The Manufacture and Properties of IRON AND STEEL BY HARRY HUSE CAMPBELL GENERAL MANAGER, THE PENNSYLVANIA STEEL COMPANY; S. B., MASSACHUSETTS INSTITUTE OF TECHNOLOGY ; MEMBER AMERICAN INSTITUTE OF MINING ENGINEERS ; MEMBER AMERICAN SOCIBTV OF CIVIL ENGINEERS ; MEMBER IRON AND STEEL INSTITUTE OF GREAT BRITAIN SECOND EDITION REVISED NEW YORK THE ENGINEERING AND MINING JOURNAL *6i BROADWAY LONDON zo BUCKLERSBURY 1904 Engineering- - Library COPYRIGHT, 1896, BT THE SCIENTIFIC PUBLISHING COMPANY COPYHIGHT, 1903, BY THE ENGINEERING AND MINING JOURNAL COPYRIGHT, 1904, BY THE ENGINEERING AND MINING JOURNAL To ALL THOSE, FAMOUS OR OBSCURE, WHO, BY THE FURNACE, IN THE SHOP, OR AT THE DESK, ARE JOINING HAND AND BRAIN TO SOLVE THE PROBLEMS OF THE METALLURGIC ART, THIS VOLUME is FRATERNALLY DEDICATED. PREFACE TO SECOND EDITION. There are many engineers who wish a brief statement of the art of making steel. It is quite impossible to do this and at the same time to discuss the metallurgical details, for this involves much shop language that is not understood by any one except the metal- lurgist. The great electrician whose genius has been crowned with the laurels of two hemispheres referred to the first edition of this book and laughingly, but earnestly, declared that the chapter on the open-hearth was too abstruse for his intellect, while an unedu- cated open-hearth melter told me he had learned, from that same chapter, how to build a furnace, how to run it, and how to make a good livelihood. The melter understood my language, but to Edi- son it was a foreign tongue. In this edition I have tried to give in Part I a sort of Introduc- tion for those who are not metallurgists. It does not pretend to give all the qualifying conditions, but simply the main principles. Part II embraces the ground covered by the first edition of Struc- tural Steel, but many chapters have been entirely rewritten and a great deal of new matter added. Much of the text relating to the chemical history of the open-hearth furnace has been condensed from certain papers which I contributed to the Trans. Am. Inst. Mining Engineers, Vol. XIX, pp. 128 to 187; Vol. XX, pp. 227 to 232, and Vol. XXII, pp. 345 to 511, and 679 to 696, while por- tions of Chapters XVI, XVII and XVIII appeared in the Trans. Am. Soc. Civil Engineers, April, 1895. In many cases the present book is an amplification of previous work. The experiments and investigations have been conducted at the works of The Pennsyl- vania Steel Company, of Steelton, Pa., and all the details of manu- facture and treatment have been under my direct observation. In Part III I have entered into a more comprehensive compari- son of the industrial situation and have compared the salient points of foreign and American practice. Each country has something to learn from every other. There are still many small economies VI PREFACE TO SECOND EDITION. to effect in the art; there will be a constant cheapening as the cost of all supplies and of transportation is lowered by the natural progress in engineering skill ; there are certain important improve- ments that are in plain view; and there may be still more radical changes not yet foreseen. Every dollar taken from the cost of a ton of steel increases the consumption by opening new markets; by rendering possible, for instance, the extension of railways and telegraphs to the uttermost corners of the earth, and in this way the metallurgist becomes not only a giver of dividends to his em- ployer, but a philanthropist whose benefactions reach to the valleys of the Himalayas and to the sources of the Nile. I have compared at some length the condition of the industry in each separate country. These descriptions of the various dis- tricts or provinces are not intended as complete investigations. It would be impossible for instance to describe the American districts so fully that every engineer and metallurgist of our country would find all the information he might wish, or even find a record of all that he already knows. It would also be impossible to tell an Eng- lish engineer much about those parts of his own country with which he is acquainted. It may be possible, however, to give some facts for the benefit of travelers; to clear the way for a foreigner visit- ing America, or an American visiting other lands. It is for this purpose only that these articles have been written and their end will be accomplished if they furnish certain fundamental facts on which to base such a journey. Some readers might prefer that less space should be devoted to theoretical matter and more to descriptions and drawings of fur- naces and apparatus, but in my opinion the place for such informa- tion is in the trade periodicals. It takes so long to print a book like this that the drawings are antiquated when the issue appears, and every year that it stands upon the shelf it becomes more and more a catalogue of discarded devices, while on the other hand the fundamental principles of metallurgy remain the same from year to year, and their value knows no depreciation. A book just issued in England refers very courteously to the former edition of this work, but states that little information is given concerning the practical details of operation. That same book sets forth that an open-hearth furnace is charged by putting the pig-iron in first ; that in a twenty-five-ton furnace not over nine men can be employed, even when there are doors on both sides, and PREFACE TO SECOND EDITION. VU that with rapid work it takes two hours to charge a heat. Now those figures are true for the district with which that writer was familiar, but in America the pig-iron is put in last, while at Steel- ton on a furnace of the size mentioned we use twice the number of men and with good scrap finish the work by charging, by hand labor only, in a period ranging from thirty minutes down to eleven minutes. Of equal value is much of the so-called practical infor- mation given in metallurgical treatises. In many places in these pages I have tried to give credit to the many friends who have rendered assistance in divers ways. It only remains to thank them as a whole, both those at home and abroad, for aiding in this work which has been accomplished in the intervals of what I trust is not otherwise an entirely idle life. H. H. CAMPBELL. Steelton, Pa., December, 1902. PREFACE TO THIRD EDITION. It is only a few months since the second edition was issued, and it -is a source of satisfaction to find that the supply is so soon exhausted. It is also a source of regret since any such book must contain mistakes, and some time must elapse before they are all discovered. Every attempt has been made to corroborate the ac- counts given of the iron industry in foreign countries and copies of the book have been sent to many foreign metallurgists with duplicate sheets for alterations. The replies indicate that the de- scriptions were correct; in the case of Germany there were certain errors in statistics, but a personal visit to Westphalia gave an opportunity to get more accurate information, and the chapter on that country has been revised accordingly. This visit also gave a chance for a further study of the practical details of the basic Bessemer process and furnished material for some changes in the treatment of that subject. Some late statistics have been added, but as a rule no attempt has been made to bring all the records of production down to date. No other book has ever published in detail the output of each producing district, and the only object in doing so was to com- pare the relative importance of the different parts of each country. Much of the information was collected with great difficulty, and it is impossible to get later figures in each case. I have, however, looked over various statistics from different sources, but fail to find any change in the relative conditions of the different countries or the separate districts. For purposes of comparison, therefore, it is unnecessary to get the very latest data. H. H. CAMPBELL. STEELTON, PA., October, 1903. TABLE OF CONTENTS PART I. The Main Principles of Iron Metallurgy. PAGE The making of pig-iron 3 The making of wrought-iron 5 A definition of steel Q The making of crucible steel 7 The acid Bessemer process 7 The basic Bessemer process . 9 The open-hearth furnace H The acid open-hearth process 12 The basic open-hearth process 15 Segregation 17 The influence of hot working on steel 18 The effect caused by changes in the shape of the test-piece ..... 19 The influence of certain elements upon steel 21 Specifications on structural material 24 Welding , 26 Steel castings . . 26 Inspection s 28 PART II. The Metallurgy of Iron and Steel. CHAPTER I. THE ERRANCY OF SCIENTIFIC KECORDS. SECTION la. Difficulties in obtaining comparative data 37 Ib. Errors in chemical methods 39 Ic. Necessity of uniformity in chemical work . . . . 42 Id. Variations in the parallel determinations of practicing chemists , y= .' , J . * 44 Ie. Methods of deducing metallurgical laws ...... 46 CHAPTER II. THE BLAST FURNACE. SECTION Ha. Iron ores used for smelting . 48 lib. Fuel used for smelting 51 lie. Flux 52 lid. Construction and operation of furnaces 5ft lie. Chemical history 64 Ilf. Utilization and waste of heat 7ft ix TABLE OF CONTENTS. PAGE SECTION Ilg. Metallurgical conditions affecting the nature of the iron 82 Hh. Blast 85 a the amount of air required b the heating of the blast c the water vapor in the atmosphere Hi. Tunnel head gases 90 IIj. Utilization of tunnel head gases 103 a use of potential heat in 'stoves and boilers b use of sensible heat in stoves and boilers c use in gas engines d preheating the air going to the stoves Ilk. Relation between the chemical and physical qualities of cast-iron . . ... . ..... 124 CHAPTER III. WROUGHT-!RON. SECTION Ilia. General description of the puddling process .... 129 Illb. Effect of silicon, manganese and carbon 130 IIIc. History of sulphur and phosphorus 132 Hid. Effect of the temperature of the furnace . ... V 133 Hie. Effect of work upon wrought-iron 135 Illf. Heterogeneity of wrought-iron 136 Illg. Conditions affecting the welding properties .... 139 CHAPTER IV. STEEL. SECTION IVa. Definition of steel . 140 IVb. Cause of failure of certain proposed definitions . . . 142 FVc. The American nomenclature of iron products .... 146 CHAPTER Y. HIGH-CARBON STEEL. SECTION Va. Manufacture of cement and crucible steel 147 Vb. Chemical reactions in the steel-melting crucible . . . 148 Vc. Chemical specifications on high steel ....... 149 Vd. Manufacture of high steel in an open-hearth furnace . 151 CHAPTER VI. THE ACID-BESSEMER PROCESS. SECTION Via. Construction of a Bessemer converter 155 VIb. Chemical history of an acid- Bessemer charge .... 158 Vic. Variations in the chemical history due to different contents of silicon 160 VId. Swedish Bessemer practice 161 Vie. History of the slag in the converter .... ;' . . 162 Vlf. Calorific history of the acid- Bessemer converter . . . 164 VIg. Use of direct metal . .... ., ;. 168 Vlh. Use of cupola metal . . . .... ..... .... 170 Vli. Certain factors affecting the calorific history . \2^'. . 171 VIj. Kecarburization . 174 TABLE OF CONTENTS. CHAPTER VII. THE BASIC-BESSEMER PROCESS. SECTION Vila. General outline of the basic-Bessemer process . . . 175 Vllb. Elimination of phosphorus iff VIIc. Amount of lime required 173 Vlld. Chemical reactions in the basic converter 179 Vile. Elimination of sulphur in the basic converter . . . 180 Vllf. Calorific equation of the basic converter 183 VEIg. Recarburization lg^ CHAPTER VIII. THE OPEN-HEARTH FURNACE. SECTION VEIIa. General description of a regenerative furnace . . . 186 Vlllb. Quality of the gas required 187 VIIIc. Construction of an open-hearth furnace 188 Vllld. Tilting open-hearth furnace 205 Vllle. Method of charging 211 VIIH. Ports 213 Vlllg. Valves 214 Vlllh. Regulation of the temperature 217 VHIi. Calorific equation of an open-hearth furnace . . . 218 CHAPTER IX. FUEL. SECTION IXa. The combustion of fuel 233 IXb. Producers 237 a bituminous coal b hard coal IXc. Miscellaneous fuels 245 a natural gas b petroleum c water gas IXd. Heating furnaces 249 a soaking pits b regenerative furnaces c reverberatory furnaces d continuous furnaces IXe. Coke ovens 256 IXf. Coal washing 263 IXg. General remarks on fuel utilization 267 CHAPTER X. THE ACID OPEN-HEARTH PROCESS. SECTION Xa. Nature of the charge in a steel melting furnace ... 269 Xb. Chemical history of a charge during melting .... 270 Xc. Chemical history of a charge after melting 272 Xd. Quantitative calculations on slags 273 Xe. Reduction of iron ore when added to a charge ..... 274 Xf. Pig and ore process 275 Xg. Conditions modifying the product 277 TABLE OF CONTENTS. SECTION Xh. History of sulphur and phosphorus ....... 278 Xi. Method of taking tests ............ 279 Xj. Recarburization ............ 279 Xk. Advantages of the process in securing homogeneity . . 281 . CHAPTER XI. THE BASIC OPEN-HEARTH PROCESS. SECTION XIa. Construction of a basic open-hearth bottom .... 282 Xlb. Functions of the basic additions ........ 283 XIc. Use of ore mixed with the initial charge ..... 284 Xld. Chemical history of basic open-hearth charges when no ore is mixed with the stock ........ 285 Xle. Elimination of phosphorus during melting ..... 286 Xlf. Composition of the slag after melting . . . *" : * 286 Xlg. Relative value of different limes ........ 287 Xlh. History of basic open-hearth slags ....... 288 Xii. Automatic regulation of fluidity in slags ..... 290 XI j. Determining chemical conditions in slags ..... 292 Xlk. Elimination of sulphur .......... ', . 294 XII. Removal of the slag after melting ..... ; ** . 297 Xlm. Automatic formation of a slag of a given composition 298 XIn. Recarburization and rephosphorization ...... 299 CHAPTER XII. SPECIAL METHODS OF MANUFACTURE AND SOME ITEMS- AFFECTING THE COSTS. SECTION XHa. The manufacture of low phosphorus acid open-hearth steel at Steelton ..... ../.''. . . . 302 XHb. The pig and ore basic process ... ...... 806 XIIc. The Talbot process . . . . V ........ 310 XHd. The Bertrand Thiel process .......... 315 XHe. The heat absorbed by the reduction of iron ore . . . 319 Xllf. The amount of ore needed to reduce a bath of pig- iron ............. .... 324 XHg. The gain in weight by reduction of iron ore . . . . 329 XHh. The increment in cost due to waste in the Bessemer process ................ 333 XHi. The increment in the open-hearth process ..... 335 XII j. The increment in the rolling mills ....... 336 XHk. The duplex process ............. 337 CHAPTER XIII. SEGREGATION AND HOMOGENEITY. SECTION XIHa. Cause of segregation ............ 34Q XIHb. Examples of segregation in steel castings .... 344 XIIIc. Examples of segregation in plate ingots ..... 345 XHId. Attainment of homogeneity in plates ...... 347 Xllle. Homogeneity of acid rivet and angle steel .... 354 XIHf. Homogeneity of high-carbon steel ..... ... 357 TABLE OF CONTENTS. Xiii PAGE SECTION XJIIg. Homogeneity of acid open-hearth nickel steel . . . 359 XHIh. Investigations on Swedish steel 362 CHAPTER XIV. INFLUENCE OF HOT WORKING ON STEEL. SECTION XlVa. Effect of thickness upon the physical properties . . 364 XlVb. Discussion of Riley's investigations on plates . . . 365 XIVc. Amount of work necessary to obtain good results . 366 XIYd. Experiments on forgings 370 XlVe. Tests on Pennsylvania Steel Company angles . . . 371 XlVf. Comparison of the strength of angles with that of the preliminary test-piece 373 XlVg. Physical properties of The Pennsylvania Steel Com- pany steels of various compositions ...... 374 XlVh. Properties of hand and guide rounds 375 XI Vi. Effect of variations in the details of plate rolling . 376 XIVj. Physical properties of plates and angles 378 XlVk. Effect of thickness on the properties of plates . . . 379 CHAPTER XV. HEAT TREATMENT. SECTION XVa. Effect of annealing on rolled bars 381 XVb. Annealing bars rolled at different temperatures . . 385 XVc. Effect of annealing on bars rolled under different conditions of work and temperature 386 XVd. Effect of annealing on plates of the same charge which showed different physical properties . . . 387 XVe. Effect of annealing eye-bar flats 389 XVf. Methods of annealing 389 XVg. Further experiments on annealing rolled bars . . . 391 XVh. General remarks on the determination of temperature 392 XVi. Definition of the term "critical point" 394 XVj. Definition of the different structures seen under the microscope 403 XVk. Effect of work on the structure of soft steel and forging steel 409 XVI. Effect of work upon the structure of rails 410 XVm. Effect of heat treatment upon the structure of cast- ings 412 XVn. Effect of heat treatment upon the structure of rolled material 416 XVo. Theories regarding the structure of steel 417 CHAPTER XVI. THE HISTORY AND SHAPE OF THE TEST-PIECE. SECTION XVIa. Difference in physical properties between the surface and the interior of worked steel ' 420 XVIb. Physical properties of strips cut from eye-bar flats . 421 XVIc. Comparison of longitudinal and transverse tests . . 422 xiv TABLE OF CONTENTS. PAGE 424 424 426 427 431 434 SECTION XVId. Comparison of parallel and grooved tests . . . XVIe. Effect of shoulders at the ends of test-pieces . . XVIf. Use of the preliminary test-piece as a standard . XVIg. Comparative properties of rounds and flats . . . XVIh. Effect of diameter upon the physical properties . XVIi. Influence of the width of the test-piece .... XVIj. Influence of the length of the test-piece 435 XVIk. Tests on eye-bars 440 XVII. Effect of rest after rolling 448 XVIm. Errors in determining the physical properties . . . 448 XVIn. Effect of variation in the pulling speed 452 CHAPTER XVII. THE INFLUENCE OF CERTAIN ELEMENTS ON THE PHYSICAL PROPERTIES OF STEEL. SECTION XVIIa. The quantitative valuation of alloyed elements . . 455 PART I. EFFECT OF CERTAIN ELEMENTS AS DETERMINED BY GENERAL EXPERIENCE AND BY THE USUAL METHODS OF INVESTIGATIONS. SECTION XVIIb. Influence of carbon 456 XVIIc. Influence of silicon 456 XVIId. Influence of manganese 463 XVIIe. Influence of sulphur 467 XVIIf. Influence of phosphorus 469 XVIIg. Influence of copper 472 XVIIh. Influence of aluminum 475 XVIIi. Influence of arsenic 478 XVIIj. Influence of nickel, tungsten and chromium . . . 479 XVIIk. Influence of oxide of iron 480 PART II. EFFECT OF CERTAIN ELEMENTS AS DETERMINED BY SPECIAL MATHEMATICAL INVESTIGATIONS. SECTION XVIII. Investigations by Webster 482 XVIIm. Investigations by The Pennsylvania Steel Company 486 XVIIn. Quantitative valuation of the elements by the method of least squares 487 XVIIo. Application of the method of least squares .... 496 XVIIp. Effect of carbon, manganese, phosphorus and iron . 497 XVIIq. Values of carbon and phosphorus 500 PART III. EFFECT OF CARBON, MANGANESE AND PHOSPHORUS UPON THE; TENSILE STRENGTH OF IRON, AS DETERMINED BY SPECIAL MATHEMATICAL INVESTIGATIONS. SECTION XVIIr. Values of carbon, manganese, phosphorus and iron in a new series of acid steels 505 TABLE OF CONTENTS. XV PACK SECTION XVIIs. Values of carbon, phosphorus and iron in acid steel when manganese is neglected 507 XVIIt. Values of carbon, manganese, phosphorus and iron in a new series of basic steels 51g XVIIu. Values of carbon, manganese, phosphorus and iron in basic steel, as determined from the old and the new series combined 517 XVIIv. Meaning of the term "pure iron" 523 XVIIw. Synopsis of the argument and conclusions .... 524 CHAPTER XVIII. CLASSIFICATION OF STRUCTURAL STEEL. SECTION XVIIIa. Influence of the method of manufacture .... 529 XVIIIb. Chemical specifications 532 XVIIIc. Use of soft steel in structural work 535 XVIIId. Tests on plates 537 XVIIIe. Standard size of test pieces 538 XVIIIf. The quench test 540 XVIIIg. Classes of steel proposed 541 CHAPTER XIX. WELDING. SECTION XlXa. Influence of structure on the welding properties . . 583 XlXb. Tensile tests on welded bars of steel and iron . . . 584 XIXc. Influence of metalloids upon welding 588 CHAPTER XX. STEEL CASTINGS. SECTION XXa. Definition of a steel casting 591 XXb. Methods of manufacture . 592 XXc. Blow-holes . 594 XXd. Phosphorus and sulphur in steel castings 595 XXe. Effect of silicon, manganese and aluminum .... 595 XXf. Physical tests on soft steel castings 596 XXg. Physical tests on medium hard steel castings . . . 600 PART III. The Iron Industry of the Leading Nations. CHAPTER XXI. FACTORS IN INDUSTRIAL COMPETITION. SECTION XXIa. The question of management 603 XXIb. The question of employer and employed 614 XXIc. The question of tariffs 623 CHAPTER XXII. THE UNITED STATES. SECTION XXIIa. General view 629 XXIIb. Coal 639 XXIIc. Lake Superior 647 XVI TABLE OF CONTENTS. SECTION XXIId. Pittsburg 657 XXIIe. Chicago . . 664 XXIIf. Alabama . 668 XXIIg. Johnstown 675 XXIIh. Steelton 675 XXIIi. Sparrow's Point 679 XXIIj. Cleveland 684 XXIIk. Colorado 687 XXIII. Eastern Pennsylvania 688 XXIIm. New Jersey, New York and New England .... 689 CHAPTER XXIII. GREAT BRITAIN. SECTION XXIIIa. General view 692 XXIIIb. Northeast Coast 700 XXIIIc. Scotland 710 XXIIId. South Wales 714 XXIIIe. Lancashire and Cumberland 717 XXHIf. South Yorkshire 721 XXIIIg. Staffordshire 722 XXIIIh. North Wales 723 XXIIIi. The Eastern Central District: Lincoln, Leicester and Northampton 723 CHAPTER XXIV. GERMANY. 'SECTION XXIVa. General view 727 XXIVb. Lothringen and Luxemburg 730 XXIVc. The Ruhr 742 XXIVd. Silesia 751 XXIVe. The Saar 755 XXIVf. Aachen 756 XXIVg. Ilsede and Peine 757 XXIVh. Saxony . . 758 XXIVi. Siegen . . . 758 XXIVj. Osnabruck . , 759 XXIVk. Bavaria 760 XXIVI. The Lahn . ... ....-;. ....... 760 XXIVm. Pommerania 760 XXI Vn. Other districts 760 CHAPTER XXV. FRANCE. SECTION XXVa. General view 762 XXVb. The East . . ... . 764 XXVc. The North 767 XXVd. The Centre 769 XXVe. The (South 770 XX Vf. The Northwest and the Southwest ...... 771 TABLE OF CONTENTS. CHAPTER XXVI. KUSSIA. PAGE SECTION XXVIa. General view 772 XXVIb. The South 773 XXVIc. The Urals ....". 779 XXVId. Poland 782 XXVIe. The Centre 733 XXVIf. The North 734 CHAPTER XXVII. AUSTRIA. SECTION XXVIIa. General view 785 XXVIIb. Bohemia 788 XXVIIc. Moravia and Silesia 789 XXVIId. Styria - 791 XXVIIe. Hungary 794 CHAPTER XXVIII. BELGIUM. 796 CHAPTER XXIX. SWEDEN. 803 CHAPTER XXX. SPAIN. 812 CHAPTER XXXI. ITALY. 816 CHAPTER XXXII. CANADA. 818 CHAPTER XXXIII. STATISTICS. 821 APPENDIX. Value of certain factors used in iron metallurgy . 838 Content of metallic iron in pure compounds of iron 838 Reactions in open-hearth furnaces 838 Properties of air 838 Comparison of English and metric systems 839 Gravimetric and calorific values 839 INDEX TO TABLES. ERRANCY OF SCIENTIFIC RECORDS. HO. PAGE I- A Variations in pieces of the same rolled bar 38 I-B Errors in the work of different national committees 40 I-C Variations in determinations of Carbon and Phosphorus ... 40 I-D Results on the same steels by The Pottstown Iron Co. and The Pennsylvania Steel Co ; , 44 * % BLAST FURNACE. II-A Slags made by smelting ores without lime 52 II-B Comparison of furnace practice at Middlesborough and Pitts- burg 76 II-C Distribution of calorific energy 77 II-D General equation of the blast furnace 79 II-E Blast furnace slags 85 II-F Specific heat of CO, 0, H and N 90 II-G Temperatures produced by burning 91 II-H Vapor in the atmosphere as affecting the blast furnace ... 95 II-I Volume and composition of tunnel head gases 99 II-J Percentages of CO 2 and O in products of combustion .... 104 II-K Loss of heat by CO in products of combustion 106 II-L Data on products of combustion 108 II-M List of gas engines on blast furnace gas 115 II-N Composition of pig-iron and spiegel 127 WROUGHT-IRON. III-A Elimination of the metalloids in the puddling process .... 133 III-B Analyses of puddle or mill cinder 135 III-C Wrought-iron plates from shear and universal mills .... 136 III-D Requirements on wrought-iron in the United States .... 137 III-E Irregularity of wrought-iron . . 138 DEFINITION OF STEEL. IV-A Effect of quenching different soft steels 144 IV-B Effect of quenching at different temperatures ....... 145 HIGH STEEL. V-A High steels not in accordance with specifications 150 V-B Composition of clippings taken from the top and bottom blooms of each ingot of a high-carbon heat 152 xix XX INDEX TO TABLES. MO. V-C Variations in Swedish steel 153 V-D Variations in one lot of crucible steel rounds ....... 154 ACID BESSEMER. VI-A Chemical history of an acid Bessemer charge 157 VI-B Calculations on weights of Bessemer slags 158 VI-C Manganiferous Bessemer pig-irons and slags 161 VI-D Bessemer steel made from high-manganese pig-iron 162 VI-E Composition of American Bessemer slags 163 VI-F Calorific history of the acid Bessemer converter 167 VI-G Loss of combined iron in cupola slag 170 BASIC BESSEMER. VII-A Metal, slag and gases from the basic converter 180 VII-B Reduction of manganese from slag 181 VII-C High sulphur iron in basic converter 182 VII-D Calorific equation of the basic Bessemer process 184 OPEN-HEARTH FURNACE. VIII-A Distribution of heat in the producer 228 VIII-B Distribution of heat in the furnace 230 VIII-C Distribution of heat in producer and furnace combined . . . 232 FUEL. IX- A Products of combustion of hard and soft coal 23 i IX-B Loss of heat in products of combustion 236 IX-C Heat lost in producer ash 241 IX-D Heat lost by CO 2 in gas 243 IX-E Waste gases from reverberatory furnaces . 253 IX-F Calculations on waste gases from reverberatory furnaces . . . 254 IX-G Coke ovens in England 263 IX-H Otto Hoffman and Semet Solvay ovens in the United States . 263 ACID OPEN-HEARTH. X-A Elimination of metalloids in an open-hearth charge 271 X-B History of metal and slag in an acid furnace 273 X-C Reduction of iron ore 274 X-D Data on open-hearth slag and metal at different periods ... 275 BASIC OPEN-HEARTH. XI-A Composition of slag and metal from seventeen heats .... 285 XI-B Elimination of phosphorus and carbon during melting .... 286 XI-C Relative value of limes with 3.0 and 7.0 per cent, of SiO, . . 287 XI-D Relation between Si0 2 and FeO in basic slags 291 XI-E Maxima and minima in the heats composing Table XI-D . . . 291 XI-F Unstable basic open-hearth slags . 293 INDEX TO TABLES. XXI NO. XI-G Normal basic open- hearth slags ............ 293 XI-H Basic open-hearth slags after melting ... ...... 295 XI-I Basic open-hearth slags before adding recarburizer ..... 295 XI-J Elimination of sulphur by calcium 1 chloride ........ 296 XI-K Detailed data on the use of calcium chloride ....... 296 XI-L Average slag analyses of twenty- seven basic heats ..... 299 CONSIDERATION OF CERTAIN SPECIAL METHODS AND SOME ITEMS AFFECTING THE COST OF MANUFACTURE. XII-A Composition of metal and slag in making transfer steel . . . 305 XII-B Comparison of data in Tables X-B and XII-A ...... .. 306 XII-C Record of "all pig" basic open-hearth heats at Steelton . . . 309 XII-D Reactions in the Talbot process ............ 312 XII-E Rate of production and elimination of sulphur in the Talbot furnace ................... 314 XII-F Representative heats under present practice at Kladno . . . 318 XII-G Oxygen needed for a pig-iron charge ...... .... 325 XII-H Oxygen used in the Talbot furnace ... ....... 326 XII-I Silica in the Talbot furnace ............. 327 XII-J Oxygen in the Talbot furnace ............ 327 XII-K Distribution of the metallic iron in the Talbot furnace . . . 330 SEGREGATION. XIII-A Example of extreme segregation in pipe cavity ...... 344 XIII-B Composition of a twenty-inch steel roll cast in sand .... 344 XIII-C Examples of segregation in plate ingots ........ 345 XIII-D Examples of segregation in large ingots ........ 346 XIII-E Results on plates rolled from ordinary plate ingots .... 348 XIII-F Results on universal mill plates rolled from slabs ..... 349 XIII-G Physical and chemical properties of annealed bars cut from plates rolled from basic open-hearth slabs ....... 350-3 XIII-H Showing that variations in the carbon of the test-pieces given in Table XIII-G are due to analytical errors . . . 354 XIII-I Tests on rounds from different parts of the same heats . . . 355-6 XIII-J Composition of rods from heat 10,168 ......... 357 XIII-K Chemical composition of angles rolled from 26"x24" ingots of acid open-hearth steel .............. 358 XIII-L Distribution of elements in high carbon ingot ...... 359 XIII-M Distribution of elements in high carbon blooms ..... 360 XIII-N Composition of the liquid interior of an ingot ...... 360 XIII-O Homogeneity of acid open-hearth nickel steel ...... 361 XIII-P Segregation in Swedish ingots ............ 362 HOT WORKING. XIV-A Results on different thicknesses of steel plates . . . XIV-B Physical results on plates from different sized ingots XX11 INDEX TO TABLES. NO. PAGK XIV-C Influence of thickness on the physical properties, the per- centage of reduction in rolling being constant 368 XIV-D Influence of thickness upon the physical properties, all pieces being rolled from billets of one size 369 XIV-E Effect of hammering rolled acid open-hearth steel 369 XIV-F Physical properties of thick and thin angles 371 XIV-G Comparison of angles and preliminary test 372 XIV-H Physical properties of steel angles 373 XIV-I Effect of flats finished at different temperatures 375 XIV-J Comparison of hand rounds and guide rounds 375 XIV-K Changes in the properties of plates by variations .in the methods of rolling; classified by preliminary test .... 376 XIV-L Changes in the properties of plates by variations in the methods of rolling; classified by finished plate 377 XIV-M. Comparison of angles and sheared plates 378 HEAT TREATMENT. XV-A Effect of annealing on rounds and flats 382 XV-B Comparison of the Bessemer bars in Table XV-A 383 XV-C Comparison of the open- hearth bars in Table XV-A .... 384 XV- D Effect of annealing acid open-hearth rolled steel bars .... 385 XV-E Effect of annealing bars of different thickness, the percentage of reduction in rolling being constant 386 XV-F Effect of annealing bars of different thickness, all pieces being rolled from billets of one size 387 XV-G Rolled plates which show wide variations in their physical properties are made alike by annealing 388 XV-H Comparative tests of eye-bar steel 389 XV-I Comparison of natural and annealed flat bars 390 XV- J Effect of annealing at about 800 C 391 XV-K Comparison of natural and annealed bars in Table XV-J . . 392 XV-L Theoretical microstructure of carbon steels . . 407 XV-M Microstructural composition of some quenched carbon steels . 407 HISTORY OF TEST-PIECE. XVI-A Comparison of three-quarters-inch rolled rounds in their nat- ural state, and seven-eighths-inch rounds of the same heats turned down to three-quarters-inch 421 XVI-B Properties of test-pieces cut from forged rounds 421 XVI-C Properties of test-pieces cut from rolled flats 422 XVI-D Comparison of eye-bar flats with the preliminary test . . . 423 XVI-E Comparison of longitudinal and transverse tests 424 XVI-F Comparison of parallel and grooved tests 424 XVI-G Comparison of the ultimate strength of two-inch tests with shoulders and eight-inch parallel sided tests 425 XVI-H Comparison of angles with the preliminary test 426 INDEX TO TABLES. XX111 NO. v. PACK XVI-I Comparative physical properties of rounds and flats .... 428 XVI-J Comparative physical properties of round and flat bars in the natural and annealed state 429 XVI-K Physical properties of rounds of different diameters .... 431 XVI-L Effect of changes in the width of the test-piece 433 XVI-M Influence of the width upon the elongation (Barba) .... 435 XVI-N Effect of width upon the elongation (Custer) 435 XVI-0 Influence of the length of the test-piece 436 XVI-P Influence of the length upon the elongation (Barba) . . . 438 XVI-Q Physical properties of eye-bars, classified according to method of manufacture, thickness and width 441 XVI-R Physical properties of eye-bars, classified according to length, width and thickness 443 XVI-S Properties of eye-bars, classified according to length .... 444 XVI-T Proportion of rejections caused by applying a sliding scale of elongation to the eye-bar records in Table XVI-Q .... 446 XVI-U Physical changes in steel by rest after rolling 447 XVI-V Physical properties of the same bars of steel, as determined by different laboratories 449 XVI-W Parallel determinations of the elastic limit by the auto- graphic device and by the drop of the beam 451 XVI-X Effect of the pulling speed of testing machine 453 INFLUENCE OF ELEMENTS. ^VII-A Physical properties of silicon steels 457 XVII-B Influence of silicon upon tensile strength 458 XVII-C Physical properties of steels containing from .01 to .50 per cent, of silicon 459 XVII-D Comparison of low-silicon and high-silicon steels 460 XVII-E Effect of manganese upon the physical properties .... 465 XVII-F Properties of steel with 1.00 per cent, of manganese .... 466 XVII-G Physical properties of forged steel with high manganese . . 468 XVII-H Effect of phosphorus upon the physical properties .... 471 XVII-I Effect of copper upon the physical properties 475 XVII- J Physical properties of aluminum steel 476 XVII-K Effect of aluminum upon the physical properties 477 XVII-L The physical qualities of nickel steef 479 XVII-M Records of heats composing Group 63 in Table XVII-N . . 481 XVII-N List of groups used in determining the effect of certain elements upon the tensile strength of steel 488-90 XVII-O Effect of certain elements upon the strength of steel ... 495 XVII-P Effect of carbon, manganese and phosphorus 499 XVII-Q Values of carbon, manganese, phosphorus and iron obtained by arbitrarily dividing the list in Table XVII-N ... 501 XVII-R Effect of carbon and phosphorus .... 503 XVII-S Ultimate strength of the steels given in Table XVII-N as compared with the results from certain formulae .... 504 X.X1V INDEX TO TABLES. XVII-T Values of carbon, manganese, phosphorus and iron from the normal acid steels in Table XVII-U 506 XVII-U List of groups of acid steels of old and new series .... 510-2 XVII-V Average error of groups in Table XVII-U 512 XVII-W Values of carbon, manganese, phosphorus and iron from the basic steels in Divisions I and II of Table XVII-X ... ,516 XVII-X List of groups of basic steels of old and new series . . . .518-20 XVII-Y Average error of groups in Table XVII-X 522 CLASSIFICATION OF STEEL. XVIII-A Rise in elastic ratio with fall in ultimate strength . . . 536 XVIII-B Calculation of 12 yTfor different diameters 539 WELDING. XIX-A Tensile tests on welded bars of steel and wrought-iron . . . 585-6 XIX-B Welding tests by The Royal Prussian Testing Institute . . . 587 CASTINGS. X-A Comparison of castings and rolled bars 598 XX-B Physical properties of castings of medium hard steel . . . 599 AMERICAN VS. EUROPEAN PRACTICE. XXI- A Miles of railway in operation in 1899 609 UNITED, STATES. XXII-A Production of pig-iron and steel in 1900 by districts . . . 631-2 XXII-B Production of steel from 1867 633 XXII-C Annual production of Bessemer, open-hearth and rail steel in the United States and Great Britain 634 XXII-D Percentage of various kinds of steel made in the United States and Great Britain 634 XXII-E Imports of iron ore 636 XXII-F Productions of coal and coke in 1900 644 XXII-G Output of coal from the principal coal fields in 1900 . . . 645 XXII-H Production of soft coal in Pennsylvania in 1900 and amount ' used for coke 645 XXII-I Coke statistics for Pennsylvania and West Virginia in 1900 646 XXII-J Sources of American ore supply 649 XXII-K Movement of lake ore 652 XXII-L Production of pig-iron and steel in Pennsylvania in 1901 . 658 XXII-M Distribution of works in the Pittsburg district 663 XXII-N Production of pig-iron in Alabama 673 XXII-O Production of ore in Cuba 683 XXII-P Distribution of works in New Jersey, New York and New England .'.-.. .^/v''. .../. ...... 691 GREAT BRITAIN. . XXIII-A Imports of iron ore from different countries 696 XXIII-B Production of coal, ore, iron and steel in 1900 698 XXIII-C Production of pig-iron 699 INDEX TO TABLES. N0 - PAQB XXIII-D Production of iron ore .... O QQ ' Vt7u XXIII-E Imports of iron ore at different ports 700 XXIII-F Iron and steel plants on the Northeast Coast 799 XXIII-G Production of ore and pig-iron and imports of ore on th Northeast Coast 709 XXIII-H Imports of ore on the Northeast Coast 710 XXIII-I Production of pig-iron in Scotland 712 XXIII-J Iron and steel plants in Scotland 712 XXIII-K Production of ore and pig-iron and imports of ore in Scot- land 713 XXIII-L Imports of ore into Scotland 713 XXIII-M Iron and steel plants in South Wales 716 XXIII-N Production of pig-iron and imports of ore on the Bristol Channel 71g XXIII- Imports of ore on the Bristol Channel 717 XXIII-P Iron and steel plants on the West Coast 719 XXIII-Q Production of ore and pig-iron and imports of ore on the West Coast 720 XXIII-R Imports of ore on the West Coast 720 XXIII-S Iron and steel plants in South Yorkshire 721 XXIII-T Production of pig-iron in South Yorkshire 721 XXIII-U Production of ore and pig-iron in Staffordshire ..... 723 XXIII-V Production of ore and pig-iron in Eastern Central England 725 XXIII-W Production of pig-iron in Central England , 725 GERMANY. XXIV-A Production of coal, coke, ore and iron 729 XXIV-B Movement of ore 729 XXIV-C Production of steel 730 XXIV-D Composition of minette ores 733-4 XXI V-E List of works in Lothringen and Luxemburg 741-2 XXIV-F Production of coke in Germany 743 XXIV-G List of works in Westphalia 750 XXIV-H Composition of Silesian ores 753 XXIV-I List of works in Silesia 754 XXIV-J List of works in Saar District 755 XXI V-K Composition of Ilsede ores 757 FRANCE. XXV-A Production of fuel, ore, iron and steel in France in 1899 . 764 XXV-B List of works in the East of France 767 XXV-C List of works in the North of France 768 XXV-D List of works in the Centre of France 769 XXV-E List of works in the South of France 770 XXV-F List of works in the Northwest and Southwest of France . 771 XXVI INDEX TO TABLES. HO. XXVI-A XXVI-B XXVI-C XXVI-D XXVII-A XXVII-B XXVII-C XXVII-D XXVII-E XXVII-F XXVII-G XXVII-H RUSSIA. Imports of iron, steel and fuel Production of coal, ore, iron and steel List of works in South Russia V .. . ^ . . . . . Imports of iron and fuel at St. Petersburg . AUSTRIA. Production of coal, ore and pig-iron in Austria and Hun- gary in 1900 . . r ., Production of steel in Austria List of works in Bohemia Output of the Silesian coal fields List of works in Moravia and Silesia List of works in Styria . Production of coal, ore and pig-iron in Hungary in 1899 . Production of steel in Hungary BELGIUM. XXVIII-A Production of coal, coke, iron and steel in Belgium XXVIII-B List of important blast furnace plants in Belgium PAGE 773 776 779 784 787 787 789 790 791 793 795 795 79,7 798 SWEDEN. XXIX-A Production of coal, ore, iron and steel in Sweden . . . 803 XXIX-B List of works in Sweden . . . ...... v'V. 810 SPAIN. XXX-A Spanish ore production and exports 814 ITALY. XXXI-A Exports of ore from Elba in 1899 817 CANADA. XXXII-A Composition of fuel and ore at Cape Breton 819 XXXII-B Canadian bounty on iron and steel 820 THE IRON INDUSTRY. XXXIII-A Discordant Data in Steel Output in Germany .... 822 XXXIII- B Key to numbers denoting source of statistical information 823 XXXIII- C Production of pig-iron per capita 824 XXXIII-D Pig-iron producing districts of the world 832 XXXIII-E Steel producing districts of the world 833 XXXIII-T? Production of coal, ore, pig-iron and steel in 1900 . . . 834 XXXIII-G Production of coal by the leading nations 834 XXXIII-H Production of iron ore by the leading nations .... 835 XXXIII-I Production of pig-iron by the leading nations 835 XXXIII- J Production of steel by the leading nations 836 XXXIII-.K Production of wrought-iron by the leading nations . . . 836 XXXIII-L Imports and exports of fuel and iron 837 XXXIII-M Import duties on iron staples 838 INDEX TO FIGURES. NO. II-A Blast furnace at Jones & Laughlin's, Pittsburg 60 II-B Bosh construction at Steelton, Pa 61 II-C Bertrand blast furnace top 63 II-D Blast furnace reactions as determined by the temperature . 65 II-E Chemical reactions in blast furnace c . . . . 71 II-F Indicator cards gas and steam engines 119 II-G Oechelhauser gas engine . 121 II-H Koerting gas engine 123 VI-A Section of 18- ton converter, two views 156 VIII-A Bad type of open-hearth furnace 189 VIII-B 40-ton acid furnace at Steelton, Pa., two views 191-2 VIII-C 50-ton Campbell basic furnace at Steelton, Pa., three views . 194-7 VIII-D 30-ton basic furnace at Donnawitz, Austria, six views . . 198-203 VIII-E 50-ton basic furnace at Duquesne, Pa., two views 204 VIII-F 50-ton basic furnace at Sharon, Pa., two views 204 VIII-G 50-ton Wellman furnace at Ensley, Ala 209 VIII-H Method of charging a tilting furnace 211 VIII-I Wellman charging machine, two views 212 VIII-K Valves used at Steelton, two views 214-5 VIII-L Forter valve 216 IX-A Water seal producer, two views 237 IX-B Frazer Talbot producer 238 IX-C Semet Solvay coke oven, two views 260 IX-D Otto Hoffman coke oven 261 XV-A Variations in the critical points in different steels 395 XV-B Micro-photographs Nos. 1 to 9 397 XV-C Micro-photographs Nos. 10 to 18 398 XV-D Micro-photographs Nos. 19 to 24 399 XV-E Micro-photographs Nos. 25 to 30 400 XV-F Micro-photographs Nos. 31 to 36 401 XV-G Micro-photographs Nos. 37 to 45 402 XV-H Graphical representation of the phase doctrine 419 XVI-A Curves showing elongation with varying length 437 XVI-B Expansion of curves in Fig. XVI-A 439 XVI-C Curves showing law of elongation of eye-bars 445 ixvii XXV111 INDEX TO FIGURES. NO. PAGE. XVII-A Curves showing relation of the chemical composition of acid open-hearth steel to the ultimate strength, as shown in Table XVII-N 491 XVII-B Curves showing relation of the chemical composition of basic open-hearth steel to the ultimate strength as shown in Table XVII-N 492 XVII-C Curves showing relation between the composition of acid open-hearth steel and its ultimate strength as shown in Table XVII-U 508 XVII-D Curves showing the relation between the composition of basic open-hearth steel and its ultimate strength as shown in Table XVII-X 521 XVIII-A Eight-inch test specimen 552 XVIII-B Two-inch test specimen 570 XXII-A Map of United States, eastern half . 637 XXII- A Map of United States, western half 63& XXII-B Pennsylvania, West Virginia, Ohio, etc., eastern half . . . 641 XXII-B Pennsylvania, West Virginia, Ohio, etc., western half . . . 642 XXII-C Map of lake region 653 XXII-D Mesabi, Vermilion and Gogebic ranges 654 XXII-E Marquette and Menominee ranges 655 XXII-F Map of Allegheny County, Pa 656 XXII-G Bessemer plant at Edgar Thomson . 664 XXII-H Bessemer plant at South Chicago 666 XXII-I Rail mill at South Chicago . . . . 667 XXII-J Birmingham ore deposit ....... 670 XXII-K Bessemer plant at Steelton . . 680 XXII-L Open-hearth plant at Steelton 681 XXII-M Rail mill at Sparrow's Point 685 XXIII-A Map of Great Britain 69S XXIII-B Coal fields of Great Britain 694 XXIII-C Durham coal field 701 XXIII-D Cleveland ore deposit 702 XXIII-E Rolling mill of Northeastern Steel Company 708 XXIII-F Works at Cardiff 715 XXIV-A Map of Germany 728 XXIV-B Minette District 732 XXIV-C Rombach Steel Works 73& XXV-A Map of France . . . . 763 XXV-B Coal and ore fields of France 765 XXVI-A Map of Russia . .,..,. . . >. 775 XXVII-A Map of Austria 78ft INDEX TO FIGURES. XXIX NO. PAGE XXVTII- A Map of Belgium 799 XXIX-A Map of Sweden 804 XXIX-B Swedish blast furnace 806 XXX-A Map of Spain 813 XXXIII-A Production of coal in the leading nations 828 XXXIII-B Production of ore in the leading nations 829 XXXIII-C Production of pig-iron in the leading nations 830 XXXIII-D Production of steel in the leading nations 831 INDEX TO ABBKEYIATIONS. A. I. M. J7. American Institute of Mining Engineers. Am. Soc. Civil Eng. American Society of Civil Engineers. A. S. Mech. Eng. or Am. Soc. Mech. Eng. American Society of Mechanical Engineers. Journal Frank. Inst. Journal of the Franklin Institute. Journal I. and S. /., or I. and S. I. Journal. Journal of the Iron and Steel Institute of Great Britain. Proc. Inst. Civil Eng. Proceedings of the Institute of Civil Engineers (England). Proc. English Inst. Mech. Eng. Proceedings of the English Institute of Mechanical Engineers. Trans. A. I. M. E. Transactions of the American Institute of Mining Engineers. Trans. A. S. Mech. Eng., or Trans. Am. Soc. Mech. Eng. Transactions of the American Society of Mechanical Engi* neers. Trans. Am. Soc. Civil Eng. Transactions of the American So* ciety of Civil Engineers. C by comb. carbon as determined gravimetrically. C by color. carbon as determined by the color method. Graph. graphite. Tr. trace. Und. or undet. undetermined. PART I. INTRODUCTION. The Main Principles of Iron Metallurgy. INTRODUCTION. THE MAKING OF PIG-IRON. The process of making steel begins by making pig-iron from iron ore. This iron ore is natural iron rust. It is a combination of iron and oxygen, and if we take away the oxygen the iron is left alone. Charcoal or coke or carbon in any form will rob iron ore of its oxygen, and it will do this at a very moderate tempera- ture, the action taking place if the ore and coke are mixed and heated red hot. But it is necessary to do more than this. The iron must be melted and the earthy parts of the ore and coke must be separated from the iron. The operation is conducted in a fur- nace about one hundred feet high, filled with a mixture of coke, iron ore and limestone, and superheated air is. blown in at the bot- tom. A portion of the coke is burned by the oxygen of the air and serves to maintain the furnace at a high temperature, while another portion is employed in robbing the iron ore of its oxygen. The air that is blown into the furnace is first heated to a dull red heat by passing it through "stoves." These stoves are in turn heated by burning in them the gases escaping from the top of the furnace. In ancient days these gases were allowed to escape freely, but now the tops are closed tight and all the gas is taken down to the level of the ground, part being used under boilers to generate steam to run the blowing engines, and part in the stoves to preheat the blast. As the air is red hot when it enters the tuyeres, and as it imme- diately meets glowing coke which has been heated by its downward passage through the furnace, it follows that a very high tempera- ture must be caused at this point. This region, therefore, imme- diately about the tuyeres is called the "zone of fusion." It is here that the real melting occurs, but a great deal of the work is done higher up in the furnace, for the gases from this hot zone of fusion ascend through the overlying 70 or 80 feet of stock and heat it to a high temperature, and under these conditions there is a reaction 3 4 - INTRODUCTION. between the carbon of the gas and the iron ore, whereby the oxygen of the ore unites with the carbon and leaves the iron in the finely divided metallic state known as "spongy iron." The reaction is not complete and a great deal of ore reaches the zone of fusion in a nearly raw state, but in this zone the extremely high temperature quickly completes all reactions ; the raw ore is rapidly reduced, the earthy impurities unite with the limestone and are fused into slag, while the metallic iron melts and is collected in the hearth below the tuyeres. The metal so produced is not pure iron, for while it is in contact with white-hot coke in the furnace, it absorbs a certain amount of carbon. This amount is quite constant, and it is safe to assume that any piece of ordinary pig-iron, no matter what its appearance may be, contains from 3.5 to 4.0 per cent, of carbon. Some of this carbon is chemically combined with the iron, and some is held in suspension as graphite. If a large proportion is combined, the fracture of the iron looks white and the metal is hard and brittle. If a large proportion is in the free state, the fracture will be gray or black, with loose scales of graphite, and the iron is soft and tough. Very slow cooling tends to put the carbon into the con- dition of graphite, while sudden chilling from the liquid state tends to keep it in combination and give a hard and white iron. The iron also contains silicon, which is absorbed in the furnace from the ash of the coke. Sometimes this silicon will amount to only one-half of 1 per cent, and sometimes it will be 3 per cent. Usually there will be from 1 to 2 per cent. A certain small proportion of sulphur will also be present. It is not wanted at all, but there is seldom less than two-hundredths of one per cent., while there may be one-quarter of one per cent., and even more. When there is over one-tenth of one per cent, the iron is apt to be hard and brittle and to have a close and white fracture. In such iron, the silicon is usually low and this contributes to the closeness of the grain. The percentages of silicon and sulphur that are present in the iron depend in great measure upon the conditions in the blast fur- nace, and hence may be controlled by the furnaceman. But there is one element which is universally present, over which he has no control. This element is phosphorus. Whatever quantity is pres- ent in the ore and fuel will be found in the pig-iron, so that the only way to get an iron low in phosphorus is to get ore and coke INTRODUCTION. 5 which contain only a small percentage. In irons used for making steel by the usual Bessemer process, the iron is not allowed to con- tain over one-tenth of one per cent, of phosphorus. For basic steel and for foundry work no fixed limit can be given. Where great toughness is required in iron castings it is well to use what is called "Bessemer pig-iron/ 5 by which term is meant an iron containing not over one-tenth of one per cent, of phos- phorus. Such an iron costs very little more than ordinary foundry grades. In other cases a high percentage is desired to confer great fluidity, and irons carrying 3 per cent, of phosphorus are in demand, a certain proportion of such metal being used in making intricate castings where the metal must accurately fill every corner of the mold. Pure iron itself is very difficult to melt; it is soft, tough and malleable both hot and cold, but the elements above described, preeminently the presence of nearly 4 per cent, of carbon, change its character completely in the following ways : (1) It is more fusible. (t) It is brittle. (3) It cannot be forged either hot or cold. Thus we have what the general public calls cast-iron. In the trade, however, this term is applied to it only after it has been melted again and cast into some finished form. The product of the blast-furnace is always spoken of as pig-iron. It is the founda- tion stone of all the iron industry; it is one of the great staples in the commerce of the world. The foundryman makes from it his kettles and stoves; the puddler refines it and supplies the village blacksmith with bars for chains and horseshoes; the steel maker transmutes it into watch-springs and cannon. THE MAKING OF WKOUGHT-IROK When the Bessemer process of steel making was invented it was confidently predicted that it sounded the death-knell of the puddling furnace, but although there have been several announce- ments of the funeral, the great event has never actually occurred. There seem to be a few places where wrought-iron- is needed, and there are many more places where the blacksmith and the machinist find steel unsatisfactory, because they do not know anything about the metal and refuse to learn, usually stating that they have been "working long enough to know." 6 INTRODUCTION. Wrought-iron is made by melting pig-iron in contact with iron ore and burning out the silicon, carbon and phosphorus,, leaving metallic iron. This iron is not in a melted state when finished, for the temperature of the furnace is not sufficiently high to keep it fluid after the carbon has burned. It is in a pasty condition and is mixed with slag and when taken out of the furnace is a honey- comb of iron, with each cell full of melted lava, and this honey- comb is squeezed and rolled until most of the slag is worked out and the iron framework is welded together into a compact mass. The bars are rough and full of flaws and are regarded as an intermedi- ate product. This "muck bar" is then cut up and "piled" and heated to a welding heat and rolled again, and this time the bar is clean and becomes the "merchant iron" of commerce. The previous description refers to the use of pig-iron only, but in many works this practice is modified by using scrap of various kinds, especially steel turnings from machine shops. Oftentimes almost the entire charge is made of cast-iron borings and steel turnings, although a certain amount of larger steel scrap is gener- ally used to make the ball hold together. In making the pile for the second rolling a certain proportion of soft steel scrap is often used, as this welds up with the rest, so as to be practically the same, and this increases the tensile strength of the bar. The main principles of the process, however, remain the same in all its forms. A DEFINITION OF STEEL. In the olden time all kinds of steel, whether made in the crucible, in the cementation chamber, or in the puddle furnace, contained carbon enough to make them suitable for cutting tools when hard- ened in water, and the steels that were made in the Bessemer con- verter during the early days of its history were all more or less hard, much of it being used for tools ; consequently the metal made in the converter was rightly called Bessemer steel. As time went on and the cost of the operation was reduced below that of making wrought-iron, a great deal of very soft metal was made in the converter and in the open-hearth furnace. This new metal did not fill the old definition of steel, but it was impossible to draw any line between the steel used for rails and that used for forgings, and it was impossible to draw a line between the metal used for forgings and that used for boiler plate, and as it was impossible to do this, practical men in America and England did INTRODUCTION. 7 not try to do it, but called everything that was made in the Bessemer converter, or in the open-hearth furnace, or in the crucible, by the name "steel/* A few scientific committees tried to make new names, but their labors came to naught in England and America. In Germany the committees had their way for many years, and the soft metals of the converter and the open-hearth were called ingot-iron. This term still survives in metallurgical literature, but in the German works where the metal is made, it is called steel, and the plant itself is called a stahl werke (steel works), so that we have the peculiar anomaly of a steel works making what is called steel by the work- men, while the official reports declare that it makes no steel at all. It seems inevitable that Germany must soon give up this outgrown system. The current usage in our country and in England in regard to wrought-iron and steel may be summarized in the following defini- tions : (1) By the term wrought-iron is meant the product of the puddling furnace or the sinking fire. (2) By the term steel is meant the product of the cementation process, or the malleable compounds of iron made in the crucible, the converter or the open-hearth furnace. THE MAKING OF CEUCIBLE STEEL. Most of the hard steel in the market to-day is made in the open- hearth furnace. Enormous quantities are used for car springs and agricultural machinery, and both the acid and basic furnaces fur- nish a share. There are some purposes, however, which call for a steel entirely free from the minute imperfections often present in open-hearth metal. Such is the case in watch-springs, needles and razors; and it is found that the old crucible process gives in the long run the most satisfactory metal for such work. This process consists in putting into a crucible a proper mixture of scrap, pig-iron, or charcoal and heating it until everything is thoroughly melted, the crucible being kept tightly closed to prevent the admittance of air. This process is a century old, but bids fair to round out another with little change. THE ACID BESSEMEE PROCESS. The Bessemer process consists in blowing cold air through liquid 8 INTRODUCTION. pig-iron. Sometimes the pig-iron is brought directly from the blast-furnace while fluid, and sometimes it is remelted in cupolas. In the early plants in England and America the lining of the vessel which held the iron was of ordinary silicious rock and clay, and this is still the universal practice in America. In other countries it has been necessary to develop a modification of the process, the linings being made of basic material, whereby the chemistry of the opera- tion is greatly changed. The growth of the basic Bessemer practice made it necessary to have a distinguishing name for the old way, and it is therefore called the acid process, the word being used in a chemical sense rather difficult to explain to any one not versed in chemistry. In the acid process, the air passing through the iron burns the silicon and carbon, while the heat caused by their combustion fur- nishes sufficient heat to not only sustain the bath in a liquid state, but to increase its temperature, and to oftentimes necessitate the addition of scrap or steam as a cooling agent. This increase in temperature is due principally to the silicon, which is of great calorific power, while the burning of the carbon gives barely sufficient heat for the bath to hold its own. It is necessary, therefore, that the iron contain sufficient silicon to raise the temperature to the point where steel will remain perfectly fluid. In the old days when operations in a steel works were slow and converters were allowed to cool off between charges, it was neces- sary for the pig-iron to have about 2 per cent, of silicon to get sufficient heat, but with the rapid methods of to-day, it is found that 1 per cent, is enough. When the silicon and carbon are all burned, a certain amount of manganese is added in order that the steel shall be tough while hot, and be able to stand the distortions it is subjected to in the rolling mills. If soft steel is wanted, this manganese is obtained by using a rich alloy called ferromanganese, containing 80 per cent, of man- ganese, while if rail steel is being made, the usual method is to make a liquid addition of spiegel iron a pig-iron containing about 12 per cent, of manganese. For every ten tons of steel about one ton of this spiegel will be added, and this at the same time gives enough manganese to make it roll well, and enough carbon to confer the necessary hardness. When the rich alloy is used to make soft steel, as before explained, INTRODUCTION. 9> tne amount added is very small and the carbon thus carried into the bath is trifling. The resulting steel is poured into a ladle, and the slag, being very light, floats on the top. The steel is then tapped from the bottom,, the separation of metal and slag being perfect. Minute cavities of slag are often found in steel, but these come from internal chemical reactions, or sometimes from dirt in the mold. They do not arise from mixture of the metal and slag when poured in the way that is, almost universally used in Bessemer and open-hearth works. In this acid process there can be no removal of phosphorus or sul- phur, and as no steel is allowed to contain over one-tenth of one per cent, of either, it is plain that the pig-iron must not contain more than this allowable amount. It has been shown, in the discussion of the manufacture of pig-iron, that the phosphorus in the ore will appear in the metal. Consequently if the ores of any district con- tain more than one-twentieth of one per cent, of phosphorus, which will give one-tenth of one per cent, in the iron, that district cannot possibly use the acid Bessemer process. If they do contain as little as this, then this process is the cheapest method of making steel that has ever been discovered or probably ever will be. THE BASIC BESSEMER PROCESS. The basic Bessemer process is similar to the acid Bessemer, both being founded upon the general truth that if cold air be blown through pig-iron, the combustion of the impurities in the iron will furnish sufficient heat. to keep the metal in a fluid state. In the acid process it has been shown that only two elements are thus burned, viz., silicon and carbon, and that the silicon supplies most of the heat. In the basic process the lining is made of basic material, usually of hard burned dolomite, which is a limestone containing from 30 to 40 per cent, of magnesia. When the linings are basic, it is a bad thing to have much silicon in the iron, because when silicon is oxidized it forms silica (Si0 2 ), and this attacks the lime lining. The percentage of silicon is therefore kept as low as possible, and this makes it necessary that some other source of heat be provided. This is the more necessary because more heat is needed in the basic process than in the acid, on account of the lime which is added in the converter and which must be melted during the operation. The element used to. take the place of silicon and supply heat is 10 INTRODUCTION. phosphorus. In the acid process phosphorus is not eliminated at all, but when the linings are basic it is possible to add lime and make a basic slag in which phosphorus can exist as phosphate of lime or phosphate of iron. In the acid process it is not feasible to add lime, because the lining of the converter would be eaten away and the slag could not remain basic enough to hold the phosphorus. As already stated, the basic Bessemer process requires more heat than the acid process, because considerable time must be added to give a basic slag, and because the lining of the vessel is eaten away much faster. It has also been explained that silicon is not allowed in the iron to any extent, because the more silicon there is present, the more lime must be added to counteract it. Inasmuch as silicon is the principal source of heat in the acid process, and as still more heat is required in the basic converter where silicon is not allowed, it is evident that phosphorus, which replaces silicon as a heat producing agent, must be present in con- siderable quantity. In most basic Bessemer works the iron con- tains about 2 per cent, of this element. If it falls below 2 per cent, the heat produced is not sufficient to give the proper tempera- ture to the fluid metal at the end of the blow. With very fast work and a short time between charges this percentage could doubtless be reduced considerably. Thus it happens that the Bessemer process is applicable to only two kinds of ores : (1) Those containing only a trace of phosphorus, giving an iron suitable for the acid process. (2) Those containing a high percentage giving an iron contain- ing 2 per cent, of phosphorus, suitable for the basic process. There are many deposits of ore in different parts of the world which are intermediate between these classes, and which give a pig- iron ranging from one-tenth of one per cent, up to one and one- half per cent. These irons are not suitable for either form of the Bessemer process, although it often happens that an iron which contains too little phosphorus for the basic vessel can be used in admixture with an iron that contains a surplus. When this is impracticable, such irons can be used for steel only in the basic open-hearth furnace. When the air is blown through the melted iron in a basic con- verter the silicon is first oxidized, and the carbon next. Thus far the operation is the same in both the acid and the basic vessel. INTRODUCTION. 11 At that point the acid process ceases, but in the basic process the blast of air is continued and the phosphorus is oxidized and passes into the slag. The slag therefore contains a considerable per- centage of phosphorus and this makes it valuable as a fertilizer. The demand for it is unlimited and the revenue derived from it is a very important matter to all plants using this process. The cost of labor, however, and the greater waste and diminished output of a basic Bessemer render this process out of the question except where suitable pig-iron can be had at a much lower price than iron fit for the acid process. In the United States this condition does not exist and there is no plant in operation in this country. The final operation of adding spiegel iron or ferromanganese is conducted in practically the same way in the basic Bessemer vessel, as has already been described in the account of the acid process. THE OPEN-HEARTH FURNACE. An open-hearth furnace really means a furnace having a hearth exposed to the flame, so that any piece of steel or other material placed upon the hearth is exposed openly to the action of the burning gases. The term has been narrowed by custom to denote such a furnace where steel is melted. A furnace for this purpose must be regenerative in order to get the requisite intense tempera- ture. Regenerative furnaces are also used very generally for heat- ing steel in rolling mills, but they are not called open-hearth fur- naces except when the steel is actually melted. By a regenerative furnace is meant one in which the heat carried away in the stack gases is used to warm the air and gas before they enter the furnace. Strictly speaking, a furnace would be regen- erative if air pipes were put into the stack and the air blast were passed through these pipes. But by custom the term means only a furnace which is heated by gas, and where both gas and air are heated before they enter the furnace by being passed through chambers filled with bricks loosely laid, these bricks having pre- viously been heated by the waste gases. By having two sets of chambers, one set can be used to absorb the heat in the waste pro- ducts and the other set to warm the incoming gases. By proper systems of reversing valves these two sets of chambers can be used alternately for uicii purpose, and in this way the gas and air are heated to a yellow heat before they unite, and it is quite evident that yellow-hot air and yellow-hot gas will give a very intense heat. 12 INTRODUCTION. The problem in an open-hearth melting furnace is not to reach the- desired temperature, but to control the temperature and preyent the roof and walls from melting down. THE ACID OPEN-HEARTH PROCESS. The term acid open-hearth furnace means a regenerative gas furnace used for melting steel, and lined with silicious material (sand). It has been shown that the Bessemer process can be con- ducted in a vessel lined with silicious material, or in a vessel lined with basic, material, and it has been shown that this difference in lining makes a radical difference in the process. In the same way the manner in which a steel melting furnace is lined profoundly influences the subsequent operations. Contrary to popular belief, the bottom in itself plays very little part and has very little influ- ence, but the character of the bottom determines the character of the elag that can be carried, and the character of the slag deter- mines the chemistry of the process. In the acid open-hearth process a mixture of pig-iron and scrap is charged into the furnace and melted. Nothing is added to form a slag, as the combustion of the silicon and manganese, together with some iron that is oxidized, and some sand from the bottom, affords a sufficient supply. The slag is about half silica (Si0 2 ), while the other half is composed of oxides of iron and manganese. When the mass is melted it is fed with iron ore, and the oxygen in the ore oxidizes the excess of carbon until the required com- position is attained, whereupon the steel is tapped, the proper addi- tions of manganese being made at the time of tapping. Melted spiegel iron, so generally used in Bessemer practice, is not used in open-hearth work, but the manganese is added in the form of a rich ferromanganese, which is generally thrown into the ladle as the heat is tapped. Sometimes a spiegel iron is used, but this is put into the furnace a little while before tapping and allowed to melt. It is necessary for the highest success of the operation that the slag should be kept within certain limits in regard to its chemical composition, for if it contains too much silica it is thick and gummy, and the operation will be much retarded, while if it con- tains too much oxide of iron it will be sloppy and the metal will be frothy and over-oxidized. It would seem at first sight that there would be considerable difficulty in regulating the composition of a slag that is constantly receiving iron ore and constantly absorbing- INTRODUCTION. 13 silica from the bottom. Moreover, the amount of ore is not con- .stant nor the rate at which it is added, for on some heats scarcely any ore is thrown in, on others there may be 500 pounds added in three or four hours, and on others there may be 3,000 pounds used in the same period of time. As a matter of fact, there is very little difficulty in maintaining a very regular chemical composition if moderate judgment be exer- cised and the additions of ore are regulated by the temperature of the furnace and the condition of the metal. Many an open- hearth melter has never heard of silica, and yet can keep a constant percentage of it in his slag. This is due to the fact that the slag regulates itself to a great extent. The pig-iron used in the charge always contains silicon and this furnishes silica. If the amount is not sufficient, there will be a cutting away of the sand bottom to supply more. We thus have by the wearing of the bottom an inexhaustible source of supply of silica. In the same way we have a similar supply of iron oxide by the oxidation of the iron of the bath. If iron ore is added, this is the easiest way for the slag to get the oxide, since it simply appropriates it to its own use. Iron ore is a compound of two atoms of iron with three atoms of oxygen, expressed in chemistry thus Fe 2 :i , wherein Fe is iron and is oxygen, and the figures represent the proportions. . If the slag contains too high a percentage of silica, and needs more iron oxide, and if under these conditions iron ore is added, then only one of these atoms of oxygen goes toward oxidizing the silicon and carbon of the bath. This leaves two atoms of iron and two atoms of oxygen, and these unite together to form two parts of a different oxide, FeO, or since there are two atoms of each, thus 2FeO. The extra atom of oxygen has united with carbon and formed a -gas in which one atom of carbon unites with one atom of oxygen. In chemistry this action is expressed thus: C+0=CO. The symbol C stands for carbon, and for oxygen, and when united in equal proportions, they form CO, which is the chemical symbol for carbonic oxide. The whole operation of adding iron ore to an open-hearth bath, when only the extra atom of oxygen is given to the carbon, and the rest of the oxide stays with the slag, may be expressed by the fol- lowing simple chemical formula: F e 2 3 +C=2FeO+CO. 14 INTRODUCTION. This concentrates in one line all the explanation we have just gone through. Sometimes the slag has a sufficient supply of oxide of iron and needs no more. In this case, when ore is added, all the oxygen goes to the carbon of the bath so that there are three atoms of oxygen calling for three atoms of carbon. This leaves the iron, alone in its metallic state and it is instantly dissolved in the bath,, and the weight of the charge is increased by just so much. The chemical symbol expressing this is as follows: Fe 2 3 -f-3C=2Fe+3CO. Generally it will happen that the truth lies between these two con- ditions; that the slag keeps part of the -oxide and the rest is re- duced, part of the oxygen uniting with carbon and part of the iron being dissolved in the bath, the remainder of the oxide of iron entering the slag. Still another condition exists whenever iron ore is not added to the bath. Under this state of affairs, it may be necessary for the slag to have more oxide of iron, and there is no place for this to come from except the bath. Therefore, when there is need of oxide of iron, the iron of the bath unites with the oxygen of the flame and goes into the slag. Thus it is clear that if no iron ore is used, a certain equivalent amount of good stock must be oxidized, and that if iron ore is used the weight of metal tapped will be greater than if it had not been added. The amount of carbon in the steel, and therefore the tensile strength, depends entirely on the conduct of the operation, but the amounts of phosphorus and sulphur depend upon the kind of stock which is put into the furnace. If a superior quality of steel is required the original stock should contain only small percentages of these elements. Such stock, however, costs more money than common scrap. If an ordinary quality is required then ordinary pig-iron and scrap are used. It is a common belief that it is an easy thing to distinguish between open-hearth steel and Bessemer steel. It is usually very easy to tell basic open-hearth steel from acid Bessemer, or acid open-hearth from basic Bessemer, but it is impossible by any ordi- nary means to tell acid Bessemer from acid open-hearth or basic Bessemer from basic open-hearth. Most American metallurgists INTRODUCTION. 15 and engineers, however, agree that open-hearth steel of a given composition is more reliable, more uniform, and less liable to break in service than Bessemer steel of the same composition. And there are many metallurgists and engineers both in this country and abroad who believe that acid open-hearth steel is more reliable than basic open-hearth steel of similar composition. In Chapter XVII it will be shown that there is mathematical evidence to support this opinion. There are many who disagree with this proposition, but almost every American who disputes it will confidently assert that open- hearth steel is superior to Bessemer steel, and he will just as un- qualifiedly put basic Bessemer steel in a lower place, yet his opinion on these two steels is no more capable of complete logical demon- stration than my opinion in favor of acid steel. The reasons for this opinion, founded on an experience extending over a score of years, may not be written in the compass of this chapter or this book. THE BASIC OPEN-HEARTH PROCESS. The term basic open-hearth furnace means a regenerative gas furnace, used for melting steel and lined with basic material, usu- ally either magnesite or burned dolomite. It has been stated in discussing the acid open-hearth that the bottom itself takes very little part in the operation, but that it determines the character of the slag that can be carried. When the bottom of the furnace is made of silica (sand) the slag must be silicious ; but when the bottom is basic the slag must be basic. Con- sequently in the basic open-hearth furnace the charge is composed of pig-iron and scrap, just as in the acid furnace, but, in addition to this, a certain amount of lime or limestone is added. The whole mass of iron, scrap and lime is melted down by the action of the flame. The silicon and carbon of the pig-iron are oxidized, just as in the acid process; the manganese of the scrap and some of the iron are both oxidized just as on the sand bottom; but the silica and the oxides of iron and manganese do not make a slag by them- selves, for they unite with the lime that has been added. This gives a basic slag and when the slag is basic the phosphorus in the pig-iron and scrap will be oxidized and enter the slag as phosphate of lime or iron, just as it does in the basic Bessemer vessel. Thus the basic open-hearth furnace will allow the purification of iron con- 16 INTRODUCTION. taining phosphorus, and for the same reason, but in very much less measure, sulphur can be eliminated. After the charge of pig-iron and scrap is melted, iron ore is added as fast as necessary to oxidize the excess of carbon, and when the metal has reached the desired composition it is tapped into the ladle, the additions of manganese being made in the same manner as in the acid furnace. The principles underlying the reactions in a basic furnace may briefly and incompletely be stated as follows: (1) Silicon oxidizes readily at a high heat under almost all conditions. Its oxide is sand (Si0 2 ), which acts as an acid, by which is meant that it will combine if it has a chance with one of the bases or earths, like lime, iron or manganese. (2) Phosphorus oxidizes readily, but it will not stay in the form f oxide unless the conditions are favorable. Its oxide is phos- phoric anhydride (P 2 5 ), which acts as an acid like silica; but silica when formed is stable and will stay where it is put, but the oxide of phosphorus must have something to unite with, and this something must be one of the bases or earths like lime, iron or manganese. If oxide of phosphorus is formed and there is no base for it to unite with, the metallic iron robs it of its oxygen, and then we have oxide of iron, while the phosphorus is left alone, dissolved in the bath. (3) The oxide of phosphorus requires a considerable quantity of bases to unite with. If the quantity is limited, the phosphorus may stay for a time, but will then leave. If a slag contains all the phosphorus it can hold at a certain temperature and the furnace gets hotter, some of the phosphorus will go back into the metal. If, with the same slag the carbon begins to burn faster from any oa B d 8 +CO=2Fe s O 4 +C0 2 (complete) Fe+OO a =FeO+CO 400C Fe,0 3 +3C=2Fe-h3CO (begin) 3Fe a O 8 -FCO=2Fe,O 4 +OO t (rapid) 350 300C Fe+OO a =FeO+CO (begin) 250 2Fe 2 O 3 +80O=7C0 8 -f 4Fe+0 (begin carbon deposition) 3Fe 2 O 8 +CO=2Fe,O 4 -fCO 2 (begin) 2000 Carbon begins to reduce Fe 2 3 at about 400 C. (750 F.). The reactions between carbon and the usual oxides are as follows: 66 METALLURGY OF IRON AND STEEL. (4) Fe 2 3 -f-3 C 2 Fe+3 CO. (5) Fe 3 4 +4 C=3 Fe+4 CO. (6) FeO+C Fe+CO. Each of these reactions is endothermie i.e., it absorbs heat. The carbonic acid (C0 2 ) formed by the reduction of iron oxide by carbonic oxide (CO), or by carbon, is an oxidizing agent, and by a change in temperature there may be a complete reversal and undoing of the reduction just performed, according to the follow- ing reactions: (7) 2 FeO+C0 2 =:Fe 2 3 +CO. (8) 2 Fe+3 C0 2 :=Fe 2 3 +3 CO. The first creating a large amount of heat and the second absorbing energy. These reactions depend upon both the temperature and the dilu- tion of the gas with carbonic oxide. At high temperatures the action is strong and considerable carbonic oxide must be present to avoid reoxidation. The main landmarks of the relations may be thus summarized: (a) Carbonic acid (C0 2 ) begins to oxidize spongy iron at 300 C. (570 F.). (b) Carbonic acid (C0 2 ) begins to unite with carbon at 550 C. (1020 F.), and the reaction is complete at 1000 C. (1830 F.). (c) The reduction of metallic iron depends upon the percentage of carbonic acid (C0 2 ) in the gases, but the critical content of C0 2 depends upon the temperature, as follows: At a white heat a gas containing C0 2 =10%, C0=90%, will not reduce metallic iron from the oxide. At a full red heat a gas containing C0 2 =32%, C0=68%, will not reduce metallic iron. At a low red heat a gas containing C0 2 =60%, C0=40%, will not reduce metallic iron. A mixture of C0 2 =50%, CO =50%, passed over spongy iron at a white heat oxidizes it to FeO, while if passed over Fe 2 3 reduces it to FeO. It is essential to remember that the reactions in the upper part of the blast furnace are not made up of simple processes of reduc- tion like reactions (1) to (6) or oxidations like (7) and (8). While THE BLAST FURNACE. 67 these actions are progressing there is a deposition of carbon accord- ing to relation (9), (9) 2 Fe 2 3 +8 CO=7C0 2 +4 Fe+C, It is stated by high authority that carbon deposition cannot take place without a contemporaneous oxidation of metallic iron by carbonic acid (C0 2 ), or by carbonic oxide according to the relation (10) or (11), (10) Fe+CO=rFeO+C, (11) 2 Fe-fC0 2 =r2FeO+C, but it is very difficult to understand how these reactions can pos- sibly take place in the upper zone of the blast furnace, since at the temperatures existing at the point under discussion the reactions (1) and (9) are the only ones possible, and it follows therefore that no metallic iron can exist except through reaction (9), which calls for carbon deposition, and this reaction produces metallic iron instead of oxidizing it. It may be perfectly true that at higher temperatures the great bulk of carbon deposit is dependent upon, or at least is associated with, an oxidation of metallic iron by carbonic acid (C0 2 ) or carbonic oxide (CO), but the testimony Indicates that the first of the carbon deposit is formed where the temperature is insufficient for the formation of metallic iron save by the simultaneous formation of impregnating carbon. More- over, if metallic iron were formed it could not be oxidized by carbonic acid (C0 2 ), since reaction (12) does not begin until a tern- (12) Fe-f C0 2 =FeO-fCO. perature of 300 C. (510 F.) is reached and does not become rapid until a still higher altitude is attained. On the other hand, it is well known that carbon deposition does not take place with rapidity until the temperature is from 400 C. to 500 C. (say 840 F.), and this would indicate that such deposi- tion might depend upon reaction (12) between metallic iron and carbonic acid (C0 2 ), but it may also depend upon the reduction of iron oxide by carbon, as shown in reactions (4), (5) and (6). These latter reactions are all endothermic i.e., they absorb heat, while the reduction of iron oxide by carbonic oxide (CO) is exothermic i.e., it creates heat. Eeaction (4) begins to take place at about 400 C. (750 F.), BO 68 METALLURGY OF IRON AND STEEL. that at this temperature a supply of metallic iron is provided, and since carbonic acid (C0 2 ) is able at this point to oxidize metallic iron according to reaction (12), it .follows that there may coexist all the factors necessary for any reactions, since by interchange there may be present Fe 2 3 , Fe 3 4 , FeO, Fe, CO and C0 2 . Two of the reactions occurring are (13) and (14), .(13) 2 FeO+C0 2 :=Fe 2 3 +CO, (14) 2 Fe+3 C0 2 =Fe 2 3 +3 CO, the first creating a large amount of heat and the second absorbing energy. Some interesting experiments on carbon deposition were carried on by Laudig.* He passed blast furnace gas over different ores, the gas containing about 7.5 per cent. C0 2 , and 29 per cent. CO, the temperature being just above the melting point of zinc. The following list shows the results obtained, the figures being the weight of carbon deposited in per cent, of the weight of ore : Min. Max. Old range soft hematites 4.48 35.13 hard hematites. 2.16 12.88 blue ores 1.56 4.72 brown ores 0.98 24.92 magnetites nil nil Mesabis 10.20 36.40 Scale and cinder 0.08 0.74 It was assumed by Laudig that the reducibility and value of an ore depended upon two conditions : (1) That it should be of such a character that carbon would be deposited throughout the mass ; (2) That it should not be too readily disintegrated or too much increased in volume by this action. Cases were cited in tests on some of the Mesabi ores where the mass increased to four or five times its volume after exposure to the gas, thus explaining the choking and scaffolding encountered when smelting these fine varieties. I believe that much remains undis- covered in this field. Thus it is a matter of record that Cuban ore * Trans. A. I. M. E. t Vol. XXVI, p. 269. THE BLAST FURNACE. 69 has been smelted at Steelton with a consumption of less than a ton of coke per ton of iron, and this was done moreover in a furnace only 65 feet high, the practice being continued for a long time. This ore is mostly magnetite, in hard lumps, containing 10 per cent, silica and from 0.25 to 0.50 sulphur, and on account of this latter impurity it was essential to maintain a good temperature, but this was done so successfully that the iron produced ran from a trace up to .04 per cent, in sulphur. This experience does not agree with the current belief that magnetites are hard to smelt, and it does not agree with the theory about the necessity of carbon deposition since Laudig states that no carbon was deposited in the magnetites, a fact which I have verified by experiments. It is also quite certain that the smelting values of the old range ores do not vary in proportion to their absorption of car- bon, and it is well to keep in mind the fact that hematite ores when charged into a blast furnace are very quickly converted into a magnetite, although it is quite possible that this conversion gives an opportunity for the permeating power of the gases which would be absent in the case of magnetites where no such reaction takes place. I have commented above on the necessity of invoking something beside the oxidizing influence of carbonic acid upon iron to explain the beginning of the carbon impregnation, but the question is so puzzling and it is so difficult to investigate that in the present state of metallurgy there seems to be about as much darkness as light sur- rounding the matter. It is certain, however, that the subject is of great importance, as it is known that carbonic oxide alone is unable to remove the last traces of oxygen from iron oxide, this office being performed by deposited carbon in the lower region of the blast furnace, and it is also known that carbon deposition ceases at about 600C and that carbonic acid (C0 2 ) then acts upon and dissolves carbon, so that in the lower and hotter portions of the furnace there is probably no carbon deposit except what is so to speak associated with the iron, waiting for a chance to unite with it as carbide. Howe* has reviewed the work of Bell and others very thoroughly in respect to carbon impregnation, and concludes thus: "The exact nature of the reactions is not known. Metals which like iron are reduced by carbonic oxide, but which unlike it are not * Metallurgy, p. 122. 70 METALLURGY OF IRON AND STEEL. oxidized by this gas or by carbonic acid, do not induce carbon deposition as far as known : this suggests that it is connected with the oxidation of iron by one or both of these gases by reactions like the following : Fe+xCO=FeO x +xC, FeO x +yCO=FeO x+y +yC, rather than to mere dissociation of carbonic oxide, thus : 2 CO=C+C0 2 which indeed may be regarded as the resultant of either of these two reactions :" FeO x +yCO=FeO x _ y +yC0 2 . FeO x +yCO=FeO x+y +yC. The chemical phenomena of a blast furnace have been repre- sented graphically by Bell and also in a book by Prof. Robt. H. Richards for the use of students in the Massachusetts Institute of Technology, but I believe that no attempt has ever been made to show them with quantitative accuracy. From what has gone before and what will appear in the rest of this chapter it may be seen that it is possible to map out the progress of the reactions, after assuming certain working conditions. This task has been performed for me by Mr. John W. Dougherty, Superintendent of The Pennsylvania Steel Company, and the results are shown in Fig. II-E. It must be understood that the curves are drawn very carefully and express quantitatively the exact relative amounts of each ele- ment or substance, as nearly as our knowledge admits, for the special conditions under consideration. The height is taken to be 90 feet, and information is given as to the temperature to be expected at different distances above the hearth, these temperatures being given in degrees Centigrade. The conditions assumed are as follows: Temperature at tuyeres 1500" C. Ore=60 per cent. Fe ; no water. Coke=87 per cent. C ; 1888 Ibs. per ton of iron. Stone=100 per cent. CaC0 3 ; 1010 Ibs. per ton of iron. Pig-iron=4 per cent. C ; 1 per cent. Si. Ratio of tunnel head gas by volume, 1 C0 2 to iy 2 CO. Temperature of tunnel head gases 260 C. Height of furnace, 90 feet. THE BLAST FURNACE. 71 72 METALLURGY OF IRON AND STEEL. It is also assumed upon the authority of Bell that the carbon needed for the carburization of the pig iron is deposited in the iron oxide, in the upper portion of the furnace, and that the amount so deposited is just sufficient for the work. In the absence of positive data an estimate is made of the amount of cyanogen present. No data are given on the diagram concerning silicon, sulphur, phos- phorus and other similar elements, as it is evident that their graphic representation when shown on so small a scale would be a straight line. In the case of alumina, the amount is considerably greater, but it has not been shown on the diagram, as it undergoes no change and affects no other constituent of the charge until it reaches the zone of fusion just above the tuyeres. It will be readily understood that the isothermal lines in a blast furnace are not horizontal, as they will vary with the irregularities in the rate of the descent of the stock in different parts of the furnace, but it seemed unnecessary to attempt to show these complications. From this diagram we may learn the following: At the tunnel head the ore (Fe 2 3 ) is plunged into an atmosphere of CO 24: per cent., C0 2 =16 per cent., N=60 per cent., and a temperature of about 260 C. (500 F.), and there is immediately a reduction of part of the ore to Fe 3 4 , this action increasing as the ore descends and reaches a higher temperature. By the time a depth of 10 feet is reached, all the Fe 2 3 has been converted into Fe 3 4 and the temperature is 450 C. (890 F.). Before this reduction is completed, and even before it is well under way, there begins the peculiar reaction of carbon deposition by which the gases react upon the ore and deposit carbon throughout the pores of the oxide, and this carbon so deposited remains asso- ciated with the iron, finally furnishing the proportion needed for its conversion into pig iron. This carbon deposition begins at a tem- perature of about 300 C. (570 F.), very soon after the first stages of reduction are under way, rapidly increases until all the Fe 2 3 is reduced to Fe 3 4 at a temperature of about 450 C. (840 F.) and then continues at a slower rate until the Fe :! 4 is all reduced to FeO at a temperature of about 600 C. (1110 F.). The mixture of carbon and metallic iron then descends until the zone of fusion is reached, when the mixture is converted into iron carbide. As above stated, the gases reduce the Fe 2 3 and at a temperature of 450 C. the iron is nearly all present as Fe 3 4 . This descends THE BLAST FURNACE. 7$ unchanged until at 13% feet it meets a temperature of 500 C. (930 F.), when it is strongly acted upon and converted into FeO, the transformation being complete when a temperature of about 580 C. (1080 F.) is reached at a depth of 19 feet. This FeO so formed, impregnated with deposited carbon, descends quite a dis- tance unchanged until a temperature of 700 C. (1290 F.) is encountered at a depth of 26 feet, when the last atom of oxygen is taken by the carbonic oxide, and spongy iron begins to form. This reaction is completed when the temperature reaches 800 C. (1470 F.) at a depth of 32 feet. The limestone comes down through the furnace until it encount- ^rs the temperature of 800 C. (1470 F.), at which the last of the FeO is reduced to spongy iron, at which place it is decom- posed and the carbonic acid is driven off to rise through the stock, while caustic lime (CaO) descends to the zone of fusion to flux the silicious ingredients of the charge. The carbonic acid (C0 2 ) so driven off from the limestone plays an important and objection- able part in its passage from its place of birth to the tunnel head. It has elsewhere been stated that at all temperatures above 550 C. (1020 F.) the following reaction occurs : C0 2 +C=2 CO, and as the limestone is not decomposed until a temperature of 800 C. is reached it follows that during the passage of this carbonic acid from the point where it is made at a depth of 32 feet until it reaches a temperature of 550 C. (1020 F.) at a depth of about 17 feet, which is to say, during the travel of the gas through a vertical distance of 15 feet, it is constantly reacting upon the coke. Experi- ments show that a quantity of carbonic acid equal to the entire amount liberated from the limestone is thus destroyed in the upper portions of the furnace, with the production of an equivalent amount of carbonic oxide (CO). The potential energy of this carbonic oxide may be subsequently utilized under boilers or in the stoves, but it is totally lost as far as the economy of the furnace itself is concerned. It is not strictly correct to say that all the carbonic acid from the stone is decomposed, for alongside of this amount so produced is a certain quantity arising from the reaction between the ferrous oxide (FeO) and the carbonic oxide (CO), and there is no warrant for supposing that a molecule of gas derived from the stone has any 74 METALLURGY OF IRON AND STEEL. history different from a molecule derived from the reduction of the ore, but it may be said for the sake of simplicity, as represent- ing quantitative values, that the reactions in the upper portion of the furnace consist of the reduction of iron oxides (Fe 2 3 , Fe 3 4 , FeO) by carbonic oxide (CO) and the simultaneous oxi- dation of coke by the carbonic acid (C0 2 ) of the limestone. With the exception of this last reaction, and thu formation of a small amount of carbon deposit, the coke charged at the top goes down through the furnace unchanged in quantity or condition until it reaches the immediate neighborhood of the tuyeres, the presence of so large a proportion of carbonic oxide rendering the oxidation of carbon out of the question. Below the place where the last of the FeO is reduced, at a tem- perature of 800 C., at which point the limestone is entirely decomposed, there are practically no reactions whatever occurring, and the whole history is one of heat absorption preparatory to the intense concentration of energy at the tuyeres. The temperature, therefore, rises steadily and regularly as the tuyeres are approached. This rise in temperature is shown upon the diagram as being per- fectly uniform throughout the entire height of the furnace, which, of course, is not strictly true, for the bosh region is cooled by water, and, being at a high temperature, the chilling effect at this point must be more rapid than will be found a little higher up, where there is little radiation and no heat absorbing reactions. There is still another zone where the limestone is decomposed, and this portion would show a considerable variation from a regular increase in temperature, while above that point considerable heat is absorbed by the union of carbonic acid from the stone with coke (C0 2 -|-C 2 CO), and a considerable amount created by the reduc- tion of the iron oxides by carbonic oxide (CO). Inasmuch as any attempt to equate these conditions would involve many assump- tions, it may be just as well to presuppose a uniform rate of pro- gression. The reactions in the immediate neighborhood of the tuyeres differ very materially from the reactions occurring higher up, on account of the facilitation of chemical action by the intense temperature. The entering blast is composed of nitrogen and oxygen; the nitro- gen passes unchanged through the zone of fusion and through the upper zones of reduction, and escapes in its original state and quantity with the tunnel head gases. A very small and uncertain THE BLAST FURNACE. 75 quantity combines with carbon to form cyanogen, which in turn combines with potassium or sodium to form cyanides, but these are constantly undergoing decomposition in their passage upward through the ore, according to the reaction : 2 KCN+3 FeO=K 2 0+2 CO+3 Fe-f 2 N. The oxygen, immediately upon entering, unites with the glowing coke to form carbonic acid (C0 2 ), but by contact with other pieces of incandescent coke this is all changed into carbonic oxide (CO), and from a distance of about four feet above the tuyeres to the point where limestone is decomposed and ferrous oxide reduced, there is no carbonic acid in the furnace, the entire gaseous atmos- phere being composed of nitrogen and carbonic oxide (CO). As before stated, the coke comes down through the furnace unchanged and unaffected in quality or quantity, save for the oxida- tion of a small amount by the carbonic acid (C0 2 ) driven off from the limestone. No other action takes place until it reaches a point about four feet above the tuyeres, when it meets the carbonic acid (C0 2 ) formed at the tuyeres, and there then occurs the reaction: C0 2 +C=2 CO. At the same time other particles of incandescent carbon, possibly only a fraction of an inch away from where the foregoing reaction is taking place, are coming in contact with molecules of free oxygen from the blast and there occurs the following reaction : C+2 0=C0 2 , the carbonic acid so formed being doomed to immediate destruction on its first meeting with the next molecule of incandescent carbon. The final result of this combustion is the formation of carbonic oxide (CO) with no admixture of carbonic acid (C0 2 ), and this carbonic oxide rises in unchanging quantity to the point where it meets unreduced ferrous oxide (FeO). Here begins the formation of carbonic acid (C0 2 ) from both the reduction of the ore and the decomposition of the limestone, and in spite of the destruction of some carbonic acid (C0 2 ) by the coke with formation of carbonic oxide (CO) the proportion of carbonic acid (C0 2 ) in the gases increases all the way to the top. 76 METALLURGY OF IRON AND STEEL. It need hardly be stated that all the figures relating to vertical' distances must be changed for every variation in the height of differ- ent furnaces, nor that the temperature of the tunnel head gases is quite different at every furnace, while the horizontal measurements on the drawing must be made to accord with the furnace practice on coke, ore, etc., but it has been deemed worth while to solve one definite problem as an example of the method which seems applicable to all similar investigations. SEC. Ilf. The Utilization and Waste of Heat. Any discussion of the distribution of heat in a blast furnace must base itself on the investigations of Sir Lowthian Bell. One of the last contributions TABLE 1KB. Comparison of Furnace Practice at Middlesborough and Pittsburg. Middles- borough. Pittsburgh.. General conditions- Height of furnace, feet 80 80 Cubic contents, feet 25600 18200 Per cent of metallic iron in ore .... 39 59 Weekly product per 1000 feet cubic content tons 21 57 12800 Temperature of blast degrees cent 704 593 250 171 Ratio of CO to CO 2 in gases 2 11 2 35 Data per ton of pig iron- Coke, pounds 2239 1882 Limestone pounds ... . ....... 1232 1011 Ore, pounds 5376 3613 Weight of blast pounds 9761 7974 Weight of tunnel head gases pounds . 13 381 11 211 Slag, pounds 3136 1200 Calories used in the furnace per ton of pig iron- 1.681,887 1.681,887 Reduction of metalloids in pig-iron . 212 039 133 655 Dissociation of CO 73 152 74 168 Fusion of pig-iron 335280 335,280 Evaporation of water in coke ... . 13970 4216 120904 118.516 Expulsion of CO a from limestone 206 756 157,175 Reduction of this CO 2 to CO 214 579 177 190 782.320 299,212 Radiation, cooling water etc 494 792 298,145 4,135,679 3,279,444 Calories in tunnel head gases per ton pig iron- Sensible heat 364 000 254 700 Potential as CO .... 3 810 000 3 137 000 Total in tunnel head gas 4 174.000 3.391 700 Summary per ton of pig iron (a) Calories used in furnace (as above).. 4 135 679 3 279 444 (b) Calories in tunnel head gases (as above) . 4 174000 3 391 700 Sum of (a) and (b) , 8309679 6 671.144 (c) Less calories from blast included in (a) 738 632 626 872 Calorific power produced per ton of iron. . . 7 671 047 6,044.272 Calorific power produced per ton of coke 7 574 400 7 1% 000 THE BLAST FURXACE. 77 made by him was a discussion of a paper by Gayley.* In his re- marks he compared the working of a typical Pittsburgh furnace with the practice in the Cleveland district in England. In Tables II-B and II-C the results are tabulated, so as to show the way the heat is utilized under two entirely different sets of conditions. In Table II-B I have calculated what I believe are the correct figures, being merely an expansion of the data given by Bell. In TABLE II-C. Distribution of Calorific Energy on the Assumption of the Same Coke for Middlesborough and Pittsburg. Table II-B shows that the English coke was 5 per cent, better than American coke. Hence with the same coke, the fuel in Pittsburg would have been only 1788 Ibs. per ton. Equivalent in Pounds of Coke. Per cent, of total Calo- rific Value English. American. English American. Constant factors- 452 90 452 90 20.2 4.0 25.2 5.0 Fusion of pig 1 iron Total 542 58 66 58 210 542 36 41 49 80 24.2 2.6 2 5 2.6 9.4 30.2 2 2 3 2 7 45 Factors beyond the control of the smelter- Reduction of the metalloids Expulsion of CO 2 , from limestone Fusion of slag . Total 382 20 5 34 134 206 20 2 33 80 17.1 0.9 0.2 1.5 6.0 11.5 1.1 0.1 1 8 4.5 Factors more or less under control- Evaporation of water in coke Decomposition of water in blast Radiation cooling water etc Total 193 99 1023 135 68 837 86 4.4 45.7 7.5 3.8 47.0 Tunnel head gases- Sensible heat Potential as CO Total 1122 905 50.1 50.8 Grand Total 2239 1788 100.0 100.0 * Trans. A. I. M. E. t Vol. XIX, p. 957. In the figures as given here some changes are made. Following the system in his previous writings, the learned investigator has used a unit of 20 kilo- grammes as being readily convertible into 20 cwt. Unfortunately, it is too easily convertible and in one case the figure given for calories produced per ton of iron is really the value per 20 kilogrammes, and a column headed pounds does not refer to pounds at all. These errors have no bearing on the funda- -mental questions, but attention is called to them to save trouble for others. 78 METALLURGY OF IRON AND STEEL. Table II-C I have departed from his line of calculation in finding the equivalent amount of coke in the American furnace. The object of the investigation is to account for the larger amount of fuel used in England, and Bell sums up every way in which the lean and silicious ores of Cleveland increase the work to be done, but although he mentions the fact that Connellsville coke contains more ash than the coke of Durham, he makes no allowance for this at all. It is quite certain that a pound of ash in the fuel will have just as much effect as a pound of similar earth in the ore, and it is. just as certain that the furnaceman cannot get calorific power out of this ash, and for this reason I believe that the calculation by Bell on the heat developed per unit of coke (p. 958 loc. cit.) is entirely misleading. The difference of 7.00 per cent, (not "71/2 per cent.") is almost entirely accounted for by the extra ash which the Ameri- can coke contains, for Durham coke is given as 5 to 7% per cent, in ash, while Connellsville will run at least 5 per cent, higher. The exact composition of the gases from the Cleveland furnace is not given, but the ratio is recorded and the weight produced per ton of iron, and from these data I have made calculations of the composition. (In the case of both the English and American fur- naces no allowance was made for an unknown quantity of steam in the escaping gases and a certain small error is caused in this way.) By thus determining all the factors, we are able to tabulate the figures in a more logical way. Bell views the gases simply as a vehicle of sensible heat, with the exception of the calorific power returned in the blast, but I believe it is more correct to calcu- late all the potential energy in the coke and find how much is accounted for, either as potential or chemical energy, or as sen- sible heat. Bell has done this in some cases in his previous writ- ings and showed that in one case 74 per cent, of the entire heating power of the fuel was employed in useful work, but this counted the energy developed in the boilers and in the hot stoves. I believe it is better to keep this separate under the name of "potential heat in gas," as the economical use of such gas is a problem entirely dis- tinct from the metallurgy of a blast furnace. It may or may not be possible to improve radically on the economy of energy in the interior of a furnace, but it is certainly possible to improve on the power plant and the oven plant in use at many places. The treatment of the energy used in heating the blast is a rather confusing problem. It cannot be neglected, as the hot blast pro- THE BLAST FURNACE. 79 duces an increase in the calories developed in the furnace; and it cannot be treated alone, as this same energy is included in the po- tential heat of the unburned tunnel head gases. This potential heat becomes kinetic when the gases are burned in the stoves and in the boilers, but it is impossible to make a full account of it and put it all into the equation of the furnace, because only a portion is used to heat the blast, the rest being burned under the boilers and dissipated in losses having no direct bearing upon the calorific history of the furnace proper. I have tried to cover the general heat equation in Table II-D, which gives on the one side the total heat developed in the furnace and on the other side the distribution of this heat. TABLE II-D. General Equation of the Blast Furnace. Middles- borough. Pittsburg. Per ton of pig iron- Calories from formation of CO 2 2 42? 000 1 Qfi'2 000 Calories from formation of CO 1 336000 1 025 000 Calories potential in gas as CO 3 810000 3 137 000 Total per ton of iron 7 573 000 6 144 000 Per ton of coke- Calories from formation of CO* . . .2428 000 2 360000 1 342000 1 220 000 Calories potential in gas as CO . 3 812 000 3 735 000 Total per ton of coke 7582000 7.315 000 Distribution by per cent, of total energy- 32.1 32.2 17 6 167 50.3 51.1 Total 100.0 100.0 The item of potential heat includes all the energy of the escaping gases, except the sensible heat. This potential heat appears later in four places: (1) Heat utilized in stoves in heating the blast. (2 ) Heat utilized in boilers in making steam. (3) Heat lost in ovens by incomplete combustion, in the stack gases, and by radiation. (4) Heat lost at boilers by incomplete combustion, in the stack gases, and by radiation. It would be possible to verify the conclusions if the exact calorific 80 METALLURGY OF IRON AND STEEL. value of the coke were known, but this is not given in either case. Bell assumes that Durham coke contains 10 per cent, of earthy and volatile materials, but some of this volatile matter is hydrogen, which appears as potential heat in the gases. It is probable that the heat value of Durham coke is about 7400 calories per kilogramme, or say 7,500,000 calories per ton. The coke of Connellsville will probably give about 7,120,000 calories per ton. The figures given in Table II-D, as found by theoretical calcula- tions, show a value for Durham coke of 7,582,000 calories, being about 1 per cent, greater than the foregoing assumption, and for Connellsville 7,315,000 calories, being about 3 per cent, more, while in Table II-B a somewhat different method gave 7,574,000 calories for Durham and 7,196,000 calories for Connellsville. This is a sufficiently close approximation, considering the inaccuracy of the data. The coke, the ore and the stone vary in composition from day to day. The moisture in coke, ore and blast will depend upon the weather ; and so, throughout the whole list, it is impossible to make more than an approximation of what we call the general practice, but it is possible, by careful investigations like those conducted by Bell on the Cleveland furnace, to find the values of each factor under an assumed or actual set of conditions, and from these re- sults may be deduced the relative importance of the factors in- volved. Even if the total calories developed vary somewhat from the heat value of the coke, the ratio of one factor to the whole is not necessarily greatly in error. We may consider that the Middlesborough and Pittsburgh fur- naces represent two extremes of good practice; one with lean ores and slow running, and the other with rich ores and fast running, and from Tables 1I-C and II-D the following conclusions may be drawn : (1) Of all the heat energy contained in the coke charged in a blast furnace, almost exactly one-half goes away in the tunnel head gases, a small part as sensible heat, but most of it as unburned CO. (2) This proportion of heat so lost is about the same whether the furnace is working on lean ores with a high consumption of fuel or on rich ores with a low fuel ratio. (3) The other half of the energy is used in reducing the iron ore, in melting the iron and slag, in losses from conduction and radia- tion, and in minor chemical reactions. THE BLAST FURNACE. 81 (4) The proportion of the total energy used for each one of these items depends upon the special conditions; as, for instance, the proportion needed for the reduction of C0 2 and the proportion needed for the melting of the slag both depend on the amount of limestone needed, and this in turn depends on the impurities in ore and fuel. In the case of the reduction of the ore and the fusion of the pig iron, both of which take a given amount of heat, the proportion which this given amount bears to the total will depend solely upon what the total is, being greater with a small fuel ratio. (5) The proportion lost in radiation and through the cooling water will decrease as the output of the furnace is increased, either by the use of rich ores or by rapid driving, or both. (6) The heat needed for the reduction of the ore calls for between 20 arid 25 per cent, of all the energy delivered to the furnace. (7) The fusion of the pig iron requires from 4 to 5 per cent. (8) The fusion of the slag requires from 4.5 to 9.4 per cent., increasing with the amount of impurities and the quantity of stone. (9) The heat lost by radiation and in cooling water varies from 4.5 to 6.0 per cent., decreasing with a larger output of pig iron. (10) The reduction of the metalloids, the expulsion of C0 2 from limestone, and the reduction of this C0 2 to CO, each require from 2 to 3 per cent. (11) The dissociation of CO, and the decomposition of water in the blast, each call for from 1 to 2 per cent., while the evapora- tion of the water in the coke takes a small fraction of 1 per cent. (12) Some factors are beyond the control of the smelter, as for instance all those depending on the limestone, this being determined by the impurities to be fluxed. In the American furnace before described the factors beyond the control of the smelter required only 206 pounds of coke, while in the English furnaces 382 pounds were needed, a difference of 176 pounds. Inasmuch as fifty per cent, of all the energy is lost in the escaping gases, it is. evident that these factors alone account for an extra 352 pounds of fuel in the English furnace. (13) The factors which are more or less under control are practically the same in both cases, giving a total of 7.5 per cent, in Pittsburgh and 8.6 per cent, in Cleveland. (14) The loss in the tunnel head gases is the only great item presenting any hope for future economies. In the Cleveland prac- tice the ratio of CO to CO, was 2.11. In Pittsburgh it was 2.35. 82 METALLURGY OF IRON AND STEEL. It has been stated by Bell that a ratio better than 2 to 1 cannot be hoped for, but instances are given elsewhere showing that much better practice is possible. SEC. Ilg. Metallurgical Conditions Affecting the Nature of the Iron. The composition of the slag and the temperature of the furnace are the two great forces at work determining the quality of the product and much remains to be learned concerning their mutual relation. A slag is necessary for two reasons: (1) To carry away the silica and earthy matters contained in the ore and fuel. (2) To carry away the sulphur contained in the ore and fuel. It must be liquid enough to be fluid 'at the temperature of the furnace and run freely from the cinder notch, and it must be viscous enough so that it does not act too readily on the linings and destroy them. In other words, acids and bases must be in such proportion that they are mutually satisfied with each other, and it is plain that this satisfaction depends in great measure upon the temperature, since a high heat renders a slag active that might otherwise be inert. It is rather difficult to determine just what constitutes an acid and a basic slag, as the function of alumina is not thoroughly under- stood. It is stated by Elbers* that "if the percentage of silica be low, it acts as an acid and hence increases the fluidity of the slag, but if high it acts as a base and lowers the fusing point." Phillips,* in discussing furnace slags, says, "for every-day practice and with slags of 33 and 36 per cent, silica the alumina is considered as silica. In calculating furnace burdens the error thus caused is comparatively slight." It is seldom that an increase in the proportion of lime in the slag gives trouble by erosion of the walls, since a hot furnace usually pro- tects itself by a deposit of hard carbon upon the inner surface of the bosh and hearth, but trouble does arise in other ways. If the slaT is too basic it will not run out, and therefore fills the hearth, while if it is too acid it will not absorb the sulphur. If the ore and fuel contain only a small amount of this impurity the slag may be able to dissolve it, even though the composition vary through very wide limits, but if sulphur be present in excess it may be necessary to keep the slag within very narrow bounds to make it capable of Berg- und Huttenmannische ZeUung, Vol. XLVIT, p. 253. *Ala. Geol. Survey, 1898, p. 45. THE BLAST FURNACE. 83 holding the sulphur in solution, and it will often happen that it will be necessary to increase the amount of slag so as thereby to have more latitude. With rare exceptions, the ores used in the large iron districts of the world contain only a small proportion of sulphur, but the coke almost always carries a very considerable quantity varying from one-quarter of 1 per cent., which is very low, to over 2 per cent., a fair average of good coke being about 1 per cent., so that, ordinarily, the question of removing sul- phur resolves itself into handling the sulphur in the fuel. It may often happen that special provision must be made to accomplish this; thus some of the Lake Superior hematites contain so little silica that they do not produce sufficient slag to carry away the sulphur from Gonnellsville coke, and it is found necessary to mix them with more silicious ores in order to produce a greater volume of cinder. Some ores contain sulphur up to 2 per cent., as, for in- stance, the Cornwall deposit in eastern Pennsylvania. Part of this can be expelled by roasting, but although the ore is rich in silica, it is found advisable at times to still further increase the volume of cinder to carry away the double burden of sulphur in ore and coke. When much sulphur is present in either coke or ore it may be removed by running the furnace very hot, thereby making pos- sible a very basic slag, but it is difficult to do this without making an iron high in silicon, and this is considered a disadvantage in America, as with rapid work in the Bessemer a content of 1 per cent, of silicon is quite sufficient. High silicon can be used, how- ever, if necessary, and if plenty of scrap be added and plenty of blast supplied, the blow is not very long. At Steelton we have many times put a mixture into the vessels containing 3 per cent, of silicon and have blown the heats in about twelve minutes, when low silicon iron would take about nine minutes, no difference being found in the life of the linings or the bottoms. The amount of silicon reduced, and hence the percentage of this element in the iron, depends on several conditions, being aided by: (1) A rise in temperature; for at high thermal altitudes the oxygen has a greater affinity for carbon than for silicon, and, there- fore, carbon can reduce silica with production of silicon. (2) A decrease in lime additions; for lime tends to hold silica in proportion to its needs, so that the higher a slag is in silica, the less firmly is any one molecule fastened in that slag. (3) An increase in the total amount of silica present; for, when 84 METALLURGY OF IRON AND STEEL. all other 'things are equal, the greater the exposure, the greater is the opportunity for its reduction, so that if one furnace working on ores low in silica makes three-quarters of a ton of slag to' every ton of iron, and another furnace working on ores high in silica makes one and one-half tons of the same composition, the tendency will be toward twice the percentage of silicon in the second iron that would be found in the first. It will be noticed that one of the conditions favorable to the production of high silicon pig iron, viz., high temperature, is also favorable to the elimination of sulphur, while another condition an acid slag is opposed to it. This complication gives rise to variations in practice whereby these factors are arrayed against each other for the attainment of certain ends. Thus it is possible to make : (1) An iron with high silicon and low sulphur, by running the furnace at a high temperature with a slag sufficiently ba,sic to hold the sulphur, but not basic enough to keep silicon from being reduced. (2) An iron with low silicon and low sulphur, by using a lower temperature with a somewhat more basic slag, or a high temperature with a much more basic slag. (3) An iron with low silicon and high sulphur, by using a low temperature with a slag not sufficiently basic. (4) An iron with high silicon and high sulphur, by using a high temperature with a slag not sufficiently basic. t Manganese is another element which is found in many ores, and which occasionally plays an important part in the operation. A content of only 1 or 2 per cent, in the ore will nearly all be carried away in an ordinarily acid slag, but if a greater quantity of lime be added, there is less demand for metallic oxides in the cinder and the manganese is reduced and alloyed with the iron. A high temperature seems to favor this reaction, but part of this effect may be due to the corresponding increased fluidity in the extra-basic slag. The specifications of high temperature and a limey slag, which favor the presence of manganese in the pig iron tend also toward the elimination of sulphur. When the slag is made more basic, as it should be in the production of spiegel, to prevent the loss of oxide of manganese in the cinder, the conditions are evidently opposed to the reduction of silicon, so that high-manganese iron generally THE BLAST FURNACE. 85 contains low silicon, and almost always low sulphur. It is possible, however, by special care, to make a silico-spiegel containing as much as 11 per cent, of silicon and 18 per cent, of manganese, this alloy being used as a recarburizer in steel making. Table II-E shows the composition of blast furnace slags as taken from various sources. SEC. Ilh. The blast, (a) The amount of air required. The usual way of measuring the amount of air that enters the furnace is to calculate the cubical displacement of the pistons in the WOW- TABLE II-E. Composition of Blast Furnace Slags. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 ?6 Slag. Iron. Remarks. SiO a Al,0 3 CaO MgO FeO s Total not in- cluding S. Si S 33 10 32.27 24 26 32.68 32.28 34. -50 84.98 34 70 33.68 29.86 28.95 30 62 32.55 30 08 31.46 36.08 37.19 36.86 32.06 33.57 35.38 36.35 33.70 35.11 35.10 35.84 14.92 14 57 11.53 13.50 9.38 7.94 12.05 11 44 11.93 12.04 12 04 10.47 11.13 11 44 11.50 12.85 12.65 10.74 11.97 10.65 11.76 10.21 12.56 14.21 14.75 14 34 40.76 41.02 40.25 43 28 46.95 46.47 41.33 41 27 45.96 45.20 49 30 49 13 47.16 46 36 44.85 41.69 35.47 42.46 42.46 44.11 38.19 40 10 38 12 22 48 27. 95 32.71 9.67 10.30 13.28 9.44 9.52 10.47 9.62 9.96 6.69 11.41 8.46 7 49 6.61 8.76 10.41 7.25 11.32 6.62 10.25 8.55 12.32 10 95 11 60 22.38 22 28 17.46 98.45 98.16 98.32 98.90 98.13 99 38 97.98 97 37 98.26 98.51 98.75 97.71 97.45 96.64 98 22 9*. 41 97.53 97.31 97.37 97.69 98.53 98.60 98.30 100.12 100.08 100.35 3.37 3.18 4.81 1.25 0.70 0.69 2.60 2.32 1.27 1.27 .57 .26 .15 .58 .20 2.15 1.92 1.50 1.59 0.94 1.18 0.66 0.50 1.37 1.85 1.60 tr. tr. .01 .06 .11 .05 .03 .02 .02 .02 tr. .02 .03 .03 .07 .020 .029 .028 .032 .017 .040 095 .101 .048 .038 .034 Cuban ore, hot furnace. " war " cool Spanish ore, hot coo Lake ore and part an- thracite coal : most- ly Connells- ville coke. 1 Lake ore and y Connells- I ville coke. m. furnace. 1 furance. Hot furnace. Fairly hot. Normal. Cool. 'Av.of 8 weeks 7 weeks " 7 weeks .... O.M' 0.90 0.63 0.63 81 0.90 0.99 0.32 Y62' 1.70 1 54 1.76 1 74 1.60 1.28 96 Averages for hot furnaces 33.21 34. 84 31.77 35.55 13.67 It. "5 11.98 12.05 40 68 41.30 45.58 40.52 11.08 9.79 9.05 1.66 98.64 97.68 98.38 97.66 3.791 tr. Cuban ore. 2.46 .025 Spanish ore. 1.27 .020 1.79 .027 Lake ore. Averages for moderate or cool furnace 33 151 10 271 45 57 9 si 98.80 10.88 .071 Cuban ore. 30 73 11.32 47 36 34.76 11.30 40.12 35 35 14 43 29 69 8.35 10.86 20 71 'T.26 "i'46 97.75 0.35 98.29 0.81 100.18 1.61 .03 Spanish ore. .063 Lake ore. ' NOTE All slags are from Steelton furnaces except Nos. 24, 25 and 26. The ore mixture was the same in all the cases where Spanish ore Avas used. 86 METALLURGY OF IRON AND STEEL. ing cylinders, but this is not accurate, as the losses from leaks and from inefficiency of inlet and exit valves cannot be measured. It may be well to calculate the theoretical amount of air indicated by the results obtained on tunnel head gases. In Section Hi will be found Table II-I, which gives the weight of nitrogen and oxygen contributed by the blast per ton of pig-iron under different condi- tions of furnace practice. Selecting practice D as representing a consumption of 1900 pounds of coke per ton of iron and a good efficiency as shown by the ratio in the tunnel head gases of 2 CO to 1C0 2 , we find by calculation that such a furnace when making iron at the rate of 300 tons per twenty-four hours will require about 19,700 cubic feet of air per minute. The correctness of this result is indicated by the figures obtained by Bell,* who calculates in an entirely different way and gives the weight of the air blast as 103.74 kg. per 20 kg. of iron, a ratio of 5.187 kg. to 1 kg. of iron=5270 kg. per 2240 pounds of iron, which, for a furnace making 300 tons in 24 hours, is at the rate of 1098 kg.=849 cubic metres=:29,983 cubic feet per minute. It is to be noted that the consumption of coke in Middlesborough was 22.32 units per 20 units of iron=2500 pounds per ton of iron, while I have assumed for American practice a consumption of 1900 pounds, and correcting for this, the figures according to Bell would indicate that 22,790 cubic feet of air was supplied per ton of iron, which is a moderately close agreement to 19,700 cubic feet, the result just obtained by entirely different methods of cal- culation, and under radically different conditions. (b) The heating of the blast. In the foregoing calculation it has been shown that in round numbers a furnace making 300 tons of pig-iron per day will re- ceive 19,700 cubic feet or 558 cubic metres of air per minute, equal to 803,500 cubic metres per 24 hours. It will produce 3551 cubic metres of tunnel head gases per ton of iron (see Sec. Hi) equal to 1,065,000 cubic metres in twenty-four hours and about one-third, or 355,000 cubic metres of this gas will be sent to the stoves. The specific heat of the air is .307 calories per cubic metre and the blast must be warmed from its natural temperature to a dull red heat, say 1300 F. or 700 C. so that the heat required for this operation will be 803,500X.307X?00=172,670,OOQ. * Manufacture of Iron and Steel, p. 204. THE BLAST FURNACE. 87 The gases from the tunnel head enter the stoves quite a little warmer than the atmospheric temperature, say about 170 C. (300 F.), and their sensible heat will be utilized in heating stoves. The specific heat of tunnel head gas is about .320 calories per cubic metre, so that the sensible heat thus carried to the stoves will be 355,OOOX.320X170=19,312,000, and the net amount which must be supplied by the combustion of the gas will be the total amount to heat the blast minus this sen- sible heat carried in by the tunnel head gases, which, is therefore, 172,670,00019,312,000153,358,000 calories. It has been assumed that one-third of the tunnel head gas is sent to the stoves, and it is shown in Table II-I that Gas D has a calorific value of 823 calories per cubic metre, after allowing for a small proportion of hydrogen. The theoretical value, therefore, of this will be 355,000X8^3=292,165,000 calories. Thus we find that the gas furnished to the stoves has a theo- retical heating value of 292,165,000 calories, while the heating of the blast calls for only 153,358,000 calories, showing an efficiency of 52 per cent. The low temperature of the gases, their varying quality and the difficulty of properly regulating the quantity of air for combustion will account for this low percentage of efficiency, while the presence of large quantities of dust in the gas render impracticable the use of small passages fof the more perfect absorp- tion of the heat. In this calculation no account has been taken of the moisture in the atmosphere, or of difference between summer and winter temperatures. This matter will be discussed later. It may be interesting to compare the results of calculations by Bell,* although conducted on entirely different lines, and by entirely different methods. He states that the heating to 500 blast for 18.83 kg. of pure carbon in coke required 11,345 calories. In the foregoing paragraph it has been found that heating the blast for 300 tons of iron to 700 C. required 153,358,000 calories, or * Iron and Steel Manufacture, p. 143. 88 METALLURGY OF IRON AND STEEL. 511,193 calories per ton. It is shown in Table II-I, Section Hi, that under the practice assumed, giving Gas D, there will be 768 kg. of carbon in the tunnel head gases per ton of pig-iron. If a rough allowance be made for the heating to 700 C., instead of 500 C., it will be found that 18.83 kg. will require: 511,193 = 8952 cals. Thus Bell gives 11,345 calories, while our figures show 8952 calories. We have not made any allowance for oxygen contained in the gases, nor for moisture, but have taken simply the quantity of air theoretically necessary to" burn the carbon to a gas containing a low ratio of CO to C0 2 . After allowing for various losses and for leaks, it is probable that this amount in practice would be increased 20 per cent, and that a furnace making 300 tons of pig iron in twenty-four hours will call for over 23,000 cubic feet of air per minute, under which assumption our figures would agree with those given by Bell. It has just been shown that when the blast is heated to 700 C. it contains over 500,000 calories per ton of iron produced, and it was shown in Section Ilf that under American practice the full value of the coke charged represented 6,000,000 calories per ton of pig iron, of which one-half is utilized in the furnace itself, the other half escaping in the gases. The heat in the blast, therefore, represents 17 per cent, of all the heat that escapes from the tunnel head, and as the amount utilized is just equal to the amount escaping, it follows that the heat in the blast represents also 17 per cent, of all the heat utilized in the furnace. If this is true when the air is at 700 C., it is possible to say that each 100 C. in the blast represents 2.4 per cent, of the fuel utilized, or if the coke consumption is 1900 pounds per ton, it represents 46 pounds of coke, so that it would seem that an increase of 100 C. (180 F.) in the temperature of the blast should save 46 pounds of coke per ton of iron. Such a conclusion, however, is not warranted by either theory or facts. It was long ago explained by Bell that the great gain in hot air is found in the first increments of heat, and that when a tem- perature of 700 C. is reached the gain by further superheating is comparatively slight. It is hardly necessary to pursue the calcula- THE BLAST FURNACE. 89 tion on theoretical lines, as many assumptions must be made, and because general experience has corroborated the foregoing statement. In calculating the amount of fuel needed for any metallurgical operation it is necessary to consider two things: (1) The amount of energy needed. (2) The intensity of heat required. A pound of coal produces a certain amount of energy and heat when burned and this amount is constant whether the coal is burned slowly or fast. It is the same whether it is burned in an open grate by natural draft or in a furnace under forced blast, but under forced draft the coal burns in a shorter time, and this means that there is a greater amount of heat produced per unit of time, and since the loss by conduction and radiation is about the same, it follows that this rapid combustion produces a higher temperature. If only low temperatures are required, as, for instance, in the evaporation of steam in boilers, the efficiency of the coal is about the same whether the fires be forced or not, but when cast iron or 'copper or other difficultly fusible substances are to be melted it is almost necessary to use a blower. Thus making the arbitrary assumption that a coke fire without blast will give a temperature of 1000 C. and that a fire with blast wiU give 1400 C., it is evident that no increase in the amount of fuel or length of time will melt a substance requiring a temperature of 1200 C. unless forced blast be used, but that with forced blast the melting can easily be accom- plished. In the same way the use of hot blast renders possible a higher temperature than with cold blast, and with this high temperature the blast furnace may readily smelt what was done with difficulty and with a great quantity of fuel when cold blast was used, but by the very same course of reasoning it will be clear that, once a sufficient temperature is attained, any increase beyond this may be of comparatively little value. It has been shown in Section He that at the moment the hot blast of air strikes the glowing coke a certain amount of carbonic acid (C0 2 ) is formed, but that this is immediately transformed into carbonic oxide (CO), so that the first reaction and the equa- tion may be written as follows : 1 kg. C+4.45 c.m. air=1.87 c.m. CO+3.25 c.m. N r=5.39 c. m. products of combustion. 90 METALLURGY OF IRON AND STEEL. The burning of 1 kg. of carbon to CO produces about 2450 calories when the carbon and the air are both cold, but the produc- tion of energy is much greater with hot carbon and hot air, just in accordance with the extra energy in these two factors, and it is possible to find the temperature that will be created under any set of conditions by dividing the total number of calories by the sensible heat of the gaseous products of combustion. This calcu- lation is not perfectly simple because the specific heat of these pro- ducts varies with every change in temperature. Table II-F gives the specific heat of the common gases at different temperatures. TABLE II-F. Specific Heat of Gases at Different Temperatures, between C and t C. Formulae j N, CO and O = . 306 + . 000027 t Formulae -j COa 0.374 + 0.00027 t Specific. Specific. Specific. Temp. Temp. Temp. N, etc. C0 2 N, etc. CO.. N, etc. CO, .306 .374 800 .328 .590 1600 .349 .806 200 v .311 .428 1000 .333 .644 1800 .355 .860 400 .317 .482 1200 .338 .698 2000 .360 .914 600 .322 .536 1400 .344 .752 2200 .365 .968 We are thus confronted with the fact that we should know the resulting temperature in order to find the specific heat, and should know the specific heat to find the temperature. This may be done quite readily by the method of successive approximations, but I am indebted to Prof. J. W. Richards for a method by which accurate results can be obtained by direct processes, and with assumptions which give rise to unimportant errors. I have adopted his method and have worked out the answer for the range of available tempera- tures. It will suffice to explain the details of one calculation, by which we find the temperature produced by the combustion of the carbon at the tuyeres of the blast furnace, with air at 700 degrees Centigrade. The specific heat of carbon above 1000 degrees C. is 0.5, but below 1000 C. it is less, so that the total heat in 1 kg of C. at t (when t is above 1000) is approximately 0.5120. Assuming that the THE BLAST FURNACE. 91 heat value of 1 kg of carbon is 2450 calories, the calculation for a temperature of 700 C. will be as follows : Heat in air 700x4.45x0.325= 1012 Heat in carbon 0.5 1 120 Heat in carbon and air 0.5 t+ 892 Heat of combustion 2450 Total heat in 5.39 c. m. of products 0.5 1+3342 Heat per c. m. . . 623.8-f 0.0928 t Therefore . t= 623 ' 8 +<>.0928 1 0.306+0.000027 t from which we have : 0.2132 t+0.000027 t*=623.8 t=2273 C. In this calculation no allowance has been made for the dissocia- tion of the water vapor in the air, but taking the amount usually present in the atmosphere, it is found that from 200 to 300 calories will be absorbed per kg. of carbon, and this will reduce the temper- ature at the point of combustion about 115 C., so that it is neces- sary, to subtract this from each result. It is not supposed that this will by any means give accurately the temperature of the zone of fusion, but it is believed that it is an approximation; and it is still further believed, what is of great importance, that the results in Table II-G are comparative and show the relative temperatures caused by changes in the temperature of the blast. TABLE II-G. 'Temperatures Produced by Burning Carbon with Air at Different Temperatures. Temp. of air. Resulting temperature. c. ( 30 F.) 1559 C. (2840 F.) 100 c. ( 210 F.) 1641 C. (2990 F.) 200 o C!. ( 390 F.) 1724 c. (3135 F.) 300 0. ( 570 F.) 1808 c. (3290 F.) 400 C. ( 750 F.) 1893 c. (3440 F.) 500 r. ( 930 F.) 1978 r. (3590 F.) 600 r. (1110 F.) , 2062 c. (3740 F.) 700 r. (1290 F.) 2146 c. (3895 F.) 800 e C 1 . (1470 F.) .... 2232 c. (4050 F.) 900 r. (1650 F.) .... 2316 c. (4200 F.) 1000 e c. (1830 F.) , 2400 c. (4350 F.) It will be found by inspection that the increase in temperature Is constant for each increment in the temperature of the blast, 92 METALLURGY OF IRON AND STEEL. which is to say that the same increase in the resulting temperature of the zone of fusion follows the heating of the blast from 600 to 1000 as from to 400 ; hut, as before pointed out, an increase in temperature of the zone of fusion has nothing whatever to do with the amount of heat produced in the furnace as a whole, and the calculation as to how much saving is effected is very complicated and admits much difference of opinion. There can be no question of how much heat is contained in a given amount of air, or in the air for a given amount of coke, but it is a question whether this should be compared with the total value of the fuel, or with the amount util- ized in the furnace proper, or with the amount developed in the neighborhood of the tuyeres. Moreover, any one of the assumptions is wrong, for it is necessary to take into account the fact that the fuel can never be reduced below a certain point on account of the necessity of having free carbonic oxide in the tunnel head gases to act upon the ore. The exact proportion of this gas necessary is much lower than formerly supposed, but there is some limit, and as this limit is approached each gain is made at a greater sacrifice. Experience has shown that there is a practical limit in heating the blast, and in practice it is usually from 1000 F. (540 C.) to 1400 F. (760 C.). In further elucidation of this point I give the following remarks of Prof. J. W. Eichards on reading the manuscript of the forego- ing discussion: NOTE BY PROF. J. W. RICHARDS. The conclusion is correct that the increase in the temperature of the zone of combustion is proportional to the increase in the temperature of the blast. I have made a formula for the temperature at the point of combustion, using the temperature of the blast as a variable, and by differentiating and taking the first differential coefficient have obtained the relative rate of in- crease of the two temperatures, from which it appears that when the tempera- ture of the blast is O C. the rate of increase in the furnace is 0.86 for 1 in the blast and at 1000 it is 0.85. Theoretically, therefore, the maximum temperature attainable increases about 85 for every 100 increase in the blast. Actually, however, the tem- perature of the whole zone of fusion depends on the ratio of burden to the coke burned, or rather to the heat available in the zone, and as the furnace Is burdened heavier when hot blast is used, the temperature of the whole zone of fusion, and of the fused materials, will be lower than theory would call for. The heat developed by combustion and absorbed mostly by the CO and N raises these gases to a certain temperature. As they ascend they cool off by transmitting their heat to the ingredients of the pig-iron and slag. The maxi- mum temperature to which the burden can be heated at the zone of fusion is the heat which the CO and N lose in ascending through the furnace, divided by the calorific capacity of the pig-iron and the slag-forming materials. Whatever be the temperature of the gases, these conditions will determine the maximum THE BLAST FURNACE. 93 temperature of the fused materials. This explains why in the use of hot blast the temperatures of the fused iron and slag are not proportional to the theo- retically calculated temperature of the gases, for, as stated above, more burden is carried with the hotter blast. (c) The Vapor in the Atmosphere. The vapor in the atmos- phere is everywhere recognized as seriously interfering with the operation of a blast furnace, but accurate information on the sub- ject is not always obtainable. The Pennsylvania Steel Works is situated only three miles from a station of the United States Weather Bureau, at Harrisburg, Pa., and I have obtained the data, from this source, of a district one hundred and fifty miles from the ocean and still farther from any great fresh water lake. The district is not mountainous, and has an annual rainfall of about 40 inches, which is about the same as most -places in the northern and eastern portion of the United States. The average humidity throughout the year, for three successive years, was 68 per cent., 75 per cent, and 76 per cent., and this percentage did not vary as much as might be supposed in differ- ent parts of the year. Selecting January, April, August and No- vember in one year as typical months, there were eight days in January and one day in November when the humidity was 100 per cent., or, in other words, when the atmosphere was saturated, while in April the highest humidity was 96 per cent, and in August 93 per cent. The minimum figures showed one day in each month as follows: January, 40 per cent.; April, 33 per cent. ; August, 54 per cent. ; November, 40 per cent. There were 17 days in January when the humidity was 80 per cent, and over, April having 6 days, August 9 days and November 11 days. There were 3 days in- January when the humidity was 60 per cent, or less, April having 13 days, August 4 days and November 9 days. Thus August has less than the average number of days of high humidity and much less than the average of low humidity, while November shows a large proportion with high humidity and a large proportion with low humidity. In other words, the humidity in August remained steadily at about the average, while in Novem- ber it varied widely, but averaged about the same as in the summer. The early spring-time showed the largest number of days with a low humidity, while January had the largest number with high humidity. These facts are recorded, as they differ quite a good deal from popular belief. 04 METALLURGY OF IRON AND STEEL. A general error arises from confounding the percentage of humidity with the amount of vapor. One cubic foot of air at 32 F. (0 C.) will hold, at 100 per cent, humidity, only .000304 pounds of water per cubic foot, while at 92 F. (33 C.), it will hold .00225 pound, or seven times as much, and it follows that a cubic foot of air at 90 F., with only 50 per cent, humidity, will carry between three and four times the vapor that will be held in saturated air of only 32 F. In the previous discussion it has been shown that a blast furnace, making 300 tons per day, will need over 20,000 cubic feet of air per minute, or about 100,000 cubic feet per ton of pig-iron. It will be shown in Table II-I, Section Hi, that a furnace producing gas (D) which has been the basis of previous calculations, requires 806 kg. of oxygen per 1900 pounds=862 kg. of coke, equivalent to 750 kg. of carbon. This proportion is somewhat different from that in the tunnel head gases as the limestone contributes carbon and oxygen, and the ore contributes oxygen, but at the base of the furnace the weight of oxygen will almost exactly equal the weight of carbon. This pre- cludes entirely the formation of any C0 2 so that the higher oxide must be formed higher up in the furnace by the action of the ore. Therefore, the heat reaction arising from the setting free of oxy- gen from the steam will consist simply of the union of 8/9=0.89' kg. of oxygen with sufficient carbon to form CO. .67 kg. C.+.89 kg. 0, which will produce 1650 calories. It would seem, therefore, thafr the true refrigerating effect of the decomposition of H 2 will be the heat absorbed in setting free 1/9 kg. of hydrogen, which will equal the heat produced by burning 1/9 kg. of hydrogen ^^f 1 - =3333 cals. minus the heat produced by the union of the oxygen with carbon=3333 1650=1683 calories. There is, however, another point to be considered. We may view the reaction not in the light of the dissociation of steam, but as the oxidation of carbon, and this carbon, had it not been burned by water, would have been burned by air, and in this case would have produced a positive gain in heat. It may be correct and it may be a fallacy to view this hoped for heat as part of the problem. If we do so view it, it would tend to counterbalance the heat pro- duced by the oxygen of the steam, but it cannot entirely counter- THE BLAST FURNACE. 95. balance it since the steam carries no nitrogen with it, while the oxygen of the air carries a heavy load of inert matter. The ques- tion is very puzzling, but the answer is of considerable importance. In Table II-H the refinements just elaborated have been omitted and the dissociation of one kilogramme of steam is considered to absorb the same amount of heat as the oxidation of the hydrogen contained therein. TABLE II-H. Vapor in the Atmosphere as Affecting the Blast Furnace. Degrees Fahr. Cubic Feet of Air Need- ed per Ton of Pig Iron. Pounds of Water in One Cubic Foot of Satur- ated Air. Pounds of Water in Air Needed per Ton of Pig Iron Calories Ab- sorbed in Disso- ciating this Steam. 1kg. =3333 cals. 1 lb.=1510 cals. Pounds of Coke Representedby this absorption. 1 kg. coke=4200 cals. 1 Ib. coke=1900 cals. 100 Per Cent. Humidity. 40 Per Cent. Humidity. 32 100,000 .000304 30.4 45,900 24 10 42 102000 .000440 44.9 67,800 36 14 52 101,000 .000627 65.2 98500 52 22 62 106,000 .000881 93.4 141,000 74 30 72 108,000 .001221 131.9 199,200 105 42 82 110,000 .001667 183.4 276,900 146 58 92 112,000 .002250 252.0 380,500 200 80 From this it will be seen that a saturated atmosphere of 92 F., which sometimes exists during the day in America, calls for an expenditure of 200 pounds more fuel per ton of iron than dry air at 32 F. It also shows that at low temperatures, it matters very little whether the air is saturated or not, as the content of vapor is so small in either case, and it shows that a saturated atmosphere of 60 F. will demand no more fuel than a dry air of 85 F., as the content of vapor is the same in either case. A summer temperature of 90 F. means that the blowing engines must run one sixth faster to give the same wind, and that the coke consumption will be from 70 to 200 pounds higher per ton of iron than on a moderately cool winter day. NOTE : On reading the manuscript of the foregoing discussion, Professor J. W. Eichards offers the following : The carbon burnt to carbonic oxide at the tuyeres produces the heat of formation of carbonic oxide, no matter where the oxygen comes from, oxygen comes in as air, the above heat is generated and is available; of the oxygen conies in as steam, the above heat is also generated but not is available, and a deduction must be made for the heat required to decompose 96 METALLURGY OF IRON AND STEEL. the steam and set the oxygen free. The chilling effect of the steam is there- fore 29,000 calories per kilo of hydrogen thus liberated. To keep the zone of fusion at the same temperature while this chilling effect is being produced, requires that more carbon be burnt there per unit of burden to be fused ; therefore the chilling effect at the tuyeres can only be counteracted by either decreasing the burden or increasing the fuel ratio. If the burden is considered constant, then more carbon must be burnt at the tuyeres, enough more to make up for the chilling effect ; and since carbon burns at the tuyeres only to carbonic oxide, the extra amount to be burnt at the tuyeres will be the chilling effect in calories divided by the heat effect (generated and introduced in hot blast) per kilo of carbon consumed at the tuyeres. Assuming that coke contains 90 per cent, of fixed carbon, of which 90 per cent, is burned at the tuyeres, and that the hot blast brings in one-half as much heat as is generated by combustion, one kilo of coke will represent (90X90X2,450) X f =2,977 calories, and the increased amount of coke required is equal to the chilling effect divided by 2,977 (using kilos and kg. calories). SEC. Hi. Tunnel head gases. The volume and the quality of the tunnel head gases are becoming more and more a matter of moment as progressive steel works managers are no longer content to merely raise sufficient steam at the furnace for the furnace itself, but are making all the steam possible and supplying power to other departments. The question also appears important in view of the development of gas engines driven by blast furnace gases. ISTeedless to say that no provision is ever made at furnaces to measure the volume of these gases. A rough calculation can be made from the amount of air blown, but this in turn is gen- erally an unknown quantity. Furnacemen habitually speak of the number of cubic feet blown, when they mean the cubical displace- ment of the air pistons, without knowing accurately the amount lost by leaks in the piston packing, at valves, at tuyeres, and at joints. With engines in fair condition and blowing against or- dinary pressures, this way of speaking does very well to compare one furnace with another, but it will hardly suffice as a basis for a determination of the gas produced at the tunnel head. ' The composition of the gas varies considerably, but usually within well denned limits. It is composed almost entirely of five sub- stances, nitrogen, hydrogen, carbonic acid, carbonic oxide and steam. In any complete investigation of the blast furnace the weight of this steam must be taken into consideration, for it carries off a considerable amount of sensible heat, and in burning the gas either in the stoves or under boilers allowance must be made for the sensible heat carried away by this steam in the products of com- bustion going to the stack, but except as a vehicle ^of sensible heat it hardly affects the work on hand. In determining the composition THE BLAST FURNACE. 97 of the gases, steam is seldom taken into account, for it condenses in the cooling tubes and therefore does not appear in the volumetric operations. Moreover, the amount of water present varies so greatly and depends so much upon accidental or temporary conditions that it is impossible to say what is a fair average. In wet weather the coke and the ores may both be saturated, while in dry weather they may both contain very little moisture, so that the quantity of water or .steam present in the gases will vary through a wrde range. When it is considered that, as above stated, the effect of this moisture is very slight, it may be well to ignore its presence altogether. Hydrogen is present in very variable quantity and the experiments of Bell shed very little light on the conditions surrounding its creation. The moisture in the blast is without doubt all disso- ciated in the zone of fusion, but most of the hydrogen caused thereby is oxidized again in the upper parts of the furnace. A certain amount of hydrogen comes from the small proportion of volatile matters in the coke. From these and possibly other causes the gas is usually found to contain anywhere from five tenths of one per cent, to three per cent, of hydrogen by volume. The weight of this hydrogen, however, is so small that it represents a very small amount of oxygen and in the following calculation no attention will be paid to the reaction by which it is produced. It will, however, be assumed that the tunnel head .gases contain five-tenths of one per cent, of free hydrogen, since the heating power of this small quan- tity is worthy to be taken into account. - The nitrogen which constitutes about sixty per cent, by volume of the total gases comes from the blast and from nowhere else. The carbon comes from the coke and from the carbonic acid in the limestone. The oxygen comes from the blast, from the ore and from the carbonic acid in the limestone. Most of these factors are known accurately and it is possible to calculate just what the volume of tunnel head gases will be when the weights of the dif- ferent materials going in at the top are known, as the weight of the carbon in the coke and the stone is known accurately and all this carbon, with the exception of what is combined in the pig iron, must be contained in the gases. Thus if we know the ratio of C0 2 to CO in these gases we may know just how much carbon exists as C0 2 , and how much as CO, and from this we may calculate the weights of these two gases and the amount of oxygen. 98 METALLURGY OF IRON AND STEEL. Taking then the total amount of oxygen thus determined and subtracting the oxygen added by the ore and the stone we have the amount of oxygen added in the blast. The amount of oxygen in the stone is easily found as it is only necessary to account for the oxygen in the carbonic acid, since the oxygen combined with cal- cium will remain in combination with the slag. The amount in the ore is also accurately known, for no matter how poor or how rich the ore may be, every ton of pig iron contains about 95 per cent, metallic iron, provided it is a low phosphorus pig iron, the remain- ing five per cent being carbon and silicon, and this 95 per cent, of metallic iron existed in the ore in the form of iron oxide, either as ferrous or ferric or magnetic oxide. In either case the amount of oxide per unit of iron can be determined. In the present case it is assumed that hematite ore is used and the iron is of course in the form of Fe 2 3 . Calculating in this way I have given in Table II-I the amount of tunnel head gases made under different methods of practice. Thus, for instance, in practice A it is assumed that 1600 pounds of coke and 600 pounds of stone are used per ton of iron and that the gases contain 1.5 per cent. CO to 1 per cent. C0 2 . In practice B, 1600 pounds of coke and 1000 pounds of stone are used with the same ratio of 1.5 and so on up to practice I which represents conditions with a very lean and very sulphurous ore requiring a hot working furnace with Jarge lime additions calling for 3000 pounds of coke and 2000 pounds of stone, this assumption not being theoretical at all, but being matched in practice. It is assumed that the ratio in this case is 2.5. Calculating these different conditions we find the volume per ton of iron with the heat value per cubic metre and by multiplying these together, we get the heat value of the gases per ton of iron. It will be seen that the heat value per cubic metre changes very little, for the percentage of CO stays reasonably constant and it is the percen- tage of C0 2 that varies, but the value of the gases varies very nearly in proportion to the amount of fuel used and consequently a furnace using a large proportion of fuel has a chance to recover some of the energy that is wasted in the large quantity of gases escaping from the tunnel head, since these gases can produce a large amount of power if properly used. It has been stated by Bell that we cannot hope^that the tunnel head gases will contain a ratio of less than 2 of CO to 1 C0 2 , but THE BLAST FURNACE. TABLE II-I. 99 Volume' and Composition of Tunnel Head Gases under Different Conditions. (Coke=87 per cent, carbon ; Limestone =97 per cent CaCO. Assumptions :1 Pig Iron=95 per cent. Fe and 3.75 per centC ' (Tunnel Head gas contains 0.5 per cent. H. Per Ton Pig Iron Lbs. Carbon Per Ton Iron. Carbon in Gases Per Ton Iron, Kg. Weight Per Ton Iron, Kg. Oxygen Per Ton Iron, Kg. 8 _o 3 & to ft? . O 3 g A M g if 1 O r^4 o> "o " O 1 8 8 i O 8 s !' a OQ 5 G i i d 3 03 8 O c I Sr A 1.5 1600 600 1464 84 1380 627 251 376 920 877 1170 415 88 667 B 1.5 16001000 1512 84 1428 649 259 390 950 910 1211 415 147 649 C 1.7 19001000 1773 84 1689 768 284 484 1041 1129 1402 415 147 840 T) 2.0 19001000 1773 84 1689 768 256 939 1195 1368 415 147 806 E F 1.7 2.0 22001000 22001000 2034 2034 84 84 1950 1950 886 886 328 295 558 593 1203 1082 1302 1384 1619 1580 415 415 147 147 1057 1018 G 2.25 25001000 2295 84 2211 1005 809 696 1133 162 1752 415 147 1190 H 2.5 25001000 2295 84 2211 1005 287 718 1052 1675 1722 415 147 1160 I 2.50 3000 2000 2850 84 2766 1258 360 898 1320 2095 2157 415 296 1446 1- Volume of Gases Per Ton Iron; Cubic Metres. Composition of Gases ; Per Cent. Volume and Heat Value Per Ton Iron. ogl~$ Heat Heat o ugH 2") Volume Value; Value 1 fifft CO 2 . CO. N. CO 2 . CO. N. Cubic Metres. Cals. Per Per Ton Iron. * Cu. M. Cals. A 2208 467 702 1752 15.99 24.03 59.98 2921 '36 2 150,000 B 2148 482 728 1705 16.54 24.97 58.49 2915 764 2,227,000 C 2780 528 903 2206 14.52 24.83 60.65 3637 760 2.764,000 D 2668 477 956 2118 13.43 26.92 59.65 3551 823 2,922,000 E 3499 611 1042 2778 13.79 23.51 62.70 4431 720 3,190,000 F 3370 549 1107 2675 12.68 2.5.^6 61.76 4331 782 3,387,000 G 3939 575 1299 3126 11.50 25.98 62.52 5000 794 3,970,000 H 3840 534 1340 3048 10.85 27 23 61.92 4922 826 4,066,000 4786 670 1676 3798 10.90 27.28 61.82 6144 818 5,026,000 there are plenty of instances in America where the results show a better record than this. Thus Whiting* records the continuous operation of a furnace where the ratio was 1.5. He does not give the percentage composition of the gases, but as he gives all the other data, I have calculated that it probably ran as follows : C0 2 15.8 per cent. CO 23.7 per cent. N 60.5 per cent. * Trans. A. I. M. E., Vol. XX, p. 280. j 100 METALLURGY OF IRON AND STEEL. At one of the large steel works in America known for its low fuel consumption, I am told that the average composition of the gases gives C0 2 14.5 per cent., CO 27 per cent, and N 58.5 per cent. This is a ratio of 1.88. f At a 65-foot furnace at the Pennsylvania Steel Works the average composition of the gases for one and a half hours showed as follows : C0 2 13.7 per cent. CO 23.7 per cent. N" 63.1 per cent, giving a ratio of 1.7 per cent. Samples taken on three other days gave very nearly, the same ratio, one being less and the other two somewhat more. i By referring to Table II-I it will be seen that practice A ap- proaches very close io the data given by Whiting and according to his figures it is probable that he used about 1650 pounds of coke per ton of iron and about 600 pounds of stone, his ratio being 1.5. The large steel works referred to with a ratio of 1.88 corresponds very closely to either practice C or D, while the furnace at the Pennsylvania Steel Works, with a ratio of 1.7 corresponds very well with practice C. It will be found that in every case the heat value of the gases gives approximately 50 per cent, of the heat value of the fuel charge, which was the conclusion arrived at in another section. In making these calculations, it is recognized that certain errors are unavoidable and that certain conditions have been omitted that have an influence on the result. Thus there, is a certain amount of silicon produced from the silica of the ore and coke and this silicon when it is reduced gives up its oxygen to the gases. In the same way a small proportion of calcium oxide is reduced, the calcium uniting with the sulphur as sulphide and the oxygen escaping with the gases. A certain amount of water may be decomposed and the hydrogen escape in the form of free hydrogen, while the oxygen goes off in the gases, and the oxygen formed by these three reactions is not accompanied by any nitrogen, while in our calculation we have assumed that the oxygen not coming from the ore and the stone was accompanied by the proper atmospheric proportion of nitrogen. But these refinements are not really necessary for a practical determination as the results are much more accurate than would be supposed at first glance. The carbon comes from the stone and from the fuel and from them alone, and the important point in all investigations of tunnel head gases is to find the amount of carbon THE BLAST FURNACE. 1Q1 escaping as C0 2 and as CO. It is a matter of very little moment how much or how little nitrogen accompanies these two gases, for the only gas of anyinterest after it leaves the tunnel head is the carbonic oxide. If a wrong calculation is made concerning the nitrogen, the figures will merely show that this percentage of carbonic oxide is either too high or too low. If our error shows too low a percentage of carbonic oxide we shall have a reduced calorific value per cubic metre and a larger volume, while if our calculations have erred in the other direction, we shall have too high a calorific value per cubic metre and too small a volume. In either case the product of the two, which will give the value of the tunnel head gas per ton of iron, will be a constant. Having thus found the heat value per ton of iron escaping in the waste gas we may find the horse power represented by that gas, and it is shown in Table II-I that according to the amount of fuel used the value will vary through very wide limits, according to the amount of fuel used. It is always best to assume that there will be progress in fuel economy, and this is the same thing as saying that there will be less and less heat value per ton of iron escaping in the tunnel head gases. Taking therefore the minimum, which is practice A in the table, we have 2,150,000 calories produced per ton of iron, which is equivalent to a total production of 645,000,000 calories in a furnace making 300 tons of iron in twenty-four hours. It is an accepted fact that one horse power used steadily throughout 24 hours represents 61,080 British thermal units or 15,394 calories, so that the total energy represented by the tunnel head gases will be 645,000,000, divided by 15,394 or about 42,000 horse power if every unit of force could be put into action. As a matter of fact fully one third of the gas goes to the stoves, leaving about 28,000 horse power for the boilers. It is a well known fact that the best boilers, when fired with coal under the most favorable conditions, can absorb 80 per cent, of the energy in the fuel, but it is seldom that good boilers, under ordinary conditions, utilize more than 70 per cent. In blast- furnace work, the results are often much worse than this, since the gases vary very much and it is impossible to supply the air in just the right quantity. Moreover, the gas does not burn readily at all times and it is impossible to avoid either a loss from unburned carbon or a carrying away of heat by an excess of air. The dust also deposits on the surface of the boiler and retards the absorption 102 METALLURGY OF IRON AND STEEL. of heat, while the low temperature of the gases as they enter the fire chamber preclude the best utilization of energy. From all these causes it is probable that the boilers at some plants do not appropriate more than sixty per cent, of the energy supplied to them, and this reduces the effective energy to 17,000 horse power, and as a compound steam engine utilizes only 10 per cent, of the energy delivered to it in steam, it follows that such an engine can develop 1700 horse power from a blast furnace making 300 tons of iron per day. The modern blowing engines for such a furnace will require not over 1500 horse power, and there will therefore be a slight excess of steam if the foregoing assumptions are correct, and a considerable excess if the boilers are more efficient than before assumed. It will be granted that actual results prove the calculations just elaborated, and that the available engine power is almost exactly as shown. This indicates that the figures are correct, and they may be summarized as follows : (1) From 3000 to 4000 cubic metres or 106,000 to 141,000 cubic feet of tunnel head gases are made per ton of pig-iron, when the fuel consumption is from 1600 to 2000 pounds per ton of iron. (2) About one third of the gas is needed to heat the stoves. (3) The boilers absorb and utilize only from 60 to 80 per cent, of the real heating value of the gases going to them. (4) The blowing engine absorbs only ten per cent, of the energy in the steam going to it. (5) If a gas engine be used, its efficiency must be compared not with the steam engine alone, but with the boiler and steam engine together. (6) With a higher coke consumption, the heating value per cubic metre will be increased somewhat and in addition the total volume of gases will be increase'd nearly in proportion to the weight of fuel. (7) The heat value of the tunnel head gas is about 50 per cent, of the total heat value of the coke, whether the consumption of fuel is high or low. (8) This calculation takes no account of wasteful furnaces or of those running on exceptionally bad ores, or on coke containing a large amount of hydrocarbons. Thus in Stahl un# Eisen, Nov. 1, 1901, Lurmann gives the composition of gases from different fur- THE BLAST FURNACE. 103 naces in Germany. A Westphalian furnace gave a ratio of 2.9 with 4.0 per cent, hydrogen; one in the Minette district a ratio of 2.75 with 3.0 per cent, hydrogen; one in Silesia gave a ratio of 5.5 and another contained 6.3 per cent, of hydrogen. With very poor ores the value of the tunnel head gases must be much greater than with a rich burden and if they are entirely utilized the power obtained from them will atone in some measure for the greater amount of fuel needed to smelt the leaner mixture. For this reason the use of gas engines is more important in the Minette district of Germany than in the United States. SEC. II j. The Utilization of the Tunnel Head Gases. (a) Use of the potential heat in stoves and boilers. It must always happen that in the combustion of the tunnel head gases, a great deal of heat is lost. One cause of this is the low tem- perature of the gases as they enter the combustion chamber of either the boiler or the oven and, as a consequence, the flame is very long and, as it is cooled by contact with the surfaces to be heated, either some CO will go to the stack unburned, or there will be a consider- able excess of air, giving a certain amount of free oxygen in the escaping gases together with its attendant nitrogen; sometimes there will be both a certain amount of CO and an excess of air. Under ordinary conditions of fuel consumption it is possible to calculate quite accurately how much heat is lost by either or both of these conditions, for a piece of coal or coke, if burned with just- the right amount of air, must ultimately give a certain per- centage of C0 2 and a certain percentage of nitrogen, and it makes no difference whether this combustion is all done in one place, as for instance the shallow fire of a cook stove, or whether it is parti- ally done in a gas producer and completed in a heating furnace. In either case the final result will be the same. In the blast furnace we have certain complicating circumstances, for oxygen is supplied by the ore without nitrogen, and carbonic acid is supplied by the limestone, so that the ratio of carbon to oxygen in the ultimate products of combustion is entirely different from the ratio that will result from the combustion of carbon under usual conditions, and it will be evident that this ratio will depend upon the amount of limestone used per ton of coke, and upon the amount of air. In this way the composition of the gas will vary in different districts, and with different furnaces, for if one uses 2300 pounds of coke per ton of iron the amount of carbon to a 104 METALLURGY OF IRON AND STEEL. pound of oxygen in the products of combustion will be greater than in a furnace running on the same ores and with the same limestone and using 2000 pounds of coke per ton of iron. TABLE II-J. Percentages of C0 2 and in Products of Combustion when Gases A and I (Table II-I) are Burned with Varying Amounts of Air. Excess of Air. Per cent. CO 2 Per cent, free O Gas A Gas I Gas A Gas I 27.55 19.75 25.21 17.64 5.93 6.30 A little consideration, however, will show that in a furnace using a regular amount of coke and a regular amount of limestone per ton of iron, it matters not at all how complete or incomplete the reactions may be in the furnace, and how much as a consequence the tunnel head gases may vary in composition from day to day, the ultimate products of combustion from the burning of these gases will be the same. It will also be shown in Table II-J that although different conditions of practice give unlike gases and that these may give unlike products of combustion, the variations in these products are so small that they may be neglected in all practical investigations into the question of heat utilization. In this Table II-J, we have taken gases A and I from Table II-I as representing two extremes of furnace practice and tw T o very different types of tunnel head gases. The first line gives the composition when the exact theoretical amount of air is supplied, assuming perfect com- bustion, while the second line gives the composition when double the needed amount of air is used. It will be seen that as far as prac- tical purposes are concerned the composition of the products is the same for both gas A and gas I, and it is therefore unnecessary to go into the refinement of calculating each individual blast furnace gas to find out what the composition of the products will be, for if two extreme cases give results so closely alike, we may safely as- sume that all gases will bear a close resemblance. THE BLAST FURNACE. 105- It Ml always happen in burning blast furnace gas or any other fuel, that a certain amount of excess air must be added to insure perfect combustion, and for this reason the composition has just been given of the products with a large excess of air. It is very often desirable to know just how much excess is added, and it has been the custom in making experiments in the com- bustion of coal under boilers to estimate the amount from the percentage of C0 2 in the products of combustion. In the case of burning coal this method is not far in error, for, as before explained, the products of combustion must always be the same for a given excess of air; but blast furnace gas is not constant, and the products of combustion are not exactly the same for differ- ent gases. A very much better exponent of the amount of excess air present is the percentage of free oxygen. This, of course, varies somewhat with different gases, but in Table II-J it is shown that a certain percentage of excess air gives about the same percentage of free oxygen in the products of combustion, even though the initial gases were quite different. It often happens that a certain amount of CO escapes unburned, whereby not only is there a loss of energy, but the composition of the products of combustion is changed somewhat, as less air is needed for what combustion takes place and therefore the volume is decreased and the ratio of the different components is altered. It also will happen under these circumstances that a given per- centage of free oxygen will represent a slightly different percentage of excess air than when no free CO is present, but I have found by calculation that the error thus caused is so slight it may be disre- garded. It is necessary, however, to consider the amount of CO which escapes in this way, and in Table II-K are shown the results of calculation on Gas D in Table II-I, which is chosen as being of average composition. The general conclusions to be drawn from these results are as follows : , (1) The products of combustion of all tunnel head gases are of approximately the same composition, and, therefore, the volume and weight produced per unit of coke charged will be the same. (2) The percentage of C0 2 in the products is not a good meas- ure of the amount of excess air. (3) The percentage of free oxygen in the products is a good measure of the amount of excess air. (4) When CO escapes unburned the composition of the prod- 106 METALLURGY OF IRON AND STEEL. ucts is altered not only by the presence of CO, but on account of the smaller amount of air needed for the imperfect combustion. (5) This alteration in composition is not sufficient to affect ma- terially the accuracy of the estimation of the amount of excess air from the percentage of free oxygen present, since the change in the proportion of oxygen is not great enough to invalidate the result. (6) The proportion of the ' unhurried CO in the products is a measure of the proportion of the original CO escaping. (7) In thus estimating the proportion of CO lost in the pro- ducts it is unnecessary to make any allowance for the percentage of excess air, since this does not cause sufficient variation within usual limits, to seriously affect the accuracy of the result. (8) The presence of CO in the products indicates a loss of com- bustible matter amounting to the proportions of the original energy .shown in Table II-K. TABLE II-K. Loss of Heat by Presence of CO in Products of Combustion. Per cent. CO in products. Proportion of energy lost. 0.65 to 1.00 100 to 2.00 2.00 to 2 80 2.80 to 3.80 3.80 to 5.00 5 per cent. 10 per cent. 15 per cent. 20 per cent. 25 per cent. The lower percentages apply to cases where no excess air is present, and the higher to those where there is 100 per cent, excess. The unburned CO in the products of combustion represents a certain loss of heat without any regard to the temperature at which these products escape to the stack, but in addition to this there is a cerain loss of heat from excess air, this loss depending entirely on the temperature at which the gases escape. If the products of combustion go to the chimney at exactly the same temperature as the air and gas entered the combustion chamber, or if they escape at the temperature of the atmosphere, then there will be no loss of heat, no matter how much air is used in excess ; but the products of combustion always do escape at a higher temperature and, by virtue of this, they carry away a certain amount of sensible heat, and that loss is greater just in proportion to the temperature of the escaping gases, and therefore each cubic metre of air which is THE BLAST FURNACE. 107 admitted in excess of the theoretical requirements carries away in the chimney a certain amount of sensible heat for every degree of temperature. Referring to Table II-I in Section Hi, giving the composition of different tunnel head gases, we may take Gas D as representing a very fair coke consumption and a very good carbon ratio. This gas is as follows : C0 2 13.43 CO 26.92 N 56.95 Calculating this gas as burning with different amounts of air we have Table II-L. The volume of the products does not increase exactly in proportion to the volume of air supplied, because there is a certain constant shrinkage due to the combustion of the origi- nal CO. Perfect combustion, without any excess of air, produces a certain volume of gases and any excess of air beyond this dilutes the gases by an exactly similar amount, and this excess air carries with it a certain amount of sensible heat. It is possible to calcu- late the loss carried away by this excess air alone; but the method adopted in this table is to give the total loss as carried away by the products of combustion, including the excess air. It will be seen, for instance, that when the gas is burned with just sufficient air and the products escape at 200 C.=390 P., the products of combustion carry away 12.5 per cent, of all the heat produced, while when 100 per cent, excess air is present, which is to say that air is supplied in double the quantity theoretically necessary, the products of combustion at 200 C. carry away 17.4 per cent, of all the heat supplied. At 600 C., which is just below a red heat, the combustion with the theoretical amount of air indicates that the products of combustion carry away 41.8 per cent, of all the heat produced, while with 100 per cent, excess air the products carry away 56.9 per cent. Thus 100 per cent, excess air means an additional loss of 5 per cent, when the products escape at 200 C. and 15 per cent, when they escape at 600 C. By comparing these results with the figures which have been given in Table II-K, it will be found that as long as the products escape to the stack at a moderate temperature, a very large excess of air is to be preferred to the escaping of a small quantity of CO ; thus it was shown that the presence of less than one per cent. CO indicated a loss of 5 per cent, of all the heat value, while Table II-L indicates that the escaping gases at a temperature of 400 C. carry away 26.6 per 108 METALLURGY OF IRON AND STEEL. cent, of aii the heat with no excess air and 30.6 per cent, of all the 'heat when 40 per cent, excess air is present, thus showing that 40 per cent, excess air is responsible for a loss of only 4 per cent, of the heat produced. Consequently it would be necessary to have 50 per cent, excess air and a stack temperature of 400 C. in order that the loss from such excess air should equal the loss from the presence of one per cent, of CO in the products of combustion. TABLE II-L. Data on Products of Combustion of Gas D (Table II-I)'. NOTE The specific heats of the gases were calculated for and 600 and for no excess of air and for 100 per cent, excess. Intermediate points are interpolated. Calorific value of gas = 823 cals. per cu. m. 4 a s & L i 20 40 60 80 100 SL 11 "Sfe i* t Per 100 volumes of gas burned. Per cent, of total heat generated which is carried away by the sensible heat of the products when these products escape at different temperatures ; also the specific heat of the products at these temperatures. Volume of air Supplied. Volume of products. 200C (390F) 400C (750F) 600C (1110F) Specific heat of gases. Per cent, of heat lost. Specific heat of gases. Per cent, of heat lost. Specific heat of gases. Per cent, of heat lost. 0.00 1.64 3.05 4.27 5.33 6.26 64 77 90 103 116 128 151 164 176 189 202 215 .342 .340 .338 .337 .335 .333 12.5 13.5 14.5 15.5 16.5 17.4 .362 .360 .357 .354 .351 .349 26.6 28.6 30.6 32.6 34.6 36.5 .380 .377 .373 .369 .366 .363 41.8 44.8 47.8 50.8 53.8 56.9 It will be found that the gases often contain both an excess of air and a certain amount of CO. If the mixture during its com- bustion could be passed through an indefinite length of hot pas- sages, it would hardly be possible that free oxygen and free CO' could remain uncombined in any large proportion, but there is a. limit to the completeness of combustion under practical circum- stances, as in burning a gas under boilers the flame comes in contact with cold metallic surfaces which check or retard combus- tion in the same way that a cold piece of iron put in a candle flame will stop the chemical action and will cause carbon and carbonaceous compounds to be deposited upon the metal. In this way a certain amount of carbon and oxygen escape from the stack without uniting one with the other. The way to prevent this is to cause combustion to be more thoroughly accomplished before it comes in contact with the water cooled surface. THE BLAST FURNACE. 109 It is necessary to note that in comparing analyses of products of combustion we should add together the losses shown by the excess air and the unburned CO ; thus we may have a loss of seven per cent, caused by excess air as indicated by free oxygen, and at the same time a loss of two per cent, indicated by the presence of a certain proportion of CO. Summarizing the foregoing conclusions and interpolating in the tables we may say that of all the heat produced, the losses in the products of combustion will be in round numbers according to the following schedule: Ten per cent, will be lost by any one of the following condi- tions : (a) By the sensible heat of the products of perfect combustion escaping at 160 C. 320 F. when no free oxygen is present. (b) By the sensible heat of the products escaping at 120 C. =250 F. when they contain from 6.0 to 7.0 per cent, of oxygen showing 100 per cent, excess air. (c) By the presence of from 1.3 to 1.9 per cent. CO. Twenty per cent, will ~be lost: (a) By the sensible heat of the products escaping at 300 C. =570 F. when no free oxygen is present. (b) By the sensible heat of the products escaping at 250 C. =480 F. when they contain from 6.0 to 7.0 per cent, of oxygen showing 80 per cent, excess air. (c) By the presence of from 3.0 to 4.0 per cent. CO. Thirty per cent, will be lost: (a) By the sensible heat of the products escaping at 450 C. =840 F. when no free oxygen is present. (b) By the sensible heat of the products escaping at 330 C. =630 F. when they contain from 6.0 to 7.0 per cent, of oxygen showing 100 per cent, excess air. (c) By the presence in the products of 6.0 per cent, of CO. Forty per cent, will be lost: (a) By the sensible heat of the products escaping at 600 C. =1110 F. when they have been burned with the theoretical amount of air. (b) By the sensible heat of the products escaping at 44( 820 F. when they contain from 6.0 to 7.0 per cent, of oxygen -'showing 100 per cent, of excess air. It will be understood that a large percentage of loss may occur 110 METALLURGY OF IRON AND STEEL. by a combination of any two of these factors, as for instance, when the products contain both free oxygen and unburned CO, under which conditions the total loss is the sum of the two factors. (b) The Use of Sensible Heat of Gas in Stoves and Boilers. When the tunnel head gases are taken directly to the stoves or to the boilers without scrubbing, all the sensible heat of the gas is used as the temperature of the resulting combustion is just that much higher and its efficiency just that much greater. When the gas is scrubbed, this sensible heat is lost and there is an addi- tional disadvantage in the water vapor that will be carried to the stoves or boilers. When the gases are taken directly from the tun- nel head to the combustion chamber there is considerable steam present, but it is in the form of a gas, and if this is subsequently dissociated with absorption of heat, the hydrogen produced is again oxidized into steam and therefore there is no heat lost, but in the passage through a scrubber there is a considerable quantity of water carried along in the shape of fog and all this moisture must be converted into steam in the stove or in the boiler. It may very likely be advantageous in many cases to scrub the gas in spite of this for there is no doubt that if the scrubbing were perfectly successful, in other words if every particle of for- eign matter were to be eliminated, there would be a great advan- tage in having a clean gas, for instead of the crude appara- tus in use for burning these gases, we could then substitute some- thing in the form of a Bunsen burner and get almost perfect combustion in exactly the place wanted, but the great difficulty is that we cannot remove the last traces of the sublimate and these clog the action of any Bunsen burner or anything approach- ing its structure. The future will doubtless see a very much better arrangement for burning these blast furnace gases than now exists, and it is possible that thorough scrubbing will be a prerequisite to the introduction of such methods. It has already been stated that Gas D, in Table II-I, represents a very good fuel consumption and a good carbon ratio, and it was shown that when a furnace is running under these conditions it produces 3551 cubic metres of gas per ton of pig-iron. Calcu- lating the amount of air needed to burn this we find that 2060 cubic metres are called for theoretically, while with 100 per cent, excess of air, just double that quantity, or 4120 cubic metres, will be required. It is shown in Section Ilh that a furnace producing THE BLAST FURNACE. Ill such a gas requires 2687 cubic metres of air per ton of iron to be supplied by the blowing engines, and it is clear, therefore, that if the tunnel head gases are burned with 30 per cent, excess air, the amount of air needed for their combustion in the stoves and boilers equals the amount of air required from the blowing engines. It is probable that more than this proportion of excess air is generally used so that the air needed for combustion exceeds the amount sup- plied in the blast. It is probable that few f urnacemen appreciate this fact, or will even believe it. If one-third of the gas is taken to the stoves then the stoves are receiving more than one-third of the amount of air delivered by the blowing engines, and if the boilers are re- ceiving two-thirds of the tunnel head gases then the air inlets at the combustion chamber are receiving two-thirds as much air as the blowing engines deliver to the tuyeres. It is almost out of the question to pre-heat all this air for, by the nature of the case, if the volume of air required by the tunnel head gases is as great as the volume required by the furnace, it would require as large an outfit of stoves as is required by the furnace, and there is no available place for the heat to come from except from the com- bustion of the gases themselves, and this would be wasting at one end and gaining at the other. It will be shown later that the in- troduction of gas engines may render possible the preheating of this air by the heat of the waste gases escaping from the cylinders of the engines. (c) Use of tunnel head gases in gas engines. It is a well-known fact that blast furnace gas can be used in gas engines for developing power, and it is just as well known that a given amount of gas will develop about twice as much energy in a gas engine as it will if burned under boilers and the resultant steam be used in a steam engine. High authority has stated that the available power is 3.6 times as great, under prac- tical conditions. I prefer for purposes of illustration to make the conservative assumption that the gas engine will give twice the power. It is highly probable that there will be less irregularity if the gas is burned in gas engines than if it is burned under boilers, because the real calorific power of blast furnace gas does not vary as much as is generally supposed. It does often for a considerable period possess a strong disinclination to burn under a boiler, this 112 METALLURGY OF IRON AND STEEL. being particularly noticeable when the furnace is very hot, for the f urnacemen then say that the gas is "gray" and that it is "poor," because it will not burn with a clear flame; but this gas is of the same composition as free burning gas, and if it is mixed with a proper amount of air in the cylinder of a gas engine and ignited by an electric spark, it should give the normal amount of energy. In Section Hi it was shown that under certain assumptions of rather low fuel the tunnel head gases contained sufficient energy to produce 42,000 horse power if every unit of force could be utilized. It was also stated that under usual conditions at least one-third of the gas was used in heating the stoves, leaving an equivalent of 28,000 horse power in the gas going to the boilers, but that owing to the losses in boilers and engines we found that very little more power was developed than was necessary to run the blowing engine. It is possible to increase this surplus somewhat by having a better boiler plant than was assumed, and it has been shown that a furnace using a greater proportion of fuel will fur- nish a much greater surplus of power, but it was considered best to presuppose a reasonable economy of fuel with a fair outfit of boilers. In. order to compare a given set of conditions where steam engines are used with similar conditions where gas engines are installed, it will be assumed that the gases available after the stoves are supplied contain 28,000 horse power. If the plant is equipped with an extra good boiler plant, the steam will represent 75 per cent, of the energy in the gas, or 21,000 horse power, and if good compound engines are used it will be possible to develop from this about 2100 horse power, so that if the blowing engine calls for 1500 horse power there will be a surplus of 600 horse power available for pumping and for outside uses. If it is supposed that just enough steam is produced to run the blowing engines and the surplus gas is diverted to gas engines, and if it is supposed that twice as much surplus power is developed in this way, it follows that each 300-ton furnace will furnish 1200 excess horse power. This increase is important, and seems to fill the minds of many men as one of the coming economies, but as a matter of fact it is merely the beginning, for the first step in true economy is the operation of the blowing engine by gas, since in this way instead of developing a total of 2100 horse power there will be a total of 4200 horse power, and after subtracting the 1500 necessary for blowing there will remain a surplus of 2700 horse THE BLAST FURNACE. 113 power for outside uses. Thus the use of gas engines for auxiliaries only gives just double the amount of power available for outside uses, but the use of gas for blowing engines gives four and one-half times as much surplus as furnished by a steam plant. This calculation presupposes that the steam engine utilizes 7.5 per cent, of all the energy contained in the gases supplied to the boiler, and that the gas engine utilizes 15.0 per cent. The best steam plants do better than this, but it must be considered that blast furnace gas is not the most desirable kind of fuel and that the operation of a blowing engine against a varying load does not pre- sent the best conditions for steam economy. For the same reasons the efficiency of the gas engine is taken considerably below what has been done under favorable conditions. . Taking the figures just found it is shown that for each 300 tons of pig-iron produced there will be a surplus of 2700 horse power, and in a steel plant making two thousand tons of pig iron per day this is equivalent to 18,000 horse power, which is ample to run all the converting plants and rolling mills necessary to finish this quantity of pig-iron into rails, or into the ordinary forms of finished material. In order to utilize this source of supply to the best advantage, it will doubtless be necessary to install a central electric station in which all the gas is used to develop electric power which is then distributed to motors that drive the rolling mills. If this plan can be carried out, no boilers will be used in the entire steel works, the only fuel being that used for heating. The importance of this problem has been long recognized and it may be well to record the steps that have been taken to reach a solution, and then explain why the introduction of gas engines is so long delayed. The historical facts may be thus summarized : ^ In May, 1894, B. TL Thwaite applied for a patent in England which was granted in May, 1895 (No. 8670), for a method of purifying blast furnace gas for use in gas engines; acting along the lines laid down by Thwaite, the first gas engine driven by furnace gas was set to work in February, 1895, by James Riley, manager of the works at Wishaw, Scotland. This motor was a success and was in operation four years later. At this time the importance of the work was understood, calculations being made on the saving to be expected, and from that time until now, various gas engine builders have experimented in this field. With one 114 METALLURGY OF IRON A3TD STEEL. exception the cleaning of the gas has been considered necessary, this exception being the Cockerill Co., which announced that the gas could be used in its engines without scrubbing, but the results have not been entirely satisfactory and the washing of the gas is now looked upon as a necessity by the company at Seraing. In 1899 it was promulgated far and wide that the whole problem was solved and American engineers were looked upon as being behind the times for not equipping their plants with gas engines. During that year I visited every gas engine in Europe which was operated by furnace gas. Every builder was anxious to show his engine as an example of successful construction, and most managers of works were willing to exhibit their plants as evidence of their pro- gressiveness, but nevertheless I put on record in an official report the following conclusions : (1) That there was not a thoroughly satisfactory installation in existence. (2) That some engines then in operation and construction were structurally weak, while others were too complicated and would easily be deranged by dust. (3) That in spite of all assertions, the gas must be cleaned to give good results, and that no method then in use did wash the gas satisfactorily or sufficiently. (4) That gas engines could be made simple in construction, and strong in design ; that some way would soon be found to wash the gas; and when this was done, gas engines would come to stay. Having confidence in the future, we operated a gas engine at Steelton for some months in the year 1900, but the dust gave rise to troubles which might easily be obviated with a different type of construction. This was the first engine in America driven by furnace gas, and the only other engine up to the present time is one of small size operated in the early part of 1902 by the Mary- land Steel Co., very satisfactory results being obtained. I believe that history has proven the correctness of the above judgment of European engines in 1899, an opinion shared by other American engineers who saw the facts just as clearly and decided to wait. Those who rushed into the breach, on the Continent, deserve the thanks of the engineering world, but they have paid dearly for their glory. At times when the papers have been giving drawings and pictures of new installations, and when these plants have been held up as examples for American engineers to follow, these same THE BLAST FURNACE. 115 plants, almost before the ink on the pictures was dry, have been shut down with their cylinders cut to destruction, or with parts crippled by breakage. ^HCW^OOOW^ ~^ *' ^T 1 C Q c^fD O ef- ''I "^Ocoo fif! . IllifS'i <-i o> ^ I 00 : : ft: g: : : : 8 O ^ Kind of Engine. Germany. Luxemburg. France. Belgium. Austria. England. Spain. Russia. Italy. Total. Total for Each Type. O HI Q <. l_l S. 5' It was not until the latter part of 1901 that the gas engine could be called a success, breakages having been so frequent that builders were obliged to replace with stronger constructions while 116 METALLURGY OF IRON AND STEEL. the destructive work of dust has led to the development at Dude- lingen and Differdingen of the cleaning device where the gas is drawn into a centrifugal fan provided with an internal spray of water. Table II-M gives a list of the engines now in operation in Europe, while America has none. It is not a proud position that American engineers have occupied in waiting for others to do the work, but it may safely be stated that we are richer than if we had been building gas engines. A most important point which bears upon the matter to-day is the fact that up to the present time a thoroughly well built gas engine, with its scrubbers, its reserve units, and reserve producers, has cost so much more than a steam engine that the fuel saved would no more than pay the interest and depreciation on the extra investment. These conditions are changing and the price of engines will inevitably decrease as makers adapt their shops to the new work and as the risks of loss in starting new machines becomes less formidable. It is now expected that before many months, one of the new American plants will follow the lead of some of the foreign works and will offer something more than mathematical calculations on the benefits of blowing engines driven by gas. In view of the possibility of such developments it may be well to review briefly the fundamental principles of gas engine construction. -^When the piston of a steam engine arrives at the end of its stroke, the valves open and a connection is thereby made directly with the boiler, and with what may be considered an inexhaustible supply of power. That is to say, a steady pressure is immediately put upon the piston head, and no matter how fast or slow the piston moves, this pressure follows, like a perfect spring, just as far as desired. In practice, the cut-off is about one-third the length of the cylinder, and during that time, and for that space, the pressure in the cylinder and against the piston is nearly equal to the pressure in the boiler, while beyond that point, the piston is carried forward by the expansive force of the steam and finally at the end of the stroke by the momentum of the flywheel. The point of cut-off in modern engines is controlled by the governor, so that the amount of steam admitted to the cylinder is exactly in proportion to the work to be accomplished. In older and more wasteful types the same end is reached by the throttle valve, which indirectly regulates the pressure of the steam admitted, but in either case the initial pressure, by which is meant THE BLAST FURNACE. 117 the maximum pressure at the beginning of the stroke, can never exceed the boiler pressure, unless we imagine a completely dis- ordered condition of valves, whereby the cylinder is filled with steam at high pressure on the wrong side of the piston, creating a great compression. A gas engine differs radically in its principles from this descrip- tion. It is a cannon, with its projectile fastened to a crank shaft, and this cannon is required to explode every second and keep ex- ploding indefinitely, without getting hot or deforming even a valve. In addition to the structural problem concerned in this state- ment, there are certain thermal and chemical questions : (1) There must be something corresponding to a governor, whereby the speed is controlled, and this must regulate either the amount of gas entering on each stroke, or the number of admis- sions per minute. The latter plan, the "hit or miss," is a common one, it being arranged that when the engine runs over a certain speed, the gas valve fails to open, and the fly wheel does the work. >_. (2) In using gas of poor quality, like producer or blast furnace gas, it is necessary, in order to get much power out of an engine, that the explosive mixture should be compressed before ignition. (3) The pressure obtained after ignition will evidently depend very much upon the pressure before ignition and as the cubical content of the exploding chamber is a constant, it is evidently impos- sible to have a constant pressure before explosion, if there is any variation in the volume of gas and air added. It is for these and other reasons that the "hit or miss" system has been generally adopted. (4) The "hit or miss" system is wrong, because it produces irregularities in speed of revolution. Supposing that the engine is a mere shade too fast and the admission "miss," then the whole cycle must be completed of perhaps two complete revolutions be- fore another explosion can occur, and the flywheel must do all the work in that time. If the work is variable it may reach its maii- mum during this idle period and the speed decrease far below what would be allowable for many purposes, as for instance, in the production of an alternating current. (5) The above mentioned period of two revolutions is not true of all engines, but in order to understand any gas engine it is necessary to keep in mind this original Otto cycle. 118 METALLURGY OF IRON AND STEEL. (a) Explosion, high initial pressure, forward stroke of piston, ending with a cylinder full of dead products of com- bustion which will not condense, but must be removed before the next supply of gas enters. (b) The backward stroke of piston sweeping out the dead (c) Forward stroke of piston, sucking in a new supply of gas and air in measured quantities. (d) Backward stroke of piston with all valves closed, com- pressing the mixture of gas and air just admitted, the result- ant back pressure being dependent upon the cubical content of the space left for the exploding chamber, and the amount of gas and air admitted. Thus in a single cylinder engine, working on the Otto cycle, there is only one impulse for two complete revolutions, and this impulse is an explosion throwing a great strain on all the working parts. (6) The very high pressure caused by the explosion is accom- panied by a very high temperature, and it is difficult to make valves which will stand the work, while cylinders are always water jack- eted and even pistons are sometimes so cooled. (7) If too high a back pressure be attempted, the explosive mixture may spontaneously ignite before the piston reaches the end of the stroke, with the production of enormous strains on all parts of the mechanism. (8) If too low a pressure be used the gas may fail to ignite, and the igniters be covered with dust, which is pretty sure to cause other failures to ignite in subsequent admissions. (9) The presence of mineral dust in blast furnace gas increases these difficulties, not simply by the wear on sliding surfaces, but by the interference with all valve adjustments and seats, giving rise to leakages and back explosions. (10) The limitations just described concerning the admission of varying amounts of gas and air, and the control of compression, render it impossible in most engines to get good fuel economy under varying loads, although some of the later types attempt to attain this end. With a modern steam engine rated at 1000 horse power, the consumption of steam is nearly proportional to the load whether the engine is developing 1200 or 800 horse power, while the waste THE BLAST FURNACE. 119 will not be prohibitory even if the load falls to 500 or rises to 1500 horse power. On the other hand most gas engines, under such variations, FIG. II-F. INDICATOR CARDS FROM GAS AND STEAM ENGINES. . 100 Ibs. 32"x 48"CORLISS CONDENSING ENGINE. 100 Ibs. 40 x 48 PORTER-ALLEN ENGINE. 15"x22"THREE-CYLINDER WESTINGHOUSE GAS-ENGINE. ;show a much greater consumption of fuel than with their normal load, and they give an unsatisfactory speed regulation. It may also be said that no overload is practicable, for the rating is the maximum capacity. The indicator cards given in Fig. II-F, will 120 METALLURGY OF IRON AND STEEL. exhibit the difference between the work of a steam engine and an ordinary gas motor. The term "ordinary" gas motor is used as the Letombe engine aims to overcome this difficulty ; under a light load this engine takes a small quantity of gas and a very large quantity of air, say to a total volume of 100, and compresses the mixture to a pressure of say 300 before ignition. Under 'a full load it takes a larger amount of gas and the proper amount of air to give the best explosion, the total volume being say 70, and this is compressed to a pressure of say 200. These figures are not accu- rate, but they will illustrate the principle of getting a higher com- pression for the poorer mixture, and thus always obtaining a sharp explosion. (11) In a gas engine there is probably an accentuation of a con- dition existing to some extent in heavy steam engines. When the weight of the reciprocating parts is very great, the force of the steam at the beginning of the stroke is absorbed by the inertia of the reciprocating parts and the effect upon the crank pin may sometimes be negative. It would seem probable that in a gas engine this condition should be more strongly marked, as the part* are very heavy and the ratio of crank to connecting rod is larger than in the steam engine. From what has been said it will be seen that there are many difficulties, but the foreign engineers have struggled with them. The greatest bugbear is the old four cycle system, giving only one impulse in two revolutions, thereby reducing the horse power of the engine and giving a variable speed. The most radical depar- ture is in what is known as the Oechelhauser motor, first installed at Horde and shown in Fig. II-Gr. In this construction the cylinder is open at both ends and is a true cylinder throughout, save the opening near either end for gas, air and exhaust. There are two pistons working in opposite directions, the piston rods projecting out through the two open ends. When they are nearest together the space between them is the ignition chamber, and the explosion forces one piston in one direction and the other in the opposite way, nothing being exposed to the force of this explosion cave the smooth walls of the cylinder and the heads of the pistons. When the pistons reach the end of their stroke they uncover pas- sages in the walls of the cylinder which connect with the exhaust and then with both air and gas, the latter being under pressure. Air is blown through from one end to the other to wash out the THE BLAST FURNACE. FIG. II-Gr. OECHELHAUSER GAS ENGINE. 121 122 METALLURGY OF IRON AND STEEL. dead products of combustion,, and furnish air for the next explo- sion, and then a measured quantity of gas is forced in. All this is done quickly and then the two pistons on the return stroke close these openings and pass over them and slide toward each other, compressing the mixture between them ready for the electric spark. An impulse on every revolution is thus obtained and the valves are removed from all heat and all shock. The one inherent fault in this type of machine is the system of crank shaft and connecting rods. It is evident that both pistons must be connected with the same shaft, and this makes necessary that one piston rod .must be supplied with a cross head and two very long connecting rods, and that the main shaft itself be of a very complicated construction with a number of bearings. The earlier engines of this kind were not strong enough and the later examples have been made much heavier. The Koerting engine, shown in Fig. II-H, is designed to take an impulse on each and every stroke, a compressor being used to force the gas and air into the cylinders. (d) Preheating the air going to stoves. Under steam engine practice the sensible heat of the tunnel head gases is completely used except what is lost by radiation, for a warm gas entering the stoves or boilers means a correspondingly increased production of heat. When the gas goes to scrubbers on the way to the gas engines, this sensible heat is wholly lost, and it may be worth while inquiring whether this heat can be used to preheat the air going to the stoves. If one third of all the gas goes to the stoves and 3551 cubic metres of gas are made per ton of iron, then 1284 cubic metres of gas go to the stoves per ton of iron, and if 30 per cent, excess air be added, then about 1060 cubic metres of air must be supplied. The other two-thirds of the gas will go to the gas engines and we will, therefore, have the sensible heat of about 2270 cubic metres of gas available for heating 1060 cubic metres of air. Assuming that the air be heated to practically the same temperature as the gases, i. e. from 16 C= 60 F. to 120 C=250 F. the heat thereby given to the air will be 1060X.307X104=33880 calories while the total heat created by the combustion of the gas in the stoves will be 1284X823=1,067,000, THE BLAST FURNACE. 123 FIG. II-H. KOERTING DOUBLE-ACTING GAS ENGINE. 124 METALLURGY OF IRON AND STEEL. so that the gain from thus preheating the air is a little over 3 per cent, of the total heat produced in the stoves. If the tunnel head gases were much hotter the gain would be cprrespondingly in- creased, but with a cold top the gain will not warrant any expendi- ture of capital. It is quite possible, however, that the exhaust from gas engines can profitably be employed in this work. If two-thirds of the gases are used in engines the products of combustion will far ex- ceed the volume of air going to the stoves, and if these products escape at a high temperature the air for the stoves could be heated very nearly to that temperature by a suitable system of pipes and a great improvement made in the efficiency of the ovens, while the amount of gas needed by them could be decreased. SEC. Ilk The Relation Between the Physical and Chemical! Qualities of Cast Iron. The pig iron used in the great steel works of the country is valued entirely according to its chemical com- position, and little or no account is taken of it's physical appear- ance, commonly known as its "fracture/ 5 save as a rough and ready way of estimating in advance its chemical formula. Within com- paratively few years there has been a strong movement among pig-iron users and manufacturers to adopt the same system through- out the general trade, but it is difficult to alter the prejudices of generations, and it is hard for uneducated foundrymen to cast away all their knowledge of fractures gained by years of obser- vation, and rely on tables of analyses with mystic decimals show- ing the proportions of elements of whose very existence they have been ignorant. The matter is not made better by the fact that there are many things not fully understood concerning the relation between the chemical and physical qualities, one instance in point being the superiority of charcoal cast-iron, and the better quality obtained by melting in air furnaces. As long as such phenomena are not fully explained by the scientists, or as long as they disagree in their explanations, so long must the aforesaid foundrymen be par- doned for clinging to their convictions. The trouble is that most of the deductions concerning cast-iron have been made without complete data, and by men who did not know that the data were incomplete; who, for instance, took no account of manganese since it was not given in the report of the chemist ; or who accepted glaring palpable errors like those pointed THE BLAST FURNACE. 125 out by Prof. Howe, where an average of a whole class of iron is reported as containing nearly 15 per cent, of carbon, with one speci- men holding over 16 per cent, of graphite. When such absurdities are put into the hands of unscientific foundrymen it is no wonder that the conclusions are slightly erratic. The most scientific discussion of the constitution of cast-iron has been contributed by Prof. Howe. His opinions are not neces- sarily right because they are enunciated in scientific language, or because they embody the latest results of microscopic investiga- tion, but they are very likely to be right, as the reader may feel quite sure he is not being misled by any fallacy. In reading any such paper on abstruse subjects, it is easy to be sidetracked and to overlook the continuity of the line of thought, for we are asked to concentrate into a few minutes the work of months, but the investigator who has worked for months or years is supposed to consider every sidelight and every difficulty, and the weight of his conclusions oftentimes depends fully as much upon his repu- tation for clear thought as upon the extent of his practical experi- ence. The argument of Prof. Howe is that pig-iron and steel form a continuous series; that, from one point of view, steel is a grade of cast-iron and cast-iron a grade of steel. This is an assumption which needs no justification to the open-hearth melter, who is ac- customed to see a bath of pig-iron change by insensible grada- tions through a thousand intermediate stages from the richest pig to the condition of finished steel. It is shown in Chapter XV that steel is a mixture or alloy of two components, ferrite and cemeniite, but that these two sub- stances combine together in one definite proportion and in one proportion only to form pearlite. The proportion is seven parts of ferrite to one of cementite, so that pearlite contains neces- sarily about 0.80 per cent, of carbon. It follows that steel or iron containing more than 0.80 per cent, of carbon cannot all be pearlite, but that the pearlite which is present will contain, if the metal is cooled slowly, the full quantity of carbon represented by 0.80 per cent, of the mass, and that the rest of the carbon will exist in some other form. Part may exist in combination with \ the iron as cementite, and part may exist in the free state as graphite. Steel containing 0.90 per cent, of carbon if cooled slowly will be mostly pearlite, but will usually contain a trace of 12G METALLURGY OF IRON AND STEEL. graphite and a certain amount of cementite. Metal containing- 4 per cent, of carbon cannot contain any more pearlite than the steel just mentioned, but there will be just so much more carbon to form either graphite or cementite. The amount of graphite will depend upon several conditions. A. hot blast-furnace will give a higher percentage than a cold furnace, and high silicon will also cause the separation of free carbon, while manganese and sulphur will cause the carbon to remain combined. After subtracting the graphite from our cal- culation, the remaining carbon and iron form a matrix which may be assumed to follow the laws that hold good for all the grades that are usually known as steel. Thus, as stated by Prof. Howe, cast-iron with 1.25 per cenfl combined carbon is really steel of 1.25 per cent, carbon, but weak- ened and embrittled by graphite. In the same way he regards cast-iron with 3 per cent, of combined carbon plus 1 per cent, of graphite as essentially a mechanical mixture of two substances; (1) 99 parts white cast-iron, containing 3 per cent, of combined carbon, and (2) 1 part of graphite. The contention that graphite "weakens and embrittles" cast- iron is directly opposed to the views of most practical men, but it seems as if he has made a good argument, for his reasoning is founded on the undeniable fact that ordinary pig-irons, when con- taining about the same proportion of silicon, manganese and sul- phur, carry the same proportion of total carbon, no matter whether they are gray or white. It follows, therefore, that an increase in the proportion of graphite means a corresponding decrease in the proportion of combined carbon, and since one quarter of the total carbon is in the form of pearlite, and since cementite must contain 6.57 per cent, of carbon, it follows that if much carbon exists as graphite, the proportion of cementite present rapidly decreases and the proportion of soft ferrite rapidly increases, with a consequent toughening of the mass. This toughening is usually ascribed to graphite, when in reality the graphite weakens the iron by destroying its continuity, but the injury caused in this way is entirely overshadowed by the fact that as long as it exists as graph- ite, it cannot at the same time exist as cementite. Thus an element like silicon will toughen iron because it drives the carbon into the condition of graphite, while manganese will make it brittle, because it causes it to combine. It is a generally THE BLAST FURNACE. 127 accepted theory, although not undisputed, that charcoal pig-iron contains less carbon than coke-iron, and if this is true, the better quality of charcoal-iron could easily be explained by a low propor- tion of cementite and also a low proportion of graphite, two con- ditions which can seldom be found in iron. This would also explain why melting charcoal-iron in cupolas takes awav its superi- ority, for the iron absorbs carbon in melting until it is of the same composition as irons made in a coke furnace, so that to retain its quality it is necessary to melt it in an air furnace. It is necessary, however, to consider that the lower proportion of carbon in char- coal-iron is not an established fact, for some authorities, like Stead, aver that the opposite is the case, and Prof. Howe, in a private com- munication, after reading this manuscript, states that the evidence on this point is inconclusive, and that the lower content may be- assumed only as a probability. TABLE II-N". Composition of Various Pig-irons and Spiegels. ow ss II II C 32 if 82400 51800 11.2 18.9 81180 80775 80400 49760 50200 49050 14.2 15.5 16.0 22. 22.5 22.4 32100 31050 81100 81470 51000 50650 50530 50830 52570 13.0 14,6 17l8 17.2 19.0 19.9 21.6 24.6 With less careful work there is a constant retrogression in quality as the size of the finished piece increases, and this is usually recog- nized in specifications, as will be seen by Table III-D, which is copied from a paper by A. E. Hunt.* SEC. Illf. Heterogeneity of wr ought-iron. The most com- plete investigation on the subject of wrought-iron is a report by Holleyf on the work of a Board appointed by the United States Government to test material for chain cables. It was found that the tenacity of 2-inch bars for chain cables should be from 48,000 to 52,000 pounds per square inch, while 1-inch bars should show 53,000 to 57,000 pounds. This conclusion is reached after very careful reasoning, and it illustrates the profound influence of this one item of reduction in rolling. It will be evident that unless the history of the bar is known, ordinary chemical analysis will fail to give any information as to whether it has been rolled from a pile 4 inches square or from one 7 inches square. In the making of rounds, which was the only shape tested by the Board, there is op- *On the Inspection of Materials of Construction in the United States. Journal I. and 8. I., Vol. II, 1890, p. 299. t The Strength of Wrought-Iron as Affected by its Composition and ty its Re- duction in Rolling. Trans. A. I. M. E., Vol. yi, p. 101. WROUGHT IRON. 137 portumty for very bad practice in beginning to form the piece too early in the operation, for there is a much better chance to work and weld the iron in closed rectangular passes than in the CD J2.I 2*H $?f 5: .? j i_j i * >tCT nil Hill a "! | R C EC i punr|sau| "SSB""""'* 21 8 8 S. I d- c- A SB ^ * P 1 1 1 1|! I 6 I |a| 2 o o" a? fill p P P | g 8 8 5 I* P P r r Limit of elasticity, Ibs. per square in. Ultimate strength, Ibs. per square in. Elongation in 8 in., per cent. Reduction of area, per cent. Angle of bend. l! sr ir ^ w f B s- 5 is P formation of round sections. Usually, a bar which has not re- ceived sufficient work will contain an abnormal percentage of slag, and this can be determined in the laboratory; but a slight excess- 138 METALLURGY OF IRON AND STEEL. does not necessarily imply that the iron has not been well worked, for it may arise from viscosity of the cinder, rendering its expulsion difficult. In any event, it will be seen that, although a certain quantity may benefit the metal by preventing crystallization, any- thing beyond this must decrease the cohesion of the particles of iron. In the investigation just mentioned, it was found that the slag varied from 0.192 per cent, to 2.262 per cent, of the total weight of the iron ; and it must be remembered that these tests were made on material destined for a service calling for the best product of the mill. Some makers may have supposed that the presence of slag would facilitate welding, but the investigation did not bear this out, for it is distinctly stated in the report that, while "slag should theoretically improve welding, like any flux, its effect in these experiments could not be definitely traced." On the con- trary, the iron which was highest in slag (2.26 per cent.) "welded less soundly than any other bar of the same iron, and below average as compared with the other irons." TABLE III-E. Variations in the Character of Different Specimens of the Same Brands of Wrought-Iron and of Different Irons as Submitted to the United States Board for Testing Chain Cables. Subject. Same Iron. All Irons. Min. Max. Min. Max. Carbon, per cent., .026 .042 .064 .512 .015 .512 Phosphorus, per cent., .065 .095 .232 .250 .065 .817 Silicon, per cent., .028 .182 .182 .821 .028 .821 Manganese, per cent., tr. .021 .059 .097 tr. .097 Slag, per cent., 0.674 1.248 1.738 2.262 0.192 2.262 Ultimate strength, pounds per square inch, 56201 47478 69779 57867 47478 69779 Elongation in 8 inches, per cent., 11.7 14.1 20.6 82.5 6.5 82.7 Reduction of area, per cent., 27.7 16.0 59.8 81.5 7.7 59.8 The percentage of slag not only varied in different brands of iron, but in pieces of the same make. This was true also of all WROUGHT IROX. 139 the factors investigated. Table III-E shows the variations in the same make of iron, two extreme cases being given under each head. It also gives the maximum and minimum individual records. SEC. Illg. Conditions affecting the welding properties of wrought-iron. These conditions of varying work, percentages of slag, and irregularity of the same irons, not to mention the possible overheating of piles in the laudable effort to produce a perfect weld, complicate so fundamentally the relation between the chemical com- position and the physical properties, that it need not be wondered that the committee could not find the exact influence of each chemi- cal component. There was formulated, however, the following very valuable conclusion: "Although most of the irons under consid- eration are much alike in composition, the hardening effects of phosphorus and silicon can be traced, and that of carbon is obvious. Phosphorus up to .20 per cent, does not harm and probably im- proves irons containing silicon not above .15 per cent, and carbon not above .03 per cent. None of the ingredients, except carbon in the proportions present, seem to very notably affect the welding by ordinary methods/' Regarding this last clause it should be said that the highest ivilphur in any sample was .015 per cent., which is very low; but that copper was present in one instance up to .43 per cent. ; nickel up to .34 per cent., and cobalt up to .11 per cent. Moreover, the high percentages of these three elements were coincident in one "bar, yet welding gave fair results, notwithstanding that phosphorus was higher than was found advisable. A careful reading of the evidence, however, indicates that the experiments were far from conclusive as to these elements. This matter of welding power was of special moment in iron for chain cables, but it is also the very root of the entire process, for the integrity of the finished bar depends upon the completeness of the welds between the different particles. In Chapter XIX the welding of iron and steel will be discussed at greater length. CHAPTER IV. STEEL. SECTION IVa. Definition of steel. Although it seems a per- fectly simple matter to give a definition of steel, the task has never yet been accomplished to the satisfaction of all concerned. A true formula must apply not only to all the metals commonly designated by the term, but to all compounds which ever have been, or ever will be, worthy of the name, including the special alloys made by the use of chromium, tungsten, nickel and other elements intro- duced to give peculiar qualities for special purposes. Moreover it has been shown in Sec. Ilk that the latest researches show no dividing line between the softest steel and the ordinary grades of pig-iron. Prior to the development of the Bessemer and open-hearth pro- cesses there was little room for disagreement as to the dividing line between steel and iron. If it would harden in water, it was steel ; if not, it was wrought-iron. When the modern methods were in- troduced, a new metal came into the world. In its composition and in its physical qualities it was exactly like many steels of commerce, and naturally and rightly it was called steel. By de- grees these processes widened their field, and began to make a soft metal which possessed many of the characteristics of ordinary wrought-iron, and which was not made by any radical change in methods, but simply by the use of a rich ferromanganese. Not- withstanding this fact, some engineers claimed that the new metal was not steel, but iron. The makers replied that it was made by the same process as the hard steel, and that it was impossible to draw a line in the series of possible and actual grades of product which they made. The problem rppidly became of great importance, since the filling of engineering contracts and the interpretation of tariff schedules depended upon the application of the one term or the other to the soft product of the converter and the melt'ing-f urnace. 140 STEEL. At this juncture an international committee was appointed from the leading metallurgical societies of the world, and a list of the members shows us a formidable array of well-known names : Hoi- ley, Bell, Wedding, Tunner, Akernian, Egleston and Gruner. This committee reported in October, 1876, to the American In- stitute of Mining Engineers, the following resolution : (1) That all malleable compounds of iron with its ordinary ingredients, which are aggregated from pasty masses, or from piles, or from any forms of iron not in a fluid state, and which will not sensibly harden and temper, and which generally resemble what is called "wrought-iron," shall be called weld iron. (2) That such compounds, when they will from any cause harden and temper, and which resemble what is now called "puddled steel/' shall be called weld steel. (3) That all compounds of iron with its ordinary ingredients which have been cast from a fluid state into malleable masses, and which will not sensibly harden by being quenched in water while at a red heat, shall be called ingot iron. (4) That all such compounds, when they will from any cause so harden, shall be called ingot steel. The Institute, in accordance with its rules, declined to promul- gate any official opinion on the subject, but did recommend that the proposed nomenclature be used in all future papers presented at its meetings. It is fortunate that no more positive action was taken in forcing into use a system which was radically wrong. This classification disregarded a primal necessity of business, for it is necessary to have a name for the material while in process of manufacture. As a practical maker of a certain material used in the arts, I wish a title by which to call it. I cannot give orders to make a heat of - and wait until it is made, rolled, chilled in water, and tested for hardness before it can have a generic name. The word "steel" was in use for this very purpose in every Bessemer and open-hearth plant in the country and when the name was once given at the converter or the furnace, it clung throughout its his- tory in the rolling mills and shops just as the term is used in the steel works of Germany in defiance of the official classification. To-day nothing is heard about this proposed nomenclature, its sole panegyric together with an unwilling eulogy having been writ- ten by Professor Howe. He opens his great work, published four- 142 METALLURGY OF IRON AND STEEL. teen years after the committee had issued its manifesto, by saying this:* "The terms Iron and Steel are employed so ambiguously and inconsistently that it is to-day impossible to arrange all vari- eties under a simple and consistent classification." And he adds, with some triumph in the memory of forensic victories, but more pathos over the record of disappointed hopes, that the result would have been quite different "could the little band, which stoutly op- posed the introduction of the present anomaly and confusion into our nomenclature, have resisted the momentum of an incipient cus- tom as successfully as they silenced the arguments of their oppo- nents." He closes by completely surrendering to the enemy in these words: "So firmly has this (generic) sense of the word become established that, unfortunately, it were vain to oppose it." It is a pity that after this acknowledgment of the final judg- ment of the metallurgical world, he should commend the practice of calling malleable-iron castings by the name of steel, f simply because they coincide with a definition he has just branded as obso- lete, for in so doing he sanctions what is to-day one of the greatest frauds in the business. Steel castings are made by pouring melted steel into a flask. This steel must be made either in a crucible, an open-hearth furnace, or a Bessemer converter, for it is impossible to melt scrap in a cupola and have good steel run from the taphole. It is either ignorance or crime to call by the name of steel castings the hybrid metal made by melting a mixture of pig-iron and steel scrap in a cupola, and it is just as far from truth to apply the term to malleable iron. Any definition of steel which gives room for these mistakes writes its own epitaph as erroneous and absurd. SEC. IVb. Cause of failure of certain proposed definitions. One reason has already been given why the projected renaissance of a decayed nomenclature was a failure, but although the lack of any other general term to denote the product of the converter was a most formidable obstacle, it is easy to believe that this could have been overcome. The whole structure, however, lacked a foundation, because there can be no satisfactory definition as to what constitutes hardening. It will not do to prescribe any test with a file, for there is too much chance for personal equation in such a trial, not to mention the impossibility of having every file of exactly the same hardness. It will not do to make a quench bend, for the success of such a test is determined in too great a * Metallurgy of Steel, p. 1. t Loc. cit. STEEL. measure by certain variable conditions of the preheating not fully understood, and by the manipulations of the smith. All these points were fully understood by practical men at the time the committee was at work, and the arguments were ably presented by Park and Metcalf.* They asked for a definition as to what constituted hardening, and received the answer that a divid- ing line is unnecessary. Prof. Akermanf recommends that it be placed where the quenched piece cannot be scratched by feldspar. He recognizes that small variations in many elements other than carbon will determine the amount of hardening, and also mentions the difference caused by the temperature of the water and the way in which the piece is immersed, and whether it is held still or moved. If the learned professor had wished to condemn his case, he could have done little more. Laboratory experiments on quenching and scratching with feldspar are well enough for some purposes, but when these must be performed before the material can have a name, and when such work gives us simply the name and no other information at all, then, surely, the matter presents itself in the form of a reductio ad absurdum. It is true, as argued by Prof. Howe, that many of the common products of metallurgy and art shade imperceptibly into one an- other ; but it is surely extraordinary when the dividing line cannot be drawn even in theory, much less in practice; when, wherever it falls, it must divide, not intermediate, but finished products, used in enormous quantities, and blending into one another by insensible gradations, and when every shade of these variations is the subject of rigorous engineering specifications. It is customary and necessary in ordering steel to give a certain margin in filling the specifications, and it will be evident, no matter how close this margin is, that if a line could be drawn, it would not infrequently happen that he who ordered ingot iron would receive steel, and he who ordered steel would receive ingot iron. Many different tests have been proposed at various times for determining the mechanical properties of steels, but although some of them are of value in special cases, the one method of investiga- Can the Commercial Nomenclature of Iron 6e Reconciled to the Scientific Terms Used to Distinguish the Different Classes t Metcalf. Trans. A. I. M t'on Hardening Iron and Steel; Its Causes and Effects. Journal I. and 8. ?.. Vol. II, 1879, p. 512. 144 METALLURGY OF IRON AND STEEL. tion which has become well-nigh universal is to break by a tensile stress and measure the ultimate strength, the elastic limit, the elongation, and the reduction of area. Strictly speaking, none of these properties has any direct connection with hardness, and it is also true that in special instances, as with very high carbons, hard- ening may reduce the tensile strength by the creation of abnormal internal strains; but in all ordinary steels, it is certain that hard- ening is accompanied by an increase of strength, by an exaltation of the elastic limit, and a decrease in ductility. Now, if it is conceded that no practical test defining hardening has ever been devised, and if it can be shown that sudden cooling produces a very marked increase in ultimate strength, an exaltation of the elastic limit, and a decrease in ductility even in the softest products of the converter and the open-hearth furnace, then we are partially justified in assuming that hardening has occurred on the ground that the more easily recognized correlated phenomena con- tinue in unbroken order down the scale of the various iron products. The conclusion is weak in logic, I will admit, but from the stand- point of the engineer of to-day, who grades everything by the ten- sile test, and who makes "strong" steel and "hard" steel inter- changeable terms, I claim good ground for my position in calling isteel hardened when it is strengthened. TABLE IV-A. Effect of Quenching on the Physical Properties of Different Soft Steels. NOTE. Bars were 2"x%" flats, rolled from a 6"x6" ingot, and were chilled at a dull yellow heat. Number of test-bar. 1 2 3 4 5 6 Composition^per cent- Carbon, Manganese, Phosphorus, Sulphur, .09 .44 .011 .033 .12 .82 .004 .027 .11 .43 .010 .010 .12 .82 .004 .027 .09 .89 .017 .031 .10 .16 .010 .019 Ultimate strength; pounds per sq. inch, Natural, Quenched, 49390 66080 48960 65670 48960 66300 48260 63640 49760 62280 46250 58380 Elastic limit ; pounds per square inch. Natural, Quenched, 83220 47310 83390 und. 33010 und. 82340 50170 81040 46580 29830 40500 Elastic ratio, per cent. Natural, Quenched, 67.26 71.60 68.20 und. 67.42 und. 67.01 78.83 62.88 74.79 64.50 69.38 Elongation in 8 in. ; ) per cent. Natural, Quenched, 29.75 18.75 81.00 16.25 32.50 15.00 82.50 17.75 81.25 23.75 87.75 27.50 Reduction of area; per cent. Natural, Quenched, 50.80 56.50 52.50 63.27 54.10 63.47 55.75 64.47 49.00 65.15 68.88 68.97 STEEL. 145 The fact that common soft steel is materially strengthened by Chilling has been widely recognized for many years, but the extent of the alteration in physical properties in the softest and purest metals is not generally understood. Table IV-A gives a series of tests that I have made, which may shed some light. on this point. As the bars were rolled from a small test ingot, the elongation is much less than the normal, but the consequences of the quenching are well marked. Additional tests were made on another sample of soft basic open-hearth metal. The original piece was a rolled flat, 4 inches wide and 5-16 inch thick. This was cut lengthwise into two strips 1% inches wide by 5-16 inch thick, and these strips were again cut into 18-inch lengths, so that the whole bar gave 12 test-pieces. Six of these were taken from alternate sides of the original bar throughout its length and tested without treatment, while the other six were broken after chilling at different tempera- tures. The results are given in Table IV-B. TABLE IV-B. Effect of Quenching the Same Steel at Different Temperatures. Bars l%"x T y; Composition, per cent.; C (by combustion) .057; Mn .33; P .006 ; S .019. ^1 id si- Is o Heat treatment. Iff j S |f 1*4 .28 .53 H-& III o a -dS 0> 03 O II & H 3 W H Natural state; average of 6 bars Chilled at a dull red heat 46098 49740 33825 33800 35.37 70.00 70.00 78.37 67.95 " dark cherry red . . " medium cherry red " cherry red " " low bright red ... . " " bright red 56500 51100 57240 58200 62640 38830 84570 89060 89930 88860 Ir III 63.80 70.80 66.10 64.80 68.10 68.73 67.65 68.24 68.61 62.04 There is possibly a mixture of tests in the case of the " dark cherry red" and "medium cherry red/' or perhaps an error in estimating temperature, but I give the results as they were recorded. The elongation is not given, for the pieces persisted in breaking near the grips. This may have arisen from the fact that the ends of the bars as they lay in the muffle were not as hot as the middle, and hence did not receive so severe a chilling, but the difference is not enough to invalidate the nature of the results. The reduction of area is lessened somewhat, but this seems to be affected much hy chilling than the other properties, a fact which is also shown in Table TV-A. 146 METALLURGY OF IRON AND STEEL. The untreated bars show that the metal was of extreme softness, while the chilled specimens prove that each change in the quench- ing temperature is reflected in the physical condition of the chilled bar. SEC. IVc. The American nomenclature of iron products. The classification by hardening is a dead issue in our country. It had quietly passed away unnoticed and unknown before the Committee of the Mining Engineers had met, and the best efforts of that bril- liant galaxy of talent could only pronounce a kindly eulogy. Strictly speaking, some mention must be made of hardening in a complete and perfect definition, for it is possible to make steel in a puddling furnace by taking out the viscous mass before it has been completely decarburized ; but this crude and unusual method is now a relic of the past, and may be entirely neglected in practical discussion. No attempt will be made here to give an ironclad for- mula, but the following statements portray the current usage in our country: (1) By the term wrought-iron is meant the product of the puddle furnace or the sinking fire. (2) By the term steel is meant the product of the cementation process, or the malleable compounds of iron made in the crucible, the converter, or the open-hearth furnace. This nomenclature is not founded on the resolutions of com- mittees or of societies. It is the natural outgrowth of business and of fact, and has been made mandatory by the highest of all statutes the law of common sense. It is the universal system among engineers not only in America, but in England and in France. In other lands the authority of famous names, backed by conservatism and governmental prerogative, has fixed for the pres- ent, in metallurgical literature, a list of terms which I have tried to show is not only deficient, but fundamentally false. The foregoing discussion has taken no cognizance of the micro- scopical structure of steel, because the investigations thus far made in this field of research do not give any limits by which we can form a definition. It is rather indicated, as pointed out in Sec. Ilk, that there is no dividing line between the softest steel and the hardest pig-iron. In Chapter XV will be found further informa- tion on this subject. CHAPTER V. HIGH-CARBON STEEL. SECTION Va. Manufacture of cement and crucible steel By the use of reasonably pure ores and by skillful puddling, it is quite possible to produce wrought-iron in which the phosphorus shall not exceed .02 per cent. This bar of soft pure iron may be con- verted into hard steel by placing it in fine charcoal and exposing it to, a yellow heat. By a slow process, called cementation, the carbon penetrates the metal at the rate of about one-eighth inch every 24 hours, so that a bar five-eighths of an inch thick is satu- rated about 48 hours after it arrives at a proper temperature. This operation is carried on in a large retort where many tons of bars are treated at one time, so that it will always happen that some parts of the furnace arrive at a full heat much sooner than others, and remain longer at that temperature. Consequently, when such a retort is opened, it is necessary to break all the bars and grade them by fracture according to their degree of carburiza- tion. The point of saturation is about 1.50 per cent, of carbon, but the average of the whole will be about one per cent. The steel thus produced is known as blister or cement steel. Its use is limited by the fact that it always contains seams and pits of slag which were present in the wrought-iron, and these defects are of fatal moment in the manufacture of edged tools. To avoid this trouble, cement steel may be melted in crucibles, out of contact with the air, and, being thus freed from the intermingled slag, can be cast into ingots and hammered or rolled into any desired shape. This double process is expensive, and a cheaper and more common method is to put a proper quantity of charcoal into the crucible with crude bar iron, the absorption of carbon progressing with great rapidity when the metal is fluid. This practice is almost universal in America, and it is claimed by men whose word must carry weight in the metallurgical world, that it gives a steel equal in every respect to the older method ; but against this must be put the work of firms whose name is synonymous with most excellent pro- 147 148 METALLURGY OF IRON AXD STEEL. duct and who, at much extra cost, use a certain proportion of cemented bar for the most expensive steels. It is difficult to say how much of the extra quality is due to the method of manufac- ture and how much to the strictest care in working and inspecting, and it is also hard to find out whether the conservatism does not arise from the laudable desire to supply old customers with ex- actly the same metal, in name as well as in fact, that has been furnished them in the past. In deference to time-honored tradition, it may be well to quote without approval or further dispute the following dictum of See- bohm,* which expresses the ancient doctrines: "The best razor steel must be melted from evenly converted steel. It will not do to mix hard and soft steel together, or to melt it from pig let down' with iron, for it will not then possess the requisite amount of body and the edge of the razor will not stand." A third variation is the melting of wrought-iron with a proper proportion of pig to raise the carbon to the desired point, while in still another, used in Sweden, the charge of the crucible consists of pig and iron ore. The aim of all methods is to obtain a malleable metal containing from .60 to 1.40 per cent, carbon, and free from blowholes. For certain purposes some special element like chrom- ium, or tungsten, may be used as an alloy, but with this exception every other ingredient may be regarded as an impurity. SEC. Vb. Chemical reactions in the steel-melting crucible. The best tool steel must be as tough as possible, and, therefore, the phosphorus should not be over .02 per cent. Sulphur, which does not appreciably affect brittleness, but which does decrease forge- ability, is not quite so important, but should not exceed .04 per cent. Manganese may be present in larger quantity, and it is not an uncommon practice to put into the pot a mixture of manganese ore and carbon so that metallic manganese may be reduced and confer better forging qualities. If the percentage does not exceed .20 it has very little bad effect; if much above this, it will cause brittleness and liability to crack in quenching. As in every branch of industry, a simple outline of operations such as is given above may be elaborated indefinitely by the descrip- tion of the variations in practice which have been developed in dif- ferent works. Some such details seem absolutely essential to the * On We WctiKlacture of Crucible Cast-Steel. Journal I. and 8. I., Vol. II, 1884, p. 372. HIGH-CARBON STEEL. 149 originators, but they may be unknown at other equally successful establishments. There is one feature, however, known as "killing," which is in universal use. Just after the steel is melted there is more or less action in the crucible, since there are several rearrangements to be consummated. Thus, in addition to the iron and charcoal in the pot, there is a small amount of glass or similar material to give a passive slag; there is also a little air, some slag and oxide of iron in the puddled bar, the scale and rust on the surface of each piece of metal, and silica, alumina and carbon from the scorification of the walls. A little time is necessary after fusion for the various reactions to occur between these factors and for the attainment of chemical equilibrium. Aside from these general reactions, the spe- cial work of the "killing" epoch is the reduction of silicon from the slag and lining in accordance with the following equation : Si0 2 +2C=Si+2CO. The carbon is drawn either from the charcoal, from the metal, or from the walls of the crucible. In the case of graphite pots the supply from the latter source will be more than ample, while even clay pots furnish quite an amount from the coke which is mixed with the clay in their manufacture. This process of reduction goes on until the steel contains from .20 to .40 per cent, of silicon and the metal lies quiet and "dead." The pot is then taken from the furnace by means of tongs, and the contents are cast into ingot form. The crucible lasts from four to six heats, and the weight of a melt is about 80 pounds when the crucible is new, the subsequent charges being regulated according to the strength of the scorified walls, and by the desire to lower the level of the slag line to the less affected portions. SEC. Vc. Chemical specifications on high steel. In olden times all springs, tools, dies, and the like, were made from either cement or crucible steel, but in late years large quantities of high-carbon metal have been produced in the Bessemer converter and used for many common purposes, although, ordinarily, the steel made by this process contains too much phosphorus to make it suitable for the best work. The manganese in Bessemer steel is much higher than in crucible metal, and this has a tendency to cause cracks in quenching. Formerly a content of .75 to 1.10 per cent, was not 150 METALLURGY OF IRON AND STEEL. uncommon, but the demands of the trade have forced an improve- ment in this respect, and it is now customary to keep the manganese below .80 per cent. ; it is impracticable to have it much below .50 per cent, on account of red-shortness. It is possible to make a -much better selection of the stock for an open-hearth furnace and to produce a steel which is low in man- ganese, phosphorus, and sulphur. The relative merits of open- hearth and crucible steel have been the subject of vigorous discus- sions, but, as in many similar cases, the critics who are loudest in expressing their opinions are the least competent to judge. Often- times a comparison is made between a pure crucible steel and an open-hearth metal containing about .07 per cent, of phosphorus and .60 per cent, of manganese, and on the strength of this comparison, and taking the word of some ignorant or untruthful open-hearth maker as to the quality of his product, the conclusion is formulated that crucible steel is undeniably superior. Such generalizations on insufficient evidence constitute the large majority of those made in our tool shops, but it is evident that no comparison is valid unless the steels are of the same composition, and in this latter respect it will not do to accept the unproven statements even of makers who rank as virtuous. To show that this last clause is not meaningless, Table Y-A gives analyses of three grades of steel, furnished by one of the large and well-known steel manufacturers of the country. The first column shows the name by which the maker billed it. TABLE V-A. Examples of Commercial High Steels which are not in Accord- ance with Specifications. Nature of sample as marked by maker. Composition; percent. C P Mn Si 8 "Crucible" . 1.00 .94 .80 .04 .065 .072 .83 .56 .64 .02 .23 .19 .025 .125 .155 "Pennsylvania Railroad spring" . Low phosphorus spring " Needless to say that the carbon content in these metals is right, for otherwise they would be entirely unsuitable, but each sample shows discrepancies between actual composition and name. Cru- cible steel may and often does contain as much as .04 per cent, of phosphorus, but no purchaser expects to have that amount when hp HIGH-CARBON STEEL. 151 buys the product of the pot, and when this figure is considered in connection with the high manganese, and above all with the absence of silicon, the natural conclusion is that the metal ran from the taphole of an open-hearth furnace. The second sample was sup- posed to fill the Pennsylvania Railroad specifications for springs which at that time called for phosphorus below .05 per cent., man- ganese below .50 per cent., and sulphur below .05 per cent., but a glance will show the liberties that were taken. The "low phos- phorus" spring steel contains .072 per cent, of that element, an amount slightly under the average of common rails, but which can by no stretch of words be called "low" for hard metal. The sulphur is extraordinarily high, but where there are no specifications on this element, there is not much ground for criticism, since it has little influence on the cold properties. SEC. Vd. Manufacture of high steel in an open-hearth furnace. It is perfectly possible to make regularly, in open-hearth fur- naces, a steel of any carbon desired from .05 to 1.50 per cent., with phosphorus below .04 per cent., with manganese below .50 per cent., and with sulphur below .04 per cent. During the last few years this steel has come into use in enor- mous quantities and all the car springs used in the country and almost all similar articles are of open-hearth steel. It is to-day being used very extensively under the name "cast steel," a term which is both a truth and a lie. It is the truth because the steel is cast ; it is a lie because "cast steel" is a trade name dating back a century, and meaning the product of the crucible. There are one or two minor points about this material which should be recognized by maker and user. First; there is not as good an opportunity to get a "dead melt" in the furnace as in the pot, and hence there is more liability of blowholes in the ingots and seams in the bar. For making razors, watch springs and other delicate instruments, no expense is too great in the avoiding of minute defects, but when these imperfections are few and not of such vital importance, there must inevitably be a tendency to economize in the cost of the raw material. Second ; a heavy heat of open-hearth steel must be cast in masses which are very large in comparison with the 4-inch ingot of the crucible works, and the chances for segregation are correspondingly increased, although Table V-B will indicate that with proper pre- cautions there is very little danger of trouble from this cause. 152 METALLURGY OF IRON AND STEEL. TABLE Y-B. Composition of Clippings taken from the Top* and Bottom Blooms of Each Ingot of a High Carbon Open-Hearth Heat, Made by The Pennsylvania Steel Company. Number 1 of Ingot. Part of Ingot. Composition; per cent. Carbon by Com- bustion. P Mn S Si Cu 1 Top 1.009 1.080 .030 .031 .80 .29 .027 .026 .14 .18 .10 .10 Bottom 2 Top 1.046 1.006 .029 .026 .29 .29 .027 .027 .15 .18 .10 .10 Bottom. 8 Top 1.042 0.933 .031 .030 .29 .30 .028 .029 .11 .14 .10 .10 Bottom 4 Top 1.090 1.027 .032 .034 .28 .29 .028 .025 .09 .12 .10 .10 Bottom 5 Top 0.948 1.089 .035 .036 .32 .29 .026 .027 .17 .10 .10 .10 Bottom 6 Top 1.065 1.086 .030 .033 .28 .29 .026 .026 .11 .11 .10 .10 Bottom . . 7 Top 1.073 1.043 .030 .028 .29 .30 .025 .028 .11 .15 .09 .10 Bottom 8 Top . . 0.982 0.953 .029 .032 .30 .29 .025 .026 .12 .13 .10 .08 Bottom 9 Top . 1.044 0.915 .031 .032 .29 .28 .026 .027 .11 .13 .09 .10 Bottom Test. 1.073 .030 .28 .033 .12 .07 Some very interesting experiments were made by Wahlberg, who took tests from the top and bottom of high carbon ingots made at four well known works in Sweden. The variations in the results obtained by different chemists have already been shown in Table I-C and need not be discussed here. The original paper gives full information, from which we find that one analyst found a difference in the carbon content of the outer skin of the ingot at the top and at the bottom amounting in the four different ingots to the following in per cent. : ' .13 .06 .09 .09 The differences at the center of the ingot between top and bottom were respectively .19, .05, .13 and .09 per cent. * The piece from the upper bloom was from a point corresponding to one- quarter way from the top of the ingot, and was therefore near the point of maximum segregation. The sample was the usual clipping produced in cutting a billet under the hammer. HIGH-CARBON STEEL. isa There is one important point which is not discussed in the origi- nal paper. Wahlberg gives in each case the carbon as "branded" on the bar, by which we may assume that the steel would have been sold as having that particular amount of carbon. It may be well to compare this with the results obtained by the chemists, and Table V-C gives this information, the maximum and minimum in each case being obtained by some one chemist from the top and bottom of the same ingot, and it should be stated that in each case I have selected the chemist whose results gave the widest variation. TABLE V-C. Variations in Swedish Steel. Carbon per cent. Brand. Maximum. Minimum. 50 46 49 50 53 61 50 49 55 62 59 69 90 88 106 100 88 105 110 107 119 124 114 131 In the Case of the Steelton steels, concerning which the fullest information is given, the variations in phosphorus, sulphur, man- ganese and copper are trifling, while those of silicon are unimpor- tant. In carbon the difference between extremes is 16 points, and while this may seem to be a great variation in one charge, it will be found that the variations in each separate ingot were less than in the Swedish steel. The average variation between the top and bottom of a Steelton ingot was .07 per cent. It is necessary to consider that a true comparison is not between one small ingot of crucible steel and a heat of open-hearth metal, but between equal amounts of each. In other words, the question must be asked, whether the irregularities are greater in a lot of ten tons of crucible steel than in ten tons of open-hearth. This cannot be satisfactorily answered, since so much depends upon the care with which the stock is selected, but Table V-D gives some analyses of different bars of one lot of crucible steel, sold under one mark and of uniform size by one of the leading firms in the United 154 METALLURGY OF IRON AND STEEL. States; it will be evident that uniformity can by no means be assumed. TABLE Y-D. Variations in Composition between Different Bars of one Lot of Crucible Steel Rounds. No. of Bar Composition, per cent. Carbon by color. P Mn S 1 2 3 4 6 .85 .85 1.05 .98 .90 .013 .011 .010 .018 und. .20 .20 .17 .21 .28 018 014 010 .012 .010 CHAPTER VI. THE ACID BESSEMER PROCESS. SECTION Via. Construction of a Bessemer converter. The acid ^Bessemer process consists in blowing air into liquid pig-iron for the purpose of burning most of the silicon, manganese and carbon 2.14 - FeO Fe,O, Bid, . P,O. . s : 0.36 Probable weight of liquid slag in per cent, of metal . 7 11 27 Quantitative calculation on the Sulphur. Sulphur in lime used, per cent.= 0.054 per cent. T per cent, of slag @ 0.86 per cent. S (see above columns) = per cent. . Less sulphur in lime added = 15.2 per cent, of 0.054 per cent. = per cent. Total sulphur received from metal, per cent. Sulphur removed from metal: 100 parts of initial iron contained, per cent Less 85 parts of blown metal containing 0.080 per cent. S = per cent. . . 0.087 0.008 0.089 0.160 0.068 Total sulphur removed, per cent 0.092 It will be noted that the calculation rests on "the probable weight of liquid slag" for one heat, and this can hardly be considered a final and conclusive proof that volatilization cannot occur, or that it does not often occur, or even that it does not usually occur. In another chapter (see Sec. Xlk) I have tried to show that such loss of sulphur may take place in open-hearth practice, and, if this is true, it seems probable that it will also hold good in the converter. An account by Hartshornef of the practice at Pottstown, Pa,, agrees quite well with the data above given for Horde. The cupola * On the Elimination of Sulphur from Iron. Journal I. and 8. 1, Vol. I, 1893, p. 61. t The Basic Bessemer Steel Plant of the Pottstown Iron Company. Trans. A. L M. E. % Vol. XXI, p, 743. THE BASIC-BESSEMER PROCESS. 183 mixture is of the following composition in per cent.: Si, .0.3 or less ; S, .03 or less ; Mn, 0.80 ; P, 2.50 to 3.00. It will be seen that the specification for the cupola mixture is very rigid, and that the limitations must inevitably result in an increased cost for raw material. Some years ago it was the practice at two different works in Germany to add about two-thirds of the lime at the beginning, so that when the metal was nearly dephosphorized the slag could be decanted, after which the rest of the lime could be put in and the final dephosphorizatio'n effected by a purer slag. The first cinder, which was rich in phosphorus and poor in iron, was fit for agricul- tural purposes, while the second, which was poorer in phosphorus and richer in iron, was used in the blast furnace. This practice has been discontinued and at all works the total quantity of lime is added at the beginning of the blow. The final slag runs as follows in per cent.: Si0 2 , 5 to 6; CaO, 45 to 50; P 2 5 , 16 to 20; FeO, 11 to 13; MnO, 5 to 6; MgO, 5 to 6. In some cases the Si0 2 may be higher, but the P 2 5 is then in a less soluble state and the slag is not so well suited for agricultural purposes. SEC. Vllf. Calorific equation of the basic converter. The calo- rific equation of the basic converter may be calculated by the same method that was used in the work on the acid process (see Table VI-F), but the great quantity of slag and the absorption of heat in its liquefaction render accurate results rather difficult. The silicon is much lower in the pig-iron and consequently the heat derived from this source is less, but the phosphorus more than makes up for the decrease. It was found in the calculation in Section Vlf that the net value of silicon per kg. was 4686 calories; of iron 741 cals.; of carbon 1163 cab., and by the same method we may find that the value of phosphorus is 3821 calories. Assuming an iron with Si=0.5%, P=1.5%,- C=4.09&, and assuming also that 4.0 per cent, of iron is burned to useful purpose, the heat produced per 1000 kilos of iron will be as shown in Table VII-D, the total being about 50 per cent, more than the development in the acid converter. 184 METALLURGY OF IRON AND STEEL. TABLE VII-D. Production of Heat in the Basic-Bessemer Converter. 5 kg. silicon 23,430 calories 35 kg. carbon 40,700 40 kg. iron . . 29,640 15 kg. phosphorus 57,315 Total '. . . .151,085 It is the general practice to use a pig-iron containing 1 or 2 per cent, of manganese, and about 2 per cent, of phosphorus, and such a pig would produce a still hotter blow than the one above given,, but it has been proven in the Westphaliam steel works that when a basic plant is worked up to its full capacity, the phosphorus con- tent can be reduced, just as in a In the case of a furnace which has an insufficient supply of fuel and which contains a full charge of metal, the increased radiation at high temperatures, together with the absorption of energy by the bath, may automatically prevent the attainment of too high a heat ; but in a good furnace, and more especially in an empty one, the action is so rapid that the supply of gas and air must be carefully regulated in order that radiation can maintain an equilibrium. This necessary control of temperature also places a limit on the heat of the regenerators, so that they are usually of a temperatur? of about 1800 F. (say 1000 C.). Dissociation plays no part in the practical operation of a furnace, for, with common producer gas and air, both admitted to the valves at a temperature of about 60 F. (16 C.), the melting chamber may easily be made hot enough to fuse a very pure sand into viscous porcelain. One such specimen of fused material, made under rather unusual conditions, showed the following composition in per cent. : Si0 2 , 98.82 ; A1 2 0~, 0.9;Fe 2 3 , 0.2. SEC. Vlllb. Quality of the gas required in open-hearth fur- naces. The system of regeneration, which supplies the furnace with a fuel already raised to a yellow heat, renders unnecessary any stringent specifications regarding the quality of the gas. Ordinary producer gas contains over 60 per cent, of non-combustible material, and yet is all that can be desired as far as thermal power i 188 METALLURGY OF IRON AND STEEL. cerned. Certain substances, such as sulphurous acid and steam,, are objectionable, but this arises 'rather from their chemical action upon the metal than from any interference with calorific develop- ment. With coal of ordinary quality sulphur causes no trouble, but when it is present in large amounts it is absorbed by the steel. The presence of steam causes increased oxidation of the metal- loids and a greater waste of iron. This oxidation is not always objectionable, since it is sometimes impracticable to obtain sufficient steel scrap, and, if the charge contains an excess of pig-iron, some agent must be used to burn the silicon and carbon. A gas contain- ing hydrogen, like natural gas or petroleum, will be more efficient in this work than a dry carbonic oxide flame, while an excess of steam will make the action still more rapid. Hence it would be possible to use steam in place of ore as an oxidizing agent, but the practice is not to be recommended. If the steam is used during the melting, a considerable proportion of the oxide of iron which is formed will unite with the silica of the hearth and thus become lost beyond recovery. It is advantageous, therefore, to have no free steam present during the melting of the charge, while after the melting is done the oxygen may be supplied in the form of ore with much more satisfactory results. The metal at the time of tapping should be as nearly as possible in the condition of steel in a crucible during the "dead melt," and this can only be attained by a neutral flame. In spite of the opin- ions of many metallurgists, such a -flame cannot be obtained for any length of time, since it has no active calorific power, and even when black smoke is pouring from the stack, the silicon, man- ganese, carbon and iron are absorbing oxygen from the gases. A carbonic oxide flame can be made more nearly neutral than any other, and hence is more desirable at the end of the operation. SEC. VIIIc. Construction of an open-hearth furnace. In the furnace which is exhibited in Fig. VIII-A it will be noted that the hearth sits partly upon the arches of the chambers. These arches, during the entire run of the furnace, are at a bright yellow heat and are continually subjected to strains and deformation by the alter- nating shrinking and expansion of the walls that support them. It is needless to say that a poorer foundation for a furnace would be difficult to conceive, and it is a positive certainty that some tiay there must be a long stop to make what are called "general re- pairs," this term being often used to cover the alterations con- THE OPEN-HEARTH FURNACE. 189 -sequent upon defective installation. Yet this drawing is copied from one of our leading trade papers as the design of a firm of metallurgical engineers, and, unfortunately, it is the common type erected by many such firms, both in this country and abroad, who are guided partly by ignorance and partly by the necessity of sub- mitting plans for the cheapest construction that will work satis- factorily until their responsibility ceases. D Longitudinal Section through Center of Furnace. E, E*, air chambers; F, F v gas chambers; H, gas port; 7, air port; JT, furnace hearth" L, flues to valves ; M, M, binding rods ; O, meeting place of gas and air. FIG. VIII-A. COMMON, BUT BAD TYPE OF AN OPEN-HEARTH FURNACE, It is not easy, however, to say just what the best construction is to avoid these difficulties. H. W. Lash, of Pittsburgh, devised horizontal chambers and thereby the charging floor of the furnace was brought down to the general level, and it was not necessary to elevate the stock, as it could be brought in on trucks without any lioist. There are objections, however, to horizontal chambers, for the tendency of the hot gases is to seek the upper passages and 190 METALLURGY OF IRON AND STEEL. thus the benefit of the full area is not secured. In vertical cham- bers, on the contrary, there is an automatic regulation of the cur- rent ; for, if there is a hot place, the in-going cool gases naturally seek it, and if there is a cool place, the out-going hot gases find it, and thus there is a constant tendency to equalization and to the highest efficiency of a given regenerator content. The worst fea- ture of horizontal chambers is the lack of any propelling action of the gases. . With vertical regenerators the hot gas and air rise nat- urally and force themselves into the furnace, but with horizontal passages there is only a very slight positive pressure due to the slight up-take near the furnace. The fuel will and should always leave the producer under a slight pressure, so that it will need no further assistance on its way to the furnace, but it is advisable to force the air with a fan blower. The amount of room necessary in a regenerator is something on which there is much difference of opinion, but there is no doubt that a very much larger amount is economical than is generally given, the only question being where the limit is, for it is not worth while to spend money for additional chamber area when the saving does not give a fair return on the investment. If the chambers are made large enough, every particle of heat can be intercepted, and the gases will go to the stack at the temperature of the incom- ing gas and the incoming air, but this would be carrying things to an extreme, and financially would not be true economy. It can be stated that the gases should not by any means be at a red heat, although a very large number of furnaces are running with fair fuel economy where the gases during most of the melting operation escape to the stack showing a dull red or a full red temperature. The space occupied by the air and gas checkers combined should be at least 50 cubic feet per ton of steel in the furnace, while to get the best results this figure should be at least doubled. In other words, in a 50-ton furnace the checker bricks in each chamber should occupy at least 2500 cubic feet, which is equivalent to a space 16'xlG'xlO', while if they occupy a space 20'x20'xl2' there will be a further saving in fuel. These dimensions do not include the space below the bricks to give draft area for the gases, nor the space above the bricks to allow the flame to spread over the whole surface of the chamber. In the 40-ton Steelton furnace, shown in Fig. VIII-B, the volume THE OPEN-HEARTH FURNACE. m 192 METALLURGY OF IKON AND STEEL. THE OPEN-HEARTH FURNACE. 193 occupied by the air checkers, as shown in the drawing, is about 45 feet per ton; the gas chamber is of less volume, so that the total is from 65 to 70 feet for both chambers. The double passage, however, allows a better absorption than would be given by the same volume in one mass. In the 50-ton Steelton furnace in Fig. VIII-C the total checker volume on one end is about 100 feet; in the 30-ton Donawitz furnace in Fig. VIII-D about 110 feet; in the 50-ton Duquesne furnace in YIII-E about 55 feet, and in the 50-ton Sharon furnace in Fig. VIII-F about 90 feet. In one open hearth plant I was told that the content was 100 cubic feet, but found that this was on both ends, the gas checkers on each end occupying 17 cubic feet per ton of steel and the air checkers 32 cubic feet. The products of combustion passing to the chimney from this furnace were red hot during a portion of the operation., The information just given is by no means sufficient in stating merely the space occupied by the bricks, for it is fully as important to know the amount of space left between them for the passage of the gases. The area of these channels must be far in excess of the area of the ports or of the flue leading to the chimney, since the friction caused by the small passages will retard the flow of gases, and this retardation will increase continually during the running of the furnace owing to the deposits of dust in these passages, de- creasing the size of the orifices and forming a rough surface for the current to pass over. For this reason the sum of the area of all the passages between the bricks must be several times as great as the size of the flues and ports. It is the area between the bricks which will in great measure determine the life of the checker bricks, for these bricks must be changed when the passages are clogged with dust. On the other hand, the loss of heat will also depend on these areas, for with larger orifices the gases will go down through the checkers and to the stack without giving up their heat to the bricks, so that open-hearth f urnacemen must continually arrive at a compromise between large openings to allow long life to the checkers, and small openings to allow the proper absorption of heat. There is also a third consideration, which is to arrange the bricks in such a way that they present the maximum area of heat absorp- tion with the least interference with the passage of the gases, and with the least opportunity for the deposition of dust on horizontal surfaces. It would be idle to describe any arrangement of check- 194 METALLURGY OF IRON "AND STEEL. ers, as the special conditions made necessary by the shape and size of the chambers in different furnaces determine the way in which the bricks shall be laid. The air chamber should be larger than the gas chamber, because a cubic foot of gas requires somewhat more than a cubic foot of FIG. VIII-C. 50-ToN CAMPBELL BASIC FURNACE, STEELTON, PA. air in order to attain complete combustion and to have a slight excess of oxygen; moreover,, the air enters cold, while the gas is generally rather warm; but in practice the relative values of the gas and air chambers will usually be determined much more by the difficulties of getting room than by any nice calculations on the volumes of gases. It is well, however, to keep the principle in mind that if the gas is very hot there is less work for the gas THE OPEN-HEARTH FURNACE. 195 196 METALLURGY OF IRON AND STEEL. chamber to do, and the fact that under these conditions the gases escaping to the chimney through the gas valve are at a high tem- perature has nothing to do with the case, for if the entering gases are hot the escaping gases must be hotter. Thus, with a given sized chamber, the escaping gases will always be just a certain num- ber of degrees hotter than the gases that go into it. If in a certain furnace this difference is 300, then if the entering gas is 400, the escaping gases will be 700, and if the entering gases are 700, the outgoing gases will be 1000, so that it would be useless to increase the size of the chamber just because the outgoing gases are hot, for these conditions are caused by hot entering gase;?, and the escaping products would be hot no matter how large the chamber might be. Different melters have different ideas as to how a furnace should be run, and it is sometimes better to let them have their own way than it is to change the practice radically to accomplish a very small saving. One melter may oftentimes do better work if the air is extremely hot, while another may prefer that the air be much colder than the gas. These differences also arise from the particular construction of ports so that if an attempt is made to change the relative temperature of the chambers, it might necessitate a complete change in the con- struction of the ports and very likely in the roof of the furnace. Under such circumstances the most practicable thing to do is to run the temperatures of the chambers in accordance with the con- struction of the ports and the roof. These conditions will often- times make considerable difference in the relative amounts of heat delivered to the gas chamber and the air chamber, and, therefore, will determine the relative size of the two chambers, and this may account for the difference of opinion of different melters and dif- ferent furnacemen concerning the proper area for the regenerators. In the Schonwalder construction, introduced abroad, the main point is to have very large flues underneath the checkers, so as to insure free draught in all parts of the chamber, so that the hot gases will go down and the cold gases come up, equally over the entire horizontal cross section. To make more certain, the chamber is divided into two compartments by a vertical wall, and separate flues run from the valve to each. The results seem to indicate that a saving of fuel follows this construction. It very often happens that it is impossible to build a furnace exactly as desired. This was the case in the constructions shown in Figs. YIII-B and THE OPEN-HEARTH FURNACE. 197 198 METALLURGY OF IRON" AND STEEL. THE OPEN-HEARTH FURNACE. 199 200 METALLURGY OF LEON AND STEEL. d P *^ ' CO 5. c4" r-T ? 8 75 iC ^C I-* i^t $ 2'SS'S tf if ^ OJ -*t>. : a 5S O t^ S ' c^" si (N rH- r~ ITS o j. M OCX'tc" ^ s&z %% C3 1C S S 2,?g -^ M rH ?f $1 ^5 It It t~ O5"* ^Se 3 ^SS : co O I* II I i| 1 ; l il j& Si!- s jl|B| ipSisa iEa tfife-fcia 2 . - Si 8|. ll: O Q B S 2 .S g || i li'li 5gS"-S 6C -3.^bDO^ o tg5|2 ^ rife-i'C-S S il|5 - Ill 5 I ?S?I -g So <* ca S Sia^ r3 bC ^^ | C7l= c-d )5 >> I P ^S S ^^t2i^ |g'3w" omfi il 111. : c ;< IllS go g c? "s S l" i| ^ ;*fd S 5rf CHAPTER IX. FUEL. SEC. IXa. The combustion of fuel A full definition of the word "fuel/' and the correlated term "combustion," would necessi- tate a journey into the domain of chemical physics. Such a disser- tation would not be entirely unprofitable, for in the modifications of the Bessemer process the calorific value of silicon, manganese, phosphorus and iron are uf vital importance, but in the affairs of everyday life the term "fuel" embraces only the various forms of carbon known as charcoal and anthracite coal, and combinations of carbon and hydrogen, such as natural gas, petroleum and bituminous coal, while the meaning of "combustion" is also narrowed down to the union of such substances with oxygen. In the case of complex hydrocarbons, like wood, soft coal, or oil, the full history of com- bustion would be a record of manifold dissociations and syntheses; but for practical purposes it may be considered that in all com- pounds of hydrogen and carbon there is an isolation of each element just previous to union with oxygen, and the molecular history may, therefore, be represented by the following simple equations : 1 kilo C+2 2/3 krlos 0=3 2/3 kilos C0 2 , producing 8133 calories. CO+0=C0 2 , 1 kilo CO+4/7 kilo 0=1 4/7 kilos C0 2 , producing 2438 calories. 1 cubic metre CO +1/2 cubic metre 0=1 cubic metre C0 2 , producing 3072 calories. 2 H+0=H 2 0, 1 kilo H+8 kilos 0=9 kilos H 2 0, producing 34,500 calories, including latent heat in steam. 29,040 calories, not including latent heat in steam. 1 cubic metre H+l/2 cubic metre 0=1 cubic metre H 2 0, 238 234 METALLURGY OF IRON AND STEEL. producing 2614 calories, not including latent heat in steam. 1 kilo C+l 1/3 kilos 0=2 1/3 kilos CO, producing 2450 calories. It has been questioned whether this latter action ever takes place, the theory being that carbon always burns first to C0 2 and is then reduced to CO by absorption of incandescent carbon. Whether this is true or not is of little moment, for nothing is gained or lost in calorific energy by the transmutation, and it is, therefore, simpler to assume a direct action. The above equations represent the combustion of carbon and hydrogen with oxygen. Needless to say this never occurs in prac- tice, for it is burned with air, and air is a mixture of oxygen and nitrogen, the proportion by weight being 23.2 oxygen and 76.8 nitrogen, and by volume 20.9 oxygen and 79.1 nitrogen; and it follows, therefore, that the products of combustion from burning coal are composed in great part of nitrogen. The products from burning hard coal and soft coal will vary somewhat, owing to the fact that soft coal contains about 5 per cent, of hydro- TABLE IX-A. Products of Combustion of Hard and Soft Coal. Hard Coal. Soft Coal. Excess Air. CO, CO, O Per Cent. Per Cent. Per Cent. Per Cent. No excess. 21.0 0.0 19.1 0.0 10 19.1 1.9 17.3 2.0 20 17.5 3.5 15.8 3.6 30 16.1 4.8 14.5 4.9 40 15.0 6.0 13.5 6.1 50 14.0 6.9 12.6 7.1 60 13.0 7.8 11.7 8.0 70 12.3 8.6 11.0 8.8 80 11.7 9.3 10.4 9.5 90 11.1 9.9 9.9 10.1 100 10.5 10.5 9.4 10.6 gen, and oxidation of the hydrogen produces water, and in taking a sample of the gases from the stack, this water is condensed as it passes through the tubes of the apparatus and does not appear in the analysis as usually performed, but in order to burn this hydro- gen it is necessary to supply a certain quantity of air and this air FUEL. 235 carries with it a certain amount of nitrogen, and this nitrogen does appear in the products of combustion, so that in burning soft coal the products of combustion contain a slightly higher percentage of nitrogen and a slightly lower percentage of carbonic acid than will be obtained in the burning of hard coal. Table IX-A shows the composition of the products of combus- tion of hard and soft coal when burned with varying amounts of air. The first line gives the results of theoretical combustion when just sufficient air is added to completely burn the carbon and hydro- gen and each succeeding line shows an additional 10 per cent, of air in excess of what is theoretically needed. It is found in prac- tice that such an excess is necessary to insure complete combustion. The amount of excess necessary varies with the conditions under which the coal is burned, but it is seldom possible to have complete combustion with less than 30 per cent, excess air. The percentage of nitrogen is not given, but it is easily found by difference, as whatever is not carbonic acid or oxygen is nitrogen. It will be seen that there is scarcely any difference between the products formed from soft coal and from hard coal, and that the amount of free oxygen present indicates the excess air that is present. The coal always contains a certain amount of ash, but this may be en- tirely neglected in such calculations, for the ash does not escape from the stack and the products of combustion are just the same whether the coal is pure carbon or whether it contains a large quan- tity of earthy matter. Combustion of carbon (coal) with no excess of air: 1 kg. carbon+8.87 cu. metres airr=1.86 cu. m. C0 2 +7.01 cu. m. N" Combustion with 100 per cent, excess: 1 kg. carbon+17.74 cu. m. air=1.86 cu. m. C0 2 -{- 14.02 cu. m. N +1.87 cu. m. 0. The equations given herewith represent the volume of air re- quired by each kg. of carbon and the volume of the products caused by the combustion". In one case the formula represents theoretical combustion and in the other case with 100 per cent, excess air; for any intermediate amount of air the carbonic acid will be the same, and the nitrogen and the oxygen will be pro- portional. This excess air means a considerable loss of heat. 236 METALLURGY OF IKON AND STEEL. There muso necessarily be a loss even if there be no excess of air,, for the products of combustion are so voluminous, owing to the amount of nitrogen present, that they carry off a great deal of sen- sible heat. The amount so carried away will depend upon the temperature of the waste products, but it will not be exactly pro- portional to the temperature, as has already been shown in Table II-F, in Chapter II. Using the figures there given and interpolat- ing for intermediate points, a calculation may be made on the specific heat of the gaseous mixtures shown in Table IX-A and the TABLE IX-B. Loss of Heat in Products of Combustion of Hard Coal in Per Cent, of Total Heat Produced. Temperature of Gases ; Degrees Cent. 100 200 300 400 600 Specific heat of waste gases- No excess air .328 .327 .324 .322 .320 .318 3.8 4.5 5.1 5.8 6.5 7.2 .336 .334 .331 .328 .326 .324 7.5 8.9 10.3 11.7 13.0 14.4 .344 .341 .338 .335 .332 .329 11.3 13.4 15.4 17.5 19.5 21.6 .352 .348 .345 .341 .338 .334 15.5 18.4 21.1 23.9 26.7 29.5 .367 .363 .358 .354 .349 .345 24.0 28.3 32.5 36.8 41.0 45.3 20 per cent excess 40 60 80 100 ..... Per cent, of heat lost No excess air 20 per cent excess 40 60 80 100 loss of heat determined. The results are shown in Table IX-B, from which may be learned that if the gases from a coal fired boiler escape at 200 C., a temperature which is attainable, the loss in sensible heat is 7.5 per cent, when no excess air is present, but if 100 per cent, of excess air is used the loss will be 14.4 per cent. When the temperature is 300 C. the loss with 100 per cent, excess air is 21.6 per cent, and with 400 C. it is 29.5 per cent. The figures in the table for 300 C. and 600 C. were calculated in full,, and it will be noted that they are not exactly proportionate owing to the variations in the specific heats of the gases, but they also show that for moderate temperatures the error will be small if exact proportionality be assumed. In this calculation no account has been taken of the water produced by the combustion of the hydro- gen or the moisture present in the air. These two items will FUEL. 237 increase slightly the loss of heat, but both the moisture in the air and the hydrogen in the coal vary so greatly under different con- ditions that it is hardly worth while to make any average con- cerning them. SEC. IXb. Producers. In almost all metallurgical operations where gas is used for heating, the fuel employed in the producer is a rich bituminous coal ; but in some special cases, as for instance in drying ladles and the like, anthracite coal is sometimes used. For driving ga^ engines hard coal is much to be preferred, as the gas contains very little tarry vapor, and hence needs much less scrub- bing. It is necessary therefore to consider both fuels. (a) Bituminous coal in a gas producer: FIG. IX-A. WATER SEAL PRODUCER. The conversion of soft coal into gas is performed by burning it in a thick fire and collecting the gases evolved. Air is blown in beneath the grate to force combustion, and a jet of steam is also admitted to keep down the temperature and prevent the formation of clinkers. Within the last few years the water seal producer has been very generally adopted. Many different forms have been used, but the main principles of the construction are illustrated in Fig. IX-A, while Fig. IX-B shows a special form. The space below the water level is supposed to be full of ashes, which can be removed without any interference with the operation of the' producer. The ashes will also fill the room for one or two feet above the water line. Above this will be glowing carbon, and the 238 METALLURGY OF IRON AND STEEL. FIG. IX-B. FRAZER TALBOT PRODUCER. FUEL. 239 air as it goes up forms carbonic acid (C0 2 ), and this rising through the bed of coal absorbs more carbon and becomes carbonic oxide (CO), but this action is never complete, and some carbonic acid passes through the fire unchanged. With a hot deep fire free from cavities the gas may contain as low as 2.5 per cent, by volume of C0 2 , but if the fire be thin or if it is riddled with holes, there may be as much as 10 per cent. It is also in the "zone of combustion" that the steam is broken up by the carbon with formation of hydrogen and carbonic oxide, but, as in the similar reduction of carbonic acid, this reaction is never perfect and some steam always goes through unaltered. The best decomposition is attained in a hot fire, but this is just the condition that is not desirable on account of the formation of clinkers. On the other hand, if the supply of steam be increased indefinitely the fire will get colder and colder*, producing no gas and letting steam and air pass through unconsumed. There is a mean between these extremes which is almost forced upon the operator, wherein the fire is kept at a constant temperature, and in this condition there is not much increase in hydrogen from the steam, while, on the contrary, there is quite a little steam passing away with the gases. In the upper zone of the fire, the volatile hydrocarbons of the fuel are distilled by the heat of the combustion beneath, and in this way the gaseous products contain a certain proportion of tarry vapors, some of which are condensed in the conducting tubes. The zones of combustion and distillation are not separated by any arbi- trary line, but a goodly share of the rich components of the coal are carried down into the body of the fire and exposed to a high temperature. This causes the separation of carbon, some of which, staying in the fire, is burned with the coal, while the rest is carried forward into the conducting tube. When the fire is very hot, large volumes of soot are formed in this way and soon give trouble in the pipes, but when cool there is little soot, but much tar. The worst condition is when holes form in the bed of coal. This allows air to come through and burn the hydrocarbons above the fire with a smoky soot-producing flame, cakes the coal into an unworkable mass, and increases the percentage of carbonic acid in the gas. In Sec. Vllli were discussed certain producer experiments, and the gas there given may be taken as fairly representative of ordi- nary practice, the composition being as follows: 240 METALLURGY OF IRON AND STEEL. Per cent. Siemens Gas. by volume. CO 2 5.7 C 2 H 4 0.6 O 0.4 CO 22.0 H 10.5 CH 4 2.6 N, by difference 58.2 100.0 It has been shown that some of these percentages, notably of 2 , H, and CH 4 , will vary through wide ranges according to the condition of the fire, but the content of nitrogen will always be about 60 per cent. This component remains passive throughout all the future history of combustion, but it so reduces the calorific in- tensity that the gas is applicable only to regenerative furnaces. The ordinary methods of gas anal^is fail to take definite account of any save true gaseous components, but in the products of a soft- coal fire there are certain amounts of soot and tar. Some of this material is deposited in the conduits, but this does not constitute a Tery great part of the total energy. I have elsewhere* recorded that in the case of an exposed 7-foot iron pipe, 250 feet long, the condensation of tar amounted to only three-tenths of 1 per cent, of the total heat value, while the gas itself, after passing through the tube, contained a proportion that represented from one-tenth to one-eighth of the total heating power. In spite of the low calorific power of this tar it is found that when the suspended matters are removed by scrubbing, the value of the gas is reduced very seriously, for it is the tar which gives luminosity to the flame and thereby renders it able to heat not only by direct impact, but by the no less potent action of radiation. It is by virtue of this quality that the luminous flames from the dense hydrocarbons so far surpass the clear pro- ducts of an anthracite fire. The investigation given in Sec. VHIi showed that the losses of energy in a producer as operated at Steelton were as follows : Lost as carbon in ash 2.1 Sensible heat of dry gas 13.7 Sensible heat of steam in gas 0.7 Radiation and conduction (by difference) 5.1 Total 21.6 * The Open-Hearth Process. Trans. A. I. M. E., Vol. XXII, p. 376. FUEL. 241 The total shows that over one-fifth of all the heat value of the coal is lost in one way or another. The figure for radiation and conduction is determined by difference, and hence bears all the errors in the determinations. The other items offer some ground for discussion. (1) The carbon in the ash. In Sec. VHIi reference was made to certain experiments in Ger- many by von Jiiptner in which the loss of carbon in the producer ash represented 20 per cent, of the total value of the coal, for he states that the ash coming from the producer contained 74 per cent, of carbon and only 26 per cent, of true ash, this refuse being what would be considered a very fair fuel in some localities. It is hardly right to take such a practice as representative of good methods, as such a waste is entirely unnecessary, for at Steelton it is found quite possible to run soft coal' gas producers where the ash contains less than 20 per cent, of carbon, and, in fact, may average from 12 to 18 per cent. It is possible to estimate very closely how much of the total value is lost if we know the percentage of carbon in the ash and the percentage of ash in the original coal. The lat- ter point must be taken into consideration. For instance, if the coal contains 13 per cent, of ash, and if when working this coal the waste TABLE IX-C. Percentage of the Total Heat Value Eepresented by the Presence of Varying Proportions of Carbon in the Ash. TerCe nt. of Tc Lc tal Heat St. Value Per Cent. Ash. in Coal. 4 7 10 13 20 per cent. C in ashes. . . . 40 . ... 50 60 . ... 1.5 3.0 4.0 5.5 8 2.5 5.5 7.0 10.0 15 3.2 7.0 10.0 14.5 21 4.0 8.5 13.0 20.0 on 15 25 20 material coming from the producer contains 87 per cent, of carbon, it would show that absolutely no work had been done in the pro- ducer and that, therefore, there was 100 per cent, waste, but if the coal contained only 4 per cent, ash and the ashes coming from the producer contained 87 per cent, carbon it would show that only 242 METALLURGY OF IRON AND STEEL. about 30 per cent, of the coal had been wasted. It is therefore of great importance to take the purity of the coal into consideration, and the relative losses with different proportions of ash are not exactly proportional, for they follow different curves when plotted. By calculating different coals I have found the heat value repre- sented by certain percentages of carbon in the ashes and they are given in Table IX-C. It will be seen that with a coal of 7 per cent, ash and with the producer ashes containing less than 20 per cent, of carbon the loss of heat value is less than 2 1/2 per cent, of the original value of the coal, which is a very radical difference from the loss mentioned by von Jiiptner, wherein 20 per cent, of the total value was thrown away from this cause. (2) Sensible heat in gas and steam. The sensible heat of producer gas is wholly wasted, for in a regenerative furnace it makes no difference what the temperature of the entering gas may be, as the temperature of the outgoing products of combustion on the opposite end will be just that much higher, so that the loss on one end balances the gain on the other. In the experiment before mentioned by von Jiiptner, the average temperature of the producer gas in four experiments is 267 C. I am much inclined to doubt the correctness of these temperatures, for I find that von Jiiptner's loss from radiation and conduction alone was as much as all the factors in the Steelton practice com- bined, while the loss from sensible heat of gas and steam was low on account of the low temperature of the escaping gases. It is well known that the loss by radiation is determined by difference, and it is clear that a cold fire should not give as much loss by radiation as a hot one, so that the matter may be straightened out by assuming that von Jiiptner took the temperature of the gases at some distance from the producer and that the item of radiation included a part of the sensible heat of the gas. Under this as- sumption the true radiation from the body of the producer becomes more nearly what would be expected, although a detailed compari- son of the producer calculation is useless owing to the confusing way in which von Jiiptner calculates the heat history of the hydro- gen on the basis of its full calorific value, including the latent heat of condensation. This has already been referred to at length in Sec. VIIIi. It is quite possible that the fires were at a low temperature for FUEL. 243 a short time, but I hardly believe that they could be run con- tinuously under such conditions. I have made experiments on that line and operated a fire for several hours at a black heat, but at the end of that time the whole top of the fire had become a bed of tar, so that it was .impossible to do any poking, and it was neces- sary to stop charging fresh coal and to decrease the amount of steam and to allow the fire to burn up and distill and break up the tarry matters so that the fire could be worked properly. In the experiments at Steelton the gases were at 655 C. and it is quite certain that most producers are run at this temperature. It may appear at first sight that the presence of carbonic acid (C0 2 ) in the gas must be taken as the first and most important loss, but a little reflection will show that this item is taken care of under the head of sensible heat and under radiation; for the pro- duction of an excess of carbonic acid must give rise to heat and this heat must show itself somewhere. If it is used to dissociate steam then it is not lost, for the gas will be enriched by the hydro- gen. Consequently it is not entirely right to assume that a slight increase in carbonic acid necessarily means poorer practice. The gas above quoted as made at Steelton ran as follows : C0 2 =5.7 H=10.5 It is clear that if less steam nad been used the fire would have been hotter and if properly poked would have shown a lower per- centage of C0 2 ; but it would probably also have shown a lower percentage of H, so that nothing would have been gained in the calorific value of the gas, and the heat value of the coal would not have been better conserved. TABLE IX-D. Percentage of the Total Heat Value of the Coal Represented by Varying Amounts of C0 2 in Gas. 2 per cent/CO2= 5.3 per cent, loss 8.0 10.8 " 13.7 " 16.6 " 19.6 " 23.0 " 9 " 26.5 10 " " " 30.0 244 METALLURGY OF IRON AND STEEL. Notwithstanding this theoretical fact that a higher content of carbonic acid is by no means a proof of bad practice, it remains true that under usual conditions the percentage of carbonic acid is an index of the fuel economy, and it is possible to calculate by a rather long process the percentage of heat represented by certain proportions of this gas. Table IX-D shows the percentage of the total heat value of the coal which is represented by certain propor- tions of C0 2 in the gas, provided that the heat produced by its formation is not utilized in the decomposition of steam. In the producer gas previously considered there was 5.7 per cent, of carbonic acid, which, according to this table would represent 15.7 per cent, of the total value of the coal. The calculation of Prof. Kichards in Table VIII-A shows that the formation of C0 2 in the case there under consideration produced 207,300 calories, when the total heat value of the coal was 1,405,000 calories. The carbonic acid in this case represented 14.8 per cent., while Table IX-D would indicate 15.7 per cent, for the same gas. The agree- ment is sufficiently close, since the table does not pretend to be absolute, but is constructed for purposes of comparison only. In ordinary producer practice the carbonic acid runs from 4 to 6 per cent., indicating a loss from this cause of 11 to 16 per cent, of the total heat value of the coal, but under exceptionally good prac- tice the gas will carry between 3 and 4 per cent, of carbonic acid, indicating a loss of from 8 to 11 per cent., thus causing a saving of say 5 per cent, in the amount of coal used. With bad practice the gas may contain 10 per cent, of carbonic acid, indicating a loss of 30 per cent, of the total heat value, which is about 17 per cent, more than is necessary, so that under this practice the amount of coal consumed is one-sixth greater than would be used in good practice. A high percentage of carbonic acid may usually be de- tected without the aid of a chemist, for it is bound to show itself /n a hot fire, and the sensible heat of the gases in the tube is not only the result, but the exponent and measure of the waste. (b) Hard coal: Hard coal is about equal to soft coal when used for firing boil- ers, both in facility of working and in the quantity required, and the smaller sizes are extensively used for this purpose in the east- ern portion of the United States. The smallest sizes are used, as they are not marketable for household purposes and can be had at a less cost. They are, however, more troublesome and require FUEL. 245 special grates and usually forced draft. This material has also been used successfully in producers, the gas consisting almost wholly of carbonic oxide (CO) and nitrogen. In operating such a fire it is necessary to inject steam at the grate or the producer becomes unmanageably hot. The steam rots the clinkers and cools the fire, and hydrogen is produced as in the manufacture of water gas. The gas produced is of about the following composition: Per cent, by volume. CO 27.0 H 12.0 CH 4 +C 2 H 4 1.2 C0 2 2.5 N 57.3 This is nearly the same result that will be obtained in a soft coal producer, but, when the attempt is made to substitute the one for the other, it is found that while gas from anthracite is nearly equal in producing low temperatures, such as firing boilers or drying ladles, it is far inferior, if not entirely valueless, in creating an intense heat, even when properly regenerated; it is supposed with much reason that this inferiority lies in the absence of the suspended volatilized tarry matters, which are characteristic of soft coal gas. These components have quite an appreciable heating value, but their main function is to give luminosity to the flame, and, by increasing its power of radiation, augment enormously its practical value. The absence of these components, however, makes anthracite producer gas particularly well adapted for use in gas engines, as for this work it is necessary to avoid any soot producing components on account of the dangers of premature ignition. SEC. IXc. Miscellaneous fuels. There are some fuels which are essentially local in their character like natural gas and oil; a few remarks will, therefore, suffice for them, and for water gas also, which is used somewhat in metallurgical operations. (a) Natural gas: In the favored district lying just west of the Alleghenies in Pennsylvania, West Virginia, Ohio and Indiana, natural gas has been almost universally used for all kinds of heating from about 1884 until the present time. The composition varies in different wells, but in all cases the gas is made up of members of the paraf- fine series, with not over one-half of 1 per cent, of carbonic acid 246 METALLURGY OF IRON AND STEEL. (C0 2 ) and from 2 to 12 per cent, of nitrogen. By ultimate analysis it gives about 70 per cent, of carbon and 23 per cent, of hydrogen, while, by ordinary methods, it shows from 67 to 93 per cent, of marsh gas, the remainder, when the marsh gas is low, being prin- cipally hydrogen. At first this gas was passed through regenera- tive chambers, but this was discontinued owing to the deposition of soot and to the discovery that sufficient heat was obtained by leading the gas directly to the ports and burning it with air which had been regenerated in the usual manner. Of late years the supply of gas has been decreasing and the demand has been met by the constant drilling of new wells in new territory. There is a limit to this method, and it would seem that before many years this fuel would cease to be a factor in the larger operations of a steel works. ^ (b) Petroleum: Crude oil may be transformed into a vapor by atomizing with steam and then superheating the mixture, but unless exposed for sometime to a yellow heat it remains a vapor, and hence will con- dense if carried through long, uncovered pipes or introduced into the cold valves of a regenerative furnace^ It may be put into the chambers at some point where the temperature is high, and in this way condensation will be prevented as well as the waste heat be utilized. There is also a partial molecular rearrangement with the steam, but the action is far from perfect, for, after passing through 20 feet of small brick flues at a yellow heat, the product may contain 20 per cent, of free aqueous vapor. The mixture of oil vapor and steam may be burned in a muffle, for, after the walls are red hot, there is a reciprocal sustention of heat; but the use in reverberatory furnaces is very wasteful since the action is very sluggish. Even in regenerative practice a charge of cold stock retards combustion much more with oil than with coal gas, and even at maximum temperatures, the flame is longer on ac- count of there being double work to do before the combustion is complete. Each molecule of oil, as it comes into a hot furnace, undergoes a process of dissociation, the rich hydrocarbons break- ing up under the tension of internal molecular activity. This absorbs heat and thus for an instant the action of disruption lowers the temperature below the point of ignition. Moreover, as each point of oil explodes, it makes a small balloon of gas, and it takes a moment for this to become mixed with the air necessary for its FUEL. 247 combustion. If steam is present its reduction by carbon entails a certain delay. These matters may seem trifling, but they are probably the ex- planation of the very important fact that, under the usual condi- tions of furnace operation, a flame from oil vapor is longer than a flame from coal gas. In the burning of clear carbonic oxide, or a mixture of it with nitrogen, there is no preliminary decomposition to be performed, the air being free to immediately touch and burn the molecules of the fuel. It is impossible to state the comparative economy in the use of coal and oil, since their relative values vary so widely in different localities. It often happens that the freight on fuel is three, four, five or perhaps ten times its value at the source of supply, and it will be evident, since oil contains so much more calorific power, that the freight per unit of heat value becomes less and less, com- pared with coal, as the absolute transportation charge increases ; so that if both were to be carried fifty miles, coal might be much the cheaper, while if the distance were a thousand miles, the status would be just the reverse; moreover, the price of oil is constantly varying through very wide limits owing to the discovery of new methods of utilizing what have before been subsidiary or waste products. A rough comparison may always be made by assuming that 50 gallons of oil are equivalent to about 1000 pounds of soft coal when used in regenerative furnaces or under boilers. In the latter case, the success of the practice depends upon the arrange- ments made to prevent chilling of the flame before vigorous com- bustion is in progress. (c) Water gas: NOTE : This discussion on the manufacture and use of water gas Is con- densed from a much longer article by George Lunge, in The Mineral Industry for 1901. When steam is passed over incandescent carbon (preferably in the shape of coke or anthracite) the subjoined reaction takes place : C+H 2 0=CO+H 2 That is to say, equal volumes of carbon monoxide and hydrogen are formed, the mixture possessing the caloric value of 2800 metric heat units per cu. m., an amount one-half the heat value of gas made by distilling bituminous coal in retorts. The heat produced by gram-molecules is for CO+H 2 +0 2 =C0 2 +H 2 0=68.4+57.6 248 METALLURGY OF IRON AND STEEL. =126 heat units, whereas the direct combustion of carbon, C-j-0 2 =C0 2 , produces only 97 heat units. It stands to reason that the introduction of an incombustible substance like water can- not be the source of fresh energy, and the apparent gain of energy represented by the figure: 126 97=29 heat units must be ex- plained by its introduction from an extraneous source. This is found in the heat that accumulates in the incandescent fuel. The reaction: C-f-H 2 0=CO-|-H 2 i s endothermic; i. e., it takes place with expenditure of heat. The splitting up of H 2 requires an expenditure of 57.6 heat units, of which only 28.6 are supplied by the reaction 0+0=00, so that a difference of 29 heat units has to be made good. In the long run these 29 heat units must be supplied apart from the incandescent fuel, the temperature of which constantly sinks and soon falls below the point where the reaction C+H 2 0=CO+H 2 is prevailing (assumed to be above 1000 C.). Below this tempera- ture another reaction comes into play, viz., C+2H 2 0=C0 2 -(-2H 2 , which produces a gas composed of one-third inert carbon dioxide and two-thirds combustible hydrogen. This second reaction is also of endothermic character, and if real water gas is to be made, the operation must be divided into two distinct phases or stages. Be- ginning with a stock of incandescent coal in a generator 2 or 3 m. in height and at a temperature of about 1200 C., steam, prefer- ably in the superheated state, is introduced and water gas is formed according to the reaction, C+H 2 0=CO-fH 2 . Soon, however, the temperature sinks and carbon dioxide C0 a is produced in the gas by the secondary reaction, C-f2H 2 0=C0 2 +2H 2 . Before the carbon dioxide begins to prevail, the steam must be shut off, the temperature being then below 1000 C. This whole period of "steaming" lasts 4 or 5 minutes, and the gas produced during this period is called ff blue gas," containing by volume 48 to 50% H, 40 to 45% CO, 4 to 5% C0 2 , 4 or 5% N", and having a calorific value of about 2600 heat units per cu. m. Immediately after the steam is shut off, the "blowing up" or second stage begins; air is blown into the generator, whereby car- FUEL. 249 bon is burnt and the temperature at once rises. When it has reached the required degree, the air blast is shut off, and the gen- erator is ready for another "steaming." Until quite recently the blowing-up was carried on exactly as in the manufacture of ordi- nary producer gas (Siemens gas), so that the carbon was burnt to monoxide only, thereby generating 29 heat units instead of 97 heat units, which were set free for each atom of carbon; but this was considered unavoidable, as the great bulk of fuel contained in the generator must necessarily reduce any carbon dioxide formed to carbon monoxide, and probably at such high temperatures that from the first carbon monoxide only is formed. This drawback has been overcome by the Dellwik-Fleischer process*, whereby such conditions are established in the generator that during the blows a practically complete combustion to carbon dioxide is obtained within the bed of fuel to be heated, while at the same time condi- tions are maintained favorable to the making of water gas. The radical difference between the "old" processes and the method originated by Dellwik is that in the former the gas, while leaving the generator during the "blow," contains principally carbon mon- oxide, together with the inevitable nitrogen, while in the latter it consists principally of carbon dioxide and nitrogen. Per 1 pound carbon. Dellwik Old way. method. Water gas, cu. ft 21.7 44.7 Heat units 3627 7465 Per cent, utilized 48 92.5 The difference in results is outlined herewith. In the old water gas processes the quantity of gas formed during the blows is amply sufficient to raise the steam needed for the process ; in the new pro- cess the escaping heat is only sufficient to preheat the feed water for the boiler. We must, therefore, add 12 to 15% of fuel for the steam, which reduces the theoretical quantity of gas obtained from 12 Ib. of carbon to 656 cu. ft., and limits the possible utiliza- tion of the heating value of the fuel to about 80%. SEC. IXd. Heating furnaces. (a) Soaking pits: Nothing is more interesting to an American who visits the steel plants of Europe than to find that no coal furnaces are used to heat the ingots between the Bessemer and the rolling mill, but that * Journal I. and 8. I., May, 1900. 250 METALLURGY OF IRON AND STEEL. they are allowed to heat themselves from internal heat in a Gjers soaking pit, a very small amount of coal being often used to main- tain a reducing atmosphere. This old device appears to be per- fectly satisfactory, and it is difficult to understand why it cannot be used in America, but although it has been thoroughly tried in this country, it has been put aside, probably forever. It is one of many things which are declared to be perfectly successful in Eu- rope, but which would not be so called in this country if the results were the same. There is always trouble on Monday morning, and the first round of ingots must be allowed to heat the pits and then withdrawn to be heated elsewhere. The ingots must be put in without delay and must be rolled when ready, or the pit will cool. These things do not fit into American practice, where no one factor must be allowed to retard the mill a moment. It is better to burn a little coal and have ingots always ready to roll without regard to when they were made. It is probable also that failures in this country arose in great measure from the kind of steel. The pits work much better on very soft steel, and as the carbon is raised it is necessary to lengthen the time that the ingot remains in the furnace. No foreign works makes rail steel as high in carbon as we do in America, and more than one foreign engineer will shake his head over the problem of regularly heating steel in this manner when it must often carry from .50 to .65 per cent, of carbon, with higher manganese and phosphorus than is used abroad. Whether these conditions were or were not the cause of the failure and abandonment of these pits in at least four American works many years ago, the fact remains that they were thus abandoned. (b) Regenerative furnaces: It is the universal practice in America and the general practice abroad to use regenerative furnaces for heating ingots or blooms whenever these ingots or blooms are red hot to start with. Under these conditions it requires less fuel in a regenerative furnace than in any other type, and there is no interruption in the output of a furnace. In heating ingots the amount of fuel needed is very small. The furnaces in America are invariably of the vertical type and resemble a Gjers soaking pit, and are operated in much the same manner save that small quantities of gas and air are admitted. At the works of the Maryland Steel Company at Spar- row's Point, Md., only 40 pounds of coal are used per ton of ingots, FUEL. 251 taking the average from week to week. Counting the producer labor this does not cost over 5 cents per ton, which is much better than to have interruptions of work with unfired pits and a lot of cold ingots every Monday morning. In the same way it is customary in America to use regenerative furnaces to reheat blooms coming from the blooming mill before finishing into small shapes. This costs something, but it saves also, as such furnaces serve as a sort of reservoir to receive blooms when the ingots are rolled faster than the finishing mill or to deliver them when the blooming mill is behind. In other words reheating tends toward the uninterrupted operation of the mill, which is the first requisite of economy. It also saves the wear of the rolls and the consumption of steam fuel in the finishing mill. As above stated, for giving a wash heat to hot blooms in this method of work, the regenerative furnace is most convenient and economical. (c) Soft coal in reverberatory furnaces: A reverberatory furnace is one in which the fire is at one end, the stack at the other, and the stock is placed on the hearth be- tween, the flame passing over the top of whatever is placed upon the hearth to be heated. Such a furnace is suitable for, heating cold blooms or billets. A regenerative furnace is not suitable be- cause each charge of cold material lowers the temperature of the regenerators and after about four or five successive charges, it takes longer to heat the blooms than would be required in a coal fired furnace. The operation of a reverberatory furnace is far from satisfactory. When the furnace is filled with cold blooms, the ab- sorption of heat is so great that combustion is retarded and a clear hot flame cannot be obtained. At a later period of the operation, when the blooms are hot, a clear hot flame cannot be carried, as the metal would be oxidized. During the advanced stages, it is necessary to run more or less of a smoky flame and as the blooms on the hearth are of very nearly the same temperature as the flame, it follows that very little heat is utilized in the furnace, but that most of the energy passes out the flue. After the blooms have reached their proper state and during the time that the blooms are being drawn and rolled one at a time, it is evident that all the heat entering the furnace goes out the stack, except what is lost by radiation and conduction. In the ordinary reverberatory furnace the amount of fuel actually used to heat a ton of steel is twenty times as much as theory would call for. 252 METALLURGY OF IRON AND STEEL. One common way of getting more perfect combustion is to in- troduce air at the bridge wall or just over the fire, but oftentimes, this results in more loss than gain, for the average heater will not regulate the amount of air each minute to correspond to the exact amount of smoke that comes from the fire, and if it is not so regu- lated, the flame will often be too sharp and the metal on the hearth will be oxidized. The cost of an increase in the loss of metal of only 1 or 2 per cent, will more than balance the gain in coal, and may even equal the entire cost of the fuel. In more than one European works I have been given very good figures for coal con- sumption, but have been told that the waste by oxidation was from 5 to 7 per cent. In- many localities where fuel is cheap it has been the practice to let the flame from the heating furnace escape directly into a stack, but no argument is needed to show that the hot products of com- bustion should be passed through a boiler. The amount of heat available cannot very well be calculated, but is known by experi- ence. It varies with the condition of the charge, being less after the furnace is filled with cold blooms, and greatest when they are at the full heat. It is quite evident that it is not a good invest- ment to put up a boiler big enough to absorb every particle of waste heat during the short period when the furnace is at its high- est temperature, and it is also evident that the boiler should be more than enough to handle the minimum. The exact point of economy will depend necessarily upon the price of coal, because if fuel is high, a larger boiler will be warranted than when coal is cheap. It is doubtful if any works has erred in spending too much for boilers over the furnace. Nearly all have done the opposite. Steam must be made, and if it is not made by this waste heat, which calls for no expenditure of labor in handling coal or ashes, then it must be supplied by boilers in the fire room. After considerable investigation of this subject, I would give as my opinion the following : (1) For each ton of coal used in twelve hours, the waste heat from the furnace averages from 25 to 30 horse power. (2) When the furnace is at its highest heat, it represents a con- tinuous development of 35 horse power per ton of coal burned in twelve hours. (3) When a furnace is supplied with a boiler capable of absorb- ing one-half of all the heat created at the highest temperature of 0FA*r*ICMT OF CIVIL EMCUMKC*4NCI BEKE_Y, CAU -OftNIA FUEL. 253 the furnace, the average loss throughout the day will be about one- third of the total made, or about one-half of what is utilized, this being due to the fact that this limited capacity is enough at certain periods, and that the boiler makes beyond its rated and economical capacity, as shown by the great loss of heat in the escaping gases. (4) When a furnace is equipped with ample boiler capacity, the horse power developed by each ton of coal put into the fire box will be about one-half as much as would be developed by the same coal if burned under an ordinary stationary boiler. In Table IX-E are given analyses of the waste gases from soft coal reverberatory furnaces after passing through -boilers. In the first column is given the interval from the time when the furnace was charged to the time when the test was taken, and in the second column is given the number of tests that were averaged to give the composition stated. TABLE IX-E. Waste Gases from Reverberatory Furnaces. Interval from charging furn- ace to taking tests. No. of Tests. CO, CO O Less than 20 minutes 17 10 8 4 9 4 2 20 minutes to 1 hour 18 11 9 3 a 2 a 1 hour to 2 hours . . . 6 11 8 7 5 5 2 hours to 3 hours 7 10 6 7 2 j 1 3 hours to 4 hours. . , . . . g 9 8 4 2 5 4 True average 54 11 5 3 Observations were made as to the time when fresh coal was added, but the analyses did not seem to show any relation thereto. Thus there were 14 tests showing over 6 per cent. CO and the average time since coaling for these was 13 minutes. There were 20 tests showing less than 3 per cent. CO, and the average time since coal- ing was 16 minutes. There were 8 tests with over 6 per cent, oxygen, and the average time since coaling was 16 minutes. The results are so nearly uniform for the different periods of the operation that we may take the average as representing the general history, and find the loss of heating power due to the escape of un- turned CO and also the loss of heat by the excess of air or oxygen that is present. In the same way the gases taken at the later periods of the work may be compared. The seven tests taken about two hours and a half after charging show a high percentage 254 METALLURGY OF IRON AND STEEL. of CO and a moderately low percentage of oxygen, while those taken an hour later show a smaller waste of CO, but a large excess of air. It is not necessary to take into account the actual consumption of coal. As a matter of fact this was not taken at the particular time that the tests were made, except in one case, when it ran 490 pounds of coal per ton of blooms heated. Having the composition of the gas it is easy to find the amount of each component in a given volume or weight of gas and to find what proportion of car- bon is burned to C0 2 , what proportion to CO, what oxygen is re- quired and what percentage of excess is present and the loss of heat from each cause: The results are given in Table IX-F, the loss from excess of oxygen being calculated on the assumption that the gases leave the boiler at a temperature of 250 C.=480 F., which is higher than should obtain in good boiler practice, but which is much lower than the average of fairly well-equipped furnace boil- ers. TABLE IX-F. Calculations on Waste Gases from Eeverberatory Furnaces. Kind of Gas. Average. 2 h. 30 m. 3h. 30m. &C ] CO a per cent 11.0 3.0 5.0 21.5 3.6 10.6 7.2 1.1 27.8 0.5 9.8 4.2 5.4 20.8 3.3 J.2 < CO percent Loss from CO per cent Loss from oxygen per cent Total loss per cent 25.1 28.3 24.1 It will be seen that even with gases varying through pretty wide limits the loss due to unburned combustible and to an excess of air is fairly constant. As already explained, the operation cannot be conducted for the benefit of the boiler. The proper heating of the steel is the first consideration and the boiler must take care of itself. Moreover, we cannot expect good combustion to take place after the gases have gone into the boiler, since unburned gases will go through side by side with oxygen, but it does not follow that everything has been done that can be done. There is room for im- provement when over one-fifth of all the power is wasted by non- combustion, but even under ordinarily good arrangements, it is FUEL. 255 possible to run a rolling mill with the power obtained from boilers over the heating furnaces, without any assistance from the fire room. (d) Continuous furnaces: A continuous furnace is a reverberatory furnace, but it is not charged with a whole heat of cold blooms at one time. The blooms or billets are fed in at the flue end, pushed toward the firebox and drawn when they reach the hottest part. The pieces are always hot when they reach the vicinity of the fire, and, therefore, the combustion of the fuel is facilitated, as the flame coming over the bridge wall is never cooled by a lot of freshly charged blooms, as in the intermittent furnace. As the flame goes onward to the flue end, it constantly finds colder and colder blooms and gives up its heat, so that if we conceive a furnace of indefinite length, the escaping gases will be entirely cold and every particle of specific heat utilized, except what is lost by radiation. Notwithstanding these theoretical advantages, there are certain obstacles in the road. The same rules hold good that have been before enunciated, regarding a certain necessary loss of combustible to insure against oxidation of steel, while the loss from unburned carbonic oxide and from excess of air will probably be much the same as shown in the discussion of reverberatory furnaces. One of the difficulties about a continuous furnace is to move the pieces from one end to the other. It is, of course, the natural and almost universal way to put the hearth on an angle, but some power must be applied. In Europe, where such furnaces are very com- mon, it is not unusual to roll the blooms or ingots forward by hand labor only, the pieces being tipped over by means of bars through doors at the side, but the cost of such labor would be prohibitive in America, while this practice gives rise to heavy loss, as the coating of scale falls off at every turn and exposes a fresh sur- face to oxidation. It is impossible to say how much of the heavy oxidation is due to this cause and how much to a sharper flame than is customary in America, but both causes doubtless contribute to the result. In one foreign works rails are buried in the hearth of the furnace, which are replaced when they burn away, and when the furnace is repaired, the ingots being pushed forward by power; in other cases, no rails are used, but the ingots are simply pushed along the sand bottom, which is, of course, much torn by the operation. 256 METALLURGY OF IRON AND STEEL. In America the invariable practice is to have the billets rest on water-cooled pipes. These pipes absorb considerable heat and cool the under side of the bloom somewhat, but the gain in time and labor completely overshadows this small loss. Such furnaces in this country, with few exceptions, are used for billets not over six inches square, since it seems difficult to heat larger blooms suffi- ciently uniformly on the top and bottom, and there is not time when they reach the end of the furnace to turn them over and let the under side get hot. In the exceptions before noted, the blooms are of nearly uniform size and the conditions are favorable, a fur- nace of this type being successfully operated on pieces 8 inches square and 10 feet long. Much time, money and ingenuity are being spent on this problem, and the end is not yet. It is with much hesitation and a, consciousness of rank heresy that I wish to register my doubts as to whether there is any econ- omy to be gained in thus handling heavy blooms and miscellaneous material. The labor of charging a continuous furnace is less than for any other type, but with modern machinery the ordinary fur- nace can be charged and drawn very nearly as cheaply. The con- trol of the temperature, in cases where this is important can be regulated much better in the old way, and the consumption of fuel is not very different, when all factors are considered. The argu- ment has already been made that steam must be produced in some way, and the question is whether the total coal consumed in the furnace and at the fire room is greater in the one case than in the other. I have asked that simple question of two score men in America and Europe, and have not found one who knew from actual investigation. In most cases the old furnaces had never been fitted with proper boilers. In the few cases where the* data were at hand, the only conclusion possible was that no fuel was saved by the continuous furnace. SEC. IXe. Coke ovens. Almost all the coke of America and about three-fourths of that produced in England is made in the old bee-hive ovens, whereby a pile of coal is burned slowly until the volatile matters are expelled, these volatile matters passing away in clouds of smoke. This smoke is a rich gas during the early stages of the operation, and might be used as a source of heat if it were not that such plants are seldom in the neighborhood of industrial establishments. In Belgium and Germany this system was long since discarded FUEL. 267 as wasteful and the coal is burned in retort ovens, by which is meant any construction wherein the coal itself does not burn, but where it is heated in a closed muffle by the combustion of the gases dis- tilled from itself. The gases so distilled may be taken from the tops of the retorts and carried to purifiers, where the tar and am- monia are extracted and sold, in which case they are called by- product ovens. The profits from these by-products vary very much; in some years of high prices they are very attractive; in other years they are nothing. In other cases the gas is taken directly from the upper part of the coal chamber to the combustion passages underneath. By this method the by-products cannot be obtained, but the advantage is gained that the gases reach the flues at a red heat, while in by- product work they are thoroughly cold. Consequently when no by-product work is attempted, less gas is needed to perform the coking and more heat is available for steam raising. It is also possible to use a leaner coal, containing less volatile matter. Thus we might say that if the gas be scrubbed free from tar and thor- oughly cooled, the coal should contain 18 per cent, of volatile mat- ter in order that sufficient calorific value be brought to the flues, while a coal with 15 per cent, of volatile matter would furnish sufficient gas, if this gas were brought red hot into the flues with all the tar in suspension. These figures are not to be accepted literally, as much depends on the nature of the volatile matter. I am informed by W. H. Blauvelt that some Semet Solvay ovens in Belgium are working on coal with only 17 per cent, of volatile matter, with profitable recovery of the by-products. In this coun- try some Pocohontas coal has been worked with 18 per cent, of volatile constituents. In Germany a very considerable proportion of the ovens have no by-product plant attached and some of these are new installa- tions. At many other works the chemical industry is very profit- able. This difference often arises from the great variation be- tween the coal of different seams and mines in the same locality. In general, it may be said that the retort oven without by-products is the most advantageous system where the value recoverable from these products is small, and where the retort system yields a large increased percentage of coke in comparison with the bee-hive, or where the superior density which the narrower retort oven gives to a spongy coke is of advantage. 258 METALLURGY OF IRON AND STEEL. In every prospectus of retort ovens much is said of the great excess of gas which can be reckoned upon as a by-product, but a journey through the coke plants of Europe does not bear out this argument. All the European plants burn their gas under boilers and make no attempt to use it in any other way, and most of this steam so made is used in the chemical plant and the coal washer, the excess for general use not being important in a single instance. It should be stated that in most of the Westphalian ovens the coal is selected so as to get as cheap a mixture as will give good results, and the lower the volatile matter the greater will be the yield of coke, but this reasoning, however, does not apply to the ovens in Silesia, where the percentage of volatile matter is very high, but where the excess gas is of little importance. It has been the rule in America that the surplus gas has been much less than was expected, although the plant at Ensley, Ala., furnishes gas sufficient for the heating furnaces in the rolling mill, the coal containing 32 per cent, of volatile matter. With Pocahontas coal there is no excess. It is possible to get a large amount of gas by a combination of two conditions: (1) A high percentage of volatile matter. (2) A neglect of the character of the coke, with a view of obtain- ing the greatest quantity of gas. It will be evident that the gas expelled from the coal during the first stages of the operation will be very rich and in great volume, but there follows a time when it decreases, but it is necessary to continue the distillation in order to have the coke dense. During this latter period the coal is not self-supporting, in that the gas burned in the flues is more than the gas produced, and the freshly charged ovens nearby must make up the deficit, so that if the coke is to be used as ordinary fuel, as in locomotives, or for any similar purpose, it is well to pay no attention to quality ; but for blast fur- nace work the extra time necessary may use up all the surplus gas. It is possible to keep separate the product made during the early part of the process and use this in supplying cities with illuminat- ing gas, reserving the later product, containing less illuminants, for burning in the flues, the high price commanded by illuminants making this a very attractive proposition. There are many systems in use for building .coke ovens, and it seems to the casual observer that the so-called patents are of little FUEL. 259 validity, but that the main point gained in employing any par- ticular engineer is to get the advantage of his special knowledge. Some of the ovens are regenerative, while many plants have aban- doned this arrangement, the main trouble with a regenerative con- struction being the loss of heat by leakage if the foundations give way, and in most of the plants that have come under my observa- tion, whether regenerative or not, the deformation was very marked. The general principles of coke oven construction have been dis- cussed by W. H. Blauvelt,* and the following is quoted from his paper : "While the principles of operation are the same, there are two distinct types of retort-ovens, viz., the vertical and horizontal flue types. In the former there are some thirty-odd vertical flues in each wall between the ovens. These are connected at the top and bottom by larger horizontal flues, running the length of the oven, the lower one being divided into two parts by a partition midway between the ends. The gas is burned in the lower flue, the flame rising through half the vertical flues and descending through the other half and escaping usually to regenerators of the ordinary reversing type, wiiieh heat the air for the combustion. The course of the gases is reversed about every hour and sent through the flues in the opposite direction. "In the horizontal flue oven the gas is burned in horizontal flues, usually three in number, which are connected at the ends so as to form a continuous system, the gas being admitted through small pipes at the ends of the top and middle flues, where it meets the air for the combustion. The gases travel from above downward, pass under the bottom of the oven, through a simple recuperative arrangement for heating the air, and then to boilers, where steam is made for operating the plant." In Fig. IX-C is given an example of the Semet-Solvay hori- zontal flue type, as erected at Ensley, Ala,, while Fig. IX-D shows the regenerative Otto Hoffman ovens now building at the works of the Maryland Steel Company at Sparrow's Point, Md. Of the total number of coke ovens in the United States in 1899 as given in the Census Eeport only about two per cent were of retort construction, while in Germany there were not 2 per cent, of bee-hives. This difference is due to several causes. One is that * Trans. A. I. M. E., 1898. 260 METALLURGY OF IROX AXD STEEL. the bee-hive oven makes a very superior coke from Connellsville coal, and there is a prejudice or belief that the retort coke will not be as good. Another reason is that the cost of the ovens is SECTION E-E SECTION D-D ^-Jl^ SECTION F-F CROSS SECTION LONGITUDINAL SECTION FIG. IX-C. SEMET-SOLVAY COKE OVEN. FUEL. 261 very much greater, and when the price of coke is low the com- panies have no money to spend, and when there is a boom, bee-hives are put up as being quicker to build and as paying for themselves in a year. X i 262 METALLURGY OF IROX AND STEEL. The prejudice against retort ovens crystallized around investi- gations made many years ago by the leading metallurgists of the Cleveland district in England who advised against the new method. Since then a new light has been seen, and Middlesborough is rap- idly introducing the by-product ovens into many of her works. The advantages of retorts appears very strongly in using a coal poor in volatile matter, for when such coal is coked in bee-hives, a great deal of the fixed carbon must be burned to supply heat, and the yield of coke is small ; but with the closed oven the amount of heat required is less, and a smaller amount of combustible suffices and the only loss in weight is the volatile part. Thus with a rich coal the yield of coke is about the same in the bee-hive and the retort, the latter, however, giving an excess of gas for other uses; while with poor coals the yield of coke is much greater in the retort oven. It is not correct to say that the yield of coke can be accurately estimated from the laboratory tests on fixed carbon, for there is a complicated reaction in the retort oven, and probably also in the bee-hive, whereby the dense hydrocarbons are broken up after they are distilled and deposit carbon in the mass of coal, so that it is possible to produce more coke than there was fixed carbon in the coal. The proportion so made depends upon the molecular arrangement of each particular coal. As indicated above, England has been slow in building retort ovens. They have been used for many years on the lean coals of South Wales, but it is only comparatively recently that they have come into general use in the Cleveland district and around Leeds. Rapid progress has been made within a few years. The total coke production of England is supposed to be from twelve to thirteen million tons, and the retort ovens now erected in the Kingdom have about one-quarter of that capacity. Table IX-G is taken from The Iron and Coal Trades Review, and shows the number of each type in England. In Table XXIII-F, in Chapter XXIII, will be given a list of the coke ovens in each State of the Union, while Table IX-H gives detailed information concerning the retort ovens in operation or construction in 1901. The figures for the Otto Hoffman type are from an article by Dr. Schniewind in The Iron Age, July 18, 1901, while I am indebted to a private communication from W. H. Blauvelt for the data on the Semet-Solvay. FUEL. 263 TABLE IX-G. Coke Ovens in England in 1900. Annual capacity. Tons. Coppee 2296 ovens Semet-Solvav, 450 ovens 500 000 Simon Carves 450 000 Otto 350 000 Collin 120 000 3,220,000 TABLE IX-H. List of Otto-Hoffman and Semet-Solvay Coke Ovens Erected or Projected in United States and Canada at Close of 1901. Owner. Location of Ovens. Number of Ovens. Daily Capacity. Tons. Otto Hoffman- Cambria Steel Co Johnstown, Pa 160 Pittsburgh Gas and Coke Co Glassport Pa. ... 120 New England Gas and Coke Co Everett, Mass 400 Dominion Iron and Steel Co Sydney, Cape Breton. . . . 400 Hamilton Otto Coke Co Hamilton, Ohio 50 Lackawanna Iron and Steel Co Lebanon, Pa 232 Lackawanna Iron and Steel Co Buffalo, N. Y 564 South Jersey Gas E and T Co Camden, N. J 100 Maryland iSteElCo Sparrow's Point Md 200 Total 2226 Semet-Solvay Solvay Process Co Syracuse N. Y 30 150 Dunbar Furnace Co Dunbar, Pa 110 450 National Steel Co . .... Sharon, Pa 25 80 Tennessee Coal I. and R. Co National Tube Co Enslev, Ala Wheeling, W. Va 240 120 1000 500 Solvay Process Co Detroit, Mich 30 150 Total 555 2330 SEC. IXf. Coal washing. There are many deposits of coal which contain a high percentage of ash or of sulphur or of both, and which consequently give a coke of inferior quality, quality can be much improved by washing the coal before it goes to the coke oven, it being possible in this way to materially reduce the proportion of slate and sulphur. A considerable proportion of the slate can be separated from the coal without any difficulty, the extent of the purification depending upon the fineness to which the coal is crushed and the care taken in operating the machines. 264 METALLURGY OF IRON AND STEEL. The sulphur presents greater problems. In some cases it is present in coarse grains of iron pyrites, while in other cases it is in flat, thin films, which float in the water during the process of washing and thus accompany the coal in spite of the difference in specific gravity. Sometimes the sulphur is in the form of an organic compound, and this -tannot be separated by ordinary methods. There are two different kinds of coal washing plants; one de- pending on a combination of sieves and jigs; another where a bumping table is used. These two systems are both good and the underlying principles will be separately considered. A. Sieves and Jigs. If a thousand bullets and a thousand feathers be dropped simultaneously into a wooden box ten feet high, the bullets will be found in a layer at the bottom of tihe box and the feathers will be on top, because the air obstructs very much the fall of the feathers, while its effect upon the bullets will be slight. The action of air in separating unlike substances is used in very few cases, the winnowing of wheat being the most familiar example. In the treatment of minerals, water is the agent used, but the principle is identically the same. It is inconvenient to have a high column of water and so an upward stream is sub- stituted, down through which and against which the particles must fall. Taking for instance a mixture of slate and coal, where the pieces are of uniform size, it will be evident that if a shovelful is thrown into a strong upward current of water, the slate will get to the bottom quicker than the coal, owing to its greater weight. In practice the separation is rendered easier by having a very short column of water with a sieve at the bottom, and the water comes up tlhrough the sieve in pulsations, thus making a quicksand out of the mass and allowing each particle full freedom to find its proper place. When a mixture of slate and coal has been separated in this manner, and while it is being kept in the condition of quick- sand by the continual pulsation of the water from beneath, the two minerals separate into well-defined layers and a stream of mixed coal and slate may be fed into one end of the box, while the slate may be drawn off through one orifice at the other end, and the coal through another orifice. In this description it must be remembered that it is necessary that the pieces of slate and coal shall be of uniform size, for the rate of falling in water depends upon this as well as upon the spe- FUEL. 265 cific gravity. A bullet will fall to the bottom of the ocean quicker than a pebble of equal size, but a rock two feet in diameter will fall faster than fine bird shot. This arises from the fact that both the area of resistance and the area of frictional surface in- crease only as the square of the diameter, while the weight increases as its cube. For this reason, a coal-washing plant of this kind should include a crusher to crush the coal to a certain size; it should then have an arrangement of sieves which will separate the crushed coal into several different lots, each lot being composed of pieces of nearly equal size ; it must then treat each of these lots separately in pulsating water or by some equivalent method and collect the coal and slate separately from each of these lots. There will also be a certain* proportion of very fine material which cannot be handled by any known economical method and which can only be collected in a settling basin. The separation into lots of equal size is called "sizing," and the separation into lots having equal rates of falling in water is called "sorting." In the above description, it is stated that the coal is first sized and then sorted, but it is perfectly possible to first sort and then size. In practice the separation of slate and coal is never complete, because the particles are of very irregular shape, and it will be evident that a flat disc of slate or coal will not follow the same law of falling in water as a more compact body ; that if it happens to fall edge downward it will fall faster than a sphere or cube, while if it remains flatwise, it will fall slower. For this reason, and because there are many pieces that are neither pure coal nor pure slate, but are a mixture of both, the coal will always contain some slate and the slate will always contain some coal. The greater the number of different "sizes" made, the more perfect the separa- tion, but each "-size" involves complication of the plant and in- creased cost of maintenance. In a very complete and economical coal-washing plant in Western Germany the coal in its natural state carries from 22 to 30 per cent, of ash. It is crushed and separated into six sizes on wet sieves. After the washing is com- pleted the ash in the coal runs about 10 per cent., giving a coke containing from 12 to 14 per cent. The fine dust that is collected in settling basins can be used under boilers. The loss of coal in the slate is not over 3 per cen per cent. 50.27 51.96 52.43 52.94 51.90 500 49.27 51.10 55.82 51.72 51.98 1000 52.77 60.30 55.73 52.28 62.77 1500 50.97 51.48 55.66 52.90 62.75 None. MnO, per cent. 14.91 21.65 15.61 21.84 18.50 500 15.20 19.09 15.31 20.44 17.51 1000 14.70 17.50 13.89 19.06 16.29 1500 14.22 16.72 12.40 16.36 14.92 None. FeO, per cent. 31.23 22.59 27.14 23.18 26.03 500 80.68 26.12 25.11 24.21 26.53 1000 26.96 28.26 26.20 26.26 26.92 1500 31.70 26.03 26.96 29.13 28.45 None. FeO and MnO, per cent. 46.14 44.24 42.75 45.02 44.54 500 M 45.88 45.21 40.42 44.65 44.04 1000 41.66 45.76 40.09 45.32 43.21 1500 u 45.92 42.75 39.86 45.49 43.88 COMPOSITION OF THE METAL. Heat No. Silicon, per cent. Manganese, per cent. After adding ore, as below. After adding ore, as below. None. 500 Ibs. 1000 Ibs. 1500 Ibs. None. 500 Ibs. 1000 Ibs. 1500 Ibs. 7596 7598 7606 7635 .07 .04 .04 .13 .01 undet. .05 .07 .01 undet. .03 .05 .01 .01 .02 .06 .10 .02 .08 .19 .02 .02 .05 .08 .02 .02 .03 .09 .02 .02 trace. .10 Samples were taken of metal and slag after every 5CO pounds of ore. These groups and heats were not selected to show this special action, the investigation being made for other purposes; but the wonderful regularity in results, corroborated as it is by many other records, shows that in the magnificent alembic of the melting furnace, at the highest heat we know save that of the elec- tric arc, at a temperature when wrought-iron melts like wax in the candle flame, the molecular relations are guided by fixed and unal- terable laws. It is this stability of conditions that gives to the open-hearth melter the ground on which he can work out regular and reliable results, and which makes the process peculiarly fitted for the manufacture of -structural material. SEC. Xf. Pig-and-ore process. The amount of ore required for a charge will not follow closely the amount of carbon, since the flame is constantly at work, and ore is added when the melter thinks it advisable rather than when it is absolutejy necessary. If 276 METALLURGY OF IRON AND STEEL. the charge is hot it dissolves the ore rapidly and there is little chance for the flame to do its share of oxidation, while if the charge is cold only a small amount of ore 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, but it may be broadly said that if the bath contains 1 per cent, of carbon, 1500 pounds of ore may be used in bringing it down to .05 per cent. The first 500 pounds will reduce it to about .80 per cent, of carbon, the second to .40 per cent, and the third will finish the work. If silicon and manganese should be as low during the interval between the firs* and second ore additions as at a later time, the burning of the car- bon might be the same then as later, but either the presence of these protectors or the less favorable physical condition of the slag in a high-carbon bath retards 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 and even thousands of pounds of ore may be added after melting before the carbon is affected. Therefore, when it is necessary to charge nothing but pig-iron, it is advisable to have it contain as little silicon as possible, and even then the oxidation of carbon requires several hours. The ore is not lost, for the reduced iron makes up for the metalloids which are burned, so that the weight of the steel may equal or exceed the weight of the pig-iron charged. The expense of the pig-and-ore process rests in the slow combus- tion of carbon, for it is impossible to hurry the work without caus- ing violent boiling of the voluminous slag, producing scorification of the hearth and possibly a loss of metal through the doors. The process upon an acid hearth is much slower than on a basic bottom, for in the latter case a slag rich in iron does not have such disas- trous results upon the hearth. Since the fuel consumption per hour is nearly the same during the period of oreing as it is during the period of melting, it is plain that there is a considerable de- crease in product with an increased fuel ratio. By the use of a tilt- ing furnace this difficulty may be lessened, for as soon as the silicon has been oxidized, the contents of the furnace may be emptied into the ladle and then the metal be immediately returned to the furnace with as little slag as desired. ' When most of the slag is thus re- moved, the action is much more rapid and there is no trouble from frothing. The tapping of slag from stationary hearths has always THE ACID OPEN-HEARTH PROCESS. 277 resulted unsatisfactorily, and the same is true of attempts to remove it from a tilting furnace by surface decantation, but this process of repouring requires no handling save the raising of the ladle in a vertical line so as to allow the metal to be returned to the furnace through the same hole from which it has just been tapped, and seems to solve the question of slag removal in a simple way. SEC. Xg. Conditions modifying the character of the product. If the temperature of the metal is very high, the last traces of sili- con will not be oxidized, for the affinity of silicon for oxygen is a function of the temperature. In the Bessemer converter the metal may contain as much as 1 per cent, of silicon if blown sufficiently hot, but in the open-hearth there is no chance for the bath to arrive at an intense degree of heat as long as a considerable percentage of this element is present ; for superheating is not readily attained without a lively bath, and the bath will very seldom be lively as long as it holds a high content of silicon. Thus the open-hearth 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 silicious slag, there will even be a reduction of the silica of the hearth according to the equation SiO a +2C=Si+2CO. This variation in affinity of Si for plays an important part in the production of steel castings where a higher temperature is used than for ingots of ordinary size. The constant presence of a small proportion of silicon, due to the high temperature, tends to prevent the absorption of gases, and it is stated by Odelstjerna* that if at any time the metal is allowed to cool so that the last traces of silicon are burned, the gases which are absorbed cannot be expelled by a subsequent superheating. I am of the opinion that Odelstjerna is correct in his statements, but that there may be other factors involved in a full explanation. It is certain that in the manufacture of small ingots which are to be rolled directly into plates, there are delicate adjustments of temperature and slag that are not easily explained by considering the history of silicon alone. One of these factors, which may be cited by way of illustration, * Trans. A. I. M. E., Vol. XXIV, p. 308. 278 METALLURGY OF IRON AND STEEL. is the extent and force of the oxidizing influence. It is the opinion of some metallurgists that the best quality of open-hearth steel can only be made when the burning of the metalloids is. carried on at a very slow rate, so that the bath shall not contain an excess of oxygen at any time, and it is stated by Ehrenwerth* that a certain American works makes a practice of keeping a charge in the fur- nace a very long time when a very good quality of steel is desired. As a matter of fact, the works in question did carry out such a system at one time out of respect to foreign tradition, but found no advantage in so doing, and has long since discontinued the practice. It is also an opinion, held by men of acknowledged reputation, that a high proportion of pig-iron in the original charge will give a superior product. If this is true, it probably arises from the fact that the presence of a high proportion of carbon after melting, with the consequent long exposure to the flame, will result in a thorough washing of the bath. I believe that there is a limit to this action, and that very little can be gained by raising the content of carbon in the melted bath above 1 per cent., for this proportion insures a vigorous boil. ' It is difficult to see how the condition of the bath, after it has been run down from 1 per cent, of carbon to three-tenths of 1 per cent., can be any different from the condition which would have existed if the original content had been 2 per cent. It would seem probable that one or two hours of exposure of the completely liquid bath to the flame would give ample opportunity for any reactions which could be in progress, and the old adage that "enough is as good as a feast" might be applied to the present case. It is not unprofitable, however, to consider the conclusions from practical experience, however invalid they may appear, for they may some- times represent a vital truth, albeit our point of view may not be high enough to allow a complete survey of the field. SEC. Xh. History of sulphur and phosphorus. In the above records no account is taken of sulphur or phosphorus, but numer- ous determinations and universal experience prove that the content of phosphorus in the steel will be determined by the initial content in the charge. It is true that acid open-hearth slag may contain some phosphorus, and I have found one case where it held 0.04 * Das Berg- vnd Hiittenwesen auf der Weltausstellung in Chicago. Ehren- warth, 1895, p. 276. THE ACID OPEN-HEARTH PROCESS. 27i* per cent., but it would require a higher percentage than this to make a difference in the metal that could be detected by ordinary analysis, so that for practical purposes it must be assumed that every molecule of phosphorus that is present in the pig-iron, scrap and" ore will appear in the finished metal. The percentage of sulphur cannot be predicted with so much precision. Traces of this element may be burned during melting and pass away as sulphurous anhydride, but the proportion thus eliminated is small. On the other hand, there is a tendency to absorb sulphur from the flame. With fairly good coal this incre- ment may be neglected, but with, bad coal, and especially when the slow working of the furnace renders it necessary to expose the charge to the gases for a long time, the amount thus absorbed may be ruinous. It has been suggested with some reason that the addi- tion of lime in the producer might retain at least a part of the sul- phur in the ashes of the producer, so that it would not appear in the gas, but it would also give trouble by making a fusible ash. The ore is another source of contamination, for it generally con- tains a certain proportion of pyrites. As the ore floats on the sur- ' face of the bath some sulphur may be oxidized above the surface and the products pass away with the flame, but the remainder will be absorbed by the bath. * SEC. Xi. Method of taking tests. The condition and nature of the metal and slag are determined from time to time by taking samples from the furnace by means of a small ladle and casting test-ingots with a cross-section about one inch square. These are chilled in water and broken, and the carbon is estimated from the appearance of the fracture. The reliability of such a determina- tion depends upon the constancy of the conditions of casting and chilling, and the expertness of the judge, but, roughly speaking, the content can be ascertained within 10 per cent, of the true amount. SEC. Xj. Recarburization. When the desired point has been reached the recarburizer is added, this being almost invariably used in a solid state. It is generally heated red hot, but. this is not essential, for, in making structural steel, "ferro" containing 80 per cent, of manganese is used almost exclusively, and the weight of the addition is so small that it chills the bath only slightly. The ferro may be added to the metal while in the furnace, and this method has the advantage that the bath can be thoroughly stirred 280 METALLURGY OF IRON AND STEEL. after the recarburizer has melted, but it has the disadvantage that during the time the last pieces are fusing, the portions which melted first are losing their manganese to the oxygen of the slag and flame. In a hot furnace this action is very rapid, and although the entire addition may melt in less than a minute, a considerable proportion of manganese is lost by oxidation. When the recarburizer is added in the ladle, it is evident that the latter action will not occur, but there will be a certain loss on ac- count of the oxide of iron contained in the metal, and the function of the recarburizer is to remove this oxygen. The loss of manga- nese will be the same whether the addition is made in the furnace or in the ladle, but in the latter case the effects of slag and flame are absent. Hence it follows, all other things being equal, that the loss will be more regular when recarburization is performed in the ladle, and this is equivalent to saying that the content of man- ganese in the steel can be made more nearly alike throughout a series of heats. The amount of manganese lost in recarburization not only varies with the way in which it is added, but also with the percent- age of carbon and manganese in-the bath. As would naturally be supposed, the amount of oxide in the bath is less with high than with low carbons, and so therefore the loss of manganese in recarburizing decreases as higher steel is made. It is found that the loss is less with smaller percentages of manganese, so that with the same bath, if 1.00 per cent, of Mn be added, there will be .60 per cent, in the metal, being a loss of .40 per cent., while if .50 per cent, be added the steel will have .40 per cent., being a loss of only .20 per cent. It seems as if with the lower manganese the action was not perfect, and that with each successive increment of ferro an additional atom of oxygen is removed. This fact holds good whether the recarburizer is added in the furnace or in the ladle. The fear of non-homogeneity under the practice of adding the ferro in the ladle is not entirely unfounded when small heats are made and the metal is not very hot, but when charges of 20 to 50 tons of hot steel are properly poured and recarburized,'the steel is thoroughly uniform. When metal is made very high in manganese, certain precautions must be taken, but in ordinary structural steels, when the manganese runs below .65 per cent., there is an all-per- vading action throughout the melted mass which dispels all thought of non-homogeneity. THE ACID OPEN-HEARTH PROCESS. 281 SEC. Xk. Advantages of the open-hearth process in securing homogeneity. In the low steel of the Bessemer process there is very little trouble from irregular distribution, although the more viscous slag sometimes holds pieces of the solid recarburizer and keeps them from melting until the steel is nearly all poured. The result is that when they do finally fuse, small streams of high man- ganese metal flow down into the upper part of the last ingot and form a hard spot in the steel. This does not and should not often happen, and most Bessemer soft steel is uniform throughout. In making high-carbon steel, however, the conditions of manufacture make. the hearth far superior to the converter. The metal in the Bessemer process is always blown until nearly all the carbon is eliminated, since it has been found impracticable to stop the opera- tion at any definite intermediate point. All the carbon content of the steel, therefore, must be added in the recarburizer, and abso- lutely perfect homogeneity can only be secured by absolutely per- fect mixing. In the open-hearth, on the other hand, high-carbon steels are made by interrupting the process at the desired stage, and it is plain that no mixing is required as far as carbon is con- cerned, and about the same quantity of recarburizer will be used for a given manganese whether high or low steel is being made. CHAPTER XL THE BASIC OPEN-HEARTH PROCESS. SECTION XIa. Construction of a basic open-hearth bottom. The basic process, as herein discussed, consists in treating a charge of either melted or solid pig-iron, or a mixture of pig-iron and low- carbon metal, upon a hearth of dolomite, lime, magnesite, or other basic or passive 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, the operation being so conducted that the product is cast in a fluid state. In the above specification that the slag shall be stable, no recog- nition is accorded that hybrid practice wherein a little lime is thrown into an acid furnace, near the end of the operation, with the intention of removing a part of the phosphorus by the temporary and uncertain action of a partially basic slag. Regular metal- lurgical results can only be obtained under regular conditions, and to this end the hearth should be made of material that will not be scorified by basic additions. The current belief that the lining of the bottom is the dephosphorizing agent is a complete mistake, for the highest function of the hearth is to remain unaffected and allow the components of the charge to work out their own destiny. In practice it is never possible to construct either an acid or a basic bottom so that it is entirely passive, for a slag which is viscous with silica will slowly attack a pure sand bottom, and a cinder which is mucilaginous with lime will gradually eat into a dolomite hearth. For the construction of a permanent bottom, carbon, bauxite, lime, chromite, magnesite and dolomite, have been used. Mag- nesite gives the best results but it is very costly, and well-burned dolomitic limestone answers well enough. In some places the stone is used in its n'atural state, but this is a doubtful economy, the bet- ter plan being to thoroughly roast it in a kiln or cupola and then 282 THE BASIC OPEN-HEARTH PROCESS. 283 grind and mix with tar. The roof and walls being made of silica bricks, it is necessary to have a joint of chromite or other passive material between the acid and the basic work ; but it must be under- stood that at the intense heat of a melting furnace, and in an at- mosphere charged with spray of iron oxide, almost any two sub- stances will unite if pressed together, so that the par$ of the joint which bears the weight of the superposed brickwork must be shielded from the direct action of the flame. SEC. Xlb. Functions of the basic additions. Given a hearth capable of resisting the action of metal and slag, the problem of the basic furnace is the melting and decarburization of iron as in acid practice, with the additional duty of removing a reasonable quantity of phosphorus and some sulphur. Under the oxidizing influence of the flame and ore, the phosphorus is converted into phosphoric acid (P 2 5 ) which can unite with iron oxide, but the conjunction will be only temporary, for the carbon of the bath reduces the iron, leaving the acid helpless, and then the phosphorus in its turn is robbed of its oxygen and returned to the bath. But if lime is added, the acid can form phosphate of calcium, and since the oxide of this element cannot be reduced by the carbonic oxide, the phosphorus is never left without a partner, but forms part of a stable cinder. This oxide of calcium is sometimes added in the form of com- mon limestone, the carbonic acid being expelled in the furnace. It will be evident that this entails a considerable absorption of heat, and the melting must be delayed accordingly, but it has a com- pensating advantage in that the gas, in bubbling through the metal, keeps up a motion which facilitates chemical action, and also that the carbonic acid gives up part of its oxygen to the silicon, phos- phorus, carbon and iron. This oxidizing action allows the use of a greater proportion of pig-iron, and also aids in the removal of phosphorus, so that there seems to be good ground for using the cheap natural stone. I be- lieve, however, that it is more economical to put it through a pre- liminary roasting and reduce by nearly 50 per cent, the amount of basic addition, for the rate of melting is thereby hastened, while the oxidizing -effect can be obtained by the use of ore. It is true that ore costs more than stone, but, on the other hand, its full value is returned in metallic iron, and, moreover, it is possible to use a greater proportion of pig-iron on account of the reduced 284 METALLURGY OF IRON AND STEEL. quantity of gas evolved, for the amount of oxidation done during melting, either by stone or ore, is limited by 'the frothing of the stock, and this is evidently determined by the amount of gas evolved in the reactions. Therefore, if ore produces less gas than stone in oxidizing a given quantity of carbon, then more pig can be used with ore than with stone. The reactions are as follows : Limestone, CaCO 3 +C=2 CO+CaO. Ore, Fe 2 O 3 +3 C=3 CO+2 Fe. Thus two volumes of gas are formed for each atom of carbon when stone is used, while only one volume is produced with ore. The available oxygen in the. ore is nearly twice as much as in the same weight of stone, so that by using a mixture of 500 pounds of burned lime and 500 pounds of ore, there will be the same quan- tity of basic earth, and the same oxidizing effect, as with 1000 pounds of raw stone, while there will be only half as much gas pro- duced with a contribution of 300 pounds of metallic iron. SEC. XIc. Use of ore mixed with the initial charge. The ore and lime are put into the furnace with the pig and scrap, so that the hearth will be protected during the melting and an active cinder be at work continuously. When high-phosphorus stock is used, the amount of oxidation to be done for a given, weight of pig-iron is much greater than in acid practice. Thus in 10,000 pounds of low-phosphorus iron for an acid open-hearth, the oxygen absorbing power is as follows : 1.0 per cent. silicon=100 pounds Si, absorbing 114.3 pounds oxygen. 3.5 per cent. carbon=350 pounds C, absorbing 466.7 pounds oxygen. Total oxygen absorption, 581.0 pounds If pig-iron be used in basic work with the same content of silicon and carbon, but with the addition of 1.00 per cent, of phosphorus, there will be an additional absorptive power of 129 pounds of oxygen or a total of 710 pounds. If the first mixture were put into a furnace there would be about 40 per cent, of the work done during the melting (under the conditions shown in the pre- ceding chapter), so after melting there would remain 60 per cent, of 581, or 349 pounds of oxygen to be given to the bath. In the second case, it is evident that the presence of phosphorus will not cause a greater action during melting, but that if all other con- ditions are similar, the total absorption will be the same, so that, after melting, the phosphoric bath will have an ^absorptive power THE BASIC OPEN-HEARTH PROCESS. 285 of 349+129=478 pounds of oxygen, and there will be one-third more work to do during the period of oreing with the same pro- portion of pig. These figures may seem somewhat abstruse, but they explain the very important fact that there is much more oxidation to do with phosphoric iron than with good stock, so that it is advisable to use ore mixed with the charge to perform a part of the work dur- ing fusion. On an acid hearth, when running exclusively on pig- iron, ore is sometimes added with the original charge, but there is always danger of this oxide uniting with the sand of the hearth before the metalloids can reduce it. In basic practice, on the con- trary, the ore can do no harm, for it has little effect on the dolo- mite and soon reacts upon the silicon, phosphorus, and carbon. TABLE XI-A. Average Composition of Slag and Metal from Seventeen Basic Heats. Test. Metal. Slag. Composition, per cent. Composition, per cent. C. Si. Mn. P. SiO 3 . MnO. CaO. MgO. FeO. P,0. A B C D .71 .34 .12 .10 .06 .01 .01 .01 .33 .25 .22 .49 .046 .022 .013 .018 19.21 16.37 15.08 15.75 11.12 10.36 9.01 14.11 42.16 42.78 42.16 89.05 6.64 7.87 8.45 10.40 13.68 16.29 20.34 16.65 5.149 4.848 3.850 2.961 SEC. Xld. Chemical history of basic open-hearth charges when no ore is mixed with the stock. The addition of 'ore is not neces- sary when sufficient scrap is available, for the flame will supply oxygen to the metalloids, as will be shown by Table XI-A, which gives the average history of 17 heats when no ore was used with the original charge, and when tests of metal and slag were taken at four different epochs. The heat? were all similar in character and -were operated under similar conditions, and therefore the mixing of slags and metals to obtain average results is justifiable. Each charge was made up of about one-half pig-iron and one-half steel scrap, and contained 2.00 per cent, carbon, 0.40 per cent, silicon, 0.85 per cent, manganese, and 0.20 per cent, phosphorus. Tests of slag and metal were taken as follows : (A) After complete fusion of metal without ore. (B) At beginning of boil, after the addition of 1965 pounds of ore per heat. 286 METALLURGY OF IRON AND STEEL. (C) When the bath was ready for the recarburizer, 775 pounds of ore being added per heat between tests B and C. (D) After casting. SEC. Xle. Elimination of phosphorus during melting. The elimination of phosphorus during melting is a variable, depending upon the conditions of oxidation and the ability of the slag to absorb the phosphoric acid. Table XI-B will show in a general way the proportions of carbon and phosphorus that are oxidized during melting under different kinds of practice. TABLE XI-B. Elimination of Phosphorus and Carbon During Melting Upon a Basic Hearth. Pounds of ore charged with stock, per ton of metal. Number of heats in group. Composition of metal, per cent. Composition of slag after melting; per cent. Phosphorus. Carbon. Initial. | P *J dJn-d sas F m Initial. After melting. t^ SSS fl SiO,. FeO. none, none, none, none. 800 115 140 17 4 9 9 8 6 7 0.20 1.36 0.19 0.19 2.50 0.55 0.55 .046 .594 .023 .072 .744 .274 .402 77 57 88 62 70 50 27 2.00 1.50 1.80 1.80 8.50 2.90 2.90 .71 .60 .27 .78 .59 1.00 1.48 65 60 85 67 83 66 49 19.21 14.90 15.55 19.98 11.96 80.73 34.22 13.68 und. 19.68 12.20 8.61 10.71 10.95 SEC. Xlf. Composition of the slag after melting. Neither the percentage nor the total amount of elimination during melting is a matter of vital importance, for whatever work is left undone during that period will be completed before tapping. In this removal of phosphorus after fusion, the composition of the slag is the impor- tant factor, and this will depend, primarily, upon the amount of silica, and, secondly, upon the lime added. The total supply of silica will determine the quantity of lime, and it will also deter- mine the weight of the resultant cinder. Thus, if the final slag is to contain 16.67 per cent, of Si0 2 and 50 per cent. CaO, it is evi- dent that the basic additions must contain ff=three times as much available CaO as there is Si0 2 in the entire charge, and also that the final slag will weigh six times as much. The composition of the cinder differs considerably, for when good stock is used it may contain over 20 per cent, of silica and still be capable of eliminating the impurities, but when much phos- phorus is to be removed, the silica must sometimes be as low as 12 THE BASIC OPEN-HEARTH PROCESS. 287 per cent., the proportion of CaO usually varying inversely with the silica. The amount of lime which can be taken up is limited, for at a certain point the slag becomes viscous, particularly when the scorification of the hearth supplies magnesia. Allowing for about 10 per cent, of MnO, 8 per cent. MgO, 18 per cent. FeO, and 4 per cent. A1 2 3 , etc., it may be roughly stated that with 12 per cent, of Si0 2 there will be about 48 per cent. CaO, while with 20 per cent, of Si0 2 there will be 40 per cent. CaO. In the attainment of this ratio between Si0 2 and CaO the purity of the lime is an important factor, especially when a slag low in silica is needed. Ordinary lime as it comes from the kiln contains a certain unex- pelled percentage of C0 2 , and, in the handling and exposure prior to use, it absorbs a certain amount of moisture, so that with the usual proportions of earthy impurities it will average about 80 per cent, of CaO. SEC. Xlg. Relative value of limes as determined lg their chem- ical composition. The content of Si0 2 in the lime depends entirely upon the kind of stone used and the care with which the ash of the fuel is kept separate. When a choice must be made between a cheap and impure lime and a more costly article low in silica, the value of each may be calculated by finding the excess of CaO" over what is necessary to satisfy its own acids. Two representative limes are assumed in Table XI-C, both containing 80 per cent. CaO, one with 3 per cent, and the other with 7 per cent. Si0 2 , and the computation is made for two different slags. TABLE XI-C. Relative Values of Limes with 3.0 and 7.0 Per Cent, of Si0 2 . SiO 3 in slag; percent Slag A. Slag B. Lime with 3 per cent. SiO a . Lime with 7 per cent. SiO,. Lime with 3 per cent. SiO,. Lime with 7 per cent. SfO,. 12.0 48.0 4.0 80.0 12.0 12.0 48.0 4.0 80.0 20.0 40.0 2.0 80.0 20.0 40.0 2.0 80.0 CaO in slag; percent. Ratio CaO to SiO 2 in slag Total CaO in lime; per cent. . . CaO in the liine which is needed to satisfy its own silica; per cent. 4.0x3.0 . . ... 4.0X7.0 28.0 20x30 6.0 2.0X7.0 14.0 CaO available for foreign silica; per cent . . ... 68.0 1.31 52.0 1.00 74.0 1.12 66.0 1.00 Relative value. . , 288 METALLUKGY OF IRON AND STEEL. It will be seen that the pure lime is worth 31 per cent, more than the impure kind when a calcareous slag is to be formed, but if a more silicious cinder is permissible, as in the case when very little phosphorus is to be removed, the pure lime is worth only 12 per cent, more. SEC. Xlh. History of basic open-hearth slags. The proportions of Si0 2 and CaO are the main points in the construction of a basic slag, but there are other factors which exercise an important influ- ence upon the result. Magnesia is always present from the wear of the hearth, but is rather undesirable, as it makes the slag vis- cous and has much less power to hold phosphorus than lime. Alumina comes from the impurities in the dolomite, lime and ore, but being usually in small amount may be neglected except when an analysis is expected to add up to 100 per cent. The same is true of the alkalies and small percentages of miscellaneous impuri- ties. Manganese is usually present in the stock and serves a use- ful purpose in conferring fluidity upon the slag, so that, being a base itself, the total basic content can be higher than with a slag containing only silica and lime. It is also valuable in removing sulpljur, for there is a tendency toward the formation of sulphide of manganese, which floats to the top of the metal where the sul- phur, being exposed to the flame, is oxidized and passes away with the waste gases. This action is rather uncertain, and, in fact, the explanation is somewhat a matter of supposition, but it seems quite well proven that manganese, either metallic or in the form of ore, *iids in the elimination of sulphur, and the above theory is in accord with certain well-known phenomena of liquation in the puri- fication of pig-iron by the addition of spiegel, as described by Mas- senez.* All the components thus far enumerated are in great measure fixed and determined agents in the transactions. It is true that manganese is sometimes reduced from the slag by the carbon of the bath, and also that a certain percentage may remain unoxidized in the metal, but aside from this it may be said that the oxides of aluminum, silicon and manganese exist in the slag in just the quantities that were added with the stock ; but there are three other constituents, iron oxide, phosphoric acid, and sulphur, whose pres- * On the Elimination of Sulphur from Pig-iron. Journal I. and 8. I., Vol. II. 1891, p. 7G. THE BASIC OPEN-HEAKTH PROCESS. 289 ence in the slag is determined by the conditions of manipulation and by the proportions of the other constituents. Iron oxide is always present in greater or less extent, the exact amount depending upon the reducing power of the carbon of the bath. It matters not whether ore is added before melting, after melting, or not at all; there is a certain content of FeO which is demanded by the existing conditions, and that certain content will be present. An exception must be made in the case of ore added after the carbon is nearly eliminated, but aside from this there will be just as much iron oxide lost in the slag when no ore is used as when it has been added in proper quantity, and therefore it may be assumed that all the ore is a clear gain and that its iron is all re- duced and added to the metallic bath. The presence of iron oxide in either acid or basic slag is an anomaly, for in an ordinary acid charge it seems as if the oxidation of the silicon and manganese would be sufficient to produce a slag without other aid. Nevertheless we have found in the foregoing chapter that there is a force at work in an acid furnace which is constantly creating a slag with a composition of about 50 per cent. Si0 2 and 45 per cent. FeO+MnO. If more FeO is added, the car- bon of the metal immediately seizes the oxygen and sets free metal- lic iron, but the same powerful action which so quickly accom- plishes the destruction of this excess is not able to pass much below the limit even by exposure for hours without any addition of ore. There is an automatic adjustment to a fixed status which is one of the most wonderful phenomena of chemical physics. The only ex- planation I can offer is that it is an instance of the general law that all forces tend to work along the lines of least resistance, which, being interpreted for this case, means that a slag will seek to combine with anything that promotes fusibility. If given the opportunity a silicious slag absorbs either bases or silica, but pre- ferably bases, and particularly those which impart the greatest fluidity. This action tends to continue indefinitely, and in an acid furnace, if the heat is not tapped after the carbon is burned, the formation of iron oxide will go on with great rapidity, and the fluidity of the slag will be greatly increased in spite of the cutting of the hearth. This latter action is a correcting condition, but it is not the controlling influence under ordinary circumstances, as is proven by the small amount of the scorification of the hearth dur- ing oreing. The real determinant is the carbon of the bath, and 290 METALLURGY OF IROX AND STEEL. there is an equilibrium established between the oxidizing power of the flame, the reducing power of the metalloids, and the struggle after fluidity. In the basic process there is a difficulty in making a slag com- posed entirely of silicate of lime, for this is much more viscous than a slag of the same percentage of silica containing other bases ; there is a tendency, therefore, toward the absorption of iron oxide, but this is opposed by a contest on the part of the lime for the possession of the silica, and the result is a decrease in the percent- age of iron when there is an increase in lime. Inasmuch 'as the substitution of CaO for FeO produces a more viscous slag, this would seem to invalidate the theory just advanced, but, as above indicated, the effect is due not to a change in the law but to the action of stronger forces. Thi? more bases that are present, the less necessity is there for an additional amount, since the weight of silica necessarily remains constant, and, as the reducing action of the metalloids comes into play, the slag begins to be robbed of its iron, which at the same time is its most reducible and its most fusible base. The presence of oxide of manganese in the slag modi- fies without completely changing the relations just described, for, by furnishing an additional base and imparting greater fluidity, it tends to render the presence of iron oxide less necessary. SEC. Xli. Automatic regulation of fluidity in basic open-hearth slag. This matter of fluidity is of vital practical importance, for the slag must run freely from the furnace, else the hearth will soon be filled ; furthermore, the slag must be so basic that the hearth is not scorified. The two conditions, fluidity and basicity, determine the nature and amount of the basic additions, for the sum of CaO and MgO cannot much exceed 55 per cent, without producing a viscous cinder, neither can the percentage of Si0 2 fall below 10 per cent, unless unusual amounts are present of the oxides of iron, manganese, or phosphorus. I have advanced this theory of the automatic regulation of fluid- ity with some hesitation, but it seems to account for a curious rela- tion between the content of Si0 2 and FeO in a large number of basic slags, which are grouped and averaged in Table XI-D. The phosphoric acid was not determined, but it may be taken for granted that an increased proportion of phosphorus in the charge will give higher phosphoric acid in the cinder, and the table shows that in the case of high phosphorus the combined Si0 2 and FeO THE BASIC OPEN-HEARTH PROCESS. 291 runs about 27.5 per cent., with medium phosphorus about 35 per cent., and with low phosphorus about 36 to 37 per cent. It is quite true that a difference in manipulation would change the absolute TABLE XI-D. Relation Between Si0 2 and FeO in Basic Open-Hearth Slags.* m S no o> SH 0> Composition of slag; 3 o o oT o o Limits of SiO, in slag, per cent. ~ 0&0 ||| & *ti per cent. 2 o.5 A o .fl-S o o ^ PH E SiO,. FeO. SiO,. + FeO. i 8 1.35 .068 below 10 9.20 18.45 27.651 2 10 1.35 .088 above 10 12.54 14.93 27.47 8 15 0.19 .016 8 to 12 incl. 10.71 25.31 36.02 4 16 0.19 .017 13 to 14 incl. 18.84 21.81 85.65 5 19 0.19 .020 15 to 16 incl. 15.90 18.21 84.11 6 13 0.19 .022 17 17.32 17.97 85.29 7 13 0.19 .025 18 to 19 incl. 18.94 15.50 84.44 8 12 0.19 .023 20 to 22 incl. 21.57 13.58 85.15 9 7 0.19 .059 23 to 27 incl. 25.48 9.04 84.52 10 16 0.10 .014 10 to 13 incl. 12.28 22.18 84.46 11 14 0.10 .012 14 14.47 22.78 87.25 12 15 0.10 .016 15 15.54 21.10 86.64 13 2.) 0.10 .017 16 16.46 21.32 87.78 14 19 0.10 .015 17 17.47 19.24 86.71 15 12 0.10 .012 18 18.32 20.02 88.34 16 11 0.10 .018 19 19,41 17.66 87.07 17 14 0.10 .020 20 20.53 14.92 85.45 18 21 0.10 .016 21 21.51 14.58 86.09 19 17 0.10 .019 22 22.46 13.41 35.87 20 11 ~ 0.10 .022 23 23.41 12.40 85.81 21 9 0.10 .028 24 24.48 11.05 85.53 22 12 0.10 .042 25 to 29 incl. 26.37 10.58 36.95 percentages, but the attainment of a certain definite content of FeO-f-Si0 2 seems assured. This conclusion is verified by an ex- amination of the individuals of the original records, for it is found that low Si0 2 is accompanied by high FeO and vice versa. This is TABLE XI-E. Maxima and Minima in the Individual Heats Composing the Groups in Table XI-D. Initial phos- phorus in charge; per cent. Slag showing maximum 8iO a ; per cent. Slag showing maximum FeO; per cent. SiO,. FeO. SiO,. FeO. 1.35 0.19 0.10 16.50 27.35 29.15 6.99 6.63 8.27 9.46 9.53 15.66 27.72 84.47 84.36 * The full records of the above charges will be found in Sec. 45 of my paper on The Open-Hearth Process, in Trans. A. I. M. E., Vol. XXII, p. 436 et seq. 292 METALLURGY OF IROX AND STEEL. shown by Table XI-E, which is composed of the extreme cases of high and low percentages of Si0 2 and FeO, found in the individual heats which compose the groups in Table XI-D. It would be entirely wrong to suppose that an increase in Si0 2 has reduced the FeO by simple dilution, for a reduction in FeO from 20 per cent, to 10 per cent, would imply a permanent addi- tion of Si0 2 equal to the entire volume of the slag, and this is manifestly absurd. The conclusion seems inevitable that Si0 2 and FeO replace one another in some way, and that one fulfils some function of the other. As FeO is basic and Si0 2 is acid, this func- tion cannot possibly be related to the basicity of the slag or any strictly chemical status, and the only explanation which suggests itself is that both confer fluidity and that there is an automatic regulation of this quality in accordance with the theory before elaborated. SEC. 'X.Ij.r-Determining chemical conditions in basic open- hearth slags. Just as oxide of iron exists in slag in accprdance with favorable conditions rather than with the initial character of the charge, so the content of phosphoric acid is governed by the chemical environment. As a general law it may be said that the capacity of a cinder for phosphoric acid increases with the propor- tion of bases it contains, and that lime is the most potent of these bases. The most important modification of this law is the neces- sity for a certain fluidity, since a slag which is very viscous does not seem to be as effective as one which is rendered fluid by oxide of manganese or iron. Thus, although lime is immeasurably superior to oxide of iron as a dephosphorizing agent, nevertheless, as T have shown elsewhere,* a slag containing a slightly higher percentage of FeO is more efficient. One of the more important determinants of the capacity of slag for phosphorus is the phosphorus itself. The absorption of phos- phoric acid is not a case of simple solution (if such a phenomenon exists) like that of salt in water, but it is a union of acid and base, and, therefore, each molecule of phosphoric acid which enters the slag decreases its capacity for more in exactly the same way that silica would. It is impossible to prove this conclusively by ordi- nary averages, for the additions of lime are usually regulated by the demands of the silica rather than of the phosphorus, and it is * The Open-Hearth Process. Trans. A. I. M. E., Vol. XXII, p. 446. THE BASIC OPEN-HEARTH PROCESS. 293 a coincidence if the maximum content of phosphoric acid is pres- ent. Moreover, the precise determining conditions vary with each par- ticular combination of the remaining elements, with the intensity of the reducing conditions, and with the duration of the exposure. Thus Table XI-F gives examples of slags which were produced un- der abnormal conditions; the samples were taken from an open- hearth furnace soon after melting, and before an extreme tempera- ture had been reached to give the carbon of the bath its full reduc- ing power to break up unstable compounds. TABLE XI-F. Unstable Basic Open-Hearth Slags. Slag. Composition, per cent. SiO 2 . P 2 5 . FeO. SiO 3 .+ P a 6 . 1 2 3 4 5 6 7 8 37.53 34.05 32.45 30.26 25.21 20.60 17.31 15.07 2.01 3.08 8.33 5.99 8.34 10.97 16.60 23.06 10.26 18.45 9.36 10.08 11.88 10.90 12.15 10.53 39.54 87.13 35.78 86.25 83.55 81.57 83.91 38.13 These slags are especially selected as being extreme instances of high phosphorus for a given silica, and they are therefore value- less as an indication of what may be expected in regular practice. They do show, however, that there is no such thing as a critical percentage of silica, since a cinder with 37 per cent. Si0 2 may hold 2 per cent. P 2 5 . TABLE XI-G. Normal Basic Open-Hearth Slags. Slag. Composition, per cent. SiO a . P a O B . FeO. SiO a .+ P a O 6 . 1 2 3 20.72 19.04 12.40 6.36 8.24 18.73 16.20 20.16 12.60 27.08 27.28 26.13 The slags in Table XI-G, although selected somewhat arbitrarily, are fairer examples of the results of regular work. In both Table XI-F and XI-G- there is a column headed "Si0 2 +P 2 5 ," and the constancy of this total under similar conditions, even. with 294 METALLURGY OF IRON AND STEEL. slags of widely varying character, indicates that the total acid con- tent of the slag is the measure of its power to absorb phosphorus. SEC. Xlk. Elimination of sulphur. A certain proportion of phosphorus is likely to be volatilized by the heat and carried away in the waste gases. This renders futile any attempts to make ac- curate quantitative calculations on the chemical history, but other- wise the action is of little importance since it cannot be relied on for purification of the metal. This volatilization occurs in greater measure in the case of sulphur, but here also it is entirely imprac- ticable to eliminate any appreciable proportion by this method alone, since volatilization occurs only from the slag, and the action, therefore, presupposes the transfer of sulphur from the metal to the cinder, and this in turn presupposes a condition which will purify the metal without the ex post facto intervention of volatili- zation. The removal of sulphur can be accomplished in at least four ways, which will be considered seriatim. \1) By the addition of metallic manganese and liquation of sulphide of manganese. The extent of this reaction is very uncer- tain, but usually the addition of 0.60 to 0.75 per cent, of manga- nese in the form of recarburizer reduces the sulphur content about 0.01 per cent. (2) By the use of manganese ore, which, being reduced by the metalloids of the bath, furnishes metallic manganese. The ore should be added with the original charge in order that it may be thoroughly mixed with the metal. It is very difficult to isolate the effect of this agent from the contemporaneous action of the basic slag with which it must be associated, but there is no doubt that it aids in the purification. (3) By the action of a very limey cinder. In a former paper* I gave the results of experiments in removing sulphur by ordinary lime slags. The cinder, during melting, was kept high in silica to economize lime, and part of this slag was removed after fusion, and fresh lime added. Notwithstanding the high acid content, the slag after melting held quite an appreciable proportion of sulphur. The final slag, being richer in lime, removed a greater quantity and the results seem to show that, as the silica decreases, the capacity for sulphur increases, but the relation is not as regular as might be * The Open-Hearth Process. Trans. A. I. M. E. f Vol. XXII, p. 446. THE BASIC OPEN-HEARTH PROCESS. 295 wished, and it must be acknowledged that many points are still obscure. The records are given in Tables XI-H and XI-I. TABLE XI-H. Basic Open-Hearth Slags after Melting, Arranged According to their Sulphur Content. Sulphur Charge Number. Initial sulphur, per cent. in metal after melting, Composition of slag after melting, per cent. per cent. S. SiO 2 . FeO. CaO. MnO. 1546 .43 .28 .28 37.53 10.26 34.53 4.66 1611 .20 .14 .26 32.63 10.17 86.25 und. 1608 .28 .17 .22 31.30 10.98 41.45 und. 1628 .20 .16 .21 33.20 9.45 und. und. 1648 .20 .14 .21 34.37 6.57 und. und. 1567 .28 .18 .20 30.26 10.08 45.26 5.42 1646 .20 .15 .18 83.97 11.61 und. und. 1626 .20 .11 .18 86.43 5.04 und. und. 1564 .28 .10 .17 82.45 9.36 45.05 5.49 1555 .28 .22 .14 80.63 13.41 89.17 7.15 1680 .20 .09 .14 25.57 8.01 und. und. 1606 .28 .19 .12 85.79 18.00 83.18 und. 1569 .28 .19 .08 34.05 18.45 85.09 6.25 TABLE XI-I. Basic Open-Hearth Slags before adding Eecarburizer, Arranged according to their Sulphur Content. J all Sulphur, after melting. lo~ Composition of slag before adding the recarburizer, per cent. 2; a o 25*tH ,M Slag, per ct. Metal, per ct. 3-sl 02 S. SiO a . FeO. CaO. MnO. 1608 .28 .22 .17 .095 .61 12.73 26.91 43.99 und. 1611 .20 .26 .14 .054 .58 10.45 26.19 45.85 und. 1555 .28 .14 .22 .086 .56 13.78 26.91 42.14 4.85 1606 .28 .12 .19 .100 .54 12.90 31.14 88.58 und. 1569 .28 .08 .19 .089 .48 15.90 18.63 und. und. 1630 .20 .14 .09 .062 .43 16.26 19.98 49.50 und. 1546 .43 .28 .28 .120 .36 18.67 24.84 37.28 4.44 1567 .28 .20 .18 .062 .33 14.85 23.49 45.74 4.54 1564 .28 .17 .10 .089 .33 19.18 16.11 49.98 4.58 1648 .20 .21 .14 .090 .26 17.97 23.94 44.41 und. (4) By oxy chloride of lime. A process has been devised by E. H. Saniter* whereby sulphur is eliminated from basic open-hearth metal by the use of oxychloride of lime. It is important to note, however, that "in order to attain this result it is necessary, at an early period after the charge is melted, to obtain an exceedingly basic slag, and to add a suitable quantity of calcium chloride to * On 'a New Process for the Puri/fccation of Iron and Steel from Sulphur. Journal I. and S. I., Vol. II, 1892, p. 216 ; also, A Supplementary Paper on a New Process on Desulphurizing Iron and Steel. Journal I. and 8. I. t Vol. I, 1893, p. 73. 296 METALLURGY OF IRON AND STEEL. it"; and it is further specified that "by a very basic slag is not meant what has hitherto been considered as such, but a step in advance of that with about 50 to 60 per cent, of lime." This point is also insisted upon by Stead,* who reviews the experiments and states that the chloride is used "in conjunction with an excess of lime over and above what is usually employed." He gives analyses of slag and metal for two charges, and a summary of these is given in Table XI-J. The results of a more complete investigation of one charge are shown in Table XIrK, the data being taken from a paper by Snelus.f TABLE XI-J. Elimination of Sulphur by Calcium Chloride. Beat. Composition, per cent. ..^7 Metal. Slag. Sulphur. After adding CaCl 2 . At time of tapping. Initial. In steel. Si0 2 . CaO. S. Si0 2 . CaO. S. 1 2 .37 .17 .047 .055 10.75 14.45 54.65 44.34 1.25 .53 10.20 11.75 48.08 47.86 .05 .57 TABLE XI-K. Detailed Data on the Elimination of Sulphur by Calcium Chloride. Open-hearth charge : 80 per cent, white iron, 20 per cent, scrap, the whole averaging about .30 sulphur. Time of taking sample. Composition of metal, per cent. Composition of slag per cent. C. S. SiO 2 . CaO. S. After complete fusion 1 hour after melting .20 .09 .06 .10 .320 .181 .093 .040 18.30 15.00 11.60 10.80 49.24 49.60 55.64 57.00 .315 .576 .659 .645 4 hours after melting Steel, 5% hours after melting . The sulphur after melting is higher than the calculated initial content, but this is probably due to incorrect sampling and to the absorption of sulphur from ore and gas, since the percentage of * On the Elimination of Sulphur from Iron. Journal I. and S. I., Vol. II, 1802, p. 260. t Report upon the Saniter Desulphurization Process. Journal I. and S. I. t Vol. I, 1893, p. 82. THE BASIC OPEN-HEARTH PROCESS. 297 sulphur in the slag shows that a considerable amount was taken from the metal. After melting, the carbon was reduced to .20 per cent., and one hour later it was .09 per cent., but it was necessary to hold the charge in the furnace for four and one-half hours after complete decarburization, and to dose it with calcium chloride in the proportion of 50 pounds to the ton of metal, in order to remove the sulphur, a delay which is decidedly objectionable. The oxy- chloride, however, conferred fluidity upon the cinder, and made it possible to carry as high as 57 per cent, of CaO, and it is probable that this increased mobility and corresponding activity rendered the lime more efficacious in absorbing sulphur. This point is not satisfactorily settled, for notwithstanding the learned discussions and investigations following Saniter's experiments,* the inner his- tory of desulphurization is still unwritten. A quantitative investigation that I made into the elimination of sulphur by weighing and analyzing the slags from three of the charges given in Table XI-II, showed that about 36 per cent, of the sulphur was unaccounted for, having probably been carried away as sulphurous acid (S0 2 ) in the waste gases. The fact that both sulphur and phosphorus thus escape in an intangible form and in uncertain quantities, renders quantitative work on basic slags very unsatisfactory. Moreover, a sample of slag is not always repre- sentative, for on some heats portions of the basic additions remain sticking to the hearth, while on others old accumulations of such deposits dissolve in a charge to which they do not belong. SEC. XII. Removal of the slag after melting. When the stock is properly charged, the greater part of the basic addition becomes an active agent during the melting of the charge. Especially when ore is used the intense action oxidizes a considerable proportion of the phosphorus during the melting, and the slag after fusion con- tains oftentimes a high percentage of phosphoric acid. The idea has occurred to numberless metallurgists that this first slag should be removed in order to get rid of its phosphorus and silica and thus give the opportunity for a new and purer slag having a greater dephosphorizing power. There are certain practical difficulties in the way, for the height of the metal in the hearth is always varying with the filling of the bottom and with the frothing of the charge, so that there is danger of losing metal if a taphole is opened much Report upon the Saniter Desulphurization Process. Journal I. and 8. I., Vol. I, 1893, p. 82. 298 METALLURGY OF IRON AND STEEL. below the level of the upper surface of the slag; on the contrary, if the slag is tapped from its upper surface there is no force to the stream and it is constantly chilling as it runs. In spite of these troubles the partial removal of the slag* is not uncommon. Complete removal can be accomplished by the use of a tilting furnace, for the entire charge can be poured out and only the metal returned to the hearth. Under ordinary conditions this manipulation is unnecessary, but it may not be unprofitable to con- sider the rules that apply, whether the whole or only a part of the slag be removed during the progress of the operation. Given a pig-iron containing a considerable proportion of silicon and with low phosphorus, it will be an advantage to have the first slag as high in silica as possible so as to avoid the addition of a corresponding quantity of lime. This practice, however, cannot be carried to an extreme, for if the amount of lime is re- duced to such an extent that the slag after melting contain much over 30 per cent, of silica, the hearth will be badly scorified. If melted pig-iron is used, this difficulty disappears, for ore may be added to a bath of pig at the rate of over one ton per hour and the silicon be rapidly oxidized. The slag so produced in the ab- sence of a full supply of lime may run about 30 or 35 per cent, of silica and 25 to 35 per cent, of iron oxide. This would scorify the hearth if left long enough in the furnace, but it should be removed after the silicon is oxidized, for during the oxidation of carbon from a content of 3 per cent, down to about 1 per cent, the frothing is very violent, and if the slag is not removed there will be considerable trouble and delay. If the pig carries much phosphorus or sulphur, the first slag which it is intended to remove should not be too rich in silica, for under these conditions the full content of the impurities will re- main in the metal after the tapping of the slag and it will be necessary to make a large volume of cinder to remove them during the second stage of the operation. The better way in this case is to make the first slag rich enough in lime to carry a good propor- tion of phosphoric acid and sulphur, and liquid enough to pour well. The second slag can then be made from fresh lime, and it will be evident that it will more readily absorb the impurities than a cinder which is already partly satisfied. SEC. Xlm. Automatic formation of a slag of ja given chemical composition. In such practice there might appear to be a dim- THE BASIC OPEN-HEARTH PROCESS. 299 culty in properly regulating the composition of the second slag, but the records in Tables XI-H and XI-I show that such is not the case, for, in the heats there given, a part of the slag was removed soon after melting. Quite a difference will be found between the first and second slags, but this is because the first slag was pur- posely made high in silica in order to save lime. When it is re- quired to maintain a similar composition throughout the heat, it can be done in basic as well as in acid practice, as shown by the average slag analyses of 27 heats in Table XI-L. TABLE XI-L. Average Slag Analyses of Twenty-seven Basic Open-Hearth Heats. Slag. Composition, per cent. SiO 2 . P a 6 . CaO. FeO. A.fter melting 14.35 12.40 15.53 13.73 45.07 45.40 9.00 12.60 Before tapping Four-fifths of the lime was added with the charge, and the re- mainder, together with 400 pounds of ore, was used after melting, but in spite of the incorporation of this basic material into the slag during the interval between the two stages at which the samples were taken, it will be seen that by careful supervision and through the action of the internal chemical forces, a remarkably uniform composition was maintained, which proves conclusively that the manipulations of the basic process may be as completely under control as the operations upon the acid hearth. SEC. XIn. Recarburization and rephosphorization. Kecarburi- zation is carried on in the same way as in acid work, and is subject to the same general laws. A complicating condition is often added when either the stock or the ore contains any considerable propor- tion of manganese, for the decarburized metal may then hold as much as .20 or .30 per cent, of Mn. Not only must this be allowed for in making the final addition, but it will also be found that the bath contains less oxygen under these circumstances, and therefore there will be less loss of metallic manganese during the reaction. In basic practice there is a factor not present in acid work, in the danger of rephosphorization, or the return of phosphorus from slag to metal. In the basic Bessemer this is a source of consider- able trouble, but in the open-hearth the recarburizer is almost 300 METALLURGY OF IRON AND STEEL. always added in a solid state and the metal probably contains less- oxygen, so that the reaction is less violent. Moreover, during the solution of the ferro, the slag is constantly at work with its de- phosphorizing influence, so that the sum total of the reactions may even show a decrease in phosphorus. Other things being equal, it would seem probable that a slag containing a high percentage of phosphoric acid will hold this component less firmly than a purer cinder, and I have tried to illustrate this point* by experiments, the results of which may be summarized as follows : (1) With slags containing under 5 per cent. P 2 5 and not over 20 per cent. Si0 2 , the rephosphorization need not exceed .01 nor average over zero per cent. (2) With slags containing from 5 to 10 per cent. P 2 5 and not over 19 per cent. Si0 2 , the rephosphorization need not exceed .015 nor average over .005 per cent. (3) With slags containing from 10 to 15 per cent. P 2 5 and not over 17 per cent. Si0 2 , the rephosphorization need not exceed .02 nor average over .005 per cent. (4) With slags containing from 15 to 20 per cent. P 2 5 and not over 12 per cent. Si0 2 , the rephosphorization need not exceed .02 nor average over .01 per cent. In using phosphoric stock it is not safe to presuppose the elimi- nation of phosphorus below .04 per cent, until the carbon has been lowered to about .08 per cent. Hence, to make rail steel it is neces- sary to eliminate the carbon to that point and then add the required amount of recarburizer, as in the Bessemer process. It is imprac- ticable to use melted spiegel-iron in open-hearth practice, unless there are a great number of furnaces, because the charges come so irregularly and at such long intervals that a cupola becomes chilled, but it has been found possible to add finely divided carbon in the ladle, its absorption by the metal being so rapid that the results are quite regular. Several ways of doing this have been devised, the most successful of which has been very fully described by Dr. Wedding, f Pow- dered "anthracite" coal is mixed with about 7 per cent, of burned lime and with sufficient water to make a plastic mass, and is then formed into bricks. These are dried thoroughly to expel all the * The Open-Hearth Process. Trans. A. I. M. E., VoJ. XXII, p. 484. t Stahl und Eisen. 1894, pp. 473 and 533 ; also 1895, p. 570. THE BASIC OPEN-HEARTH PROCESS. 301 uncombined water, and are then ready to be fed into the ladle as the heat is poured. The escape of the combined water in the lime causes the bricks to crumble to pieces when in contact with the melted steel, but this crumbling is gradual, so that the carbon is fed to the metal con- tinuously and the bath is able to absorb it as fast as it is set free. This moisture also creates a constant motion of the bricks and acts -as a mechanical stirrer. It should be noted, however, that the kind of coal which is re- ferred to by Dr. Wedding is a rather hard bituminous coal and not at all what is known as "anthracite" in America, and that the practice at different works leads to the conclusion that coke dust or other similar forms of carbon answer equally well. CHAPTEE XII. SPECIAL METHODS OF MANUFACTURE AND SOME ITEMS AFFECTING THE COST. SEC. Xlla. The manufacture of low phosphorus acid open- hearth steel at Steelton. The early history of the open-hearth in the United States is confined entirely to the making of acid steel, very little basic metal being made until after 1890. A large pro- portion of the output went into boiler plate and quite a quantity into forgings, while there was a considerable tonnage of high carbon steel, which was ultimately sold under the name of "cast steel," this term being perfectly truthful in one sense and entirely un- truthful in another, as it was intended to convey the idea that the metal was made in a crucible. The ordinary grades of boiler steel and forgings were made of stock running from .08 to .10 per cent, of phosphorus, while metal for fire boxes and special forgings, as well as some of the high carbon steel, was made of low-phosphorus stock, usually a mixture of Swedish pig-iron and charcoal blooms. A certain quantity of low-phosphorus pig-iron was made in America, and during the latter part of the acid epoch a considerable quantity was manufac- tured of what is known as "washed metal." This is made by treat- ing melted pig-iron in a furnace lined with iron ore and lime and eliminating most of the silicon, sulphur and phosphorus and about half the carbon. The pig-iron is the same grade as is used in the basic open-hearth furnace, and the "washed metal" process is essen- tially the same as the basic open-hearth process of to-day. It differs from it in the following particulars: (1) In the basic open-hearth furnace, the bottom is made as durable as possible and it is desired that it shall not be cut away by the action of the metal and slag. The iron ore needed to oxidize the metalloids and the lime needed to make a basic slag are both added with the charge, and the reactions take place in a 302 METHODS OF MANUFACTURE, AND COST. 303 definite way very similar to the fusions made by a chemist in a platinum crucible, the crucible playing no part in the reaction. In the washed metal process the bottom is not durable, but is intended to be the source of supply of the ore and lime needed to oxidize the metalloids and to supply a basic slag. (2) The washed metal furnace is not allowed to reach a very high temperature, because the slag is not stable and at a higher temperature the hearth would be cut away, the reactions would be more violent and the phosphorus would leave the slag and go back into the metal. In the open-hearth furnace the phosphorus does not go back, because the slag is stable, by which is meant that it contains a sufficient proportion of lime to make a permanent com- pound with the phosphorus so that it is not readily reduced by carbon. Such a slag needs a high temperature for complete fusion and this temperature cannot well be carried in the washed metal furnace. (3) The washed metal furnace is tapped when the metal con- tains about 2 per cent, of carbon, because if the carbon be run down any lower a much higher temperature would be needed, and because this kind of product suits the demands of the trade. It has been stated that the standard low-phosphorus open-hearth steel of former days was made from either low-phosphorus pig- iron and charcoal blooms or washed metal and charcoal blooms, and it has been shown that this washed metal was the product of a basic process. The charcoal blooms were also of basic origin, because they were made in a primitive Tubal Cain sort of way by the action of a basic oxidizing slag on melted metal. After the general introduction of the basic open-hearth process it became possible to buy in the open market a supply of low- phosphorus steel scrap at a very moderate price, and this steel scrap rapidly took the place of the high-priced charcoal blooms and practically stopped their manufacture. Thus while the ad- vent of the basic open-hearth furnace rendered it possible to pro- duce a low-phosphorus steel very much cheaper than it had ever been produced before, it also cheapened the cost of low-phosphorus acid open-hearth steel by giving it cheap scrap. This is true, however, only to a certain extent, for the basic furnaces themselves need scrap and use most of the available supply. Moreover, the different plants for making steel castings 304 METALLURGY OF IRON AND STEEL. are always in the market, and some of the plate mills use steel plate scrap to pile with puddled iron to make wrought-iron plate, so that it is difficult to find sufficient low-phosphorus scrap to keep a large acid open-hearth plant in continual operation, and even if this could be done, the low-phosphorus pig-iron, which must be used, costs from three to five dollars per ton more than the ordi- nary Bessemer grade. In order to overcome these commercial difficulties we have intro- duced at the works of The Pennsylvania Steel Company an adapta- tion of the old washed metal process. The pig-iron is charged, either liquid or solid, in a basic lined furnace and almost all of the silicon and phosphorus and part of the sulphur and carbon are eliminated. At this stage of the proceeding it is washed metal, and in olden times would have been run out in chills, cooled off and afterward charged into the acid furnace, but in this new practice it is poured into a ladle, and, while still fluid, is poured directly into the acid furnace. A certain amount of scrap may be used in the basic furnace, or in the acid furnace, or in both; but the main point is to have no basic slag enter the acid furnace and to be sure that the dephosphorized metal, when it goes into that furnace, shall contain as much carbon as is usually present in an acid bath after the stock is melted. We thus have the transferred charge starting off on its acid journey in just the same condition it would have been in if it had been melted in the acid furnace, bo that the reaction, the slag, and the whole history from that moment, -are the reactions, the slag and the history of the acid open-hearth furnace. This practice is not feasible in most open-hearth plants, since no arrangements are usually provided for transferring metal in this way, but the demands of engineers for pure acid open-hearth steel made it necessary to equip a plant to supply this special product at a moderate cost. In order to show that the compo- sition of the metal and slag in the transfer process is the same as in the usual acid furnace, I had samples taken from the bath during different stages of the operation. The metal was tapped from the basic furnace when it contained from 2.50 per cent, to 3.50 per cent, of carbon, and transferred in a molten state to the acid furnace. When the carbon was about 1.00 per cent, the taking of samples was begun. It is seldom that a charge in an METHODS OF MANUFACTURE, AND COST. 305 acid furnace is higher than this when it is melted, so that the records may fairly be compared with the ordinary acid heat after complete fusion. TABLE XII-A. Composition of Metal and Slag in the Acid Furnace when Washed Metal is Transferred in a Molten State from a Basic to ail Apid Furnace. Note : Samples over 1.10 per cent, in carbon omitted. Heat No. Composition of Metal, per cent. Composition of Slag, per cent. C Si S P SiO 2 MnO. FeO MnO+FeO SiO a +MnO-f- FeO A . 1 00 02 .033 .025 50 . 57 12. 16 00 f\A A A OA f\4 17-7 .71 .01 !037 !025 49^91 11.' 08 O^.UT: 32 58 44. zO 43.66 y4. // 93.57 .30 .03 .037 .029 55.76 9.75 28.05 37.80 93.56 .19 .02 .033 .025 55.44 9.22 30.15 39.37 94.81 B .80 .03 .025 009 47.71 3.46 44.64 JQ in OK 01 .31 .03 .020 .008 53^90 4^30 3?! 62 rro. U 41.92 yo.oJ. 95.82 .21 .02 .021 .008 51.50 7.67 35.55 43.22 94.72 o 95 02 .020 .019 51.08 12 94 OQ 7Q Afy f?o QQ C1 .70 .02 !020 .019 45^38 9!04 zy /y 40.05 t . /' 49.09 yo.o.L 94.47 .54 .03 .021 .022 50 01 9.10 35.55 44.65 94.66 .23 .03 .020 .021 52.61 10.92 30.87 41.79 94.40 D .77 .03 .026 .010 53.52 10.92 28.98 39.90 93.42 .45 .03 .029 .011 52.22 8.34 32.58 40.92 93.14 .31 .03 .029 .012 52.50 7.36 36.54 43.90 96.40 E .90 .02 .040 .034 51.82 6.52 37.44 43.96 95.78 .60 .01 .034 .031 50.27 7.44 38.79 46.23 96.50 .17 .02 .034 .030 51.66 5.51 39.51 45.02 96.68 F .... 1.09 02 .027 .008 42 . 50 9.89 41 76' 51 65 94 15 .72 .02 !027 !oos 51/20 io!i7 33! 75 43^92 95! 12 .24 .02 .027 .008 56.61 9.60 29.61 39.21 95.82 G .75 .01 .028 .010 46.95 11.46 39.24 50.70 97.65 .46 .01 .028 .010 51.02 10.44 33.93 44.37 95.39 .26 .01 .029 .010 54.80 11.58 28.17 39.75 94.55 H .95 .01 .022 .026 42.21 14.34 37.98 52.32 94.53 .62 .02 .024 .030 49.66 12.65 32.65 45.30 94.96 .25 .02 .023 .028 50.28 11.72 31.41 43.13 93.41 I 70 02 .030 .011 45.16 15.14 35.46 50.60 95 76 .43 .02 !028 !oio 47.65 9.' 89 36^99 46.88 94.53 .22 .03 .029 .011 57.23 9.36 26.91 36.27 93.50 The results on nine heats are given in Table XII-A, and they may be compared with figures given in Table X-B. This latter table shows, under Group I, the composition of slag and metal as found some years ago in an acid furnace running on the usual pig, scrap and ore process. A comparison of the results is shown in Table XII-B. 306 METALLURGY OF IRON AND STEEL. TABLE XII-B. Comparison of Data in Tables X-B and XII-A. Group I. Table X-B Transferred Steel. After Melting ..... .54 50.24 45.58 95.82 .13 49.40 46.29 95.69 Min. Max. .70 to 1.09 42.21 to 53.52 42.73 to 52.32 93.42 to 97.65 .17 to .31 49.40 to 50.28 36.27 to 45.02 93.41 to 96.68 Av. .88 47.95 47.13 95.08 .23 53.62 41.3u 94.92 8iO 2 in slag End of Operation FeO+MnO SiO a +FeO+MnO Carbon in meta SiO 2 in slag . FeO+MnO SiO a +FeO+MnO ' It should be stated that the last sample was not always taken just before tapping. Thus in heat D, Table XII-A, the final carbon was not .31 per cent., but the last sample was taken at that point and for the purposes of the investigation, this, was deemed sufficient. It will be seen that the composition of tho slag, both at the earlier periods and at the later epoch, corresponds closely to that taken in former experiments, and if samples had been taken with lower carbons so as to correspond with the .13 per cent, in Group I, Table X-B, it is likely that there would have been even a still closer resemblance, as the percentages of metallic oxides would probably have increased. SEC. Xllb. -The pig and ore "basic process. In the year 1901 the "United States produced 3,618,993 tons of basic, open-hearth steel, while in the year 1895, when this book first appeared, the total production of acid and basic open-hearth steel put together was only 1,137,182 tons, and in 1894 the total was 784,936 tons. The great increase was caused by an enormous expansion in the field of structural work. This field rapidly extended, owing to the cheapness of the material and to various other causes, among which might be mentioned the invention of steel skeleton office buildings, and the demand for heavier railroad bridges caused by heavier rolling stock. The introduction of steel cars also accounts for a very great demand, as well as the phenomenal growth of the tin plate business, while many smaller industries like the making of car springs constitute in the aggregate a tonnage which can hardly be credited. In early days the open-hearth furnace looked for its supply of scrap to the mills that rolled Bessemer ingots, but since 1879, when METHODS OF MANUFACTURE, AND COST. 307 the open-hearth first began to be an important producer, the output of Bessemer steel has increased only tenfold, while the product of the melting furnace has increased ninety fold. With this enormous increase in product there is naturally a demand for melting scrap, which in some localities cannot always be supplied. It is a common belief that a basic furnace can handle anything that may be picked up in a junk yard, but experience teaches that while it is undeniably true that it can do so, it teaches just as undeniably that there is no economy in using bad material unless it can be bought for a much lower price per ton. In some for- eign countries the only pig-iron available is one containing a high percentage of phosphorus. When there is plenty of steel scrap to mix with such an iron, it can be used without much trouble, but when it must be used alone the product of the furnace is lessened materially and the cost greatly increased. In America there is no incentive to use a high phosphorus mixture, except in the Ala- bama district and in Cape Breton. The ores of the Lake Superior region furnish an iron which is so low in phosphorus that this element is always eliminated in the basic furnace to below .04 per cent., which is the established standard, but in using the irons of Alabama, Tennessee and Kentucky much care is necessary or the steel may hold more than the allowable amount. The phosphorus problem is one which can be met by careful attention to the slag, by seeing that it receives sufficient lime, that it is rendered fluid by iron oxide, and that it is in sufficient quantity to hold the phosphorus in a state of stable combination. The removal of phosphorus is a local issue, in which some dis- tricts have no interest, but the question of working a large pro- portion of pig-iron is one which nearly all large works are some- times driven to face. In an ordinary stationary furnace the use of an entire charge of pig-iron is very objectionable on account of the excessive frothing of the metal and slag. From the time that the metal is thoroughly melted, when it may contain about 3 per cent, of carbon, until the proportion is reduced to about li/2 per cent., the bath resembles soda water more than pig-iron, and it tries to flow out of the doors and to occupy about twice the room it should. In Steelton we have solved the difficulty caused by this frothing by using the tilting furnace rotating about a central axis. (See Section Vllld.) The pig-iron is brought in a melted state from the 308 METALLURGY OF IRON AND STEEL. blast furnace and poured into the open-hearth furnace, a sufficient quantity of iron ore and lime being added. During the combustion of silicon no violent reaction occurs, but immediately afterward a general movement takes place, whereupon the furnace is tipped over until the metal is thrown away from the doors and up on the back side. In this way the capacity of the furnace is practically doubled, while the flame enters and goes out as usual. The furnace is kept in this position for two or three hours, or longer, until the bath has quieted down. Meanwhile the slag is trying to froth out of the ends of the furnace and down the ports, but to do so it must flow over the open joint between the port and the furnace. This joint is not wide, but special provision is made to allow the slag to run out through a small hole and fall down beneath the end of the furnace in a slag pit. In this way a very considerable quantity is removed and the time of operation considerably lessened. At some works the slag is removed by means of a small tap-hole or through the regular door, but under these circumstances the stream continually chills and must be carefully tended. In the arrangement above described there is little tendency to chill, for the flame is constantly playing back and forth through the ports and the slag opening is in the immediate course of the hottest flame. This practice of using direct metal has been in more or less continuous use for several years on furnaces of fifty tons capacity. Working in this way the iron of the ore is reduced in such quantity that the product of steel, counting both ingots and scrap, exceeds the weight of pig-iron charged by from 4 to 6 per cent, when the charge is entirely pig-iron. There is nothing new in this practice, the only feature which distinguishes it from work done at many other places at many times in the past being the use of a tipping furnace rotating round a central axis. With the Wellman furnace it would be im- possible to tip the furnace in the manner described, and while this would not prevent the use of melted pig-iron for the entire charge. it would materially increase the difficulties unless the furnace were charged to only half its capacity. It is not necessary that the iron should be brought in a melted state from the blast furnace, as the same general line of procedure can be followed when it is charged cold. Table XII-C shows the results obtained from two series of heats, in one of which most of the metal was charged cold, while in the other the metal was all fluid. In these series METHODS OF MANUFACTURE, AND COST. 309 especial care was taken to have the weights accurate and to know the composition and the weight of the slag produced. I do not con- sider that any results on loss are worthy serious study unless the exact amount of pure metallic iron put into the furnace is known and unless this equals the weight of metallic iron in the ingots, the scrap and the slag. In addition to this it is well to know the total amount of CaOput into the furnace in the form of limestone, burned lime or dolomite, and see whether this agrees with the amount of CaO which is indicated by the weight and composition of the slag. In the following two series these conditions were attained and the amount of CaO used was found to check the records of the slag, while the balance sheet of metallic iron agrees within one-fifth of one per cent. In individual heats no such accuracy can be ob- tained, and it is often impossible on a series of heats, as the wear- ing of the hearth or the accumulation of slag will give a gain or a loss. In Table XII-C the term "first slag" signifies that which flows through the port opening, and is thus removed entirely from the furnace during the progress of the operation, while "second slag" means the final cinder as it comes from the furnace at the time of tapping : ;: * ; TABLE XII-C. Eecord of "All-Pig" Basic Open-Hearth Heats at Steelton. x First Series. Pounds. Second Series. Pounds. Li Ir Ir R< ! Oi In Sc * ' Fi Se - quid metal (1.4 per cent. Si). . . 156.200 352,210 36.020 3,600 405,287 ^carburizer . 4,725 Total metal charged 548,030 144,100 551,200 13,800 410,012 116,300 429.000 1,355 e (66 3 per cent Fe) orotS rap 565,000 27,130 17,140 430,355 73,600 41,500 rst slag cond slag Total slag , 44,270 115,100 Composition of first slag ! CaO. . . . FeO.... 810,.... Composition of second slag. . 4 CaO. . . - /FeO.... 24.04 23.67 11.84 18.14 41.63 45.00 11.78 16.14 41.90 37.26 26.93 25.94 310 METALLURGY OF IRON AND STEEL. Taking as a basis the weight of pig-iron and recarburizer, the weight of ingots and scrap together was 103.1 per cent, in the case of the cold metal, and 104.95 per cent, with liquid metal. These figures, of course, neglect entirely the weight of ore charged, but it is customary to speak of such practice by saying that the gains were 3.1 per cent, and 4.95 per cent, respectively. This sub- ject will be again referred to in other sections of this chapter. In the case of the cold pig, the first and second slags together carried away 7.3 per cent, of all the metallic iron put into the furnace, including the iron in the ore. In the case of the melted iron, this loss was 7.4 per cent. The silicon in the pig-iron was 1.4 per cent., which is rather high for basic practice. Had it been lower there would have been less silica produced, less lime would have been necessary, less slag would have been produced, and less iron would have been lost in the cinder. The slag is not exactly proportionate to the silicon in the iron, as there are other sources from which silica is supplied, but it seems from calculation that had the silicon in the pig-iron been reduced one-half, to a content of 0.70 per cent., the volume of slag would have been only two-thirds as much, and this would mean that it would carry away less than 5 per cent, of the total iron in the charge, which would mean a gain of 2.5 per cent, in the weight of ingots over the actual practice and give a total gain in weight of 7.5 per cent. It is true that less ore would be required with lower silicon, but on the other hand, a lower percentage of silicon means a higher con- tent of metallic iron in the pig-iron, which is bound to show itself in a greater product. The practice of using direct metal in an open-hearth furnace is one in which the open-hearth is only hall the operation. The blast furnace is the other half, and the cost sheet of both must be considered in making up the cost of ingots. SEC. XIIc. The T allot Process. The last section described the difficulties encountered in the use of the pig and ore process in a furnace that cannot be tilted while in operation, like the ordinary stationary hearth or the Wellman type. A way of overcoming this trouble has been carried out by Mr. Talbot at the Pencoyd Iron Works, at Philadelphia.* The pig-iron is melted in a cupola and is poured into a Wellman furnace. When the charge is ready to tap, a portion of the steel, and a portion only, is poured into Journal I. and 8. I., Vol. I, 1906. METHODS OF MANUFACTURE, AND COST. 311 the ladle and cast into ingots. The remainder, which may be one- half or two-thirds of the whole, is kept in the furnace and a new supply of cupola iron is added to it. . Taking the case of a 50-ton furnace and assuming that thirty tons of low carbon metal is retained and twenty tons of pig-iron added, it is clear that the average of the new bath will contain about 1.5 per cent, of carbon, which will be quite a manageable mixture. A point in this practice which might trouble the average open- hearth man is the impossibility of repairing the lower portion of the hearth, or even of knowing what condition it is in. The slag line can be repaired after part of the charge has been removed, but the lower part of the bottom is always covered by liquid metal. It is claimed, however, that this covering of steel acts as a protec- tion by keeping away the slag and oxide of iron, and that no repairs are necessary to the "flat." Considerable stress is laid on the addition of iron oxide before the addition of pig-iron in order to create a violent reaction and quickly oxidize the metalloids, and it is even claimed by Mr. Tal- bot that this oxidation produces heat and is thus an important fac- tor in the operation. It will be shown in Section Xlle that this is a. great mistake and that the reaction absorbs much energy. Were it not so, there would be no difficulty in eliminating silicon and carbon in the open-hearth furnace by ordinary methods, for a charge can be decarburized with great rapidity by shoveling ore into the furnace continually; the reactions take place and the silicon and carbon are oxidized as fast as can be desired, but this cannot be continued because there is such an absorption of heat that the bath becomes cold and time must be given for it to get hot. It is difficult to see how the time necessary for decarburization can be shortened by preheating and melting the ore, and having a sudden and violent reaction with a consequent chilling. The de- carburization itself will take place in much less time, but the total time necessary to melt the ore, to complete the reaction, and to heat up the charge after the reaction will probably be longer than if the ore were added after the pig-iron is charged. Table XII-D is condensed from Mr. Talbot's paper showing the history of the metal and slag in the furnace. There are five heats given in full in his paper and one other heat in part, but I have moted only two, as they are fairly representative of all those de- pr-ribed. The heats given by Mr. Talbot are not consecutive, and 312 METALLURGY OF IRON AND STEEL. it is only natural to suppose that he selected those which ran along without any mishaps. It is also natural to suppose that the gen- eral average would show a somewhat less output per hour of actual operation. This supposition is corroborated by the information given in the paper on the results from two weeks' work, for while the average of the five heats indicates an output of 92 tons per day, the record for a fortnight gives an average of only 493 tons per week, which if continued would give 2136 tons in a month. TABLE XII-D. Eeactions in the Talbot Process. Note : For convenience I have started both heats at 12 :00 o'clock. Time. Sample. Weight Ibs. Composition of Metal. Composition of Slag. C S P Mn Si Fe Si0 9 P a O. MnO 12:00 12:30 1:05 1:10 1:18 1:20 1:20 1:35 1:40 1:47 1:50 1:50 3:30 3:30 3:40 4:30 4:35 4:40 4:40 12:00 12:40 1:10 1:15 1:25 1:40 1:45 2:00 2:05 2:05 3:50 4:35 4:40 4:50 4:55 Heat No. 254 Slag from previous heat Scale 10.49 11.68 13.26 7.00 3,600 90,000 23,700 113.700 2,200 1,440 113,700 12000 125,700 2,500 2,250 1 100 Bath and slag 0.06 3.80 0.49 .051 .082 .053 .026 1.012 0.132 0.08 0.26 0.15 6! 18 25.57 8.68 9.44 Bath and slag Ore 11.87 12.10 16.45 Limestone 0.38 3.80 0.71 .056 .065 .057 0.111 0.980 0.144 0.14 0.43 0.14 6. '25 10.39 12.62 17.05 Cupola iron Bath and slag . ' 10.71 12.32 15.56 Cinder Limestone Ore Limestone 1,000 Manganese Ore 800 125,700 3,000 128,700 0.07 3.80 0.11 0.16 .025 .065 .033 .050 0.035 0.980 0.041 0.036 0.17 0.43 0.18 0.50 6^25 13.95 Cupola iron 11.59 11.81 14.29 Steel and slag tapped .... 11.55 11.70 12.03 12.03 .7.83 5.12 Heat No. 306 Slag from previous heat. Sc-ile 3,800 95,000 14,000 109,000 109,000 17.200 126,200 2 300 Bath and slag .06 3.80 0.11 0.07 3.80 0.34 .053 .052 .052 .057 .057 .052 0.045 0.976 0.062 0.049 1.004 0.111 0.06 0.24 0.06 0.05 0.26 0.08 oise o.'ss 43 37 5.18 4.17 Cupola iron Bath and slag 21.17 23.16 "l8". 05 11.22 9.95 '12'. 08 10.82 9.83 12'. 45 .'.'.'.'!! Bath and slag Cupola iron Bath and slag Cinder 2,700 400 126 200 6,100 132,300 Manganese ore Bath and slag 0.07 3.80 0.07 0.14 .049 .057 .047 .050 0.022 1.004 0.030 0.038 0.08 0.26 0.10 0.45 o!35 21.54 Cupola iron Bath and sing 16 28 Steel and slag tapped ... 18.39 10.94 12.26 5.44 It is stated by Mr. Talbot that the output was decreased by the necessity of repairing the cupola at the week end, so that liquid iron was not available until Monday night, the furnace being run on cold stock meanwhile. I can hardly look upon this fact as of METHODS OF MANUFACTURE, AND COST. 313 much importance, for the rate of output with liquid metal is no greater than should be obtained from such a furnace on cold stock. The furnace in which the work was done would actually hold 70 tons, as shown by the record that 156,000 pounds were in the furnace at one period of the operations, and also by the direct statement of Mr. Talbot that it was rated at 75 tons capacity. The results therefore show that a 75-ton furnace can make steel at the rate of 2100 gross tons per month. This would hardly seem to be anything extraordinary and more than one works is now operating furnaces of less capacity and making fully as much or more on all pig heats. Moreover, it is not always that open-hearth furnaces are supplied with iron containing only 0.58 per cent, of silicon, this being the average of all the iron used in the heats cited by Mr. Talbot. The statement that there is nothing extraordinary in the output of the Talbot furnace will be questioned by some, for in the dis- cussion of the paper before the Iron and Steel Institute it seemed to be assumed that there was something unusual in the records given and the same impression is conveyed by Mr. Talbot. Thus, in some remarks on the paper, I stated what had been done with direct metal at Steelton, and Mr. Talbot asked why the practice had. not been continued when "such a splendid opportunity had been presented for increasing the output." As a matter of fact, I had not stated or intimated that the output had been increased to any wonderful extent, for we had done nearly as well on cold metal. Thus I find a time in 1896 when we were running 97.5 per cent, of cold pig-iron in a 50-ton furnace and the output was 437 tons in one week, which is at the rate of 1894 tons per month. It is not possible to give the records for long periods, because at other times a larger proportion of scrap was used. This fact may explain why no great effort was made to separate furnaces so that some would be on direct metal exclusively, as Mr. Talbot seemed to think so advisable. The use of direct metal is not revolutionary, and is not even new; it is advantageous to a certain extent, but it does not save as much time as might be expected. In the same way it will not do to lay much stress on the gain in weight from the iron ore, which is brought forward so prominently by Mr. Talbot. It is a mistake to regard this as in any way char- acteristic of the method. Section Xllg will take up at length the 314 METALLURGY OF IRON AND STEEL. discussion of this subject, while Sections Xlle and Xllf also bear upon the matter. TABLE XII-E. Data on Eate of Production and Elimination of Sulphur in Talbot Furnace. Heat. Rate of Production. Elimination of Sulphur. Weight of in- gots; Ibs. Time from tap to tap. Hours-Min. Calculated aver- age sulphur in metal charged. Sulphur in fin- ished steel. 254 S7.405 39,100 39,085 37,410 38,650 191,650 92 tons. 350 4-25 440 455 4-30 2220 .041 .048 .058 .054 .049 .038 .038 .050 .050 .054 264 285 306 408 Total Rate per 24 hours. . It will be seen from Table XII-E that there was very little elimination of sulphur in any of the heats. This shows that the slag was kept fluid and not very basic, and under these conditions the furnace will run much faster and make more product than if a better steel is made. It is not extra good practice to start with iron containing only 0.58 per cent, of silicon and .05 per cent.* of sulphur, and not eliminate any of the latter impurity. As a mat- ter of fact three out of the five heats given by Mr. Talbot would not fill the standard American specifications for boiler plate. It may be urged that there was no necessity of elimination when the content was low at the beginning. This reasoning, however, will hardly apply to the results given on pages 59 and 61,* where Mr. Talbot gives the results of two weeks' working and the com- position of fifty-five heats of steel. Of these the sulphur content was as follows : 7 heats between .040 and .049 per cent. 20 " " .050 " .059 21 " " .060 " .069 3 " " .070 " .079 3 " " .080 " .089 1 heat .090 If the slag had been made more basic, and sufficient time allowed for the elimination of sulphur, and if during all this time the LOG. cit. METHODS OF MANUFACTURE, AND COST. 315 reactions had been consummated in the presence of this more basic and more viscous and more voluminous slag, the time of the charges would have been considerably increased and the amount of fuel and all other costs correspondingly greater. In the opera- tions of the Talbot furnace as described, the iron was melted in a cupola and this tended to increase the sulphur by absorption from the coke, but on the other hand, it gave an opportunity to select the iron that was treated, and it is quite certain that a blast furnace could not be relied upon to furnish regularly a better iron than was used in the operations recited by Mr. Talbot. It is not a pleasant task to criticize a new method on the basis of results obtained in the earlier stages of practice, for improve- ments will naturally come from experience, but on the other hand it is to be remembered that a new process, when carefully tended by the eager and intelligent care of an inventor, often shows results far in excess of the average obtained in after years by alien hands. It should be said in justice to Mr. Talbot that, while my views herein expressed as to the limited value of the Talbot process are shared by a great many American metallurgists, in England it has met with great approval from eminent men. It remains for the future to decide whether there is much gained. A process or practice may be successful and yet be of no very great advantage over other similar methods. I have described a method used at Steelton for handling heats of all pig-iron. The process is suc- cessful, but the gain from it does not revolutionize anything, "and it has been worked side by side with the scrap practice as tempor- ary circumstances determine. Such conditions are understood by business men, but they are apt to be overlooked by those who devise new processes. SEC. Xlld. The Bertrand Thiel process* There has been de- veloped at Kladno, in Bohemia, a system of handling phosphoric pig-iron which has had the same misfortune that falls to the lot of most new methods. It has been over-heralded. It embodies some principles which are not new, but which have been worked out as well as the existing conditions will allow. There were two open- hearth furnaces at Kladno, and they were on two different levels, making it possible to tap from one furnace into the other by means of a runner. The higher furnace is used to remove the *This section, in an incomplete state, has been read by Mr. Bertrand. 16 METALLURGY OF IRON AND STEEL. silicon, part of the carbon and most of the phosphorus, while the second furnace completes the process. Four years ago, when the practice at Kladno had not been reduced to the precision it has reached since, Mr. Bertrand published* the results of twelve heats, which show that the metal was in the first or primary furnace an average of 4 hours and 50 minutes, and in the second furnace an average of 2 hours and 20 minutes. The proportions of pig-iron and scrap are quite unimportant, ns scrap may be used in either, or in both, or in neither of the furnaces. It is considered the best practice, however, to charge mostly pig-iron in the first furnace, using sufficient ore to give a good reaction and to oxidize the metalloids, and to charge some scrap in the second furnace. The stock in the second furnace is partly melted when the steel runs down to it from the primary furnace, and there is a quick and violent reaction. Gare is taken to allow no slag to run from the first to the second furnace, and in this way the phosphorus, which has been eliminated in the first furnace, is kept out of the operation from that time forward. The second furnace starts with a semi-purified metal and with a new and clean slag. Following is a summary of the data given by Mr. Bertrand: Metal. Slag. C P Si Mn Si0 2 P 2 6 FeO Pig iron 3.8 2.2 1.6 0.4 1.0 .05 1.0 0.5 From first furnace . .... 26.30 13.23 12.23 11.78 9.49 14.26 From second furnace The average sulphur in the finished steel is .042 per cent., but it is stated by Mr. Bertrand that all the pig-iron contained less than .05 per cent., so there would seem to be very little elimination of this element. The average phosphorus in the steel is .067 per cent. The twelve heats may be divided as follows, in their content of this element: 1 heat .021 per cent. 2 heats between .03 and .04 2 " " .04 " .05 " " 2 " .05 " .06 1 heat .075 1 " .086 1 " .098 1 " .170 * Journal I. and 8. I., Vol. I. l, co ". METHODS OF MANUFACTURE, AND COST. 317 This shows that out of these twelve heats one heat was so high in phosphorus that it could not be sold in America, while seven more were above the established standard for American basic steel. Attention is called to this fact, not so much to criticize the process, for it has been stated that the work had hardly passed beyond the experimental stage, as to illustrate that on the continent of Europe the specifications on structural steel are in no manner as severe as in America. In this country a charge known to contain .17 per G50 707 9 751 Total oxygen needed Fe 2 O 3 needed 49,530 165 100 ' 37.207 124020 Ore needed (94 per cent.) Ore used 175.640 144 100 131940 116 300 Thus it is shown that in the case of the cold pig-iron, the ore used was 82.0 per cent, of what was theoretically necessary, while in the case of the liquid metal, it was 88.1 per cent. It is quite natural that a charge of cold pig-iron should show a lessened use of ore, as part of the ^oxidation is done by the flame during the melting. The difference will be even greater than is shown here, for the series which has been called "cold pig" was really com- posed of nearly 30 per cent, of molten metal, as shown in Section Xllb. Thus in the case of the liquid metal, the amount of ore called for by theory agrees within 12 per cent, of the amount actually used. I have found a similar agreement in calculating the results of the eighty heats mentioned in Section Xlld in the discussion of the Bertrand Thiel process. The average heat contained 27.140 326 METALLURGY OF IRON AND STEEL. pounds of pig-iron, nearly all of which was charged in a molten state. The average amount of ore used was 7466 pounds, corre- sponding to an addition of 616 pounds to the ton. But it is neces- sary to note that the pig-iron used in the Bertrand Thiel process at Kladno was of the following composition in per cent. : C 3.5 P 1.5 Si 1.0 Mn 0.4 Such an iron will demand 24 per cent, more oxygen than an iron containing 1.0 per cent. Si, 3.75 per cent. C, and 0.6 per cent. Mn, and it should also be noted that in the Bertrand Thiel process much oxygen is supplied by the flame as it fuses and oxidizes the scrap in the secondary furnace, while some oxygen is furnished by the limestone. I find also a close agreement in the records published by Mr. Talbot for his process. The six heats given by him are not con- secutive, but it will be found that the composition of the metal before the first addition of pig-iron and the composition after the last addition were very similar, as shown by the following averages : c. P. Mn. First metal 06 .030 .10 Last metal 13 .035 .15 It would seem fair, therefore, to add together the amounts of pig- iron and ore for the six heats, and since these additions were of nearly uniform weight, to average the figures showing the chemical composition. The results thus obtained are given in Table XII-H, all estimated figures being enclosed in parentheses: TABLE XII-H. Oxygen used in the Talbot Furnace. Total pig iron in six heats 212,100 pounds. Average composition \ ^*;g g J-g Additions. Pounds. Per cent, metallic iron. Pounds free oxygen. Scale 22 400 74 5 4 768 Ore 15 100 58 3 754 Cinder Manganese ore... Limestone 13,800 2.500 23,240 66.8 (20.0) 2,634 620 2,700 Total 14 476 METHODS OF MANUFACTURE, AND COST. 327 The above figures show that the additions of ore and limestone account for 14,476 pounds of oxygen. This assumes that the car- bonic acid set free by the decomposition of the limestone is broken up when in contact with melted pig-iron and that one atom of oxygen is set free. The amount of silica present can be found approximately as shown in Table XII-I. TABLE XII-I. Silica in the Talbot Furnace; Si0 2 Per cent. SiO. Pounds. Scale 22 400 50 112 Ore 15 100 3 00 453 Cinder 13 800 8 00 1 104 Manganese ore Limestone 2.500 23,240 (8.00) (1 00) (200. (232) From roof and walls 2 H OD 'g^S Hi -^ ^rt^ 3 111 O fe, ffii 3 , an ne-t bon tion a s, aken on The carb ples ots were rolled in crop end, B is gregat nto of Se E T ition nt. Com pe ui < 8J9 H.up qoiqM. Com pe UI 1U931B} 9J9AV S3UJ IllJtp qoiqM. tion nt. Ut '. eaa in.ipu.oiq.Yi 33 C S3 3 3 s a 3 3 -0-3 3 a 3 3 11 fl fl 3 3 -o-d fi S3 3 3 oo S2S 8^3 11 sss. ii Ill i \M\N\C* CO CO CO' 111 jo saqom }o3ui jo o/tg joquinu found in ordinary plate ingots, but this assumption is hardly sus- tained by Table XIII-D, which gives the results obtained by drill- ing into the axial line of slabs rolled from larg ingots, made by SEGREGATION AND HOMOGENEITY. 347 The Pennsylvania Steel Works. The points just below the top crop end, and one-third way down the ingot, are assumed to include the most contaminated region. The concentration shown in these cases probably marks the extent of the action of simple crystalliza- tion, while more extreme cases would represent the liquation of small quantities of fusible impure compounds. The content of carbon is not given, for a color determination is worthless when an accurate comparison is to be made, while in the present case the probability of error is unusually great, since the condition of the carbon will not be alike in the center and on the outside of a slab, owing to the difference in the rate of cooling. On the other hand, the estimation by combustion is so tedious that it is not always practicable to make such a large number of analyses. SEC. XII Id. Attainment of homogeneity in plates. The fact that plates are not homogeneous when rolled from ordinary ingots does not become evident under the ordinary systems of inspection, since, as a general thing, only one piece is taken from the sheet, and this comes from the edge, but it will be shown by Table XIII-E that the variations are by no means unimportant. The first in- stance is taken from Pourcel,* the next three are from Cunning- h?m,t while the last two are from my own investigations. The data on heat 11,393 were obtained by rolling an ingot on a universal mill into a long plate. The upper third of this plate was sheared into 16-inch lengths, and tests taken along the center line and the edge. A strip was also cut from the bottom end of the plate in the center and on the edge. The tests of heat 10,768 were cut from a "pitted" plate. The flaws in the bars render worthless any records of elongation, but the chemical results are valuable, while the determinations of ten- sile strength are probably approximately correct. The ingot was rolled on a shear mill to a thickness of three-quarter inch. The plate was only 112 inches long after trimming, so that the seven tests represent the entire length of the sheet. A great deal of this irregularity between different parts of the same plate may be avoided by rolling from a slab as described in the previous section. It would, of course, be untrue to say that segregation can be avoided by making a larger ingot, or that it can be counteracted by a greater amount of work upon the steel, but it * Loc. Cit. t Trans'. A. 1. M. E., XXIII, p. 626, et seq. 348 METALLURGY OF IRON AND STEEL. is nevertheless true that a slab will usually give a much more uni- form plate. TABLE XIII-E. Physical and Chemical Properties of Different Portions of Plates Boiled from Ordinary Plate Ingots. Heat No. Part of ingot correspond- ing to the place from which test was taken. Ultimate strength ; Ibs. per sq. inch. sa gs 1-9 S MOO P< c 6 ***. I!!! $3 3 Composition; per cent. Author- ity. C. P. S. Not given. TOP S Bottom JJ ^e 65426 66848 59636 59310 82.0 27.0 33.0 32.5 55.9 58.6 58.7 55.0 57.9 48.1 .94 .32 .25 .25 .050 .100 .060 .060 .025 .061 .028 .022 Pourcel. iter ' rf) iter Not given. TOP c e e d j Middle JJ Bottom j^ d | rQ 53600 53000 52600 55900 55300 60200 30.7 32.0 28.2 28.5 31.5 24.5 .15 .17 .15 .16 .16 .16 .021 .023 .018 .022 .019 .024 . . . C'nning* ham. iter '6 iter ye iter Not given. Top, edge 75900 69700 64200 65700 65000 63700 66600 61400 66600 64600 9.5 20.0 25.0 25.0 27.0 25.5 23.8 26.0 24.0 23.8 - .22 .20 .18 .19 .21 .19 .20 .17 .19 .19 .064 .058 .034 .043 .036 .038 .089 .030 .040 .040 . . . C'nning* ham. Second piece Third piece, Fourth piece Fifth piece, e Sixth piece, Seventh piec Eighth piece Ninth piece, Bottom . . ,edge . . . 3dge . ... , edge . . . dge . ... 3dge . ... e, edge . . . edge ^dge Not given. Edge 59200 66600 67100 66500 22.5 24.5 23.0 20.0 60.8 59.1 54.7 52.0 .08 .08 .09 .09 J077 .151 .141 .153 .040 .063 .085 .085 C'nning- ham. 4 inches froir 8 inches froir Center L edge Ledge 11893 Preliminary Top Second test Third test Fourth test Fifth test Sixth test; from top of Bottom test . . . 56000 61600 65420 63360 61490 62020 60330 59860 59460 58940 59160 59320 58920 54660 63850 ' 28.75 25.00 27.00 27.00 25.25 28.0 26.50 29.50 28.50 27.50 27.00 28.75 84.75 29.00 ' 45.9 ' 44.6 45.8 44.3 38.6 53.7 45.8 52.5 49.9 52.0 47.5 51.2 66.4 61.0 .077 .128 .087 .110 .107 .110 .109 .098 .098 .098 .096 .097 .097 .073 .070 .045 .078 .082 .068 .063 .068 .064 .056 .045 .056 .057 .055 .042 .033 .031 Author* edge . ... center . . . edge . ... center . . . edge . ... center . . . edge . ... center . . . edge . ... center .... % way ( edge ingot \ center edge . . . center .... 10768 Preliminary Top Second test Third test Fourth test Fifth test Sixth test Bottom test edge .... center .... edge center .... edge 65600 62180 63840 61140 62900 56090 61280 63480 60620 53400 61420 56920 61000 56220 60220 .059 .088 .095 .075 .083 .051 .081 .051 .084 .051 .090 .062 .080 .065 .075 .049 .057 .058 .048 .045 .031 .045 .033 .050 .032 .051 .038 .043 .042 .038 Author. .... .... center .... edge center ... edge .... center .... edge ..... .... center .... edge center .... .... .... SEGREGATION AND HOMOGENEITY. 349 This will be shown by Table XIII-F, which gives the results obtained by testing the edge and the middle of several universal- mill plates which were made from slabs from the same ingot. A careful record was kept of the position of each, slab, and the tests" were cut from the top end of each plate. Thus the list of tests from the successive plates gives the same information as if one long slab had been rolled into one plate and had then been cut up for testing. The segregation in the central axis is shown by a slightly higher content of metalloids, and by a higher tensile strength, but the variations between parts of the same plate, and the variations TABLE XIII-F. Physical and Chemical Properties of Different Portions of Open- Hearth Universal Mill Plates, Rolled by the Central Iron Works from Pennsylvania Steel Company Slabs. NOTE. Plate No. 1 represents the bottom of the ingot, the others being numbered consecutively toward the top. Heat No. 1 ft M 6 fe Part of plate. Elastic limit; pounds per square inch. Ultimate strength; pounds per square inch. Elongation in 8 inches; per cent. Reduction of area; per cent. Composition; per cent P. S Mn. 2905 Acid. 1 Edge, Middle, 33030 35880 54040 55000 29.50 27.50 59.1 61.8 C66 .074 040 .040 -89 38 2 Edge, Middle, 33240 34870 54000 55540 29.50 2900 688 61.3 ,068 074 044 039 86 .37 3 Edge, Middle, 82570 34670 53220 65420 3100 80.50 62.5 621 .068 .074 .040 040 .37 .36 4 Edge, Middle, aS430 35240 53400 56450 31.25 ' 30.50 60.6 58.4 .054 074 040 .045 .37 35 5 Edge, Middle, 33270 846(30 54080 56840 3075 33.00 60.7 67.1 080 -088 .047 052 .86 85 6 Edge, Middle, 33520 85090 54380 57380 31.00 29.25 57.3 567 077 087 .05( .048 .87 .88 7 Edge, Middle, 33150 85110 54120 58180 2925 2625 59.5 66.2 .071 083 046 060 .36 .86 9765 Basic. 1 Edge, Middle, 34050 31900 55360 54440 29.50 31.50 63.2 64.2 .007 .007 038 .032 45 43 2 Edge, Middle, 83580 32460 55350 53780 30.50 31.75 692 63.6 008 007 .045 031 .040 035 45 43 .45 43 3 Edge, Middle, ,33210 33170 56340 55240 2875 3250 57.8 63.1 007 008 4 Edge, Middle, 33580 32550 56580 56020 80.50 30.25 665 604 .007 008 .036 036 .45 .43 5 Edge, Middle, 33580 32800 56340 57240 28.75 80.00 582 58ft 007 .008 042 .040 ,46 .44 350 METALLURGY OF IRON AND STEEL. between different plates, are much less than is shown in Table XIII-E -for plates rolled directly from ingots. The usual way of testing is to take a strip from a corner of the plate, and Table XIII-G- gives the records so obtained from one- quarter-inch sheets, which were rolled from basic open-hearth slabs made by The Pennsylvania Steel Company. The ingots from which the slabs were made varied in section from 26"x24" to 38"x32", and weighed from 6 to 10 tons each. A record was kept of the part of the ingot from which each slab came, and the corre- sponding plates were tested both in the natural and in the annealed states. The table gives only the results on the annealed bars, for by the reheating and cooling the artificial effects of cold finishing were avoided, .and all the test-pieces were brought to a common ground of comparison. The plates of any one heat are all of one thickness^ the discard of other sizes accounting for the many missing mem- bers. In each case the order in the list follows the order in the TABLE XIII-G. Physical and Chemical Properties of Annealed Bars cut from Plates Eolled from Basic Open-Hearth Slabs, which were cut from different parts of 10-Ton Ingots. NOTE. Carbon was determined by color and is therefore unreliable. Thickness of plate. : Part of ingot from which slab was cut. Ult. strength; pounds per ' square inch. Elastic limit; pounds per square inch. oo "3 S 'M 8>J! d ofl H"* 4J t% fi *"* * 0? cd Chemical composition; per cent. C. P. Mn. S. inch. 1st ingot. Top, Bottom, 49080 48330 47750 48500 47810 46970 48200 81830 31170 29980 81760 31110 80690 81000 36.75 82.00 84.50 29.50 83.00 35.00 82.50 65.3 63.6 67.0 66.3 68.1 64.5 64.3 .11 .15 .16 .13 .12 .12 .11 .015 .018 .015 .013 .015 .015 .017 .31 .32 .82 .81 .31 .31 .027 .020 .022 .023 .023 .019 .025 Average, 48091 81077 83.82 65.6 .13 .015 .81 .023 ^ I _rt 3 Top, Bottom, 49380 48010 48760 49170 49040 47670 46860 32080 28760 82030 82010 29940 30090 81380 33.00 83.00 33.75 32.00 81.75 33.00 32.50 64.2 65.7 64.9 64.2 60.7 63.8 65.3 .10 .16 .13 .13 .12 .14 .11 .016 .018 .018 .015 .014 .013 ^013 .31 .35 .31 .32 .31 .34 .32 .025 .023 .026 .024 .025 .019 .021 Average, 48413 I 80899 1 32.71 64.1 .13 .015 .82 .028 SEGREGATION AND HOMOGENEITY. 351 TABLE XIII-G Continued. 1 a s M a Part of ingot from which slab was cut. C . Chemical composition; percent. Ult. strengtl pounds pei square incl Elastic limil pounds pei square inc i Is 1- 5 ^ 5~ | "8 'of %Z ti EH C. P. Mn. S. inch. 1st ingot. Top, Bottom, 51040 51660 51620 51760 51200 50470 50260 50820 32710 asoso 82180 32230 31730 32310 83340 32820 81.00 30.50 33.00 32.50 31.50 32.75 32.50 33.00 63.8 64.1 62.8 63.3 61.1 61.8 62.6 62.1 .13 .12 .13 .14 .13 .12 .10 .10 .014 .014 .018 .on .017 .006 .012 .016 .48 .46 .42 .44 .41 .45 .45 .47 .014 .021 .025 .024 .024 .028 .020 .021 Average, 51104 82488 32.00 62.7 .12 .014 .45 022 X 3 1 Top, Bottom, 52160 52050 52240 50600 50820 50360 50530 49880 32450 31330 82940 33020 32240 32470 32240 31850 82.00 82.00 33.00 31.00 32.25 82.50 82.75 34.50 57.0 60.7 62.6 61.0 61.2 63.5 60.0 62.8 .14 .12 .11 .11 .12 .13 .12 .11 .009 .017 .018 .013 .014 .005 .018 .012 .45 .46 .47 .46 .44 .45 .46 .46 .025 .024 .023 .016 .022 .023 .016 .016 Average, 51080 32318 32.50 61.1 .12 .013 .46 .021 ch. 1st ingot. Top, Bottom, Average, 52620 52210 50940 50360 50000 3J860 86130 31780 30590 81840 81.00 82.50 32.00 28.75 31.50 60.2 65.0 65.7 60.0 56.4 .16 .16 .14 .15 .14 .019 .019 .016 .019 .016 .44 .43 .44 .44 .44 .032 .032 .028 .029 .025 51226 82640 31.15 61.5 .15 .018 .44 .029 a 5 *i a S Top, Bottom, 51880 53060 52820 52970 52870 50860 50000 50950 86380 30660 85450 32540 31340 80070 83730 31280 32.50 28.75 27.00 31.25 31.75 32.50 85.00 35.50 63.5 55.2 62.2 58.9 57.9 61.4 62.7 64.5 .14 .15 .16 .15 .15 .15 .14 .14 .017 .024 .021 .017 .019 .019 .017 .016 .42 .44 .44 .44 .44 .44 .44 .44 .029 .033 .031 i .080 .032 .029 .029 .025 Average, 51926 32681 31.73 61.0 .15 .019 .44 .030 All X inch. 2d ingot. 1st ingot. Top, Bottom, Average, 54160 53840 54460 51200 53000 51740 52420 53020 38230 38210 38070 35500 88370 37310 87200 87600 26.00 27.25 28.25 31.00 30.50 31.00 27.50 31.25 61.5 60.1 61.4 64.0 60.9 64.9 65.2 66.3 .13 .13 .12 .13 .12 .11 .11 .12 .039 .033 .038 .023 .031 .031 .030 .033 .32 .28 .32 .28 .37 .81 .29 .29 .050 .058 .060 -028 .051 .047 .046 .050 52980 37561 29.09 63.0 .12 .032 .31 .048 Top, Bottom, Average, 54070 54130 51520 52520 52980 88520 88350 36090 38130 37770 27.50 30.25 26.00 30.25 31.00 64.4 63.8 65.6 63.8 66.0 .12 .13 .13 .11 .12 .036 .037 .036 .031 .031 .34 .31 .31 .31 .29 .058 .053 .057 .048 .044 53044 37772 29.00 64.7 .12 .034 .31 .032 & i Top, Bottom, Average, 54850 54480 53960 53580 53130 87830 36560 88520 37860 37260 30.00 28.75 29.50 28.75 25.75 61.9 63.8 63.3 63.8 54.3 .13 .13 .12 .12 .12 .037 .035 .034 .033 .031 .26 .30 .32 .32 .82 .050 .048 .047 ..045 .047 54000 37006 28.55 61.4 .12 ! .034 .80 .049 352 METALLURGY OF IRON AND STEEL. TABLE XIII-G Continued. Heat No. Thick, of plate. Part of ingot from which slab was cut. Ult. strength; pounds per square inch. Elastic limit; pounds per square inch. Elongation in 8 in.; percent. Reduction of area; perct. Chemical composition; per cent. C. P. Mn. | 8. | 8238. 50-ton heat. All 1/5 inch. 1 M Top, Bottom, Average, 50:270 51630 49180 50240 53520 86880 38510 35130 38090 36150 31.75 32.00 30.75 29.25 31.00 60.3 64.0 58.3 59.2 63.4 .12 .11 .11 .11 .13 .027 .023 .019 .024 .014 .35 .36 .37 .36 .43 .082 .027 .027 .030 .081 50908 38152 30.95 61.0 .12 .021 .87 .029 i Top, Bottom, Average, 53010 53620 51520 50400 52730 37030 39140 34270 37330 36810 27.50 25.75 25.50 24.75 28.50 61.1 61.1 58.0 56.1 58.0 .12 .13 .11 .12 .13 .027 .027 .021 .025 .022 -33 .38 .38 .40 .38 .039 .035 .('32 .028 .031 52256 88916 20.40 58.9 | .12 .024 .37 .033 a i Top, Bottom, Average, 52610 51540 52760 52550 51480 36970 35700 86940 37040 40480 31.25 27.00 33.00 32.00 28.75 60.4 61.5 65.0 62.3 56.0 .13 .12 .11 .11 .11 .034 .030 .026 .028 .020 .38 .37 .37 .36 .39 .040 .033 .033 .028 .028 52188 37426 30.40 61.0 .12 .028 .37 .032 | 8234. 50-ton heat. 1 B9 Top, Bottom, Average, 58080 55580 54820 54280 54360 3.5890 34920 34450 35320 34400 30.00 28.00 31.25 31.25 30.50 60.0 59.0 62.0 63.0 62.2 .19 .14 .13 .14 .17 .025 .019 .019 .023 .022 .48 .46 .46 .46 .47 .080 .024 .023 .025 .021 55024 34996 30.20 61.2 .15 .022 .47 .025 1 Top, Bottom, Average, 55680 55210 54120 53200 54180 35380 34580 35950 34460 34700 31.50 29.50 31.25 31.25 31.75 59.2 62.3 61.2 62.7 60.9 .11 .12 .14 .12 .13 .024 .024 .021 .020 .021 .49 .48 .47 .46 .49 .027 .027 ,C26 .020 .021 54478 35014 31.05 61.3 .12 .022 .48 .024 3 fcfl 55 ~ 4 Top, Bottom, Average, 54000 55120 54180 53940 53400 85440 36310 35060 34460 33590 31.50 29.50 30.75 30.00 31.25 62.8 63.8 62.9 65.4 63.6 .14 .13 ,17 .14 .15 .020 .025 .024 .019 .019 .46 .48 .45 .46 .46 .021 .027 .028 .022 .020 54128 34972 30.60 63.7 .15 .021 .46 .024 Top, Bottom, Average, 55120 54280 53980 52720 54720 34300 34940 a5230 33400 34340 30.50 29.50 28.00 32.50 81.75 62.6 61.9 63.3 63.8 63.2 .16 .15 .18 .14 .14 .021 .024 .022 .021 .023 .47 .47 .54 .46 .46 .027 .025 .041 .024 .025 54164 34442 30.45 63.0 | .14 .022 .48 .028 Top, Bottom, Average. 53970 54640 53590 35710 34410 33210 30.25 33.00 32.00 65.3 63.9 64.9 .16 .16 .12 .023 .021 .019 .48 .47 .46 .024 .024 .021 54067 34443 31.75 64.7 62.6 64.6 60.C .15 .021 .47 .023 Top, Bottom, Average 53550 54550 55560 35420 36180 87360 31.75 32.00 28.25 .15 .12 .15 .022 .021 .024 .48 .49 .023 .026 .022 54553 36320 30.67 62.4 .14 .022 .49 .024 SEGREGATION AND HOMOGENEITY. 353 TABLE XIII-G Continued. "Thick"." of plate. Part of ingot from which slab was cut. Ult. strength; pounds per square inch. Elastic limit; pounds per square inch. Elongation in 8 in.; percent. Reduction of area; perct. Chemical composition; per cent. C. P. Mn. 8. 1 All y inch. I t Top, Bottom, Average, 49880 49150 48190 48190 29740 29680 30030 30270 31.75 33.00 33.00 30.25 58.5 63.5 57.1 60.8 .11 .10 .11 .11 .017 .017 .016 .016 .32 .85 .26 .35 .040 .041 .033 .043 48853 29930 32.00 60.0 .11 .017 .32 .039 4J S a- i Top, Bottom, Average, 50480 49030 47740 48310 28570 31880 29930 30430 30.75 33.75 33.25 33.00 61.0 62.6 63.9' 64.7 .13 .12 .10 .11 .019 .018 .017 .010 .83 .33 .88 .31 .043 .038 .085 .086 48890 30203 32.69 63.1 .12 .018 .33 .038 y *s> "* n Top, Bottom, Average, 49630 48910 30410 30510 30.00 30.50 64.0 63.0 .11 .10 .017 .017 .36 .35 .024 .083 49270 30460 30.25 63.5 .10 .017 .36 .029 Top, Bottom, Average, 48440 47600 47260 30460 80530 29850 32.00 34.00 31.25 65.9 57.2 58.0 .10 .11 .13 .019 .017 .016 .32 .35 .35 .036 .086 .034 47767 30280 32.42 60.4 .11 .017 .34 .085 A 5 < +j -u o ^ 5) C3 Top, Bottom, Average, 50660 50860 53860 32710 30480 33710 35.00 33.25 29.25 64.7 63.8 58.6 .13 .13 .11 .017 .021 .025 .45 .44 . .46 .022 .028 .087 51793 32300 32.50 62.4 .12 .021 .45 .029 I i Top, Bottom, Average, 54080 52680 51520 50750 50280 33970 34100 32140 32840 31760 80.00 31.25 33.00 33.25 31.75 59.4 63.9 61.0 64.2 65.2 .15 .15 .12 .14 .13 .024 .022 .018 .020 .013 .46 .46 .44 .44 .43 .081 .029 .026 .023 .022 51802 32962 31.85 62.7 .14 .019 .45 .02G i a s Top, Bottom, Average, 58440 51(320 505CO 492CO 32440 33400 82650 31460 32.50 32.75 31.25 31.00 60.7 65.1 61.9 65.0 .11 .13 .14 .15 .024 .019 .021 .020 .42 .42 .42 .41 .030 .029 .027 .026 51245 32488 31.88 63.2 .13 .021 .42 .028 1) fl ** Top, Bottom, Average, 52060 54260 52880 50890 32460 34450 33450 82090 31.75 30.00 29.50 33.75 64.2 59.4 62.8 61.4 .15 .17 .14 .10 .028 .026 .024 .018 .44 .44 .45 .42 .030 .028 .030 .029 52523 83113 31.25 62.0 .14 .024 .44 .029 ingot from top to bottom, and it will be seen that, as a rule, the plates from the top give a slightly higher strength than those from the bottom, but that the variations are unimportant, not being as great as will often be found in different parts of a single plate rolled from an ordinary plate ingot. The carbon determinations in Table XIII-G are inaccurate, since they were made by the color method. The work was performed by 354 METALLURGY OF IRON AND STEEL. rnen who are regularly engaged in doing nothing else, and without any attempt at extra care, but in order to see whether there really were any such differences in composition as the records would indicate, the samples showing the widest variations in three heats were reworked twice by color and once by combustion ; the results are given in Table XIII-H,. and show that the variations in any one heat are in the third place from the decimal point, which is close to the limit of experimental error. TABLE XIII-H. Showing that Variations in the Carbon Content in the Test-Pieces Given in Table XIII-Gr are Due to Analytical Errors. . - ' Group A is made up of B of those showing the lowest. Heat No. Group. Composition ; per cent. Original as given in Table XIII-G. Reworked. Carbon by color. P. Mn. Duplicate determi- nations by color. Average of group by combustion. 6633 A .15 .16 .018 .015 .32 .32 .13 .13 .14 .13 .118 B .11 .015 .31 .13 .14 .124 8284 A .19 .17 .17 .025 .022 .024 .48 .47 .45 .18 .17 .15 .19 .18 .16 .165 B .11 .12 .12 .024 .024 .020 .49 .48 .46 .17 .15 .16 .17 .16 .17 .158 8286 A .15 .17 .028 .026 .44 .44 .14 .14 .14 .15 .150 B .11 .10 .024 .018 .42 .42 .13 .14 .13 .14 .149 SEC. Xllle. Homogeneity of acid open-hearth rivet and angle steel. A very good opportunity of investigating the homogeneity of a heat of steel occurs in the manufacture of rivet rods and angles, where tests may be conveniently taken from many differ- ent members. In the case of rivet rods, the test-pieces will repre- sent the ent'ire cross section of the ingot, and thus include the segregated portions. Table XIII-I gives the records obtained from several tests taken at random from the piles of rivet rods from five different heats, without any knowledge as to what part of the heat or what part of the ingot the tests came from. SEGREGATION AND HOMOGENEITY. 355 The data on the natural bars are arranged in the order of tensile strength, while in parallel columns are given the results obtained by annealing the same bar. Although all the pieces of one heat were annealed at the same time, and with the utmost care to have all conditions uniform, it will be seen that the variations in the strength of the treated bar are entirely independent of the vari- ations in the strength of the natural bar. This would indicate that the differences are due to irregularities in rolling and to de- terminative errors rather than to any inherent variations in the character of the metal. TABLE XIII-I. Tests on "Rivet Rounds taken from Different Parts of the Same Heats. All steels were made by The Pennsylvania Steel Co. ' m Ultimate Elastic limit; Elongation in Reduction of is * P. strength; pounds pounds per 8 inches ; per area; per |-g' I per square inch. square inch. cent. cent. * |l . i . d . i . i 43 ! 0^; 1 1 g 1 1 1 1 1 .S " l=s Is a ^ 3 O d "S W d 1 1 5 o o 1 1 fc 1 fc 5 1 d 61260 55640 43960 34420 31.25 30.00 60.30 66.24 _ . "~ 60950 54760 42430 84840 32.00 29.50 62.73 65.91 S -3 25 60800 52700 42790 33800 32.X 31.50 65.25 68.88 . II P 60720 55130 43600 34700 31.25 32.50 66.76 67.87 1 O 4^ [[5 d 60210 54600 41160 84040 30.50 30.75 62.60 65.68 s; d 7? 60010 54360 41720 34040 30.50 32.50 6(5.76 67.74 ?]! 59970 54820 40770 33840 30.50 32.00 63.97 C4.92 *?. H&-!Q 59710 54340 40900 34320 32.50 32.50 63.43 63.78 2s if If 59620 54040 40920 34120 33.00 30.00 57.70 66.39 J 59300 54600 40320 34030 34.50 33.00 65.96 68.05 Avei age, 60260 54500 41860 34220 31.80 31.42 63.55 66.54 ^ H & 56040 49990 37710 30200 33.25 34.75 65.73 66.70 $H ^ CC || 56000 50520 37800 30700 35.00 34.i5 64.26 69.18 fl AT S 55520 50520 37890 31750 31.50 35.75 61.86 60:94 43 3 43* S . 55420 51000 37360 81165. 81.75 34.50 62.18 67.97 S jro" 55080 49460 36130 30910 83.00 34.75 56.03 68.70 m ^ Q!I n" 55040 51170 37980 31475 34.75 34.50 65.48 69.50 \00 ji 54980 50400 37710 30665 33.00 35.50 59.64 69.68 _j o t>\ cT5 54950 50640 37800 31345 31.75 35.00 67.02 69.28 2 ^ "~i^ 54860 50520 37980 31970 33.00 35.00 64.09 68.04 3 rH O^" 54720 50940 36830 31900 83.75 35.75 55.25 67.85 A\ er age,* 55260 50520 87520 31210 33.07 34.97 62.15 68.38 4J II 54000 48870 36230 30990 83.75 33.75 62.30 70.59 3D . 53500 494GO 35960 31220 34.50 36.25 68.32 68.27 08 ."O c^ . 53400 48520 35710 31520 83.50 35.50 64.05 69.28 43 fl J o iM 52690 48290 35880 31190 33.75 32.50 66.49 68.77 ' O o if C3 *" 53300 48460 36060 31370 83.75 34.25 61.57 68.14 a d oiS " 52620 49760 35080 82710 33.75 36.25 68.27 67.52 0, 8 iff ^ -r? 52620 48640 80490 33.75 36.25 65.29 69.43 O O_ >\ gig^ 52620 48520 35950 30590 81.25 35.00 62.04 69.49 2 ~* .& 51910 49230 36230 32580 82.25 34.50 58.68 67.98 if II ' 51900 48410 34840 80350 33.75 83.75 63.72 66.V>5 A\ r erage, 52860 48820 85700 1 31300 33.40 34.80 68.57 68.64 356 METALLURGY OF IRON AND STEEL. TABLE XIII-I Continued. Acid open-hearth. | Kind of steel. > 110,000 pounds. | Weight of charge. Diameter of bar. Composition; per cent. Ultimate strength; pounds per square inch. Elastic limit; pounds per square inch. Elongation in 8 inches; per cent. Reduction of area; per cent. Natural.- Annealed. Natural. Annealed. Natural. Annealed. Natural. Annealed. .- '= 5 A .s o - 5 t'l L s& 3IJ& Jis .--O 3^ 3 v'r'ge l;l 5 3H iSi 33 P age, 55480 55480 55430 55400 55160 54770 54750 54690 54520 54220 49460 49940 49460 49700 49700 50720 50420 50010 50880 49770 37600 36670 38400 37250 37950 37600 88400 39120 38640 38900 29870 30350 30110 31300 32730 31760 32740 32470 32220 32230 32.50 32.50 83.25 30.00- 30.00 32.50 83.75 32.75 33.00 83.75 28.75 31.25 31.75 35.00 82.50 30.00 33.75 32.50 31.25 33.75 67.45 68.22 68.40 64.67 64.97 69.68 63.15 67.35 67.17 66.57 65.80 65.30 67.70 69.22 60.22 64.62 61.12 67.96 67.25 69.47 54990 50006 38053 31578 32.40 32.05 66.76 66.72 55000 54780 54700 54180 54170 53880 53770 53770 52860 52600 50230 49170 50880 48820 48290 48930 50520 49060 50160 50640 37710 37100 86750 87450 36580 36320 35610 85960 85700 35360 81950 30310 81623 30840 80840 30730 31670 81120 81920 32400 31.75 33.75 82.50 81.75 31.25 31.00 82.50 82.75 83.25 33.00 33.75 36.00 34.00 85.00 34.00 86.00 34.00 35.50 35.50 33.00 66.31 62.83 60.11 62.30 67.83 60.20 60.02 65.73 61.39 68.49 70.77 68.77 66.70 68.77 68.77 68.43 65.08 69.76 69.57 68.62 53970 49670 36450 31340 32.35 8*i87 63.52 68.52 Basic open-hearth. % 40,000 pounds. 1 2 y s inch. Si l& O age, 48340 47380 48450 48230 49175 48560 47730 48785 48640 49440 47835 48050 48360 48400 33065 31530 83650 31600 83340 32760 83260 82130 82935 83270 82900 81920 82185 83880 84.50 85.00 85.00 87.00 86.25 33.75 35.00 34.00 84.25 84.00 34.00 33.75 86.25 33.75 ! ! ! ! 71.87 72.05 72.05 74.14 70.09 72.95 74.49 71.80 71.92 71.48 72.72 71.42 74.28 73.64 ! ! ! ! ! .' 48384 32745 34.75 72.49 In further proof of this, drillings were taken from the three annealed bars of heat ' 10,168, which showed the highest tensile strength, and from the three which were the weakest. The results of analysis are given in Table XIII-J. The ingots from which these rivet rods were made measured 16"x20" in cross section and weighed about two tons each. In the case of angles it is the practice at The Pennsylvania Steel Works to roll a larger ingot than is used elsewhere for the same purpose, the cross section being 24"x26", and the weight about 5 tons. In order to test tke uniformity of the material, the blooms from sev- SEGREGATION AND HOMOGENEITY. 367 eral such ingots were stamped so as to denote from what part of the ingot each one came, and drillings were taken from the corre- sponding finished angles. TABLE XIII-J. Composition of Kivet Eods from Heat 10,168.,' which showed the Greatest Differences in the Tensile Strength of the Annealed Bars. Nature of Sample. Ultimate strength ; pounds per sq. inch. Composition; percent. Natural. Annealed. C. P. S. Mn. Preliminary test 52280 53690 54077 50680 48680 .12 .12 .12 .013 .013 .013 .024 .019 .024 .29 .30 .30 Average of strongest three bars of % inch diameter . . . Average of weakest three bars of % inch diameter . . . The results of analysis are given in Table XIII-K, and they show that each ingot was practically uniform throughout. The drillings were taken so -as to include the center of the bar, which is the most impure portion. In each case the first bloom in the list is the top of the ingot, and the last is the bottom; the varying number of blooms in the ingots arises from the different weight of the angles to be made. SEC. XHIf. Homogeneity of high-carbon steels. It would nat- urally be expected that segregation would be most marked in ingots of high-carbon, because such metal remains liquid for a long time. It is found, however, that even under these conditions separation of the impurities does not always occur. This will be shown by Tables XIII-L and XIII-M, which give the results of certain in- vestigations by The Pennsylvania Steel Company. The data on" carbon in Table XIII-L are of little importance, for a color deter- mination is well-nigh worthless on such high steels. The determinations of carbon in Table XIII-M are made by com- bustion and are accurate, and they show a considerable variation in the distribution of this element ; this might be expected when such a large proportion is present, and its hold upon the iron corre- spondingly less firm. The sulphur and phosphorus are very regu- lar, the variations in the purer metal being almost within the limits of error. In the ingot of medium phosphorus, the percent- age of variation is no more than in the others, but the actual range 358 METALLURGY OF IRON AND STEEL. is greater. Although this would follow naturally, it is possible to show, by an incident which happened under my own observation, that concentration does not always occur,, even in the case of im- pure steels. f&i. s S &s ^ ^^"^ pC "S t^ fl\ >^^ 5? 1 S sJ g^ w i- a 33 rt 18 f* O oqqqqq jo -OK I? 32 i? 32 32 ' qqqqq< i -H O rH O (M O (M : rn (M * 10 O t rH (M W * U3 O t- 00 Oi O jo -OK I! i- ,, > ^-n ^ 4^ CO -C ^ K qqqqqqqc uaqranu A 50-ton acid open-hearth charge had been made containing .46 per cent, of carbon, together with unusually high manganese, phos- phorus, and silicon. The ingots had a cross section of 16"x20", and weighed about 4000 pounds each. In loading them, one fell SEGREGATION AND HOMOGENEITY. 359 over and "bled" at the top. The amount of liquid metal thus lost did not exceed 25 pounds, although the cavity was completely emptied, so that if segregation existed to any considerable extent it should appear in this metal which remained liquid to the last. Table XIII-N will, however, show that very little segregation had taken place. TABLE XIII-L. Distribution of Elements in a High-Carbon, Low-Phosphorus Open-Hearth Ingot, 14 inches square, 63 inches long. NOTE. Made by The Pennsylvania Steel Company. Carbon was determined by color, and is, therefore, only approximate. Part of the ingot from which test was taken. fe ^fl^aS 5 H m ,d 2fiU|q fraaa Composition; percent. C. P. Mn. Average. C. P. Mn. Four inches from bottom, 2 4 6 7 .79 .78 .79 .72 .013 .015 .013 .012 .20 .20 .19 .19 .77 .018 .20 Fifteen inches from bottom, 2 4 6 7 .77 .87 .84 .78 .011 .015 .011 .011 .20 .20 .20 .19 .81 .012 .20 Twenty-six inches from bottom, 2 4 6 7 .80 .89 .85 .81 .012 .014 .014 .009 .18 .21 .21 .20 .84 .012 .20 Thirty-seven inches from bottom, 2 4 6 7 .77 .90 .89 .83 .011 .014 .015 .012 .20 .21 .20 .20 .85 .013 .20 Forty-eight inches from bottom: all above this would be cut off as scrap when the ingot is rolled, 2 4 6 7 .79 .91 .89 .94 .011 .014 .016 .014 .21 .20 .19 .21 .88 .014 .20 Four inches from top, 2 4 6 7 .74 .90 .95 1.06 .010 .016 .017 .023 .21 .21 .21 .21 .91 .016 .21 SEC. Xlllg. Homogeneity of acid open-hearth nickel steel. > It is the current impression among manufacturers of nickel steel that the presence of this element prevents segregation. In order to have some evidence upon this point, an investigation was con- ducted on an ingot of nickel steel made by The Pennsylvania Steel Company. The cross section of the ingot was 18"x20", and the weight about 5500 pounds. This was rolled into a piece 16 inches wide, 5 inches thick, and 20 feet long, and cut into five slabs. 360 METALLURGY OF IRON AND STEEL. The top slab was rolled into a three-eighth-inch universal plate, the second slab into a three-eighth-inch sheared plate, the third slab into a half-inch universal plate, the fourth slab into a half-inch sheared plate, and the fifth slab was hammered into a bloom and then rolled into 6"x4" angles. TABLE XIII-M. Distribution of Elements in 7-inch Square Blooms Eolled from High-Carbon, Open-Hearth Ingots, 14 inches Square. A slice was cut crosswise from the rolled bloom at different places and drillings taken from the center of this slice, corresponding to the center of the ingot. Kind of ingot. Place from which slice was taken. Composition; percent. Cby comb. P. Mn S. Si. .12 .09 .10 .09 .11 .11 Low- phosphorus ingot. Ladle test -... Top of ingot after cutting off 20 per cent, as scrap .984 .941 .990 .991 .982 1.012 .013 .015 .019 .017 .020 .016 .09 .11 .11 .11 .11 .11 .022 .012 .010 .012 .010 .010 One-fourth way down the ingot One-half way down the ingot Three-quarters way down the ingot . . . Bottom of ingot Medium- phosphorus ingot. Ladle test Top of ingot after cutting off 20 per cent, as scrap . . .... 1.440 1.205 1.430 1.443 1.400 1.459 .050 .064 .059 .051 .053 .055 .28 .28 .27 .27 .27 .27 .016 .015 .015 .013 .014 .012 .12 .16 .12 .12 .18 .12 One-fourth way down the ingot One-half way down the ingot Three-quarters way down the ingot . . . Bottom of ingot Low- phosphorus ingot. Ladle test .913 .925 .024 .021 .022 .021 .025 .021 .13 .13 .14 .18 .13 .13 .019 .018 .018 .020 .021 .021 ' Top of ingot after cutting off 20 per cent, as scrap . . One-fourth way down the ingot One-half way down the ingot . .965 .948 .956 .943 Three-quarters way down the ingot . . . Bottom of ingot . TABLE XIII-K Composition of the Liquid Interior of an Ingot as Compared with the Ladle Test of the Same Charge. Origin of sample. Composition; percent. Carbon by combustion P. S. Mn. Si. Metal from interior . .480 .461 .095 .091 .047 .034 0.95 1.18 und.. .12 Ladle test Each end of each slab was marked so as to note whether it was toward the top or bottom of the ingot, and the location of each test- piece in each plate was kept of record. Table XIII-0 gives the physical and chemical results obtained from the ^different strips, SEGREGATION AND HOMOGENEITY. 361 while the diagram immediately below the table represents the entire length of the original piece produced by rolling the 18"x20" ingot to a section of 16"x5". The numbers on this diagram correspond to the numbers of the test-pieces in the table, and serve to mark the exact place in the ingot from which the corresponding test-piece was derived. TABLE XIII-0. Homogeneity of Acid Open-Hearth Mckel Steel. Size of ingot, , 18"x20" ; made by The Pennsylvania Steel Company. Coi preliminary test, per cent.: C, .24; Mn, .78; P, .032; S, .027. Composition ol i i Shape into which slab was rolled. J 03 3 3* g ^ te S (.025) (.040) u 5 C o -^ G "S K Q s .117 .140 6 | .549 .543 S 1.102 1.099 (.019) (.034) (.026) (.025) (.026) (.026) Bottom Bottom. Bottom Top Top. Top. .220 .270 .612 .675 1.234 1.262 i (.022) (.042) g- g (.030) (.034) g g (.028) (.033) ^ D fl H a L a 1 .192 .218 a | .625 .631 g 1.240 1.217 (023) (.030) (.033) (.034) (.031) (.031) Bottom Bottom. Bottom SEC. Xlllh. Investigations on Swedish steel. The experiments related in this chapter were for the most part made at Steelton; manufacturers, as a rule, do not want to discuss segregation at all, and published records are rare. Very recently, however, a careful account has been written by Wahlberg* on certain investigations on Swedish steels. He gives the determinations by three chemists * Journal I. and S. ! Vol. II, 1901. SEGREGATION AND HOMOGENEITY. 363 of the carbon and phosphorus in several different steels and Table XI II-P shows the averages made from his tables. Inspection will show that B, E, G, H, J and L, which is to say one-half of all the ingots, showed practically no segregation of either carbon or phos- phorus. F, I and K showed segregation in the center of the top of both carbon and phosphorus, but none elsewhere. C and D showed segregation in the top and a slight amount in the centre of the bottom,, while A showed quite marked segregation in the top and a very considerable amount in the bottom of both carbon and phos- phorus. It will be evident, however, that by cutting off the top of the ingot the remainder of the steel will be practically uniform, for, as before pointed out, the central axis constitutes but a small portion of the finished material. The burden of the testimony given in this chapter is to the effect that segregation is an ever present factor; that the extent of the concentration bears a certain relation to the proportion of impuri- ties that are present; that manganese, copper and nickel do not segregate to any extent, but that certain portions of the finished material will contain a higher percentage of carbon, phosphorus and sulphur than will be found in the tests cut from the edge of plates and bars, or than will be shown by an analysis of the pre- liminary test. It is also indicated that a degree of uniformity, sufficient for practical needs, may be expected if the initial metal is low in phosphorus and sulphur. CHAPTEK XIV. INFLUENCE OF HOT WORKING ON STEEL. SECTION XlVa. Effect of thickness upon the physical prop- erties. One of the fundamental difficulties in writing specifications is to decide the nature of the test-piece to be required, inasmuch as the strength and ductility will vary in pieces of different thickness, while the results will not be alike in tests cut from different struc- tural shapes, like plates, angles and rounds, even though they be /oiled from the same steel. From one point of view each piece of metal throughout a bridge should be of exactly the same strength per unit ctf section without regard to its thickness; but in taking this as a basis a serious trouble is encountered. Suppose, for in- stance, that 'a metal is required running between 56,000 and 64,000 pounds per square inch, and a charge is made which in three- eighth-inch plate gives 57,000 pounds. If this steel be rolled into seven-eighth-inch angles, or into one-inch plate, or into two-inch rounds, it is quite probable that these will run below the allowable minimum. On the other hand, if the steel gives 62,000 pounds in a preliminary test, the larger sections will give proper results, while one-quarter-inch plate will be too high in ultimate strength. Where a structure is to be made of large quantities of very large or very small sections, it is well to specify that the test shall be made on the special thicknesses needed, but in ordinary cases it seems absurd to the practical mind that a heat of steel can be perfectly suitable for one size and unsuitable for another. It was the custom in the past for inspectors to recognize the situation and make tests from the usual sizes, with a full knowledge that thicker and thinner members would give different results, but in later prac- tice there is a growing tendency to test each separate thickness, a change which has been the cause of great expense to the manufac- turer. Provisions to cover this point should be incorporated into contracts and a certain definite allowance made for variations in the dimensions of the finished material. On the other hand the 364 INFLUENCE OF HOT WORKING ON STEEL. 365 requirements should be worded so that manufacturers would be obliged to put sufficient work on large members to render them of. proper structure. There is often a confusion of terms in considering the effect of work as represented by a large percentage of reduction from the ingot, and the effect of finishing at a low temperature. This is found most often in the case of plates, for it has been quite a gen- eral practice to roll these directly from the ingot in one heat. In order that a piece shall be finished hot enough under this practice, there has been a standing temptation to use a thin ingot; but, on the other hand, it has been almost universally shown that the best results are obtained when a large amount of work is put upon the piece during rolling. SEC. XlVb. Discussion of Riley's investigations on the .effect of work. The truth of this last statement was disputed by Kiley,* who tabulated the results of testing different thicknesses of plate when rolled from ingots of varying section. In all cases the ingot was either hammered or cogged to a slab and this was reheated be- fore finishing into a plate. His analysis of the records consisted in picking out individual cases and showing that the small ingots gave some results which were equal to those from the large ones, but this method of comparison must be recognized as entirely unworthy of the subject. It is true that the number of tests is very small, and it would not be surprising if the accidental variations in the double working should produce anomalous results ; but even taking these very data and making comparisons by the proper system of averages, it will be found that they tell a story exactly opposite from the conclusions formulated by Mr. Eiley. In Tables XIY-A and XIV-B such figures are presented. In the comparison of the different thicknesses in Table XIV- A the thinner plates give much better results, the one-half-inch plate showing an increased ductility in spite of its greater strength. The one-quarter-inch plates are somewhat lower in elongation and two and one-half per cent, better in reduction of area than the one inch plates, but they possess 7600 pounds more, strength, so that less ductility should be expected. This statement is open to criticism, as no account is taken of the effect of variation in the * Some Investigations as to the Effects of Different Methods of Treatment of MiJd Steel in the Manufacture of Plates. Journal I. and 8. I., Vol. I, 1887, 121. 366 METALLURGY OF IRON AND STEEL. dimensions of the test-piece, but Table XIV-B, which is free from this error, proves that the plates made from the large sizes have a higher tensile strength and greater ductility. TABLE XIY-A. Average Physical Eesults on Different Thicknesses of Steel Plates Without Regard to Size of Ingots ; there being an Equal Num- ber of Plates of each Thickness Rolled from Each Sized Ingot.* Thickness of plate. Ultimate strength; Ibs. per square in. Elongation in 8 inches; per cent. Reduction of area; per cent. Annealed, ulti- mate strength ; pounds per square inch. One inch . . . One-half inch . One-quarter in. 62037 64534 69642 24.40 24.71 22.35 40.20 44.85 42.68 59416 61018 62989 TABLE XIY-B. Average Physical Results on Plates from Different- Sized Ingots Without Regard to Thickness of Plate; there being the same Number of each Thickness Rolled from a Given Size.* Size of ingot: in inches. Thickness of slab in inches. Ultimate strength; Ibs. per square inch. Elongation in 8 inches; per cent. Reduction of area; per cent. Annealed ulti- mate strength; pounds per square inch. 24x15 14x14 18x12 18x12 12x6 8 8 8 4 4 66155 65296 65128 65520 64923 24.14 23.91 23.77 23.68 23.68 45.79 44.13 41.38 40.00 41.58 62197 62571 60461 60461 60013 Thus these experiments which were heralded as upsetting current beliefs are found to vindicate them; they do prove that in some cases very good results may be obtained by skillful manipulation under a bad system; but manufacturers have long since learned that a large amount of reduction is essential to secure reliable re- sults in regular practice, and no short series of tests can upset this well-established fact. SEC. XIYc. Amount of work necessary. Up to within a com- paratively recent period it was a common practice in America to roll plates directly from the ingot in one heat. This was unsatis- factory for more than one reason. First, the rolling of thin plates involved either the making of small ingots, which was objection- able and costly, or it involved rolling them from a large ingot, which * From data in Journal I. and 8, I., Vol. I., 1887, p. 121, et aeq. INFLUENCE OF HOT WORKING ON STEEL. 367 was very severe on the machinery; second, when the ingot was rolled into one single plate the segregated interior of the mass con- stituted a very considerable proportion of the finished piece, and it was generally out of the question to cut this part off, as by so doing a piece would be wasted which would be a very large pro- portion of the whole and which generally would be unsuited for other purposes on account of its dimensions. Third, it is not possible io make every heat of steel just the exact composition and physical qualities desired, and if the steel be cast in ingots of a size suited for the making of certain plates, and if, on account of such variations in chemical or physical qual- ity, they are not suited to the purpose for which they are made, they may be unsuited for any other purpose. On the other hand, when large ingots are cast and bloomed in a large mill and cut up into slabs, it is possible to know before the steel is rolled just what are the chemical and physical qualities of the metal, and the slabs may be made to suit the orders on hand. Moreover, the upper part of the ingot may be put into the less important work, while the bottom portion may be used for fire box places and for other pur- poses calling for the best material. For these reasons the use of a slabbing mill has come into quite general use. The Pennsylvania Steel Company was the first works in this country to introduce this practice; the Carnegie Steel Company followed with a much larger mill; The Pennsylvania Steel Com- pany then built one of a large size handling an ingot 36 inches by 48 inches, and the Illinois Steel Company and the Lukens Iron and Steel Company have lately followed the example. It is difficult to say just what should be the size of the slab for a given plate. Theoretically it would seem immaterial whether a 15- inch ingot is cogged to 8 inches and rolled to one-half inch, or whether it is cogged to 4 inches and rolled to the same thickness. The experiments of Mr. Eiley point the same way, but they are far from being comprehensive. If a slab 4 inches thick is not heated to a full heat the plate may be finished at the same temperature as one of the same gauge rolled from a hotter slab of twice the thickness, but in regular practice the thin slabs are sometimes heated hotter than the thick ones, and consequently will be finished at a higher temperature. If carried too far this produces a coarser structure and an inferior metal, so that it is best to proportion the thickness of the slab to the thickness of the plate. The exact relation is of 368 METALLURGY OF IRON AND STEEL. little importance as long as the reduction is sufficient, for in this matter the old adage is strictly applicable : "Enough is as good as a feast." This will be shown by Tables XIV-C and XIV-D, which investigate the effect of work on billets made from ingots 16 inches square and which thus had an all-sufficient reduction to begin with. TABLE XIV-C. Influence of Thickness of Test-Piece on the Physical Properties when the Percentage of Seduction in Soiling is Constant for all Thicknesses ; the Finished Bars in each Case having a Sec- tional Area of about 8 Per Cent, of the Billet. Ultimate d 1 J3 strength; Ibs. per sq. inch. Elastic limit; pounds per square inch. Elongation ir 8 inches; per cent. Reduction of area; per cent. 1 43 Q Sta Sfi c8 ^ IB "3 S ^ 3 n a 4J 3* M 2 ! o 5^ ^5 C | 58850 50620 19.75 . . . 58.4 4x4 2x% 59540 60160 37050 89840 85.00 31.00 60.0 57.4 8%x3% 2x54 59730 60490 38100 40490 29.76 32.50 66.4 55.1 9227 8x3 2x% 60950 61390 42110 42090 30.00 30.50 60.0 55.9 2^x2J4 2x^4 62J350 62700 43070 46630 27.50 28.75 60.7 63.3 2x1% 8x$ 65130 67470 52180 57830 26.25 23.75 58.9 67.5 4x4 2x% 67860 68140 42850 44050 25.00 24.25 40.8 43.9 1509 2x^4 67550 68040 43190 45560 26.25 28.25 46.1 46.6 8x3 2x% 67470 68300 44090 46610 26.25 23.25 53.2 50.3 4x4 2x% 72840 73260 47080 49160 25.00 24.00 40.7 40,8 8%x3% 2xVjs 71230 73510 46010 50830 26.25 25.00 40.5 43.5 1440 3x3 2x% 72950 78710 48760 50540 26.25 22.00 52.1 43.1 2Jx2^ 2x% 73620 75650 51550 58280 26.25 26.75 45.9 52.1 2x1% 2x^ 78560 79260 58140 6382C 22.75 25.25 52.0 50.4 It will be found from a detailed comparison of these tables that there is little difference between the bars of the same thickness, even though rolled from different-sized billets. There is a gain in ultimate strength as the thickness decreases, this being most marked in the cold-finished bars, but the differences are not very marked -except in the case of the one-eighth-inch. The elastic limit follows the same law, but it is raised more than the ultimate as the bar gets thinner. The elongation varies irregularly, but, as a rule, it remains unaffected except in the one-eighth-inch, where it is low- INFLUENCE OF HOT WOKKING ON STEEL. TABLE XIV-D. 369 Influence of Thickness of Bar upon the Physical Properties when all Pieces are Boiled from Billets Three Inches Square. Heat number. en 1 ti I *O I 33 Ultimate strength ; Ibs. per square inch. Elastic limit; pounds per square inch. Elongation in 8 inches; per cent. Reduction of area ; per cent. Finished at usual tem- perature. Finished at dull red heat. Finished at usual tem- perature. Finished at dull red heat. Finished at usual tem- perature. Finished at dull red heat. Finished at usual tem- perature. Finished at dull red heat. 4605 1 51370 51070 50850 52960 55560 50960 52430 51970 52280 55000 32860 83200 35700 86220 47380 83760 86050 37860 40040 42500 84.50 81.50 82.50 81.25 80.00 82.75 80.00 80.00 82.50 29.00 59.6 59.2 60.8 63.2 53.2 56.7 57.2 58.9 68.3 60.4 9227 1 59690 60350 60950 62230 66340 60190 60510 61390 63970 68130 87000 88560 42110 42600 49860 40130 40470 42090 49200 56180 85.00 29.50 80.00 25.75 27.50 80.00 32.50 80.50 29.25 24.00 55.4 58.8 60.0 55.9 56.6 58.7 61.7 55.9 61.9 65.7 1509 1 65ROO 67310 67470 69210 72100 67090 67660 68300 70200 77460 40980 43090 44090 47950 54060 45830 45170 46610 53680 64430 29.50 26.25 26.25 26.50 27.75 25.50 25.50 23.00 25.25 15.25 50.9 47.1 53.2 54.1 55.0 44.8 46.2 50.3 56.9 48.2 1440 2xg 72440 72570 72950 75620 77500 74060 68150 73710 71260 80240 46440 46200 48760 51160 60920 49480 45990 50540 54660 69360 . 27.50 27.25 26.25 25.00 26.00 24.00 28.50 22.00 27.25 18.50 45.7 47.3 52.1 53.5 46.8 42.0 58.4 48.1 49.4 58.6 ' TABLE XIV-E. Effect of Hammering Boiled Acid Open-Hearth Steel. NOTE. Chemical composition in per cent. ; C, .40 ; Mn, .86 ; P, .037 ; S, .046. JJ 1 j aa * Jr Is* | 1 I 1 ||I |S E* CD oa ft O&gS O I Remarks. g 33 en C? M _O . X O & ^ O ^ LJ fM TT 3 } S 3 ~f " > h 4J "(H S 8.gg- IEJ Hp*S OG Jg S 55 H & W PH E A B 6 6 54460 41500 89240 88660 29.00 28.00 41.2 42.2 61.0 46.8 Finished at dull yellow. Annealed at bright yellow. 5 50800 89070 26.50 88.0 57.0 Finished at dull yellow. D 4 55240 87300 25.50 87.0 63.3 Finished at dull yellow. E 3 51170 86450 27.50 39.3 59.2 Finished at dull yellow. F 2 51830 89280 28.00 41.8 58.1 Finished at dull yellow. G 2 57140 92400 28.00 42.0 61.8 Finished at cherry red. H 4 45620 89900 27.00 38.9 50.8 Finished at dull yellow. I 3 47830 88800 25.00 34.3 53.9 Finished at dull yellow. K 2 51000 88760 27.50 42.7 57.5 Finished at dull yellow L 5 54020 86400 7.50 5.8 62.5 Annealed at white heat. M 2 4700 93360 24.50 84.8 58.6 Finished at cherry red. 370 METALLURGY OF IRON AND STEEL. ered to some extent. The reduction of area is also irregular, but it seems to be independent of the thickness even in the thinnest plate. The conclusion seems justifiable that if the billets have already received sufficient work, the good condition caused thereby is not destroyed by reheating, since bars rolled from them reach their standard level of quality with only a reasonable degree of reduction, as proven by the fact that further work gives no decided improve- ment. But it is also certain, as shown by all experience, that no harm can be done by increased work, and that there is a slight gain in the long run provided the finishing temperature remains con- stant. SEC. XlVd. Experiments on forgings. The persistency of a proper structure even through subsequent heating may be seen in Table XIV-E, which gives the results obtained from a series of forged billets. The original bloom was 6 inches square, being rolled from an ingot 18"x20". From this bloom several short pieces were cut and treated in different ways : A was not reheated, but a test-piece was cut from it as a standard of comparison, B was heated to a full working heat and cooled without hammer- ing. C was hammered to 5 inches square in one heat. D was hammered to 4 inches square in one heat. E was hammered to 3 inches square in one heat. F was hammered to 2 inches square in one heat. G was hammered to 2 inches square in one heat from the an- nealed bar B and was finished at a cherry red heat. H was hammered to 5 inches square, then reheated and ham- mered to 4 inches. / was hammered to 4 inches square, then reheated and ham- mered to 3 inches. K was hammered to 3 inches square, then reheated and ham- mered to 2 inches. L was hammered to 5 inches square, then overheated and cooled without hammering. M was made by reheating the burned piece L and hammering to 2 inches square in one heat, being finished at a cherry red heat. All the pieces were worked under a 4-ton double-acting hammer, and the test-bars were cut from the corner of the billet and pulled in a length of 2 inches. INFLUENCE OF HOT WORKING ON STEEL. 371 It is quite evident that the pieces which were heated twice, and which received only one inch of reduction after the second heating, must have been finished hotter, as well as have received less work after a full heat, but in spite of these differences in amount of work and temperature it is clear that the bars are practically uni- form in strength and ductility, showing that the steel was in thor- oughly good, condition originally, and that proper heating did no harm when followed by a fair amount of work. The ultimate strength is fairly uniform save in the case of the two bars which were finished at a cherry red heat. The elastic ratio varies in much greater measure, but the results are not regular since, in some cases, as in K, a high ratio accompanies heavy reduc- tion under the hammer, while in others, as in D, it appears in bars which have received very little work. TABLE XTV-F. Comparative Physical Properties of Test-Pieces of Bessemer Steel Cut from Thick and Thin Angles of Large and Small Sizes. Each figure is an average of 50 bars. Thickness of angle; inches. Elastic limit: Ibs. per sq. in. Ult. strength; Ibs. per sq. in. Elastic ratio; per cent. Elongation in Sin. ; percent. Reduction of area; per cent. Large sizes. Small sizes. Large sizes. Small sizes. Large sizes. Small sizes. Large sizes. Small sizes. Large sizes. Small sizes. ! 43002 ' 43637 41671 41080 40391 38867 44158 43060 43128 41634 41836 40944 60097' 60019 60120 59467 59360 58267 61252 60629 60239 59151 59750 59084 '71.55' 72.70 69.31 69.08 68.04 66.70 72.09 71.07 71.59 70.38 70.02 69.30 28.13' 28.16 28.58 28.65 29.03 28.r7 27.55 28.55 28.52 29.24 28.74 29.38 '58.23' 57.59 5-5.17 55.30 58.43 51 . OD S -Jl I ll Ss .2 i a - i.- ft 0> ft p< ~" o 'o'o A c, & aj -^ ? o< ^ o"* o g 1 ft s * 3 s!S %2 OP .. g H-g' ^ t* P Jflfl ** p a 1 *<-< S|| Ills It Id 111 111 || *J fl . o>.2 >-3 p, III fc M H ^ ^ Qr-H Q 02 J^ O> ^j C8irX S D Pij? CH ,j i* 2 .. II 1 it IN jd or 15 SH Q *2 CS .2 p. .28 "n ft| |2 3-S S OQr^ p t ,;j 9^ +3 >flft 5ft t " 00 $) Qi a H fc M K OH H M P3 12 more than 7500 60040 49479 10561 44659 74.4 25.94 52.9 ^ 8 T8 '18 less than 7500 56475 51177 5298 42570 75.4 26.31 52.3 5 8g 13 more than 5500 57807 50020 7787 40407 69.9 26.94 57.4 CD 5 19 less than 5500 54799 51033 3766 39675 72.4 28.78 61.1 i 94 more than 4000 59582 54096 5486 44653 74.9 26.44 59.6 /B 68 less than 4000 58823 55741 2582 43028 73.8 27.10 55.3 69 more than 3000 58705 54013 4692 40420 68.9 28.50 56.9 & i 60 less than 3000 57021 55328 1693 40266 70.6 28.37 57.8 o Jig 7 10 more than 3000 59414 53557 5857 38222 64.3 28.09 59.9 10 TB 16 less than 3000 56501 54786 1715 36525 64.6 30.58 58.5 7 more than 3000 59135 53934 5201 88078 64.4 27.90 57.9 10 less than 3000 56977 55840 1137 36770 64.5 27.13 52.5 8 more than 2000 62228 59506 2722 42687 68.6 25.69 51.0 C * f 4 less than 2000 61425 60550 875 42325 68.9 25.41 51.0 i| Is jr 11 more than 1000 61827 59706 2121 42027 68.0 25.12 63.2 st 11 IB 9 less than 1000 59022 59320 39875 67.6 24.46 55.5 j? ~ s i 19 more than 1000 61174 59573 1601 40157 65.7 24.19 50.2 fj 4 14 less than 1000 60293 60408 39693 65.8 24.69 48.7 round, and thus naturally gives a higher ultimate strength, while it also works the skin of the piece during the finishing process with- out any great reduction in diameter. It will be seen that nothing is gained by this operation, for, although the guide rounds are slightly reduced in strength, they are considerably better in elonga- tion and reduction of area. SEC. XI Vi. Changes in the physical properties r of steel by vari- INFLUENCE OF HOT WORKING ON STEEL. 377 ations in the details of plate-rolling. It has been already stated that it is the practice at The Pennsylvania Steel Works to roll a preliminary test-bar from each open-hearth heat for physical test- ing, and that the ultimate strength of this bar corresponds closely with that of angles rolled from the same charge. In the case of plates, on the contrary, there is often a considerable variation, and Table XIV-Iv investigates the effect of such differences upon the physical qualities. TABLE XIV-L. Changes in the Physical Properties of Steel by Variations in the Details of Plate-Eolling ; Classified According to Strength of Finished Plate. 0-3 d fl Ultimate strength ; i 00 "^ S 1 pounds per square oT-2 o> d da . in 2 |gaT8 inch. s2 Is 05 3 -is 03 o .ga ft CO Is Sft| 1 J!< ?! ft o ft IJ Ed 03 o> ifl ess of s. A <~ O Sj|* ft -g b d if i -^ * !, 3 m i* '-C HI * O fe d * s* o g itlll la lal s" ~ S d o || J3tH ^ s.s dftS a 11 >< d"ft ft 5 ft od *A n3 fc fe ^ H H H < 85 more than 4000 56971 51963 5008 43106 75.6 26.66 57.8 ^ IB 80 less than 4000 56652 54680 1972 41345 73.0 27.35 55.2 2 42 more than 8000 56370 52161 4209 40387 71.6 28.28 58.5 50000 to 49 less than 8000 55358 54441 1517 39759 71.0 28.66 58.2 B 58000 7 more than 1700 55963 53391 2572 87613 67.2 30.27 58.6 TB 6 less than 1700 53981 53213 768 84802 64.5 31.43 59.6 1 3 more than 1100 56633 54076 2557 36366 64.2 27.91 54.7 ft 4 4 less than 1100 55292 54843 449 36150 65.4 28.50 53.7 o 39 more than 4000 60130 54234 5896 44572 74.1 26.63 58.7 58000 TS 38 less than 4000 59344 56401 2943 44054 74.2 26.92 56.2 rS to 64000 15 more than 3000 59750 53676 6074 40928 63.5 27.87 57.6 15 less than 3000 58920 56969 1951 40855 60.3 28.07 58.7 .! 6 more than 2550 62841 59151 8690 43933 60.9 25.92 50.5 Q| 6 less than 2550 61080 60557 523 41200 67.4 25.04 52.0 p,KJ 56000 9 more than 1400 61833 59647 2186 42512 68.7 25.28 54.9 - 64000 TB 11 less than 1400 59527 59439 88 40230 67.6 24.45 53.8 3 03 T< 05 17 more than 1700 61241 59442 171:9 40110 63.5 24.38 50.7 16 less than 1700 60331 60442 39800 C'J.O 24.43 48.6 It is assumed that the preliminary test-piece is the standard, and whatever difference from this is shown in the plate is due to the conditions of rolling. On this basis it is possible to compare those plates which show a great with those which show a less variation 378 METALLURGY OF IRON AXD STEEL. from the standard, and note the corresponding ductility. In the first division, for example, it was found that the average increase in strength from the preliminary bar to the finished plate was 7500 pounds per square inch. Consequently this figure was taken as a dividing line, and a comparison was made of the heats showing more than this difference with those showing less. The same rule was followed in all the other divisions. Table XIV-L gives a different view of the same data, the groups being divided on the less logical but more practical basis of the TABLE XIV-M. Comparative Physical Properties of Angles and Sheared Plates, both being made from Pennsylvania Steel Company Steel. S-l 1 ! Sfe-S **hA g as oo t a ft .S | ft .S g J 3 M" dS C-2 05 2-S 2 "^5 g G -S S ^ M '& ti O 03 00 3 * ic^-^ s d 1 S 5 . > 03 S&$ D lal jft ^ H' I* 8 |S Angles 32 52533 3(5284 69.07 32.18 63.7 Basic open-hearth soft steel, below .04 per cent, in phos- phorus. A to | Plates 107 20 54998 38017 69.12 29.28 58.G Angles 53171 34891 65.62 32.33 62.3 " Plates 102 55017 34947 63.52 29.03 61.5 Basic open-hearth medium steel, below .04 per cent. In phosphorus. A to | Angles Plates 64 265 58865 58271 39692 40349 67.43 69.24 30.52 28.27 58.8 58.1 Angles 212 60845 40891 67.21 29.35 57.4 A to f Plates 190 60217 43278 71.87 25.98 57.4 Acid open-hearth soft steel, below .08 per cent, in phos- phorus. A to j Angles Plates 126 59 60695 60768 39415 39061 64.94 64.28 29.23 25.87 55.6 55.1 A to | Angles Plates 81 13 60558 60666 38645 37932 63.81 62.53 28.95 24.67 53.8 52.7 strength of the finished plate. It will be seen that the elongation for a given tensile strength is not seriously affected by the variations in rolling, but that the hotter finished plates are somewhat better. The elastic ratio exhibits much less variation than would be ex- pected, and this might throw some doubt on the results, but all the different groups teach the same lesson, and in some of them the number of heats is so large as to give great weight to the conclu- sion. The plates were all rolled from slabs, which in turn had been rolled from large ingots, so that there was ample work on all thicknesses. SEC. XI Vj. Comparative physical properties of plates and INFLUENCE OF HOT WORKING ON STEEL. 379 angles. It is very difficult to make a comparison of two different structural shapes, since it does not often happen that the same heat is rolled into more than one kind of section, but an attempt is made to do this in Table XIV-M. The prime requisite is that the steel in both cases shall be of the same manufacture, and this specification is satisfied in the present instance. The figures for the angles are found by combining certain groups in Table XIV-H, which was compiled from the records of The Pennsylvania Steel Company, while the plates represent the average obtained from The Paxton Boiling Mill, which was running on slabs from the same works. The one predominant feature is the lower elongation in the plates. This may be explained by the fact that the metal receives a less thorough compression in the plate train than it does in the rolling of angles, in which latter case it undergoes a certain amount of lateral thrust. SEC. XlVk. Effect of thickness on the physical properties, of plates. The effects caused by variations in rolling temperature appear in their most marked degree in the comparison of plates of different gauges. It is not customary to test the same heat in several sizes, but by long experience the manufacturer is able to judge the relative properties of each thickness. The heads of two widely-known plate mills have given me as their estimate that, taking one-half inch as a basis, there will be the following changes in the physical properties for every increase of one-quarter inch in thickness : (1) A decrease in ultimate strength of 1000 pounds per square inch. (2) A decrease in elongation of one per cent, when measured in an 8-inch parallel section. (3) A decrease in reduction of area of two per cent. W. R. Webster* gives the same data on ultimate strength, but does not mention the relation of section to elongation. It is, therefore, plain that in the writing of specifications some allowance must be made for these conditions, since a requirement which is perfectly proper for a three-eighth-inch plate will be un- reasonable for a l!/2-inch. Moreover, the effect is cumulative, since a harder steel must be used in making the thick plate and * Observations on the Relations between the Chemical Constitution and Ulti- mate Strength of Steel. Journal I. and 8. I., Vol. I, 1894, p. 329. 380 METALLURGY OF IKON AND STEEL. this will tend to lessen the ductility rather than make up for the reduction caused by the larger section. In plates below three- eighths inch in thickness it is also necessary to make allowances, since it is almost impossible to finish them at a high temperature, and the test will give a high ultimate strength and a low ductility. These conditions have now been officially recognized by the United States Government, for the rules of the Board of Supervis- ing Inspectors, issued January, 1899, contain the following clause : "The sample must show, when tested, an elongation of at least 25 per cent, in a length of two inches for thicknesses up to one- quarter inch, inclusive; and in a length of four inches, for over one-quarter to seven-sixteenths, inclusive; and in a length of six inches, for all thicknesses over seven-sixteenths inch and under 1% inches." It is to be hoped that constructive engineers will follow this example in recognizing the influence of causes over which the manufacturer has no control. CHAPTER XY. HEAT TREATMENT. Within the last few years there have been radical advances in our knowledge of the structure of steel and the influence exerted by what has come to be known as "heat .treatment." The main prin- ciples of this branch of metallurgy have been understood for quite a long time, but they were applied only in exceptional cases, such as the manufacture of guns and armor plate. To-day progressive manufacturers are using the results of research in improving the quality of their ordinary forgings and castings, and it is therefore necessary to consider at some length the general underlying prin- ciples of the science of micro-metallography. This has been done in the latter half of this chapter, the article being written by my brother, J. W. Campbell, who has- charge of the special steels at Steelton. The introduction of accurate determinations of temperatures and a better knowledge of the proper heat to use, h#s to a certain extent diminished the value of the experiments and investigations published in the first edition of this book, but I believe they may be worth recording again, as it is quite certain that many non-pro- gressive works will follow the common and ancient methods of an- nealing both at the forge of the smith and on a larger scale in the treatment of eye bars and similar material. In the following sec- tions the word "annealing" is used unless otherwise stated to signify that the piece was heated in a muffle heated by a soft coal fire, the bar being withdrawn when it had reached a dull yellow heat. The experiments were carefully performed and it is believed that the practice was fairly uniform. SECTION XVa. Effect of annealing on the physical properties of rolled bars. It is a well-known fact that annealing tends to remove the strains which are created by cold rolling and distortion, but it is not generally understood how profound are the changes 381 382 METALLURGY OF IRON AND STEEL. produced. Table XV-A will show the results obtained on rounds and flats by comparing the natural bar with the annealed specimen TABLE XV-A. Effect of Annealing on Rounds and Flats of Bessemer and Acid Open-Hearth Steel. A 4"x4" billet from each heat was rolled into a 2"x%" flat and another into a ^ round. Limits of ultimate strength; pounds per square inch. Kind of steel. Number of heats in average. Condition of bar. Ultimate strength; pounds per square inch. Elastic limit; pounds per square inch. Elongation in 8 inches ; per cent. Reduction of area; percent. Elastic ratio; per cent. 56000 Bess. 11 Natural Annealed 58869 55703 42318 37828 27.75 29.14 58.83 66.55 71.88 67.91 60000 O.H. 4 Natural Annealed 58568 54098 40300 31823 29.69 28.75 60.78 62.65 68.81 58.82 60000 to Bess. 6 Natural Annealed 62087 59372 45323 40570 27.04 80.13 55.31 65.50 73.00 68.83 1 64000 O.H. 7 Natural Annealed 62187 58364 42606 35120 28.04 28.61 62.16 63.47 68.51 60.17 A 64000 to 68000 Bess. 9 Natural Annealed 66241 61694 47568 42228 26.08 28.25 50.07 62.91 71.81 68.45 a & 68000 to Bess. 3 Natural Annealed 70457 65903 50263 44660 24.75 26.08 48.30 63.23 71.34 67.76 72000 O.H. 2 Natural Annealed 70530 65500 49000 37685 26.88 23.38 61.10 55.30 69.47 57.63 72000 to Bess. 4 Natural Annealed 77440 71548 53760 47643 24.06 25.81 42.35 57.53 69.42 66.59 80000 O.H. 12 Natural Annealed 76616 69402 51108 40505 24.52 23.04 53.73 56.54 66.71 58.36 56000 to Bess. 11 Natural Annealed 58458 54194 41698 35603 31.45 30.05 56.13 63.13 71.33 65.70 60000 O.H. 4 Natural Annealed 58130 51418 40400 30393 30.13 31.06 61.75 60.50 69.51 59.11 60000 tn Bess. 6 Natural Annealed 60825 56192 43135 87542 30.42 30.63 56.20 63.38 70.92 66.81 64000 O.H. 7 Natural Annealed 62089 55021 42441 31576 80.14 30.36 60.86 60.00 68.86 57.39 1 64000 to 68000 Bess. 9 Natural Annealed 64621 58838 45194 88476 28.42 28.36 47.80 59.01 69.94 65.39 & 68000 to Bess. 3 Natural Annealed 69773 04160 49060 43770 2C.G7 28.53 48.40 59.50 70.31 68.22 72000 O.H. 2 Natural Annealed 69420 60850 45090 84000 25.63 26.50 59.30 52.10 64.98 55.87 72000 to Bess. .< Natural Annealed 76900 68780 52240 43568 23.44 26.38 40.15 51.00 67.93 63.34 80000 O.H. 12 Natural Annealed 75865 67618 49691 39403 1 24.69 [ 26.31 54.40 51 .C6 65.50 58.27 HEAT TREATMENT. 383 when all the pieces were rolled from billets of the same size and on the same mill. The decrease in ultimate strength by annealing the Bessemer bars averaged 4175 pounds per square inch in the rounds and 5683 pounds in the flats, while the open-hearth was lowered 5134 pounds in the rounds and 7649 in the flats. In this important and funda- mental quality the two kinds of steel are very similarly affected, but in other particulars there seems to be a radical difference which is difficult to explain. TABLE XV-B. Comparison of the Natural and Annealed Bessemer Steel Bars Given in Table XV-A, which show about the same Ultimate Strength. 2 2 1 1 1 o 43 d S 3 CS I j d cr y fto 00 eg 1 *s box d 3 S|3 1! 2,3 At a$ t, Q O S fci 14 I3&3 leB d Ha* il od rt aj 0) ft || O p S S PH 56000 to 4 Natural 58568 40300 29.63 60.78 68.81 8? 60000 7 j Annealed 58364 35120 28.61 63.47 60.17 68000 to 2 Natural 70530 49000 26.88 61.10 69.47 *^ 72000 12 Annealed 69402 40505 23.04 56.54 58.36 55000 to 4 Natural 58130 40400 30.13 61.75 69.51 .d 60000 7 Annealed 55021 31576 30.36 60.00 57.39 5S 60000 to 7 Natural 62089 42441 30.14 60.86 68.36 jj^S 64000 2 Annealed 60850 34000 26.50 52.10 55.87 S V' 66000 to 2 Natural 69420 45090 25.63 59.30 64.96 70000 12 Annealed 67618 89403 26.31 51.36 58.27 further experiment would or would not corroborate these results, it is quite certain that annealing under ordinary conditions, even though very carefully conducted, may produce grave differences in physical properties in steels of similar composition which have been rolled in the same manner and treated at the same time, even when the effect upon the ultimate strength has been the same. It would also appear that in the Bessemer steel the marked increase in ductility is purchased at a great sacrifice of strength, and the question arises whether the gain is not more than balanced by the loss, and whether an equal degree of toughness could not be HEAT TREATMENT. 385 secured by using a softer steel in its unannealed state. A com- parison of the natural and annealed bars of corresponding tensile strength in Table XY-A will give the results shown in Tables XV-B and XV-C. SEC. XVb. Effect of annealing on bars rolled at different tem- peratures. These results show that the annealed bar has a very much lower elastic limit than a natural bar of the same ultimate strength, and oftentimes has less ductility. The difference between the Bessemer and open-hearth steels cannot be due to irregular TABLE XV-D. Effect of Annealing Acid Open-Hearth Boiled Steel Bars 2x% inches. to ^ t, 2 2 a c ip, 8 * S rt& a ccg 1 3 8-2d ,9 ^ d " 1 !||l iS . jj "S < Pi | ft dj O 1 M ^ o^g,| o3 ^ .2 S-S "o^ 3 aT 2d 2 ^ E o . d 5*3 o a) .S OB ft O O 111 1 111 ill !! !! fc 3 H 5 H H PH 5 56000 to 60000 Usual Nat. Ann. 58130 52323 89733 31677 80.42 30.75 61.90 60.63 68.4 60.5 1 3 C I 9 ' P 035' ' 'Mn, .56. Dull red Nat. Ann. 59857 51557 43037 33893 31.83 32.92 59.60 63.60 71.9 65.7 60000 to 64000 Usual Nat. Ann. 61703 54463 41985 30953 30.19 30.38 60.70 59.35 68.0 56.8 T-r 4 C 12 "P 086* Mn, .48. ' Dull red Nat. Ann. 63585 55058 45213 36988 30.06 30.94 57.58 61.53 71.1 67.2 72000 to 80000 Usual Nat. Ann. 75688 66584 49155 37934 24.66 26.06 54.05 50.74 64.9 57.0 III g C 9 4 * P 052 * ' Mn, .77. ' Dull red Nat. Ann. 78083 53334 67058 1 40343 27.41 26.50 52.23 53.41 68.3 60.2 finishing,, since all the bars were rolled at the same time,, and further experiments given in Table XV-D indicate that the same law holds good whether the metal is finished hot or cold. In the bars which are finished at the usual temperature there is a loss in strength due to annealing of from 6000 to 9000 pounds per square inch, and a lowering in the elastic limit of from 8000 to 11,000 pounds. In the colder finished bars the loss in strength is from 8000 to 11,000 pounds, and the elastic limit is lowered from 8000 to 13,000 pounds. Thus in both cases the elastic limit is affected much more than the ultimate strength, and the 386 METALLURGY OF IRON AND STEEL. result is seen in a lower elastic ratio. The ductility does not seem to be materially improved in any instance. The cold finishing raised the strength of the bars 1727 pounds per square inch in Group I, 1882 pounds in Group II, and 2395 pounds in Group III. Annealing lowered the strength of these cold-finished bars so that in Group I it was 766 pounds per square inch below the annealed hot-finished bar, while in Group II it was TABLE XV-E. Effect of Annealing on Bars of Different Thickness, when the Per- centage of Eeduction in Kolling had been Constant for all Pieces. a ^s SS s Ultimate strength; Ibs, per sq. inch. Elastic limit; Ibs. per sq. inch. Elongation in 8 inches; percent. Reduction of area ; percent. 4605 4x4 f 2x1* 4x4 51640 51120 50850 45870 45100 46350 46010 44960 33440 26350 25980 37.50 82.50 82.50 28570 87.50 88.00 89.50 84.00 81.25 59540 59730 51360 62350 65130 51230 54110 37050 38100 42110 43070 52180 28410 85.00 29.75 80.00 27.50 82.50 32.75 81.75 80.00 31170 60.1 56.4 61.0 64.8 64.0 64.3 67.2 60.0 56.4 60.0 60.7 58.9 59.7 60.1 56.6 62.4 64.9 1509 4x4 2x 67860 67550 67470 42850 43190 44090 38750 88810 40430 25.00 26.25 26.25 26.50 29.00 29.25 40.8 46.1 53.2 57.S 58.4 56.1 4x4 1440 72840 71230 72950 73620 78560 67060 67860 69720 74000 47080 46010 48760 51550 58140 43580 42020 43920 25.00 26.25 26.25 26.25 22.75 27.00 29.00 26.25 26.50 25.25 40.7 40.5 52.1 45.9 52.0 53.6 63.4 55.4 54.1 53.G 595 pounds above it, and in Group III 474 pounds. The effect upon the elastic limit is not as thorough, and the influence of the cold finishing may be seen in the higher elastic ratio of the an- nealed cold-finished bar. SEC. XVc. Effect of annealing on bars rolled under different conditions of work and temperature. All these results will be cor- roborated by Tables XV-E and XV-F, which show the effect of annealing on bars which have been finished under different con- ditions. In Table XV-E, where each bar was made from a billet HEAT TREATMENT. 387 of proportionate size, the pieces would be in the rolls about the same length of time, so that the only difference in character will be due to the more rapid loss in heat from a thin bar and from the more "thorough compression. In Table XV-F, where all bars were rolled from the same-sized billet, these factors are supple- mented by the extra cooling during the longer exposure in the rolls. TABLE XV-F. Effect of Annealing on Bars of Different Thickness, when All Pieces had been Boiled from Billets 3 inches Square. Heat Number. ft PQ w 1" Ult. strength; Ibs. per sq. inch. Elastic limit; Ibs. per sq. inch. Elongation in Sin. ; percent. Reduction of area; per ct. | Annealed. Natural. Annealed. 1 Annealed. Natural. Annealed. 4605 1 51370 51070 60850 52960 6556Q 45490 43280 46350 44470 45830 32860 83200 35700 86220 47380 25560 24110 25980 84.50 81.50 32.50 81.25 30.00 86.75 88.00 89.50 88.50 33.25 59.6 59.2 60.8 63.2 63.2 65.6 64.2 67.0 69.6 69.0 27780 9227 1 59690 60350 60950 62230 66340 52880 52270 52460 53500 64310 37000 88560 42110 42600 49860 29030 28460 29860 81000 30600 85.00 29.50 80.00 25.75 27.50 82.00 82.00 81.75 80.75 26.25 55.4 58.8 60.0 55.9 66.6 66.4 65.1 56.6 58.4 61.6 1509 1 65600 67310 67470 69210 72100 61480 64500 62660 65240 66940 40980 43090 44090 47950 64060 37840 41400 40430 44510 49000 29.50 26.25 26.25 26.50 27.75 29.00 29.25 29.25 80.50 27.50 50.9 47.1 63.2 64.1 55.0 57.1 66.0 66.1 52.6 52.6 1440 li 72440 72570 72950 75620 77500 69730 67980 67860 71560 70820 46440 46200 48760 51160 60920 45250 42000 43920 48250 56420 27.50 27.25 26.25 25.00 26.00 24.25 28.25 26.25 26.50 25.50 45.7 47.3 52.1 63.5 46.8 56.3 64.2 65.4 59.0 59.9 SEC. XVd. Effect of annealing on plates of the same charge which showed different physical properties. This matter of finish- ing temperature is of supreme importance in filling specifications on structural material, more especially in the rolling of thin plates, for it will often happen that different members of one heat will show wide variations in tensile strength when the metal itself is practically homogeneous. Table XV-G will illustrate this point by giving the records of test-pieces which gave the greatest vari- ations in any one heat, and comparing the natural bar with a piece of the same strip when annealed. 388 METALLURGY OF IRON AND STEEL. It will be seen that annealing has almost wiped away the vari- ations in each heat, and it is therefore quite certain that the dif- ferences lie in the rolling history. The true way of. testing the TABLE XV-G. Showing that Eolled Plates of the same Acid Open-Hearth Heat, which show Wide Variations in their Physical Properties, are made alike by Annealing. NOTE. In each case, A is the test giving the highest tensile strength of any plate in the heat, and B is the one giving the lowest. Carbon was determined by color and is therefore not reliable. Heat number. Thickness of plates. Condition of test bar. i 49 CO i Ultimate strength; pounds per square inch. Elastic limit; pounds per square 4* 3 00 03 d .- - H * is WU3 II H Reduction of area; per cent. Elastic ratio; per cent. Chemical composi- tion; per cent. C. P. Mn. S. 6633 i Natural Natural Annealed Annealed A B A B 61000 56480 47750 46970 53200 46300 29980 30690 21.50 25.25 34.50 35.00 61.9 60.0 67.0 64.5 87.2 82.0 62.8 65.3 .16 .12 .015 .015 .32 .31 .022 .019 . . . . . . . . . 5658 i Natural Natural Annealed Annealed A B A B 65370 60380 52160 50260 52560 48800 32450 33340 21.75 21.50 32.00 32.50 58.7 61.1 57.0 62.6 80.4 80.8 62.2 66.3 .14 .10 .009 .012 .45 .45 .025 .020 8217 i Natural Natural Annealed Annealed A B A B 64620 59960 52820 50000 53140 48490 35450 31840 25.00 21.50 27.00 31.50 58.1 45.5 62.2 56.4 82.2 80.9 67.1 63.7 .16 .14 .021 .016 .44 .44 .081 .025 8226 i . Natural Natural Annealed Annealed A B A B 64260 57040 54070 53960 54370 39990 38520 38520 21.00 28.75 27.50 29.50 50.6 56.6 64.4 63.3 84.6 70.1 71.2 71.4 .12 .12 .036 .084 .34 .82 .058 .047 8231 T 8 Natural Natural Annealed Annealed A B A B 64480 61100 53*30 52180 50560 45030 84870 33780 26.00 26.00 31.25 81.25 58.8 48.0 61.9 63.2 78.4 73.7 64.8 64.7 .13 .11 .021 .018 .55 .51 .048 .044 8233 i Natural Natural Annealed Annealed A B A B 66360 58160 52760 51480 59100 47630 36940 .40480 20.75 24.50 33.00 28.75 62.7 60.3 65.0 56.0 89.1 81.9 70.0 78.6 .11 .11 .026 .020 .37 .89 .033 .028 8234 A Natural Natural Annealed Annealed A B A B 66300 maso 55560 5403& 49440 47930 87360 84448 20.75 27.00 28.25 31.75 67.5 61.7 60.0 63.7 74.6 78.1 67.2 63.7 .15 .14 .024 , .021 .49 .47 .022 .023 8235 i Natural Natural Annealed Annealed A B A B 63220 58240 47740 47600 58300 47630 29930 30530 13.50 21.25 83.25 34.00 54.9- 53.5 63.9 57.2 92.2 81.8 62.7 64.1 .10 .11 .017 .017 .33 !85 .035 .034 8296 A Natural Natural Annealed Annealed A B A B 64020 58720 53860 50660 49510 42960 83710 32710 23.25 30.25 29.25 85.00 58.1 60.0 58.6 64.7 77.3 73.2 62.6 64.6 .11 .13 .025 .017 .46 .45 .037 .022 HEAT TREATMENT. 389 homogeneity of steel, or of comparing two different samples, is to make the tests on annealed bars. This practice was pursued in Chapter XIII. SEC. XVe. Effect of annealing on the physical properties of eye-bar flats. It does not follow that plates and bars should be annealed to put them into their best condition. On the contrary, the foregoing tests have shown that very little is gained in ductility, while there is quite a loss in working strength, and that it would be better and much cheaper to choose a softer steel in its natural state. Moreover, it must be considered that the bars which have been discussed in the foregoing tables have been small test-pieces which could be treated under fairly constant conditions, and even then the results are far from regular. TABLE XV-H. Comparative Tests of Eye-Bar Steel. Heat number. Longitudinal strip; cut from near the edge of eye-bar ; natural. Full-sized eye-bar; annealed. Elastic limit; pounds per square in. JH lf! 03 P< 00 Elongation in 8 inches; per cent. Reduction of area; per cent. Elastic ratio; per cent. Elastic limit; pounds per square in. i"|fl- itsg ifl 2ss |^ar Elongation in 8 inches; percent. Reduction of area; per cent. Elastic ratio; per cent. 1 2 8 4 5 6 7 8 9 40710 11370 C.7CO 40380 41480 41310 40370 41900 41070 68830 71400 69460 69400 72320 73640 72060 76700 69680 27.CO 26.25 25.75 25.00 24.50 23.75 25.60 25.75 27.00 47.18 50.C8 44.31 48.41 46.78 86.54 40.00 43.76 44.33 E9.1 58.2 57.3 58.9 57.4 66.1 56.0 54.6 58.9 86500 40400 38300 40600 42100 83700 35400 89600 35900 62100 65200 63250 67100 65000 57600 64700 67700 65200 43.70 40.00 41.85 86.00 86.60 45.60 45.62 88.43 40.00 32.60 46.55 45.95 45.00 48.40 50.00 61.80 42.65 46.40 58.8 62.0 60.5 60.5 64.8 58.5 54.7 58.5 55.1 Av. ! f 41008 71499 25.62 44.60 57.4 88056 64206 4087 46.54 59.3 These deductions will be corroborated by Table XV-H, which gives the parallel records of pieces cut from a flat bar in its natural state, and the full-sized eye-bars after annealing. The steel was made and rolled by one of our largest American works. It is plain that there is a great gain in the elongation, but the reduction of area is unaffected and there is a decided loss in elastic and ultimate strength. SEC. XVf. -Methods of annealing. A different view of the sub- ject is taken by Grus. C. Henning.* He states that steel is injured * Trans. Am. Soc. Mech. Eng. } Vol. XIII, p. 572. 390 METALLURGY OF IRON AND STEEL. by annealing if it is in contact with flame, while it is improved if it is reheated in a sealed muffle. I cannot assent to this broad con- clusion, for, while it may be true that a flame can be run too hot and the piece be burned through carelessness, it by no means fol- lows that such local overheating is necessary; nor is there any ground for assuming the absorption of deleterious gases from a proper flame. Moreover, the figures which he gives do not show a decided improvement of any kind in the bars which were heated in a retort. TABLE XV-I. Comparative Physical Properties of Natural and Annealed Flat Steel Bars; as given by Henning.* is" I . 1 o d E p, 43 W H U <9 M C 53 C ^H s S ^ H eg 00 ^0 - ^1 I Ml 43 2^0^ Ixl 9 - 73 01 ^ II fc S 4 W H N H P2 10 i to l& 1.12 Natural Annealed 88737 40299 71226 69296 23.89 25.53 47.0 53.5 54.4 58.2 16 1| tO I/B 1.41 Natural Annealed 85411 88298 684C5 67971 24.38 24.95 46.65 49.17 51.7 56.3 12 li to 15 1.62 Natural Annealed 85729 38692 63490 60411 24.25 25.28 47.27 49.85 51.4 55.7 It is stated (loc. cit., p. 577) that most of the "flats" were "properly" annealed, and so I have averaged the records which he gives of the natural and the reheated pieces, separating them into three groups according to thickness. The results are given in Table XV-I. It will be seen that the metal has undergone very little change at all, and it is impossible to see anything which can be called a radical improvement. Any attempt to carry out a general system of annealing plates and shapes will result in wide variations in temperatures and rates of cooling, for it will be impossible to have a large pile of metal heated uniformly throughout, since the outside of the lot will be at Trans. Amer. Soc. Mech. Eng., Vol. XIII, p. 586, et seq. The factor which Mr. Henning calls the "yield point" is here called the elastic limit. I havfe omitted from the averages the tests which are noted in the original as being wrongly marked, and also three tests which show such extremely low elongation that it is certain the material was not properly treated, or that there is an, error in the records. HEAT TREATMENT. 391 a full heat when the interior is unaffected. Since the manufacturer may always manipulate the operation so as to affect the test-pieces in preference to the rest of the steel, and since it will be to his interest to keep the temperature as low as possible to avoid warp- ing, there will be no certainty either that the work has been properly carried out or that it has been of the least advantage. SEC. XVg. Further experiments on annealing rolled bars. The experiments on annealing related in this chapter were per- formed by the usual method of estimating temperatures by the eye. They were, however, conducted under conditions exceptionally favorable to uniform results, as the pieces were small and were enclosed in a muffle and were carefully watched. No ordinary an- TABLE XV-J. Effect of Annealing at about 800 C. (1472 F.) on the Physical Properties of Structural Steel. (Bars are rolled flats 2"x%".) ~.s II -r f-t - - "'.; ""^.S IE f- .4 54 to 64,000 17 Annealed 57.870 35.320 36.6 57.6 61.0 55 to 60.000 Acid.* 4 Natural 58130 40400 30.1 61.7 69.5 55 to 60,000 7 Annealed 55.021 31,576 30.4 60.0 57.4 SEC. XVh.f General remarks on the determination of tempera- tures. For the commercial operation of annealing, the tempera- ture may be conveniently and accurately determined by the use of a platinum or copper ball with the usual water receiver. In more accurate work it is advisable to use a Le Chatelier pyrometer, but in either case considerable care must be taken to insure that the piece of metal which registers the temperature, whether it be the ball or the electric couple, is of the same degree of heat as the forg- ing or the casting under treatment. It is generally taken for granted that if the juncture of a Plati- num Platinum ten per cent. Ehodium couple is in contact with the steel under treatment, the temperature as registered is correct. Practically, although not absolutely, this is true, for if the con- ditions of heating are the same, that is, if the furnaces are of the same general size and plan and the pieces under treatment are * These constitute Group III In Table XV-C. t The remainder of this chapter is mainly the work of .?. W. Campbell. HEAT TREATMENT. 393 approximately the same size, the readings are relative, and being relative may be considered to be correct. Now is this true under conditions radically different ? If a small piece of steel is placed in a muffle and heated, the muffle having been at a high temperature before the introduction of the piece, it will be found even while the piece is black or very dark red, say not over 650 C., that the needle of a Le Chatelier pyrometer, the couple of which is in con- tact with the steel, will indicate a temperature some thirty degrees higher. This is probably due to the fact that while it takes some time for the mass of steel to absorb the heat from the muffle, the fine wires of the couple arrive at the high temperature in perhaps twenty or thirty seconds. Of course, the juncture, being in con- tact with the cooler steel, is considerably cooler than the furnace, but nevertheless it is some degrees higher than the piece, and this higher temperature is the one which sets up the difference of poten- tial which affects the galvanometer. This is undoubtedly the case in still greater measure with larger furnaces and larger masses, and if it is desired to compare a small piece with a large one the temperature of treatment, must be the same. There is one way of arriving at this with certainty, and this is in accordance with what Howe describes as the con- dition of invisibility. He sets forth that a certain color is indica- tive of a certain temperature, whatever the material, and proves it by stating that if pieces of several different kinds of metals be placed in a furnace and heated carefully and slowly, and held till it is certain that they are heated equally through and through, on looking into the furnace nothing can be seen but the walls of the furnace. The pieces are invisible. He then shows that since the only light is that given off by the heated surfaces themselves and since if there were even the slightest difference in color, the edges of the pieces could be seen, the whole furnace and contents must be the same color and this he calls "invisibility." Now if a large piece of metal is heated until the wires of the couple cannot be seen in contact with the piece, and if this heating be continued until the piece shows an uniform color all over its surface, and until it has been heated throughout to this color, an absolute reading is obtained at least absolute within the limits of error of the galvanometer. In this connection it should be stated that the Le Chatelier pyrometer is the best practical method of taking readings of high temperatures. That a piece 394 METALLURGY OF IRON AND STEEL. has been heated thoroughly can only be discovered by prac- tice and a knowledge of the heating capacity of the furnace. As good a way perhaps as any is to note the time of heating to a certain indicated temperature, then cool under conditions which may be duplicated and note time of cooling ; then heat to this temperature again, soak for some time and cool under previous conditions, and if the cooling takes longer the piece is heated more nearly uni- formly. After a few trials in this way the necessary time may bo estimated with sufficient accuracy. It may seem that this is an unnecessary refinement, but up to the present time, except in a limited number of grades of steel and at a few w r orks, proper atten- tion has not been given to the annealing of steel. SEC. XVi. Definition of the term "critical point/' If a piece of steel containing over 0.50 per cent, of carbon be allowed to cool slowly from a high temperature, certain peculiar phenomena will be noticed. The cooling at first proceeds at a uniformly retarded rate, but when a temperature of about 700 C. is reached there is an interruption of this regularity. In some cases the rate of cooling may become very slow, in other cases the bar may not de- crease in temperature at all, while in still other cases the bar may actually grow hotter for a moment in spite of the fact that it is free to radiate heat in every direction and that it has been cooling regularly down to that particular temperature. Moreover, it will be found that when this "critical point" is passed, the bar cools as before until it reaches the temperature of the atmosphere. It is, of course, a matter of common knowledge that a bar will cool in less time from 1000 C. to 900 C. than it will from 200 C. to 100 C. and the term "uniformly retarded/' as above used, is in- tended to cover this fact. It is quite clear that there must be some change taking place within the metal itself giving rise to heat, and any point at which such an action takes place in any steel is called a "critical point" and in metallography such a point is denoted by the letter A, the particular one just described in which there is a retardation in the cooling of a piece of steel being denoted by the term Ar. In heat- ing a piece of steel through this range of temperature, we naturally encounter an exactly opposite phenomenon, there being an absorp- tion of heat by internal molecular reaction, with a consequent retardation in the rise of temperature, and this point is called Ac. It has been shown by Prof. Howe that Ac is some 30 C. higher HEAT TREATMENT. 395 than Ar, but it is also found that in order to induce the change Ar the steel must first be heated past the point Ac. while the change at Ac cannot take place unless the steel has first been cooled to a point below Ar. It is clear therefore that these two retardations are simply opposite phases of the same phenomena. The previous discussion has considered only steels containing as much as one-half of one per cent, of carbon and mention has been made of only one critical point, when as a matter of fact it is quite certain that there are three, although it will be shown later that the three points are practically coincident in steels containing 900' 850 .60 ABSCISSAS =CARBON CONTENT ORDINATES=TEMPERATURE CENT. .70 .80 FIG. XV- A. VARIATIONS IN THE CRITICAL POINTS IN DIFFERENT STEELS. over 0.30 per cent, of carbon. At one of these points, recently proven to be the second, is the point of magnetic transformation. Below this point carbon steel is attracted by a magnet. Abo^e this point it is attracted only slightly if at all. It has been before explained that the critical points are found at a slightly different temperature according to whether the metal is being heated or being cooled, and it is evident that the point of magnetic trans- formation, which coincides with the second critical point, will vary in the same way. In soft steels these three points are readily distinguished, but as 396 METALLURGY OF IRON AND STEEL. the carbon content is increased the difference in temperature be- tween these points grows less and less, until in the harder steels the variations are hardly beyond the limits of experimental error. Moreover, there are several elements beside carbon, like mangan- ese, phosphorus, etc., which influence the location of the critical point, so that with two steels of the same carbon content, but with varying manganese, the upper critical point of one may be lower than the lower critical point of the other. The three critical points in a cooling bar are distinguished as Ar 3 , Ar 2 , Ar 1? the point Ar 3 being the one at the highest tempera- ture and Ar at the lowest. In heating a bar the same three in- terruptions take place and the points are designated Ac 1? Ac 2 , Ac ;i , it being understood that in each case the lowest numerals Ac and Arj refer to the lowest temperatures, and the highest numerals Ac 3 and Ar 3 to the highest temperatures, and that points bearing the same exponent like Ac and Ar represent practically the same degree of temperature. In Fig. XY-A is shown a diagram which aims to represent the variations in the critical points for different steels. The data given by different experimenters vary consider- ably, but the heavy lines representing Ar 1? Ar 2 and Ar 3 are found by striking a sort of average from the available information. On each side of these heavy lines are shaded areas which represent the variations in the position of the critical point caused by differences in the content of manganese, phosphorus, etc. In the case of the soft steels the critical points are so far apart that the variations caused by these elements do not cause the maximum of one point to coincide with the minimum of the one just above, but as the content of carbon increases, the range between the highest and lowest criti- cal points decreases, while the variations do not decrease, and as a consequence the maxima and minima run together so that they are indistinguishable. The nature of the change that takes place at any one of these critical points is not known, but it is known that at each such point there is a great change in the micro-structure of the steel. It is known that the structure of the metal is quite different on either side of the critical points ; that the forms, in which the iron and its- alloyed constituents present themselves, change quite suddenly at certain definite points, and the structures found under certain well understood conditions are so characteristic that they form the basis of a science, but it is not known whether the heat liberated or ab- HEAT TREATMENT. 397 No.l. No. 2. No. 3. No. 4. N ,. 5. No. 6 No. 7. No, 8. FIG. XV-B, No. 9 398 METALLURGY OF IRON" AND STEEL. No. 10. No. 11. No. 12. No. 13. No. 14. No. 15. No. 16. No. 17. FIG. XV-C. No. 18. HJ2AT TREATMENT. 399 No. 19. No. 20. No. 21 No. 22. No. 23. FIG. XV-D. No. 24. 400 METALLURGY OF IRON AND STEEL. No. 25. No. 26. No. 27. No. 2& No. 29. No. 30. FIG. XV-E. HEAT TREATMENT. 401 No. 31. No. 32. No. 33. No. 34. No. 35. No. 36. FIG. XV-F. 402 METALLURGY OF IRON AND STEEL. No. 37. No. 38 No. 39. No. 40, No. 41. No. 42. No. 43. No. 44. FIG. XV-G. No. 45. HEAT TREATMENT. 403 sorbed at a critical point is due to the change from one structure to another, or whether both the change and the heat are due to some unknown molecular phenomena. The next section will discuss the structures and forms which are best known and which must be studied to understand the effect -of heat treatment. SEC. XVj. Definitions of the different structures seen under the microscope. The microscopic examination of almost any piece of steel properly polished and etched will show that it is not entirely homogeneous, but that it is usually made up of at least two differ- ent forms of matter. It will not do to say that it is always made up of different substances, for it is generally agreed that some of these forms are allotropic,* the particular forms present in any one piece depending upon the way in which that piece has been heated and cooled. Considering all variations in heat treatment, the following forms will be encountered by the investigator: aus- tenite, martensite, pearlite, cementite, ferrite, trpostite and sorbite. Austenite is produced only by quenching steel containing more than 1.30 per cent, of carbon in ice water from above 1050 C. Its ap- pearance is intended to be represented by the white portion of No. 1, Fig. XV-B, but this may be cementite in spite of the fact that the piece was steel containing 1.40 per cent, carbon, one-quarter of an inch thick, and was quenched in melting ice from a dazzling heat. Even under these conditions it is impossible to obtain a large quantity of austenite, sirce the tendency to revert to the next form is very strong when the proper temperature is reached. The theory of austenite, as well as of martensite, will be taken up in Section XVo. At about 1050 C. a change occurs, and in this grade of steel quenched below th\s point and above A^ the second form, martensite, appears. This phase, together with a certain amount of cementite or of ferrite, depending on the carbon con- tent, is found in carbon steels containing less than 1.30 per cent, of carbon quenched at any point above Ar 1? as will be shown in Table XV-M. Martensite is the constituent which confers hardness on steel and corresponds to the maximum hardness obtainable by * The word "allotropic" is used by some of the metallographists to designate the character of the metallic aggregates. This is not strictly correct, since allotropy refers to unlike forms of the same element, while the different metallic aggregates found in microscopical investigations of masses of steel are not ele- ments and are not of the same composition. The term "phase" was introduced by Gibb and is used later in this discussion. 404 METALLURGY OF IRON AND STEEL. carbon alone. It may be compared to a sugar solution which is more or less sweet according to the proportion of sugar present. Marten- site may be easily recognized by its appearance, shown in Fig. XV-B No. 2. At the upper critical point Ar 3 , the conditions become more favorable for the production of cementite and ferrite, and variable amounts of one or the other are formed, depending on the carbon content; at the second critical point, Ar 2 , no radical change is noticeable, the only effect being an increase in the amount of ce- mentite or ferrite, but at the lower critical point, Ar 1? the marten- site disappears, and in steels cooled slowly to below this temperature the structure is composed entirely of ferrite, or entirely of pearlite, or of pearlite mixed with ferrite or cementite. Ferrite is iron free from carbon" and forms almost the whole of a low carbon steel, while cementite is considered to be a compound of iron and carbon denoted by the formula Fe 3 C, the carbon of this form being known as cement carbon. Pearlite is formed by the structural union of ferrite and cementite in definite proportions, not being a com- pound, but simply an intimate mixture. It appears in two forms, granular and lamellar, the former being seen in steel which has been worked or reheated to a low heat, while the latter is found only in steel which has been cooled slowly through the critical range. It is to the lamellar variety that its name is due, the struc- ture by oblique light giving an effect like mother of pearl. In addition to these common forms there are two others, troostite and sorbite, of which little is known at present. As steel cools through the critical range, the transition from martensite to one of the forms contained in unhardened steel is not abrupt, but appears to be in two steps. Thus by quenching during this critical change a new condition will be obtained troostite and if this quenching takes place at the end of the critical range in cooling, a second effect is noticed, which is called sorbite. Quenching in lead, or reheating quenched steel to a purple tint may also produce sorbite, and Osmond states that when small pieces are cooled in air the chilling is sufficiently rapid to prevent the complete transformation into ferrite and cementite, some sorbite being formed. Thus aus- tenite, martensite and troostite are found only in steel quenched at or above the critical range, while ferrite, cementite, pearlite and sorbite, are characteristic of unhardened steel. It is difficult to develop troostite and sorbite in the process of etching in such a way that they will be clearly visible under the microscope, and it has HEAT TREATMENT. 405 already been stated that the conditions of their existence are uncer- tain, so that for practical purposes these two forms may be neg- lected until their properties have been further studied, and since the conditions under which austenite is formed are never realized in practice, this also may be passed by. Ferrite and cementite present very nearly the same appearance, but they never occur to- gether, and as they differ very much in hardness it is easy to dis- tinguish them, for ferrite is pure iron and if the point of a needle is drawn across it the surface will be easily scratched, while cemen- tite is a compound of carbon and iron and the point will make very little impression. It is generally admitted that ferrite is structure- less even under the highest powers of the microscope. Pearlite is an "eutectic alloy," a term which may possibly not be familiar to all readers. An eutectic alloy is formed by the simul- taneous crystallization of different metals in a liquid mixture, as for example a mixture of copper and silver. These metals form an alloy in the proportions of 72% silver and 28% copper at a tempera- ture of 770 C. (1418 F.), and if a melted mixture of these two metals contain any different proportion than this, and if it be allowed to cool, the element in excess of this proportion crystallizes out, the crystals remaining uniformly distributed throughout the molten mass. When the critical point of 770 C. is reached, the alloy of 72 silver and 28 copper becomes solid, and entrains the innumerable crystals of the excess element which have separated from the mother liquid. A little consideration will show that under the microscope the element solidifying first and the eutectic alloy will occupy areas exactly proportional to the original constitution. In steel at high temperatures the same conditions exist as. in the mass of silver and copper just described, save that the elements are in what is called "solid solution," martensite at the lowest critical point going through a transition into ferrite and cementite. The element in excess separates by itself, and when the proper relation has been established the ferrite and cementite crystallize together in most intimate mixture to form pear lite. As stated pre- viously, the excess of cementite or ferrite begins to form by itself at the upper critical point, a small amount being found in steel quenched just below this, and at the second point this amount is increased, but this excess is always small except in the case of low carbon steel. 406 METALLURGY OF IRON AND STEEL. The foregoing argument may be summarized as stated by Sau- veur : (1). All unhardened steels are composed of pearlite alone, or of pearlite associated with ferrite or cementite. (2) Without taking into consideration austenite and troostite, hardened steel is composed of martensite alone, or of martensite associated with ferrite or cementite. (3) Ferrite and cementite cannot exist together in the same piece of steel. (4) The presence of the lamellar variety of pearlite is almost certain proof that the steel has been annealed. Following the proposition that ferrite is iron free from carbon and that cementite is a compound represented by the formula, Fe 3 C, it is evident that in very low steels, say ranging from .02-.10 carbon, the structure will be almost entirely ferrite, and that in steel of 2.00 per cent, carbon there will be an excess of cementite. There will therefore be one point of carbon content at which the component ferrite and cementite will both be satisfied, which is to say that the original proportion will be that of the eutectic alloy. This occurs in a pure steel containing about .80 per cent, of car- bon, the micro-structure of this grade showing no ferrite or cemen- tite. Late investigations seem to prove that in hypereutectic* steels, that is, those containing more than .89 per cent, of carbon, the upper critical point, A 3 , follows the curve, SE, in Fig. XV-H. This is the point at which cementite begins to form and, according to Howe and Roberts-Austen, progressively separates out within the martensite in cooling and forms a network whose coarseness is proportional to the temperature to which the steel has been heated. No break in the cooling curve has been noticed, but the first appear- ance of cementite is considered to mark the point, Ar 3 , while Ar 2 and Ar! are as given in diagram Fig. XV-A. Tables taken from Prof. Sauveur give results as shown in Tables XV-L and XV-M, the numerals being intended to represent per cent, of volume, since if a body containing an infinite number of particles, uniformly distributed, is cut by a plane, the ratio of the sum of the small areas to the total area is equal to the ratio of the volume of the small particles to the total volume. Theoretically, of course, this is not true of a mass of steel, but for practical pur- poses it is correct. HEAT TREATMENT. 407 The different photographs in Fig. XV-B represent the appear- ance of steels of different carbon content. No. 3 is a steel con- taining 1.39 per cent, of carbon and is from a bar in the condition in which it left the rolls. It shows a pearlite grain surrounded by walls of cementite. Nos. 4 and 5 represent lamellar and granular TABLE XV-L. Theoretical Micro- Structure of Carbon Steels. Carbon per cent. Pearlite. Fe. Cem. 100 .10 12 88 .40 50 50 .70 87 13 .80 100 1.00 97 3 1.20 93 7 2.50 71 29 TABLE XV-M. Micro- Structural Composition of some. Quenched Carbon Steels. Carbon, per cent. Quenched above Ar s Quenched between Ar 3 and Ar 2 . Qvv nched between Ar 2 and AT I . Quenched below Arj or slowly cooled Mart. Fer. Cera. Mart Fer. Cem. Mart. Fer. Cem. Pearl. Fer. Cem. 0.09 0.21 0.35 0.80 1 20 2 50 77 23 27 73 11 31 56 89 69 44 ' 10 23 50 .100 92 77 90 77 50 8 23 Quenched above Ar 2 . Martens! te. Ferrite. Cementite. 100 100 Quenched above Ar x . Martensite. Ferrite. Cementite. i-00 94 80 6 20 pearlite respectively. No. 6 is a steel containing .67 per cent, of carbon, the appearance of which is similar to No. 3, but there is really quite a difference, in that there is not a sufficient amount of carbon to form the eutectic alloy. Consequently there is an excess of ferrite and this forms the walls, whereas when the carbon ex- 408 METALLURGY OF IRON AND STEEL. ceeds .89 per cent, there is an excess of cementite, which therefore forms the walls. Nos. 7 and 8 contain very little carbon, No. 8 being especially soft, showing almost no pearlite. Index of Micro-Photographs, Figs. XY-B to G. Magnification. No. Diameters. 1 Austenite wo 2 Martensite. 175 3 Pearlite with cementite walls C=1.39 75 4 Lamellar pearlite 900 6 Granular pearlite 900 6 Pearlite with ferrite walls C=0.67 75 7 Mild steel C=0.20 showing ferrite and pearlite 75 8 Ferrite C=0.03 75 9 Cold worked steel showing lines of flow and in center actual rupture 30 ,10 Nickel steel roll, fracture in relief 1 11 Same steel as No. 10, polished and etched 50 12 Nickel steel roll shown in No. 10, annealed at 800 C 50 13 Small piece of same nickel steel roll annealed three times at 850, 800, 750 C 50 14 Special high carbon steel, unannealed 50 15 Special high carbon steel, annealed 50 16 Carbon steel casting, unannealed 20 17 Same steel as No. 16, annealed 50 18 Same steel as No. 16, annealed twice 50 19 75-lb. T rail, center of head ; broken in service 46 20 75-lb. T rail, center of head ; broken in service 46 21 85-lb. T rail, center of head ; broken on drop test 46 22 100-lb. T rail, center of head ; finished at 1000 C 46 23 85-lb. T rail, center of head ; "hot rolled" 46 This rail was one of two from the same ingot rolled under different conditions. See Section XVe, Par. 1 and 2. 24 85-lb. T rail, center of head ; "cold rolled." See No. 23 46 25 107-lb. girder rail. Sec. 228, P. S. Co 44 26 3 07-lb. girder rail, Sec. 228, P. S. Co 46 27 90-lb. girder rail, Sec. 200, P. S. Co 48 28 90-lb. girder rail, Sec. 200, P. S. Co 46 29 70-lb. T rail, Sec. 237, P. S. Co., center of head 46 30 70-lb. T rail, Sec. 237, near surface 46 31 M. S. Co. 100-lb. T rail, center of head 46 32 M. S. Co. 100-lb. T rail, near surface 46 33 M. S. Co. 85-lb T rail, near surface 46 34 M. S. Co. 85-lb. T rail, "hot rolled." See No. 23 4G 35 M. S. Co. 85-lb. T rail, near surface, "cold rolled." See No. 23. . 46 36 Bessemer steel, C=0.45. Finished at 490 to show effect of cold rolling 50 37 Ingot structure, C=0.06 20 38 Center of 1" round, C=0.06 75 39 Near surface of same piece as No. 38, showing loss of carbon by heating 75 40 Ingot structure, C=0.47 20 41 Bloom 8"x8", rolled from 32"x38" ingot ; C=.40 75 42 Billet 2"x2" hammered from bloom shown in No. 41 75 43 Section of a finished angle 75 44 Ingot structure, C=1.00 20 45 1 '' round rolled from ingot shown in No. 44 50 HEAT TREATMENT. 409 SEC. XVk. Effect of work on the structure of soft steel andj forging steel. Steel as usually cast, cooling slowly from the liquid state with no work done upon it, forms in crystals and shows in general the same structure throughout. The outer skin has a structure different from the rest of the mass, as it cools quickly and is under heavy strains as long as any of the metal is hot, and there is also an area of abnormal crystallization at the top of the ingot due to segregation, but the greater part of an ingot is of the same general crystalline character. Rolling tends to break up this grain and prevent further growth during the process, but immediately after cessation of work the formation of grains begins and con- tinues until the metal has cooled to the lower critical point. Hence it is evident that the lower the temperature to which steel is worked the more broken up the structure will be, but on the other hand if the rolling be continued below the critical point, the effect of cold work will be shown and strains will be set up which will make the piece unfit for use without annealing. Consequently it is necessary to stop the work somewhat above the critical point and in practice with large pieces it is customary to finish some 150 C. to 200 C. above this point, since the metal becomes so stiff at the lower temperature that the wear and tear on the rolls is excessive. In blooms, billets and such hard steels as are to be reheated for hardening, the need of an extremely low finishing temperature is not so evident. If the grain be reasonably fine, the metal is solid and dense, and the crystallization of the steel when put in service will be determined by the final heat treatment. This will be taken up more in detail in Section XVm. It would appear that the smaller the piece the finer the grain, and this arises partly from the necessity of finishing a large piece while the center is still hot and partly from the slower rate of cooling of the large piece. In No. 37, Fig. XV-G, is shown the micro-structure of a low-carbon ingot magnified 20 diameters and in Nos. 38 and 39 the same grade of steel rolled into I" rounds and magnified 75 diameters. These last two are the center and outside respectively of the same piece and show the effect of a high temperature in burning the carbon of the steel near the surface. The dark element in No. 38 is pearlite, the light is ferrite. It will be noticed that very little pearlite is shown in No. 39. This is in accordance with the ex- planation in Section XVm, where it is shown that if the carbon were partly burned away it would leave just so much less cementite 410 METALLURGY OF IRON AND STEEL. to mix with the ferrite to form pearlite, and consequently leave more ferrite free. In No. 40 is shown the structure of an ingot containing 0.47 per cent, of carbon magnified 20 diameters. No. 41 gives the structure of an 8" bloom rolled from a 32"x38" ingot, and No. 42 a test from the same bloom hammered to a piece 2" square. These last two are magnified 75 diameters, and it should be noted that the areas of the ingot structure shown in the photo- graphs are to the areas of the finished pieces as one to fourteen. Figs. 44 and 45 show the structure of a steel containing about one per cent, of carbon before and after rolling, the first being a section from a 16"x20" ingot, the latter a section from a piece 1" in diameter cooled on the hot bed. It will be seen that the grain is well broken up without any sign of cold work, and the bar is con- sequently 'in very good condition for the hardening and tempering to which such hard steels are usually subjected. This bar was taken at random from the hot bed at Steelton. If steel is worked below the critical point, strains are developed which injure the metal and may even rupture it. In No. 9, Fig. XV-B, is shown a piece of forging steel magnified 30 diameters. This illustrates the distortion of cold work, and the black line in the middle of the print is a crack where the tension became greater than the cohesion of the metal. SEC. XY1. Effect of work upon the structure of rails. Nos. 19 and 20, in Fig. XV-D, show the micro-structure of two rails which broke in service. No data are available as to how long they had been in use, but it is probable that it was only a short time. No. 21 is an 85-lb. T rail, which broke under the drop test. These three fractures, as well as all the other photographs, are selected not as exceptional, but as representative of what will usually be found un- der similar conditions. Fig. 22 is made from a heavy rail section finished at a temperature of 1000 C., and it will be noticed that its appearance is almost if not quite the same as that of Nos. 19, 20 and 21. In Nos. 23, 24, 34 and 35 are shown the results of some experiments performed by Mr. S. S. Martin at the works of the Maryland Steel Company at Sparrow's Point. An ingot was rolled into blooms and two adjacent blooms were rolled into rails without further heating, the first being held before rolling in order to allow it to cool so that all work should be done at as low a temperature as possible, without, of course, reaching the lower critical point, while the second was rolled as quickly as possible through all the HEAT TREATMENT. 411 passes except the last, but was then held at the finishing pass minutes, the result being that both pieces went through the finish- ing pass at the same temperature, which was about 750 C. I will designate as the "hot-rolled rail" the one which was rolled rapidly, but which was cooled' down just before the finishing pass, and as the "cold-rolled rail" the one which was rolled at a lower temperature during the whole operation. No. 34 represents the micro-structure of a portion of the hot rolled rail at a place very near the surface and No. 35 the structure of the cold-rolled rail at a similar place. It is evident that a superficial examination of photographs, without any knowledge of certain fundamental conditions, might lead to the conclusion that the two methods of rolling gave identical results, but the testimony of Nos. 23 and 24 proves quite the opposite. No. 23 is from the center of the head of the hot-rolled rail and No. 24 from the center of the cold-rolled rail, and it is clear that there is a radical and fundamental difference in the results, the reason for which is per- fectly clear. The finishing pass in almost every set of rolls does very little work, for it is unusual to have over ten per cent, of reduction upon the piece, oftentimes there being much less, while in all other passes, save one regulating the height, it is usual to have from twice to three times as much. Consequently the effect of the last pass does not penetrate to any great depth. Such a penetration is necessary if the grain is to be broken up, for the head of a heavy rail offers a thicker mass of metal than is found in almost any other structural shape, and the very fact that it is considered necessary to hold a rail before finishing proves that the grain needs to be broken. If the rail is at a sufficiently low temperature the grain will not grow coarser as the rail stands, and the rail might as well be finished at once; but if it is at a high temperature and the grain is coarse, then it will do no good to hold it before the last pass, or to shower it with water, for this will merely perpetuate the coarse crystalliza- tion that exists. The holding of the rail therefore before the last pass is a delusion ; it gives a lower finishing temperature and a low shrinkage, and it renders possible a very nice looking photograph from a piece of the outside skin, but it does not give any of the fundamental good qualities which should accompany such a finish- ing temperature, and which will accompany it if the temperature of the finishing pass is a true exponent of the rolling conditions. The 412 METALLURGY OF IRON AND STEEL. attempt to estimate the structure of the rail from the amount of shrinkage is simply putting the cart before the horse; it is much like the practice in vogue a few years ago of rolling octagon spring steel and then defacing the bar by hitting it with a hammer to make it resemble the bars turned out by the tilting hammer. This tilting consisted in a rapid succession of blows continued during the cooling of the piece until a very low temperature was reached, and by this means the crystalline structure was rendered very fine and the steel was in the very best condition. The rolls did not finish the bar as cold, nor did the effect of rolling penetrate as thoroughly as the blow of the hammer, and this lack could hardly be atoned for by duplicating an incidental accompanying condi- tion. There will always be some difference between the structure of the center of the head of the rail and the portion near the surface, but if the rail is rolled at a proper temperature during the passes when considerable work is put upon the piece, this difference will not be serious. No. 25, in Fig. XV-E, shows the center of the head of a girder or tram rail weighing 107 pounds per yard, and No. 26 shows the surface of the head. No. 27 shows the center of the head of a 90-pound girder rail and No. 28 the surface. No. 29 is the center of a 70-pound T rail and No. 30 the surface. All these were rolled at Steelton on regular orders and it will be noted that while there is a difference, the structure of the center is very good. Fig. XV-F shows the structure of T rails rolled at Sparrow's Point at the works of the Maryland Steel Company and represents the best modern practice. No. 31 is the center of a 100-pound T rail and No. 32 the surface; No. 33 the center of an 85-pound T rails, these structures representing the regular practice at the works. Nos. 34 and 35 have already been discussed as hot-rolled and cold- rolled rails. No. 36 represents the structure of a small test bar of rail steel which was rolled for the purpose of this experiment as cold as the strength of the rolls would allow, the finishing tem- perature being 490 C. (915 F.), which is considerably below the critical point, as shown by the lines of work appearing in the photo- graph. This evidently is the finest structure obtainable, and it may be used as a standard by which to estimate the condition of the other pieces. All the photographs 'in this rail steel series are cross- sections that are magnified forty-six diameters. SEC. XVm. Effect of heat treatment upon the structure of cast- HEAT TREATMENT. 413 ings. It has been proven by many investigators and is generally acknowledged that in heating steel through the lowest critical point the crystalline structure is obliterated, the metal assuming the finest condition of which it is capable. Above this point the size of the grain increases with the temperature. There is a difference of opinion as to whether the increase in size takes place during the heating or at the moment when cooling begins, but it is un- necessary to determine this question, the general proposition being true that the higher a piece of steel is heated above this point the larger the grain becomes. At the corresponding point in cooling, the structure - ceases to change, except in very soft steel, as shown by Stead, and any size of grain is retained and cannot be changed by heat treatment below this point. There is, however, a change from hardening to cement carbon, which may take place at comparatively low temperatures. This is the principle on which the tempering of steel is founded, quite a definite amount being changed at temperatures which are represented approximately by the color of the bar. Cement carbon is that form which confers the softest possible condition and great- est ductility, while hardening carbon gives the condition of greatest hardness. Hence the temper is drawn by every rise in tempera- ture. At the lowest critical point the change from cement to hardening carbon takes place almost instantly, all carbon above this tempera- ture being of the hardening variety, but the reverse change in cool- ing appears to require a certain length of time. This is the ex- planation of hardening by quenching, the more rapidly the steel is cooled through this point, the less being the chance of the carbon to change its state. A sudden cooling in ice water prevents any change, while annealing is effective only in proportion as the time of exposure to this temperature was long or short. Since fine structure and cement carbon are the principal factors of toughness and ductility, both of which are the aim in annealing, it would seem that the best method of tempering would be to heat to the lowest critical point and not higher, and quench from this heat and subsequently draw the temper. Similarly the best way of an- nealing, since the reverse change takes place several degrees below this, would be to cool at once to just above this lower point and allow several hours for the metal to cool past the critical tempera- 414 METALLURGY OF IRON AND STEEL. ture, and long enough from this point to the cold state to prevent the setting up of strains from too rapid cooling. Practically, however, it seems to be necessary to heat consider- ably above the lowest critical temperature in order to insure the thorough breaking up of the cell walls to allow the enveloping form to permeate the grain. This arises from the fact that the changes by which ferrite is formed attain their maximum effect only when the metal is subjected to a range of temperature which includes the three critical points. When steel cools slowly a certain amount of ferrite forms at the upper point, Ar 3 , an additional amount at the second point, Ar 2 , while the principal change occurs at the lowest point, Ar r Thus if the metal be considered as a solid solution, it may be said that crystallization takes place at the upper point, the solution of martensite becoming more concentrated. When the steel is heated, as in the case of annealing, the reverse phenomenon takes place, for at the lowest point the grain is broken up, the pearl- ite becoming martensite, somewhat diluted by the portion of ferrite which it takes up. If now the piece be cooled slowly without further heating, the resulting structure will be quite different from the original. The size of the grains will be much smaller and the piece will therefore be in much better physical condition, but there will still remain room for improvement, for throughout the mass will be found a certain proportion of ferrite, corresponding to the amount which, as already explained, is transformed at the higher temperatures of Ar 2 and Ar 3 . In order therefore to thoroughly disseminate the ferrite and encourage to the greatest extent the formation of martensite, it is necessary to heat to the upper critical point Ac 3 . This high tem- perature, however, gives rise to a somewhat larger grain than if the lower critical point, Ac 1? had not been exceeded, so that while there is a gain in the extent of the transformation, the grain of the resulting steel is coarser and there is consequently a loss in strength. The best result is obtained by combining the two methods, the steel being first heated to the upper critical point, Ac 3 , and allowed to cool slowly, by which complete transformation is effected, and then reheated just above the lower critical point, Ac 1? by which the grain is rendered fine and all strains obliterated. In case two heatings are out of the question, it is generally better to heat to the upper critical point, as it is preferable to have a slightly larger grain with a fine division of the microscopic forms, than to have a piece HEAT TREATMENT. 415 of metal of somewhat finer grain but much less homogeneous. Con- siderable care must be exercised in heating pieces which are not to be machined after treatment, since at a high temperature the carbon near the surface of steel is burned out to an appreciable depth by the action of the flame, unless the metal is protected in some way from oxidation. An effect of this kind may be noticed under the microscope with little difficulty. If the carbon has been driven off it follows that there is less cementite left to combine with ferrite to form pearlite when the metal is cooling through the critical point. Consequently there will be less pearlite formed in the oxidized sur- face than in the remainder of the piece. This effect is shown in Xos. 38 and 39, these being the center and the outside respectively of a soft steel bar. In Xo. 11, Fig. XV-C, is shown a large pearlite grain surrounded by a thick wall of ferrite. This represents the micro-structure of a 28-inch steel roll casting containing .25 per cent, carbon and 3.5 per cent, nickel, which was put in service unannealed and broke within a few hours. In Xo. 10 is shown the fracture in natural size, and the photograph was made from the broken specimen with- out any polishing or other treatment. It is a striking illustration of intergranular weakness, the lines of rupture following almost entirely the ferrite envelope and leaving the individual grains in- tact. Xo. 12 shows the micro-structure of this broken roll after one annealing at 800, and notwithstanding the exceedingly coarse structure of the original casting the annealed micro-structure is quite fine and shows a grain outline very much broken up. It is probable that a second annealing would have almost obliterated the crystallization, and it would have been interesting to carry this on for several more heat treatments^ [but as this was impracticable a piece was cut off and heated successively to 850, 800 and 750 Centigrade and allowed to cool slowly with a complete destruction of crystallization as shown in Xo. 13. It should be noted that Xo. 11 and Xo. 12 are results obtained with full size pieces, and not with small tests, as is too often the case, under which circumstances the results are not always com- parable with the effect on a large piece. The two pieces were taken from the same relative positions and represent, it is believed, the structure of the roll. The casting conditions, so far as could be determined, were normal. The annealing was effected at 800 C. as registered by the pyrometer, it being necessary to consider that 416 METALLURGY OF IRON AND STEEL. this does not always represent the temperature exactly unless the "invisible" condition is obtained. No. 16 represents the micro-structure of a steel casting unan- nealed, magnified 20 diameters. It is almost impossible to give an idea of the structure in a small photograph, but the illustration shows parts of three grains, and like all the other reproductions, is typical. No. 17 shows the same casting after annealing. The picture is not all it should be, but by careful examination a re- markably small grain may be distinguished; the areas of pearlite and ferrite are indicative of an insufficient breaking up of the microscopic forms. No. 18 represents the casting after a second annealing. No. 14 and No. 15 show the structure before and after annealing of a special high carbon casting used in railroad work where ability to withstand shock is of prime importance. ** As stated in Section XVi, the second critical point is character- ized by a loss of the magnetic properties in heating; this point is very easily determined by using an electro magnet, the wires of which are connected with a sensitive galvanometer. The act of moving the magnet into and away from contact with the metal moves the needle of the galvanometer as long as the metal is mag- netic. It would seem as if this should be a good point to agree upon as the temperature to which castings shall be heated for an- nealing. Sufficient data are not available to state positively that such treatment would give the best results possible, but it seems quite certain that treatment on this line would give good structure and be a great improvement on most of the haphazard methods now in use. SEC. XVn. Effect of heat treatment on the structure of rolled material. In order to determine the effect of heat treatment on the structure of rolled material, tests were taken from finished angles, the general method of procedure being as follows : A piece five feet long was sheared from the angle and cut into five equal lengths. An ordinary test bar was taken from one of the legs of each piece in the same relative place and numbered from 1 to 5. From each of the extremes 1 and 5 a section was cut for the microscope and the bars pulled in the testing machine to prove that the piece was homogeneous. The bars, 2, 3 and 4, were treated in a muffle heated by an electric coil at temperatures varying from 625 C. to 890 C., the temperature in all experiments being taken by a Le Chatelier pyrometer. No attempt was made to heat HEAT TREATMENT. 417 the pieces quickly, as it was intended to work under normal con- ditions, the operation usually occupying from one to three hours. The bars were held at the high temperature only long enough to insure uniform heating and then cooled for several hours to about 350 C. A longer annealing would probably have given slightly different physical results on account of the more nearly perfect elimination of strains and transformation to cement carbon, but the difference would have been slight, and as the object was to determine the effect of heat on the structure it was unnecessary to consider this phase of the problem. Small sections were cut from the treated pieces, as well as from the untreated, and were polished and etched. They were invari- ably taken from the same relative position and etched on the surface representing the cross section of the angle. A great majority of these specimens when examined under the microscope showed well denned structures similar to those exhibited in Nos. 8 and 43. The orientation was apparently the same in both the treated and the untreated bars, and the size of the grains did not appear to be affected by the treatment, although bars from different heats showed considerable variation. It would therefore seem probable that as finely divided a grain can be produced by rolling as by any of the usual annealing processes, although there is room for further in- vestigation on this point. SEC. XVo. Theories regarding the structure of steel. There are several theories now before the scientific world to account for the hardening and the magnetic transformations in steel and the phenomena of the so-called critical points. It would be better per- haps to call them hypotheses, as they are in each case offered tenta- tively and as lines of thought on which to base experimental re- search. It is beyond the province of this book to enter into a full discussion of these various conceptions, but it may be well to give a brief summary of the most prominent. The carbon theory considers that the effect of hardening is due entirely to a change in the carbon contained in the steel. In com- mon with the other theories, it supposes that at temperatures below the critical point the carbon is in the state of cement carbon, com- bined with iron in the proportion Fe 3 C. At the lower critical point a change in carbon is supposed to occur, and since from tempera- tures above this point carbon steels are hardened by sudden cool- ing, the advocates of this theory have devised the name "hardening 418 METALLURGY 01 IRON AND STEEL. carbon/' The cause of evolution of heat at this point in cooling is considered to be the change from hardening to cement carbon, but no satisfactory explanation is given by this theory for the changes at the second and third critical points. The allotropic theory holds that the iron of the steel is in differ- ent allotropic forms between the different critical points, and that below the second critical point the iron exists as alpha iron, but at this point I eta iron is formed, and at the upper gamma, the carbon being diffused in the iron. The cause of the evolution of heat is explained by the change from gamma to beta iron at Ar 3 , from beta to alpha at Ar 2 , while at.Ar x the carbon combines with alpha iron to form Fe 3 C. The retention of a hard allotropic state of iron, this retention being helped by the presence of carbon, is considered to be the cause of hardening. The carbo-allotropic theory is similar to the allotropic theory, except that hardening is supposed to be due to the retention by sud- den cooling of a hard carbide of iron. The Phase Doctrine. Prof. Bakhuis-Koozeboom explains* the detail of the Phase Doctrine, a phase being denned as a mass chem- ically or physically homogeneous, or as a mass of uniform concen- tration. Thus he states that a phase may be liquid or solid, may be an element or a compound, or a homogeneous mixture of vari- able concentration. Carbon, alpha, beta and gamma iron, liquid solutions, solid solutions of carbon in gamma iron or martensite, cementite and ferrite are all phases, while pearlite is a conglomer- ate of phases. He gives a diagram shown in Fig. XV-H, which is intended to show the critical changes of alloys of iron and carbon containing different percentages of carbon at different temperatures. From this it may be seen that the area, PSTX, represents the structure of slowly cooled steels containing less than .89 per cent, of carbon, and SKLT the structure of high carbon steels cooled slowly. MOSP is the region between A x and A 2 , showing alpha iron, while GOM is that between A 2 and A 3 , beta iron. Above GOS, which is the line A 3 in Fig. XV- A, the iron is in the phase gamma, the micro-structure being 100% martensite. As shown by the curve, SE, the higher the carbon in the steel the higher the heat needed to prevent the separation of cementite. Thus m in a 1.00 C steel is the temperature necessary to hold in solution the excess * Zeitschrift fur Physikalische Chemie, Vol. XXXIV, 1900. 1. and S. Inst., September, 1900. HEAT TREATMENT. 419 of cementite. At about 1050 C., however, cementite as such dis- appears even in high carbon steels and the carbon is considered as being in solution in gamma iron. This is the point above which it is necessary to heat in order to obtain austenite, from which it is argued that austenite is carbon dissolved in gamma iron. 1600 1500 I400 r 1300 1200 noo c 1000 900 C 800 600 1 2 3456 ^s/1 1 ^xxj Ca rbon I >er Ce ! it D \ \ N N, q ^ / - \ "S . Liquid / X > Liquid v^ , s ^ U / + G liquid raphit e \^- Martens! e 7 3 { ^ *^B, / V C \ ^^*s _ -1B- *. Martt nsite 2 >?. Martensite + ^ *&'' X' ^/ F' Graphite G / ' Marte nsite^ Ceine ntite . ? H o V /m 9 d; p a V / J 1 , | <* Q FQI-T ite + Pearl ite+C jment te o" Pearlite N 7 L It FIG. XV-H. GRAPHICAL KEPRESENTATION OF THE PHASE DOCTRINE. Martensite is considered as a solution of Fe 3 C in allotropic iron, being a saturated solution in steel containing about .89 per cent, carbon. Prof. Arnold has disputed the allotropic theory in several articles nnd has evolved an hypothesis of his own which he calls the "sub- carbide theory/' on the supposition that hardening is due to the, retention of a hard sub-carbide of iron Fe 24 C. These theories will be found thoroughly considered in the vol- umes of the Iron and Steel Institute of the past few years. Enough is given here to show the variety of ideas, all of which have their strong and their weak points. CHAPTEE XVI. THE HISTORY AND SHAPE OF THE TEST-PIECE. SEC. XVIa. Differences in physical properties between the sur- face and the interior of worked steel. The first question that arises in the inspection of steel is the manner in which the test- piece shall be taken. In former days it was the custom to care- fully plane or turn a piece to a standard size, with a certain length between shoulders and a certain radius for the terminal fillets; but this method is both tedious and expensive with no correspond- ing advantages. It is still used in steel castings, for it is im- possible to cast a bar of sufficiently accurate section to be fit for a tensile test, and it is also used in the case of forgings where the piece is too large to be broken in full section, and when it is deemed advisable to carve a piece from the finished material. In all other work the test is either a part of the finished bar, as in the case of small rounds and flats, or is cut from the member, as in the case of angles, channels, etc., with two sides of the piece in the condi- tion in which they left the rolls. A sufficient length is taken to allow about 10 inches between jaws, and the readings are made on an 8-inch length which is defined by marks of a center-punch. A machined piece is generally inferior to a bar as it leaves the rolls. It is true that Table XIV-J shows no gain in ductility from continued stretching or polishing of the skin, but this is an entirely different matter from the full compression which the outer surface of a bar receives in the last pass. In a series of tests made at Chester, Pa,, by the United States Government* in 1885, the ma- chine was not powerful enough to" pull a seven-eighth-inch round, so that rods of this size were turned down to three-quarter-inch in diameter. The comparative results are given in Table XVI-A r the figures in each case representing the average of 14 heats which were tested in both diameters. * Report of the Naval Advisory Board on the Mild Steel used In the Construc- tion of the Dolphin, Atlanta, Boston and Chicago; 1885, pp. 81, 82. 420 THE HISTORY AND SHAPE OF THE TEST-PIECE. TABLE XVI-A. 421 Comparative Physical Properties of 94-inch Rolled Rounds in their Natural State, and %-inch Rounds of the Same Heats Turned Down to %-inch. Condition of bar. Ult. strength ; pounds per square inch. Elongation in 8 inches; percent. Reduction of area; per cent. |4 inch natural, J| inch turned to 3 4 inch, 65764 66088 27.53 25.30 42.7 42.0 The pieces cut from the seven-eighth-inch bar are evidently in- ferior to the three-quarter-inch tests, although it will be shown in Table XVI-K that the larger bar should give the better elongation. It is probable that the inferiority is due to the removal of the best part of the piece in the operation of turning. This phenomenon is more marked in larger sizes, as will be shown by Table XVI-B, which gives the results on bars cut from forged bridge-pins. TABLE XVI-B. Physical Properties of Test-Pieces %-inch in Diameter, cut Forged Rounds. Size of Ingot, 18x20 inches. Pennsylvania Steel Company, 1893. from rt . -2 2 T M ff .^ oe d ! 1 S In Place from which test was taken. to p, 11 J aj "3 V ' ^ rt 2 CH o d 'a fi & ~ -1^ to i? 2*3 fcs B S"y IH S tc " d S2 S -^ 5 ft5 5 ft " ft 5 o P K H 03 i Bin. At a depth of 1 inch from outside. At a depth of 2 inches from outside. The central axis. 62720 58100 58100 82870 29170 81490 21.50 22.25 20.25 40.4 87.5 84.1 52.4 50.2 54.2 10 in. At a depth of 1 inch from outside. At a depth of 2^ inches from outside. The central axis. 66070 62750 60900 87080 85670 82140 19.50 18.00 19.50 33.9 82.7 23.8 56.1 56.8 52.8 Preliminary test of same heat from 6 in. ingot I 42250 I 26.25 I 41.7 I 66.1 SEC. XVIb. Physical properties of strips cut from eye-bar flats. Similar differences will be found if test-pieces be cut from dif- ferent parts of rolled bars such as are used for making eye-bars. This will be illustrated by Table XVI-C. These results display considerable uniformity in the higher strength of the test bars which were rolled from the large ingot, 422 METALLURGY OF IRON AND STEEL. but the number of specimens is not sufficient to fully establish the fact. Such a comparison is often invalidated by certain unknown factors, for if the test bar be finished hot and the "flat" cold, the relation may be reversed. Table XVI-D shows the comparative results on nine heats of steel made at one of our large steel works, and will illustrate how widely the preliminary test may differ from the finished bar in individual cases, while the average of the two is nearly the same. In the light of such facts it seems absurd to reject a heat of steel because the preliminary test falls a few hun- dred pounds below an arbitrary standard. TABLE XVI-C. Physical Properties of Test-Pieces of Different Section Cut from Boiled Flats, together with the Results on %-inch Rounds of-the Same Heats Rolled from a 14-inch Square Ingot. 1.1 = edge of bar; 2, 2=%-inch rounds cut on a machine; 3 = center of bar; 4 inch round rolled from an ingot. p. 5 '1 - 91 a.-& G * J3 H "^TJ" o3 "< S'S oog fi tH ljfl|s| O A OP, H 3 f^- 2 H* iii 5S| fccftfl ||| o ! ~A M O *s %ti 3*i s ss-g a - o o ^ ~ 6 o3 ^ p^ & aj p, jj H t 3 H P3 Cut from T V in ch and f-inch angles .... Rolled from 6-inch test ingot . . 39 39 41J300 42270 60190 60200 28.89 26.44 68.0 424 Cut from T 7 B -inch and i-inch angles .... Rolled from 6-inch test ingot 46 46 40170 43070 60660 61360 29.05 25.01 56.4 40.0 Cut from ^g-inch and f-inch angles .... Rolled from 6-inch test ingot 37 37 39710 42990 61520 62930 28.96 23.10 53.6 88.2 It is the custom at Steelton to make such a preliminary test on every charge, but this is done merely to classify the metal. If the bar is rolled under proper conditions, its ultimate strength rep- resents the ultimate strength of the finished material, and without regard to any results on elongation or other qualities, the steel is used or laid aside. We are perfectly willing that the inspectors should see all the results, but we claim that these records have nothing to do with * The Inspection of Materials of Construction in the United States. Journal I. and 8. I., Vol. II, 1890, p. 316. t See discussion of my paper on Specifications for Structural Steel. Trans. Amer. Soc. Civil Eng., April, 1895. THE HISTORY AND SHAPE OF THE TEST-PIECE. 427 the acceptance or rejection of the material. In other words, this test is our own work, while the business of the inspector is to test the material that he buys as fully and carefully as he may wish, \vithout regard to whether a small test ingot has or has not fulfilled certain requirements, or whether it has been made at all. Our experience in comparing results from the preliminary test with those from the finished material, differs radically from that recorded by Mr. Hunt,* although we agree on the important point that the ultimate strength remains nearly constant. Table XVI-H compares the data obtained from a large number of charges of acid open-hearth steel having a tensile strength between 56,000 and 64,000 pounds per square inch. They were all rolled into angles and the charges are grouped according to the thickness of the fin- ished material. The great inferiority of the tests from the 6-inch ingot is easily explained. It is very difficult to cast small ingots so that they will not be scrappy, and the bars rolled from them will oftentimes contain flaws; consequently we break down the ingot to a billet two inches square and chip out the flaws, after which the piece is reheated and gives a perfect bar. It does not receive sufficient work to ensure good elongation, but this is of no consequence, for it is only the strength of the material which is under investigation, and in this respect the results are found to be strictly comparable with the finished material. SEC. XVIg. Comparative physical properties of rounds and ftats. It has been mentioned that the properties of a flat bar are different from those of a round, and it will not be unprofitable to investigate the relation. The points involved are three : (1) The percentage of work on the piece. (2) The finishing temperature. (3) The shape of the piece. (1) The amount of reduction from the bloom or ingot should not play too great a part in the problem, for it is the duty of the manufacturer to so conduct the operation that every piece, no matter how large, shall have sufficient work. But it must be con- sidered that a large section, a 9-inch round for example, cannot possibly be finished under the same thorough and permeative com- * Loc. eft. 428 METALLURGY OF IRON AND STEEL. pression that can be put upon a bar only one inch in diameter or upon a thin flat. (2) It is the business of the rolling mill to so arrange that every piece is rolled at a proper temperature, but it will be recognized as ITP practicable to finish bars of all diameters and thicknesses under identically the same conditions. (3) The shape of the test-piece has an influence upon the nature of the results, but it is often difficult to isolate this relation from the effect of work and finishing temperature. Same bars annealed. S3 Natural bars. Oa Oa ''a> WOO MIX CC X rrss ii oa " CD WW oa oa WI oa / ns 28 oa 85 ss Limits of ultimate strength; pounds per square inch, Kind of steel. No. of heats in average. II p O p ifiS XO gil ill Wp S o 5" I CD III' S CD tr 1 to H 0*0 i I THE HISTORY AND SHAPE OF THE TEST-PIECE. 429 The separation of these three intertwining influences is a com- plicated problem, the nature of which will be illustrated by Table XVI-I, which gives the results obtained from a large number of heats by cutting two billets from the same ingot and rolling one into a round and the other into a flat. All the lessons of this table are not written on its face, but an examination discloses the following facts : (1) Taking into consideration both natural and annealed bars, there are 18 comparisons between rounds and flats. The ultimate strength. is less in the flat in every case. The elastic limit falls in 17 cases, and the gain in the exception is slight. The elongation is raised in 16 cases, while in the two exceptions the loss is small. The reduction of area is lowered in 14 cases and raised in four. The elastic ratio is lowered in 15 cases, while in the exceptions the increase is small. TABLE XVI-J. Comparative Physical Properties of Bound and Flat Bars in the Natural and Annealed States. Average of all heats given in Table XVI-I Condition of bar. Shape of bar. Gain=+ Round Flat in flat. Ultimate strength; pounds per square inch, Natural Annealed 66679 62015 65911 59567 768 2448 Elastic limit ; pounds per square inch, Natural Annealed 46588 39633 45268 87106 1320 2527 Elastic ratio; percent., Natural Annealed 69.87 63.91 68.68 62.29 1.19 1.62 Elongation in 8 inches ; per cent., Natural Annealed 26.48 27.16 28.22 28.78 + 1.74 + 1.57 Reduction of area ; percent., Natural Annealed 54.98 61.98 54.05 58.12 -O.93 3.86 (2) Comparing the loss of strength in passing from round to flat, as shown in Table XVI-J, there are nine possible comparisons between the loss in the natural bar and the loss in the annealed piece. The ultimate strength falls more in every case in the an- nealed than it does in the natural bar. The elastic limit falls in six cases and rises to a much less extent in three. The elongation rises in five cases and falls in four. The reduction of area falls in all cases. The elastic ratio falls in five cases and rises in four. It will be found also that the exceptions and irregularities are 430 METALLURGY OF IRON AND STEEL. not confined to any one kind of steel, so that it would seem proper to average the losses and gains in order to eliminate the errors due to the small number of heats in some of the groups. The results of such condensation are given in Table XVI-J, which shows the true average of all the heats and not the average of the groups. It is shown that the loss of ultimate strength from the round to the flat is very much greater in the annealed than in the natural bars and that the elastic limit more than keeps pace with it, as shown by the elastic ratio. The difference can hardly be due to the effect of varying work, for the round was reduced to 2.6 per cent, of the area of the billet and the flat to 4.7 per cent., the reduction in both cases being so heavy that the results should be uniform as far as this factor is concerned. The effect of the finishing temperature may be ignored in the case of the annealed pieces, and yet there is a difference of 2448 pounds per square inch in ultimate strength between the flat and round. The natural bars show less difference, which would indicate that the effect of the finishing temperature has raised the strength of the flat more than the round. This is contrary to the condition just noted that the reduction in rolling was less in the case of the flat, but it is in accord with the evident fact that a thin bar would cool faster than a round bar of somewhat less sectional area. The effect of the finishing temperature, therefore, was to raise the tensile strength of the flat more than it did the round, but not enough to overcome the difference in physical properties caused by the shape of the bars. The reduction of area is less in the case of the flat, and the difference is more marked in the annealed than in the natural bars. The elongation is higher in both kinds of flats than in the corre- sponding rounds, but the difference is greater in the natural bars. This appears at first sight to be an exception, but on further con- sideration it will be seen that a decrease in gain is equivalent to a loss, and this brings it in accord with the decrease in the ductility, as shown by the lessened reduction of area. The net result may be summarized as follows : (1) Flat bars differ from rounds in having less tensile strength, lower elastic limit, lower elastic ratio, greater elongation, and a slightly lower reduction of area. (2) This difference is caused not by reason of a different finish- ing temperature, but in spite of it. THE HISTORY AND SHAPE OF THE TEST-PIECE. 431 SEC. XVIh. Comparative physical properties of rounds of dif- ferent diameter. The variation in strength of bars of the same steel is not by any means confined to pieces of different shape, for it will exist in rounds of different diameters. In Table XVI-K are given the results on a large number of rivet rods where several tests were made from the same heat. All the charges were of the same quality of steel, ranging from .11 to .15 per cent, in carbon, .02 to .04 per cent, in phosphorus, and .022 to .038 per cent, in sulphur. TABLE XVI-K. Comparative Physical Properties of Bounds of Different Diameters,. Eolled from the Same Heats, Made by The Pennsylvania Steel Company. Each figure is an average of from 4 to 16 determinations. Heat No. Ult. strength; pounds per square inch. Elastic limit; pounds per square inch. Elongation in 8 inches; per cent. Reduction of area; percent. % in. %in. tn. %in. ^in. %in. %in. %in. 11478 11489 11550 11694 11796 11945 12006 12007 12519 2032 2073 60028 59170 58223 57833 57980 57456 57550 57943 58774 59670 59772 58215 57671 57707 58078 57517 56753 55878 57408 56106 56963 56425 40023 87333 39219 89373 88830 38498 38205 38752 89015 89050 39941 39433 37079 87482 38210 38288 37268 86485 37498 37485 86810 37007 29.52 29.81 29.73 32.45 30.14 29.81 29.58 30.38 29.80 29.67 30.25 30.63 81.96 30.40 30.75 31.04 30.59 80.58 81.44 ' 81.34 80.50 32.79 60.56 63.45 62.70 66.50 60.45 61.60 60.81 64.18 62.40 64.50 64.90 60.80 62.81 64.10 62.60 68.50 59.60 65.05 61.10 59.45 57.90 68.70 Av. 58582 57156 38931 37550 80.10 81.09 62.91 61.83 % in. %in. %in. %in. %in. %in. %in. %in. 11478 12007 1523 2200 60423 58120 59633 59421 60028 57943 55735 59435 41373 38200 42360 41276 40023 88752 88756 39860 29.44 30.16 30.06 30.00 29.52 30.38 31.66 30.31 65.40 64.55 64.22 64.86 60.56 64.18 65.40 64.65 Av. 59399 58285 40802 89348 29.92 30.47 64.76 63.69 % in. l^in. %in. Wn. %in. l^in. %in. l^in. 12334 57820 59813 37770 37298 30.85 82.25 63.15 61.55 %in. IT'S in. %in. 1 T % in. %, in. IT B B in. %in. l&in. 12368 62683 60480 89985 88576 30.69 31.97 62.23 53.80 l^in. VA in. l^in. 114 in. 11517 60633 ' 86770 82.02 54.8 The number of heats given in the table would not be sufficient to justify SL general conclusion if there were only a single bar of each heat, but it will be noted that each figure is the average of from 4 to 16 determinations. In the comparison of the three-quarter 432 METALLURGY OF IKON AND STEEL. and seven-eighth-inch rounds there were 112 tests of the smaller size and 94 of the larger, while in the comparison of the five- eighth and three-quarter-inch there were 32 tests of the former and 34 of the latter. No average is given where less than four tests were taken of the same size from the same heat. Comparing the seven-eighth-inch with the three-quarter-inch bars, it will be found that in the larger size the following changes occurred : (1) The ultimate strength was lowered in ten heats and raised in one, the average showing a decrease of 1426 pounds per square inch. (2) The elastic limit was lowered in all cases, the average show- ing a decrease of 1381 pounds per square inch; the elastic ratio was reduced from 66.5 per cent, to 65.7 per cent. (3) The elongation was raised in ten cases and lowered in one, the average showing an increase of 0.99 per cent. (4) The reduction of area was lowered in seven heats and raised in four, the average showing a decrease of 1.08 per cent. Comparing the five-eighth and three-quarter-inch, it will be found that in the larger size the following alterations have taken place : ( 1 ) The ultimate strength was lowered in three heats and raised a trifling amount in- one, the average showing a decrease of 1114 pounds per square inch. (2) The elastic limit was lowered in three cases and raised in one, the average showing a decrease of 1454 pounds per square inch ; the elastic ratio was reduced from 68.7 per cent, to 67.5 per cent. (3) The elongation was raised in every case, the average show- ing an increase of 0.55 per cent. (4) The reduction of area was lowered in three heats and raised in one, the average showing a decrease of 1.07 per cent. The consistent testimony of these records is corroborated by the data on the larger diameters. It is true that only one heat is given on each of these sizes, but it so happens that there were from twelve to sixteen bars in each case, and as the steel was of the same manu- facture in all particulars the results may be accepted as fairly comparable. It seems quite certain that larger bars will give a lower ultimate strength, a lower elastic limit, a lower elastic ratio, a better elongation, and a lower reduction of area. Some of these THE HISTORY AND SHAPE OF THE TEST-PIECE. 433 characteristics may be due to differences in finishing temperature, but the data on elastic limits show that the pieces were all rolled at nearly the same degree of heat, and such small variations, even if due entirely to rolling conditions, are not sufficient to account for the increase in the elongation. TABLE XVI-L. Effect of Changes in the Width of the Test-Piece upon the Physical Properties. Thickness in inches. No. of heats in av. Width of test-piece in N inches. 3 2 IK 1 % % Ultimate strength; pounds per square in. | 2 8 8 2 10 72510 72020 67945 73840 68111 73480 72220 68500 73550 G8224 73840 72420 68710 74530 67950 73250 72643 68220 73370 67890 74420 71563 68050 73520 68338 75440 73531 68940 76130 67442 True av. 30 69784 70059 70176 69968 69872 70578 .s. .Is *i."S GO f| o^*2 S ^3 33 a as -1$ 0> *^ q^ rj & S'fl 3 be be c a Sr^ 3 fl 3 M ._ PiVi > be > 0^00 IS M-i-s^ cl.5* Ilii 5&9 o ec j3 era) D.-S fetf 02 S" .3 ssa sill l s -gl sfp S * " I!li" I a 5 THE HISTORY AND SHAPE OF THE TEST-PIECE. 447 xu xable XVI-T is given an analysis of the records showing the number and percentage of bars in each division which give less than a certain percentage of elongation. TABLE XVI-U. Alteration in the Physical Properties of Steel by. Best after Boiling.* 1 Number of group. Limits of ultimate strength; pounds per square inch. Hand rounds. Guide Rounds. No. of Bars tested. Alteration. Gain= + Loss= Ij Alteration. Gala = + 'r. X (- 00 "S d d I i Less than 24 hrs. rest. More than 24 hrs. rest. Elastic limit; pounds per square inch. Ultimate strength; pounds per square Elongation in 8 inches; percent. Reduction of ares; per cent. Less than 24 More than 2 Elastic limi pounds pei square inc Ultimate st pounds pei square inc 11 d o 3 a 3*" Reduction o per cent. I II III IV V VI VII 55000 to 60000 60000 to 65000 65000 to 70000 70000 to 75000 75000 to 80000 80000 to 85000 85000 to 90000 6 10 22 24 85 16 8 10 86 86 47 80 16 + 719 453 170 166 314 165 + 92 +437 +596 +382 +688 +201 +767 +525 +.65 + .73 +.33 +.44 .81 +.42 +.46 + .32 + .99 +1.13 +1.45 +1.14 +2.33 +1.24 + .62 10 32 21 10 7 80 12 20 8 8 1207 471 + 302 809 + 213 + 885 180 + 197 + 107 + 36 +1.11 .25 + .66 + 1.06 + .29 + 2.14 + 2.07 + 2.95 + 6.76 + .44 Av. of all tests. 48 894 + 109 + .56 + 2.87 121 197 270 +507 + .99 The standards assumed are those which are specified for differ- ent grades of structural steel in Chapter XVIII. A study of the table will show that the number of rejections on longer lengths is fully as great as with the shorter bars,, and this proves that the de- crease in the specified elongation for an increase in length is not greater than should justly be allowed. In the bars made by "A" the rejections under Specification I amount to 4 per cent, in Bes- semer metal, and 10 per cent, in open-hearth; in those made by "B" they are 10 per cent, in the Bessemer and 20 per cent, in the open-hearth, while with "C" they are 23 per cent. Taking into consideration that the records cover only the products of large and well known works, and that all bars having a crystalline fracture * Notes on Results Obtained from Steel Tested Shortly after Rolling. Amer. Soc. Mech. Eng., Vol. IX, p. 38. 448 METALLURGY OF IRON AND STEEL. and those breaking in the eye were discarded, it must be acknowl- edged that the standard calls for good material. SEC. XVII. Alterations in the physical properties of steel by rest after rolling. In addition to the variations which may be caused by differences in the working of the test-piece and in its shape, there is probably another factor in the length of time which elapses between rolling and testing. This subject was investigated at The Pennsylvania Steel Works by E. C. Felton, now President of the Company, a condensation of whose work is given in Table XVI-U. The changes are not very strongly marked, but there seems to be consistent testimony of a molecular rearrangement, progressing for several hours after the bar is thoroughly cold, whereby there is a lowering of the elastic limit, and an increase in the ultimate strength, the elongation, and the reduction of area. SEC. XVIm. Probable error in current practice in determining the physical properties. It is the rule in most practical work that at least two sides of the test-piece are not machined, and hence it is impossible to make a perfectly accurate measurement. In order to find how great an effect may be caused by such errors and by differences in machines and the method of operating them, the experiment was tried of sending a bar from six different acid open- hearth heats to six different testing laboratories. The pieces were rolled flats, 2"x%", and each series was made up of one piece from each of the six bars, so that the only possible difference between the steel sent to the various places would be the difference between parts of the same bar. All pieces were tested in the shape in which they left the rolls without any machining, and although the edges were not perfectly smooth, they were so nearly true that only one operator referred to any difficulty in making a true measurement. Table XVI -V ex- hibits the results reported. The bars were tested by The Central Iron and Steel Works, Harrisburg, Pa.; The Baldwin Locomotive Works, Philadelphia, Pa.; The Pottstown Iron Company, Potts- town, Pa. ; The Carnegie Steel Company, Pittsburg, Pa. ; The Car- bon Steel Company, Pittsburg, Pa., and The Pennsylvania Steel Company, Steelton, Pa., but the identity of the different works is purposely concealed in the table under the letters A, B, (7, etc., to avoid invidious comparisons. An examination will show that there are quite important vari- ations in every one of the factors. Moreover, th& divergence is not THE HISTORY AND SHAPE OF THE TEST-PIECE. 449 the result of averaging erratic individuals, for whenever one average is higher than another, it is because the majority of the bars are higher when taken separately. TABLE XVI-V. Physical Properties of the Same Bars of Steel, as Determined by Different Laboratories. NOTE. All bars were rolled flats, 2"x%", and were not machined. Tested by Number of heat. A. B. C. D. E. F. 10027 58130 57880 58560 57710 57980 59230 10028 60790 60140 61740 60080 60660 61830 Ultimate strength; t pounds per square incti. 10030 10065 10066 10072 63560 60840 62840 61160 63330 euro 62700 62190 64530 62180 63480 61730 63180 60440 61970 61390 63450 61290 62630 61640 64280 62200 64170 62110 Average, 61220 61233 62037 60795 61275 62303 10027 42400 37350 88900 87490 39020 89730 10028 42200 37940 41400 88720 89730 41320 10030 43620 40780 42540 88940 40740 42770 Elastic limit; pounds per square inch. 10065 10066 10072 41540 42610 41400 38150 40350 37650 42250 42110 41770 88710 88905 88710 40210 40180 40950 41250 43140 89860 Average, 42295 38703 41495 38579 40138 41345 Elastic ratio, 69.1 63.2 66.9 63.5 65.5 66.4 10027 29.25 29.00 30.50 30.37 30.75 29-75 10028 30.75 30.00 32.00 29.75 81.00 29.50 Elongation in 8 inches; per cent. 10030 10065 10066 10072 29.00 29.25 29.25 30.00 29.00 28.75 32.25 33.75 31.00 30.50 30.50 34.25 28.12 80.25 29.12 29.37 29.00 29.50 33.25 80.75 28.50 82.50 29.50 29.00 Average, 29.58 30.46 31.46 29.50 30.71 29.79 10027 61.3 61.3 60.6 56.2 54.1 61.2 10028 63.1 59.7 62.9 58.9 53.3 62.3 10080 60.1 57.0 60.0 55.9 52.7 57.8 Reduction of area; 10065 61.8 58.4 60.6 56.7 55.9 61.6 per cent. 10066 61.5 59.9 60.9 54.0 52.5 60.0 10072 61.8 57.6 61.2 57.4 54.1 61.3 Average, 61.6 59.0 61.0 1 56.5 53.8 60.7 The variations in contraction of area may easily be explained, for the determination rests upon the most accurate measurements of an irregular broken body. In a bar having an original section of 2"x%", the fractured end will have a thickness of about 0.20 inch, and almost invariably will be of irregular form, the sides being concave rather than flat. A true estimation of the broken area could be made only by the most careful duplicate readings and by the aid of the calculus. These refinements are out of the question in practice, but the chances of error must always be con- sidered when a test-bar falls a little short of the requirements. 450 METALLURGY OF IRON AND STEEL. The variations in elongation may be partially accounted for by unlike methods of measurement, for if the original punch-marks be put on the outer edge of the bar, they will give a different read- ing after fracture than if they were put in the center line, owing to the unequal distortion of the bar. This complication would not occur in the case of a round test-piece. The differences in ultimate strength and elastic limit are due in some measure to slight variations in the original measurements of the bar. The elastic limit was found by noting the "drop of the beam/' this being the universal practice in American steel works and rolling mills. This method has been criticized by some investi- gators, who advocate an autographic device for registering the point where the elongation ceases to be exactly proportionate to the load. The introduction of such a system would result in endless con- fusion, since all specifications and contracts of the present day are based upon the elastic limit as now determined by the fall of the beam. The statement that the current method is especially inaccurate i& open to debate. In the series of tests given in Table XVI-V, it will be found that the elongation, as determined by different observ- ers, varies from 29.50 to 31.46 per cent., these figures being in the ratio of 100 to 106.6, or range of error of 6.6 per cent. The reduc- tion of area varies from 53.8 to 61.6 per cent., a ratio of 100 to 114.5, or a range of error of 14.5 per cent. The elastic ratio varies from 63.2 to 69.1 per cent., a ratio of 100 to 109.3, or a range of error of 9.3 per cent. Thus the determination of the elastic ratio is much more accu- rate than the results on contraction of area, and nearly as accurate as the results on elongation, both of which are determined by exact measurements made on the piece when at rest. It would be quite in order for reformers to apply their energies to the accurate deter- mination of the reduction of area and the elongation, instead of try- ing to substitute a new method for determining the elastic limit, especially when this method has been publicly branded as inaccu- rate.* As a rule the autographic device gives a slightly lower reading than is found by the drop of the beam ; thus in a paper by Gus. C. Henningf there are given the determinations of the elastic limit * Lewis. Trans. Am. Soc. Civil Eng., Vol. XXXIII, p. 351. t Trans. Am. Soc. Alech. Eng., Vol. XIII, p. 572. - THE HISTORY AND SHAPE OF THE TEST-PIECE. 451 on a series of tests, as found by the two methods. I have averaged the list of heats where both readings are given, and find that in thirty-eight cases the autographic record was 46.6 per cent, of the ultimate strength, while the beam dropped at 52.9 per cent. ; in the annealed bar the first method gave 51.6 per cent., and the second 56.9 per cent. Such a marked difference is not found in all cases, as shown by Table XVI-W, which gives the results obtained by E. A. Ouster, who at the time was connected with The Baldwin Locomotive Works, Philadelphia, Pa. In the case of the slow speed there is less difference between the two determinations of the elastic limit than is shown by Henning, while with the fast speed there is more. This matter of the influ- ence of the pulling speed upon the recorded physical properties is considered in the next section. TABLE XVI-W. Parallel Determinations of the Elastic Limit by the Autographic Device and by the Drop of the Beam.* No. of tests. Pulling speed. Ultimate strength ; pounds per sq. in. Elastic limit; pounds per square in. as determined by Elastic ratio; per ce.nt., as determined by Auto- graphic device. Fall of beam. Auto- graphic device. Fall of beam. 6 3 1 inch in 8 minutes. 4 inches in 1 minute. 56820 58870 36120 35890 37510 40530 63.6 61.0 66.0 68.8 The whole subject of the determination of the elastic limit was discussed in The Engineering News, of July 25, 1895. After re- viewing at great length the arguments presented by several engi- neers in previous issues, and after quoting from many authorities, the following conclusions were reached : "Having thus shown the impossibility of determining by micro- metric measurement the elastic limit, when it is defined as the point at which the rate of stretch begins to change, and the extreme variability of the position of the so-called' 'yield-point' with the method of running the machine and with the method of measur- ing and recording results, had we not better drop these new defi- nitions and methods of attempting to locate points whose position * From E. A. Custer, Baldwin Locomotive Works, Philadelphia, Pa. 452 METALLURGY OF IRON AND STEEL. is so extremely variable, and whose determination depends so largely upon the personal equation of the observer, and return to the good, old-fashioned definitions and methods? If for scientific purposes there is any need for determining microscopically that point at which the rate of stretch begins microscopically to change, let us call that point the 'limit of proportionality/ as Bauschinger did, and leave its determination to the college professors. "Let us keep the old term elastic limit with its old significance as that point at which a permanent set visible to the naked eye takes place, at which the rate of stretch increases so that the in- crease may be (albeit with some difficulty) distinguishable by the use of a pair of dividers and a magnifying glass, or more easily and certainly by the drop of the beam, or by the increase in the number of turns of the crank needed to produce a given increase in stretch. "For the purpose of determining this elastic limit let the testing machine be run by hand until the limit is passed and the record taken (or run by hand between the load of 30,000 pounds and the elastic limit), and then let the power gear be thrown in and the test completed in the present rapid fashion. Since the term 'yield point, is quite recent, and has no meaning essentially different from the words 'elastic limit' in time-honored practice, why need it be used at all?" These conclusions represent common sense in their summary dealing with the petty theories of enthusiasts, who are so wrapped up in the accurate determination of a micrometrical measurement that they ignore the more important variations inherent in the method itself, not to mention the still more overwhelming differences caused by changes in the history and shape of the material. I do not see, however, why it is necessary to revert to the primitiye and laborious method of driving a machine by hand when there is a power attachment with different pulleys. The speed should be lower during the determination of the elastic limit than can be used for breaking the piece, but a specification that this work must be done by hand is a confession of lack of ingenuity which is neither creditable to engineering science, nor justified by facts. SEC. XVIn. Effect of variations in the pulling speed of the test' ing machine upon the recorded physical properties. To find the effect of variations in pulling speed, ten different rivet rods were taken from an acid open-hearth heat. From each rod five bars THE HISTORY AND SHAPE OF THE TEST-PIECE. 453 were cut, and each one of these was broken at a different speed. The results are given in Table XVI-X. TABLE XVI-X. Effect of Variations in the Pulling Speed of Testing Machine upon the Kecorded Results. NOTE. Tests were made by The Pennsylvania Steel Company. Number of bars. Pulling speed; inches per minute. 4.50 8.CO 0.67 0.88 0.07 Ultimate strength; pounds per square inch. 1 2 8 4 5 6 7 8 9 10 61060 61140 61610 61500 61870 60200 60620 60520 61200 61030 C1860 60760 61230 61150 61580 59720 60140 59580 61100 60100 60640 59200 59910 58950 59960 59040 59290 58760 60000 59480 60240 59440 59680 59620 59910 58240 59380 58400 59620 59340 59660 59100 59100 59220 59760 59100 58200 58160 58870 59100 Av. 61075 60672 59523 59387 59027 Elastic limit ; pounds per square inch. 1 2 3 4 5 6 7 8 9 10 46640 44070 46920 46730 45080 44360 47500 44680 45000 46100 44930 43500 44680 45560 46300 43400 43670 44680 43440 43940 43240 44810 42220 42720 43120 41690 43090 42650 42380 43120 42650 41980 41270 41830 43430 40810 41880 41370 40860 41600 89610 89480 89250 40300 40480 89240 88950 89720 39720 89720 Av. 45708 44410 42904 417G3 89647 Elastic ratio; per ct. Av. 74.84 73.20 72.08 70.32 67.17 Elongation in 8 inches ; per cent. 1 2 8 4 5 6 7 8 9 10 29.50 82.00 81.75 27.75 81.50 80.50 29.50 31.00 80.00 29.65 28.25 30.50 82.00 27.00 80.50 80.75 30.50 28.50 82.00 81.75 31.00 80.75 27.50 28.50 80.00 29.00 81.00 29.25 28.00 29.50 28.00 29.50 29.25 28.00 29.50 80.00 81.00 28.00 80.00 30.00 84.00 81.25 81.25 82.25 80.25 82.00 82.75 82.75 80.75 8200 Av. 80.32 80.18 29.45 29.33 81.93 Reduction of area ; per cent. 1 2 8 4 5 6 7 8 9 10 66.1 67.1 62.3 64.9 63.3 66.0 66.8 62.4 64.5 66.2 65.9 66.0 62.4 65.0 64.4 66.2 66.3 62.6 63.5 66.0 66.7 66.0 63.9 64.9 64.2 66.7 67.4 68.0 64.3 66.1 67.0 66.7 63.2 65.9 63.7 67.3 67.1 63.1 65.8 67.1 68.4 67.1 68.4 67.7 65.0 66.0 67.9 64.8 66.9 67.6 Av. 64.96 64.83 65.82 65.69 66.48 It will be seen that a decrease in pulling speed is accompanied by a decrease in the ultimate strength, elastic limit, elastic ratio, 454 METALLURGY OF IRON AND STEEL. and elongation. The differences are not extreme, but their regu- larity, when viewed in connection with the uniform conditions of the experiment and the evident homogeneity of the material, makes the testimony almost conclusive. In the case of the slowest speed there is an exception to this rule in a marked' increase of extension, and an inspection will show that this does not arise from an aver- age of erratic members, but from an increase in every bar. This point is not of great practical importance, since it requires nearly an hour to break a single bar of ductile steel at this speed. The reduction of area seems to remain practically constant throughout the series. The natural result of this investigation would be a tendency toward higher breaking speeds. It is believed, however, that this may be carried too far, since with fast work it is more difficult to take accurate readings. CHAPTER XVII. THE INFLUENCE OF CERTAIN ELEMENTS ON THE PHYSICAL PROP- ERTIES OF STEEL. SECTION XVIIa. Difficulties attending the quantitative valua- tion of alloyed elements. Numerous investigations have been con- ducted to discover the influence of different elements on the strength and ductility of steel, a common method being to melt definite combinations in crucibles and ascribe the physical result to the known variables, under the assumption that all other things are equal. This system of experiment will answer in noting the effect of large proportions of certain elements, or in showing the qualita- tive influence of unusual ingredients; but it is worthless in the accurate quantitative valuation of minute proportions of the metal- loids, since small variations in the chemical equation are masked by irregularities in the detail of casting and working. The problem is also complicated by numberless combinations of different percent- ages of the various elements, so that it is difficult to obtain groups of charges where there is only one variable. It has, therefore, not infrequently happened that inconclusive data have been joined to bad logic, and the conclusions of special investigators have been at variance with all the teachings of experi- ence. It is not my purpose to enumerate all the theories or deduc- tions of experimenters, but I shall aim to give a general survey of the situation and to review the opinions and work of different lead- ing authorities. In Part I each element is considered separately, and I believe that the views therein advanced are in accord with the general consensus of opinion among metallurgists. Parts II and III give the result of special investigations into the effect of carbon, silicon, manganese, phosphorus and sulphur upon the tensile strength of steel, and a determination of the strength of pure iron. The results of this work are condensed into empirical formulae from which may be calculated with reasonable accuracy the ultimate strength of any ordinary structural steel whose composition is known. 455 456 METALLURGY OF IRON AND STEEL. PART I. EFFECT OF CERTAIN ELEMENTS AS DETERMINED BY GENERAL EXPERI- ENCE AND BY THE USUAL METHODS OF INVESTIGATION. SEC. XVIIb. Influence of carbon. The ordinary steel of com- merce is carbon-steel; in other words, the distinctive features of two different grades are due for the most part to variations in car- bon rather than to differences in other elements. There are often wide variations in manganese, phosphorus, silicon, etc., but it is rarely that the carbon content does not determine the class in which the material belongs. This selection of carbon as the one impor- tant variable arose primarily from the fact that primitive Tubal Cains could produce a hard cutting instrument with no apparatus save a wrought-iron bar and a pile of charcoal; and the natural developments in manufacture have led to the conclusion that a given content of carbon will confer greater hardness and strength, with less accompanying brittleness, than any other element. There are certain exceptions to be taken to this statement in the case of hard steels made by manganese, chromium, or tungsten, but it may be accepted as true in soft steel. It follows, therefore, that no limit should ever be placed to the carbon allowed in any structural material if a given tensile strength is specified. It is, of course, true that every increment of carbon increases the hard- ness, the brittleness under shock, and the susceptibility to crack under sudden cooling and heating, while it reduces the elongation and reduction of area, but the strength must be bought at a certain cost, and this cost is less in the case of carbon than with any other element. SEC. XVIIc. Influence of silicon. The contradictory testimony concerning the effect of silicon on steel has been well summarized by Prof. Howe,* who records many examples of exceptional steels with abnormal contents of silicon, and who fully discusses the theories advanced by different writers. He finds no proof that silicon has any bad effect upon the ductility or toughness of steel, and he concludes that the bad quality of certain specimens is not necessarily due to the silicon content, but to other unknown con- ditions. A Bessemer steel with high silicon is sometimes produced by hot blowing, but it will be entirely wrong to compare such metal * TTie MctaJlurr>y cf Steel, p. 36. , INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 457 with the common product and ascribe all differences to the chemi- cal formula, rather than to the circumstances which created that formula. Since the appearance of The Metallurgy, an able paper has been written by Hadfield,* who produced alloys with different contents of silicon by melting wrought-iron and ferro-silicon in crucibles. The metal was cast in ingots 2% inches square, and these were reduced by forging to 1% inches square and then rolled into bars 1% inches in diameter. In the list of analyses in the paper re- ferred to, there are slight differences in the composition of drillings from different bars of the same ingot, but in Table XYII-A I have averaged the results of each cast so as to show the nature of the material under investigation, and have given the physical results on the rolled bars in their natural state. TABLE XVII-A. Physical Properties of Silicon Steels, f r cent. ! I er cent. 1 QQ :rength ; ?r square j - 3 J d ^ c8 Q f|| P< 1 w - 3 CQ ft "s & - o ft fl 0< t H" o I IN H, 8-3 ll . .S i|^ S 2 GQ c8 Hi bJDXS C o 2 is d |||j I 1 c8 & .Q O ^a.a 3S^ JH S ^.s rrt 1?

o ! 1 53 Manganese; per cent. Ultimate strength; pounds per square inch. Difference in strength between each test and the barB. Difference in strength due to difference in carbon. Difference bet. the last two columns showing increase in strength due to silicon. Increase in percent- age of silicon com- pared with bar B. Increase in strength due to .01 per cent. of silicon. B 18 77 21 76160 C D E F G H .19 .20 .20 .21 .25 .26 1.57 2.14 2.67 3.40 4.30 5.08 .28 .25 .25 .29 .36 .29 84000 88480 95200 106400 109760 107520 7840 12320 19040 80240 33600 31360 1210 2420 2420 8630 8470 9680 6630 9900 16620 26610 25130 21680 .80 1.37 1.90 2.63 3.53 4.31 83 72 87 101 71 50 In the whole series it must be considered that the amount of work done upon the ingot in reducing it from 2% inches square to 1% inches in diameter was wholly insufficient to give a proper structure, so that little weight can be attached to the determination on any one bar. This renders it difficult to calculate the exact effect of silicon, especially since the bars A and B present some contradictions. Thus B contains .04 per cent, more carbon than A, .07 per cent, more manganese, and .56 per cent, more silicon, and yet has only 2240 pounds more tensile strength per square inch. Inspection shows that A is probably the erratic member, for its INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 459 strength is altogether too high for its composition. Moreover, the annealed bars show a loss in strength of 24 per cent, from the nat- ural in A } while ,bars B, C and D gives 15, 12 and 14 per cent., re- spectively, so that it is likely that A is finished at too low a tem- perature and has a higher strength than really belongs to it. For this reason it will be set aside as abnormal, and in Table XVII-B the bar B is taken as a basis from which to investigate the differ- ences in tensile strength. Xo allowance is made for manganese, since this element is fairly constant in all the specimens, but a value of 1210 pounds per square inch is given to carbon in accord- ance with the formula given in Table XVII-U. After this allow- ance the remaining variations are ascribed to silicon, but this is not strictly correct as no data are at hand concerning the content of phosphorus, so that the answer is open to question. TABLE XYII-C. Physical Properties of Steels Containing from .01 to .50 Per Cent. Silicon.* NOTE. All bars rolled well; they bent well both hot and cold except No. 11, which broke cold at an angle of 50; they all welded perfectly; the differences in hard- ness were scarcely perceptible. o> 43* d 8 1 1 h I 1 |i (H ft J 1 o o h ft **" o> ^ M c *. o g i 1 g ,C o ll III ill 11 II II n O 02 PH s w P H N H 1 .010 .16 .050 .WO .550 49280 66394 74.3 23.1 48.8 2 .061 .16 .028 .058 .619 49750 70806 70.3 20.4 40.7 8 .070 .15 .084 .051 .500 47152 66102 71.3 22.9 61.5 4 .092 .21 .084 .064 .634 60243 75398 66.6 19.4 44.1 5 .102 .18 .028 .066 .662 47622 75197 63.4 20.6 61.4 6 .121 .19 .064 .068 .576 60848 71367 71.2 21.9 43.7 7 .815 .13 .028 .057 .480 47690 65901 72.4 24.8 66.6 8 .247 .19 .028 .074 .642 49795 77728 64.0 17.6 49.6 9 .320 .15 .040 .081 .490 49997 74435 67.1 16.7 86.1 13 .882 .16 .042 .087 .583 55373 79901 69.3 18.0 80.7 11 504 .18 .094 .121 .455 f9024 822S) 71.7 19.4 84.8 This table cannot be called a conclusive equation of the effect of silicon, for the carbon was determined by color instead of com- bustion, the number of tests is altogether too limited, and no ac- count is taken of phosphorus, but there seems to be a strengthening effect of about 80 pounds for every .01 per cent, of silicon up to a content of 4 per cent., while be3 r ond this there is a deterioration * Report of British Association, 1888. 460 METALLURGY OF IRON AND STEEL. of the metal, as shown in Table XVII-A. This would mean an increase of only 1600 pounds for .20 per cent, silicon, being one- third more than that produced by .01 per cent, of carbon. (See Table XVII-U.) It has already been noted that A, which was the only bar containing an ordinary percentage of silicon, gave abnor- mal results in tensile strength, but this cannot be due to silicon, for the elastic ratio is quite normal, the elongation fair, and the reduction of area very good. An investigation into the effect of ordinary proportions of silicon was conducted by Turner under the auspices of the British Asso- ciation. Table XVII-C gives the results as published in Journal I. and S. L, Vol. II, 1888, p. 302. There are considerable varia- ations in the elements other than silicon, and the bad character of No. 11 may well be explained by its high content of phosphorus. Tor better comparison Table XVII-D gives the averages of the first four tests, all of which are below .10 per cent, in, silicon, and the last three, which are above .30 per cent. TABLE XVII-D. Comparative Physical Properties of Low-Silicon and High-Silicon Steels ; from Data in Table XVII-C. d ' Composition; percent. 2 a . OB rj u s I c3 .. a d * ^ -^ t* *" cT *O & 3J 05 "^ 8| ft* ID'S Sf +J 03 S Q d f S ^ ~ rj ri ^ c3 b-C3 o ^ 2 s| si.o 1 ll c rt -OJ fc SI. C. F ,. P. Mn. H P H S tf 1 4 .056 .170 .0 61 .0-33 .576 49106 69">75 70.5 21.5 46.3 2 3 .402 .160 .0 59 ,096 .493 54798 78863 69.5 18.0 33.9 The effect of the difference caused by elements other than silicon may be calculated from the formula given in Table XVII-U, car- bon being taken at +121 pounds for .001 per cent., and phosphorus at +89. The result is as follows : Lbs. per sq. In. Group II should be stronger than Group I. On account of phosphorus, 38 X 89 3382 Group II should bo weaker than Group I. On account of carbon, 10 x 121 1210 Net strengthening from constituents other than silicon . 2172 Strengthening from all constituents including silicon . . 9188 Strengthening due to .35 per cent, of silicon .7016 Strengthening due to each .01 per cent, of silicon r . 200 INFLUENCE OF CERTAIN ELEMENTS QN STEEL. 461 This signifies that .20 per cent, of silicon would give an increase in ultimate strength of 4000 pounds per square inch, which is only a little more than would be given by .03 per cent, of carbon. (See Table XVII-U.) The influence of silicon upon the tensile strength is often con- founded with that of carbon. It is well known that the addition of high-silicon pig-iron to a charge of low steel strengthens the metal more than a similar addition of ordinary pig-iron. But the fact is lost sight of that this silicon prevents the burning of carbon, both by the absorption of oxygen and by the deadening of the bath, so that the resultant metal is of higher carbon. If the ordinary color method were reliable, this would be de- tected and proper credit given to it, but it often happens that an increment of .03 per cent, of carbon is not shown by analysis, so that its effect upon the ultimate strength, which will amount to about 3500 pounds per square inch, will be incorrectly ascribed to whatever small percentage of silicon has survived the reactions during recarburization. This criticism on the determination of carbon applies to the data given in Tables XVII-A and XYII-C, and renders the calculations thereon of limited value. . These conclusions are corroborated by the testimony of Groups 49, 52, 54, 55, 56, 57, GO and 61 in Table XVII-N, as shown in Pig. XVII-A. All of these groups contain high silicon, but they do not seem to differ materially from the normal steels. Between the limits of 82,000 and 100,000 pounds ultimate strength there are seven groups in Table XVII-N", Nos. 48, 49, 50, 51, 52, 53 and 54, some containing high silicon and some with a low percentage, but the great variations do not seem to have any decided effect in altering the trend of the curve, although the contents of sulphur, phosphorus and manganese are fairly constant. (This question is discussed more fully in Section XVIIp.) It is well known that many continental works have habitually made their rails with from .30 to .60 per cent, of silicon, and that all requirements of strength and ductility have been met. All the authorities do not approve this practice, and it is stated by Ehren- werth,* that the latest results are rather in the opposite direction in the case of low steels, f but I was told some years ago, by the manager of one of the French establishments, that the only way * Dos Berg- und Htittenwesen auf der Weltausstellung tn Chicago, 1895. t See page 78, ante. 462 METALLURGY OF IRON AND STEEL. in which he was able to fill one contract with particularly severe specifications, was by making the rails contain from .30 to .40 per cent, of silicon, since a less proportion would not stand the drop- tests. It is not necessary to question whether this conclusion was warranted or not; it is enough to know that the steel was of the best quality, whether on account of the silicon or in spite of it. The fact that silicon may be allowed in rails has been acknowl- edged by Sandberg, who writes as follows:* "Silicon up to .30 per cent., with carbon .30 to .40 per cent., does not harden steel or make it brittle, and diminishes .its strength in such small degree as not to imperil the safety of the rail." The italics in the quota- tion are my own, and are to call attention to the implication that silicon lowers the strength rather than raises it. Exceptional cases have been recorded of soft steels with high silicon, like the very tough rail mentioned by Snelus,f with carbon below .10 per cent, and silicon .83 per cent. It must be considered, however, that although this might have been very tough for a rail, it does not follow that it was very tough for soft steel, but it is quite certain that it could not have been bad or brittle. With the knowledge possessed concerning the relative effect of impurities upon hard and soft steels, the assumption would almost be justified that low-carbon metal might be allowed to contain a larger percentage of silicon than higher steel. There is no need, however, of such an admission, for structural steels do not often contain over .05 per cent, of silicon, while usually they hold less than .03 per cent. Tool steel is subjected to the most severe of all tests in the ex- posure of a hardened edge to the blows of a hammer or the shocks of a planer. It was not the laboratory but the requirements of general practice from which was unconsciously evolved the formula for such metal, requiring low phosphorus, low sulphur and low manganese. In this process of natural selection no mention was made of silicon. It is true that some makers try to keep it as low as possible, but a large part of the best steel has regularly contained, year after year, from .20 to .80 per cent, of this element. Notwithstanding all this testimony as to the harmlessness of sili- con, it is firmly believed by many practical metallurgists that the * Proc. English Inst. Mech. Eng., 1890, p. 301. t On the Chemical Composition and Testing of Steel Rails. Journal I. and S. /.> Vol. II, 1882, p .583. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 463 presence of even .03 per cent, materially injures the quality of soft steel, such as is used for fire-boxes. I cannot positively assert the contrary, but I believe that the effects ascribed to silicon may be due to the conditions of manufacture which gave rise to it, or to the conditions of casting which it produces. These conditions might be fatal under one practice, as, for instance, when ingots are rolled directly into plates, while they might be harmless, or even beneficent, when an ingot is roughed down- and reheated. The opinions of practical men are sometimes of more value than the learned conclusions of theorists, and must never be ignored, but they are not always inerrant. SEC. XVIId. Influence of manganese. Spiegel-iron or ferro- manganese is added to a heat of steel at the time of tapping in order that it may seize the oxygen, which is dissolved in the bath, and transfer it to the slag as oxide of manganese ; but this reaction is not perfect, as shown in Section Xj, and there is reason to believe that all common steels contain a certain percentage of oxygen.* Steel low in phosphorus and sulphur requires less manganese than impure metal, although it is difficult to see why there should be less oxygen to counteract, and this indicates that the function of the manganese is to prevent the coarse crystallization which the impurities would otherwise induce. Besides conferring the quality of hot ductility, manganese also raises the critical temperature to which it is safe to heat the steel, for just as it resists the separation of the crystals in cooling from a liquid, so it opposes their formation when a high thermal altitude augments the molecular mobility. These two manifestations of the same general force render manganese one of the most valuable and essential factors in the making of steel, although there is no doubt that it has been used too freely in some cases. Years ago some of the railmakers of the country looked upon it as a panacea for all bad practices in the Bessemer and the rolling mill, and steel often contained from 1.25 to 2.00 per cent, of manganese, but it was soon discovered that such rails were brittle under shock, so that the permissible maximum has been gradually lowered, and the standard product of the present day contains from .70 to 1.00 per cent. In higher steels the same lesson has been learned, but in this case the necessity of a low content is far * See Section XVIIk for a further discussion on this point. 464 METALLURGY OF IEON AND STEEL. more marked, since a percentage which is perfectly harmless in unhardened steel will cause cracking if the metal be quenched in water. For this reason it is advisable to reduce the proportion of this element in hard steel to the lowest possible point. In structural metal there is no quenching to be done and the line of maximum manganese need not be drawn too low. It is much more convenient for manufacturers to produce a higher tensile strength by the use of spiegel-iron, which contains manganese, than with ordinary pig-iron, since the presence of manganese dead- ens the metal and prevents the oxidation of the carbon. Thus it happens that an increased tensile strength resulting from the addition of more recarburizer is usually accompanied by an increase in the content of manganese, and it is currently as- sumed that a considerable part of the extra strength is due to the higher percentage of this element. In great measure this is an error, for the increase in carbon is often sufficient to account for the change. Ferro-manganese containing 80 per cent, of manganese holds about 5 per cent, of carbon, and since about one-third of the man- ganese is lost during the reaction while very little carbon is burned, it follows that about 2 /VX 80=53 points of manganese will be added to the steel for every 5 points of carbon. Thus, if the con- tent of manganese in any heat be raised .20 per cent, by an in- .crease in the amount of the recarburizer, there will at the same time be an increment of .02 per cent, of carbon. This slight change in carbon will not always be detected by the color method, particularly as an increase in manganese interferes with the accuracy of the comparison by altering the tint of the solution, and so the effect of this carbon, representing an increase in tensile strength of about 2400 pounds per square inch, is often ascribed to the increment of manganese. It is necessary, there- fore, to carefully compare steels where the composition is thor- oughly known to find the effect of this element, and this is done in Parts II and III of this chapter. It is also currently believed that manganese reduces the ductility of steel to a great extent, but Table XVII-E will show that the effect is not well marked. This table is made by grouping together heats of the same general character and of about the same tensile strength, and separating them into two classes according to their manganese content. Xo arbitrary line is drawn between a high INFLUENCE OF CERTAIN ELEMENTS ON STEEL. ' 465 and low percentage, but each group is divided so that the number is as nearly equal as possible on each side. An unequal number is due solely to the fact that several heats may have exactly the same content, and these must all be placed either on one or the other side of the line. TABLE XVII-E. Comparative Physical Properties of Open-Hearth Steel with Dif- ferent Contents of Manganese. Made by The Pennsylvania Steel Company. 3 |I o3 g rt I rt Ei & ^ bo c8 5 00 3 tH * i~| 33 rt 03 s 1 ij *J 33 O Sft fi is 3 1 03 05 aj^ 2 * O r^ 2 t-i d 'o o&c-a .Srtrt^ 38 t> . 8 P rt s & "o o So a g 73 1 I* El ji - O rt g || III HI rt o o a rt tn rQ 0) 11 tl o M 3 i, P5 & p K P5 QQ i Acid 55000 to 60000 .08 Low High 7 6 .30 .37 57922 58881 38692 38598 29.91 28.08 59.02 57.07 66.8 65.6 % diam. ii Basic 55000 to 63000 .03 Low High 11 11 .44 .57 58005 59563 38547 40133 30.16 30.36 60.21 58.55 66.5 67.4 2*% in Acid 60000 to 65000 .08 Low High 16 14 .35 .51 62180 62605 41308 41169 28.00 27.65 50.89 54.66 66.4 65.8 % diam. IV Acid 65000 to 70000 .08 Low High 26 32 .51 .78 67421 68192 43923 45854 25.96 25.82 51.29 51.50 65.1 67.2 X diam. V Acid 70000 to 75000 .08 Low High 18 25 .60 .91 72353 72115 46836 48359 24.23 24.63 47.79 47.73 64.7 67.1 % diam. VI Acid 75000 to 80000 .08 Low High 11 11 .65 .84 77520 78083 49411 50226 22.34 23.63 44.42 48.49 63.7 64.3 % diam. VII Acid 80000 to 85000 .08 Low High 9 9 .68 .82 81747 81860 51219 52231 20.63 22.67 41.04 47.75 62.7 63.8 % diam. VIII Acid 85000 to 90000 .08 Low High 5 5 .75 .83 86460 88034 54517 55409 20.41 20.66 40.56 41.92 63.1 % diam. It will be evident that there is no marked difference between the steels of high and low manganese, and the results of the eight dif*- ferent groups are so uniform in their testimony that the work of chance must be almost absent. These records of ductility,, how- ever, do not take into account the very important quality of resist- ance to shock. It has always been a problem to devise some way of applying a satisfactory test in this direction, but the method is yet to be found. A few crude experiments which I performed on steel of high manganese, to see how it would act under shock, are given in Table XVII-F. The bar was struck while in tension with a copper hammer, each 466 METALLURGY OF IRON AND STEEL. blow being powerful enough to have permanently bent the bar if it had not been continually straightened by the action of the machine.. One of the effects of this hammering is to momentarily loosen the bar in the grips and make a sudden jar upon the piece. This action coupled with the stress upon the outside fibres and the direct vibra- tion, make the test quite exhaustive, although from the difficulty TABLE XVII-F. Resistance to Shock of Steel Containing about 1.00 Per Cent, of Manganese. All tests %-inch rolled rounds, made by The Pennsylvania Steel Company. g J d s jf 0*1 u Q) Cl *^* P H t. t numbe iganese; r cent. Conditions under which test was made. 5 Its ** o> K CO ^ C o uction o r cent. 8 *A is A* ,3 AST CC! w \ P H H P5 6960 1.00 Average of two tests, pulled quietly Average of two, hammered from start to 71040 47055 25.87 55.05 finish 70770 46380 2612 61.40 Average of two tests, pulled quietly 72175 48075 27.00 54.98 6961 1.03 Average of two, hammered from start to finish . . 71120 47330 26.00 59.20 6962 0.94 Average of two tests, pulled quietly Average of two, hammered from start to 74020 48165 25.62 52.60 finish 74490 48340 23.50 55.70 One bar, pulled quietly One bar, hammered from elastic limit to 81070 52880 22.50 43.60 fracture 80460 52760 23.50 48.30 6963 1.18 One bar, hammered from failure to fracture, One bar, began hammering at 72000 pounds, and moved scale weight back as the bar weakened 78050 69040 51800 52760 19.25 21.00 55.30 47.80 6981 0.82 One bar, pulled quietly One bar, hammered from failure to fracture, 67340 65940 46030 44430 28.12 28.00 55.00 57.90 6982 0.91 One bar, pulled quietly One bar, hammered from failure to fracture, 66700 67240 46310 46090 26.00 31.25 55.98 55.60 One bar, pulled quietly . . ........ 69700 47650 26.00 51.70 1.03 One bar, hammered from failure to fracture, 70080 46360 27.12 53.70 of measuring the force of impact it can hardly be called practical. Some of the bars were not struck until "failure," or until the maximum stress had been reached. This was on account of the trouble from slipping or jumping above noted which followed the hammering at earlier periods, and it was taken for granted that if a bar would break at all from shock, the fracture would be likely to occur about the time when the piece was under destructive tension. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 467 It will be evident that the hammering did not in any case deter- mine the time of breakage, for each piece gave as good an elongation and reduction of area as a part of the same rod which was pulled in the usual manner. It is not the intention to advocate the use of such a high content of manganese as is shown in Table XVII-F. The general conclu- sion of metallurgists,, evolved from experience, seems to point to as low a proportion as will ensure good working in the rolls. In the case of such ingots as are rolled directly into plates, the allowable content is limited by the requirement that the steel shall boil in the molds, but it does not follow because bad results accompany higher manganese in such practice, that the quality of the product is pro- portionally deteriorated when the ingot is roughed down and re- heated. The effect of large proportions of manganese upon steel is one of the most curious phenomena in metallurgy. As the content rises over 1.5 or 2.0 per cent, the metal becomes brittle and almost worth- less, and further additions do not better the matter until an alloy is reached with about 6 or 7 per cent, manganese. From this point the metal is not only extremely hard, but possesses the rather pecu- liar property of becoming very much tougher after quenching in water, without any great change in hardness. The physical properties of manganese steel are shown in Table XVII-G, which is taken from an article by Hadneld.* This alloy is used in the making of car wheels, dredger links and pins, and other articles where the maximum of hardness must be com- bined with toughness. Its great disadvantage is the difficulty of doing machine work upon it, for the best of hardened tools will rapidly crumble and wear out. In cases where finishing is essen- tial it is necessary to grind by emery wheels. SEC. XVIIe. Influence of sulphur. Nothing is better estab- lished than the fact that sulphur injures the rolling qualities of steel, causing it to crack and tear, and lessening its capacity to weld. This tendency can be overcome in some measure by the use of manganese and by care in heating, but this does not in the least disprove that the sulphur is at work, but simply shows that it is overpowered. The critical content at which the metal ceases to be malleable and weldable varies with every steel. It is lower with * See also The Mineral Industry, Vol. IV, for an essay on Alloys of Iron, by R. A. Hadfleld. 468 METALLURGY OF IRON AND STEEL. each associated increment of copper, it is higher with each unit of manganese, and it is lower in steel which has been cast too hot. TABLE XVII-G. Physical Properties of Forged Steel Containing from .83 to 19.00 Per Cent. Manganese.* Composition; per cent. Natural. Quenched in water. Annealed. "ft 9 s ... o s s|i gJL ... o sJL , *. o = Q d 1 | 1 jffl 02 , jfjl M III 3 o rt fl ** HI fc 3 03 s P H P H p , .20 .03 .83 73920 31 2 40 15 230 125440 6 g 40 09 8 89 85120 1 4 .52 .87 6.95 56000 2 51520 2 47040 2 5 .47 .44 7.22 60480 2 56000 2 60480 5 6 .61 .30 9.87 73920 5 87360 15 85120 16 7 .85 .28 10.60 76160 4 89600 17 91840 17 8 1.10 .16 12.60 87860 2 120960 27 82880 11 9 .92 .42 12.81 87860 5 136640 37 107520 20 10 .85 .28 14.01 80640 2 150080 44 107520 14 11 1.10 .82 14.48 87360 1 141120 87 109760 5 12 1.24 .16 15.06 109760 2 136640 81 105280 2 13 1.54 .16 18.40 114240 1 118720 10 87360 1 14 1.83 .26 18.55 96320 1 123200 5 -15 1.60 .26 19.10 116480 1 13?1SO 4 91840 1 In the making of common steel for simple shapes, a content of .10 per cent, is possible, and may even be exceeded if great care be taken in the heating, but for rails and other shapes having thin flanges it is advantageous to have less than .08 per cent., while every decrease below this point is seen in a reduced number of defective bars. It is impossible to pick out two steels with differ- ent contents of sulphur and say that the influence of a certain mi- nute quantity can be detected, but it is none the less true that the effect of an increase or decrease of .01 per cent, will show itself in the long run, while each .03 per cent, will write its history so that he who runs may read. The effect of sulphur upon the cold properties of steel has not been accurately determined, but it is quite certain that it is unim- portant. In common practice the content varies from .02 to .10 per cent., and within these limits it seems to have no appreciable influence upon the elastic ratio, the elongation, or the reduction of area. It is more difficult to say that it does not alter the tensile Condensed from Hadfleld, Journal I. and S. I., Vol. IT; 1888, p, 70. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 469 strength, for a change of one thousand pounds per square inch can be caused by so many things that it is a hold venture to ascribe it to one variable. Webster* has stated that sulphur probably increases the ultimate strength at the rate of 500 pounds per square inch for every .01 per cent. I am inclined to think his conclusion is not founded on sufficient premises, and shall try to prove this in Sec- tions XVIIs and XVIIu. In rivets, eye-bars and fire-box steel, the presence of sulphur is objectionable, for it will tend to create a coarse crystallization when the metal is heated to a high temperature, and reduce the strength and toughness of the steel. In other forms of structural material the effect of this element is probably of little importance. SEC. XVIIf. Influence of phosphorus. Of all the elements that are commonly found in steel, phosphorus stands preeminent as the most undesirable. It is objectionable in the rolling mill, for it tends to produce coarse crystallization and hence lowers the tem- perature to which it is safe to heat the steel, and for this reason phosphoric metal should be finished at a lower temperature than pure steel in order to prevent the formation of a crystalline struc- ture during the cooling. Aside from these considerations its influ- ence is not felt in a marked degree in the rolling mill, for it has no disastrous effect upon the toughness of red-hot metal when the content does not exceed .15 per cent. ^ The action of phosphorus upon the finished material may not be dismissed in so few words. Prof. Howef has gathered together the observations of different investigators, and the evidence seems to prove that the tensile strength is increased by each increment of phosphorus up to a content of .12 per cent., but that beyond this point the metal is weakened. Whether this last observation be cor- rect or not is of little practical importance, for it would be criminal to use a metal for structural purposes that contained as much as an average of .12 per cent, phosphorus. Below this point it is abso- lutely certain that phosphorus strengthens low steels, both acid and basic, and a quantitative valuation of its effect will be found in Parts II and III of this chapter. The same* certainty does not pertain to any other effect of this metalloid. Prof. Howet has ably discussed the whole matter, and * Further Observations on the Relations "between the Chemical Constitution and Physical Character of Steel. Trans. A. I. M. E. t Vol. XXIII, p. 113. t The Metallurgy of Steel, p. 67, et seq. I Loc. cit. 470 METALLURGY OF IRON AND STEEL. I herewith make quotations from The Metallurgy of Steel, and place them in the form of a 'summary. (1) The effect of phosphorus on the elastic ratio, as on elonga- tion and contraction, is very capricious. (2) Phosphoric steels are liable to break under very slight tensile stress if suddenly or vibratorily applied. (3) Phosphorus diminishes the ductility of steel under a gradu- ally applied load as measured by its elongation, contraction and elastic ratio when ruptured in an ordinary testing machine, but it diminishes its toughness under shock to a still greater degree, and this it is that unfits phosphoric steels for most purposes. (4) The effect of phosphorus on static ductility appears to be very capricious, for we find many cases of highly phosphoric steel which show excellent elongation, contraction and even fair elastic ratio, while side by side with them are others produced under apparently identical conditions but statically brittle. (5) If any relation between composition and physical properties 'is established by experience, it is that of phosphorus in making steel brittle under shock; and it appears reasonably certain, though exact data sufficing to demonstrate it are not at hand, that phos- phoric steels are liable to be very brittle under shock, even though they may be tolerably ductile statically. The effects of phosphorus on shock-resisting power, though probably more constant than its effects on static ductility, are still decidedly capricious.. The difficulty of detecting a high content of phosphorus by the ordinary system of physical tests, will be shown by Table XVII-H, which is constructed by comparing the acid open-hearth angles in Table XIV-H, which are of the same ultimate strength and of the same thickness, but which contain different percentages of phosphorus. Analyzing this record, it will be found that the higher phos- phorus gives a higher elastic ratio in all six groups, the difference ranging from 0.45 per cent, to 1.59 per cent., but the elongation and the reduction of area are almost exactly the same in the two kinds of steel. It is the difference between static and shock duc- tility that makes phosphoric steel so dangerous. In the ordinary testing machine there is no important difference between a pure steel containing less than .04 per cent, of phosphorus, and a com- mon steel with .08 per cent., or a bad steel with .10 per cent. Not only constructive engineers, but men calling themselves INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 471 metallurgists, have staked and have lost their reputations in pro- moting new processes designed to make good finished material out of steel containing high phosphorus. Many a time the metallurgical world has been stirred by some new discovery whereby such metal was induced to show a high ductility in the testing machine, and each time the new process has passed away unwept, unhonored and unsung as it was rediscovered that static and shock ductility were totally different properties, and that the high phosphorus metal gave lamentable failures as soon as it went beyond the watch- ful care of its parents and its nurses. TABLE XVII-H. Comparative Physical Properties of Low-Phosphorus and High- Phosphorus Steels; being a Comparison of the Acid Open- Hearth Angles given in Table XIV-H, that are of the Same Ultimate Strength and of the Same Thickness, but with Dif- ferent Contents of Phosphorus. \ * w i" . o S 0"S L i imits of ulti- mate strengt Ibs per squa inch. o. of group. er cent. umber of hca verage ulti- mate strong! Ibs. per sq. ii verage elasti limit; Ibs. p sq. inch. verage elast : ratio; perce fc! OP ^ O >" ft verage rcduc tion of area; per cent. EH PH K ^ < 2 2 <5 I T 5 U tO i .05 to .07 .07 to .10 212 50 60845 60064 40891 41143 67.21 68.50 29.35 28.82 07.4 58.4 .05 to .07 126 C0695 39415 64.94 29.23 55.6 56000 to .07 to .10 50 60583 40170 66.30 29.05 56.3 64000 III I 9 * to | .05 to .07 .07 to .10 81 50 0,0558 61049 38645 39656 63.81 64.96 28.95 28.98 53.8 54.8 IV H to 2 .05 to .07 .07 to .10 121 50 59906 59763 37478 3833S 62.56 64.15 29.32 29.CO 51.3 55.3 .05 to .07 40 65656 43713 66.58 27.90 .55.0 64000 to Tre to i .07 to .10 25 66365 44486 67.03 27.19 55.4 72000 .05 to .07 29 65631 42191 64.28 27.83 53.7 TS to i .07 to .10 33 65777 42817 65.09 27.49 53.2 It is true that numerous cases can be cited of rails, plates, etc., containing from .10 to .35 per cent, of phosphorus, which have withstood a long lifetime of wear and adversity ; but in the general use of such metal there has been sucji a large percentage of mys- terious breakages that it seems quite well proven that the phos- phorus and the mystery are the same. Much information on the effect of phosphorus may be gathered 472 METALLURGY OF IRON AND STEEL. from a study of high steels. A very severe trial is put upon a cokl- chisel or similar tool in the resisting of the continued shock on the sharpened edge, and it is undeniable that each increment of phos- phorus has its effect in rendering such a tool brittle. It is true that in this case the steel is quenched and also that it contains a con- siderable proportion of carbon, but there is no evidence to show that the effect of phosphorus is different when the carbon is high, even though it be true that it is more marked. Neither is there any reason to suppose that the quenching changes its nature, for experiments with high phosphorus steel of low carbon indicate that sudden cooling would rather counteract the influence of phosphorus than enhance it, since it tends to prevent the formation of coarse crystals. It would seem therefore that the regularly increasing baneful- ness of phosphorus as the carbon is raised does not portray any change in nature, but that although the effect of the metalloid in lower steels is obscured, its character is the same. ISTo line can be drawn that can be called the limit of safety, since no practical test has ever been devised which completely represents the effect of in- cessant tremor. For common structural material the critical con- tent has been placed at .10 per cent, by general consent, but this is altogether too high for railroad bridge work. All that can be said is that when all other things are equal safety increases as phos- phorus decreases, and the engineer may calculate just how much he is willing to pay for greater protection from accident. SEC. XVIIg. Influence of copper. The iron made from the ores of Cornwall, Pa., contains from .75 to 1.00 per cent, of copper, and large quantities of rails have been made from this iron alone, but it has oftener been the custom at eastern steel works to use from 25 to 50 per cent, of this iron in the mixture. Other deposits contain considerable quantities of this element, notably some beds in Virginia, while the ores of Cuba give an iron with about .10 per cent, of copper. Not only has such metal been put into rails, but into all kinds of steel, both hard and soft, and large quantities have been worked in puddle furnaces and in foundries, so that the miscellaneous cast- iron, wrought-iron and steel scrap, throughout the East, is very apt to contain quite an appreciable quantity of copper, and as steel- makers will continue to have more or less of this element to handle, it is of pressing importance that its effect be understood. The INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 473 necessity for such knowledge is the more marked as it is the custom in certain favored districts to intimate that copper is injurious, although definite proof is always lacking. Most of the Bessemer steels which are recorded in this book con- tain from .30 to .50 per cent, of copper, while much of the open- hearth steel is of the same character, and this will be sufficient proof that the best of steel may contain a considerable proportion. If, therefore, it appears from a set of experiments that copper exerts a bad effect, then one of two things follows : (1) That the experiments have left some factor out of the ques- tion. (2) That the maker of good steel has some trick by which he overcomes the enemy. It would be a cause for satisfaction if we could boast that the latter supposition were true, but, as a matter of fact, we have never known that copper injured the cold properties of steel in any way, and it is unnecessary to add that no system has been devised to obviate its influence. Hard and soft steels of our manufacture have found their way into all channels of trade, and although many failures have come, as they have - everywhere, from high carbon, high manganese, or high phosphorus, there have been no cases where it was necessary to invoke the aid of copper. This fact outranks and transcends in value any limited series of tests that might be given. In the same way there is no evidence that copper segregates, experience pointing rather to perfect uniformit} r . A story has been the rounds of the trade journals of a copper wire which crystallized out in the head of a rail, but, unfortunately, no method is known by which the phenomenon can be duplicated, since such rails might be of great value in electrical work. Steel may contain up to one per cent, of copper without being seriously affected, but if at the same time the sulphur is high, say .08 to .10 per cent., the cumulative effect is too great for molecular cohesion at high temperatures and it cracks in rolling. This teai- ing occurs almost entirely in the first passes of the ingot, so that it is of little importance to the engineer who is concerned only with perfect finished material. In the purest of soft steels containing not more than .04 per cent, of either phosphorus or sulphur, the influence of even .10 per cent, of copper may be detected in the less ready welding of seams during the process of rolling, but ordi- 474 METALLURGY OF IRON AND STEEL. narily when the sulphur is below .05 per cent, the copper injures the rolling quality very little, even if present in the proportion of .75 per cent. In all cases the cold properties seem to be entirely unaffected. These conclusions are not founded on any limited series of tests on special alloys ; they are the fruit of years of experience in the making of millions of tons of cupriferous steels, and it is quite certain that any baneful influence of this constant companion would have been felt in the many investigations which have been made into the mechanical equation of structural metaL The only facts ever brought out against copper as far as I am aware are in a paper by Stead,* who shows that steels containing from 0.46 to 2.00 per cent, of copper do not give good results in drawn wire when a high percentage of carbon is also present, but in the same paper it is stated that there is nothing to show that rails or plates are affected injuriously. The quantitative effect of copper upon the tensile strength of steel was the subject of a paper by Ball and Wingham,f in which they showed that as much as 7 per cent, could be alloyed to iron, and that a specimen with 4 per cent, forged well both hot and cold. It was found also that the alloys were very hard, so that when the content was over 7 per cent, the metal could not be cut by a good tool. The experiments showed a considerable increase in tensile strength in the case of higher copper, but no great weight can be given to the determinations, for the methods used in making the alloy and in cutting the tests were too crude for conclusive results. It is not easy to make a comparison between the ductility of high-copper and low-copper steels, for at works using such material it is customary to keep a fairly constant percentage in the mixture rather than to vary between wide limits. A limited number of heats have been grouped together in Table XVII-I, and although the list is not as long as might be desired, it should be considered ,that the heats were all made within a short period in the same Bessemer, and were all rolled in the same mill. It will be noted that no difference is to be found in the ultimate strength between steels with high and low-copper, although all the heats were made in the same way as nearly as possible, the work- * Jour. I. and S. I., Vol. II, 3901, p. 122. t On the Influence of Copper on the Tensile Strength o Steel. Journal I. and 8. I., Vol. I, 1889, p. 123. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 475 men not knowing either in the Bessemer department or in the roll- ing mill what kind of iron was in use. Moreover, the high-copper gives a slightly higher elastic ratio, which is a benefit, and also a better elongation and reduction of area. These results can hardly be called conclusive, for the num- ber of heats is too limited, but as the data on high-copper steels are uniform with the much larger number of similar angles given in Table XIVH, and as the two separate averages for low-copper correspond so closely to one another after allowance is made for the two different thicknesses, it seems quite justifiable to conclude that the high-copper is not in any way harmful. TABLE XVII-I. Comparative Physical Properties of Low-Copper and High-Copper Steel Angles. Made by The Pennsylvania Steel Company, 1898. 1 | 1 bo g j | rt o D A 1*1 Zi.ja MS VH o bickness i p< umber of jji i^i i? "3 CO & ! dj S 11 QJ QJ it 428 11 H 8 fc H W PH H T B B .10 .35 11 17 61376 60283 44152 43841 27.52 27.88 56.30 59.01 71.9 72.7 .10 10 58965 42218 28.85 55.50 71.6 ' .85 11 59630 43478 29.02 57.86 72.9 A notable investigation into the effect of copper was conducted by Mr. A. L. Colby at the Bethlehem Steel Works, and was de- scribed in The Iron Age, November 30, 1899. He relates that steel containing 0.57 per cent, of copper was forged into crank shafts for the United States battleships and stood every test required by the government specifications. Another ingot was forged into gun tubes for 6-inch guns for the United States Navy and fulfilled every requirement of the department. Other exhaustive tests were made on plates and all the results pointed the same way. SEC. XVIIh. Influence of aluminum. It is hardly necessary to discuss at length the effect of aluminum upon steel, for although it is often used to quiet the metal, it unites with the oxygen of the bath and passes into the slag. Sometimes a very small percentage 476 METALLURGY OF IRON AND STEEL. remains in steel castings, while it is quite conceivable that other steels may receive a small overdose by mistake, so that Table XVII-J will be of interest as giving the results of an investiga- tion by Hadfield.* TABLE XVII-J. Physical Properties of Aluminum Steel. NOTE. Size of bars |J x | inch ; all samples forged either very well or fairly well except No. 10 which was very shelly. The fractures from Nos. I to 7. inclusive, were granular, but Nos. 8, 9, and 10 showed increasing coarse crystallization. All bars bent double cold after annealing except No. 10. Attempts at welding were unsuccessful on samples Nos. 3, 5, and 8. 2 c3 I c8 3 C? . oo tft=* g? a I I Composition; percent. II fll V- O j> IS 5 w 73 2 - Op ^ c ^ "cfi ^ "cB S lj 0) o . d !=* Ss-S CO 3 t. '11 1 C. Si. S. P. Mn. Al. ^ S ga.s o a u'~ 1^ 3 1 .22 .09 .07 .15 47040 64960 36.70 62.9 72.4 2 .15 .18 '.10 ' '.04 ' .18 .88 61520 67200 87.85 58.18 767 8 .20 .12 .11 .61 48160 62720 88.40 64.50 76.8 4 .18 .16 '.09' '.03' .14 .66 45920 64960 83.85 49.86 70.7 6 .17 .10 .18 .72 49280 62720 40.00 60.74 78.6 6 .26 .15 '.08' '.04' .11 1.16 51520 73920 32.05 61.46 69.7 7 .21 .18 .18 1.60 44800 69440 32.70 52.14 64.5 8 .21 .18 '.09" '.03' .18 2.20 47040 69440 22.75 27.80 67.7 9 .24 .18 .82 2.24 48160 72800 20.67 2464 661 10 .22 .20 .08 .03 .22 5.60 85120 896 After making allowances for the variations in other elements, it will be found that the aluminum has little effect upon the tensile strength, while it does not materially injure the ductility until a content of 2 per cent, is reached. These conclusions do not agree with the results which I have found by casting different alloys in the form of 6-inch square ingots. The aluminum was added in a solid state and it is quite possible that it was not disseminated uniformly, but the analysis was made on the test-bar itself, and the fusible nature of the metal makes it probable that the piece would be reasonably homogeneous. Either two or three ingots were cast from each heat, the first containing either no aluminum or only a trace, while the others were made so as to give fairly rich alloys. The results are given in Table XVII-K. The casting and working of such ingots is a regular operation at the works where these experiments were made, and perfect uni- Aluminum Steel. Journal I. and 8. I., Vol. ll f 1890, p. 161. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 477 formity is always obtained in respect to tensile strength, so that it is probable the variations in bars of the same heat are due to the different contents of aluminum. These changes are as follows : (1) The addition of one-half of 1 per cent, of aluminum in- creases the tensile strength between 3000 and 8000 pounds per square inch, exalts the elastic limit in about the same proportion, and injures .very materially the elongation and contraction of area. The effect both upon strength and ductility is more marked in the case of low than in high-steels. TABLE XVII-K. Effect of Aluminum upon the Physical Properties of Steel. 3-inch square ingots, made by The Pennsylvania Steel Company, rolled to 2x% inch. Heat number. Composition; percent. imate strength; mnds per uare inch. Elastic limit; pounds per square inch. Elastic ratio; percent. 1! l! li C o 3"* Reduction of area; per cent. C. P. Si. Mn. s. Al. gu Soft basic open-hearth steels. 1791 .11 .11 .024 .022 .48 .45 .035 .035 .00 .58 48800 56880 83190 41150 68.0 72.4 31.25 18.25 48.6 29.8 1792 .11 .11 .010 .011 .45 .41 .019 .023 .00 .45 40440 53440 31040 3(5900 68.1 69.1 30.00 22.50 49.9 81.5 1793 .11 .11 .013 .35 .00 .50 47160 53900 &3490 38530 71.0 71.5 31.25 27.00 45.8 33.7 Soft acid open~ hearth steels. 8681 .17 .16 .14 .035 .61 .025 .04 .473 .899 58560 63440 64160 39310 42100 39100 67.1 66.4 60.9 30.00 23.00 17.50 45.7 86.3 25-4 . . . 3686 .14 .14 .12 .059 .58 .021 .03 .46 1.171 55030 67810 67420 43260 47950 48850 66.5 707 72.5 24.00 20.00 8.00 46.2 34.0 15.0 3688 .12 .12 .13 .034 .51 .021 .013 .45 .80 55700 59880 61470 39550 39100 43710 71.0 65.3 71.1 28.7 21.7 21.2 51.8 40.5 34.2 Hard acid open-hearth steels. 3682 .47 .44 .43 .048 .21 ,70 .018 .00 .571 1.135 107450 110550 105100 65930 72420 68080 61.4 65.5 64.8 10.0 92 12.5 201 17.5 21.0 . . . 8683 .54 .47 .43 .044 .31 .75 .020 .00 .37 .94 124040 122080 128040 47830 47680 47440 88.6 89.1 87.0 10.0 '7.5' 18.0 8.2 9.4 . . . 3684 .40 .36 .38 .040 .26 .67 .028 .01 .54 .90 95010 98375 98720 42740 43050 43150 45.0 43.8 43.7 18.7 14.0 12.5 41.0 24.5 20.4 3685 .40 .38 .34 .046 .30 .63 .031 .00 .52 .73 94700 100055 98480 44610 47240 46910 47.1 47.2 47.6 16.2 13.7 12.5 81.8 24.1 17.5 22.0 89.7 25.4 8689 .42 ,40 .34 .046 .21 .71 .025 .00 .81 .66 90900 94560 96680 63550 59190 59460 58.9 62.6 61.5 15.5 15.0 14.7 478 METALLURGY OF IRON AND STEEL. (2) The addition of another half of one per cent, does not have much effect upon the ultimate strength or the elastic limit, but it still further decreases the ductility of the metal. It is stated by Odelstjerna* that the use of aluminum, in the manufacture of steel castings, gives an inferior metal, even though the addition amount to only .002 per cent., and that such steel presents a peculiar fracture, the faces of the crystals -being large and well defined. It must be kept in mind, however, that these conclusions apply to one particular kind of practice, and that the use of aluminum, under certain conditions, may produce a most harmful effect, while under other possible conditions the result would be much less marked. Nothing is more difficult than to isolate one factor in a metallurgical equation, and to discover its real value, when it is always associated with complicating and equally powerful agencies. SEC. XVIIi. Influence of arsenic. The effect of arsenic upon steel was .quite 'fully investigated several years ago by Harbord and Tucker, f The conclusions given by them may be summarized as follows: Arsenic, in percentages not exceeding .17, does not appear to affect the bending properties at ordinary temperatures, but above this percentage cold-shortness begins to appear and rapidly in- creases. In amounts not exceeding .66 per cent., the tensile strength is raised very considerably. It lowers the elastic limit, and decreases the elongation and reduction of area in a marked degree. It makes the steel harden much more in quenching, and injures its welding power even when only .093 per cent, is present. These results have been corroborated by J. E. Stead,} who found that between .10 and .15 per cent, of arsenic in structural steel has no material effect upon the mechanical properties; the tenacity is but slightly Increased, the elongation and reduction of area ap- parently unaffected. With .20 per cent, of arsenic, the differ- ence is noticeable, while with larger amounts the effect is decisive. When one per cent, is present, the tenacity is increased, and the elongation and reduction of area both reduced. This increase in strength and diminution in toughness continue as the content of The Manufacture of Open-Hearth Steel in Sweden. Trans. A. I. M. E., Vol. XXIV, p. 312. t On the Effect of Arsenic on Mild Steel. Journal I. and 8. I., Vol. I, 1888, p. 183. t The Effect of Arsenic on Steel. Journal T. and S. I., Vol. I, 1895, p. 77. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 479 arsenic is raised to 4 per cent., when the elongation and reduction in area become nil. These experiments are of considerable practical importance, since a great many steels carry an appreciable proportion of arsenic. Some chemists take little cognizance of this fact, and their phos- phorus determinations are often too high on account of the presence of arsenic in the phosphorus precipitate. Other analysts take spe- cial precautions to avoid this contamination. TABLE XVII-L. The Physical Qualities of Nickel Steel as Compared with Carbon Steel of Similar Tensile Strength. NOTE. All steels were made in an acid open-hearth furnace by The Pennsylvania Steel Company. Composition; percent.. Kind of steel. C. Mn. P. S. Ni. Nickel .24 0.78 .032 .027 8.25 Hard forging . . .30 to .35 Forging 25 to .80 .60 to 1.00 .60 to .80 .03 to .05 .03 to .06 .03 to .05 .03 to .07 nil. nil. Kind of steel. It "3 d | M ... Mi G ,_ ,H W cd Shape of member. 1! ||1 O in 3 m I 1 - S OJ O in C8 IH 0*2^ o 1 l o3 U> eg +3 O ! 33 4 G o C3 o 2 *^ ^ *H cB 5 C3* ii I OG O fl ^ S P " 1 H W g ft Nickel . . 86015 63575 73.9 20.19 34.00 46.3 Rounds, Hard forging . 87663 58055 66.2 16.70 24.44 80.3 Forging . 78066 51793 66.3 23.94 52.0 Nickel . . 86960 58553 67.3 21.75 89.67 60.5 Angles, Hard forging . Forging .... 87820 76970 54153 49544 61.7 64.4 19.25 34.83 43.3 49.0 Universal plates, longitudinal, Nickel Hard forging . Forging .... 85773 82773 78996 58410 50163 46654 68.1 60.6 59.1 21.08 20.50 26.78 89.25 87.67 52.0 47.0 52.1 Universal plates, transverse, Nickel Hard forging . 86417 85173 58203 (50000)* 67.4 (58.7)* 16.50 18.83 28.92 23.17 36.1 27.4 Sheared plates, longitudinal, Nickel Hard forging . Forging .... 85337 85012 78918 58169 (50000)* 49128 68.2 (58.8)* 62.3 19.00 22.10 22.03 85.50 89.40 48.8 48.4 50.8 Sheared plates, transverse, Nickel Hard forging . Forging .... 84377 84327 57260 (50000)* 67.9 (59.3)* 17.13 21.71 82.50 87.00 43.4 41.3 SEC. XVIIj. Influence of nickel, tungsten and chromium. The first public presentation of the effect of nickel upon steel was * Approximate ; could not determine accurately. 480 METALLURGY OF IRON AND STEEL. a paper by Jas. Kiley.f Since that time the properties of nickel steel have become widely known through the experiments by the United States Government on the armor plate manufactured by The Bethlehem Iron Company,, and by the Carnegie Steel Com- pany. As often happens in the case of a new metal, the tendency is to exaggerate its importance. In a paper read before the Ameri- can Society of Civil Engineers, in June, 1895, I gave the detailed results found by testing nickel steel when rolled into rounds, angles and plates, and compared them with the records of carbon steel of about the same tensile strength. A condensation of the work will be found in Table XVII-L. It will be noted that the nickel steel is superior, but in less measure than may be generally supposed. It must be kept in mind, however, that in armor plate, as in many another field, there is sometimes but a* very small distance between absolute success and absolute failure, and that it matters little how much margin there is above success, provided there is a margin at all. There are other elements used to make special alloys with iron, some of these metals being of considerable importance. Tungsten and chromium are both employed to give tool steels extreme hard- ness, their peculiar characteristic being that no quenching or tem- pering is required. These alloys, however, do not come under the head of structural material, and will therefore not be considered hero. SEC. XVIIk. Influence of oxide of iron. The last step in the making of a heat of steel is the addition of the recarburizer to wash the oxygen from the bath, but this action is not perfect, and the exact relation is not generally understood. The amount of oxygen taken from the metal will evidently be measured in a rough way by the amount of manganese and other metalloids that are burned during the reaction. This is particularly true of acid prac- tice. In basic work there is oftentimes a very considerable loss of manganese through the presence of a large amount of free oxygen in the slag. This occasionally occurs in the acid furnace, but less frequently. It was shown in Section Xj that the loss of man- ganese in recarburization is a function of the quantity which is added. In other words, if there is a reduction in the percentage of manganese which is added to an open-hearth bath at the time of * Alloys of Nickel and Steel. Journal I. and 8. I. t Vol. I, 1889. p. 45. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 481 tapping, there will be a reduction in the amount of manganese which will be oxidized, and this proves conclusively that the reac- tion is not perfect, and that an increasing amount of oxygen must remain in the metal as the content of manganese decreases ; but a reasonable proportion of this oxygen can hardly exert any marked deleterious influence, else the fact would long ago have been known in some more definite form than the suppositions and theories which are occasionally founded on exceptional phenomena. Assuming as certain that high oxygen will more likely be found in steels both low in manganese and in oxidizable metalloids, it may reasonably be expected that any bad effect it may exert will be seen in the softest products of the basic open-hearth and in the purest of acid steel. On the contrary, it is well known that the reverse is true, and that the ductility increases as the condition of pure iron is approached. TABLE XVII-M. Individual Eecords of Heats Composing Group .63 in Table XVII-K ^* . a ^ o 1 d 8 S a g 9, ir .~ 9 t-, 2%% *i CD O 5 Heat numl Carbon by < bustion; Carbon by < per cent. Phosphoru per cent. Manganese per cent. Sulphur; p Copper; pe Elastic lim pounds p square in Ultimate s pounds p square in Elastic rat per cent. 4669 .04 .007 .02 .024 .10 28420 45620 62.3 4809 .04 .007 .05 .019 .05 80640 46310 66.2 4930 .04 .007 .04 .021 .06 24370 46000 53.0 4932 .04 .011 .04 .029 .04 25810 46480 55.5 4971 .03 .010 .05 .032 .14 26780 47140 56.8 4972 .04 .010 .04 .021 .10 27920 47000 69.4 Average, .025 .04 .009 .04 .024 .08 27323 46425 58.9 Some people imagine that it is not well to take all the impurities out of iron, their thesis having been forcibly, though somewhat inelegantly, expressed in the saying that a shirt can be ruined by too much scrubbing. Unfortunately, the simile is entirely worth- less, for the purification of steel is not a process of washing, al- though often so called. Dephosphorization does not consist in mechanically removing certain foreign ingredients, but in placing the metal in contact with a slag of such a character that the metal- 482 METALLURGY OF IRON AND STEEL. loids find in it a more congenial home, and although it is true that over-oxidation assists the purification, it is not at all a necessary adjunct, since the transfer of allegiance may be effected by a slag moderately rich in lime, combined with the normal oxidizing influ- ences. In a discussion of a paper by Webster, which will be referred to at length in Part II of this chapter, H. D. Hibbard* deduced the fact that oxide of iron reduces the tensile strength of very soft metal by several thousand pounds. I cannot indorse this conclu- sion, but offer Table XVII-M as evidence to the contrary. These heats were made in a basic open-hearth furnace, and their regularity both in chemical and physical character shows that we are dealing with a normal and definite metal and not with an acci- dental product. They were purposely made with the lowest pos- sible content of manganese, and it seems positively certain that the steel must be saturated with oxygen. These six heats constitute Group 63 in Table XVII-N, and by the most casual inspection, as well as by a glance at Curve AA in Fig. XVH-B, it will be plain that these steels are much stronger than would be expected as compared with those containing more carbon. It may be that the first increments of carbon have less strengthening effect than further additions, or it may be that the first increments of manganese have a marked weakening effect. but it is more probable that the oxide of iron increases the ultimate strength. PART II. EFFECT OF CERTAIN ELEMENTS AS DETERMINED BY SPECIAL MATHE- MATICAL INVESTIGATIONS. SEC. XVIII. Investigations by Webster on the influence of the metalloids. A very comprehensive and systematic study of the physical formula of steel has been carried out by W. R. Webster, f He has used the long and laborious method of successive approxi- mations, and by "cutting and trying" has found the effect of each element upon the ultimate strength, as well as the effect of the * Trans. A. I. M. E., Vol. XXI, p. 999. t Observations on the Relations beticeen the Chemical Constitution and Phy- sical Character of Steel. Trans. A. I. M. E., Vol. XXI, p. 766, and Vol. XXIII,, p. 113; also Journal I. and 8. I., Vol. I, 1894, p. 328. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 483 thickness and finishing temperature. The results are given by him as follows : .01 per cent, of sulphur increases the tensile strength 500 pounds per square inch. .01 per cent, of manganese has an effect which varies with each increment as follows, the values being expressed in pounds per square inch: An increase in percentage gives an increment of making a total increase in strength over metal with no manganese of from .00 to .15 3600 8600 .15 to .20 1200 4800 .20 to .25 1100 5900 -25 to .30 1000 6900 .30 to .35 900 7800 .35 to .40 800 8600 .40 to .45 700 9300 .45 to .50 600 9900 .50 to .55 500 10400 .55 to 60 500 10900 .60 to .65 500 11400 .01 per cent, of phosphorus has an effect which varies according to the amount of carbon present : With .08 p .09 " .10 " .11 " .12 " .13 " .14 " .15 " .16 .17 er cent, o f carbon i t i s 800 po 900 1000 1100 1200 1300 1400 .1500 1500 1500 ' inds p < t t er sqv t < tare in < ' ch. < < < < < < n i K t ti u Carbon is credited with a constant effect of 800 pounds for each ,0i per cent. Mr. Webster has constructed, from these values, a table showing the strength of metal containing different proportions of carbon and phosphorus, from which, as a basis, the strength of a given steel may be found by allowing for the content of manganese and sulphur. This table presents a curious anomaly, as will be shown by the following excerpt :* Estimated Ultimate Strengths ; Pounds per Square Inch ; per Webster. Carbon ; percent. .07 .08 .09 .10 .11 .12 .13 .14 .15 .16 .17 .18 P = .00perct. P = .03perct. P t 40350 42750 45150 48350 41150 43550 45950 49150 41950 44650 47350 50950 42750 45750 48750 52750 43550 46850 50150 54550 44350 45150 47950 49050 51550 52950 56350' 58150 45950 50150 54350 59950 46750 51250 55750 61750 47550 52050 56550 62550 48350 52850 57350 63350 49150 53650 58150 64150 Journal I. and 8. I., Vol. I, 1894, p. 338. 484 METALLURGY OF IRON AND STEEL. An examination of these figures reveals two absolutely irreconcil- able conditions, for Mr. Webster takes as his starting point the dictum that carbon is a constant, and proceeds to construct a table in which it is not a constant at all, and in which it is not even constantly irregular. By his own calculation a steel of .06 per cent, phosphorus and .10 per cent, carbon is strengthened 1400 pounds by the addition of .01 per cent, of carbon, while with .10 per cent, phosphorus it is strengthened 1800 pounds by the same addition. Assuredly, this is not a constant effect. Moreover, car- bon does not even have a constant effect with the same content of other metalloids, for, with .10 per cent, of phosphorus, an increase in carbon from .07 to .08 per cent, raises the strength 800 pounds, while an increase from .08 to .09 per cent, strengthens it 1800 pounds. It would be just as correct to conclude from these results that phosphorus is a constant and carbon a variable, as to say that car- bon is a constant and phosphorus a variable. The changing values which it would be necessary to assign to carbon to fulfill the first assumption would be no more arbitrary and hypothetical than the changing values assigned to phosphorus by Mr. Webster, or the changing values which he has assigned to manganese. Thus the table which has been given is entirely indecisive, since it can be translated into two diametrically opposite readings, and it must be acknowledged that one empirical formula is as good as another, provided the same answers are obtained from both. This curious contradiction of the premises by the conclusion can only arise from some erroneous hypothesis in the values as- signed to the different elements, for in the construction of such equations it is plain that an error in one factor must be atoned for by an opposite and equal error in another factor. If this rea- soning be true, then very little faith can be attached to the formula as an expression of fundamental laws, however accurately the mathematical results may coincide with observations. It is to be regretted that the earnest endeavor of Mr. Webster to write the physical formula should have been hampered by the necessity of working on sheared plates, which are finished under greater variations of temperature than angles or bars, and further- more, that these plates were of basic Bessemer steel, a "material which would not be chosen for its regularity. By correcting for thickness and finishing temperature, Mr. Webster has shown that INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 485 about 90 per cent, of the heats investigated came within 5000 pounds per square inch of what his equation calls for. This is a very satisfactory result, and it is not in a spirit of hypercriticism (for my own results, to be given later, display ex- amples of the same character), but from a strictly scientific point of view, that attention is called to the very unpleasant corollary that one charge out of every ten does not give results within 5000 pounds. Some of these undoubtedly are vitiated by wrong chemical determinations, for the carbon was determined by color, and this gives only approximate results; on others there might well be an error in estimating the finishing temperature;^ on others there would be mistakes in measuring and testing; while some pieces, perhaps, did actually show those peculiarities which we call abnor- mal, which are ascribed sometimes to oxide of iron, sometimes to nitrogen, and not infrequently to the devil, but which grow less numerous as we learn more of our art. I cannot believe that the complicated formula of Mr. Webster represents actual conditions, and the remainder of this chapter will attempt to show that a reasonably accurate empirical equation of steel may be written without the introduction of such manifold variations, and by the use of constant values for each element within the limits usually obtaining in structural metal. It will also be shown that the first increments of manganese do not add greatly to the strength of steel, since low-manganese metal is stronger than would be indicated by a formula that applies to steels containing higher percentages of this element. INTRODUCTORY NOTE. The remainder of Part II of this chapter may be omitted by the general reader. It discusses the first investigation made upon a series of steels by the method of least squares, but the results which are given later on a second series shed much light on points that are not made clear by the first investigations and give authority for more positive statements. This increased knowledge does not arise from any superiority of the second investigation, but simply from the fact that the basis of work was doubled and the validity of the results correspondingly enhanced. The careful student will find it necessary to read all that is written to understand the steps involved, and to know why certain 486 METALLURGY OF IRON AND STEEL. elements have been omitted from the formula, but those less curi- ous may pass to Part III, which embraces the latest investigations on both series, while Section XVIIw gives a synopsis of the whole argument and the conclusions drawn therefrom. SEC. XVIIm. Investigation on Pennsylvania Steel Company steels ~by the method of averaging groups of preliminary tests. I believe that the true way to investigate the influence of the metalloids upon the physical qualities of steel is to make groups of heats so as to avoid the determinative errors in any one charge. It is essential that the components of each group should be very nearly alike in chemical composition, or the whole purpose of the investigation may be frustrated by the intermingling and the inte- gration of factors which should be differentiated and equated. If, however, the members -of each group are as nearly uniform as pos- sible, we may thereby reduce the effect of determinative errors and render possible an accurate determination of carbon, for it is out of the question to make, a combustion on each separate charge, and I do not consider a color determination as any fit ground for scientific work. The method of forming the groups in the following investiga- tion is of such importance that it is necessary to give a full descrip- tion. It is the custom at Steelton to make a preliminary test of every open-hearth heat, and it is found that this test is almost invariably a reliable exponent of the charge from which it comes. In the rolling of plates, angles and miscellaneous shapes, the thick- ness of the piece and the finishing temperature have a great effect upon the result, but in this test-piece the conditions of heating and working are so constant that the results as shown by the test- ing machine reflect only the influence of variations in the chemical equation. Having preserved the broken bars for a considerable time, there were at hand 575 pieces of acid steel below 80,000 pounds ultimate strength, 1160 pieces of basic steel below 70,000 pounds, and 145 pieces of acid steel above 80,000 pounds. In addition to the ulti- mate strength, the content of manganese, sulphur and phosphorus was on record for each piece. Taking the low-acid steels as one basis of work, a further separa- tion was made according to the tensile strength; for example, in the case of the low-acid steels there were 148 heats between 58,000 INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 487 and 60,000 pounds, and these were considered a sub-division. This was again divided, the heats being arranged according to their chemical composition. Thus there were 18 heats which showed lower manganese than the rest, and these were averaged to give Group 8 in Table XVII-N. There were 13 heats showing high manganese, and these gave Group 16. The low-sulphur heats gave Group 24, the high-sulphur Group 17. The low-phosphorus gave Group 25, the high-phosphorus Group 19, while there were 72 heats which did not show a high content of any element, and these form Group 15. Oftentimes it would happen that a charge which contained high manganese would show either a low or a high percentage of some other element, and hence would appear in two or more groups, so that the total number of heats in the table is larger than the num- ber of test-pieces. After forming these groups, the average manganese, sulphur, phosphorus and ultimate strength of each were calculated from the records, while the average carbon, silicon and copper were deter- mined by weighing an equal quantity of drillings from each bar and making a chemical analysis, the carbon being determined by combustion. Since each member of a group contained nearly the same percentage of carbon, it is evident that very little error is introduced by this system of average, while it assuredly tends to hide the idiosyncrasies of any one heat. By this system of combination the low-acid steels gave 47 groups, which are given in Division I, Table XYII-N, and are plotted in Curve AA, Fig. XYII-A. The high-acid steels gave 15 groups, which are given in Division II, and are plotted in Curve BB, Fig. XVII- A. The basic steels gave 75 groups, which are given in Di- vision III, and are plotted in Curve AA, Fig. XVII-B. In these graphic representations the ordinates are the ultimate strength per square inch, and the abscissas the percentage of carbon, the latter element being selected because it is universally recognized as the controlling component. , SEC. XVUn. Quantitative valuation of the elements by the method of least squares. It is certain that carbon increases the strength of steel when present in small proportions, but that after a certain content is reached (say about 1.00 per cent.) there is no increase in cohesive power from a further addition. It will also be granted that this point is not a sudden break in the line, but 488 METALLURGY OF IRON AND STEEL. TABLE XVII-N. List of Groups Used in Determining the Effect of Certain Elements upon the Tensile Strength of Steel, together with the Formulae Obtained therefrom by the Method of Least Squares. NOTE. All figures relating to ultimate strength are expressed in pounds per square inch. Number Effect of .001 per cent, upon the ultimate strength. of Kind of formula steel. Carbon.' Manganese. Phosphorus. Iron. 1 Acid, +152.9212 3.902156 +131.6955 +0.3432669 2 Basic, +103.4560 +5.298315 + 94.08509 +0.3899613 ft ti 2 Composition; percent. ft 0> 3 s * Formula No. 1. M 9 c Q CO IS *j O < |ij s 3 i 1 o 3 c?^ ^ 08 g A !> o o 1 Is || | I iH O ft U> if -11 |||| Ifi| 3 g 2 * o .s d ft ft fi$3 c3 3*3 5c c^i; -Hi ^ 3 be 1-^ ,3 o o ^ t> 03 S CS O> cc & Zl 33 a rfi 1 1 2 ^ o i 6 .082 .006 .290 .034 .034 .120 99.434 52090 50018 2072 46672 2 12 .105 .009 .380 .059 .074 .180 99.193 57375 58369 + 994 50106 8 11 .109 .008 .310 .036 .066 .140 99.331 57310 58248 + 938 507(55 4 12 .109 .007 .380 .048 .082 .150 99.224 57430 60045 +2615 50729 5 38 .113 .009 .430 .038 .061 .130 , 99.219 57140 57694 + 554 51339 6 11 .118 .007 .480 .046 .096 .180 99.078 62870 62060 810 51290 7 5 .115 .007 .490 .029 .037 .090 99.232 55450 54610 840 51649 8 18 .115 .013 .300 .043 .069 .170 99.290 58780 59585 + 805 51669 9 12 .116 .005 .590 .025 .034 .100 99.130 56830 53942 2888 51767 10 19 .116 .015 .500 .069 .082 .190 99.028 60870 60580 290 -61732 11 9 .116 .013 .470 .057 .089 .170 99.085 62610 61638 972 51752 12 18 .117 .018 .330 .039 .073 .200 99.223 61190 60278 912 51952 "5 18 17 .117 .005 .450 .049 .099 .160 99.120 61430 63198 +1768 51916 J> 14 19 .118 .005 .590 .030 .035 .100 99.122 56990 54377 2613 52070 5 15 72 .118 .007 .420 .045 .075 .140 99.195 59110 60333 +1223 52095 ,13 16' 13 .118 .008 .560 .044 .063 .140 99.0(57 58163 1187 52051 17 15 .118 .007 .450 .064 .081 .170 99.110 H^so 60977 +1717 5206(5 18 15 .118 .014 .570 .056 .076 .180 98.986 00900 59808 1092 52023 ^ 19 21 .119 .009 .420 .051 .090 .140 99.171 5:>310 62453 +3143 52240 A 20 15 .119 .017 .430 .028 .065 .160 99.181 6102J 59125 1895 52243 M O 21 96 .119 .009 .440 .043 .077 .160 99.152 61130 60657 473 52233 s 22 19 .123 .014 .440 .030 .063 .160 99.170 59110 59431 + 321 52851 o 23 6 .129 .008 .490 .050 .118 .160 99.045 05020 67354 +2334 53720 2 24 11 .131 .012 .470 .033 .051 .130 99.173 60690 58958 1732 54075 25 13 .134 .015 .480 .035 .045 .150 99.141 58820 58577 243 5452:? k 26 12 .138 .021 .360 .041 .077 .140 99.223 62940 63899 + 959 551(53 BE 27 88 .140 .016 .480 .042 .077 .180 99.065 02S90 68682 + 792 55415 1^ 28 10 .143 .006 .390 .045 .086 .200 99.130 64880 65700 + 820 55891 29 10 .147 .012 .540 .024 .056 .160 99.061 63210 G1752 1458 50484 M 80 12 .151 .012 .640 .033 .051 .130 98.983 62650 61288 1362 57009 81 7 .151 .005 .490 .055 .088 .160 99.051 64930 6(5709 + 1819 57092 82 12 .156 .008 .570 .035 .070 .170 98.991 65180 64830 350 57830 1 83 8 .171 .011 .630 .026 .036 .100 99.023 (52850 62125 425 60142 | 34 4 .178 .008 1.000 .043 .076 .140 98.555 71930 67157 4773 61051 85 8 .183 .014 .680 .030 .027 .100 98.906 65100 62859 2241 61956 86 9 .185 .008 .760 .028 .038 .130 98.851 65590 64201 1329 6222*3 87 6 .193 .009 .670 .020 .036 .100 98.972 052SO 6.5614 + 381 634HS 88 5 .198 .013 .610 .032 .060 .140 98.947 69970 69765 205 64244 89 8 .207 .012 .790 .045 .067 .150 98.729 71210 71286 + 76 65545 40 8 .212 .010 .820 .039 .073 .140 98.706 71870 72710 + 846 60302 41 4 .213 .012 .700 .019 .046 .140 98.870 09750 08687 + 87 66511 42 5 .225 .015 .990 .048 .077 .220 98.425 78700 74471 4221) 68193 43 5 .235 .016 .750 .027 .037 .140 98.795 71170 71796 + 626 69850 44 12 .240 .009 .760 .030 .054 .140 98.767 7232.) 74754 +2434 70605 45 7 .242 .010 .860 .049 .076 .190 98.573 78020 77497 523 70844 46 6 .282 .009 .660 .033 .053 .160 98.8013 76830 81444 +4614 77040 47 6 .282 .010 .770 .023 .043 .140 98.732 76940 79673 4-2733 77015 INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 489 TABLE XVII-N (Continued). d Composition; percent. A Formula No. 1. gi * **-> O o d i! 3 o 33 yiijL .'S i 5 >> IP (H So -3 C o C3 Jj J ^ 2 bc 2^<2 1^*3 s* - 5 5 3 O d 5 W> ^H C2 ci '5 C3 Q> "y "33 ^ % o 33 N OC PH H <1 H> r 48 7 .306 .010 .790 .034 .050 .090 98.720 8268C 841& +1503 80681 5 49 7 .333 .220 .650 .026 .041 .080 98.650 8741C 8764$ + 239 84786 a 50 11 .341 .020 .850 034 .045 .110 98.600 8698C 88605 +1622 85992 .1 51 8 .374 .030 .830 .035 .057 .120 98.554 9075C 95291 +4541 91023 52 14 .890 .220 .680 .023 .034 .080 98.573 9263C 9530C +2670 93476 tH i 53 6 .427 .028 .860 .026 .027 .100 98.532 9930C 9932C + 20 99120 fig~ 54 17 .428 .220 .650 .023 .036 .080 98.563 97270101485 +4218 99284 o ft< 55 16 .438 .220 .690 .026 .033 .130 98.463 102900 10243v 468 100779 22 ~* " 56 14 .477 .240 .690 .025 .080 .080 98.458 107300 107998 + 699 106741 15 2 57 20 .480 .230 .690 .022 .032 .060 98.486 111740108731' 3009 107209 fi 58 13 .480 .090 1.120 .044 .106 .190 97.970 121210 116621 4589 107032 X} 59 18 .507 .061 1.185 .047 .110 .180 97.910 126800 121003 5797 111140 60 10 .527 .250 .720 .027 .032 .070 98.874 116980 11576;] 1217 114358 M 61 10 .554 .230 .680 .022 .032 .090 98.392 122950 120054 2896 118493 62 9 .555 .090 1.130 .042 .109 .190 97.884 123620 128417 +4797 118472 1 Formula No. 2. 63 6 .025 .005 .040 .024 .009 .080 99.817 46420 42570 3850 41511 64 4 .045 .006 .270 .045 .010 .110 99.514 47550 45834 1716 43462 65 4 .050 .009 .330 .026 .007 .190 99.388 47060 46J337 723 439SO 66 4 ..050 .005 .360 .031 .022 .150 99.382 47610 47905 + 295 43928 67 16 .052 .012 .350 .054 .019 .140 99!373 49010! 4777 1237 44131 68 6 .055 .015 .340 .019 .008 .100 99.463 47130! 4703 99 44477 69 7 .055 .005 .220 .030 .012 .140 99.538 475701 4680 769 44506 70 12 .058 .005 .340 .029 .011 .140 99.417 47010 47606 + 596 44769 71 8 .061 .006 .460 .025 .016 .140 99.292 47300 4897 +1673 45031 72 18 .062 .008 .210 .036 .015 .120 99.549 48980 47758 -1222 45235 73 6 .065 .008 .360 .080 .014 .180 99.293 49770 4867 1100 45445 J"" 1 74 17 .070 .013 .350 .034 .031 .140 99.362 49250 5076 +1510 45989 75 22 .074 .005 .360 .023 .007 .130 99.401 48830 4898 + 154 46418 09 76 19 .074 .009 .390 .018 .013 .100 99.396 49150 49700 + 556 46416 g 77 13 .076 .011 .410 .062 .018 .180 99.243 50880 5042 451 46564 78 94 .078 .003 .880 ,031 .016 .110 99.382 49090 50343 +1253 46825 3 79 15 .081 .005 .540 .031 .016 .130 99.197 49220 5142Q +2200 47063 2 80 17 .083 .005 .420 .029 .008 .130 99.325 50910 502H8 -612 47320 i 81 16 .083 .006 .570 .035 .017 .110 99.179 51060 51881 + 822 47263 Q 0} 82 26 .084 .009 .250 .033 .021 .140 99.463 50900 50777 123 47477 ft 83 23 .085 .014 .380 .032 .036 .140 99.313 51140 522i +1782 47522 O 84 21 .090 .006 .400 .018 .015 .100 99.871 51200 51592 + 392 48062 O 85 121 .093 .006 .400 .032 .019 .130 99.320 5103!) 52255 +1229 48352 1 86 17 .093 .006 .400 .038 .040 .160 99.263 53020 54218 +1193 48330 PQ 87 21 .094 .011 .430 .036 .046 .180 99.203 54800 '55016 + 216 48410 88 14 .096 .007 .440 .065 .023 .160 99.209 53000 53115 + 115 48619 M 89 19 .099 .012 .280 .035 .029 .160 99.385 52950 ' 53210 + 260 48998 M 90 14 .100 .009 .660 .029 .019 .150 99.033 53380 54249 + 869 48965 H 91 5 .102 .010 .470 .087 .027 .150 99.154 53600 54249 + 649 49219 92 15 .108 .013 .440 .064 .027 .130 99.223 54950 54221 729 49349 "53 93 15 .108 .008 .420 .019 .018 .110 99.317 52910 53822 + 912 49903 "C 94 125 .109 .010 .430 .031 .021 .120 99.279 52380 54246 +1266 49992 S 95 03 .112 .005 .420 .034 .025 .160 99.244 54880 54866 14 50288 H 96 23 .115 .009 .430 .031 .009 .130 99.276 52750 53736 + 986 50611 97 13 .117 .007 .460 .035 .053 .130 99.198 57210 58211 +1001 50788 98 15 .118 .014 .490 .057 .033 .140 99.148 56980 56573 407 50872 99 18 .120 .004 .430 .018 .020 .120 99.288 54860 55293 + 433 51133 00 7 .121 .008 .540 .032 .056 .140 99.103 60580 59294 -1286 51165 01 11 .125 .012 .670 .038 .025 .130 99.000 56680 57440 + 760 51538 02 10 .125 .019 .540 .060 .036 .110 99.110 58790 57829 961 51581 03 16 .126 .008 .620 .028 .024 .140 99.054 55090 57206 +2116 51663 04 19 .131 .008 .300 .029 .022 .130 99.380 54690 55966 + 1276 52307 05 20 .132 .006 .390 .027 .009 .130 99.306 54890 55295 + 405 52382 06 63 .132 .010 .470 .033 .028 .190 99.137 56870 57440 + 570 52316 07 9 .134 .016 .510 .036 .055 .110 99.139 59110 60400 +1290 52528 08 15 .136 .009 .310 .029 .024 .130 99.362 57010 6671P -292 62817 490 METALLURGY OF IRON AND STEEL. TABLE XVII-N (Continued). p, rt 2 Composition; percent. ft Formula No. 2. 1 1 ti o 00 r? O Us jjlU 3 ~ s ! *1 C5,a d ! 3 I h J It -j ^ d opd . 03 s S5 ill ,Q a *o ~ as 3 bo 11 8 M d i A Sf | | o df 2 o>,S > 03 to' ^-S S <3 oo lea r.''~ & 55 Jj w 3 M g. 133 6 .231 .029 .360 .025 .012 .120 99.223 67530 65628 1902 62591 s 134 5 .233 .008 .490 .020 .021 .130 99.098 67560 67322 238 62750 135 5 .260 .060 .810 .025 .014 .100 99.231 68470 68554 + 84 65595 136 5 .311 .080 .440 .029 .020 .070 99.050 73010 75013 +2008 70800 187 5 .338 .025 .620 .026 .017 .100 98.874 77950 78410 + 460 7 *?'"*") that the effect of each unit of carbon decreases as it is approached. If this relation holds good throughout the whole series of alloys, then each successive increment of carbon will have a less effect from the starting point of pure iron. It is also possible for the same reasons that every other metalloid will follow the same rule, so that the influence of each separate alloyed element will be represented by a curve. This may be an arc of a circle, or a parabola, or a cycloid, or a broken line ; it may be different in degree or different in nature in the case of each element ; and it may vary in degree or even in nature with changes in the proportions of the associated elements ; but it will be as- sumed in this investigation that within the narrow limits of the divisions of the table, the effect of a regular increase in the per- centage of each metalloid would be represented by a straight line. In other words, that an increase of carbon from .20 to .21 per cent, gives the same increment in strength as an increase from .10 to .11 per cent. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 491 If this last assumption be true, then the seemingly erratic devi- ations of the curves in Fig. XVII-A and Fig. XVII-B from a straight line are due to variations in the associated percentages of silicon, manganese, sulphur, phosphorus and copper. It seems pos- sible to find the effect of these elements by the method of least squares. Each group may be regarded as an equation containing seven unknown quantities, the combined effect of which produces a certain ultimate strength. If A is written for the effect of .001 per cent, of carbon upon the ultimate strength, B lor silicon, C for manganese, D for sulphur, E for phosphorus, F for copper, and G for iron, then Group I will take the following form : 82 A+6 B+290 C+34 D+34 E+120 F+99,434 G=52,090. Curve sh owing- Relation of ;he Chem cal Composition o / \^ ^B Acid Curve_A Curve B Curve C Open-He A=frora B=froir C= strei irth Stee Division Division gth due 1 to its Ult I, II, o Carbon mate Sti and Iron ength. only,. V^ 'C 100;OQO 90,000 80,000 70IQPO 0,000 80,000 0#X> L^ Abscissa s Carb yn, per ce nt. i A Ordinatt s = Ultin ate Stren gth,ibs.] >er sq. in, I S s A t M \ C tM jy / A/ rj S > 3 S _S I0\ S 20< *> . > ,5 _"" IV/N S ^ * u \ ^ ..^ I6 N > 24 \ -*** *> > "^ 2I > < " < 12 > '> tt 4? 13 CHART SHOWING METHOD OF DETERMINING SEVEN UNKNOWN QUANTITIES IN SEVEN EQUATIONS. After the determination of the final factor in equation No. 28, the value of each element is successively determined by substitution in ISTos. 26, 23, 19, 14, 8 and 1. After the last unknown is thus INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 495 found the values are substituted in No. 7, and if it is then found that the results agree to the seventh or even to the sixth logarithmic place, it may confidently be asserted that the values are correct to the third and sometimes to the fourth integer, and this is amply sufficient for the work in hand. Notwithstanding such methods of proof and the reasonable, although in some respects the unexpected, nature of the results just given from Division I, it is with no little disappointment that I am forced to confess that further investigation throws grave doubts on the validity of this method of least squares when applied to such a number of unknown quantities, and when any one of these quantities is of very little importance. The reasons for this con- clusion will appear in the results shown in Table XVII-0, which were obtained from the normal equations derived from the groups composing Division II. TABLE XVII-0. Effect of Certain Elements upon the Strength of Steel as Deter- mined from Division II in Table XVII-N. Order of solution. Effect of .001 per cent. Strength of pure iron. C. Si. Mn. S. P. Cu. Forward . . . 1 +148.403 Backward . . | +148.402 +36.030 +36.012 +29.069 +29.056 +867.923 +367.4(57 34.340 34.203 29.110 29.088 42347 42530 NORMAL, EQUATIONS FROM THE HIGH ACID OPEN-HEARTH HEATS, CONSTITUTING DIVISION II OF TABLE XVII-N Equation from A; 3,008,187 A + 990,763 B + 5,453,835 C + 202,211 D + 351,674 E + 743,270 F + 650,950,560 G = 708,894,410. Equation from B; 990,763 A + 441,605 B + 1,596,765 C + 57,835 D + 91,526 E + 207,280 F + 212,514,226 G = 230,460,700. Equation from C; 5,453,335 A + 1,596,765 B + 10,427,125 C + 391,865 D + 704,040 E + 1,447,300 F +1,201,479,570 G = 1,298,675,100. Equation from D ; 202,211 A + 57,835 B + 391,865 C + 14,854 D + 26,895 E + 54,730 F + 44,851,610 G = 48,331,270. Equation from E ; 851,674 A + 91,526 B + 704,040 C + 26,895 D + 52,914 E + 102,250 F + 76,070,701 G = 84,275,880. Equation from F; 743,270 A + 207,280 B + 1,447,300 C + 54,730 D + 102,250 E + 208,300 F + 162,236,870 G = 177,275,000. Equation from G; 650,950.560 A + 212,514,220 B + 1,201,479,570 C + 44,851,610 D + 76,070,701 E + 162,236,870 F + 145,264,790,433 G = 154,504,087,640. After laborious attempts to find any mathematical error, I am certain that the discrepancies between the results found by solving in reverse order are due solely to logarithmic errors, and could 496 METALLURGY OF IRON AND STEEL. only be lessened by using logarithm tables of more than seven places. But these errors are of no importance, and it is certain that the values are approximately correct, mathematically speaking, al- though they are absurd from a practical point of view. If .001 per cent, of sulphur did actually cause an increase of 367 pounds, then .06 per cent., which is a very common content, would increase the strength 22,000 pounds, when in reality its effect is very slight, if it is even appreciable. Phosphorus is shown as a minus quantity, which is entirely wrong, and copper is given at 29 pounds, which is equivalent to saying that one-half of one per cent, would reduce the strength 14,500 pounds, when, in fact, a content of even one per cent, does not seem to have any effect at all. These ridiculous values place in question the validity of the method of least squares, by which they were determined, and the next section will attempt to survey the territory over which it has jurisdiction. SEC. XVIIo. Application of the method of least squares as Urn- ited by the conditions of the problem. The fundamental difficulty in the solution of Division II is the fact that the iron is not self- determining. The highest percentage of iron in any group of the division is 98.720, and the lowest is 97.884, being a ratio of less than 101 to 100. It is true that the ratios in Divisions I and III are very little higher than this, but in both these cases there is a determining condition in the fact that there are a number of groups which are nearly pure iron, and it will evidently be less probable that a wrong result will be found under such circumstances. The only way, therefore, of obtaining an intelligent result for Division II is to make the iron self-determining, and since this cannot be done within the limits of the division, it is necessary to combine it with Division I. This combination may be regarded as unjustifiable, since the effect of carbon decreases after a certain point is passed, but it can be answered that the curve in Fig. XVII-A gives no sign of falling, and that the value of carbon just found for Division II is greater than for Division I. Moreover, it will be shown in Table XVII-P that the value of carbon as found by the combination of I and II is higher than for I alone, so that there is good warrant for the union of the two. This conjunction will tend to prevent an absurd result in the case of iron, and will give a better value for carbon ; but it will not prevent a wrong estimation of an element like copper, which has INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 497 very little influence upon the tensile strength. It is certain that the equations of condition are not absolutely accurate, owing to the limitations of chemical research and the variations in the rolled test-bars. These errors are incorporated into the normal equations, and are distributed in the final solution so as to give the best mathe- matical result. It does not follow that the values so found will accurately repre- sent the actual practical state of affairs, for a purely fanciful result is not an unusual phenomenon in mathematics ; thus, in the solution of every quadratic equation, two values are always produced by the plus and minus roots, and one of these values is often inapplicable to the original conditions. This occurred in the derivation of the curves given in Fig. XVI-B, for there were two possible conic sec- tions discovered in each case. . One of them fitted the problem, while the other was a reverse curve exactly similar to the first, but situated for the most part in minus territory, and having an exist- ence only as a mirage of the true solution. To prevent such a purely mathematical answer to the present practical problem it is necessary to discard two sets of variables: (1) Those which are known to have very little effect. v (2) Those which are present in very nearly constant proportion. If an element has no effect, then it cannot be self-determining, but may be forced to bear 'all the results of analytical errors. If it is present in nearly constant quantity, then the slight variations can have very little determining effect. From one point of view these limitations beg the question, for it becomes necessary to know in a general way the influence of an element before its value can be quantitatively determined. The ultimate logical consequences of such a provision need not be dis- cussed, for, in the problem under consideration, it is known that copper has scarcely any influence upon the tensile strength, and that the same is true of sulphur when present in ordinary propor- tions. In the case of silicon there is a chance for greater hesitation, but it will be noticed that in only eight groups is the content of this metalloid above .20 per cent., while in only three other groups, or 11 in all, is it over .03 per cent. Within the limits of .00 and .03 per cent., which thus includes five-sixths of the groups, the power of silicon is not enough to disturb the calculations. SEC. XVIIp. Effect of carbon, manganese, phosphorus and iron 498 METALLURGY OF IRON AND STEEL. upon the ultimate strength. Having thus decided to neglect the- effect of silicon, sulphur and copper, the equations of condition are simplified so that they take the following form : EQUATIONS OF CONDITION. From Group I ; 82 A +290 C + 34 E + 99434 G = 52090. From Group II ; 105 A + 880 C + 74 E + 99193 G == 57375. From these may be deduced the following normal equations : NORMAL EQUATIONS, DIVISION I. Equation from A; 1,210,191 A + 4,298,830 C + 450,670 E + 710,516,809 G = 471,142,635. Equation fromC; 4,298,830 A + 15,861,200 C + 1,644,430 E + 2,581,030,930 G = 1,697,750,700. Equation from E; 450,670 A + 1,644,430 C + 215,997 E + 300,954,795 G = 194,090,210. Equation from G; 710,516,809 A + 2,581,030,930 C + 300,954,795 E + 460,910,659,759 G = 296,665,604,300. NORMAL EQUATIONS, DIVISION II. Equation from A ; 3,008,187 A + 5,453,335 C + 351,674 E + 650,950,560 G = 708,894,410. Equation from C ; 5,453,335 A + 10,427,125 C + 704,040 E + 1,201,479,570 G; = 1,298,675,100. Equation from E ; 351,674 A + 704,040 C -f 52,914 E + 76,070,701 G = 84,275,880. Equation from G; 650,950,560 A + 1.201,479,570 C + 76,070,701 E + 145,264,796,463 G = 154,504,087,640. NORMAL, EQUATIONS, DIVISION III. Equation from A; 1,505,996 A -f 4,700,050 C + 225,664 E + 954,850,000 G = 574,293,000. Equation from C; 470,005 A + 1,723,710 C + 83,790 E + 340,994,800 G = 198,609,150. Equation from E; 225,664 A + 837,900 C + 48,942 E + 169,769,400 G Equation from G; 9,548,500 A+ 34,099,480 C + 1,697,694 E + 7,882,188,000 G = 4,206,995,000. NORMAL, EQUATIONS. DIVISIONS I AND II COMBINED. Equation from A; 4,218,378 A + 9,752,165 C + 802,844 E + 1,361,467,000 G = 1,180,037,000. Equation from C; 9,752,165 A + 26,288,830 C + 2,348,470 E + 8,782,511,000 G- = 2,996,426,000. Equation from E; 802,844 A + 2,348,470 C + 268,911 E + 877,025,500 G = 278,366,090. Equation fromG; 18,614,670 A +37,825,110 C + 8,770,255 E + 6,061,755,000 G = 4,511,697,000. These equations, when solved, give the values shown in Table XVII-P. In two cases the elimination has been performed in the order G f E, C, A, and has then been repeated "backward" in the order A, C, E, G. The comparison of results shows the degree of INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 499 accuracy obtained. In the other two cases the work was not re- peated in this manner, but the table gives two values of iron. These two determinations are the result of substitution in the ex- treme equations, as shown by the chart on page 494, and the almost perfect agreement of the two proves that the work is correct within the limits of logarithmic error. TABLE XVII-P. Effect of Carbon, Manganese and Phosphorus upon the Strength of Iron, as Determined from Table XVII-N by the Method of Least Squares. NOTE. All values are in pounds per square inch. No. of division. (See Table XVII-N.) Order of solution. Effect of .001 per cent. Strength of pure iron. - Carbon. Manganese. Phosphorus Division I. Forward, +141.4929 3.086216 +109.3771 37139.65 87139.67 Division II. Forward, Backward, +166.8914 +166.8939 +3.921577 +3.928512 + 97.28167 + 97.24250 23236.27 23231.43 Divisions I and II combined. Forward, Backward, +152.9212 +152.9203 3.902156 3.901182 +131.6955 +131.6965 84326.69 84326.22 Division III. Forward, +103.4560 +5.298315 + 94.08509 88996.13 88996.14 The values are given for Division I in order that they may be compared with those found by combining Divisions I and II. They are given also for Division II separately, in order to corroborate what was said in Sections XVIIn and XVIIo on the worthlessness of any solution of this division by itself. The value of 23,236 pounds for the strength of pure iron is absurd, and, of course, this renders worthless all the other factors, but the coincidence of the results when the equations were worked in opposite directions proves conclusively the accuracy of the work. Moreover, I have applied these values to the separate groups of Division II, and the greatest discrepancy in any one group between the actual and the calculated strength is 6784 pounds, while the sum of the plus and minus errors is only 4.2 pounds, being an average error of only 0.28 pounds for each group. This shows again, what has been insisted upon elsewhere, that perfectly correct mathematical results may be inapplicable to the practical conditions unless the factors are self-determining. The values found by the combination of Divisions I and II, and 500 METALLURGY OF IRON AND STEEL. the values given for Division III, are those which have been applied to the groups in Table XVII-N under the titles of Formulae No. 1 and No. 2. The antepenultimate column gives the tensile strength as calculated from the formula,, while the penultimate shows the error, or the difference between this calculated value and the result found by the testing machine. The accuracy of the formulae may be judged from the fact that the sum of the plus and minus quantities for the 'acid steels, comprising Divisions I and II, is 29 pounds, being an error of half a pound for each group. In the case of the basic steels the error is only 5 pounds, or only one-fifteenth of a pound for each group. SEC. XVIIq. Value of carbon and phosphorus when manganese is neglected. In the preceding section it has been shown that man- ganese is a plus quantity in basic steels, and a minus quantity in acid metal. These contradictory values nray seem improbable, although they are by no means impossible. In order to get a little more light on the subject, I have arbitrarily divided the list of groups, given in Table XVII-N, into two sets, and have determined the most probable values of carbon, manganese and phosphorus for each set. It would naturally be expected that the results from one-half the number of groups would be less valid and less uniform than from the complete list, but they may nevertheless be of use as corroborative evidence. The method of dividing the list was to take the odd numbers for one set and the even numbers for the other. Inasmuch as the original arrangement is on the basis of carbon content alone, it will be evident that this insures a fair division without any chance of selection in aid of any preconceived theory. It would have been much better if a calculation could have been made on those groups showing low manganese, and those with high manganese, but as the low steels did not offer any examples of a high content of this element, and the high steels did not offer any examples of a low content, the result would have been of no value. In the case of acid steel a mistake was made in taking for this arbitrary division the original list of groups, which, of course, was made up before the determinations of carbon were made by combustion. On comparing the numbers on this original list with the new arrangement, it was found that the two sets of so-called "odd" and "even" numbers really embraced the following groups, as given in Table XVII-N, after they had been renumbered : INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 501 Odd numbers: Groups, 1, 2, 3, 8, 9, 10, 13, 15, 18, 22, 23, 25, 26, 27, 29, 31, 33, 36, 38, 41, 42, 43, 44 and 45. Even numbers: Groups 4, 5, 6, 7, 11, 12, 14, 16, 17, 19, 20, 21, 24, 28, 30, 32, 34, 35, 37, 39, 40, 46, and 47. Inasmuch as one arbitrary division seems to be as good as another and as the calculation is very laborious, it was deemed unnecessary to repeat the work simply for the sake of uniformity, but this explanation is made for the sake of -any mathematician who might wish to test the accuracy of the solution. In the case of the basic steel, the odd and even figures were taken as they stand in Table XVII-K The results are given in Table XVII-Q. TABLE XVII-Q. . Values of Carbon, Manganese, Phosphorus, and Iron, obtained by Arbitrarily Dividing the List in Table XVII-X According to Odd and Even Numbers and Solving Each Division by the Method of Least Squares. Factor. Kind of steel. Value in pounds per sq. inch. Odd. Even. Combined. .01 per cent, of carbon Acid, Basic, + 1554 +1069 +1502 + 992 +1529 + 1035 .01 per cent, of manganese Acid. Basic, 0.18 +20 107 + 85 89 + 53 .01 per cent, of phosphorus Acid, Basic, +1451 + 799 + 1032 + 1100 +1317 + 941 Pure iron . . . Acid, Basic, 80824 40303 40519 37749 84327 88996 It will also be seen that in each case the "combined" value, which is the original value given in Table XVII-N, is very close to an average of the odd and even. This is by no means a foregone con- clusion, and would not follow if the factors were not self-deter- mining to a great extent. It will also be seen that there are variations in the values of each one of the factors, but that manganese shows the widest range. In the acid steel the figure for the even numbers is 107, while in the odd numbers it is only a small fraction. The varia- tions in phosphorus are very small when compared with this, while those of carbon are insignificant. The value of iron must neces- sarily change with the other elements, since it is less self-determin- ing than carbon or phosphorus. .502 METALLURGY OF IRON AND STEEL. The great differences found in the values of phosphorus in the odd and even subdivisions of the basic heats are easily explained. An examination of the table will show that of the odd numbers there are only four groups showing more than .04: per cent, of phosphorus, and only three groups in the even numbers. There is therefore too little variation for the phosphorus to have an over- powering self-determining effect. The combined figures are sub- ject to the same criticism,, but the larger number of groups gives the results a greater validity. Taking into consideration the fact that manganese is indicated as positive in basic and negative in acid steels, and that it gives wide differences in value between the odd and even lists, it would seem reasonable to suppose that it has very little effect at all when present in usual proportions, since the method of least squares should give a reliable result for an element which has a strong and positive action, when such an element is present in widely varying proportion. Accepting such a conclusion, it remains to be seen whether a true formula can be deduced by omitting manganese altogether, and ascribing all the variations in tensile strength to the carbon, phos- phorus, and iron. On this new basis the following normal equations -are formed, the solutions of which are given in Table XVII-E. NOBMAL EQUATIONS, OMITTING B, C, D, AND F. DIVISIONS I AND II COMBINED. Equation from A; 4,218,378 A + 802,344 E + 1,361,467,000 G = 1,180,037,000. Equation from E ; 802,344 A + 268,911 E + 377,025,500 G = 278,366,090. Equation from G; 13,614,670 A + 8,770,255 E + 6,061,755,000 G = 4,511,6OT,000. DIVISION III. Equation from A; 1,505,996 A + 225,664 E + 954,850,000 G = 574,293,000. Equation from E ; 225,664 A + 48,942 E + 169,769,400 G = 98,593,980. Equation from G; 9,548,500 A + 1,697,694 E + 7,382,138,000 G = 4,206,995,000. The data in Table XVII-E may be expressed in simple formula?, an allowance being made for the fact that there is never quite 100 per cent, of iron in any steel. In Table XVII-S these formula? are applied to the groups of metals given in Table XVII-N. In order to see whether these formula? satisfy all the classes of steels under consideration, the results in Table XVII-S may be analyzed by the following method : Silicon : Referring to the acid steels in Table XVII-N, it will be found that there are eight groups containing .22 per cent, or INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 503 over of silicon. Four of these, 49, 55, 56, and 60, show an error in Table XVII-S of less than 2000 pounds. There are two groups, 52 and 54, having an aggregate plus error of 6390 pounds, and two groups, 57 and 61, with an aggregate minus error of 7080 pounds. Thus there is .no evidence that silicon influences the result. TABLE XVII-E. Effect of Carbon and Phosphorus upon the Strength of Iron. NOTE. All values are expressed in pounds per square inch. Kind of steel. Effect of .001 per cent. Strength of pure iron. Carbon. Phosphorus. Acid steel ; Divs. I and II, Basic steel; Division III, 148.495 108.542 126.449 119.707 83212.2 40196.5 Sulphur: There are six groups in the acid steel, 2, 10, 11, 17, 18, and 31, which contain .055 per cent, or more of sulphur, and none of these shows an error in Table XVII-S of over 2000 pounds, In the basic steels there are eight groups, 73, 77, 88, 91, 92, 98, 102, and 110, showing over .055 per cent., and the greatest error in any of them in Table XVII-S is 1680 pounds. Thus the sulphur does not seem to affect the situation. Manganese: There are sixteen acid groups containing .75 per cent, or more of manganese. Of these there are six, 36, 39, 40, 43, 45, and 53, which have an error in Table XVII-S of less than 2000 pounds, while group 48 is only 60 pounds above this figure. Of the remainder there are five groups, 44, 47, 50, 51, and 62, giving an aggregate plus error of 19310 pounds, and four groups, 34, 42, 58, and 59, with an aggregate minus error of 13720 pounds. This would indicate, if it indicates anything, that manganese has a minus value in the acid steels, which is in accordance with the mathematical deductions of the last section. Among the basic groups there are only two, 120 and 128, which contain more than .75 per cent, of manganese. These two show an aggregate minus error in Table XVII-S of 5750 pounds. There are six other groups with a content of manganese between .65 and .75 per cent. Five of these, 90, 101, 109, 115, and 127, show an error under 2000 pounds, while the remaining group gives a minus error of 2340 pounds. There is, therefore, a slight indi- 504 METALLURGY OF IRON AND STEEL. cation that manganese strengthens basic steel, as was discovered in the last section. TABLE XVII-S. Ultimate Strength of the Steels Given in Table XVII-X as Com- pared with the Eesults Obtained from the Following Formula?. Formula for Acid Steel; 83,000+1485 C+1260 P = Ultimate Strength. Formula for Basic Steel; 40,000+1085 C + 1200 P = Ultimate Strength. P. Ultimate strength. d Ultimate strength. p. Ultimate strength. 2 flj 2 dj 2 a bfi g a Hit 4-t O r-< ^ Ilsl ha f 3 11" I 1 1 la ,2 * 1 a 1 li -> 03 o to umbci "3 p w II TtG Sl|| fc 5 O^ P fe < 0^ q < Q 1 52090 49460 2630 2-> 59110 59200 + 90 4;: 71170 72560 + 1390 2 57375 57920 + 515 23 65020 67030 +2010 4-1 72320 75410 +3120 8 57310 67500 + 190 21 60690 58880 1810 45 78020 78510 + 490 4 57430 59520 + 2090 25 58820 58570 250 46 76830 81560 -1730 5 57140 57470 + 330 26 62940 63200 + 260 47 76940 80300 -3360 6 62870 61880 990 27 62890 63190 + 600 48 82680 84710 +2060 7 55450 54740 710 28 64880 65070 + 190 49 87110 87620 + 210 * 8 58780 58770 10 29 63210 61890 1320 50 86980 89310 -2330 9 56830 54510 2320 30 62650 61850 800 51 90750 95720 -1970 10 60870 60560 310 31 64950 66510 +1500 52 92630 95200 -2570 on 11 62610 61110 1170 32 65180 61990 190 53 99300 99810 + 510 2 12 61190 59570 1620 33 62850 62930 + 80 64 97270 101090 -3820 j 13 61430 62850 . 1120 31 71930 69010 2920 55 102900 102200 700 *< 14 56990 51930 -2060 35 65100 63580 1520 56 107300 107620 f 820 15 59110 59970 + 860 33 65590 65260 330 57 111740 108310 3130 16 59350 58160 890 37 65280 66200 + 920 58 121210 117610 3570 17 59260 60730 +1170 38 69970 69960 10 59 126800 122150 4650 18 60900 60100 800 39 71210 72180 + 970 60 116980 115290 1690 19 59310 62010 + 2700 40 71870 73680 +1810 61 122950 119300 3650 20 61020 58860 2160 41 69750 70130 + 680 62 123620 129150 -5530 21 61130 60370 760 42 78700 76120 2580 63 46120 43790 2630 88 53000 53180 + 180 113 57030 58600 -1570 64 47550 46080 1170 89 52950 54220 + 1270 111 57060 57270 + 210 65 47060 46270 790 90 53880 531:30 250 115 60870 59260 1610 68 47610 48070 + 460 01 63600 54310 + 710 110 63480 61890 1590 67 49010 47920 1090 92 54950 54420 530 117 58970 58010 - 930 63 47130 46930 200 52910 63880 + 970 118 60770 60200 570 69 47570 47410 160 91 52980 54350 + 1370 119 59110 59220 + 110 70 47010 47610 + 600 95 64880 55150 + 270 120 63400 60380 -3020 71 47300 48540 +1240 93 62750 63560 + 810 121 63710 61510 2230 __. 72 48980 48530 450 97 57210 59050 +1840 122 60810 612CO + 480 73 49770 48730 1010 98 66980 56760 220 l*^* 63110 62650 460 71 49250 51320 +2070 99 54860 55420 + SCO 121 60710 6Q06Q + 220 OB 75 48830 48870 + 40 100 60580 59850 730 125 60870 61920 -105.0 2 70 49150 49590 + 440 101 66680 56560 120 12f 67570 65230 -2340 3 77 50880 50110 470 102 58790 57880 910 127 66480 66260 220 78 49090 503X0 + 1290 103 65090 56550 + 1460 128 67180 64750 2730 79 49220 60710 +1490 101 51690 56850 +2160 12S 66820 64590 2*30 80 60910 49970 910 10) 54890 55400 + 510 130 63600 63330 270 81 51060 51050 10 103 66870 57680 + 810 131 63740 64910 -^1200 82 50900 51630 + 730 1C7 59110 61140 +2C30 132 63470 64650 -t-iWO 83 51140 53510 + 2100 103 57010 57640 + 630 188 67530 66500 1030 81 51200 51570 + 870 109 59110 58830 280 131 67560 67800 + 240 85 51030 52370 +1310 110 60570 58890 KJ80 68470 69890 + 1420 86 6302-0 51800 +1870 111 58860 58740 120 13< 73010 76140 +3130 a POO 55720 + 920 112 58970 58980 + 10 137 77950 78710 f 760 Phosphorus: There are thirteen acid groups containing .08 per cent, of phosphorus or more, and seven of these, 6, 10, 11, 13, 17, INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 505 28, and 31, have an error in Table XVII-S of less than 2000 pounds. Of the remainder, four groups, 4, 19, 23 and 62, give an aggregate plus error of 12,330 pounds, and two groups, 58 and 59, give a minus error of 8220 pounds. This would indicate that the value of phosphorus in the acid steels is nearly correct but that it may be a trifle too high. The basic metals contain no examples of high phosphorus, and hence the value found cannot be corroborated. It will be found that these deductions must be materially modified on account of the investigations chronicled in Part III. In the later work the value of iron is nearly the same in acid and basic metal. This assuredly seems more in accord with reason, and gives greater force to the values found for the metalloids. The above calculation will be of interest to show how nearly an arbitrary equation can fit the case. PART III. EFFECT OF CARBON, MANGANESE, AND PHOSPHORUS UPON THE TENSILE STRENGTH OF IRON, AS DETERMINED BY SPECIAL MATHEMATICAL INVESTIGATIONS. INTRODUCTORY NOTE. A general synopsis of the argument and conclusions of both Parts II and III is given in Section XVIIw. SEC. XVIIr. Values of carbon, manganese, phosphorus, and iron in a new series of acid steels. In the introductory note to Part II of this chapter it was stated that a second series of steels had been investigated. The method employed in the formation of the groups was the same as described in Section XVIIm, and all the details of the work were performed by the same men that conducted the previous examination. The two series, which we may call the "old" and the "new," are therefore of equal force and virtue, and the testimony of the one must always be considered in connection with the testimony of the other. It was proven in Section XVIIo that the influence of silicon in small proportions was so slight that it did not make a satisfactory working factor for the method of least squares. The same was found true of sulphur and copper. In plotting the records of acid steel of the new series, however, it was found that the groups that contained high silicon seemed to show a greater tensile strength 506 METALLURGY OF IRON AND STEEL. than steels of low silicon with the same content of carbon. As this was not the case in the old series, the groups were all put together in the former calculation, but in the light of this new evidence it would seem proper to separate them on the basis of their silicon content. This is easily done, since in both cases the high-silicon! heats were put together in separate groups. In the low-silicon groups neither the total content nor the variations in this element seem sufficient to materially disturb the result. The normal acid steels of the new series are shown in Division I of Table XVII-U, and the normal acid steels of the old series in Division II. They are both combined to give the line A A in Figure XYII-C. The high-silicon steels of the new series are given in Division III, and those of the old series in Division IV. They are both combined to give the line BB in Figure XVII-C. The high-manganese and high-phosphorus steels of the old series are placed in Division V, but are not shown in the diagram. Considering only the normal acid steels of both the old and the new series, as enumerated in Divisions I and II, a calculation was made by the method of least squares to find the values for carbon, manganese, phosphorus, and iron, which would most nearly satisfy the conditions of the problem. The results are shown in Table XYII-T. TABLE XVII-T. Values of Carbon, Manganese, Phosphorus, and Iron, as Determined by the Method of Least Squares from the Normal Acid Steels in Divisions I and II in Tables XVII-U. Series. Influence of .01 per cent, in pounds per sq. inch. Carbon. Manganese. Phosphorus. Iron, New series.Division I +1126 +1868 + 3 -23 + 716 +1068 +4.0439 +8.7544 Old series, Division II In the old series as originally formed, including the abnormal steels, the value of .01 per cent, of manganese was 39 pounds. In the revised series, after omitting these groups, it is 23 pounds, while in the new series the value deduced is + 3. It would appear, therefore, that manganese does not make a satisfactory working fac- tor in the calculations on acid steels, while the .values obtained for INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 507 it, in addition to being contradictory, show that it does not have a very important influence. In the following section, therefore, I have computed a formula from carbon, phosphorus, and iron alone, and have then compared the ultimate strengths as calculated from this formula with the actual tensile tests. SEC. XVIIs. Values of carbon, phosphorus, and iron in acid steel when manganese is neglected, as determined from the normal steels of the old and the new series combined. Considering only the normal steels as given in Table XVII-U, and omitting man- ganese from the problem, we shall have by the method of least squares the following equations, in which A = the influence of .001 per cent, of carbon, expressed in pounds per square inch, B = the influence of .001 per cent, of phosphorus/ and C = the influence of .001 per cent, of iron. ACID STEELS;* DIVISIONS I AND II IN TABLE xvn-u. Equation from A 3,227,256 A+1,065,433 B+1,676,848,333 C=l, 130,441,385. Equation from B Equation from C =670,977,073,9 1,065,433 A+488,892 B+689,328,873 0=441,177,250. 1,676,848,333 A+689,328,872 B+1,043,135,334,268 C 0. The solution of these equations gives the following values : Lbs. per sq. In. Effect of .001 per cent, of carbon -t- 121.6 Effect of .001 per cent, of phosphorus -f 88.9 Strength of pure iron. 38,908 There is never quite 100 per cent, of iron in any steel, so that it would not be right to take the above determination of iron as a starting point. Theoretically it would be necessary to calculate the value of iron for each separate metal, and this was done in Table XVII-N"; but for practical purposes it will be assumed that struc- tural steel contains 99.2 per cent, of iron, which by the above de- termination should confer a strength of 38,600 pounds per square inch for acid metal. It then becomes practicable to write the following formula by which the strength of acid steel may be calculated when the per- centages of carbon and phosphorus are known, the answer being expressed in pounds per square inch. Acid Steel ; 38600+121 Carbon+89 Phosphorus-}- R=Ultimate Strength. * The sum total of the coefficients in these equations is not quite 1,060,000,- 000,000, as it should be theoretically, because the factors in the old series relat- ing to silicon, sulphur and copper have been omitted. 508 METALLURGY OF IROis &ND STEEL. The unit for carbon and phosphorus is .001 per cent. The factor R represents an allowance for the conditions under which the piece is rolled, whether finished hot or cold. In the present series of groups it is zero. itmc T^o5T 190.000 Curv< ical and 1 ?s showing the relation between the cl composition of acid open-hearth stee its ultimate strength, normal steels, Divisions I and II, nigh-silicon steels, Divs. Ill and IV, mreiroii-f pure carbon, calculated froir nula 38600+ 121 carbon=ultimate strenj ssas=carbon, per cent. iates=ultimate strength, Ibs. per sq. ii S ' B iem- 1 Hf /> BB=1 CC=p fori "Absci Ordir i the jth. "S f /V X x c ich. | V l> / 0.00 A\ * ^ r M.OtQ 1 JA ^ > 70,000 i/lr > ^ J\J 0,000 *j m X' 30.000 c^X X 40,000 .03 JO a 20 ,25 .90 .99 A9 ,49 M M TIG. XYII-C. CURVES SHOWING EELATION BETWEEN THE CHEMI- CAL COMPOSITION OF ACID OPEN-HEARTH STEEL AND ITS ULTIMATE STRENGTH AS GIVEN IN TABLE XVII-U. y In Table XVII-U this formula has been applied to all the steels, both normal and abnormal, and the differences between the actual and the calculated ultimate strength have been placed in the last column. This difference will sometimes be spoken of as the "error" in subsequent remarks, as being the discrepancy between the re- corded results and those obtained by calculation. An examination of this column reveals several notable points. First : Group 54 is entirely abnormal. It is almost identical in composition with Group 53, and yet differs from it by 4200 pounds in strength. The fact that No. 53 is an average of twelve heats and conforms to the formula, while No. 54 is an average of only four heats, points to the latter as an erratic member which has INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 509 gome bar-sinister in its history. Out of numerous possibilities it is only necessary to mention that one of the test-bars might have been wrongly marked. This group will be neglected in the following observations. Second : There is a decided difference between the old and the new series. The sum of all the plus values in Division I of Table XVII-U, after omitting Group 54, is 49,310 pounds, while the sum of the plus values in Division II is only 7290 pounds. The sum of the minus values in Division I is 10,930 pounds, while in Division II it is 57,780 pounds. The individual records corroborate these totals, for in Division I there are 39 groups where the error is plus, and only 16 groups where it is minus. On the other hand, Division II furnishes only 11 groups where the error is plus, while it has 38 groups where it is minus. This seems too decided a record to be the result of chance, yet, as before stated, the two investigations relate to steels which were made in the same furnaces and handled by the same men, while the physical and chemical determinations were made on the same apparatus and by the same operators. In the light of this evidence it is not remarkable that results from different sources are some- times inconsistent. i Third : There are seven groups among the normal acid steels where the actual strength is more than 2000 pounds below the cal- culated, and six of the seven, Nos. 29, 36, 45, 46, 49, and 50, show no striking peculiarity. The other group, No. 55, is low in phos- phorus and sulphur, and rather high in manganese. On the other hand, there are ten groups where the actual strength. is more than 2000 pounds above the calculated, and six of these, Nos. 90, 94, 98, 101, 104, and 105, show a high content of man- ganese. Of the others, No. 1 is low in manganese and high in sulphur, No. 62 is high in phosphorus, No. 67 is normal, and No. 76 is low in sulphur. Thus the only point that is gained by a review of those heats that display a difference of more than 2000 pounds between the calculated and actual strengths, is that high manganese seems to increase the tenacity. The figure 2000 pounds is chosen arbitrarily, since this seems a sufficiently close approxi- mation to attain 'by any formula. Fourth : The influence of manganese may be investigated by put- ting together the groups that show a similar content of this element. Thus there are twenty-nine groups that hold from .30 to .39 per 510 METALLURGY OF IRON AND STEEL. TABLE XVII-TJ. List of Groups of Acid Open-Hearth Steel of Old and New Series, Used in Determining the effect of Certain Elements upon the Tensile Strength of Steel, Together with the Formula Obtained Therefrom by the Method of Least Squares. NOTE. All figures relating to ultimate strength are expressed in pounds per square inch. Formula; ; the unit for carbon and phosphorus being .001 per c being expressed in pounds per square inch. 88600+121 Carbon -f 89 Phosphorus = Ultimate Strength. cent., and the result Composition; percent. si ti bc ti Q) J3 a I &> & 5*1* P. jj 1 O w 03 2p SJ X ~ .o-oj . g be 08 0> .^02 . 00 i 3 g .3 h U 0> 3 "c3 ^ 1-^3 to O 6 II ll e3 O | & CS "3 ft 8 X! I ft o >* fc &** .53 ^ GC PH P q 1 5 .001 .007 .29 .071 .069 54 54700 52120 2580 2 4 .073 .009 .24 .034 .050 .18 52850 51880 970 3 7 .075 .005 .27 .054 .070 .21 54880 53910 970 4 5 .076 .003 .81 .052 .080 .16 55190 54920 270 5 6 .078 .006 .81 .036 .057 .22 53020 53110 + 90 6 7 .087 .008 53 .043 .060 .15 55470 54470 1000 7 9 .090 .017 .33 .062 .088 .16 57310 57320 + 10 8 4 .091 .006 51 .024 .035 .08 52180 52730 + 550 9 5 .091 .006 .36 .027 .059 .15 54760 54860 + 100 10 7 .095 .005 .39 .036 .051 .18 54960 54630 830 11 8 .096 .011 .25 .060 .075 .22 57200 56890 810 12 11 .103 .007 .85 .077 .074 .26 57260 57650 + 890 13 35 .106 .007 .35 .052 .070 .20 57140 57660 + 520 14 18 .116 .009 .29 .055 .070 .15 58860 58870 + 10 15 15 .116 .005 .82 .051 .075 .18 60770 59310 1460 16 9 .117 .007 .89 .028 .056 .08 56950 57740 + 71)0 17 14 .117 .006 .36 .057 .090 .11 58860 60770 +1910 18 5 .119 .005 .88 .030 .041 .14 55030 56650 +1620 19 11 .120 .006 57 .080 .071 .24 58920 59440 + 520 20 12 .121 .008 .41 .026 .065 .06 58820 59030 + 210 21 11 .125 .007 .41 .032 .047 .08 56960 57910 + 950 22 15 .125 .016 .39 .080 .076 .24 61080 60490 590 Division I. Normal 23 15 .126 .005 .48 .039 .063 .26 59060 59450 + 890 acid open-hearth steels. New series. 24 25 49 11 .127 .128 .013 .006 59 59 .055 .057 .075 .095 .14 .13 60850 61110 60640 62540 210 + 1430 26 16 .130 .009 .40 .042 .053 .10 59170 59050 120 27 69 .131 .007 .38 .051 .072 .18 58940 60860 +1920 28 14 .132 .009 .47 .053 .076 .10 61260 61340 + 80 29 12 .134 .009 .44 .042 .058 .05 57080 59980 +2900 30 14 .140 .007 .41 .074 .087 .20 62910 63280 + 370 81 10 .143 .006 59 .065 .099 .14 62830 64710 +1880 82 11 .148 .006 f .49 .051 .071 .13 62860 62830 80 83 16 .151 .003 .40 .045 .060 21 60920 62210 + 1290 54 12 .151 .006 53 .055 .084 J8 63030 64350 +1820 35 12 .152 .007 .41 .029 .072 .13 62910 63400 + 490 36 16 .155 .012 .41 .034 .069 .09 60940 63500 +2560 37 40 .155 .003 .41 .051 .073 .19 62910 88850 + 940 88 12 .159 .011 .40 .045 .056 .17 02930 62820 110 89 8 .166 .007 .42 .048 .094 .18 65430 67050 + 1620 40 8 .170 .011 .46 .035 .070 .11 64640 65400 + 760 41 18 .170 .012 .43 .046 .074 .17 64840 65760 + 920 42 8 .171 .007 .41 .065 .078 .24 65260 66230 + 970 43 7 .180 .004 .53 .044 .072 .12 65320 66790 +1470 n'\> < t ;> 46 7 .207 .009 .41 .047 .088 .12 69410 71480 +2070 47 5 .207 .011 .57 .042 .066 .14 69950 69520 480 INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 511 TABLE XVII-U. Continued. Composition ; per cent. * di fl| 2 *! * g-d III P. 00 d O' en 'lo S| III s c3 >v3 OP C ^* ^ ,3 t fl-) fl +3 rt g W -O t* | ?* % 3 ^3'g > fcJD to 23 o3<2 5 P^ on & O 02 s i Pn D Q 48 5 .214 .013 .47 .068 .077 .18 69700 71350 +1650 49 7 .218 .009 .43 .049 .070 .17 68410 71210 +2800 50 7 .224 .008 .37 .045 .079 .09 69440 72730 +3290 Division I. Contin'd. 51 5 .229 .011 .50 .032 .065 .07 70810 72090 +1280 Normal acid open- 52 5 .244 .008 .46 .038 .044 .15 70360 72040 +1680 hearth steels. New 53 12 .330 .035 .52 .029 .039 .05 82810 82000 810 series. 54 4 .331 .018 .51 .032 .039 .03 78610 82120 +3510 55 11 .406 .060 .54 .030 .035 .07 86990 90840 +3850 56 4 .424 .060 .57 .031 .043 .05 94470 93730 740 57 6 .082 .006 .29 .034 .034 .12 52090 51550 540 58 12 .105 .009 .38 .059 .074 .18 57380 57890 + 510 59 11 .109 .008 .31 .036 .066 .14 57310 57660 + 850 60 12 .109 .007 .38 .048 .082 .15 57430 59090 + 1660 61 38 .113 .009 .43 .038 .061 .13 57140 57700 + 560 62 11 .113 .007 .48 .046 .096 .18 62870 60820 2050 63 5 .115 .007 .49 .029 .037 .09 55450 55810 + 360 64 18 .115 .013 .80 .043 .069 .17 58780 58660 120 65 12 .116 .005 .59 .025 .034 .10 56830 55660 -1170 66 19 .116 .015 .50 .069 .082 .19 60870 59930 940 67 9 .116 .013 .47 .057 .089 .17 62610 60560 2050 68 18 .117 .018 .33 .039 .073 .20 61190 59250 1940 69 17 .117 .005 .45 .049 .099 .16 61430 61570 + 140 70 19 .118 .005 .59 .030 .035 .10 56990 55990 1000 71 72 .118 .007 .42 .045 .075 .14 59110 59550 + 440 72 13 .118 .008 .56 .044 .063 .14 59350 58490 860 73 15 .118 .007 .45 .064 .081 .17 59260 60090 + 830 74 15 .118 .014 .57 .056 .076 .18 60900 59640 1260 75 21 .119 .009 .42 .051 .090 .14 59310 61010 +1700 76 15 .119 .017 .43 .028 .065 .16 61020 58780 2240 77 96 .119 .009 .44 .043 .077 .16 61130 59850 1280 78 19 .123 .014 .44 .030 .063 .16 59110 59090 20 79 6 .129 .008 ,49 .050 .118 .16 65020 64710 310 80 11 .131 .012 .47 .033 .051 .13 60690 58990 1700 Division II. Normal 81 13 .134 .015 .48 .035 .045 .15 58820 8320 acid open-hearth steels. Old series. 82 83 12 38 .138 .140 .021 .016 .36 .48 .041 .042 .077 .077 .14 .18 62940 62890 62150 62390 ' ' 790 500 84 10 .143 .006 .39 .045 .086 .20 64880 63560 -1320 85 10 .147 .012 .54 .024- .056 .16 63210 61370 -1840 86 12 .151 .012 .64 .033 .051 .13 62650 61410 1240 87 7 .151 .005 .49 .055 .083 .16 64950 64700 250 88 12 .156 .003 .57 .085 .070 .17 65180 63710 1470 89 8 .171 .011 .63 .026 .036 .10 C2850 62500 850 90 4 .178 .008 1.00 .043 .076 .14 71930 66900 5030 91 8 .183 .014 .68 .030 .027 .10 65100 63150 1950 02 9 .185 .008 .76 .028 .038 .13 65590 64870 1220 93 6 .193 .009 .67 .020 .030 .10 65280 C5160 120 5 .198 .013 .61 .032 .000 .14 69970 67900 2070 C5 8 .207 .012 .79 .045 .OC7 .15 71210 C9C10 1600 CG 8 .212 .010 .82 .039 .073 .14 71870 70750 1120 4 .213 .012 .70 .019 .046 .14 60750 68470 1280 98 5 .225 .015 .99 .048 .077 .22 78700 72680 6020 99 5 .235 .016 .75 .027 .037 .14 71170 70330 840 100 12 .240 .009 .76 .030 .054 .14 72320 72450 + 130 101 7 .242 .010 .86 049 .070 .19 78020 74650 3870 102 6 .282 .009 .66 .033 .053 .16 76830 77440 + 610 100 6 .282 .010 .77 .023 .043 -14 76940 76550 890 104 7 .306 .010 .79 .034 .050 .09 82680 80080 2600 105 11 .341 .020 .85 .034 .045 .11 86980 83870 3110 100 8 .374 .030 .83 .035 .057 .12 ! 90750 88930 1820 512 METALLURGY OF IRON AND STEEL. TABLE XVIT-TJ. Continued. Composition; percent. d 3 -6 3 * < 3 38 "S> ft a a >J | II II ||| ^ o ^3 o> t ^5 o> 3 d O CS^^ be 2 -3 C-0 8 he 3 & 1 ft 1 1 p, ff 111 Hit bo S 3 ^H ^ 3 o ^M ^H o5 C< -< o3^t OP fc 5J % CO cu P |Q 107 14 .253 .150 .55 .026 .033 .03 73710 72150 1560 108 7 .816 .200 .65 .025 .033 .06 82-240 79770 2470 Division III- High- silicon acid open- hearth steels. New 109 110 111 112 7 7 9 9 .342 .366 .392 .408 .190 .170 .210 .230 .61 .60 .63 .70 .020 .022 .022 .021 .029 .028 .029 .029 .04 .04 .06 .04 87860 91580 98180 102430 82560 85380 88610 90550 5300 6200 9570 11880 series. 113 9 .461 .230 .64 .021 .029 .05 106560 96960 9600 114 7 .470 .230 .65 .021 .031 .07 111830 98230 136CO 115 7 .5555 .200 .72 .022 .030 .07 120590 106010 14580 116 7 .333 .220 .65 .026 .041 .08 87410 82540! 4870 117 14 .81)0 .220 .68 .023 .034 .08 88820 3810 Division IV. High- silicon acid open- 118 119 17 16 .428 J88 .220 .220 .65 .69 .023 .026 .036 .033 .08 .13 97270 102900 93590 94540 3680 8360 hearth steels. Old 120 14 .477 .240 .69 .025 .030 .08 107300 98991* 8310 series. 121 20 .480 .230 .69 .022 .032 .06 111740 99530 12210 122 10 .527 .250 .72 .027 .032 .07 116980 105220 11760 123 10 .554 .230 .68 .022 .032 .09 122950 108480 14470 Div. V. High-manga- nese and high-phos- 124 13 .480 .090 1.12 .044 .106 .19 121210 106110 15100 phorus acid open- hearth steels. Old 125 126 18 9 .507 .555 .061 .090 1.19 1.18 .047 .042 .110 .109 .IS .19 126800 123620 109740 115460 17060 8160 series. cent. Nineteen of these have a plus error with a total of 21,650 pounds, while ten groups have a minus error with a total of 8,030 pounds. The difference between these totals, or rather their alge- braic sum, is + 13,620, which, divided by twenty-nine, gives the average error for one group. Table XVII- V has been constructed by this process of differential synthesis for each increment of man- ganese, Group No. 54 being omitted for reasons given above. TABLE XVII-V. Average Error of Groups in Table XVII-U Arranged According to their Manganese Content. Manganese; ] per cent. No of Total Total Net Average Limits. Average. heats minus error. plus error error. error .20 to .29 .80 to .39 .40 to .49 .50 to .59 .60 to .69 .70 to .79 .80 to .89 .90 to 1.00 .27 .36 .44 .55 .65 .76 .84 1.00 6 29 88 18 6 7 4 2 5370 8030 10660 10520 5730 7930 9420 11050 + 10 +21650 +27600 + 6600 + 610 + 130 5360 + 13620 +16940 3920 5120 7800 9420 11050 -898 + 470 -f 446 801 858 1114 -2355 -5525 INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 513 It should be noted that most of the groups that contain high manganese are in Division II, and it has been remarked that there is some occult cause why the actual strengths of this division are slightly above the formula. The error arising from this condition is not sufficient to invalidate the records, but when most of the members of the division are slightly above the formula it is not extraordinary if the high-manganese groups follow the rule. Passing over this reasoning, the table teaches that in steels containing about .25 per cent, of manganese, the actual ultimate strength is 893 pounds greater than would be indicated by the for- mula. With an increase in the content to .36 per cent, the actual strength is 470 pounds less than the formula, and with .44 per cent/, it is 446 pounds less. From this point an increase in manganese gives an increase in strength. It is important to notice that the figures + 446 for .44 per cent, of manganese, and-j-470 for .36 per cent., are consistent with the minus values for the higher percentages, since manganese was en- tirely omitted in the derivation of the formula, and the result may therefore be looked upon as strictly applicable only to the average content of an element which it neglects; and if such an element does have an effect, it should make itself evident in a plus error on one side of the average and a minus error on the other. This reasoning, however, is inconsistent with the fact that man- ganese did not make a good working factor in the method of least squares. This inconsistency is explained by the values obtained in the first three lines of Table XVII-V. With a content of .27 per cent., the actual strength is more than the calculated, and this is directly opposed to the law of higher strength with higher manganese. Moreover, the figure for .36 per cent, is practically the same as that for .44 per cent., being + 470 in one case and + 446 in the other. Considering the fact that three-quarters of all groups were below .50 per cent, in manganese, and that the results on such metal were indecisive, it is not strange that manganese did not form a proper determinative member of the equations. It is indicated, therefore, that with less than .50 per cent, of manganese the effect is not so uniformly marked as with larger proportions. Whether this is due to the different physical or mole- cular condition of soft metal, or to the presence of oxide of iron, or whether it arises from abnormality of^the steels, or determinative errors in the records, cannot be satisfactorily demonstrated. 514 METALLURGY OF IRON AND STEEL. The results, as a whole, justify the use of a formula for normal acid steels without any factor representing manganese. With con- tents above .60 per cent, of this element, it is necessary to make allowance for an increased strength, while above .80 per cent, the tenacity will rapidly increase. It may also be necessary to allow for a very low content of man- ganese, since it was found in Table XVII-V that when there is less than .30 per cent, the actual strength was 893 pounds more than was indicated by the formula. This fact will be considered in Section XVIIv in connection with other information from the basic steels. Fifth : The high-silicon steels all show a much greater strength than is given by the formula. The natural inference would be that silicon strengthens steel, but it is necessary to notice a curious and important fact, viz., that the differences between the calculated and actual strength vary in proportion to the content of carbon, and not in proportion to the content of silicon. In the new series, as given in Division III, the lowest carbon is .253 per cent., and the error is 1560 pounds. As the carbon increases to .316 per cent., the error rises to 2470 pounds, and with .342 per cent, it is 5300 pounds. The old series starts at .333 per cent, as the lowest carbon, and the error is 4870 pounds, so that the two series agree perfectly at the starting* point. They also agree at their highest point, for the maximum carbon is .535 per cent, in the new series, and .554 in the old, the error being 14,580 pounds in one case and 14,470 in the other. Between these two extremes there are considerable variations, but in the main the law holds good that the error steadily rises with higher carbon. A glance at the table will show that the content of silicon is practically constant throughout both series, and hence it is mathe- matically impossible to find any constant value for this element which will account for the variations in ultimate strength. In explanation of this it may be urged that the formula by which the strength is calculated gives a wrong value to carbon. The answer to this criticism will be found in the line CC in Figure XVII-C. The most casual inspection will show that this line is very nearly parallel to the trend of the line AA. It is impossible to decide exactly what that trend is, but the line CC seems to follow the average direction as near as it can be estimated. If any criticism were to be made, it would be that the tangent of CC was too great rather than too small. Bearing in mind that INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 515 the carbon determines the tangent of these lines, and that the linear distance between them represents the effect of the other metalloids, it will be seen that the graphic delineation bears almost conclusive proof of the mathematical deductions. The general trend of the line BB in Figure XVII-C, which repre- sents the high-silicon steels, forms a decidedly greater tangent to the horizontal axis than the line AA or its counterpart CC f and it would be impossible to draw a line which would be parallel to the trend of BB, and which at the same time would be parallel to the trend of (7(7, and since it has been remarked that the tangent of CC is fully as great as it can be to fall parallel to AA, and is possibly a step beyond, it will be evident that a different law is indicated for the metals with high silicon. This law may be stated in two different ways : First: .JJiat a constant percentage of silicon exerts a greater effect with each increment of carbon. Second: That when a constant percentage of silicon is present, each increment of carbon exerts a greater influence. It will be granted that this law has an upper limit, since the ultimate strength does not increase after a certain content of carbon is attained. It also appears that there is a lower limit, for, by referring again to Figure XVII-C, it will be seen that the line BB joins AA at a point corresponding to about .25 per cent, of carbon, and it is therefore indicated that silicon has very little effect below this point, even when present in considerable propor- tions. These high-silicon groups were all composed of heats made for steel castings, and it seems possible that the different conditions of -casting temperature might exert an influence on the result. If this were true, it would also seem as if soft steel, made for castings, should show different physical properties from heats made in the ordinary way. Such does not seem to be the case, for Groups 9, 16, 20, 23, 85, 86, 89, 91, 92, 93, and 99, were composed almost entirely of casting heats, and yet conform very closely to the formula. Sixth: The influence of sulphur has not been taken into account in the formula, and accordingly an investigation was made on the steels of Divisions I and II of Table XVIT-U by the same process of synthetical differentiation that was used to discover the effect of manganese in Table XVII- V, Group No. 54 being omitted as 516 METALLUKGY OF IRON AND STEEL. before. The results are given herewith, it being evident that no law is indicated. 16 groups bet. .019 and .03 per cent, sulphur gave an average error of 485 Ibs. 80 " .03 " .04 " " " " " 260 " 27 " .04 " .05 " " " " " -188 20 " .05 " .06 " " " " " +819 ' 7 " .06 " .07 " " " " " +584 " 5 " .07 " .081 " " " " " 378 " Seventh : A similar table, which is given on the following page, shows the average error for the different percentages of phosphorus. As there seems to be no law of error, the value given to phosphorus is probably approximately true. The foregoing conclusions are summarized in Section XVIlw in connection with a similar study of basic steel. 1 group bet. .02 and .03 per cent, phosphorus gave an av. error of 1950 Ibs. _ 117 ' 159 173 11 .08 " .04 8 ' ' .04 " .05 16 ' .05 " .08 16 .06 " .07 84 .07 " .08 11 .08 " .09 7 .09 " .10 1 .11 " .12 3iO SEC. XVIIt. Values of carbon,, manganese, phosphorus , and iron in a new series of basic steels. The steels considered in Sec- tions XVIIr and XVIIs were all made by the acid process, but at the same time that they were under investigation, similar series of basic steels were being studied. The groups were formed in the same way as described in Section XVIIm, and a list of them is given in Division I of Table XVII-N, while the old series of basic steels is shown in Division II. The numbers given to the groups are continuous with those in Tables XVII-U to avoid confusion in references. The members of both series are combined to give Curve AA in Fgure XVII-D. TABLE XVII-W. Values of Carbon, Manganese, Phosphorus, and Iron, as Determined by the Method of Least Squares from the Basic Steels in Divisions I and II of Table XVII-X. Series. Influence of .01 per cent, in pounds per square inch. Carbon. Manganese. Phosphorus. Iron. New series. Division I, Old series, Division II, + 935 + 1085 -H14 + 53 +939 +941 + 3.6335 + 8.8996 INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 517 The solution of the new series by the method of least squares is given in the first column of Table XVII-W, while the second column shows, for comparison, the determinations on the old series of basic steels as given in Table XVII-X. The results indicate that manganese has a decided strengthening effect upon basic steel, although it was found that in the case of acid steel no positive relation could be proven. This conclusion is corroborated by a calculation which was made by combining the old and new series,, and solving the resultant equations by the method of least squares, without taking any account of manganese as a factor. In the case of acid steel this process gave a satisfactory formula, but in the basic steel it gave the following results : .01 per cent, of carbon=+998 pounds. .01 per cent .of phosphorus=+1444 pounds. Pure iron=39,987 pounds. This value of phosphorus is not sustained by any other evidence. Referring to Table XYII-W, it will be seen that the corresponding figure was +939 for the new series, and -f- 941 for the old series. Thinking that there might be a clerical error, the solution was repeated in reverse order, as described on page 494, but the answers were found to be mathematically correct to five places. This high value of phosphorus, when manganese is omitted, may be explained in the following way : (1) It has been shown that carbon is self -determining in every series investigated, and that it gives fairly accurate results. (2) The iron is less self -determining, but with basic metal, where some groups are so nearly pure iron, the chance for variations in this factor is less than in the case of acid steel. (3) It is evident, therefore, that if manganese is a positive factor, and if it is neglected, its effect must be forced upon some other factor by the method of least squares, and phosphorus is the only factor available. (4) The responsibility falls on phosphorus rather than on carbon, because the variations in phosphorus are very small and it is there- fore less self -determining than carbon, and less than in acid steel where it is present in large proportions. SEC. XVIIu. Values of carbon, manganese, phosphorus, and iron in basic steel, as determined from the old and the new series combined. Accepting as proven the conclusion of the foregoing section that manganese has a decided influence upon the tensile 518 METALLURGY OF IRON AND STEEL. TABLE XVII-X. List of Groups of Basic Open-Hearth Steels of Old and New Series, used in Determining the Effect of Certain Elements upon the Tensile Strength of Steel, Together with the Formula obtained therefrom by the Method of Least Squares. NOTE. All figures relating to ultimate strength are expressed in pounds per square inch. The group numbers are made continuous with those of Table XVII-U to avoid confusion in references. Formula: the unit for carbon, manganese, and phosphorus being .001 per cent., and the result being expressed in pounds per square inch. 87430+95 Carbon-f 8.5 Manganese +105 Phosphorus = Ultimate Strength. ti Composition, per cent. d 3 c*i ti P) ^riv-3 D * 2 bo 1 g II -^ o 3) OP 3w> "o ss o . o **! 2g-' 8|jl Numbei Numbei group, Carbon, combi Silicon. 1 3 CC Phosph Copper. c3 g) .5 > 00 1 h>a 1 g oa jS . 1 cfi C G? B O ft ^ &H Ill l"3s| fa a I! II e3 ^ 8 q A o 49 1 >-to III !! 5 33 s m PH ft ft II Hi Sill a CO Q jg O 33 i 3 be - l-H ^ r] o > w ~ o3 5 03 3 & ^ . 33 ^ 02 PH 8 Q 235 7 .121 .008 .54 .032 .056 .14 60580 59400 1180 236 11 .125 .012 .67 .038 .025 .13 56680 57620 + 940 237 10 .125 .019 .54 .060 .036 .11 58790 57680 1110 238 16 .126 .008 .62 .028 .024 .14 55090 57190 +2100 . 239 19 .131 .008 .30 .029 .022 .13 54690 54730 + 40 240 '20 .132 .006 .89 .027 .009 .13 54890 54230 660 241 63 .132 .010 .47 sm .028 .19 5(5870 56900 + 30 242 9 .134 .016 .51 .036 .055 .11 59110 60270 +1160 243 15 .136 .009 .81 .029 .024 .13 57010 65500 1510 244 11 .137 .020 .72 .037 .033 .18 59110 60030 + 920 245 6 .142 .017 .53 .058 .029 .12 60570 58470 2100 246 10 .144 .008 .50 .020 .026 .12 58850 58090 770 247 37 .144 .015 .52 .034 .028 .13 5S970 58470 500 248 14 .146 .015 .44 .019 .023 .11 57030 57460 + 430 249 21 .147 .005 .43 .027 .011 .10 57060 5(5200 860 250 7 .151 .016 .68 .029 .024 .18 60870 ('0080 790 251 9 .152 .008 .64 .034 .045 .17 63480 62040 1440 252 10 .153 .011 .46 .027 .012 .10 58970 57130 1840 Division II, con- 253 13 .153 .008 .53 .034 .030 .1(5 60770 59620 1150 tinued; old series. 254 12 .155 .012 .39 .029 .020 .12 59110 57570 1540 255 6 .158 .012 .82 .032 .027 .17 63400 62250 1150 256 8 .164 .018 .57 .046 .031 .16 63740 61110 2630 257 7 .173 .009 .53 .021 .02? .11 60810 60570 240 258 11 .180 .012 .56 .029 .026 .15 63110 62020 1090 259 10 .181 .006 .48 .031 .011 .10 60740 59860 880 260 8 .181 .011 .37 .028 .019 .07 60870 59770 1100 261 5 .185 .039 .72 .049 .043 .11 67570 65640 1930 262 5 .190 .008 .72 .037 .047 .17 66480 66530 + 50 263 5 .196 .025 .86 .032 .029 .17 67480 66400 1080 264 10 .199 .012 .62 .030 .025 .12 6(5820 64230 -2590 265 7 .204 .007 .45 .028 .010 .12 63600 61690 1910 266 8 .210 .010 .53 .020 .018 .13 63740 63770 + 80 267 6 .215 .005 .42 .024 .011 .16 63470 62580 890 268 6 .231 .029 .36 .025 .012 .12 67530 63700 3830 269 5 .233 .008 .49 .020 .021 .13 675(50 65940 -1620 270 5 .260 .060 .31 .025 .014 .10 68470 66230 2240 271 5 .311 .080 .44 .029 .020 .07 73010 72820 190 272 5 .338 .025 .62 .026 .017 .10 77950 76600 1350 BASIC STEELS;* DIVISIONS I AND II, TABLE XVII-X Equation from carbon: 3,503,736 A + 9,353,710 B + 423,710 C + 2,049,800,569 D = 1,230,544,020. Equation from manganese; 9,353,710. A + 29,555,000 B + 1,350,180 C + 6,301,464,560 D = 3,660,255,100. Equation from phosphorus; 423,710 A + 1,350,180 B + 74,634 C + 290,433,400 D = 169,202,400. Equation from iron; 2,049,800,569 A + 6,301,464,560 B + 290,433,410 C + 1,439,974,511,304 D = 822,329,462,810. The solution of these equations gives the following values : Lbs. per sq. in. Effect of .001 per cent, of carbon + 94.9 Effect of .001 per cent, of manganese + 8.5 Effect of .001 per cent, of phosphorus t 105.4 Strength of pure iron 37733 *The sum total of the coefficients in these equations is not quite 1,460,000,000,000, ftsit should be theoretically, because the factors in the old series relating to aili- oon, sulphur, and copper, have been omitted. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 521 strength of basic steel, the groups of both the new and the old series as given in Division I and II of Table XVII-X were com- bined, the resultant equations being given at the foot of page 519, in which A = the influence of .001 per cent, of carbon, B = the influence of .001 per cent, of manganese, C= the influence of .001 per cent, of phosphorus, and D = the influence of .001 per cent, of iron. Following the same line of argument as in acid steels, it is necessary to make allowance for the fact that there is never 100 per cent, of iron in any steel. The figure 99.2 per cent, was taken as a basis in Section XVIIs, and it will also be taken in the present case. It is true that the phosphorus is generally lower in basic steel, but, on the other hand, the carbon is usually higher. On this assumption the strength given by the iron itself in an average basic steel will be 37,430 pounds per square inch. 120.000 C urves showing rels composition of ba its ultimate streng A=normal steels, 1 C=pure iron 4 pure formula 37430+ 95 c* bscissas=carbon, p rdinates=ultimate i ition b sic op 'th. Divisio carboi etwee en-het nsl ai i, calci ultim b. ?th, It a the c irth s id II, ilated ate sti s. per 1 hemic beel ai from tl ength sq. inc al 110 000 -ioblSb- -ic^oT A C > A o ie irbon= er cen stren h. J A , JflJ ^S C 60,000 ,J M^ ^ A-" ^ JVv * 40.000 C <. .0* JO .! .2.0 .25 .90 .AS .4* A SO .35 FIG. XVII-D. CURVES SHOWING EELATION BETWEEN THE CHEMICAL COMPOSITION OF BASIC OPEN-HEARTH STEEL AND ITS ULTIMATE STRENGTH AS SHOWN IN TABLE XVII-X. 522 METALLURGY OF IRON AXD STEEL. Constructing a formula in the same way as for acid metal, we have the following, the answer being expressed in pounds per square inch. 37,430+95 Carbon+8.5 Manganese+105 Phosphorus+R=Ultimate Strength. The factor E represents an allowance for the conditions under which the piece is rolled, whether finished hot or cold. In the present series of groups it is zero. In each case the unit is .001 per cent., but since manganese is seldom determined beyond two decimal points, it will be convenient in calculation to use a unit, of .01 per cent, and a value of 85 pounds per unit, but it would be very confusing to so write the formula. In Table XVII-X this formula has been applied to the basic steels of the old and the new series, and the differences between the actual and the calculated ultimate strengths have been placed in the last column. An inspection of these differences or "errors" as they have been called, brings to light one or two points of interest. First: The difference, which was found between the two series of normal acid steels, exists also between the two series of basic products. In Division I there are fifty-six groups that give a plus error, with a total of 57,130 pounds, while there are only fourteen groups that are minus, with a total of 18,080 pounds. On the other hand, Division II offers only 24 groups having a plus error, with a total of 18,330 pounds, while it has 51 groups with a total minus error of 65,350 pounds. The net error of Division I is +39,050 pounds, and that of Division II is 17,020 pounds. The reason for this difference is unknown. Second : An investigation was made into the effect of manganese in the same way as was done for acid steel in Table XVII- V, and the results are shown in Table XVII-Y. t TABLE XVII-Y. Average Error of Groups in Table XVII-X, Arranged According to their Manganese Content. Manganese ; per cent. Number of heats. Total minus error. Total plus error. Net error. Average error. Limits. Average. .20 to .29 .30 to .39 .40 to .49 to .59 .60 to .69 .70 to .79 .80 to .89 .26 .36 .43 .53 .65 .72 .84 9 47 52 24 8 3 2 13950 20190 16850 13230 9720 1930 2230 + 690 -f 30680 + 24680 +14240 + 4200 + 970 13260 +10490 + 7830 + 1010 5520 960 2230 1473 + 228 + 151 + 42 690 820 1115 INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 523 There is no such regular progression as was shown in the former case. This is readily explained by the fact that manganese is given a value as part of the formula, and it is indicated that the value determined must be a very close approximation to the truth. In the case of the steels containing between .20 and .29 per cent, manganese, the actual strength is 1473 pounds above the calculated. This will be again referred to in Section XVIIv. Third : The influence of sulphur was investigated in the same way as manganese. The results, given herewith, agree with those found from acid steel, in showing that sulphur exerts no regular influence upon the tensile strength. 13 groups bet. .01 and .02 per cent, sulphur gave an av. error of +317 ibs. 87 " " .02 " .03 69 " " .03 " .04 9 " " .04 " .05 12 " " .05 " .06 4 " " .06 " .07 2 " " .08 " .09 598 " + 251 " 397 " + 81 " 855 " 645 " Fourth : A similar table, which is here given, shows the average error for the different percentages of phosphorus. This is done as corroborative evidence that the value of phosphorus in the formula is correct, for it may be assumed that if the value was too high or too low, the fact would be made manifest by a large error in the groups containing either high or low phosphorus. The fact that no regular relation exists seems to indicate that the deduced value is practically correct. 21 groups bet. .00 and .01 per cent, phosphorus gave an av. error of 20 Ibs. 63 " .01 " .02 89 " .02 " .03 13 " .03 " .04 7 " .04 " .05 3 " .05 " .06 56 " 168 " + 261 " 234 " + 263 " SEC. XVIIv. Meaning of the term "pure iron!' In the fore- going investigation, a slightly different value was found for "pure iron" as derived from acid steels, and "pure iron" as derived from basic metal. This contradiction is solely a matter of words. Abso- lutely pure iron never has been, and, in all probability never will be made. The steels given in Table XVII-M are about as near to pure iron as can be found. Heat No. 4932 in that table contains .011 per cent, of phosphorus, .04 per cent, of manganese, .029 per cent, of sulphur, and .04 per cent, of copper. The carbon was not deter- mined by combustion, but it must have been about the same as the 524 METALLURGY OF IRON AND STEEL. average sample of the six heats, which was .025 per cent. This would leave a total content of impurities of 00.145 per cent. If copper is omitted from the total, as having no appreciable effect, the total will be 00.105 per cent. Notwithstanding this purity, the tensile strength of this heat is 46,480 pounds, which is practically the same as the average of the group. The great strength of this metal, as compared with steel containing a larger proportion of impurity, has already been discussed in Section XVII-E, but must again be considered here. It is easy to imagine that oxide of iron is present in this de- carburized and dephosphorized product, and that it may confer an abnormal cohesive power. This supposition is corroborated by Tables XVII- V and XVlI-Y, which indicate that both acid and basic steels, when low in manganese, are somewhat stronger than would be accounted for by their content of carbon and phosphorus, and it will be acknowledged that such steel holds a considerable quantity of oxygen. It is true that these abnormal metals may contain unusual pro- portions of certain substances like hydrogen, nitrogen, or carbonic oxide, but since the effect of these constituents is entirely hypothet- ical, the most reasonable assumption is that oxide of iron increases the ultimate strength. Whether this theory is perfectly true or not is of little importance so -far as the present investigation is concerned, for the results obtained from absolutely pure iron would be utterly valueless as a guide in creating a proper formula. From one point of view there is no more real necessity of knowing the strength of pure iron than of knowing the strength of pure carbon or pure phosphorus. There may be no connection at all between the tensile strength of a carbide or phosphide of iron and the tensile strength of its separate components, since a chemical compound often has nothing in com- mon with its parents. In the foregoing pages,, therefore, the term "pure iron" is arbi- trary, and is intended to express simply the datum plane from which it is most convenient to start in order to find the strength of steel by a simple formula. SEC. XVIIw. Synopsis of the argument and conclusions in the foregoing investigations. The argument involved in the foregoing calculations is so complicated, and the conclusions are so scattered throughout the text, that it will be convenient^ to give a general INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 525 synopsis of Parts II and III of this chapter. As far as the conclusions are concerned, it is conceded that no one series of experi- ments can make a valid foundation for universal generalizations, but it has been deemed proper to put the discovered relations into the form of statements, which are to be accepted subject to the limitations of the premises. Basis of the investigation. The object of the investigation was to discover the influence upon the tensile strength of open-hearth steel, of the metalloids that are usually found therein. Both acid and basic metals were investigated, but the two kinds were kept separate throughout the work. The preliminary tests of several hundred heats of each kind of steel were at hand, with a record of the ultimate strength of each, together with the content of sulphur, phosphorus, and manganese. These .tests were made into several divisions on the basis of their ultimate strength, and these divisions were again subdivided so as to produce groups that would show high and low phosphorus, high and low sulphur, and high and low manganese. These groups were analyzed by taking an equal quantity of drillings from each bar, and determining the carbon by combustion, and also the silicon and the copper. The iron was calculated by difference. Each one of these groups was then considered as a unit, and an equation was constructed from its chemical composition. On one side of the equation were the carbon, silicon, manganese, sulphur, phosphorus, copper and iron, and on the other side was the ultimate strength. The coefficients of the factors were the percentages found by analysis, while the factors themselves were the unknown quanti- ties whose values were to be sought. Mathematical calculations. The only method which seemed to meet the case was the method of least squares, but the first applica- tion of this very complicated and laborious mathematical agent gave results which were palpably incorrect. It was demonstrated that the error arose from using silicon, sulphur, and copper as fac- tors in the equations, when, as a matter of fact, they exerted no controlling influence. Neglecting these elements, a solution was made by which values were found for carbon, manganese, phosphorus, and iron. Differ- ences existed between the basic and acid steels in the values of all these factors, but the most striking variation was in manganese, it 526 METALLURGY OF IRON AND STEEL. being found that it was a minus quantity in acid, and a plus quan- tity in basic steel. After completing these calculations, the same line of work was repeated on an entirely new series of acid and basic steels. The results corroborated the former records in most respects, but the value of manganese was found to be very nearly zero in the case of the acid steel. Certain computations showed that this element gave very discordant results when the acid steels were separated into two arbitrary divisions, while the figures for the other metalloids pre- served their general character, and the conclusions were drawn that manganese was an unsatisfactory factor in acid metal, that its effect upon the tensile strength was very small when present in ordinary proportions, and that a working formula could be constructed with- out it. Finally the old and the new series of steels were put together and a solution was made of the combined list to find the most probable values of the metalloids. Manganese was neglected in the case of the acid steel, but it was found to have a decided influence upon basic metal. From the values so determined, formulae were constructed, and these were applied in Tables XVII-U and XVII-X to the groups from which they were derived. Against each group is placed the strength as given by the formula, as well as the difference between this figure and the actual strength. This column of differences was then analyzed in the case of both acid and basic steels, and it was sought to find whether there was any law of error ; for instance, whether high-sulphur groups would always give a plus difference and low sulphur groups would always give a minus difference, thus indicating that the formula did not fit the facts, and that the values were not correct. From this series of steps the following conclusions were drawn : Conclusions. (1) The strength of pure iron, as far as it can be determined from the strength of steel, is about 38,000 or 39,000 pounds per square inch. (2) An increase of .01 per cent, of carbon raises the tensile strength of acid steel about 1210 pounds per square inch, and of basic steel about 950 pounds. This difference between the effect of carbon upon acid and basic steels, as found by mathematical analysis, is corroborated by the graphic records in Figures XVII-C and XVII-D. INFLUENCE OF CERTAIN ELEMENTS ON STEEL. 527 (3) An increase ^ of .01 per cent, of manganese has very little effect upon acid steel unless the content exceeds .60 per cent., but it raises the strength of basic steel about 85 pounds per square inch. (4) An increase of .01 per cent, of phosphorus raises the tensile strength of acid steel about 890 pounds per square inch, and of basic steel about 1050 pounds. (5) The following formulae will give the ultimate strength of ordinary open-hearth steel in pounds per square inch, the carbon, manganese, and phosphorus being expressed in units of .001 per cent., and a value being assigned to E in accordance with the con- ditions of rolling and the thickness of the piece. FOKMULA FOB ACID STEEL. 38,600+121 Carbon+89 Phosphorus+R=Ultimate Strength. FORMULA FOB BASIC STEEL. 37,430+95 Carbon+8.5 Manganese+105 Phosphorus+R=Ultimate Strength. (6) The metals, from which these data were derived, were ordi- nary structural steels ranging from .02 to .35 per cent, of carbon, and it is not expected that the formulae are applicable to higher steels. or to special alloys. (7) A considerable difference may be found between steels which apparently are of the same composition, and which, as far as known, have been made under the same conditions. (8) In the case of acid steel, an increase in manganese above .60 per cent, will raise the tensile strength above the amount indicated by the formula, the increment being quite marked when a content of .80 per cent, is exceeded. (9) In steels containing from .30 to .50 per cent, of carbon, the value of the metalloids is fully as great as with lower steels, while the presence of silicon in such metal in proportions greater than .15 per cent, seems to enhance the strengthening effect of carbon. (10) In steels containing less than .25 per cent, of carbon the effect of small proportions of silicon upon the ultimate strength is inappreciable. (11) Sulphur, in ordinary proportions, exerts no appreciable influence upon the tensile strength. (12) Both acid and basic steels containing less than .30 per cent, of manganese give an actual strength greater than is shown by the formula, and when this is taken in connection with the abnormal strength of the unusually pure metal shown in Group 198 of Table 528 METALLURGY OF IRON AND STEEL. XVII-X, it is indicated that oxide of iron raises the ultimate strength. NOTE. Several years have elapsed since the foregoing formulae were deduced. During that time every open-hearth heat made at Steelton has been calculated according to formula and almost every one, acid and basic, has come out within 2500 pounds of the actual strength as determined by the breaking test, except the steels con- taining manganese in excess of .60 per cent. Usually the calcu- lated strength is within 1000 or 1500 pounds of the actual. Our experience proves that the formula represents something and it is used as a check and as a guide in the practical and commercial disposition of hundreds of thousands of tons of steel. The exception noted in the case of high-manganese contents is in exact accordance with conclusion No. 8, just given. It was not possible for the mathematical method to give a correct answer for this kind of metal, because such steels were not represented in sufficient proportion in the groups taken and because no simple formula of this kind so determined, could express a varying func- tion. I have more than once met with the objection that these formulae do not allow for the variations in thickness and finishing tem- perature. This criticism is unfounded. The only way in which these can be allowed for is by adding a certain sum for thin pieces and cold finishing, or subtracting for heavy pieces and hot finishing. This was explained in the conclusions given and the factor R covers this ground, as it may be either plus or minus. Mention is made of this because it does not seem to have been made sufficiently prom- inent. CHAPTEE XVIII. CLASSIFICATION OF STRUCTURAL STEELS. SECTION XVIIIa. Influence of the method of manufacture on the properties of steel. The first problem in the writing of speci- fications for structural steel is the advisability of prescribing the method by which it shall be manufactured. Some engineers, with commendable fairness, hold that the way in which a bar or plate is made is a matter entirely beyond their dominion. Logically this position is impregnable, but it is not so practically, for although there is no essential difference in the results obtained from open- hearth and Bessemer steel in the ordinary testing machine, there is good testimony to show that the product of the converter is an inferior metal which gives way in a treacherous manner under shock. It is granted that in a strict sense there is no such thing as treachery or mystery, but these are convenient terms to cover an undiscovered law. The evidence concerning the unreliability of Bessemer steel is made up for the most part of scattered individual opinions, many of which have been made on insufficient evidence, but they are too numerous to be entirely ignored, and they are fortified by the carefully considered statements of men whose words are weighed, and who are absolutely disinterested in their decisions. Thus A. E. Hunt, whose long experience as the chief of The Pittsburg Testing Laboratory gives much force to his opinion, wrote as follows:* "Xumerous cases have come under our obser- vation of angles and plates which broke off short in punching, but although makers of Bessemer steel claim that this is just as likely to occur in open-hearth metal, we have as yet never seen an instance of failure of this kind in open-hearth steel." Mr. Hunt also quotes (loc. cit.) from a paper by Wailes before the British Association to the effect that "these mysterious failures * The Inspection of Materials of Construction in the United States. Journal I-. and 8. I., Vol. II, 1890, p. 316. 529 530 METALLURGY OF IRON AND STEEL. occur in steel of one class, viz., soft steel made by the Bessemer process." There is also the testimony of W. H. White, Director of Naval Construction, Royal Navy.* "With converter steel riveted samples have given less average strength, greater variations in strength, and much more irregularity in modes of fracture than similar samples of open-hearth steel." My own experience leads me to think that Bessemer steel requires more work for the attainment of a proper structure than open- hearth metal, so that a thick bar is more apt to have a coarse crys- talline fracture. This may be ascribed in any particular case to improper heat treatment, but if it is true that open-hearth metal would not be injured under a similar exposure, then it is proven that there is a difference between the metals, and, if this be acknowl- edged, then there is no necessity for further argument. It is true that Bessemer metal has been used for rails, and that these are exposed to great stress and shock, but it also true that a large number of rails break in service, and that the use of ordinary rail steel for bridges was long ago given up as dangerous. More- over, it is quite probable that the number of broken rails would be considerably reduced if they were made of open-hearth steel. The question therefore arises why rails are not made of this ma- terial, and railroad engineers occasionally come forward with in- quiries to that end. It may be well to say therefore that the making of open-hearth rails is purely a commercial question," but it involves immense sums of money. All rails made to-day in America are made by the Bessemer process, and each rail-making plant must be regarded as a unit. The converting department is one factor of this unit, its capacity and whole scheme of operation being de- signed for the one purpose of supplying the blooming mill with just the right quantity of ingots of just exactly the right size. It may be that at a given rail-making works there is no open-hearth furnace plant at all. In such a case if open-hearth rails are wanted they can be made only by some such changes as the following : (1) Bring cold blooms from other works, entailing much expense and the erection of a mammoth plant of bloom heating furnaces. (2) Bring cold ingots from other works, with the same necessity for heating furnace equipment. In both cases the extra fuel con- sumption and waste in heating would be very serious matters. * Experiments with Basic STCCIP. Jcimial 1. ann N. /.. >'ol. I, 1892, p. 35^ CLASSIFICATION OF STRUCTURAL STEELS. 531 (3) The foregoing propositions are merely temporary on their face and the only true solution is an open-hearth plant. This calls for a very large amount of capital, and when the plant gets into operation the Bessemer plant will become a scrap heap of no value whatever, for in order that it shall be of any value it must run, and in order that it may run, it would be necessary to build a complete plant of rolling mills to handle its product, and this would seldom be desirable even if it were feasible at all in some cases. (4) Having written off the value of the Bessemer outfit as a dead loss, it is necessary to guarantee business to the open-hearth department in sufficient quantity to keep it in steady operation at a price in proportion to the increased cost. It is out of the question to operate the open-hearth plant on certain orders for open-hearth rails at a slightly higher price, and then start up the Bessemer plant on other orders and let the open-hearth lie idle. Such a proposition is clearly out of the argument. (5) It may seem possible to have a number of mills and have the open-hearth and Bessemer plants both operating continuously and distributing their product as orders demand. One or two works in the country are able to do this to a greater or less extent, but it is impossible to do it and maintain the proper coordination of dependent factors and keep the operating costs in each department at a minimum. We may conclude therefore that small lots of open-hearth rails may be made, but their production on a large scale means a plant laid out with that end in view, and if this plant is not guaranteed a regular line of business extending over many years at an increased price, it will be a losing venture. Such an innovation is hardly justified by the present knowledge of the rail business. Within the last few years it has been clearly shown that a great improvement may be made by certain modes of heat treatment. Much care is now taken to finish the rails colder than formerly and to do a great deal of work upon them while they are at a moderately low heat. By so doing a much better grain is attained, and this renders pos- sible the use of a higher content of carbon than was formerly thought advisable. This question of finishing temperature and all the associated problems of wear and toughness are being thoroughly threshed 'out, and it may be well to await the results of experiments liow under way before starting out into untried fields. In the case of structural shapes there is no difficulty in obtaining 532 METALLURGY OF IRON AND STEEL. at moderate cost all needed sections in open-hearth steel, and it would seem to be the safer way to prescribe that it shall be used in all structures, like railroad bridges, where the metal is under constant shock, and where life and death are in the balance. In this connection it should be stated that the method by which the steel is made cannot be discovered by ordinary chemical analysis. Certain experiments indicate that there is a difference between Bessemer and open-hearth steel in the character of the occluded gases, but this system of analysis is never resorted to in practice, and no provision is made for it in laboratories. Moreover, it is doubtful if any expert would risk his reputation by asserting posi- tively, from any such evidence, that a certain steel was made by either one or the other process. Consequently, when open-hearth metal is specified, a careful watch should be kept in the steel works that there is no substitution of the inferior material. SEC. XVIIIb. Chemical specifications. Another point concern- ing which there is room for discussion is the propriety of limiting the chemical composition. Some engineers contend that as long as the physical tests are fulfilled, the making of the metal is an entirely foreign matter. This position is untenable, for it would be possible to make a steel with 0.25 per cent, of phosphorus which would satisfy the ordinary tests of strength and ductility, and although such a content could usually be detected in the shops, a considerable proportion of the bars might be able to pass muster. It is impossible to fix a limit of phosphorus below which there is no danger of treacherous breakage, but it is quite certain that, as the content is reduced, the danger of such disaster disappears. On this account it becomes not only the province but the duty of the engineer to specify the chemical composition of the metal that he buys. In the construction of ordinary roof-trusses and similar work there is no necessity for stringency, and Bessemer steel with a maximum content of .10 per cent, of phosphorus may be allowed; but in railroad bridges^ traveling cranes, and other structures where the steel is exposed to moving loads and continued shock, and where the consequence of failure may not be measured in money, the speci- fications should require the use of open-hearth steel with a maximum phosphorus of .06 per cent. The common limit at the present day is .08 per cent., but the time has come for another step in advance, CLASSIFICATION OF STRUCTUEAL STEELS. 533 since the difference in the cost of the purer metal has been reduced to an unimportant figure. In addition to thus limiting the chemical content of phos- phorus, it is necessary to specify the manner in which the sample shall be taken for analysis. There are four methods of doing this of which only one is correct, and this correct one is seldom or never used. Taking for illustration a rolled billet of steel three inches square, its cross-section may be mentally divided into nine equal squares, each having an area of one square inch. Eight of these squares are next to the surface, while only one is in the interior. This central square will include almost all the segregated portion of the mass. Ordinarily a sample of such a billet would be taken by drilling to a depth of half an inch, but it is evident that this does not take cognizance of the interior core, and that the chemical deter- minations on the drillings will show too low a content of certain segregrating metalloids. Another method is to drill all the way to the center, and to cake all the drillings that are made. Two-thirds of these drillings will come from the outside square and one-third from the inside, or a ratio of two from the outside and one from the interior, while the true ratio is eight to one; hence the content of segregating metalloids found by this method is higher than the true average. A third method which is sometimes used, although manifestly inaccurate, is to take only those drillings that come from the central portion, but this will give a very much higher content of certain elements than will be found throughout the bar. The fourth way is to plane the entire surface and thus get a true average, but, as before stated, this practice is seldom carried out. In the case of angles, a very fair sample can be obtained by drilling into the bar as far as the center, the results so obtained being only slightly higher than the true values. In plates it is much more difficult to take a fair sample, since the segregated portion is in the body of the sheet, and it is usually impracticable to drill a hole without injuring the strength of the member. It is easy to see that great injustice may be done by insist- ing on unusual methods of sampling. It would be perfectly right to state in the contract that drillings were to be taken from the center of the plate, but it is not right to take them in this way 534 METALLURGY OF IRON AND STEEL. in the absence of a previous understanding. On the other hand, -the engineer has an indisputable right to investigate the homo- geneity of any plate, and to reject those members that show exces- -sive segregation. It is necessary, therefore, to take some account of these variations, and in the following specification it is provided that when drillings :are taken from the center of plates, the allowable maximum of phos- phorus and sulphur shall be raised 25 per cent. ; e. g., from .04 to .05 per cent., or .08 to .10 per cent. The engineer who has been calling for steel containing less than ..08 per cent, of phosphorus, may deem it a step backward to allow the center of plates to contain .10 per cent., but it is necessary to consider that the new provision is merely a formal recognition of a fact, and that the higher phosphorus has always existed in the center of plates, particularly if they have been rolled directly from ordinary plate ingots which have not undergone a preliminary "roughing" and "cropping." It is also well to consider that less careful engineers, who have specified a maximum of .10 per cent, of phosphorus, have received many a plate that contained .12 per cent., and even .15 per cent., of this impurity. The fact of non- homogeneity in plates is a strong argument in favor of the further lowering of the allowable maximum, for, when all other conditions are the same, each decrease in the average content diminishes the increment due to segregation. Usually it is specified that basic metal shall show a still lower phosphorus. There does not seem to be any proof that basic open- hearth steel of a given composition is more unreliable than acid metal of the same character, but in order to meet any possible danger, and because the cost of a little extra purification is not excessive, it is not amiss to require that the best basic steel shall not show over .04 per cent, of phosphorus. The other elements need not be rigidly limited, for many com- binations are possible, and some discretion should be left to the maker in the attainment of definite results. It is not uncommon for engineers of limited knowledge to write specifications that give an upper limit for every element, and require a tensile strength which cannot be obtained by the formula. The carbon should always be left open, so that if the maker wishes to reduce the phosphorus he may use carbon to get strength. Manganese may be limited to .60 per cent, on the steels under CLASSIFICATION OF STRUCTURAL STEELS. 535 64,000 pounds per square inch, and to .80 per cent, on harder metal. This will ensure a safe material, and will not be a burden on the manufacturer. Silicon is of little importance, but the maximum /nay be placed at .04 per cent, for soft steel, this proportion being seldom, if ever, exceeded. Sulphur, in most cases, concerns the manufacturer more than the engineer, for if it is too high the bar will crack in rolling and be imperfect, while it seems to have" no marked effect on the ductility of the finished piece. In material for eye-bars, however, there is danger that high sulphur may cause coarse crystallization during the heating necessary to form the eye. Copper may be entirely neglected, for no ill effect upon the cold properties of low steel has ever been traced to its action, while thousands of tons of excellent metal have been made with a content of .75 per cent. Eivet steel, like eye-bar flats, stands on an entirely different footing from other structural metal, for this must be heated and worked after leaving the place of manufacture. Only the very best of material should be used, and it should be so soft that it will not be injured by cold working or crystallized by overheating. The phosphorus should not be over .04 per cent., the sulphur not over .05 per cent., and the tensile strength not over 56,000 pounds per square inch. These limits should be insisted upon whether acid or basic open-hearth metal be used. SEC. XVIIIc. Use of soft steel in structural work. It is not the intention of this chapter to arbitrarily state just what should or should not be given as the best tensile strength for every purpose, but it is my opinion that a softer metal should be used for bridges than is often employed, because, although a slight sacrifice is made in the ultimate strength, there is a gain in the working strength due to the higher elastic ratio, and a decided increase in toughness and resistance to shock, so that the calculations may be made on the same basis for the working load as with a harder metal. The fact that the elastic ratio rises as the ultimate strength decreases is not generally recognized, but will be shown in Table XVIII-A. This is constructed by comparing the groups of angles in Table XIV-H, which are made by the same process, and are of the same thickness, and which contain the same percentage of phosphorus. It will be found that in every case the stronger steel gives a lower elastic ratio. 536 METALLURGY OF IRON AND STEEL. TABLE XVIII-A. Else in Elastic Eatio with Decrease in Ultimate Strength. Com- parison of 'the Angles Given in Table XI V-H which are Made by the Same Process, of the Same Thickness, and with the Same Content of Phosphorus. Kind of steel. Content of phos- phorus; percent. Thickness of angle, in inches. Harder steels. Softer steels. Rise in elastic ratio in softer steels; percent. 2 S-g %* A H&Sg ~ fl C!c3 >P,OQ <1 Average elastic limit; pounds per square inch. Average elastic ratio; per cent. Av. ultimate strength; pounds per square inch. o vi lit *ft Iff, *-! fl ** O g.S2fl $-~ Average elastic ratio; per cent. Basic O. H. below .04 & to | <* 1 t A to | ii to I 58865 58538 59235 59125 39692 37827 37487 36035 67.43 64.62 63.28 60.95 52533 53171 51903 51923 36284 34891 84026 32356 69.07 65.62 65.56 62.31 1.64 1.00 2.28 1.36 Acid O. H. .05 to .07 SSI 65656 65631 43713 42191 66.58 64.28 60845 60695 40891 39415 67.21 64.94 0.63 0.66 Acid 0. H. .07 to .10 tlol 66365 65777 44486 42817 67.03 65.09 60061 605S3 41143 40170 68.50 66.30 1.47 1.21 Acid Bess, .07 to .10 tlol 66277 65940 46422 45280 70.04 68.66 60659 59882 43417 42518 71.58 7100 1.54 2.34 The tendency in the first epoch of steel structures was toward a hard alloy, but later practice has been a continual progress toward toughness. There was a halt in this movement at a tensile strength of 60,000 pounds, not entirely on account , of any magic virtue in the figure, but because the ordinary mild steels gave that result, and a much higher price was charged for a softer metal. The conditions to-day are somewhat different, for the reduced cost of low-phosphorus pig-iron, and the introduction of the basic hearth, have altered the economic situation. A steel with a tensile strength of 50,000 to 58,000 pounds per square inch is a most attractive material, possessing all the good characteristics of wr ought-iron with greater strength and toughness. With this recommendation for the adoption of softer metal, cer- tain classes are proposed from which the engineer can choose. In some cases the option is given between acid and basic open-hearth steel, but it must not 'be forgotten that it costs more to make low- phosphorus metal by the acid than by the basic process, so that the terms of the specification should be enforced after the contract is awarded, out of justice to the other bidders who have based their calculations on the letter of the law. In steel above .08 per cent. CLASSIFICATION OF STKUCTURAL STEELS. 537 of phosphorus, this difference in cost disappears and there is no economy in the use of the basic hearth. The option is sometimes given between open-hearth and Bessemer metal, but it will be understood that whenever the former is specified the latter is not admissible, although as a matter of course the manufacturer may supply open-hearth in place of Bessemer, if for any reason he wishes to use the better and more expensive material. SEC. XVIIId. Tests on plates. In the specifications for plates it will be noticed that a variation of 10,000 pounds per square inch is allowed, and that concessions are made for thick and wide sections. All this may seem to some engineers to be a step backward, but in reality these provisions have been in force for many years. The engineer who writes a new specification calling for a better elonga- tion, never knows that he receives exactly the same steel that has been made before. The plate rollers have been driven to expedients which are -not dishonest, but which are dangerously near the line of deception. Thus, if it is required that a test must be cut from one plate out of every ten, the manufacturer will leave a coupon on every plate and test strips are cut from immediately next to them ; after finding which plates fill the requirements, the coupons are cut from the others and the inspector is told that the pile is ready for him. If every plate is to be tested, then a coupon is left upon each corner and a contiguous strip is privately tested by the maker. After finding which corner gives the best results, the other coupons are cut off and the plate submitted to the inspector. This is not dishonest, for any one corner represents the plate just as much as any other corner, and it would manifestly be absurd to designate from which corner the test is to be taken. It is also quite certain that no one corner represents the center of the plate, for the edges are always finished colder than the center, and it is just as certain that in a plate rolled direct from an ingot with only the usual amount of scrap, the corners in no way represent the part of the plate which corresponds to the segregated portion of the ingot. It is by care in the preliminary testing rather than by improve- ment in the quality of material that advances have been made, and it is time that the fact be made known to engineers. The mill managers have been aided by the inspectors for most of these men (to their credit be it said) are anxious to pass material which they 538 METALLURGY OF IRON AND STEEL. know to be good. They allow the manufacturer to put part of a heat into thick plates and part into thin, and make the tests OD three-eighths or one-half inch gauge ; they pass over the sheets that are 100 inches wide, and cut the coupons from plates that are less than 70 inches. These concessions have been tacitly made in the past; I have merely put them into print. On the other hand, I have called for higher tests on plates under 42 inches wide. This is because they can be made on a universal mill, and since better results can be had in this way, it is right to demand what there is a perfectly simple way of obtaining. It will be seen that no allowance is made for a variation in tensile strength for different shapes, while concessions are made for differences in thickness. This inconsistency arises from the fact that it is generally known beforehand whether a certain heat of steel is to be rolled into angles, or plates, or eye-bars, and it is seldom that it is necessary to put part of a heat into one shape and part into another. On the other hand, it is almost always necessary to roll a charge into more than one thickness and more than one size of angles, plates, etc., and it is evidently an onerous restriction if proper allowance be not made for the normal variations due to different thickness. SEC. XVIIIe. Standard size of test-pieces. In all the tensile tests a length of eight inches is taken as the standard for all sections, allowance being made for variations in shape and size. For several years there have been conferences held in foreign lands to establish uniform methods of testing, and it has been officially recommended that in the case of rounds the length of the test-piece shall be pro- portional to the square root of the sectional area, the formula being given as follows : / == 12.0V / when I = the length in inches and f = the sectional area in square inches. In Table XVIII-B I have calculated from this formula the proper length for rounds from one-half inch to 114 inches in diameter. It will be seen that the length is greatly reduced as the diameter grows less, and this, of course, is equivalent to demanding less elongation, while on larger sizes the length is increased, this being the same thing as demand- ing more elongation. It is rather difficult to compare this system, in which the elonga- tion is constant and the length varies, with the system wherein the length is constant and the required elongation varies ; but an attempt is made to do this by obtaining the proportional elongation for the CLASSIFICATION OF STRUCTURAL STEELS. 539 different lengths from Curve AA in Figure XVI-A. The results are given in the last column of the table, and it will be found that the allowances for changes in sectional area, given in the following pages, are in line with the formula just mentioned. A long time has been spent in arriving at the general adoption of an international standard length of eight inches, and it would be very unfortunate if a complicated substitute were introduced. Such a change, how- ever, is very unlikely from present indications. TABLE XVIII-B. Calculation of 12.0 V / f r Different Diameters, together with the Proportional Elongation for the Given Lengths as Deter- mined by Curve AA in Figure XVI-A. .1963 .3067 .4417 .6018 .7854 .9940 1.2271 .443 .554 .665 .775 .886 .997 1.108 ?r K 5.32 6.65 7.98 9.30 10.63 11.96 13.30 33.2 81.5 80.2 29.8 28.7 27.8 27.1 It is understood in these specifications, as well as throughout this book, that the elastic limit is determined by the drop of the beam, for this is the universal method in American steel works and rolling mills. I have no sympathy with that group of agitators who are trying to introduce new meanings to old terms, and to apply old terms to new factors. It matters not whether the drop of the beam does or does not mark the spot where the elongation ceases to be exactly proportionate to the load. It is certain that it represents a critical point of failure, and this is acknowledged by the agitators before mentioned, who recommend its determina- tion on all tensile test-pieces. Moreover, it is shown in Section XVIm that this is a definite point which can be determined more accurately than the reduction of area, and nearly as accurately as the elongation. If a new point is desired, such as is shown by an autographic device, then this 540 METALLURGY OF IRON AND STEEL. new point should be given a new name. The term "elastic limit" has been preempted by general use., as part of a system of trade nomenclature, to designate the point where the beam drops. Upon this determination all specifications and contracts are based, and any attempt to ascertain the elastic limit in any other way is a change in the contract requirements which would not be sus- tained in a court of equity. All calculations upon factors of safety in existing bridges are based upon this "drop of the beam," and there seems to be no good reason why one arbitrary point should be substituted for another and no reason why future work should not be carried on under the present established and well-understood system. SEC. XVIIIf. The quench-test. In these specifications there is nothing said about a quench-test, for I am of the opinion that it is an absurdity when applied to ordinary structural material. It was defended by Mr. Hunt* on the ground that it would guard against material that would be injured by careless heating and cooling in the mill or shops, but this suggests the query why such carelessness should be tolerated. It is assumed that the work is done by mills and shops that understand their business, and the steel should be made to fit the work in hand and not the ignorance of middle- men. It is right to make the most severe tests on the cold properties, for the derailment of a train will subject certain members- to great deformation; such an accident is always a possibility which human foresight seems powerless to avoid, but the carelessness in the shop stands on a different footing, for it is caused by posi- tive and unnecessary acts in error. Moreover, the quench-test depends very much upon slight differ- ences in the methods of heating and cooling, differences which are almost imperceptible and unexplainable, and the same steel may be made to pass or fail under modes of treatment which seem to be inherently identical. It would appear, therefore, that no warrant exists for the imposition of this test upon the material for a rail- road bridge, which is not calculated to withstand a conflagration followed by a flood. This position is being taken by a very large number of engineers, and a quench-test is rapidly becoming a thing of the past. * The Inspection of Materials of Construction in the United States. Journal I. and 8. I., Vol. II, 1890, p. 312. CLASSIFICATION OF STRUCTURAL STEELS. 541 SEC. XVIIIg. Classes of Steel Proposed. In the former edition of this work I included several general provisions concerning methods of testing. They are omitted here, not because of any change of views, but because they appear in the American Standard Specifications. I may retain my own views regarding some minor points of special metallurgy,, but there should be agreement by vote on the method of testing materials. I also recommended several classes of steel for different purposes, and gave the speci- fications which they should be called upon to meet. I have seen no reason to change any views or any figures in the tables, save that I have specified that manganese in rivet steels shall not fall below a certain minimum. I do not offer these specifications for general adoption and never considered that they would be adopted. The specifications drawn up by the American Society for Testing Materials will be given later, and these are already the recognized standard and should be used, but I offer these tables as detailed data representing what changes take place in the physical qualities as the chemical composi- tion changes and as the thickness or shape of the rolled member varies. These pages have been of use to engineers and authors in the past in obtaining such information. They do not represent random guesses, but the condensation of many experiments and much work, albeit they are too complicated for the demands of those who wish to read as they run. CLASS I. Extra Dead Soft; for Common Eivets, Wire Cables, and other Purposes where Exceptional Toughness is Required. Method of manufacture. Basic open-hearth process. Chemical composition, in per cent. P below .04 ; S below .06 ; Si below .04 ; Mn .35 to .50. Physical requirements as follows : Shape. Diameter in inches. . Ultimate strength ; pounds per square inch. Elastic ratio. Elonga- tion in 8 inches; percent. Reduc- tion of area; per cent. Minimum. Maximum. Hi vet rods, 1 1 /8 % 46000 46000 45000 45000 44000 44000 55000 54000 54000 54000 54000 54000 C4.0 C3.0 61.5 60.0 58.5 57.0 28.0 29.0 29.25 29.50 2975 30.00 52 58 56 54 52 50 A rolled round about three-quarters inch in diameter, after being nicked about- one-quarter way through, shall bend completely ,542 METALLURGY OF IRON AND STEEL. double without fracture, with the nick on the outer curve of the bend. Heats rolled into bars less than five-eighths inch in diameter may be tested in trial rods of three-quarters inch. If any bar fails to pass the physical tests, four more pieces shall be taken from the same heat, and the average of all five bars shall be considered the true record. Rivets, when cut out of the work into which they have been put, shall show a tough silky structure, with no crystalline appearance. CLASS II. Bridge Rivets; for Rivets in Railroad Bridges. Method of manufacture. Acid or basic open-hearth process. Chemical composition, in per cent. P below .04 in acid steel ; below .03 in basic ; S below .05 ; Si below .04 ; Mn .35 to .50. Physical requirements as follows : Shape. Diameter in inches. Ultimate strength; pounds per square inch. Elastic ratio. Elongation in 8 inches ; per cent. <$ O f-> S23 alS. fill >TS a o <1 Minimum. Maximum. Average. Minimum. Rivet rods, if u ii ? i 48000 48000 47000 47000 46000 46000 57000 56000 56000 56000 56000 56000 66.0 65.0 63.5 62.0 60.5 59.0 29.0 80.0 80.5 81.0 81.0 31.0 27.0 28.0 28.5 29.0 29.0 29.0 60 60 58 56 54 52 Two tons of bars from the same heat shall constitute a lot, and two specimens, each from a different bar, shall be tested from each lot. The above table gives the average required of these two- bars, and the minimum below which no bar shall fall. If the average elongation or reduction of area on any one lot shall fall below the requirement, two additional bars shall be cut from the same lot, and the average of the four pieces shall be considered the average of the lot, provided that no concession be made in the minimum. Heats rolled into sizes less than five-eighths inch may be tested in trial rods of three-quarters inch. A rolled round about three-quarters inch in diameter, after being nicked one-quarter way through, shall bend completely double without fracture, with the nick on the outer curve of the bend. A piece of three-quarter-inch rod cut one-half inch long shall be upset while cold into a disc one-eighth inch thick, without developing extensive flaws or showing signs of cold shortness. CLASSIFICATION OF STRUCTURAL STEELS. 543 Rivets, when cut out of the work into which they have been put, shall show a tough silky structure, with no crystalline appearance. CLASS III. Hard Bridge Rivets; a Substitute for Class II, Giving Greater Strength with Less Toughness. Method of manufacture. Acid or basic open-hearth process. Chemical composition, in per cent. P below .04 in acid steel ; below .03 in basic ; S. below .05 ; Si below .04 ; Mn .35 to .60. Physical requirements as follows : Shape. Diameter in inches. Ultimate strength ; pounds per square inch. So 11 K .Elongation in 8 inches; per cent. <5> O * JJJ& 5*fa 3 t-, >"d a o * Minimum. Maximum. Average. Minimum. Rivet rods. 6/ si 54000 64000 53000 53000 52000 52000 63000 09000 62000 62000 62000 62000 61.0 60.0 58.5 57.0 55.5 54.0 28.0 29.0 29.5 30.0 30.0 30.0 26.0 27.5 27.5 28.0 28.0 28.0 55 55 53 El 49 47 Two tons of bars from the same heat shall constitute a lot, and two specimens, each from a different bar, shall be tested from each lot. The above table gives the average required of these two bars, and the minimum below which no bar shall fall. If the average elongation or reduction of area on any one lot shall fall below the requirement, two additional bars shall be cut from the same lot, and the average of the four pieces shall be considered the aver- CLASS IV. Common Hard Rivets; for Roof Trusses and othe* Structures not Exposed to Shock. Method of manufacture. Acid or basic open-hearth process. Chemical composition, in per cent. P below .06 in acid steel; below .04 in basic; S. below .05 ; Si below .04 ; Mn. .35 to .60. Physical requirements as follows : Shape. Diameter in inches. Ultimate strength ; pounds per square inch. Elastic ratio. Elonga- tion in 8 inches; per cent. Reduc- tion of area; per cent. Minimum. Maximum. Rivet rods, fa 64000 54000 53000 53000 52000 52000 63000 62000 62000 62000 62000 62000 61.0 60.0 58.5 57.0 55.5 54.0 27.0 28.0 28.5 29.0 29.0 29.0 55 55 53 51 40 47 544 METALLURGY OF IRON AND STEEL. age of the lot, provided that no concession be made in the minimum. Heats rolled into sizes less than five-eighths inch may be tested in trial rods of three-quarters inch. Rivets, when cut out of the work into which they have been put, shall show a tough silky structure, with no crystalline appearance. CLASS V. Soft Bridge Steel ; for Angles, Plates, Bars, etc., for Bridges, Cranes, and Similar Structures Exposed to Shock. Method of manufacture. Acid or basic open-hearth process. Chemical composition, in per cent. P below .06 in acid steel, below .04 in basic; S below .07 in plates and angles, below .06 in eye-bars; Si below .04; Mn below .50. Physical requirements as follows: 8 Ult. str.; Ibs. per -t> oT sq. inch. 00 g 03 "a .2 n CD 8 a a g 35 Sj Remarks. QJ ^ S g a JM 1 a .2 *a C,d "o o * _o "2 H 1 O 3 > A H S 1 S" (S* . ^/ 50000 58000 63.0 29.0 55 o> i| 50000 58000 61.5 29.0 53 One piece of 24-inch angle must open out "Sb % 49000 58000 60.0 290 51 flat and another close shut without sign of A 3/ 49000 58000 58.5 29.0 49 fracture. ^ % 48000 58000 57.0 29.0 47 On plates under 42 inches wide the required elongation shall be raised 1.5 per cent, and the reduction of area 2.0 per cent. On plates over 70 inches wide the elongation shall be lowered 1.5 per cent, and the reduction of area 2.0 per 6 53000 63000 65.0 23.0 44 cent. On tests cut crosswise from the sheet, 00 |^ 51000 61000 63.0 26.0 50 the minimum tensile strength shall be low- 1 i 50000 49000 60000 59000 62.0 60.0 26.0 25.0 50 48 ered 3000 pounds, the elongation 3 per cent., and the reduction of area 10 per cent. On A* i 4 48000 58000 58.0 24.0 46 universal-mill plates the allowance for trans- i^ 47000 58000 t 56.0 23.0 44 verse tests shall be 5000 pounds, 5 per cent, and 15 percent. Both longitudinal and transverse strips cut from plates shall bend double flat. When every plate in the heat is tested, the minimum elongation and reduction shall be \ lowered 5 per cent. S4 50000 58000 57.0 1 i 50000 49000 49000 48000 58000 58000 58000 58000 56.0 54.0 53.0 52.0 The elongation in full length shall be 15 per cent, in bars from 10 to 20 feet long, 14 per cent, in 21 to 25 feet, 13.5 per cent, in 26 to 30 feet, and 13 per cent, in 31 to 35 feet. ::; SHAPES. In channels, beams, etc., the requirements on tests cut from the web shall be the same as for plates between 42 and 70 inches wide, with the same allowance for difference in thickness. In tests cut from the flange the minimum tensile strength shall be lowered 3000 pounds, the elongation 3 per cent., and the reduction of area 10 per cent.. Four tests shall be taken from each heat, and the average of these four shall conform to the above table. If the average elonga- tion or reduction of area of any heat shall fall below the require- CLASSIFICATION OF STRUCTURAL STEELS. 545 ment, four additional bars may be cut from the same heat, and the average of the eight pieces shall be considered the average of the heat. Heats rolled into sizes less than five-eighths inch may be tested in trial rods of three-quarters inch. Rivets, when cut out of the work into which they have been put, shall show a tough silky structure, with no crystalline Appearance. CLASS VI. Medium Bridge Steel; a Substitute for Class V when Greater Strength and Less Toughness are Required. "Method of manufacture. Acid or basic open-hearth process. Chemical composition, in per cent. P below .06 in acid steel, below .04 in basic; S below .07 in plates and angles, below .06 in eye-bars; Si below .04; Mn below .60. Physical requirements as follows: 1 Ultimate 1 strength ; * "of d Ibs. per sq. inch. *! \ d . "o X a a g M tS f/J Ofi Remarks. i d 1 1 3 to a ** S 05 3 4 d OQ 2 H 1 3 08 H W~ ft , % 56000 64000 63.0 27.0 50 s 56000 55000 64000 64000 61.5 60.0 27.0 27.0 48 46 One piece of angle, not over % inch thick, shall open out flat, and another close shut C3 % 55000 64000 58.5 27.0 44 without sign of fracture. %) elonga- tion in fifteen feet and the tensile strength specified, it shall not be cause for rejection, provided that not more than one-third (1/3) of the total number of eye-bars tested break in the head. 7. The three classes of structural steel for bridges and ships shall conform to the following bending tests ; and for this purpose the test specimen shall be one and one-half inches Bending Tests. w{ ^ if possible ^ and for all mate rial three-fourths inch (%") or less in thickness the test specimen shall be of the same thickness as that of the finished material from which it ia cut, but for material more than three-fourths inch (%") thick the bending test specimen may be one-half inch (%") thick: Rivet rounds shall be tested of full size as rolled. (d) Rivet steel shall bend cold 180 flat on itself without frac- ture on the outside of the bent portion. (e) Soft steel shall bend cold 180 flat on itself without frac- ture on the outside of the bent portion. (f) Medium steel shall bend cold 180 around a diameter equal to the thickness of the specimen tested, without fracture on the outside of the bent portion. 552 METALLURGY OF IRON AND STEEL. TEST PIECES AND METHODS OF TESTING. 8. The standard test specimen of eight-inch (8") gauged length, shall be used to determine the physical properties specified in paragraphs Nos. 4 and 5. The standard shape r of the test specimen for sheared plates is shown in Fig. XVIII-A. For other material the test speci- men may be the same as for sheared plates, or it may be planed or turned parallel throughout its entire length, and in all cases where possible, two opposite sides of the test specimens shall be the rolled surfaces. Eivet rounds and small rolled bars shall be tested of full size as rolled. PARALLEL SECTION M I NOT LESS THAN 9228.60 i,,^* r^CTjj , , MM .70 ft -i- ~f- *-- MM 9t 25.40 r MM ^ 18457.20- ABOUT PIECE TO BE OF SAME THICKNESS AS THE PLATE. FIG. XVIII-A. EIGHT-INCH TEST PIECE. 0. One tensile test specimen shall be taken from the finished Number of material of each melt, but in case this develops Tensile Tests. fl aws ^ or breaks ou t s ide of the middle third of its gauged length, it may be discarded and another test specimen sub- stituted therefor. 10. One test specimen for bending shall be taken from the finished material of each melt as it comes from the rolls, and for Tests specimen material three-fourths inch (%") and less in thick- ness this specimen shall have the natural rolled sur- face on two opposite sides. The bending test specimen shall be one and one-half inches (l 1 /^") wide, if possible, and for material more than three-fourths inch (%") thick the bending test specimen may be one-half inch {%") thick. The sheared edges of bending test specimens may be milled or planed. CLASSIFICATION OF STRUCTURAL STEEL. 553 (g) The bending test may be made by pressure or by blows. 11. Material which is to be used without annealing or further treatment shall be tested for tensile strength in the condition in Annealed Test which it comes from the rolls. Where it is imprac- specimens ticable to secure a test specimen from material which has been annealed or otherwise treated, a full-sized section of ten- sile test specimen lengthy shall be similarly treated before cutting the tensile specimen therefrom. 12. For the purpose of this specification, the yield point shall be determined by the careful observation of the drop Yield Point. of the beam or halt in the gauge of the testing ma- chine. 13. In order to determine if the material conforms to the Sam ie for chemical limitations prescribed in paragraph No. 2 chemical herein, analysis shall be made of drillings taken from a small test ingot. VARIATIONS IN WEIGHT. 14. The variation in cross section or weight of more than 2% per cent, from that specified will be sufficient cause for rejection, except in the case of sheared plates, which will be covered by the following permissible variations : (h) Plates 12!/2 pounds per square foot or heavier, up to 100 inches wide, when ordered to weight, shall not average more than 2% per cent, variation above or 2% per cent, below the theoretical weight. When 100 inches wide and over, 5 per cent, above or 5 per cent, below the theoretical weight. (i) Plates under 2 l / 2 pounds per square foot, when ordered to weight, shall not average a greater variation than the following : Up to 75 inches wide, 2% per cent, above or 2% per cent, below the theoretical weight. 75 inches wide up to 100 inches wide, 5 per cent, above or 3 per cent, below the theoretical weight. When 100 inches wide and over, 10 per cent, above or 3 per cent, below the theoretical weight. ********** (/) For all plates ordered to gauge, there will be permitted an average excess of weight over that corresponding to the dimensions on the order equal in amount to that specified in the following table: 554 METALLURGY OF IKON AND STEEL. TABLE OF ALLOWANCES FOR OVER WEIGHT FOR RECTANGULAR PLATES WHEN ORDERED TO GAUGE. Plates will be considered up to gauge if measuring not over 1/100 inch le than the ordered gauge. The weight of 1 cubic inch of rolled steel is assumed to be 0.2833 pound. PLATES 14 INCH AND OVER IN THICKNESS. Width of plate. Thickness of plate. 'UP to 75 inches 75 to 100 inches. Inch. Percent. Per cent. Over 100 inches. Per cent. 1/4 10 14 18 5/16 8 12 16 3/8 7 10 13 7/16 6 8 10 1/2 5 7 9 9/16 4% 6% 8% 5/8 4 6 8 Over 5/8 3% 5 6% PLATES UNDER % INCH IN THICKNESS. Width of plate. Thickness of plate. Up to 50 inches. 50 inches and abor. Inch. Per cent. Per cent. 1/8 up to 5/32 10 15 5/32 " 3/16 8% 12% 3/16 " 1/4 7 10 FINISH. 15. Finished material must be free from injurious seams, flaws or cracks, and have a workmanlike finish. BRANDING. 16. Every finished piece of steel shall be stamped with the melt number, and steel for pins shall have the melt number stamped on the ends. Eivets and lacing steel, and small pieces for pin plates and stiffeners, may be shipped in bundles, securely wired together, with the melt number on a metal tag attached. INSPECTION. 17. The inspector representing the purchaser, shall have all reasonable facilities afforded to him by the manufacturer to satisfy him that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment . CLASSIFICATION OF STRUCTURAL STEEL. 555 AMERICAN STANDARD SPECIFICATIONS FOR STRUCTURAL STEEL FOR BUILDINGS., PROCESS OF MANUFACTURE. 1. Steel may be made by either the open-hearth or Bessemer process. CHEMICAL PROPERTIES. 2. Neither of the two classes of structural steel for buildings shall contain more than 0.10 per cent, of phosphorus. PHYSICAL PROPERTIES. 3. There shall be two classes of structural steel for buildings, classes. namely : RIVET STEEL and MEDIUM STEEL which shall conform to the following physical qualities : Rivet Steel. Medium Steel. Tensile strength, pounds per square inch 50,000 to 60,000 60,000 to 70,000 Yield point, in pounds per square inch shall not be less than 1/2 T. S. 1/2 T. 8. Elongation, per cent, in eight inches shall not be less than 26 22 5. For material less than five-sixteenths inch (5/16"), and more than three-fourths inch (%") in thickness, the following Modifications in modifications shall be made in the requirements for SZSfSS elongation: material. (a) For each increase of one-eighth inch (%") in thickness above three-fourths inch (%") a deduction of one per cent. (1%) shall be made from the specified elongation. (&) For each decrease of one-sixteenth inch (1/16") in thick- ness below five-sixteenths inch (5/lte") a deduction of two and one-half per cent. (2%%) shall be made from the specified elon- gation. (c) For pins the required elongation shall be five per cent. (5%) less than that specified in paragraph No. 4, as determined on a test specimen, the center of which shall be one inch (1") from the surface. 656 METALLURGY OF IRON AND STEEL. i 6. The two classes of structural steel for buildings shall con- form to the following bending tests; and for this purpose the Bending test specimen shall be one and one-half inches Tests - (l l /2 ff ) widej if possible, and for all material three- fourths inch (%") or less in thickness the test specimen shall be of the same thickness as that of the finished material from which it is cut, but for material more than three-fourths inch (%") thick the bending test specimen may be one-half inch (%") thick: Rivet rounds shall be tested of full size as rolled. (d) Rivet steel shall bend cold 180 flat on itself without frac- ture on the outside of the bent portion. (e) Medium steel shall bend cold 180 around a diameter equal to the thickness of the specimen tested, without fracture on the outside of the bent portion. TEST PIECES AND METHODS OF TESTING. 7. The standard test specimen of eight-inch (8") gauged length shall be used to determine the physical properties specified in para- graphs Nos. 4 and 5. The standard shape of the test s P ec i m en for sheared plates shall be as before shown by Fig. XVI 1 1- A. For other material the test specimen may be the same as for sheared plates, or it may be planed or turned parallel throughout its entire length, and in all cases where possible two opposite sides of the test specimen shall be the rolled surfaces. Rivet rounds and small rolled bars shall be tested of full size as rolled. 8. One tensile test specimen shall be taken from the finished material of each melt or blow, but in case this develops flaws, or breaks outside of the middle third of its gauged Number of . , , . , , -,. , , , . . & . Tensile Tests. length, it may be discarded and another test speci- men substituted therefor. 9. One test specimen for bending shall be taken from the fin- ished material of each melt 01* blow as it comes from the rolls and Test specimen for material three-fourths inch (%") and less in Bending. thickness this specimen shall have the natural rolled surface on two opposite sides. The bending test specimen shall be one and one-half inches (iy 2 ") wide, if possible, and for material more than three-fourths inch (%") thick the bending test speci- men may be one-half inch (i/ 2 ") thick. The sheared edges of bend- ing test specimens may be milled or planed. CLASSIFICATION OF STRUCTURAL STEEL. 557 Eivet rounds shall be tested of full size as rolled. (/) The bending test may be made by pressure or by blows. 10. Material which is to be used without annealing or further treatment shall be tested for tensile strength in the condition in Annealed Test which it comes from the rolls. Where it is imprac- specimena. ticable to secure a test specimen from material which has been annealed or otherwise treated, a full-sized section of tensile test specimen length shall be similarly treated before cutting the tensile test specimen therefrom. 11. For the purpose of this specification, the yield point shall Yield be determined by the careful observation of the drop Point - of the beam or halt in the gauge of the testing machine. 12. In order to determine if the material conforms to the samples for chemical limitations prescribed in paragraph No. 2 chemical herein, analysis shall be made of drillings taken from a small test ingot. VARIATION IN WEIGHT. 13. The variation in cross section or weight of more than 2^/2 per cent., from that specified will be sufficient cause for rejection, except in the case of sheared plates, which will be covered by the following permissible variations : (g) Plates 12y 2 pounds per square foot or heavier, up to 100 inches wide, when ordered to weight, shall not average more than 2% per cent, variation above or 2~y 2 P er cent, below the theoretical weight. When 100 inches wide and over, 5 per cent, above or 5 per cent, below the theoretical weight. (h) Plates under 12% pounds per square foot, when ordered to weight, shall not average a greater variation than the following : Up to 75 inches wide, 2% per cent, above or 2% per cent, below the theoretical weight. 75 inches wide up to 100 inches wide, 5 per cent, above or 3 per cent, below the theoretical weight. When 100 inches wide and over, 10 per cent, above or 3 per cent, below the theoretical weight. ******** (i) For all plates ordered to gauge, there will be permitted an average excess of weight over that corresponding to the dimen- sions on the order equal in amount to that specified in the follow- ing table : 558 METALLURGY OF IRON AND STEEL. TABLE OP ALLOWANCES FOB OVERWEIGHT FOR RECTANGULAR PLATES WHEN ORDERED TO GAUGE. Plates will be considered up to gauge if measuring not over 1/100 inch less than the ordered gauge. The weight of 1 cubic inch of rolled steel is assumed to be 0.2833 pound. PLATES }4 INCH AND OVER IN THICKNESS. Width of plate. Thickness of plate. Up to 50 inches 75 to 100 inches. Over 100 inches. Inch. Per cent. Per cent. Per cent. 1/4 10 14 18 5/16 8 12 16 3/8 7 10 13 7/16 6 8 10 1/2 5 7 9 9/16 4% 6% 8% 6/8 4 6 8 Over 5/8 3% 5 6% PLATES UNDER % INCH IN THICKNESS. Width of plate. Thickness of plate. Up to 50 inches 50 inches and above. Inch. Per cent. Per cent. 1/8 up to 5/32 10 15 5/32 " 3/16 8% 12% 3/16 " 1/4 7 10 FINISH. 14. Finished material must be free from injurious seams, flaws or cracks, and have a workmanlike finish. BRANDING. 15. Every finished piece of steel shall be stamped with the melt or blow number, except that small pieces may be shipped in bundles securely wired together with the melt or blow number on a metal tag attached. INSPECTION. 16. The inspector representing the purchaser shall have all reasonable facilities afforded to him by the manufacturer to sat- isfy him that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. CLASSIFICATION OF STRUCTURAL STEEL. 559 AMEKICAN STANDAKD SPECIFICATIONS FOR OPEN-HEAKTH BOILEE PLATE AND EIVET STEEL. PROCESS OF MANUFACTURE. 1. Steel shall be made by the open-hearth process. CHEMICAL PROPERTIES. 2. There shall be three classes of open-hearth boiler plate and rivet steel, namely : FLANGE OR BOILER STEEL, FIRE BOX STEEL and EXTRA SOFT STEEL, which shall conform to the following limits in chemical composition: * Flange or Fire box Extra soft Boiler steel. steel. steel. Per cent. Per cent. Per cent. Phosphorus shall not exceed.... ( Basic 0.04 Basic 0.03 ( Acid 0.06 Acid 0.04 0.04 Sulphur shall not exceed 0.05 0.04 0.04 Manganese . 0.30 to 0.60 0.30 to 0.50 0.30 to 0.50 3. Steel for boiler rivets shall be of the EXTRA SOFT class, as 8^^ RiV6t specified in paragraphs Nos. 2 and 4. PHYSICAL PROPERTIES. 4. The three classes of open-hearth boiler plate and rivet steel,, namely: FLANGE OR BOILER STEEL, FIRE BOX STEEL and EXTRA SOFT STEEL, shall conform to the following physical qualities: Flange or Fire box Extra soft boiler steel. steel. steel. Tensile strength, pounds per square inch 55,000 to 65,000 52,000 to 62,000 45,000 to 55,000 Yield point, in pounds per square inch shall not be less than 1/2 T. S. 1/2 V. S. 1/2 T. S. Elongation, per cent, in eight inches shall not be less than 25 26 28 5. For material less than five-sixteenths inch (5/16"), and more than three-fourths inch (%") in thickness, the following Modifications in modifications shall be made in the requirements for SSS35S elongation: material. (a) For each increase of one-eighth inch (%"} 5GO METALLURGY OF IRON AND STEEL. in thickness above three-fourths inch (%"), a deduction 01 one per cent. (1%) shall be made from the specified elongation. (b) For each decrease of one-sixteenth-inch (1/16") in thick- ness below five-sixteenths inch (5/16") a deduction of two and one-half per cent. (2y 2 %) shall be made from the specified elon- gation. 6. The three classes of open-hearth boiler plate and rivet steel shall conform to the following bending tests and for this purpose Bending the test specimen shall be one and one-half inches Test s. (!%") wide if possible, and for all material three- fourths inch (%") or less in thickness the test specimen shall be of the same thickness as that of the finished material from which it is cut; but for material more than three-fourths inch (%") thick, the bending test specimen may be one-half inch (%") thick. Rivet rounds shall be tested of full size as rolled. (c) Test specimens cut from the rolled material as specified above, shall be subjected to a cold bending test, and also to a quenched bending test. The cold bending test shall be made on the material in the condition in which it is to be used, and prior to the quenched bending test, the specimen shall be heated to a light cherry-red as seen in the dark and quenched in water, the tem- perature of which is between 80 and 90 Fahrenheit. (d) Flange or boiler steel, fire box steel and rivet steel, both before and after quenching, shall bend cold one hundred and eighty degrees (180) flat on itself without fracture on the outside of the bent portion. 7. For fire box steel a sample taken from a broken tensile test specimen, shall not show any single seam or cavity more than one- Homogeneity fourth inch (y") long in either of the three frac- tures obtained on the test for homogeneity as de- scribed below in paragraph 12. TEST PIECES AND METHODS/ or TESTING. 8. The standard test specimen of eight-inch (8") gauged length shall be used to determine the physical properties specified in paragraphs Nos. 4 and 5. The standard shape of Test Specimen for ,-, Tensile Test. the test specimen for sheared plates shall be as be- fore shown by Fig. XVIII-A. For other material the test specimen may be the same as for sheared plates, or it may CLASSIFICATION OF STRUCTURAL STEEL. 561 be planed or turned parallel throughout its entire length and in all cases where possible two opposite sides of the test specimens c-hall be the rolled surfaces. Kivet rounds and small rolled bars shall be tested of full size as rolled. 9. One tensile test specimen will be furnished from each plate as. it is rolled, and two tensile test specimens will be furnished Number of from each melt of rivet rounds. In case any one of Tensile Tests. these develops flaws or breaks outside of the middle third of its gauged length, it may be discarded and another test specimen substituted therefor. 10. For material three-fourths inch (%") or less in thickness, the bending test specimen shall have the natural rolled surface on Test specimens two opposite sides. The bending test specimens cut for Bending. f rom plates shall be one and one-half inches (l 1 /^") wide and for material more than three-fourths (%") thick the bending test specimens may be one-half inch (%") thick. The sheared edges of bending test specimens may be milled or planed. The bending test specimens for rivet rounds shall be of full size as rolled. The bending test may be made by pressure or by blows. 11. One cold bending specimen and one quenched bending spec- imen will be furnished from each plate as it is rolled. Two cold bending specimens and two quenched bending speci- mens will be furnished from each melt of rivet rounds. The homogeneity test for fire box steel shall be made on one of the broken tensile test specimens. 12. The homogeneity test for fire box steel is made as follows : A portion of the broken tensile test specimen is either nicked with Homogeneity Tests a ^isel or grooved on a machine, transversely about for Fire BOX a sixteenth of an inch (1/16") deep, in three places about two inches (2") apart. The first groove should be made on one side, two inches (2") from the square end of the specimen; the second, two inches (2") from it on the oppo- site side; and the third, two inches (2") from the last, and on the opposite side from it. The test specimen is then put in a vise, with the first groove about a quarter of an inch (%") above the jaws, care being taken to hold it firmly. The projecting end of the test specimen is then broken off by means of a hammer, a number of light blows being used, and the bending being away from the groove. The specimen is broken at the other two grooves in the same way. The object of this treatment is to open and 562 METALLURGY OF IRON AND STEEL. visible to the eye any seams due to failure to weld up, or to foreign interposed matter, or cavities due to gas bubbles in the ingot. After rupture, one side of each fracture is examined, a pocket lens being used if necessary, and the length of the seams and cavities is determined. 13. For the purposes of this specification, the yield point shall Yield be determined by the careful observation of the drop Point. of the beam or halt in the gauge of the testing machine. 14. In order to determine if the material conforms to the- chemical limitations prescribed in paragraph No. 2 herein, analy- sam i for s * s sna ^ be ma( le of drillings taken from a small chemical test ingot. An additional check analysis may be made from a tensile specimen of each melt used on an order, other than in locomotive fire box steel. In the case of locomotive fire box steel a check analysis may be made from the tensile specimen from each plate as rolled. VARIATION IN WEIGHT. 15. The variation in cross section or weight of more than 2%- per cent, from that specified will be sufficient cause for rejection, except in the case of sheared plates, which will be covered by the following permissible variations: (e) Plates I2y 2 pounds per square foot or heavier, up to 100 inches wide, when ordered to weight, shall not average more than 2!/2 per cent, variation above or %y 2 P er cen t- below the theoretical weight. When 100 inches wide and over, 5 per cent, above or 5 per cent, below the theoretical weight. (/) Plates under 12% pounds per square foot, when ordered to weight, shall not average a greater variation than the following : Up to 75 inches wide, 2% per cent, above or 2% per cent, below the theoretical weight. 75 inches wide up to 100 inches wide, 5 per cent, above or 3 per cent, below the theoretical weight. When 100 inches wide and over, 10 per cent, above or 3 per cent, below the theoretical weight. ******** (g) For all plates ordered to gauge, there will be permitted an average excess of weight over that corresponding to the dimensions on the order equal in amount to that specified in the following table: CLASSIFICATION OF STRUCTURAL STEEL. 563 TABLE OF ALLOWANCES FOB OVERWEIGHT FOE RECTANGULAR PLATES WHEN ORDERED TO GAUGE. Plates will be considered up to gauge if measuring not over 1/100 inch less than the ordered gauge. The weight of 1 cubic inch of rolled steel is assumed to be 0.2833 pound. PLATES % INCH AND OVER IN THICKNESS. Width of plate. Thickness of plate. Up to 75 inches. 75 to 100 inches. Over 100 inches. Inch. Per cent. Per cent. Per cent. 1/4 10 14 18 5/16 8 12 16 3/8 7 10 13 7/16 6 8 10 1/2 5 7 9 9/16 4% ey 2 8% 5/8 4 6 8 Over 5/8 3% 5 6% PLATES UNDER ^4 INCH IN THICKNESS. Width of plate. Thickness of plate. Up to 50 inches. 50 inches and above. Inch. Per cent. Per cent. 1/8 up to 5/32 10 15 5/32 " 3/16 sy 2 12 y a 3/16 " 1/4 7 10 FINISH. 16. All finished material shall be free from injurious surface defects and laminations, and must have a workmanlike finish. BRANDING. 1.7. Every finished piece of steel shall be stamped with the melt number, and each plate, and the coupon or test specimen cut from it, shall be stamped with a separate identifying mark or number. Rivet steel may be shipped in bundles securely wired together with the melt number on a metal tag attached. 0, INSPECTION. 18. The inspector representing the purchaser, shall have all reasonable facilities afforded to him by the manufacturer to satisfy him that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. 564 METALLURGY OF IRON AND STEEL. AMERICAN STANDARD SPECIFICATIONS FOR STEEL RAILS. PROCESS OF MANUFACTURE. 1. (a) Steel may be made by the Bessemer or open-hearth process. (&) The entire process of manufacture and testing shall be in accordance with the best standard current practice, and special care shall be taken to conform to the following instructions. (c) Ingots shall be kept in a vertical position in pit heating furnaces. (d) No bled ingots shall be used. (e) Sufficient material shall be discarded from the top of the ingots to insure sound rails. CHEMICAL PROPERTIES. 2. Rails of the various weights per yard specified below shall conform to the following limits in chemical composition : 50 to 59+ 60 to 69+ 70 to 79+ 80 to 89+ 90 to 100 pounds. pounds. pounds. pounds. pounds. Per cent. Per cent. Per cent. Per cent. Per cent. Carbon 0.35--0.45 0.38--0.48 0.40--0.50 0.43--0.53 0.45--0.55 Phosphorus shall not exceed 0.10 0.10 0.10 0.10 0.10 Silicon shall not exceed 0.20 0.20 0.20 0.20 0.20 Manganese 0.70-1.00 0.70--1.00 0.75--1.05 0.80-1.10 0.80-1.10 PHYSICAL PROPERTIES. 3. One drop test shall be made on a piece of rail not more than six feet long, selected from every fifth blow of steel. The rail Drop shall be placed head upwards on the supports and the various sections shall be subjected to the follow- ing impact tests: Weight of rail. Height of drop. Pounds per yard. Feet. 45 to and including 55 15 More than 55 65 75 M 85 65. 75. 85. 100. 16 17 18 19 CLASSIFICATION OF STRUCTURAL STEEL. 565 r- If any rail break when subjected to the drop test, two additional tests will be made of other rails from the same blow of steel, and if either of these latter tests fail all the rails of the blow which they represent will be rejected, but if both of these additional test pieces meet the requirements, all the rails of the blow which they represent will be accepted. If the rails from the tested blow shall be rejected for failure to meet the requirements of the drop test as above specified, two other rails will be subjected to the same tests, one from the blow next preceding, and one from the blow next suc- ceeding the rejected blow. In case the first test taken from the preceding or succeeding blow shall fail, two additional tests shall be taken from the same blow of steel, the acceptance or rejection of which shall also be determined as specified above, and if the rails of the preceding or succeeding blow shall be rejected, similar tests may be taken from the previous or following blows, as the case may be, until the entire group of five blows is tested, if necessary. The acceptance or rejection of all the rails from any blow will depend upon the result of the tests thereof. TEST PIECES AND METHODS OF TESTING. 4. The drop test machine shall have a tup of two thousand (2000) pounds weight, the striking face of which shall have a radius Drop Testing of not more than five inches (5"), and the test Machine. ra j]i ^n De placed head upwards on solid supports three feet (3') apart. The anvil block shall weigh at least twenty thousand (20,000) pounds, and the supports shall be a part of, or firmly secured to, the anvil. The report of the drop test shall state the atmospheric temperature at the time the tests were made. 5. The manufacturer shall furnish the inspector, daily, with carbon determinations of each blow, and a complete chemical analy- Sampiefor s ^ s ever j twenty-four hours, representing the aver- chemicai age of the other elements contained in the steel. These analyses shall be made on drillings taken from a small test ingot. FINISH. 6. Unless otherwise specified, the section of rail shall be the American Standard, recommended by the American Society of Civil Engineers, and shall conform, as accurately as possible, to the templet furnished by the railroad company, consistent with paragraph No. 7, relative to specified 566 METALLURGY OF IRON AND STEEL. weight. A variation in height of one sixty-fourth of an inch (1/64") less and one thirty-second of an inch (1/32") greater than the specified height will be permitted. A perfect fit of the splice bars, however, shall be maintained at all times. 7. The weight of the rails shall be maintained as nearly as pos- sible after complying with paragraph No. 6, to that specified in contract. A variation of one-half of one per cent. (1/2%) for an entire order will be allowed. Rails shall be accepted and paid for according to actual weights. 8. The standard length of rails shall be thirty feet (30'). Ten .per cent. (10%) of the entire order will be accepted in shorter lengths, varying by even feet down to twenty-four feet (24:'). A variation of one-fourth of an inch (i/i") in length from that specified will be allowed. 9. Circular holes for splice bars shall be drilled in accordance with the specifications of the purchaser. The holes shall accurately conform to the drawing and dimensions furnished in every respect, and must be free from burrs. 10. Rails shall be straightened while cold, smooth on head, sawed square at ends, and, prior to shipment, shall have the burr oc- casioned by the saw cutting, removed, and the ends made clean. Number 1 rails shall be free from in- jurious defects and flaws of all kinds. BRANDING. 11. The name of the maker, the month and year of manufac- ture, shall be rolled in raised letters on the side of the web, and the number of the blow shall be stamped on each rail. INSPECTION. 12. The inspector representing the purchaser shall have all rea- sonable facilities afforded to him by the manufacturer to satisfy him that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. No. 2 RAILS. 13. Rails that possess any injurious physical defects, or which for any other cause are not suitable for first quality, or No. 1 rails, CLASSIFICATION OF STRUCTURAL STEEL. 567 shall be considered as No. 2 rails, provided, however, that rails which contain any physical defects which seriously impair their strength shall be rejected. The ends of all No. 2 rails shall be painted in order to distinguish them. AMEEICAN STANDAED SPECIFICATIONS FOR STEEL SPLICE BAES. PROCESS OF MANUFACTURE. 1. Steel for splice bars may be made by the Bessemer or open- hearth process. CHEMICAL PROPERTIES. = , 2. Steel for splice bars shall conform to the following limits in chemical composition : Per cent. Carbon shall not exceed 0.15 Phosphorus shall not exceed 0.10 Manganese 0.30 to 0.60 PHYSICAL PROPERTIES. 3. Splice bar steel shall conform to the following physical Tensile qualities : Tests. Tensile strength, pounds per square inch 54,000 to 64,000 Yield point, pounds per square inch 32,000 Elongation, per cent, in eight inches shall not be less than 25 4. (a) A test specimen cut from the head of the splice bar shall Bending bend 180 flat on itself without fracture on the out- Tcsts - side of the bent portion. (b) If preferred the bending tests may be made on an un- punched splice bar, which, if necessary, shall be first flattened, and shall then be bent 180 flat on itself without fracture on the outside of the bent portion. TEST PIECES AND METHODS OF TESTING. 5. A test specimen of eight-inch (8") gauged length, cut from Test specimen tor the head of the splice bar, shall be used to determine Tensile Tests, ne phy s i ca l properties specified in paragraph No. 3. 568 METALLURGY OF IRON AND STEEL. 6. One tensile test specimen shall be taken from the rolled splice bars of each blow or melt, but in case this develops Number of flaws, or breaks outside of the middle third of its Tensile Tests gauged length, it may be discarded and another test specimen substituted therefor. 7. One test specimen cut from the head of the splice bar shall Test specimen be taken from a rolled bar of each blow or melt, or for Bending. jf preferred the bending test may be made on an un- punched splice bar, which, if necessary, shall be flattened before testing. The bending test may be made bv pressure or by blows. 8. For the purposes of this specification, the yield point shall Yi e id be determined by the careful observation of the drop Point - of the beam or halt in the gauge of the testing ma- chine. 9. In order to determine if the material conforms to the chemi- sampiefor ' ca ^ limitations prescribed in paragraph No. 2 herein, chemical analysis shall be made of drillings taken from a small test ingot. FINISH. 10. All splice bars shall be smoothly rolled and true to templet. The bars shall be sheared accurately to length and free from fins and cracks, and shall perfectly fit the rails for which they are in- tended. The punching and notching shall accurately conform in every respect to the drawing and dimensions furnished. A vari- ation in weight of more than 2% per cent, from that specified will be sufficient cause for rejection. BRANDING. 11. The name of the maker and the year of manufacture shall be rolled in raised letters on the side of the splice bar. INSPECTION. 12. Tht inspector representing the purchaser, shall have all reasonable facilities afforded to him by the manufacturer, to satisfy him that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. CLASSIFICATION OF STRUCTURAL STEEL. 569 AMERICAN STANDARD SPECIFICATIONS FOR STEEL AXLES. PROCESS OF MANUFACTURE. . 1. Steel for axles shall be made by the open-hearth process. CHEMICAL PROPERTIES. 2. There will be three classes of steel axles which shall conform to the following limits in chemical composition : Car, engine truck Driving wheel Driving wheel and tender truck axles. axles. axles. (Carbon steel.) (Nickel steel.) Per cent. Per cent. Per cent. Phosphorus shall not exceed. .. 0.06 0.06 0.04 Sulphur " " " 0.06 0.06 0.04 Nickel .... 3.00-4.00 PHYSICAL PROPERTIES. 3. For car, engine truck, and tender truck axles no tensile test TeTts ile shall be required. 4. The minimum physical qualities required in the two classes of driving wheel axles shall be as follows : Driving wheel Driving wheel axles. axles. ( Carbon steel. ) ( Nickel steel. ) Tensile strength, pounds per square inch. . 80,000 80,000 Yield point, pounds per square inch 40,000 50,000 Elongation, per cent, in two inches 18 25 Contraction of area per cent . . 45 5. One axle selected from each melt, when tested by 'the drop test described in paragraph No. 9, shall stand the number of blows Drop at the height specified in the following table with- Test - out rupture and without exceeding, as the result of the first blow, the deflection given. Any melt failing to meet these requirements will be rejected. Height of Deflection. Inches. 81/4 81/4 81/4 8 8 7 51/2 Diameter of Height of axle at center. Number of drop. Inches. blows. Feet. 41/4 5 24 43/8 5 26 4 7/16 5 281/2 45/8 5 31 43/4 5 34 53/8 5 43 57/8 7 43 570 METALLURGY OF IRON AND STEEL. 6. Carbon steel and nickel steel driving wheel axles shall not be subject to the above drop test. TEST PIECES AND METHODS OF TESTING. 7. The standard turned test specimen one-half inch (%") di- ameter and two-inch (2") gauged length, shall be used to determine the physical properties specified in paragraph No. 4. It is shown in Fig. XVIII-B. For driving axles one longitudinal test specimen shall be cut . from one axle of each melt. The center of this test Number and Loca- tion of Tensile specimen shall be half way between the center and specimens outside of the axle. Tt Specimen for Tensile Tests. 8. FIG. XVIII-B. TWO-INCH TEST PIECE. 9. The points of supports on which the axle rests during tests must be three feet apart from center to center ; the tup must weigh Drop Test 1640 pounds ; the anvil, which is supported on springs, must weigh 17,500 pounds ; it must be free to move in a vertical direction; the springs upon which it rests must be twelve in number, of the kind described on drawing; and the radius of supports and of the striking face on the tup in the direction of the axis of the axle must be five (5) inches. When an axle is tested it must be so placed in the machine that the tup will strike it midway between the ends, and it must be turned over after the first and third blows, and when required, after the fifth blow. To measure the deflection after the first blow prepare a straight CLASSIFICATION OF STRUCTURAL STEEL. edge as long as the axle, by reinforcing it on one side, equally at each end, so that when it is laid on the axle, the reinforced parts will rest on the collars or ends of the axle, and the balance of the straight edge not touch the axle at any place. Next place the axle in position for test, lay the straight edge on it, and measure the dis- tance from the straight edge to the axle at the middle point of the latter. Then after the first blow, place the straight edge on the now bent axle in the same manner as before, and measure the distance from it to that side of the axle next to the straight edge at the point farthest away from the latter. The difference beween the two meas- urements is the deflection. The report of the drop test shall state the atmospheric temperature at the time the tests were made. 10. The yield point specified in paragraph No. 4 shall be deter- Yieid mined by the careful observation of the drop of the Point - beam, or halt in the gauge of the testing machine. 11. Turnings from the tensile test specimen of driving axles, or drillings taken midway between the center and outside of car, Sam lefor engine, and tender truck axles, or drillings from chemical the small test ingot if preferred by the inspector, shall be used to determine whether the melt is with- in the limits of chemical composition specified in paragraph No. 2. FINISH. 12. Axles shall conform in sizes, shapes and limiting weights to the requirements given on the order or print sent with it. They shall be made and finished in a workmanlike manner, and shall be free from all injurious cracks, seams or flaws. In centering, sixty (60) degree centers must be used, with clearance given at the point to avoid dulling the shop lathe centers. BRANDING. 13. Each axle shall be legibly stamped with the melt number and initials of the maker at the places marked on the print or in- dicated by the inspector. INSPECTION. 14. The inspector representing the purchaser, shall have all rea- sonable facilities afforded to him by the manufacturer to satisfy him that the finished material is furnished in accordance with these 572 METALLURGY OF IRON AND STEEL. specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. AMERICAN STANDARD SPECIFICATIONS FOR STEEL TIRES. PROCESS OF MANUFACTURE. 1. Steel for tires may be made by either the open-hearth or crucible process. CHEMICAL PROPERTIES. 2. There will be three classes of steel tires which shall conform to the following limits in chemical composition : Passenger Freight engine Switching engines. and car wheels. engines. Per cent. Per cent. Per cent. Manganese shall not exceed ____ 0.80 0.80 0.80 Silicon shall not be less than... 0.20 0.20 0.20 Phosphorus shall not exceed ____ 0.05 0.05 0.05 Sulphur shall not exceed ....... 0.05 0.05 0.05 PHYSICAL PROPERTIES. 3. The minimum physical qualities required in each of the three classes of steel tires shall be as follows : Pas- Freight engine Switch- . senger and car Ing en- engines. wheels. gines. Tensile strength, pounds per square inch. 100,000 110,000 120,000 Elongation, per cent, in two inches ....... 12 10 8 4. In the event of the contract calling for a drop test, a test tire from each melt will be furnished at the purchaser's expense, pro- Drop vided it meets the requirements. This test tire shall stand the drop test described in paragraph No. 7, without breaking or cracking, and shall show a minimum deflection equal to D 2 -f-(40T 2 +2D), the letter "D" being internal diameter and the letter "T" thickness of tire at center of tread. CLASSIFICATION 6F STRUCTURAL STEEL. 573 TEST PIECES AND METHODS OF TESTING. 5. The standard turned test specimen, one-half inch (%") di- Tist specimen for ameter and two-inch (2") gauged length, shall be Tensile Tests. used to determine the physical properties specified in paragraph No. 3. It has been already shown in Fig. XV1II-B. 6. When the drop test is specified, this test specimen shall be cut cold from the, tested tire at the point least affected by the drop test. If the diameter of the tire is such that ^ ne wn l e circumference of the tire is seriously af- fected by the drop test, or if no drop test is required, the test specimen shall be forged from a test ingot cast when pour- ing the melt, the test ingot receiving, as nearly as possible, the same proportion of reduction as the ingots from which the tires are made. 7. The test tire shall be placed vertically under the drop in a running position on a solid foundation of at least ten tons in weight Drop Test and subjected to successive blows from a tup weigh- Described. j n g 2240 pounds, falling from increasing heights until the required deflection is obtained. 8. Turnings from the tensile specimen, or drillings from the small test ingot, or turnings from the tire if preferred by the in- chTmica'r spector, shall be used to determine whether the melt Analysis. is within the limits of chemical composition speci- fied in paragraph No. 2. FINISH. 9. All tires shall be free from cracks, flaws, or other injurious imperfections, and shall conform to dimensions shown on draw- ings furnished by the purchaser. BRANDING. 10. Tires shall be stamped with the maker's brand and number in such a manner that each individual tire may be identified. INSPECTION. 11. The inspector representing the purchaser, shall have all rea- sonable facilities afforded to him by the manufacturer to satisfy him that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. 574 METALLURGY OF IRON AND STEEL. AMEEICAN STANDAKD SPECIFICATIONS FOR STEEL FOKGINGS. PROCESS OF MANUFACTURED 1. Steel for forgings may be made by the open-hearth, crucible or Bessemer process. CHEMICAL PROPERTIES. 2. There will be four classes of steel forgings which shall con- form to the following limits in chemical composition : II! a fl a 5 * b "6 -H eS O fa 5 d a|l a a S a |I5 S r cS ~ fc. fa o O O a " ' Phosphorus shall not exceed. Sulphur Nickel fa a o o Per cent. Per cent. Per cent. Per cent. 0.10 0.06 0.04 0.04 0.10 0.06 0.04 0.04 3.00-4.00 3. Tensile Tests. PHYSICAL PROPERTIES. The minimum physical qualities required of the different sized forgings of each class shall be as follows: Pounds per square inch. 58,000 29,000 75,000 - 37,500 | S tJ 1 ss g^ flS S a o o Per cent. 28 35 SOFT STEEL OR Low CARBON STEEL. For solid or hollow forgings, no diameter or thickness of section to exceed 10". 18 CARBON STEEL NOT ANNEALED. For solid or hollow forgings, no diameter or 30 thickness of section to exceed 10". Elastic CARBON STEEL ANNEALED. limit For solid or hollow forgings, no diameter or 80,000 40,000 22 35 thickness of section to exceed 10". For solid forgings, no diameter to exceed 20" 75,000 37,500 23 35 or thickness of section 15". 70,000 35,000 24 30 For solid forgings, over 20" diameter. CARBON STEEL, OIL TEMPERED. For solid or hollow forgings, no diameter or 90,000 55,000 20 45 thickness of section to exceed 3". CLASSIFICATION OF STRUCTURAL STEEL. 575 S a Pounds per square inch. ^5,000 50,000 80,000 45,000 80,000 50,000 80,000 45,000 80,000 45,000 95,000 65,000 90,000 60,000 85,000 55,000 g fl o o CARBON STEEL, OIL TEMPERED. Per For solid forgings of rectangular sections not cent. exceeding 6" in thickness or hollow forgings, the walls of which do not exceed 6" in thick- 22 45 ness. For solid forgings of rectangular sections not exceeding 10" in thickness or hollow forgings, the walls of which do not exceed 10" in thick- 23 40 ness. NICKEL STEEL ANNEALED. For solid or hollow forgings, no diameter or 25 45 thickness of section to exceed 10". For solid forgings, no diameter to exceed 20" 25 45 or thickness of section 15". 24 40 For solid forgings, over 20" diameter. NICKEL STEEL, OIL TEMPERED. For solid or hollow forgings, no diameter or 21 50 thickness of section to exceed 3". For solid forgings of rectangular sections not exceeding 6" in thickness or hollow forgings, the walls of which do not exceed. 6" in thick- 22 50 ness. For solid forgings of rectangular sections not exceeding 10" in thickness or hollow forgings, the walls of which do not exceed 10" in thick- 24 45 ness. 4. A specimen one inch by one-half inch (I"xy 2 ") shall bend Bending cold 180 without fracture on outside of bent por- Test - tion, as follows: Around a diameter of y 2 ", f r forgings of soft steel. Around a diameter of 1%", for forgings of carbon steel not an- nealed. Around a diameter of 1%", for forgings of carbon steel if 20" in diameter or over. Around a diameter of 1", for forgings of carbon steel annealed, if under 20" diameter. Around a diameter of I" for forgings of carbon steel oil-tem- pered. Around a diameter of %", for forgings of nickel steel annealed. Around a diameter of 1", for forgings of nickel steel oil-tem- pered. 576 METALLURGY OF IRON AND STEEL. TEST PIECES AND METHODS OF TESTING. 5. The standard turned test specimen, one-half inch (i/ 2 ") di- ameter and two-inch (2") gauged length, shall be used to deter- Test specimen for mine the physical properties specified in paragraph Tensile Test. No. 3. It has already been shown in Fig. XYI1I-B. 6. The number and location of test specimens to be taken from a melt, blow, or a forging shall depend upon its character and im- d portance and must therefore be regulated by indi- Locationof vidual cases. The test specimens shall be cut cold Tensile specimens. from the f org i n g or full-sized prolongation of same, parallel to the axis of the forging and half way between the center and outside, the specimens to be longitudinal, i.e., the length of the specimen to correspond with the direction in which the metal is most drawn out or worked. When f orgings have large ends or col- lars, the test specimens shall be taken from a prolongation of the same diameter or section as that of the forging back of the large end or collar. In the case of hollow shafting, either forged or bored, the specimen shall be taken within the finished section pro- longed, half way between the inner and outer surface of the wall of the forging. 7. The specimen for bending test one inch by one-half inch Test specimen (r'x 1 /^") shall be cut as specified in paragraph No. for Bending. Q rpk e k en( }j n g ^ e st may be made by pressure or by blows. 8. The yield point specified in paragraph No. 3 shall be deter- Yieid mined by the careful observation of the drop of the beam, or halt in the gauge of the testing machine. 9. The elastic limit specified in paragraph No. 3 shall be deter- mined by means of an extensometer, which is to be attached to the Elastic test specimen in such manner as to show the change in rate of extension under uniform rate of loading, and will be taken at that point where the proportionality changes. 10. Turnings from the tensile specimen or drillings from the sample for bending specimen or drillings from the small test A^aTlrif i n ot > if preferred by the inspector, shall be used to determine whether or not the steel is within the lim- its in chemical composition specified in paragraph No. 2. FINISH. 11. Forgings shall be free from cracks, flaws, seams or other CLASSIFICATION OF STRUCTURAL STEEL. 577 injurious imperfections, and shall conform to dimensions shown on drawings furnished by the purchaser, and be made and finished in a workmanlike manner. INSPECTION. 12. The inspector representing the purchaser shall have all rea- sonable facilities afforded to him by the manufacturer to satisfy him that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. AMERICAN STANDARD SPECIFICATIONS FOR STEEL CASTINGS. PROCESS OF MANUFACTURE. 1. Steel for castings may be made by the open-hearth, crucible or Bessemer process. Castings to be annealed or unannealed as specified. CHEMICAL PROPERTIES. 2. Ordinary castings, those in which no physical requirements rdinary are specified, shall not contain over 0.40 per cent, of castings. carbon, nor over 0.08 per cent, of phosphorus. 3. Castings which are subjected to physical test shall not con- Tested tain over 0.05 per cent, of phosphorus, nor over 0.05 Castings. p er cen ^ O f sulphur. PHYSICAL PROPERTIES. 4. Tested castings shall be of three classes : "HARD/' "MEDIUM" Tensile and "SOFT." The minimum physical qualities re- Tests - quired in each class shall be as follows : Hard Medium Soft castings. castings. castings. Tensile strength, pounds per square inch 85,000 70,000 60,000 Yield point, pounds per square inch 38,250 31,500 27,000 Elongation, per cent, in two inches 15 18 22 Contraction of area, per cent 20 25 30 5. A test to destruction may be substituted for the tensile test, 578 METALLURGY OF IRON AND STEEL. in the case of small or unimportant castings, by selecting three cast- Drop ings from a lot. This test shall show the material Test - to be ductile and free from injurious defects, and suitable for the purposes intended. A lot shall consist of all cast- ings from the same melt or blow, annealed in the same furnace charge. 6. Large castings are to be suspended and hammered all over. Percussive No cracks, flaws, defects, nor weakness shall appear Test - after such treatment. 7. A specimen one inch by one-half inch (l"x%") shall bend cold around a diameter of one inch (1") without fracture on out- Bending side of bent portion, through an angle of 120 for Te8t - "SOFT" castings, and of 90 for "MEDIUM" castings. TEST PIECES AND METHODS OF TESTING. 8. The standard turned test specimen, one-half inch (%") Di- ameter and two-inch (2") gauged length, shall be use( ^ *o determine the physical properties specified in paragraph No. 4. It has already been shown in Fig. XVIII-B. 9. The number of standard test specimens shall depend upon the character and importance of the castings.. A test piece shall Number and be cut cold f rom a cou P on to be molded and cast on Location of some portion of one or more castings from each melt Tensile Specimens. ,, ., .,, -, ,. , -, or blow or from the sink-heads (in case heads of suf- ficient size are used). The coupon or sink-head must receive the same treatment as the casting or castings, before the specimen is cut out, and before the coupon or sink-head is removed from the casting. 10. One specimen for bending test one inch by one-half inch (l" xl /2") shall be cut cold from the coupon or sink-head of the Test specimen casting or castings as specified in paragraph No. 9. or Bending. The bending tegt may be made by press ure, or by blows. 11. The yield point specified in paragraph No. 4 shall be deter- Yield mined by the careful observation of the drop of the beam or halt in the gauge of the testing machine. 12. Turnings from the tensile specimen, drillings from the CLASSIFICATION OF STRUCTURAL STEEL. 5i9 bending specimen, or drillings from the small test ingot, if pre- sampiefor f erred by the inspector, shall be used to determine chemical whether or not the steel is within the limits in phos- phorus and sulphur specified in paragraphs Nos. 2 and 3. FINISH. 13. Castings shall be true to pattern, free from blemishes, flaws or shrinkage cracks. Bearing surfaces shall be solid, and no poros- ity shall be allowed in positions where the resistance and value of the casting for the purpose intended, will be seriously affected thereby. INSPECTION. 14. The inspector representing the purchaser, shall have all rea- sonable facilities afforded to him by the manufacturer to satisfy him that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. AMEEICAN STANDAED SPECIFICATIONS FOR WKOTJGHT IKON. PROCESS OF MANUFACTURE. 1. Wrought-iron shall be made by the puddling or by the char- coal hearth process or rolled from fagots or piles made from wrought-iron scrap, alone or with muck bar added. PHYSICAL PROPERTIES. 2. The minimum physical qualities required in the four classes Tensile o f wr ought-iron shall be as follows : Merchant Merchant Merchant Stay-bolt iron. iron. iron. iron. Grade "A." Grade "B." Grade "C." Tensile strength, pounds per square inch 46,000 50,000 48,000 48,000 Yield point, ' pounds per square inch . '. 25,000 25,000 25,000 25,000 .Elongation, per cent, in eight Inches . 28 25 20 20 580 METALLURGY OF IRON AND STEEL. 3. In sections weighing less than 0.654 pound per linear foot, the percentage of elongation required in the four classes specified in paragraph No. 2, shall be 21 per cent., 18 per cent., 15 per cent, and 12 per cent., respectively. 4. The four classes of iron when nicked and tested as described Nicking -in paragraph No. 9 shall show the following frac- Test - ture : (a) Stay-bolt iron, a long, clean, silky fiber, free from slag or dirt, and wholly fibrous, being practically free from crystalline spots. (ft) Merchant iron, Grade "A," a long, clean, silky fiber, free from slag or dirt or any coarse crystalline spots. A few fine crys- talline spots may be tolerated, provided they do not in the aggregate exceed 10 per cent, of the sectional area of the bar. (c) Merchant iron, Grade "B," a generally fibrous fracture, free from coarse crystalline spots. Not over 10 per cent, of the fractured surface shall be granular. (d) Merchant iron, Grade "C," a generally fibrous fracture, free from coarse crystalline spots. Not over 15 per cent, of the fractured surface shall be granular. 5. The four classes of iron, when tested as described in para- cold Bending graph No. 10, shall conform to the following bend- Te8t - ing tests : (e) Stay-bolt iron; a piece of stay-bolt iron about 24 inches long shall bend in the middle through 180 flat on itself, and then bend in the middle through 180 flat on itself in a plane at a right angle to the former direction, without a fracture on outside of the bent portions. Another specimen with a thread cut over the entire length shall stand this double bending without showing deep cracks in the threads. (f) Merchant iron, Grade "A," shall bend cold 180 flat on itself without fracture on outside of the bent portion. (g) Merchant iron, Grade "B," shall bend cold 180 around a diameter equal to the thickness of the tested specimen, without fracture on outside of the bent portion. (h) Merchant iron, Grade "C," shall bend cold 130 around a diameter equal to twice the thickness of the specimen tested, with- out fracture on outside of the bent portion. CLASSIFICATION OF STRUCTURAL STEEL. 581 6. The four classes of iron, when tested as described in para- Hot Bending graph No. 11, shall conform to the following hot Te8t - bending tests : (i) Stay-bolt iron shall bend through 180 flat on itself, with- out showing cracks or flaws. A similar specimen heated to a yel- low heat and suddenly quenched in water between 80 and 90 F. shall bend, without hammering on the bend, 180 flat on itself without showing cracks or flaws. (/) Merchant iron, Grade "A," shall bend through 180 flat on itself, without showing cracks or flaws. A similar specimen heated to a yellow heat and suddenly quenched in water between 80 and 90 F. shall bend, without hammering on the bend, 180 flat on itself without showing cracks or flaws. A similar specimen heated to a bright red heat shall be split at the end and each part bent back through an angle of 180. It will also be punched and expanded by drifts until a round hole is formed whose diameter is not less than nine-tenths of the diameter of the rod or width of the bar. Any extension of the original split or indications of frac- ture, cracks, or flaws developed by the above tests will be sufficient cause for the rejection of the lot represented by that rod or bar. (Te) Merchant iron, Grade "B," shall bend through 180 flat on itself, without showing cracks or flaws. (I) Merchant iron, Grade "C," shall bend sharply to a right angle, without showing cracks or flaws. 7. Stay-bolt iron shall permit of the cutting of a clean sharp thread and be rolled true to gauge desired, so as not Threading Test. . . to- jam in the threading dies. TEST PIECES AND METHODS OF TESTING. 8. Whenever possible, iron shall be tested in full size as rolled, to determine the physical qualities specified in paragraphs Nos. 2 and 3, the elongation being measured on an eight- inch < 8 ") g au S ed kagto- In flats and sha P es to large to test as rolled, the standard test specimen shall be one and one-half inches (!%") wide and eight inches (8") gauged length. In large rounds, the standard test specimen of two inches (2") gauged length shall be used ; the center of this specimen shall be half way between the center and outside of the round. Sketches of 582 METALLURGY OF IRON AND STEEL. these two standard test specimens have been already shown in Fig. XVIII-A and Fig. XVIII-B. 9. Nicking tests shall be made on specimens cut from the iron as rolled. The specimen shall be slightly and evenly nicked on one side and bent back at this point through an angle of 180 by a succession of light blows. 10. Cold bending tests shall be made on specimens cut from the cold Bending bar as rolled. The specimen shall be bent through Tests. an angle of 180 by pressure or by a succession of light blows. 11. Hot bending tests shall be made on specimens cut from the bar as rolled. The specimens, heated to a bright red heat, shall be Hot Bending bent through an angle of 180 by pressure or by a Tests. succession of light blows and without hammering directly on the bend. If desired, a similar bar of any of the four classes of iron, shall be worked and welded in the ordinary manner without showing signs of red shortness. 12. The yield point specified in paragraph No. 2 shall be deter- Yieid mined by the careful observation of the drop of the Point - beam or haft in the gauge of the testing machine. FINISH. 13. All wrought iron must be practically straight, smooth, free from cinder spots or injurious flaws, buckles, blisters or cracks. In round iron, sizes must conform to the Standard Limit gauge as adopted by the Master Car Builders' Association in November, 1883. INSPECTION. 14. Inspectors representing the purchasers, shall have all reasonable facilities afforded them by the manufacturer to satisfy them that the finished material is furnished in accordance with these specifications. All tests and inspections shall be made at the place of manufacture, prior to shipment. CHAPTER XIX. WELDING. SECTION XlXa. Influence of structure on the welding proper- ties. Wrought-iron may be welded so that the point of union is as strong as the rest of the bar, for by upsetting the piece there can be an extra amount of work put- upon the metal,, and since the strength of the original bar was dependent upon the perfection of a great number of welds,, it follows that the additional local heating and hammering may give a superior strength. Unfortunately, this is rarely the case, and it is seldom that failure does not take place in the neighborhood of the weld under destructive tests. It often does happen that a rod will break a short distance away from the actual point of union, but in spite of current supposition this by no means shows perfect workmanship, for it usually arises from the overheating of the iron at the point of fracture, without sufficient subsequent work to develop a proper structure. In working steel the conditions are fundamentally different, for the bar is not a collection of fibres and welds, but a thing complete in itself, so that it is impossible to make any improvement in a properly worked piece by cutting it in halves and putting it together again. It is quite conceivable that a bar may originally be under- worked or overheated, and that additional local work can enhance the strength at the point of welding, but this assumption of a bad material to start with may be neglected. 'It is also possible to finish the hammering on a welded piece at a very low temperature and thereby exalt the ultimate strength beyond the true value, but inasmuch as this will give a' less ductile and unreliable material, it will not be considered. It is also possible, much more than with wrought-iron, to have the weld stronger than a certain adjacent part of the bar, for the best of steel will be crystallized by high heat somewhat more readily than wrought-iron, and hence it can and often does happen that the metal in the neighborhood of the weld has a bad structure due to 583 584 METALLURGY OF IRON AND STEEL. lack of hammering after high heating. The higher the critical temperature necessary to produce crystallization, the less is the danger from this source, so that, aside from the mere facility of welding at point of contact, the freedom from phosphorus and sulphur is a matter of prime importance, since both of these elements render the metal less able to withstand high temperatures. The fundamental difference in crystallizing power between wrought-iron and steel makes a close comparison of the two im- possible, but nevertheless it may be profitable to quote from Holley the following conclusions concerning iron :* "(1) None of the ingredients except carbon in the proportions present seems to very notably affect the welding by ordinary meth- ods. [The maximum percentages were P, .317 ; Si, .321 ; Mn, .097 ; S, .015; Cu, .43; Ni, .34; Co, .11; Slag, 2.262.] "(2) The welding power by ordinary methods is varied as much by the amount of reduction in rolling as by the ordinary differences in composition. "(3) The ordinary practice of welding is capable of radical im- provement, the most promising field being in the direction of weld- ing in a non-oxidizing atmosphere." SEC. XlXb. Tensile tests on welded bars of steel and iron. A glance at the allowable contents of metalloids, as given in the fore- going synopsis, will show the wide gulf that separates iron from steel, and this will be still further indicated by Table XIX-A, which gives the tensile tests on a series of welded steel bars of different compositions, the investigation having been conducted un- der my own direction. The total lack of certainty and regularity in the results is evident, and it should therefore be said that the smiths were men of long experience in handling steel, and they fully understood that the individual results were to be compared. The bars were of a size most easily heated and quickly handled, but not- withstanding these most favorable personal and physical condi- tions, the record is extremely unsatisfactory. In the case of the rounds, each workman has at least one bad weld against him, while there is only one heat which gave uniformly good results. Picking out the worst individual weld of each work- man, blacksmith "A" obtained only 70 per cent, of the value of the original bar, "B" 54 per cent., "C" 58 per cent., and "D" onlv * The Strength of Wrought-iron as Affected fy/ its Composition and "by its Re duction in Rolling. Trans. A. I. M. E., Vol. VI, p. 101. WELDING. 585 44 per cent. The forging steel showed one weld with only 48 per cent., the common soft steel 44 per cent., while even the pure basic steel gave one test as low as 59 per cent. In some cases where the break took place away from the weld, the elongation was nearly up to the standard, this being true of the four tests of the seventh group, and it should be noted that this metal contained .35 per cent, of copper, but in the other pieces the stretch was low and the fracture so silvery that it was plain the structure of the bar had TABLE XIX-A. Tensile Tests on Welded Bars of Steel and Wrought-Iron. Figures in parentheses indicate that the bar broke in the weld. N=natural bar; W=welded bar. * denotes that elongation is measured in 2 inches. K*ind of steel. Conditions of-test Composition; per cent. N=natural. W= welded. Elastic limit; pounds per square inch. Ultimate strength; pounds per square inch. Elongation in 8 inches; per cent. Reduction of area; per cent. Name of smith. C. Mn. P. S. Cu. Acid O H. forging III! *a " .20 .89 .089 .03 .33 N W W w AV 46670 45890 45580 70450 (60940) (55090) (40840) (42190) 26.25 *10.00 *9.00 *7.00 *3.00 53.50 19.73 7.55 8.12 4.04 A B C D . . . . . . Acid Bess, forging .fill *a * .25 1.86 .083 ' .05 .35 N W W w w 56140 56750 86600 (68810) (55020) (62060) (41930) 22.25 *4.00 *6.00 *5.00 *3.00 85.40 29.29 0.78 6.50 2.10 'A* B c D Acid O H. soft. ifti sr* .09 .46 .08 .35 N W w w w 40980 38230 44660 45030 60680 61060 60380 65610 (26640) 30.00 *60.33 *36.00 '*i2.bo 53.20 56.51 58.50 56.28 4.53 'A ' B C r> Acid O. H. soft. Ifii *8 * .09 .39 .076 . . . .35 N. W w w w 38940 37550 37400 40910 39220 56900 57650 (42740) (43910) 58790 28.75 *39.00 *9.00 *10.50 *34.00 59.89 62.18 13.48 14.55 62.29 'A' B c D Acid H. soft. Iffi *i * .09 .40 .08 .'. . .35 N W w w w 41670 33740 88300 34460 56300 (89490) (30550) 53880 50020 80.00 *6.00 *7.00 *37.00 *16.00 62.56 8.63 10.79 65.46 23.22 ' A' B C D Basic 0. H. soft. li'&s iP* KM .08 .55 .019 .35 N W w w w 33880 87660 35370 31820 51760 58650 (30640) (51850) 49690 82.75 *32.00 *8.00 *27.00 *41.00 65.85 59.55 18.88 46.77 67.85 'A' B c D . . . Basic O. H. soft. fi'e-2 %$v ^ * .06 .30 .014 .35 N W w w w 32580 41930 a5470 38280 89720 48990 64530 52100 54200 55110 81.75 *36.00 *39.00 **41.00 71.56 66.68 70.81 72.81 70.01 'A' B C D 586 METALLURGY OF IRON AND STEEL. TABLE XIX-A. Continued. Kind of steel. Conditions of test. Composition; percent. N=natural. W= welded. Elastic limit; pounds per square inch. Ultimate strength; pounds per square inch. .J ll s! li C o od S"" 4 Reduction of area; percent. Name of smith. C. Mn. P. S. Cu, Basic 0. H. soft. If si *r .08 .50 .027 .35 N W W W w 39820 37330 40880 44510 62000 (49210) 69460 68380 (55560) 80.00 *9.00 *30.00 '*i7.'od 55.96 8.22 48.15 48.54 A B C D Acid Bess, soft. I I! I s .06 .36 .032 .054 .C9 N W 40780 42780 59140 60560 29.50 7.50 47.65 21.60 .06 .40 .032 .054 .69 . . . N W 42020 45150 61370 65780 25.00 8.50 46.89 24.78 .06 .45 .032 .054 .C9 N W 40740 46720 60730 58540 26.25 5.00 46.72 19.48 . . . .06 .86 .032 .054 .69 N W 42680 43350 60780 48740 28.75 1.25 47.23 20.20 . . . Basic O. H. soft. 2x%-inch flats; scarf v/eld. .08 .17 .008 .016 .10 N W 30300 31690 -45070 43290 39.00 11.25 69.70 42.16 .11 .32 .011 .029 .08 N W 33600 50190 45900 83.75 8.50 58.48 34.11 .11 .32 .006 .018 .11 N W 35730 32120 49580 45280 83.00 10.00 56.92 22.18 .09 .29 .005 .021 .10 .N w 36390 37400 50050 45280 83.00 7.50 59.82 41.08 . . . Basic 0. H. soft. la 1 f .12 .13 .36 .005 .022 .08 N .W 34580 30840 51080 41600 28.50 7.50 48.63 26.34 . . . .39 .005 .025 .10 N W 85470 50770 87000 83.75 7.50 51.50 29.88 .12 .29 .005 .016 .10 N W 86830 33300 51300 43530 81.25 7.00 52 62 29.31 .12 .51 .005 .021 .09 N W 37650 35200 64770 48280 26.25 7.00 41.94 21.74 Wrought- iron -. ar3 *i s M N W W W w w 33390 32950 34060 82700 32040 32760 50080 89320 40620 45140 44730 88430 23.50 6.00 6.25 11.75 11.00 4.00 27.26 15.52 22.26 20.98 19.25 i.3fi ! ; '. been ruined. In most cases where the test-bar broke in the weld, the pieces parted at the surfaces of contact, showing that no true union had taken place ; one or two fractures were homogeneous, but they showed the coarse crystallization that follows overheating. The lap welds represent the method ordinarily used in making pipe, and are really a better criterion of the welding quality of the steel than the round pieces, for in making the union the pieces were WELDING. 587 simply laid together with no upsetting, and hence there was less chance for the manipulations of the smith. All of this steel, both Bessemer and open-hearth, had been pronounced suitable for the making of pipe, although it will be a revelation to most metallurg- ists that such a high content of copper could possibly be allowed. In all cases the bars broke across the weld with a more or less crys- talline fracture, there being no instance where the separation was at the plane of union, so that, while thorough welding was proven, it was also evident from the lessened ductility that the metal was overheated during the operation. TABLE XIX-B. Welding Tests by the Eoyal Prussian Testing Institute. Kind of metal. Ult. strength; pounds per square inch. Per cent, elonga- tion in 200 m. m. =7.87 inches. Per cent, reduc- tion of area. fa Av. 9 tests, welded. 2*3 X 7t > rt Av. 9 tests, welded. Av. 6 tests, natural. Av. 9 tests, welded. Medium O. H. steel . Soft O. H. steel 72110 64570 57890 41820 45800 47080 20.8 25.1 22.2 8.2 5.1 7.7 84.9 44.7 89.5 4.5 10.5 14.0 Puddled iron The figures on the iron bars show that the situation is no better than with steel, for the welded bars are far inferior to the natural piece both in strength and ductility. The general truth of these experiments is corroborated by Table XIX-B, which gives a con- densation of the results on a series of tests made by the Eoyal Prussian Testing Institute, the data being translated into Ameri- can form.* The average tensile strength of the welded bars of medium steel was 58 per cent, of the natural, the poorest bar showing only 23 per cent. In the softer steel the average was 71 per cent, and the poorest 33 per cent., while in the puddled iron the average was 81 per cent, and the poorest 62 per cent. The complete destruction of ductility is conclusively shown in the case of all three metals, even the wrought-ir6n being hopelessly wrecked. As above stated, the flat bars given in Table XIX-A were such as had been used successfully in making pipe which would stand * Journal L and S. I., Vol. I. 1883, p. 425, et seq. 568 METALLURGY OF IRON AND STEEL, all the ordinary tests of distortion, while the soft basic metal, made to fill the stringent requirements of the United States Government, would meet the most severe tests. Such metal is used regularly in certain branches of manufacture where the best welding qualities are required, and the users are firmly convinced that "the weld is perfect/' It may be possible to produce better results by special arrange- ments, but it must certainly be acknowledged that a weld as per- formed by ordinary blacksmiths and by the usual methods on the best metal whether iron or steel, is not nearly as good as the rest of the bar; and it is still more certain that welds of large rods of common forging steel are entirely unreliable and should not be employed in any structural work. Electric methods do not offer a solution of the problem, for during the process the metal is heated far beyond the "critical temperature of crystallization, and only by heavy reductions under the hammer or press can much be done toward restoring the ductility of the piece. In many cases this subsequent hammering is impracticable owing to the consequent deformation of the piece. SEC. XIXc. Influence of the metalloids upon the welding prop- erties. The way in which the impurities of the metal affect the welding power has been a matter of discussion, it having even been supposed that they act simply by interposition, and, again, that they increase the susceptibility of the iron to oxidation. I believe both of these theories are wrong. If the first were tme, then one per cent, of carbon would have the same effect as one per cent, of sulphur, which is manifestly not the case. The second theory does not hold, since sulphur, which is notoriously one of the worst enemies of welding, is not oxidized either in the acid Bessemer or open-hearth furnace, and there is no ground for assuming that it oxidizes in welding. It will also be seen that as phosphorus, car- bon and manganese protect iron from burning in the Bessemer and open-hearth, so they must also tend to be preferentially oxidized in a blacksmith's fire, and thus by preventing the formation of iron oxide, as well as by the formation of a liquid flux containing phos- phoric acid and oxide of manganese, they should, as far as oxidation is concerned, assist rather than retard the welding. A third theory is advanced that the impurities affect the mobility. When half of one per cent, of carbon is added to the metal, it pro- duces a compactness or hardness, even when the steel is hot, that WELDING. 589 must prevent the easy flowing together that follows a pressure upon two pieces of white-hot wrought-iron or soft steel. A higher tem- perature cannot be used, because every increase in carbon reduces the safe working temperature at the same time that it increases the stiffness. This decrease in mobility doubtless plays an important part in the explanation, but I believe that a greater influence is to be found in what may seem at first sight to be the same thing, but which in reality is a different quality, viz. : The power, or property, of pass- ing through a viscous state on the road to liquidity. There are other metals, lead and copper for instance, which are malleable and ductile, but which do not go through a history of slow softening under the application of heat, the change to a liquid state being sudden and without any marked intermediate stage. Pig-iron is of the same character, for no matter how low the other metalloids may be, the presence of three per cent, of carbon produces a metal which changes suddenly from a solid to a liquid state, and it is reasonable to suppose that each increment of carbon, phosphorus and manganese tends in the same direction. In addition to this effect, I believe that an equally important factor exists in the action of carbon, phosphorus, sulphur and cop- per in destroying the quality of cohesion by increasing the tendency to crystallization, for it is well known that these metalloids lower the point at which the steel becomes what is incorrectly, but quite naturally, called "burned/' When the steel is overheated it crumbles under the hammer, and it is plain that it cannot be easily united to another piece when it is incapable of remaining united to itself. This theory also explains what seems to be a fact, that a small proportion of manganese aids in welding, for although it does decrease the mobility at any particular temperature, it allows a higher heat to be put upon the metal without the creation of a destructive crystallization, and thus indirectly renders possible a greater mobility and maintains a more favorable internal mole- cular structure. The following conclusions summarize what has just been given and seem to fit the theory and the facts : (1) With the exception of manganese in small proportion, the usual impurities in steel reduce its welding power by lowering the critical temperature at which it becomes coarsely crystalline. 590 METALLURGY OF IRON AND STEEL. (2) A small content of manganese aids welding by preventing crystallization. (3) Only the purest and softest steel can be welded with any reasonable assurance of success. (4) The confidence of a smith in his own powers and his belief in the perfection of the weld, is no guarantee that the bar is fit ta use. CHAPTER XX. STEEL CASTINGS. SECTION XXa. Definition of a steel casting. Within the last few years steel castings have come into general use in the structural world, but there is still a lamentable ignorance concerning their na- ture. A steel casting by very definition must be made of steel which is cast in a fluid state into the desired shape. This leaves open to discussion the great question considered in Chapter IV as to what is included in the term "steel" but although the making of a general definition is complicated by the possibility of producing "puddled steel," there is no necessity of introducing this qualifica- tion into remarks on castings, since fluidity is an essential feature. As for the distinction between "steel" and the so-called "ingot iron," it is needless to say that endless confusion would be intro- duced in the trade if the soft products of the open-hearth were to be styled "iron castings." Notwithstanding the plain limits which have been set by metal- lurgy and common sense, there is a cloud of error hanging around the term "steel castings," which is due partly to ignorance and partly to deliberate fraud. It has been the practice of some persons to make castings from a mixture of pig-iron and steel melted in a cupola, although every metallurgist and every foundryman of in- telligence knows that the metal is altered very much by remelting, and that the changes in silicon, manganese and carbon depend on all the varying and uncertain factors of temperature and exposure. In melting ordinary pig-iron, the carbon usually changes very lit- tle,, for, by the nature of the case, the content of this metalloid was adjusted in the blast-furnace to about the absorptive capacity corre- sponding to the manganese and silicon, and as the conditions in the cupola are similar to those in the blast furnace, it follows that a metal which is the normal product of one will not be fundamentally altered by passing through the other. But a mixture of steel and iron is not a normal product of any 591 592 METALLURGY Of IRON AND STEEL. furnace, and in its treatment in the cupola there is a tendency to make radical changes in the composition by the absorption of car- bon. Thus, by the unnatural union of pig and scrap, and by the uncertain changes in silicon, manganese and carbon, there is pro- duced a hybrid metal which is useful for special purposes, but which is fundamentally different from any kind of steel. It is true that scrap and iron are melted together to make open-hearth steel, but this is done under an oxidizing flame and, either during the melting or afterward, the metalloids are almost entirely eliminated, giving a definite starting point from which a known and regular metal can be made by th.e addition of proper recarburizers. Sometimes castings of cupola metal, made either with or without scrap, are heated in contact with iron oxide in order to burn the contained metalloids. The product is a more or less tough metal, known as malleable iron, which is extensively employed in making small, thin, or complicated shapes that could scarcely be poured in steel, but which can be made of the more liquid iron. The attempt has been made to call these "steel," and the claim has been fortified by analyses showing that the composition resembles that of some steel. The argument is too shallow for consideration, since, on the same basis, the product of the puddle furnace or the charcoal bloom- ary might be termed "mild steel." Malleable iron must always be inferior to steel, because any oxides of silicon, manganese, phos- phorus or iron which are formed remain diffused throughout the mass, thereby breaking to some extent the bond of continuity. Such castings are useful in a certain field, for they are far tougher tl\an cast-iron, and they may even enter into competition with steel castings, but they must always bear a different name, since steel castings must necessarily be made by pouring into finished shape the melted product of a crucible, a Bessemer converter, or an open- hearth furnace. SEC. XXb. Methods of manufacture. The crucible process is sometimes employed for small castings, since the conditions of the "dead-melt" give a much more quiet metal, evolving less gas in con- tact with cold surfaces, and the casting is more apt to be free from blow-holes. In certain special cases, as in the manufacture of big guns at Krupp's, the crucible has been used in making large masses of metal, but its great cost must prohibit its adoption for general structural work. The Bessemer has been used to some extent in the past for mak- STEEL CASTINGS. ing steel castings, but it is utterly unfitted for the work on account of the great cost of the operation when only two or three heats are required during the day. One way of obviating this is by taking an occasional heat from a Bessemer plant which is running regu- larly on other products, but this supposes, what is seldom the case, that the mixture is low in phosphorus. The day has passed away when a casting could be made of ordinary steel, and as it is now necessary to make a careful selection of the stock so that the content of phosphorus shall not exceed .04 per cent., the melting furnace is the cheapest as well as the most efficient instrument of production. Within the last few years there has been a revival of Bessemer castings due to special developments along certain lines of pro- cedure which have been practiced in exceptional cases for many years. After the drop of the carbon flame, a certain amount of melted ferro-silicon is added to the bath and the blowing resumed. The silicon is oxidized and produces a very high temperature, and the advocates of the small converter lay great stress upon this fea- ture. Thus in an article in the Iron Age, June 5, 1902, a writer who claims to be skilled in open-hearth practice, states that the small converter will make a steel containing only .10 per cent, of carbon, with a trace of manganese and silicon, while "an open- hearth furnace cannot make this grade at all and it could not be kept liquid in the ladle." This is a complete mistake, for several open-hearth plants have made large quantities of such metal. In fact, it is not up-to-date to talk about a steel containing .10 per cent, of carbon as being extra soft, for The Pennsylvania Steel Co., as well as other works, stands ready to deliver .any amount of blooms or billets with carbon below .04 per cent, with a trace of manganese and silicon and low in phosphorus and sulphur. It has been deemed necessary to refer to this communication be- cause it is the most recent, and because it is characteristic of a thousand similar advertisements continually appearing in the news columns of the technical press. It is essential to keep in mind that there is no difficulty at all in a good open-hearth furnace in making steel just as hot as can be wanted; in fact, considerable care must be exercised to keep the metal from being too hot. On some kinds of work an excess of temperature may not cause trouble, but in other cases the open-hearth furnace offers far better opportunities for that complete control of temperature and casting conditions which is so desirable and so essential. 594 METALLURGY OF IRON AND STEEL. The open-hearth furnace also allows more perfect control over the casting conditions. A basic hearth is sometimes used and has an advantage in the ability to make low phosphorus without much extra cost, but basic metal seems to be more "lively" in casting, and hence there is greater danger of honeycombs. It is, however, a fact which is worth a hundred arguments that basic furnaces, both in this country and abroad are making good castings, -and it is econ- omy to do so when there is a radical difference in the cost of the raw material. SEC. XXc. Blow-holes. The use of good stock determines to a great extent the nature of the product, but it does not in the least influence the solidity of the castings. This depends partly on the temperature and composition of slag and metal before tapping, and partly on the quantity and nature of the recarburizing additions. An increase in these latter agents covers up the errors in furnace manipulations, but shows itself in a higher content of metalloids. Honeycombed metal may arise from bad casting conditions or it may come from a laudable desire to reduce to the lowest possible point the proportions of silicon and manganese, for the manufac- turer well knows that the blow-holes decrease only slightly the strength and toughness of a casting, while the complete removal of them by overdoses of metalloids gives a brittle metal. It is the current impression that during the last few years all the difficulties in making sound castings have been completely over- come by the introduction of metallic aluminum and certain alloys of silicon. It is true that great progress has been made, but there is no magic wand for sale which can be waved over a ladleful of steel to "kill" it "dead." Hadfield,* in an able article on the use of aluminum, says : "There is no rapid or royal road to the pro- duction of sound steel castings; this is only attained by long ex- perience combined with specialized knowledge." . Some engineers specify that the cavities shall not exceed a certain percentage of the total area, but the common-sense method is to clothe the inspector with discretionary power, for a flaw may be perfectly harmless on the under surface of a base-plate when it would be fatal in the rim of a wheel. In this connection it should be noted that there is a radical difference between a "blow-hole" and a "pipe." The cavities which may often be seen where the "sink-head" or "riser" is cut off, are not evidence of unsoundness * Aluminum Steel. Journal 7. and S. I., Vol. II, 1890, p. 174. STEEL CASTINGS. 595 faut exactly the opposite, for they show that feeding has continued after the riser was exhausted, and that the hidden interior has been rendered solid at the expense of the visible surface. SEC. XXd. Phosphorus and sulphur in steel castings. In writ- ing the specifications for steel castings, the most important point is to state that phosphorus shall not exceed .04 per cent. An excess of the other elements may be guarded against by requiring a proper ductility, but phosphorus, although influencing to some extent the ordinary testing history, is often masked by other factors, and mani- fests itself only at a later time in that brittleness under shock which is its inherent characteristic. This is an important matter in the case of rolled metal, but it is of much more vital moment in steel castings, for these will generally fail, not by being pulled and stretched to destruction, but by sudden strain and shock. The content of sulphur is of little importance to the user, for it affects the cold properties very slightly, but it will do no harm to specify that it shall not be over .05 per cent., good castings generally containing less than this proportion. Copper need not be men- tioned, for there is no evidence that it has any influence upon the finished casting. SEC. XXe. Effect of silicon, manganese and aluminum. The elements used to procure solidity are silicon, manganese and alumi- num. Their value to the steelmaker is due in great measure to their power of uniting with oxygen, the action being as follows : 3.44 parts manganese unite with 1.00 part of oxygen. 3.44 parts aluminum unite with 3.01 parts of oxygen. 3.44 parts silicon unite with 3.93 parts of oxygen. Hence the aluminum is three times, and the silicon four times, as efficient as manganese, weight for weight, while they have an additional value from their greater affinity for oxygen, since this enables them to seize the last traces from the iron and wash the bath so much the cleaner. Another function which may play a part in the operation is the increase in capacity to dissolve or occlude gases, and as far as the value of the casting is concerned this will be equivalent to destroy- ing them. It is not known how far this determines the situation, but it is evident that it has no connection with the power to unite with oxygen. It was once thought that aluminum increased the fluidity of steel. by lowering the point of fusion, but experiments 596 METALLURGY OF IRON AND STEEL. with a Le Chatelier pyrometer* gave the same melting point of 1475 C. for ordinary soft steel as for an alloy with five per cent, of aluminum. The tendency of both aluminum and silicon is to make the steel creamy and sluggish ; it is true that such metal will run through small passages without chilling better than ordinary steel, but this is because the latter foams and froths when in con- tact with cold surfaces, and the flow is thereby impeded and suffi- cient surface exposed to chill the advance guard of the stream. The percentage of manganese should not exceed .70 in soft cast- ings nor .80 in harder steels, since more than this may render the metal liable to crack under shock. Silicon can be present up to .10 per cent, in the mild steels and .35 per cent, in the hard without any appreciable diminution in toughness. Aluminum is seldom present except in traces, and should not be over .20 per cent., for it decreases the ductility. The carbon must vary according to the de- sired tensile strength and the use to which the casting is to be put. When it is over .70 per cent, the steel becomes so hard that machin- ing is slow, and there is danger of lines of weakness from shrink- age in complicated shapes. SEC. XXf. Physical tests on soft steel castings. Since the fail- ure of cast-work is almost always due to sudden strain, it is the safer plan to have the metal for common purposes between .30 and .50 per cent, in carbon, but when great toughness is required it should not be over .15 per cent. This latter specification also pre- supposes a low content of manganese, silicon, and, above all, of phosphorus ; with this composition the casting displays all the char- acteristics usually associated with the toughest of rolled shapes. A test on an unannealed gear-wheel of such metal, manufactured by The Pennsylvania Steel Co., was made by cutting the rim between the spokes and then bending one arm to a right angle, twisting another through more than 180 without sign of fracture, while a third was hot-forged from a star-shaped section of about 2 inches by iy 2 inches into a bar 1*4 inches by three-eighths inch, and after being cooled was twisted into a closed corkscrew. Similar pieces were exhibited by Krupp in his magnificent exhibit at Chicago, but we stand ready in America to duplicate any such metal x>n regular contracts. Such trials, made on castings taken at random, are far preferable * See article on Pyrometric Data, by H. M. Howe, Engineering and Mining' Journal, October 11, 1890, p. 426. STEEL CASTINGS. 597 to tensile tests from sample bars, since the small pieces will not be in exactly the same physical condition as the larger castings. The results have a certain value, however, and avoid the necessity of" spoiling good finished work. It is well to keep in mind that a flaw or blow-hole in the small test does not necessarily imply that the casting contains similar imperfections, and also that while an open cavity, however small, which is visible on the surface of a machined test will have a disastrous effect upon the strength and ductility, it might be of slight importance if buried in the interior. This neces- sity of having a perfect surface makes it difficult to conduct a long series of tests with exactly the same dimension of test-pieces, for if five-eighths inch in diameter is the desired size, it may be neces- sary to turn some of the pieces to one-half inch, while the length must sometimes be reduced to 6 or 4 inches. It is also a strong argument against the use of an 8-inch test piece, for the chance of pinholes and a consequent bad record is thereby multiplied four- fold when the presence of such holes has practically no effect upon the casting. This test piece should not be annealed unless the castings them- selves are to be treated in the same manner, and although it is cus- tomary to anneal most structural work, the trouble is not necessary in a great many cases if the very best of stock is used. This state- ment will be called heretical by many engineers, but the tests that have just been recorded upon an unannealed gear-wheel will show that the metal can be exceptionally tough in its original state. In the case of castings of complicated shape and those exposed to shock, annealing should be specified, but it must be remembered that there is no magic charm in this word. It is not sufficient to simply say that they shall be annealed and make sure only that they are covered with soot or fresh oxide. The heat treatment of steel is no longer a mere heating to remove strains, with the hope that some unknown change may occur to toughen the mass ; it is or should bb and always can be a scientific procedure, by which the metal is raised to an accurately determined critical temperature, whereby certain molecular rearrangements occur. If these rearrangements are properly guided, the result will be seen in a fine grained struc- ture and a tough metal. If they are not properly guided the last condition may be as bad as the first. Up to within a few years most steel castings were made of hard metal containing from .30 to .50 per cent, of carbon, and having a 598 METALLURGY OF IRON AND STEEL. tensile strength of 80,000 to 100,000 pounds per square inch, but just as engineers have long since learned that the strongest and safest bridge is not built of rolled steel with .30 per cent, of carbon, TABLE XX-A. Comparative Physical Properties of Bars Cut from Annealed Soft Steel Castings and Unannealed Bars of the same Heats Rolled from 6-Inch Square Ingots, together with Results of Similar Bars made from Large Ingots. Steel manufactured by The Pennsylvania Steel Company. jd * Si 3 C rrt I a a Composition; percent. S,| ill iP ftw w C*^. CD CD ction of j cent. 49 - o s tr 2 3 C! G o o 3 H C. P. Mn. S. pp, M s a " H c '"' I* 3552 .17 .027 .65 .034 58190 34290 24.00 82.1 8555 .17 .027 .66 .056 56030 32440 14.90 19.7 8557 .17 .032 .60 .029 55880 32750 27.13 42.3 8559 .17 .027 .65 .038 55350 30350 23.10 42.5 8563 .17 .024 .62 .024 59390 34790 20.10 84.5 8565 .23 .029 .65 .025 60060 33130 20.65 26.8 8568 .14 .029 .70 .032 58320 31750 17.25 20.8 8571 .18 .033 .58 .028 56700 30670 26.88 46.7 8573 .17 .028 .67 .027 57440 31430 21.66 36.7, 8573 .17 .036 .70 .027 68860 34260 22.04 29.8 8577 .17 .037 .59 .029 57980 33220 23.00 39.3 8578 .17 .045 .67 .026 58810 33510 22.16 30.4 8579 .15 .037 .63 .028 54940 32190 22.75 47.0 8580 .18 .038 .71 .017 68970 34180 22.25 36.7 8582 .17 .036 .63 .024 56380 31520 13.00 25.5 8583 .18 .032 .61 .022 59400 35330 14.13 18.8 8584 .18 .027 .60 .027 55970 29690 22.38 32.1 8586 .17 .027 .60 .027 55630 30300 18.50 31.4 8588 .16 .043 .63 .031 56950 32530 26.50 42.7 8592 .18 .027 .69 .028 59050 32940 20.00 33.0 Average of annealed .17 .032 .64 .029 57515 82564 21.12 83.44 east bars. 2x%-inch bars rolled from 6-inch square ingots cast from the same heats and tested in 63523 42700 24.74 43.80 natural state Average of 2x%-inch bars rolled from 4-inch billets made from 16- Natural 62089 42441 30.14 60.86 Inch ingots of 7 different heats of Annealed 65021 31576 80.36 60.00 about the same tensile strength as the above castings so they must learn that in still greater measure it would be better to use a softer metal in castings.. Table XX-A gives the results of tests made on sample bars of cast steel, showing the composition and physical qualities. STEEL CASTINGS. 599 The silicon is not given, but it was below .05 per cent, in every case. The test piece was not cut from the casting itself, but from a small coupon which is much more likely to contain blow-holes, and this will explain why it was often necessary to pull the piece in a six-inch length. The test was round in every case, and there- fore gave slightly worse results than a flat, but this is far from explaining the great inferiority of the casting when compared with the preliminary test, or the much more marked difference from what should be expected in properly rolled steel of similar tensile strength. TABLE XX-B. Physical Properties of Annealed Bars cut from Castings of Me- dium Hard Steel ; all Bars %-inch in Diameter. Manufactured by The Pennsylvania Steel Company. Heat number. Composition; percent. ength; square S 43 a gj 1 jj" H Ultimate sti pounds pei inch. Elastic limit pounds pej inch. H" Reduction oJ per cent. C. Mn. P. S. Si. 921 .20 .54 .026 .022 .30 60580 60680 60830 61480 62420 83710 82380 32750 30740 82460 30.50 86.50 86.00 82.00 88.00 88.57 51.90 44.34 89.80 50.90 55.6 53.4 63.8 50.0 52.0 953 .22 .56 .035 .034 .30 63320 64880 65500 65845 65930 67010 37400 34170 44850 83595 82290 48630 36.00 24.50 29.00 26.00 80.00 26.00 46.33 28.57 89.40 32.40 83.37 82.40 59.1 52.7 68.5 51.0 49.0 72.6 974 .38 .75 .029 .023 .35 72630 75240 44940 45880 16.00 23.00 20.70 81.63 61.9 61.0 966 .35 .68 .038 .034 .34 73090 75160 45390 45510 17.50 20.50 21.25 27.64 62.1 60.6 The results show what has so often been mentioned in these pages that the ultimate strength and elastic limit are altered very little by the amount of work upon the piece as long as it is not finished at a low temperature. Thus, in the annealed casting the elastic limit is 56.62 per cent, of the ultimate strength, while in the -annealed bars rolled from the ingot it is _57.39 per cent. This approximation is remarkable because the factors relating to duc- tility show that the physical state of the two metals must be radic- different. 600 METALLURGY OF IRON AND STEEL. SEC. XXg. Physical tests on medium hard steel castings. It has just been shown that the average elastic ratio in annealed cast- ings is about the same as in annealed rolled bars, but there will be much greater variations between individual tests in the case of the unworked metal owing to local imperfections, and there will also be greater variations with a stronger steel. This will be shown by Table XX-B, which gives the results on duplicate bars from four different heats of harder metal. It will be seen that the ultimate strength is fairly regular, and this indicates that the metal itself is homogeneous, but that minute imperfections give rise to the variations in the elongation, reduction of area, and elastic ratio. In the body of a casting these defects exert little influence, but they seriously affect the integrity of a small machined piece. This will emphasize the statement already made that the safest way, whenever practicable, would be to make a drop test on a sample casting rather than to cut a small bar from the piece or from a separate coupon. PART III. The Iron Industry of the Leading Nations. CHAPTEE XXI. FACTORS IN INDUSTRIAL COMPETITION. NOTE. In the summer of 1899, I visited many of the large steel works in Eng- land, Belgium and Austria, and most of the large plants in Germany. I was received everywhere with unvarying courtesy and hospitality, and was given every facility to inspect methods and results. I trust that nothing here written will be construed to be more than fair scientific criticism of my hosts. SECTION XXIa. The question of management: It is a common thing in America to smile over the non-progress- iveness of our foreign friends and to congratulate ourselves that we are not as other men. There are many people here who believe that foreign engineers are not quite up to our standard and that we are especially commissioned by Providence to illuminate the whole 'world with our spare energy. I will take away no glory from my fellow countrymen. They need no spokesman and they will be sure to get all that is due them, as the progressiveness of American metallurgists and Engineers is well known in foreign lands, but it is well to remember that there is one vital difference between metallurgy abroad and metallurgy here. The direct management of a steel works in America has practically its own way. If a mill is out of date and a new system of rolling or manipulation is needed, it does not take long to get authority to make the change. It is called extraordinary repairs, it is called improvement, or it is not mentioned at all. The directors leave much to the management; they feel that they pay men to attend to the operation of the works and constant improvements are looked upon as necessary and in- evitable. As for the stockholders, they are not considered, for a stockholder in America is not supposed to rise in meeting and question the wisdom of spending any reasonable sum upon improve- ment and then find out whether the improvement is paying for itself. In England, especially, the very reverse is the case. The stock of many of the older steel works is very widely distributed and a large number of shareholders do not know anything about improve- 604 THE IROTs T INDUSTRY. inents and do not care. They want their dividends, and if any money is taken from profits and spent on new machinery, it must be fully explained why this was done, and it must be shown that this expenditure has been justified by results. If American man- agers had to go through such an inquisition whenever they proposed an improvement, and if, on the other hand, they could satisfy the shareholders by inventing nothing, it is possible they would lead a less strenuous life. According to the usual financial custom in England, no improve- ments are made out of profits, new capital being authorized and obtained for this purpose when deemed necessary. There are many exceptions to this system, but it is certain that it is quite generallv carried out to an extent that will hardly be credited by Americans. An instance may be cited of an English works in South Russia managed entirely on English lines. The capital stock is $6,000,- 000, and in the last eleven "years annual dividends have been de- clared ranging from 15 to 125 per cent, and averaging over 32 per cent. In 1900 the disbursements were only 20 per cent., or $1,200,- 000, the decrease being due to low prices and bad markets. Just at this juncture it is found necessary to build certain railway lines, etc., and an issue of bonds is made of $750,000 to pay for this, just as nearly double this sum goes out in dividends. Nothing can con- vince our friends on the other side of the water that this is any- thing but the true and only right way, just as it would be im- possible to convince men here that it was anything but wrong. The English manager has also to contend against very strong labor organizations, their ignorant and tyrannical control being a hopeless bar to the progress of English industries. There was a time when such societies regulated affairs in many American steel works, but it was soon discovered that progress and labor organiza- tions do not sail in the same boat. This has lately been discovered in England, but it is not easy to fight against established customs. In the summer of 1899 I visited the Cleveland district of England. Everything indicated prosperity in the iron trade and new work was underway that would add to the output and general business of the place. The firm in charge of this new plant stated that their boiler-makers and riveters would work only three days in the week. Their wages had been advanced to offer extra inducements, but this did not help matters in the least, for by working hard and well during Thursday, Friday and Saturday these riveters were able to FACTORS IN INDUSTRIAL COMPETITION. 605 earn over seven dollars per day, or twenty-two dollars in the week. When work ceased on Saturday evening a drunken carouse began, which lasted until Wednesday night. In a short walk through tho streets of Middlesborough on Monday forenoon I found at least a dozen men lying drunk upon the sidewalk a condition which can- not be paralleled in any American city. It is impossible to reform. this state of affairs since the unions control the entire situation and do not consider any offense of this kind as ground for a discharge, and a strike would follow any attempt to interfere with the God- given right to get drunk once a week. All over England Blue Monday is something more than a name ; it is a costly factor in all industries, standing side by side with the tyranny of the labor unions that are fighting with bulldog obstinacy against improve- ments that would ultimately be for the benefit of all concerned. We have had much of this sentiment brought to America by the English and the Welsh, but although they have acquired much power in certain localities and in certain trades, they have never been able to control the whole American iron industry in the way they con- trol the great English producing districts. tt In England there is a tendency to have the management of an enterprise descend from father to son, and this transference of power is often gradual, the son being perhaps the assistant of his father for many years. It is evident that such a system tends to conservatism and the perpetuation of old conditions. This tend- ency is, of course, acccentuated by the general sentiment of the country to bow to the opinions of older men and accept their de- cision as final. In America, we bow to the decision, but we reserve the right to differ. The conservative influence of a management controlled more di- rectly by the stockholder and by family and local traditions must inevitably result in retarding the advancement of progressive young men in the establishment, and in pushing forward the conservative element, so that the man who finally reaches the top of the ladder will be more often a conservative than in America where progress- iveness must be shown before promotion is possible. Add to all this the continental opposition to change v^aich is especially marked in England, the magnifying of every cusiom and tradition into a law of nature, the opposition to self-evident improvements from a simple disinclination to be different from others, and it is easy to 506 THE IRON INDUSTRY. understand why our European friends do not move as fast as we do in America. Thus we have proved why, in many respects, our friends across the water do not keep pace with us in the race,|but it remains to explain why, in many respects, they are ahead. It is not necessary to discuss the development of the Bessemer and open-hearth pro- cesses, because when these methods came to light the iron industry in America was a small affair compared with the old established plants of Europe; but in the manufacture of coke, for instance, Germany has been using the retort oven for twenty years, while America has just discovered its existence. England was very slow in adopting it, owing to the opposition of venerable authority, but our country is the most behindhand in accepting the benefits of the invention. In the matter of gas engines driven by blast furnace gas the Con- tinent has completely distanced us. Engines have been operating successfully in France and Germany for four years, while long before this, Eiley, at Glasgow, in Scotland, put in operation an en- gine built by Thwaite which has been running since 1895. Im- mense machine shops all over the Continent are busy turning out engines whose aggregate horse-power runs into many thousands, whereas in America nothing of any consequence has been done in this direction. Much of this forwardness in so using blast furnace gas arose from the fact that gas engines driven by illuminating and producer gas have been used much more extensively abroad. The visitor to any English city is struck with the puffing of these engines in the lumber yards and the cellars of numberless work- shops, while in America this economical motor is little known. In another respect the European works are ahead of America and that is in the use of the unfired soaking pit. This practice is almost universal on the Continent and is common in England, while in America it has been a failure. This arises from the curious fact that acid steel does not give good results when heated in unfired pits, and the Bessemer plants in America make acid steel only, while almost all the plants on the Continent are basic. Moreover, in America it is the rule that nothing must interfere with tihe regular sequence of operations, and that if anything happens to the tools, nothing else shall be behindhand when they are ready. It is evident that a gas fired furnace is more thoroughly under control, and capable of holding ingots ready, than one which has no supply of heat v The unfired pit FACTORS IN INDUSTRIAL COMPETITION. 607 is simply one example of a very important truth, which may be stated thus: a method or device or improvement which is voted a success in Europe will oftentimes be voted a failure in America if it gives the same results. In other words, we will not bestow upon it as much intelligent care as our German friends, and will not consent to the delays and interruptions which they regard as of little importance. This statement may be questioned by some per- sons, but there are many engineers of prominence and of experi- ence who will agree with me in this broad statement. America has developed along its own lines, but the lines on which England and Germany have worked have not been as capable of rapid development. In the Bessemer process England has faced a continually lessening ore supply, decreasing both in quantity and quality and increasing in price. Immense progress could hardly be expected under such circumstances. Germany has had to adopt the basic vessel almost exclusively and has been much more successful with it than any works in America. In rolling mills, our friends across the ocean have used generally the two-high reversing mill, and it is quite evident that the possibilities of expansion in amount of product with this system are less than with the three-high train. This one item, the capacity for expansion, is the great dividing line separating European and American practice and the reasons for the difference are not thoroughly appreciated. Taking the case of railroads for an illustration, we have on this side of the water a new country. The lines that form a network all over the western half of our domain and most of the lines in the older half have been built and equipped within the memory of young men. There were no old and obsolete patterns to copy. The new roads began with apparatus which to a greater or less extent was standard as far as America was concerned. The style of rail was practically uni- form with all roads and the amounts ordered by many railroads were enormous. One railroad would order fifty thousand and an- other sixty thousand tons for delivery in one season, and all rails were so nearly alike that it usually sufficed to change one set of rolls to go from one order to another. The differences in sections that did exist arose from a desire of the railroad engineer to experi- ment and get a better rail, although this sometimes resolved itself down to an egotistical desire to have his name associated with a par- ticular design. This latter state of affairs in America seldom pro- duces the kind of glory desired, and within the last few years a (J03 THE IRON> INDUSTRY. concerted action on the part of the steel manufacturers and engi- neers has resulted in the general acceptance of one set of standard sections. In England the conservatism and importance of the railroad engineers render such standardizing apparently impossible. Not only that, but the use of chairs and their associated paraphernalia makes a change very expensive, while the much smaller mileage of their lines, due to geographical limitations, makes it impossible to have a very large order either for replaceals or new track. ' One road at least in the British Isles with a high sounding name is only two hundred miles long, and it is laid with half T rails and half bullheads, and in order to make renewals it is necessary to order two kinds of rails, splice plates and accessories, and each order will be just half what it would be if the road were laid with one kind of section. Needless to say such a road must inevitably pay a higher price per ton for its rails and fastenings. I.n America it is now possible to keep in stock the fish plates for the standard Ameri- can Society sections. The rolls do not have to be put in for a few tons and taken out and put away for a year. The railroad engineer may think this matter of roll changes is no concern of his, but it is his concern and his railroad must pay the bill in the end, and if the English railroads would unite on certain standard sections of rails and joints and abolish and forget certain details of inspection and testing that have come down from the dark ages and are perpetuated by the red tape of Boards of Trade, they would get their material at a much lower price, they would get it in half the time, and it would be just as good. The responsibility for these conditions, however, does not fall upon the English steel works. They have had to meet certain business demands and they and the railroads are prevented from making any changes by the regulations of the Board of Trade, this latter institution being practically a govern- ment commission whose hands are tied by Parliamentary legislation. All experience proves that progress is either slow or impossible when a legislative body ha? to be moved. It is also to be borne in mind that England cannot extend her domain as the United States has extended, and that the increase in the cost of raw materials seems to put a limit upon future possi- bilities. Under these circumstances, it would have been of doubt- ful expediency to build a counterpart of one of our immense Amer- ican mills, for the total production of steel rails in all En~!nntf is FACTORS IN INDUSTRIAL COMPETITION. 609 only about 800,000 tons a year, while in America a capacity of 600,000 tons is considered about right for one mill. England has, therefore, clung to the two-high reversing engine, which for smaller products has certain advantages. In the first place, it is much better adapted for rolling directly from the ingot, for with a revers- ing mill, the bar can be backed out if there is a tendency to split. It also renders it easier to roll many different sections on one mill in steady operation instead of having several mills each adapted only to its own specialty, and it is also easier to make certain diffi- cult sections on a two-high mill on account of the ability to vary the speed. TABLE XXI-A. Miles of Railway in Operation in 1899. FROM STAHL UND EISEN, JULY 1, 1901. United States ..... ..... ^ 190,360 Germany .............. 31,570 Austria-Hungary ........ 22,670 Great Britain and Ireland J 21,670 Canada ..................................... 17,350 Italy ....................................... 9,830 Spain ....................................... 8,300 Sweden ..................................... 6,700 Belgium .................................... 3,870 Europe except above ........................... 13,730 Asia ........................................ 35,140 South America ................ .......... ..... 28,030 Australia ................................... 14,760 Africa ...................................... 12,570 Mexico, Central America and Newfoundland ....... 9,800 Total ................................... 482,480 This matter of small orders will be better understood by com- parison of the total mileage of railroads in the different countries, as shown in Table XXI-A. The United States has 40 per cent, of all the railroads in the world, Germany coming next with less than 7 per cent., and if we omit those nations that make their own rails and take all the rest of the world, including Canada, the total "markets of the world" do not include as many miles of track as are laid within our borders. Thus if we can assume that Germany, which ranks next to the United States in length of track, should monopolize the rail trade of the world with the exception of the United States, Russia, France, Austria, Great Britain and Belgium, 610 THE IROX INDUSTRY. each of which is self-supporting, she would not have as much trib- utary track as stretches out before the doors of American steel works. These reasons have influenced the development of rolling mills all over Europe, and the newest and most thoroughly equipped plants have not copied from America, but have simply enlarged and expanded the old two-high construction. In this connection, it is worthy of note that one of the newest American rail mills is of this. type. In making the usual sections of structural material and rail- way splices, it is the custom in America to cut the ingot into several blooms or billets and reheat these for finishing, this being done in order that the bloom or billet mill shall run steadily at its maxi- mum capacity. In Europe little thought is given to this argument. The question everywhere heard is this: "What could we do with all the steel if we should run continuously ?" It is therefore much more common abroad than in this country to roll many different sections in one reversing mill, the stuff being finished in one heat from the ingot, the finished bar being very long; in one mill a 2" square billet is finished 475 feet long and a 3"x3" angle 425 feet. Oftentimes the finishing is done on a different mill, and frequently the finishing mill is three-high, the blooms being cut up and trans- ferred without reheating. The Germans use many three-high trains for finishing, and these are often of large size and are run fast. In more than one place 15-inch beams are rolled directly from the ingot without cropping the ends and without reheating, the work being done by hooks and tongs without any machinery except a steam cylinder to raise the swinging support of the hooks used to catch the piece. Such a lift- ing motion is necessary when the rolls are 30 inches in diameter and the mill runs 110 to 120 revolutions. I have seen a mill of this size and speed handling 8-inch blooms weighing about 1200 pounds, and few American workmen would care to work as fast and as hard as these hookers, although American workmen would have smiled at the idea of a man being able to do anything when wearing wooden shoes. In rolling beams by hand in a train of that size an army of men is required, and the average visitor can hardly understand why some simple labor-saving de- vices are not introduced. It is related of an American at a German works that he offered to spend a certain reasonable sum in machinery and save so many dollars every month. The manager FACTORS IN INDUSTRIAL COMPETITION. 611 answered by showing him the cost sheets and proved that the total expenses for labor in the mill did not equal what he proposed to save. Such an answer, however, cannot possibly be true of all places where labor is thrown away. In one of the famous steel works of the world are two blooming mills, three-high, and exactly alike, turning out a combined product of ten thousand tons per month. In America one such mill would take care of from forty to sixty thousand tons per month and two men on each turn would operate it, while in this place it took fourteen men on each mill. The fundamental difference was that the table rollers were not driven, and it would be safe to say that the introduction of ma- chinery to drive those rollers would have paid back the money every three months, to speak moderately. It will not do, however, to suppose that the management was entirely contented with this condition. On the contrary, plans were drawn for an entirely new works, which involved immense expendi- ture of money, and it seemed to be the accepted law not only at this particular works but elsewhere, that an old plant should not be im- proved when a new one was contemplated. The reasons for this are, of course, self-evident and must have force everywhere, but in America such improvements do go on constantly even under ex- actly those conditions, because with our high priced labor and almost unlimited demand for steel, it is often easy to pay for new apparatus in a year, while in Germany with cheap labor and a much smaller product, it would take a much longer time. At another works, in another district, there were four mills under one roof, the building being large enough to cover all the room necessary for handling and shipping the product of all the mills, making it one of the largest buildings in the world. The total output of these four mills was about 400 tons each twenty-four hours. In Amer- ica the same outlay would be expected to produce from five to ten times that amount. This condition, however, is not universal, and the American visitor will find other plants more to his mind. It is impossible to obtain the same output from a basic converter that can be obtained from an acid lined vessel, as the addition of the basic materials, the greater amount of oxidation to accomplish, and the much greater wear of the linings, render it out of the question. Nevertheless there are several German works, among which may be mentioned Eothe Erde, Phoenix, Hoesch and Hoerde, which can make from 612 THE IRON INDUSTRY. 32,000 to 35,000* tons of steel per month from a plant of three basic converters ranging from 11 to 18 tons capacity, and there is no need to say that this cannot be done without an all-sufficient supply of "push." The diversity of product in a German mill and the intermittent work, arise oftentimes from the universal control by syndicates of all the items of production, and there is no incentive to rush out a heavy tonnage for a week and then shut down with an empty order book, but it would seem difficult to ever get a mill up to its maxi- mum output and efficiency with workmen who wear wooden shoes. It would be good business to pay for a leather outfit simply for the moral effect. ^ There are some American writers and metallurgists who ascribe the forwardness of steel manufacture in America to the ingenuity and brilliancy of a little group of men who developed our great plants a quarter of a century ago. It is an unkind act to disparage the work of our predecessors, but I am actuated not in the slightest degree by any personal feeling in expressing the opinion, which is not simply of my own creation, that no one man should be lifted upon a pedestal of fame unless the foundation stones bear the names of many others almost if not quite equal to him in worthiness. It was the custom twenty years ago to pick out as an idol one who could deliver a witty after-dinner speech, and to forget that something was due to the cook. Nothing is easier than to join a mutual admiration society and gradually have every member be- come in his own estimation more and more indispensable to the daily routine of the universe. In my judgment the char- acteristics of American metallurgy have been developed by so many minds that it is invidious to name a few, and these minds were not creators, but creatures; they were carried forward in the flood of "push," which has been and is to-day the predominant feature of our countrymen, who will forgive mistakes if they are made while going forward. Much of the difference between the two sides of the Atlantic is due to the fact that no spirit of rivalry has ever entered into Euro- pean steel works. Men do not go from one place to another; they do not brag of outputs; they do not challenge every one to enter the race. It is beyond question that many of the great advances that America has made have been due almost wholly to vainglory * SchrSdter, private communication. FACTORS IN INDUSTRIAL COMPETITION. 613 and a simple desire to "beat all creation." Another factor was the desire to increase outputs when the margin of profits justified the most lavish expenditure, and it is doubtful if in every case it was foreseen that these outlays would result in such a great decrease in the operating cost per ton. In foreign countries this argument of beating a competitor has absolutely no place. In one of the old works in Germany there are blast furnaces in operation which are only 48 feet high, but as they are running on a fuel consumption of from 1790 to 1840 pounds of coke per 2240 pounds of iron, the management sees no possible justification for destroying existing plant and starting in on new construction involving immense ex- penditure. The facts that the furnaces are out of date, that they are slow, that they are curiosities, have no bearing on the matter at all and are not considered for a moment. In our country we might keep such furnaces if we had no money to build others, but we would apologize for them; we would say they were not worth looking at, but in Germany this sentiment is entirely unknown. It is open to debate whether a little of the foreign spirit would not be as valuable an acquisition for the American, as a little American spirit is valuable for the European. In America we enter contests from the time that we are born, and we always play to win. ljurope does not know this feeling and she will not make two thousand tons of rails in a day from one rail mill until she acquires it. She has engineers who are as bright and smart as any in America ; they are as progressive as any of our nation; they are working along many lines; they are introducing many labor saving devices; they are building mammoth mills and great machine shops ; they are not narrow ; they are copying Amer- ica where America is good; they are filling their machine shops with American tools and they are taking a fresh start. In Austria a grand transformation scene is in progress : a syndicate controlling most of the iron and steel works of the kingdom is dismantling and abandoning most of them and concentrating the work into a few plants, which are being reconstructed and rebuilt with a thorough- ness that we egotistically think is American. In Germany indi- vidual works are doing and have done the same thing; but in Europe these improvements are not always announced to the world with a flourish of trumpets. There is a district in Germany which is said to possess financial advantages over any other, and the question is naturally suggested THE IRON INDUSTRY. why the works at that place do not expand and monopolize the business. There are two answers, each of which is deemed suffi- cient in Germany. First, it would be difficult to get the necessary labor to move to a new place. Second, there is no inducement for the stockholders to spend money, as they are quite satisfied with dividends at the rate of seventy-five per cent., and there is no use of exhausting ore to pay dividends to new capital. These two rea- sons are equally incomprehensible to Americans ; they represent the difference between Europe and America. Each land has much to give to the other. Perhaps we can teach them how to work, but they can teach us how to save up just a little of our surplus energy and use it in enjoying the fruits of labor. SEC. XXIb. The question of employer and employed. This is usually called the "labor question," and is often spoken of in much the same way that the consumption of fuel would be discussed, but although it may be convenient to treat it thus in books, it cannot be so handled in actual life. There are three distinct methods of arranging relations between the employer and the employed. The first is the paternal system, where the employer does everything for the workmen, as at Pullman in our own country, and at Creusot in France. This is probably the worst thing possible and breeds a servile lot of men, whose highest thought is expecting the next spoonful of gruel. It is soup-house charity when there is no neces- sity for philanthropy. The second method is the treatment of men as men. The self- respecting man does not ask charity ; he wishes to pay one dollar for one dollar's worth of goods. There are exceptions to this rule, but there are also many other objectionable people in the world. This self-respecting man should be the one for whom all rules are made and the others may do as they like. This man is a free agent, able to make his own contracts, to work or to leave, and as a rule he generally has a job and is too busy to make speeches on the labor question. The third system is the labor organization where men bind them- selves together and appoint a committee of those who can talk longest and whose duty it is to get all they can for "labor." These unions declare that every man is the equal of every other man when he is not; that a fast workman shall not be allowed to do any more work than a slow workman which would seem to be an attempt to upset the decree of Providence: that a good workman FACTORS IN INDUSTRIAL COMPETITION. 615 shall not receive more than a lazy dummy which is absurd; that labor saving devices shall not be introduced unless the money saved is distributed among the workmen ; and, worst of all, that dealings with the men shall be done through certain intermediary officers, when it is notorious that in some cases the men chosen to such office have gained power by cajolery, bribery and the lowest methods of ward politicians. It must be acknowledged that the same class of men achieve political success in both small and large cities under our system of popular sovereignty, and it would certainly be unwise to change our government in order to prevent the election of demagogues to office ; but it must be remembered that no demagogue nor Board of Aldermen is given authority over the freedom of the individual nor over the operation of great industries. The Czar of Kussia might hesitate to order one hundred thousand men out of employment, and practically expose to mob rule great industrial establishments and ruin the trade of a million people. There is only one power on earth in any civilized land which has such authority, and this is a committee of men chosen by a small fraction of the community and often by a minority of the interested parties. It is of record that the disastrous decisions of such committees have often been condemned by the greater bodies of which they form a part, al- though such condemnation generally does about as much good as an apology for hanging the wrong man. These faults are recognized by the labor unions themselves, and many well meaning persons advocate "compulsory arbitration" as the panacea for all ills ; but it is impossible to see how a manufac- turer can be forced to take orders and to operate his mill if he chooses to shut down. To compel him to do so would be condemna- tion of property, and the slightest consideration of fairness would lead the state or the community to make good any loss he might sustain by the continuance of operations. On the other hand it is impossible to see how a workman can be compelled to work at any wage which is not satisfactory to him, when perhaps he is offered more elsewhere, and no manufacturer would ask for such an unconstitutional infringement upon the personal rights of his workmen. Moreover the labor unions themselves, while anxious for a law to compel employers to abide by an award, recognize the injustice and the impossibility of forcing a workman to labor for less than he considers his due. It would therefore seem that the THE IKON INDUSTRY. best way is the simplest : it is to let each man exercise the rights given him by our laws of working for the highest wage he can get, and of leaving when he is not treated rightly. If a man. is un- reasonable he may be discharged, while if an employer is unjust he will be unable to find laborers to do his work. Under the system of labor unions the men who perform some par- ticular line of work may often be entirely unrepresented on the committee. The works with which I am connected has in opera- tion seven different rolling mills and each one is essentially differ- ent, both in amount and character of product. In some of these mills there are over thirty different kinds of positions where the men are paid by the piece or ton, not counting the work done by the day or hour, and each of these tonnage positions has a special rate agreed upon. Under any system of committees it is plain that the great majority of positions will have no representative, and that there will always be an incentive on the part of a committeeman to look after his own job and his own friends, while, on the other hand. the management of the works will be only too glad to give such a committeeman anything he may ask if he will agree to a low rate for those who are not present at the conference. A few years of such work will generally bring on a strike, and a thousand well- meaning humanitarians will then advocate "arbitration," by which is usually meant a reference to some men who do not know a pair of tongs from a straightening press, and who probably will recom- mend that the difference be split, the whole question of dispropor- tionate rates being left exactly as it was. To what extent this dis- proportion can obtain has been shown by sworn testimony before a Congressional committee, where it was proved that men who joined the disastrous strike at Homestead drew thirty thousand dollars a year. It might be of advantage for the manufacturer to pay still higher bribes to the leaders of the workmen, since such wages for rollers cannot be called earnings, if it were not for the fact that there is a limit to what the members of a union will stand, for it is necessary to keep in mind the all important point that the action of the com- mittee is not final. The signature of the company hears with it the highest responsibility, but the signature of the committee is worthless. It may or may not be agreed to by the union. bu f whether it is or whether it is not, the decision does not carry with it the slightest financial responsibility. It does not bind and cannot FACTORS IN INDUSTRIAL COMPETITION. 617 bind any individual to work for the company a single day longer than he chooses, and if the industrial situation brightens and men find other more remunerative employments it is the privilege of each and every man to leave, and if they choose to go out on a sym- pathetic strike, as unions have done before, there is absolutely no redress for a violated contract. I do not believe in such inequitable arrangements, nor do I believe in arbitration on many of the questions arising, or in a sys- tem of committees so organized. I believe that each man who thinks himself ill treated should have access to the office of the manager. It is the right of appeal to a higher court and it is my experience that it is the rare exception that a body of men appear to discuss a question unless there is some ground for their action. Investigation generally shows that their statements are correct, and while, of course, the workmen are trying to get all that they can, and while the manager is naturally trying to give as little as he may, it generally happens that the level-headed men lead in the argument, and that a fair and equitable arrangement can be made, and no man feels that he is outwitted by a committeeman. He has stated his case; he has heard the reply; he remains a free citizen to accept the offer or to decline it, and no works can long operate if the offer is not just and right. I do not know whether these rules are universal, and there may be cases where different conditions govern and where large bodies of skilled men of one trade may join for mutual protection ; but in a steel works where hardly any two positions are alike either in nature of work or in rate of pay, the labor organization as at present constituted has no place. Moreover, under no condition will it ever be more than an unworthy and petty factor in the universal labor problem until it gives up once and for all the tenet it now holds to be fundamental, that a limit of production should be set for each man. If labor unions will drop this primal error, reason may find a basis for discussion, while with this dictum as a premise there can be no reconcilement with the spirit of progress. They must also drop the tyrannical theorem that non-union men may not work with union men, and the anarchistic conception that non-union men must not deliver goods to union shops. Many modern strikes are based on these ideas, and it is quite evident that arbitration is utterly out of the question since the answer is either yes or no. Any board of arbitrators, by the mere act of considering such claims, THE IKOX INDUSTRY. thereby acknowledge that they have a standing in equity, when a moment's consideration will show that they subvert the principles of our government. Almost all of the large steel plants of America manage their own affairs. The result is that the introduction of labor saving devices creates no trouble, the more so because such devices, while they decrease the number of men, demand a higher grade of workmen, so that it often happens that the man who operates the new machine will earn a higher rate of wages than any man made before at the same kind of work. Another reason why labor saving machines are not entirely contrary to the interests of the skilled workman lies in a fact which seems to be unknown to the average social economist. In the manufacture of steel, there is a great deal of very hard and heavy work. Formerly, when most of the work was done by hand, a skilled man was almost necessarily one who was superior physically, and as soon as he reached middle life he was obliged to accept some less arduous and less remunera- tive employment. With the introduction of machinery the skilled employee may often retain his position during the remainder of his life, and the ability to keep an old and trusted employee in a posi- tion where his experience is of value to himself and to his employer is not merely a question of sentiment ; it is an advantage as great to the employer as to the workman. There is a singular paradox in regard to labor saving machinery in that such improvements always tend to an increase in the number of men employed, for the inevitable result is a cheapening of the product and the usual result is an increase in capacity. The cheap- ening brings more business and the establishment taken as a whole employs more men than before. The progressive works grows while the others disappear. The current argument in favor of labor unions may be stated thus: (1) Capital is allowed to organize; (2) Labor must have the same rights as capital; ( 3 ) Labor must be allowed to organize. It is impossible to dissent from the truth of the premises, and it is impossible to escape from the conclusion; but it is necessary to define the terms used. It is essential to the argument that we know just what is meant by "organize." Capital is allowed to organize into corporations, but the rights and privileges of these bodies are regulated by law. They may not overstep whatever regulations FACTORS IX INDUSTRIAL COMPETITION. 619 may be made, and the people can make or change these rules. In only one case in America can a corporation interfere in any way with the private rights, property or freedom of the individual. That exception is the right of eminent domain possessed by railroad companies, and it is well known that the conditions under which this right may be exercised give to every injured party more than sufficient compensation for the trespass. Nevertheless, it is an in- fringement of a personal right, and for this reason such corpora- tions have always been regarded as quasi public and subject to leg- islative control. This control moreover has not been entirely theo- retical, for some of our socialistic Western States have enacted laws that have brought ruin to all the capital invested. Taking into consideration simply manufacturing corporations as being the only ones pertinent to our inquiry, it may be safely as- serted that in no particular do their corporate rights allow any trespass upon the private rights of individuals. It is true that they may use their money to injure men or communities, but so may any private person. Any multi-millionaire might buy a factory and shut it down and ruin a village, and it is difficult to see what could be done about it. He might discharge all his old and trusted serv- ants and the law could hardly touch him. He might commit all the sins charged against corporations and there would be no redress. It is wrong to condemn corporate laws for allowing acts which a private individual may legally do, and it is quite certain that manu- facturing corporations have been given no rights of eminent do- main, no privilege to infringe upon the private estate of the citi- zen. They have the power to issue bonds, to issue stock, to conduct business under a perpetual name, and in return have certain duties, certain taxes to pay, certain regulations under which they must con- duct their business and protect the interests of the minority. This is the extent of their powers as granted by the State. All other powers are inherent as vested in general constitutional prerogatives. This then is the definition of "organize" and the right of men, whether so-called "laborers" or not, to so unite has never been questioned. They may form organizations for pleasure, for im- provement or for business ; but it is another matter when they "or- ganize" to restrict personal liberty. That a band of men may agree among themselves not to work more than a certain number of hours per day is as certain as that they may agree not to smoke, or not to eat meat. Their right to do so is unquestionable. It is their privi- 620 THE IRON INDUSTRY. lege to agree that they will only handle two shovelfuls of earth per hour, or one shovelful per day. It is their right to refuse to work for less than five dollars per day or twice that amount. It is their right to ask their employer to sign a scale and agreement to that effect for one year or ten years, but it is also the right of the em- ployer to ask what guarantee is given that they will stay in his employ, no matter what other inducements are held out to them in other places, and it is also his inalienable right to tell them that such agreements are not according to his wish and that he will try and get men who will work without them; and if such "organiza- tion" should reach the last stage and the "organizers' 7 should de- mand that no one should work in the shop except those subscribing to the union and paying the salaries of the officers, the only possible answer is that such a rule is contrary to the fundamental tenets on which this government rests. It has already been stated that certain matters cannot be arbi- trated. Thus it is of record that a certain "union" works in America was shut down several times, not on account of any dis- agreement between employer and employe, but on account of dis- putes between two rival labor unions. It is quite comprehensible why under such conditions or under many other circumstances a manufacturer might conclude to employ only non-union men. His right to do so is as unquestionable as the right of a farmer to employ only colored laborers or to employ only white men, or to employ both as he chooses. Granting that the manufacturer has concluded to run his place non-union, it is evidently impossible to submit the matter to arbitration. If his conclusion is unwise, he will suffer most, for if men will not work for him then he will lose money, and if he can get only the scum of the streets then also will he lose ; but if he can obtain good men in sufficient num- bers, then it is quite certain that the conditions are acceptable to them and to him and that therefore his position is just and equit- able. It is impossible to conceive how a decision to employ only non- union men can be susceptible to arbitration, and it would seem unnecessary to more than state the theorem were it not that poli- ticians and certain lecturers at Chautauqua are advocating com- pulsory arbitration. It must always be remembered that no em- ployer ever entertained a prejudice against a labor union on gen- eral grounds alone. The opposition arose from the plain fact that FACTORS IN INDUSTRIAL COMPETITION. 621 labor unions regularly develop into the most tyrannical and out- rageous violators of individual rights. It has happened many times that a hundred union men have left a shop because one non-union man was at work. Is it possible that any employer with a grain of self-respect, or any intelligent person, will say that such a matter is open to arbitration. Our common law recognizes prose- cution and imprisonment, but it recognizes the arbitration of crime as the compounding of a felony and calls this a crime in itself. The proposition has been made by a President of the United States that employers should not discriminate against union men, but that union men on the other hand should not interfere with non-union men working beside them. This is a most excellent solution from an academic standpoint, but in nine cases out of ten where such an arrangement is attempted, it is overthrown by the union element, and in places where the troubles have developed into riot and murder we have yet to hear of any assistance given by labor leaders to the legal authorities to punish the instigators of crime. Labor organizations are a form of socialism. In the same cate- gory stand the comprehensive paternal laws of Germany and other European countries and the less radical measures either proposed or enacted in our own land. This fact does not necessarily brand them as wrong, for socialism may contain elements of right and justice. I do not make the senseless generalization that since trades unions are socialistic and socialism wrong, that therefore the unions are wrong, for it would be necessary to prove that all phases of socialism are wrong. But I do make the point that if socialism is right at all, it is right for all ; there must be no classes in America. There is no stone wall between the humblest laborer in a steel works and the manager. The pathway is wide open from the workshop or the mill to the sanctum of the administrative head. The rule that applies to one must apply to the other. If eight hours is the maximum time the laborer is allowed to work, then the same law must govern the manager. If the humblest workman must not work except within certain hours, then the manager must not think except during the same interval. The mechanic must not go home and think how a job can be done better, the superintendent must not try to improve the plant, nor make one more ton of steel to-day than was made yesterday. Moreover, if no man is to do work except at his own trade, then no man must work in his own garden, 622 THE IRON INDUSTEY. raise his own flowers, or mend a broken fence. Such is the inevi- table logic of the labor union. The labor leaders will hardly wish to say that there are classes and castes in America, and if there are no classes then the same rules should govern all ; and if these rules are to be made for all, then they must be laws, made by the regular law-making bodies; made by the people through their chosen representatives. This has been done in New Zealand; it may be well to await the result. It is necessary, however, to remember that in this great experi- ment success will not be measured solely by freedom from strikes, for the industrial peace compelled by arbitration is not necessarily the best thing, any more than political and social peace compelled by the superior force of an autocratic monarchy betokens the high- est triumph of government. The excitement of a political cam- paign in America is more desirable and more truly an exponent of a healthy condition than the sullen passivity with which servile subjects might view a change of masters. The current views of many^ political leaders in interfering with industrial freedom re- semble the medieval notion that a decree of the king could fix the price of wheat, prohibit the export of gold, or exalt the value of a debased currency. The success or failure of such measures cannot be determined by the immediate effect; some people imagine that when the arbitration laws of New Zealand have prevented a strike by the easy method of splitting the difference, a great triumph has been won. They forget that this is a backward step; that it is abandoning the business method of fixing a price, and substituting the ancient Jew practice of asking twice as much as is expected in order that an intermediate price may be secured. If the public supposes that the truth is a compromise between extreme demands, it is easy to keep business in a ferment by asking for an advance. It will take a generation for New Zealand to discover the result of her innovations, but even at this early day the situation is not entirely happy. The employers in three provinces have come out strongly against the present system of compulsory arbitration, while on the other hand the labor union of one of these same provinces is up in arms at the unexpected and strange phenomenon of an award against the workmen, and the Labor Council is asking "why should we obey an adverse award, when no jail in large enough to hold us all?" Not until the regulations made in this distant island have had time to produce their proper fruit, not FACTORS IN INDUSTRIAL COMPETITION. 623- until New Zealand becomes thickly settled and possessed of the complex industrial life existing in those countries which are fac- tors in the business of the world, not until the new schemes of labor regulation have proven their efficacy under international competition, can the laws of this much-discussed country become more than an interesting experiment to be watched rather than to be copied. SEC. XXIc. The question of tariffs. In the minds of many of my readers this discussion will not be in the least complete if I do not place upon record my unqualified belief that the present condi- tion of the American iron manufacture is solely due to the operation of the high protection system, which has been in force for so many years. Let me say therefore that there are some men in the iron trade who believe that the entire business of this country is not represented by a tariff measure, just as on the other hand there are men not connected with the iron business at all who fail to appre- ciate that the tariff is robbing them of their last cent. During the period that high tariffs have been in force our iron industry has expanded, to most wonderful proportions, but that such expansion is due to the tariff is not a necessary conclusion. That such ex- pansion has from time to time been interrupted by periods of panic and disaster is unquestioned, but it is rash to say that such disasters are the inevitable results of protective tariffs. It is quite true that American manufacturers have sometimes sold a part of their products to foreign customers at a lower price than the ruling market quotations at home, and this fact is immediately grasped and spread broadcast by petty politicians and by so-called economists, who seem always to be climbing out on the scale beam in one direction as far as they can go to balance the equally erratic high tariff promoters who are climbing the other way. Nothing can so quite keep in countenance the fallacies of fanatics as counter fallacies gravely argued. Nothing could more please the advocates of free trade than to see protectionists trying to prove that iron ore is not raw material. My mind is not broad enough to grasp all the complex conditions that surround the industrial progress of Amer^ ica, and I cannot see as clearly as some men that no steel would ever have been made here had it not been for certain divinely in- spired orators in Congress ; neither can I see as clearly as others that the nation would have been richer and greater had no duty ever been imposed on foreign manufactures. It is possible that the rea- 9 ^ THE IRON INDUSTRY. son why I cannot see so clearly is that my information is gained at first hand, and is not made up of partisan statements. An able and honest President of the United States publicly announced that a tariff was a tax, and that the price of an article here was the price abroad plus the tariff. If the statement concerning the price had been true then undoubtedly the tariff would have been a tax, but unfortunately for the reputation of the said President, the statement was not true, as he might easily have found and should have found by the most casual inspection of the regular trade papers. In the case of steel rails, for example, the price in the United States is not equal to the foreign price plus the tariff, and has not been for fifteen years, while there have been many times when they were sold here much cheaper than they could be bought at European works. Such free trade nonsense is matched by many protectionist pamphlets declaring that high tariffs mean high prices and high wages, when on the one hand we have seen the United States selling steel cheaper than any other country in the world, and we may see Austria and France, both high tariff nations, paying starvation wages to their work-people, and using women in great numbers as laborers in the roughest kinds of work. ^ The following conclusions may be wrong, but I trust they are not fanatical or entirely unfounded: (1) A high tariff on a certain article hastens very much the establishment of factories to produce that article. (2) The establishment of a new industry like making steel, cot- ton or woolen goods, carpets, etc., etc., requires at least ten years before all the social and industrial conditions have become so corre- lated that the cost of production reaches an economical footing. (3) During this period the general public pays a somewhat higher price for this article, the excess depending on the amount of protection an'd the amount of domestic competition. (4) In some cases and in industries not requiring very large in- vestments of capital or the creation of communities of special work- men, this period during which the public is so taxed may be very short, and the price may soon drop even below that paid to foreign manufacturers. (5) If the profits to the protected manufacturer are large, new works will be erected, and if these combine to extort an unreason- able profit, still other works will be built, the end being the same FACTORS IN INDUSTRIAL COMPETITION. 625 in any event in that the needs will be met and internal competition ultimately bring about a price representing in the long run not much over a fair profit. (6) Whether this price, the cost plus a fair profit, is or is not more than the price abroad will depend upon the natural advantages of the situation. If an article cannot be made here as cheaply as abroad, then the question must be answered whether the public should pay the premium." If it can be made as cheaply, then com- petition will force it to be so made. (7) The "price abroad" is a term which must be used carefully, for the price at which standard articles can be bought from time to time for delivery beyond the borders of the home market does not in the least represent what would be the price under a greater demand; such a demand, for instance, as would be made on Ger- many and the United States if all the steel works of England should shut down. Neither do these quotations represent the real cost of manufacture. (8) The real cost of manufacture includes many things which are usually overlooked, but which are of immense importance. The main items are as follows, it being assumed for the sake of sim- plicity that a steel works owns its own ore and coal mines and coke ovens : (a) Actual operating costs at all mines and works, including labor, fuel, repairs, etc., etc. (b) Freight charges on all raw materials and incidentals. (c) Interest at 6 per cent, on all money actually invested in mines and plant, and on all floating capital needed to carry on the business. (d) Expenses incident to superintendence, selling agencies, taxes, bad debts, pensions, damages, etc., etc. (e) Depreciation, by which is meant a class of items generally overlooked. The ore and coal must bear not only the interest on the money invested, but a sum sufficient to pay for an equal quan- tity of material when the beds are exhausted. The depreciation of the steel plant itself is still higher, for it is almost safe to say that to keep a steel works up to its value, to keep it as a factor in the great strife of competition, requires an annual expenditure of ten per cent, of its cost. Engines, boilers, rolling mills, cranes, shears and all the manifold equipment may last that time, may last longer, or may be outlawed before that period expires. A mill not up to 626 THE IRON INDUSTRY. date cannot compete with one that is, and if it cannot compete,, then it loses money; and if it loses money, then it is worth nothing, absolutely nothing, no matter how new it is or how much it cost. (9) This item of depreciation is often represented on the cost sheets by new equipment and machinery, but sometimes these are erroneously or falsely put into the capitalization account. Whether ten per cent, is or is not the correct figure for a steel plant, it is quite certain that a very considerable amount must be included in the true cost of manufacture. Assuming that the plant cost ten million dollars, a depreciation of ten per cent, is equal to one million annually; and if the pro- duction during the year is five hundred thousand tons, then this charge amounts to two dollars on every ton of steel made. It may be more in some works and may be less in others. (10) When business is slack it is necessary that the manufac- turer ignore this item altogether, for he will assuredly operate his plant if he can cover his actual running expenses. If therefore he does not earn his depreciation during a period of one, two or three years, then he must earn a double amount for an equal period when good times return, and this must not be considered as profit. He must also ignore the interest on the money invested in plant and in floating capital, as well as the expenses of selling agencies, taxes, insurance, etc., since all these items, like depreciation, will go on whether steel is made or not. (11) During this era of low prices, the actual cost sheets and the annual reports may show no loss or even a margin of profit, and the average observer might conclude that these figures represent the proper selling price, a conclusion which would be entirely erroneous. (12) It is the part of common sense for rival manufacturers to get together and agree to prevent cutthroat competition, by which not only are all profits thrown away and all depreciation and in- terest charges ignored, but even operating costs encroached upon. A fair price under such an arrangement would include depreciation and interest as fundamental parts of the cost. (13) Having made such an agreement for home trade it becomes good policy to ignore these items on competitive business for foreign deliveries, since they are both fixed quantities, not depending on the amount of steel produced, and the extra output caused by such foreign deliveries cheapens the cost to the manufacturer. More- over certain lines of foreign trade cannot be held if prices are varied FACTOKS IN INDUSTRIAL COMPETITION. 627 with every local advance. Having secured for instance the business of a certain railway in Australia, it is evidently quite impossible to retain it if the price quoted follows every boom in the home market; and it is certainly good policy to keep the trade. of this railway for future business, in spite of the hue and cry about lower prices to foreign buyers. (14) This argument is not new, but has been an accepted com- mercial and industrial maxim in every country, under both protec- tion and free trade, and all the "prices abroad," so freely quoted, are based on this rule as existing in foreign lands. It is even true that bounties are actually paid in some instances to encourage export trade. (15) The payment of a bounty for export trade is directly in line with the maintenance of a protective duty after the incubative period has passed. Practically it must be looked upon as out of the question owing to the impossibility of arriving at a complete knowledge of just what would be equitable, but although such a system would breed many wrongs, it is theoretically justifiable to a certajn limited extent. A steel works, in common with every manufacturing plant, is a benefit to the general public in many ways. It contributes to the payment of taxes and thus saves an equivalent amount of individual expenditure. It is the foundation of large communities which influence and increase the general prosperity of the country by giv- ing a market for all kinds of commodities. It supplies freight to the railroads in enormous quantities. In the twelve months imme- diately preceding the time of writing, the works with which I am connected received 54,903 cars and sent out 17,471 cars loaded with finished material. Allowing for some empty cars received and estimating the average of the whole to be 30 tons per car, we have a total of 2,170,000 tons of traffic and a total train length, of 500 miles. The average length of haul is unknown, but was over two hundred miles. This business brings in an enormous income to the railroads, the gross receipts from a steel works being four or possibly six times as much as though a similar amount of material were imported from abroad, and there were no raw materials or incidental supplies to assemble. It will be evident on the slightest consideration that the cost of moving other freight is reduced by this increased business, and the establishment of other industries 628 THE IRON INDUSTRY. thereby made possible, which, in turn, react by further lowering the cost of transportation by their contribution to tonnage moved. A nation would lose no money if a bounty were paid to support manufactures, provided such support were necessary, and provided that the bounty did not exceed the sum directly and indirectly paid or saved by the manufacturer to the state and to the public. If German steel is laid down in England at one shilling per ton cheaper than English steel works can make it, and if that shilling represents the dividing line of business, then it would be money in the pocket of the taxpayers of England if a protective duty of one shilling were levied upon foreign steel, since the amounts contrib- uted by works in operation must be much more than this. It is impossible to give the upper limit of such a tariff, for the conditions are too various and include all the correlated conditions, down to the higher value of farm products in industrial communities. Within this range, whatever the limits may be, a protective tariff is not illogical ; beyond the limit, it is uneconomical. Such are my opinions. They may not embrace absolute truth. Few things have ever been written that were beyond need of change, but it has been deemed advisable to revise the first chapter of Genesis and it is barely possible that some alteration may be neces- sary in the Wealth of Nations by one Adam Smith. CHAPTER XXII. THE UNITED STATES. SECTION XXIIa. General view. It is impossible to treat the iron and steel industry of the United States with the same com- pleteness that the different nations of Europe will be discussed. This arises from the fact that the scale upon which our country is constructed is so entirely out of proportion to the scale by which other countries are always considered. It also arises from the absence of any tariff restrictions between parts of our country ; thus it is quite conceivable that if New York State were a separate em- pire she might have a high tariff on steel and a low tariff on raw ma- terials, and have long since created within her own limits a per- manent steel industry on a considerable scale. In the absence of such a great center of production the output of the State is not only left undivided but is combined with that of New Jersey. From one point of view, however, it is wrong to consider as a unit a dis- trict as large as New York alone. The iron industry of the State is made up of two parts, entirely independent of each other. In the northeastern section are the ore mines of the Lake Champlain district and in the extreme west are the furnaces of Buffalo and Tonawanda, smelting Lake Superior ores. These two districts have no relation whatever to one another ; they are 250. miles apart in a straight line, farther than from the ore mines of the Cleveland dis- trict to the coal of Cardiff; as far as from Prague in Bohemia to Gleiwitz in eastern German Silesia. It would be more logical after the erection of the new steel plant at Buffalo to make a group em- bracing the works upon the southern shore of Lake Erie, although this would be combining two entirely independent producers like Buffalo and Cleveland, which are as far apart as Dortmund and Saarbrucken. The State of Virginia is always considered as a whole, but it covers an area nearly as great as England. It is not looked upon as one of the great centers of pig-iron production, but it makes half 629 30 THE IRON INDUSTRY. as much as the whole of Belgium, half as much again as South Yorkshire, with Leeds and Sheffield, and nearly double the output of Aachen and Ilsede combined. In any book treating exhaustively of pig-iron it would hold a prominent place, but it is discussed here simply as proof of the vastness of the subject, when this State may be neglected as having little bearing on the general business of making steel. One of the fundamental differences between American and Euro- pean conditions arises in this geographical separation and the dis- tances through which the raw material must be assembled. In Europe a haul by railroad of 200 miles is considered very long and the cost is excessive, while in America it is not unusual at all. Coal and coke are carried this distance in several instances, while Chi- cago draws its blast furnace fuel from West Virginia and the Con- nellsville field, the distances ranging from 500 to over 600 miles. The most magnificent disregard of distance, however, is seen in the official publication of the American Iron and Steel Association, wherein the steel production of Colorado is combined with that of East St. Louis in Missouri. These are entirely independent pro- ducing centers and they are 800 miles apart in a straight line ; a;; far as St. Petersburg is from the coal fields of South Russia; as Middlesborough is from Upper Silesia ; Westphalia from Styria and Paris from Warsaw. The fault however does not rest with Mr. Swank, but in the secrecy enforced upon him by certain interests. The statistical reports of this country are by no means what they should be and this is due to disinclination on the part of some manufacturers to give information. The data on pig-iron are quite full as a general thing, but the records concerning steel production are very meagre. It is impossible to make out from the usual sources of information any accurate statement of the amount of steel made in the various well known and most important districts. The Directory to the Iron and Steel Works of the United States, published by the American Iron and Steel Association, 261 South Fourth street, Philadelphia, gives details of almost every plant in the United States. This information is so complete that it is moro than useless to give a list of works for each district, but I have compiled, with some labor, the number of converters, open-hearth furnaces and rolling mills in each district, and have calculated from this basis, an from several private sources of information and from official statistics, the output of iron and steel in each locality as THE UNITED STATES. 631 nearly as possible. The private information was in some cases confidential and is used only in groups, as for instance, the data concerning a portion of the Pittsburgh district. The results are given in Table XXII-A. In Table XXII-B are given the records of production of steel for the whole country from 1867. This has been condensed to make Table XXII-C in order to show, for both the United States and for Great Britain, the amount of the different kinds of steel made, while Table XXII-D gives the percentage of each product. TABLE XXII-A. Output of Pig-Iron and Steel in 1901 in the United States, together with Data on Producing Capacity ; estimates in parantheses. Note : See text for boundaries of districts ; thus "Pittsburgh" Includes parts of three States and output of pig-iron for "Steelton" includes the product of two counties. District. Blast Furnaces. Pig Iron. No. of works having roll- ing mills. M 1 No.of works making cruci- M ^ 1 ble steel. Bessemer Converters. Small, mostly for steel castings Standard size. 7 to 20 ton. Coke. Char- coal. Output; tons. Per cent, of total No. Aver- age capac- ity. No Aver- age capac- ity. Pittsburg Illinois 82 20 6,880.000 1,597.000 1,225,000 783,000 695000 512,000 481,000 478,000 449,000 439,000 337,000 309.000 303.000 208,000 185,000 171,000 398.000 301,000 68000 18,000 | 27 : 000 12.000 43.3 10.1 7.7 4.9 4.4 3.2 3.0 3.0 2.8 2.8 2 1 2.0 1.9 1.3 1.2 1.1 25 1.9 0.4 0.1 0.2 0.1 137 21 10 15 12 3 11 36 6 38 2 4 6 7 2 6 45 27 9 5 30 5 10 11 3 2 39 8 18 6 29 17 22 27 15 11 4 6 3 6 1 4 3 1 8 ""l" '"9" 3 '"i" ""4" 7 Cleveland, Ohio Steelton. Pa Johnstown, Pa Lehigh Valley, Pa... Southeastern Penn- sylvania 4 3 4 4 11 10 12 7 Y 3 5 2 Virginia New York and New Jersey 8 1 Tennes-ee Hanging Rock. Ohio. Sparrow's Point. Md. Wisconsin and Minn. Colorado.- Michigan 2 20 4 "2 3 3 '"2" 2 ' 2" 2 5 Other parts Penn Other parts Ohio Kentucky 17 9 8 1 2 1 1 1 2 2 4 5 Missouri ... 2 2 North Carolina Georgia New England 15 28 7 3 1 2 2 Delaware .... Other States 6 2,000 8 45 1 2 Total . 345 54 15,878,000 100.0 460 19 58 632 THE IRON INDUSTRY. TABLE XXII-A. Continued. District. Open Hearth Furnaces. Steel; all kinds. Acid. Basic. Steel castings not included in foregoing. No. Aver- age capac- ity. No. Aver- age capac- ity. No. Aver- age capac ity. Output; tons. Per cent, (f toial. Pittsburgh 35 3 3 2 12 3 30 25 25 35 30 45 84 9 10 8 33 9 40 40 30 40 40 40 20 13 2 2 14 2 18 15 10 4 20 15 (7.317,000) 1,750,000 870,000 656000 629,000 427,000 352,000 352.000 173,000 (150000) (150,000) 107,000 69,000 (50,000) 15,000 165.000 53000 ]. i- (189,000) 54.3 13.0 6 4 4.9 4.7 3.2 2.6 2.6 1.3 1.1 1.1 0.8 0.5 0.4 0.1 1.2 0.4 1.4 Illinois Cleveland, Ohio Johnstown Pa Southeastern Penn Steelton Pa Sparrow's Point, Md .... Scran ton Pa 5 15 6 13 40 45 6 1 15 20 Alabama New York and New Jer- sey 2 6 25 30 8 2 20 40 10 '"3" 10 '"20"' Lehigh Valley Pa Missouri Hanging Rock Ohio 4 30 Other parts of Ohio >ther parts of Penn Tenneessee . 4 4 11 20 15 4 1 3 15 15 3 20 3 15 Wisconsin and Minn.... Michigan Kentucky 1 1 1 1 1 20 15 7 30 50 3 15 7 1 4 25 30 50 7 20 Delaware Total 84 204 103 I 13,474,000 100.0 The grouping in irregular periods may seem arbitrary, but the lines of division were found by calculating each year separately and tak- ing the years that seemed to mark a change in practice. These tables, when taken in conjunction with a knowledge of the condi- tions that have ruled the steel industry of the country, tell a very clear story which may be related as follows : In 1867 the production of Bessemer steel in the United States was 2679 tons. Some small quantities were made before this, but the industry was put on a permanent footing by the establishment of an entirely new Bessemer plant at Steelton, Pa., a plant which has continued to make steel from then until now. This was fol- lowed soon afterward in the same year by Troy, while Cambria, at Johnstown, was the next to enter the field, this latter plant having also continued to be an important producer to the present time. From 1867 to 1871 about 20,000 tons per year, or about half the steel of all kinds made in the country, was made by the Bessemer THE UNITED STATES. 633 TABLE XXII-B. Production of Steel in the United States in Gross Tons from 1867 to 1900. Year. Bessemer In- gots. Open Hearth Ingots. All Kinds of Steel. Bessemer Rails. Bessemer Steel ; per cent, of total Steel. 1867 2,679 19 643 2 277 14 1868.... 7589 26 786 6 451 28 1869 10 714 893 31 250 8 616 34 1870 37500 1 339 68*750 30 357 53 1871 40 179 1 785 73 214 34 152 55 1872 107 239 2 679 142,954 83*991 75 1873 1874 152,368 171 369 3,125 6250 198,796 215,727 115,192 129 414 77 79 1875 335 283 8 080 389 799 259 699 86 1876 469,639 19 187 533,191 368269 88 1877 1878 500,524 653 773 22,349 32255 569,618 731 977 385,865 491,427 88 89 1879 . . 829 439 50 259 935 273 610 682 89 1880 1,074,262 100851 1,247 335 852,196 86 1881 . 1 374 247 131 202 1 588 314 1 187 770 87 1882 '.. 1,514,687 143 341 1,736,692 1,284,067 87 1883. . 1 477 345 119 356 1 673 535 1 148 709 88 1884 1,375,531 117,515 1,550,879 996,983 89 1885 .. 1 519 430 133 376 1,711 920 959 471 89 1886 2 269 190 218 973 2 562 503 1 574 703 89 1887 1888 2,936,033 2,511 161 322,069 314,318 3,339|071 2,899,440 2,101,904 1 386,277 88 87 1889 2 930 204 374 543 3385 732 1 510 057 87 1890 3 688 871 513 232 4 277 071 1 867 837 87 1891 3 247 417 579 753 3904 240 1 293 053 83 1892 4 168 435 669 889 4 927 581 1 537 588 85 1893 3 215 686 737 890 4 019 995 1 129 400 80 1894 b 571 313 784 936 4 412 03 9 1 016 013 81 1895 4 909 128 1 137 182 6 114 834 1 299 628 80 1896 .... 3 919 906 1 298 700 5 281 689 1 116 958 74 1897 5 475*315 1 608 671 7 156 957 1 644 520 1898 6 609 017 2 230 292 8 93 857 1*976 702 74 1899 1900 .... 7,586,354 6 684 770 2,947,316 3 398 135 10,639,857 10 188 329 2,270,585 2 383 654 71 66 1901 8 713 302 4 656 309 13 473 595 2 870 816 65 process, and all of this went into rails. From 1872 to 1874 the annual production was about 140,000 tons, all of which was rail steel, and, in spite of the development of the open-hearth process, this represented about three-quarters of the total steel output. From 1875 to 1879 the output of Bessemer increased nearly fivefold over the period just previous, and averaged about 560,000 tons per year. A great part was made in the eastern portion of Pennsylvania, at Steelton, Johnstown, Bethlehem and Scranton; but the then new works of Edgar Thomson, at Pittsburg, and the plants at Chicago and Cleveland had by this time become factors of great importance. The Bessemer output during this time was 88 per cent, of the total steel output of the country and all of it was rolled into rails. From 1880 to 1882 the output more than doubled, averaging 1,320,000 tons, which constituted 87 per cent, of all the steel made, 634 THE IRON INDUSTRY. TABLE XXII-C. Production per Year during Certain Periods of Bessemer and Open-Hearth Ingots and Rail Steel. It is assumed that 100 tons of ingots=83.3 tons of rails. Note: United States. Great Britain. Bessemer Bessemer Period. Total Steel. Bessemer Ingots. Open Hearth Ingots Rails plus 20 per cent. = Rail In- Total Steel. Bessemer Ingots. Open Hearth Ingots. Rails plus 20 per cent. = Rail In- gots. gots. 1867-1871 incl. 44,000 20,000 800 19,000 180,000 Iftfi flOO 143 000 4000 131 000 482000 1875-1879 1880-1882 1883-1887 1888-1890 632,000 1,524,000 2,167,000 3,521,000 558,000 1,320,000 1,910,000 3,040,000 26,000 125,000 182,000 401,000 508.000 1,330,000 1,627,000 1,906,000 963,000 1,808,000 2,280,000 3,585,000 742,000 1,387,000 1,563,000 2,063.000 141,000 342,000 638,000 1,429,000 *564,000 1,196,000 1,042,000 1,172,000 1891-1893 1894-18% 1897-1899 1900 4,284,000 5,269,000 8,910,000 10,188,000 3,540,000 4,130,000 6,560,000 6,685,000 663,000 1,074,000 2,262,000 3,398,000 1,584,000 1,373,000 2,357,000 2,861,000 3,109,000 3,611,000 4,751,000 5,050,000 1,545,000 1,629,000 1,823,000 1,745,000 1,463,000 1,883,000 2,813,000 3,156,000 712,000 808,000 1,004,000 912,000 1901 13,474,000 8,713,000 4,656.000 3,445,000 4,904,000 1,606,253 3,297,791 878,000 *1875 is estimated. TABLE XXII-D. Proportion of Various Kinds of Steel made in the United States and Great Britain. Period. Bessemer Steel. Open Hearth. Per cent, of Total. Rail Steel per cent, of Bessemer. Per cent, of Total. United States. (Great Britain. United States. Great Britain. United States. Great Britain. 1867-1871 inclusive 45 77 88 87 89 87 83 78 74 66 65 95 92 91 100 85 63 45 33 36 44 40 2 2 4 8 9 11 15 20 25 33 35 1872-1874 1875-1879 77 77 70 58 50 45 38 35 33 76 86 67 57 46 50 55 52 55 15 19 28 40 47 52 59 62 67 1880-1882 1883-1887 1888-1890 1891-1893 " 1894-1896 " .... 1897-1899 " 1900 1901 and almost all was put into rails. A small amount was made at Steelton of high carbon special steels, and Cambria also made some for use in her Gautier Department for agricultural tools. During this period there was a marked increase in the make of open-hearth steel, a start having been made by the building of a furnace at the works of the New Jersey Steel and Iron Co. in 1868, but the intro- THE UNITED STATES. 635 duction of the process was slow and it was not until 1880 that the output reached 100,000 tons per year. Up to this time the steel industry was largely dependent upon Spanish ores, and the works near the eastern seaboard were in the most advantageous position; but during the period from 1880 to 1890 the development of the Lake Superior deposits and the establishment of cheap methods of transportation made the United States practically independent of foreign ore, while the exploitation of the Mesabi range in 1893 transferred the command of the steel market to a point west of the Allegheny Mountains. From 1883 to 1887 the production of Bessemer steel was 1,900,- 000 tons per year, being 89 per cent, of the total, the open-hearth furnaces making about one-tenth as much. Only 85 per cent, of the Bessemer steel was rolled into rails, for about this time at Steel- ton, Cambria, Bethlehem and elsewhere, considerable high carbon steel was being made, as well as some soft steel. Some Bessemer plants not connected with rail mills were operated to make steels' for special purposes and supply the general trade, and this develop- ment became more pronounced in the next period from 1888 to 1890, when only 63 per cent, was put into ra*ils, while in the period from 1891 to 1893 more than half the Bessemer output went into miscellaneous work, and from 1894 to 1896 only one-third was used for rails. This great change was brought about by many causes, prominent among which was the general use of the reversing mill for rolling four-inch square billets directly from the ingot, and the immediate acceptance by the trade of that size as the one standard. By the economies following this innovation wrought-iron was driven from the market and was superseded by steel. One of the most impor- tant fields affected by this change was the making of railway joints or splices, which amount to from five to seven per cent, of the weight of tke rails themselves. A still greater change was the rapid and almost complete substitution of steel for plates and sheets of all kinds. During all these years, however, the open-hearth process has been making very hsavy strides and narrowing the field of the Bessemer converter. One of the first acts of trespass was in the making of high carbon steels ; it was found that the steel made in the regenera- tive furnace gave better results, and to-day very little high steel is made by the pneumatic method. The next great encroachment was 636 THE IRON INDUSTRY. in structural shapes, where the Bessemer product found a great out- let in the years from 1885 to 1893 or thereabout. The proportion of converter product going into bridges is very small at present, while it is becoming less for ships and buildings. This growth of the open-hearth furnace is shown by the fact that in 1901 the steel made in the converter formed only 65 per cent, of the total output, while in the period from 1875 to 1890 it was about 88 per cent. It is also shown by the fact that in the two years of 1900 and 1901 the proportion of Bessemer steel used for rails increased to an aver- age of 42 per cent., it being only 33 per cent, in 1894 to 1896. To-day two-thirds of the steel made in the United States is Bes- semer and one-third open-hearth. Practically all the rails are Bes- semer, but open-hearth steel is used for almost all other work where the material is subject to physical and chemical specifications. One- quarter of this open-hearth steel is made on an acid hearth, the re- mainder on dolomite or magnesite linings. The use of the basic furnace is rapidly spreading both in small and large plants, but very few new Bessemer plants are being erected. No fuel is im- ported for the making of iron and steel, but a considerable quantity of ore is brought from Cuba and elsewhere to points on the Atlantic seaboard, as shown by Table XXII-E. TABLE XXII-E. Iron Ore Imported into the United States. U. S. Geol. Survey, John Birkinbine. Imported from 1896 1897 1898 1899 1900 Cuba... 380551 383 820 165 623 360 813 431 265 Spain 121 132 66 193 13 335 145 206 253 694 French Africa 79 661 3 504 22 233 20 000 Italy 29,882 43.363 18.951 Greece Newfoundland and Labrador 33,750 20 800 29 250* 7,200 16.765 77 970 23.350 140 535 United Kingdom . 8 528 358 683 172 397 Colombia 3 150 3 000 Quebec. Ontario, etc 5588 Other countries 5 352 6 845 367 7 560 1 051 Total 682 806 489 970 674 082 A map is given in Fig. XXII-A, which is taken from the U. S. Geol. Survey. This shows in the shaded portions the principal coal fields of the United States, the anthracite deposits of eastern Pennsylvania being represented by solid black. The crosses denote THE UNITED STATES. 637 localities which are important producers of ore, the only ones deemed worthy of note as determining factors being the Lake Su- perior deposits, and those of Alabama, Colorado and Cornwall, Pa. FIG. XXII-A. UNITED STATES; EASTERN HALF. 638 THE IKON INDUSTRY. The circles indicate the position of the important steel producing centers and in the following pages will be given a detailed descrip- tion of each of these districts. FIG. XXII- A. UNITED STATES; WESTERN HALF. THE UNITED STATES. 639 SEC. XXII-b. Coal: The United States may be said to import no coal. This is per- fectly true as far as the general iron industry is concerned, but as an explanation of certain facts given in the official statistics, it is necessary to note that a considerable quantity is shipped from Brit- ish Columbia to points on the Pacific Coast, while a lesser quantity is brought from Cape Breton, Nova Scotia, to Boston, Mass., for the manufacture of illuminating gas in by-product coke ovens. Within the last few years a very considerable trade has grown up in the export of coal, mostly to Canada and Mexico, but a great deal to places oversea. In 1900 about 635,000 tons were shipped to Europe, a considerable amount going to Mediterranean ports, at- tracted by the phenomenally high prices ruling in France. Accom- panying is a statement showing the foreign trade, including Canada and Mexico: IMPORTS AND EXPORTS or COAL :IN 1900 IN LONG TONS. Production 239,567,000 Imports 1 ,909,000 Exports 7,917,000 Anthracite. In the consideration of the fuel supply of the United States a word should be said concerning anthracite, because there is much misunderstanding among foreign metallurgists as to the amount of this coal used in iron smelting. Many years ago lump anthracite was very commonly used in Eastern Pennsylvania, New Jersey and other neighboring districts as the only fuel put into the blast fur- nace, but this practice has become the exception, and coke from Connellsville has for a long period been carried to the furnaces that are situated in the very heart of the hard coal region. Some fur- naces do use anthracite alone, and at many plants it is not unusual, in cases where coke cannot be obtained or when it is very high in price, to use a certain proportion of hard coal, but this hardly war- rants the misleading classification of many of the Eastern plants under the head of "anthracite furnaces." There is a great amount of hard coal used in firing boilers in in- dustrial establishments of all kinds, but only the small sizes are available for this purpose, the larger kinds commanding a higher price for household use. Except "in the immediate neighborhood of the mines it is more economical to use bituminous coal brought from a long distance than to use the sizes that can be sold for do- mestic purposes, while the smaller grades will not burn readily and 640 THE IRON INDUSTRY. require a blast when used under boilers. Every few years the price of the smaller sizes advances and the manufacturer must either change to soft coal or alter the grates to handle still smaller pieces. This arises from the fact that the small pieces are a by-product produced in crushing, and the mines produce as little as possible of the less valuable product, while on the other hand there has been much progress in devising grates and stokers to handle the fine sizes. In many Eastern cities the community demands a smoke- less stack, so that factories are practically compelled to use hard coal. The demand is founded on aesthetic considerations, the claim that smoke is unhealthful being rather amusing. Aside from this consumption of anthracite for steam making, hard coal may be con- sidered simply as the fuel which is universally used for household purposes in the northeastern part of the country, all of this district being supplied from the mines in Eastern Pennsylvania. A cer- tain amount is also raised in Colorado and New Mexico, but the quantity is trifling compared with the output of the Appalachian field. The value of a short ton of anthracite at the mines in Penn- sylvania in 1900 is given as $1.79, while in Colorado it was $3.00, and m New Mexico $2.75. The hard coal district of Pennsylvania is divided usually into three parts, which are shown in Fig. XXII-B as Nos. 14, 15 and 16. Following is a description of each division: No. in Fig. XXII B. Name. Local Districts. Situation in Counties of Penn- sylvania. 14 Wyoming. Carbondale, Scranton. Piltston Wilkesbarre, Plymouth. Kings ton. Luzerne and Lacka wanna. 15 Lehigh. Green Mountain, Black Creek, Hazleton, Beaver Meadow Luzerne and small parts of Car- bon. Schuylkill and Colum- bia. 16 Schuylkill Panther Creek. Lorberrv, Fast SchuylkilLWestSchuylkill. Ly- kens Vallev. Shamokin. East Mahanoy West Mahanoy. Carbon, Dauphin, Schuylkill, Columbia and Northumber- land. All of this region is in the eastern center of the State. The total production of anthracite in 1900 was as follows in short tons : Pennsylvania 57,363,396 Colorado 59,244 New Mexico 41,595 Total 57,464,235 THE UNITED STATES. 641 SCALE OF MILES 10 20 40 60 80 100 FIG. XXII-B. PENNSYLVANIA, WEST VIRGINIA, OHIO, EASTERN HALF. 642 THE IRON INDUSTRY. Port Hu M I C H I G A FiG. XXII-B. PENNSYLVANIA, WEST VIRGINIA, OHIO, ETC. ; WESTERN HALF. THE UNITED STATES. 643 Bituminous. In the production of anthracite coal Eastern Pennsylvania not only is first, but stands almost alone, while in bituminous coal Western Pennsylvania stands not quite alone, but pre-eminently first. In 1900 she made over three times as much as any other State and more than one-third of the total of the country. The leading counties are Westmoreland, Fayette and Allegheny, with Cambria, Clearfield, Jefferson and Washington following with heavy outputs. The Clearfield coal is one of the best coals in the world for steam purposes, and, together with the Pocohontas and New Eiver coals of West Virginia, is carried in great quantities to East- ern points. Some of the Westmoreland coal is exceptionally rich, and as it is sold at about the same price as leaner coals, and as the freight rates are not always proportional to the distance, it follows that it is economical to use it in the manufacture of fuel gas in producers not only in neighboring districts, but in places quite re- mote from where it is raised. The foregoing remarks concerning the use of the best gas coal apply to many other things in America. On account of the com- paratively low freight rates the tendency is to obtain the best, while in Europe the high rates compel the use of local inferior raw ma- terials; by the American system the railroads do a much greater business and thereby reduce costs. In some parts of Europe the steel works have a score of different mines from which coal is drawn, and a score of places from which ore comes, and the sources of supply are constantly changing with local conditions, with perhaps periodical reversions to the utilization of poor supplies near at hand. The limited capacity of certain ore and coal fields will account for a portion of this difference. The coal deposits of the United States are divided into seven iields, which are shown in Fig. XXII-A, but only four are of any importance : (1) The Appalachian, extending from New York to Alabama, a length of 900 miles, and a width varying from 30 to 180 miles. (2) The Central, embracing parts of Indiana, Illinois and West- ern Kentucky. (3) The Western, including the coal west of the Mississippi Hiver, east of the Rocky Mountains and south of the forty-third parallel. 644 THE IRON INDUSTRY. (4) The Rocky Mountain, including the basins in that range. The smaller fields include a deposit in Northern Michigan, one in Virginia and North Carolina, and one in Washington, Oregon and Northern California, the latter claiming attention owing to the absence of a good supply on the Pacific Coast. The coal from the Central and Western divisions, including a very considerable part of the Mississippi Valley, is of importance from a general economic standpoint for industrial and domestic purposes, but need not be considered here, as it has little bearing on the iron industry; but it is necessary to discuss the beds of the Appalachian and Rocky Mountain districts, which supply practic- ally all the coal and coke used in the iron industry. TABLE XXII-F. Production of Coal and Coke in the United States in 1900 (1 ton =2000 pounds; taken from U. S. Geol. Survey for 1900.) The number of ovens given is the total number standing, less those that are marked abandoned in the report Coal. Coke. Anthracite. Bituminous. No. of Ovens. Production. Pennsylvania 57,363,000 79,318,000 25,154,000 6,358.000 21,153,000 20.671.000 8,504.000 5,436,000 32,464 154 14 10,142 369 6,529 1,488 204 13,798,893 2,631 2,358.499 72116 2,110,837 | 618,755 Illinois Indiana West Virginia Alabama Colorado 59,000 Utah Iowa 5,090,000 4,991,000 4,507,000 4,129.000 3.924.(X'0 3,904,000 2.505,000 Kentucky Kansas 458 91 74 95,532 5,948 14,501 Wyoming Maryland Tennessee 2,106 2,331 400 480 342 230 376 475,432 685,156 * 73,928 54,731 38.141 128,248 Virginia Massachusetts Georgia 333.000 1.662,000 1.922 000 11,261,000 Montana Indian Territory Others 42,000 Total 57,464,000 210,822,000 58,252 20,533,348 * Massachusetts and New York are included in Pennsylvania. Table XXII-F shows the production of coal and coke in the United States in 1900 by States, and Table XXII-G the output of the different coal fields. There are also given in Table XXII-H the records for each county in Pennsylvania for coal and coke, and THE UNITED STATES. TABLES XXII-G. 645 Output of Coal from the Principal Coal Fields of the United States in 1900 (Mineral Resources U. S. Geol. Survey for 1900.) Field. Product, tons. Per cent, of Total. Appalachian (including Alabama) Central 142,497,208 35 368 164 67.1 16 6 Western 17 ? 549'528 8 3 Rocky Mountain 13 398 556 6 3 Pacific Coast 2 704 665 1 3 Northern 849 475 4 TABLE XXII-H. Production of Bituminous Coal in Pennsylvania and Amount used for making Coke. (Mineral Resources, U. S. Geol. Survey for 1900). One ton=2000 pounds. County. Total Coal Mined. Amount Coked. Favette 15 055 000 9 421 500 Westmoreland Allegheny Cambria 14,980,000 10,052,000 8 190 000 7,001,000 ""413606 Clearfield . . . 6 621 000 30 1 000 Jefferson 6 199 000 1 034 000 Washington Somerset 4,856.000 4 779 000 ' 32 500 Armstrong Center 1,313,000 932 000 Tioga. . 931 000 Elk 926 000 2 500 Indiana 925 000 PI ooo Bedford 570 000 164 000 Mercer 528 000 Blair . . 497 000 108 000 Clarion 405 000 Huntingdon 369000 Bradford ] 321 000 Clinton 262 000 Butler 222 000 Lawrence 187 000 Lycoming 119 000 McKean Others . . . 600 000 Total 79 842 000 18 571 500 in Table XXII-I the coke production in the different fields of Penn- sylvania and West Virginia, the leading States. The division into fields is in accordance with the recognized usage of the Geological Survey, and I append a condensation of their descriptions, taken from the reports on both coal and coke. The numbers refer to Fig. XXII-B, on which the location of these fields is shown. 646 THE IRON INDUSTRY. TABLE XXII-L Coke Statistics for Pennsylvania and West Virginia for 1900. (Mineral Resources, U. S. Geol. Survey) ; one ton=2000 pounds. State and District Coke Ovens. Production. Built, Building. Pennsylvania 21,061 2,096 2,010 2,203 1,341 568 476 532 1,498 697 5,290 1,563 2,569 -827 686 10 039.000 827,000 1,067,000 754,000 557,000 135,000 133,000 113,000 111,000 62,000 1,209,000 356,000 507,000 287,000 Pittsburgh Reynoldsville and Walton* Upper Connellsville Allegheny Mountain .... Clearfield Center Lower Connellsville 1,112 West Virginia- Flat Top (Pocahontas) Upper Monongahela New River and Kanawha.. 666 640 * The figures for the Reynoldsville and Walton district are worthless. They Include the production of the coke ovens in New York and Massachusetts "for want of a better classification." It is elsewhere stated that this is done in order that "individual information (for Massachusetts) may not be divulged," which Is hardly sufficient ground for vitiating the statistics of Pennsylvania. Pennsylvania Coke Districts. No. 1. Connellsville: The County of Fayette and the southern half of Westmoreland. Pittsburgh : Vicinity of Pittsburgh, the coke being made from coal brought down the Monongahela Kiver. No. 2. Keynolds and Walton : All the ovens on the Rochester and Pittsburgh Railroad, those on the Low Grade Division of the Allegheny Valley Railway, and the mines on the New York, Lake Erie and Western Railway. No. 3. Upper Connellsville : The region around and north of La- trobe, the coal here being somewhat different from the deposit farther south. No. 4. Allegheny Mountain : Ovens along the line of the Penn- sylvania Railroad from Gallitzin to beyond Altoona, and those in Somerset County. This includes also the coke ovens near Johnstown. No. 5. Clearfield Center: The two counties of Clearfield and Center. THE UNITED STATES. 647 No. 6. Greensburg : Near the town of Greensburg, in the central part of Westmoreland County. No. 7. Broad Top: The Broad Top coal field ii Bedford and Huntingdon counties. No. 8. Lower Connellsville : A new district, first appearing in the U. S. reports in 1900. Known also as the Klondike dis- trict, a southwest extension of the Connellsville Basin. No. 9. Irwin: The neighborhood of the town of Irwin on the Youghiogheny River, in the western part of Westmoreland County. The Beaver, Allegheny Valley and Blossburg districts, formerly recognized, are no longer of importance. West Virginia Coke Districts. No. 10. Pocahontas: The ovens in West Virginia in the Poca- hontas coal field; this embraces the counties of McDowell and Mercer in West Virginia and Tazewell County in Virginia. Most of the output comes from the West Virginia side. This district is traversed by the Norfolk and Western Railroad. No. 11. Upper Monongahela: This is also called the Fairmount district; it is the northern part of the State, drained by the Monongahela, and sending its coal to market by the Baltimore and Ohio Railroad. It embraces Preston, Taylor, Harrison and Marion counties. The statistics include the ovens located at Wheeling, at the Riverside Iron Works. Xo. 12. New River and Kanawha : These two are named from the rivers draining them, and embrace Fayette and Kanawha counties. The coal is shipped partly by the Chesapeake and Ohio Railroad and partly by the Kanawha River. No. 13. Upper Potomac : Also called the Elk Garden district, in- cludes Mineral, Tucker and Randolph counties and is the southern extension of the Cumberland district of Maryland. The West Virginia Central and Pittsburgh Railway runs through this field. SEC. XXIIc. Lake Superior: NOTE : I am indebted to Mr. G. F. Knapp, of Ogleby, Norton & Co., for a careful read'ng of this manuscript. Up to 1880 the State of Pennsylvania was the greatest producer of iron ore in the Union, but the amount raised was entirely in- 48 THE IRON INDUSTRY. sufficient to supply its blast furnaces, and large quantities were im- ported from Spain, some from the west coast of England, and some from other countries like Algeria, Greece and even Ireland. For many years Michigan had been mining ore, the Marquette deposits having been opened in 1845, but it was not until 1856 that as much as 5000 tons was shipped to the furnaces of Pennsylvania. The cost of transportation was high and Spanish ores were taken to Pittsburgh as cheaply as the Western ores could be laid down at that point. The Menominee beds were opened in 1877, the first shipments from Escanaba being made in 1880, and in about the year 1881 the output of Michigan exceeded that of any other State. In 1884 the Gogebic range was opened, all three districts being in Northwest Michigan, and this still further added to its prominence ; but in the same year the Vermilion mines in Northeastern Minne- sota began to produce, and when finally, in 1892 and 1893, the Mesabi range was exploited, Minnesota became a dangerous rival. In 1901 the Mesabi mines produced 9,303,541 tons and the Ver- milion 1,805,996 tons, a total of 11,109,537 tons, while Michigan raised only 9,654,067 tons, this giving first rank to Minnesota. The cause of this enormous increase is not simply the opening of new mines, for this is but one factor in the work, the other fac- tor being the great decrease in cost of transportation. These two conditions are interdependent, since the lessening in the cost of freight could not have come about without the transport of enor- mous tonnages. In no other part of the world has there been such a complete system of handling material worked out on such a gigan- tic scale; the steam shovels in the mines, the railroads to the ports, the mammoth docks and arrangements for loading vessels in a few hours, the special fleet of ore carriers, the improvement of the locks, the unloading machinery at the lower lake ports, and the storage yards and handling apparatus at the Eastern furnaces, each one of these is a link in a chain of specialized machinery, by which it has become possible to transport ore a thousand miles and make pig-iron' for less than half a cent a pound. Table XXII-J shows the production of the different ranges in 1901, and gives figures for comparison with the other large pro- ducers. The three States of Michigan, Wisconsin and Minnesota, constituting what is known as the Lake Superior region, raised 21,445,903 tons of ore. The only competitor is the Minette dis- trict of Germany, France, Belgium and Luxemburg, which mined THE UNITED STATES. 649 17,000,000 tons, while Northern Spain raised less than half as much, its output being only 7,740,000 tons. TABLE XXII-J. Sources of American Ore Supply in 1901. U. S. GEOL. SURVEY. Lake Superior Ranges. Location. Date when opened. Output ; tons. Fe. P. S. SiO 2 . CaCO 3 . H,0. Mesabi N E Minn 1892 9 303 541 61-64 .03-. 08 .01 3-5 5 8-12 Menominee.... Marquette Gogebic Vermilion Total L ^u N.W.Mich. N.W.Mich. N.W.Mich. N. E. Minn. 1877 1855 1884 1884 3,697,408 3.597,089 3,041,869 1,805,996 56-62 60-67 58-62 61-67 .01-. 75 .02-. 15 .04-. 08 .04-J5 .01 .02 .01 tr. 36 2-6 3-7 3-5 1.0 0.5 0.3 0.4 5-10 1 12 10-12 1-6 perior . 21,445,90S Other States. Other States. Alabama 2 SOI 732 401 98 Pennsylvania . . . 1 040 6b4 Rocky Mountains 12 469 638 14 024 673 18 251 804 1 9 059 393 20 589 237 Lake Erie Receiving Port Ashtabula 3 001 Q14 2 684 563 3 341 526 3 700 486 3 981 170 Cleveland 2 456 704 2 645 318 3 223 582 3 376 644 3 831 060 Conneaut. . . 495 327 1 404 169 2 320 696 2*556' 631 s' 181 019 Buffalo. . . . 797 446 1 075 975 1 530 016 1 616 919 1 475 386 Erie 1 311 526 1 092 364 1 309 961 1 240*715 1 379*377 Loraiii 355 188 536 086 l'l!2 946 l'090'235 721 662 Fairport 1 008 340 912 879 1 241 013 1 085 554 1 181 776 Toledo 416 438 414 012 792 348 645 147 798 298 Huron... 198 231 126 755 263 600 qoi Q-M 431 311 Sandusky 79 792 136 200 87 499 154 542 33 017 Total 10 120 06 11 028 321 1 5 222 187 1 5 7Q7 787 17 014 076 On docks Dec. 1 5 923 755 5 136 407 5 530 283 . K. 1886 1V> 074 112 t 1887 94 240 94 2-0 1888 206 061 206 0")1 1889 260 291 260 21 1890 363 842 363 842 1891 264 262 264 262 1892 335 236 6 418 341 654 1893 .... 337 155 14 020 351 175 1894... 156,826 156 826 1895 307 .503 74 991 382 494 1896 298 885 114 110 412 995 1897 248 256 206029 454 285 1898 . . 83 696 84 643 168 339 1899 161 783 215 406 377'l89 1900 154 871 292 001 446 872 1901 199 764 334 833 17 651 552 248 Total 3,690,756 1,322,013 20438 17 651 ' 5050858 Total to foreign ports 70 160 Aver composition of cargoes. Fe .. 57 00 63 30 65 85 62 80 s 288 092 037 211 p .... 025 032 015 036 The most radical change however was in placing the molds on trucks ready for casting, these trucks with the molds being then taken to the rolling mill while the steel is solidifying. A me- chanical stripper then removes the molds from the ingots in close proximity to the heating furnaces, all the exhausting labor of the "pit" being abolished and the ingots charged hotter in the rolling mill furnaces. The consumption of fuel for heating at Sparrow's G84 THE IRON INDUSTRY. Point has been as low as 20 pounds per ton of ingots rolled. This arrangement of casting on trucks, which was first put in operation here, is now the standard construction not only in America but in the most progressive plants of Europe. A minor novelty in this plant, but an advance in line with more recent progress, was the installation of the Bessemer blowing engine near the blast furnace boilers in order to use the excess power developed at the smelting plant. During the last few years the Maryland Steel Company, or, as it is often known from its location, "Sparrow's Point," has furnished a great proportion of the rails exported from America. This is quite a natural result of its situation, and also of the fact that the United States Government exacts no duty on the iron ore which goes into articles of export. Following is a statement showing the amount of steel rolled in the last four years with the amount of material exported. There is also given in Fig. XXII-M a plan of the rolling mill at Spar- row's Point, while Fig. XXII-K gives a cross section of the Bes- semer plant at Steelton, Pa., showing the above described method of casting on trucks as applied at a later time. 1898 1899 1900 1901 Production 130,804 225,645 225,618 277,853 Exported 63,972 85,976 102,254 83,673 Per cent, export 48.9 38.1 45.3 30.1 ' SEC. XXIIj. Cleveland: It has been shown that the supply of ore for the furnaces of Pennsylvania comes down the Great Lakes and is unloaded at ports on the southern and eastern shore of Lake Erie. It is quite evident that a furnace at the port of entry will have no land freight to pay on the ore, and will haul less than one ton of coke, while the fur- naces near the fuel must haul ! 2 / 3 tons of ore. The proposition is quite simple from a mathematical standpoint, but a glance at the map will show that there are some circumstances which disturb the calculations, for a position on the shores of Lake Erie does not increase the sphere of commercial influence as much as might l:e expected. On the north the tariff of Canada, as well as her limited needs, bars the way, while on the west is the competition of Chicago. There is no reliable communication eastward; the falls at Niagara have given rise to two canals, one on American territory to New York by way of the Hudson Eiver, and one in THE UNITED STATES. 685 Ul I o o f (536 THE IRON INDUSTRY. Canada, the Welland Canal, connecting with the St. Lawrenue. Great sums have been spent by Canada to create an economical way of shipping by water from her western provinces to the ocean, but she is struggling not only with a commercial but a political com- plication. The navigation of the St. Lawrence from Quebec to Montreal is not satisfactory, but the latter place will not allow Quebec to get all the trade. Consequently much money is spent to improve the river channel which can be used only a part of the year, when there already exists a subsidized government rail- way which can carry the freight to Quebec at less cost. The same condition exists to some extent in the United States, where the people are urged to make a ship waterway out of the present Erie Canal, when the interest on the money needed to do this would probably pay the freight by railroad on all the material brought down. In both the case of the Canadian and American canals there is the serious objection that traffic is entirely suspended for three or four months in the winter, while in the case of the St. Lawrence River there is the additional disadvantage that the navi- gation of the lower bay for several hundred miles js very dangerous on account of the prevailing fogs. Of late years the question of marine insurance has become a serious matter. All of these matters have an important bearing on the question of locating a steel plant on Lake Erie, as proven by the stress laid on water transportation by canal and by the St. Lawrence when each new project is started. These objections., however, arc by no means prohibitory. The advantages are self-evident, nnd it may be said that the trend of new enterprises is toward this district. One of the first to make the journey was the Lorain Steel Company. There had been for some years a rolling mill near Johnstown, Pa , which bought blooms from the Cambria Company and made rails for street railways. A new company was formed and a new works built near Cleveland, equipped not only for street or "girder" rails, but for standard rails, a complete blast furnace and Bessemer plant being erected on entirely new ground. Since that time Lorairi ha> been one of the centers of steel production in the United States. It divides with Steelton the work of making all the rails and most of the equipment for the street railways of the United States, and both of these plants have taken a part in foreign trade in this line of work. The more immediate vicinity of Cleveland has played a very THE UNITED STATES. 687 Important part in the steel industry of this country for a long period. The Otis Steel Company was one of the pioneers in the- manufacture of open-hearth fire-box steel, and its name has been known all over the land. The Cleveland Boiling Mill Company was a factor in the rail situation twenty years ago, but has long since turned its product into different forms of special work, it being one of the largest producers of wire rod in the country. SEC. XXIIk. Colorado: The only great iron or steel producing district west of the Mis- sissippi Eiver is centered in the Minnequa Works at Pueblo, Colo., but its tributary mines cover an area which would overshadow a European empire. The Colorado Fuel and Iron Company owns over 30 mines in the State and 5 mines in New Mexico. The coke used at the steel works all comes from Southern Colorado, about 90 miles from Pueblo, the coal containing about 30 per cent, of volatile matter, and occurring in beds about 6 feet thick. It is washed and then gives a good hard coke containing about 16 per cent, of ash. The steam and gas coals are brought about 50 miles. In Colorado can be found coals of every description from anthracite to lignite, the beds having been exposed to severe geologic disturb- ances and to the heat of numerous volcanic intrusions. The iron ore comes mainly from three sections. At Sunrise, Wyo., 350 miles from Pueblo, there is an enormous deposit of red hematite running as high as 62 per cent, in iron, which can be mined with a steam shovel. At Fierro, N. M., 600 miles from Pueblo, is a large deposit of hard magnetic ore running up to 61 per cent, in iron. At Orient, Colo., which is 125 miles from the works, is a deposit of easily reducible limonite containing about 50 per cent, of metallic iron. All of these ores are well within the Bessemer limit of phosphorus. At Leadville, about 100 miles away, there is a deposit running about 30 per cent, in manganese and in Eastern Utah, about 400 miles distant, one with 50 per cent, of manganese. The spiegel for the steel plant is smelted at the Minnequa plant at Pueblo. A glance at the map will show that this district is protected by a great distance, and a consequent high transportation charge, from the competition of Eastern works, and that it.has an enormous area as its natural market. Unfortunately, most of this country is very sparsely settled and contains few industrial centers, but with the constant westward trend of population, the wants of railroads and THE IRON INDUSTRY. of miscellaneous users have increased, and there is a demand not only for a large works but for the local production of a large variety of finished articles. In answer to this demand very extensive improvements, amount- ing practically to a new plant, are now under way at Pueblo, and when completed there will be five blast furnaces, a Bessemer plant equipped with two 15-ton converters, an open-hearth plant with six 50-ton basic furnaces, one 40-inch blooming mill, 24-inch re- versing structural mill, rod, sheet, tin plate, wire and nail mills. SEC. XXIII. Eastern Pennsylvania: In addition to the Steelton district, already described, there are several seats of industry which should be mentioned in the eastern portion of Pennsylvania. Up to the present year the city of Scran- ton was the center of two old established plants concentrated under one management, and they were a very considerable factor in the rail trade. The whole plant is now being moved to Buffalo, N". Y., where it can receive Lake Superior ores without any charge for railroad transportation. The Scranton Company owns a consider- able share in the Cornwall ore property and will make iron at thik latter point and transport it to Buffalo for remelting to mix with the iron from lake ores. The Bethlehem Works was formerly one of the great rail pro- ducers, but has not rolled rails for many years. It is now engaged almost exclusively in making open-hearth steel forgings and has the most complete plant in the country for this work. It divides with the Carnegie Steel Company the work on armor plate for the war vessels of the United States, and turns out guns and shafts of the largest size. This plant is now enlarging its open-hearth de- partment and intends to remodel its old rail mill and enter the field as makers of angles, ship shapes and other structural material. In the neighborhood of Philadelphia are the Midvale Steel Com- pany and the Pencoyd Works, the Phcenixville Iron and Steel Com- pany and the Tidewater Steel Company. The first of these does a large amount of work in the line of special steels and forgings, while Pencoyd and Phcenixville are especially known as bridge and structural shops, making all forms of structural materials for their own use and also for the outside trade. The Pencoyd Works came into general notice beyond the boundaries of the United States on account of its delivery of the well known Atbara bridge in the Soudan. THE UNITED STATES. 689 There are a large number of blast furnaces scattered throughout Eastern Pennsylvania, mainly in the Lehigh and Schuylkill val- leys, and a very considerable amount of pig-iron is made. Most of this goes into the general foundry trade, but some is used in the neighboring steel plants. During recent years these furnaces have quite generally used the ores of Lake Superior with Connellsville coke. In the neighborhood of Chester, Pa., not far from Philadelphia, there is a marked concentration of steel-casting plants, this being one of the greatest centers in this line of work, while Coatesville, Pa., is prominent for its plate mills. In Table XXII-A I have divided Eastern Pennsylvania in a way somewhat different from that followed by Mr. Swank. He has always put the Schuylkill Valley separate, but has not included Philadelphia, which lies on both sides of this river. I have com- bined, under the title of Southeast Pennsylvania, the plants of the Schuylkill Valley with those of Philadelphia, Chester and Delaware counties. This is a logical arrangement and brings out more forcibly the importance of this region as a producer of iron and steel. SEC. XXIIm. New Jersey, New York and New England: On the shores of Lake Champlain and in the northern basin of the Hudson Eiver there are very considerable deposits of magnetite, which played quite an important part in the early history of the American iron industry, being the base of supplies for the Bes- semer plant formerly operated at Troy, N". Y. Owing to the lack of fuel it was necessary to transport either coke or anthracite coal from Pennsylvania, and with the advent of cheap Lake Superior ores the manufacture of steel at this point was abandoned many years ago. An attempt was made in recent years to operate a basic Bessemer plant, but the conditions were not such as to warrant a continuance of the operations. Some of the rolling mills at Troy have been working on stock from Western steel works. The ores are rather difficult to mine, and the annual output has been decreasing save as a higher price in boom years encourages an abnormal activity, so that the amount raised in New York is only about one-third of the quantity turned out twenty years ago. There are many large beds besides those already developed, but nearly all the ores of this district contain a very considerable proportion of titanium, which gives trouble in the blast furnace as well as in THE IROX INDUSTRY. ihe Bessemer vessel, on account of the infusibility of slags contain- ing titanic acid. This substance is so seldom found in prohibitory quantities in iron ores in other districts that prospectors and in- vestors have many times sunk large sums of money in properties which have proved worthless. This line of magnetic deposits ex- tends in a southwesterly direction across the northern portion of New Jersey and into Pennsylvania, where it appears as the Corn- wall ore hills. The character of the ore varies very much through- out its length, its main point of resemblance being in its magnetic property, the titanium being entirely absent in the more southern fields. A great many mines have been worked in New Jersey in years gone by, but either from the exhaustion of the deposits or from the inferior quality or from the high cost of mining, many of them have ceased operation, so that the amount now produced in the State is only half what was raised in 1880. Taking the whole magnetic field from Northern New York to Southeastern Pennsylvania, it may roughly be said that the Cornwall deposit, which is described under the Steelton district, produces half the total, while New York and New Jersey divide the remainder with an annual production of about 300,000 tons each. The iron made in these two States enters to a limited extent into the steel industry, some of it being sold to open-hearth fur- naces, but most of it is used in the general foundry trade. Very much money has been spent on electric concentrating plants throughout this whole region, the most extensive outfit having been erected in -Northern New Jersey by Edison, who spent several years in experiments. The ore used by him contained only about 18 per cent, of iron and was a hard compact rock, so that the expense per ton of finished concentrate was very heavy. The operation of bricking was not entirely satisfactory and the whole work was dis- continued about two years ago, but in other places less ambitious installations have been worked with more or less success from time to time. Some of the steel plants of this district are of considerable im- portance, although some are legacies from the days when the East held the supremacy in the iron trade. (In the iron world the term "East" means the region along the Atlantic seaboard east of the Allegheny Mountains.) A few kept pace with modern im- provements, but in no works east of Pennsylvania is there to-day a complete plant of blast furnaces, steel producers and rolling mills, THE UNITED STATES. 691 neither is there a Bessemer converter in steady operation, the works being engaged principally in the production of specialties or of supplying the local markets with structural and other material. Table XXII-P gives information concerning the distribution by States. TABLE XXII-P. Iron and Steel Plants in New England, New York and New Jersey. Blast Furnaces. Bessemer Plants. Open Hearth Plants. State. Coke. Char- coal. Works having standard con- Works having special con- No. of works. No. of furn- aces. Works making crucible steel. Works having rolling mills verters. verters. 1 3 1 4 14 1 7 Rhode Island. . . . 1 2 2 4 1 1 1 2 5 New York 16 3 * 6 11 3 21 New Jersey 11 4 9 5 17 Total 27 10 2 16 37 11 53 *The Troy works is idle. CHAPTER XXIII. GREAT BRITAIN. SEC. XXIIIa. General "View: As far as the coal and iron industry is concerned, the term Great Britain may be considered to embrace only England, Wales and Southern Scotland. These divisions of the Empire cover about 88,000 square miles, an area almost exactly the same as that cov- ered by the States of Pennsylvania and Ohio combined. The pop- ulation of this island, however, is from three to four times as great as that of these two States, while the pig-iron production in 1900 was about the same, the output of the blast furnaces of Great Britain in that year being 8,960,000 tons, while Pennsylvania and Ohio made 8,837,000 tons. In 1901 Great Britain fell to 7,761,- 000, while Pennsylvania alone made 7,343,000 and Ohio 3,326,000 tons. In both cases a great part of the ore was brought a long distance by water, to England by the ocean, and to Pennsylvania by the Great Lakes ; but Great Britain was compelled to find a for- eign output for nearly half her product, while the home demand in America offered a market for all except a small portion of the output. In Fig. XXIII-A are shown the districts into which the country may conveniently be divided. The statistics of output as given in the figure and in these pages do not agree with the reports of the British Iron Trade Association because I have taken the data from the Home Office Reports, which are published later than the Asso- ciation Reports and are made up from the sworn statements of the manufacturers. The difference is over one hundred thousand tons in the total production of pig-iron. It must be kept in mind that the figures shown in the enclosures embrace the surrounding dis- trict. The enclosure in Durham represents also Northumberland, and the latter division raised one-quarter of the coal credited to the two counties. The lack of room makes it difficult to locate the squares upon the map exactly as statistics would require; it must 692 GREAT BRITAIN. 693 therefore be remembered that Barrow is in Lancashire, and hence the product of the Barrow Steel Works is included in the enclosing lines shown in the southern portion of the county where there was EMLAXD AO WALES [[|[ SCALEC[F MILES 10 10 20 30 40 50 STATISTICS OF PRODUCTION} 1 Unit = WOO Tons per Year. Imports - ^ FIG. XXIII-A. room for the figures. The map is thus intended as a general guide, but not as an accurate scientific graphical diagram. The statistics shown on the map are for 1899, but the figures for 1900 are given 694 THE IRON INDUSTRY. in Table XXIII-B. The changes do not in any way alter the gen- eral outline of results, but in the later table I have tried to improve somewhat on the method of grouping. COAL FIELDS OF GREAT BPJTA1N 46 60 80 100 FIG. XXIII-B. Fig. XXIII-B shows a map of the coal fields of Great Britain, taken from an exhaustive treatise on the subject.* A glance will show that fuel is distributed widely throughout the island. More- * Les Charbons Brittaniques ; Loz6 ; Paris, 1900. GREAT BRITAIN. 695 over, most of the coal gives a good coke, that of Durham being especially noted for its quality. The Home Office reports show that in 1900 the total exports of coal were 44,089,197 tons, of which 18,460,070 tons came from the ports in South Wales, 15,315,091 tons from the ports on the Northeast Coast, and 7,377,094 tons from Scotland, these three districts supplying over 93 per cent, of all the coal exported. There were 985,365 tons of coke sent over sea, and of this South Wales contributed 112,918 tons, Scotland 131,273 tons, while the Northeast Coast shipped 624,317 tons, the product of Durham and Xorthumberland. The Durham district therefore supplies only one-third of the coal exported, but furnished five-eighths of the coke. The coal was shipped to all parts of the world, France taking the most, 8,314,697 tons; Germany next with 5,938,178 tons, Italy 5,115,125 tons, Eussia 3,116,099 tons, almost all to her northern ports ; while Belgium received 1,152,109 tons. The Pacific Coast of the United States took 34,880 tons, while even the Atlantic coast had 5,265 tons. The coke also was spread all over the earth ; thus out of a total of 985,365 tons, the best customer was Spain and the Canaries with 155,561 tons, probably as return cargo for the ore vessels; next comes Norway with 93,683 tons, Holland 89,293 tons, Northern Eussia 86,950 tons, Sweden 79,879 tons. Of the leading iron pro- ducing nations Belgium took 39,409 tons, Germany 44,444 tons, France 47,832 tons, Austria 10,203 tons, and the Pacific Coast of America 15,367 tons. The shipments to Spain and to Northern Eussia are important, since these two districts depend upon out- side sources for their fuel. It will be noticed that Holland re- ceived 93,000 tons, but it is quite certain by comparing the statis- tics of neighboring countries that this coke went mostly to Ger- many and Belgium. The same confusion is found in the reported exports of pig-iron. Similar serious errors are found in the record of American exports, where shipments to the interior of Europe appear against the port of entry. The steel industry of the country is largely dependent upon its supply of foreign ore. It was about 1865 that the imports of ore were worth mentioning, but according to Bell* they probably were not over 10.000 tons per year. In 1867 they had risen to 86,568 tons; in 1870 to 400,000 tons, and in 1880 to 3,000,000 tons. Some * Principles of Manufacture, p. 453. 696 THE IRON INDUSTRY. ore comes from Greece, Algeria, Italy, Sweden and other countries ; but 90 per cent, of the imported ore comes from Spain, where some of the largest English companies have their own ore prop- erties. This ore goes impartially to the north, south, east and west. Scotland gets one million tons a year; the West Coast re- ceives an equal quantity, and both the northeast district around Middlesborough, and Glamorganshire in the southwest, receive double this amount. Table XXIII-A shows the origin of the ores imported into the Kingdom in 1882, 1886, 1890, 1895, 1899 and 1900. TABLE XXIII-A. Imports of Iron Ore into Great Britain from Different Countries. 1882 1886 1890 1895 1899 1900 3 072,955 2,533,939 3,627.646 3,807.188 6,186 022 5,551,559 Greece 17 969 79007 193 353 319 759 304 648 Algeria 91 097 201 601 205670 162 525 231 361 141 624 80 904 105 193 98 055 Italy... 89231 35546 79 312 127 317 94 771 88 532 France . . 38 274 4H 165 Other Countries 31,663 33,543 W,6i(jS 79.024 79,198 65,383 Total. .. 3 284 946 2 822 598 4 031 265 4 450 311 7 054 578 6 97 963 Almost all this imported ore is transformed into acid steel either by the Bessemer or open-hearth processes. The native ores pro- duced in the Yorkshire North Hiding (the Cleveland district), in Lincolnshire, Staffordshire and elsewhere, go into basic steel, or wrought-iron, or into the general pig-iron supply. It must not be forgotten in studying the map that the distances are all small in comparison with those familiar to American conceptions. From the Scotch iron works south of Glasgow to the coal mines of Gla- morganshire in South Wales is less than three hundred miles, while across the island from the steel works at Barrow to the coke fields of Durham is only seventy miles. On this account the great works in England have arranged themselves not so much with rela- tion to their raw material as with regard to a market for their output and to subsidiary conditions. Cardiff and Glasgow bring ore across the sea to their coal beds, while Middlesborough brings the fuel to the ore, and Barrow pays freight on a part of both fuel and ore ; but in each of these cases the steel works is on tidewater, a most important factor in a nation that depends on foreign trade. GREAT BRITAIN. 697 In other cases there are local conditions, as in Staffordshire and South Yorkshire, where, during long years and even centuries, there have grown up industries like those of Sheffield and Bir- mingham that call for large quantities of steel and iron to be worked up into finished articles of commerce. In considering the short distances covered by raw material it is necessary to remember that freight rates are much higher in Eng- land than in America. The normal charge for carrying a ton of pig-iron from South Staffordshire to London, a distance of 120 miles, is from $2.40 to $2.90, and for carrying coke 100 miles from South Durham to Cumberland the rate is $1.80 per ton.* In the United States the rate on pig-iron from Pittsburg to Philadelphia, a distance of 353 miles, is $1.77. On coke between the same points it is $1.95. It will be found that the rate on coke is considerably over three times as high as in America, while on pig-iron it is four to five times as much. Both Scotland and Middlesborough have specialties in furnish- ing supplies to the great shipbuilding industries on the Clyde and on the Northeast Coast. The vessels launched in 1900 in Great Britain footed up about 1,500,000 tons, and we may make a rough estimate that this took about 500,000 tons of steel and iron. This would mean one-twelfth of all the wrought-iron and steel made in the Kingdom and the large share of this business goes to the two districts mentioned. In Table XXIII-B is given more detailed information concern- ing the distribution of the iron industry" in the year 1900. The statistics of steel output are taken from the report of the British Iron Trade Association, while the figures for coal, ore, blast fur- naces and pig-iron are from the Home Office Reports. The tables at the end of each section giving the number of blast furnaces, converters and open-hearth furnaces, are from a supplement of the Iron and Coal Trades Review, issued July 5, 1901. A slight but unimportant disagreement may be found in one or two in- stances between the two sources of information. In Tables XXIII-C, D and E are given the results of an inquiry into the history of the iron trade during the last twenty years. Through the courtesy of Mr. Swank, of the American Iron and Steel Association, of Philadelphia, I was able to get a file of the Home Office Reports from 1882 to 1900, with the exception of * Report of Commissioners of British Iron Trade Association, p. 95. 698 THE IRON INDUSTRY. 1885, and the absence of figures for that year in the following pages is thus explained. Under the separate districts I have given de- TABLE XXIII-B. Production of Coal, Ore, Iron and Steel in Great Britain in 1900. Data on Coal, Ore and Pig Iron from Home Office Reports; on Steel from British Iron Trade Association. District. Coal. Ore. Pig Iron. Blast Furnaces. Wrought Iron. Tons. P.C. Tons. P.C. Tons. P.C. Total. Active. Tons. P.C. No'h Yorkshire (Cleveland).. Durham and Northumberl'd Scotland 3,742 46,315,240 33,112,204 39,083,973 28,246,937 26,865,193 14.227.076 3,109,615 2.106,443 23,871,544 6,292,012 1,822,622 124699 21 14 17 13 12 6 1 1 11 3 1 5,493,733 19,124 849031 29,472 56,944 1,733,791 1,084,797 843 4,298,145 2,858 359,506 323 99,641 39 *6 '12' 8 31 3 2,136,584 973 010 1,156,885 841,528 290,601 1,585925 596,807 66,586* 636,653 561,626 113,486 24 11 13 9 3 18 7 1 7 6 1 82 39 102 68 28 81 75 5 47 54 17 6 65 30 82 30 I8- 60 39 4 32 43 8 4 198,000 17 206,000 18 South Wales... So'h Yorkshire West Coast Staffordshire. . . North Wales... Eastern Central 137.000 175000 379,000 31,000 12 15 33 2 Dt- rby and Not- tingham Other parts of. . England Other parts of 36,000 3 Ireland 1 Total 225,181,300 100 14,028,208 100 8,959,691 100 604 405 1,162,000 100 *Output in 1899. District. Production of Steel. Bessemer. Open Hearth. Bessemer and Open Hearth. Acid. Basic. Acid. Basic. Tons. P.C. Tons. P.C. Tons. P.C. Tons. P.C. Tons. P.C. 27 No'h Yorkshire (Cleveland).. Durham and Northumberl 'd Scotland 64,000 1.3 269,000 5.5 975,000 20.0 25,000 0.5 1,333,000 960,000 520,000 210,000 118,000 77,000 19.6 10.6 4.3 2.4 1.5 3,000 ' 47,000 39,000 14S1000 30,000 To 0.8 3.0 0.6 963,000 960,000 588,000 659,000 367,000 30000 20 20 12 14 7 100 South Wales.... So'h Yorkshire West Coast 440.000 250.000 502,000 9.0 5.1 10.3 " 8l',o66" "iie 142000 2.9 North Wales. . . Total 1,256,000 25.7 492,000 10.0 2,860,000 58.4 292,000 5.9 4,900,000 tailed statistics, but in the three tables just mentioned I have taken the average for four periods. The striking fact is hereby made plain that England in regard GREAT BRITAIN". 699 TABLE XXIII-C. Production of Pig-iron in Great Britain; one unit=1000 Tons. Data for 1830, 1860, 1870 and 1880 from Bell ; later figures from Home Office Reports. District. 1830. 1860. 1870. 1880. Average 3882 to 1884 incl. Average 1886 to 1890 incl. Average 1891 to 1895 incl. Average 1896 to 1900 incl. Northeast Coast . .... 5 659 1 627 2 416 2,666 2,642 2 638 3 194 West Coast 169 678 1 541 1 676 1,589 1 284 1 576 Scotland 37 937 1,206 1,049 1,081 922 826 1 128 South Wales 278 1 019 1 073 927 897 807 734 770 Eastern Central 8 75 386 434 505 494 641 Staffordshire 213 617 892 610 551 542 506 586 Central 18 126 180 367 435 388 417 521 South Yorkshire Shropshire 29 73 98 145 78 112 307 88 291 70 197 50 213 45 296 10 North W r ales 25 49 43 58 42 48 40 59 Others 166 69 48 108 Total 678 3,827 5,964 7,749 8,309 7,759 7,245 8,889 TABLE XXIII-D. Production of Iron Ore in Great Britain; one unit=1000 tons. Data for 1860, 1870 and 1880 from Bell ; later figures from Home Oflice Reports. District. 1860. 1870. 1880. Average 1882 to 1884 incl. Average 1886 to 1890 incl. Average 1891 to 1895 incl. Average 1896 to 1900 incl. Northeast Coast 1 484 4298 6 528 6 439 5416 4 700 5 639 Eastern Central 118 1 048 2 765 2 824 2 897 2 974 4 018 West Coast 990 2,093 2 759 2861 2 569 2 199 1 943 Staffordshire . . . 1 543 1 378 1 798 1 898 1 341 925 l'025 Scotland .... 2 150 3,500 2 664 2 172 1 226 785 887 Bristol Channel 828 865 534 314 197 160 120 Central 376 385 153 17 15 11 4 Shropshire . ... 166 338 227 226 92 51 g 256 308 287 172 78 72 fiC North Wales 85 59 43 3 1 1 32 99 268 253 193 178 330 Total 8 024 14 371 18 026 17 184 14 025 12 055 14 031 to her iron industry seems to be in a stationary condition. Her output of ore has decreased in the last twenty years, but is now in- creasing, owing mostly to the development of the lean ore beds of Eastern Central England, including Leicestershire, Lincolnshire and Northamptonshire. There has been a very decided increase in the amount of ore imported and the production of pig-iron has been thus sustained, but the second period shows a smaller product than the first, the third less than the second, and the great increase in the fourth period does not bring the output very far beyond the rate of production from 1882 to 1884. '00 THE IRON INDUSTRY. TABLE XXIII-E. Imports of Iron Ore into Great Britain at Different Ports. District. Average 1882 to 1884 iucl. Tons. Average 1886 to 1890 incl. Tons. Average 1891 to 1895 incl. Tons. Average 1896 to 1900 incl. Tons. Northeast Coast . . . 948,000 1,488,000 1,920,000 2,354 000 - 1,434,000 1,347.000 1,183,000 1,387,000 382,000 575,000 694,000 1 394,000 West Coast 294,000 317,000 166,000 882000 Others 11,000 15,000 15,000 31,000 Total 3 069,000 3,742,000 3,978,000 6 048,000 This stationary character in the total output is true of almost every district, the Eastern Central region being the only one that has increased its output of pig-iron to any notable extent. The other districts have held their own by importing more and more ore, and have maintained a remarkable regularity of tonnage, the order of precedence to-day being practically the order of twenty years ago. In this lack of development England stands alone. By referring to the tables in Chapter XXXIII it will be found that since 1880 Russia has increased her output of pig-iron 5.66 fold, the United States 4.14, Austria-Hungary 3.13, Germany 3.09, Belgium 1.68, France 1.58, Sweden 1.30, while England in 1901 made almost exactly the same tonnage of pig-iron as she smelted in 1880. The output of steel has been increased as follows : The United States 10.80 fold, Sweden 10.35, Germany 9.62, Austria-Hungary 8.46, Belgium 4.96, Russia 4.94, France 4.03, and England 3.57. It is not my purpose to discuss the causes why England has stood still, but it is necessary to keep the plain fact in mind and to know that it is as true of each district as for the whole country. SEC. XXIIIb. The Northeast Coast: I am indebted to Mr. Arthur Cooper, manager of the Northeastern Stfel Works, for a careful reading of this section. The Northeast Coast is the great iron and steel producing dis- trict, making more than one-third of all the pig-iron and more than one-quarter of all the steel of the Kingdom, and nearly one-fifth of all the puddled iron. Middlesborough is the center where the coke of Durham meets the ore from Spain, or from the Cleveland Hills, and the finished steel finds an outlet either in the shipyards along the Tees, or by water to other ports of the Kingdom, or of GREAT BRITAIN. 701 other countries. The Cleveland beds produce 40 per cent, of all the ore raised in the island. This is all smelted in the immediate neighborhood of the mines, and in the annual report of C. E. Muller & Co. of January 14, 1902, it is stated that out of a total of 79 blast furnaces in operation in the Northeast in 1901 there were 43 smelting Cleveland ore, the others presumably being on imported material. A small proportion, about one-seventh, of the Cleveland iron is converted into steel, mostly by the basic Bes- semer process, but almost all of the steel made in the district is DURHAM COAL FIELD 'E A ^ Newcastle) 'Gateshead Durham ! 'Bishop # mth rath Shield^ mderland Norton c tockton m rest Hartlepool x ilbuisn Inborn FIG. XXIII-C. made from Spanish ore. The Cleveland deposit is not rich enough in either phosphorus or manganese to give a proper iron for the basic Bessemer, and it is necessary to add to the burden a certain proportion of other ores which are richer in these elements; con- sequently most of the product goes into foundry and forge pig for use both at home and abroad. The output of Middlesborough fur- naces, especially those of Bell Brothers, forms the foundation of foundry practice throughout the northern part of the continent; it is often used alone, but is mixed with iron of lower phosphorus 702 THE IRON INDUSTRY. to make the better class of castings. On another page, in the dis- cussion of the ore deposits of Lincolnshire, Leicestershire and Northamptonshire, further remarks will be made on the recent de- velopments in the lean ore deposits of England. Fig. XXIII-C shows the relation of the coal field of Durham to the district around Middlesborough, while Fig. XXIII-D shows the Cleveland ore deposits.* . Q Wynyard Seaton Carew CLEVELAND ORE BEDS ESTON Ormesby Nonnanby MINES N FIG. XXIII-D. The Cleveland ore is a carbonate and the composition is given by Kirchhoff as follows : Per cent. Protoxide of iron 35.37 Peroxide of iron 1.93 Protoxide of manganese 1 .00 Alumina 6.95 Lime 6.63 Magnesia 3.73 Silica 10.22 Carbonic acid . 22.02 Sulphur Phosphoric acid Organic matter . Moisture . . 0.1 Of 1.15 1.20 9.80 * These maps are taken from certain letters written by C. Kirchhoff for The Iron Age, of which he is editor, and he has kindly granted permission for their reproduction here. I am also indebted to the same letters for much informa- tion concerning this district t From other sources of information I believe that the average content of sulphur is nearer 0.25. GREAT BRITAIN. 703 Metallic iron 28.85 Phosphorus 0.50 Loss by calcination 29.58 " Iron in calcined stone 40.96 The composition of calcined stone is given by the same writer as follows : Per cent. Peroxide of iron 59.77 Oxide of manganese 0.99 Alumina 9.28 Lime 9.23 Magnesia 5.41 Silica 13.66 Sulphur 0.12 Phosphoric acid 1.41 Total 99.87 Metallic iron 41.84 Phosphorus 0.62 The ore varies considerably in different parts of the field, the above being a fair average. In many cases the content of iron .is less and there is consequently a greater proportion of silica and earthy matter so that a larger quantity of fuel and stone is re- quired. For this reason considerable differences in practice and in cost will be found between furnaces very near together in Mid- dlesborough. The ore deposit, at its northern edge, sometimes contains as much as 32 per cent, of iron and in exceptional cases even 33 per cent. The thickness of the bed is also greatest at this point, measuring 15 feet 7 inches at the mines of Bolckow, Vaughan & Co. Toward the south it grows thinner and at the points marked with a cross upon the map it divides into two seams of about four feet each. The quality also falls, and at the extreme outcrop at Whitby there is only 25 per cent, of metallic iron. The ore is calcined to expel carbonic acid, and this removes also the water and organic matter, so that the roasted product contains about 40 per cent, of iron. The figures above quoted from Kirch- hoff give 41.84 per cent, of iron and 13.66 per cent, of silica. Information from other sources leads me to think that the figures quoted are rather roseate and refer to the best records rather than to the average supply. I have been told that the general run of ore after calcining will carry only 40 per cent, of iron with silica up .to 19 per cent. 704 THE IRON INDUSTRY. The average selling price of this ore from 1870 to 1883 is given by Bell as $1.02 per ton at the mines, with 30 cents freight, mak- ing a total of $1.32 per ton delivered at the furnace. The value in 1899 is given in the Home Office Reports at $1.01 per ton at the mine. Counting a short haul and the cost of calcining, it can hardly be less than $1.15 per ton for a 30 per cent, ore; this is 3.83 cents per unit, and if the Cleveland pig contains 92 per cent, of iron, the cost of the ore per ton of pig will be $3.52. Kirchhoif gives the cost of the ore delivered at the furnaces of Bolckow, Vaughan as 85 cents per ton, to which must be added the cost of calcining. For a 30 per cent, ore this means a little over 3 cents per unit or about $3.00 per ton of pig-iron. The composition of the coal from Durham varies somewhat ac- cording to the seams from which it comes, but the beds are very much alike and the coals are often mixed. The average of four samples quoted by Bell is as follows : Per cent. C 80.51 H 4.49 O+N 8.03 S 1.26 Ash 5.16 Water 1.01 100.46 The fixed carbon was 70.32 per cent, and the loss in coking is given by Bell and by Kirchhoff as usually over 40 per cent, in bee- hive ovens. By far the greater quantity of Durham coke is made in this type of oven, although progressive works in Middlesborough are now introducing the by-product process. Bell states that the coke runs 6.60 per cent, in ash and 0.96 per cent, in sulphur. Kirchhoff gives the detailed composition of four samples, an aver- age of which is as follows : Per cent. Carbon 88.16 Sulphur 1.11 Ash 9 33 Water 1. 40 100.00 The distance from the mines in South Durham to the furnaces in Middlesborough is from 20 to 30 miles, and the freight is about 50 cents per ton. GREAT BRITAIN. 705 The coke is hard and strong and is in demand abroad, very con- siderable quantities being exported. Over 60 per cent, qf all the coke sent abroad by England in 1900 was shipped from the North- east Coast. There were also heavy shipments of coal, the propor- tion coming from this district being one-third of the total exports. As above shown, the ash in Durham coke is considerably less than is found in some other first-class cokes and this decreases to a slight extent the amount of silicious material entering the blast furnace. The amount of fuel needed for a ton of Cleveland iron is given by Bell as 1% tons, and in exceptional cases it may be lower, but from information received from most excellent authority, I believe this is more often the hope than the actuality. Taking the whole cam- paign of the furnace and considering the amount actually paid for on board cars, there are probably few furnaces at Middlesborough getting along with less than 1*4 tons, and there are many using more. The cost of this coke is given by Kirchhoif as $1.82 to $2.20 per ton at the mines, and the cost therefore at the furnaces at Middlesborough will be from $2.30 to $2.70 per ton. The sell- ing price is considerably above this, running from $3.15 to $3.50 per ton. When smelting the Cleveland iron stone, the amount of lime- stone necessary varies with the character of the ore. Bell gives the amount needed as 1175 to 1350 pounds per ton and the cost as 80 cents per ton delivered at the furnace. The cost of stone under these conditions would be from 43 to 49 cents per ton of iron. Kirchhoff gives "also about 1300 pounds of stone per ton of iron, but gives the cost of the stone at $1.20 per ton, making an item of about 70 cents per ton. My own information from authoritative sources agrees with the amount of stone above given, but Cochrane, in a detailed investigation of Cleveland practice and the use of lime, shows a consumption of about 1600 pounds. In this case, however, the ore contained only 26.9 per cent, of iron. From an- other source I have been given the figure of 1900 pounds of stone at a cost of $1.10 per ton of stone, representing about 95 cents per ton of pig-iron. We may therefore estimate the cost of Cleveland pig-iron for those who own their own coal mines and ore beds, counting noth- ing for the money invested, and also the cost for those who do not own their own supplies. 106 THE IRON INDUSTRY. Minimum. Fair Complete. practice. Per ton Pig-iron. ownership. Market prices. Fuel 1% tons @2.40 $2.70 " 1% tons @3.30 $4.10 Stone 1300 Ibs 70 .95 Ore 3.00 3.50 $6.40 $8.55 If we add to these items 60 cents for labor and 25 cents for sup- plies, which are figures given by Kirchhoff, we "have a total of $7.25 for the best managed and best equipped plants owning their own coal and ore mines, and $9.40 for other plants buying their raw material and using somewhat more fuel. There are still other works which show a considerably higher cost. In these totals are- not included the item of general expenses and administration, andi it does not include the interest and depreciation account, so that they by no means represent the actual cost of making pig-iron in Cleveland. They may, however, be compared with many similar calculations where the cost of pig-iron in different localities -is con- fidently predicted, as in such cases these latter items are always ignored. It may also be pertinent to the question to record that the selling price of Cleveland iron in the winter of 1900-01 was $11.20 per ton, and there is no reason to suppose that money was lost in the transaction. Thus it is quite clear that Cleveland iron can be made cheaply, but it is also true that it is an undesirable metal. It contains so much phosphorus that it is hard to use in a basic open-hearth fur- nace, although it is perfectly certain that it can be so used. On the other hand it contains so little phosphorus that it is not well fitted for the basic Bessemer. In order to get iron for the basic converter it has been customary to enrich the phosphorus content by adding a certain proportion of puddle cinder, and to raise the manganese by using manganiferous imported ores. With the di- minution of the supply of puddle cinder it is necessary to use a certain amount of basic converter slag in the blast furnaces, and no matter what the mixture may be, the silicon must be kept low, thus requiring a very large amount of lime to flux the high silica in the ore. Taking everything together, the cost of making iron fit for the basic converter is given by Kirchhoff at from $1.00 to $1.50 per ton above the figures just recorded for the ordinary product. For open-hearth work the manganese is not necessary GREAT BRITAIN. 707 and the phosphorus an injury. It would seem therefore as if a cheap iron could be made for this purpose, while the phosphorus might be lessened if necessary by mixing with foreign ores. The price of Spanish ore in the winter of 1900-01 was about $2.61 at Bilbao, with the low ocean freight of $1.03, making a total of $3.64 per ton at Middlesborough. As the ore contains about 49 per cent, of iron this gives a cost of 7.43 cents per unit, or about $7.06 per ton of iron. The assumption that the ore con- tains only 49 per cent, of iron may seem rather pessimistic, but the decrease in the quality of the Spanish ores has been a serious matter. This subject was discussed in 'the presidential address of William Whitwell before the Iron and Steel Institute, and he gave the composition of Eubio ores as imported at Middlesborough in 1890 and 1900. The comparison is as follows : 1890 1900 Fe dry 55.50 52.80 Water 9.00 9.10 Fe as received 50.50 47.99 Silica 7.10 10.09 The ocean freight on ore is usually 30 cents higher than the figures just given, which would make the ore cost $3.94 per ton, or a trifle over 8 cents per unit, or about $7.60 per ton of iron. The silica in this ore runs about one-half as high as in the Cleveland stone, and the quantity of limestone needed is much less, and the amount of fuel will be about 0.95 tons per ton of pig-iron. The cost therefore of the ore, fuel and stone for a ton of hematite pig- iron will be as follows: Low freight. Usual freight. Ore $7.06 $7.60 Coke 2.66 2.66 Stone (about) 50 .50 $10.22 $10.76 Adding to this the same amount for labor and supplies as in the case of Cleveland iron, viz., 85 cents, we have the cost of hema- tite iron from $11.10 to $11.60, not reckoning the items of general expense or interest. In the winter of 1900-01 the selling price was about $13.85 per ton. The most important steel works on the Northeast Coast are given in Table XXIII-F. The works of Bell Brothers have not been large producers of steel in the past, but they have lately put in an extensive open-hearth plant. Fig. XXIII-E shows a plan 708 THE IRON INDUSTRY. O O W g I g XI GREAT BRITAIN. 709 of the works of the Northeastern Steel Company, at Middlesbor- ough. In Tables XXIII-G and H are given data concerning the industrial history of the district. TABLE XXIII-F. Iron and Steel Plants on the Northeast Coast. Name of Works. Location. Blast. Furnaces. Bessemer Converters. Open Hearth Furnaces. Acid. Basic. Acid. Basic. Middlesbro'. ... 30 4 4 6 4 10 Durham 7 27 11 10 8 23 6 ""2" 6 Middlesbro' Palmers Shipbuilding Co Jarrowon Tyne. 5 Annstrong,Whitworth & Co. (Elswick) Newcastle 12 4 8 6 51 Edw Williams 8 123 4 10 107 8 TABLE XXIII-G. Production of Ore and Pig-iron and Imports of Ore on the North- east Coast. See also Tables XXIII-C and XXIII-D for data before 1882. Year. Ore Raised. Ore Im- ported. Pig Iron. North Yorkshire. Dur- ham. Total. North Yorkshire. Durham. Northum- berland. Total*. 1882.. 1883.. 1884.. 1885 6,326,314 6,756,055 6,052,608 83,726 50,248 49,091 6,410.040 6,806,303 6,101,699 1,098,000 951,000 795,000 834 000 1,803,508 1,867,329 1,725,823 815,671 823,659 779,131* 93,422 88,535 2,712,601 2,779,523 2,504,954 1886.. 1887.. 1888.. 1889.. 1890.. 1891.. 1892.. 1893.. 1894.. 1895. . 18%.. 1897.. 1898. . 1899.. 1900.. 5,370,279 4,980,421 5,395,942 5,657,118 5,617,573 5,128,303 3,411,400 4,625,520 5,048,966 5,285,617 5,678,368 5,679,153 5,730.413 5,612,742 5,493.733 1,759 2,506 40,233 3,991 11.488 7,715 9,275 5,372,038 4,982,927 5,436,175 5,661,109 5,629,061 5,136,018 3,420,675 4,625,520 5,051,645 5,304,681 5,697,645 5,696,005 5,751,281 5,629,702 5,512,779 1,015,000 1,475,000 1,426,000 1.660,000 1,864,000 1,569,000 1,522,000 2,061,000 2,410,000 2,037,000 2,360,000 2,337,000 2,283,000 2.457,000 2,330,000 1,735,885 1,841,444 1,856,274 1,915,050 1,961,328 1 769,492 1,333,656 1,943,404 2,088,299 2.058,279 2,209,074 2,134,507 2,095,131 2.211,222 2,136,584 700,836* 682,797* 774,984* 802,267 792,932 814,875 610,892* 770,510* 885,568* 867,878* 1,002,852* 1,063,134* 1,103,495* 1,040,174* 973,010* 2,436.721 2,524,241 2,631,258 2,782,466 2,837,599 2,631,183 1,944,548 2,713,914 2,973,867 2,926157 3,211,926 3,197,641 3,198,626 3,251,396 3,109,594 65,149 83,339 46,816 2,679 19,064 19,277 16,852 20,868 16,960 19,046 'Including Northumberland. 710 THE IRON INDUSTRY. TABLE XXIII-H. Imports of Iron Ore at Ports on the Northeast Coast. Year. Middles- borough. New- castle. North and South Shields. Stockton. Hartle- pool. Sunder- land. Others. Total. 1882. 1883. 1884. 1885. 1886. 1887. 1888. 1889. 1890. 1891. 1892 498,000 444,000 398,000 397,000 507,000 844,000 724,000 855,000 947,000 778,000 886000 307,000 295,000 179,000 179,000 240,000 276,000 357,000 374,000 513,000 388,000 242000 82,000 39,000 42,000 53,000 38,000 48,000 21,000 25,000 80,000 58,000 86000 114,000 56,000 64,000 69,000 89,000 116,000 116,000 125,000 129,000 148,000 123000 11,000 30,000 64,000 58,000 76,000 113,000 104,000 181,000 147,000 126,000 111000 64,000 79,000 47,000 77,000 64,000 69,000 95,000 97,000 47,000 71,000 74 000 22,000 8,000 1,000 1,000 1,000 9,000 9,000 3,000 1,000 1,098,000 951,000 795,000 834,000 1,015,000 1,475,000 1,426,000 1.660,000 1,864,000 1,569,000 1 522 000 1893. 1894. 1895 1,269,000 1,444,000 1 273 000 268,000 385,000 358000 89.000 149,000 154000 251,000 237,000 164000 90,000 86,000 46 000 94,000 108,000 42 000 "i',666' 2,061,000 2,410.000 2 037 000 1896. 1897. 1898. 1899. 1900. 1,391,000 1,193,000 1,103,000 1,334,000 1,251,000 345!000 319,000 325.000 300,000 252,000 2181000 413,000 352,000 377,000 402,000 190,000 235,000 252,000 206,000 258,000 120,000 94,000 178,000 151,000 116,000 94,'000 81,000 67,000 87,000 49,000 2,000 2,000 6,000 2,000 2,000 2,360,000 2.337,000 2,283,000 2,457,000 2,330,000 SEC. XXIIIc. Scotland (Ayrshire and Lanarkshire) : I am indebted to Mr. James Riley, formerly general manager of the Steel Company of Scotland and of the Glasgow Iron and Steel Company, for a careful review of this section. The iron industry of Scotland dates back about one hundred and fifty years, and has played an important part for half a century. It was well along in the last century before there was any appre- ciation or knowledge of the value of the blackband from the coal measures which at that time existed in great quantities throughout Ayrshire and Lanarkshire. This blackband was roasted and gave an ore making 63 per cent, of pig-iron, and it was raised very near the furnaces. In 1870 Scotland produced 3,500,000 tons of ore, but in 1880 this had dropped to 2,660,000 tons. Half of this was blackband, but the price had risen to $3.60 per ton at the pit. In 1900 only 597,826 tons of ore were raised from the coal measures, the price being officially given as about $2.40 per ton at the pit mouth, and this constituted 70 per cent, of all the ore raised in Scotland. The ore production in 1900 was less than 6 per cent, of the total for the Kingdom, while in 1870 it was about 25 per cent. The figures given on the map refer only to the counties of Ayr and Lanark, which produce two-thirds of all the coal and ore mined in Scotland, and smelt practically all the pig-iron, but in the tables I have used the totals for Southern Scotland. GREAT BRITAIN. 711 The pig-iron industry, in spite of the disappearance of the black- band and the importation of foreign ores to take its place, still retains a distinctive characteristic in the use of raw so-called "splint" coal in the blast furnace. The composition of good Lan- ark coal is as follows : Per cent. C 66.00 H 4.34 O+N 12.03 S 0.59 Ash 5.42 Water 11.62 100.00 Fixed carbon 53.4 This coal when charged in a raw state into the furnace will not fuse and get sticky, provided the furnace is not more than 70 feet high. The heating value of this coal is only about 80 per cent, of Durham coal, but counting the loss of fuel value in -the coking process, there is a slight advantage, ton for ton, in the Scotch coal charged in the furnace over the Durham coal, which must first be coked. When using this raw coal the furnace gases contain quite a quantity of hydrocarbons, and it is found profitable to put up scrubbers and collect the tar and ammonia before the gas passes to the boilers and stoves. The best beds of Lanarkshire coal are approaching exhaustion, and recently some plants have experi- mented in the making of a poor coke from the local coal and using it as a mixture with the inferior splint coals, but this practice seems to make no progress. A very considerable amount of coke is made in the Kilsyth district, but this is used for foundry pur- poses. The district of Ayrshire and Lanarkshire produces 9 per cent, of all the coal raised in the Kingdom, and exports large quantities. In spite of the great decrease in the supply of native ore, the production of pig-iron has been sustained by the use of Spanish ores, but there has been very little increase, the amount smelted having remained nearly constant during the last forty years. This statement concerning the stationary production of the district was questioned by Mr. Eiley, and I therefore append the statistics in Table XXTTT-T ; the figures prior to 1885 are taken from a paper by Mr. Eiley,* and the later data from the Home Office Reports. * Jour. I. and 8. I., 1885. 712 THE IRON INDUSTRY. TABLE XXIII-I. Production of Pig-Iron in Scotland. Period. Production per year. Inclusive. Tons. 1861 to 1865 1,122,600 1866 to 1870 1,089,800 1871 to 1875 1,021,600 1876 to 1880 993,600 1881 to 1885 1,084,400 1886 to 1890 922,217 1891 to 1895 826,128 1896 to 1900 1,128,161 Scotland now makes 12 per cent, of the pig-iron and 20 per cent, of the steel made in the Kingdom. As before stated, most of the ore is imported from Spain, and the pig-iron is used on the spot to make acid open-hearth steel for shipbuilding and other purposes. TABLE XXIII-J. Iron and Steel Plants in Scotland (Ayrshire and Lanarkshire )\ Name of Works. Location. Blast Furnaces. Bessemer Converters. Open Hearth Furnaces. Basic. Acid. Basic. Steel Co. of Scotland j David Colville & Sons (Dal- zell) Newton.... j Glasgow j 30 18 6 12 8 3 9 9 8 8 1 Glasgow I. and S. Co Wishaw 4 Flemington Giengarnock Clydebrldge . . Ayrshire Cambuslang. Mossend 12 4-10 tons Clydesdale Summerlee & Mossend Co. Mossend 7 Wm Baird & Co Scattered.... Coltness Scattered... 26 9 11 '26 Win Dixon . Total 95 4 111 1 Scotland makes only a small amount of Bessemer steel and hardly any basic open-hearth, but she makes as much acid open-hearth steel as Cleveland, each of them making one-third of all that kind of metal made in Great Britain. Table XXIII-J gives a list of the principal plants in Scotland. Most of the steel plants make plates and miscellaneous structural bars. In Tables XXIII-K and L are given certain items of statistical information; the importa- GREAT BRITAIN. 713 or ore come mostly to ports on the western shore, but a con- siderable quantity is brought to Grangemouth and other ports on the Firth of Forth. TABLE XXIII-K. Production of Ore and Pig-iron and Imports of Ore in Scotland. See also Tables XXIII-C and XXIII-D for data before 1882. Year. O re. Pig Iron. Raised. Imported. Ayr. Lanark. Total. 1882 2 404 177 385000 350423 775 577 1,126 000 1883 2 228 851 356 000 356 751 772 249 1 129000 1884 1*883*158 406000 292 287 695,713 988,000 1885 .. 487 000 1886 1506 731 418000 301,464 634,337 935,801 1887 ... 1*321 899 545000 301 652 630'588 932 240 1888 1 '238' 597 552000 320 374 707 400 1 027 774 1889.... l'06l'734 647000 320'654 657*549 978,203 1890 '998 835 714 000 240 848 4% 218 737066 1891 748336 360000 201 063 473 013 674 076 1892 872 435 840 000 276 788 695 705 972 493 1893 847*406 654000 246 939 546 116 793,055 1894 1895 631,304 824 673 598*000 1020000 182,546 326,454 459,697 722,320 [642,243 1 048 774 1896 .... 983 670 1 296000 360247 753,791 1 114038 1897 936 850 1 403 000 369 836 766 671 1 136 507 1898 824' 219 1*444*000 314302 748245 1 062*547 1899 843 585 1456000 345488 825342 1 170 830 1900 849,031 1,372,000 376,498 780,387 1,156,885 TABLE XXIII-L. Imports of Iron Ore at Ports in Scotland. Year. Glasgow. Ardrossan. Ayr. Troon. Otherg. Total. 1882... 251 000 55000 9000 14000 56000 385000 1883 280000 25000 7000 2000 42000 356 000 1884, 265,000 40,000 9000 3000 89000 406000 1885 292,000 15000 29000 5000 146000 487 000 1886 246000 8000 23000 2000 139000 418000 1887 303 000 15*000 34000 6000 187 000 545000 1888 323,000 38*000 17000 2000 172*000 552 000 1889 358,000 43000 44000 11000 191 000 647 000 1890 330,000 60000 91 000 31 000 202000 714*000 1891 241.000 17000 36000 4000 62 000 360 000 1892.... 516000 114000 59000 31000 120 000 840 000 1893 355,000 149000 59000 42000 49000 654 000 1894 302000 171000 36000 31 000 58 000 598 000 1895 521,000 252,000 80000 51 000 116000 1 020000 1896 589,000 410000 96000 77000 124000 1 296 000 1897 730000 438000 100 000 52000 83 000 1 403 000 1898 655000 487 000 15 000 71 000 79 000 1 444 000 1899 730000 402000 112 000 102 000 110 000 1*456 000 1900 698 000 372 000 92 000 117 000 93 000 1 372 000 714 THE IRON INDUSTRY. SEC. XXIIId. South Wales: In this district I have included Glamorganshire and the Eng- lish counties of Monmouth and Gloucester. It is in the latter that we find the ancient district still bearing the title of the Forest of Dean, which was once famous as an iron district, but which, in 1900, produced only 9885 tons of ore, no pig-iron being made in its borders. The iron industry of South Wales was founded on a local supply of lean clay band running about 30 per cent, in iron. In 1860 the above mentioned counties, together with two or three neighboring ones that are no longer producers, raised 830,000 tons of ore and in 1870 the amount was a trifle larger. From then the production rapidly decreased, being only about half as much in 1880, while now it is a negligible quantity. The production of pig-iron has remained nearly stationary from 1860 until now. Before the local ores failed the hematites of the West Coast were brought in, and then by almost providential dispensation the mines of Northern Spain were developed, and from that time South Wales has run almost exclusively on this imported supply. In former times the coal from certain districts at works near Merthyr was used directly in the furnace in the same way as in Scotland, but this practice has been discarded and a somewhat richer coal is now coked. The volatile matter in this coal is rather low, running from 16 to 22 per cent., and some seams contain 30 per cent, of ash, but, by washing, this may be reduced so that the coke contains only about 10 per cent, and very good results are obtained. The Spanish hematites imported at Cardiff in 1899 contained only about 50 per cent, of iron and from 7 to 14 per cent, of silica, but they were smelted with about one ton of coke per ton of iron. Some of the older iron works are situated in the interior, a legacy from ancient times, but new plants are being placed on tidewater, thus reducing the freight on both raw material and finished product. The northern shore of the Bristol Channel produced almost exactly the same quantity of steel in 1900 as Scotland. Unlike Scotland, half of the output is Bessemer, but like Scotland, it is all acid, both Bessemer and open-hearth. This district in 1900 raised 17 per cent, of all the coal mined in the island and fur- nished 42 per cent, of all the coal exported from the Kingdom, and 11 per cent, of all the export coke. It made about 9 per cent. GREAT BRITAIN. 715 of all the pig-iron and 20 per cent, of all the steel. The amount of puddled iron made is very small. This arises from the fact that Gas Producers Siemens Furnaces ^...^.^ **,**,**,. QFeed, ODDDDD oaQOIj[j ertical ru FIG. XXIII-F. DOWLAIS WORKS, CARDIFF, WALES. there are no cheap native ores and it does not pay to put iron from Spanish ores into puddled bar. 716 THE IRON INDUSTRY. Fig. XXIII-F shows a ground plan of the new open-hearth plant and plate mill of the Dowlais Iron Company at Cardiff, this being TABLE XXIII-M. Iron and Steel Plants in South Wales. Name of Works. Location. Blast Furnaces. Bessemer Con- verters. Open Hearth Furnaces. Acid. Basic. Acid. Basic. Blaenavon Iron Co CrawshayBros. (Cyfarthfa) Ebby Vale S. and 1. Co. ... Guest Keen & Co., form- 1 erly Dowlais Iron Co . . f Nettlefolds Blaenavon Merthyr Tydfil Ebby Vale Dowlais 9 9 { I 2 4 6 6 2 2 2 2 8 6 Cardiff. ... Tredegar 5 8 5 6 5 5 37 Landore 2 2 Pontardawe Steel Works Morriston. . Other open hearth plants.. 9 8 Other blast furnace plants. Total 69 84 TABLE XXIII-K Production of Pig-Iron and Imports of Ore on the Bristol Channel. See also Tables XXIII-C and XXIII-D for data before 1882. Year. 1882 Ore Imported. Pig Iron. Glamorgan- shire. Monmouth- shire. Gloucestershire and Wiltshire. Total. .481,000 .575,000 .247,000 .275,000 ,134.000 ,335.000 ,342,000 ,447,000 .478,000 ,091.000 ,170000 ,164.000 ,268,000 .221,000 ,297.000 1,554,000 910.000 404,350 384,128 378,275 530,084 522,135 473.116 934,434 906,263 851,391 1883 1884 :. 1885 1886 1887 268.828 310,000 424,681 426.854 416,874 424,533 420,710 444,356 454,363 447,715 464,486 470,443 319,280 397.768 457,448 446.259 399.538 407,848 336,033 263.297 236,089 254,551 '266,961 315,935 334,373 176,035 666,596 767,448 870,940 864,501 863,826 798,510 718,650 714,818 734,373 704,676 780,421 804,816 495,315 929,415 841,528* 1889 38.109 39,104 37,944 34,643 34,373 25,459 1890 . . 1891 1892 1893 1894 1895 1896 1897 1898 1900 1,704000 1,471,000 578,741 350,674 In the statistics of the Home Office for 1900 the product of Glamorgan- lire is combined with Denbigh, and Monmouth with Flint, both of which ombinations are questionable. To get the total for 1900 I have subtracted the output of Denbigh and Flint for 1899 from the total given for 1900 GREAT BRITAIN. 717 one of the best arranged plants in Great Britain. Table XXIII-M gives a list of the principal plants in the district, and Tables XXIIKN" and give certain statistics. TABLE XXIII-0. Imports of Iron Ore at Ports on the Bristol Channel. Year. Cardiff. Newport. Swansea. Others. Total. 1882 599 000 738000 144000 1 481 000 1883 656 000 749000 170000 1 575000 1884 481 000 627' 000 139000 1 247 000 1885 1886 1887 440,000 443,000 472000 673,000 571,000 742,000 159,000 113,000 112,000 3,000 7,000 9,000 1,275,000 1,134,000 1,335,000 10OU 544 ooo 694 000 104000 1 342000 1889 633,000 678,000 135IOOO 1,000 1417,000 1890 1891 647,000 486 000 778,000 469000 151,000 135,000 2,000 1,000 1,478,000 1,091,000 1392 583000 439,000 146,000 2,000 1,170,000 1893 644000 377,000 137,000 6,000 1,164,000 1894 ........ 640,000 448,OfO 178,000 2,000 1,268,000 1895 1896 653,000 655000 415,000 453 000 152,000 189000 1,000 1,221,000 1 297 000 1897 723000 620,000 210,000 1,000 1,554,000 1898 478 000 247 000 185000 910000 1890 832,000 619,000 251,000 2,000 1,704,(00 1900 780,000 435,000 255,000 1,000 1,471,000 SEC. XXIIIe. Lancashire and Cumberland: I am indebted to Mr. J. M. While, general manager of the Barrow Works, for reading the manuscript relating to this district. The county of Lancaster reaches across Morecambe Bay and includes Barrow-in-Furness and the Barrow Steel Works. It is in this detached portion of Lancashire and the neighboring portion of Cumberland that all the ore is raised and a great part of the iron and steel made. It is the custom, however, to keep the records by geographical rather than by natural lines, and the output of Barrow-in-Furness is combined with the output of South Lan- cashire and sometimes with that of Derby. This last named county produces no ore, but its output of both coal and pig-iron is about two-thirds as muoli as Lancashire. The figures on the map therefore give a somewhat wrong impression, as it would naturally be inferred that the ore of Lancashire was produced in the southern portion. The enclosure is so placed to indicate the seat of the iron manufacture in that part of the county and in Derbyshire. In Table XXIII-B the county of Derby is joined to Nottingham to make a separate district. The especial feature of Cumberland and Northwest Lancashire is 718 THE IRON INDUSTRY. the deposit of what is known as West Coast hematites. Up to 1830 these beds were little known and no pig-iron was smelted in either Cumberland or Lancashire. In 1854 the production of ore was 579,000 tons, but this was sent to South Wales and South Staf- fordshire. In 1860 the output had increased to 990,000, in 1870 it was 2,093,000,- and in 1880 it reached 2,759,000 tons. With this great development of the ore beds, blast furnaces sprang up both in Cumberland and Northwest Lancashire, and in 1860 there were 169,000 tons of pig-iron smelted. In 1870 this had increased to 678,000 tons, while in 1880 the record was 1,541,000 tons. It will be found that in 1880 the amount of ore mined in these two coun- ties, which as above stated was 2,759,000 tons, is just sufficient to account for the production of 1,541,000 tons of pig-iron, since the ore contained about 54 per cent, of metal; so that although there were over 3,000,000 tons of foreign ore unloaded that year at Brit- ish ports, it would not seem as if foreign ore was needed in this vicinity. Nevertheless the Home Office Eeports for 1882 show that some 300,000 tons were imported on the West Coast. One- third of this came to Chester and Liverpool and hence need not be considered as directly competing with the local ore, but another third was unloaded at Fleetwood, just across the bay from Barrow, while one-third was taken to Barrow, Workington, Whitehaven and Maryport, on the very borders of the ore region. It was at this time that these hematites were a most Important factor in the iron industry. A large quantity of the pig-iron was exported, much of it to America, its low phosphorus content, often about .04 per cent., rendering it especially valuable for acid Bes- semer work. That day has passed away and the deposits are thin- ning out. In 1900 there were only 1,733,791 tons of ore mined, or only five-eighths of the output in 1880. The pig-iron produc- tion in the two counties is maintained by the use of Spanish ores. The coke is brought from Durham, a distance of from 60 to 100 miles, or from West Yorkshire. The supply of ore at one mine has been prolonged by building a sea wall through an arm of a bay and pumping the pond dry. The success of this undertaking led to a larger project along the same line when the newly won territory showed signs of exhaustion. The value of the iron ore is given in the Home Office Eeports as $3.95 per ton for a 51 per cent, ore, equal to 7.74 cents per unit, and at this rate the ore will cost $7.35 for each ton of pig-iron GREAT BRITAIN. 719 containing 95 per cent, of iron. This does not include the trans- portation from mine to furnace. It must also be noted that the Home Office Eeports for 1899 gave the metallic content of the ore as 53 per cent, on the average, while the figure for 1900 is ahout 51 per cent. The value per ton however is given at a higher figure in 1900, notwithstanding the poorer quality. The two counties of Lancaster and Cumberland in the year 1900 produced 26,86.5,193 tons of coal, or about 12 per cent, of the total, almost all of this coming from Lancashire. The production of pig-iron was 1,585,925 tons, or about 18 per cent, of the total, while the steel constituted 14 per cent, of the outturn of the King- dom. There were also produced 175,000 tons of puddled bar, being 15 per cent, of the total output of the Kingdom. Almost all of this was made in Lancashire. TABLE XXIII-P. Iron and Steel Plants in Cumberland and Lancashire. Name of Works. Location. Blast Fur- naces. Bessemer Con- verters. Open Hearth Furnaces. Acid. Basic. Acid. Basic. Barrow Hem. S. Co London & Northwestern. Moss Bay .... Barrow in Furness... Crewe 12 4 4 7 10 Workington ... 4 3 Cammell, Chas., & Co.. . . Bolton I. & 8 Co Bolton 5 Wigan C & I Co Wigan 10 6 Salford Manchester 2 Millom & Askam Co Carnforth Hem. I. & S.Co. N'rth Lonsdale I. & S. Co. Cammell & Co ... j N'rth west'rn H.I.& S. Co. Others Askham 9 4 4 Derwent ) 8 5 25 Solway J Total 81 24 6 The principal plants are given in Table XXIII-P, the Barrow Works being in Northwest Lancashire, in Barrow-in-Furness, and the other large works in Cumberland. The furnaces of Millom and Askam Company make iron for the open market, and one of them, started in August, 1901, is built on the most modern Ameri- can lines. Tables XXlII-Q and E give statistics concerning this district. The imports of ore at Chester,, Liverpool and Manchester are: 720 THE IRON INDUSTRY. grouped separately, as these ports supply quite a different region from the northern points. It is likely that a considerable propor- tion of the imports at these more southern harbors goes to fur- naces outside of Lancashire, TABLE XXIII-Q. Production of Ore and Pig-Iron and Imports of Ore on the West Coast. See also Tables XXIII-C and XXIII-D for data before 1882. < )re Raised. Ore Im- Pig Iron. Veai-. Cumberland. Lancashire. Total ported Cumberland. Lancashire Total. 1882.. 1,726,235 1,410,111 3,136,351 302.000 790,999 1,001,181 1.792,180 1883.. 1,478,062 1,372,815 2 850 877 302,000 796.770 876,445 1 673.215 1884.. 1XR.S 1,358,090 1,237,285 2,595,375 279.000 223,000 715,328 845.792 1.561.120 1886.. 1,261,655 1 216.193 2.477,848 308.000 695.048 715.228 1.410.276 1887.. 1.480,553 U92.467 2,673.020 368.000 755.441 945,258 1.700.699 1888.. 1,67380* 1.106.013 2.679,817 232.000 745.740 854.238 3.599.978 1889.. 1.594!461 1021.990 2,616.451 272,000 761.748 900.433 1.662.181 1890.. 1.431.159 968.467 2.399.626 404.000 737.026 832.614 1,569 640 1891: 1,417,860 977.130 2,394.990 148,000 715.305 724,750 1.440.05;* 1892.. 1.355,007 845395 2.200.402 237,000 593,245 605,478 1.198.723 1893. . 1,352.410 876.672 2.229.082 180,000 584.401 713,052 1,297.453 1894.. 1286.590 870,617 2.157.207 122.000 606,899 688,744 1295643 1896.. 1215,410 798,325 2.013.735 143,000 540.298 648,740 1,189,038 1896.. 1 279.558 816.570 2.096.128 458.000 680,001 771,420 1,451,421 1897.. 1 294,160 783.427 2 077,587 643,000 706.893 819,475 1,526.368 1898.. 1,251,764 749.427 2,001,191 806.000 732,853 886,210 1.619,063 1899.. 1,137,750 670,924 1.808,674 1.402.000 744,065 954.637 1.698.702 1900.. 1,103,430 630,361 1,733,791 1,102,000 729,074 856,851 1,585.925 TABLE XXIII-R. Imports of Iron Ore at Ports on the West Coast. Year. Barrow. Mary port. Workington. Chester, Liv- erpool, and Manchester. Others. Total. 1882.. . 1883.. . 1884 . 26.000 5.000 13000 6.000 ' 12 000 51.000 41.000 27 000 97.000 129.000 141 000 115.000 121,000 on om 302.000 302.000 1885. . 1886. - 1887. . 1888. . 1889. . 1890. . 1891. . 1892. . 10,000 21,000 9,000 19,000 21,000 99.000 27.000 47.000 27,000 60.000 125,000 126.000 113.000 185,000 61.000 75000 25.000 35,000 48.000 12.000 14,000 7,000 1,000 138,000 156.000 151,000 56,000 111.000 88,000 51.000 105000 23,000 36000 35.000 19,000 13.000 25.000 8.000 10 000 223.000 308.000 368.000 232.000 272.000 404.000 148.000 237 000 1893. . 24,000 67000 8^000 4 000 1 80 000 1894. . 16,000 55000 46 000 5 000 1 Vi flflfl 1895. . 1896. . 1897. . 1898. . 1*99. . 1900. . 33,000 154,000 126,000 203,000 450.000 304,000 61.000 188,000 381.000 357,000 523.000 482,000 15,000 37.000 44.000 118.000 219.000 145,000 32.000 68,000 81.000 83.000 103,000 70.000 2.000 11.000 11.000 45.000 107,000 101.000 143,000 458.000 643000 806.000 1,402 000 1 102.000 GREAT BRITAIN, 721 SEC. XXIIIf. South Yorkshire: The district of South and West Yorkshire includes the historic iron works of Bradford, Leeds and Sheffield. It has never been a great producer of iron ore or of pig-iron, but the town of Sheffield was known five hundred years ago as a maker of steel, and it was here that the crucible process had its birth. The present impor- tance of the district comes from the old established works and the subsidiary steel-using establishments and finishing mills that have grown up around some of the landmarks of the iron trade. TABLE XXIII-S. Iron and Steel Plants in South Yorkshire. Name of Works. Location. Blast Furn- aces. Bessemer Con- verters. Open Hearth Furnaces, Acid. Basic. Acid. Basic. Brown, Bay ley & Co., Attercliffe. Bessemer, H., & Co . Bessemer Fox Samuel & Co Sheffield.. 2 2 2 2 '."..'.'.'.'. 4 2 Steel, Peach & Tozer, Phcenix Cammell & Co Cv clops i% 3 6 t; Scott, Waiter. Leeds Steel Works, Parkgate Iron Co Leeds Sheffield.. 3 5 3 4 1 5 4 5 Brown, J . & Co , Atlas Firth & Sons Norfolk Vickers Sons & Maxim " 4 3 7 Hadfield St Fdy Co " Others ... W. Yorkshire Iron and Coal Co 5 4 6 Lowmoor Co Others . Total 5 26 39 i TABLE XXIII-T. Production of Pig-Iron in South Yorkshire (Sheffield.) See also Tables XX11I-C and XX11I-D for data on ore production, and for years before 1882 Year. Output. Year. Output Year. Output. Year Output 1882..... 321 ,430 1887 178.455 1892... . 261.537 1897 294.846 1883 304.381 1888 190.846 1893... . 155.027 1898 297.490 1884 248,313 1889 229.029 1894... . 225.185 1899 305.583 1885 1890 248,581 1895... 195123 1900 290.601 1886 137.307 Ib91 2^8.354 1896... . 289.497 In 1900 it raised 13 per cent, of all the coal produced in Great Britain. It produced very little iron ore and made only 290,601 722 THE IRON INDUSTRY. tons of pig-iron, or 3 per cent, of the total output; but it 588,000 tons of steel, this being 12 per cent, of the total of the Kingdom. It also made 137,000 tons of puddled bar, or 12 per cent, of the total made. The principal steel works in the district are shown in Table XXIII-S, and the yearly output of pig-iron is given in Table XXIII-T. SEC. XXIIIg. Staffordshire: It is customary to divide this county into a northern and south- ern portion. Forty years ago the south produced more ore than the north and three times as much pig-iron. The ore was a poor ironstone imbedded in the shale of the coal formations, but the deposit has slowly become exhausted and it is necessary to excavate so much shale that the selected ore is very expensive. For these reasons the mining of ore has almost ceased in this southern por- tion and the furnaces run on hematite from Lancashire, or Spain, blackband from North Staffordshire, or the cheap but silicious ores of Northamptonshire, which need only be hauled 60 miles. In North Staffordshire the ore consists mainly of blackband. Bell gives the details of the occurrence in one mine as follows : (1) Blackband 14 inches thick lying on the top of 18 inches of poor coal. (2) "Red slag ironstone," 16 inches thick, lying above 2 feet of poor coal. (3) "Red mine stone" 20 inches thick with 18 inches of coal. There is also a bed of clay ironstone 3% feet in thickness. The yield of pig-iron from the calcined blackband is about 50 per cent, and the value in 1900 is officially reported as $1.82. The amount raised in that year was 1,083,421 tons, so that this deposit is of no small economic interest. The whole county in 1900 produced 14.227,076 tons of coal, or 6 per cent, of the total output ; 1,084,797 tons of ore or 8 per cent, of the total, almost all being in the northern portion as above stated; 596,807 tons of pig-iron or 7 per cent, of the total, this being nearly equally divided between north and south. It made 367,000 tons of steel, or 7 per cent, of the total. Of this amount 142,000 tons were Bessemer steel, all made in basic vessels, The county also made 379,000 tons of puddled bar in 530 fur- naces, which is one-third of the entire output of Great Britain. Two-thirds of this is made in South Staffordshire. This is the GREAT BRITAIN. 723 -only district in Great Britain where the puddling industry is hold- ing its own. Table XXIII-U gives the annual output of ore and pig-iron. TABLE XXIII-U. Production of Ore and Pig-Iron in North and South Staffordshire. See also Tables XXIII-C and XXII I D for data before 1882. Ore Pig Iron. North. South. Total. North. South. Total. 1882 1 887 120 135 4QQ 2 022 529 275 577 247667 523 244 1883 l'(582 600 114 644 l'797 244 267 911 285 325 553*236 1884 l'783'800 89' 945 l' 873' 745 296256 279' 737 575*993 1885 1886 1 499300 91 755 1 591 055 233500 ' 236 137 469 637 1887 840400 97 618 938 018 260201 240 724 500 925 1888 1 629' 277 60*491 1 689'768 279 169 310 451 589 620 1889 1 211 496 51 182 1 262 678 276 219 328489 604 708 1890 1 1*3'447 41,063 l'224'510 255,777 289,648 545 425 1891 . . 1 023 885 47 236 1 071 121 232254 311 816 544 070 1892 '99fl'895 49*745 l'040'640 241,416 308 194 549 610 1893. . 770607 38 172 808779 199 010 285531 484 541 1894 815 368 31 147 846515 210,069 282302 492' 371 1895 828 856 31 205 860061 193647 265411 459058 1896 901 356 30096 931 452 236 176 308 459 544 635 9897 . . 892' 421 34' 100 926' 521 242688 324' 059 566*747 1898 1 058 349 53 363 1 111 712 268357 332 869 601 226 1899 1 '020' 932 48'826 1 '069*758 283212 338'283 621*495 1900. . . . 1 040 605 44 192 1 084,797 272 617 324 190 5% 807 SEC. XXIIIh. North Wales: In the Home Office Report for 1900 the statistics for North and South Wales are combined in a very curious way, for the pig-iron output of Denbigh in North Wales is included in Glamorganshire in the south, while that of Flint, adjoining Denbigh, is combined with the southern English county of Monmouth. In Tables XXIII-C and D will be found data on the output of ore and pig- iron. In making up the averages I have assumed that the output in 1900 for Denbigh and Flint was the same as in 1899 and have corrected the figures for the southern counties accordingly. SEC. XXIIIi. The Eastern Central District; Lincoln, Leices- ter and Northampton; and the Central District; Derby and Not- tingham : The eastern shore of England, just south of the Humber, is not usually regarded as one of the great iron centers of the world, but it is of considerable consequence. The three counties of Lincoln, Leicester and Northampton in 1900 produced over 30 per cent, of 724 THE IRON INDUSTRY. all the ore raised in Great Britain, and they made nearly as much pig-iron as South Wales and more than Staffordshire. The ore of Lincolnshire is an oolite, occurring in a bed measur- ing from ten to twenty feet thick, and is very easily mined. It is only two or three feet below the surface and is worked in open quarry. It varies very much and Bell gives the composition for each foot in depth for eight successive feet, stating that the results are typical. The figures show that in the wet state the iron was anywhere from 21 to 37 per cent., and in the dry state from 21 to 45 per cent. The ore is sorted more or less by hand-and-eye inspection, and the average product in a dry state carries 34 per cent, of iron with about 6 per cent, of silica and 28 per cent, of carbonic acid and lime, the latter making the ore self-fluxing. It is even a little too calcareous and needs mixing with a silicious ore. Its value is given as 75 cents at the mines. The ore was once undoubtedly a carbonate, but by exposure it has been changed to a hydrated peroxide and it is therefore used without calcining. Northampton raises an increasing amount of a very lean and silicious iron ore, some of which is smelted nearby, and the rest sent to Staffordshire and elsewhere. The ore gives about 38 per cent, in the pig-iron, and is worked in the open from a bed 18 feet thick. After paying royalty the ore can be delivered at nearby furnaces for 65 cents per ton. This gives a cost of $1.70 for the ore per ton of pig-iron, but the high silica renders the smelting costly. The deposits in this part of England are related geologically to the Cleveland beds and may be looked upon as the southern out- crop. The use of these lean ores is a rather recent development, just as in Luxemburg the Minette deposit has come only recently into great prominence. In 1830 there were only 5300 tons of iron made from the lean ores of Cleveland and Lincolnshire. In 1860 Cleveland mined 1,480,000 tons of ore and by 1870 this had risen to 4,300,000 tons, and by 1880 to 6,260,000 tons. The in- crease has not continued in Cleveland, which in 1900 mined only 5,493,733 tons, but the mines of the southern district are coming to the front. In 1860 this region raised only 118,000 tons, in 1870, 1,048,000 tons, in 1880, 2,766,000 tons, while in 1900 the output of the three counties of Lincoln, Leicester and Northamp- ton reached 4,298,145 tons. Thus, although the production of the Cleveland district has fallen since 1880, the total production GREAT BRITAIN. 725 of the lean ores from this geological horizon has increased from 9,026,000 to 9,818,000 tons. Estimating the average iron content of the ore at 32 per cent, and the iron in the pig at 93 per cent, this amount of ore represents about 3,300,000 tons of pig-iron, or about 37 per cent, of the total pig-iron made in the Kingdom. TABLE XXIII-V. Production of Ore and Pig-iron in Eastern Central England. See also Tables XXIII-C and XXIII-D for data before 1882. Or e. Pig Iron. Yc&r. Leicester. Lincoln. Northamp- ton. Total. Lincoln and Leicester. Northamp- ton. Total. 1882.. 267,802 1.287,289 1,333,085 2,888,186 201,561 192,115 393,676 1883.. 294,825 1,107,793 1,290,087 2,692,705 237,068 216,641 453,709 1884.. 261,837 1,348,693 1 279,783 2,890,313 259,398 196,212 455,610 1885 1886.. 390,687 1,193,621 996,440 2,580,748 242,342 197,853 440,195 1887.. 372,773 1,305,929 935,473 2,614,175 251,869 236,390 488,259 1888.. 535.831 ,345,101 1,066,746 2,947,678 298,673 236,841 535,514 1889.. 582,858 ,560,690 1,257,080 3;400,548 336,175 230,820 566,995 1890 609,964 1,052,409 1,278,381 2,940,754 268,405 225,046 493,451 1891.. 6i6,125 ,214,131 1,043,541 2,903,797 284,766 194,395 479,161 1892 680,985 ,459,404 1,120,365 3,260,754 279,556 177,817 457,373 1893.. 471.098 ,039,112 719.071 2,229,281 216,575 143,815 360,390 1894.. 568,026 ,554,286 1,130,773 3.253,085 343,616 223,348 566,964 1895.. 598,551 ,554462 1,082,252 3,225,265 349.232 254,744 603,976 18%.. 702,842 ,576779 1,263,650 3,543,271 361,029 274,462 635,491 1897.. 714,651 ,765,365 1,264 915 3,744,931 363,487 249,824 613, 311 1898.. 696.015 848.404 1,406150 3,950.569 381,824 250,835 632.659 1899.. 677,667 2,094,330 1,779.710 4,551,707 408,989 279,301 688,290 1900.. 750,708 1,924,898 1,622,539 4,298,145 388,745 247,908 636,653 TABLE XXIII-W. Production of Pig-iron in Derbyshire and Nottinghamshire (Central England). Statistics formerly kept separate, but now combined. See also Tables XXIII-C and XXIII-D for data before 1882. Year. Derby. Nottingham Total. Year. Derby and Not- tingham. 1882 372 650 73085 445735 1892 481 449 1883 1S84 353,474 359338 68,740 78 175 422,214 437-513 1893 1894 343! 115 376 726 1835 1895 413 454 1S86 296213 50119 346,332 1896 455 487 1887 296 118 1897 488 472 1383 362.744 1898 529*208 1339 378 464 91 650 470 114 1899 571 994 2890 387.760 75-300 463 660 1900 561 626 1891 387.127 83821 470.951 726 THE IRON INDUSTRY. In the counties of Lincoln, Leicester and Northampton there are 47 blast furnaces, of which 32 were active in 1900. In Derby and Nottingham there are 54 furnaces, 43 being active. It might seem from a glance at the map that Nottinghamshire should be com- bined with Lincolnshire and Leicestershire, but in the Home Office Keports its output of pig-iron is joined with that of Derbyshire. Neither Derby nor Nottingham produces iron ore in quantity worth mentioning, so that the apparently arbitrary division is founded on good reason, Tables XXIII-V and W give detailed information concerning these two districts. CHAPTER XXIV. GERMANY. SEC. XXI Va. General View: The discussion of the German iron industry as it appeared in the former edition was founded principally on knowledge gained by personal inspection. There were also at hand a most valuable series of letters by Kirchhoff, which were printed in The Iron Age, of which he is editor. They began in May, 1900, and later were issued in book form. The manuscript for the former edi- tion was submitted both to Dr. Wedding, of Berlin, and Herr Schrodter, editor of Stahl und Eisen at Dusseldorf. Since this book was published it has been read by other friends in Germany, and I ami indebted particularly to Mr. Franz J. Miiller, General Director of The Rheinische Steelworks at Ruhrort, and to 0. von Kraewel, Superintendent of the same Company, for a very critical review, and their information has been used in this edition. Much matter has also been derived from a paper by Brugmann.* It need hardly be said that none of my friends can be held responsible for personal or political opinions. In the matter of statistics Germany is lamentably weak. The Government regularly gathers an immense mass of figures, which are duly published in all the journals and technical papers, but they are worthless. I spent much time and money collecting the records for the former edition, but must state that they were in error. Germany recognizes three kinds of product: (1) Ingots for sale; (2) half finished product; (3) finished product; but if one works sells ingots to another, and the second works makes billets and sells them to a third mill for rerolling, then this steel is put in the total three separate times. A very large amount is actually added twice, because almost all the wire mills in Germany are independent. Within the last two or three years, the total production of ingots in the whole of Germany has been collected. Before that time no statistics were reliable, and even now tihere are no data published as to the output of separate districts. I am able however to present in tihis edition for the first time, a reasonably * Jour. I. AS. I., Vol. II, 1902. 728 THE IRON INDUSTRY. accurate estimate by high authority of the output by districts for the year 1902-03. The figures will be found in Table XXIV-C. They differ widely from the former statement, but it will be noted t that the present figures cover the output of ingots, while the pre- vious edition took cognizance only of finished material, fhus ex- cluding billets for export either abroad or outside the district. GERMANY. 729 The data on steel works and blast furnaces and puddling plants have been taken from the Gemeinfassliche Darstellung des Eisen- hiittenwesens for 1900. The boundaries of each district have been faithfully reproduced from a drawing by Dr. Wedding, but it is 'impossible to take these limits as true for all the statistics given. For instance, the map shows the area of the Ruhr coal basin, which TABLE XXIV-A. Production of Pig Iron, Ore, Coke and Coal in Germany. Note : Districts are in the order of their pig-iron output. Data for 1899 from Wedding ; for 1900 from Schrb'dter ; details for pig-Iron, ore and coal for 1900 are not at hand in the same grouping as given here, but the totals as published for each province indicate that the output is about the same for each division in 1899 and 1900. District. Pig Iron, 1899. Ore, 1899. Coke, 1900. Coal, 1899. Tons. Per Cent. Tons. Per Cent. Tons. Per Cent. Bituminous. Tons. Lignite. Tons. Ruhr 3,186,704 1,290,264 982,930 744,672 656,942 596,565 152,7^6 124,1*3 115,200 83,321 80.342 21,012 none 95,811 39 16 12 9 8 . 7 2 2 1 1 1 ""2" 212,794 6,972,758 6,014,394 476,823 2,119,145 none 16,584 799,728 128.430 184,020 none 756.758 8,108 28,558 122,981 1 39 34 3 12 j{ 9,644,000 none none 1.947,000 none 894,000 267,000 all others 33.000 75 is" 55,184,138 1,071,103 none 27,959,689 none 9,589.636 1.764,398 none 547.822 638.153 none none 4,546,756 338,058 none none none none 609,515 none none 3.927,257 1.544,805 none 37,277 none 277,337 1,292,348 1,851.542 24,664,585 Lothringen Luxemburg Silesia Siegen Stiar 7 2 ii Aachen Ilsede Osnabruck .... Bavaria Pomerania Lahn Saxony 4 74,000 i Others Cent. Germany Total . 1 8,130,655 100 17,989,635 100 12,859.000 100 101,639,753 34,204,666 TABLE XXIV-B. Movement of Ore in Germany in the Year 1899 in the Districts Importing or Exporting Across the Frontier. District. Lothringen and Lux- emburg. Ruhr. Silesia. Pomerania. 12,987 152 212 794 476,823 none 1 807 421 1,271,052 33,787 1.884,769 124,200 1 384 447 275 406 329,705 1,337 000 Brought from the Siegen, the Lahn and 4,734,600 730 THE IRON INDUSTRY. is the foundation of the prosperity of the district, and it shows the outline of the ore deposits in Siegerland, but in the interval along the Khine are blast furnaces and steel works and countless works, both large and small, some making their own steel and some buying their raw material, but all turning out some of the ten thousand articles "of German manufacture," each shop special- ized for something, for bolts, for scissors, for scythes, or for needles, and all contributing to the prosperity of the Rhine province. The general statistical situation is shown in Tables XXIV-A, B and C. TABLE XXIV-C. Estimated Output of Ingots (including castings) in Germany for Twelve Months, 1902-03 ; metric tons. District. Acid Bessemer. Basic Bessemer. Acid Open Hearth. Basic Open Hearth, Total. The Ruhr 240000 2 246,000 176,000 1 667000 4 329 000 Silesia 55 000 242000 292 000 589 000 Lothringen 953000 45000 998 OOQ Luxemburg 408000 408 000 The Saar 867 000 10 000 160 000 1 037 000 Saxony 10800 40 000 7200 85'000 143 000 Siegerland 154,000 154000 Aachen 287000 46000 333 000 Ilsede-Peine 239000 239 000 29,000 30666 59 000 Bavaria 100000 30 000 130 000 Total 334 800 5 382 000 193 200 2 509 000 8 419 000 i SEC. XXIVb. Lothringen and Luxemburg: The province of Lothringen is the old French Lorraine, so famil- iar to every one as a great arena of war. Following its incorpora- tion into the Empire of Germany not only was its name changed so as to be in accordance witfo the German language, but almost every town and village received either a new name or a German prefix or suffix. As a matter of fact, this was quite natural, for it as quite impossible for the German or the English speaking people to pronounce correctly many of the French names, and it would have been absurd to have a German city called by a name that nine- GERMANY. 731 tenths of the German inhabitants could not pronounce. English and Americans have committed worse sins without excuse in changing the spelling of Napoli to Naples, Venezia to Venice, and Wien to Vienna. Moreover, it is urged by my German friends that the new spelling is really the original, and that the French were the real offenders in changing the names during their tem- porary occupation. At present many maps of Lothringen contain the odd names, and these are used exclusively in France and Bel- gium for obvious reasons, and also very widely in England and America, while the term Lorraine is probably known to a hundred Americans where Lothringen is known to one. This change, nat- ural though it is, entails endless confusion upon the traveler, who might be supposed to guess that Hayange means Hayingen, and Differdange, Differdingen, but can hardly be expected to know that Diedenhofen and Thionville are the same. Lothringen is a fundamental part of the Empire, unlike Luxem- burg, which is merely connected with it through a tariff treaty (zollvereki'). Both districts have the same general characteristics, and rely on the enormous bed of iron ore which extends beyond their borders into France and Belgium, and whose known contents will supply enough iron for many generations. This ore goes by the term "Minette," a contemptuous diminutive once given it by French workmen ; this is also the name of one of the French, prov- inces in whicih it occurs. It is an oolite, consisting of small grains, each one of which is made up of concentric shells of silicious or cal- careous matter, and hydrous ferric oxide. The beds throughout the greater part of Lothringen carry an excess of lime, but near the Luxemburg border is a deposit running high in silica and carrying 40 per cent, of iron, so that by proper mixing a self-fluxing burden can be obtained. Table XXIV-D shows the composition of different grades of ore according to different authorities. .The map of the Minette region shown in Fig. XXIV-B was originally made by Dr. Wedding, but was much extended and com- pleted by Kirchhoff. The formation is made up of many different beds, and these vary greatly in thickness, the deposit in the north being 180 feet thick, while in the south it is only 20 feet ; but there is no regularity at intermediate points, either in thickness or in tine arrangement of interstratified rocks, and there is much fault- 732 THE IRON INDUSTRY. THE MIJTETTE DISTRICT OF LOTHRIXGEX, LUXEMBURG AND ERAOTCE / Nueres M Limits of Iron District Dots indicate Bloat Furnaces Elanta S . SCALE OF MILES f 612346678910 FIG. XXIV-B. GERMANY. 733 TABLE XXIV-D. Composition of Ores from Lothringen and Luxemburg and Data showing the Thickness of the Beds, and Thickness of Inter- mingled Strata of Earth and Limestone, arranged from Schrodter, Stahl und Eisen, March 15, 1896. Also data from Wedding, Eisenhiittenkunde, Zweite, 1897, p. 59; Kohlmann, Stahl und Eisen, Vol. XVIII, p. 593 ; and Stahl und Eisen, Vol. XX, p. 1266. Note ; the boreholes are at different points in the Aumetz Arsweiler district. Strata and Thickness in Feet. Fe Mn P SiO a CaO Al a O, Schrodter Depth Thickness Character Borehole from of of Surface Layer Deposit A 16 Red sand 25.6 33.3 9 4 16 10 Red sand 26 6 31 3 9 5 26 41 Lime & clav. ...... 67 9 Red Minette 30 7 7 5 21 5 57 76 i Lime ...... 77 1 Red Minette. 38 5 9.2 12 1 6 9 78 3 Red Minette 32 4 10 19 8 5 8 81 7 Red ore 39 4 7 7 11 6 4 9 88 19 Earth ... . 107 13 Gray ore 33 7 7 6 20 4 1 120 16 Earth . ... 39 15 1 (,-g o 4 1 150 3 Blk Minette 21 21 3 l ^5 3 15 7 153 12 Black ore... 41.1 10.7 4.6 6 165 3 Black ore... 33.0 168 3 Black ore.. . 171 2 Black ore. . . 37.0 7 B 13 S limestone. 13 5 R. sandy ore 21.0 15.0 18 25 S limestone. 43 4 Red ore 47 17 S. limestone. 24.0 0.53 24.0 64 5 Red ore 69 6 S. limestone. 27.0 0.59 22.5 75 7 Re 1 ore 82 18 Marl 28.0 20.0 100 17 Gray ore.... 117 3 Earth 38 0.84 12.0 6.0 120 7 Gray ore . . . 127 19 Earth 35.0 0.91 12.9 6.3 146 10 Brown ore . . 156 9 Earth 89.3 0.82 6.3 7.7 165 6 Black 170 4 Earth 36.9 0.86 6.8 6.7 174 4 Ore . . . 36 4 57 6 2 4 5 CO 9 Limestone. . . 9 6 R sandy ore 26.9 20 15 27 L'stone.marl 42 4 Yellow ore.. 21.3 19.5 46 8 Blue marl . . . 54 2 Gray ore 35.0 12.0 56 6 Gray ore .... 42.6 6 5 62 7 Gray ore... 31 4 15 2 69 3 Gray ore oo o 12.3 72 2 Gray ore .... D 81 R.sand.marl 29.8 11.7 81 12 Red lime ore 44.5 11.6 6 3 93 14 Poor M. & marl 734 THE IKON INDUSTRY. Table XXI V-D Continued. Strata and thickness in feet. Fe Mn P SiO 2 CaO A1 2 8 107 20 Gray ore .... 127 12 Blue marl . . . 45.6 12.5 4 .5 "139 16 Brown ore . 39.6 25.5 B 95 Lime ores. . . 95 20 Gray ore .... 37 e 12.3 18 2 115 15 Marl 35 8 21 1 6 4 148 14 Black ore... 42.0 17.0 3.0 F 8 Red sand . . . CO, 8 38 Earth 29.4 8.3 30.4 5.9 16 49 19 Earth .... 68 2 Yellow .... 34 7 8 7 15 7 5 g IJJ 9 70 12 Earth 82 5 Yellow 28.3 17.9 14 4 8 3 11 3 87 6 Earth 93 13 Gray 34.1 10.7 14.2 6 6 106 21 Karth 127 7 Brown 38.8 16.2 4.7 7.8 134 8 Earth 142 3 Black 32.7 21.8 6.9 6.1 Wedding. Red Calcareous 42.9 tr 54 9 9 14 8 4.7 H,O 6 3 Red Silicious . . 34 5 7 32 23 6 12 5 8 8 6 Gray 38 9 92 9 5 1<5 3 2 3 17 5 Brown Green 21.5 33 4 "6*4 0.71 88 16.5 24 4 21.0 2 7 6.4 10 3 25.1 15 Stahl und Eisen. Rumelange Dudelange.. Esch 33.2 40.7 39.5 0.6 0.4 0.4 0.80 1.00 1.00 6.8 7.5 13.4 16.3 7.7 6.4 5.2 4.7 6.1 Differdange la Madelaine Kohlmann. Black ; thickness 18 feet 27.6 39.2 18.2 32 to 45 0.3 0.4 0.2 0.72 0.81 0.53 42.0 16.1 8.5 11 to 22 4.9 5.3 33.3 2to7 4.6 6.4 2.3 6 Brown ; 6 to 12 feet . . 36 to 45 5 to 21 4 to 9 Gray calcareous 32 to 41 5 to 15 4 to 14 4 to 6 Yellow calcareous ; 15 feet 32 to 36 7 to 9 10 to 15 Red calcareous : 6 to 12 feet 34 to 40 8 to 9 9 to 15 Red silicious 36 26 to 27 2 to 3 ing, in' some eases the throw being 200 feet. It is roughly true, however, that as we go southwest in/to France the beds go down into the ground, get less in thickness 'and higher in silica. In Luxemburg the ore mines are owned partly by companies that acquired ownership many years ago, partly by railroads, built in order to get subsidies in the shape of ore lands, partly by farmers and private individuals, while part is still controlled by the govern- ment. Much of the ore in Luxemburg is bought and sold in the open market, while in Lothringen nearly all the property is in the hands of iron producers, and the great steel works in both Belgium and Westphalia have acquired title to mineral lands, some of these acquisitions being quite recent, while some date back many years. The ore supply in Luxemburg is calculated as good for ajxrat GERMANY. 735 one hundred years, at the present rate of consumption, but in Lothringen the beds are considered good for eight hundred years. The mineral domain of this latter province covers about one hun- dred thousand acres, half of which is owned by the local steel companies. A good part of the remainder is owned by the com- panies operating steel works in Westphalia. Kirchhoff mentions the following as having mines in Lothringen and works in the Rhenisih district : Aachener Hiitten Act. Verein, Gutehofrnungshiitte, Friederich Wilhelmshiitte, Phoenix, Union, Horde, Hoesch, Rheinische and Krupp. In the Saar district we have Gebriider Stumm, Rb'chlings, Bnrbach and Dillengen. Belgium is represented by the Angleur Company and by Cockerills. This list of course omits the local steel companies of Lothringen, all of which have their own prop- erties. As above stated, there is a very considerable quantity of ore sold in the open market in Luxemburg, but very little in< Lothringen, so that the selling price in the former province will be a better measure of the market. The figures given by Dutreux show that in the five years from 1895 to 1899 the average market price varied from 49 to 57 cents per ton, with a general average for the whole period of 52 cents. The cost of the ore to those wiho possess their own mines must be less than this, but it is hardly likely that it is less than 40 cents, after allowing for a sinking fund. The Tun of mine will average about 31 per cent, in iron, but the ore carried to Westphalia is richer than the average. It will run about 35 per cent, in iron* and costs about 75 cents per ton at the mines. The new freight rate is $1.40 per ton, giving a total of $2.15 per ton of ore delivered in Westphalia, or 6.14 cents per unit. If the ore is smelted at the mine it is necessary to carry nearly 1J tons of coke from the Ruhr to Lothringen, at a cost of about $1.82 per ton of coke at present rates, as the freight on fuel in Germany is about one cent per ton per mile, This, of course, does not include the cost at the ovens, which is estimated by Kirch- hoff to be about $2.00 for those who have .their own collieries, so that the cost of fuel delivered in Lothringen will be $3.82 per ton of coke or $4.80 per ton of iron. The ore for a ton of pig * Jour. 1. & S. /., Vol. II, 1902, p. 17. 736 THE IRON INDUSTRY. will cost about $1.30, so that the total for ore and fuel sums up $6.10 in Lothringen and $9.10 in Westphalia. I am afraid that this estimate of Kirchhoff on the cost of coke assumes that a good profit is made on the by-products, but allows nothing for the in- terest and depreciation of the plant. Against the obvious advantage of transporting 1J tons of coke instead of 3 tons of ore is the disadvantage that Lothringen is not a great market. To the southwest is the frontier of France and the French steel works working on the same deposit, while on the northwest are the cheap labor and fuel of Belgium tapping the orefield in Luxemburg. To the south is the mountain barrier of Switzerland, to the east the coal field and iron works of the Saar, and to the north the smoking valleys of the Rhine and the Ruhr. All this means that the steel must be carried a long dis- tance and past the doors of active competitors. A great part of the output of Germany is sent oversea and a large part is consumed in finishing mills in the northern districts, and inasmuch as the coal of Westphalia is right on the road between the mines and the market, it is evident that the northern works are not necessarily destined to succumb to the competition of the Minette district. There is a chance for both ends working together, since cheap transportation must include ore going in one direction and coke in the other, and there is also great opportunity for reductions in charges. The German railroads are owned by the government, and they offer a very good argument against state control. Like all German official work, they are conducted with perfect honesty, but with an immense amount of red tape. As a consequence of the honesty and of the high freight rates, they pay a handsome profit, but on account of the red tape this money goes into the general treasury and defrays the expenses of the military establish- ment instead of being used to improve the transportation service. A great deal of money is spent on immense stations for passenger traffic, but the freight service is not what it ought to be, and the transportation of ore from Lothringen to Westphalia costs 1 cent per ton per mile, while coke and finished material are from 30 to 50 per cent. more. Private ownership of railroads in America has resulted in spending money for improvements, for larger cars and heavier engines, and has cut down the rates far below the German GERMANY. 737 tariff, even though the American roads traverse districts much more sparsely settled than the western provinces of Germany. In addition to the questions of freight which have just been dis- cussed, we have the very important fact that Westphalia pos- sesses large and old established works surrounded by communities of skilled workmen. The task of starting a steel works in a part of the country where such an industry has not existed before is hard enough in America, but in any other part of the world it is still harder, for in our land men are accustomed to move, and very readily break away from old associations. A still more important matter is the absolute destruction of capital involved in a transfer of the iron industry, for a works in Westphalia cannot be trans- ported bodily to Lothringen. If the attempt were made it is doubt- fud if twenty per cent, of the money would be utilized, and this being so it becomes cheaper to destroy the old and to build anew rather than to attempt to move, and it may be shown by calculation that the interest and depreciation on a steel works, including the blast furnaces, is more than the cost of transporting the ore supply a considerable distance. In the case of a Westphalian works, which perhaps is all paid for and has no outstanding bonds, the depreci- ation account may be neglected and the interest charges looked upon as profit, while in a new works in Lothringen these items be- come a direct load upon the cost sheet. From these considerations it ihappens that we find many different ways of working. The old plants in the Ruhr are buying prop- erties in Lothringen and are bringing ore to their furnaces and so also are the steel works in the valley of the Saar. Other plants are making pig-iron at the mines and sending it to Westphalia and to Aachen, while still other works are being built at the ore bank, the coke being brought from the Ruhr. The production of the whole Minette district, including Loth- ringen, Luxemburg and France, was less than three million tons in 1872, but in 1895 it had risen to eleven million tons. In 1898 it was fifteen million and in 1899 about seventeen million, of which France contributed four millions, Luxemburg six millions and Lotlhringen seven millions. Of the thirteen million tons mined in Lothringen and Luxemburg about one-fourth was shipped to Bel- gium and France, leaving about ten million to be used in the Em- pire. About one-eighth of this latter was sent to the Saar and the 738 THE IKON INDUSTRY. Kuhr, while the remainder, between eight and nine million tons, was smelted at the mines, Lothringen in 1899 producing 1,290,264 tons of pig-iron and Luxemburg 982,930 tons, all this iron being made from local ores. It has been pointed out by Kirchhoff that the importance of the Minette district is concealed by the accident of its situation. The total output of ore from the whole deposit in 1899 was about seven- teen million tons, which would make about six million tons of pig-iron, but this is divided between three different nations and between different provinces, and even the portion which we have considered as German can hardly be called so rightly, since Lux- emburg is not an integral part of the Empire. The two provinces together raised very nearly three-quarters of all the ore mined in Germany, Siegerland standing next with 12 per cent, of the total, but the combined production of pig-iron- in the Minette field was only two-thirds as much as in the Ruhr. In 1899 there were seventeen active blast furnaces in Lothringen and twenty in Luxemburg, which were not connected with steel works in those provinces, but which sold their iron in the open market or shipped it to the Saar or the Euhr, many of these fur- naces being owned and operated by steel works in these two dis- tricts. The Minette ores give a pig-iron running quite regularly about 2.00 per cent, in phosphorus, and very considerable quanti- ties are sold for foundry work and for puddling. There were twenty-two furnaces in Lothringen and nine in Luxemburg con- nected with adjacent steel works, so that less than half the fur- naces in the district were owned by local steel plants. The total number of active furnaces as above given' was sixty- eight, and the production of pig-iron was 2,273,194 tons for the two divisions, representing an average of a little over 90 tons per day for each furnace. Such a calculation of average capacity is not usually of much value, as an old district is very likely to have a number of small and antiquated plants, but in the official list published by the Verein Deutscher Eisenhiittenleute, from which most of these data are taken, there are no very small furnaces men- tioned in these two provinces. The capacity as published in the above mentioned list is considerably in excess of the results above calculated, but it would seem as if the statistics would be more accurate than estimates, and we may say therefore that the average GERMANY. 739 furnace in the Minctte district, most of the plants being of rather modern construction, turns out between ninety and one hundred tons per day, some of them of course exceeding this considerably. It is necessary for American metallurgists to consider that this is done on an ore running only 31 per cent, in iron, but on the other hand the mixture is usually self -fluxing, so that for a comparison we must take the ore and limestone together in non-calcareous ores, and figuring in this way we will find that Lake Superior ores when 740 THE IRON INDUSTRY. mixed with the usual amount of stone give about 45 per cent, of iron, so that the furnaces working on Minette ores smelt about 50 per cent, more material than American plants, without taking into account the ash in the fuel. It will be noted that the mixture is not always self-fluxing, for near the Moselle River the calcareous beds are scarce and it is necessary to use limestone as a flux. Most of the blast furnaces in this district use Westphalian coke, the shipments in 1899 from the Ruhr ovens amounting to nearly three million tons, which was nearly 40 per cent, of the total coke output of the northern coal field. Some coke is imported from Belgium by plants in Luxemburg, but the German article is far superior in quality. There are three steel works (in Lothringen and two in Luxemburg having between them twenty-six converters, ranging from ten to twenty tons capacity, and averaging about fifteen tons. There were only two open-hearth furnaces, one acid and one basic. All the converters are basic. Three new plants were started in the year 1900, at Rombach, Kneuttingen and Differdingen. In Fig. XXIV-C will be found a drawiing of the first of these. It is representative of the best Ger- man engineering practice and is entirely new, having been started in 1900. The engineer is Bergassessor Oswald, of Coblenz, to whose courtesy I am indebted for the drawings. There are seven blast furnaces in the Rombach plant, three of them new, the latter being 90 feet by 23 feet with a 13-foot hearth. The blowing en- gines are ample, but it is intended to eventually use gas engines for this purpose and thus save the steam for driving the reversing rolling mills. To this end the boiler capacity was made very large, the steam pressure being 140 pounds and economizers and super- heaters installed, it being hoped that by this means the rolling mills can be driven by the blast furnace gases. There are two mixers for the iron, each of 200 tons, feeding 4 basic 17-ton con- verters. The pig-iron runs from 1.5 to 2.0 per cent, phosphorus and 0.5 per cent, manganese, this latter element being obtained from ores from Spain, the Caucasus and from the Lahn district. The miixture is self-fluxing and runs about 31 per cent, in iron. The blooming mill is a 48-inc'h reversing, 9 feet 6 inches be- tween housings, and this feeds two large mills without the blooms being reheated. The larger finishing mill is 36-inch with four stands, and rolls large beams, while the smaller is 30-inch for GERMANY. 741 billets, and will finish a bar 400 feet long, this not being an ex- traordinary length in Germany. There are three other smaller mills, 26-inch, 22-inch and 14-inch, for rails and miscellaneous structural work, and these are to be driven by shunt motors, a very large power plant being provided which eventually is to be run by gas engines. All the machinery is of massive type and the labor saving and mechanical handling devices are worked out with thoroughness. The capacity is now 35,000 tons per month, but this is soon to be much increased. The Difterdingen plant was also constructed with lavish expendi- ture and a very extensive outfit of blowing engines driven, by blast furnace gas was installed. Much trouble was experienced through dust, although these difficulties have since then been in great meas- ure overcome. The plant operated by De Wendel at Hayingen is an extreme example of the system of spare mills, as four complete mills, each with its modern German multiple cylinder engine, stand waiting their turn to run, for there are only men enough to run at most two mills and only steel enough for that number in spite of the fact that they are operated in a very slow manner. The building covering these mills includes all the hot beds, finishing machines, storage and loading yards, and as a rough guess I should say it is 700 feet by 1000 feet, not including the converting department. The output is about 400 tons per day. Table XXIV-E gives a list of the steel works and blast furnaces in the district. TABLE XXIV-E. List of Steel Works with Blast Furnaces in Lothringen and Lux- emburg. District and Works. Location. No. of Blast Furnaces and Daily Capacity in Tons. Bessemer Converters, Number and Capacity in Tons Open- Hearth Furnaces Number and Capacity in Tons. Acid. Basic. Acid. Basic. Lothringen Aumetz Friede.. Rombacher, etc. . DeWendei & Co. . Luxemberg Diidelingen. etc.. Differdingen Kneuttingen .... Rombach 1 Hayingen 3-130 7-140 7110 6-110 6-110 4120 420 418 612 312 6 10 3-20 1-15 115 1 Gross-Moyeuvre . Diidelingen Differdingen 742 THE IRON INDUSTRY. List of Blast Furnaces without Steel Works. Location. Owner. District. Blast Furnaces. Owned by Steel Feutsch Aumetz Friede Lessee . . 3_ 120 Redingen . Dillengen Saar.. . . 2 90 Diedenhofen Rochling Saar 2 150 Ueckingen Gebruder Stumn Saar 4 120 Deutsch Oth.... j Esch Acieries Angleur Rothe Erde Belgium Aachen 2 90 5 190 Luxemburg. . . . \ Esch Burbach Saar 2 120 Unattached 7 120 13 120 SEC. XXIVc. The Ruhr: The Ruhr district embraces most of the province of Westphalia and includes a little of the western shore of the Rhine. It is here that we find the coal that gives the best coke on the continent of Europe, though it is far from being equal to the coke of Durham or of Connellsville. The Rulhr coal district proper is included in an irregular space measuring about fifty miles east and west and a little less north and south, this field being shown on the map in black witih) Ruhrort on the western end and Horde on the east, but as a matter of fact, coal is found east of Horde as far as Hamm and also extends westward across the Rhine, several new mines having recently been opened on the western bank. The great works of Krupp at Essen are almost in the center. The deposit covers an area about equal to the county of Westmoreland in Pennsylvania or the Dunham coal field in Northeast England, but Westmoreland raises only about ten million tons of coal per year, Durham about forty-six million- and Westphalia over fifty million. The production of coke in the Ruhr is about the same as in Fayette County, Pennsylvania, which includes flhe Connellsville beds. The output of Durham is not known accurately, as no sta- tistics are kept in England of this material. The Ruhr raises one-half of all the bituminous coal raised in Germany, and makes two-thirds of the coke, and, in addition to supplying the wants of Western Germany, sends some coke to other countries. In 1899 Germany exported 750,000 tons of coke to France and 135,000 tons to Belgium, almost all of this coming from Westphalia, Austria received 600,000 tons, but part of this GERMANY. 743 was sent from Silesia. The product of the Westphalian ovens, '(however, is so much better than the eastern supply that it is carried in large quantities as far as Styria in Southern Austria. In 1892 tfhe Ruihr district made 66 per cent, of all the coke made in Ger- many, but in' 1900 its share had risen to 75 per cent. This increase in relative rank as a coke producer has gone on with remarkable regularity, as will be shown in Table XXIV-F. TABLE XXIY-F. Production of Coke in Germany, by Districts. Data from Schrodter ; private communication. One unit=1000 metric tons. District 1892 1893 1894 1895 1896 1897 1898 1899 1900 Ruhr 4560 4780 5 398 5562 6266 6 872 7,374 8202 9 644 1 060 1 060 1 122 1 190 1 269 1 399 1 455 1516 1 411 Lower Silesia 325 366 416 431 *443 424 430 460 536 Saar 587 574 695 713 744 821 887 876 894 259 219 207 212 310 251 259 269 267 Oberkirchen 26 27 24 27 27 31 30 33 33 82 73 79 70 77 78 72 74 74 Total 6 899 7 099 7 941 8 205 9 136 9 876 10507 11 430 12 859 Per cent, made in the Ruhr. 66 67 68 68 69 70 70 72 75 The exports of coke to Belgium are counterbalanced by coke brought into Luxemburg from that country, the amount so im- ported being greater than the amount going from Westphalia to Liege. It is only a small proportion of the furnaces in Luxemburg that tihius import coke, and the amount sent from the Ruhr to Lothringen and Luxemburg in 1899 amounted to 2,783,000 tons, or nearly 40 per cent, of the total coke production of Westphalia, The coal occurs in a great number of beds, of varying thickness, the number of workable seams being over two hundred, but none of them is over six feet thick and the average only about half that. The total thickness of the coal measures is between seven and eight thousand feet and they are much, folded and faulted. In the south- ern portion of the field the outcropping beds have been nearly worked out, and as mines have been opened more and more to the north it has been necessary to sink deeper to reach the coal, one shaft going down 2500 feet, and all through strata heavily charged with water. When it is considered that there is more trouble from gas in the deeper mines it will be evident that conditions do 744 THE IRON INDUSTRY. not indicate any future decrease in the price of coal or any likeli- hood of any extraordinary development in capacity. The upper beds give a coal containing from 35 to 45 per cent, of volatile matter, the middle region from 15 to 35 per cent, and the lowest seams not over 15 per cent. It is from the so-called "fat" coals of the middle region that most of the coke is made, the ash in the product running about 10 per cent. The sale of coal and coke is controlled by a syndicate wihich embraces 90 per cent, of the coal output, and the price of fat coal has risen during the last few years from $2.00 in 1895 to $2.44 in 1900, these figures being at the mine. Kirohhoff gives quotations from the annual reports of many collieries, and I find from these figures that the larger collieries, producing between them one-third of all the coal and coke of the district, show a cost ranging from $1.31 to $1.69 per ton of coal, with an average of about $1.55, some of the smaller collieries run- ning up to $2.00 and even to $2.50. The wages of miners have advanced very much in recent years. In 1878 day laborers received only 56 cents and the miners 67 cents, but there was then an advance through many years so that in 1891 the wages were 71 cents for common labor. A reaction followed and then another rise, and in 1898 common) labor com- manded 76 cents per day and the miners earned $1.14. The min- ing situation in Westphalia is much as it is in the United States, for the rapid development of industry has gone ahead of the nat- ural increase in population and nearly one-third of the working force in the mines come from Poland, Eastern Prussia and Italy. These alien communities are less common in Europe than in our own land. The average selling price at the oven of blast furnace coke in the Ruhr basin varied from $1.96 per ton in 1887 to $4.95 in 1890. It dropped to $2.75 in 1893, 1894 and 1895 and then rose to $3.50 in 1900 and $4.25 in 1901. A great part of this coke is made in by-product ovens, and it is well known that responsible coke oven builders will agree to build ovens and operate them free of cost for a term of years, taking their pay in the by-products, and turn over the plant at the end of the period to the party of the second part. This being so, it is quite evident that the price of coke in Westplhalia includes a very good profit, and the figure given 1 GERMANY. 745 is no measure of the cost of fuel to those steel works tifoat own their own mines and ovens, among which are the following: Hoerde, Union, Hoesch, Schalke, Bochumer Verein*, Krupp, Gutehoffnungshutte, Phoenix, Rheinische, and Deutsche Kaiser. In the matter of iron ore, Westphalia occupies a very subordinate position. A small amount of blackband is raised, containing about 35 per cent, of carbon iand about 28 per cent, of ironi, mainly in the form of carbonate, but the quantity is inconsiderable compared with the output of pig-iron and steel. Sixty per cent, of the ore supply comes from the Siegen, the Lahn and Lothringen, and the remainder from over sea. Spain contributes over 20 per cent, of the total ore smelted in the district, and Sweden about 15 per cent. The supply brought from the Siegen is spathic ore, which is roasted before using; it contains about 35 per cent, of iron -and is more fully described in the account of that district. Tihe ores from the Lahn and from Lothringen are also described in the proper place, but it has already been stated that the Minette ore brought to the Ruhr is richer than the average. The composition runs about as follows: Fe, 32 to 38 per cent. ; Si0 2 , 6 to 8 per cent. ; CaO, 10 to 18 per cent. The usual blast furnace burden in Westphalia carries from 35 to 40 per cent, of this ore, about 35 to 40 per cent, of Swedish (Grangesberg or Gellivare) and about 10 per cent, of spathic ore from Siegerland or brown ore from Nassau, the re- mainder being cinder, pyrites residue, etc. Many of the well known steel works of this part of the country are not of the type familiar to American metallurgists. They are produced by slow accretions rather than by one comprehensive plan, and it is seldom that any contemplated improvement involves the destruction of any part of the existing plant. Oftentimes there is complete discordance between the equipment or the management of separate departments of the same plant, and a new and up-to- date blast furnace will be running alongside a legacy of 1840. A massive new blooming mill will be found supplying small finishing mills that hold together only by the force of habit, while the most carefully built and most economical steam engine, equipped with every possible fuel saving device, will be operated in conjunction with one abandoned by James Watts. These conditions obtain sometimes in America, but they are merely incidental and tem- porary, existing only during a period of reconstruction, while on 746 THE IRON INDUSTRY. the Continent they are typical and are almost universal in the old plants of Westphalia. The contrast between the new and the old is oftentimes a journey from the sublime to the ridiculous, and in a steel works on the Ruhr there is not the excuse for such con- ditions that exist in some other sections. In the newer plants of Lothringen it is openly stated that complicated methods of work and new machinery cannot be introduced owing to the stupidity of the local laborer, but in Westphalia generations of steel making have bred a class of workmen quite superior to those of the country districts, and it is probable that they would handle new machinery in a short time. The work turned out of the machine shops at Essen show that the workmen and -the foremen could use better apparatus than they have, and that possibly a little less patting on the back and a little more shaking up would be a good thing, and the engineering skill and thoroughness evinced in the new armor department render it difficult to understand how the same minds can patiently contemplate from day to day the heirlooms of Tubal Cain that are on every side. It should be stated, however, that a revolution is in progress, for it is recognized that the Essen works are a back number. There are no blast furnaces there and we have the singular phenomenon of the largest works in the country with ancient blast furnaces scattered all over the region and bring- ing iron together from all directions to be converted into steel. This is all to be changed, however, for a new works is now con- structing on the banks of the Rhine, where -water transportation will cheapen the costs of both incoming and outgoing material, and where new methods and mills will be up-to-date and in accord with modern German engineering. This new plant is at Rheinhausen near Ruhrort, and ocean going vessels of 2000 tons burden now come up the river to this latter port, and the advantages of what is practically an inland tidewater situation will be manifest when we consider the large quantities of Spanish and Swedish ores used and the amount of steel exported. The cost of pig-iron made from Spanish ores is given hy Kirch- hoff at $13.75 per ton. The large quantity of ore imported of this kind would lead to the conclusion that the cost of basic pig-iron is nearly as higih, but as a matter of fact this ore is used almost en- tirely by two works, Krupp's and Bochum, these being the only GERMANY. 747 large producers of acid Bessemer steel in Germany. The product is used for special steels, the acid metal being considered preferable. Kirchhoff gives the detailed figures obtained from the annual reports of several companies to show the profits of the industry, It is of course impossible to make any clear statement of profits and losses for these old plants, which have their own sources of raw material and sell everything from coal to machinery, but I have made a rough calculation that in the year 1898-99 the profits of Gutehoifnungshiitte represented $6.00 per ton on a production of 300,000 tons of steel. At Phoenix with an output of 330,000 tons, and at Bochum with 227,000 tons, the profit was $4.00 per ton. The taxes at Gutehoffnungshutte amounted to 44 cents per ton, and the funds put aside for workmen's pensions, etc., footed up 48 cents, per ton, while at Phoenix the taxes were 53 cents and the pensions 30 cents. It must again be remarked that these taxes and pensions include the mines, coke ovens, etc., and that the profits include all the subsidiary branches of the plant, but I have calculated the results on the output of steel, as these plants are miscellaneous steel producers and may rightly be compared with many works in America and other countries. In Krupp's works there are fifteen acid-lined Bessemer convert- ers, each of 5 tons capacity, and at Bochum there are 3 of 8 tons, a total of 18 acid vessels with an average of 5J tons capacity. The output of acid Bessemer steel in 1899, in the Ruhr district, was 118,000 tons. It is quite certain that all these converters were not worked to their full capacity and this is particularly true of those in works outside of Essen, but if we assume that all the acid Bessemer steel was made at Krupp's the production will be only 660 tons per converter per month. In America we do not have many converters of this size, as they have been relegated to the scrap heap, but twenty years ago, when the steel industry was in its infancy and when the old methods of hydraulic cranes and pit casting were in vogue, it was considered that 120.000 tons per year was just about the proper output for two converters of this size, supplied with one ladle crane and pit. In other words, the product for each acid converter in Westphalia to-day is just one- tenth what it was in America twenty years ago. The reasons for 7-18 THE IRON INDUSTRY. this condition may be sufficient or may not be, but the facts are of record. The works of Krupp are not the only ones by any means that are branching out in improvements, for the Rheinische has built what is practically a new works, and the Deutscher Kaiser is a completely new establishment. No attempt has been made, how- ever, either in Westphalia or in Lothringen to change the general system of operation,, there being little tendency to specialization and little thought of steady operation for large production, the controlling idea being that it is impossible to change rolls quickly, and that it is necessary to have spare mills lying idle, ready to start on a different section. The weak point of this plan is that it is almost out of the question to have the same heating furnaces sup- ply two or three different mills and handle the stuff economically, and quite difficult to arrange the hot bed and finishing part of the mills so as to serve two different trains of rolls. In one of the new plants working on different structural shapes, at the time of my visit in 1899, the chaotic condition of the hot bed and cold bed and load- ing department was something which cannot be described. This branch of rolling mill work is the weakest feature of German prac- tice, while the operation of heavy blooming and reversing mills is the strongest. There are a large number of steel works not possessing blast fur- naces at all and one of these at least operates Bessemer converters, but the greater part of the steel, as might naturally be expected, is made by the steel works having blast furnaces either near the steel works or elsewhere, this being true of both Bessemer 'and open- hearth product. All the basic Bessemer plants use "direct metal." The output of acid Bessemer steel is small as explained above, being only one-tenth part of the basic tonnage and the acid open hearth also contributes only one-tenth part as much as the basic furnaces. About half the steel is made in what we may call the large steel plants, meaning by this that they operate both blast furnaces and a Bessemer plant, while the rest was made in small plants and in steel casting works, the latter having 21 furnaces averaging 9 tons each. I am informed by Mr. Schrodter that "there are several works which turn out 32,000 to 35,000 tons in a month, from eitiher two GERMANY. 749 or three basic converters of 18 to 20 tons capacity, using one vessel at a time." I have received personal communications from four German works giving me the actual output of their converters and the data are given herewith. The first four plants in the list are in the Euhr district, while Rothe Erde is at Aachen. Size of Tons per month Works. converter. per converter. Phosnix 12^ tons 7,000 Hoesch 11 tons 8.000 Horde 18 tons 8,000 Rheinische 15 tons 6,500 Rothe Erde 15 tons 7,500 A basic lining in a converter is considered to do well if it lasts 220 heats, while the bottoms average from 45 to 50 heats. It is the practice to run one vessel at a time, and this one vessel will make three heats per hour, since the actual time of blowing is about .twelve minutes. Every sixteen hours the bottom must be changed, while delays occur occasionally from repairs to tuyeres. When such a delay does occur, another vessel is immediately brought into use until the repairs are completed. Sometimes the vessels are used alternately when tihe iron is blowing very hot, and sometimes heats are made out of turn to keep the lining hot on an idle vessel, as a basic lining suffers from becoming too cold. At the end of three days the first vessel will be worn out and the relining takes fifteen hours and the firing about six hours more. While this is going on the second and third vessels must be work- ing and of course there are many times when a fourth unit is needed, the best and newest plants being designed on tlhis basis. Under this system it is easy to see that the output will not increase in proportion to the number of the converters, but each unit ren- ders possible a more uniform output per hour, which tends to economies in the rolling mills. This regularity is of more importance in Germany than in America on account of the use of unfired soaking pits, the use of coal for heating being almost unknown. In some works the Sunday iron is melted in the blast furnaces during the week, no cupolas being provided. The first round of ingots on Monday morning is kept in the pits only twenty minutes, and then rolled into blooms, as it is not hot enough to finish into rails or billets. The next round stays forty minutes, and the next sixty minutes, after which 750 THE IKON INDUSTRY. the mill goes on throughout the week finishing billets, rails, beams, or other shapes at one operation. During a roll change in the -finishing mill, the blooming mill may make blooms or large billets. Moreover it is the general practice to have at least two finishing mills supplied from the same blooming mill, and these run alternately so that one is* always ready. One of these is generally equipped to roll small billets. In this way the converting department and the soaking pits are kept running steadily and the loss from oxidation in the heating furnaces, wlhich is so costly a thing in America, is un- known. To iihe average observer a German plant, turning out from 1000 to 1500 tons per day, seems to be operating at a very low TABLE XXIV-G. List of Westphalian Steel Plants and Blast Furnaces, Giving the Number of Furnaces and Converters and Their Eated Capacity. Note : Figures on blast furnaces are estimated daily capacity; all the steel plants having blast furnaces at the steel works, use direct metal. Name of works. Location. Blast Fur- naces. Bessemer Converters. Open Hearth Furnaces. Acid. Basic. Acid. Basic. Bessemer steel works with fur- naces at works Horde Bergw Horde Dortmund 7160 9-160 3-200 4140 8140 3100) 3 150 f 3270 4-300 418 418 3 It 3-5% 4 12 118 j 718 )2-7 J4 25 11-8 418 7^25 j 615 11-4 14 20 1 112 410 7 15 Union Hoesch Dortmund Bochum 3-8 Gutehoffnungshiitte Oberhausen Phoenix 3-12 415 4-18 112 "i is' 9-10 Rheinische Ruhrort" Deutcher Kaiser . Bruckhausen.... Bessemer steel works with blast furnaces elsewhere Krupp 156 18-21 Furnaces at - Duisburg Hochf eld 3100 3200 280 4 75 Rheinhausen Neuwied Miiihofen Furnaces at Eschweiler 414 Berge Borbeck . . Kupferdreh 2150 1 125 Bessemer Plants without blast furnaces Haspe 36 Stahl Industrie Bochum ..... 28 212 6415 Steel works without blast fur- naces 612 Blast furnaces without steel works 20110 GERMANY. 751 cost, in spite of there being a few more men than would be found in America. There were 147 basic open-hearth furnaces in the Ruhr district in 1899 with an average rating of about 17 tons. Three-fifths of the estimated capacity was in the plants operating Bessemer con- verters, the remainder being scattered in many different establish- ments, six furnaces being used for steel castings. The district is also the great producer of wrought-iron, there being nearly 500 puddle furnaces at work, or nearly half the total number in the empire. Table XXIV-G gives a list of the principal producers of steel and iron, but it will be understood that the estimated capacity of blast furnaces represents a maximum hoped for, rather than a regular production. Thus the seven furnaces at Horde are rated at 160 tons when the figures for 1898 show an average product of 90 tons, and the same reports give 90 tons for the furnaces be- longing to the Union Works, 130 tons for the Hoesch, and 110' tons for Gutehoffnungshiitte. The data for both bl'ast furnaces and steel producers are taken from official sources. SEC. XXI Yd. Oberschlesien, Upper Silesia: In the extreme southeastern end of Germany, surrounded on the north, east and south by Russia and Austria, lies a little district about fifty miles square, which produces half as much coal as the Ruhr Valley, one-fourtfh as much coke, and which stands second among German districts in the production of steel. Isolated by the political frontier lines and by the mountainous character of the country, it forms a factor not only in the industrial world, but in the general political situation, for tariff measures and expendi- tures for internal improvements by railway or canal must be ar- ranged to give this district its share in the benefits, in order that it may not pay taxes to assist a competitor. Coal is found in both Upper -and Lower Silesia, by which is meant both eastern and western, but the iron industry exists only in the east. The character of the population is quite different from that of Western Germany, for Eastern Silesia formed part of the old dismembered province of Poland, as might be inferred from the names of the towns. It is more provincial; wages are lower; the standard of living is not as high, and the proximity of Russian Poland, Austria and Hungary gives rise to a great deal of floating foreign labor. The primitive character of the population 752 THE IKON INDUSTRY. is indicated by the traveling bazaars, temporarily established in the market places of the towns. The wares are the crudest hand- made articles, ranging from shoes to augers, and could not be sold in an up-to-date community except to a museum. Gangs of Rus- sian women travel around in search of work exactly a,s Croatian or Austrian workmen go from one place to another in America, and these women as well as others from Austria and from the home villages, work in the steel works, on the railroads, or any place where there is work to be done, beginning this drudgery at the age of sixteen. Their wages are about 25 cents per day, while men earn from 50 to 62 cents. The principal advantage possessed by Silesia is its coal supply. In 1899 it raised nearly 28,000,000 tons of coal, which was over half as much as Westphalia produced, and it made 1,777,000 tons of coke, nearly one-quarter of the amount turned out in the Eu'hr. The coal is very rich in volatile matter, running from 30 to 35 per cent., but it gives a very poor coke. The quality has been much im- proved in some places by stamping the coal, this being done both wet and dry at different works, but it is even questioned by some whether any good is done by this compression, the burden of evi- dence, however, seeming to be in its favor. The Silesian coal field reaches over the boundary into Moravia and Poland and will be further referred to in the discussion of Austria and Russia. For- merly considerable ore was mined in Silesia, but the supply is de- creasing, for in 1894 there were 600,000 tons raised, while in 1899 there were only 477,000 tons. This ore is very poor stuff of the following composition : Per cent. Iron 25 Manganese 2 to 3 Silica 30 to 40 Zinc 0.8 Water 30 In the dry state this would figure out Fe, 36 per cent. ; Silica, 43 to 57 per cent. ; Zn, 1.1 per cent. The foregoing data were given me on, the spot by the manager of one of the blast furnace plants, and they agree with results recorded by Bremme, Stahl und Eisen, Vol. XVI, p. 755. The figures given by Wedding are as shown in Table XXIV-H. GERMANY. 753 TABLE XXIV-H. Composition of Ores from Upper Silesia. Wedding : Ausfuhrliches Handbuch der Eisenhiitten Kunde, 1897 ; Zweite Auflage; Braunschweig, Fr. Vieweg & Sohn, p. 59. Tarnowitz. Tarnowitz (very rich.) Trockenberg Fe w O, 47.05 50.43 49.06 H.O 12.00 8.11 13.01 MnO 4.30 4.52 7.23 giO, 24.89 25.47 21.29 AlnO 8 88 7.80 5.99 cab.:::::::::::: 96 1.02 1.18 MeO .... 0.04 0.50 0.36 P O 49 76 63 zno :::."::. 2.20 2.21 1.50 Total 100.81 100.82 100.25 Metallic iron wet Metallic iron dry 32.9 37.4 35.3 38.4 34.3 39.4 The ore is very fine and there is an immense amount of flue dust mixed with much troublesome sublimate containing the zinc. About 35 per cent, of lime is needed as a flux. The local furnaces are gradually ceasing to use this ore, but I found the works at Donnersmarckhiitte carrying it to the extent of 50 per cent, of the burden. Foreign ore is now used in the blast furnaces, the amount brought to the district in 1899 being 330,000 tons from Hungary and 275,000 tons from Sweden, the amount of foreign ore smelted being 40 per cent, greater than the domestic product. The Hun- garian ore is a carbonate and is roasted before using. It comes from Kotterbach, south of the Tatra Mountains, some of the mines there being owned by the works at Friedenshutte. A small amount of ore is sent across the border into Austria, but this is a mere local condition. It is rather singular that Friedenshiitte should ihave been one of the first works to install gas engines driven by furnace gas, when the local conditions of dust would make the trial almost a crucial test, and when coal for firing boilers can be had for $1.00 per ton. The steel works of this district are of the usual German type. They are troubled like a large proportion of Continental and Eng- lish plants for lack of water. In America most works 'have been placed in some advantageous position, but in Europe they "just grew," and they seldom are near a sufficient water supply, as a good 754 THE IRON INDUSTRY. sized river, according to foreign standards, carries just about enough water to cool two or three blast furnaces, and condensers are a luxury. This disadvantage is overcome partly by the use of central condensing plants, which are much more common than with us, and by cooling towers, where the water is pumped up about fifty feet and allowed to trickle down over brush or similar devices. The cooling is not enough to give a good vacuum, and the clouds of water vapor are a nuisance in summer and winter, but it is the best that can be done. Many plants use the condensed water to return to the boilers and elaborate settling and skimming tanks are installed to separate the oil, but much remains to be done to give clean water. The statistics for 1899 show that there were 33 blast furnaces in operation, making 745,000 tons of iron, which is an average of 22,600 tons per furnace, or 62 tons per day. There were two acid Bessemer converters of 8 tons capacity, and 7 basic vessels of 10 tons capacity. There were 30 basic open-hearth furnaces, averag- ing 16 tons capacity, in the larger steel works, and a few others in steel casting plants. There are no acid open-hearth furnaces in the district. Silesia is a large producer of wrougiht-iron, there TABLE XXIV-I. List of Steel Works and Blast Furnaces in Upper Silesia. Location. Blast Fur- naces. Bessemer Converters. Open Hearth Furnaces. Acid. Basic. Acid. Basic. Steel works with blast furnaces Friedenshutte Friedenshutte. Konigshiitte. . . f Schwientoch- ( lowitz 4-110 780 3-75 412 2-8 217 /4 12 tl 10 2-15 j 415 M 20 220 I 215 1 120 J3-15. 1120- (1-20 (315 Konigshiitte 1-8 Bethlen Falva Borsigwerk Borsigwerk.... Oberlagiewnik. Gleiwitz 375 370 Hubertushiitte Steel works without blast fur- naces Huldschinsky'che 18 18 Baildonhiitte Kattowitz Bismarckhutte. j Schwientoch- 1 lowitz Blast furnaces without steel works Julienhutte. Bobreck 7-60 375 Three others, one each Zabrze GERMANY. 755 being 287 puddle furnaces in operation, or 30 per cent, of the total for Germany. In Table XXIV-I is a list of the steel works and blast furnaces. SEC. XXI Ve. The Saar: The Saar district is about 40 miles square, with an underlying bed of coal. It includes the neighborhood of Saarbrucken and the western extremity of Bavaria, The coal is not of the best quality and gives a poor coke, which would hardly be used in America, but that it can be used is proven by the steel works at Volklingen. and Burbach. There are four plants in the valley, and three of them make most of their pig-iron at the steel works, but these three, and the fourth also, operate furnaces in Lothringen or Lux- emburg and bring the pig to the Saar. The coal varies considerably, and Wedding states that it con- tains about 7.7 per cent, of ash, but at one works which I visited it ran from 22 to 30 per cent, of ash, and in another from 18 to 20 per cent. In both places it was crushed and washed and the ash TABLE XXIV-J. List of Steel Works and Blast Furnaces in the- Saar District, with the Number of Furnaces and Eated Capacity. Location. Blast Fur- naces. Bessemer Converters. Open Hearth Furnaces. Acid. Basic. Acid. Basic. Steel works with blast furnaces Burbach Burbach 5130 2 -120 411 815 also at Esch Luxemburg 1 RochLng'sche also at Didenhofen Lothrin- Volklingen 5120 2180 6- CO 4 30 4-15 Gebruder Stumm Neunkirchen.-. 412 112 also at Ueckmgen Lothrin- : Steel works with furnaces else- where Dillingen . Dillingen . . . 315 1-15 j 130 1 225 3-15 3-15 Furnaces at Redingen Loth- 2 60 Steel works without furnaces- Weber . Hostenbach . St Ingbert 3-12 Blast furnaces without steel works Halbergehutte 430 756 THE IRON INDUSTRY. reduced to about 10 per cent., giving a coke with 12 to 14 per cent. The coal is charged into the coke ovens in a saturated state holding about 11 per cent, of water, and is rammed with an electric ram- mer before charging, this preliminary ramming compressing the mass so that the coke is much more dense and the amount used for smelting is decreased 10 per cent. By-product ovens are used and the yield of coke is about 70 per cent, of the weight of dry coal. Scarcely any of this coke is carried outside the valley of the Saar, but the local blast furnaces use it exclusively. The ore is all brought from the Minette district, and the mixture is self -fluxing, containing about 31 per cent, of iron, and the pig carries about 2 per cent, of phosphorus, the practice being the same as in Lothringen, save that the coke is inferior to the West- phalian fuel used in the latter place. There are 20 blast furnaces in the Saar, and in 1899 they smelted nearly 600,000 tons of pig- iron, or 30,000 tons eacih, being a little over 80 tons per day, reckoning them as all in operation. There were no acid converters and only three acid open-hearth furnaces, two of these being used for steel castings. There were four basic Bessemer works with 18 converters of an average capacity of 13 tons, and 16 basic open- hearth furnaces of an average capacity of 16 tons. Three of these basic furnaces were in steel casting plants. Table XXIY-J gives a list of the steel works and blast furnaces. SEC. XXIVf. Aachen (Aix la Chapelle) : The immediate neighborhood of Aachen possesses a bituminous coal field which in 1899 raised 1,764,000 tons of coal. Some of this gives a fair coke and the output of the ovens in the above year was 337,000 tons. There is also a deposit of lignite from which nearly 4,000,000 tons were mined. The output of this kind of coal is rapidly increasing for use in making steam and similar purposes, a large proportion of the total being made into briquettes. The ore production is very small, being only 16,580 tons in 1899. There are some scattered blast furnaces which made 153,000 tons of iron during the year. The district is important a.s a steel maker on account of the works at Rothe Erde, on the outskirts of Aachen, This plant makes no pig-iron at its works, but operates five furnaces at Esch in Luxemburg, all the pig-iron going to Rothe Erde for remelting. There are three basic converters of 15 tons each, which made 287,000 tons in the year 1902, or 8000 GERMANY. 757 tons per month for each vessel. There are also three open-hearth furnaces of 25 tons capacity. The Rothe Erde works are very progressive and have a very extensive system of cranes, commanding their storage and ship- ping yards, quite unusual in foreign works and not at all common in American plants. A conspicuous feature is a very high, crane covering traveling cranes of ordinary height and span and transfer- ring material or the smaller and lower cranes themselves. SEC. XXIVg. Ilsede and Peine: In the southeast corner of the province of Hannover, between the towns of Hannover and Brunswick, is a deposit of brown iron ore which is mined by open cut, the bed varying from 6 to 41 feet in thickness. The composition of this ore is given in Table XXIV-K, the material called "washed ore" being obtained by washing the clay from the fine ore produced in mining, thus obtain- ing clean grains of ore. The ore is used raw and is self-fluxing, giving a pig-iron con- taining about 3 per cent, of phosphorus, which is the best for basic Bessemer practice of any iron made in Germany. It is smelted at Ilsede in three blast furnaces of 200 tons each, and the fuel ratio is about 1 to 1. The records of manufacture for 223,000 tons of pig show that 2.925 tons of ore were used per ton of pig-iron, while the coke was 1.008 tons. The coke is brought from the Ruhr, a distance of over 150 miles, with a freight rate of $1.58 per ton, but it has been estimated by Sdhrodter that the cost of pig-iron TABLE XXIV-K. Composition of Ilsede Ores. (Wedding: Eisenhiitten Kuude ; 1897, Zweite ; p. 33.) Aluminous. Calcareous. Washed Ore. Phosphoric. Fe,O,.. 58.26 44.16 62 73 16.41 MnO 7.31 4-72 6.26 1.00 SiO 10 70 3 90 4 87 3 00 A1.O 4.76 1.00 1 02 1.16 Cat) 5 09 21.61 8.90 31.50 MgO 44 91 P O 6 2 46 2 16 4.08 25 96 H,O-J-COo 10.98 22 46 13 14 19.97 Total . . 100 00 iOO 00 100 00 t 00.00 Metallic Iron wet. 40.8 30 9 31.3 11.5 758 THE IRON INDUSTRY. was only about $6.75 per ton, in an era of low prices a few years ago. In 1899, owing to high cost of fuel and supplies, the pig- iron cost $9.10 and in 1900 it was $10.10. A local supply of lignite helps keep the wolf from the door. In 1902 the output of ingots was 239,000 tons, about 20,000 tons per month. The pig-iron is converted into steel at Peine, only .about three miles away, where there are four basic converters of 15 tons capacity. SEC. XXIVh. Kingdom of Saxony: The Kingdom of Saxony, which must not be confounded with tihe province of the same name, is on the border of Austria, touch- ing Silesia on the east, while Bavaria lies on the west. Leipzig is in the extreme northwest and Dresden, the capital, is in the center. It contains a very good supply of fuel, and in 1899 raised 4,500,000 tons of bituminous coal and 1,300,000 tons of lignite. Some of this coal will make coke, and 72,000 tons were so used in the year mentioned. There are some deposits of ore, but the amount raised is unimportant. No pig-iron is smelted, but pig- iron is brought in from outside and the district around Chemnitz shows quite a development of the steel industry. A very small amount of puddled iron- is also made in the Kingdom. There are four steel works altogether. One of them has two acid converters of six tons capacity, which in 1902 made 11,000 tons of steel, and another works has three basic converters of 15 tons, which made 40,000 tons. There is one acid open-hearth furnace of eight tons and eleven basic furnaces of thirteen tons. There are also some small steel casting plants. SEC. XXIVL The Siegen: Siegerland includes the southern portion of Westpihalia and the eastern arm of the Rhine province. It has no coal within its bor- ders, but raises a large amount of ore, most of this latter being a carbonate occurring in mammoth fissure veins, the limits of which are unknown, but which are certainly of great extent. The ore is mined by shafts averaging about 700 feet in depth, and is roasted before smelting, the loss in weight being about 30 per cent. About two-thirds of the output is smelted in the district. the rest going to the furnaces in the Ruhr or along the Lower GERMANY. 759 Rhine. In 1899 there were 2,120,000 tons of ore raised, which was about one-eighth, of the total for Germany. The composition of the ore as given by Wedding indicates about 38 per cent, of iron and 7.5 per cent, of manganese in the raw ore, but this is richer than the average. The calcined ore accord- ing to Brugmann* runs from 47 to 48 per cent, in iron, 8 to 10 per cent, in manganese and 9 to 12 per cent, in residue. Tihe distance to the Ruhr is about 90 miles and the freight 70 cents per ton. The cost delivered is about $4.40, the low phosphorus and high manganese making the ore desirable. There are 32 blast furnaces in the district, four of them being operated by steel works. These four have a daily capacity ranging from 70 to 110 tons, but the others are much smaller, the average rated capacity being only 60 tons. The total pig-iron production in 1899 was only 657,000 tons, which is only about 30 tons per day for each furnace, but it should be said that many of the old furnaces are making spiegeleisen, a considerable proportion of the output running 20 per cent, in manganese. Much pig is used for puddling, there being over one hundred furnaces in the district, or 10 per cent, of the total for Germany. There are four steel works in the district, concerning one of which the German records give no information beyond a question mark. The other three make only basic open4iearth steel, having 12 furnaces of an average capacity of 13 tons. The output of steel in 1902 was 154,000 tons. SEC. XXI Vj. Osnabruck: The district of Osnabruck lies at the junction of Western Hann- over and Northern Westphalia; being only 50 miles in a straight line from the Ruhr it might be included in that district, but it possesses its own coal and ore beds and thus stands partly by itself. In 1899 it raised 550,000 tons of bituminous coal and 128,000 tons of ore. The ore comes from the Hiiggel and though low in phosphorus is very friable. Wedding states that it carries from 37 to 47 per cent, of iron., but Brugmann gives its content as from 15 to 25 per cent., with much moisture. The iron industry is centered in the Georgs-Marien-Bergwerks, at Osnabruck. There are four blast furnaces, and in 1899 the 'Journal I. <& S. /., Vol. II, 1903. 760 THE IRON INDUSTRY. production of pig-iron for the district was 115,000 tons, which would give about SO tons per day for each. There are two acid converters of seven tons, and three basic open-hearth furnaces of twenty tons each. SEC. XXI Vk. Bavaria: The iron and steel industry of Bavaria consists mainly of the Eisen. Ges. Maximilianshutte, at Rosenberg in Oberpfalz. It has two blast furnaces, three basic converters of five tons capacity and two basic open-hearth furnaces of fifteen tons. A small amount of wrought-iron is also credited to this province. The quantities of coal and ore raised and the amount of finished material made are unimportant. SEC. XXI VI. The Lahn: The district known as the Lahn begins at Coblenz and stretches northeastwardly through Hessen Nassau, south of the Westerwold range. It has no good coal, but a small deposit of lignite, and produces a little foundry pig-iron, wrought-iron and steel from local ores and Westphalian coke. It is known, however, for its deposits of red and brown hematites, large quantities being sent to Westphalia. In 1899 the Lahn raised over 750,000 tons of ore, this being about one-third what was mined in the Siegen. The average commercial run of red hematite is about 50 per cent, in iron. Tihe ore is carried 130 miles to Westphalia, with a freight rate of 97 cents;. the delivered price is $3.80 or about 7.6 cents per unit. This neighborhood also supplies ore, carrying 22 to 38 per cent, of iron, 7 to 8 per cent, of manganese, and 18 to 25 per cent, of residues. This is laid down in Westphalia for $3.50 per ton. SEC. XXI Ym. Pommerania: This district is mentioned on account of the new tidewater plant of three blast furnaces of the Eisenwerk Kraft, near Stettin on the Baltic Sea, which is built to smelt imported ore. Coal is brought from England and coked in by-product ovens, the ammonia forming a source of revenue. The iron is all for foundry use and by its situation this plant has easy access to Berlin, this city being one of the greatest markets in the world for such iron on account of the immense business done in miscellaneous castings. SEC. XXI Vn. Other Districts: In the table herewith, showing the output of fuel and iron> GERMANY. 761 figures are given for Central Germany, indicating the large amounts of lignite raised in Merseburg and Magdeburg in Saxony, and in Frankfurt in Brandenburg, but these mines have little bearing on the iron industry. The lignite of this region is not as good as that mined in the Rhenish district, for it contains a very large amount of water, the vaporization of which absorbs such a large amount of heat that the calorific value per ton is greatly re- duced. Ordinary lump bituminous coal will drain after being soaked so that it will carry only two or three per cent, of moisture, but this lignite can and does carry over one-half of its total weight ini water, and yet feel reasonably dry to the hand. A carload may contain four tons of fuel and six tons of water, and this fact ren- ders it far from economical to transport it any distance. CHAPTER XXV. FRANCE. I am Indebted to my friend, Mr. August Dutreux, of the Tie. des Forges ds Chatillon, Commentry et Neuves-Maisons, for a careful reading of the manuscript of this article. SECTION XXVa. General View: The iron industry in France is spread over the whole country, as will be seen in the map in Fig. XXV-A ; many of the seats of in- dustry date back a great many years, but viewed from the stand- point of to-day the control of the situation rests in the ore beds of the Minette district on the borders of Luxemburg and Lothringen. This deposit has been fully described on another page in the dis- cussion of the latter province in the article on Germany, and it was there stated that the ore extended into French territory and is found in the province of Meurthe et Moselle, but other parts of the country must be considered either as furnishing the fuel, or as being the seats of old established industries. The position of the iron business was discussed in Journal I. & S. I., Vol. II, by H. Pinget, secretary of the Comite des Forges de France. This article is principally a condensation of his work, with such addi- tions as have been suggested by subsidiary information; but through the courtesy of M. Pinget I am in possession of the sta- tistics for 1900, and he has also given me in detail the number of converters and open-hearth furnaces in each province and their output. I have grouped these provinces in the way usually fol- lowed by French writers, the results being shown in Table XXV-A. The map in Fig. XXV-A gives the output for 1899, as the later data were not available when it was made. Early in 1900 I was able, through the intercession of Hon. M. E. Olmsted with the State Department, to enlist the services of the American Chamber of Commerce in Paris in the collection of very full statistics concerning the production and consumption of fuel and iron in the different provinces of France. The information so collected was deemed of great value by the Department, and 762 FRANCE. 763 I very gladly agreed that the report should appear as a government publication, which may be consulted by those desiring fuller in- formation on many points. I sent to the Chamber of Commerce a large map of France and requested that the coal and ore fields be marked upon it. The results are shown in Fig. XXV-B, while the following pages embody many facts obtained through the same source. 764 THE IRON INDUSTRY. TABLE XXV-A. Production of Fuel, Ore, Iron and Steel in France; metric tons. Data on fuel, ore and pig-iron, private communication, Struthers, Eng. and Mining Jour. Data on steel and rails from Pinget. Comite des Forges. Data marked thus* are for 1898. Totals disregard data for 1898 and are official. Production in 1899. Coal Coke. Ore. No. of Blast Furnaces in Operation. Pig Iron. Wrought Iron. 4 224,000 61* 1 576 000 213 ooo Nnrth 19 861 000 1 357 000* 12* 297 000 350 000 Centre 6,516,000 '362,000* 148,000 16* 247,000* 80,000* &5outh 3.066,000 233,000 204,000 11* 136,000* 12,000 24,000* 8* 106,000* 9,000 2* 75000* Others 3,421,000 377,000 1* 61,000 Total 32863,000 1,952,000 4,986,000 111* 2,578,000 834000 13 370 000 1,951,000 Exports 1,026,000 Production, in 1900 No. of Steel Works. Bessemer. Open Hearth. Total Steel. Rails in 1901. Total. With Bessemer Con- verters. No. of Con- Iverters. Product. No. of Fur- naces. Product. East 9 4 10 3 1 2 5 34 6 3 1 1 1 1 19 9 2 2 2 3 554.890 232.329 52.128 33,326 45,579 32,909 8 13 43 10 2 5 10 71,104 138,548 261,788 59,769 15,4S4 54.602 68,542 625,994 370,877 313,916 93,095 61,013 87,511 68,542 119.873 72,289 North Centre South Southwest Northwest Others 48793 33,<00 17.859 Total 13 37 951,161 91 669,787 1,620,948 291,814 'SEC. XXVb. The East: Much of the information regarding this district is appropriated from Kirch- hoff's letters, before referred to in the discussion of Lothringen. The eastern division embraces the great ore deposit in the prov- ince of Meurthe et Moselle and the neighboring districts of Haute Marne, Ardenne and Meuse. The map of the Minette district, given in connection with Lothringen, will indicate the position of both mines and steel works. All of the basic Bessemer plants in the Minette district are in the province of Meurthe et Moselle, but the other three make the greater part of the open-hearth product, and their output is constantly increasing. The fuel must be FRANCE. 765 brought quite a distance and a glance at the map will show that the Belgian coal fields are as near as those of Northern France, and since the coke made from the French deposit is not of the best, and since it has been impossible during recent years to get a suffi- cient supply, there is a large amount of coke brought from Ger- many and Belgium in spite of the tariff. The Pompey Company TT 3 - 9 C H has coke ovens at Seraing, Belgium, but as a rule the companies do not control their fuel supply, although very lately the furnaces r. round Lorigwy have united to form a coke company, a plant of r>00 ovens being projected. 766 THE IRON INDUSTRY. In 1898 this district produced 60 per cent, of all the basic Bes- semer steel made in France, and at that time there were only four works in operation, the Longwy, Micheville, Joeuf and Pompey. Other works are building or have already started which will over- shadow these completely, from which some idea may be formed of the complete supremacy of this district as the great producing center. It is customary to consider Meurthe et Moselle as made up of three districts, Longwy, Joeuf and Nancy ; but in reality they are exactly alike in metallurgical conditions. In the Longwy division there are three steel plants of moderate capacity as follows : (1) The Longwy Company, which in 1899 produced 186,463 tons of pig-iron and 158,910 tons of ingots. (2) The Micheville Company, which in 1899 made 172,138 tons of pig-iron and 156,989 tons of ingots. (3) The Societe des Forges de Montataire, with a new works at Frouard, with three eight-ton converters. In the Joeuf district are two steel works : (1) The Soc. An. de Vezin-Aulnoye has a new plant at Home- court, near Joeuf, with six blast furnaces and four eighteen-ton converters, with an estimated capacity of 1200 tons per day. (2) The old plant of Do Wendel, in which Schneider & Co., of Creusot, are interested, has a rated capacity of 500 tons per day, but is of an antiquated type. Owing to the relations existing be- tween France and Germany no railroad connection is allowed with the works, since it brings its ore by rail- from German territory, and all its products are hauled by cart to the existing French rail- road. The third district of Nancy has two steel plants : (1) The Pompey Company at Pompey. (2) A new works being built at Neuves-Maisons by the Com- pagnie des Forges de Chatillon, Commentry et Neuves-Maison. This company is one of the oldest and largest in France and has operated works for many years in the central district at Monti u- Qon, Commentry and elsewhere, and it is very significant when such a new departure is taken and a very large works projected in a district so entirely disconnected with all preceding operations. The new plant is intended to include five blast furnaces and four 18- ton converters. In addition to the blast furnaces connected with steel works FRANCE. 767 above mentioned, there are many others making iron for the gen- eral market and on January 1, 1900, there were 65 furnaces com- pleted, with 54 in blast, the total capacity of all being estimated at 5000 tons per day. It is unnecessary to discuss the metallurgical situation in this locality as it has been covered by the description of Lothringen. Table XXV-B gives a list of the works in this district. TABLE XXV-B. List of Steel Works in the East of France. Those marked (B) have Bessemer converters. Province. Companies. Location. Meurthe-et-Moselle Socite anonyme des Acieries de Longwy (B) Mont-Saint-Martin Societe anonyme des Acie"ries de Micheville (B) Micheville MM. de Wendel et Cie, Maitres de Forges (B) Joeuf Societe anonyme de Vezin-Aulnoye (B) Homecourt Socie"te" anonyme des Hauts-Four- neaur, Forges et Acie"ries de Pom- pey (B) Pompey Socie"te anonyme des Forges et Fon- deries de Montataire (B) Frouard Meuse Societe anonyme des Forges et Aci- e"ries de Commercy Commercy Haute-Marne Compagnie des Forges de Cham- pagne et du Canal de Saint-Dizier Marnaval-Saint- a Wassy Dizier Ardennes MM. Boutmy et Cie, Maitres de Messempre"- Forges Carignan MM. Lefort et Cie, Maitres de Forges Mohon SEC. XXVc. The North: The great coal field of France lies in the provinces of Nord and Pas-de-Calais. It is an extension of the Belgian deposit and ex- tends from the border to beyond Bethune ; the city of Valenciennes may be regarded as a center. The coke made is not of the best quality, but the Belgian is little better, if at all, and the demand has been far ahead of the supply owing to the remarkable devel- opment of the iron industry in Meurthe et Moselle, so that although there are now 2000 coke ovens in operation and many more in process of erection, the price of fuel in France has been almost prohibitive. In the year 1900 coal retailed in Paris at $15.00 per ton and coke for foundry use as high as $10.00. These prices, which were exceptionally high even for France, of course encouraged imports in spite of a duty of 25 cents per ton, and coal 768 THE IRON INDUSTRY. from the United States entered Mediterranean ports, while Eng- land sent 6,000,000 tons of fuel, including coal and coke, and Ger- many supplied considerable coke. Much Belgian and English fuel is imported into the coal region itself, for in 1899 the foreign coal used in the provinces of Nord and Pas-de-Calais amounted to one- sixth of the total consumption. In the province of Calvados in the northwest, a comparatively short distance from the French coal fields > nearly all the fuel consumed was brought from Eng- land. It is the intention of French coke makers to increase the number of ovens so as to render foreign imports unnecessary, but it is doubtful if this increase can affect some of the northwestern and southwestern works, which are close to the sea and which will find English coke cheaper, as well as better. The cost of mining in the Nord and Pas-de-Calais field is enhanced by the depth of the shafts and by the numerous dislocations and contortions of the strata, and the coal must compete on the east with the product of Belgium and Germany and on the west with English fuel. ^ A certain amount of iron has been made in this district, but the great drawback has been the absence of any ore deposit, the supply having been drawn from Meurthe et Moselle, or from Spain and Sweden. For many years there has been a small amount of hema- tite mined in the province of Calvados, but the amount produced has been unimportant. I am informed that there has now been discovered the mother lode of spathic ore in large quantities and of good quality. The freight on this will always be low owing to the continual march of empty cars returning northward to the coal districts, and it is thus possible to establish an iron center in the District of the North. To what extent this may develop remains to be determined. Table XXV-C gives a list of the steel works in the district. TABLE XXY-C. List of Steel Works in the North of France. Those marked (B) have Bessemer converters. Province. Companies. Location. Nord Socie*te" anonyme des Hauts-Four- neaux, Forges et Acie'ries de De- nain et d'Anzin (B) Denain Socie"te" anonyme des Forges et Aci- e'ries du Nord et de 1'Est (B) Trlth-Salnt-Leger Socie'te' anonyme des Usines de la Providence Hautmont Pas-de-Calais SocietS anonyme des Acie'ries de France (B) Isbergues FRANCE. 769 SEC. XXVd. 7% Center: The central district embraces the provinces of Loire, Saone et Loire, Allier, Rhone, Cher, Isere and Mevre. It includes the works at Creusot, Montlugon, Commentry, St. diamond, Firminy and St. Etienne. Notwithstanding this array of names familiar to metallurgists, the output of this part of France may be briefly passed over. It is of small amount and the existing works have gradually become specialized, making certain lines of finished high grade products for a limited market, as, for instance, armor plate, guns and tool steels. The fuel supply is not good, the blast fur- nace coke of St. Etienne in the Loire basin containing an average of 14 per cent, of ash. The supply from Allier, which goes to TABLE XXV-D. List of Steel Works in the Center of France. Note : Those marked (B) have Bessemer converters. Province. Companies. Location. Allier. Isere.. Loire . Compagnie des Forges de Chatillon. Ccmmentry et Neuves-Maisons MM. Ch. Pinat et Cie, Maitres de Forges Gompagnie des Forges et Acieries de la Marine et des Cheminsde fer Compagnie des Fonderies, Forges et Acieries de Saint- Etifnne MM. Claudinon et Cie, Maitres de Forges Nievre Saone-et-Loire... Societ6 anonyme des Acieries et Forges de Firminy . . MM. Jacob Holtzer et Cie, Maitres de Forges Societe anonyme de Commentry-Fourchambault et Decazeville MM. Schneider et Cie, Maitres de Forges. (B) MM. Campionnet et Cie Montlucon Allevard Saint - Chamond et Assailly Saint-Etienne Le Chambon-Feu- gerolles Firminy Unieux Imphy Le Creusot Gueugnon Commentry, Montlugon, etc., is no better, while much of the fuel for the Creusot works comes from the Burgundy basin in Saone et Loire, and for the making of coke must be mixed with one-third of the coal from St. Etienne. Ore is wanting, over one-third the supply being brought from Spain, and there seems to be no future development possible as far as international metallurgy is con- cerned. The whole district in 1899 made only 4000 tons of rails, which was but a little more than one per cent, of the total output of steel. The Creusot works still turn out a very fair product, but much of their pig-iron is brought from more favored districts. This plant makes almost all the few rails made in this part of the 770 THE IRON INDUSTRY. country, and quite a little material for ships, and claims attention on account of its miscellaneous business in machinery, ordnance and structural work; but there is little danger that the establish- ments of Central France will make many conquests in international trade in the lines of heavy machinery or structures until their present methods of hand labor are completely revolutionized. In the southern part of this division Algerian ore is used, as well as some from the Pyrenees. In 1888 there were 24 blast furnaces reported in blast, but ten years later in 1898 only 16 were in opera- tion. Table XXV-D gives a list of the steel works in this district. SEC. XXVe. The South: The southern district covers the provinces of Gard, Aveyron, Ardeche, Bouches du Rhone and Ariege, and includes the coal field of Alais in Gard, which gives a coke that is used in the blast fur- naces of Besseges and Tamari. There is also a deposit in Aveyron, which, though poorer than the Alais coal, will run over 18 per cent, in volatile matter and will give a marketable coke in Coppee ovens. In the southeast there are deposits of lignite, the province of Bouches du Ehone raising 490,000 tons in 1899, and neighbor- ing districts contributing 117,000 tons. Some of this is sent to TABLE XXY-E. ' List of Steel Works in the South of France. Note : Those marked (B) have Bessemer converters. Province Companies. Location. Ariege Societe Metallurgique de 1'Ariege .. . Pamiers Soci6t6 anonyme de Commentry-Fourchambault et Decazevil'le Decazeville Gard Switzerland and Italy. The quality of this fuel, however, is not good and the supply is scant, so that about one-quarter of all the coal consumed in this part of the country is imported from Eng- land, principally for steam purposes. The iron industry has re- ceived an impetus from quite recent developments in the Pyrenees ; these mountains have long supplied ore in moderate quantities, but it is likely that the output will be largely increased. Some ore is also brought from Algeria. In 1888 there were nine blast fur- naces in operation, while in 1898 there were eleven in blast, some FRANCE. 771 of these in the region near the Pyrenees being small and using charcoal for fuel. Table XXV-E gives a list of the steel works in the district. SEC. XXVf. The Northwest (Loire Inferieure) and the South' west (Landes) : Both of these divisions fall under the same head, as both of them import Spanish ores from the north of Spain and smelt with Eng- lish coke. The works in Loire Inferieure also bring some pig- iron from other provinces of France. The production of neither district is of great importance from a general point of view, al- though both contribute quite largely to the rail output. At the works at Trignac, near St. Nazaire, there are three blast furnaces, three 10-ton converters and four open-hearth furnaces, the pro- duction of Bessemer steel being about 2500 tons per month. The names of the works in the two districts are given in Table XXV-F. TABLE XXV-F. -* List of Steel Works in the Northwest and Southwest of France. Note : Those marked (B) have Bessemer converters. Province Companies. Location. Loire-Inferieure Societe anonyme des Acieries. Hauts-Fourneaux, Forges et Acieries de Trignac. (B) Societe anonyme des Forges et Acieries de Basse-Indre Compagnie des Forges et Acieries de la Marine et des Trignac Basse-Indre Le Boucau CHAPTEE XXVI. RUSSIA. I am Indebted to Mr. A. Monell, formerly of the Carnegie Steel Company, for a careful reading of the manuscript in conjunction with a naval attache" of the Russian Government, whose services he kindly requisitioned. Mr. Monell has visited many of the Russian works and his approval of the text renders it of much greater value. The manuscript has also been read by Mr. Julian Kennedy. The information has been gathered from many sources. Much of it has been taken at first hand from the Russian Journal of Financial Statistics and The Mining Industries of Russia, and some from Consular Report No. 555 of the British Foreign Office. A paper by Bauerman, Jour. I. and S. I., Vol. I., 1898, and articles in "Stahl und Eisen," by Neumark and Gouvy, furnished very much in the way of detail, and many other papers were consulted. In the matter of statistics, it often happens that the published figures are contradictory, but the data given here are in accord with those issued from official sources. In consulting statistics concerning Russia, the weights are usually given in poods and the values in roubles. It may be convenient therefore to record that one pood is about 36.14 pounds, and hence 62 poods are one gross ton, or for practical purposes, 60 poods=l ton. A rouble is 51.5 cents and this is one hundred kopecks or copecks. SEC. XXVIa. General View: Within the last ten years Eussia has trebled her production of pig-iron and increased her output of steel fourfold. No other nation can show such a record. The reason, however, is not hard to find, for all the force of an autocratic government has been ap- plied to the building up of home industries in the same way that America and Germany have developed the manufacture of iron by high tariffs. In Eussia ore is admitted free, a bounty is paid on all pig-iron exported, and the freight rates arc VTV low. The government owns the railways, and their requirements, to- gether with the supplies for the war equipment of both army and navy, absorb four-fifths of all the iron produced. This abnormal condition arises from the fact that the one hundred million peas- ants in Bussia use scarcely any iron implements or tools of any kind. They are an undeveloped, mediaeval people, and like the rest of the human race, must learn to know their own needs. As a result there is a very low limit to the capacity of Eussia to absorb 772 RUSSIA. 773 iron and steel and the government may fix its own price in buying material. The policy in the past has been to encourage manufacture, espe- cially in South Russia, and the large dividends regularly pro- claimed attracted large amounts of foreign capital. The New Russia Company, for instance, the oldest and largest steel works, has declared dividends since 1889 of from 15 to 125 per cent. In 1899 the aggregate share capital of foreign companies in Russia was over seventy million dollars, more than half of this being in mining interests, foreign money representing one-quarter of all the mining industry of the Empire. In addition to this proportion in mines there is a very large investment in iron and steel works ; the Belgians especially, and the French also, have built many ex- tensive plants, oftentimes without inquiring into local conditions at all and relying on the government to buy whatever was made at such a price that big dividends could be declared. The Bourses of the Continent swallowed anything with a Russian name, and the public contributed from its hoardings. The inevitable crisis came in 1899 and 1900, the government refusing to pay exorbitant prices, and a process of natural selection is now in progress. The situation of many concerns is indicated by the official report of a French company, which pathetically but almost humorously states that the plant they have built in the lonely forests of the Ural is suffering from "the absence of mines and railways near the works." Naturally, this great crisis has had its effect on the imports of iron and steel and this will be shown in Table XX VI- A. TABLE XXVI-A. Imports of Iron, Steel and Fuel into Russia ; tons. 1897 1898 1899 1900 Pig iron 100000 113000 139000 50000 Iron . ... 300 000 320 000 270000 97000 Steel 90 000 74 000 48000 20 POO Iron and steel goods. Coal .... 270,000 2 150000 280.000 2 500 000 300,000 4000 000 220'.000 4 000 000 Coke 400000 450000 '550000 540'000 It will be noted that importation of iron and steel fell off re- markably owing to the necessity of finding a market for the home production. The imports of coal and coke did not decrease, be- 774 THE IRON INDUSTRY. cause they are brought in to the plants in Northern Eussia and Poland which depend entirely on outside sources of supply. Everywhere in Russia the iron manufacturer has two great troubles : If he is near coal, the ore is uncertain or being rapidly exhausted. If he is near good ore, there is no fuel. In either event the available labor is unreliable and inefficient, for the great majority of the men come from the agricultural class and seldom break off all connection with their native village, many of them working in the factories only during the winter and going back to the farms in the spring. The government watches over them with paternal care. No man can work continuously for twelve hours, and if he works at night the hours must not exceed ten. On days preceding holidays the day work must not be over ten hours, and work must cease at noon the day before Christmas. There are fourteen holidays, in addition to all Sundays, obligatory on all members of the Russo-Greek Church, and there are many other regulations about the making of individual written contracts with each laborer, to violate which is a serious offense for either work- man or employer. For joining a strike a man may serve more than a year in prison, as this would involve a violation of a writ- ten agreement. It should be stated, however, that these rules, although enforced with autocratic completeness, are tempered by regulations that allow for accidents and for extraordinary repairs. The government also insists on very complete arrangements re- garding the health and welfare of the workmen in their home life. The New Russia Company, in Southern Russia, employs 14,500 workmen. Only 150 of these are women, a showing which com- pares more than favorably with conditions across the Austrian border. The company supports a hospital with 106 beds and a dis- pensary with, six doctors, five surgeons' assistants, two midwives, one apothecary and two assistants, the cost of this department amounting to $36,000 per year. It also supports a system of schools costing $75,000 per year, and tea houses, baths, etc., etc. That all this is good cannot be questioned, but that it is a regula- tion of the State bespeaks a paternal government, and bespeaks also a people who need a paternal government, and this is a people who are in a certain stage of sociological evolution and who must de- velop for more than one generation before the common peasant becomes the industrial equal of the artisan of America. As might be expected in a country so great, there are several RUSSIA. 775 different centers of production, and owing to the undeveloped con- dition of transportation the distances intervening between these centers acts as a sort of commercial protection. This is true in every country to a greater or less extent, but Eussia presents ex- RUSSIA '\* _r~^~sj < .X ^T" ~~ZT / ^n" n / /O *t* < sr-a H * ( -4L^_J FIG. XXYI-A. treme examples. -The Moscow district, in the center of Eussia, is 600 miles from the works in Poland, or from those in Ekatennos- lav, while Poland and South Eussia are separated by an equal dis- tance. The Ural 'district is still more isolated, being nearly 900 776 THE IRON INDUSTRY. miles from Moscow, 1200 miles from the Sea of Azov and more than that from Poland. Fig. XXVI-A will give an idea of the general distribution of the iron and steel industry, while Table XXVI-B gives more definite statistics. The total output of steel in 1899 was 1,939,000 tons, one-third of this being made in the Bessemer converter and two-thirds in the open-hearth furnace. The output of rails was 530,000 tons, about one-quarter of the total being made by the New Kussia Company. TABLE XXVI-B. Production of Coal, Iron Ore, Iron and Steel in Russia. Data for 1899 from The Mining Industries of Russia, published by The Mining Scientific Committee of St. Petersburg, 1901. Data for 1900 from British Con- sular Report, No. 555, June, 1901. District. Coal in 1900. Ore in 1900. Blast Fur- naces in 1899. Pig Iron in 1900. Steel in 1899. Wr. Iron in 1899. Tons. tn** o Tons. n-8 o> p-l u stf fij PP 4 "5 g Tons. *i Tons. 1 Tons. *1 3 South.... Ural 10,479,000 355,000 3,950,000 129,000 69 2 26 3 3,117,000 1,612.000 488,000 649,000 32,000 91,000 52 27 8 11 1 1 100 3 33 2 9 4 3 37 102 83 45 5 17 40 135 35 54 9 20 1,474,000 791,000 263,000 239,000 35,000 30,000* 52 28 9 9 1 1 982,000 291,000 282,000 190,000 178,000 16,000 50 15 15 10 9 1 57,000 246,000 66,000 54,000 87,000 19,000 11 47 12 10 18 4 Poland... Moscow. . North Siberia.) and j- Finland] Total . . Imports. . 14,913,000 4,000,000 100 5,989,000 54 239 293 2,832,000 50,000 100 1,939,000 48,000 100 529,000 270,000 100 * Output in 1899. SEC. XXVIb. The South: The predominant factors in Eussian development to-day are the South Russian coal fields in the basin of the Don and the ore beds of Krivoi Rog and Kertsch. The coal deposits cover an area of about 90 miles by 200 miles and are estimated to contain fourteen thousand million tons of fuel. There are nearly three hundred mines opened, some shafts going down 1300 feet, but over three- quarters of the total product comes from fifteen openings. The seams are of only moderate thickness, not exceeding seven feet, rarely over five feet, and as a rule from twenty-four to thirty inches. One seam which is worked is only sixteen inches. The cost of mining is therefore high and during recent years the supply has been far behind the demand. "In the year 1900 RUSSIA. 7<7 the price of coal at the nearest railroad station to a Donetz mine was $5.00 per ton, although good authority gives the average cost at $1.00 to $1.70 per ton at the pit's mouth. The district in 1888 produced 2,205,000 -tons, 6,686,000 in 1897 and 10,479,000 tons in 1900, this being 69 per cent, of all that was raised in Russia. The coal varies from lignite to anthracite the same seam being quite different in places a few miles apart. The anthracite beds are much more extensive than those furnishing the soft coal, but the furnaces at Salin are the only ones using hard coal for smelting. The bituminous varieties are generally high in sulphur, ranging from 1 to 4 per cent. The coke is of poor physical structure and most of the coal needs to be washed, several plants for this purpose having recently been put in operation. The best beds give a coke containing 8 per cent, ash and 1.1 per cent, sulphur, but other coals give a product up to 25 per cent, ash and 4 per cent, sulphur. In 1900 there were made 2,500,000 tons of coke, but not more than one-third the coal used for this purpose could be called true coking coal. The percentage of volatile matter at some plants is 18 to 21 per cent., while in other places the proportion is higher. In 1900 there were 4000 ovens, two-thirds of which were of the Coppee type, no by-product plants being in use. The ore found in the basin of the Don is poor and of little importance, the nearest deposits of any size being in Krivoi Rog in Kherson, on the border of Ekaterinoslav. The deposit is of limited extent and varies greatly in composition and character, the richest ore being pulverulent and giving considerable trouble in the blast furnace on account of this fine condition. Ores below 40 per cent, are considered worthless, the composition of eight samples in the general market supply varying as follows : Fe. 47 to 66 P 01 to .04 SiO 2 . 4 to 26 CaO 0.5 to 2.7 S 0.36 Water 5.0 Neumark gives the following as an average : Fe 60 SiO 2 5 to 8 A1 2 O 3 1 to 2 P. 03 to 1.01 The most striking feature is the great variation in the content 778 THE IRON INDUSTRY. of phosphorus. The amount of ore in sight is very limited and most of the good deposits are owned by companies that smelt their own output and sell no ore. At times the end has seemed very near, but it is now estimated that there is a -supply for the next twenty or thirty years at the present rate of production. It is evi- dent that this is not a bright outlook, as a diminishing ore supply always means a higher cost. To help out in this time of trouble, very large deposits of ore have been opened at Kertsch, about 300 miles to the south across the Sea of Azof, the beds being near the surface so that they can be worked open cut by steam shovels. The layer is about 30 feet thick, but the upper and lower portions are poor and only the middle strata, constituting two-thirds of the whole, are used. The composition is as follows : Fe 40 to 46 Mn 0.3 to 3.0 SiO 2 15 A1 2 O 3 5 to 6 S : 0.1 to 0.2 P 1.5 Neumark considers that this will give the cheapest iron in Eus- sia, and places the cost of pig-iron at from $11.00 to $12.50 per ton. The ore will be used in works now building at Kertsch, and it is also carried to the furnaces in the Krivoi Eog district in spite of its low content of iron. In 1899 the production of ore in South Eussia was as follows : Tons. Krivoi Rog 2,650,000 Local Donetz 180,000 Kertsch 190,000 Manganese ores 100,000 Total 3,120,000 South Eussia in 1887 produced only 161,000 tons of iron ore, but in 1897 the output had risen to 1,898,000 tons, and in 1899 to 3,120,000 tons or over half the entire output of the Empire. In 1900 it was estimated that the Kertsch peninsular would raise 600,000 tons. The tonnage of wrought-iron and steel produced in 1899 was twelve times what it was ten years before. In 1888 this district made only 13 per cent, of the pig-iron and 18 per cent, of the steel made in Eussia ; in 1899 it made over 50 per cent, of both pig-iron and steel. RUSSIA. 779 In 1900 there were 17 works, the most important being given in Table XXVI-C, the new works in Kertsch not being included. The plants are scattered considerably, one large works, the Providence Russe, with three blast furnaces, being at Mariupol, one at Tagan- roth and others nearer the ore in the vicinity of Ekaterinoslav. TABLE XXVI-C. Principal Iron and Steel Works in South Russia in 190Q. Pig Iron. Finished Iron and No. of Men Employed. Tons. Tons. N6W Russian Company Limited 270000 153000 8,319* 210000 170000 6636 Petrovski, Russo-Belgian Met. Co 148,000 107,000 2.689 Alexandrovski, Briansk S. Russian Co.. 146,000 90,000 7,174 Donetz-Yurieff Met Co .... 110000 30 000 3240 95000 76000 2,371 80060 65000 3 122 80000 458 Nikupol-Mariupol Min & Met Co. . . . 76'iX'O 23000 1,619 " Russian Providence " at Mariupol .... 70.000 40,000 1,841 Bulinski (Pastukhoff) 40000 25000 3091 *It has been previously stated, on the authority of the Russian Journal of Financial Statistics, that the number of workmen in 1899 in all the works of the New Russian Co. was 14,500. It is stated in a British Consular Report that the number is 8,319. It is probable that the latter figure omits some of the mines or associated industries. SEC. XX Vic. The Urals : - /, The Ural district presents some problems of peculiar interest to the metallurgists. The ores have long been known and the iron made from the beds of Mount Tagil has been famous all over the world. The deposits are scattered over quite a distance north and south, both on the eastern and western slopes of the range, and lie mostly between 54 and 60 north latitude and 56 and 62 east longitude, an area about 240 by 420 miles. Some of the beds are brown ore, occurring in strata 130 feet thick and containing 60 per cent, of iron after roasting, while other deposits are of mag- netite and are among the most important in the world. The chief center of the Eastern Urals is near Nisjne Tagual, where the hill known as Wissokaia Gora offers a deposit about a mile square, in which the best ore runs from 60 to 65 per cent, in iron. The famous iron mountain of Blagodat is thirty miles north of Nisjne Tagual and three miles from the Kouchwa Station on the Ural Railway. This mountain is seamed with ore running from 52 to 58 per cent, in iron. The more northern deposits in 780 THE IRON INDUSTRY. the Ural district are difficult of. access, but the southern, as above indicated, are on the line of the railway from Perm to Ekatevin- burg. In 1888 this district produced over one-half of all the pig-iron made in Russia. Since then the proportion of both has decreased, but the production of pig-iron has doubled in tons and the output of steel increased nine fold. This development has gone on in spite of the fact that good fuel is scarce. There are very large deposits of coal, but the quality is very bad, the ash running from 17 to 23 per cent, and it gives a very poor coke. The whole dis- trict in 1900 raised only 339,000 tons, or much less than half a ton for each ton of pig-iron. From this it may easily be seen that the almost universal fuel is charcoal, and this is not always of the best. In the southern part pine wood is used and the blast fur- naces are built as much as 59 feet high, this being considered the maximum allowable, but as we go northward the charcoal becomes poorer and the possible height of the furnaces less, so that in the Central Urals they are only 50 feet and in the northern part only 42 feet, the average production for one furnace per day being twenty tons. To the average metallurgist it may seem impracticable to carry on metallurgical operations on a vast scale when charcoal is the only available fuel, but certain things must be taken into account. First : The great iron district of South Russia is 1200 miles away as the crow flies, rather far for Russian railways, and when it comes to water transportation the advantage is all the other way, for the Ural iron works would be shipping down stream. This is an important matter in Russia where there is an immense com- .merce in the transportation of products down river on rafts and barges which are broken up for lumber at the end of the journey, there being no need of a return cargo. Second: The Russian government prohibits the destructive de- foresting of lands, so that the same area may be reckoned as afford- ing a sure supply of charcoal in a given number of years. Third: After allowing for the growth of population and its needs, the Urals will have 40,000,000 acres of pejpetuat forest land, equal to a space 250 miles square, and it is estimated that this will produce charcoal sufficient to make 4,700,000 tons of pig-iron per year. It is also calculated that this charcoal can be made for $4.25 per ton. RUSSIA. 781 Fourth : The ore is abundant and some of it of the best quality. These facts are not disputed and it therefore becomes a question why there is not a more rapid development in the region. This subject was made the occasion for an investigation by the govern- ment. It was shown that onerous restrictions and routine im- posed by branches of the government itself were responsible for much of the trouble, these being in great contrast to the encourage- ment given to industries in South Russia. Possibly quite as seri- ous a matter was the system of land tenure, for it was pointed out that a great part of the land has not yet been allotted to the serfs set free a generation ago, and as no man knows what land he will have or how much he will get, it can hardly be expected that he will take much interest in any part of it, or spend money on im- provements. Another factor in the problem is the law providing that the landed proprietors must furnish steady work to the people living on the estate, and under these circumstances it can hardly be expected that labor saving machinery will be introduced. A most peculiar feature in the situation is the status of what are styled "Possession Works." These are owned by the govern- ment and are leased to individuals or companies. These properties embrace 6,000,000 acres of forest land, equal to an area 100 miles square, and the blast furnaces produce 200,000 tons per year, or one-third the total production of the Urals. The terms of the lease prohibit the proprietor from making any improvements or changes without special authority from the State. There are also numberless petty prohibitions, as for instance, the sub-letting of leaseholds, etc., that render an efficient liberal management entirely out of the question. Coupled to these conditions is the natural opposition of mediaeval feudal landed proprietors to changing the existing order. Some day the spirit of enterprise which is now transforming Russia may take hold of this remote corner of the empire, and when the great plains of Siberia and Eastern Russia are more thickly peopled we may have the curious condition of an immense iron and steel producing district with charcoal as the only fuel. It may also be possible that some of the best ores may be trans- ported 1200 miles to the Donetz coal basin, or that the coal may be taken to the ore. The prohibitive distances intervening between outside countries and the center of the Continent make many things possible when the time comes that the plains of Asia are covered 782 THE IRON INDUSTRY. with cities, or when they will be laid out with railway systems as the Great Desert of our own West has been reconstructed in a gen- eration. At the present time one solution to the transportation problem in the Urals is being given by a company which is building a plant of six 15-ton open-hearth furnaces at Tsaritain on the Volga. The pig-iron will be made in charcoal furnaces in the Urals and be brought 900 miles on barges by river, and it must all be brought on the summer freshet, as the upper tributaries are only navigable at that time. The fuel is naphtha, which will be brought 700 miles from Batoum by way of the Caspian Sea and the Volga. One of the principal works in the Urals is the Nijni Tagual, owned by Demidoff, Prince San-Donato. This is near the ore de- posits of Blagodat and Vissiokaia and has eleven blast furnaces, twelve open-hearth furnaces and a Bessemer plant. The largest works in the Southern' Urals is near the ore mine of Komarowo, but its output is only 2000 tons of pig-iron per month. This ore deposit is a brown hematite, but a little distance 'to the eastward is an immense deposit of magnetite at Magnitnaja or the "Iron Moun- tain." SEC. XXVId. Poland: The prominence of Poland as an iron center rests solely on the fact that with the exception of Ekerinoslav it is the only part of Russia where extensive deposits of coal are found. In 1888 the Dombrova field, in the Bendzin district, province of Petrokov, in Poland, produced 2,376,000 tons of coal, being slightly more than Southern Russia, but in 1899 Poland had increased only to 3,950,- 000, while South Russia raised 6,686,000 tons. The coal of the Dombrovski basin is an extension of the Silesian deposit and gives a much poorer coke than is made from the coal in German and Austrian territory. The blast furnaces therefore bring almost all their supply from Austrian Silesia and Moravia. This condition has caused a very slow development of the coal industry, the in- crease in output in the three years from 1897 to 1900 being only 6 per cent. In this latter year Poland produced 26 per cent, of all the coal raised, the South contributing 69 per cent, and all other portions of the Empire onlv 5 per cent. A small amount of lignite is raised, but in 1900 the output was only 95,000 tons. There are some deposits of iron ore in Poland, and there are nearly one hundred mines where brown hematite and spherosiderite RUSSIA. 783 are found, but the ore is very lean and variable, holding 20 to 50 per cent, of iron and the amount produced is unimportant. In 1899 only 488,000 tons were raised, half of which came from the province of Radom. The composition reported was 30 per cent, of iron in the raw stone and 35 per cent, when roasted. In recent years the ores of the Krivoi Rog have been brought 700 miles to replace the local supply. There are about 40 iron plants in the district, but they are as a rule very small and almost all the iron is made in four works, of which the principal is the Huta Bank- owa, operated by French capital, possessing three blast furnaces making together about 250 tons of iron per day and eleven open- hearth furnaces. There is also quite a forge and tube plant at War- saw, which has had open-hearth furnaces running on imported pig- iron, though blast furnaces are now building. The Briansk Com- pany> which has already been mentioned as having a works in South Russia at Ekaterinoslav, also has a plant in Poland at Grodno. In 1888 Poland produced 51,000 tons of steel and in 1899 it made 282,000 tons, and yet owing to the great advance in South Russia the percentage of total production made in this province was much less at the later period. SEC. XXVIe. The Center: The district of Central Russia is one of the oldest in the Empire and includes an area about two hundred miles square, with Mos- cow at its northwest corner. There is a little coal found here, but it is the worst in Russia, being high in ash and sulphur and of poor structure and of little use in the iron industry. Formerly there were large forests, but two-thirds of this area is now denuded and charcoal has risen to prohibitory prices. There is a limited amount of brown and spathic ores, the latter in the best beds averaging about 50 per cent, of iron, giving 59 per cent, in the roasted ore. The silica is about 10 per cent. The home supply of raw material is so poor that coke is now brought 350 miles from the Donetz basin, and ore from the Krivoi Rog and Kertsch, the distance for the latter being about 600 miles. The principal works are at Tula, about 75 miles south of Moscow, and at Lipetzk, about 100 miles southeast of Tula. At the first named place there are three blast furnaces, each making 120 tons per day, while Lipetzk has two furnaces of larger capacity. 784 THE IRON INDUSTRY. SEC. XXVIf. The North: The district of North Kussia includes the provinces of Peters- burg, Olonetz and Courland. There are some deposits of magne- tites and lake ores, and works have been operated here for a long time, using charcoal as fuel. The present output of ore and pig- iron is small, but by the importation of fuel and pig-iron, mostly from England, a very considerable amount of steel is made. TABLE XXVI-D. Imports at St. Petersburg in 1899. Tons. Pig-iron 9,000 Coke 128,000 Coal 1,639,000 There are several works of some size in the north, the Poutiloff, Nevski, Alexandrovsky, Kolpino and Obeuhof? being in the neigh- borhood of St. Petersburg. The Poutiloff is the largest of these, having two converters and twelve open-hearth furnaces. Another works, the Petrozavodsk, is situated about one hundred miles away at Ladogua. CHAPTEK XXVII. AUSTRIA-HUNGARY. I am indebted to my friends, Ernest Bertrand, general manager at Kladno, and Carl Sjogren, engineer at Donawitz, for reviewing this manuscript and giving much information. SECTION XXVIIa. General View: The dual Empire of Austria-Hungary is often treated as a unit and often as two distinct entities, and it is sometimes difficult to tell whether statistics relate to Austria proper or to Austria-Hun- gary. This is due to the peculiar political relations existing be- tween the two countries, which it is beyond the scope of this article to discuss. The steel production of Austria demands attention on account of the energetic way in which improvements have been made in recent years, and because her metallurgists have always been progressive. It was as far back as November, 1863, that acid Bessemer steel was made at Turrach, in Styria, and this was followed in the next year by Neuberg, and by eight others soon afterwards. The Thomas Gilchrist basic Bessemer process was ushered into the world in 1878 and only one year later the first charge was made at Kladno, in Bohemia. In the same year both Teplitz and Witkowitz adopted the practice. The steel industry of Austria, as far as it is here necessary to consider it, exists in three districts shown in Fig. XXVII-A: Moravia and Silesia in the north and east ; Bohemia in the north- west, and Styria and Carinthia in the southwest. Not one of these possesses all the essentials for cheap production, for Bohemia and Styria have no coke, and Moravia no ore. Moreover, the situation of Austria does not facilitate international trade, especially as Kus- sia, which would be a natural outlet for manufactures, has adopted a very decided protective tariff svstem. For this reason the Aus- trian industry is not specialized and cannot tend toward a heavy production of one line of work, but toward a diversified output, and 785 786 THE IKON INDUSTRY. for this reason also the basic open-hearth is rapidly becoming the- general method of manufacture. Quite a considerable amount is made by the basic Bessemer, but very little by the acid open-hearth, while during January, 1901, there was blown what will probably be the last heat of acid Bessemer steel. The statistics of produc- tion as far as available are given in Table XXVII-A and XXVII-B the latter showing how the basic process has supplanted the work on acid linings. AUSTRIA-HUNGARY. 787 TABLE XXVII-A. Production of Fuel, Ore, Iron and Steel in Austria-Hungary in 1900; metric tons. Index of Authorities ; see Table XXX- A. Province. Bituminous Coal. Lignite. Ore. Pig Iron. Steel. Bohemia 3 590,670 s 17 359,952 9 667 946 281 639 9 214 OOO 19 Styria . . . 2 802 891* 1 151 173 9 275 901 9 205000** Moravia , 1,478 957 9 190 213 9 8 582 9 271 304 9 Silesia 4 697 091 9 1 101 9 70 9 41 81 9 e } 235,000 1B Gallicia 1 166 633 9 76 792 9 2*062 9 Trieste 54 604 9 Other Provinces . 59 194 9 1 108 968 9 66 687 9 72 878 9 127 000 Austria ... .... 10 992 545 21 539 917 1 894 458 1 000 207 781 OOO 1 * Hungary 1 238 855 9 4 292 584 9 1 567 860 9 451 647 9 353 OOO 14 Austria Hungary 12,231,400 25,832,501 3,462,318 1,451,854 l,134,000 l * TABLE XXVII-B. Production of Steel in Austria (not including Hungary). From Kupelweiser; Oesterreichischer Zeitschrift, XLIX., 1901. In 1879, the first basic steel was made. In January, 1901, the last acid Bessemer heat was blown. Year. Bessemer Steel. Open Hearth Steel. Total Steel. Acid. Basic. Total. Acid. Basic. Total. 1879 76,348 3,500 79,848 19,697 19,697 99,545 1880 75,027 17,835 92,862 20,481 20,481 113.348 1881 88',279 31,889 120,168 29,846 29J846 150,014 1882 101,230 57,714 .158,944 39,740 39,740 198,684 1883 101,254 88,429 - 189,683 43,797 43*797 233,480 1884 86,855 70,987 157 842 40,009 40,009 197,851 1885 88,288 76,821 165,109 41,021 41 021 206,130 1886 60,016 105,839 165,855 25,861 11,204 37^065 202,920 1887 67,620 118,379 185,999 18,309 29,631 47,940 233,939 1888 76,533 139,127 215,660 25,572 50,962 76.534 292,194 1889 72,849 126,502 199,351 32,121 77,516 109,637 308,988 1890 76,684 103,180 179,864 29,204 133,808 163,012 342,876 1891 60,713 95,061 155,774 27,800 150,493 178,293 334,067 1892 50,379 100,841 151,220 20,114 180,951 201,065 352,285 1893 48,657 108,104 156,761 19,794 203,894 223,688 380,449 1894 47,784 133,131 180,915 17,729 254,835 272,564 453,479 1895 46,502 127,816 174,318 18,576 304,747 323,323 497,641 1896 46,931 157,216 204,147 21,587 356,973 378,560 582,707 1897 38,713 167,688 206,401 14,764 405,098 419,852 626,253 1898 41,963 184,650 226,613 15,952 480,125 496,077 722,690 1899 38,538 186,643 225,181 18,314 540,894 559,208 784,389 1900 18,214 182,809 201,023 23,196 557,110 580,306 781,329 Owing to the high freight rates and the long distances from the northern coal districts to the southern parts of the Empire a large quantity of coal is imported at southern ports. In the year 1899 ihe total coal raised was 41,000,000 tons, but only 11,450,000 was 788 THE IRON INDUSTRY. bituminous, the remainder being lignite. In the same year the imports amounted to 17,000,000, so that much more bituminous coal was imported than was used. The local gas works at Trieste sell coke for domestic use at $9.30 per ton. A large quantity of Westphalian coke is brought to the blast furnaces of Bohemia and even to Styria, since the coke districts of Moravia and Silesia are as yet unable to meet the demand. There is one large blast furnace at Trieste which uses coke from England and sometimes ocean- borne coke from Westphalia, and some of the smaller charcoal fur- naces in the south often use a certain proportion of imported coke. I was informed by one of the great iron masters of Austria that he had seriously considered the use of American coke in the blast furnaces of southern Austria, but the high prices and high freights of the last two years have been prohibitory. The total production of coke in Austria in 1900 was 1,227,918 tons, almost all of which was made in Moravia and Silesia. The production of Hungary was only about 10.000 tons. To balance the very considerable quantities of coke coming into Austria from Germany, there are large amounts of brown coal (lignite) carried from the region around Teplitz in Bohemia into Germany. It goes northward by water transports on the Elbe to Magdeburg, and even to Hamburg, meeting there the competition of English and Westphalian fuel. SEC. XXVIIb. Bohemia (see No. 1 on Map) : This province is well supplied with fuel although there is no good coking coal. In 1889 it raised 4.070,000 tons of bituminous coal, or nearly as much as Austrian Silesia, while it produced 17,960.000 tons of brown coal (lignite), or over 82 per cent, of the total for Austria, most of the latter coming from the immediate vicinity of Teplitz. Bohemia also has a supply of iron ore which is quite well suited for the basic Bessemer. It carries from 0.6 to 0.8 per cent, of sulphur and is roasted and leached with water to dissolve the sulphates, after which treatment it averages about as follows : Per cent. Pe 42.00 to 48.00 P 1.2 Mn 0.1 S.. 0.8 . The coke is brought from Silesia and Westphalia. The principal works are those of the Prager Eisen Industrie AUSTRIA-HUNGARY. 789 Gesellschaft at Kladno and Teplitz, and the Bohmische Montan Gesellschaft at Konigshof. Kladno has four large modern blast furnaces, a basic Bessemer plant with three converters of 13 tons capacity, a basic open-hearth plant and mills for rolling rails, structural shapes, wire, etc. The blooming mill is strong and in- gots of three tons are rolled into rails and beams in one heat. Teplitz has three basic converters, two heavy plate mills and a beam mill. It receives pig-iron from Konigshof, where there are four modern blast furnaces, a foundry and one basic converter. Until quite recently there was considerable business done in making small ingots and great quantities were made only four inches square, which were rolled directly into small shapes, but this practice is now carried on only at Konigshof and in very small amount. It is found more economical to roll billets from large ingots than to cast small pieces, this being the trend of experience throughout Europe where many plants are giving up the old practice. It is at Kladno that Mr. Bertrand has developed the Bertrand Thiel open-hearth process which has been discussed in Chapter XII. The ore used in the open-hearth furnaces is partly Gellivare (Swedish), and some of this is also used in the blast furnace to reduce the content of phosphorus in the pig-iron to about 1.5 per cent. It is also necessary to mention the steel casting plant of the Skoda Company at Pilsen, which has a high reputation for difficult stern posts, etc., for large ships, and is also equipped with hydraulic presses for guns and armor. Table XXVII-C gives a list of the principal works in Bohemia. TABLE XXVII-C. List of Steel Works in Bohemia (Bb'hmen). This district is marked on the map as No. 1. Name of Plant. Location. No. of Bessemer Converters. No. of Open Hearth Furnaces. Annual Output ; tons. Acid. Basic. Acid. Basic. Prager Eisenindustrie . . . j Boemische Montan etc Kladno... 3 3 2 2 } 160,000 40,000 14,000 Teplitz Skoda Steel Works Pilsen . . 4 SEC. XXYIIc. Moravia and Silesia (see No. 2 on Map) : The coal field which has already been described as covering a 790 THE IRON INDUSTRY. large part of Upper German Silesia extends over Austrian Silesia and into Moravia. As before explained, the coal is rich, but does not give the best of coke. Immediately around Ostrau, where Witkowitz is situated, the quality of the coke is quite good, but in Silesia it is poor. It is however the only coke district east of Westphalia, and forms the nucleus for a considerable iron indus- try. The coke is used not only in Moravia, but in adjoining Bohemia and is shipped across the Eussian frontier to the blast furnaces in Poland, which are almost entirely dependent upon this district for their supply. Some is sent to Styria, but the southern works use much coke from Westphalia on account of the better quality of the German product. The relative importance of the Silesian coal district as it affects the different nations will be seen from Table XXVII-D, which shows the output of bituminous coal from this international field. TABLE XXVII-D. Output of the Silesian Coal Field. Tons In 1899. Germany ; Silesia .' . . . 23,527,000 Austria; Moravia and Silesia 6,252,000 Russia ; Poland 3,905,000 The province of Silesia produced three times as much coal as Moravia, but the latter division made the most coke, as the south- ern portion seems to give the best material for smelting. The one predominant iron and steel producer in this region is the works at Witkowitz in the province of Moravia. This plant draws much of its ore from its own mines in Hungary, the deposit being a carbonate, which is roasted. It makes about one-quarter of all the pig-iron that is made in Austria, the output being about 25,000 tons per month. There are six blast furnaces and two acid lined converters and eight twenty-ton basic open-hearth furnaces, which are operated by the duplex process, the pig being first blown in an acid converter, and then transferred to a basic open-hearth furnace. In this way the wear on the converter lining is minimized and the output of the open-hearth furnaces is about doubled ; the blast furnaces are not confined to narrow limits in the composition of the iron and the whole process is a very attractive solution of a metallurgical problem. Apparently also the financial solution is attractive, but I believe that the work is more expensive than other AUSTRIA-HUNGARY. 791 methods. The works produces large quantities of all forms of rolled steel and has a large steel casting plant which has a wide reputation. In the coal region of Silesia are the works at Trynietz with two acid converters, and seven basic open-hearth furnaces, and mills for the making of rails, structural shapes and merchant iron. Table XXVII-E gives a list of the principal works in Moravia and Silesia. TABLE XXVII-E. List of Steel Works in Moravia (Mahren) and Silesia (Schlesien). This district is marked on the map as No. 2. Name of Plant. Location. No. of Bessemer Converters. No. of Open Hearth Furnaces. Annual Output ; tons. Acid. Basic. Acid. Basic. Witkowitz Bergbau, etc.. | Archduke, Frederic Witkowitz . Witkowitz . 2* 8* 4 7 150,000 25.000 60,000 Teschen . . . 2 SEC. XXYIId. Styria (see No. 8 on Map) : A journey to a steel plant, whether it be in America, Germany or Eussia, is not usually looked upon as a pleasure from an. aesthetic point of view, but there is one exception in a visit to the beautiful valley where the ancient town of Leoben and the steel works of Donawitz lie peacefully hidden in the shadow of the Alps. At the end of the valley, only a few miles away, is a mountain towering in a huge cone nearly 5000 feet above the sea and 3000 feet above the hamlet below. This is the Erzberg or Ore Moun- tain. The whole surface is a layer of spathic ore from 200 to 500 feet thick and it is mined by a succession of terraces all the way up the mountain side. This deposit has been known from the most ancient times, the present province of Styria being a part of the Eoman province of Noricum, from whence came a large portion of the weapons of the Eoman legions and other iron instruments of the Empire. In fact, Styria and Carinthia both claim the "rather doubtful honor," as Tunner expresses it, of supplying the nails for the cross thai; was erected on Calvary. Certain it is that the mines were worked *These converters and furnaces are worked by the "combined" or "duplex" process. 792 THE IRON INDUSTRY. tens of thousands of years before that, for the remains of primeval man have been found beside the unburned charcoal of prehistoric forges. To-day some of the ore is brought to Donawitz, near Leoben, while a large amount is smelted in a new furnace plant erected at Eisenerz, nearer the Erzberg, and there are furnaces also at Hieflau. It is a spathic carbonate of about the following composition: Crude. Roasted. 38.93 51.80 2.15 2.84 FeO I. Crude. . 34.97 II. Roasted. Fe PP O. 16 75 74 04 Mn Mn 3 O 4 . . . 2.98 4.01 SiO 2 8.20 11.04 AloO* 2.09 2.81 CaO 3.06 4.12 MgO 2.92 3.93 COo 27 60 P 2 O 5 0.04 0.05 SO, . tr. 98.61 100.00 When weathered it is a brown hematite containing about 54 per cent, of iron, but the proportion of weathered ore is small. The ore is roasted in kilns, giving an average of about 50 per cent, in iron. It is smelted with coke brought from Westphalia and Aus- trian Silesia, the first of these being 500 miles away in a straight line. The transportation moreover is very expensive from both fields owing to the very heavy grades on the picturesque route through the Steiermark Alps. Many of the blast furnaces of Austria are built upon a plan which is different from the usual American construction. The whole structure rests not upon solid ground on the general level, but on a pier formed of arches, so that one may walk directly un- derneath the bottom. At Donawitz the tap hole is at least fifteen feet above the general level. The mere elevation is nothing unusual, as many American furnaces are built high in the air to 'allow the iron and slag to be carried away in cars, but in Austria it is claimed that the bottom of the furnace must be kept cool in order to pro- vent the cutting away of the lining and the breaking out of the iron. This difference in construction is due very much to a difference in the work to be done. When running on ordinary Bessemer iron for the acid converter, the temperature is high, and graphite is deposited as a protective covering in the interior of the hearth ; but AUSTRIA-HUNGARY. 793 when a very low silicon iron is desired the conditions are quite the reverse. It is safe to say that no American furnaceman will agree to make iron regularly with as low a content of silicon as that which is considered the standard product at Donawitz. I have been given the following as typical : c 4.00 Si 0.10 to 0.30 S tr to 0.03 P 0.08 to 0.10 Mn.. 2.0 to 2.5 This iron is taken to a basic open-hearth furnace in a molten state, and the value of the low silicon need not be dwelled upon. The linings are of magnesite, for in Styria this mineral is abund- ant and it is as cheap as almost any other refractory material. Taken all in all, it may be considered a fortunate thing for the rest of the world that good coking coal does not exist in the Steier- mark. There is a deposit of brown coal nearby, and Styria in 1899 raised 2,624,000 tons or about ten per cent, of the total output of Austria. It is the only province besides Bohemia that does pro- duce a large quantity, but there is no bituminous coal found in the empire except in the northern provinces. The predominant steel producer in the district is the Alpine Montan Gesellschaft and mention has already been made of the fur- nace plants smelting the ore of the Erzberg. The one great steel works is at Donawitz, near Leoben, which has lately been entirely rebuilt. There are also modern plate and universal mills at Zelt- weg. Table XXVII-F gives a list of the principal works in Styria. TABLB XXVII-F. List of Steel Works in Styria (Steiermark). This district Is marked on the map as No. 3. Name of Plant. Location. No. of Bessemer Converters. No. of Open Hearth Furnaces. Annual Output ; tons. Acid. Basic. Acid. Basic. Oesterreichische 13 2 2 160,000 20.000 25,000 Alpin Montan etc 3 THE IRON INDUSTRY SEC. XXVIIe. Hungary: The iron industry of Hungary is considerably scattered, but more than half of all the pig-iron of the country is made in the northern portion in the counties of Szepes, Gomor, Borsod and in their immediate neighborhood. (See No. 4 on map.) Consider- able ore is found in this district, the deposit being a spathic car- bonate which must be calcined. In 1899 there were 1,337,000 tons of ore raised in this field, about 30 per cent, of this being ex- ported. The steel works at Witkowitz in Moravia owns mines here and in 1899 took 200,000 tons of ore from Borsod County, which was nearly all it produced, while a considerable quantity is sent from other mines to Bohemia and German Silesia, the works at Friedenshiitte owning mines near Kotterbach. Out of 67 blast furnaces in all Hungary there are 37 in this Szepes Iglo district. Most of them are small, some use charcoal, but many bring coke from Silesia, as good coking coal is not found either here or in any other part of Hungary. There is a very considerable steel plant of the Eimamurian Salgo Tarjan Ironworks Company at Salgo-Tarjan, this company owning mines in Gomor County and having blast furnaces and rolling mills. About 75,000 tons of steel are made per year from three 7-ton basic converters. There are also smaller works at Ozd, while the Austrian-Hungarian State Railway operates a plant of two basic converters and several open-hearth furnaces, making to- gether about 50,000 tons per year. Another small Bessemer plant it situated at Sohl. In the South is the old established plant at Reschitza, where there are three basic converters and three 20-ton open-hearth furnaces with a capacity of about 40,000 tons per year. The iron for this is made in the immediate neighborhood. These two districts in the north "and in the south make three- quarters of all the pig-iron smelted in Hungary and a larger pro- portion of the steel. The only other district worth mentioning is ^in the southeast in Transylvania, where a larger amount of pig- iron is made than in Reschitza. The great drawback throughout all Hungary is the absence of coking coal and only 10,000 tons are produced per year, this being made in the vicinity of Buda Pest. The Hungarian works therefore are on a moderate scale, and being protected by the government in every way content themselves with supplying the wants of the state railways and of the general home AUSTRIA-HUNGARY 795 market. Table XXVII-G gives the output of fuel and iron in 1899, while Table XXVII-H gives the records of steel production. TABLE XXVII-G. Production of Coal, Ore and Pig-Iron in Hungary in 1899 by Dis- tricts; in metric tons. From Struthers : Eng. & Min. Journal ; private communication. of .? i f* i! Is i eg >> > V m |9 if of 1 s "3 I I I i o I Designation in Fig. XXVII-A. Active blast furnaces . . . 4 32 5 9 6 7 Q 54 Idle blast furnaces 5 2 2 4 13 Pig Iron... 259 698 107 575 76 060 8 314 451 647 1 33?'451* 270 882 135 793 186 230 22 823 1 567 860 Bituminous Coal 7 648 470 018 761 189 1 238855 Coke 10036 10,036 785 010 53 819 1 883 114 1 570 641 4292584 TABLE XXVII-H. Production of Steel in Hungary. From Kupelweiser; Oesterreichischen Zeitschrift ; XLIX, 1901. Year. Bessemer Steel. Open Hearth Steel. Total Steel. Acid. Basic. Total. Acid. Basic. Total. 1880 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 12,854 61,269 51,106 47,163 72,687 60,152 72,976 57,475 54,030 68,493 69,968 80579 73,172 66,567 66,081 41,894 49,842 12.8E4 61,269 51,106 47,163 72,687 75,066 107,817 98,737 99,478 119,806 127,464 146,097 139,714 132,345 137,391 104,030 112,178 8,021 11,384 3,201 4,199 3,100 3,800 4,700 525 8,021 11,384 5,941 18,090 27,932 32,458 48,907 53,234 59,380 69,421 79,483 100,809 154,976 170,965 194,160 228,605 240,586 20,875 72,653 67,017 65,253 100,619 107,524 156,724 151,971 158,858 189,227 206,947 246,906 294,690 303,310 331,551 332,635 352,764 2,740 13,891 24,832 28,658 44,207 52,709 59,380 69,421 79,483 100,809 153,563 176,436 189,862 226,195 229,199 14,914 34,841 41,262 45,448 51,313 57,4% 65,518 66,542 65,778 71,310 62,136 62,336 1,413 3,529 4,298 2,410 11,387 * Of this total, 385,319 tons were exported, mainly to Moravia, but some to Bohemia and German Silesia. CHAPTEE XXVIII. BELGIUM. This article has been submitted to M. H. de Nlmot, secretary of the Association des Maitres des Forges, at Charleroi, Belgium, who has been kind enough to go through it very carefully and give me some figures for 1900 not otherwise obtainable. M. de Nimot objects to my statement that the working people of Belgium are "bound to the vocations of their fathers." I deem it merely justice to him to offer his protest, but I believe that while it may not be absolutely and universally true, as no such generalization can ever be, the argument as herein given portrays a real condition and a real difference between the workmen of Belgium and America. Belgium is essentially a fuel producing country. In 1900 she raised 23,462,817 tons of coal, which is about one-tenth of the production of the United States or of Great Britain. The produc- tion of coke was 2,434,678 tons. Table XXVIII-A gives the main facts about the country, from which it may be seen that about three- fourths of all the coal and also of the coke comes from the province of Hainaut on the border of France, and almost all the remainder from Liege. The Belgian coal mines have reached a great depth, which materially increases the cost of operation, and there is much trouble from the great amount of gas in the beds, causing fearful explosions which seemingly no care can prevent. The average working depth in Hainaut is 1600 feet, while some mines run from 3400 to 3800 feet. It is estimated that the coal supply will last from one hundred to two hundred years longer, this period being the same as that assigned to the deposit of Central France, the North of England and Central Bohemia. The average cost of coal at the mines for the whole country for 1899 was officially given at 10.72 francs =$2.07 per ton, and the average selling price $2.40. In 1900 the cost was $2.78 and the selling price $3.48. The average price of coke was $3.96 at the ovens in 1899, but in 1900 the price averaged $4.18, although blast furnace coke was sold at an average of $3.40 per metric ton. About one-fifth of all the coal raised, and over one-third of all the coke made, is exported, most of these shipments going to France. On 796 BELGIUM. 797 the other hand, the imports of coal amount to one-seventh as much as is raised and a very considerable quantity of coke is brought in, these imports coming from Westphalia across the eastern border while the exports go southward. The Westphalian coke is far su- perior to the Belgian product, but it is economical for the French works to buy the poorer article on account of the short haul. TABLE XXVIII-A. Production of Coal, Coke, Iron and Steel in Belgium in 1900; metric tons. Hainaut. Liege. Namur. Luxem- burg. Total. 16,532 630 6190892 739295 23 462 817 Imported from Germany 1 573 697 1 173 917 " " France 497 088 Exported to France 3 917 765 Coke made 1 748 450 , 686 , ' 228 2 434 678 Imported from Germany . . . '220 753 '* " England. . . 40 559 25*688 Total exports 1 073 313 Exported to France 646 369 Ore raised 247890 247 890 Imported from Luxemburg 1 564 579 321 478 291 783 98539 Others . 2.*>2 236 Pig iron made 1 018 561 Imported from England .... 155 873 France 78603 Germany . . . 53 674 Un States 12*259 Steel made 225,945 429254 655 199 Rails 184 428 Puddled iron 333 981 Finished iron 358 163 Exports of finished iron & steel 415808 Total number of blast furnace- 16 17 6 39 Active in 1901 g 12 5 Number of Bessemer converters 47 Number of open hearth fur naces 18 Av. wage in steel works per day 77 cents 78 cents Belgium formerly raised quite a considerable quantity of iron ore, but her maximum production was reached in 1865 with a total of 1,019,000 tons, the output since then having decreased until now it is only about one-fifth of that amount. Some ore is raised in the province of Luxemburg, which just touches the great Minette deposit that spreads out over the adjoining Grand Duchy of Lux- emburg, now in commercial alliance with the German Empire. It is from the Grand Duchy and from Rhenish Prussia that Belgium 798 THE IEON INDUSTRY. draws most of her ore, although a very considerable amount is brought from Spain to Liege, very little foreign ore going else- where in the country except some containing manganese. The pig-iron from these Spanish ores makes about one-sixth of all the iron produced in Belgium, and it is used almost entirely for the manufacture of acid Bessemer steel. The ores from the Minette district give an iron running from 1.3 to 2.0 per cent, in phosphorus and large quantities are used for puddling and for foundry purposes. In making iron for the basic Bessemer it is a common practice to use a certain proportion of manganiferous ores and slags so that the iron will contain from 1.5 to 2.0 per cent, of manganese. The pig-iron used in Belgium is almost all of domestic manu- facture, about one-sixth of the total output being made in the province of Luxemburg, the remainder being equally divided be- tween Liege and Hainaut. The total production of the country at its maximum is about one million tons per year or just about what would be made by ten furnaces making three hundred tons per day. Three-quarters of all the pig-iron made in the Kingdom is smelted at eight plants, a list of which is given in Table XXVIII-B. TABLE XXVIII-B. List of Important Blast Furnace Plants in Belgium. Province. Name of Works. Location. Number of Blast Furnaces. Capacity per Furnace per Day. ( la Providence. Marchienne .... 3 Hainaut J de Coullet Near Charleroi 4 90 de Monceau Sur Sambre. . . Near Charleroi 2 90 . f Soc. John Cockerill Seraing 6 L'Esperance Longdoz Seraing 2 Liege < Angleur . . Tilleur 4 I 4 Luxemburg. . . . d'Athus Athus . 2 70 The steel is made almost entirely in the two provinces of Liege and Hainaut. The production in 1899 was 718,000 tons or about 60,000 tons per month, but in 1900 this fell to 655,000, while in 1901 it was about 500,000 tons, owing to the great depression in business throughout Europe. Out of 47 converters only 25 are in operation and only 12 open-hearth furnaces are working in the whole country. Over 60 per cent, of the steel was made at Liege, BELGIUM. 799 and the works of John Cockerill made most of the rails that were rolled, amounting in 1900 to 134,000 tons, or about 11,000 tons per month. The great advantages possessed by Belgium are the short dis- tances through which material must be carried, as will be shown in HOLLAND AND BELGIUM SCALE OF MILES 6 10 20 30 40 60 STATISTICS OF PRODUCTION: 1 Unit = 1000 Tons per Year. Distances are in Straight Lines. NOR Mouth of Maas SEA D R E N T H E ' V NORTH, ZUYDER OVERYSSEL 1 / L AU..N D ,i FIG. XXVIII-A. 800 THE IRON INDUSTRY. Fig. XXVIII-A. A circle of less than a hundred miles radius takes in the coal and ore mines and a seaport, while the average haul is much less than this. The wages of labor are also very low, and although it is a common saying that a man works just in pro- portion to the way he is paid, this saying is not always mathe- matically exact. It is perfectly true that a man working for 60 cents a day in Liege does not do as much work as an American laborer receiving twice as much, but it does not follow that he is only half as efficient. It is true that a woman loading coke and ore buggies and pulling them on the blast furnace hoist for thirty cents a day may not do the work done by a buggy puller in Pitts- burgh receiving six times as much pay, but it does not follow that she only does one-sixth as much. There is a chance for a large margin of profit for the manufacturer, particularly in the very great number of cases where some human intelligence and some human hand must be at a certain post, and where the grade of the intelligence and the strength of the hand are matters of little moment. There are multitudes of positions in a steel works where this condition obtains, and in Belgium women fill such positions, receiving a mere pittance. As before stated, they do a very large share of the work that we call "general labor." About ten years ago Belgium passed laws regulating the employment of women and children in mines, and there has been a very marked advance in this direction. In 1870 there were from 8000 to 9000 women and girls working underground in the coal mines. In 1889 there were 3700. In 1891 the women and girls constituted four per cent, of all the working force under ground, while in 1899 they formed only a fraction of one per cent. Of the over ground workers the women and girls constituted 25.1 per cent, in 1891, 24.3 per cent, in 1899, and 23.1 per cent, in 1900. Of the over ground workers at these mines in 1900, in a total of 34,075 people, there were 3787 girls between the ages of 16 and 20, or 11.1 per cent, of the whole. In addition to these there were 2589 girls between 14 and 16, a pro- portion of 7.6 per cent., so that 18.6 per cent, of the entire force was made up of girls between 14 and 20 years of age. Considering the works above and below ground together for the year 1899, concerning which I have the full official statistics, there was a total of 125,258 people, of whom there were 6522 girls from 14 to 20 years of age, or 5.2 per cent. A little calculation from the mortality tables will show that this represents over half of all BELGIUM. 801 the girls of that age that would be found in a community contain- ing that number of people, and after allowing for the infirm it will be 'seen that in the coal mining communities of Belgium almost all the girls between the ages of 14 and 21 work around the coal mines or coke ovens.* It is difficult for an American to appreciate what this means until he sees the conditions on the spot and until he has known what it is to work day and night shift out doors in all weather and in all seasons. It seems inevitable that the same law of pro- gress which has just led Germany to abolish woman labor in steel works, which emancipated woman in England a generation ago, and which never allowed her to consider drudgery in America, will extend its power over Belgium and Austria. When this happens the wages of the men must be increased, as there will be but one wage earner in the houshold. The spread of general intelligence will also have its effect upon even the remote districts. At present the working classes in many places seem bound to their home and to the vocation that their fathers knew before them. This is a sort of mediaeval and pro- vincial idea not entirely absent in other parts of Europe, and it may even be detected in America, but in England and in the United States it cannot be reckoned with in the labor situation. These ideas must disappear and with them will disappear the cheap labor of Belgium, although all history shows that an increase in the wages of the day laborer need not necessarily raise the cost of manufactures. In addition to her production of steel, Belgium turns out a large quantity of puddled iron. In the year 1900 her production of steel was 655,000 tons and of wrought-iron 358,000 tons, a great deal of the latter being exported in the form of structural shapes. Belgium covers an area of only 11,370 square miles and had a population in 1899 of 6,744,532, so that her output of steel and wrought-iron is greater per inhabitant than any other nation. As a result she must seek an outlet and her exports of iron and steel wares amount to nearly one-half her total production. The actual tonnage so shipped, however, is comparatively small, being only one-quarter of the exports of Great Britain. The area of Belgium is only one-fourth that of Pennsylvania, * I have calculated these figures from the official report of the Directeur General des Mines for 1899. 802 THE IRON INDUSTRY. but if we take the southwestern part of the latter State, compris- ing the great coke and iron districts in the counties of Allegheny, Westmoreland and Fayette and as far east as Indiana, Cambria and Blair, we find that this section of the State, though having the same number of square miles as Belgium, contains less than one- fourth of her population. Or if we take the most thickly settled three States in the Union the New England States, Massachu- setts, Ehode Island and Connecticut we find that these three have an area thirty per cent, greater than Belgium and yet have only half the population. These figures may give some idea of the density of population in this ancient state. CHAPTER XXIX. SWEDEN". For the Information herein given concerning Sweden I am principally Indebted to my friend, Hjalmar Braune, metallurgical engineer of the Mining School at Filipstad, who has carefully read, corrected and twice reread the manuscript, a'nd I feel sure that there can be no errors in the text. I have also consulted the Swedish official publication, Kommerscollegii beriittelse, for 1900 for the statistical data in Table XXIX-A and Fig. XXIX-A. Much general information has been taken from L'Industrie Miniere de la, Suede, 1897, by Nordenstrom, and the paper by Akerman in the Journal of the Iron and Steel Institute for 1898. Compared with the greater nations, the quantity of steel turned out by Sweden is of little importance when measured by tons, but she cannot be omitted from special consideration on account of her increasing importance as a source of iron ore, on account of the ancient prestige of her products, and on account of the care and skill with which that prestige is maintained. TABLE XXIX-A. Production of Coal, Ore, Iron and Steel in Sweden in 1900 and 1901; metric tons. Data for 1901 from private communication from Richard Akerman. South 1900. Southeast 1900. Centre 1900. North 1900. Total 1900. Total 1901. Coal... 250,000 250 000 271 509 Ore .... 1 000 1 563 000 1 044 000 2 608 000 2 795 160 Pig 24000 503 000 527 000 '523' 375 Wrought Iron 23 000 165 000 188 000* 165 000* \ Bessemer Steel 9l!(XX) 91,000 77,231 Open Hearth Steel 19000 188 000 207 000 190 877 Total Steel 19000 279000 298 000 269 897 * The classification of wrought iron products is very imperfect and the figures are not Accurate. The chief characteristic of Sweden in the iron industry is her lack of coal and her supply of forests for the manufacture of char- coal. It is quite a safe assertion that had coal existed in Sweden 804 THE IRON INDUSTRY. to any extent the manufacture of iron would be far greater, but her steel would never have achieved its present reputation or com- NOKWAY AND SWEDEN A T PRUSSIA FIG. XXIX-A. manded its present price. There are two or three ore beds of exceptional purity as far as phosphorus is concerned, and the fame of Swedish iron rests on these deposits at Dannemora, Norberg and SWEDEN. 805 Persberg. It is well known that charcoal contains no sulphur, and if the ore after roasting contains none the pig-iron can contain none, even though the blast furnace be working cold. This is a proposition rather startling, but decidedly attractive to the average furnaceman. Up to the year 1895 Sweden produced more wrought-iron than steel, but since then the output of iron has remained stationary, while the output of steel has increased. Ninety per cent, of this iron has always been made on the Swedish Lancashire hearth, an improved form of the ancient device, wherein a mass of pig-iron is caused to melt on the top of a charcoal fire and the melted mass again brought to the top and remelted, all the time being exposed to the blast, by which the silicon, manganese and carbon are eliminated under the influence of a slag of about the following composition: Si0 2 =10 per cent.; FeO=78 per cent; Fe 2 3 = 12 per cent. This gives the softest product that can be made by any steel or iron-making process, and when a charcoal pig-iron, low in phosphorus, sulphur, manganese and silicon, is used with charcoal, the latter being free from phosphorus and sulphur, the product must necessarily be pure. In order to get the proper kind of pig-iron, it is necessary to have an ore free from phosphorus. The usual Swedish ore is a very hard magnetite; the blast furnaces are small, ranging from 40 to 60 feet in height and 7 to 10 feet bosh, with a diameter at the tuyeres of from 3.5 to 6.5 feet. When making pig for the Lancashire hearth the blast is kept at about 300 C. (570 F.) in order to keep the furnace cool, and for the same reason a diameter of over five feet at the tuyeres is not considered good practice, for a larger diameter even with cold blast will produce so high a tem- perature that manganese and silicon will be reduced. A drawing of a Swedish blast furnace for making pig-iron for the Lancashire hearth is shown in Fig. XXIX-B. The pig-iron used in the Lan- cashire hearth runs about as follows in per cent. : . Si 0.10 to 0.50, usually 0.25 to 0.30 Mn 0.10 to 0.30 P 0.01 to 0.03 S 0.00 to 0.02 The composition of a very soft Lancashire wrought-iron, used for electrical purposes, is as follows in per cent. : 806 THE IrfON INDUSTRY. FIG. XXIX-B. SWEDISH BLAST FURNACE. SWEDEN. 807 C 0.05 0.06 Si... 0.023 Mn 0.03 P 0.025 S 0.005 In making Bessemer iron a somewhat higher temperature is allowable and the diameter may be 6.5 feet, at the tuyeres, and the blast may be from 400 C. to 500 C. (750 F. to 930 F.), but even under this practice, and still more surely in the making of pig for the Lancashire process, the temperature of the zone of fusion in the blast furnace is so low that sulphur cannot be eliminated in the slag, and it is therefore necessary to always roast the ores even though they contain but a small quantity of pyrite. This roasting also changes the condition of the iron from Fe 3 4 to Fe 2 3 , and thereby reduces the consumption of fuel in the blast furnace. In making Bessemer iron the aim is to get about 1.00 per cent, silicon and from 1.50 to 3.00 per cent, manganese. The charcoal contains about 85 per cent, of carbon, 3 per cent, of ash, 12 per cent, of moisture and 0.01 per cent, of phosphorus, and the consumption of fuel is such that from 600 to 1000 kg. of carbon are burned per 1000 kg. of pig-iron. In 1897 an accurate calculation showed 144 active furnaces, and allowing for the actual time in blast there was an average produc- tion of 13.1 tons per day. There were 130 works making wrought- iron and steel, and they averaged 12 tons per working day, which may give some idea of the scale of operations in Sweden. It is, of course, true that the average is no measure of the best, but in 1897 the largest blast furnaces were reckoned at 40 tons per day. In 1901 there were 139 blast furnaces giving an average daily product of 13.96 tons for the time they were in operation. In 1893 the produc- tion of Bessemer steel was 84,400 tons, being a trifle more than the open-hearth, which was 81,890 tons. The Bessemer output in- creased to 114,120 tons in 1896, but it is decreasing and in 1901 was only 77,231 tons, while the open-hearth product meanwhile steadily increased, until in 1900 it was 207,450 tons, there being a falling off in 1901 to 190,877 tons. During the year 1900 about one-third of the Bessemer and one-fifth of the open-hearth steel was made by the basic process, the basic Bessemer being used in only one works. The production of crucible steel amounts to a little over 1000 tons per year. Sweden exports large quantities of her iron and steel, the pro- 808 THE IRON INDUSTRY. portion sent to foreign countries varying very much according to general business conditions, but on the whole there has been a tend- ency for the proportion to be less as the growth of basic processes has enabled other nations to make the purer grades of metal. In 1840 she exported 86 per cent, of her wrought-iron and steel; in 1870, 62 per cent, and in 1897, 45 per cent. In 1890 the exports amounted to 225,000 tons and in 1897 to 210,000 tons. In 1900 she exported 356,080 tons of wrought-iron and steel, or about 73 per cent, of her output, showing the effect of the general revival in the iron industry. Having regard to the coal and iron industry alone, we may arbitrarily divide the country into .seven parts. In the extreme south there is the district of Malmohus, which produces about' 250,000 tons of bituminous coal per year, but this has no bearing at all on the iron trade. On the southwest is the district of Elfs- borgs, where two open-hearth furnaces make about 3000 tons of steel per year. In the immediate vicinity of Stockholm, in the districts of Stockholm, Upsala and Sodermanland, a small quantity of ore is mined, and there are eighteen works producing about 7 per cent, of the iron and steel output of the country. In the southern central portion, comprising the districts of Ostergot- land, Jonkoping, Kronoberg, Kalmar and Blekinge, are 21 works making about 8 per cent. A little north of Stockholm is the dis- trict of Gefleborg making 15 per cent. The western central portion, including the district of Vermland, Orebro, Vestmanland and Kopparberg, is the great center of manu- facture. This district in 1900, notwithstanding the great develop- ment in the extreme north in the Gellivare mines, raised 55 per cent, of all the ore produced in Sweden, nearly one-half of this coming from the mines at Grangesberg. This last named ore runs about 55 per cent, in metallic iron and .08 per cent, in phosphorus, and most of it is exported. It is in this region that the old mines of Dannemora, Norberg and Persberg are located, some of which have been worked, for six and seven hundred years, and which have made Sweden famous for the quality of her products. There are 56 iron works in this western central section and in the year 1900 they made 74 per cent, of all the pig-iron and nearly 70 per cent, of all the iron and steel. There were 179 Lancashire hearths, 17 converters making a total of 58,392 tons in the year, and 34 open-hearth furnaces, making 156,110 tons of steel. The SWEDEN. 809 Bessemer converters averaged a little over 3400 tons per year or less than 300 tons per month. The capacity of Swedish converters is from three to six tons. The iron is taken to them directly from the blast furnace and only three to five heats are blown per day. To the outside world, one of the most important features of Sweden to-day is the exploitation of the great iron mines recently opened beneath the Arctic Circle. At present the Gellivare mines are the only ,ones that are well developed. The ore is carried by rail to Lulea on the Baltic Sea, but a railroad is now under construc- tion in a westerly direction across Norway to Ofoten. This port, although so far north, is open all the year, while Lulea is inac- cessible in winter. The railroad is now constructed as far as the great deposits of Kirunavaara and Luossavaara, where surveys indi- cate the existence of over 200,000,000 tons of ore above the water level, and it is expected to complete the line to Ofoten in the year 1903. The Swedish government has limited the amount for ex- port to about 1,500,000 tons per year. The ore run's from 57 to 70 per cent, in iron, the A grade being guaranteed between 67 and 70 per cent, with phosphorus below .05 per cent., but unfortunately there is comparatively little of this kind. The next class runs from 66 to 69 per cent, with phosphorus from .05 to .10 per cent., and so on down to the poorest with 57 to 61 per cent, of iron and 1.50 to 3.00 per cent, of phosphorus. The field has been only partially explored, but it is quite certain that the phosphorus is scattered haphazard throughout the whole deposit, so as to make careful selection necessary, and it also seems certain that the greater part will run from 0.7 to 1.0 per cent, in phosphorus and possibly from 1.0 to 2.0 per cent. The ore is very hard and must be blasted. The sulphur is almost always below 0.10 per cent., the manganese about 0.30 per cent., but titanic acid is present in varying quantities from 0.3 to 1.0 per cent. In the immediate neighborhood are the Routivare deposits, of great extent, but as. they contain only 50 per cent, of iron and carry 11 to 13 per cent, of titanic acid, they can hardly be looked upon as of great value. Some of the older iron mines in Sweden can offer ores of only moderate quality. The great deposit at Grangesberg has been al- ready mentioned as being from 50 to 58 per cent, in iron, from .06 to .27 per cent, in phosphorus and from .03 to .25 per cent, in sulphur. These beds have only lately come into prominence being 810 THE IRON INDUSTRY. made valuable by the development of the basic process. The far- famed Dannemora mines produce about 47,000 tons per year. The phosphorus is extremely low, about .002 per cent., but the iron is about 50 per cent, and the silica from 9 to 15 per cent. The Norberg mines, producing 138,000 tons, give about 52 per cent, iron and from 2 to 32 per cent, of silica. Mention is sometimes made of the famous iron mountain of Taberg, but it is merely a rock carrying 30 per cent, of iron with 14 per cent, silica and 6 per cent, titanic acid. The total exports of ore in 1900 were 1,619,900 tons, of which Northern Sweden, principally the Gellivare district, contributed two-thirds, the rest coming from the districts of Yestmanland, the Kopparberg and Gefleberg. Out of this total 1,390,000 tons went to Germany, 103,000 tons to Great Britain, 99,000 tons to Bel- gium, and 9000 tons to France, while about 19,000 tons were sent across the border into Finland. A large proportion of the ore shipped to Germany was really intended for trans-shipment to Austria, it being impossible to determine the exact amounts. TABLE XXIX-B. List of Largest Works in Sweden. Districts. Name of Works. Nearest Large Town. Steel Output in 1900; tons. f Iggesund Hudiksvall 6.000 Forsbacka Gefle 12,000 (jreneborg ( Hofors Sandviken A vesta Gefle Gefle Falun 20,000 25,000 20 000 Kopparberg j Domnarf vet Falun 50,000 Vertnland -< Munkfors Hagf ors Filipstad Filipstad 6,000 14,000 Nykroppa Filipstad . . . 15,000 Orebro -1 Bofors* Ohnstinehamn .... 5,000 Vestmanland Degenfors Christinehamn 23,000 15000 Upsala Gefle 5000 Motola 6000 Ostergotland -j Finnspang Norrkoping 7,000 * Mainly steel castings, guns, armor, etc. In Fig. XXIX- A I have combined the districts before described and have shown (1) the extreme north, a forest-covered, unsettled country, producing ore alone; (2) the extreme south, producing SWEDEN. 811 COL( alone and the southern central portion, making a small amount of *ron; (3) the central district west of Stockholm in which the iron industry of Sweden is centered. Some readers may inquire concerning the production of Norway, so that it may be well to say that there is no iron made in Norway, and the amount has always been small ; but a great deal of Swedish Lancashire product has been taken to that country and worked into finished articles and exported under the very incorrect name of "Norway iron." This term may now be a fixture in the trade, but has no place in a metallurgical treatise. In Table XXIX-B is a list of the principal steel works in Swe- den, showing their location and production of steel in 1900. CHAPTER XXX. SPAIN. The Information concerning Spain is taken from a paper by Alzola, Jour. I. & S. I., Vol. 11, 1896, and from miscellaneous sources. The only claim held by Spain to our consideration as an iron nation is her position as a source of supply for ore. It has been announced many times that the mines were exhausted, and it is a fact that the ore exported is growing leaner. At some mines con- siderable spathic ore is shipped, which was not considered of any value fifteen years ago, but in spite of the immense amounts of ore produced for so many years the total output has steadily increased, and the year 1899 saw by far the greatest record, the output of the mines being 9,400,000 tons, four-fifths of which was raised in the region around Bilbao. Quite a considerable quantity of this is smelted in the neighborhood of the mines, and there are a few steel works of considerable magnitude in the district, the fuel being drawn from coal mines in Asturias, about 200 miles west of Bilbao. The local steel works, however, use but a small pro- portion of the ore output, and in 1900 over 90 per cent, was ex- ported, the port of Bilbao sending out two-thirds of the whole. England claimed nearly three-quarters of the shipments and Ger- many the greater part of the rest. Detailed figures are shown in Table XXX-A and are illustrated by Fig. XXX-A. The Bilbao ore proper comes from an area about 15 miles in length and 2% miles in width. Four classes are distinguished :* ( 1 ) Vena, a soft purple compact and often powdery hematite. (2) Campanil, a compact and crystalline red hematite, often ac- companied by rhombohedra of carbonate of lime. (3) Rubio, a brown hematite usually mixed with silicious ma- terial. (4) Carbonato, a grey granular and silicious or a creamy white laminated and crystalline spathic iron ore. * Brough, Cantor Lectures Soc. Arts, Man. and Commerce, Feb., 1900. 812 SPAIN. 813 Vena is the purest of these and was the only one used in the ancient local Catalan forges. Carnpanil, on account of its low phosphorus, is the most valuable, but is now nearly exhausted. Rubio is the most abundant, but is likely to be mixed with veins of iron pyrites. Carbonate is found usually below the other ores. The district is divided into seven parts, of which the Sommorosto 814 THE IRON INDUSTRY. produces half the. total from the beds of Triano and Matamoros. The other districts are Galdames, Sopuerta, Ollargan, Abondo, Alonsolegui and Guenes, each of which yields a supply for ship- ment. The Vena ore runs about 56 per cent, in iron; Campanil about 54 per cent., and the spathic ore from 40 to 45 per cent,, giving 55 to 60 per cent, after roasting. The composition of Kubio ore, which is the great bulk of the hematite shipments, was the subject of discussion by William Whitwell, in his presidential TABLE XXX-A. Spanish Ore Production and Exports. 1899. 1900. Production- (Province of Vizcava 6 495 564 5 317 920 1 158 169 1 117 017 Northern part. ' Oviedo 65 944 6l'oOO ' Guipuzcoa ... . 27 618 17 476 f Province of Murcia . 668 947 806609 Southern part . ' Almeria and Grenada } ' Sevilla 537,144 309 688 562,758 365 434 1 Malaga and Jaen 66 575 68 691 Northwest 4 Lugo H'OOO 104 no Others 54085 99 131 Total 9 397 734 8 520 146 Exports- {From Bilbao in Vizcaya 5512067 4 556 317 Northern ports. Santander in Santander 673,807 612,109 Castro Urdiales in Santander. . (From Carthagena in Murcia 662,715 430,255 674,690 436,462 ' Porman in Murcia 120 120 128,180 Southern ports ' Garrucha in Almeria 405,153 312,087 ' Almeria in Almeria 188,858 246,351 Sevilla 319 026 339,432 Destination Great Britain 6 224 229 5 484 323 Germany via Holland 1 416 198 1 268 623 Germany direct 128 251 172 496 443818 450,749 Belgium . 254 860 247 351 United States 32,422 195,961 Other countries 13 359 3 758 Total 8 613,137 7 823,270 address before the Iron and Steel Institute. He compared the analyses reported at his own works at Thornaby, near Middles- borough, during eleven years, and they showed a constant decrease in quality. Since the determinations are averages of a very large number of cargoes in each case, and are given under such author- ity, they must be accepted as representative. SPAIN. 1890 1900 Fe in ore as received 50.50 47.99 SiO 2 in ore as received 7.10 10.09 Moisture 9.00 9.10 Fe in dry state 55.50 52.80 The spathic ore, which has been lately considered of much value,, runs from 40 to 45 per cent, in iron, giving from 55 to 60 per cent, after roasting. In addition to the well known deposits of Northern Spam, there are very extensive deposits on the Mediterranean, the principal ore centers being in the provinces of Murcia, Almeria and Malaga. It is from Murcia that the well known Porman ore comes, the mines being near to Carthagena. This is a brown hematite rather high in silica and containing a certain amount of lead, which is not a desirable thing around an iron furnace. There are other deposits farther inland, the deposits of Morata being ten miles from the coast and those of Calaspara about 85 miles, the latter ore being a red hematite running about 57 per cent. Some magnetite of poorer quality is also found. Almeria produces the Herrerias ore, contain- ing on the average of about 52 per cent, of iron and 8 per cent, of manganese, which is used for the manufacture of spiegel, and it also furnishes the Sierra de Bedar ore from the mines of Jupiter, Por- fiado and San Manuel. Some of the Bedar ore is fine and runs about 60 per cent, in iron when dry, while other mines give a purple lump ore running about 50 per cent, in the dry. The Sierra Alha- milla deposits at Los Banos, Alfaro and Lucainena are also in this province. They are remarkably low in phosphorus and are in the form of big hard lumps, and command an extra price for use in open-hearth furnaces. In the provinces of Malaga are found the ores of Marbella, the mines lying about three miles from the coast and about thirty miles southwest of Malaga. This is a magnetite containing about 60 per cent, of iron. There are other deposits in the vicinity 'of Estepona and Eobledal. The province of Sevilla also produces a considerable quantity from the mines of Pedroso and Guadalcanal, but the ore must be carried over fifty miles to Sevilla and this port cannot accommodate vessels of a large size. The province of Huelva furnishes the Eio Tinta ore, which is a hard and lumpy, but sulphurous deposit. CHAPTER XXXI. ITALY. A certain amount of iron and steel is made in Italy, the whole country in 1899 having in operation 21 open-hearth furnaces, two Bessemer and two Robert converters. Most of the steel was made from imported pig-iron and scrap. The Terni works is the largest plant, and in 1899 it imported 90,000 tons of material, converting this principally into supplies for the railways and the navy. As the amount of pig-iron imported into the country is from six to eight times as much as is melted within its borders little need be said regarding this industry. It is necessary, however, to make mention of the mines of Elba, which have been famous for cen- turies and which have supplied America with large quantities of low phosphorus ores. These deposits are controlled by the Italian government, which has leased them for short periods to contrac- tors, but now has followed the wiser plan of giving a long lease. The terms of the contract, made in 1898, are intended to encour- age the manufacture of iron and steel at home. The government is to receive a royalty of ten cents per ton on all ore smelted in Italy, but it must receive $1.50 on all ore shipped to other coun- tries. The company securing this lease is made up of home capital in the Island of Elba, and it is developing coal mines across the ocean in Venezuela for a supply of fuel. The lease runs twenty years, and not over 160,000 tons per year may be exported, while at least 40,000 tons must be offered to Italian furnaces. An important point in the general problem is that in the past the ore has been taken away from Elba as return cargo in vessels carrying coal to Italy, and if such exports cease the cost of coal and coke will be higher. A still more important matter is the ap- proaching exhaustion of the deposit. The government has care- fully surveyed the remaining supply and has limited the output so that it will last twenty or thirty years at the rate of about 250,000 tons per year. Needless to say the working of the lessening and 816 ITALY. 817 narrowing beds, scattered over a considerable area, will be done at a considerably increasing cost. It is safe to say therefore that the mines of Elba can hardly be viewed as an important factor in the international iron trade. TABLE XXXI-A. Exports of Ore from Elba in 1899. Tons. Great Britain 102,700 Germany via Holland 53,300 United States 41,700 France 29,000 Total 226,700 CHAPTEE XXXII. CANADA. Up to the year 1901 the iron and steel industry of Canada was of little importance, but it has now come to the front as the land of new enterprises of very considerable magnitude. The Cramp Steel Company is erecting a plant at Collingwood, Ontario, for making Bessemer and open-hearth steel,, while a very extensive system of industries, of which a steel works is only a part, is de- veloping on the Canadian side of the Sault, between Lake Superior and Lake Huron. The Bessemer plant connected with this latter enterprise consists of two six-ton converters and was started in February, 1902. It is for the future to say just how great all these works will become, but it is the intention now that they will follow the current American practice of smelting the rich ores of the Canadian Lake Superior region with coke brought from the- coal fields of Pennsylvania or West Virginia. Another plant is on different lines and presents points of inter- est to the metallurgist. The Dominion Iron and Steel Company has built a steel works at Sydney, Cape Breton, at which point the company owns very extensive fields of rich coal, giving a coke- which has been successfully worked in blast furnaces. The high percentage of volatile matter leads to the hope that a large excess of gas will be available for use in open-hearth furnaces. The coal varies considerably and some beds are quite high in sulphur, so that for the production of coke it has been found necessary to- wash the coal. Table XX XII- A shows the composition of the raw- material as publicly stated by the management. The ore, which goes by the name of Wabana, comes from Great Bell Island in Concepcion Bay, Newfoundland, about 35 miles from St. Johns, and about 400 miles from the steel plant at Syd- ney. It is easily mined, being in well defined thin layers and of a. brittle nature, but it is not of the best quality, as shown in the table just given. It will give a pig-iron running about 1.5 per 818 CANADA. 819 cent, in phosphorus, which is rather low for basic Bessemer prac- tice and rather high for economical working in an open-hearth furnace. TABLE XXXII-A, Composition of Fuel and Ore at Cape Breton. Raw Coal. Reserve Mine. Caledonia Mine. Dominion Mine. Moisture 1 45 1.54 1.21 Volatile Matter.. Fixed Carbon . . . Sulphur 32 45 60.45 1 64 30.86 62 91 1 50 31.89 61 49 1 56 Ash 5.65 4.69 5.41 Washed Coal Moisture 1 01 1.08 0.84 Volatile Mattter. Fixed Carbon Sulphur Ash 32.99 62.21 1.11 3 79 33.92 61.69 1.07 3.31 37.86 62.60 1 17 4.50 Retort Coke- Sulphur 0.91 0.78 1.01 Ash 6 07 5 38 6 24 Bell Island Ore. Best. Worst. Moisture 1 50 2.50 Fe 54 43 51 84 SiO, 9 34 13 00 p . 744 835 s 05 03 There will be four blast furnaces 85 by 20 feet and ten 50-ton open-hearth furnaces of the Campbell type. The first steel was made on December 31, 1901, and the plant has been com- pleted during the summer of 1902. There are to be 400 Otto Hoffman by-product ovens, which will be similar to those which have been in operation near Boston, Mass., for making gas for city use from Cape Breton coal. The steel plant of Sydney is in a good har- bor, but this is closed by ice a part of the year, during which time traffic can be carried on by way of Louisburg, about forty miles by railroad on the south coast. The ore deposit at Bell Island is also on good water, but is likewise ice-bound for three or four months in the year. One of the great arguments advanced in favor of new works in 820 THE IRON INDUSTRY. Canada is the bounty offered by the government on pig-iron and steel manufactured within the Dominion. The bounty is to grow less in the future and expires completely in 1907. The schedule appears in Table XXXII-B, by which it appears that a company making steel from native ores receives a bounty of $2.70 per ton of pig-iron and $2.70 per ton of steel, or say about $6.00 per ton of finished product. From this it declines to nothing in July, 1907. TABLE XXXII-B. Canadian Bounty on Iron and Steel, per ton. Pig Iron. From From Steel. Native Foreign Ore. Ore. ToApril21 1903 $3 00 $2 00 $3 00 April 21, 1902, to July 1, 1903. July 1, 1903, to July 1, 1904. . . 2.70 2.25 1.80 1.50 2.70 2.25 July 1, 1904, to July 1, 1905. . . July 1, 1905, to July 1, 1906. . . July 1, 1906, to July 1, 1907... 1.65 1.05 .60 1.10 .70 .40 1.-65 1.05 .60 CHAPTER XXXIII. STATISTICS OF THE IRON INDUSTRY. In Tables XXXIII-D to M, inclusive, are given statistics of the production of coal, iron ore, iron and steel in the leading nations, and the imports and exports for the most recent year for which complete statistics are available, the official reports for different countries and for different branches of the same Government often appearing at different times. In the case of some countries certain information can hardly be obtained at all, as, for instance, in regard to the production of wrought iron or of lignite in the United States. In other cases there is much difference in the way the figures are usually given. In the United States the production of steel is always given in the ingot weight. We do have a figure of finished rolled material, but this includes all the wrought iron. In Eng- land the ingot is also used, but in some other countries the data are given for the finished bar, while in Belgium the records show the weight of the blooms or billets in the intermediate stage. Any one of these systems has its good points, but comparisons are difficult. Judging from my own ignorance in the matter, it is doubtful if most people appreciate the difficulty of obtaining accurate statistics of production, and it may be well to illustrate by referring to an attempt to get data for Germany, which is supposed to have a complete system. In Table XXXIII-A are given the various fig- ures encountered. The data from Wedding were collected exclu- sively for this book and as they disagreed with some other records, an investigation was made for me by Consul General Mason in Berlin, with the results accredited to him in the table, the divisions used being the customary items given in German statistics. The dif- ferent figures were then sent to Mr. Schrodter and I asked for an explanation of what is meant by finished steel, and whether the same metal could appear twice in Mason's tabulation. Mr. Schrodter states that not until the year 1900 were any records kept 821 822 THE IRON INDUSTRY. of the output of ingots, but he does not cast any light on the ques- tion of duplication. He does state, however, that the amount of fin- ished material in 1900 was 6,361,650 tons, which is the amount given by Mason as the total output. He also states that the total production of ingots and castings was 6,645,869. Now this is the same thing as saying that the weight of finished material was 95.72 per cent, of the weight of the ingots, a difference of only 4.28 per cent, to account for all scrap and oxidation, and while the losses from these causes may be much less in Germany than here, I can hardly believe that the figures are correct. TABLE XXXIII-A. Discordant Data on Steel Output in Germany. Source of Information. 1898 1899 1900 1901 Swank Am I. & S Ass , 1901 6 328 666 6 365 259 Miner*! Industry 19 1 5 734 307 6 290 434 6 645 869 Rentzoch. 5 0% 896 5 667 050 6 645 869 Gemeinfass, Darstel 1901 4 352 831 4 791 022 4 799 000 Wedding* 4 967 770 Mason * ingots 441 601 467 721 352 935 Blooms, billets, etc .... 986 572 1 040 670 1 183 128 Finished steel 4 352 831 4 820 275 4 825 587 Total 5 781 004 6 328 666 6 361 650 Schrodter; * steel castings . . 107 210 Bess and O H ingots 6 287 OL2 Total . 6 645 S69 6 394 222 * Private Communication. A great deal of confusion is caused by differences in nomencla- ture and classification in different countries by different statisti- cians. The term "iron and steel productions" may include pig-iron and it may not. The term "bar-iron" may mean wr ought-iron, or it may include steel, as soft steel is called ingot-iron on the Con- tinent. Sometimes steam engines are included in "iron and steel exports," and sometimes they are classed under machinery. It is difficult to find the truth without a detailed analysis of the original records, and if professional statisticians are guilty of grievous er- rors, I trust I may be pardoned for any that may creep into the data herewith given. In almost every case I have indicated the authority by putting a small distinguishing numeral. The key to these numbers is given in Table XXXIII-B. STATISTICS OF THE IliON INDUSTRY. 823 TABLE XXX11I-B. Key to Numbers Denoting Source of Statistical Information. * Swedish Offlc. Stat., 1900. 2 Swank : Am. I. and S. Assoc., 1900, 1901 and 1902. 3 Gemeinfass. Darstell. des Eisenhiit- ten, 1901. 4 TJ. S. Geol. Survey. 5 Min. Ind., 1900. 8 British Iron Trade Assoc., 1900. '.British Home Office Reports, 1900. Russian Journ. Financial Stat., 1899. l ^truthers : Sci. Pub. Co. ; private communication. 10 Nimot : Belgium ; private communi- cation. 11 Wedding : Berlin ; private communi- cation. 12 Iron and Coal Trades Rev., Jan. 5, 1900. 13 Min. Ind., 1893. " Oesterreich. Zeitschrift, XLIX, 1901. 15 A. von Kerpeley, Vienna ; private communication. 16 Bertrand, Kladuo ; private communi- cation. "Verein Deutscher, E. and S.. Ind., 1882. "Iron, Vol. XXXIII, p. 376. 18 Stahl und Eisen, Vol. IX, p. 445. 20 Stahl und Eisen, Vol. X, p. 164. 21 Kintzle; Journ. I. and S. L, Vol. II, 1690. -- Stahl und Eisen, Vol. XI, p. 428. 2:5 Swedish Offic. Stat., 1890. 14 British I. T. Assn. Bulletin, No. 20. 25 Swedish Offic. Stat, 1892. 2 " Stahl und Eisen, Vol. XII, p. 1007. - 7 Swedish Offic. Stat., 1893. 28 Verein ' Deutscher, E. and S., Ind., 1893, No. 17. 20 Verein Deutscher, E. and S., Ind., 1894, No. 21. 30 Journ. I. and S. I., Vol. II, 1894. 31 Verein Deutscher, E. and S., Ind., 1895, No. 20. 32 Stahl und Eisen, Vol. XVI, p. 395. 33 Journ. I. and S. I., Vol. II, 1896. 34 Swedish Offic. Stat., 1897. s - Stahl und Eisen, Vol. XVIII, p. 38. 35 Swedish Offic. Stat., 1898. 37 Stahl und Eisen, Vol. XIX, p. 32. 38 Comite des Forges Bulletin, 1458. 39 Stahl und Eisen, Vol. XX, p. 39. 40 Akerman ; private communication. 41 Comite des Forges. 42 Mining Industries of Russia, 1901. 43 Schrodter ; private communication. 44 British Iron Trade Ass'n, 1901. 45 British Consular Report, No. 555. A complete statistical digest cannot be attempted in such a lim- ited space, but it is desirable to find the main conditions in order to know the internal economy of each nation and its relation to the world at large at the beginning of the new century. The tables show that the iron producers may be divided into three classes according to the quantity of pig-iron and steel they produce. First, and almost in a class hv itself, is the United States; next come Germany and Great Britain, the latter producing slightly more pig-iron than Germany, but verv much less steel. These three nations produce eigfttv per cent, of all the coal, pig-iron and steel made in the world, and nearly seventy per cent, of the iron ore. In the next class are France, Russia, Austria and Belgium. These four nations produce about eighteen per cent, of all the pier- iron and steel made in the world, and about fifteen per cent, of all the coal and iron ore. "i-o fi,;-n^ n i ncs inolri^ps Sweden and Spain, which are important 824 THE IRON INDUSTRY. as sources of the iron ore supply for the greater nations, but which have no coal for smelting. In the same list, but of less importance, are Greece, Algeria, Cuba and Italy, which are widely known for their ore mines, but produce very little or no iron. Another way of comparing the nations is according to the amount of pig-iron produced per inhabitant. This is done in Table XXXIII-C. TABLE XXXIII-C. Production of Pig-iron per Capita in 1899, Pounds. Great Britain 505 United States : 405 Germany ... 330 Belgium 322 Sweden 244 France 145 Austria- Hungary 67 Russia 46 Italy 1 The United States may be looked upon as self-contained, pos- sessing within its borders all the material necessary for the iron industry. A certain amount of ore is imported for use in plants near the seaboard, and some small lots of foreign pig-iron find their way into distant portions of the country; but the proportion of imports to the total consumption is very small for either fuel, ore, iron or steel. This condition arises in great measure from the geographical isolation of America and the almost prohibitory dis- tances from sources of supply. To understand the totally different conditions in Europe it is only necessary 'to consider that the boundary of France touches the coal fields of Belgium, and the boundary of Belgium touches the ore fields of Luxemburg. The close geographical relations of the countries in Northwestern Europe naturally give rise to inter-traffic in raw materials, when unhampered by foolish tariff restrictions on such article?. The iron industry of Belgium is founded on imported ore, while France, Germany and England bring from one-fifth to one-third of their ore supply from beyond the boundary. With coal also it is neces- sary to disregard the political limits, and in some cases the figures seem contradictory, as when a nation both imports and exports large quantities. This may be explained by local conditions, as for in- stance on the eastern boundary of Germany we may find coke going into Austria and brown coal returning into Germany, this brown STATISTICS OF THE IRON INDUSTRY. 825 coal being cheap and perfectly suitable for heating, but not fit for smelting. There is room for difference of opinion as to just how percentages should be calculated, but I have compared the quantity of imports with the quantity actually used. Thus the amount of iron ore used in a country is the tonnage raised plus the imports minus the ex- ports, and I have found the proportion of imports to this tonnage so smelted. In calculating the exports, however, I have taken into account only the quantity raised and the quantity exported, so as to find the proportion of the home production which was sent away from the country. Any other basis of calculation will be found to give curious results in the case of countries that both import and export large quantities. Taking up the question of fuel supply it will be found that France imports one-third of all she uses. Austria imports one- third of all her bituminous coal, but produces large quantities of brown coal and is a heavy exporter of this inferior fuel. Russia and Belgium import about fifteen per cent, of their consumption, while Sweden is almost wholly dependent upon other countries for her coal. The figures for iron ore show that Belgium imports almost all her supply and that Great Britain and France import about one- third of all that is used. On the other hand, Germany exports almost as much as she imports, ^fhile Sweden sends most of her ore abroad. Spain is also a factor in the ore question, but is not included in the table as she has no bearing on international com- merce in any other line of iron products. The statistics for pig-iron show that Belgium and Germany are the only nations that import any considerable portion of their sup- ply, while Great Britain is the only one that exports any important amount. In 1899 and 1900 the latter nation exported 15 per cent, of her pig-iron. In these two years the United States exported only two per cent, and Germany about the same, while in 1901 the United States sent abroad only one-half of one per cent, of her pig- iron. In wrought-iron and steel, Great Britain, Russia and Belgium import quite a considerable proportion of their total production, while the United States imports a very small percentage. Singu- larly enough, the nations that import the greatest proportion also export the greatest, for England exports one-third of her finished 826 THE IKON INDUSTRY. iron and steel, and Belgium nearly one-half of her output. The United States up to the present time has shipped away only a small proportion of her output, but in 1900 it reached 12 per cent, of the total. In 1901 there was quite a falling off in exports owing to the extraordinary home demand. This comparison gives some idea of the character of the busi- ness of these nations, but it does not convey any definite informa- tion about the extent to which these nations influence the com- merce of the world. Thus, although the United States sent abroad only a small proportion of her products, the actual tonnage so ex- ported in 1900 was nearly three times the over-sea shipments of Belgium, although the latter nation, as above stated, sent nearly half of her products to other countries. The overshadowing factors in over-sea commerce are Great Britain, Germany and the United States in the order named, and in this calculation the commerce of England with her own colonies is not included. Other nations play a very small part in the general international iron trade. There are some people who may look for a table giving the rate of wages in each country, and possibly it would please some of my political friends to have some figures duly tabulated to prove some tariff theories. It would be quite easy to give statistics on either side. From personal knowledge I could quote the earnings of boiler-makers in free-trade England at over $7.00 per day and the wages of skilled rolling mill men at $1.50 in protectionist Germany and Austria. It is thoroughly well known to manufacturers and practical employers of labor that the information collected by our Government at so much cost and trouble is hardly worth the trouble of printing, but statisticians are constantly quoting the records for want of better information. The weak points are recognized by the Department itself, but there are great difficulties in the way of obtaining really valuable data. Thus, for instance, it is of little use to record that the wages of bricklayers are $5.00 per day in a certain city and only $2.50 in a certain town, for it is quite prob- able that in the city the work is intermittent, made up of short jobs interrupted by weather, so that from inclement days and intervals between jobs, the annual earnings will be no more than in the town where perhaps a steel works offers perfectly steady work under shelter in rough weather throughout the whole year, and where the rent and cost of living is much less than in the greater community. It is also of little value to give -the average amount STATISTICS OF THE IRON INDUSTRY. 827 of money drawn by an employee, for it is necessary to know whether every man worked full time. The information so gathered is, how- ever, of more value than the usual statements of the number of men employed in good times and bad times. As a matter of fact, a rolling mill always employs the same number of men whether it runs six days per week or one day a month. The men are on the pay roll and never replace other men in other departments, but work when the mill works and are idle when it is idle. Their earn- ings are a measure of the industrial situation, but their number is constant. A decrease in the actual working force in a steel works generally signifies a stoppage of certain portions of the plant, as, for instance, a certain number of blast furnaces, or it indicates a cessation of new work on improvements, which in America we regard as an inherent part of the general plan of operation. It is not in the province of this book to discuss the future, since prophecies are only guesses ; but it may be well to call attention to the serious inroads now being made upon the supply of iron ore. I make no mention of the exhaustion of coal beds, because this is a hackneyed subject and a long supply is assured. The ore question is seldom considered, but it would seem to merit consideration. In 1865 the world mined about 18,000,000 tons of ore, and in 1900 about 87,000,000 tons. If this rate of increase continues during the coming years it will be found that in 1935 the consumption will be so rapid that in a period of five years, say from 1935 to 1939 inclusive, as much ore will be smelted as was used from 1880 to 1900. This is true of the United States in particular as well as of the world in general, and I believe that few American iron masters can view with equanimity such a prospect. We are to-day eating up the hoardings of untold geologic ages at a rate which will exhaust the known rich deposits during the present century. When these are gone it may be that others will be discovered, and it may be that the eastern part of the United States will depend upon the concentration of the lean beds of New York, New Jersey, Pennsylvania and Alabama, while Europe will work the mammoth beds of Luxemburg and Lothringen. It is to be expected that the Eocky Mountains will furnish new fields, while Africa and the unknown corners of the earth may be relied on to. prevent a catastrophe. 828 THE IRON INDUSTRY. *4U 2-30 220 210 200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 COAL PRODUCTION. 1 UNIT= 1 MILLION TONS. / I x 1 is lg- CO pf> 2 .X x^ ^ li x ^ ^ \ / / / ""-X OR *.$> * \ 7 x ^ 1 ^ ir / \ r / / / ^ ? / \ / -, / ( F # / \ ! X x 4 f / x -^ ^^ /s ^ 2/ ^x x 1 < ^^ ^- * / C ^ jS^ F* siuivf - ' -?*" - ~- == === ' -" _Al - STR . *. vjvj! =^- " ; . SN *L _BE ~~~~ -GIL M_ ^. ^ * ^., ' ^ . ^ = - .. - AUS1 RU; RIA (tU^ C SSI A FIG. XXXIII-A. STATISTICS OF THE IRON INDUSTRY. 829 FIG. XXXIII-BJ 830 THE IRON INDUSTRY. PIG IRON PRODUCTION ONE UNIT=1 MILLION TONS. FIG. XXXIII-C. STATISTICS OF THE IRON INDUSTRY. 831 STEEL PRODUCTION. ONE UNIT=1 MILLION TONS. 832 THE IRON INDUSTRY. TABLE XXXIII-D Pig Iron Producing Districts of the World. cS 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 20 27 28 29 30 31 32 33 34 35 3*i 37 38 39 40 41 42 43 44 4.5 40 47 48 4'J 50 District; see foregoing chapters for further information. g > Output. Tons. Estimates in paren- theses. Per. Cent of total Pittsburg parts of Pa Ohio and W Va U S A 1902 1899 1900 1899 1902 1900 1899 1900 1902 1900 1900 1902 1900 1900 1899 1901 1899 1900 1900 1899 1902 1900 1902 1902 1902 1900 1901 1900 1902 1902 1900 1902 1899 1900 1900 1900 1900 1902 1900 1899 1901 1900 1902 1902 1899 1899 1899 1897 1900 1899 1902 1900 1899 1899 1900 1900 1900 1900 (7,852,000) 3,187,000 3,110,000 2,273,000 1,730,000 1,586.000 1,576,000 1,474,000 1,472,000 1,156,000 1,019,000 860,000 818,000 791,000 743,000 695,000 657,000 637,000 597,000 597 000- 593,000 562,000 537,000 518,000 512,000 503,000 478,000 452,000 393,000 327,000 313,000 303,000 297,000 294,000 291,000 282,000 276,000 274,000 263,000 247,000 245,000 239,000 223,000 155,000 153,000 136,000 124,000 58,000 24,000 20,000 899,000 203,000 396,000 181,000 35,000 130,000 24,000 (100,000) 17.89 7.26 7.09 5.18 3.94 3.61 3.59 3.36 3.35 2.63 2.32 .96 .86 .80 .69 .58 .50 .45 .36 .36 .35 .28 .22 .18 .17 .15 .09 .03 .90 .75 .71 .69 .68 .67 .66 .64 .63 .62 .60 .56 .56 .55 .51 .35 .35 .31 .28 .13 .06 .05 2.05 .46 .90 .41 .08 .30 .06 .23 Cleveland northeast coast of Kngland . Lothringen and Luxemburg, the Minette district of Germany . . Illinois USA Alabama USA Scotland . . . Cleveland Ohio USA South Wales The Urals Russia Silesia, Germany Steelton; Dauphin and Lebanon counties, Pa., U. S. A The Siegen, Germany Eastern Central England, Lincoln, Leicester and Northampton Staffordshire, England The Saar Germany New York and New Jersey, U. S. A Central England Derby and Nottingham Virginia, U. S. A Lehigh Valley, Pa., U. S. A Johnstown, Pa., U. S. A Central Sweden (95 per cent of total for Sweden) Southeastern Pennsylvania, U. S. A. (the Schuylkill Valley, Philadelphia, Delaware and Chester counties) Hungary Tennessee USA Hanging Rock, Ohio, U. S. A Moravia and Silesia Austria Sparrow's Point, Maryland, U. S. A Northern France . Spain ' South Yorkshire (Sheffield) England Styria, Austria Wisconsin and Minnesota Poland, Russia Central France Canada Moscow, Russia . Michigan, U S A Aachen (Aix la Chapelle) Germany Southern France .... . . lisede (Peine) Germany Japan . . .' Italy Kentucky, Missouri, Washington, North Carolina, Georgia, Texas, Massachusetts, Connecticut and parts of Pennsylvania and Ohio not included above Great Britain, parts not included above Germany, parts not included above France, parts not included above Russia, parts not included above Austria, parts not included above Sweden, parts not included above. . . Other countries TOTAL , 43,890,000 100.00 STATISTICS OF THE IRON" INDUSTRY. 833 TABLE XXXIII-E Steel Producing Districts of the World. M & District ; see foregoing chapters for further information. Output tons ; esti- mates in parentheses Per Cent of total 1 2 3 4 5 6 I 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 Pittsburg; parts of Pa., Ohio, and W. Va., U.'S. A The Ruhr western Westphalia Germany 1901 1902 1901 1902 1900 1902 1899 1900 1900 1901 1900 1901 1900 1901 1900 1902 1900 1901 1900 1900 1900 1901 1901 1902 1900 1899 1899 1900 1902 1902 1900 1900 1900 1899 1899 1901 1902 1901 1900 1902 1902 1901 1900 1900 1901 1900 1902 1900 1901 1901 1900 1900 1900 1900 1900 (7.317 000) 4,329,000 1,750,000 1,406,000 1,333,000 1,037,000 982,000 963,000 960,000 (870,000) 659,000 656,000 655,000 629,000 626,000 589,000 588,000 427,000 371,000 367,000 353,000 352,000 352,000 333,000 314,000 291,000 282,000 279,000 240,000 238.000 235,000 214,000 205,000 190,000 178,000 173,000 154,000 (150,000) 150,000 143,000 130,000 107,000 93,000 88,000 69,000 61,000 59,000 58,000 26,000 472,000 30,000 69,000 127,000 19,000 16,000 22.33 13.21 5.34 4.29 4.07 3.17 3.00 2.94 2.93 2.65 2.01 2.00 2.00 .92 .91 .80 .79 .30 .13 .12 .08 .07 .07 .02 .96 .89 .86 .85 .73 .73 .72 .65 .62 .58 .54 .53 .47 .46 .46 .44 .40 .33 .28 .27 .21 .19 .18 .18 .08 1.44 .09 .21 .39 .06 .05 100.00 Illinois USA Lothringen and Luxemburg, the Minette district of Germany. . The Saar Germany South Wales Cleveland Ohio USA ohnstown Pa USA. Belgium Southeastern Pennsylvania, U. S. A. (the Schuylkill Valley, Philadelphia, Delaware and Chester counties.) Eastern France the Minette district . . Silesia, Germany South Yorkshire (Sheffield) England Steelton Pa , U S. A taff ordshire , England parrow's Point Maryland USA cranton Pa U S A Aachen (Aix la Chapelle) Germany The Urals Russia . . Central Sweden (94 per cent of total for Sweden) Isede (Peine), Germany Colorado, U. S. A Moravia and Silesia, Austria Northern Russia New England, U. S. A The Siegen Germany Alabama, U. S. A Lehigh Valley Pa USA Osnabruck Germany Italy ; Missouri, Delaware, Kentucky, Tennessee, Indiana, Michigan, Wisconsin, Minnesota, California and parts of Pennsylvania and Ohio not included above . Great Britain, parts not included above France parts not included above Austria parts not included above Sweden parts not included above * 32,764,000 834 THE IRON INDUSTRY. TABLE XXXIII-F. Production of Coal., Ore, Pig-iron and Steel in 1900. United States and Great Britain, 1 unit = 1000 gross tons ; other countries = 1000 metric tonst Index of Authorities (see Table XXXIII B). Country. Coal. Iron ore. Pig iron. Steel. Tons. Per cent, of total. Tons. Per cent, of total. Tons. Per cent, of total. Tons. Per cent- of totaL 239,567 B 225 181 2 149.551 2 3:3.270 2 14,913 8 38,064 9 23,463 2 252? 2,773 2 480 2 5,598 2 6.095 2 31.5 29.6 197 44 1 9 50 3.1 4 01 0.7 0.8 27,553* 14.028 2 18.964 2 4 986 1 * 5.880 1 * 3,462" 248 1 2.610 2 8.480 2 247 2 148* 63 2 4392 485 2 f 551 2 27 2 g 308 15.7 212 66 6.6 3.9 03 29 95 0.3 01 01 0.5 05 07 13,789 2 8.909 2 8,520 2 2,699 2 2.821 2 1.462* 1,019 2 527 2 294 2 24* 86 2 20 2 * 34.2 221 21.1 67 70 36 25 13 07 01 02 01 10,188 2 4,9016 6.646 6 1,660 2 1.468* 1,146" 655 2 301 2 1506 58 6 24 8 87.8- 18.0 24.4 61 5.4 42- 2.4 11 0.6 02 0.1 Germany and Luxemb'g fggr ...:....... TnfHa 17f 6.722 2 5,507 2 1,830* 388f l,938f 5,OC0 2 0.9 07 02 0.1 0.2 07 58 2 t 01 Natal 8. African Republic. . . Others 1200 2 13 100 2 0.3 16 2 Total 760,609 1000 89,345 100.0 40,318 100.0 27,207 100. a * 1899. g 1896. 1 1898. J 1897. TABLE XXXIII-G. Production of Coal (all kinds) by the Leading Nations. United States and Great Britain, 1 unit = 1000 gross tons ; other countries = 1000 metric tons. Index of authorities; see Table XXXIII-B. Year. United States. Great Britain. Germany and Lux- emburg France Russia Austria- Hunga'y Bel. gium. Swe- den. Italy. Spain. 1880 1881 1882 1883 63,S23* 76,865* 92.219* 102 868* 146,969* 164,184* 156 500* 163,737* f9,118* 61,540* 65,378* 7C.443* 19 362* 19.766* 20,604* 21 334* 3238* 3,440* 3673* 3.916* 14,800* 15, 80S* 15,555* 17.048* 16 867* 16,874* 17 591* 18.178* 984 115* 140* 149* 139* 136* 165* 214* 847* 1,210* 1,196* 1,071* 1884 1885 1886 1887 106,906* 99,069* 101 500* 116 652* 160,758* 159 351* 167.518* 162 120* 72,114* 73.676* 73,683* 76 233* 20 024* 19,511* 19,910* 21,288* 3,870* 4208* 4,506* 4464* 18,000* 20,435* 20,779* 21 879* 18 051* 17 438* 17,286* 18 379* 161* 170* 166* 165* 228* 190* 243* 328* 979* 946* 1,001* 1,0?8* 1888 1889 132.733* 126 098* 169,936* 176 917* 81 960* 84 789* 22,603* 24,304* 5.187* 6,216* 23,860* 25 328* 19 218* 19,870* 169* 187* 367* 390* 1,037* 1,154* 1890 1891 1892 1893 1894 1895 1898 1897.... 1898 1899. . . . 1900.... 1901.... 140.867* 150,5061 160,115* 162,815* 152.448* 172,426* 171.416* 178,769* 191 9415 225.103'' 239,5675 261,8742 181.614* 185,479* 181,787* 164,326* 188 278* 189.661* 195.361* 202.1195 202,0126 220085& 225,1812 219.047 2 89^057* 94 252*3 9^ 544*3 95.426*3 98 806*3 103 958*3 112 471*3 120.474*3 127.959*3 135,844*3 149.788*3 152,6292 26.083* 26,025* 26179* .25 651* 27,459* 28,020* 29.1905 30.7985 82,3565 328635 33,2702 32,3252 6,017* 6,233* 6.816* 7.535* 8.629* 9079* 9229* 11 207* 12.2425 13.5582 14 918*5 16.2702 27.504* 28,823* 29,038* 30,449* 31.492* 32,655* 33.6765 35.939' 87,786-' 38,7385 38.0649 41,203 20 866* 19 676* 19.583* 19.411* 20.459* 20 415* 21 2525 21 4925 22 088 5 22 07: 5 23,4632 22,2132 188* 198^ 199* 200* 214* 224* 226* 224* 236* 239* 252* 272*o 376* 289* 296* 317* 271* 250* 276* 314* 341* 389* 4802 426 1 212* 1.288* 1,461* 1,485* 1,657* 1,784* 1.878* 1,939* 2.4675 2.6005 2.7732 2,7482 STATISTICS OF THE IRON INDUSTRY. 835 TABLE XXXIII-H. Production of Iron Ore by the Leading Nations. United States and Great Britain, 1 unit = 1000 gross tons ; other nations = 1000 metric tons. Index of authorities ; See Table XXXIII-B. Year. United States. Great Britain. Germa- ny and Luxem- burg. France. Russia and Finland. Aus- tria. Hun- gary. Bel. ;ium. Swe- den. Italy. 02 OS 1 1880... 7,120i 18,0267 7,239i 2,874i l,024i 6971 4461 2531 7751 2891 3,5651 6141 1881... 8400* 17.4167 7,6001 30^21 l,017i 6191 4651 2231 8261 4211 3.5031 6571 1882... 9'154i 18 ',27 8,2*531 34671 1.0771 9031 5461 2091 8931 2421 47261 5671 1883... 9,1141 17,383' 8.7571 3298 997i 8821 5981 2161 8851 2041 4 5261 5571 1884... 7,6401 16,138' 9,006' 2.977 1 0151 9741 6511 1761 9101 2251 3,9071 4931 1885... 7.600 15 418' 91581 2,318' 1 0941 9311 6511 1871 8731 2011 3,9331 4191 1886. 10 OOO 1 141107 84861 22861 1,0891 7961 6351 1531 8721 2091 41671 4331 1887... ll.SOOf 13 098? 93511 25791 l,356i 847 1 566* 1721 9031 2311 6.7961 4381 1888... 12,060' 14,5917 10 6641 2842' 1.4011 1,009! 6341 186* 9601 1771 56101 3841 1889... 14 518* 14.5467 11 0('2i 30701 1,6401 1,1151 6661 1821 9861 1731 57111 3521 1890... 16,036* 13,7817 11,4101 3,472i 1 7961 1362* 7921 1721 9411 2211 6,5461 4751 1891... 14,591* 12,7787 10 6581 3,579i 1.9991 12311 8761 2021 9871 2161 4,8821 4051 1892... 16,297* 11,3137 11,5391 3 7071 2,0441 9931 9211 2101 1,2941 2141 5.4361 4531 1893... 11,588' 11.2037 11,4581 3,517' 2,0951 l,109i 9771 2391 1,4841 1911 54981 3941 1894. . . 11.880' 12 3677 12 S92i 3.7721 2.4881 1,2151 900' 3111 l,927i 1881 5.3971 3441 1895... 15,958' 12,6157 12,3501 3680' 2,9271 1.38.-)i 9551 3131 1,9051 183 5 5141 3181 1896... 16005 1 13,7017 14.16-'i 4.0621 32051 1,4491 l,270i 3071 2 0391 2041 6.7631 374 1 1897 .. 17,518* 13.788' 15 4661 4,582 4,1121 1,6141 14211 2411 2,0871 20H 7.4201 4411 1898... 19.434* 14,1777 15.893* 4731 1 4,8711 1,7341 1,6671 2171 2.3031 190' 7,1971 4741 1899... 24 683* 14.4617 17.990 1 4986 58801 1,9001 1,9531 2011 24361 2371 93981 6511 1900... 27.553* 34,028' 18 < 642 5.448* 5,989*2 1,8949 1,668' 24810 2.6102 2472 8,4809 602 a 1901 28,887* 12,275 a 16 570 a 4,791 2 5,663 a 219 2 2,795*0 232 a 7907 2 514 a TABLE XXXIII-I. Production of Pig-iron by the Leading. Nations. United States and Great Britain, 1 unit = 1,000 gross tons ; other countries = 1,000 metric tons. Index of authorities ; see Table XXXIII-B. Year. United States. Great Britain Germany and Lux- emburg. France. Russia and Finland. Aus- tria. Hun- gary. Bel- gium. Swe- den. Italy. Spain 1880.. 3,8352 7,7497 2,7291 1,7251 4711 3201 1441 6081 4061 17i 86i 1881.. 4,1442 8,1447 2,9l4i 1,8861 4921 3801 1641 6251 4301 281 114i 1882 . . 4,6232 8,5877 3,3811 2,039i 6011 4351 r.6i 7271 3991 251 1201 1883.. 4.5962 8,5297 3,470i 2.0691 5001 5221 1761 7831 4231 241 1401 1884.. 4,0982 78127 3.6011 1,8721 5321 5401 1951 7511 4311 181 1211 1885.. 4,0452 7.4157 3,6871 1,6311 5521 4991 2:61 7131 4651 161 1591 1886.. 5,6832 7,0107 3,5291 1.6171 5491 4851 2351 7021 4421 121 1181 1887.. 64172 7.5607 4,0241 1,5681 6331 5121 1931 7561 4571 121 1651 1888.. 6,4902 7,9997 4,3371 168 Ji 6871 5861 2041 8271 4571 131 1651 1889.. 7,6042 8 3237 4,5251 1,7341 7551 6171 2391 8321 4211 131 1981 189J.. 9,2C3-' 7.9047 4.6581 1,9621 9501 6661 2991 7881 4561 141 17H 1891 . . 8,2802 7.4(,67 4.641*3 1.897* t,028i 6171 3051 6841 4911 121 1491 1892... 9,1572 6,7097 4,937*3 2,0571 ,0381 6311 3101 753* 4861 131 1341 1893... 71252 69777 4,953*3 2,0031 ,1811 6631 3191 7451 4531 81 1351 1894... 66572 7,4277 5,5?9*3 2.0701 ,3331 6901 r 33 'l 8191 4631 101 l*a 1895... 9,446-' 7,7037 5,789*3 2,0041 ,4541 759'- e 3491 8291 4631 91 18t,i 1896... 8.6232 8,6607 6361 2,340i ,8671 817*6 4011 9591 4941 lU 24 -f 0.000027 1 H.O = 342 + 0.00015 1 CH 4 = 0.418 -4- 0.00024 1 C 5 H 4 = 0.424 -f 0.00052 1 Mariotte's Law. The volume of a gas is directly proportional to the absolute temperature and inversely proportional to the pressure upon it. Note : Absolute zero = 273-5* C. Law of Dulong and Petit. The product of the atomic weight of an elementary substance by its specific heat is always a constant quantity. INDEX. PAGE Aachen iron industry 755 Absolute zero 839 Acid open hearth process 12, 269 Acid steel, low phosphorus at Steelton 211 Acid vs. Basic Steel .. 14, 23, 25, 29, 534 Air, composition of 234 Air, properties of 838 Air needed in combustion 235 Akerman, on definition of steel 143 on Swedish Bessemer work 161 et seq. Alabama, iron industry 49, 668 open hearth 307 Algeria, statistics 834 Allegheny County 657 Allotropic forms, microscopic 403 et seq. Allotropic theory 418 Alpha iron 418 Alumina, in blast furnace slag 82 Aluminum, effect on physical properties 475 in castings .595 Alzola, on Spanish ores 812 American practice 660 American Society for Testing Materials 25 American Steel Manufacturers' Association 24 Angles, physical properties 371 et seq. Annealing 381 et seq., 415 et seq., 429 Anthracite, combustion of 234, 235, 236 for recarburization 301 gas in gas engines 245 in blast furnace 51, 639 in producers 244 in Russia 777 mining districts in United States 640 Appleby, on tests of rounds 433 Arnold, on sub-carbide theory 419 Arsenic, effect on physical properties 478 Ash, from producer 241 in coal 241, 266 Atomic weights 839 (841) 842 INDEX. PAGE Austenite -. 403 et seq. Australia, statistics 834 Austria-Hungary, iron industry 785 imports of coke 788 statistics 834 et seq. Axles, specifications 569 Ayrshire, see Scotland Bahnis-Roozeboom, on phase doctrine 418 Ball, on effect of copper 474 Barba, on tests of steel : 434 et seq. Barrow-in-Furness 717 Basic vs. acid steel 14, 23, 25, 29, 534 Basic linings, functions of 282 Basic open hearth process 15, 282 Bauxite for basic hearths 282 Bessemer, acid process 7, 155 American practice 83 basic process 8, 176 at Steelton 180 at Troy 689 basic steel, quality 14 calorific history, acid 164, 171 basic 183 for steel castings 27 gases from 164 increments in cost 333 in Sweden 168 iron burned, acid 166 basic 183 lime used, basic 176 pig iron 5 slag, acid 163 basic 179 steel , 6 vs. open hearth 14 Bavaria, iron industry 759 Belgium, coal field 765 iron industry 796 labor question 800 statistics 834 et seq. Bell, on blast furnace reactions. .53, 55, 64, 69, 70, 76, 77, 78, 80, 87, 88, 97, 98 on heat of gasification 223 Bertrand, blast furnace top 63 on Austria 785 on reduction of ore 319 Bertrand-Thiel process 315 et seq., 325, 331, 789 Beta iron 418 Bethlehem works. . . .688 INDEX. 843 PAGE Bilbao ore 49 Birmingham; see Alabama Bituminous coal 643 in gas producer 237 Black band 48, 710, 722 Blast, for blast furnaces 85 et seq. heating of 78 Blast furnace 3, 4, 5, 48 et seq. boilers 79, 103, 110 chemical reactions 55, 64 gas combustion .* 107, 108, 109, 110 gas in gas engines Ill height of 57 stoves, air needed 122 use of anthracite in South Russia 777 charcoal in Urals 780 raw coal 711 Blauvelt, on coke ovens 257, 259, 262 Blister steel 147 Blow holes 594 Bohemia, iron industry 48, 788 Boilers, blast furnace 79, 103, 110 gas needed 101 over heating furnaces 252 Boiler plate specifications 559 Bounties .' 627 Canadian 820 Bridge steel, specifications 550 Brown hematite 49 Braune, on Sweden 803 Bumby, on pig iron in Scotland 711 Buildings, steel for, specifications 555 Burned lime in basic Bessemer 176 Burning of steel 198 By-products 257 By-product coke ovens in United States 663 Calorie, definition 839 Calorific equation of acid converter 164 basic converter 183 open hearth furnace 218 et seq. Campbell, tilting furnace 205 et seq., 307 Campbell, J. W., on heat treatment 381 Canada, iron industry 818 statistics 834 Cape Breton, iron industry 819 open hearth furnaces 307 Carbo-Allotropic theory 418 Carbon, calorific value in converter 167 844 INDEX. PAGE Carbon, calorific value in open hearth 322 combustion of. 233 determination of . 40 effect on pig iron 126 on steel 22, 46, 456, 487 et seq., 527 on wrought iron 139 for basic hearths 282 in pig iron 4 in producer ash 241 in puddle furnace 131 in tool steel. /. . . , 151 in wrought iron 138 protective power of 271 segregation of 340 et seq. theory (metallography) 417 use as recarburizer 154 Carbon deposition 68 Carbonic acid and iron 66 et seq. Carbonic acid in blast furnaces 54, 73 et seq. in producer gas 243 Carbonic oxide, combustion of 233 in products of combustion 105 Carbon Steel Company, open hearth practice 208 Carnegie Steel Company, slabbing mill 367 Cast iron ; see pig iron. Cast steel 151 Castings, specifications 577 Cement carbon 413 Cementation 147 Cementite, in cast iron 125 in steel 403 et seq. Cement steel 147 Central I. & S. Co., plates 349, 423 wrought iron 135 Chambers for open hearth furnace 190 et seq. Charcoal in blast furnaces 51, 780 Charcoal blooms in United States 302 Charge in open hearth furnace 269 Charging open hearth furnace 211 Checkers in regenerators 190 Chicago, iron industry 665 Chromite for basic hearths 282 Chromium, effect on physical properties 480 Clay band 48 Clay iron stone 48 Cleveland (England) coke ovens 262 iron industry 48, 76 et seq., 684 et seq. labor conditions. . . . . 604 INDEX. 845 PAGE Cleveland (England) ore deposits. 701 Cleveland (U. S.) iron industry 684 Coal fields; see Table of Contents. Coal production : see Table of Contents. Coal, imports and exports; see Table of Contents. Coal, international trade 825 Coal washing . .... 263 et seq. Cobalt, effect on welding 139, 584 Covhran, on use of lime in blast furnace 56 Cockerill Co., gas engines 114 Coke, districts of United States 646 exports from N. E. coast (England) 705 imports and exports ; see Table of Contents. in blast furnace 52 production ; see Table of Contents. Coke ovens 256 by-product in United States 663 by-product, use abroad 606 Combustion, general view 233 et seq. of blast furnace gas. . . 104, 107 Colby, on influence of copper .475 on specifications 547, 548 Colorado, iron industry 687 Colored labor in Alabama 674 Connellsville, coke 52, 76 et seq. coke ovens . . 260 district 657 Continuous furnaces 255 Converter ladle . . . .210 Cooper, on Northeast Coast 700 Copper, effect on welding 139, 584 in Cornwall ore 472 influence on physical properties 22, 472, 475 Cornwall ore deposit 83, 676, 690 copper in 472 Cost of manufacture...... ... 604, 626 Crucible steel. . , 7, 147 Crystallization by heat .584 Critical point 394 et seq. Cuba, ore '. 49, 472, 636, 682 statistics 636, 834 Cuban ore, copper in 472 smelting of 68, 69 Cumberland, iron industry 717 Cunningham, on segregation 181, 347 Cupola castings 591 Cupolas, practice 166, 170 Custcr, on tests of steel 435, 451 846 INDEX. PAGE Cyanogen in blast furnace 72, 75 Dead melt 188 Depreciation 625 Derbyshire, iron industry 725, 726 Diameter, influence on physical qualities 431, 433 Direct metal at Steelton 208 in open hearth 308, 310, 313 in Sweden 168 ore needed 324 Dissociation 187 Distances in America 630 in Russia 630 Dolomite in basic Bessemer 176 in basic open hearth 282 in blast furnace 671 for basic hearths 9, 282 Don, basin of 49, 776 Donawitz, iron industry 792 open hearth furnace 193 Dougherty, J. W. } on blast furnace 64, 70 Dowlais Iron Company, plan of works 715 Drillings, method of taking 533 Drop of beam 451 Duplex process 337 Duquesne, open hearth furnace 193 Durham coal and coke. , 52, 76, 77, 78, 80, 704 Dutreux, iron industry of France 762 Edison, on ore concentration 50 Ehrenwerth, on Bessemer practice (acid) 172 on open hearth work 278 Elastic limit 540 Elastic ratio, errors in measuring 450 higher in soft steel 535 Elba ore 48, 49, 816 Elbers, on blast furnace slag 82 Electric concentration 690 Electric welding 588 Elongation 20 errors in measuring 450 influence of diameter 431, 433 of length 435 et seq. of width 434 et seq. England; see Great Britain. Ensley, Ala., coke ovens 258 Erie Canal 686 Errors in metallurgical records 39 et seq. Erzberg; see Styria 791 Essen, machine shops 746 INDEX. 847 PAGE European methods 255, 256 Eutectic alloy 405 Exports from Sweden , 808 of leading nations 837 of ore from Germany 729 Excess air 104 Eye bars, annealing 389 physical properties 421 tests on 440 et seq. Fawcett, on ore transportation 651 Felton, on rest after rolling .448 Ferrite in cast iron 125 in steel 403 et seq. Ferro-manganese 8, 12, 463, 464 Ferro-silicon, composition of 127 Finishing temperature, effect of 409 et seq. Firmstone, on dolomite 53 Flats vs. rounds 422, 427, 430 Fluidity of basic slag 290 Flux in blast furnace 52, 73, 74 use of dolomite 671 Foreign practice 610 Forest of Dean 714 Forgings, physical properties 370, 421 specifications 574 Formulae for tensile strength 21, 23 Forter valve 216 France, iron industry 762 et seq. statistics 764, 834 et seq. Frazer-Talbot producer 239 Freights 627 Fuel 233 et seq. economy of 661 ratio, blast furnace 86 Gain from ore in open hearth 313 Gamma iron 418 Gas, blast furnace 102, 111 et seq. engines Ill for open hearth furnaces 187 from basic converter 179 from tunnel head 107, 108, 109, 110 producer 240 use abroad 606 Gas scrubbers, at Scotch blast furnaces .711 Gayley, on blast furnaces 77 German nomenclature 6 Germany, Bessemer practice (acid) 160 coke exports to France 768 848- INDEX. PAGE Germany, iron industry* ....*...*...... 727 railroads . K ....... \ . 736 rolling mill practice. 610 statistics 834 et seq. errors in. 821 Gjers pits .-.-..-. 249, 250 Gogebic ; see Lake Superior. Graphite in pig iron 4 Great Britain, coal fields 694 engineering practice 603 exports of fuel 695, 768, 784 imports of ore . . 696 iron industry 692 production by districts 698 of rails 634 of steel 634, 692 statistics 692, 834 et seq. Greece, statistics ..-.- 834 Grooved tests vs. parallel sided 424 Guide rounds vs. "hand rounds 375 Hadfield, on effect of aluminum. 476 on effect of silicon 457 on manganese steel 467 on steel castings. 594 Hand rounds vs. guide rounds 375 Harbord, on effect of arsenic < 478 Hard coal ; see anthracite. Hardening carbon 413 Hardening of steel, definition 142 Hard structural steel 546 Hartshorne, on basic Bessemer 185 on Bertrand-Thiel process. . .. 331 Heating furnaces 249 Heat lost in open hearth furnace 218 et seq. Heat treatment 381 et seq. Hematite 48, 669, 682, 718 Henning, on elastic limit 451 on methods of annealing 389, 390 Hibbard, on oxide of iron 482 High carbon steel 147, 634 homogeneity of 357 Hofman, Prof. H. 0., on coking. . .' 226 Holley, on wrought iron 136 Homogeneity of open hearth steel 152, 281, 347 Horde, basic Bessemer practice 183 Horse-power of blast furnace gas 101, 1 12 Hot working 18 Howe, on Bessemer practice (acid) 158, 159 INDEX. 849 PAGE Howe, on carbon deposition 69 on critical point 394 on definition of steel 142, 143 on effect of carbon 46 on effect of phosphorus 469 on effect of silicon . 456 on invisibility ... 393 on micrometallurgy 406 on phosphorus in acid slag 159 on structure of pig iron 125 on temperature of melting 596 Humidity '. . . 93 Hungary, iron industry 48, 794 statistics 834 et seq. Hunt, A. E., on influence of method of manufacture on physical properties, 529 on preliminary tests 426 on quench test 540 on wrought iron 136 Huston, on plate tests 424 Hydrogen, combustion of. ' 233 in blast furnace gas 97, 99, 100 in producer gas 227, 242, 248 Illinois Steel Company; Bessemer practice : 157 Bessemer slag (slag) 163 slabbing mill 367 Ilsede, iron industry 755 Imports of leading nations 837 of ore into Germany . . 729 Increments in cost, duplex process 338 open hearth process 335 rolling mills 336 India, statistics 834 Indicator cards, gas and steam engines 119 Influence of elements upon steel. 21, 455 et seq. Ingot iron 141 Ingot steel 141 Inspection 28 et seq. Iron, calorific value in converter 167 Iron oxide; see Iron ore. in basic slag 289, 291, 292 in open hearth 311 Italy, iron industry 816 statistics 834 et seq. Japan, statistics 834 Joeuf district 766 Johnstown, iron industry. ... . . 675 Joliet, steel works. ... .'.'.."/. .'.Y.Y.Y. '.'.'.. 664 Jones and Laughlinsy blast furnace: .;.... 60 850 INDEX. PAGE Jones mixer 169 Julian, on Bessemer practice 158 von Juptner, on open hearth practice 218 et seq. on producer work 241, 242 Jurugua, mine in Cuba 682 Kennedy, Julian, on blast furnace 64 on Russia 772 Kertsch, ore beds 778 Killing, crucible steel 149 Kirchhoff, on Cleveland district 702 et seq. on Westphalia 727, 735 Kladno 789 blast furnace 63 open hearth 315 et seq. Knapp, on Lake Superior deposits 647 Koerting gas engine 123 Krivoi Hog, ore beds 777 ore taken to Poland 783 Krupp's works 747 Labor in Alabama. 674 in Belgium 800 in Cleveland (England) 604 Labor organizations 604, 614 Lahn, iron industry 760 Lake Champlain, ore deposits 689 Lake Erie, iron industry 686 Lake Superior ore 49, 58, 83, 647 statistics. . . 652 Lake transportation of ore 648 Lanarkshire; see Scotland. Lancashire hearth 805 Lancashire, iron industry 717 Langley, on carbon determination 39 Large ingots, homogeneity of 346 et seq. Latent heat of fusion of steel 173 Laudig, on carbon deposition 68, 69 Least squares, use of method .23, 487 Lebanon, blast furnaces 678 LeChatelier pyrometer 391, 416 Ledebur, on furnace slag 53 Leicester, iron industry 724, 725 Length, influence on physical properties 435 et seq., 445, 539 Letombe, gas engines 120 Lignite, in Germany 761 Lime in basic Bessemer 176, 178, 179 in basic open hearth 286, 287 in blast furnaces 55, 83 Limestone in basic open hearth 283 INDEX. 851 PAGE Limestone in blast furnace 52, 73, 74, 98 Limonite 49, 669 Lincolnshire, iron industry 724, 725 Linings, absorption of iron oxide 166 Liquation of sulphide of manganese 294 Liquid interior of ingot 360 Longitudinal vs. transverse tests 422 Longwy district 766 Lorraine; see Lothringen. Loss from carbon in ash 241 of heat, combustion of blast furnace gas 107 of heat in blast furnace 76 et seq. Lothringen, iron industry 730 Lukens Iron and Steel Co.; plate tests 424 Luminosity of flame 240 Lunge, on water gas 247 Lnrmann, on blast furnace gas 102 Luxemburg iron industry 730 Magnesia in basic open hearth 288 Magnesite for basic hearths 282 Magnetic concentration 50 Magnetic properties, effect of heat 416 Magnetite 50 in Cuba 682 in United States 690 Mahoning Valley. 657 Manganese, allowable content 463 determination of .40, 44 effect on steel 22, 463, 483, 487 et seq., 513, 522, 527 effect on welding 584, 589 in acid Bessemer ' 162, 174 in acid open hearth 271 in basic Bessemer 181 in basic open hearth 298 in blast furnace 80 in castings 595 in crucible steel 148 in pig iron 80 in puddle furnace 131 in wrought iron 131 loss in recarburization 280 protective power 271 reduction from slag 294 segregation 340 et seq. use in removing sulphur 340 Manganese ore in open hearth 294 Manganese steel 467 Markets of the world . . . . 609 852 INDEX. PAGE Marquette; see Lake Superior.. Martensite , 403 et seq. Martin, on micro-metallography .410 Maryland Steel Company; see Sparrows Point. Bessemer plant 155 coal consumption 250 coke ovens 259 gas engines 114 miscroscopic work 410 rail manufacture 412 Mason, on German statistics 821 Medium steel 545 Menominee ; see Lake Superior. Merchant iron 130 Mesabi; see Lake Superior. carbon deposition 68 Mctcalf, on definition of steel 143 Method of least squares 23 Method of manufacture, influence on physical properties. 14, 23,25, 29, 529, 534 Metric system 839 Meurthe et Moselle 764 Microscope, use on steel fractures .403 ct .vn/. Middlesborough; see Cleveland 700 Mill cinder 135 Milwaukee, steel works 664 Minette district 49, 731, 764, 797 furnaces 103 Mixer; see Receiver. Monell, on open hearth practice 205, 330 on Russia 772 Moravia, iron industry 790 Muck bar 6, 130 Natal, statistics , 834 Natural gas 245, 660 Necking of test piece 439 Neutral joint. . . : 283 Newfoundland, ore 49, 818 New England, iron industry 689 New Jersey, iron industry 50, 690 New South Wales, statistics 834 New York, iron industry , 629 ore 50 Nickel, effect on physical properties 479 effect on steel 23 effect on welding 139, 584 Nickel steel, homogeneity of 359 Wimot de, on Belgium 796 Nord, coal and iron industry 767 INDEX. 853 PAGE Northeast Coast of England 700 et seq. Northeastern Steel Company 700, 708 Northamptonshire iron industry 724, 725 North Wales iron industry 723 "Norway iron," so called 811 Nottingham iron industry 725, 726 Oberschlesien ; see Silesia. Odelstjcrna, on effect of aluminium 478 on effect of silicon 277 Oechelhauser, gas engine 121 Oil as fuel 271 Oolite , , , 49 Open hearth furnace. 11, 186 et seq. with natural gas 660 process, acid 12, 269 basic 282 metal, for rails 530 for tool steel 151 manufacture in United States 302, 306, 635 Ore j see Statistics. cost of transportation 651 imported into United States 636 in acid open hearth furnace 272, 274 in basic open hearth 284 in Bessemer converter. 324 in open hearth. , 13, 272, 274, 284, 311 international trade 825 reduction, absorption of heat 319 supply of America. 649 supply of the world 827 Osnabruck, iron industry 759 Oswald, on Rombach works 267, 740 Otto cycle, gas engines 118 Otto Hoffman, coke oven 259, 261, 262, 26$ Overheating; see heat treatment. Oxidation in open hearth 324 et seq. Oxide of iron, effect on physical properties 480, 523 Oxides of iron ; reactions 65, 66, 67 Oxychloride of lime 295, 296 Oxygen in products of combustion 104, 105, 106 in steel .480, 523 Park, on definition of steel 143 Pas de Calais, coal and iron industry 767 Pearlite in cast iron 125 Pearlite in steel 403 et seq. Peine, iron industry 755 Pencoyd Iron Works 310 Pennsylvania; see Table of Contents. 85 1 INDEX. PAGE Pennsylvania Steel Works; see Steelton; see also all tables and tests where other sources of information are not mentioned; basic Bessemer 180 blast furnace construction 59 gas engines. 114 low phosphorus acid steel 304 open hearth furnace 193 slabbing mill 367 Petroleum 246 Phase doctrine 418 Phases, miscroscopic 403 et seq. Phillips, on Alabama 668 on blast furnace slag 82 on dolomite 53 Phosphorus, allowable content 469 et seq. calorific value 10 combustion of 8 determination of 40 effect on steel 22, 469, 483, 487 et seq., 516, 523, 527, 531 effect on welding 139, 584 in acid open hearth 278 in basic converter 9, 177 in basic open hearth 15, 283, 286 in Bertrand-Thiel process 316 in blast furnace 4 in castings 595 in crucible steel 748 in pig iron 4 in puddle furnace 132 in tool steel 150 in wrought-iron 132 segregation of 340 et seq. volatilization of : ... 181 Physical properties; see chapters XIV and XVI. Pig and ore process at Steelton 208, 275, 278, 306 et seq. Pig iron; see Statistics. composition and physical qualities 124 international trade. 825 manufacture 3, 4, 5 production in leading nations 830, 832, 834, 835 production, per capita 824 Pinfjel, on statistics of France 762 Pipes in castings 594 Pittsburg, blast furnaces 76, 81 iron industry ; 657 Plates, allowances for over-weight 554 from ingots 18 from slabs . .18 INDEX. 855 Plates, physicial properties 366, 377 tests on. 537 Poland, iron industry 782 Pomerania, iron industry 760 Ports, open hearth furnace 213 Possession works in Urals 781 Pottstown Iron Company 44, 434 Pourcel, en segregation '.39, 343, 344, 345, 347 Preliminary tests 426 groups of 486 Producer, ash from 241 operation of 222, 227, 237 et seq. temperature of gas 217, 218 Products of combustion 104, 105, 234 et seq. from blast furnace gas 107, 108, 109, 110 Production; see Table of Contents. Production of steel in Great Britain 634 in United States 631, 634 Protective power of elements 271 Puddle cinder 135 Puddling furnace 5, 129 et seq. Pueblo, steel works 687 Pulling speed, effect on physical properties 452 Pure iron, definition 522, 526 Pyrometer, LeChatelier 391, 392 Quenching, effect of manganese 464 effect on steel 144 Quench test 540 Radiation, loss from in open hearth 224 Railroads in Germany 736 Rails, method of rolling 411, 412 of open hearth steel 530 sections 607 specifications 564 Railways, miles of 609 Raw coal in blast furnace 51 Recarburizer, function of 8, 463, 464 in acid Bessemer 174 in basic converter 184 in acid open hearth 279, 280 in basic open hearth 299 Receiver 169, 170, 740 Red hematite 49 Reduction of area, errors in measuring 450 Reduction of ore, heat absorption 319 et seq. in open hearth furnace 274, 275, 276, 313, 329 Regenerative furnaces 11, 186, 190, 250 Removal of slag from open hearth 297, 307 ' 856 INDEX. PAGE: Rephosphorization, in basic Bessemer 184 in basic open hearth 299, 300 Rest after rolling 448 Reverberatory furnaces 251 Reversing valves, open hearth furnace 214 Richards, Prof. J. W., on Bessemer work (acid) 165, 166 et seq. on open hearth practice 223, 226 et seq. on specific heat 90 on zone of fusion 92, 95 Richards, Prof. R. H., on blast furnace phenomena 70 Riley, on effect of nickel 480 on effect of work on steel 365, 367 on gas engines 113 on pig iron in Scotland 711 Rivet steel 541, 542, 543 Roberts, Austin, on micrometallurgy 406 Rombach, steel works 267 Rounds, influence of diameter 431 Rounds vs. flats, physical properties 422, 427 Royal Prussian Institute, welding tests 587 Ruhr district, iron industry 742, Russia, distances 630, 775 imports 77 iron industry 772 et seq. labor problem 774 ore ; 49 production of fuel, ore, iron and steel 776- statistics 834 et seq. works owned by foreigners 773: Saar district, coke 52 iron industry 753 Sandberg, on influence of silicon 462 Saniter, on use of oxychloride of lime 295, 297 Sauveur, on micrometallurgy 406 Saxony, iron industry 756 Schonwalder, open hearth furnace 196 Schrodter, on German statistics '. 821 on Germany 727 on steel output 748 Scotland, black band 710- coal 711 coal in blast furnace 51 iron industry 710 Scranton, iron industry 688 Scrap in open hearth 306 Seebohm, on crucible steel 148 Segregation 17, 19, 39, 152, 340 et seq. in ingots cast in iron molds _, 345 INDEX. 857 PAGE Segregation in steel cast in sand ; . . . . 344 in Swedish steel 152 Semet Solvay coke ovens 226, 257, 260, 263 Sensible heat in producer gas 242 Shape of test piece 19. 20, 25 'Sharon, open hearth furnace 193 Sheffield; see South Yorkshire. Shenango Valley 658 Shipbuilding in England 697 Ship steel specifications 550 Shock, influence on physical properties 465, 466 Shoulders on test piece 424 Siegen, iron industry 757 Silesia, coal deposits 790 coke 52 . furnaces 103 iron industry, Austrian 789 German 750 Silica in basic slag 291 in open hearth furnace 327 Silicon, calorific value in converter 8, 167 in open hearth 321 change of affinity with temperature 160 determination of 40 effect on steel 22, 456, 487 et seq., 514, 527 effect on welding 139, 584 in acid converter 160, 171 in acid open hearth 271 in basic converter 177 in basic open hearth 298 in blast furnace 84 in castings 595 in crucible steel 149 in foreign steel 461 in pig iron ; , 480 in puddling furnace 130 in recarburization 184 in wrought iron 130 protective power of 271 reduced in blast furnace 84 reduced in crucible 149 ISilico spiegel, composition of 127 Sink heads . . . . 27 Sjdfjren, on Austri 785 on open hearth furnace 205 Slabbing Mill; Carnegie Steel Co 367 Illinois Steel Co 367 Pennsylvania Steel Co 367 $58 INDEX. PAGE: Slag, acid, phosphorus in 159 acid Bessemer 163 acid open hearth furnace 273 automatic regulation 17 basic Bessemer 179 basic open hearth 287 et seq. blast furnace. 82, 84, 85 cupola ; 170 effect on welding 584 open hearth 12 et seq. removal in open hearth 307" Snelus, on influence of silicon 462: on use of oxychloride of lime 296. Soaking pits 249, 606 Soft coal; see Bituminous coal. Soft steel 544 Soot in producer gas 240 Sorbite '.403 et seq. South African Republic, statistics 834 South Russia, iron industry 776 South Wales, iron industry , 714 South Yorkshire, iron industry 721 Spain, iron industry 48, 812: statistics 636, 834 et seq. Spanish American, mines in Cuba 682" Spanish ore, composition 707 cost of in England 707 in Germany 746- in South Wales 714 Sparrow's Point, export of rails 684 iron industry. 679 et seq* Spathic ore 48 Specific heat of gases '. 90, 839 of steel 173' Specifications on steel 24, 532 et seq. Specular ore ' 48 Speed of test, influence 440 Spiegel, composition of 127 use of .8, 463, 464 Splice bars, specifications 567 Stable basic slags 29S Stafford, on Chicago industries 664 on open hearth ports 214 Staffordshire, iron industry 722" Standard specifications 548 et seq. Standard test pieces 538 Statistics 821 et seq. errors in. . . .692 INDEX. 859 >AGB Stead, on composition of pig iron. ..., 127 on effect of arsenic V8 on micrometallography 413 on use of oxychloride of lime 296 Steam in producer gas 188 Steel; see Statistics. definition 6, 140, 146 production in leading nations 831, 833, 834, 836 Steel castings 26 Steelton, blast furnace gas 100 iron industry 675 weather records 93 Stoves, blast furnace 3, 78, 79, 103, 110, 122 air needed 122 gas needed 101 Structural work, use of soft steel 535 Structure of steel, theories 417 Styffe, on tensile tests 37 Styria, coke from Germany 788 iron industry 48, 788, 791 Sub-carbide theory 419 Sulphur, determination of 40 effect on steel 22, 467, 483, 487, 515, 523, 527 elimination in open hearth 277 in acid open hearth 278 in basic Bessemer 180 in basic open hearth 283, 294 et seq. in blast furnace 66, 50, 677 in Cornwall ore 677 in crucible steel 148 in pig iron 4 in producer gas 188 in puddle furnace '. 132 in steel castings 595 in Talbot furnace 314 in tool steel 151 in wrought iron 132 segregation of 340 et seq. volatilization in basic Bessemer 181 in basic open hearth 294 Sweden, Bessemer practice 160 et seq., 168 coal 808 crucible steel 148 iron industry 803 ore. 50, 808 statistics 803, 8i Jt seq. Swedish ingots, segregation 382 Swedish pig iron in United States 302 860 INDEX. PAGE Tafna, ore 49 Talbot, on gain from ore 332 Talbot process 310, 326 et seq., 330 Tar in producer gas 240 Tariff 623 rates on coal, ore, iron and steel 838 Temperature, determination of 392 effect on combustion of silicon 160 of Bessemer converter 168 of burning carbon 91 of melted steel 596 of puddle furnace 133 of open hearth furnace 217, 593 Test ingots, physical properties 372 Test pieces, from castings 596 method of taking 420 Tests, chemical and physical 548 et seq. Thackray, on phosphorus 45 Thickness, effect on physical properties 364, 538 Thwaite, on gas engines 113 Tilting furnace 205, 307 Tires, specifications 572 Titanium in acid open hearth furnace 271 Traces, persistence of 160 Transferred steel 302 et seq. Troostite 403 et seq. Tropenas process 26 Tucker, on effect of arsenic ; 478 Tunnel head gases 3, 87, 96 et seq. in gas engines Ill Tungsten, effect on physical properties 480 Turner, on influence of silicon 460 Tuyeres, reactions at 74 Two inch tests 425 Union Bridge Co., eye bars 440 U. S. Govt. specifications on plates 380 United States, iron industry 629 production of rails 633, 634 production of steel 631, 634 statistics 832 et seq. Unstable basic slags 293 Urals, iron industry 49, 779 possession works 781 Valves, open hearth furnace 214 Vauclain, on boiler plate '. 44 Vermilion; see Lake Superior. Virginia, iron industry 629 Virginia ore, copper in 472 INDEX. PAGE Volume tunnel head gases 99 Wages, in different nations 826 in England. 604 Walilberg, on segregation 40, 152, 362 Wailes, on failures of steel . 529 Wales, North; see North Wales. Wales, South; see South Wales. Washed metal 302 Washing of coal 263 et seq. Waste gases; blast furnace 3, 87, 96 et seq. from heating furnaces 253, 254 heat lost in open hearth 224 in gas engines Ill Water gas 247 Water vapor in air 93 in blast furnace 94 et seq. Webster, on boiler plate 44 on effect of sulphur 469 on elongation 440 on influence of metalloids 482 et seq. on physical properties 45, 379 Wedding, on basic Bessemer 181 on German statistics 821 on Germany 727 on recarburization 300, 301 Weld iron, definition 141 steel, definition 141 Wendel works, in Germany 741 Westphalia; see Ruhr district. coke industry 52, 258, 735 iron industry 735 West Coast of England 49, 717 Welding 26, 138, 583 Weld iron 141 Weld steel 141 Wellman, charging machine 211 furnace. 205, 308, 310 West Virginia 657 While, on West Coast. 717 White, on influence of method of manufacture on physical properties 530 Whiting, on blast furnaces 99 Whitwell, on Spanish ore 707 Width, influence on physical qualities. 434 et seq., 441 Wingham, on copper in steel 474 Woman labor in Belgium 800 Woodbridge, on Lake Superior ore deposits. 650 Work, effect on steel 360, 409, 410 862 INDEX. PAG IS Wrought iron, definition 146 manufacture 5, 129 physical qualities 135 et seq. production in leading nations 836 specifications 579 welding of 138, 139, 583 Yield point; see Elastic limit. Yorkshire, South, iron industry 721 Zone of fusion. 3, 91 UNIVERSITY OF CALIFORNIA LIBRARY BERKELEY Return to desk from which borrowed. This book is DUE on the last date stamped below. ENGINEERING LIBRARY I 26 1948 W\ 4 1953 is LD 21-100m-9,'47(A5702sl6)476 T/V705 an. /? 04 THE UNIVERSITY OF CALIFORNIA UBRARY