction of high-speed automatic air-ends for blowing-engines to assure engineers that they are a success, and that at no distant day the old-fashioned mushroom- valve will be a thing of the past. There have been the usual difficulties and anxieties in reaching the present state of the art ; but we now have no doubt as to the practicability of running the air-end of blowing-engines as fast as it is advisable to run the steam-end.

The improvements shown in these illustrations are so novel as to have made it possible for the inventors to obtain patents with prac- tically no references by the Patent Office.

As to blowing-engines, the writer hopes to see the day when

Blowing-Engines. 721

the vertical engine, driven from the fly-wheels, will also be a thing of the past, as the cross-head of necessity becomes a long beam. This engine is commonly used because of the small amount of room occupied in its installation, and the seeming lower price. I say ''seeming,' because in my opinion an engine of this class in five years time costs any one using it much more than a quarter-crank, either horizontal or vertical, after the interest on extra cost for build- ing and ground is added, owing to the increased cost for maintenance and repairs. The breakage of a cross-head in the first-mentioned engine sometimes leads to a very large expense, as there is a possi- bility of the destruction of the most expensive parts of the engine.

David Baker, Sparrow's Point, Md. : The question of the best type of blowing-engine is of vital importance to the blast-furnace manager, and should be thoroughly discussed. There is room also for discussion on another point, namely, the steam-valve setting best suited for either or all types. By more careful attention to this point many costly break-downs might be avoided.

In many engine-rooms the indicator is unused, and if the engineer keeps the air-valves in fairly good condition, he thinks that, together with the general care of the engine, is sufficient. This is true mostly at small furnace-plants.

The experience of the Maryland Steel Company in this matter may be interesting and helpful. The blowing-engines used at this plant are of the double vertical condensing-type, the cylinders being 44 by 60 by 84 inches, and fitted with piston steam- and exhaust- valves. The steam-valves are controlled by the Allen link adjust- able cut-off. The air- valves are of the regular pop|)et-type, 7 inches in diameter, and were originally provided with leather faces.

These engines were built by the Southwark Foundry and Machine Company, of Philadelphia, from designs furnished by the Maryland Steel Company. In their construction no metal was spared to make a very rigid vertical engine, a result which has been realized in their working.

The Bessemer blowing-engines in the same house are from the same patterns as the furnace-engines, differing only in the size of the cylinders, tie-rods, piston-rods and foundation-bolts, the cylinders measuring 54 inches instead of 44 inches. The exhaust- valves in these engines were set to close when the piston was 10 inches from the end of the stroke, and could be set at no nearer point.

When the engines were erected the steam- valves were set with 3- inch lead. This was sufficient to make the compression in the VOL. XXII.— 46 ,

722 Blowing-Engines.

cylinders very high, so that the work of compressing the air was doubled at the end of the stroke,*by the compression in the steam- cylinder.

After a while this doubled load at each end of the stroke began to tell on the engine. The first trouble was a broken crank, and then beam-cap bolts began to break. At the same time we noticed that the bed-plate under the main shaft-bearings was lifting. We broke main cap-bolts, and finally a foundation -bolt in the bed-plate near the main shaft-bearing. After repairing, and before starting the engine again, the steam-valves were set to give some lap. The change in valve-settings showed its efiect at once. We had no further trouble with the caps or with the lifting of the bed-plate, A similar change in the No. 2 engine produced the same result and prevented any damage to bed-plate or foundation-bolts.

The amount of lap which we have scuttled upon as giving the best results is -inch. When running the engine condensing, this gives practically no compression above atmospheric pressure, and at the same time gives an economical card, considering the increased life of the engine.

On the furnace-engines we have given the steam- valves about inch lap ; as the exhaust- valves in these engines close when the piston is 3 inches from the end of the stroke. The furnace-engines run more smoothly, and there is not so much trouble in keeping the engine from knocking.

Mr. Kennedy speaks of the vexation due to the use of leather, gum, and such short-lived materials for valves. We experienced that, especially with the discharge- valves of the Bessemer blowing- engines. After trying a number of different compounds for the faces of the discharge- valves, we decided upon disks made of valca- beston as giving the best results.

This material, which is manufactured by the H. W. Johns Manufac- turing Company, New York City, is made of asbestos and rubber, vulcanized and pressed into moulds to the shape and thickness desired.

We used disks f inch in thickness and pressed to such a density as to be easily drilled with an ordinary powr twist-drill. The first of these disks has been in use over a year and shows about inch wear ; the leather ones lasted two weeks.

The Separation Of Blende Fbom Pyriteb. 723

The Sepabation Of Blende Fbom Ptbite8,

Discussion of the Paper of Prof. Blake. (See p. 669.)

C. Q. Payne, New York City: Prof. Blake's inference that magnetic separation may be successfully epi ployed upon smithsonite and iron oxide, after a preliminary roasting, is confirmed by the fact that this separation has been carried on with entire success for over a year at the works of the Wythe Lead and Zinc Mine Co., Austin- ville, Va.

The ore mined on this property is a mixture consisting chiefly of galenite, smithsonite and liraonite. The crude material is crushed, screened and washed in log washers for the removal of clay, and then jigged for the separation of the galenite.

The zinc- and iron-ores are then roasted in reverberatory furnaces, being turned and moved towards the bridge-end of the hearth by manual labor. The roasted material contains from 20 to 26 per cent, of metallic iron. The slight difference in the specific gravities and the friable condition of the ores after roasting prevent the sepa- ration of the oxides by means of jigging. It was therefore the prac- tice, previous to the use of magnetic separation, to charge the roasted material directly into the retorts of the zinc-furnaces. The rapid cutting of the retorts under this practice proved a serious in- convenience and expense.

In the spring of 1892, the writer built for the Wythe Lead and Zinc Mine Co. a magnetic separator which embodied several novel features and which was designed to treat the roasted zinc- and iron- oxides. This separator has proved entirely successful for the pur- pose for which it was designed. During the first nine months of its operation, 1415 tons of 2240 pounds of roasted material were passed through the separator, producing about 675 tons of heads carrying from 50 to 55 per cent, of metallic iron, which are sold to a neigh- boring blast-furnace. The tailings, which contain the zinc, carry from 4 to 6 -per cent, of metallic iron.

For magnetizing the iron-ore previous to separating it from the zinc-ore, no change whatever was made in the form of the reverbe- ratory furnace previously employed. It was found that, by charg- ing near the bridge-end of the hearth, toward the end of the roast-

724 SPECIFIC GRAVITY OP GOLD IN GOLD-SILVER ALIiOYS.

ing-operation, from 300 to 400 i>ounds of fine coal-dust per ton of ore, the iron-ore could be rendered quite strongly magnetic.

After drawing the charge from the furnace, it is partially quenched with water and passed through the separator. The pro- cess here employed for rendering the iron oxide magnetic is interest- ing from its simplicity, and also from the fact that, as far as my knowledge goes, this a)rapany was the first in this country to mag- netize an iron-ore by roasting, on a commercial scale, in the rular metallurgical treatment of its ores.

The Specific Qbavitt Of Gold Contained In Gold' Silver Allots,

Discussion of the Paper of Mr. Louis. (See p. 117.)

C. A. Stetepeldt, San Francisco, Cal. (communication to the Secretary) : In view of Mr. Louis's statement that the balance and weights employed in his experiment were " by no means first-rate," the results can scarcely be entitled to weight. And if a slight vari- ation in specific gravity were to be accepted as proof of the exist- ence of allotropic modifications, we should be obliged to conclude that every noetal existed in such modifications. When the observed specific gravity of an alloy (as is often, and perhaps generally, the case) differs from the calculated one, it seems more reasonable to ex- plain the difference (apart from errors of determination) by assuming that actual chemical combination takes place between some atoms of the metals.

Mr. Louis (communication to the Secretary): In reply to the criticisms with which Mr. Stetefeldt has favored me, I to point out to him that, in the first place, I do not myself attach very much weight to the results I have obtained. I merely cliEiim to have pointed out that amorphous gold obtained from dissolving up the alloy with silver has a higher specific gravity than common gold— a result corroborated by the quoted results of Rose. At any rate, it is easy to repeat my observations, and I hope this may be done by men who have better opportunities than myself for this investigation. Surely it is worthy of attention from a&sayers at the mints.

I am not prepared to say that I should regard slight variations in specific gravity as absolute proof of allotropism, but I do consider

Ic

Detection And Measurement Of Fire-Damp In Mines. 725

them as valuable evidenoe, taken in conjunction with variations in other physical characteristics (in the case in question : color, luster, ' tenacity, etc.). Is Mr. Stetefeldt prepared to say that every metal does not assume, under some condition, allotropic forms? I am personally inclined to believe that allotropism plays a far more im- portant part in metallurgy than is usually assigned to it

With regard to the state of existence of metals in alloys, I have little doubt that some form of feeble chemical union does exist between them, and I look upon the rare circumstance of the specific gravity of an alloy being exactly the mean of those of its constitu- ents as an arithmetical accident. At the same time I suppose it will not be asserted as a proved fact that allotropism itself is not a form of feeble chemical combination between atoms of the same substance.

The Detection And Mea8Ubement Of Fibedamp In

M1Ne8,

Discussion of the Papers of M. Chesneau (see p. 120), and of Professor Clowes

(see p. 606).

C. Le Neve Foster, Llandudno, Wales : I regret that Prof. Clowes is not here, and also that I cannot produce one of his lamps for inspection. The lamp has but just passed the experimental stage, and is now on sale in England for the purpose of testing it in practice. The only complaint I have heard made against it is the high price. It costs six or seven guineas. Possibly that would not frighten you on this side of the water as much as it does us.

R. W. Raymond, Brooklyn, N. Y. : So far as my observations and experience go, we have not yet reached a state of affairs in the coal-mines of this country which calls for the general use of safety- lamps by common miners. Certainly such a practice is to be avoided if possible, on account of the recklessness of miners, which it seems almost impossible to control by warning or discipline. It is much better, if practicable, to confine the use of safety-lamps to inspectors or " fire-bosses,*' and not to permit the miners to work or to go where naked lights cannot be carried. In a colliery near Pottsville, Pa., with which I was connected for five years, and which had the reputation of being highly fiery " (jets of fire-damp bub- bled continually through the puddles in the gangways, and a thin sheet of it lay along the roof), we managed to avoid all serious j

726 Detection And Measurement Op Fire-Damp In Mines.

trouble by a thorough daily inspection and an abundant mechanical ventilation by suction-fans. No miners were allowed to enter any part of the workings until after such an inspection. I should add that we were not subjected to sudden inbursts of large quantities of the gas, and our heavy air-current sufficiently diluted and swept away the steady, slow leakage of fire-damp, so that the air was kept safe. Careleaness in approaching a naked light to the roof some- times resulted in a flash ; but there were no large accumulations of gas anywhere, and no explosions occurred. I think this is, in the main, the system pursued in American collieries. For the ose of inspectors the best apparatus would not be too dear.

H. S. MuNROE, New York City : M. Chesneau, in his paper, gives full credit to Mr, Thomas Shaw, of Philadelphia, for his method of determining the percentage of fire-damp or other ex- plosive gas when mixed with air. Mr. Shaw was the first to dem- onstrate that a definite minimum |>ercentage of gas must be present to make the mixture explosive, and that the line between an explo- sive mixture and one that is non-explosive is sharp enough for quantitative tests. Mr. Shaw finds, for example, that six per cent of pure marsh-gas in air gives an explosion; and one-tenth of one per cent, less is non-explosive. M. Le Chatelier's experiments, as cited by M. Chesneau, fully confirm Mr. Shaw's results, and show thut his method of determining fire-damp is accurate and reliable.

M. Chesneau, however, remarks that the Shaw apparatus is much too complicated, and, therefore, too costly for general use in mines. Moreover, it does not lend itself easily to the analysis of samples taken underground.*'

A Shaw gas-tester has recently been purchased for the use of the School of Mines, and my experience with the apparatus thus far has shown that M. Chesneau's criticism is not well-founded. The ap- paratus, costing upwards of $500, is not cheap ; but it is by no means too costly for general use in mines,*' when its many advan- tages are considered. It is not as complicated as would at first sight appear, and it is admirably adapted to the analysis of samples taken underground, making the test in much less time and demanding much less skill on the part of the operator than the simpler appara- tus of M. Le Chatelier.

In the hands of a skilled chemist the graduated tube of M. Le Chatelier will doubtless give as good results as in the hands of its distinguished inventor ; but in the less skillful hands of our mine- officials, who find no difficulty in using the Shaw gas-tester, I fancy the Le Chatelier tube would prove quite impracticable:

oogle

Detection And Measurement Of Fire-Damp In Mineb. 727

s

o

M 0/

H

728 Detection And Mb/Isubement Of Fire-Damp In Mines.

Fig. 1 is a perspective view of the Shaw gas-tester. It is an air- pump with two cylinders, A and B, operated by the arm, C, by means of the crank, N. The large air-cylinder delivers 800 c.c. at each stroke. The smaller gas-cylinder, B, delivers a variable quan- tity, depending on its position, which can be adjusted at will. The graduations on the bar, S, upon which the small cylinder rests, and upon the oscillating arm, to which the piston-rod is clamped, deter- mine the percentage of the gas from the small cylinder, as compared with the total from both cylinders.

There are no valves either in the cylinders or in the pistons of the pum|>s, but, in their place, a rotary slide-valve, L, operated by the connecting-rod, W, regulates the flow of gas and air.

The larger cylinder, A, is intended to be supplied with the air- mixture to be tested. The smaller cylinder is supplied with pure gas, marsh-gas, or illuminating-gas, as may be most convenient.

By shifting the small cylinder back and forth until an explosive mixture is produced, we determine by ditTerence the amount of gas already present in the mixture tested.

The air and gas from the two pump-cylinders pass through a " mixer (not shown), and, by means of the valve, K, the mixed gases pass into the cylinder, X, or the cylinder, Z, at will.

The cylinder, X, is used for a preliminary qualitative test of the mine-air, and for the determination of CO,, etc.

The cylinder, Z, serves to determine whether the mixture is explo- sive or not. If explosive, the mixture is ignited by a gas-jet, which is kept burning at one side, and the force of the explosive drives out a piston and rings the gong. The test takes no longer than is nec- essary to make one or two strokes of the pump; and, when this is compared with the very elaborate manipulation described by M. Chesneau, the advantage of the Shaw apparatus will be evident The five or six tests necessary to determine the proper |>osition of the gas-pump, B, and, by difference, the amount of gas in the sample tested, can be made in three to four minutes.

The sample of mine-air to be tested is brought to the surface in rubber bags, which are filled underground by means of a small dia- phragm hand-pump furnished with the machine. The test shoald l)e made within a few hours after reaching the surface, since, accord- ing to Mr. Shaw, there will be a loss of about one-half of one per cent, in the hydrocarbon gas if the mixture remains 24 hours in the rubber bags.

The apparatus can also be used for determining the percentage of CO, COj, II2S, and other gases in air. j

Detection And Meaburement Of Fire-Damp In Mines. 729

For these and other applications, liowever, the reader is referred to more elaborate descriptions published in Mechanics, for September, 1891, and in the Colliery Engineer y for August, 1893, and to circulars of the patentee.

EcKLEY B. CoxE, Drifton, Pa. : It is difficult to get even a fire- boss to appreciate and remember the danger. In one of our col- lieries we had a couple of places containing a little fire-damp. Our boss, a very competent Irishman, went up one morning into one of these breasts, tested the air, came down again, and told the two miners waiting lelow that it was not safe for them. As he was going on, he suddenly remembered that he had lefl something up in the breast, and he actually went up again, taking with him one of the men he had just warned, who carried a naked lamp in his hat !

In an adjoining mine, where there was a certain somewhat fiery place, the fire-boss was in the habit of taking his naked lamp from his hat, hanging it on a prop, and taking his safety-lamp in his hand while he went up into the dangerous working. On one occasion, he went through the usual programme, only, instead of hanging up his nake<l lamp he absent-mindedly put it back on the front of his hat, and ascended, safety lamp in hand, to test the gas. While he w&s making his careful examination, the flame on his hat blew him up! Such things are occurring all the time. We want an apparatus to be as simple as possible, and then we must not trust either the ap- paratus or the men any more than we are forced to. Another in- stance, which occurred within a few hundred yards of where I live, illustrates the thoughtlessness of workmen, though it is not a case of dealing with fire-damp. We had a dangerous piece of blasting to do, and employed for the purpose one of our very best men. He went into a heading about fifteen feet, and loaded and fired a shot. Returning, after the smoke had cleared away, he looked into the place, and judged it to be in too dangerous a condition to be entered until the loose roof should have fallen. But just as he was going away, he looked again and saw that he had lefl in the heading his ''blasting-barrel '' (a piece of pipe, worth about five cents); and it occurred to him that, when the roof fell, it might bend or jam that tool so that it could not be used. So he went back and got his blast- ing-barrel. The roof fell immediately afterwards, and it was the merest accident that he was not crushed. I had a talk with the man, and asked him how he could have been so reckless. All he could say was, I saw the blasting-barrel, and so I went in 1" If anybody had offered him five cents to risk his life in that way, he would have

730 DETECTION AND MEiSDREMENT OF FIRE-DAMP IN MINES.

refused with scorn; but the sudden impulse to prevent a merely passible loss of five cents, entirely overpowered his prudence and common sense.

This is a fundamental difficulty in the use of all particular ap- paratus. We assume that the miner is afraid of danger, and that, in his own interest, he will be careful, whereas, he gambles contin- ually on the risk of danger, and comes to consider disaster as a mat- ter of fate. We assume, also, that the apparatus we employ will be handled and used with care. This may be true for a while; bnt in the course of time, it will become an old story, 4ind the practice will grow careless.

This principle must be borne in mind concerning all new im- provements. Arrangements by which a certain percentage of fuel is to be saved in raising steam, for instance, are likely enough to show that saving during the experimental period, when they are operated with special vigilance by men who are interested in their success, and in comparison with the ordinary daily practice, to re- place which they are proposed. But the gain generally disappears when the new practice has become ordinary.

The Shaw apparatus, to which reference has been made, has been employed, I believe, in Wilkes-Barre.

Herbert W. Hughes, Dudley, England : So far as I can judge from the discussion, the English practice of examining mines differs from. the American one in not allowing the fire-boss to carry with him any naked light. Our examination commences from the sur- face with a locked safety-lamp, and, consequently, the accidents which Mr. Coxe described cannot occur. I quite agree with the statement made by that gentleman, to the effect that if* the carefal- ness and steadiness of the men using the apparatus could be im- proved, more good would result. Familiarity breeds contempt ; and the ordinary miner often neglects the most elementary precautions, not from ignorance or rashness, but because the oonditiims have grown so familiar that a thought is not given to the danger. My own experience with delicate indicators for detecting small quanti- ties of gas is that they work well and satisfactorily when new and in the hands of the chief officials, but that they are practioaliy use less when employed by the ordinary working miner. I have per- sonally used most of the indicators on the market, and obtained reasonably satisfactory results, but on handing the appliances over to miners for everyday use, they have been failures. If the Pieler lamp is used as it should be — that is, only after the ordinary safety-

Detection And Measurement Op Fire-Damp In Mines. 731

lamp has been tried, and has failed to detect gas — it gives good and delicate indications of small quantities of inflammable gas. Its dis- advantages are pointed out by M. Chesneau in his paper; but in addition to these, it possesses another one in the fact that it gives practically no light, and the miner is obliged to carry a second lamp. While M. Chesneau has improved the safety of the Pieler, his lamp is still open to the above objection. This point is of such import- ance that all Prof. Clowes' experiments have been conducted with a view of providing a safety-lamp which can be transformed at will into a delicate indicator. He has succeeded in two ways: 1st. By a modification of the Ashworth-Gray lamp, burning benzine, by means of which 0.50 per cent, of firedamp could be detected with difficulty ; 2d. By the hydrogen-flame lamp described in the pres- ent paper. Theoretically, the latter lamp is perfect The objection raised by M. Chesneau, that as the pressure in the hydrogen-cylin- der diminishes, the height of the flame diminishes, does not hold good. The cylinders are provided with a regulating-valve, and the testing-flame is always regulated by the operator to the standard height of 10 mm. Unfortunately, the lamp is not only costly, but delicate, and I am afraid that if it is to be used in everyday mining, a different class of miners will have to be invented. I have previously pointed out that when a miner is at work, his hands are generally not so clean as those of persons whose avocation lies at the surface, and that a miner usually handles things in a rough-and- ready way. The Ashworth-Gray lamp, burning benzine, appears to me to be the most practical form for everyday use. Respecting the means of damping dust referred to by previous speakers,* I may point out that it is now generally recognized in Europe that the chief danger from accumulations of coal-dust is not from that reposing on the floor, which is more or less coarse and mixed with dirt and slate, but from the minutely subdivided particles which rest on the timber supporting the roof and sides. The most successful method of watering is that adopted at Dowlais, in South Wales, where a combined jet of air and water is employed. The water is projected in the form of a minute spray and is carried great distances before settling. The air-current is damped, and, consequently, wherever this circulates, the dust is damped also. Introducing a jet of steam into the air-current has been tried, but has not met with any large degree of success.

Not reported.

732 Geological Distribution Of Thb Useful Mbtal9.

Obolooical Distribution Of The Useful Metals In The United States,

Discussion of the Paper of Mr. Emmons. (See p. 53.)

John A. Church, New York City : It requires some courage to appear as a critic of a theory which is not only the fashion among American geologists but is usually presented by them in terms which imply that any other views are an exhibition of ignorance. Still, I am obliged to say that the theory of lateral secretion as it is stated in this and other writings of Mr. Emmons and other geologists has not added much to our real knowledge or clearness of view. Id the earlier and less developed stages of the theory, when it was used as Sandberger used it, to show that certain veins were probably derived from the rocks in which they lie, or which are adjacent, it was valu- able in pointing us to an immediate source of ore-deposition. When we are driven to assume the existence of undiscoverable rocks at an unknown but certainly a considerable distance and in an unknown direction from the vein, I do not see that we have improved upon the despised unknown source in depth " with which our igooranoe has been covered so long. The new theory may suffer from adoles- cence, and these points may be cleared up by further study, but I speak of it as it is.

Differentiation in a magrna, by which a metal is concentrated in one member of a series of outflows, may explain why certain ores have favored a given locality with their presence; but it is not a necessary precedent to ore-formation. Concentration in the source of supply cannot be a requirement, for the forces that have been able to take up four or five tons of gold from an extensive body of rock must be able to collect four or five thousand tons of lead, copper or nickel from a proportionately more extended body of rock. That is to say, concentration is no more essential for these metals than it is for gold.

In fact, differentiation, as it is now explained, is not an advance upon old ideas, but a retreat from them. It was noticed long ago that violent eruptive phenomena, however long continued, died away in solfataras ; and when a vein came to be looked upon as an extinct solfatara the inference was ready that veins are eruptive in the sense that the solfataric waters collected the metals from the unerupted residue of the magma and carried them to the veins. The early views carried differentiation further than the modern school.

Geological Distribution Of The Useful Metals. 733

It is the fashion of the new school, to which I believe all profes- sional geologists in this country adhere, while all professional mining engineers keep themselves aloof from it, to speak of these old ideas as if they were very erroneous, and were necessarily brnshed aside by the advance of experimental knowledge. It seems singular to me that the new school should recognize no other origin for ores than thfe leaching of rocks by comparatively shallow water-currents and yet recognize no other origin for the fissures that carry the ores than cataclysmic action ! If it had been found that bed-planes were com- monly the channels by which the ore-solutions entered, we might accept the fact as evidence of lateral secretion ; but when I find the adherents of this theory declaring, as does Mr. Emmons (and as do all the others), that every ore-deposit lies in a plane of faulting, or bad been filled from a fault, it seems to me hardly logical to carry one branch of the volcanic theory to such an extreme, and totally reject the other branch, with which this view is undoubtedly in sympathy.

The older geologists looked upon a vein as a channel esiablished between the surface and the interior K the. earth. Into its lower termination poured solutions, the character of which was determined by the pressure and heat normal to the depth at which they may have been forme. The almost uniformly siliceous filling of veins shows that this depth was uniform in its conditions of solution. It may have been the whole barysphere or only that upper portion within which we may imagine a comparatively lively circulation. At least it was lower than the vein.

The crevice was supposed to be an open chamber, or series of chambers, with occasional points of support. Through this cham- ber the waters rose to the surface, where they were dischaied. As they rose, they necessarily passed through zones of continually de- creasing pressure and temperature, and " relief of pressure " and "lowering of temperature" were the potent agents which were supposed to effect precipitation, of the dissolved solids, discharge of gas being another. The idea that the rocks enclosing the crevice could act as precipitants received early attention, and has led directly to some of our most widely-accepted modern ideas.

Undoubtedly, these are plausible views ; and the agencies invoked are real agencies of precipitation, as we know from the action of hot telluric waters discharging upon the surface. One of our most noted veins — the Comstock — was studied and explained in the light of this older theory by Richthofen, in 1865 ; and these ideas have not been entirely abandoned there.

734 Geological Distribotion Op The Useful Metals.

Mr. Becker, within the last ten years, has gone into an elaborate argument to prove that there has been almost no erosion of the Corn- stock rocks. If his argument is sound, the outcrop now is within 25 or 50 feet of where the original outcrop was formed. It is true, his views contradict each other; and, if the dynamical conduct of the rocks had been what he describes, there would never have been an outcrop where the Comstock was found. Still, I believe 4M writers upon that noted vein, except myself, have represented it as a solfatara, in the sense that it was formed by hot waters from a deep unknown source discharging into the atmosphere. The attempts to connect lateral secretion with the lode have been failures ; and the Comstock still represents the old theories in their advanced form. It seems to me, that the solfataric origin of ores is a more reasona- ble explanation of the observable facts, in some cases, than the theory of lateral secretion. That minute quantities of the metals are found in all or many rocks, is true ; but the crucial question of their origin has never been determined. Do they form an original source, or a secondary deposition like the vein? is a question that has not been conclusively answered; buti agree, and I think most mining engi- neers agree, with Posepny, in believing them to be the latter.

We owe to the distinguished author of this paper one of the most striking and valuable contributions to the discussion of the lateral secretion theory. From his description we may say that, in his view, the ores at Leadville were not exotic, since the rocks in which they lie, the rocks from which they were leached, and the water-cur- rents that formed them, were all at substantially the same depth. These features are essential to lateral secretion ; for, if we allow that the circulating waters sink deep enough to reach the unerupted resi- due of the magma before they take up their metallic contents, we have the old solfataric theory of origin.

Sandberger's original idea that veins are filled by leaching from the rocks that contain them, has been so expanded by the discussion of ore-bodies formed under a cover. of three or four miles of rock, that it is brought, not into conflict, but into close symimthy with the solfataric theory. The question whether the origin was in erupted or in non-erupted magma is interesting; but it is not controlling when the action is acknowledged to have taken place at very great depth, far within the " barysphere " in either case.

Having reached that amount of agreement, it seems to me that the next task of the structural geologists is to determine critically

hether any vein has really been formed in a cievice discharging M3tly into the atmosphere. The conditions found at Steamboat

Digitized by —

Geological Distribution Of Th£ Useful Metals. 735

Springs and elsewhere seem to me to point to vein-action (if vein- action there is) at some other point. The discharging waters may be regarded as the filtrate derived from metasomatic precipitation lower down, or as a mixture of waters from the upper and lower regions. As yet, I doubt if we have any proof that ore carried by the deep circulation has been retained long enough to be deposited at the surface. Mr. Becker entertained that view, but I believe his conclusions upon the geology of the Comstock to be radically wrong.

Arthur Winslow, Jefferson City, Mo.: I think that others will join me in expressing thanks to Mr. Emmons for his admirable rmmi of our ore-deposits, and for the many valuable suggestions embodied in it. The ground is so well covered that there remains little room for additions, yet I should like to make a few remarks concerning some subjects touched on of which I have personal knowledge.

Mr. Emmons refers to the specular ores of southeastern Missouri collectively, as probably of Algonkian age. There is, however, a distinction to be made. Those of Iron Mountain occur in Aichsaan porphyry, in tongue-like masses or veins which taper out with depth. The ore of Pilot Knob, on the contrary, occurs as a bed interstrati- fied with Algonkian elastics composed of debris of the pre-existing Archaean porphyries. These ores have recently been made the sub- jects of renewed study by the Geological Survey of the State, through Mr. Nason. He thinks that the facts at Iron Mountain are such as to favor the theory that the ores there are derived from the decay of great thicknesses of porphyry, accompanied by a leaching out of the abundant iron content and its deposition elsewhere in crevices and openings of the same rock, at times i>o88ibly replacing the de- composed porphyry adjacent to these crevices. These deposits would thus be examples of chemical concentration from older basic ernp- tivep. In the case of Pilot Knob, Mr. Nason concludes that the iron-ore body is probably the result of replacement of certain mem- bers of the Algonkian series of strata. This would again be an example of chemical concentration from an older basic eruptive, though if we allow that the Archiean specular ores were formed prior to the deposition of the Algonkian series here, it is possible that this Pilot Knob bed is of direct mechanical origin from the abrasion of these earlier ore-masses.

In the iron deposits of central Missouri, which consist of a mix- ture of blue specular and red hematite ores, Mr. Nason concludes that we have instances of accumulation in cavities and depressions produced by subterranean erosion of limestone. The disturbed con-

736 Geological Dbbtribution Op The Usepdl Mbtalb.

dition of the adjacent strata, their converging dips and other facts corroborate this. That there was some replacement of limestone by the iron polutions is also undoubted. This is well illustrated by the recent discovery of crinoid remains in the body of the ores, replaced entirely by blue ore. An interesting fact about these fossil remains, and one which adds support to the theory that the ore accumulated in depressions or cavities, is that they are not fossils found in the Cambrian country-rocks, but they belong probably to a Lower Car- boniferous fauna. All over the Cambrian area of the Ozark uplift patches and fragments of such later rocks are found, indicating that a thin covering once existed there, of which portions are preserved in depressions and pockets in the older dolomites.

Remote as these ore-bodies are from eruptive rocks, we are obliged to seek for their source in the surrounding sedimentaries. Mr. Nason has fixed upon the sandstones of this area as the probable contribu- tors. These are often highly ferruginous, and are readily leached by percolating waters. It is probable that the decaying dolomites also contributed a share.

In referring to the zinc- and lead-ores of Missouri, Mr. Emmons has brought forward for discussion a series of most interesting, and at the same time, most perplexing deposits, so far as a satisfactory theory of their origin is concerned. Those of the southwestern portion of the State occur essentially in the Mississippian or Lower Carboniferous limestones. The statement that they extend into the Coal-Measures should be made with limitations. They are found in shales of that age in Jasper county, but those shales are in isolated patches, which occupy depressions in the older ore-bearing Missis- sippian rocks. Hence, the metallic contents of the shales may be derived, by some secondary process of transfer, from adjacent ore- bodies. In any case, the Coal-Measures in the State, as a whole, are practically destitute of these ores, which, therefore, can hardly be declared to belong to that formation, whether their general absence from it be due to their prior formation, or to limitations in their distribution determined by physical causes.

Cross-fissures or fault-fissures in these Mississippian rocks, to which Mr. Emmons alludes, if they exist, are not very apparent The strata are, undoubtedly, very much shattered in certain limited areas, and have been subjected to extensive subterranean erosion and corrasion and great silicification. Of a system of extensive or con- siderable faults, recent stratigraphic work in this region has, however, revealed nothing.

In the Cambrian limestones of the eastern part of theState, Ihe

GEOLOGICAL DISTRIBUTION OF THE USEFUL METAIi. 737

conditions are somewhat different. Crevices and fissures are there plainly developed, and evidence of considerable faulting is indubit- able. In Franklin county, such vertical crevices have supplied large quantities of ore. In that portion of the southeast to which Mr. Emmons especially refers, however, and which has produced by far the greater part of the lead, the crevices are of insignificant dimensions, and the experience has been that they contain, them- selves, little or no ore. On the contrary, the great ore-masses con- sist of galena disseminated through a thickness of the country-rock often of 50 feet and more. At Bonne Terre, a tract 1300 feet long by 800 feet wide has been mined out of such diffused ore.

The crevices which traverse this ore-body are frequently almost blind, and can only be detected by the drip of roof-water. These are such as traverse almost any massive rock. They have possibly had their influence upon the ore-deposition ; but, to picture them as veins and channels of direct ore-supply, seems hardly justifiable even under the existing differences of opinion among geologists as to just what a vein is. The questions of the source and mode of accumulation of these ores are very broad, and involve considera- tions of the ore-deposits, and of the geological history of the whole Mississippi valley. Only through studious, detailed analytic methods can a satisfactory solution be reached. I hope, at a later date, to have something further to contribute to the Institute on this subject.

Concerning the distribution of barite, to which Mr. Emmons re- fers on page 84, 1 would add, that it is not confined to the older Cambrian rocks of Missouri, but is found in Cooper and adjacent counties, at a number of localities, in Lower Carboniferous lime- stones, sometimes associated with lead ores. In a previous paper, presented to the Institute,'*' Mr. Emmons has suggested the use of the apparent fact of the limited distribution of barite in determin- ing the origin of the ores. This occurrence in the Lower Carbonif- erous lessens the value of the suggestion, though it remaius still locally serviceable.

Mr. Emmons : I am very glad to have the details which Mr. Winslow has given us with regard to the ore-deposits of Missouri; and as my knowledge of the greater part of them is derived, not from personal observation, but from the descriptions of others, I am quite willing to accept his corrections of my facts since his information is of later date than that to which I had access. I was aware that barite occurs at times in the Lower Carboniferoup

Trans*f xxi., 41. VOL. xxiL— 47 rn,n]o

? IvL

738 Origin Of Gold-Bearing Quartz Op Bendioo Reefs.

limestones of the Mississippi valley ; indeed, small amounts have been found in the fluorspar-deposits of Rosiclare, 111., since I wrote my previous paper to which he refers ; but I think it still remains true that fluorspar is characteristic of the one horizon and barite of the other. I wish to take this opportunity to correct an error into which, as I have recently learned, I was led by incor- rect information, viz., in stating (p. 70) that the nickel-ores of Churchill county, Nev., are silicates. It appears that they are sul- phides and arsenides.

Theobioin Of The Qold-Bearino Quartz Of The Bendioo Beefs, Au8Tbalia.

Discussion of the Paper of Mr. Rickard. (See p. 239.)

Richard Pearce, Argo, Colo.: This contribution upon the very interesting subject of the origin of the gold of certain lode-for- mations oflers many subjects for thoughtful consideration. The explanation, for instance, of the perplexing features of thin lava dikes seems to me to be in harmony with facts as I have myself observed them elsewhere.

At Marshall basin, near Telluride, Colorado, there is an extensive lava-flow. The material of this formation contains a large propor- tion of fragments of foreign matter in the form of angular piec of rock. The latter show no signs of having undergone metamorphism, such as would be expected to be due to the direct effects of heat; and the evidence in this particular case would certainly lead me to suppose that the flow was due to causes analogous to those suggested by Mr. Rickard. It was, I believe, a moving mass of material, hav- ing a temperature considerably below the actual melting-point of the rock of which it is composed.

In regard to the theory which the author has built up from the existence of the iodide of gold in sea water, I cannot consider it un- tenable; but I would put the question whether any investigations have been made with a view to determining whether gold and silver occur in the beds of rock-salt, formed, as is generally allowed, by the evaporation of bodies of salt water. At Stassfurt, in (Jermany, there is a very remarkable series of alternating beds of rock-salt and anhydrite, the uppermost members of which are covered with de- posits of potash salts. The question occurs to me, why should not gold and silver have been discovered in these deposits? Iodine and bromine, I believe, have been found in the salts last deposited by

Origin Of Gold-Bearing Quartz Of Bendigo Reefs. 739

evaporation ; and surely there was the organic matter present, able to precipitate the precious metals if in solution.

In connection with this question, it would be very interesting to determine whether the salts contained in the mineral springs of cer- tain mining regions, such, for example, as Glenwood Springs and Idaho Springs in Colorado, contain any traces of the precious metals ?

I see no reason why iodine should have so much of the credit for dissolving the gold. It may have been originally a solvent agent; but it is quite possible that chlorine may also have Splayed a part in the process of solution in more ways than the particular one sug- gested by the author. Chloride of sodium is usually present in the waters circulating through vein-fissures.

The liberation of free chlorine can easily be explained by the action of dilute sulphuric acid, obtained by the oxidation of pyrite, in the presence of oxide of manganese, itself nearly always associated with the hydrated oxide of iron in the gossan of mineral veins. The solution of the gold by such chlorine takes place readily, as I have showh by certain experiments, the results of which were communi- cated several years ago to the Colorado Scientific Society.* The precipitation of the gold from such chloride solutions can be effected by any reducing agent, such as ferrous sulphate, or even, as Daintree showed, by certain metallic sulphides.

In reference to the so-called "lateral-secretion" theory, I find myself in agreement with the remarks made by Mr. Kickard. My experience in Cornwall and in Colorado more particularly, would lead me to reject so narrow an interpretation of the origin of ore- depoeits as that they were derived from the leaching of the immedi- ately adjoining and encasing rocks.

In Cornwall, more especially, my early investigations indicated that the rocks themselves were affected by the action of solutions circulating through the fractures and joints, and that those solutions mineralized and impregnated the rock adjoining the vein rather than that the reverse action took place, viz., that the vein was fed by the leaching of the adjacent rock.

The association of bodies of igneous rocks with ore-deposits has afforded grounds for deductions which do not seem to be always in

Address of retiring president, January 7, 1885. In a small flask were placed some crystals and wires of native gold from the Ontario mine ; some finely -divided hydrate of binoxide of manganese was added, and the flask filled with water contain- ing about 40 grains of salt per gallon and some few (6 or 6) drops of sulphuric acid. The whole was kept quite hot for twelve hours, and at the end of that time the solution showed a strong reaction for gold." j

740 Origin Of Oold-Bearinq Quartz Of Bendioo Reep8.

harmony with facts. Mr. Rickard's suggestion that the favorable nature of this association may be due to the hot solutions aooompa- Dying or following the extrusion of such rock-masses is not out of accord with my own experience. Here also underground observa- tion would suggest to me that the igneous rocks adjoining the vein may have been mineralized by hot solutions percolating through the fractures where the ore is now deposited, rather than that the lodes were beneficiated by the leaching of the adjoining rooks. A particular illustration occurs to me. At the Seven Thirty mine, at Silver Plume, Colorado, I remember seeing a small (2 to 3 inches) vein of very rich silver-ore, traversing a granite country. The vein had well-defined boundaries; but for a limited distance to one side of the main streak the conntry-rock had a black, sooty appearance, due to the presence of argentite, and was nearly as rich as the vein proper.

The mineralized rock was but little decomposed, while that be- yond it was still less changed, and was barren of silver.

The evidence seemed to indicate that a certain amount of silver in the original vein had been transferred to the adjacent rock by solu- tion and reprecipitation. In Gilpin county, similar instances are common. Very frequently a rich streak has a larger width of less rich ore adjoining it. The former is the vein proper, and contains nearly solid iron and copper pyrites, called "smel ting-ore" by the miners, while the outer, wider, less rich portion is essentially min- eralized country, often part of a porphyry dike, which comes under the heading of " mill-dirt," and being too poor to ship to the smelters, is profitably treated at the stamp-mills.

These and a host of other instances, which occur to me as I con- sider the subject, have led me to conclude that the agency which pro- duced the vein proper had concentrated its action along a certain line of fracture, and that the outer deposit (no less important, bow- ever, to the miner), was due to impregnations from the inside of the vein to the outside country, and not from the outside country into the vein.

Philip Argall, Denver, Colo. : I have read with very much interest Mr. Rickard's descriptive paper of the Bendigo gold-field. The present paper, however, deals more in the realms of theory than of fact, and believing that an examination of some theories advanced may provoke further discussion, thereby aiding in the solution of the problems presented, I have ventured to submit the following remarks on some of the points where I am led to disagree with the author's conclusion. j

Origin Of Gold-Bearing Quartz Of Bendiqo Reefs. 741

Argument in support of the generally accepted doctrine of the hydrothermal fusion of certain granites is unnecessary at this late date. While, as Mr. Rickard says, the Bendigo granite does not exhibit a vesicular structure like that of the lavas to which he refers, yet his suggestion that " any such structure would long ago have been obliterated by those infinitely slow physical and chemical changes which are forever bringing about the decay and repair of rocks," is not applicable. Granite does exhibit a sort of vesicular, what might be called a micro-vesicular structure, characteristic of rocks consolidated under great pressure in the presence of water- vapor and gases. The minute vesicles of this structure uudoubt- edly differ from those found in rocks cooled at the surface ; but the difieren<3e is not the result of slow obliteration. It has existed from the beginning, as the effect of different conditions of consolidation. It was, in fact, chiefly to account for the vesicles containing water, etc., in the quartz of granites, that the theory of hydrothermal fusion was framed. The iateresting character of the subject may warrant a glance at its literature, a part of which has been mentioned in Mr. Rickard's paper.

In 1847, Theodore Scheerer's famous researches on granitic rocks were published by the Greological Society of France.* He advanced the theory that the granites were highly heated plastic masses con- taining water, which oozed out under great pressure, filling cavities and fissures, and penetrating along the bedding-planes of the strati- fied rocks. Scheerer pointed out that the more fusible feldspars and micas crystallized before the almost infusible quartz, and that the minerals allanite and gadolinite, found in granite, undergo, when heated below redness, permanent molecular and chemical changes. On these grouuds he rejected the notion of pure igneous fusion for granitic rocks, and adopted that of hydrothermal fusion. Elie de Beaumontf and Daubree pursued the subject, and showed that small quantities of water in eruptive igneous rocks may have contributed to suspend their solidification and to promote the crystallization of their silicates at temperatures considerably below their fusion- points, and in a succession different from their relative order of fusi-

bility.'t

H. C. Sorby§ adopted similar views, for the same reasons, and also because of the almost innumerable inclusions of water, gas, etc., in

Bulletin, 1847, vol. iv., p. 468. f id., 2d series, vol. iv., p. 1340.

1 Geikie's Tezt-Bookqf Oeology, 3d edition, 1893, p. 308.

2 Quart. Jour, Oeol, Soc, vol. xiv., p. 480.

742 Origin Op Gold-Bearing Quartz Of Bendigo Reefs.

the quartz of granite. The liquid inclasions are frequently water, or aqueous solutions of chlorides and sulphates of sodium, potas- sium, and calcium ; sometimes they are liquid carbonic acid, or water holding carbonic acid in solution ; occasionally, they are supersat- urated solutions of sodium chloride, containing minute crystals of rock-salt

J. Clifton Ward* says, that of these inclusions a thousand milh'on may easily be comprised within a cubic inch of quartz, and that the contained water sometimes amounts to at least 5 per cent, of the volume of the containing quartz.

Sorby, experimenting with the liquid inclusions and bubbles in granite, and taking as data their volume, and the relation of the substances to heat and pressure, estimated the pressure and temper- ature at which certain granites had been consolidated. The Cornish granites gave him a temperature of 216° C, and a pressure of 50,000 feet of rock ; those of the Highlands, a mean temperature of 99° C, and a pressure of 76,000 feet of rock. Mr, Wardf arrived at some- what similar results with regard to the Cumberland and Westmore- land granites.

F. Zirkel, however, has shown J that the method adopted by Sorby and Ward is not accurate, since, in contiguous cavities where there is no evidence of leakage, the relative size of the vacuoles varies within wide limits ; whereas, if the vacuole had been due merely to the con- traction of the liquid in cooling, it should be always proportional to the size of the cavity.

Messrs. De la Vall6e Poussin and Renard,§ experimenting with the quartziferous diorite of Quenart, Belgium, obtained, by an entirely different method, a temperature of 307 C, and a pressure of 87 at- mospheres.

Rev. Samuel Houghton has shown that the quartz of granite has a specific gravity of 2.6, identical with that of silica derived from aqueous solution, while the specific gravity of fused silica is only 2.2. It appeared to him that the differences in the specific gravity of naturally and artificially fused rocks greatly strengthen the argu- ment of those chemists and geologists who believe that water has played a much more important part in the formation of granites and

fbid.f vol. xxxi., p. 569, and vol. xxxii , p. 1. t Ifnd.t vol. xxxi., p. 668.

X Mikroscopisehe Beschaffenheil der Minercdien u. OeaUine p. 46. J Memoire sur lea Kocbes dites Plutoniennes de la Belgique,*' Aead- Bff- Belg., 1876, p. 41. II " On the Origin of Granite," Oeol. Soc Dabliny 1862.

Origin Op Gold-Bearing Quartz Of Bendigo Reefs. 743

traps than in that of trachytes basalts and lava and that the for- mer class owe to its agency their high specific gravity. He assumed the metamorphism of rocks to be two-fold: hydro-metamorphism, by which rocks originally fused, and poured, while in liquid fusion, into veins and dikes in pre-existing rocks, are subsequently altered in specific gravity and arrangement of minerals, by the action of water at a temperature which, though still high, would be quite inadequate to fuse the rock ; and pyro-metamorphism, by which rocks, orig- inally stratified by mechanical deposition from water, are subse- quently transformed by heat into what are commonly called meta- morphic ro<!ks. Granite he considered to be, generally, a hydro- metamorphic rock, but, perhaps, occasionally the result of pyro- metamorphic action.

J. J. Harris Teall,* discussing the distribution of fluid inclusions in the different classes of rocks, says :

We are struck by the fact that thej are characteristic of the Plutonic rocks, such as gabbro, diorite, grauite, and crystalline schist. They are rare, or absent, in rocks of the volcanic group. Speaking generally, we may say that fluid and glass inclusions bear a sort of inverse ratio to each other, so far as distribution is con- cerned. When glass inclusions are common, fluid inclusions are either very rare, or altogether absent 'Fluid inclusions are equally abundant in the Plutonic rocks of all ages, and glass inclusions are equally abundant in the volcanic''

Pfafff determined the amount of water mechanically enclosed in the minerals of certain rocks, such as granites and syenites from several localities, lavas from Vesuvius and Etna, and obsidians from Iceland. He found that granite rocks and schist yielded from 0.11 to 1.8 per cent, of water, while the lavas and obsidian gave none — thus confirming the results of microscopic examination.

Apart from precise details of temperature, pressure, etc., Sorby's general conclusions have been accepted by geologists, and the hy- drpthermal theory has been adopted for certain rocks of the granitic class : first, because their constituent minerals have crystallized from the original magma in an order different from their relative fusion- points, the most infusible mineral crystallizing last; secondly, be- cause the quartz is vesicular, and contains abundant water- and gas- inclusions ; thirdly, because they contain minerals that cannot exist at a reii heat, and, therefore, are not the direct products of dry fusion ; fourthly, because the specific gravity o£ fused silica is only

British Petrography,

t Ueber den Gehalt der Gesteine an raechanish eingeschlossenem Wasser und Eochsalz," Poggendorflf's Annalen der Physik und Chemie, Leipzig, vol. 143, p. 610.

744 Obi6In Of Gold-Bearino Quartz Of Bendigo Reefs.

2.2, while the quartz of granite has a gravity of 2.6, identical with that of silica derived from aqueoas solutions.

The thin lava dikes penetrating nearly through 3000 feet of con- torted rock, truly present, as Mr. Rickard says, a marked and in- structive feature of the Bendigo mine-workings. Their origin has naturally been a subject of much conjecture, particularly to those who have not had an opportunity to study the peculiar behavior of ba- saltic dikes traversing sedimentary rocks.

Basaltic lava is an almost perfectly fused rock, which usually reaches the surface of the earth in a state of great liquidity, and at a high temperature* Dana'*' gives the temperature of th9 lava in the crater of Kilauea at 2200 F., within two or three feet of the surface. Davey, experimenting with Vesuvian lava, found that copper wires were readily melted in the flowing stream, at some distance from the crater, indicating, as he says, a temperature of 2426° F. Greikief says, that experiments made at Vesuvius by Scacchi and St. Clair Deville, in 1855, by thrusting thin wires of silver, iron, and copper, into the lava, indicated a temperature of scarcely 700° C. (1292° F.). But he concludes from all the data, that the initial temperature of Vesuvian lava, as it issues, is considerably above 2000° F. Prest- wich;]; says the temperature of lava varies with its composition, but is generally supposed to be from 2500° to 3000° F. Judd§ has managed to write a popular — though none the less instructive — book on volcanoes, without mentioning the temperature of lavas. He, however, deals in such expressions as incandescent liquid rock," molten rock," etc.

The evidence is, I think, fairly convincing, that basic lavas usually have a temperature of, at least, 2000° F. on reaching the surface of a freely flowing crater, and must, necessarily, have a higher tempera- ture deep down in the earth's crust. It seems, therefore, unnecessary to apply a hydrothermal fusion theory to rocks that present such abundant and clear proof of igneous fusion and high temperature.

Scropell advanced, long ago, the view that the fusion of lava is due to the combined action of moisture and heat, but it has not been generally accepted. It is, however, well known that a small amount of absorbed water-vapor, in molten rock magmas, plays an import- ant part in their liquefaction. Iddingsl has shown, that oWdian

Manual of Oeology', p. 743. f Text-book of Otology, 3d ed., 1893, p. 225.

X Chemical and Physieal Otology, p. 203. 2 Voleanoe*.

II QtMrt Jour, Oeol, Soe,, London, vol. xii., p. 326. IT U, S. Oeol. Sur, Bull, No. 66, p. 26.

Origin Of Gold-Beabin6 Quabtz Of Bendigo Beefs. 745

containing but 0.5 per cent, of water is easily melted, and becomes quite liquid, while, after driving off this water, it melts only at an intense white heat to a very viscous glass.

The lava of the Bendigo dikes is, according to Mr. Rickard, iden- tical in lithological character with the basaltic sheets at the surface. Judd says, that, while now much altered, the dike-lava was orig- inally basaltic in character, and contained free crystals of olivine of considerable size. The augite and magnetite are abundant and well preserved.

Mr. Rickard tells us that the lava now filling the Bendigo dikes was originally '' more in the condition of a boiling mud than what we oiinarily imagine as the state of liquid basalt;" and again, that on reaching the surface, it probably welled forth water, steam, and mud, overspreading the older rocks and the later gravel with one of those sheets of basalt which form so marked a feature of the sur- face geology of the Colony of Victoria."

I have already discussed some of the physical characteristics of rocks consolidating from a state of hydrothermal fusion and under great pressure. There is nothing in Mr. Rickard's paper that would indicate any such concomitant conditions for the dike-lava at Ben- digo. In the absence of complete microscopic examinations or chemical analysis of the rocks, both from the dikes and from the surface-flows, we must accept his statement that these rocks are 'identical in lithological character/' without discussion. I will, however, point out that volcanic mud does not consolidate as basalt, but as tuff. On this point, Judd says, volcanic muds have often set in their natural positions, so as to form a rock, which, though light and porous, is of tolerably firm consistency. To this kind of rock, of which Naples and many other cities are built, the name tuff or tufa is applied." Geikief shows that volcanic mud eventually con- solidates into one of numerous forms of tuff. The other authorities take similar views.

Mr. Rickard, however, adopts for these lavas the hydrothermal theory, not because of the physical structure or lithological character of the lava in the dikes, but for the following reasons :

1. The dikes, but a few inches wide, penetrate enormous thick- nesses of rock.

2. The formation of a continuous fissure is necessary for ignus injection, and this is unwarranted by the facts in the case.

♦ Op. eU.y pp. 89, 90. t Op, eit,, p. 135, and dseihfira.

746 Origin Of Gold-Bearing Quartz Op Bendigo Reefb.

3. The force which was able to shoot the lava upward, through the tortuous fractures, would eject it with great violence and to a great height into the air.

4. The dikes do not fill a clean-<;ut, continuous fissure, and rarely preserve a straight line for any great distance.

6, The feeble changes produced by the dike-lava upon rock-sur- faces, in contact, suggest themselves as due rather to the action of water than to the effect of heat.

With regard to these reasons, I would say :

1. In one mine a dike, nine inches wide, has been traced to a depth of 2600 feet ; but this is no proof that it ever reached the tertiary-rock surface, much less, that it helped to extrude any portion of the basaltic sheets. It is not uncommon, in a basaltic country, to find small dikes running for considerable distances and dying out before reaching the surface of the first lava flow ; while the large dikes can be traced, in places, not only to the surface, but sometimes to the sheet or flow which covers them, or which they helped to form. Mr. Judd says, that some of the cracks (now filled with lava) pro<luced in the rocks by the heaving force of im- prisoned vapor, never reach the surface. We have suflScient proof of this in many mining regions. Again, the same author says, that comparatively narrow dykes of the more liquid basalt often " ex- tend to the distance of hundreds of miles from the central vent." It is clear that small basaltic dikes, extending for enormous dis- tances, and through considerable depth of strata, are common oc- currences, and that many small dikes never reach the surface at all.

2. The formation of a continuous fissure is, I believe, a common occurrence.

Sir Charles Lyell,t describing the formation of fissures, etc., on Etna, says, that in the eruption of 1669 a fissure, six feet wide and of unknown depth, opened with a loud noise, and ran in a somewhat tortuous course to within a mile of the summit of the mountain. Its length was twelve miles, and it emitted a vivid light. Two other parallel fissures of considerable length subsequently opened, one after another.

We have many examples of fissuring in recent earthquake phe- nomena. That continuous fissuring has occurred at Bendigo is clearly shown in the monograph J published, since the appearance of

Op. cit., pp. 189, 210.

t Principles of Oeoiogyj 10th ed., p. 21. See also, Geikie, op. eiL,p. 208.

t Dunn's Report on the Bendigo Gold- Fields 1893.

Obioin Op Gold-Bearing Quartz Of Bendigo Reefs. 747

Mr. Eickard's paper, by the Victorian government, in which it is demonstrated that the lava dikes occar along the course of every anticlinal axis.

3. Assuming that the basalt in these small dikes did reach the surface, it does not necessarily follow that it must have been ejected to a great height in the air. But, granting that it was so ejected, what then ? Could that in any way affect the dikes now exposed in the Bendigo mines? Clearly not.

4. The sedimentary rocks at Bendigo are contorted into series of anticlinal and synclinal formation, the anticlinal axes being only 300 to 1300 feet apart. Is it not too much to expect a clean-cut fissure under such conditions? And is it not a physical impossibility to produce such a fissure in folded sedimentaries? Clearly, then, the fissures, such as they are, were determined by the structure of the rocks, and could not have been gouged out by the peculiar action of boiling mud, during a long series of ages, as Mr. Rickard would have us believe.

5. The feeble change produced upon the rock-surfaces in contact, is precisely the phenomenon that is exhibited where small dikes are injected into fissures in sedimentary rocks, while the long contact of boiling mud would necessarily have resulted in the metamorphism of the rocks. Water is the most active agent in all such meta- morphic action ; therefore, mud saturated and surcharged with im- prisoned steam and water would necessarily have greatly altered the rocks in contact with it. This is clearly proved by Mr. Rickard's ob- servation, on page 292, where he shows that granite veins but a few inches in thickness have so altered the beds that '' it is diflScult to recognize which of them were originally slate." Again, he says, page 294: 'The slates are also baked and their ordinary cleavage is largely obliterated." The question at once suggests itself: How can the hydrothermal fusion theory account for the intense action of these granite veins on the country-rock on the one hand, while on the other it is brought forward to prove the feeble action of the dikes?

If it is permissible to give my own experience, I would like to say that while engaged in mining in County Antrim, Ireland, I was for some time much puzzled at the behavior of the dikes encoun- tered in the mines. Some of them had considerably altered and* in- durated the beds traversed, while others had not altered the rock- surfaces in contact to the slightest extent. An extended examina- tion in the mines and glens showed me that it was only the larger

748 Origin Op Gold-Bearing Quartz Op Bendigo Rbef8.

dikes occupying fissures from which some of the outflows of basalt had apparently occurred that had altered the enclosing rock to any extent; the smaller dikes often proved ito be merely injections from the larger ones and did not usually reach the surface, where outflow could have occurred. In the iron-ore measures some dikes did not penetrate the basaltic roof of the ore-veins and these had no effect on the bounding rocks, while others, which penetrated the roof, and in all probability assisted in the extrusion of some of the numerous basaltic sheets that now form the Antrim plateau had a very con- siderable effect on the rocks in juxtaposition.

Disclaiming any attempt to theorize, I will merely say that my observations in Antrim proved to my satisfaction that dikes of injection," unless they are very massive, had little, if any, percepti- ble action on the enclosing rock, while the " through dikes " that reached the surface, and through which a basaltic flow was ex- truded, had usually a marked effect on the rocks in contact ; from which I deduce that the metamorphic action induced by basaltic dikes on the rocks in juxtaposition is very oftii in direct proportion to the mass of lava in the dike, or that may have passed through it.

Maccullochf has shown that in places in western Scotland lai trap dikes traverse sandstone rocks without disturbing or contorting the beds or materially altering the texture of the rock. It has been noticed that granitic intrusions in coal-seams convert the coal into anthracite or graphite, but never into coke,;]: as when in contact with volcanic or trappean rocks. This alone shows the difierent action of these rocks, due, no doubt, to the comparatively dry fu- sion and high temperature of the traps, and the low temperature, with accompanying steam and water, of the granites.

Hardened basaltic lava is a bad conductor of heat. Dana§ says, that at Kilauea, in 1840, lava, sufficiently fluid to clasp and hang in pendant stalactites from branches of trees, had, nevertheless, barely scorched the bark. Again, it is recorded that an outflow of lava from Etna covered a large sheet of snow, and that the heat given off from the lava did not melt it during the space of one hundred years. Thus we see that the action of molten lavas on the earth's surfiuje is much less than is generally supposed. That the same effects occur below ground through the same causes, there is not only no doubt,

♦ P. Argall, The Tertiary Iron-Ores of Antrim."— 5fci. Proe, JUyy, DubL Soc, vol. iii., part iv. t WeaUm leUmcUj London. J Prestwich, Chemical Geology p. 406.

§ Manual of Geology, p. 744.

Orioin Op Gold-Bearing Quartz Op Bendigo Beeps. 749

bat we have abundant proof, direct and to the point ; for it is found that, owing to the variable conductivity of the strata, the amount of water present, etc., the effect of the igneous dikes on adjoining rocks varies considerably. Again, it is found that most dikes have seams of hardened basaltic glass (tachylite) on their sides, which substance is an excellent non-conductor of heat, and may be assumed to have a similar action to that of the skin of slag that forms on the water-jacket of a blast-furnace. So that, with a limited supply of water circulating in the rock, a thin skin of tachylite would permit a long and continued flow of molten basalt through a dike-fissure, without any perceptible effect on the rocks in contact.

Coming again to the Bendigo dikes, we find that Mr. Richard says they are about 1 foot in width ; and we have previously seen that they occur along parallel anticlinal axes averaging about 800 feet apart. Therefore, the mass of the lava is to that of the rock as 1 to 800 — a quantity far too minute to have any material effect on the rocks as a whole. Taking all the facts into consideration, I believe one is warranted in the assumption that the Bendigo dikes are injections of basic rock ; few, if any of which, reached the sur- face or played any part in the formation of these basaltic sheets, which form so marked a feature of the surface geology/' and that a complete examination of rocks both from the dikes and sheets may possibly prove them to be of somewhat different character. Having carefully considered and discussed the evidence and arguments sub- mitted by Mr. Rickard in support of the hydrothermal fusion of the lava in these dikes, I cannot but feel that he has made out a clear case of igneous fusion. That may be, in his view, a catastrophic theory; but, afler all, it would not require much of a catastrophe to account for all the lava dikes in Bendigo.

On the other hand, the authorities attribute explosive, not to say catastrophic, eruptions to the saturation of the lava with vapors and gases. Reyer,* for example, claims that molten rock, highly im- pregnated with vapors and gases, gives rise to fragmentary dis- charges, while, when feebly impregnated, it flows out tranquilly. Geikie takes a similar view, contrasting the quiet outwelling of the biic lavas of Hawaii with the lavas of Vesuvius and most modern volcanoes, which, issuing, " so saturated with vapor as to be nearly concealed from view in a cloud of steam, are accompanied by abundant explosions of fragmentary materials."! Le ConteJ makes the same

Beitrag tur FysUc der Eruptionerij p. 77, quoted by Geikie, op. eiL, p. 222.

t Op, ciU, p. 223. X ElemeniB of Geology, p. 84.

760 Origin Of Gold-Bearixg Quartz Of Bendigo Reefs.

comparison, and says that the quiet type of eruption is characterized by igneous, the explosive type by aqueous fusion, the heat being greater and the amount of water smaller in the former than in the latter. It appears, then that vapor-saturated lavas are intimately associated with explosive and catastrophic eruptions.

Mr. Rickard shows that water cannot remain a liquid, in spite of increasing pressure, at a temperature above that of its critical point. Bischof pointed out in 1839 that the elastic force of steam cannot surpass a certain maximum, which it reaches when its density is equal to that of water. Dr. Andrews, experimenting in this di- rection, placed the critical point* of water at 484° C, the critical pressure at 6481 pounds per square inch, and the critical volume at 2.62 cubic centimeters per gramme. Applying these figures, we find that the maximum effect of saturated steam in a crater would be balanced with a head of lava of about 3000 feet. Superheated steam would, however, have greater power; but here, again, a limit is quickly reached, as when water-vapor is raised to a high tempera- ture, the heat begins to do the work of breaking up the molecules of water into its components, oxygen and hydrogen, thus doing work against chemical attraction,! and storing up potential energy in the separated gases. Geikie, discussing the presence of water-vapor in a column of molten lava, says that it must exist under enormous pressure, even at a white heat, and therefore its component gases may exist dissociated.;];

From which it would appear that if Greikie's view is poshed to a logical conclusion, then neither water nor the vapor of water form the actual agents that extrude lava from volcanoes, but hydrogen, oxygen, and other gases — provided our present knowledge of the physical phenomena involved is correct; for it is manifest that if the water vapor is dissociated, it is the component gases we have to consider, and not the compound from which they were evolved. The temperature at which water is completely dissociated must be less than 3000° C, for we know that at this temperature no union of oxygen and hydrogen is possible. Remsen§ shows that the decomposition of water begins at 1000° C, and is half completed at 2500° C. The dissociation takes place gradually, and its extent depends on the temperature and pressure, in accordance with the kinetic theory of gases.

Recent experimenters place the critical point at a lower temperature.

t See Hugh Robert Mills, Tke Eealms of Nature. t Op. ciL, p. 226.

2 Text-book of Inorganic Chemistry , pp. 61, 444.

Origin Op Gold-Bearing Quartz Of Bendigo Reefs. 751

The theory of the origin of the gold, advanced by Mr. Bickard, 18 (if I understand hira correctly) the precipitation of the gold, not from the ocean, but from that infinitesimal part of the ocean water entangled amid the silt and sand '' deposited on the floor of the Silurian sea. Let us see what this would amount to. The most porous rock does not contain over 25 per cent, of its volume of water. Assuming the Bendigo rocks as laid down in the ocean con- tained 25 per cent, of their present volume as entangled sea water/* and that all the gold in the water, 5 milligrams per ton, or say 26 tons per cubic mile, was precipitated in the strata. This would amount to 6.5 tons of gold in a cubic mile of the now consolidated Silurian rocks.

The part of the Bendigo district containing all the richest mines is about 8 square miles in area. The deepest shaft is but 2800 feet below the present surface, while the mines are not, by any means, explored to this depth, much less exhausted. However, to give a liberal allowance for the ground, the denudation of which yielded the rich store of alluvial gold, we will assume that the gold output of Bendigo has been derived from the exhaustion of 8 cubic miles of strata. On this basis we have 8 X 6.5 52 tons of gold de- rived from the ocean ; whereas the district has already produced over 360 tons— or, say 150 tons more gold than an equal bulk (8 cubic miles) of ocean water would account for. Or, reversing the calculation, 360 6.5 55 cubic miles of strata required to furnish the gold already mined in this district. The widest pos- sible lateral secretion could not be so far-reaching as to account for such unlimited leaching. But until the 47 miles outside of the central area are explored, no one can tell to what extent they are auriferous. They may even be richer than the ground now worked. That Mr. Dunn, the provincial geologist, advises his government not to part with an inch of this ground except for min- ing purposes, is a sufficient proof that he has good reason fbr the belief that it is valuable.

Without going into the subject of the iodide of gold theory of Sonstadt and Sterry Hunt, applied in this case by Mr. Rickard for redissolving the gold precipitated from sea-water, and for collecting it in mineral veins, I will merely quote a statement of Daintree's, published in 1866,* on the precipitation of gold from sea-water, etc., as applied to the gold veins of Victoria. " We must never lose

On the Age and Origin of Oold, Melbourne, 1866.

752 Origin Op Gold-Bearing Quartz Op Bendigo Reefs.

sight of the fact that the first agent raiist have been potent to pre- cipitate gold and silver, and the second to re-dissolve the united pre- cipitate/' This test Mr. Eickard's theory does not bear ; for it does not account for the silver alloyed with the gold in the mineral veins.

The origin of the gold found in mineral veins and placers has ever presented a fascinating study and fruitful field for investigation, alike to the chemist and to the geologist. Believing that a review* of the literature of the subject is not only advisable in discussing the present paper, but may also be useful to members of the Insti- tute for reference, I shall very briefly describe as much of it as I am familiar with.

Bischof early suggested silica as the medium for the transmis- sion of gold to the quartz reef. His researches and experiments with the silicate of gold are well known. Silicate of gold is, how- ever, extremely insoluble in water. On this point, Cosmo New- beryt says that if further experiments prove that alkaline solu- tions favor the solubility of silicate of gold, this silicate theory will be open to but few objections.

Sir Roderick Murchison;]; claimed that gold was of igneous origin, the last-formed of metals, and that it was only found in paying quantities in Paleozoic rocks ; but the discovery that the California gold veins were of Jurassic age disproved Murchison'sassertion, and it is now known thi!f old is found in rocks of every age.

Prof. H. Wurtz§iield that all goU was in solution in a Prozoio thoroughly oxidated ocean, together with per-sulphates of iron. When, however, in this lifeless ocean, organic exbtence commenced, together with the accompanying deoxidizing process, the ferric be- came the ferrous salts, and these being incapable of retaining gold in solution, it was deposited.

J. A. Phillips,!! i° 1868, discussing the chemical geology of the California gold-quartz vein, claimed that the gold may have been deposited from the same solutions which give rise to the formation of the inclosing quartz. He shows that finely divided gold is sol- uble in sesqui-chloride of iron, and more sparingly in sesqui-snl- phate. He suggests that sulphate of iron in a solution carrying gold, was transformed by a reducing agent into pyrites, the gold at the same time being reduced to the metallic state.

Chemical Oeology, f Vorw. Bay, Soc Victoria, vol. ix.

t Siluria, Eissler, Metallurgy of Oold, Proc Eoy. Soe., vol. xyi., p. 294.

Origin Of Gold-Bearing Quartz Op Bendigo Reefs. 753

Coming to Australia, we find that A. R. C. Selwyn* first pro- posed the solution and precipitation of gold to account for the nug- gets in the drift, though at that time the solution of gold by natural agencies was deemed well-nigh impossible.

Henry Rosalesf and, soon after. Belts]! advocated the sublimation theory for the filling of the Australian gold veins.

Daintrce§ sums up his conclusions on the quartz veins of Victo- ria in the following memorable words :

" I had long ago come to the conclusion that most, if not all, the gold in the quartz reefs was derived from the rocks in which these reefi) occur; that the strata themselves received their supply of gold, at the period of their deposition, from the ocean in which they were deposited ; that organic matter and the gases gener- ated therefrom on decomposition, sulphuretted hydrogen, etc., was the cause of the precipitation; and that the amount of metallic deposit was in proportion to the amount of organic matter deposited with the oceanic sediment ; that subsequent plication and desiccation of the sediment caused fissures, into which the mineral waters percolating the boimdary rocks flowed, and were decomposed, and their mineral contents were precipitated, possibly by magnetic currents, thus causing mineral reins."

Daintree cites Sterry Hunt on the reducing power of organic mat- ter in the formation of sulphides, etc. He shows that gold has been in solution in the drift;, and was precipitated with pyrites in tree- stumps. He further states that in operations on the St. Arnaud silver-ores with hyposulphite of soda, he usually found that an ap- preciable amount of gold had been dissolved j'th the silver, thus indicating, as he claimed, that gold may exist aslan ore. It has since been shown that metallic gold is soluble in hyposulphite of soda.

Russellll states that 1000 c.c. of ordinary solution will dissolve 0.0C2 gm. of gold-leaf in 48 hours.

In proof of the important part played by sulphur compounds in the reactions resulting in the deposition of gold, Daintree shows that scarcely ever has pyrites taken from the Silurian slates of Sandhurst failed to yield gold.

WilkinsonTf records a number of experiments showing that gold can be precipitated from chloride solutions, in the presence of or- ganic matter, on iron, copper, arsenopyrite, antimony, galena, etc. He cites Daintree's experiments (quoted by Mr Rickard,. page 312,

♦ Otology and Minemlogy of Ftctorto. t Prize Essay written for the Australian Grovemment,. 186(X X MinercU Fet'fis, London. Op eii,

II C. A. Stetefeldt, The LixivicUum of Sihr-Ores. % Theory and Formation of Ootd NuggeU. VOL. XXII.— 48

64 Origin Of Gold- Bearing Quartz Of Bendigo Reefs.

footnote), and concludes that organic matter is absolutely necessary to precipitate gold on pyrites, etc.

H. A. Thompson* writing in 1868, on the gold veins of Victoria, claims that the gold was mostly derived from the enclosing rock, and the veins were formed by replacement of the country -rock. He further shows, that the deposition of gold in slate is not unoommoD, instancing a very rich deposit at Clunes, where the quartz in the veins was replaced by rich auriferous slate for some distance, the gold being deposited, like fine gilding, on the cleavage-planes of the slate.

Cosmo Newberyt observes, that the existence of gold in the saline waters of the present day has been proved by several analyses, and that Daintree found gold in solution in mine-water. What the gold-salt was, whether chloride, silicate, or sulphide, he had no means of as- certaining, but suggests that it may have been in the same solution that deposited the pyrites, which probably contained its iron in the form of proto-carbonate with sulphates. He shows that chloride of gold, in this connection, can be held in very dilute solutions in the presence of alkaline carbonates and a large excess of carbonic acid, both of which are oommon constituents of mineral waters in Victoria. If the sulphide of gold is required, it is only necessary to charge the solution with an excess of HS, and in this manner both sulphides may be retained in the same solution, depositing gradually with the escape of carbonic acid. He records an interesting series of experi- ments oo the deposition of gold on pyrites, etc., from a solution of its sulphides. In all thse experiments organic matter was necessary, the action ceasing when it was removed, and starting again imme- diately with a fresh addition.

W. Skey ;]; has shown that gold can be reduced from solution wUh- oxd the addition of organic matter. He assumes that a crystal of pyrites in the drift may form, with a speck of gold, a voltaic pair, and thus by oxidation of the pyrites gold would be deposited. In this way, he claimed that 12 pounds weight of pyrites would be suffi- cient to form the" Welcome nugget of 152 pounds avoirdupois. He says that silver is dpositeil with great rapidity and certainty Bolntions which are alkaline from the presence of the fixed

Tk$ GM'fidds mid Mnerid DUirida o/ Vidona by R, Brongh Smith, London

tritero maj fotind

iirodncUon of Gold nnd the Formntion of Nuggeta in the Aariferoita Drifts.'* jr- Roy. Soe Victoria, vol. iic. Qa tbejormation of Gold nggela in Drifts."— IVaw. Nob Z. In&t v- 377. .

Origin Op Gold-Bbarino Quartz Of Bendigo Beefs. 765

alkalies or alkaline earths ; and that as such solations are passed from this condition to an acid one, the silver present in them is re- tained in solution ; any gold, however, that may be mixed with such silver is deposited upon the reducing agent, no matter in which of these conditions the solution is. This alkalinity would especially manifest itself in those reefs which traverse rocks of basic nature, such as diorite or serpentine ; hence, by the way, the large propor- tion of silver alloying the gold found in such reefs, as compared with that alloying the gold found in lodes traversing schist or older for- mations.

Skey has also shown, that sulphuretted hyclrogen attacks gold at ordinary temperatures, forming a sulphide; and we know that all the sulphides of gold are soluble in alkaline sulphides; therefore, as both these agents are generally present in waters situated at some depth in the rocks, we may very reasonably suppose that a portion, if not all, the gold has been brought into solution by these agents.

Cosmo Newbery has conducted a very interesting series of experi- ments on mine-waters, with a view to determine if they carry gold in solution and deposit it where conditions are favorable.

The scale from a boiler using mine-water at Maryborough, after the removal of all soluble matter, showed no free gold under the microscope, but by assay it gave an easily recognizable button of gold. The incrustation consisted of chlorides of sodium, sulphate of magnesia, and sulphate of lime.

Sound mine-timbers from Sandhurst, Ballarat, and other mines, were burned in clean muffles, and the ash, after washing with hydro- chloric acid, did not show gold under the microscope, but on being assayed, gave gold readily. The experiment was repeated with sim- ilar results. A piece of sound wood from a Sandhurst mine 8 pounds in weight, gave 0.002 grains of parted gold ; while some external soft wood gave from 4 pounds' weight O.Ol grain of gold. Pyrites was also found in the wood. A quantity of chips, boiled in water, yielded sulphates and chlorides of the alkalies and of magnesium. These were completely removed by successive waters and the wood was oxidized so as to form sulphuric acid from any pyrites that might be present. The acid and iron were then detected in the so- lution. These tests prove that gold has found its way into the sound, unchanged wood of mine-timbers, and must have got there in solution.

Newbery sums up the, evidence of the presence of gold in solu- tion in Australian telluric waters as follows: (1) Gold found with

756 Origin Of Gold-Beabing Quabtz Of Bendigo Beefs.

pyrites in the wood of the deep leads, first noted bj Daintree ; (2) gold foand in the tranks of trees converted into pyrites in the same leads, and in pyrites cementing together the quartz pebbles of the drift, as well as the pyrites forming stalactitic masses on the lower side of the basalt over the gravel leads ; (3) gold foand in the in- crustation of a steam-boiler using mine- water; (4) gold found in mine-water by Daintree ; (5) gold and pyrites in mine-timbers ; (6) gold in the secondary pyrites coating the joints and filling fissures in the quartz of the lodes.

Ulrieh early pointed out the relation between eruptive dikes and auriferous veins in Australia. He showed that in the Upper Silu- rian the veins increased in richness as they approached the dikes; and he found that the hornblende diorites were favorable for gold.

C. B. P. Searer* states that in Australia it is only where sedi- mentary rocks have been intersected and disrupted by igneolis ones, that veins of auriferous quartz exist to any extent in the former, proving beyond doubt that the near vicinity of igneous rocks is condu- cive to the formation of quartz veins containing gold. He shows that some of the richest mines in New South Wales are in diorite, instan- cing those at Temora. He has but little doubt that all igneous rocks that are imprnated with iron pyrites have more or less gold in them, most particularly those of the diorite and granitic classes. Wilkinson has noticed that in some of the New South Wales gold- fields, homblendic granite and intrusive greenstones or diorites are the original sources of the alluvial gold ; while Daintree and other writers have noticed and described the influence of diorite dikes on the rich reefs of Australia. J. A. Phillips! hs shown that in the colony of Victoria gold is not only found in veins traversing granite and diorite, but is also SQmetimesdisseminated through the rocks them-, selves.

David Forbes came to the conclusion, after seven years' study of South American gold-deposits, that the gold was introduced at two periods : (1st) in granite of an age not earlier than much, if not all, of the Silurian, but probably not later than the Devonian ; (2d), in the eruptive diorite (greenstone) rocks, composed o( hornblende and feldspar without quartz, breaking through strata as late as Oolitic Forbes considered that gold is a normal ccmstituent of granite. He subsequently extended these generalizations of the

Jour. Boy, Soe, N, S, Wales, 1887, vol. xxi.

t Ore Dtpositi, p. 419. t OtoL Magtmne,

Origin Of Gold-Bearing Quartz Op Benbigo Reefs. 767

origin of gold to the other parts of the world — more particularly to its 'association with diorite in Italy , the Urals, California, Aus- tralia, etc.

Prof. J. 8. Newberry,* summing up a review of the origin of gold, says, first, that gold exists in the oldest known rocks, and has been thence distributed through all strata derived from them. Sec- ond, in the metamorphosis of the derived rocks it has been concen- trated into segregated quartz veins by some process not under- stood. Third, it is a constituent of fissure-veins of all geological ages, where it has been deposited from hot chemical solutions which have leached deeply buried rocks of various kinds, gathering from them gold with other metallic minerals.

F. R. Carpenter,t describing the Black Hills ore-beds, states as his belief that large quantities of proto-sulphides of iron were formed through the agency of decaying organic matter as the rocks were laid down ; that this proto-sulphide was not, to any great ex- tent, gold bearing. Subsequently, by the action of ferric salts in solution, this proto-sulphide was changed to bi-sulphide, and the iron solutions, which wrought this change, also brought the gold which is now found in these deposits.

SorbyJ found that the fluid in the cavities of vein quartz often contains a very considerable quantity of chlorides of potassium and sodium, the sulphates of potash, soda and lime, and sometimes free acids, and he clearly gives it as his opinion that every peculiarity in the structure of the quartz of veins can be completely explained by supposing that it was deposited from water holding various salts and acids in solution.

Richard Pearce§, discussing the change from pyrites to limonite, which usually results in a purer gold in the oxidized material, states that he believes the gold is dissolved and is re-precipitated, and that this change can be explained without even taxing one's belief, to any great extent, in the solvent action of certain solutions, said to attack gold. The presence of alkaline chlorides (a natural constituent of most mineral waters), and manganese peroxide, which always occurs in gossans, will be sufficient, with the aid of free sulphuric acid (which is always formed by the oxidation of pyrites), to bring about a change.

School of Mines Quarterly, vol. iii., p. 5, 1881. f xviL, 674.

t Quoted by A. W. Howitt, Tram. Boy. Soe, Via, TroM., xviii., 448.

768 Origin Op Gold-Beabing Quartz Of Bendigo Reefb.

R. C. Hills,* describing the oxidation and impoverishment of sur- face-oreSy and the concentration of the gold in bonanzas, or the en- richment of a lower zone of a vein at the expense of the upper one, shows that pyrite was originally disseminated through the entire mass of vein-quartz, and that iron sulphate, and finally limonite, re- sulted from its oxidation. The latter mineral he found rich in gold along a circulating channel near the center of the vein, at some depth from the surface, usually 260 feet. He infers from the asso- ciation of gold with oxide of iron, and its known behavior with iron salts, that they were in solution together, and were precipitated about the same time. He notes the solubility of gold in ferrous sulphide, and its greater solubility in ferric sulphate, and shows that both salts might be formed during the oxidation of the pyrites, and con- siders that sulphate of iron is the natural solvent of gold.

Franklin Guitermanf, describing the quartzite formation of Battle Mountain, shows that the upper beds of Cambrian quartzite are trav- ersed by a series of small ore-chimneys, or ''pipes," conforming closely to the bedding. These pipes were originally filled with py- rite, but now contain oxidized ores for a considerable distance along their dip — about — into the mountain. The solid pyrites found filling tliese pipes below the zone of oxidation, assays from 1.60 to 1.76 ounces of gold per ton, and contains about 2.6 per cent silica, while the oxidized ore carries, on the average, about 7 ounces gold and 60 ounces of silver to the ton (in the Ground Hog mine). Sul- phates are present in the oxidized ores in appreciable quantities — noticeably, ferric sulphate, which occurs as the crystalline minerals, coquimbite and copiapite. He gives the following analysis as rep- resentative of this oxidized ore :

Per cent.

Hydrated seequi-sulphato of iron, 12.00

Hydrmted sesqui-ozide of iron, 54.30

Silica and alumina, 82.20

Barium sulphate, . . . . ' 2.70

The most remarkable feature of these interesting ore-deposits is, however, the occurrence of nuggets of crystalline gold in troughs in the quartzite floor, imbedded in clay, intimately associated with horn-silver.* With the nuggets, and in their vicinity, are also found

" Ore-Depoeits of Summit District, Bio Grande County, Colorado,'' Proe, Coh- rado SeL Soc., vol. i., p. 20, 1883.

t On the Oold-Deposite in the Quartzite Formation of Battle Mountain,*' Proc Colorado Sci. Soe vol. iii., part iii., 1890.

Obiqin Of Gold-Bearikq Quartz Of Bendigo Beefs. 769

lumps of olay which show high gold-results on assay, but from which it is almost impossible to obtain a color by paoDing.

One of these trough-deposits gave about 14 pounds of gold nuggets, the largest of which weighed 9 ounces :

''This deposit was imbedded in a mass of clay and ore which was carefully taken out, and which weighed 630 pounds. From this ore was obtained, by screening, 182 ounces of coarse gold, the screenings yielding about 426 ounces of gold, and some- thing over 1000 ounces of silver to the ton. The fineness of the nugget gold was about .900.

While it has been assumed by both and Le Conte, that the secondary deposition of gold, in the form of crystals and nuggets, was accomplished through the medium of a solution of per-sulphate of iron, derived by the slow oxidation of iron pyrites, I think I am safe in saying that it has remained for the discoveries in the Ground Hog mine to demonstrate, almost beyond question, the correctness of that theory."

Figs. 1 and 2 are illustrations of a magnificent specimen of crystalline gold from the Oround Hog mine;* it weighs 8.6 ounces, and is beyond doubt, one of the finest gold specimens ever dis- covered. The illustration is about three-fourths the actual size of the crystal.

J. P. Pratt,t epitomizing the results of some experiments, says that gold can be readily oxidized andsalified by ox-acids ; that there exists a liquid and volatile chloride of gold, containing more chlo- rine than the sesqui-chloride ; that there exist, likewise, a sesqui- oxideand a carbonate of gold; and, lastly, , that gold behaves, in many instances, like some of the other metals.

Prof. T. Egleston| conducted an interesting series of experi- ments on the solution and precipitation of gold, and found that gold was soluble in the following liquids: '

Potassium bromide, when heated to 150 to 200' C.

Ammonium nitrate, with ammonium chloride as impurity, at ordinary tem- perature and pressure.

Potassium snlphide, at ordinary temperature and pressure.

Sodium sulphide, u u u

Potassium cyanide, m

Ammonium sulphide, when heated to 145® to 180® C. for honrs.

He concludes : that gold is not only not insoluble, but that, in nature, it is constantly being dissolved out of the rocks and pla-

' This specimen is now in the possession of Messrs. Haberl Bros , jewellers, of Denver, to whom I am indebted for the privilege of being able to examine and photograph it. t Jour, de Pharm. et de Chimie, August, 1870. t Tram , ix., 639.

760 ORIGIl OF GOLD-BEARING QUARTZ OF BEKDIGO REEFS.

cers ; the waters of filtration dissolving oat of the rocks, in their passage through them, all the materials necessary for the solution of the gold, and carrying it, in very dilute solutions, until it meets some substance that precipitates it."

Fig. 1. Fio. 2

Crystalline Gold from the Ground Hog Mine, Colorado.

E. A. Schneider* mentions the fact that L. Hoffman and O. Kruss Imve definitely established the proportion in which gold and sulphur combine. They found that gold can form with sulphur only aurous sulphide (AusS) and auro-auric sulphide (AuSi), but not auric sul- phide (AujSj). Schneider conducted some very interesting experi- ments with colloidal aurous sulphide, and concludes :

" The separation of free gold in the upper strata of the earth's crust may have heen effected by the action of sulphuretted hydrogen on chloride of gold — for the

Bulletin No. 90, TJ. S. Oeol, Surveff, p. 56.

Origin Of Gold-Bearino Quartz Of Bendigo Reefs. 761

constant associate of gold (pyrites) owes its formation, as Doelter (Zeittchr.f KryU, ti. Min.f vol. zi., p. 30) has shown in his investigation on the subject, to the action of sulpharetted-hydrogen water on the various oxides of iron at temperatures be- low 100* C."

J. F. Kemp* mentions that, in Chili Dr. Mdrickef has recently found native gold in pearlstone (obsidian) from Guanaeo, in skeleton crystals in the glass, as inclusions in perfectly fresh plagioclase and sanidine crystals, and in spherulites. Kemp very forcibly says that the metallic contents of the minerals which constitute ores, must logically be referred to a source either in the igneous rocks or in the ocean.

The precious metals have been found in the igneous rocks of the United States in many places, as, for example, gold and silver in the diabase near the Comstock lode in Nevada, and silver in the Lead- ville porphyry.

A recent writer in Science, on the origin of gold, has shown that in South India, where the distribution of gold has been carefully worked out, it is found that the country-rock is gneissic and gran- itoid rock, covered by a number of bands of schist, lava-flows, hematite beds and conglomerates, which belong to a system which is distinct from, and newer than, the gneiss, and which is called Dharwar. The gold is found in quartz veins, which are only au- riferous in the Dharwar bands. It was therefore concluded that the gold which is now found in the auriferous reefs of Southern India was derived from the rocks of the Dharwar system, and that it was originally brought up from the depths of the earth by lava-flows, which form so large a part of that system.

Finally, certain allotropic forms of gold may be soluble in water. Prof. Roberts-Austen gives the following extremely interesting ex- ample :

" If a fragment of an alloy of potassium and gold, containing about 10 per cent of the precious metal, be thrown upon water, the potassium takes fire, decomposing the water, and the gold is released as a black powder ; this black or dark-brown gold is an allotropic modification of gold, as there is evidence that it combines with water to form auric hydride.''

In summing up this scanty review of the origin of gold, we find : 1st. That Australian geologists have long held that part of the

♦ Ore Deposits of the United States, p. 25.

t Tschermak's Min. u Petr, Mitth,, zii., p. 195.

X Introduction to the Study of MetaUurgy London, 1S92, p. 74.

762 Origin Of Gold-Beabinq Quabtz Of Bendigo Beefb.

gold found in the reefs was derived from the ocean, and was depos- ited with the strata now enclosing the veins.

2d. That gold was also derived from the eruptives, particalarly the diorite (greenstones).

3d. That gold, throughout the world, is of somewhat similar origin, and that it is, in all probability, mostly derived from igneous rocks.

4th. That gold readily enters into solution, and that it is so found in mine-waters, impregnating mine-timbers in association with the alkalies and alkaline earths combined with sulphur.

6th.' That these mine-waters chiefly consist of the alkaline chlor- ides and sulphates.

6th. That waters holding these salts in solution, are found in the inclusions of vein-quartz carrying the precious metals.

7th. That, as the natural waters now circulating in mineral veins are, under certain conditions, capable of dissolving the precious metals and holding them in solution, there is no necessity for a hypothetical solvent, consisting of a comparatively rare element combined with the largest possible amount of conjecture.

Mr. Rickard says that the occurrence of certain graphitic slate in lodes has been pointed out by others,* but the frequent association with gold quartz has not been fairly recognized. If he will take the trouble to look the matter up, he will find that it has been fairly recognized ; and, as I believe, too much has been made of it — par- ticularly if we consider the small amount of organic matter neces- sary to precipitate all the gold in the richest vein we know of.

In 1883, William Nicholas read a paper before the Royal Society of Victoria,t in which he says : Being deeply impressed with the idea that the laminated quartz reefs are, for the most part, simply transmutated slate beds, I looked for an explanation of the presence of gold in these reefs.'' He cites Newbery's experiments of precip- itating gold with organic matter, etc., and declares that the mine- water was the source of the gold in the quartz reefs of Bendigo. He further says that the organic matter in the soft, black, carbon- aceous slate, when the slates were decomposing and changing into quartz, became both the precipitant and nucleus for the extraction of the gold from the mineral waters that then circulated in the reef."

Nicholas made solutions of chloride of gold in water taken from

Instancing Sandberger. f Origm amd Ormrik of Odd, etc

Origin Of Gold-Bearing Quartz Of Bekdigo Reefs. 763

the Bendigo mines ; introduoed pieces of black slate, taken from the Bendigo reefs, into these solutions, and received on them precipitates of gold. He believed that the evidence he produced warranted the conclusion that the auriferous laminated quartz reefs are simply altered auriferous carbonaceous slates, and that, in the change, quartz and gold were formed by one operation in nature.

In discussing the causes of precipitation, Mr. Rickard remarks that the circulating solutions would meet with portions of loose or crushed rock, giving larger space, increased surface, and diminished pressure. It is difficult to conceive that water, circulating in sat- urated rocks, can have its pressure diminished by traversing a cavity, however large it may be.

Mr. Rickard (communication to the Secretary) : The interesting remarks contributed by Mr. Richard Pearce do not call for any lengthy comment on my part. That gold and silver, being known to exist in sea- water, should not be found in the bodies of rock salt due to the evaporation of such sea- water is a fact which it is indeed difficult to explain. It is, however, possible that the conditions which compelled the deposition of the salt were not such as at that time favored the precipitation of the gold. It is interesting in this connection to point out that a noteworthy part of the iodine of com- merce is derived from the nitrate beds of Chili. Whether these deposits contain any traces of the precious metals I do not know.

Mr. Argall disagrees with many of the explanations contained in my paper. In ransacking the literature of the subject he has given us a summary of the views of various writers, some authoritative, others discredited, which will be convenient for reference in the future, yet I cannot but express my regret that he did not give more extracts from the fresh leaves of Nature's open book and fewer quo- tations from the musty shelves of the library.

That '' argument in support of the generally accepted doctrine of the hydrothermal fusion of certain granites is unnecessary at this date " is a matter of opinion. It has appeared to me that it is not so generally accepted" as is suggested, that it is very frequently overlooked by writers on ore-deposition, however well recognized by petrographers, and that its proper recognition is very essential to the intelligent study of the subject which I was discussing — the origin of the gold and quartz in the reefs of Bendigo.

Mr. Argall next objects to the suggestion that if the Bendigo granite had ever had a vesicular structure it must have become ob- literated by the physical and chemical changes to which all rocks

764 Origin Of Oold-B£Arinq Quaettz Of Bbndigo Beefs.

are subject daring those enormous periods of time with which ge- ology has to deal ; and in his comment he gives the idea that this was the only explanation offered for the absence of a vesicular structure. He states that I said that '' the granite does not now ex- hibit a vesicular structure like that of the lavas/' which I did not say, because it suggests a comparison with another rock, which was beside the question. In proceeding to correct a statement of his own creation, he offers some remarks as to the consolidation of the gran- ite under great pressure, neglecting to note that this particular point had not been overlooked by me, as he suests, but particularly emphasized.* This kind of criticism is neither instructive nor de- structive, but purely and simply obstructive.

The origin of the thin dikes of lava which are so characteristic of the Bendigo mines is, I confess, an extremely difficult subject It has long seemed to me, however, as the result of much time spent in the mines, that the ordinarily accepted views on the subject of dike-formation are out of accord with underground observation. That any suggested modifications of the theories now current would be readily accepted was scarcely to be expected, and many of the ob- jections raised by Mr. Argall will be of service in promoting further inquiry in this direction. I cannot but feel, however, that my critic, in his desire to uphold those accepted views which seemed to me to be fairly isharacterized by the term " catastrophic," has given utterance to objections largely founded upon a hypercritical verbal commentary of my paper.

When I spoke of the lava being " more in a condition of a boil- ing mud than what we ordinarily imagine as the state of a liquid basalt," I desired to draw the distinction between the '' ordinary " idea of a fluidity due entirely to a molten condition arising from the action of intense heat, and a mobility largely due to the presence of particles of superheated steam. The homely illustration left room for verbal criticism which would not have been suggested if I had spoken of " a viscid material in a state of incipient ebullition." At Kilauea to-day the basalt owes much of its mobility to the steam which it contains.

What 1 actually said was as follows: The graDite, it is true, does not now ex- hibit a vesicular structure such as would ordinarily characterize a rock whoee mass had been interpenetrated by steam ; but it must be remembered that it cooled Qn<! der the pressure of a great thickness of overlying formations and was extruded at so early a period in geological time that any such structure would long ago have been obliterated by those infinitely slow physical and chemical changes which are for- ever bringing abou the decay and repair' of rocks." — Ante page 295.

Origin Op Gold-Bearing Quartz Op Bendigo Reefs. 765

Mr. Argall takes exception to every one of the reasons mentioned by me as having suggested the explanation which I offered. I will consider his criticisms seriatim.

1. That the dike in the 180 mine did not reach the surface is possible, but the burden of proof must come from the other side; since it is rendered very probable by the fact that in 'the adjacent mining districts the dikes are known to have reached the surface and to have overspread the Silurian rocks. It was scarcely nec- essary to quote from geological text-books to prove that dikes occa- sionally do fail to reach the surface. This has been shown to be the case in the particular district under discussion by the under- ground drawings which I have myself contributed.*

2. Here again we have a reference to the library without proper reference to the facts of the case. The careful examination of the behavior of the lava-dikes and of the irregular fractures which they occupy convinced me that in this case the formation of an instantane- ous and continuous fissure was untenable. It is useless to instance the formation of a fissure at the surfaccy in Sicily, against facts ob- served far underground, in Australia. The conditions are entirely dissimilar.

I must differ with Mr. Argall when he claims that Mr. Dunn's monograph contains proof that ''continuous Assuring" has taken place at Bendigo. That the dikes occur along most, but not all, of the anticlinal axes of the folds of the slate and sandstone rocks may be true; but why should this prove that the fissures which they occupy were continuously and instantaneously formed ? My under- ground drawings, which I had hoped would be more useful than much writing, must testify in my behalf. See Figs, on page 297 of the present volume; also pages 699 and 701 of vol. xxi., and pages 485, 487, 489, 497, 511, 513, 516 and 5J7 of vol. xx. of the 7ran8acti<m8.

3. This is trifling and unnecessary.

4. Here Mr. Argall is contradictory of himself. In a previous paragraph he has been arguing for a '' continuous " fissure, but now he argues against a ''clean-cut'' fracture. The two do not seem to me to give expression to ideas which are opposed to each other. It must be remembered in this case that the dikes in their general direction do not follow the bedding-planes of the Silurian rocks, but penetrate through them. Conditions, such as the structure of the rocks, which tend to prevent the formation of a clean-cut fracture

See Fig. 13 on page 297.

766 Origin Of Gold-Be A Rinq Quartz Of Bendigo Reefs.

will also tend to prevent it from being oontinned veiy far at any given moment. The irregularity of fracture indicates a resistance to flssuring greater than is the case when the fissure follows some distinct structural line of weakness. Mr. Argall now argues that a clean-cut fissure is '' a physical impossibility " in folded sediment- aries. I agree with him that the character of the fissuring is depen- dent upon the structure of the rocks ; and therefore it seems to me that evidence gathered on the slopes of Mt. Etna is hardly pertinent in this case. But he is in error when he charges me with asking my readers to believe that the fissure could have been gouged out by the peculiar action of boiling mud." Reference to the wording of my explanation* will show how unwarranted is the peculiar phrase he has employed in this ludicrous misrepresentation.

5. can the hydrothermal fusion- theory account for the intense action of these granite veins on the country-rock on the one handy while on the other it is brought forward to prove the feeble action of the dikes?"

The metamorphism of the Silurian rocks observed near the granite contact at Big Hill differs entirely from the feeble changes caused by the lava-dikes, because the effects produced in the former case were due to the influence of an enormous mass of intrusive rock (now forming the Mt. Alexander range), while the absence of any noteworthy changes in the latter case can be explained by the small bulk of the lava-dikes which, in comparison, are only thin threads. In the former case there was a store of heat close at hand ; in the latter the main body of lava, of which the dikes were minute branches, was at a depth so great as not to be able to causcany metamorphism of those rocks into which the mine-workings have penetrated.

The observations contributed by Mr. Argall from his Irish ex- perience are to the point and very interesting. It is curious lo note, however, that in County Antrim, Ireland, the dikes " in all prob- ability assisted in the extrusion (>f some of the numerous basaltic sheets that now form the Antrim plateau," but that at Bendigo, Australia, it is in his opinion highly improbable that the dikes ever reached the surface, much less, that they helped to extrude any portion of the basaltic sheets."

That large dikes have an effect on rocks in contact, such as is not observable in the case of small dikes, is a fact which on various oc- casions I have myself noted ; but the force of the distinction depends

See page 299, paragraph beginning, " The course of the dikes and the effect npon the lava," etc.

ORIGIN OP dOLD-BEARING QUARTZ OP BENDIGO REEFS. 767

upon the meaning given to the adjectives employed. Mr. Argall does not mention the size of the dikes in Antrim ; but, referring to the article from which he quotes, I find that they are stated to be from a few inches to several fathoms in thickness."* At Bendigo I saw none larger than about 2 feet, but many less than 1 foot thick. Mr. Dunn mentions one as large as 5 feet occurring in the Great Britain mine.;|l

That basaltic lava is a bad conductor of heat, and that dikes often have an outside shell of less crystalline rock due to more rapid cool- ing of their exterior, are facts which do not admit of dippute. But a lava which cools rapidly at the surface and in the open air on the slopes of Kilauea does so under conditions entirely dissimilar to those obtaining where another lava cools slowly at a great depth and under enormous pressure. In this case, as in many of the observations quoted by Mr. Argall, I deny the correctness of comparing the con- ditions obtaining at or near the surface with those which physical facts warrant us in believing to exist at great depths.

The comparison made by Mr. Argall between the volume of the lava in the dikes and that of the rocks which they penetrate is not pertinent; since I have not suggested that the dikes should ''have any material effect on the rocks as a whole.'' I did sup[)08e that the dikes, if injected in the form of a white-hot molten material through the slates and sandstones would produce some marked changes along the planes of contact ; but I did not expect any regional metamor- phism from so small a mass of material.

Mr. Argall quotes several distinguished authorities to prove the occurrence of " explosive eruptions due to the saturation of the lava with vapors and gases.'' In his argument against my proposition that the lava of the Bendigo dikes owed its mobility largely to the presence of superheated steam, and that it found its way to the sur- face slowly and gradually through the manifold fractures produced by the strain to which the rocks were being subjected, not for a few seconds but for a long period of time, he has fallen back on such observations as he could find in the text-books describing the sur- face and open-air phenomena of volcanic activity, while I have re- lied upon what I have seen underground in the mines and upon the conviction thereby induced that surface and deep underground action are very dissimilar in kind and in character.

Notes on the Tertiary Iron-Ore Measures, Glengariff Vallej," Seieniifie Pr(h ceedingi JRcyal Dublin Society April, 1881, voL iii., p. 11. I Special Eepori on the Bendigo Oold-Field, 1893, p. 14.

768 Origin Of Gold-Bearing Quartz Of Bemdigo Reefs.

The study of volcanio action at Eilanea and Mauna Loa, by Dana, and at Vesuvius and Stromboli, by Judd, has emphasized the fact that the projectile and more violent results are due to the force ot vapors expanding by reason of diminished pressure as they reach the surface. In his interesting book,* one of the most recent contribu- tions to volcanology. Prof. Dana frequently recognizes the distinction between the violent surface-outbreaks and the slow action which goes on at great depths. He speaks f of " the ordinary work " of Mauna Loa and Kilauea as being shown to be carried on by : '' (1) The ascensive force of the conduit lavas ; producing (a) a slow rise in the liquid rock from depths below ; and (b) a raising of the crater's bottom. (2) The elastic force of rising, expanding and escaping vapors." Elsewhere he observes that there is no question that the chief working vapor is the vapor of water.'* The violent eruptions of the volcano are due to different agencies. " The chief cause is no doubt the elastic force of suddenly generated vapor."

Prof. Judd's testimony is to the same effect I may quote from his description of Stromboli where § he describes the lava rising in the crater in the following words: ''The fluid mass in this way appears to be gradually impelled upwards till it approaches the lips of the aperture, when vast bubbles are formed upon its surface, and to the sudden bursting of these the phenomena of the eruption are due.''

The recorded observations of volcanic action .appear to agree on this point, that it is tl e sudden relief from pr6sure, brought about in various ways, but always comparatively close to the surface, which causes the violence of eruptions ; and the inference is suggested that at greater depths the conditions are such as to compel a more quiet and orderly behavior.

Since the date when my paper was prepared I have been put in possession of certain information on the subject of the lithological character of the Bendigo lava which, while it in some ways weakens the hypothesis suggested by me, I beg leave to contribute now.

My friend, Prof. George Ulricb,|| disagrees with much that I have written. Writing from New Zealand he says:

As to the hydrothermal origin of the basalt dikes, that the rock was injected into the fissures as a watery mud-lava, this is a hypothesis which I cannot bring into agreement with my own observations. Before leaving Melbourne, I made for

Charaderisliet VckanoeSf James D. Dana, 1890.

Volcanoes, page 19. In a letter dated Dnnedin, April 15, 1893.

Origin Of Gold-Bearikg Quartz Of Bendigo Reefs. 769

the Melbourne Museum a number of thin sections of the rocks of several of the Bendigo dikes, and found them throughout (in agreement with Prof. Judd) of basal- tic character. Although more or less affected bj decomposition, they agree well with samples of dense compact basalts occurring amongst the flows of the Kyneton, Ballarat and other districts, but most closely with the quickly-cooled side-or-wall- parts of some larger basalt dikes (4 to 8 feet wide) which run out from points of eruption in the Castlemaine and Yandoit districts. These dikes are dense, nearly glassy, at the sides, but become granular-crystal line towards the center. The Ben- digodike-rock consisted, in my opinion, originally of a great preponderance of glassy matter (now devitrified and altered) with hardly any feldspar but ferro-magnesian silicates (olivine, augite) and magnetite in some abundance distributed through it (the magma-basalt of Boricky), the glassy condition pointing to rapid cooling, as in the case of the wall-parts of the before- mentioned larger dikes. The rock had no resemblance whatever to a mud-lava, such as I have seen here in New Zealand ; and, reasoning backwards, as we have basalts in other places in Victoria which it closely resembles, and these have doubtless been ejected in a fiery-fluid state, so the Bendigo rock must have been the same. I can see no difficulty in the proposition that enormous intra-telluric pressure fissured the Silurian rocks and nearly simulta- neously forced or squeezed the thin fluid molten basalt matter into the fissures, where it rapidly cooled to a partly glassy state. It was not a gradual, but as I pic- ture to myself, a sudden mighty efibrt of pressuie from below, capable of squeezing the fluid-matter, as it were, instantly miles up into the narrow fissures, without giv- ing it any time for cooling during the passage. Once having filled the fissures, it cooled, of course, rapidly, and owing to its small bulk — the thinness of the dikes — could produce hardly any effect upon the bounding slate and sandstone walls. In the walls (also slate and sandstone) of the larger basalt dikes before mentioned, I could not see any marked change compared to the rock farther ofi*. At one place I thought the slate was more siliceous and somewhat harder; but then at another place I found it softer and clayey ; so I concluded that if some slight change at first existed it must have been obliterated in course of time by chemical influences.''

Prof. Le Conte, writing from the University of California,* says that my view is tenable with some modifications, but he points out several diflBculties which appear to him to stand in the way of its adoption. The character of the rock, "an altered basalt," appears to him incompatible with true hydrothermal fusion, though it doubt- less contained water. As to the force required to inject fluid rock through several thousand feet of stratified rocks, he considers that it is overestimated, and that only hydrostatic pressure is necessary, the strata being supposed to rest on liquid matter. The fissuring of the rocks would cause " the solid strata to sink into the liquid and the liquid to rise into the fissures in proportion as they were formed, like water rising into the cracks of ice resting on its sur- face." Prof. Le Conte also points out that both explanations, the accepted ordinary one and that advocated in my paper, are open to grave objections.

Dated Berkeley, October 6, 1893. VOL. XXII.— 49

p

770 Origin Of Oold-Beariko Quartz Of Bendigo Reefs.

" It seems impossible that so slender and irrelar Bssnres should have been filled (whether rapidly or slowly) unless the liquid was super-fused, for otherwise it would have cooled below fusing-point by contact with the walls. If it was igneous super- fusion, such intense heat ought to have had more effect on the wall-rock. If it was aqueo-igneons super-fusion, then a far more coarsely crystalline rock ought to have been formed."

After my paper was in print, I received, through the courtesy of the author, the Special Report on the Bendigo Grold-field, prepared by Mr. E. J. Dunn, and published by the Mines Department at Melbourne. Needless to say, it is a most valuable publication, and throws further light on the geological structure of a most interesting region. Mr. Dunn has not mentioned any facts contradictory to those which I placed before the Institute ; but in his explanations of the mode of formation of the dikes he is orthodox, and uses the theories which are to be found in most of the text-books. He says :

" The injection through thousands of feet of such thin dikes as penetrated the strata could only have been accomplished by forces which caused the molten rock to fill the fissures instantaneously."*

He speaks of the dike-rock as being " doleritic in character, so that the miners are not very far out in calling it ' lava.' "

Still more recently, Mr. A. W. Howittf has issued a special re- port descriptive of the rocks found in the " 180 " mine.§ Of the rock forming the dikes, he gives the following interesting descrip- tion:

"An inspection of the lavas in situ leads to the belief that they are dikes of a

very basic igneous rock, which, at some period after the formation of the quartz

reefs, were injected into fissures, and have there consolidated. Little is to be learned

from injMcUn to their composition, even with the aid of the pocket-lens. They

are very geiicj.iHj much decomposed, and it was only after protracted search that

Mr. Dunn wa -Me to find samples which were sufficiently fresh to afford good ma-

(ertal for hiilto topical examination, and for the determination of the mineral oon-

tituentA wU\ch the nomenclature of the rock must be based. The first fact

bec< evident on inspection under the microscope is the absence of -

um; the next, the preponderance of a form of monoclinic pyroxene with the

irer occurreneG of olivine, all set in a more or less plentiful glassy ground-mass.

♦ Eeil on tJi Bendigo Gold-field, by E. J. Dunn, F.G.S., Melbourne, 1893, p. 14. f The SeereJary for Mines to the Government of Victoria. X Copiefl by (lie Bendigo Advertiser of November 3, 1893, where I read it. 1 It iaia the 180 mine that the dike occurs, which has been traced for over 2600

Origin Of Gold-Bearing Quartz Of Bendigo Reefs. 771

The following particulars will make these general observations clearer : The glass which forms the latest consolidation is either almost colorless or is, in some sam- ples, pale-brown in tint In this are numerous micro! iths and skeleton crystals of amphibole, usually in parallel or stellate arrangement around grains of magnetite. There are also larger amphibole microliths, which are pleochroic in shades of green. Associated with these are somewhat large| but still minute prisms, cross- sections of which show the characteristic prismatic cleavage angle of about 124°. The marked feature of the ground-mass is the very plentiful occurrence of augite as the mineral generation next preceding the microliths and following the por- phyrite crystals of augite and olivine, which possibly are the earliest formed min- erals excepting the magnetite. As already stated, there are no feldspars, and the augite forms a loose network analogous to the mesh of triclinic feldspars common in basaltic rocks. The prismatic cleavage is more or less well marked, and there is cross-separation in the prisms, which are much extended in the direction of the 0 axis. In crystals, which are faintly tinted with shades of purple, there is a slight pleochroism. The intergrowth of amphibole with augite is common by the at- tachment of narrow of the former. I have observed in some cases a pecu- liar hourglass" form of amphibole, longitudinal and cross-sections. In the ground-mass thus constituted, there are colorless crystals of olivine more or less imperfect, fractured, or rounded. Alterations can be traced in various examples from a mere lining of the fissures to complete conversion into serpentine. Oc- currences which I can best describe as vacuities occur in this rock, and are far more numerous us the rock is more altered. They are irregular in form, rarely indicat- ing any definite crystalline outlines. Where the glassy basis ends, there commences usually a layer of carbonates in grains, or in more or less well-marked scalenohedra. Planted on this layer, and in most cases entirely filling the remainder of the vacuity, is a fibrous, usually sphaerolitic mineral, in which, with crossed nicols, the char- acteristic black cross is present. The component fibres are always more or less white in color, and, so far as I could observe in such minute individuals, extin- guish parallel to the prismatic axis. The mineral is undoubtedly a zeolite, having a positive character. In the absence of micro-chemical tests, it is not possible to speak with certainty, but I am inclined to refer it to a seolite of the mesotype group. The composition of these so-called ' lavas' is a more or less plentiful glassy basis with crystals of augite and olivine and subordinate amphibole. Very excep- tionally there appear abnormal occurrences of minerals, as, for instance, biotite. Veins of quartz have been mentioned to me by Mr. Dunn as occurring in some of these dikes. An igneous rock of such a composition would be named, according to the classification which is now generally accepted, limburgite. This somewhat. ' rare species was first established as a distinct rock by Rosenbusch, and was de- scribed and named by him from its occurrence near Heidelberg in 1872. Limbur- gite has been recorded subsequently from other localities, as, for instance, by v. Werveke, as a vesicular lava in the Island of Palma; by Blum, in 1881, from Foya, and elsewhere by other observers. According to Kalkowski, limburgite is merely a form of basalt in which augite, olivine, and magnetite preponderate over or ex- clude feldspars. The extreme freshness of some samples of this rock from Bendigo suggests that it is of comparatively late geological age. Up to the present time, however, no occurrence of limburgite has been recorded as among the sub-aerial volcanic rocks of Victoria. It may be that the limburgite of Bendigo is an extreme form of the Tertiary feldspar basalts which are so common here; but I am not ac- quainted with any analyses of sub-srial basalts in Victoria of which this limburgite might be (nsidered as the subterranean representative."

772 Origin Of Gold-Bearing Quartz Op Bendigo Beefs.

It will have been noticed that both Prof. Ulrich and Prof. Le Conte consider it very probable that the molten lava was injected by a mighty eflTort of nature through the fissures penetrating thou- sands of feet of rock, and the force which accomplished the work of propelling the lava is supposed to have been hydrostatic pressure. Mr. Dunn's views are very similar: he considers it "certain they (the dikes) have filled up fissures along the lines of least resist- ance, and where some convulsions have fractured the rocky mass right through/'*

These explanations are based upon the old theory that the earth has a solid rock-crust, lying upon liquid matter, and that the breaking of the crust causes the liquid underneath to run up into the cracks, " like water rising into the cracks of ice resting on its surface." The investigations of astronomers, physicists and certain geologists during recent years have, I believe, led to very serious modifications of the old theories, and the " term crust of the globe ' is employed by geologists as a convenient way of referring to that portion of the earth which is accessible to their observation,"t and not as referring to a thin shell coating a molten interior. In reviewing the argu- ments. Prof. Judd says : That the earth is in a solid condition to a great depth from the surface, and possibly quite to the center, is a conclusion concerning which there can be little doubt."J Other authorities who have given their attention to this subject of late years are in accord with tliis view.§ The comparatively recent in- vestigations of Prof. Dana in the Hawaiian islands point to the con- clusion that hydrostatic pressure may play some slight part in the erupting action of Kilauea and Mauna Loa, but there " is evidence that pressure from this source is the least efficient agent'* — the chief cause being, "no doubt, the elastic force of suddenly generated

" This is a question, howeveryto which only brief allusion can be- made at the present time. It still offers a magnificent field for in- quiry and thought.

Coming to the theories regarding the origin of the gold in the reefs, Mr. Argall founds his first argument, condemnatory of my suggested explanation, upon the fact that the Bendigo district has

Report on Bendigo Qold-Field p. 14.

t Judd, in Volcanoes, p. 308. t 327.

§ The question is most ably discussed in the Rev. Osmond Fisher's iyncs cf ike EariK% Crust. II Characteristics of Volcanoes, p. 235.

Origin Op Gold-Bearing Quartz Op Bendigo Reep8. 773

produced 360 tons of gold, while a calenlation, on the basis of cer- tain figures mentioned in my paper, would only account for 52 tons. I quit believe, with Mr. Dunn, that only a small part of the gold- bearing portion of the district has as yet been explored — not to say worked out — so that if my theory of gold-deposition from sea-water is to be put to such a test as Mr. Argall suggests, then the figures may be increased from 360 to 3600 tons, with every probability of being underestimated. But such a test is a delusion. Sterry Hunt has pointed out that the ancient seas probably contained more mate- rial in solution than is the case now. It is very probable, for in- stance, that the waters of the ocean were warmer in the Silurian age than at the present day, and that they therefore exercised a more powerful solvent action. It is also possible that certain precipitants were less abundant. However, apart from any such j>urely theoret- ical consideration, it seems to me a queer mode of reasoning on such a subject as ore-deposition to assume that the gold of the reefs trav- ersing the Bendigo rocks must have necessarily come from the par- ticular series of beds enclosed within the limits of the territory pene- trated by the existing mine-workings. Is not this taking a very narrow view of the formation of ore-deposits ?

As indicated in my paper,* I believe that the gold was derived from the mass of the surrounding formation, and more particularly from that portion which is beneath the area now being mined, be- cause in depth there is obtained a maximum of those conditions (increased pressure and heightened temperature) which favor solu- tion. The geological examination of the region f and the study of the changes in the microscopic structure of the rocks X have indi- cated that the sedimentary formation which encloses the quartz reefs extends to a depth far exceeding the horizon to be reached by the most profound of future mine-workings.

The lengthy series of quotations given by Mr. Argall does not call for special aimment. In my paper I made a brief reference to such of them as seemed important. It is a waste of time to quote exploded authorities like Roderick Murchison § or to refer to the buried imaginations of Rosales.|| It is to be regretted that Mr.

Page 317. f Bj Mr. Dunn and by the writer. J Bj Mr. Howitt.

i On the question of gold-occurrence Murchison was guilty of hasty generaliza- tions from insufficient facts, as, for instance, when he stated that gold mining in the solid rock, as distinguished from gravel, could not be profitable. — Silurian p. 496.

II The essay of Rosales is very much out of date. It is full of bad errors of fact, and supposes the formation of veins to be due to molten silica,'' which was spread "

774 Origin Of Gold-Bearing Quartz Of Bendigo Beefb.

Argall, quoting at seeond-baDcl from the compilatioD of Lock, has perpetuated in many instances the annoying and unscholarly vague- ness of Lock's references. His implied admission, in a single foot- note on page 754, that be has never personally examined the Austra- lian authorities he quotes, leaves us no guaranty of accuracy except the reputation of Lock. And when Mr. Argall, on the strength of a mere scrap-book like Lock's huge volume, undertakes to give the views of different authors, it seems as if the force of repeated com- pilation could no further go.

Mr. Argall gives me credit for quoting Daintree, which, thanks to Prof. Ulrich and Mr. Pearce, I was able to do in the words of eye-witnesses of his experiments ; but he overlooks the fact that I also made such references as the exigencies of space permitted to the work done by Wilkinson, Newberry, Ulrich, Le Conte, Pearce, Gui- terroan and others." It would be unnecessary to emphasize this matter were it not that my critic seems to suggest that I had attacked a very difficult subject without paying proper tribute to those who had already written upon it. Such was far from my intention, and such also, I trust, was not the fact.

As to the occurrence of graphitic slate in association with gold- bearing quartz, Mr. Argall's suggestion that I had not taken the trouble to look the matter up,'* is a matter which I leave to the readers of my paper to decide. The review of the literature of the subject of gold-deposition to which Mr. Argall has treated us has led him to conclusions very dissimilar to those which I adopted in the paper under criticism. His conclusions would form the text of a very interesting debate. In closing, I may express the hope that the present discussion will lead to the carrying out of new experi- ments and to the accumulation of a fresh stock of carefully-recorded observations. It is in that way, and that way only, that progress lies.

and "pressed" through the rocks and formed "extensive sheets" between them; and BO on ad nauseam, See pp. 208 to 316.

Index.

[Note. — In this Index authors' names are printed in small capitals and the titles of papers in italics. Casual referencesi giving bat little information, are usually indicated by bracketed page-numbers.]

Errata.

Page 15, line 13 from top, for FiCh read FetCh, Page 15, line 14 from top, for F%(h read FetCh. Page 15, line 20 from top, for Ft(h read FetOi, Page 118, line 4 from bottom, for shortage read shrinkage. Page 321, line 15 from bottom, for beHmee read at Umes. Page 324, lino 6 from top, for height (of the stamp) read weight. Page 328, line 2 from bottom, for R. N, Terhune read B, H. Terhune. Page 329, line 6 from bottom, for Hoffman read Hofmann, Page 335, line 10 from bottom, for Crook read Orooke, Page 337, line 16 from bottom, for Maiheson read Mathewson. Page 337, line 5 from bottom, for MaXheson read Mathewson. Page 631, line 2 from top, for Raisbeeh read Raitheck, Page 654, lino 16 from top, for sUver-mxCls read some silver miUs, In the table on p. 117 the percentage of carbon in the interior of Martin boiler- plate, No. 5, given as 0.028, should be 0.280.

Acid open-hearth process, 347, 396, 490.

Adirondack region, iron-ores of, 58.

Akbrman, Prof. Richard; The Bessemer Process as Conducted in Sweden [xvi], 265;

discussion, 661. Alabama, copper-deposits of [75]. Alcohol, use of, in miners' safety-lamps, 141 et seq. Algonkian rocks, 57 et seq.

Alloys (see also the metals) : copper, 261 ; gold-silver, 117. Alumihates of calcium, rapid setting of, in water, 12.

Aluminum: Bernard process [342]; Cowlee process, 341; Hall process, 341; metal- lurgy of, in America, 341 ; works at Mulhausen am Bhein [342]. Aluminum-steel, 114.

Anaconda smelting-works, Butte, Mont. [329, 333, 575]. Analyses of (see also assays of) : Basic charges in open-hearth practice, 431.

Boiler-plates, 111, 113.

Cements: grappier, 20; Portland, 16; quick-setting, 18; slag, 20.

Cherokite, 196.

Coal, small sizes of anthracite, 603.

Gtoses for fuel in open-hearth process, 366 et seq.

Hydraulic lime, 17.

776 Index.

LimeBtone (dolomite), 195.

Margarodite, 240.

Ore from Ground Hog gold- and silver-mine, Battle Mountain, Colo., 758.

Pig-iron, Bessemer, from Swedish blast-furnaces, 277.

Slag: in open-hearth process, 411 et 9eq.; Swedish Bessemer, 275, 277, 669.

Steel : Bessemer steel rolls, 109; Martin steel, 109; Martin steel-plate, 117; ingots,

Topaz crystals, 240.

Zinc spelter from Bertha zinc-mines, Va., 536. Ansell's apparatus for detection of fire-damp, 136. Anthracite (see also coal): analysis of small sizes, 603; furnace for burning small

sizes of, 581 ; results of tests of pea and buckwheats, 603. Antimony-ores : of Sevier county. Ark., 207 ; treatment of, in the United States,

Anzin coal-mines, France, detection of fire>damp at, 147 et §eq. Appalachian areas: copper-deposits in, 75; gold and silver in, 87; lead- and zinc- ores of, 80, 81 ; magnetites and hematites in, 59; manganese-bearing minerals of, 68. Appirratus : for detection of fire-damp, 123 et ieq. 726 ; for dressing- works, 225 et

eeq. ; for measuring cap-heights of hydrogen flame for detection of gas, 616. Archaean, or non-clastic, rocks, 56. Arents's siphon-tap for lead-Airnaces [332, 337]. Aboall, Philip, remarks in discussion of Mr. Bickard's paper on the origin of

gold-bearing quartz of Bendigo reefs, 740. Argentiferous lead- and zinc-deposits of Ouachita uplift, 206, 213. Arizona : copper-ores, 74, 75 ; treatment of copper at carbonate-mines, 334. Arkansas: association of magnetites with basic igneous rocks [59]; fiasures in zinc-mines of Marion county, 187 ; lead- and zinc-deposits, 172 ei seq. ; manga- nese-ores [68] ; zinc-ores [81]. Ashworth-Gray safety-lamp, 731. Aspen district, Colo., silver-production, 87. Assays of (see also analyses of) copper-matte, 677. Associates: election of, xviii ; made members, xviii.

Atlantic copper-mine, Lake Superior, Mich., stamp-mUl practice at, 326, 648. Auriferous quartz of Bendigo reefs, Australia, 303. Austria, quicksilver-deposits of Idria, 85.

Bakbb, David, remarks in discussion of Mr. Kennedy's paper on blowing-engines,

Ball steam-stamp, 322 et teg., 651. Bangbro blast-furnace, Sweden, 275 et teq. Baraboo axis of Wisconsin, 624.

Barite (heavy spar) associated with blende in southwest Wisconsin, 567. Bartlett, F. D., treatment of zinc-ores by, at Cafion City, Colo 661. Basic furnace, 353.

Basic open-hearth process, 347, 419, 499. Battle Mountain, Eagle county, Colo., ore-deposits, 758. Bayles, J. C, on microscopic analysis of iron and steel [263]. Beaumont, Elie de, on water in eruptive igneous rocks, 741. Becker, G. F., on Comstock rocks, 734. Beck's vertical illuminator for microscopic work, 247. Behrens, H., on microstructure of \jteel [256, 264]. Belleville, Jasper county, Mo., lead- and zinc-deposits at, 645. Belts on filling of Australian gold veins [753].

Index. 777

Bendiiro gold-field, Victoria, AastraliA : lava dikes, 296 H Mg. ; origin of gold-bearing

quartz, 289; rock-formations, 290. Bennett Brothers lead- and sine-mines, southwest Wisconsin [550], Bernard process for producing alnminnm [342]. Bertha Zinc and Mineral Co., Va., 511.

Bertha Sne-Mine$ at Bertha, Va, (Case) [xy], 611 ; discussion, 696. Bessemer blowing-engines, 539.

o

Beeeemer Proceu as Conducted in Sweden (Akebman) [xvi], 265; discussion, 661.

Bessemer process : account of, by Sir Henry Bessemer, 265 ; blow-holes influenced by temperature, 271 : Caspersson's converter-ladle, 284 : direct use of pig-iron in Sweden, 267 et eq. ; experiments with, by Goransson, 266; influence of man- ganese, 276; influence of silicon, 273 et seq. ; in Sweden, 265.

Bessemer, Sir Henry, account of Bessemer process by [266].

Bessemer steel (see also steel) : analysis of, 109; blow-holes in ingots influenced by temperature, 272 ; special treatment of ingots at Terre-Noire, 106; ftx)m Swedish pig-iron, 267.

Bessemer steel-works in Sweden, 265 ei teq. ; weight of change in, 668.

Bethlehem Iron Co., blowing-engines used by, 537.

Bethlehem Zinc Co. [697].

BiLHARZ, O. : General and Special ObeervaOona Concerning Ore-Dremng [xv], 225; dis- cussion, 699.

Bilharz-Bittinger ore-classifier, 227 [648].

Bischof, theory of deposition of gold by, 752.

Black Hills. Dakota, ore-beds, origin of, 757.

Black Jack " (blende), deposits of, in southwest Wisconsin, 566.

Blaine and Logan lead- and zinc-mines, southwest Wisconsin [559].

Blake ore-crusher, 322, 323 [660].

Blake, William P.: The Mineral Depotiie of SotUhtceet Wisemsin [xiy], 658; The Separation of Blende from Pyrites: A New Metallurgical Industry [xvi], 569; dis- cussion, 723; remarks in discussion of Mr. Case's paper on the Bertha rinc- mines, 698 : of Mr. Douglas's paper on American improvements and inventions in ore-crushing and concentration, 660 ; of Dr. Jenney's paper on lead- and zinc- deposits of the Mississippi valley, 621.

o Blast-furnaces : Sweden : Bangbro, 275 et aeq. ; Domnarfvet [275] ; Edske, 266 ; Hag-

o

fors [275] ; Langhyttan, 275 et'seq., 668 ; Nykroppa, 275 et acq. ; Sandviken, 275

et acq. ; Westanfors, 275 et aeq, Blauvelt, W. H., manufacture of p]:oducer-gas from anthracite (in paper of Mr.

Campbell) by, 380. Blende : deposits of, in Southwest Wisconsin, 562 ; separation of, from pyrites, 569. Blende lead- and zinc-mine, southwest Wisconsin [559]. Blister-steel, 254.

Blow-holes in steel ingots, 271, 671. Blowing-engines: Cambria Iron Co. 's, 710 ; indicator cards taken at Homestead

steel-works, 720; Maryland Steel Co.'s, 721. Blowing-Enginea (Kennedy) [xvii], 537; discussion, 709. Blue limestone or glass-rock," 633. Bog iron-ores, 62, Boiler-plates: analysis of, 111, 113; defects in steel, 106; selection of metal for,

112 ; of steamer Lividia, 106, 111 ; tests of, 114 et aeq. Boilers at Pennsylvania collieries, 588 et aeq. Bonanza lead- and zinc-mine, southwest Wisconsin [559]. Bonne Terre lead-mines, St. Francois county. Mo. [178], 186 et aeq. Boston and Colorado smelting- works, Argo, Colo., 333, 334. Boston and Montana Co., Great Falls, Mont. [333],

778 Index.

Bricks : slag-bricks, 575 ; use of hollow, for collecting Aimes at GraDt and Omaha works, Denver, Colo., 657.

Bridge metal, character of, 115.

Bridgmau ore-sampling machinery [656].

Bbooks, James C, remarks in discussion of Mr. Kennedy's paper on blowing- engines, 711.

Brown- Allen (improved O'Hara) roasting-fnmace, 330.

Brackner furnace, 329 aeq.

Brnnton ore-sampling machinery [656].

Buncombe Hill lead- and zinc-mine, southwest Wisconsin [558].

Bunsen burner, character of, 682.

Butte district, Mont, silver-production, 87.

Byrne's lead- and zinc-mines, southwest Wisconsin [550].

Oalcining- furnaces, 328.

Calcium : aluminates of, rapid setting of, in water, 12; ferrites of, under action of water, 13 ; silicates of, 11 ; silico-alumino ferrites of, behavior in water, 13 ; tests relating to aluminate of, 52.

California: gold-deposits, geological location of, 90 ; iron-ore in [62] ; manganese- ores of [68] ; occurrence of copper in [77] ; quicksilver-deposits, 85.

California steam-stamp, 322, 323, 654.

Calumet and Hecla copper-mine. Lake Superior, Mich. [73] ; Heberle mill at, 647.

Cambria Iron Co., steam-engines of, 710, 714.

Cambrian limestone : gold in, 89 ; lead-ores in, 203.

Campbell, H. H. : The Open-Hearth ProceaSj [xvi], 345 ; discussion 679 ; Mr. OBkmp-

o

bell in discussion of his paper, 692; remarks in discussion of Prof. Akerman's

paper on the Bessemer process in Sweden, 667. Canada, nickel-deposits of Sudbury, 70.

Candlot, M.: on disintegration of cements, 52; on slacking of lime [11]. Cap-measurements for hydrogen and oil flames for detecting gas, 615. Carbon in open-hearth process, 390, 464.

Carbonic acid and water, determination of, in hydraulic material, 27. Carboniferous limestone, silver in, 89. Carbon-steel, constituents of, 250. Carnegie Steel Co., Homestead steel-works of, 720 Carpenter. F. B., on Black Hills ore-beds, 757.

Case, William H. : The Bertha Zine-Minet at Bertha Va, [xv], 511 ; discossion, 696. Caspersson, C. A. : Converter-ladle, 284 ; on influence of heat on blow-holes in steel- ingots, 271. Caspersson ladle, use of, in Swedish Bessemer steel works, 284 [664]. Cassiterite in the Black Hills, 71. Cast-iron (see also iron), constituents of, 261. Cementite, a constituent of carbon-steel, 251, 253 et aeq. Cements : analyses of, 16 et eeq. ; chemical compounds in, 10; effect of temperature

on sotting, 36; grappierf 18 et $eq.; hydraulic tests of, 3 eteeq; mixed, 20;

natural, 18 : pastes, methods of measuring and testing, 37 ; Portland, 11 et eeq, ;

quick -setting natural 18; slag, 20; slow-setting natural, 19; tempering-

water, 35. Central copper-mine. Lake Superior, Mich., crushing ore at, 324. "Centrifugaled" steel, 674.

Chamberlin, T. C, on the geology of the Mississippi Valley, 172 et teq. Chemical composition of hydraulic products, 5. Chemical compounds in cements, 10. Chemical tests of hydraulic products, 26. Cherokee limestone, lead- and zinc-deposits in, 191 et eq.

Index. 779

Cherokite, analysis of, 196.

Chesneau fire-damp indicator: description of, 151; anderground experiments with, 166 ; practical tests of, 169.

Chesnsau, O. : The DeteeHon and Meantrement of Fire-Damp in Mines [xv], 120 ; dis- cussion, 725.

Chilean mills [660].

Chili, gold in eruptive rocks of [92].

China, Kwei-Chau quicksilver-deposits, 85.

Chubch, John A., remarks in discussion of Mr. Emmons's paper on geological dis- tribation of nsefnl metals in the United States, 732.

Cincinnati anticlinal, 172 ei aeq. ; 623.

Clark oxidizing and desnlpharizing apparatus, 329.

Classification of hydraulic products, 14.

Clowes, Prof. Frank : The Hydrogen-Oil Safety-Lamp for Lighting and for Accurate and Delicate Detection and Measurement of Inflammable Oat and Vapor in the Air [xv], 606; discussion, 725.

Coal : analysis of small sizes of anthracite, 603 ; bituminous, converted into gas for fuel, 371 ; cleansing of, in Great Britain and continent of Europe, 705; in Moni- teau county, Mo., 188 ; works for preparation of, 231.

Coal-gas, detection nd measurement of, with hydrogen-oil lamp, 615.

Coal-mines: Pennaylvania : Schuylkill county; No. 3, 588 eteeq.; Luzerne county ; No. 6, 588 seq.; France: Anzin, 147 et seq.; Houillidres de St £tienne, 169; Lens, 150 et seq.; Livin, 168, 170; Bonchamp [124]; Scotland: Motherwell (near Glasgow), Miny and Cunninghame [706].

Colby iron-mine (manganiferons). Gogebic range, Mich. [68].

Collom jigs [326], 650.

Colorado: copper-ores, 74, 76, 335; iron-ores, 60 eteeq; lava-flow at Marshall basin, near Telluride, 738; manganese-ores, 68.

Colorado steam-stamp, 324, 654.

Coloration of safety-lamp aureoles for detection of fire-damp, 157.

Columbia reverberatory Aimace, Argo, Colo., 655.

Coltman's lead- and zinc-mines, Southwest Wisconsin [559].

Comet ore-crusher [322], 323.

Compass, improved hanging, 543.

Comstock lode, Nevada, 91 ; silver production, 87.

Concentrating machinery, 324, 647.

Conoentrating-works : Montana: 651.

Concentration: American improvements in, 321 ; dry, in Bocky Mountains region, 327 ; machinery for, 324 ; mechanical, in Lake Superior region, 325.

Constant ore-sampling machinery [656].

Copper: in Algonkian rocks, 73; American improvements in the metallurgy of, 333 ; in Mesozoic rocks, 76; in Palteozoic rocks, 75; production of, in the United States from 1845 to 1890, 72; in Tertiary rocks, 77.

Copiier-alloys, microscopic examination of, 261.

Copper-deposits, genesis of, 77.

Copper Falls copper-mine, Lake Superior, Mich. [323].

Copper-lead matte, treatment of, in American smelting-works, 335.

Copper matte : assay value of, 677 ; constituents of, 676.

Copper-mines: Aritona: Graham county; Longfellow, 331; Tavapai county; Verde [334]; Michigan: Houghton county; Atlantic, 326; Calumet and Hecla [73 647]; Pewabic, 323; Quincy, 323; Keweenan county; Central, 324; Copper Falls [323].

Copper-ores: American treatment of, 333 ; of the United States, 72; Arizona: car- bonate-mines, 334 : Montana : Butte,' 334.

Coquillion, apparatus for detecting fire-damp constructed by, 123 [135].

780 Index.

Cornwall ore-bankSf Lebanon county, Pa., production of magnetite, 60.

Ck)st of sinking and lining shafts at sinc-miues, Bertha, Va., 531 et aeq.

Cowles electric process for producing alaminam, 341.

CoXB, ECKLEY B. : A Fttmace with AtUomaHc Stoker, TraveUing Orate and VariahU

Blastf Intended Especially for Burning amall Anihraeite Ooala [xv], 581 ; remarks

in discussion of the papers of M. Chesnean and Prof. Clowes on fire-damp in

mines, 729. Creede district siWer-veins, Colo. [92]. Cripple Creek gold-deposito, Colo. [92]. Crooke method for treatment of copper-lead matte, 335. Crushing machines: American improvements in, 322; Heberle [647]; Schrans

[647]. Crystalline rocks: classification of, 56 ; gold and silver in, 87. Crystallization of steel, 546.

Cuba City lead- and zinc-mine. Grant county. Wis. [559, 633]. Cupola-furnaces, 331. Cylinder-furnaces, 329.

DahleruSt C. G., analyses of Bessemer pig-iron by, 277.

Daintree, Bichard : experiments on precipitation of gold by, 32 ; on quartz veins of Victoria, Australia, 753.

Dakota, South, geological relation of Black Hills, to Ozark and Wisconsin uplifts,

Dana on lava of the crater of Kilauea, 744, 748, 768.

Daubre on water in eruptive igneous rocks, 741.

Davy safety-lamp, 120, 140, 144 et aeq.

Density of hydraulic products, 29.

Deposition : of lead- and zinc-ores in Cherokee limestone, 191 et aeq, ; of minerals constituting ore-bodies of Mississippi valley, 199, 210, et aeq.

Deposition of iron-ore: concentration from descending currents, 64; in Lake Superior region, 63 ; of titaniferous ores in Sweden and Norway, 65.

Desilverization of lead, improvements in, 658.

Detection and Meaaurement of Fire-Damp in Minea (Chbsnkau) [xv], 120; discussion,

Detrital deposits, gold and silver in, 92.

Diamond Joe lead- and zinc-mine, southwest Wisconsin [559].

Dikes: in iron-mines of County Antrim, Ireland, 747: lava, of Bendigo gold-field, 296 et aeq., 744, 749 , traversing sandstone rocks in western Scotland, 748.

Dinoire safety-lamp, 150.

Dodge ore-crusher [322].

DoUiak, O., on microstrncture of metals [263].

Dolomite : analysis of, 195; ores in solid, 628.

Domnarfvet blast-furnace, Sweden, analysis of slag from, 275.

Douglas, James: Summary of American Improvementa and IwvenUona in Ore-Orukmg and Concentration and in the Metallurgy of Copper, Lead, Goid, Silver, Nickel, Alu- minum, Zinc, Mercury, AnHmony, and Tin [xv], 321 ; discussion, 647.

Dressing- works (see ore-dressing), design and operation of 225.

Dry-Bone zinc-mine, Lafayette county. Wis. [560].

Dudelingen steel-works, Luxemburg, Belgium, improved practice at, 691.

Dudley, P. H., on microstrncture of steel [260, 264].

Dunn, E. J., report of, on Bendigo gold-field [770], 772.

o

DuBFEE, W. F., remarks in discussion of Professor Akerman's paper on the Besse- mer process in Sweden, 665.

Eakins, L. G. : analysis of cherokite by, 196; analysis of dolomite by, 195.

Index. 781

Eccles, Herbert, experiments by, to determine defects in steel, 110.

Edske blast-famace, Sweden, exxteriments with Bessemer process at, 266.

Egleston, Prof. T., experiments on the solution and precipitation of gold, by, 759.

Election of members and associates at Chicago, Aagnst, 1893, xyii.

Electricity applied to measurement of fire-damp, 138.

Emmons, S. F. : Geological Distribution of ihe UteftU Metals in the United States [xiv],

53 ; discussion, 732 ; Mr. Emmons in discussion of his x>aper, 737. Engines, blowing, 537, 709.

Eruptive rocks: association of lead- and zinc-deposits with, 84; origin of, 55. Etna, Mount : formation of fissures, 746 ; preservation of sheet of snow by lava

at, 748. Eureka district, Nevada, silver-production, 87. Eurich tapping-jacket [657]. Evans ore-separator, 326, 648. Excursions and entertainments, xviii.

Experiments on the specific gravity of gold contained in gold-silver alloys, 117. Explosives, safety, for use in mines, 120.

Falling Cliff (Bertha) zinc-mines, Va. [536].

Fault-fissures, lead-bearing, in Missouri, 82.

Faults: In Cherokee county, Kan., 186; in lead- and zinc-regions of the Missis- sippi valley, 622, 625; locating ore-deposits by, 628; in Ozark uplift, 178.

Ferrite, or crystalline iron, a constituent of carbon-steel. 251 et seq,

Ferrites of calcium under action of water, 13.

Fire-damp: detection and measurement of, in mines, 120, 606, 725; flame aureoles as indicators of, 144 ; indicators for underground use, 133 ; laboratory methods for the determination of, 123 ; Le Chatelier's apparatus for detecting, 125 et seq. ; limits of combustibility, 130; platinum wire electric-indicator, 137.

Fissures: ore-bearing, 82, 185; in zinc-mines of Marion county. Ark., 187.

Flames, luminous and non-luminous, experiments with, 682, 692.

Flint beds in vicinity of lead-deposits, 632.

Forbes, David, on South American gold-deposits, 756.

Forbes method of detecting fire-damp [136].

Forchhammer, analysis of topaz crystals by, 240.

Foster, C. Le Nkve : Minitigand Mineral Statisties [xiv.], 95; remarks in discussion of Mr. Case's paper on the Bertha zinc-mines, 697 ; of the papers of Ches- neau and Professor Clowes on fire-damp in mines, 725.

Foster ore-crusher [323].

Fossils, layers of, in oil-rock, 630.

Franklin ite deposits of northern New Jersey, 80.

Freiberg barrel* process for silver-ores [339].

Frue vanner, 327.

Fuel for open-hearth process [346], 370.

Furnaces : basic, 358; Brown-Allen, 330 ; Bruckner, 329 et seq. ; for burning small anthracite coals, 581; dimensions and proportions of, 602; calcining, 328; cupola, 331 ; Gerstenhofer [328] ; Herreshoff water-jacket [332] ; Hofftaian [329] ; Hnttener and Scott, for roasting cinnabar at New Almaden, Oil., 343 ; O'Hara, 330 etseq.; open-hearth compared with converter, 349; Pernot, 363; Piltz [337] ; Baschette [337] ; regenerative furnace and its machinery [346], 356 ; reverberatory, 333; size of hearth of furnaces at Argo, Col., 655 ; Spence, 330; stationary hearth [328], 330; Stetefeldt, 659; Stetefeldt shaft-calciner, 328; water-jacketed, 331; Wellman, 681; Whelpley and Stover, 328; White- Howell, 329.

Furnace with Automatic Stoker, Travelling Orate, and Variable Blast, Intended Especially for Burning Small Anthracite Coals (Coxe) [xv], 581.

782 Index.

Galena: in Cherokee county, Kan., 178; deposits in Cambrian limestone, 205 ; in Jasper coanty. Mo., 188; at Lansing, Iowa, 211; silver-bearing, in Lincoln county, Nov., 217.

Galena Level lead- and zinc-mine, southwest Wisconsin [559].

Gap nickel-mine, Lancaster county. Pa., 69 [340],

Garrison, F. Lynwood, on microstructure of car-wheel iron, 261 [263].

Gas : comparison of anthracite and bituminous, 380; dynamic equation of, in open- hearth process, 366; measurement of inflammable, 606, 728 ; natural, 366, 386 ; producer, 366; Siemens, composition of, 374; water, in open-hearth process, 367,384.

Gas-producers : Siemens, 371 ; Wellman, 371.

Gates ore-crusher [322], 323.

Gauge for testing fiery gases, Le Chatelier's, 132.

Geikie on causes for outflow of lava, 749, 750.

General and Special OhaervatioM Concerning Ore-Dreeemg (Bilhabz) [xv], 225; discus- sion, 699.

Genesis of ore-deposits : copper, 77 ; gold and silver, 92 ; iron, 63 ; lead and zinc 83; manganese, 69; nickel, 70 ; quicksilver, 85.

Geological Disiribuium of the Useful MetaU in the United Statea (Emmons) [xiv],53; discussion, 732.

Geology of: Bertha zinc-mines, Va., 513; Mississippi valley areas of uplift, 177 Southwest Wisconsin, 559.

Georgia: copper-deposits [75]; mangauese-ores [68].

German ia smelting- works, Jordan valley, Utah [329].

Gerstenhofer furnace [328].

" Glass-rock " (blue limestone), 633.

GoETZ, George W., remarks in discussion of Mr. Campbell's paper on- the open - hearth process, 679.

Gold : in Appalachian areas, 87 ; association of, with silver, 86 ; combination oC with sulphur, 760; deposits of, in Chili, 761 ; deposits in limestone pockets in the Urals and in California [698] ; deposits of South America, 756 ; in detrital de- posits, 92; genesis of deposits, 92; in Mesozoic rocks, 90; metallurgy of, in America, 338; Newberry's review of origin of deposits, 757; occurrence in Silu- rian and Cambrian limestones, 89 ; in Palozoic rocks, 89 ; precipitation of, from solution, causes of, 312; production of, in the United States, 86; in quartz veins of southern India, 761 ; in sea-water, 307, 738 ; solubility of allotropic forms of, in water, 761; solubility of, in different liquids, 759; specific gravity ot, con- tained in gold-silver alloys, 117; in Tertiary rocks, 91; theories concerning precipitation of, from sea- water, 751 ; treatment of gold-bearing ores, 235.

Gold-fields : Aueiralia : Victoria ; Bendigo, 289.

Gold- mills (see also stamp-mills, etc.) ; Ala$ka: Treadwell [330] ; California, 33a

Gold-mines: Colorado: Gilpin county; St. Louis, 313; Atutralia: Victoria; '*180," 296 et seq., 755, 770.

Gold-ores, treatment of, in America and Australia, 699.

Gold- and silver-mines : Colorado: Eagle county ; Ground Hog, 758.

Goransson, G. F., experiments with Bessemer process by, 266.

Gordon, F. W., remarks in discussion of Mr. Kennedy's paper on blowing-engines, 709. "

Granite Mountain silver-mine, Deer Lodge county, Mont, value of product, 87.

Grant and Omaha smelting-works, Denver, Colo. [578, 657].

Grappier cements, 18 et eeq. ; analysis of, 20.

Grate, travelling, for furnace burning small anthracite coals, 590 et eeq,

Greenland: metallic iron in eruptive rocks at Ovifak,65; occurrence of nickel at Ovifak, 71.

Ground Hog gold- and silver-mine, Battle Mountain, Colo., analysis of oxidized ore from, 758.

Index. 783

Gnillemin on xnicrographic examination of copper-alloys, 261 [265].

Gniterman, F., on ore-deposits of Battle Moantain, Colo., 758.

QUBLT, Dr. Adolf : On a Remarkable Deposit of Wolfram- Ore in the United States [xiv],

Qypsnm, calcined more soluble than natural, 7.

Hagfors blast-furnace, Sweden, analysis of slag from, 275.

Hall, C. M., process for producing aluminum, 341.

Hardening of hydraulic materials, 7.

Habtshorne, Joseph, remarks in discussion of Prof. Akerman's paper on the Bes- semer process in Sweden, 661.

Hartz jig [326], 652.

Hearth of furnace in open-hearth process, 367.

Hebele mill [647].

Helena lead- and zinc-mine, and mill, southwest Wisconsin [559, 568, 574].

Hempstead (Old Elevator) lead- and zinc-mine, southwest Wisconsin [559].

Hbnnin, Alphonse, remarks in discussion of Mr. Gampbell's paper on the open- hearth process, 689.

Herreshoff water-jacket furnace [332].

HiBBARD, H. D. : remarks in discussion of Mr. Gampeirs paper on the open-hearth process, 687 ; on open-hearth practice, 485 et seq.

Hills, B. C, on origin of gold-deposits, 758.

Himmelfahrt silver-mine, Freiberg, Germany [227].

Hofmann roasting-fumace [329].

HoFMAN, Pbof. H. O. : remarks in discussion of Mr. Douglas's paper on American improvements and inventions in ore-crushing and concentration, 656 ; on con- stituents of copper-matte, 676.

Holden stamp-mill, Aspen, Pitkin county, Ck>lo., 659.

Homestead steel-works, Allegheny county. Pa., indicator cards on blowing-engines taken at, 720.

Home silver- mine, Beaver county, Utah, value of product, 87.

Houghton, Samuel, on theory of hydrothermal fusion of granitic rocks, 742.

Houillidres de St. !tienne collieries, France, experiments with fire-damp at, 169.

Howe, H. M. : remarks in discussion of Mr. Campbell's pai)er on the open-hearth process, 691 ; constituents of steel named by, 251; on metallurgy of steel, 251 etaeq.

Howitt, A. W., on lava-dikes of Bendigo gold-field, 770.

Hubbard's iron-mine (wolframite), Fairfield county, Conn., 239.

Hiittener and Scott furnace for roasting cinnabar at New Almaden, Cal., 343.

Hughes, Herbert W., remarks in discussion of the papers of M. Chesneau and Prof. Clowes on fire-damp in mines, 730.

Huntingdon mill [324].

Huronian, or iron-bearing rocks, 57 et eeq.

Hydraulic limes, 16; analysis of, 17.

Hydraulic materials; fineness' of grinding, 28; hardening of, 7; swelling by slack- ing, 9 ; tests of; 3.

Hydraulic mining, 324.

Hydraulic ore-separators, 650.

Hydraulic products : agents of disintegration, 21 ; density of, 29 ; chemical compo- sition of, 5 ; chemical tests of, 26 ; classification of, 14 ; general laws of solution, 5; methods of testing, 25; physical tests of, 28; tests for invariability of vol- ume, 31.

Hydrogen-flame: advantages of, for detecting gas in mines, 613; measurement of caps, 615.

784 Index.

Hydrogen- Oil Safety-Lamp for lAghting and for Accurate and Delieatg Detection and Measurement of Inflammable Qa and Vapor in the Air (Clowes) [xt], 606 ; dis- coBsion, 725.

Hydrogen-oil safety-lamp iDTented by Prof. Clowes, 147 et ieq.f 606.

Hydrothermal fusion, literatare of theory of, 741.

I'Anson, J. C, remarks in discussion of Oberbergrath Bilharz's paper on ore-dress- ing, 705.

Ida Blende lead- and zinc-mine, Southwest Wisconsin [559].

lies, Dr., method of, for separation of matte and slag [657].

Illinois: lead- regions, 172 et eq. ; zinc-ores [81].

Improved Hanging Compose (Johnson) [xv], 543.

Improved Slag-Pots (Kelleb) [xvi], 574 ; discussion, 675.

Indian Territory lead- and zinc-deposits, 172 et seq.

Indicator-cards taken from air-cylinder of blowing-engines, 720.

Indigenous origin or deposition of ores, 627.

International Engineering Congress, Chicago meeting of the Institute in connection with, xiii.

Iowa: lead-regions, 172 etseq; zinc-ores [81].

Iron (see also cast-iron, pig-iron, etc.) : combustion of, in open-hearth process, 391 ; direct use of pig-, in Sweden, for Bessemer steel, 267 et seq. ; German nomen- clature of, 691 ; microscopic examination of, 260; occurrence of, in the United States, 57 ; segregation and its consequences in ingots of steel and, 105 et'seq. ; separation of pyrites from blende, a new metallurgical industry, 569.

Iron-mines: Connecticut: Fairfield county: Hubbard's, 239; Midiigan: Gogobic range, Colby (manganiferous) [68]; Missouri: Iron county; Pilot Knob [59], 735 ; St. Franyois county ; Iron Monntain [59] ; Pennsylvania : Lebanon county ; Cornwall [60] ; Virginia : Alleghany county ; Longdale, 543.

Iron Mountain iron-mine, St. Franiois county. Mo. [59].

Iron Mountain, St. Francois county. Mo., iron-ores, 735.

Iron-ores : in Algonkian rocks, 58 et seq. ; bog-ores in Oregon and Washington [63] ; from carbon ifero limestone, 61 ; in crystalline rocks, 57 ; deposits in eruptive rocks at Ovifak, Greenland, 65; in eruptive rocks, 63, 65; formation of Lake Surior deposits, 64 ; genesis of deposits, 63 ; manganifer- ous, in Colorado, 68; in Mesozoio rocks. 62; occurrence of magnetites with ig- neous rocks in Arkansas [59]; in Palceozoic rooks, 60; ''pipe-ore" (limonit'e) in Missouri, 639 ; product of Lake Superior region, 58; in Tertiary and recent de- posits, 62 ; titaniferous, in the Bocky Mountains, 60. Locxitieb : CaUfomia [62]; Colorado met seq. Maryland [62]; Minnesota [58, 62]; Misaowri: south- eastern, 59, 637, 646, 735; New Jersey : Highlands, 58 ; iron-zinc (manganiferous) [68]; New York: Adirondack region, 58; North Carolina [62]; Virnia, 59; Wythe county ; Austinville, 723; Wiscmisin [62] ; Mexico [62].

Iron-works: France: Sireuil, 105; Terre-Noire, 105 et seq., 268 [481, 661].

Irving and Van Hise: classification of rock-series by, 57; researches concerning origin of iron-ore deposits by, 63 et seq.

Jenney, Walter P. : The Lead- and Zinc-Deposits of the Mississippi Valley [xiv], 171 ; discussion, 621; Dr. Jenney in discussion of bis paper, 642; on lead- and zinc- deposits of the Mississippi Valley [83].

Jigs: Bilharz, 228 etseq.; Collom [326], 650: Hartz [228, 326], 652; Krom, 327, 653; Paddock, 327, 653; of Lake Superior copper-mills, 701.

Johns, H. W., Manufacturing Co., New York City [722].

Johnson, Guy B. : An Improved Hanging Compass [xv], 543.

Joplin, Jasper cminty. Mo., lead- and zinc-mines [178], 188 et seq.

Judd, Prof., on volcanic action, 298, 746, 768, 772.

Index. 785

Kansas : faalt-fissnres in Cherokee county, 186 ; lead- and zinc-deposite, 172 et seq. ; zinc-ores [81].

Keller, H. A.: Impf<jmed Slag-Pots [xvi], 574; discussion, 675.

Kellogg lead- and zinc-mines, Pulaski county. Ark. [206].

Kemp, J. F., on origin of gold-deposits, 761.

Kennedy, Julian : Blowing Engines [xvii], 537; discussion, 709.

Kentucky, lead- and zinc-deposits of, 172 et eq.

Keweenawan, or copper-bearing, rocks, 57, 73, 75.

Kilauea, Dana on volcanic action at, 744, 748, 768.

Kimball, J. P., on replacement of limestones by iron-ores [64].

Kroehnke process for lixiviation of silver-ores [339].

Kromjig, 327, 653.

Krom rolls [324], 327, 652.

Kbom, S. B., remarks in discussion of Mr. Douglas's paper on American improve- ments and inventions in ore-crusbing and concentration, 652.

Ladle, Caspersson converter, in Swedish Bessemer steel-works, 284 [664].

Ladle-crane for handling metal and slag, 369.

Lake Superior copper-mills, jigs in, 701.

Lake Superior region: copper deposits, 73; iron-ores of, 58, 64; manganese-ores [68].

La Louvire steel-works, France [108].

Langhyttan blast-furnace, Sweden, 275 et teg., 668.

Lanyon Zinc Oxide and Paint Co., Wankegan, Lake coutity, Dl. [565].

Lateral secretion, theory of, 220, 317, 732, 739.

Lava: Section of, from volcanoes, 295, 749, 750; flow of, at Marshall basin, Colo.. 738 ; temi)erature of, in crater of Kilauea, 744; dikes in Bendigo gold-field, Vic- toria, Australia, 296 et seq., 744, 764.

Lead: American practice in smelting, 336; bag-process for collecting fames [337] : improvements in desilverization of, 658 ; production in the United States, 79.

Lead-fumacee : Missotm: Jasper county; Lone Elm, 3b..

Lead-mines (see also lead- and zinc-mines) : Colorado : Clear Creek county ; Terrible [80] ; Missouri: Madison county; Mine La Motte [178j, 186 et seq. ; St. Francois county; Bonne Terre [178], 186 et seq.; Washington county ; Palmer, 640.

Lead-ores : in fault-fissures in Missouri, 82 ; in Lower Magnesian limestone, 211 ; of the Mississippi valley, 172 ei seq.; 558, 621, 622; in Paleeozoic limestones, 82; sampling argentiferous, 656 ; treatment of, 656.

Lead-regions of Wisconsin, faults in, 625.

Leadville district, Colo., silver-production, 87.

Lead-works : Pennsylvania : Allegheny county ; Pittsburgh, Pennsylvania Lead Co. [564].

Lead- and Zine-Deposits of the Mississippi VdUey (Jenny) [xiv], 171 ; discussion, 621.

Lead- and zinc-deposits : genesis of, 83; in Subcarboniferous Cherokee limestone, 191 et seq.

Lead- and zinc-mines (see also lead-mines) : Arkansas: Montgomery county; Silver city [206] ; Pulaski county ; Kellogg [206] ; Sevier county ; Antimony City [206] ; Silver Hill [206]; Kansas: Cherokee county; Galena [178, 190], 193 et seq. ; Missouri: Jasper county; Belleville, 645; Joplin [178], 188 et seq.; Lawrence county ; Aurora, 178 et seq.; Newton county; Granby [178], 194 ff seq. ; Wisconsin : Grant county ; Cuba City [559, 633] ; Lafayette county ; Bennett Brothers [559] ; Blaine and Logan [559]; Blende [559]; Bonanza [559], Buncombe Hill [559]; Byrne's [559] ; Ctoltman's [559] ; Diamond Joe [559] ; Galena Level [559] ; Helena [559, 568, 574]; Hempstead (Old Elevator) [559]; Ida Blende [59]; Leary and Coulthard [559] ; Little Giant [559], 632; McCarty [559] ; Monte Christo [559] ; Oakland Level [559] ; Eaisbeck [559, 631] ; Sallie Waters [559] ; Stop-line [559] ;

786 Index.

Von Dusko [559] ; Wagner (McFeeley) [559] ; Wisconsin Lead and Zinc Co. [559] ; Zinc Carbonate Co. [559] ; Sardinia : Monteponi, 573.

Lead- and zinc-ores : deposits in the United States, 79; local names of varieties at Wisconsin mines, 563; in Mesozoic and Tertiary rocks, 83; minerals constitnt- injf ore-bodies, 198; of Missouri, 736 ; in older crystalline rocks, 80; in Oua- chita uplift, 206 ; in Ozark uplift. 187 : in Palseozoic rocks, 80 ; runs " or fissure-fed impregnations, 189 et seq, ; in Wisconsin uplift, 208.

Lead- and zinc-regions of Mississippi valley, faulting in, 622; not glaciated, 634.

Lead- and zinc- works : Virginia: Wytjie county ; Austinville, Wythe, 723.

Leary and Coulthard lead- and zinc-mine, southwest Wisconsin [559].

Le Ch atelier, H. : Test$ of HydraiUie Materials [xv], 3; apparatus for detecting fire-damp [122], 125 H seq, ; apparatus for determining percentage of fire-damp, 726 ; on hydrogen flame for detection of gas [607].

Le Conte, Prof., on origin of gold of Bendigo reefs, 769.

Legrand safety-lamp [149].

Lens collieries, France, experiments with fire-damp at, 150 ei seq.

Jjcwis and Bartlett bag-process of collecting lead-fumes [337].

Lieberkiihn mirror for microscopic study of metals, 247.

Livin coal-mines, France, use of Chesneau fire-damp indicator at, 163, 170.

Limes : analyses of hydraulic, 16, 17; slacking and swelling of, 10.

Limestones: analysis of dolomite, 195; blue, or "glass-rock," 633; Cambrian, lead- ores in, 203 ; Cherokee, of Mississippi valley, 191 ei seq. ; chimneys in Bertha zinc-mines, Va., 513 et seq.; gold -deposits in pockets of [696]: magnesian, of Lower Silurian, ore-deposits in, 202 ; ore-bearing rocks of Trenton and Galena, 209. ,

Little Giant lead- and zi no-mine and mill, Lafayette county. Wis. [559], 632.

Liveing*s apparatus for detection of fire-damp, 139.

Loewel, French chemist, on solubility of salts [6].

Lone Elm lead-smelting works, Jasper county, Mo., 337.

Longdate iron-mines, Alleghany county, Va., 543.

Longfellow copper-mine, Graham county, Ariz., smel ting-practice at, 331.

Louis, Henry : Note on Experiments on the Specific Gravity of Oold Contained in Gold-Silver Alloys [xvi], 117; discussion, 724; Mr. Louis in discussion of his paper, 724.

Liibrig system for handling and cleansing coal, 705, 709.

Lyell, Sir Charles, on formation of fissures on Mount Etua [746].

Macculloch on trap dikes of western Scotland, 748.

Machinery: blast-producing, 332; for concentration, 324 ; jigs, 326; ore-crushing, 322 ; ore-sampling, 656 ; vanners, 327.

Magnesia, effect of. on hydraulic materials, 27.

Magnesian limestone, lead- and zinc- deposits in, 202.

Magnetic separation at Monteponi zinc-mine, Sardinia, 573.

Magnetic separators at Wythe Lead and Zinc Mine O/s works, Anstinville, Va.,

Maine, copper-deposits of [75].

Manganese : genesis of deposits, 69; geological distribution of, in the United States, 67; influence on Bessemer process, 276 ; in ocean depths [69] ; in open -hearth process, 391 et seq. ; percentage of, in iron and steel at Westanfors blast-furnace, Sweden, 280 ; total product for the United States, 1880-90,

Manganese-steel, Mukai on microstructure of [259, 265)].

Manning & Squier zinc-mines, Wythe county, Va., 511.

Marcasite in lead- and zinc-mines of southwest Wisconsin, 568.

Margarodite, analysis of, 240.

Marsac stamp-mill, Park City, Utah [328], 310, 659.

Index. 787

Marsant safety-lamp, 150.

Martens, Prof. A., methods for microscopic stady of metals of, 246 ei seq.

Martin steel : analysis of, 109, 117; physical tests of, 116.

Maryland : deposits of native copper [75] ; iron-ores of [62] ; occurrence of zinc [81].

Maryland Steel Co., blowing-engines of, 71.

Mathewson slag-tap, 337 [657].

Mattieasen and Hegeler zinc-works, Lasalle, 111., 661.

McCarty lead- and zinc-mine, southwest Wisconsin, [559].

Measurement of fire-damp in mines, 120, 606, 725.

Mechanical tests of hydraulic material, 35.

Meeting of the Institute, Chicago, August, 1893, proceedings of, xiii.

Meier, J. W., magnetic separation of iron from blende by, 573.

Members, election of, xvii.

Mercury (see also quicksilver): Hiittener and Scott roasting-furhacc, at New Almaden, Cal., 343; mining and metallurgy of, in America, 342.

Mesabi iron-range, Minn., iron -ores of, 58.

Mesozoic rocks : gold and silver in, 90 ; iron-ores in, 62 ; lead and zinc in, 83.

Metallic substances, works for preparation of, 233.

Metallography, microscopic, 243,

Metallurgy (see also smelting) : American improvements in, 321 ; Crooke method for treatment of copper-lead matte, 335 ; Patero process for silver-ores, 340 ; Patio process for silver-ores [339] ; of quicksilver in the United States, 342 ; separation of blende from pyrites, 569.

Mexico, iron-ore in [62].

Microscopic Metallography (Osmond) [xvi], 243. (See Discussion, " Physics of Steel," vol. xxiii.)

Microscopic metallography : bibliography of, 262 ; history of, 245 ; methods of prepa- ration for, 246 etaeq.; polishing metal-plates for, 246.

Mierotiructufe of Steel (Sauveue) [xvi], 546. (See Discussion, Physics of Steel," vol. xxiii,)

Mine Hill zinc-mines, Sussex county, N. J., 342.

Mine La Motte, Madison county, Mo. : concentrating lead-nickel matte at, 676 ; lead-mines [178], 186 et seq.

Mineral Deposits of Southwest Wisconsin (Blake) [xiv], 558.

Mineral Point Zinc Oxide Works, Iowa county. Wis. [565].

Miners' safety-lamps, 120 et seq.

Mining and Mineral Statistics (Foster) [xiv], 95,

Mining : hydraulic, 324; zinc-ore at Bertha, Va., 523.

Minnesota, iron-ores of [58, 62].

Miny & Cunninghame's coal-plant, near Glasgow, Scotland, 706.

Mirabel (Bradford) quicksilver-mine. Lake county, Cal., 86.

Mississippi valley : dynamic geology of uplifts considered with reference to forma- tion of ore-deposits, 183; geology of areas of uplift, 177; lead- and zinc-deposits of, 79, 81, 171 et seq., 621 ; lead and zinc-regions, faulting in, 622 ; minerals asso- ciated with ores, 198, 210, 213 et seq.

Missouri : coal-basins, 188 ; iron-ores, 59, 637, 735 ; lead-ores, 82 ; lead and zinc-ores 172 et seq,, 736 ; lead and zinc regions, 622 ; occurrence of nickel-ore at Mine La Motte, 70; zinc-ores, [81].

o

Moen, Philip W., and Trotz, Emanuel, translation of Akerman's paper on the Besse- mer process in Sweden by, 265. Montana : concentrating works, 651 ; copper-ores, 74. Monte Cristo lead- and zinc-mine, southwest Wisconsin [559]. Monteponi lead- and zinc-mines, Island of Sardinia, 573.

788 Index.

MoBOAN, Joseph, Jb.. remarks in discussion of Mr. Kennedy's paper on blowing- engine, 710.

Mortars, tests of (see tests of hydraulic material), 35 et seq.

Mount Etna : formation of fissures, 746 ; preservation of sheet of snow by lava at,

Mulhansen am Bhein, aluminum-works at [342].

Mueseler safety-lamp, 144 et seq.

Mukai, Tetskichi, on manganese-steels [259, 265].

MuNBOE, Pbop. H. S. : remarks in discussion of the papers of M. Chesnean and Prof. Clowes on fire-damp in mines, 726; of the paper of Mr. Douglas on Ameri- can improvements and inventions in ore-crushing and concentration, 647.

Murchison, Sir Boderick, theory of deposition of gold by, 752.

Murday's thermo-electric fire-damp detector, 140.

Murray, B. A. F., on geology of Bendigo gold-field, Australia, 291, 300.

Nason, Fbank L., remarks in discussion of Dr. Jenny's paper on the lead- and zinc-deposits of the Mississippi valley, 636.

Natural cements, 18.

Natural gas in open-hearth process, 366, 386.

Neill, James W., remarks in discussion of Mr. Keller's paper on improved slag- pots, 675.

Nevada, nickel-ores of [70, 738].

New Almaden quicksilver-mine, Santa Clara county, Cal., 85, 86.

Newbery, Cosmo : on existence of gold in saline waters, 752, 754; experiments de- termining presence of gold in mine-waters by, 755.

Newberry, Prof. J. S., on origin of gold-deposits, 757.

New Hampshire, copper-deposits of [75].

New Jersey: iron-ores of Highlands, 58 ; iron-zinc ores (manganiferous) [68].

New Jersey Zinc C5., Newark, N. J.. 342.

New metallurgical industry, separation of blende from pyrites, 569.

New Mexico, copper-deposits of [76].

New York State, occurrence of zinc in, [81].

Nicholas, William, on origin of gold-bearing quartz of Bendigo reefs, 762.

Nickel : genesis of deposits, 70 ; geological distribution of, 69 ; metallurgy of, in America, 340; occurrence of, in Greenland, 71; in Missouri, 70; in the United States, 69 ; production in the United States, in last decade, 69.

Nickel-mines: Pennsylvania: Lancaster county; Gkip, 69 [340] ; Caiiada: Sudbury, 70.

No. 3 colliery, Oneida, Pa., 588 et aeq.

No. 6 colliery, Eckley, Pa., 588 seq.

North Carolina: copper-deposits [75] ; iron-ores [62] ; nickel-ores, 70.

Note on Experiments on the Specific Gravity of Gold Contained in Oold-SUver AUoys (Louis) [xvi], 117; discussion, 724.

Nykroppa blast-furnace, Sweden, 275 et seq.

Oakland Level lead- and zinc-mine, southwest Wisconsin [559].

Officers of the Institute : elected February, 1894, vii ; for year ending 1894. vii.

O'Hara roasting-furnace, 330 et seq.

Oil-rock in zinc-regions of the Mississippi valley, 629.

On a Remarkable Deposit of Wolfram-Ore in the United States (GtTBLT) [xiv], 236.

80" gold-mine, Bendigo gold-field, Victoria, Australia. 286 et seq., 765, 770.

tario silver-mine. Summit county, Utah, value of product, 87.

rio stamp-mill. Park City, Summit county, Utah, 659.

m-hearth furnace of Pennsylvania Steel Co. [333].

ii'Hearth Process (Campbell) [xvi], 345; discussion, 679.

iifDEX. 789

Open-hearth piocees: acid proceas, 347, 396, 490; analyses of basic charges, 431 basic process, 347, 419, 499 ; carbon in, 390, 464 ; dynamic equation of gases, 366 fnel, 370 ; hearth, constraction of, 367, 397. 420 ; manganese in, 391 et aeq. metal, variations in composition of, 411, 415, 416, 432; open-hearth furnace compared with converter, 349; petroleum as ftiel, 386; phosphorus in, 427 et aeq. ; recarburization, 467; regenerative ftimaoe and its machinery, [346], 356 ; regulation of temperature [346], 387 ; silicon in, 390, 462, 479 ; slags, analyses of, AlXaseq.; sulphur in, 446; theoretical losses, 492, 499 ; thermal equation of ftimace, 395; use of Siemens gas in, 374 ; of water-gas, 367, 384.

Oqnirrh mountain silver-district, Juab county, Utah, [90].

Ore-crushers, 322.

Ore-crushing, American improvements in, 647.

Ore-deposition : discussion of theories, 219, 626 ; from ascent, or descent, of solu- tions? 622; influence of organic remains, 629; "lateral secretion" theory, 220, 317, 732 M?.; vertical distribntion, 197.

Ore-deposits: brecciated, 632; classification of, 190; common characteristics of lead- and zinc-deposits of Mississippi valley, 212; localization of, by Dftulting, 628 ; occurrence of, in Ozark uplift, 189 ; in solid dolomite, 628 ; workable, as- sociated with faulting fissures, 184, 209.

Ore-dressing : classification of plants, 703 ; general and special observations con- cerning, 225.

Oregon : Bog iron-ores [63] ; nickel-ores [70].

Ores containing combined gold, treatment of, 235.

Ore-separators: Evans, 326, 648; hydranlic, 650 ; Bichards-Ooggin [648] ; V-separ- ator, 327, 648.

Orford smelting-works, Bergen Point, N. J. [332], 334,*340.

Organic remains, infiuence of, on ore-deposition, 629.

Origin of (Ke Gotd-Bearing QuarU of the Bendigo Eeefi, Auttralia (Bigkabd) [xiv.] 289; discussion, 738.

Osmond, F. : Mieroacopie MetaUograpky [xvi.], 243. (See Discussion, '' Physics of Steel," vol. xxiii.)

Ouachita uplift : argentiferous lead- and sinc-oree, 206, 213; location, description, and geology of, 172, 175, 178, 623.

Oxygen, absorption of, by different elements in open-hearth process, 402.

Ozark uplift: 82; iron-ores of, 637 el feg. ; lead- and zinc-ores of, 187; location, de- scription, and geology of, 172, 173, 177, 623.

Paddock jig, 653.

Paleozoic limestone, lead-bearing, 82.

Paleeosoio rocks : gold and silver in, 89 ; iron-ore in, 60.

Palmer lead-mine, Washington county. Mo., 640.

Parrott smelting-works, Butte, Mont, 330, 334 [575, 576].

Patera process for silver-ores, 340.

Patio process for silver-ores [339].

Payne, C. Q., remarks in discussion of Professor Blake's paper on the separation of

blende from pyrites, 723. Peabce, Bichabd: remarks in discussion of Mr. Douglas's paper on American

improvements and inventions in ore-crushing and concentration, 655 ; of Mr.

Bickard's paper on the origin of gold-bearing quartz of Bendigo reef 738 ; on

origin of gold-deposits, 757. Pearlyte, a constituent of carbon-steel, 251 et eeq,

Pennsylvania: deposits of native copper [75] ; zinc-deposits, 81 ; zinc-mines, 697. Pennsylvania Steel Company: open-hearth furnaces [333]; open -hearth practice

of. 345. Percival, on geology of Wisconsin lead-regions, 621 et aeq.

790 Index.

Pemot furnace, 363.

Peru, Haancavelica, qnicksilyer deposits, 85.

Petroleam shales, or oil-rock, in ziDC-regioos of Mississippi valley, 629.

Petroleum, use of, as fuel in open-hearth process, 386.

Pewabic copper-mine, Lake Superior, Mich., crushing ore at, 323.

Pfaff, determination of amount of water in granite rooks by, 743.

Philadelphia smelting-works, Pueblo, Colo. [576].

Phillips, J. A., on gold deposition, 752 [756].

Phosphorus in open-hearth process, 427 et aeq.

Physical tests of hydraulic products, 28.

Picher Lead Ck>mpany, Joplin, Mo., 661.

Pieler's alcohol safety-lamp, 146 et aeq.

Pig-iron (see, also, iron and cast-iron) : analyses of Bessemer, from Swedish blast- furnaces, 277; chemical composition of Swedish, 269; direct use of, in Sweden for Bessemer steel, 267 et teq. ; in open-hearth process, 402, 496; Swedish, for Bessemer steel, 267 et aeq.

Pilot Knob iron-mine. Iron county. Mo. [59], 735.

Piltz furnace for smelting lead-ores [337].

Pittsburgh Beducing Company, Pa., 341.

Placer deposits of gold and silver, 92.

Platinum wire heated by electricity, as fire-damp indicator, 138.

Plattner process for gold -ores [330].

Port Clarence steel-works, England, 115.

Portland cements, 11 et ieq. ; composition of, 14 ; specific gravity of, 30.

PoUBCEL, Alexandre : Segregation and it$ Consequences in Ingots of Sted and Irofc, [xvi], 105. (See Discussion, " Physics of Steel," vol. zziii.)

Poussigne, M. L., apparatus for detecting fire-damp constructed by, 124.

Poussin, De la Vall, experiment with quartziferous diorite by, 742.

Pozzuolana mortars, influence of atmospheric dryness upon, 23.

Pratt, J. P., experiments on oxidization of gold by [759].

Producer-gas in open-hearth process. 366.

Production : of lead- and zinc-ore in Ozark area, 1889, 188 ; of magnetite from Corn- wall iron-mines. Pa., 60.

Production of iron-ore: in Adirondack region, 58; in Lake Superior region, 58; in New Jersey Highlands, 58; in southeastern Missouri, 59.

Production in the United States : copper, from 1845 to 1890, 72 ; gold, 86 ; manga- nese, 68; mercury, 342; metallic lead from 1825 to 1890, 79; nickel, 69 ; silver. 86 i zinc, 79.

Prospecting for ore, rules for, 224.

Publications of the Institute, ix.

Pueblo Smelting and Refining Co., Colo. [337].

Quartz, gold-bearing, of Bendigo reefs, Australia, 289, 303.

Quicksilver: genesis of deposits, 85; metallurgy of, in the United States, 342; work- able deposits of, 84.

Quicksilver-mines: California: Lake county; Mirabel (Bradford), 86; Santa Clara county ; New Almadeu, 85, 86. '

Quincy copper-mine. Lake Superior, Mich., crushing ore at, 323.

Baisbeck lead- and zinc-mine, Lafayette county. Wis. [559], 631.

Baschette furnace for smelting lead-ores [337].

Raymond, R. W., remarks in discussion of the papers of M. Cbesneau and Prof.

Clowes on fire-damp in mines, 725. Recarburization : direct process substituted for, in Swedish Benemer works, 273,

287; in open-hearth process, [348], 467.

Index. 791

Begenerative furnace and its machinery [346], 356.

Benard, experiment with qnartziferons diorite by, 742.

Beverberatory furnaces and practice, 333.

Bichards-Coggin ore-8ei)arator [648].

BiCHABDB, ]pROF. B. H. .' remarks in discussion of Oberbergrath Bilharz's paper on ore-dressing, 700 ; of Mr. Douglas's paper on American improvements and in- ventions in ore-crushing and concentration, 650.

BiCKARD, T. A. : The Origin of the Oold-Bearing Quartz of the Bendigo Reefs Auatralia [ziv], 289; discussion, 738; Mr. Bickard in discussion of bis paper, 763; remarks in discussion of Oberbergrath Bilharz's i)aper on ore-dressing, 699; of Mr. Douglas's paper on American improvements and inventions in ore-crushing and concentration, 654.

Bittinger's rules : for designing dressing-works, 225 et aeq. ; different interpretations of, 700.

Boasting-furnaces (see also furnaces), types of, 328.

Boberts-Austin, Prof., on solubility of gold in water, 761.

Bocky Mountains, iron-ores of, 60, 62.

Bonchamp collieries, France [124].

Bosales, H., on filling of Australian gold-veins [753].

Botator for steel ingots, 673.

Bules: for prospecting for ore, 224; Bittinger's, for designing dressing- works, 225 ei $eq., 700.

" Buns," or irregular ore-bodies, 189 et 9eq,

Safety-lamps, 120 ei seq. 725; Ash worth -Qray, 731; aureole-heights as fire-damp in- dicators, 144 et aeq.; elongation of flame in presence of fire-damp, 140; hy- drogen-oil, for detecting inflammable gas, 606 ; sensitiveness of alcohol -flame,

Sallie Waters lead- and zinc-mine, southwest Wisconsin [559].

Sandberger on ore-deposition [732], 734.

Sandviken blast-furnace, Sweden, 275 et aeq.

San Juan silver district, Colo. [92].

Sauveub, Albert: Microatnidure of Steel [xvi], 646. (See Discussion, Physics of Steel," vol. xxiii.)

Scheerer's theory of origin of granitic rocks, 741.

Schild, H., on microscopic metallography [264].

Schneider, £. A., experiments with auroos sulphides, 760.

Schondorf, experiments for detection of fire-damp by [124].

Schranz mill [647]. .

Searer, C. B. P., on gold-deposition in Australia, 756.

Sea- water, iodide of gold in, 306, 738.

o

Sebenius, J. S., remarks in discussion of Prof. Akerman's paper on the Bessemer

process in Sweden, 671. Segregation and ita Conaequencea in Ingota of Steel and Iron (Pourcel) [xvi], 106. (See

Discussion, " Physics of Steel," vol. xxiii.) Selwyn, A. B. C, on geology of Bendigo gold-field, Australia, 291 [753]. Separation of Blende from Pyrites: A New Metallurgieal Industry (Blake) [xvi], 569 ;

discussion, 723. Settling-pot, improved, 677.

Seven-Thirty silver-mine, Clear Creek county, Colo., 740. Shaft calciuing-furnaces, 328. Shafts, sinking, at zinc-mines. Bertha, Va., 531. Shaw gas-tester, method of determining the percentage of fire-damp with, 130 et

aeq., 726. Shepherd Mountain, Mo., specular iron-ores of [59].

792 Index,

Ship-plates, character of steel for, 115.

Siemens, F., on luminoas and non-lnminoos flames [682].

Siemens gas-prodacer, 371 ; composition of gas, 374.

Silica in solution in seas, rivers and thermal springs, 307.

Silicates of calcium, 11.

Silico-alnmino ferrites of odcinm, behavior of, in water, 13.

Silicon: in open-hearth process, 390, 462, 479: percentage of, in iron and steel at Westanfors blast-furnace, Sweden, 280; percentage in steel Ax>m Swedish Bes- semer works, 273 et aeq.

Silurian limestones: gold in, 89: magnetic iron-ores in, 61.

Silver: in Appalachian areas, 87; association of, with gold, 86; in carboniferous limestones, 89; in detrital deposits, 92; genesis of deposits, 92; gold-silver alloys, 117, in Mesozoic rocks, 90: metallurgy of, in America, 338; in pal- ozoic rocks, 89 ; production of, in the United States, 86; in Tertiary rocks, 91.

Silver-bearing ores of Ouachita uplift, 206.

Silver City lead- and zinc-mines. Montgomery county, Ark. [206].

Silver-mills (see also stamp-mills) : American improvements in, 339.

Silver-mines: Colorado: Clear Creek county: Seven-Thirty, 740; Monttma: Deer Lodge county ; Granite Mountain, 87; Utah : Beaver county; Horn Silver, 87, 91; Summit county: Ontario, 87, 90 ; Germany: Freiberg; Himmelfshrt [227].

Silver-ores, treatment of: Patera process, 340 ; Patio process [339].

Sireuil iron-works, France [105].

Skey, W., on deposition of gold, 754.

Slag-cements, 20 ei aeq.

Slag-pots, improved, 574

Slags: analyses of, in open-hearth process, 411 et $eq,; analyses of, from Swedish Bessemer blast-furnaces , 275, 277, 669; bricks, 575; from Swedish Bessemer blast furnaces, 275 ei $eq.

Slate, graphitic, enclosing gold-bearing quartz veins, 314.

Smelting, cupola, 331.

Smelting- works (see also gold-mills, silver-mills and stamp-mills) : CMorad : Ara- pahoe county ; Argo, Boston and Colorado, 333, 334 ; Denver, Grant and Omaha [578], 657; Pueblo county: Philadelphia [576]; Mitsouri: Madison county; Mine La Motte, 676; Montana: Silver Bow county: Butte, Anaconda [329, 333. 575] ; Parrott, 330, 334 [575, 576]; New Jersey: Hudson county; Orford [332], 334, 340; Utah: Salt Lake county ; Germania [329].

Smithsonite (" dry-bone ") deposits in southwest Wisconsin [559], 564 [574].

Solubility of hydraulic materials, 6 et aeq.

Solution of hydraulic products, laws of, 5.

Sorby, Dr. H. C, methods for microscopic study of metals, 246 et $eq, ; on theory of hydrothermal fusion, 741 et aeq.

South Dakota, geological relation of Black Hills to Ozark and Wisconsin uplifts,

Southwark Foundry and Machine Co., Philadelphia, Pa. [710], 712 ; blowingengines built by, 721.

Spain, Almaden, quicksilver-deposits of, 85.

Specific gravity of gold in gold-silver alloys, 117.

Spence roasting-furnace, 330.

Spilsbuby, K G., remarks in discussion of Mr. Case's paper on the Bertha zinc- mines, 696.

Spitzkasten, 235, 649.

itzlutie (syphon-classifier), 227, 648.

Stamp-mills : Colorado : Pitkin county, Holden, 659 ; Mtekigan : Lake Superior re- gion, 323 et aeq., 647 et aeq. ; Utah; Summit county, ICarsac [328], 340, 359, 658 ; Ontario, 659.

Index. 793

Statistics, mining and mineral, 95.

Steel (See also Bessemer steel) : accidental defects in, 258 ; alaminam, 114 ; analyses of Bessemer steel rolls, 109; analysis of ingots, 107 et ieq.; analyses of Martin, 117; blister-steel, 254; blow -holes in, 671 ; boiler-plates, defects in, 106; bridge metal, 115: causes of defects in soft steel, 110; centrifogaled," 674; conse- quences of segregation, 106 ; constituents of carbon, 250 ; crystallization of, 546; duplex process 680; effect of temperature in casting, 662; microstructure of, 546 ; moment of segregation, 108 ; Mnkai on -microstructure of manganese [259, 265] ; order of segregation of the principal elements of, 107 ; on the rela- tion between physical properties and microstructure, f51 ; physical tests of Mar- tin, 116; preparation of, for microscopic examination, 250; results of microsco- pic examination, 250 ; rotator for ingots of, 673 ; segregation and its consequences in ingots of, 105 et aeq.; ship-plates, 115; tests of, 106 et aeq,; wolfram, early manufacture of [237].

Steel rails : microstructure of, 548 ; physical tests of, 550.

Steel-works (See also Bessemer steel-works). Maryland : Baltimore county; Sparrow's Point Maryland Steel Co., 721. Penntylvania : Allegheny county ; Homestead, 720; Dauphin county; Steelton, 345. England: Port Clarence, 115 ; Sheffield, Sir Henry Bessemer & Co., 267 ; France : Creusot [491] ; La Louvire [108] ; Terre-Noire [481] ; Belgium: Luxemburg; Dudelingen, 691.

Steitz syphon [658],

Stetefeldt, C. a. : remarks in discussion of Mr. Douglas's paper on American improvements and inventions in ore-crushing and concentration, 659 ; of Mr. Louis's paper on the specific gravity of gold in gold-silver alloys, 724.

Stetefeldt ftirnace, 328, 659.

Stirling boilers at No. 3 colliery, Oneida, Pa., 588 et uq.

St. Louis gold-mine, Gilpin county, Colo., 313.

Stop-line lead- and zinc-mine, southwest Wisconsin [559].

Structural relations of iron-ore deposits in the United States, 57 et ieq.

Sturtevant mill [324].

Sulphuric acid, effect of, on hydraulic material, 27.

Sulphur in open-hearth process, 446.

{Nummary of American Improvements and Inventions in Ore-crushing and ConcentraOmi and in the Metallurgy of Copper, Lead, Gold, Silver Nickel, Aluminum Zinc, Mer- cury, Antimony and Tin (DouGLAs) [xv], 321 ; discussion, 647.

Summit gold district, Colorado [92].

Sweden, Bessemer process in, 265.

Swelling by slacking of hydraulic materials, 9.

Syphon-classifier {Spitzlutte), 2S7, 648.

Tables : of aureole-heights as fire-damp indicators, 166 ; of comparative results of

Chesneau fire-damp indicator and of chemical analysis, 168, 170. Tamm, Dr. A., analyses of Bessemer pig-iron by, 277. Teall, J. J. Harris, on theory of hydrothermal fusion of granitic rocks, 742. Temperature: effects of, in casting steel, 6; effects of, on cement, 36; influence

on blow-holes in steel ingots, 272 ; of lava in crater of Kilauea, 744 ; regulation

of, in open-hearth process, 387. Tempering- water for cement, 35.

Tennessee : copper-deposits [75] ; occurrence of zinc [81], Terre-Noire iron-works, France, 105 et seq. 268 [481, 661]. Terrible lead-mine. Clear Creek county, Colo. [80],

Tertiary rocks : gold and silver in, 91; iron-ores in, 62; lead and zinc in, 83. Tests : of hydraulic materials, 3 et seq. ; for mechanical properties of mortars, 49 ;

of open-hearth steel, 352 ; physical, of Martin steel, 116 ; physical, of steel

794 Index.

rails, 550; relating to alaminate of calcinm, 52; for setting of mortars 40; of small anthracite coals for furnace, 603; tensile, of steel plates, 114.

Tests of Hydraulic Materials (Le Chatelier) [xv], 3.

Texas : geology of Organ Mountains, 182 ; magnetic iron-ores [60] ; manganese-ores of [68] ; occurrence of copper in [77].

Thompson, H. A., on gold- veins of Victoria, Australia, 754.

Thumb-test of cement, 39.

Timbering drifts at Bertha zinc-mines, Va., 529 et seq.

Tin-ores: deposits of, in the United States, 71; in the Black Hills, 71; in Bolivia, 72; treatment of, in the United States, 343.

Titaniferous iron-ores: in Adirondack region, 58; in the Rocky Mountains, 60.

Topaz crystals, analysis of, 240.

Treadwell gold-mill, Alaska [330].

Trotz, Emanuel, and Moen, Philip W., translation of Akerman's paper on the Besse- mer process in Sweden, by, 265.

Ulrich, Prof. George: on lava-dikes of Bendigo reefs, 768; on relation between

dikes and veins in Australian gold-field [756]. United States, geological distribution of useful metals in the, 53, 732. Uplifts of Mississippi Valley, correlation of, 622. Utah : copper-ores, 76; occurrence of tiemannite in [85].

Valves, open-hearth furnace, 368. Vanner, 327.

Van Hise, Prof.: on Baraboo quartzite ranges of central Wisconsin, 625; on erup- tive rocks of Wisconsin Island, 184 ; on origin of iron-ore deposits, 64 et seq. Ventilation of mines, 120.

Verde copper-mines, Yavapai county, Ariz. [334]. Vermont, copper-deposits of [75]. Vesuvius, volcanic action at, 744, 749, 768.

Virginia: iron-ores of, 59; manganese-ores of [68]; occurrence of zinc [81]. Von Dusko lead- and zinc-mine, southwest Wisconsin [559]. V ore-separator, 327, 648.

Wagner, John B., experiments by, in burning small anthracite coals, 582.

Wagner (McFeeley) lead- and zinc-mines, southwest Wisconsin [559].

Ward, J. Clifton, on hydrothermal fusion of granitic rocks, 742.

Washington : gold-deposits, 92 ; bog iron-ores [63].

Water-gas: detection and measurement of, with hydrogen-oil lamp, 615; in open- hearth process, 367, 384.

Water-jacketed furnaces, 331.

Webb City, Jasper county. Mo.: thickness of ore-horizons at, 188; zinc-mines [178], 190 ei seq.

Wedding, Dr. H. : remarks in discussion of Mr. Campbell's paper on the open- hearth process, 691 ; on microscopic metallography, 246 et seq.

Wellman gas-producer, 371.

Wellman Iron and Steel Co., 681.

Werth, J., onmicrostructure of metals, 249 et seq,

Westanfors blast-furnace, Sweden, 275 et seq.

West Virginia, iron-ore deposits of [61].

Whelpley and Stover shaft calcining furnace, 328.

White-Howell furnace, 329.

Whitney, Prof. J. D., on geology of Wisconsin, 621 et seq.

Wilkinson, experiments by, on precipitation of gold [753].

Index. 795

WiNSLOW, Arthur: remarks in discassion of Mr. Emmons's paper on geological

distribution of nsefal metals in the United States, 735; of Dr. Jenney's paper

on the lead- and zinc-deposits of the Mississippi Valley, 634. Wisconsin: Baraboo axis of, 624; faalts in lead -regions, 625; iron-ores of [62];

lead-regions, 172 et seq., 621; lead- and zinc-deposits [81], 559; lead and zinc

regions, 622; mineral deposits, 558. Wisconsin Island (uplift): geology of, 181, 623; lead- and zinc-ores, 208; location

and description of, 172, 176. Wisconsin Lead and Zinc Co., Helena, Wis. [570], 632, 660; mines of southwest

Wisconsin [559], Wolframite, character of, and etymology of name, 237. Wolfram-ore, remarkable deposit of, in the United States, 236. Wolfram -steel, early manufacture of, in Austria [237]. Wolfs safety-lamp (benzine) [147], 148.

Wollastonite, production of, against walls of blast-furnace, 11. " Wootz," or Indian Steel, 236. Wrought-iron (see also iron), microstructnre of, 260. Wurtz. Prof. H., theory of deposition of gold by, 752. Wyandotte, Mich., Bessemer steel -works at, 665. Wyoming: geological relation of hilly regions to Ozark and Wisconsin uplifts, 182;

iron-ores of [60]. Wythe Lead and Zinc Co., Austinville, Va., 723 ; zinc-mines of, 511.

Ziervogel process for extraction of silver [334].

Zinc : deposits in crystalline rocks, 80 ; deposits of, in the United States, 79 et aeq. ; genesis of deposits, 83; metallurgy of, American improvements in, 342; in palaeozoic rocks, 80; production in the United States, 79.

Zinc Carbonate Co.'s lead- and zinc-mines and mill, southwest Wisconsin [559].

Zinc-mines (See also lead- and zinc-mines): Arkansas: Marion county, 187; New Jersey: Sussex county; Mine Hill, 342; Pennsylvania: Blair county; Tyrone [697] ; Liincaster county ; Landisville [697] ; Saucon Valley ; Bethlehem Zinc Co. [697]; Virginia: Wythe county; Austinville, Wythe Lead and Zinc Co., 511; Bertha, 511; Manning and Squier, 5il; Wisconsin: Lafayette county; Dry-Bone [560] ; Raisbeck [559], 631.

Zinc-ores: character of, at Bertha, Va., 513; deposits near Bethlehem, Pa., 81; of the Mississippi Valley, 172 ei seq., 621, 622; of southwest Wisconsin, 564 ; treat- ment of, 661 ; of Wythe county, Va., 723.

Zinc-works: Illinois: Lake county; Lanyon Zinc Oxide and Paint Co. [565]; Neto Jersey : Essex county ; New Jersey Zinc Co., 342 ; Wisconsin : Iowa county ; Mineral Point Zinc Oxide Co. [565].

Zirkel, F., on hydrothermal fusion of granitic rocks, 742.

A

If]

Munner Earth Soences Library

h

D gitized by

t

i:

-

t

4