: *< EESE LIBRARY >!') r UNIVERSITY OF CALIFORNIA. ^ , ! nLJ Shell No. i, No ^* ^-^ 333- 3 METALLURG THE ART OF EXTRACTING METALS FROM THEIR OR] \\D ADAPTING THEM TO VARIOUS PURPOSES OF MANUFACTURE. BY JOHN PERCY, M.D., F.R.S., * t LECTCJRER ON METALLURGY AT THE GOVERNMENT SCHOOL OF MINI: UNIVERSITY FUEL; FIRE-CLAYS; COPPER; ZINC; BRASH, ETC. WITH ILLUSTRATIONS CHIEFLY FROM ORIGINAL DRAWINGS CAREFULLY LAID TOWN TO SCALE. LONDON: JOHN MURRAY, ALBEMARLE STREET 1861. The right oj Translation is reserved. LONDON: PRINTED BY W. CLOWES AND SONS, STAMFOKD STREET, AND CHARING CROSS. TO MICHAEL FARADAY, WITH THE SINCERE RESPECT AND AFFECTIONATE REGARD OF THE AUTHOR a 2 PREFACE. IN no country are the operations of Metallurgy conducted on so vast a scale as in Great Britain ; and yet the contributions which have been made to the literature of the subject by British metal- lurgists are few and scanty in the extreme. But this should not lead to the erroneous conclusion which some persons are inclined to draw, that our smelters are too ignorant of chemistry to under- stand the theory of the processes under their direction, or too illiterate to be able to record the results of their experience. I have the pleasure of knowing many of these men intimately, and I will venture to affirm that with respect to knowledge, both of the theory and practice of the special departments of the Art in which they are engaged, they are not excelled by any metallurgists in Europe. The chief writers on Metallurgy are the Germans, to whom we owe two of the most remarkable works on the subject, namely, the treatise of Agricola, in Latin, which appeared in 1555 ; and the System of Metallurgy of Karsten, in German, published in 1831. The monographs, contributions to periodicals, and compendious treatises relating to the science and practice of Metallurgy which have been published in the German language, are very numerous. \\V are, probably, indebted to the Germans, to a greater extent than is commonly supposed, for the development of our mineral resources, since the introduction of German miners and metallur- gists into England, about three centuries ago, through the wisdom of Elizabeth. The Swedes, who, in the persons of Scheele and Berzelius, have played so distinguished a part in raising chemistry to the dignity of a science, have not been behind with respect to Metallurgy. Many valuable monographs and original papers on metallurgical subjects exist in the Swedish language, which unfortunately is but little ^own to Englishmen. The 'Jern- Kontorets Annaler,' or Annals of the Board of Iron-Masters, now consist of about forty volumes, which contain theoretical and ac- tical papers of the highest interest to Miners and Me. vi PREFACE. The French have published extensively on metallurgical sub- jects ; and the * Annales des Mines ' form a repertory of metallurgi- cal knowledge of great value. The chief contributors to this work have been the graduates and students of the Ecole des Mines ; and their contributions, to a large extent, consist of the descrip- tions of processes which they have witnessed out of France. It is not, a little surprising, that, considering the skill in arrange- ment and the precision of language which usually distinguish French scientific writers, no complete treatise of Metallurgy should yet have appeared in France. In the exercise of my duties as Lecturer on Metallurgy at the Government School of Mines, during the last ten years, I have constantly felt the want of a comprehensive treatise on the subject in the English language, to which I might refer the students ; and I have heard repeated expressions of regret from many of our practical Metallurgists that no such treatise existed. I now attempt to supply this deficiency ; and it is only after some years of deliberation that I decided to take this step. It is obvious that a work of this kind must be in great measure a compilation ; .never- theless, through the following pages will be found interspersed original analyses, and the results of numerous experiments and investigations, which are now published for the first time. I may be permitted to state that, although educated for the pro- fession of Medicine, and for some years engaged in the actual practice of it, I long ago acquired a strong predilection for the study of Metallurgy, to which I have almost exclusively devoted my attention during the last twenty years. I have availed myself of every opportunity of visiting metallurgical works, and have recorded the results of my observations ; but I have never entered an establishment without the consent either of the proprietor or of the responsible manager ; and never will I disclose what I have witnessed, or what may have been communicated to me in confi- dence, without the fullest authority to do so. The method of arrangement which I have adopted in this work is nearly the same as that which I have followed in my lectures. So far as relates to the instruction of students, it has, I believe, succeeded. It is almost a matter of indifference in what order the metals are considered ; for, with few exceptions, the metal- lurgy of each metal forms a subject distinct, independent, and complete iivitself. PREFACE. vii The descriptions of nearly all the processes occurring in this volume have been revised in type by the managers of the various works ; and in every case I have been careful to specify the names of those gentlemen who have thus rendered me such essential service. Amongst foreign metallurgists, my thanks are most especially due to those of Sweden for their prompt and hearty co- operation. It affords me much pleasure to acknowledge my obligations to the friends who have assisted me in my investigations. While this is only an act of simple justice, it is also one of policy ; for men work with a much better heart when they know that the credit of their labours will not be appropriated by others without acknowledgment. I cannot refrain from mentioning in this place the name of one friend, who has rendered me most willing, and I may add, most valuable aid, it is that of my colleague Mr. Richard Smith. The descriptions of the methods of assaying the ores of copper and zinc are by Mr. Smith, who has been almost daily occupied, during the last ten years, in instructing students in the art of assaying. I must not pass over in silence the names of the engraver and draughtsman Mr. James Cooper and Mr. Richard W. Mallett. They have entered heartily upon their task. I impressed upon them that their first consideration should be accuracy; and I believe they have succeeded in producing wood-engravings which, in that respect, can hardly be surpassed. The illustrations for the second volume, especially those relating to Iron, will exem- plify the skill of these artists in a far more striking manner than any contained in this. The woodcuts might have been made more attractive to the eye by the free use of shadows, or by means of perspective; but I believe that they would have been thereby rendered less useful to practical men. As it is, almost every woodcut may be regarded as an accurate, though small, mechani- cal drawing; and it is only measurable drawings of this kind which are of real utility in practice. I have taken every precaution in my power to avoid errors ; and yet doubtless some I hope not many will have escaped detection. I shall be obliged to any one who, in the course of reading this volume, may discover errors, to communicate with me on the subject. This work will be completed in one more volume, containing the viii PREFACE. subjects of Iron, Lead, Silver, Gold, Platinum, Nickel, Cobalt, Arsenic, Bismuth, Antimony, Tin, Mercury, etc. Nearly the whole of the material for this volume is collected, and almost all the illustrations are executed, so that I hope it will be ready for publication before the end of 1862. GOVERNMENT SCHOOL OF MINES, Lmdon, Nov. 1861. CONTENTS. INTRODUCTION. ON CERTAIN PHYSICAL PROPERTIES OF METALS: Physical state Action of Heat Specific Gravity Crystallization Varieties of Fracture Malleability Ductility Tenacity Toughness Softness Conduction of Heat and Elec- tricity Capacity for Heat Expansion by Heat Opacity Lustre Colour. Page 1-12 GENERAL CONSIDERATIONS ON METALLURGICAL PROCESSES. ORES Page . 13 METALLURGICAL PROCESSES : Classification of Processes 14 Reduction 14 Reduction by Carbon 14 Smelting 18 Flux and Slag 18 Regulus 19 Speise 19 Roasting 19" Distillation 20 Sublimation 20 Liquation 20 SLAGS 20 Atomic constitution of Silicates ... 21 Constitution 23 External characters 25 Brittleness and toughness 27 Page Colour 27 Fusibility 29 Experiments of Berthier on the fusi- bility of mixtures consisting of silica and various bases 33-37 Sefstrom's experiments on the for- mation of certain silicates of lime, magnesia, and alumina 3'J On the fusibility of certain com- pounds not containing silica ; alu- minates, &c 42 Sesquioxide of iron and lime 43 On fluorspar as a flux 43 Plattner's experiments on the melt- ing-points of slags 40 Objections to the method 47 Melting-points of silicates as indi- cated by the fusion of alloys of gold and platinum 48 Supposed sulphosilicates 49 FUEL. General remarks 50 ( )n the calorific power of fuel 53 Experiments of Rumford 53 Researches of Favre and Silber- mann 55 Berthier's process of estimating tin; calorific power of fuel 57 Tttl )le of Calorific Powers 58 On the calorific intensity of fuel ... 59 CLASSIFICATION OF FUELS 02 WOOD 02 Kinds of wood employed as fuel 02 Elementary composition of dry wood 03 Proportion of water in wood ... 05 Specific gravity of wood 08 Proportion of ashes yielded by wood ( '8 CONTENTS. Page Composition of the ashes of wood 69 On the rapidity of growth of wood 71 Weight of wood 72 Practical directions for the cut- ting and storing of wood in- tended as fuel 72 PEAT OR TURF 73 Specific gravity of peat 73 Composition of peat 74 Composition of the ashes of peat 75 Complete analysis of peat 76 Desiccation of peat 77 Extraction, and preparation of peat 77 COAL 78 Impossibility of proposing an exact definition of Coal ' 79 Analyses of coals 80 Ashes of coal 82 Errors in analysis of coal 84 Lignites 85 Classification of lignites ac- cording to external cha- racters 85 Composition of lignites 86 Composition of the ashes of lignite 90 Bituminous coals ,.' 91 Caking coal 91 Free-burning coal 96 Cannelcoal 96 Anthracite 96 Fibrous and granular matter in coals 96 Composition of the coals used in copper-smelting 97 Composition of bituminous coals 99 British caking coals 99 Foreign caking coals . . . k 100 ' British non-caking coals 102 Foreign non-caking coals 104 Caunel coals 105 British and foreign anthracites 105 Composition of the ashes of coals 106 On the occurrence of certain metals iii coals 106 Fremy's chemical researches on combustible minerals 106 CHARCOAL 107 Specific heat and specific gravity of charcoal 108 Analyses of Faisst and Violette 110 Various modes of charcoal-burn- ing Ill Charcoal-burning in piles or stacks Ill Chinese methods of charring in pits 123 Charcoal-burning in ovens or kilns at Dalfors Iron-works, Sweden ,. 125 Page Yield of charcoal 128 Yield by volume 129 Yield by weight 130 Influence of temperature upon yield 131 Illustrative results of charcoal- burning in piles 131 Summary of practical directions in charcoal-burning 133 Theory of charcoal-burning in circular and rectangular piles 134 Cost of charcoal-burning in cir- cular piles 141 Peat charcoal or coke 142 Carbonization by super-heated steam 143 COKE 144 History 144 Properties of coke 146 Composition of coke 146 Presence of water in coke 146 General considerations on the preparation of coke 147 Coking in piles 149 Circular piles 149 Coking in long piles or ridges 152 Coking in large open rectangular kilns -. 152 Coke-ovens 157 Cox's coke-oven 159 Jones's coke-oven 162 Coke-oven of the Brothers Appolt..... 167 - Composition of the waste gases of coke-ovens 175 Economic application of the waste gases of coke-ovens 178 Coke-ovens at Seraing, of which the waste heat and gases are applied to the raising of steam 182 Davis s breeze-oven 186 Mineral charcoal 188 Coking of non-caking coal slack by admixture with pitch 189 Collection of products of econo- mic value generated during the process of coking 191 Desulphurization of coke 194 Cost of coking 196 Combustible gases 198 Carbonic oxide 198 Hydrogen 203 Hydro-carbons 203 Concluding observations on Fuel 203 Comparison of fuels in regard to calorific power 203 Calorific power calculated from ultimate composition 204 Calorific pow r er of other kinds of fuel 205 Evaporative power of coals 206 Stacks or chimneys 207 CONTENTS. XI NATURAL REFRACTORY MATERIALS EMPLOYED IN THE /CONSTRUCTION OF CRUCIBLES, RETORTS, FURNACES, Ac. Page FlRE-CtAtS ......... 208 Composition of foreign fire-clays. . . 213 Composition of British fire-clays. . . 214 CRUCIBLES. 216 Earthen or clay crucibles 216 Stourbridge clay crucibles 220 Cornish crucibles 221 London crucibles 222 Hessian crucibles 223 French crucibles 224 Belgian crucibles 225 < } ra.pl lite, black-lead, or plumbago crucibles... .. 225 Page Composition of graphite from different localities 226 Mould for making very small cru- cibles 228 Lining crucibles with carbon 229 Covers of crucibles 230 Crucible stands 231 Tongs for crucibles 231 Sefstrom's blast furnace 231 Deville's blast furnace 232 Fire-bricks 2:{f> Dinas fire-brick 236 Sand and sandstones 288 Blue bricks .. 240 COPPER. History 241 Colour lustre crystalline system -malleability and ductility te- nacity --specific heat linear dila- tation by beat action of heat. 241 Atomic weight 242 Action of oxygen 242 Dioxide of copper 242 Protoxide of copper 243 Dioxide of copper heated with silica 243 Protoxide of copper heated with silica 244 1 Moxide of copper heated with silica and alumina 244 Trot oxide of copper heated with si- lica and alumina 245 Borates of copper 245 Disulphide of copper 246 Disulphide of copper heated with other sulphides 246 Disulphide of copper heated with ac- cess of air 247 Theory of the process of heating di- sulphide of copper with free access of air, or roasting experiments of Plattner 247 Disulphide of copper heated in ad- mixture with dioxide, protoxide, or sulphate of copper 249 Copper heated with protoxide of lead 250 Copper heated with sulphide of lead 252 Copper he; i led with sesquioxide of iron 252 ( 'opper heated with peroxide of man- ganese 252 Protoxide of copper heated with me- tallic lead 253 Dioxide of copper heated with pro- toxide of lead 253 Protoxide of copper heated with pro- toxide of lead 253 Protoxide of copper heated with sul- phide of lead 254 Dioxide of copper heated with proto- sulphide of iron and silica 254 Disulphide of copper exposed to the action of hydrogen at high tem- peratures 2">f> Disulphide of copper exposed to the vapour of water at a high temper- ature 255 Metallic copper exposed to the action of the vapour of water at high tem- peratures 256 Disulphide of copper : Heated with carbon 257 iron 257 zinc 258 lead 258 tin 259 antimony 260 Copper heated with tersulphide of antimony 260 Disulphide of copper : Heated with protoxide of lead 261 sulphate of lead 262 nitre 262 caustic soda 262 earbonatc of soda .... 26:5 baryta or lime 26:! cyanide of potassium 263 Xll CONTENTS. Page Copper and dioxide of copper 264 Copper and carbon 269 Overpoled copper 273 Copper and nitrogen 278 Copper and phosphorus 279 Copper and arsenic 281 Copper and silicon .' 282 Specific gravity of copper 283 Electric conductivity of copper 287 HISTORICAL NOTICES ON COPPER- SMELTING IN GREAT BRITAIN 289 Description of the mode of dressing and sampling copper-ores in Corn- wall 300 The sale of copper-ores 301 Swansea Copper-ore Circular 303 Standard 304 Copper-smelters in England and Wales 307 Associated Copper-smelters 307 ORES OF COPPER 309 Native copper 309 Red oxide of copper .. 309 Black oxide of copper "... 309 Green carbonate of copper, or ma- lachite 309 Blue carbonate of copper 310 Vitreous, or grey sulphide of copper 310 Purple copper 310 Copper-pyrites, or yellow copper- ore 310 True grey-copper ore, orfahlerz... 311 Chrysocolla 312 Atacamite 312 Copper-ores of Cornwall and De- von 313 THE WELSH PROCESS or COPPER- SMELTING 314 Furnaces employed : Calciner 314 Melting furnace 318 Copper-smelting in six operations : At the Hafod Works in 1848 ... 322 Modifications of the Welsh process of copper-smelting 320 The process of making " best se- lected " copper 329 Modifications of the process of making " best selected " copper at different works in 1859 330 Copper-smelting in Chili 331 Fusion for regulus 331 Roasting for spongy regulus 332 Roasting for blister-copper 332 ON THE KE- ACTIONS WHICH OCCUR IN THE WELSH PROCESS OF COP- PER-SMELTING 332 Calcination 333 Composition of the gaseous pro- ducts which escape from the ore-calciner 337 Total amount of sulphur annually evolved from the copper- works of Swansea and its vicinitv 338 Important practical conclusion concerning calcination Melting of the calcined ore External characters of coarse-metal Composition of coarse-metal External characters of ore-furnace slag Composition of the ore-furnace slag Specific gravity of the coarse -me- tal and ore-furnace-slag C oncluding observations Calcination of the granulated coarse-metal Melting of calcined granulated coarse-metal White-metal ., Puge 342 342 342 342 342 343 345 345 346 347 347 347 349 350 359 361 361 362 364 366 368 Blue-metal Slag Moss-copper Roasting Blister-copper Roaster-slag Best selected process Refining Specimens illustrating the operations of copper-smelting at Hafod ON THE ELIMINATION OF CERTAIN FOREIGN METALS DURING THE WELSH PROCESS OF COPPER- SMELTING 369 Elimination of arsenic 370 ,, antimony 372 tin 374 , , nickel and cobalt .. 375 , , gold and silver 378 Alleged superiority of the copper made by the Welsh process as for- merly practised 380 MISCELLANEOUS NOTICES OF VA- RIOUS IMPROVEMENTS IN COPPER- SMELTING. Furnaces 381 Napier's process 382 Method of smelting proposed by MM. Rivot and Phillips 385 Smelting rich copper-slags in a blast-furnace 386 COPPER-SMELTING IN BLAST FUR- NACES 387 Copper-mine and smelting-works in Sikkim 388 Copper-smelting at Singhana, in India 391 Copper-smelting in Japan 392 Copper-smelting in Sweden 395 Ore-furnace 395 Black-copper furnace 398 Refining-hearth 399 Roasting, or calcination 401 Fusion of the roasted ore 402 Roasting of the regulus from the last operation. 405 Fusion for black-copper 405 CONTENTS. xni Page Refining 406 Toughening 409 Copper rain 410 Consumption of fuel 411 Loss in smelting 411 Copper-smelting at Roraas in Norway 411 Results of copper-smelting at Atvidaberg 412 Smelting of copper-schist in Prus- sian Saxony 413 Analytical data 418 Smelting of copper-schist in Hesse 426 Other accessory products 432 Copper-smelting in Perm in Russia 433 Cupriferous pig-iron 434 Theory of the process 437 KERNEL-ROASTING AT AGORDO ... 439 Composition of the ore 440 Roasting 44 1 Styrian kilns 441 Mode of charging 443 Loss of copper 444 Changes which the ore undergoes 444 Analysis of kernel and shell 445 Theory of the process 446 WET METHODS OF EXTRACTING COPPER : Precipitation of copper from solu- tion by iron 447 Bankart's process 447 Wet process by M. Escalle 450 Hahner's patent 451 Remarks on the Patent Laws 452 ASSAYING OF COPPER-ORES BY THE CORNISH METHOD : Furnace and implements 454 Fluxes, reagents, &c 458 Sampling 461 Preliminary examination 461 Chief characteristics of the process 462 Proportions of fluxes, &c 463 Assay classification of ores and cu- priferous products 467 Page Practical directions for conducting the process 468 Special modes of assaying 476 Foreign metals occurring in copper- ores, &c 477 Methods of estimating copper by wet assay " 478 By cyanide of potassium 479 By precipitation with hyposul- phite of soda 485 By a standard solution of hypo- sulphite of soda 486 Coloration-test 487 Inaccuracy of the Cornish method of dry assaying 489 Comparative results by Cornish and Wet methods 490 Loss of copper 494 Commercial details concerning cop- per-smelting 496 Freights 496 Weights by which copper-ore is sold 497 Cost of the Welsh method 497 Sir W. Logan's formula 499 Cost of copper-works and capital required 499 Turning over of capital 501 Profit of copper-smelters in the last century 501 Cost of copper-smelting at Atvida- berg 502 Composition of commercial copper from various localities 503 Copper Sheathing 505 Amount of corrosion in sheathing made from different kinds of copper 508 Extracts from Dockyard Reports .. 509 Corrosive susceptibility of different qualities of copper 512 Treatment of the cupriferous resi- dues of the pyrites imported for the manufacture of sulphuric acid 517 ZINC. History 518 Analytical evidence ..; 521 Descriptive evidence 523 Physical properties 528 Certain chemical properties of zinc : Atomic weight Action of oxygen 531 Oxide of zinc 532 Action of water on zinc 533 Reduction of oxide of zinc by car- bon and carbonic oxide 534 Reduction by hydrogen 535 Silicate of zinc heated with carbon 537 Heated with lime and charcoal 53H Oxide of zinc heated with lx>racic acid 538 Heated with alumina '539 protoxide of lead .. 539 the fixed alkaline carbonates 539 cyanide of potas- sium 539 Sulphide of zinc 539 Heated with other sulphides 541 ,, access of air 541 , , oxidr of /inc 541 543 XIV CONTENTS. Page Heated with various metals 543 Heated in the vapour of water .. 544 Heated with carbonic acid 544 , , protoxide of copper 544 , , protoxide of lead .. 545 peroxide of man- ganese 545 nitre or nitrate of soda 545 , , . . carbonate of potass or soda 545 lime 546 Zinc and carbon 546 Zinc and phosphorus 546 Zinc and arsenic 547 ORES OF ZINC 548 ENGLISH PROCESS OF EXTRACTING ZINC : Boasting or calcination of the blende 550 Pots and condensing tubes 551 Reduction house 552 Mode of making the pots 555 Mode of charging the pots, and management of the furnace 557 Treatment of the rough zinc 558 Cost of production 558 SILESIA N PROCESS OF EXTRACTING ZINC : Retorts and appendages 558 Annealing-oven 560 Clay nozzles or condensers 561 Laggins or. stoppers 562 Iron appendages 562 Page Description of the furnace 562 Tools employed 566 Nature and preparation of the ore 567 Calciner ... 567 Distillation of zinc 569 Melting of distilled zinc 570 Yield of zinc 571 Consumption of fuel 571 Repairs , 572 Modifications of details of the pro- cess 572 Cost of production 575 BELGIAN PROCESS OF EXTRACTING ZINC : Retorts and appendages 578 Description of the furnace 580 CARINTHIAN METHOD OF EXTRACT- ING ZINO 585 Zinc fume 586 Montefiori furnace 587 Foreign matter in commercial zinc 588 Various methods of extracting zinc compared 593 Alleged improvements in the extrac- tion of zinc 595 METHODS OF ASSAYING ORES OF ZINC : By sulphide of sodium 598 By solution of ammonia and car- bonate of ammonia 601 Results obtained 602 By a standard solution of bichro- mate of potash, &c 604 BRASS. Definition 606 Historical notice 606 Valuable qualities 607 Malleability 607 Crystals 609 Process of stamping 609 Dead-dipping 610 Qualities of various alloys of copper and zinc " 611 PREPARATION OF BRASS : Manufacture of calamine brass ... 612 Cost of making calamine brass in the last century 616 Direct preparation of brass 618 Muntz's metal 619 Miscellaneous observations on brass 621 APPENDIX : Analysis of a bad sample of copper sheathing 624 INDEX ..625 UNIVERSITY METALLURGY, INTRODUCTION. METALLURGY, as at present understood, is the art of extracting metals from their ores, and adapting them to various purposes of manufacture. The etymology of the word " metallurgy " (jueVctXXo)', ore, metal, and t'pyoi', work) would seem to imply a more extended meaning, and to include all the arts in which metals are wrought into objects of utility or ornament ; but this is not the present meaning of the word. It is the province of the miner to extract ores from the earth, and by the mechanical processes of dressing to free them more or less com- pletely from foreign matter, and so render them fit for treatment by the metallurgist.. The knowledge of the principles of metallurgy is the science of metallurgy. And as the phenomena observed in metallurgical pro- cesses relate to physics and chemistry, and as mechanical appliances of various kinds are employed in these processes, it follows that the sciences of chemistry, physics, and mechanics must be the foundation of the science of metallurgy. In order, therefore, that the student may with advantage enter upon the study of metallurgy, it is essential that he should possess a considerable amount of preliminary knowledge, As the word science in relation to a manufacturing art is often vaguely used, it may be well to give the following illustration of its meaning. When an ore of copper, consisting essentially of copper, iron, sulphur, and silica, is subjected to a series of processes, such as heating with access of air under special conditions, melting, &c., copper is separated in the metallic state. The sum of these processes is termed the smelting of copper. In this operation of smelting, certain chemical changes take place : the sulphur combines with the oxygen of the air, and is evolved chiefly as sulphurous acid ; the iron is similarly converted into oxide, which combines with the silica present to form a fusible compound or slag. There are thus several facts which are proved on chemical evidence. These facts, when systematically arranged, may be said to constitute the scientific know- ledge of copper-smelting ; and that knowledge implies necessarily a knowledge of the chemical relations of copper, iron, sulphur, oxygen, and silica to each other. There are many other facts connected with copper-smelting, but those mentioned suffice for the present purpose of illustration. The man who conducts the process of copper-smelting in ignorance of these facts, has simply an empirical, in contradistinc- tion to a scientific, knowledge of the art. B PHYSICAL PROPERTIES OF METALS. IN TROD. The history of metallurgy dates from the remotest antiquity ; and, as Le Play correctly observes, " most of the fundamental phenomena of metallurgy were discovered and regularly applied to the wants of man before the physical sciences properly so called existed." l The term metal, like the term acid, is rather conventional than strictly scientific. Formerly, when science was much less advanced than at present, the metals constituted a well-defined class of elements. The properties which were regarded as specially characteristic were physical, and were not founded on chemical relations. Thus lustre, sui generis, and high specific gravity were considered to be essential characters of all metals. But we are now acquainted with metals which have a lower specific gravity than water, and with non-metallic elements which present a strong metallic lustre : sodium and lithium are examples of the former, while carbon in the state of graphite and a variety of crystallized silicon are examples of the latter. By far the greater number of the elementary bodies at present known are metals. ON CERTAIN PHYSICAL PROPERTIES OF METALS. PHYSICAL STATE. They are all solid, excepting mercury, at the ordinary temperature of the atmosphere. Action of Heat. The following is a convenient practical classification of the metals, founded on their degree of fusibility : 2 a. Fusible below redness. Tin, lead, &c. b. Fusible above redness, but at temperatures easily attainable in furnaces. Copper, gold, &c. c. Fusible only at the highest temperatures attainable in furnaces. Nickel, manganese, &c. d. Practically infusible, at least in ordinary furnaces. Platinum, indium, &c. > instrument by which high temperatures ly measured. The ordinary language em- peratures, such as red heat, white heat, &c., , in judging of temperature by the eye, much erver as well as upon the degree of illumina- ervation is made. an air-thermometer provided with a bulb of rmine with precision high degrees of temperature ; ns he has arrived at the following results : 3 Unfortunately t curately indicate atisfac L 1 NvlllC : li I Precedes Metallur- ; shown that, when certain metals and le Pays de Galles j other bodies are subjected to great pres- i,, n,,: Paris, sure, they require a higher temperature for fusion. Brit. Assoc. 1854. 3 Comptes Rendus, 1836, 8, p. 782. du Cuivre. of Cambridge, has INTROD. PHYSICAL STATE SPECIFIC GRAVITY. Incipient red heat corresponds to 525 C 977 F. Dull red 700 .. 121)2 Incipient cherry red Cherry red Clear cherry red Deep orange Clear orange White Bright white Dazzling white 800 1472 900 3052 1000 1832 1100 2012 1200 2192 1300 2372 HOO 2552 1500 to HiOO 2732 to 2912 Metals are either fixed or volatile by heat. a. Fixed metals. Gold, copper, nickel, &c. b. Volatile metals. After fusion : Cadmium, zinc, &c. ; without fusion, passing directly from the solid to the gaseous state : Arsenic. It is important to remark that the term fixed is necessarily compa- rative and conventional. A metal, which may be considered as prac- tically fixed, may yet be volatilized at very high temperatures, such as are attained by concentrating the solar rays in the focus of a mirror or lens, by the voltaic current, or by the combustion of oxygen and hydrogen. M. Despretz has published some interesting experiments on the production of intense heat, by employing in conjunction the heat derived from the solar rays, the oxy-hydrogen blowpipe, and a powerful voltaic battery. By this triple source of heat, magnesia, it is stated, was immediately volatilized in the form of white vapour, and anthra- cite melted. By the action of a very powerful Bunsen's battery alone, even carbon was melted, volatilized, and condensed in the state of a black crystalline powder ; silicon, boron, titanium, and tungsten were melted ; and 250 grins. (3858 grs.) of cuttings of platinum were melted in a few minutes. 4 Deville and Debray have recently succeeded in melting not less than 25 kilogrammes (55 Ibs. avoird.) of platinum at a time by the heat resulting from the combustion of coal-gas by oxygen. 8 Iron becomes soft, and remains so, through a considerable range of temperature below its fusing-point. In this soft state two pieces of iron may be made to unite by compression, or, in other words, may be welded together. In the process of welding the metal is never in a state of absolute fusion. SPECIFIC GRAVITY. The specific gravity of metals at the ordinaiy temperature ranges between 0-6 6 and 21 '5. 7 It varies within certain limits with the special molecular condition of the metal consequent on previous treatment. 8 The processes of hammering, rolling, and stamping tend to increase the specific gravity of metals in the state in which they exist after fusion. Marchand, however, asserts that pp 4 Comptcs Rendus, t. 28, p. 755 ; t. 29, Ib'id., t. 50, p. 1038. June, 1S(5(). Sp. gr. of lithium as determined by Bunsen. 7 Sp. gr. of fused platinum, Deville, Ann. de Chira. et de Phys. 1859, 3. s. 56, p. 4 20. 8 Vid. specific gravity of copper in the sequel. n 2 4 PHYSICAL PROPERTIES OF METALS. INTROD. bismuth is an exception, and that by subjecting the metal to great pres- sure its specific gravity is diminished. But Dr. Tyndall informs me that he has not been able to verify this assertion by experiment with a small hydraulic press. In recording the specific gravity of a metal, it is important in every case to specify the treatment which it may have previously received, as well as the temperature at which the observation was made ; for otherwise the results of different observers cannot be satisfactorily compared. CRYSTALLIZATION. The brittle metals in common use always exhibit a well-marked crystalline structure. This is especially the case with zinc, antimony, and bismuth, upon the fractured surfaces of which distinct crystalline faces may be seen. The soft metals, such as lead and tin, may also be readily obtained well crystallized. The metals, so far as observations have extended, have been found to crystallize either in the cubic or rhombohedral systems : thus gold, silver, lead, and copper occur either in cubes, the regular octahedron or rhombic dodecahedron; while bismuth, which was long supposed to crystallize in cubes, belongs to the rhombohedral system. The conditions under which metals generally crystallize are as follow : a. On solidification after fusion. b. By condensation from the state of vapour. c. By electrolytic decomposition of metallic solutions. a. Solidification after fusion. Slowness of cooling is, as might be expected, the condition favourable to crystallization. Thus an ingot of zinc, when allowed to cool very slowly, presents much larger crystalline faces on the fractured surface than when cooled quickly. The same fact is also strikingly exemplified in grey pig-iron when similarly treated. If one portion of this metal be allowed to run from a furnace upon a cold slab of iron, so as to be cooled with ex- treme rapidity, and another portion be allowed to run under a mass of hot slag, so as to be cooled with extreme slowness, the fracture of the latter will be much more largely crystalline than that of the former ; so much so that it may be difficult to believe that the two portions of metal should have flowed simultaneously from the same furnace. When it is desired to prepare a metal in a well-crystallized state after fusion, the usual process is to melt a considerable quantity of it in a crucible, and when the surface has solidified to a slight depth, to break the crust, and as rapidly as possible pour out the still liquid metal. ^ Pure bismuth may be thus procured in beautiful crystals ; but some dexterity is required to perform the operation successfully. I have beautiful specimens of copper, brass, and pig-iron, which have been accidentally crystallized on this principle. When melted lead is allowed to cool slowly, and is stirred from time to time, a period arrives when, owing to its partial solidification in the form of small crystals, it becomes a semi-fluid mass. The crystals, which have a higher specific gravity than the liquid metal, INTROD. CRYSTALLIZATION VARIETIES OF FRACTURE. 5 ami consequently tend to subside, may be taken out and drained by means of an iron ladle perforated with holes. This is actually done on the large scale, as will be described in the sequel^ in the well known process of Pattinson. Crystals of tin may be obtained in a similar manner with equal facility. If the lead had been allowed to solidify without being disturbed, it would have presented no visible evidence of crystallization, and yet it must have consisted of an aggregation of crystals. It seems reasonable to suppose that while slow cooling is favour- able to crystallization, extremely rapid cooling may, in some instances at least, determine a non-crystalline condition, or a molecular arrange- ment similar to that of bodies in the vitreous state. b. Condensation from vapour. When arsenic is volatilized, it con- denses in the form of a crystalline crust. The vapour of zinc may likewise condense in a distinctly crystalline form. c. Electrolytic decomposition. Metals which may be separated in the metallic state from a solution of their salts by the voltaic current, generally occur crystallized in a more or less distinct form. A current of low intensity, so long as it is capable of effecting the decomposition, is the condition favourable to the development of distinct crystals. On the contrary, a metal may be precipitated in the state of amorphous powder by a current of great intensity. The metal will appear at the pole at which hydrogen would be evolved on the electrolytic decom- position of wat er. VARIETIES OF FRACTURE. Except when otherwise stated, the frac- tured surface is supposed to be that produced in the metals at the ordinary temperature of the atmosphere. (Jryst-allme. Characteristic examples of this fracture are presented by zinc, antimony, bismuth, and the variety of pig-iron termed tipiegeleisen by the Germans, from the fact of its fractured surface pre- senting large, bright, mirror-like crystalline faces. Granular. Grey forge pig-iron affords an instance of this kind of fracture, which may be divided into coarse or fine according to degree. Fibrous. When an ordinary bar of iron is broken while cold by bending it backwards and forwards or, in the case of a thick bar, by cutting a nick across one surface, and then bending it so that rupture may take place along the line of the nick, the fractured surface pre- sents a fibrous appearance. Much stress is laid upon the " fibre " of iron thus manifested on fracture; and according to the character of the fibre a judgment is formed as to the quality of the iron. /''iiii'ly fibrous or silky. When a small piece of tough copper is broken by nicking it slightly on one side, and then bending it backwards and forwards until it breaks, the fractured surface will appear finely fibrous and present a silk}^ lustre. The fibre in this case seems to be entirely produced by the repeated bending which is necessary to cause rupture ; for when a large bar or ingot of tough copper is broken in the usual manner by nicking it sufficiently across one surface, sup- porting it at its ends with the nick downwards and then striking it 6 PHYSICAL PROPERTIES OF METALS. INTBOD. with a sledge-hammer on the part opposite the nick, the fractured surface is granular and not fibrous. But in the case of bar-iron it may be. shown that a fibrous structure pre-exists, and, therefore, that the fibrous appearance of its fractured surface can only be partially due to the process of bending backwards and forwards. When a bar of wrought-iron is exposed during a sufficient length of time to the action of dilute sulphuric or hydrochloric acid, it will be, as it were, dissected, and will then present the appearance of being com- posed of a bundle of parallel fibres ; but if the bar is melted and the melted metal similarly acted upon, a crystalline and not a fibrous appear- ance will be produced. Columnar. When some of the malleable metals are heated to a certain degree and then struck with a hammer or allowed to fall on the ground from a sufficient height, they easily break into columnar pieces. The peculiar form of the grain-tin of commerce is produced in this manner. Conchoidal. The fracture of certain very brittle alloys is distinctly conchoidal and glass-like. An alloy consisting of 2 parts of zinc and 1 of copper furnishes a good illustration of this fracture. MALLEABILITY : the property of permanently extending in all directions, without rupture by pressure (as in rolling}, or by impact (as in hammering'). It is opposed to brittleness, which is the property of more or less readily breaking under compression, whether gradual or sudden ; gold and copper are malleable, antimony and bismuth brittle. Malleability may be much aifected by temperature : thus ordinary copper is malleable when cold as well as when heated below a certain degree ; but beyond that degree it becomes so brittle, that it may be readily reduced to powder. Zinc in ingot is only malleable at about 150 C. Iron continues malleable even when near its point of fusion. Malleability is also affected by molecular condition. A metal may lose its malleability by being hammered or rolled, and can only regain it by being heated to a certain point. The method of restoring mal- leability by means of heat is constantly employed in the arts, and is termed annealing. In rolling a metal that is, in subjecting it to pres- sure between strong revolving metallic cylinders, technically termed "rolls," of which the axes are parallel, horizontal, and in the same vertical plane annealing from time to time is necessary ; otherwise, not only would the process proceed with difficulty, but the metal would crack, especially at the edges. With some metals, for example copper, it is immaterial whether the cooling after annealing takes place slowly or rapidly ; but with others it is very material : thus, a certain alloy of copper and tin is rendered most malleable by rapid cooling, while steel can only be rendered malleable by slow cooling. The malleability of a metal may be affected by the character of its crystal- line structure. Thus, when once the crystalline structure of an ingot of zinc is, as it were, broken down by the process of rolling at the proper temperature, the sheet of metal obtained may be much further reduced in thickness by rolling it while cold. The hardness which INTROD. MALLEABILITY DUCTILITY TENACITY. 7 it may thereby acquire may be removed by annealing at a low tem- perature ; but if the sheet is heated to a degree bordering on its point of fusion, it becomes extremely brittle, whether the subsequent cooling takes place slowly or rapidly. The sheet before having been heated to this point emits no sensible crackling sound by being bent back- wards and forwards ; but, afterf- having been thus heated, it emits a very audible crackling sound, due, probably, to the disruption of the- crystals, which have been reproduced, by the exposure of the metal to a temperature even below its point of fusion. Some alloys or mixtures of metals with each other undergo mole- cular changes in process of time, which affect their malleability. Brass wire will occasionally become very brittle, especially when kept in a state of tension. An alloy of tin and lead, which is used in pattern- casting for brass-foundry work, and which at first is com- paratively hard, becomes after a time so soft as to be no longer fit for use. DUCTILITY : the property of permanently extending by traction, as in wire- drawing. Although all ductile metals are necessarily malleable, yet they are not necessarily ductile in the exact ratio of their malleability. Thus, iron is very ductile, and may be drawn out into very fine wire ; but it cannot, like some other less ductile metals, be hammered or rolled out into extremely thin sheets. The following table of metals, arranged in the order of their malle- ability and ductility, 9 is usually found in books ; but since the recent and more exact examination of many metals by Deville and others, it Avill have to be considerably modified. Nickel, for example, in a state of much greater purity than hitherto prepared, has been found to pos- sess greater malleability and tenacity than was formerly believed. Malleability. Ductility. 1. Gold. 1. Gold. 2. Silver. 2. Silver. 3. Copper. H. Platinum. 4. Tin. \. Iron. o. Platimini. 5. Nickel. <>. Lead. 6. Copper. 7. Zinc. 7. Zinc. 8. Iron. 8. Tin. 9. Nickel. 0. Lead. TKNAOITY : the property of resisting rupture by traction. It is pro- portionate to the weight which the wire of a given metal is capable of sustaining. To determine the tenacity of different metals, wires of exactly the same diameter, or which have passed through the .same drtuc-plate, must be prepared; and the utmost weight, which each wire is capable of suspending without breaking, must be exactly determined. This weight is the measure of the resistance to rupture, or the tenacity. Tenacity is much affected by molecular condition, especi- 9 Eegnault, Cours ek ; mcnlaiiv a]MT " Sur la :>. s. 4, ]>. :57<>. 1S4U. des silicate* vitivux, 't sur 26 CRYSTALLIZATION OF SILICATES. in the pots in consequence of repairs being required in the furnace. He analysed a specimen of crystallized or devitrified glass, and a specimen of transparent glass composed of the same materials and in the same proportion. The results are as follow : Crystallized glass. Transparent glass. Silica. 55-85 56-84 Liine 24-14 21-15 Magnesia 7-63 6'37 Alumina 222 3'64 Oxide of iron 1-06 2'59 Soda 8-47 8'69 Potash 63 0-40 Manganese..., .. traces .. . traces 100-00 100-00 Specific gravity... 2-824 2-724 He analysed all the materials from which the glass was made, and calculated the composition from the proportions in which they were mixed, and found it to agree with that obtained by analysis ; so that there was nothing lost during the process of devitrification. The oxygen of the silica is to that of the bases nearly as 9 : 4 ; whence he deduces the formula 4KO, 3SiO a . R = CaO, MgO, NaO. Calculating from this formula, and neglecting the alumina and oxide of iron, the composition would be Silica 55-97 Lime 23'04 Magnesia 8-23 Soda 12-76 The crystals may be compared to natural augite, in which a portion of the magnesia is replaced by soda. Terrell also analysed the crystallized and transparent portions in the same piece of glass. The results are as follow : Crystallized glass. Transparent glass. Silica 63-67 63-43 Lime 18-65 18-14 Magnesia 6-12 4-47 Alumina 4-98 7-21 Oxide of iron 0-71 2-66 Alkalies 5-87 5-12 Manganese traces traces 100-00 100-00 Specific gravity. .. 2-610 2-857 The mixture used in the making of this glass differed from that used for the first. A fact first observed by Le Blanc is confirmed by these analyses, namely, the alumina and oxide of iron appear to become concentrated in the transparent glass or, as it might be termed, COLOUR OF SLAGS. 27 mother-liquor. In the last analyses the relation between the oxygen of the silica and that of the bases rather exceeds 9 : 4. 4 Slags may be more or less porous or vesicular. I have found pieces of blast-furnace slag in South Staffordshire presenting a cellular structure, which in regularity might almost be compared to honeycomb. When blast-furnace slag is allowed to flow from the furnace into water, it swells up immensely, forming a white, very light, pumice-like mass. Occasionally specimens of slag may be obtained from blast-furnaces in the form of spun-glass. Owing to some accidental condition, the melted slag has actually been spun, as it were, by the blast, just as glass is spun by a wheel. I have received beautiful specimens of this kind from my friend Mr. Levick, of the Blaina and Cwm Celyn Iron- works, and also from Prussia. In respect to brittkness and toughness there is great variation in slags. A slag will generally be tough in proportion to its slowness of cooling, just as devitrified glass, which is the result of slow cooling, is ex- tremely tough as compared with the original glass. The slag which has been previously mentioned, and of which the formula is R 2 3 , SiO 3 + 2 (3RO, SiO 3 ), is very brittle both in the crystallized and glassy state; whereas another crystallized slag from the same furnaces, of much less frequent occurrence, is extremely tough : its formula is R 2 3 , SiO 3 + 3 (2RO, SiO 3 ). In both formula* K*0 3 = A1 2 3 and KO = CaO, MgO, FeO, and MnO. I have not seen the last slag in a glassy state. When it is required to reduce slags to powder, as is frequently the case, it follows from the preceding considerations that a simple and effectual way of rendering them as brittle as possible is to cool them with great rapidity by allowing them to flow into water. In respect to colour, slags of common occurrence are generally grey, blue, green, red, brown, or black, of various shades. Occasionally the same slag is beautifully veined or marbled, with varying shades of colour ; and an attempt has been made to apply such slags to orna- mental purposes, though unsuccessfully. The cause of the beautiful blue colour, which is not unfrequently seen in slags from iron-smelting furnaces, has excited much attention. It was ascribed to an oxide of titanium by Kersten, who also referred to the same oxide the blue coloration of the Silesian zinc-retorts, which is very similar to that of the slags in question. He found this oxide in the substance of which the retorts were made. He passed the vapour of zinc over titanic acid heated to redness, and the acid became blue. He then prepared mixtures of the ingredients of which slags are composed, and melted them ; but the blue colour was not developed. When, however, they were kept melted at a strong heat in covered crucibles, with the addi- 4 Note Hiu- mi verrc a bouteille cristallistf. Par M. A. Terrell. Comp. Rend., -lf>, 093. 1857. 28 BLUE COLORATION OF CERTAIN SLAGS. tion of a little titanic acid and zinc, tin, or iron, they became blue. The specimens thus produced were inspected by Berzelius, who regarded the proof as conclusive. 5 Fournet opposed the explanation of Kersten on the ground that certain slags, which were remarkable for the amount of titanium which they contained, were not blue, but grey in the interior and pale yellow on the surface ; and that other slags, in which there was no reason to suspect the presence of titanium, were blue. He observed, moreover, that when common green bottle- glass is kept heated during a sufficient length of time at a temperature considerably below its melting-point, it was rendered opaque and acquired a blue colour similar to that of the slags. I had previously obtained the same result by experiment. Fournet had thin sections made of the glass so coloured, and found that when seen by transmitted .light it had a greenish-yellow tint. D'Artigues had before ascertained the same fact. Fournet showed that the powder of the blue slags and blue glass had only a dirty green tint. By melting a silicate of iron and alumina Berthier obtained a glass which, by reflexion, pre- sented a green, almost black tint, but which appeared resin-yellow by transmitted light. 6 From these considerations Fournet inferred that the blue coloration both in the slags and bottle-glass was entirely due to the same cause, namely, the same change in molecular arrangement which occasioned opacity in the bottle-glass. 7 Bontemps, who has had great experience in the manufacture of glass, and especially in its coloration, attributes the greenish colour of bottle-glass to oxide of iron combined with carbonaceous matters contained in the mixture. When the temperature is not very high, as, for example, in the covered pots in which flint-glass is made, oxide of iron gives a green colour, more nearly approaching yellow than blue ; but when the temperature is high, as in the manufacture of window-glass, the addition of a small proportion of oxide of iron to the mixture produces a glass of a bluish colour. He also remarks that it is known to the manufacturers of bottle- glass, that when the glass is cooled in the pot it becomes opaque blue before being de vitrified. He concludes from actual observation that glass may acquire all the colours of the spectrum from oxide of iron alone ; and that these colours are produced in their natural order in proportion as the temperature increases. Thus the manufacturers of china and earthenware obtain a ^r/>feA-rec/fromsesquioxide of iron at a certain temperature ; and at a higher temperature the same oxide yields an orange colour. These temperatures are low as compared with that at which glass is melted; and, as has been stated above, the oxide renders glass green at one temperature and blue at a higher tem- perature. 8 The blue colour of slags has also been referred to vanadium 9 and artificial ultramarine. In respect to vanadium, it may be true that certain blue slags contain this metal, but no satis- 5 Jahres-Bericht, 2nd part, 20, p. 97. 1841. 6 Traite des Essais, 1, p. 448. 7 Op. cit. 4. p. 870. 8 Inquiries on some modifications in the Colouring of Glass by Metallic Ox- ides. By G. Bontemps. Phil. Mag. 30, p 489. 1849. 9 Kersten, Ann. des Mine's, 4. s. 2, p. 483. 1842. FUSIBILITY OF SLAGS. 29 factory proof has been advanced that it is the cause of their Line colour ; and, in respect to ultramarine, I may state that in numerous experiments I have always found that the colour of ultramarine is destroyed at a temperature much below the fusing point of the compound of silicate of soda and alumina, which is termed ultra- marine base. On a review of the evidence I am inclined to the belief that oxide of iron is the essential element of the blue colour of the slags. Slags from iron-smelting furnaces have occasionally a very dark colour in mass, which might lead to the supposition that they contain a large proportion of iron. Analysis, however, has proved, as will be shown in the sequel, that this is not necessarily the case. It seems not improbable that this deep colour may in some cases be due to sulphur. In the French department of the Great Exhibition of 1851 were exhibited some vessels of glass remarkable for their intensely black colour, which Dumas assured me was produced entirely by the addition of sulphur. This is a subject which requires investigation. Berthier obtained a red glass of silicate of soda, the colour of which he supposed to be due to sulphide of sodium. 1 Slags are occasionally met with which present exquisitely beautiful iridescence, quite equal to that so highly prized in certain kinds of pottery. In sonie, consisting chiefly of silicate of protoxide of iron, crystals of the tribasic silicate are imbedded, which are marked with coloured bands of great distinctness and beauty. In my collection I have many illustrations of this kind. Fusibility. This is a subject of great practical importance. When melted some are as liquid as water, and others are in a greater or less degree viscous. A particular degree of consistency may be essential to the success of a process in which the slag is separated from the subjacent metallic matter by skimming : if too thin, metallic matter is liable to be drawn out of the furnace along with the slag ; and if too thick, it may not have properly subsided. In all cases in which the ores of valuable metals like copper, tin, &c. are smelted, the slags should be carefully inspected in order to ascertain that shots of metal are not retained in them. A slag is said to be clean or not clean, according as it is free or not free from mechanically- ditfused metallic matter. Experiments were made some years ago by Smith and myself on the fusibility of various mixtures of silica with lime, and silica with magnesia. Generally plumbago crucibles brasqued 2 with anthracite powder were employed, as clay crucibles were found to be rapidly attacked, so that no certain results could be obtained. In experiments of a similar kind made long before by Berthier brasqued crucibles were also used. The silica was in the state of fine white sand, such as is used by glass-makers ; the lime was prepared from Carrara marble slaked and re-heated so as to expel the water of hydration ; it was thus obtained in a state of very fine divi- 1 Tr. des Essai.s, 1. p. 425. I the carbonaceous lining of a crucible. - Brusque is a French term applied to | (See article on Crucibles in the sequel.) 30 FUSIBILITY OF SLAGS. sion; the ingredients were well mixed by trituration in a mortar. The furnace employed, when not otherwise stated, was an air-furnace having a stack somewhat exceeding 60 feet in height ; the fuel was anthracite; and a temperature sufficient to melt manganese could easily be obtained in this furnace. No. Numbers of the same Silicates in Formula. Ratio between the Oxygen of the Lime and Weight in grains em- ployed in each expe- riment. Composition of the Mixture per cent. the tables succeeding. of the Silica. Lime. Silica. Lime. Silica, 1 25 3CaO, SiO 3 .... 1 1 1260 690 64-60 35-40 2 26 3CaO, 2SiO 3 .. ! 1 2 12GO 1380 47-72 52-28 3 2CaO, SiO 3 ... ! 2 3 560 460 54-90 45-10 4 9CaO, 4Si0 3 . 1 3 4 1008 736 57-79 42-21 5 23 6CaO, SiO 3 .... 2 1 896 245 78-50 21-50 6 9CaO, 2SiO 3 .. 3 2 756 276 73-25 26-75 7 27 CaO, SiO 3 1 3 840 1380 37-83 62-17 1. Exposed to a white heat during three hours ; not melted. Some gelatinous silica was separated by the action of hydrochloric acid. Re-heated during an hour in Sefstrom's blast-furnace. Still not melted. Some dark grey melted particles were found in the mixture near the bottom of the crucible. A portion of this mixture was again exposed to a white heat during three hours ; not melted. Experi- ment repeated with 840 of lime and 460 of silica. Exposed to a strong white heat during two hours ; not melted. Experiment repeated. Exposed to a white heat during four hours and a half; not melted. Experiment repeated. Exposed to a white heat during three hours ; not melted. 2. Exposed to a white heat during two hours and a half; well melted ; compact, with a few large cavities in the centre ; fracture glassy, lower part slightly crystalline. Experiment repeated in a Cornish crucible. Exposed to a white heat during two hours. The mixture formed a fused mass with the crucible. Experiment repeated in a Cornish crucible enclosed in another of the same kind. Exposed to a white heat during two hours and a half. The outer crucible was only little acted upon, but the inner one was melted into a clear glass with the mixture. Experiment repeated with 840 of lime and 920 of silica. Exposed to a white heat during four hours and a half; melted. The upper part contained large cavities lined with imperfectly denned crystals ; lower part compact with a crystalline fracture. Experiment repeated with the same quantities. Exposed to a white heat during an hour; melted; crystallized in thin plates running in various direc- tions ; the transverse fracture presented the appearance of fibrous crystals radiating from the centre and intersecting each other at nu- merous points. The formula of this mixture is the same as that of tabular spar or Wollastonite. 3. Exposed to a white heat during two hours ; melted. Com- FUSIBILITY OF SLAGS. 31 pact and slightly crystalline. Experiment repeated with 840 of lime and 01)0 of silica. Exposed to a white heat during four hours and a half. About one-sixth of the mixture near the bottom was melted; fracture crystalline. 4. Exposed to a white heat during an hour and a quarter ; partly melted into lumps, which contained cavities lined with small imper- fect crystals. Experiment repeated with 750 of lime and 552 of silica. Exposed to a white heat during two hours. A small portion near the bottom melted into a compact glassy slag; the rest not melted. Experiment repeated with the same quantities. Exposed to a white heat during four hours and a half; not melted. Experiment repeated with the same quantities. Exposed to a white heat during two hours and a half; not melted. Mixture adhered together, but crumbled under very slight pressure. Experiment repeated. Exposed to a white heat during two hours and a half ; about one-third melted at the bottom. Compact, with crystalline fracture. 5. Exposed to a white heat during an hour and a half; not melted. Experiment repeated. Exposed to a white heat during two hours; not melted. 0. Exposed to a white heat during an hour and a half; not melted. Experiment repeated. Exposed to a white heat during two hours ; not melted. Experiment repeated. Exposed to a white heat during t\v<> hours ; not melted. - 7. Exposed to a white heat during an hour and twenty minutes ; about two-thirds melted into a compact glassy slag, the upper part con- taining small cavities lined with little crystals ; surface of mixture not melted. Experiment repeated. Exposed to a white heat during two hours and a half ; not melted. The same mixture was further exposed to ;i white heat during five hours ; about half the mixture melted into a porous crystalline mass ; the rest not melted, but adhered toge- ther. Experiment repeated. Exposed to a white heat during three hours; not melted. The same mixture was further exposed to a white heat during two hours and three-quarters; well melted. Com- part, with a few small cavities near the surface. The magnesia which was employed in the following experiments prepared by calcination of the carbonate. * Numbers of t.lie same Silic.ites in the taMrs succeeding. Formula. Ratio between the Oxygen of the Magnesia & the Oxygen of the Silica. Weight in grains em- ployed in each expe- riment. Composi tion of the mixture per cent. Magnesia. Silica. Magnesia. Silica. 1 30 :rtr*o, siO' 1 .. 1 : 1 756 552 57-79 42-21 2 81 :;M-o,L'Si0 3 .. 1 2 750 1104 40-64 59-36 8 :>}[-< ), SiO 3 ... 2 3 756 828 47-72 52-28 4 9MgO,4SiO 3 . :; 1 75(5 736 50-67 49-33 5 29 <;M-o,Sio :t ... 2 1 756 276 73-25 26-75 6 !i^r-4-o,2Si0 3 .. 3 2 756 308 07-25 32-75 7 32 .M-0, SiO 3 ... 1 :3 756 1650 31 34 68'66 32 FUSIBILITY OF SLAGS. 1. Exposed to a white heat during two hours; imperfectly fused into a white, hard, porous, slightly-crystalline mass. 2. Exposed to -a white heat during an hour ; about one-sixth of the mixture near the bottom was melted into a hard crystalline mass ; the rest consisted of a partially-melted, hard, slightly-crystalline mass, passing into a, frit near the top that is, a mass more or less sintered or fritted together. The crucible, with its contents, was again ex- posed to a white heat during an hour; the portion previously unmelted was now melted, yet not so completely as the portion first melted, of which the pieces could be readily recognised. 3. Exposed to a white heat during two hours ; melted. Compact, with a few cavities in the centre ; fracture crystalline. 4. Exposed to a white heat during two hours; melted. Hard, porous, with traces of crystallization. 5. Exposed to a white heat during two hours ; not melted, but fritted. 6. Exposed to a white heat during two hours; melted. Largely porous. 7. Exposed to a white heat during two hours; melted. Hard, porous, slightly crystalline. The result of the very numerous experiments of Berthier on the fusibility of mixtures consisting of silica and various bases will be found in the following tables (pp. 3337). The tabular form has Tuco^ oolootocl in. order to facilitate reference. In the description of the character of the products I have, as far as practicable, ad- hered to the precise language of Berthier. Berthier has interspersed through the descriptions of his experi- ments a series of deductions, which, on account of their interest or practical importance, I now introduce. 1. Weight for weight, soda fluxes 3 more than potash. 2. The alkaline silicates never acquire a stony aspect, and always produce glasses without any indication either of crystallization or of lamellar structure, whether rapidly cooled or left to cool slowly as in a porcelain-furnace. 3. Only those silicates of baryta fuse well which contain less baryta than 3BaO,2Si0 3 , and less silica than BaO,4Si0 3 . 4. Strontian is less fluxing than baryta. 5. No silicate of alumina is completely fusible at the highest temperature of our furnaces, but some of them soften, and all are more or less strongly agglomerated. The silicates of alumina Al 2 3 ,2Si0 3 , and Al 2 3 ,3Si0 3 appear to soften most of all. Their tendency to fuse may be diminished by the addition either of silica or alumina. 6. Amongst the alkalies, alkaline earths, and earths, the fluxing pro- perty in respect to silica increases or decreases with the chemical force of the base. It is to be observed that the solubility in water follows 3 The verb to flux is in common use amongst metallurgists and the workers in metals. EXPERIMENTS OF BERTHIER. 33 . ce capable of producing 150 ter of Wedgwood (p.) in 2 hrs. -furnace at Sevres, which r estimates at 140 p. that Berthier's work was pub- equivalent weights which he d by Berzelius at the time. erthier. ir-furnac pyrome porcelain Berthie ne in mind t nd that the Ber air- , rid od M 3S -J > " ^ " O^>-HC5lOr-ia5TH-HlOOCMCOt>CiGOOlOr- lOCO'-fMOiiOCM *ooooosa&o5-fiai>i-ic*io *j -> I !; S go u>~ 1 I s ~--s i::::::::::::::::::::::::: g g >> 1 | *B 1^^:77:77^^7:^77777777777^"" 1 1| " ^ li ^ ^ i@ |ii-s * : ill JS| JioococMCM ! "38.Ss^ ::::::: : ::::::: : ::: i:^^ 6 ^ t- ( ^ Mm = H2 J OO t t> CD * ^!>>r: t >^ ^^ IM -g 3 g ^ I I rf*ISfli : : : : : : : : : ^^ : : : : : ' * _J _^___ 3 CC t> CD O 00 O : : ' CMOrHTH'^'OO g 00 l> CO TM CO ,O g A fill s g-^g, I ! i 1 1! ^^rH XX ... ^-r^^XXXX^ 3 S,J ra '^^ll MOj-Sl ....__.._......_.. fl^P S 1) 34 EXPERIMENTS OF BERTHIER ON THE FUSIBILITY "g " 5? a " -5 % - ^ a ' . " rtH |> 00 CO ' CDOCCtM i I r-( i-t ^^^SO^aOOD,.- ,1 || II HI sjli^i'* I |s ^=.al^ i il!ll! a s-ia-J! (M T-I CO O t> CO t-o5o>OJi-4Ooeowooosoi> "CO T-HCOC5Oi i 10 >o 10 o ic 10 10 CD :o W . I P -a 1 1 g 1 - =*-, gS.a3 ^^lsllfl-8 s s 36 EXPERIMENTS OF BERTHIER ON THE FUSIBILITY 5 g oj IfSSSo- 6 ooooo p.-=^Ea&jO [v, FO rv, fc rv. .^~ -* o SP'S s c ^ 3 -^ ^o *- "S S S-rtcS j r:i^2 ^^|15 12?s|gi 1-S&& if^i|B&|ffji -u^ccSaJ) -T-(!-C< ?. ^^S^raoo.a'S^S iiiiisi:itli OF MIXTURES OF. SILICA AND VARIOUS BASES. 37 5 ba b ~ fc 3 i -2 3 "2 > 8 8 x s?^l|jljif I-1JT| n g .3 oT o> > S~. 5 73 >H -r -^ ^ 2 ^ "S rJ S ^i'"22bw'j;s; > jl*a).2a; tM '^Po -~bO l||^|||Mit|fc|l||s|| | I - p :, < S.2 ** * a -5 o k "3 .a o a i9..^ = 5'^rts t yi r ?i cs cst-"'p -C3.2 !==ii^i:li!li!ll:lii "- 1 as 6c . - '5 ^ 'i - S -^ '3 .,.7: ff^*3 - S - o= s^^^s-^'Bil -S - > "S c . .ti < i ) rt-r-r'uiJ'-;;s^a>a) tt '^ 3 I 5 fi;ilb-l^l""- =s ^"'" : '" s ' :r I'sslall^., o u ra s .^^ ^ .a ""* > rH CO O5 l> O 14 | 'F O O CO O CO i ;?Egas*g*n 5 sS9"c-9oill?;SS i 16S6lSSl1 a T-H 17-1 CC -H '^ "O t^- CO Cl O r- 1 CM CO -tl O O cccococococococococicr. ciOicr. o cs B *;8B^lil5-i ^|islCilili& =!=. -i S]U!!8M!litH^]}id r^fla -I Slag Sil ^ II gu igo opQ^g -g 8 fe S-=5 : -2 * t5 -35 ^ a "c g 5 P(4. b -3 S . s .s ^S-s-^s a i . ^.5-2c = D '* '' E S S ^ 5f 1 *=> ^.TI C aJi "2 ^ S"-s '-3 cL.22 * ^ B - ^ Bellim^hH ?^=1 il.tlllll&llii fslfi 38 FUSIBILITY OF SILICATES. follows the same order, and probably the same is true of the fusibility proper to each base. 7. Amongst the simple metallic silicates the fusibility is propor- tionate to the chemical energy of the oxide. But this relation ceases when we compare bases of different classes ; for example, the alkalies, alkaline earths, and earths with metallic oxides. Thus oxide of lead is infinitely more fluxing than baryta, although it is separated from all its combinations by baryta. The relation, however, does not always hold good with metallic oxides not very remote from each other. Thus oxide of zinc is a powerful base, and produces a comparatively infu- sible silicate, whereas protoxide of iron produces a remarkably fusible silicate. 8. The fusibility of simple silicates appears to depend on three causes the fusibility proper to each base ; its chemical energy ; and the proportion in which it enters into the compound. 9. With respect to double and multiple silicates, their fusibility depends upon that of the elementary silicates. A silicate which is infusible per se may always be melted by combining it with a proper quantity of a fusible silicate. It appears, evn, that the fusibility of multiple silicates is greater than that of the mean of the component silicates ; for many infusible, or difficultly fusible, silicates may, in combining with each other, form very fusible double silicates; for example, silicate of lime and magnesia, of lime and alumina, &c. 10. The alkaline silicates always give great fusibility to the com- pounds which they form with other silicates. It is remarkable that compounds containing a somewhat considerable proportion of alkali preserve, like the alkaline silicates, their vitreous aspect when cooled very slowly, and they do not then acquire the stony aspect and crys- talline structure, as is the case with most of the other silicates. 11. Potash and soda, when mixed together, are more fluxing than each alkali separately. 12. When a simple or multiple silicate, containing a somewhat large proportion of alkali, is heated with a fixed and irreducible base, this base sets free part of the alkali, which volatilizes. Thus 15 grammes of silicate of soda containing, of silica 1O35, and of soda 4-65 were heated (G. f.) with 5-6 of lime : a button was obtained which weighed only 19-2; so that 1-4 of soda had been volatilized. The button was compact, free from bubbles, opaque, with a stony fracture, somewhat shining. There is no doubt that volatilization of alkali is effected on this principle in blast-furnaces. 13. A clay, of whatever kind, always melts into glass when heated to 150 p. with half its weight of carbonate of potash or soda; part of the carbonate infiltrates into the brasque before combination occurs, and there remains in the melted product only about 12 to 15 per cent, of its weight. 14. Lime, which forms only infusible or very difficultly fusible compounds with silica, may produce, with a great number of infu- sible or slightly fusible silicates, compounds which melt easily. 15. Amongst the compounds which silica may form with lime and EXPERIMENTS OF SEFSTEOM. 39 alumina the most fusible are comprised between those in which the oxygen of the silica is double the sum of that of the lime and alumina and those in which the oxygen of the silica is half the sum of that of the lime and alumina ; and these compounds are fusible in proportion as the relation between the bases approaches that of 6CaO : APO a . They still melt well when the relation is 3CaO : APO 3 , but they become much less fusible when the relation is 3CaO : 2AP0 3 . The composition of clays which are richest in alumina may, with some exceptions, be expressed by the formula APO 3 , 2SiO 3 . Hence it follows that, by the addition of an amount of carbonate of lime inter- mediate between 3CaO and 6CaO, or the equivalent in carbonate of lime, they ought always to melt well; and that the fusion ought to be still more easy when in addition to lime silica is also added in a proportion ranging from SiO 3 to 4Si0 3 . But the addition of silica is almost always superfluous, because clays are rarely free from admix- ture with quartzose sand. Any clay whatever may be rendered suffi- ciently fusible by the addition of from half to three-quarters of its own weight of carbonate of lime to allow shots of metal to sink through the mass and collect into a button .at the bottom. When, as sometimes happens, clays are mixed with hydrate of alumina, it becomes necessary to add at the same time silica and lime. The addition of a small proportion of different bases greatly increases the fusibility. Thus the silicate composed of Silica 38-0 Lime 50' Alumina 6*5 Magnesia 2'0 Protoxide of manganese 3 5 melts into a compact enamel-like mass, of a greenish colour, here and there lamellar, but which is sufficiently liquid to allow shots of metal to pass through it. 16. Magnesia may, like lime, cause the fusion of the silicates of alumina, but it is much less fluxing. Sefstrom's experiments on the formation of certain silicates of lime, magnesia, '///'/ dlinnina* These experiments were made by the students of the School of Mines at Fahlun under the direction of Sefstroin. The mix- tures were exposed in crucibles lined with charcoal to the heat of a Sefstrom's blast furnace, in which wrought iron, manganese, and pure nickel could be melted. The numbers in the following description refer to those in the tables pp. 33-37. 3CaO, 2Si0 3 . No. 26. Pure white marble and white quartz were employed. Seven experiments are recorded, in all of which the pro- duct was well melted after an hour's blast. Both externally and on the fractured surface there were undoubted signs of crystallization. The fracture was alternately crystalline, splintery, or glassy. The thin edges of pieces were slightly transparent. The colour was more or 4 Jonrn. fiir Teclm. u. okonom. Chemie. Erdmanu. 13. 10. 1831, p. 145. Ex- tracted from the Jern-Kontorcts.Annaler, 1828. 40 FUSIBILITY OF SILICATES. less blue-grey, occasionally passing into sea-green. In most cases the product resembled the common blue agates of which mortars are made. They all had an iron-grey coating, which was not attacked by hydro- chloric acid, and which, therefore, could not be due to iron derived from the carbon crucible. By exposure to a red heat in an open cru- cible this film disappeared, but it came again on remelting the slag, during which there was so little loss that 6 grms ,000 of slag were only reduced to 5 grms ,997. The average specific gravity of the product obtained in five experiments was 2*861, the extremes being 2-781 and 2-893. CaO, SiO 3 . No. 27. This silicate generally melts far more easily than the last, so that if both are placed in the furnace at the same time it will be perfectly melted when the other is only just fritted. The two silicates so closely resemble each other that they cannot be distinguished in external appearance. This silicate seems to be some- what more brittle than the last. The results of five experiments are re- corded, and the average specific gravity of the products obtained was 2-744, the extremes being 2'731 and 2-755. 3CaO, SiO 3 . No. 25. The mixture could not be melted. An ex- periment was then made to ascertain whether the silicate 3CaO, 2Si0 3 (No. 26) could be made to combine with more lime. Pieces of this sili- cate were placed in carbonate of lime and exposed to as much heat as the crucible could support. On opening the crucible the whole mass appeared to be sintered together, but on taking it out it immediately fell to pieces : only the internal part where the silicate lay remained entire, the other part of the mass being in the state of white powder. The solid nucleus remained in a coherent state during 24 hours, but it became disintegrated when moistened with water, and sometimes with the evolution of heat. Other experiments were next made by heating the silicate with less lime, and in one case the product was actually melted ; but when it was taken hot out of the furnace it fell to pieces in less than a minute. By allowing the mass to cool in the crucible it remained in a coherent state during several days. This slag was analysed'and found to consist of 41-10 per cent, of silica and 58-77 of lime. It might, consequently, be regarded as a mixture of two sili- cates of the respective formulae 3CaO, 2Si0 3 and 3CaO, SiO 3 . Sesqui- basic silicate of protoxide of iron (3FeO, 2Si0 3 ) was mixed with suffi- cient caustic lime to form the silicate of lime, 3CaO, SiO 3 , after reduc- tion of the protoxide of iron. After an hour's firing was produced a tender mass of a gray spongy glass, which dissolved in hydrochloric acid with the evolution of hydrogen but not of carbonic acid and the separation of gelatinous silica. Distinct metallic particles were not perceived until after repeated melting of the mass at a still higher temperature. The proportion of silica was next increased, so that equal parts of the silicates 3CaO, SiO 3 and 3CaO, 2Si0 3 might be formed, After an hour and a half's firing a little finely foliated cast- iron and a well-melted slag were obtained, but the latter immediately fell to powder. The last experiment was repeated with a sustained blast during two hours. A button of grey fine-grained, cast-iron was EXPERIMENTS OF SEFSTROM. 41 produced and a well-melted glassy slag, which no longer crumbled to powder. 3CaO, 4Si0 3 . No. 28. The mixture became fluid in a certain degree and appeared to separate into two layers, of which the lower one was darker and more compact, and resembled CaO, SiO 3 , while the upper one, on the contrary, had a light and spongy aspect. The lower layer was found to be composed of 64 - 97 of silica and 35'52 of lime. CaO, 2Si0 3 . This silicate does not occur in the table. A melted mass was obtained, which, nevertheless, appeared to be homogeneous, especially when cut. Its colour was pearl-grey, passing into blue. oMgO, SiO 3 . No. 30. The product was a milk-white, glassy, porous mass. 3MgO, 2Si0 3 . No. 31. The product was a well melted, pearl- coloured, almost white enamel, which was crystalline externally. It seemed to melt more readily than Nos. 30 and 32. MgO, SiO 3 . No. 32. Product similar to No. 31 ; more crystalline, but not so well melted. 3CaO, 3MgO, 2Si0 3 . No. 46. This was prepared by heating a mix- ture of 3CaO, 2Si0 3 and magnesia. The product was a well-melted light blue-green glass having a granular fracture. 3CaO, 3MgO, 4Si0 3 . No. 47. Product a well-melted glassy slag, translucent like opal; in some places crystalline. CaO, 2MgO, 2Si0 3 . No. 48. Product a well-melted partially-crys- tallized enamel. L'CaO, MgO, 2Si0 3 . No. 49. Product a well-melted opalescent glass having a granular fracture. CaO, MgO, 2Si0 3 . No. 51. Product a pearl-coloured enamel, which appeared to be easily fusible. 5 3CaO, 2AP0 8 , 3Si0 3 . No. 58. Product a well-melted green glass. Its specific gravity in one experiment was 2-67 and in another 2-77. 3CaO, 2AP0 3 , 4Si0 3 . Not in the table. This mixture melted readily into a sooty glass, which was dichroic. Before the blowpipe it swelled up into white froth. Two experiments were made ; in one the specific gravity of the product was 2 -65 and in the other 2 '79. 3CaO, 2A1 2 3 , CSiO 3 . Not in the table. Product similar to the last, but somewhat difficult to melt before the blowpipe. Its specific gravity was 2'56. ."CjiO, A1 2 3 , 4Si0 3 . No. 56. Product a blue-green well-melted glass, of which the specific gravity was 2 '55. In another experiment, in which the blast was scarcely continued half an hour, the product wa also melted, but it was porous. .".MgO, APO 3 , 2 SiO 3 . No. 66. Product well melted after an hour's blast. In fracture it resembled compact dichroite. ' lu Krdmann 's Journal the formula j print for the one which I have intro- SC 3 + 1MS :< is given, and the same ap- duced, as in every other case the silica pean in the Jern-Kontorets Ammler (S) is placed after, and not before, the < V 18'28, p. 187). It appears to be a mis- base. 42 FUSIBILITY OF CERTAIN COMPOUNDS The green and blue colours of some of the foregoing compounds were probably due to the presence of oxide of iron. Oti the fusibility of certain compounds not containing silica ; aluminates, $c. That alumina may act the part of acid as well as that of base is proved by the composition of the minerals termed spinels, which occur well crystallized, and consist exclusively of alumina and a base of the EO type combined in definite atomic proportions. Beautifully crystallized compounds of this kind have been arti- ficially prepared by Ebelmen. On the other hand, in various mine- rals, such as staurolite, cyanite, and kaolinite, alumina must be regarded as acting the part of base, silica being the acid. But in many silicates containing both alumina and bases of the EO type the function of alumina cannot, in the present state of our know- ledge, be determined with certainty. From the foregoing experiments it appears, that when a mixture consisting of silicate of lime of the formula 3CaO, 2Si0 3 and sufficient lime to form the silicate of the formula oCaO, SiO 3 is exposed to a very high temperature, a fritted mass may be produced ; but this mass becomes speedily disintegrated, and much of the lime remains in a caustic state, as though it had only been mechanically diffused through the silicate. Now Sefs- trom found that when alumina was added to this mixture of silicate of lime and lime in the proportion requisite to form an aluminate of the formula 3CaO, 2A1 2 3 that is, a mixture represented by the formula 3CaO, 2Si0 3 +3CaO, 2A1 2 3 the eflect of heat was quite different. A well-melted mass was then obtained, of which the surface was rendered uneven by fine acicular crystals. At the upper part it was blue-grey, but underneath, where it was in contact with the charcoal, it had the lustre of iron, which is often produced in slags by the presence of the smallest amount of iron in the reagents em- ployed. In the colour and appearance of the fracture it resembled phosphate of lime ; its specific gravity was 2-888. The following ex- periments were made by Sefstrom on the formation of aluminates of lime in the furnace. 3CaO, 2A1 2 3 . From a mixture composed according to this formula a porous dirty-yellow coloured slag was obtained. In another trial the product was compact, black, and of the specific gravity 2*76. When heated before the blowpipe it became yellow, and then exactly resem- bled yellow wax. In a third experiment, after an hour's blast, a pro- duct similar to the last was obtained. Of all the aluminates which Sefstrom prepared this was the most fusible. CaO, A1 2 3 . Two experiments gave the same results. The product was a melted compact mass, of which, internally, the colour was be- tween grey, yellow, and brown ; in fracture it had a waxy lustre ; it contained small white particles of unfused matter. 3CaO, A1 2 3 . The product was a melted, glassy, yellowish-green slag. 6 Before the blowpipe it was light grey, and infusible. This aluminate had the same remarkable property as the corresponding 6 Von einer Mittelfarbe zwischen isabellgelb und grasgriin. NOT CONTAINING SILICA. 43 silicate, namely, that of falling to fine powder after a time. However, in the case of the aluminate this did not occur until after some months. Sesquioxide of iron and lime. I find that a mixture of these substances in certain proportions yields a well-melted slag. A mixture consisting of 160 grains of pure sesquioxide of iron and 100 of white marble ( = 50 grains of lime) that is, in the ratio of Fe' 2 3 : CaO was exposed in a covered clay crucible to a high temperature. It was perfectly melted, and when broken across resembled a black, opaque, vitreous slag : the crucible had one large perforation. In a second experiment a mixture, according to the same formula, of 40 grains of sesquioxide of iron and 25 of carbonate of lime was heated in a clay crucible lined with platinum foil. It was perfectly melted and escaped through the crucible. Berthier experimented upon the following mixtures : Alumina. Lime. Magnesia. 1. GCaO, 3MgO, A1'-O :! 19-9 56-5 23'6 2. 3CaO, 3MgO, Al-0 3 27'5 39-3 33-2 :!. GCaO, 3MgO, 4A1-O 3 49-9 35-3 14-8 4. 4CaO, iiMgO, 3APO ' 47-0 33'8 19'2 5. CaO, MgO, Al-O 3 53'5 25'5 21-0 1. (S.) Product granular, dull, fissured, very light, gritty between the fingers. It had a little diminished in volume. 2. (S.) Product granular, dull, sufficiently coherent, but gritty under the finger-nail. Considerably contracted. 3. (G. f.) Button well melted, bubbly, pale olive green, strongly translucent, fracture even, shining, and waxy, presenting no sign of crystallization. 4. (G. f.) Button well rounded, compact, stony ; fracture uneven, slightly shining, transparent in some parts. 5. (S.) Product granular and porous ; the interior of the pores rounded, which proves that softening had occurred. On fluor-spar as a flux. Fluor-spar occurs not unfrequently in ores. Its formula is CaF ; its composition is Per cent. Calcium, 1 equiv. = 20 51 '54 Fluor ... 1 equiv. = 18-8 48'45 At G0j>. it neither melts nor softens, but contracts much. At a higher temperature it melts sufficiently easily into a transparent liquid, which crystallizes on cooling. Heated in the porcelain-furnace at Sevres in a brasqued crucible it produced a compact mass, free from bubbles, and of a crystalline structure. The grains were very small, but transparent and well defined, and under the microscope their form could be recognised. 7 The letter B will be attached to those of the following experiments which were made by Berthier. 1. 2. grammes. grammes. Fluor-spar 100 100 Quartz in powder 30 47 hrriliiri, Tr. (les Essais, 1, p. 480. 44 FUSIBILITY OF CERTAIN COMPOUNDS 1 B. (S.) Button perfectly rounded, compact, without the slightest bubble ; fracture even, finely granular, and crystalline, resembling white statuary marble. It weighed 114 grms., so that there was a loss of 1 6 grms. The edges of the crucible were lined especially towards the angles with bubbly, colourless, and transparent glass, arising, according to Berthier, from the action of fluoride of silicon evolved upon the substance of the crucible. This gas, Berthier states, was produced by the action of the gases containing hydrogen which were present in the furnace. 2 B. (G. f.) Button compact, free from bubbles, white, opaque, very hard ; fracture stony, uneven, resembling compact quartz. It weighed 120 grms., so that there had been a loss of 21 grms. The edges of the crucible were coated with glass. 3. grammes. per cent. Fluor-spar 100 30'3 Quartz in powder 190 57 '5 Alumina 40 12'2 3 B. (S.) Heated in a brasqued crucible. Button well rounded, com- pact, without the least bubble ; fracture partly lamellar, partly con- choidal, translucent, clear grey. It weighed 270 grms., so that the loss amounted to 60 grms. 4. grammes. Fluor-spar 100 Quartz in powder 130 Kaolin washed and calcined... 100 4 B. Heated in a brasqued crucible (G. f.) Button well melted, com- pact, translucent, white; fracture uneven, very hard. It weighed 287*5 grms v so that the loss amounted to 42-5 grms. Berthier remarks that fluor-spar acts as a flux in two ways : by combining directly with silicates and forming fusible compounds : but chiefly by acting upon silicates and causing an evolution of fluoride of silicon. Fluor and silicon are thus removed, and the lime is propor- tionately increased. The agency of gases containing hydrogen does not seem to be necessary to determine the reaction, for the calcium may be oxidized at the expense of the oxygen of the silica,, and the silicon, reduced, escape in combination with fluor. Fluoride of calcium does not appear to form fusible combinations with oxides. B. 5. 6. 7. '"a! b. S grains. grammes. grammes. grains. Fluor-spar 1 equiv. 195 ... 9 87 ... 2 equiv. 19 "74 ... 1 equiv. 97 -5 Sulphate of bary ta, 1 equiv. 585 ... 29*16 ... 1 equiv. 29'16 ... 2 equiv. 585 5a. By Smith in my laboratory. Fused at a bright red heat, but not very liquid. Product hard, brittle ; fracture indistinctly crystal- line. There was a cavity in the centre lined with indistinct crystals ; it had a pink greyish-white colour, due, probably, to the fluor-spar. ob. B. Although heated very strongly it did not become perfectly NOT CONTAINING SILICA. 45 liquid. Product puffed out in some parts ; fracture granular crystal- line. The sides of the cavities were polyhedral, and some small pris- matic crystals might be perceived here and there. 6. B. Strongly heated ; completely melted. Compact ; fracture slightly crystalline, a little translucent, colourless, no indication of crystals. 7. By Smith. Does not melt so easily as No. 5. Product hard, brittle ; fracture compact, with traces of indistinct crystalline plates ; colour pink greyish-white. 7. 8. 'a. 6. grammes. grains. grammes. Fluor-spar 2 eq. 19-74 1 eq. 390 ... 1 eq. 9-87 Sulphate of lime, 1 eq. 21 '64 (cryst.)... 1 eq. G80 (dry) ... 1 eq. 21 64 (cryst.) 10. grains. grammes. grammes. Fluor-spar, 1 eq. 195 4'93 1 eq. 2'47 Sulphate of lime, 2 eq. 680 (dry) ... 21'64 (cryst.) ... 4 eq. 21-64 (cryst.) 7. B. Melted at a rather strong heat. Product compact; fracture uneven, with only faint traces of crystallization. Sa. By Smith. This mixture melts at a lower temperature than that with sulphate of baryta, and becomes very liquid. Product hard, fragile, opaque, pinkish- white ; fracture fibrous in the outer portion, but the centre part consists of shining indistinct crystalline plates. b. B. Very quickly and completely melts and becomes extremely liquid. Product crystalline, consisting of large plates crossing each other in different directions ; translucent, white, slightly pearly ; it contained cavities in which were deterniinable crystals.^ It is the most fusible compound of fluor-spar and sulphate of lime. 9a. By Smith. Eesult similar to 8a, but the crystallization was more distinct, b. B. Becomes very liquid. Product compact, free from bubbles, white, slightly translucent ; fracture granular, in plates, very brilliant. 10. B. Although heated very strongly, it did not completely melt, but was only softened. Product very bubbly, opaque, white ; fracture finely granular ; the internal surfaces of the cavities polyhedral. 11. 12. grammes. grammes. Fluor-spar 1 eq. 19-74 2 eq 19-74 Anhydrous sulphate of soda, 1 eq. 35 '68 1 eq 17*84 11 B. Melted and became as liquid as water. Product contracted very much on cooling; compact; fracture granular, crystalline, and strongly translucent, but there were no isolated crystals. 12 B. Melted with a slight bubbling, but it became extremely liquid. Product resembled No. 1 1 , but was harder and more tenacious. 13. 14. grammes. Fluor-spar ) ( 4 eq 19-74 Bone-ash [ c l ual wei g hts ' { 1 eq 27-67 MELTING-POINTS OF SILICATES. 13. By Smith. This mixture fuses at a red heat. Product compact, hard, brittle, opaque, white ; fracture slightly conchoidal. The cen- tral part consisted of small, interlacing, acicular crystals. 14. B. Contracted, so as no longer to touch the sides of the crucible ; softened, but without melting. Product very coherent ; bubbly, espe- cially on the lower part ; fracture dull and stony. According to Berthier fluor-spar has no action on sulphides ; when melted with them there is usually only simple mixture, but when the sulphide is very fusible and heavy it separates and subsides to the bottom of the crucible. Planner's experiments on the melting-points of slags. In these experiments Plattner endeavoured to determine the precise thermometric melting- points of various silicates occurring in slags, and to this end he availed himself of the method proposed and employed by Prinsep for the mea- surement of high temperatures. 8 This method consists in the applica- tion of what Prinsep terms pyrometric alloys of gold and platinum. From the fusion of pure gold to that of pure platinum he assumed 100 degrees, adding one per cent, of the latter metal to the alloy which measured each successive degree. The melting-point of gold is the zero of this scale. Prinsep did not suppose that these hypothetical degrees represented equal increments of heat throughout the scale, but, as he observes, they always indicate the same intensity. The alloys are rolled out and cut into pieces of the size of a pin's head ; eight or ten of these alloys are placed at a time in a small pyrometer cupel, each in a separate cavity, as shown in the annexed woodcut (a). In three of these cavities the metal is shown melted. The cupel is provided with a nicely fitting lid (6). Now, as Prinsep observes, when a series of such alloys has once been prepared, the heat of any furnace may be expressed by the alloy of least fusibility which it is capable of melting ; but Plattner has attempted to assign precise temperatures in Centigrade degrees to the melting-points of these alloys. He adopted the principle of a process which had been previously employed by B. de Saussure in the estimation of high temperatures. 9 The process consists in determining the greatest amount of any substance which can be melted before the blowpipe and comparing the diameter of the bead with that of the greatest amount of silver which can also be melted in the same manner. Daniell, by means of his platinum pyro- meter, found the melting-points of silver and gold to be respectively 1023 C. and 1102 C. ; and, accepting these as correct, Plattner pro- ceeded to deduce the melting-point of platinum. He assumed that the melting-point of alloys of silver and platinum would be the mean of Fig. 1. 8 Phil. Trans, part 1, p. 79. 1828. 9 Die Auwendmig dor envarmten Geb- laseluft im Gebiete der Metallurgie, F. Th. Merbach, 1848, p. 288 et eeq. Platt- ner's experiments are recorded in extenso in this work, from which I have derived my knowledge of his results. MELTING-POINTS OF SILICATES. 47 those of the two metals. He found by experiment that an alloy com- posed of 9*5 per cent, of platinum and 90'5 of silver had the same melting-point as gold. Let x be the melting-point of platinum. L^.-> 3 = 1102;* = 18550 in round numbers. He next ascertained the maximum amount of gold which could be melted in a given time by a given blowpipe flame produced by a blast of constant but very slight pressure. The fusion was effected in small cla} 7 crucibles of the same shape and size (-$ in. high and f in. in diameter at the top, Prussian measure). In order that the heat might be uniformly applied, each crucible was placed in a cavity in the end of a piece of good soft char- coal about four inches long and one inch square. The maximum amount of gold thus melted was 2290 millegrammes and 1990 mille- grammes of an alloy composed of 1760 of gold and 230 of platinum. When more gold, or ever so small a quantity of platinum, was added the fusion was rendered imperfect. Now in the case of the gold the number of degrees of heat may be estimated as 2290 x 1102 (the melt- ing-point of gold), but in the case of the alloy the number of de- grees represented by the fusion of the gold is 1760 x 1102. It is in- ferred that the number of degrees required for the fusion of 230 of platinum is equal to that required for 530 of gold, that is, the differ- ence of the weight of gold fused in the two cases. Hence the melting- point of platinum should be - - -2539. In a similar manner 230 Plattner ascertained that 100 millegrammes of an alloy composed of 70 of gold and 30 of platinum could be melted under the same condi- tions and in the same time as 100 millegrammes of cast-iron, of which the melting-point, according to Daniell, is 1530 C. From these latter data the number 2534 C. was deduced as the melting-point of pla- tinum. Objections to the method. This method of estimating high temperatures involves an assumption that the melting-point of alloys is the mean of the melting-points of their component metals, and this assumption, as will hereafter be shown, is entirely opposed to numerous well ascer- tained facts. Besides, it is obvious that, notwithstanding all the pre- cautions which appear to have been taken by Plattner in conducting his experiments, it must be extremely difficult, not to say impossible, to ensure the necessary identity of conditions in successive experiments, and to determine the exact moment at which perfect fusion is effected. The result of the experiment with cast-iron, irrespective of the objec- tions just stated, must be regarded as worthless, because, under the term cast-iron, may be comprised varieties of metal which widely differ, both in chemical composition and physical characters. Although we may not admit the correctness of the principle of Plattner's mea- surement of high temperatures, yet we may accept his experimental results as affording practical information of value. The melting-points 48 MELTING-POINTS OF SILICATES. of metals and their alloys are fixed and unvarying, except under extra- ordinary conditions of great pressure : and, as they extend through a very wide range of temperature, they may be conveniently employed in the determination and comparison of high temperatures. Melting-points of silicates as indicated by the fusion of alloys of gold and pla- tinum. It is stated that the fusion of a silicate may be effected at a lower temperature than that of the simple mixture of its components ; but I doubt whether this statement has been satisfactorily proved. That a longer time may be necessary to effect the fusion at a given tempera- ture in the latter case than in the former must be admitted, but it is probable that fusion would be equally well effected in both cases at the same temperature provided sufficient time be allowed. Plattner made numerous experiments to ascertain the melting-points of various silicates as indicated by the fusion of alloys of gold and platinum, and for this purpose he employed a pyrometer cupel made of well-burnt fireclay and an air-furnace. He also effected the fusion of these sili- cates before the oxyhydrogen blowpipe. Nos. 15, 16, 26, 27, 31, 32, 42, 44, in the table of Berthier's results, were melted at the same tem- perature as an alloy consisting of 42 per cent, of gold and 58 of pla- tinum ; Nos. 35, 36, at the same temperature as an alloy of 41 gold and 59 platinum ; Nos. 47 and 56 at the same temperatures respectively as alloys of 45 gold and 55 platinum, and 43 gold and 57 platinum ; and Nos. 69, 70, at the same temperatures respectively as alloys of 52 gold and 48 platinum, and 49 gold and 51 platinum. Other results obtained by Plattner are recorded in the following table. 1. 2. 3. 4. 5. 6. Silica 50*0 58-0 48-0 50-0 36-5 32-7 Alumina 17-0 6-0 9-0 ^6*0 8-5 7*0 Baryta 1-5 Lime 30*0 22-0 4-5 3-0 4-0 Magnesia 10-0 1-5 1-5 3'0 Protoxide of iron. . . Protoxide of man-) ganese ) 3-0 2-0 2-0 37-0 38-0 40-5 .. 60-3 Protoxide of lead. . . " 7-5 Temperature at which the silicate was formed, as indicated by the melting-points of alloys of gold and platinum. Au. PI. Au. PI. Au. PI. Au. PL Au. PI. An. PL In a carbon crucible 46+54 46+54 In an iron crucible .. 59 + 41 56+44 75+25 67+23 In a clay crucible 60+40 57+43 76+24 75 + 25 Temperature at which the silicate melts after its formation, as indicated by the melting-points of gold and platinum. Au. PL Au. PL Au. PL Au. PL Au. PI. Au. PL In a carbon crucible 77+23 76 + 24 In an iron crucible .. .. 84+16 82+18 85 + 15 83+17 In a clay crucible 84+16 82+18 85 + IT) 84+16 MELTING-POINTS OF SILICATES. 49 Supposed sulphosilicates. Plattner heated a mixture of 29*1 parts of silica and 108'0 of sulphate of baryta (heavy spar) in a carbon crucible. In this mixture the ratio between the silica and baryta is the same as in the silicate of the formula 3BaO, 2Si0 3 . The product was a greyish- yellow, compact, melted mass, of which the fracture was uneven, ap- parently crystalline, but without lustre. Its surface was polyhedral, iron-black (probably due to carbon), and of a semi-metallic lustre. It evolved a strong hepatic odour. When the powdered mass was treated with water sulphuretted hydrogen was liberated, and a small quantity of sulphide of barium was dissolved ; and when acted upon by hydrochloric acid it was only very imperfectly decomposed with the evolution of the same gas. This melted product was produced at a far lower temperature than that required for the formation of the corre- sponding silicate of baryta, and Plattner inferred that it consisted of a combination of silica with oxy sulphide of barium. A similar experi- ment was made with sulphate of lime (calcined gypsum). The silica and lime were in the proportion necessary to form the silicate of the formula CaO, SiO 3 ; the mixture was exposed in a carbon crucible to a temperature corresponding to the melting-point of an alloy of 42 gold and 58 platinum. The product was not melted, but consisted of a greyish -white, easily-pulverizable, sintered mass, which evolved a tolerably strong hepatic smell. In neither of these experiments was the product analysed, and, consequently, there is no certain evidence to prove the precise nature of the reactions and to justify a belief in the existence of sulpho silicates. Sulphate of baryta is so easily re- duced at a low temperature that Mr. Sewell of Nottingham, many years ago, obtained a patent for the production of carbonic acid by heating a mixture of sulphate of baryta and carbon in a common gas- retort. I visited Mr. Sewell's works and saw the process in operation. As in Plattner's experiments carbon crucibles were employed, the formation of sulphides is readily explained, and the temperature at which their formation would occur is without doubt much below what would be necessary to effect the combination of silica with either lime or baryta. Le Play * admits the existence of a sulpho-silicate of iron, but, as it appears to me, on insufficient chemical evidence. Description des Precedes Metallurg. etc, p. 212. 50 FUEL. THE term is used to denote substances which may be more or less completely oxidized or burned by means of atmospheric air, and evolve heat capable of being applied to economical purposes. There are only two elements which are thus applied, namely, carbon and hydrogen. All fuel consists either of vegetable matter or of the products of the decomposition of that matter naturally occurring or artificially induced. Vegetable matter, which chiefly consists of woody tissue, may prac- tically be regarded as composed of carbon, hydrogen, and oxygen, to- gether with a small quantity of so-called earthy matters. The former constitute the organic part, and the latter the inorganic part of vegetable matter. The original source of the organic part is water and the car- bonic acid of the atmosphere, both of which are decomposed in the economy of plants through the agency of solar light. The sun, there- fore, is really the source of the heat-producing power of all fuel. From the preceding considerations it will appear that the hydrogen employed as fuel must always be in association with carbon ; but the converse is not true, for carbon, which may be regarded as practically free from hydrogen, may be procured abundantly and employed as fuel. Thus the combustible part of anthracite and well- burnt charcoal or coke consists essentially of carbon with a small proportion of hy- drogen, which may be practically neglected. In all fuel containing carbon, hydrogen, and oxygen, the proportion of hydrogen may be equal to, or greater, but never less than, that required to form water with the oxygen. It may be shown that it is only the hydrogen in excess which is available as a source of heat, so that in the combustion of a substance of which the composition may be represented by carbon and water, the carbon alone is the source of heat. The hydrogen, indeed, in this case causes a loss of a large amount of otherwise available heat, in as much as it may be regarded as virtually existing in combination with oxygen in the state of water, and the carbon cannot be burned without the evaporation of this water at the ex- pense of the heat developed by its combustion. It is true that the hydrogen may perform an important function in such a fuel in gene- rating/ame, but the proposition is, nevertheless, true that it not only does not contribute to the actual amount of heat produced, but con- sumes, so to speak, no inconsiderable portion of it. The products of the perfect oxidation or combustion of carbon and hydrogen are carbonic acid and water respectively. Both products are obtained on the perfect combustion of any compound of carbon and hydrogen, or of these elements associated with oxygen. The amount of heat which any element in the same allotropic condition FUEL. 51 developes on complete combustion is perfectly definite, and is the same whether combustion be slowly or rapidly effected ; it is as definite as the amount of electricity evolved in the voltaic battery by the oxidation of a metal like zinc, for example. It has been ascertained that the amount of heat produced by the complete combustion of car- bon and sulphur varies in a small, though very sensible, degree with their allotropic condition. The term perfect, or complete, which has been used to qualify the degree of oxidation or combustion of carbon and hydrogen, requires explanation. In respect to carbon it expresses the maximum of oxygen with which it is capable of combining, but in respect to hydrogen it does not, for oxygenated water, or peroxide of hydrogen, contains twice as much oxygen as water. Yet in respect to both elements it expresses the highest degree of stability in the product of oxidation, the affinity by which the second atom of oxygen is held in combination in oxygenated water being extremely feeble. The intensity, or pyrometric degree, of heat, must not be confounded with the quantity of heat developed on combustion. The quantity of heat generated by the perfect combustion of a body (A) may be much greater than that of another body (B), weight for weight, but the intensity of the heat derived from B may far exceed that from A. However, the intensity of the heat developed on the combustion of the same body will, cceteris paribus, be proportionate to the rapidity of combustion ; or, in other words, it will be inversely as the time in which combustion is effected. The term calorific intensity will be employed in contradistinction to calorific power, which expresses the relative quantity of heat. When a piece of well-burnt dry charcoal is ignited and exposed freely to the air, it burns without any sensible flame, and the pro- duct is carbonic acid. When, on the other hand, a piece of light dry wood is ignited, it burns with much flame, and the products, if the combustion is perfect, are carbonic acid and water. Ordinary flame is gas or vapour of which the surface, in contact with atmos- I >li eric air, is burning with the emission of sensible light. The truth of this proposition may be easily demonstrated by experiment upon the flame of a candle or gas-jet, as is stated in every treatise on chemistry. Indeed, if it were not so, the gas and its supporter of combustion must be mixed ; but in that case there would be an explo- sion, attended only with the instantaneous production of flame. The piece of charcoal contains nothing from which any sensible amount of inflammable gas can be produced by the application of heat, and the solid carbon, in ignition, passes directly to the state of carbonic acid. Hence there can be no flame. But when a piece of wood is ignited, the case is altogether different. The inner substance of the wood in. contact with its burning surface is precisely in the condition of wood which is heated in a close vessel and evolves various combus- tible gaseous and liquid volatile products. Hence there must be flame. However numerous these products may be, it should be re- membered that, so long as they are ultimately converted into carbonic E 2 52 FUEL. acid and water, the proposition previously announced is correct in respect to the quantity of heat developed. When a piece of charcoal smoulders away in atmospheric air, or when it is burned in oxygen gas, light is evolved. In the former case it is only a dull red light, but in the latter it is intensely brilliant, yet there is no sensible flame. This statement must be made with a certain degree of reservation. It is customary to refer the production of flame to the so-called com- bustible gases ; but it would be equally correct to refer it to the so- called supporter of combustion. The luminosity of flame is caused by the presence of particles of solid matter within, or in immediate contact with, the gas in active combustion. In the flame of a candle or jet of coal-gas this matter is carbon in a fine state of division, the existence of which may be shown by holding a cold glass rod for a second or two across the flame, when the portion within the burning surface will be covered with a black soot-like deposit. The flame of a candle or of gas burning in atmos- pheric air is highly luminous, whereas that developed by the combus- tion of hydrogen by oxygen is very feebly so ; yet the intensity of the heat of the former is very small as compared with that of the latter. The luminosity, therefore, of flame affords no certain indication of its temperature. When the flame resulting from the combustion of hydrogen and oxygen in admixture, in the proportion in which they exist in water, is projected upon a piece of lime, or some other sub- stance, an intensely brilliant light is produced, and, under these con- ditions, there is a correspondence between the light and heat of flame in respect to intensity. As the temperature of a gas in active combustion under ordinary circumstances is much higher and its specific gravity, consequently, much lower than that of the surrounding atmosphere, flame will necessarily tend to rise. The length of flame will depend on the rapidity with which the combustible gas is generated, and the velo- city with which flame rises will be proportionate, cceteris paribus, to the difference of temperature between the gas in combustion and the surrounding atmosphere. It has been stated that when a piece of charcoal is ignited and ex- posed to the atmosphere it burns without flame, but the result may be different when charcoal is burned in mass, as in a furnace. In this case flame may be copiously produced by the combustion of car- bonic oxide, which is generated when carbonic acid comes in contact with charcoal heated to bright redness; and this condition always occurs in a furnace of which the fire-place contains charcoal to the depth of a few inches. When atmospheric air impinges upon incan- descent charcoal carbonic acid is formed, but as this gas rises through the superincumbent mass of charcoal heated to bright redness, it is converted, more or less perfectly, into carbonic oxide, which, after- wards coming in contact with atmospheric air, burns with its beau- tiful and characteristic blue flame. EXPERIMENTS OF RUMFORD. 53 ON THE CALORIFIC POWER OF FUEL. Although there is no means of estimating the absolute amount of heat evolved by the combustion of a body, yet the relative amounts of heat evolved by the combustion of different bodies may be accurately determined, liumford estimated the calorific power of a body by the number of parts by weight of water which one part by. weight of the body would, on perfect combustion, raise one degree in temperature. Thus 1 part by weight of charcoal in combining with 2| of oxygen to form carbonic acid will evolve heat sufficient to raise the temperature of about 8000 parts by weight of water 1 C. Similarly, 1 part by weight of hydrogen in combining with 8 parts by weight of oxygen to form water will raise 34000 parts by weight of water 1 C. The relative calorific powers, therefore, of carbon and hydrogen are as 8 : 34. The amount of heat reqriired to raise 1 gramme (15*432 grains) of water from to 1 C. is conventionally taken as the unit of heat or calorie of the French. 8 It is not a matter of indifference whether any portion of the thermometric scale be selected, because a greater amount of heat is necessary to increase the temperature of water by 1 ; ncar the boiling-point than at lower temperatures. Rumford employed in his experiments a rectangular vessel of thin sheet- copper containing a worm of three horizontal coils of the same metal. The vessel was 8 inches long, 4^ broad, and 4| deep. The worm consisted of a flat tube ^ inch in depth and 1 in breadth. One end protruded through the top of the box and the other was fitted to a circular hole in the bottom, 1 inch in diameter, and in this hole was inserted a funnel, of which the mouth was 1^ inch wide. The lower end of the worm was situated near one of the short sides of the vessel, and the other end issued vertically near the opposite side. A tube was inserted in the top of the box, through which could be introduced a thermometer, having a cylindrical reservoir in length equal to the depth of the vessel, so that by this means the mean temperature of the water with which it was filled could be ascertained. The substance of which the calorific value was required was burned under and within the funnel-mouthpiece, when a current of air would circulate upwards through the worm and escape at the opposite end. The heated gaseous products of combustion would thus convey the heat developed through the worm, from which it would be communicated to the surrounding water. In order to counteract the error arising from loss by radia- tion the temperature of the water with which the vessel was filled was reduced, just before the commencement of an experiment, a few degrees say 5- -below that of the surrounding atmosphere, and the 8 As English writers on this subject usually employ the French gramme for the unit of weight, I have followed their example. It is quite immaterial what unit is selected. The reader may substitute Crts by weight, the English grain or pound * gramme. 54 CALORIFIC POWER OF FUEL. combustion was continued until the temperature of the water was exactly 5 degrees above that of the surrounding atmosphere. By this arrangement it was estimated that the vessel would receive as much heat by radiation and conduction as it would lose during the experi- ment. With a view to diminish the effect of conduction as much as practicable the vessel was supported on pillars of wood. In order to test the power of the instrument to extract the whole of the heat from the gaseous products of combustion, they were made to traverse a second vessel, similar in all respects to the first, when it was found that the temperature of the water in the second vessel was not increased. 9 The data required in the use of this instrument, or calorimeter, are as follow : The weight of the substance consumed (n). The weight of the water (w). The weight of the copper (c) and the specific heat of copper (.?). The initial temperature of the water, or that at the beginning of an experiment (). The final temperature of the water, or that at the close of an experiment ('). Other corrections for the glass of the thermometer, &c., would be necessary in experiments of great precision, but in Eumford's appa- ratus, which is comparatively rude, they would be superfluous. By multiplying the weight of the copper used in the instrument by the specific heat of copper the weight of water is found, which, in respect to absorption of heat, would be exactly equivalent to the weight of copper in the instrument. Let x represent the amount of heat produced by the combustion of 1 part by weight of any given body in atmospheric air ; the fol- lowing formula will then express the calorific power of the body. nx = (t' t) (w -f cs] (f For example Let n = 10 parts by weight. w = 8900 ditto. c = 1000 ditto. * = 0-09515 (Regnault). t = 11 C. t' = 20 C Then _ (20 - 11) (8900 + 1000 X 0-09515) ~lo~~ That is to say, 1 part by weight of the substance on perfect combustion in atmospheric air raises 8095 parts by weight of water 1 C., or in round numbers 8000 times its weight. This is nearly the calorific power of charcoal. Vide Encyclop. Metropolitana, 1830. Mixed Sciences, v. 2, p. 266. RESEARCHES OF FAVRE AND SILBERMANN. 55 The calorific power of various bodies has been investigated by La- voisier, Dulong, Despretz, and Grassi ; but we are indebted to Favre and Silbermann, 1 and to Andrews, 2 for the most recent researches on the subject. The apparatus employed in these researches was founded on the same principle as that of Kumford's, but constructed so as to ensure far more accurate results. All necessary corrections in the calculations have been made, and every precaution seems to have been taken in conducting the experiments ; and, in general, the results of these later observers are in close accord. Researches of Favre and Silbermann. Calorific power of carbon. Favre and Silbermann experimented on carbon in the different allotropic states of diamond, graphite, and char- coal. Andrews and they ascribe the discrepancy in the results of previous observers to the ignorance of the fact first announced by Dumas and Stas that, during the combustion of carbon, even in oxygen gas, a certain amount of carbonic oxide is always produced. And when carbon is only oxidized to the degree of carbonic oxide, much less heat is evolved than when it is oxidized to the maximum, so as to form car- bonic acid. As it was not found possible to prevent the formation of some carbonic oxide during the combustion of carbon, eve^under the most favourable conditions, the amount of carbonic oxide produced in each experiment was accurately determined. This was done .by pass- ing the products of combustion first through a solution of potasfc which absorbed the carbonic acid, and afterwards through a tube containing incandescent protoxide of copper. By this means the carbonic oxide was completely converted into carbonic acid, which was collected by a solution of potass and weighed. The total amount of carbon con- sumed may thus be found, as well as the relation between the carbonic acid and carbonic oxide produced. ( '((/or/jlc power of carbonic oxide. To effect the perfect combustion of (Mi-bomc oxide it was found necessary to adopt the plan of Dulong, and burn it in admixture with one-third of its volume of hydrogen. The relative proportion of the two gases was ascertained in each experiment by passing some of the gaseous mixture in its course to the combust ion- chainber of the calorimeter over incandescent protoxide of copper, and determining the weight of the carbonic acid and water thereby pro- duced, as in the process of an ordinary organic analysis. From the mean of two experiments, 1 gramme of carbonic oxide evolves 2402-7 (say 2403) units of heat by conversion into carbonic acid. Calorific power of wood-charcoal. Correction was made in the manner already described t for the amount of carbonic oxide formed in each experiment. The charcoal operated on was freed from associated impurities by different methods, which all furnished a product yielding the same amount of heat on combustion, provided it was completely deprived of hydrogen. The same results were obtained when the 1 Ann. de Ch. et .1*- Pliys., 3, s. 1852. ] 2 Philos. Magazine, 1848. 32, p. 321, p. l. i>. 5.J07; Hi;, i. 5; 87, p. 406. I 426. 56 CALORIFIC POWER OF WOOD-CHARCOAL. charcoal was heated to the temperature of iron-assays or at about 1000 C. during a long time or when heated successively at incipient redness in a current of chlorine, hydrogen, and nitrogen. After this treatment it was again calcined. Special arrangements were adopted to determine the amount of hydrogen which might be present in the charcoal, and a correction was made accordingly. From the mean of a considerable number of results, 8080 was deduced as the calorific power of carbon in the state in which it exists in purified Avood-charcoal. The calorific power of carbon in other states will be found in the table below. It has been stated that by the conversion of 1 gramme of carbonic oxide into carbonic acid, 2403 units of heat are evolved. The amount, therefore, of carbonic oxide containing 1 gramme of carbon will evolve 5607 units of heat. But 1 gramme of carbon, in passing to the state of carbonic acid, evolves 8080 units. Hence, 1 gramme of carbon, by conversion into carbonic oxide, will evolve (8080 - 5607 = ) 2473 units. This is a very striking fact, that carbon should, on passing only to the state of carbonic oxide, evolve less than half the amount of heat which it evolves on passing to the state of carbonic acid. The probable explanation is, that when carbon combines with the first equivalent of oxygen to form carbonic oxide, much heat is rendered latent by the passage of the carbon from the solid to the gaseous state. It' was formerly propounded that the heat developed in combustion is propor- tionate to the oxygen consumed ; but in the case of carbon at least this law is assuredly erroneous. There is no exact relation, as will appear from the following table, between the calorific power and the specific heat of carbon in different allotropic states. Calorific power. Specific beat (Regnault). Wood-charcoal 8080 0-24150 Carbon of gas-retorts 8047-3 0-20360 Artificial graphite 7762-3 0-19702 Native graphite 7796'6 0-20187 Diamond 7770-1 0-14687 A remarkable fact was observed in respect to diamond, namely, the change eifected in its calorific power by heating it to 400 C. or 500 C., and then allowing it to cool before burning it. Thus, the calorific 'power, before the preliminary heating, was 7770-1, and, afterwards, 7878-7, the difference being 108-6. Calorific power of hydrogen. From the mean of six determinations, which all closely approximate, the calorific power of hydrogen was found to be 34462. The weight of hydrogen consumed in each expe- riment was deduced from the weight of water collected. Calorific power of Marsh-gas (CH 2 ). The gas was made by heating baryta with crystallized acetate of soda. Its calorific power by com- bustion in oxygen was found to be 13063, as deduced from the mean of three determinations. The relation in weight between the carbon and hydrogen in this gas is as 3 : 1. If, therefore, its calorific power were the mean of that of its elements, the number would be (8080 x o BERTHIEK'S PROCESS. 57 4- 34462) -f- 4 = 14675-5. Assuming that the calorific power of the hydrogen is the same as in its uncombined state, the calorific power of carbon (V) as it exists in Marsh-gas is B k ^ /^ ' V- " <} THE "f 8* +34462 130fis -- 13 63 - \^t' A * ., Vt* Calorific power of Olefiant gas (C 2 H' 2 ). From the mean of two detenu i tions the number deduced was 11857-8. The relation hi weight between the carbon and hydrogen in this gas is as 6 : 1. If, therefore, its calo- rific power were the mean of that of its elements, the heat developed by the combustion of the carbon existing in 1 gramme of the gas would be - = 6925-7 ; and that of the hydrogen would be 1 * 34462 = O+l 6+1 4923-1. The sum of these numbers is 11848-8, or nearly the number found by experiment. Favre and Silbermann make the remarkable statement that when carbon is converted into carbonic acid by oxygen, as it exists in protoxide of nitrogen, more heat is evolved than by its combustion in pure oxygen. Thus its calorific power by combustion in the former gas was found to be 11158, or 3078 in excess of 8080, its calorific power when burned in oxygen. If this be correct, and oxygen and nitrogen could be directly combinedso as to form protoxide of nitrogen, cold should be produced ^during the combination. />V//> A.;/ '* process of estimating the calorific power of fuel. In the erro- neous belief that the amount of heat evolved by combustion is propor- tionate to the amount of oxygen consumed, Berthier proposed to determine the calorific power of fuel by burning it by means of the oxygen contained in protoxide of lead. When charcoal, for example, is heated in admixture with a sufficient quantity of protoxide of lead, it is converted into carbonic acid at the expense of the oxygen of the protoxide, with the reduction of an equivalent proportion of lead. In respect to pure carbon, or matters containing carbon without any other reducing agent, this process might be employed with advantage, as it may be easily practised, and would yield correct results in the com- parison of one carbonaceous matter with another. But when hydrogen is present, as is nearly always the case in fuel even in charcoal and co ke it may lead to erroneous conclusions, as will clearly appear from the following considerations. Three parts by weight of carbon reduce the same quantity of protoxide of lead as one part by weight of hydrogen. But the calorific powers of carbon and hydrogen respectively in round numbers are as 8 : 34. The calorific power, therefore, of 3 of carbon : 1 of hydrogen is as 24 : 34. Hence', the same weight of lead obtained by reduction would in the case of carbon indicate a calorific power of 24, and in that of hydrogen 34 ; so that the process is inap- plicable to the determination of the calorific power of fuel containing variable proportions of carbon and hydrogen. Berthier thus describes his process. 3 Mix intimately 1 part of the 3 Traite des Essais, 1, 228. 58 FUEL. TABLE OF CALORIFIC POWERS. One gramme of each Substance. Supporter of Combustion. Product of Combustion. Number of grammes of water heated 1C. Name of observer. Diamond Oxygen.. j ? > ? 5 > > > Carbonic acid ... > > > > > j > > > > 5 > > Carbonic oxide ? Carbonic acid "1 and nitrogen f Carbonic acid... Water 7770 7811-5 7781-7 7787-5 7737-1 8047-3 7237 7167 7912 7714 8080 7900 8039-8 2473 2227 11158-2 2402-7 34462 33808 34743 34666 23783-3 13063 13108 11857-8 11942 7183-6 6850 2220-9 2249 2216-8 2258-6 2253-2 2307 3400-5 5747 1301 4134 (Favre and | Silbermann. > > > > ? Lavoisier. Duloug. Despretz. Grassi. ( Favre and \ Silbermann. Andrews. ( Favre and I Silbermann. > Andrews. J Favre and \ Silbermann. > > > Andrews. Dulong. Grassi. J Favre and \ Silbermann. 5 Andrews. J Favre and \ Silbermaun. Andrews. ( Favre and \ Silbermann. Andrews. ( Favre and 1 Silbermann. Andrews. JFavre and \ Silbermann. Andrews, j > Graphite native , , , , another spe- j , , artificial from blast- 1 furnaces / , , , , another specimen Carbon of coal-gas retorts . . . Charcoal from wood from wood > > (Protox- ) ide of ( nitrogen) Oxygen .. Carbonic oxide Hydro ' en ^as j > ( Hydrochloric ) " Chlorine Oxygen . Marsh-gas (CH 2 ) 1 aciu ) ( Carbonic acid) \ and water . . . J > > > > > > > > Sulphurous acid > > > 5 J ) ) > > (Carbonic and 1 1 sulphurous acids J Phosphoric acid Oxide of zinc . . . Magnetic oxide ? defiant gas (C 2 H 2 ) Sulphur, native, in fine crys- 1 ! tals very pure . ' / > 5 1 native opaoue , , melted seven years) previously ..... \ . , melted an hour after 1 crystallization ... j , , in the soft state 5 an) hour after melting ... J , , in the state of flowers Bisulphide of carbon Zinc " Iron fuel in the finest state of division with more litharge than it can reduce 20 parts at least, but not more than 40. Charcoal, coke, or CALORIFIC INTENSITY OF FUEL. 59 coal may be readily pulverized ; but in the case of wood the fine saw- dust produced by a fine saw or rasp must be employed. The mixture is put into the bottom of a close-grained conical crucible, and covered with 20 or 30 times its weight of pure litharge. The crucible, which should not be more than half full, is covered and heated gradually until the litharge is melted and evolution of gas has ceased. At first the mixture softens and froths up. When the fusion is complete the crucible should be heated more strongly for about ten minutes, so that the reduced lead may thoroughly subside and be collected into one button at the bottom. Care must be taken to prevent the reduction of any of the litharge by the carbonic oxide in the gases of the furnace. The crucible may now be taken out of the fire and left to cool. When cold it is broken, and the button of lead detached and weighed. If preferred, the melted contents of the crucible may be directly poured into a conical ingot-mould of metal. The accuracy of the result should be tested by repetition. Forchhammer recommends the use of a mixture of 3 parts by weight of litharge and 1 of chloride of lead, instead of litharge only, as this mixture fuses at a much lower temperature than pure litharge, and does not corrode the crucible so much as litharge. 4 ON THE CALORIFIC INTENSITY OF FUEL. Suppose 1 gramme of carbon in the state of charcoal and 2*67 grammes of oxygen that is, the exact proportion in which these elements combine to form carbonic acid be brought together and combine ; and suppose, further, that there is no loss of heat either by radiation or conduction. Now, if the specific heat of carbonic acid were the same as that of charcoal, its temperature after combustion would be found by dividing the number given as the calorific power of charcoal by the weight of the carbonic acid produced (1 -f- 2'67), and adding to the quotient the original temperature of the charcoal and oxygen that of both being assumed to be the same. But the specific heat of carbonic acid is not the same as that of charcoal. It will therefore be necessary to multiply 3-67, the weight of carbonic acid produced, by its specific heat, in order to ascertain its temperature after combustion. For the sake of simplicity, in the following con- sideration, the temperature of all the elements concerned will be taken as at C., at a constant pressure O m 760 of mercury. Let c =. weight of charcoal in grammes, p = calorific power of charcoal, 8 = specific heat of carbonic acid, T = temperature after combustion. When the product of combustion is carbonic acid, the oxygen will always be c x 2-67. pc = number of units of heat developed by combustion. T - (c + 2-67c) s Herir. u. Hiittenm. Zcitunp:, 1846, p. 465. 60 CALORIFIC INTENSITY OF FUEL. This formula, it must be borne in mind, indicates the theoretical maximum temperature which charcoal is capable of producing by combustion in oxygen under conditions which can never occur in practice. Let us now suppose that the combustion of charcoal is effected under the same conditions, by oxygen mixed with nitrogen, as in atmospheric air. The nitrogen is inert, and the effect of its presence would be simply to lower the temperature of the carbonic acid pro- duced. Let n represent the weight of nitrogen in grammes, and s its specific heat. Then T = pc (c + 2 67c) s + ns' Let us next suppose that water is present. The temperature would be much further reduced, especially by the abstraction of the heat which becomes latent in the conversion of water into steam, the state in which it must necessarily exist in the circumstances supposed. Let w represent the weight of water in grammes, s" the specific heat of steam, and I its latent heat (see note 5 , p. 61). Then T - pc id (c + 2- 67c) s + ns 1 + In the combustion of hydrogen by oxygen, let ;/ represent its calorific power and h its weight in grammes. Then T = (h + 8/0 s" In the determination of the calorific power of hydrogen, the water produced is in the state of vapour, which is subsequently condensed, so that its latent heat becomes sensible, and is estimated in the calo- rimeter. But in calculating the theoretic temperature resulting from its combustion, the latent heat must be deducted, for the reason already given. In the combustion of charcoal and hydrogen conjointly, by oxygen mixed with nitrogen, m _ pc + p'h 9hl When the combustion of the charcoal is imperfect and some carbonic oxide is formed, the temperature will suffer a corresponding reduction. Let G' represent the weight of charcoal converted into carbonic oxide, p" its calorific power ( = 2473), and s'" the specific heat of carbonic oxide. Then m _ _ pc + p"c' + p'h Qhl _ (c + 2-67c) s + (c' + l-33c') s'" + 9As" + ns' When a solid inert body is present, such as ashes in coal, its weight multiplied by its specific heat must be added to the divisor. In the combustion of fuel in actual practice by atmospheric air, the CALORIFIC INTENSITY OF FUEL. 61 maximum theoretical temperature can never be attained, for the fol- lowing reasons. Heat must be ~~ last by radiation. In air-furnaces, which are chambers having a grate^BL the bottom and a flue at the top, much heat is lost by radiation fronrfce grate. In blast-furnaces, which are chambers open at the top and ^fc|ed at the bottom, with the exception of one or more small apertures tSfeough which air is blown in, the loss by radiation may be diminished t^jithe absence of an open grate. But as the material of which furnaces Mte constructed 'must in the course of time become sensibly heated, so Must there necessarily be loss of heat by radiation in/all. There is losf heat by conduction . This occurs not only throuan the matter forming ^he furnace, but also by means of the gaseous current which must be kept constantly flowing through the inc/m descent fuel. The loss will be unnecessarily increased if too much air is allowed to pass through the fuel ; and there is reason to believe that in many cases loss from this cause may be greater than is generally supposed. r If we compare charcoal and hydrogen in respect to calorific intensity, \ we shall find that charcoal exceeds hydrogen, notwithstanding the converse is true in respect to calorific power. Let the calorific powers of charcoal and hydrogen be taken as 8080 and 34000 respectively, and the latent heat of steam at 53 7, 5 and let it be required to find the calorific intensity of 1 gramme of each of these elements. In the case of charcoal T = _--_ ._, = 10173-9 3 O ' t> / X U ' Z 1 OTC In the case of hydrogen T = ^x of^5 J) = 6822 ' 7 In the case of carbonic oxide T = ] 57 2164 = 7072<8D From these calculations, in which it is assumed that the specific heat of carbonic acid and the vapour of water is constant at all tempera- tures, it follows that in respect to calorific intensit}' the value of fuel is, cceteris paribus, great in proportion to the carbon which it contains. It is on this account that charcoal, coke, and the highly carbonaceous coals of South Wales may be so advantageously employed in the smelting of iron, which requires a very high temperature. The com- bustibility, however, of the charcoal or carbonaceous matter is an impor- tant element in this consideration. Graphite is pure carbon, yet, owing to its extreme incombustibility as compared with charcoal, its pyro- metric effect would practically be very inferior to that of charcoal. The longer the time required for the combustion of any given fuel, the greater the loss of heat by radiation and conduction, and, con- sequently, the less the calorific intensity. 5 The total heat of steam at 100 is, according to Regnault, 637, that is, inclusive of 100 from to the boiling- 100, so that the actual amount of heat to be deducted is 637 - 100 = 537. Re- gnault, Chemical Reports and Memoirs, point. But, as has been already men- Cavendish Society, 1848, p. 273. tioned, the temperature under the con- 6 Regnault. ditions supposed can never be below 2 CLASSIFICATION OF FUELS WOOD. Wood Peat. CLASSIFICATION OF FUELS. j Soft. 1 Hard. / T . . , f Bituminous wood. Lignite { Brown coal. Non-caking, rich in carbon. ( Anthracite. Non-caking, rich in oxygen. )ai < Bituminous coal < Caking. Products of car- bomzation Combust^ bomzation ... ^ i -, Coal coke. Hydro-carbons. WOOD. Wood is essentially composed of organic tissue and a small proportion of inorganic matter ; and, in its ordinary state, it contains a large quan- tity of water, which may be completely expelled at a temperature much below that at which the decomposition of the organic, part would occur. This tissue has the same elementary composition in all kinds of wood, though it may be associated with widely different kinds of organic matter in different species of trees. Thus, fir-trees contain turpentine, and oaks tannin. The organic tissue is essentially the combustible part of wood, as the associated organic matters are too small in quantity to produce any calorific effect of practical importance. In external characters there is great variation in different kinds of wood : some are light, soft, and loose in grain, while others are heavy, hard, and close in grain. They have, accordingly, been divided into two classes light and soft woods, like deal, and heavy and hard woods, like oak. The manner of burning of wood must, obviously, be connected with its external characters. Every one is familiar with the difference in this respect between deal and oak. Kinds of wood employed as fuel. Tt is but rarely that wood is directly employed as fuel in metallurgical operations requiring high tempera- tures, as the heat which it produces on combustion in its ordinary state is insufficient. It is, therefore, generally converted into char- coal. The choice of wood intended for burning must depend upon the nature of the trees which grow in the vicinity of the smelting works. In the following table is a list of the trees which most fre- quently occur in Europe : ELEMENTARY COMPOSITION OF DRY WOOD. 63 Botanical Name. English Name. French Name. German Name. Acer Pseudo-platanus, L Sycamore Sycomore Ahom Betula alba L White birch Bouleau Birke Alnus glutinosa Gaertn ... Alder Aune Erie Carpinus Botulus L. Hornbeam . Charme Fa^us sylvatica, L Beech Hetre Buche Fraxinus excelsior L . . Ash Frene . ... Esche Populus tremula L Poplar or aspen Tremble Espe garded as a variety of > P. niTa Lombardy poplar d'ltalie (Ttalienische \ Pappel. Quercus Robur, L Oak Chene... Eiche Ilex L Yeuse Tilia Europjea L (Lime or linden- 1 Tilleul Linde \ tree .. . J .ZEsculus Hippocastanum, L. Salix alba L Horse-chestnut \Vhite willow . . . Marronier d'lude Saule ... . Rosskastanie. Weisse Weide -. caprea L ( Great round- } Saule Marceau Saalweide TJlmus campestris L { leaved willow/ Elm Orme TJlme Abies excelsa, D.C. ; syn. ) Spruce fir . . Faux sapin Edeltanno Ptnus Abies, L 1 Silver fir Sapin commun Fichte. syn Ptnus Plcea, L J LiiriK EuropsBa D C Larch Meleze .. . Larche Piuus sylvestris L Scotch fir i Pin sauvage, pin'l Kiefer t de Geneve J Elementary composition of wood. The following table has been com- piled from the results of Chevandier. 7 In every case a sample was prepared for analysis by collecting and mixing the sawdust produced by sawing billets in two from the top, middle, and bottom of the trunk ; so that the mixture represented the average composition of every part of the trunk, inclusive of bark and alburnum. The sawdust was dried at 140 C., and placed in a dry vacuum until it ceased to lose weight. ELEMENTARY COMPOSITION OF DKY WOOD. 1 2 3 4 5 Name of Tree. Age and Part of Tree. Exclusive of Ash. Ash per cent Mean Composition exclusive of Ash. Carbon. Hydrogen. Oxygen. Nitrogen. Carbon. Hydrogen. Oxygen. Nitrogen Beech > j > 9 > 70 years 58 ,, 69 ,, Branch wood Shoots 49-89 49-96 49-75 50-49 49-62 6-13 6-02 6-04 6-11 6-12 43-09 42-79 43-09 42-64 43-58 0-88 1-23 1-12 0-76 0-67 0-86 1-00 0-88 2-15 1-29 49-89 6-07 43-11 0-93 7 Recherches sur la composition ele- inentaire des diffe'rents bois, et sur le rendemcnt annuel d'un hectare de forets. Par M. Eugene Chevandier. Ann. do Ch. et de Phys., 3, s. 10, p. 129. 1844. 6-4- FUEL. ELEMENTARY COMPOSITION OF DllY WOOD Continued. Name of Tree. Age and Part of Tree. Exclusive of Ash. Ash percent Mean Composition exclusive of Ash. Carbon. Hydrogen. Oxygen. Nitrogen. Carbon. Hydrogen. Oxygen. Nitrogen. (Faggots from 6 Beech young stems of 25 to 30 51-15 6-31 41-74 0-80 1-50 years old ... 'Faggots from i 7 8 > 5 J > the branches of trees 70 to 51-24 6-15 141-35 1-26 51-06 6-22 41-75 | 0'97 1-94 1-71 51-08. 6-23 | 41-61 1-08 80 years old | 'Faggots from the branches of a tree 120 50-88 6-25 41-60 1-27 1-93 | years old. 10 Oak... 120 years 50-97 6-02 41-96 1-05 2-43J From the n branches of No. 10 51-01 6-00 41-72 1-26 2-03 50-64 6-03 ! 42-05 1-28 From young 12 shoots of No. 50-09 6-07 42-31 1-52 1-68 10 j 13 j Faggots from 1 shoots 30 50-82 6-23 41 -.98 0-97 1-45" years old ... Faggots from j 14 J > the branches I of a tree 50 50-73 lost. lost 0-99 1-56 , years old . . . 50-89 6-16 41-94 1-01 Faggots from] j 15 J a tree 70 > 50-93 6-15 41-91 1-01 2-10 years old ...) j Faggots from) 16 5 ) a tree 130 51-08 6-10 41-74 1-08 2-16 years old . . . J 17 Birch 60 years old. . . 50-59 6-21 42-16 1-03 0-711 (From the 18 , , branches of 50-79 6-29 41-48 1-44 1 1-03 No. 17 ' > 50"61 ft-92 4.9-04. 1-12 19 > 5 From the] young shoots > 50-48 6-20 42-43 0-89 0-60 of No. 17...) / 20 5 > (Faggots from) shoots 30 J52-21 6-36 40-24 1-19 1-161 years old ... Faggots from 21 5 shoots 35 51-61 6-32 40-95 1-12 1-54 >; 51 93 6-31 40-69 1-07 years old . . . Faggots from' 22 > ? the branches of trees 50 to 51-97 6-25 40-89 0-89 1-26 60 years old Branches and 1 2 r Poplar,) * 6 \ Aspen f stem of a tree 25 years -50-31 6-31 ; 42-39 I 0-98 1-86 old. Average Faggots from 24 the branches of stems 25 51-02 6-28 41-65 1-05 2-98 years old ... 25 Willow (From a shoot ( 20 years old : 51-75 6-19 41-08 0-98 3-67 ELEMENTARY COMPOSITION OF DRY WOOD. 65 ELEMENTARY COMPOSITION OF DRY WOOD continued. Name of Tree. Age and Part of Tree. Exclusive of Ash. Ash per cent. Carbon. Hydrogen. Oxygen. Nitrogen. 26 Willow. . . Faggots from branches of shoots 20 years old ... 54-03 6-56 37-93 1-48 4-57 * TVIea.ii . . . 51-215 6-237 41-449 1-098 1-772 The next table is compiled from the results of Petersen and Schodler, 8 and those of Heintz: Nos. 27, 29, 31, 33, 34, 40, 41 are by Heintz. 9 The former were made in Liebig's laboratory, under his own inspection. A minute but unimportant error exists in these analyses, arising from the presence of a little carbonic acid in the ashes ; but at the most it cannot exceed 0-2 per cent. Every specimen of wood analyzed was taken from the trunk, ELEMENTARY COMPOSITION OF DRY WOOD. Name of Tree. Exclusive of Ash. Carbon. Hydrogen. Oxygen. 'Nitrogen. 27 28 29 30 30& 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 Oak, var. pedunculata Oak 48-94 49-43 48-29 48-18 48-53 48-89 48-60 48-08 48-63 49-20 49-35 49-08 49-70 49-41 49-87 50-62 49-94 49-95 49-59 50-11 48-90 49-37 49-11 5-94 6-07 6 00 6-28 6-30 6-19 6-37 6-12 5-94 6-22 6-07 6-71 6-31 6-86 6-09 6-27 6-25 6-41 6-38 6-31 6-23 6-52 6-44 43-09 44 50 45-14 45-54 45-17 43-93 45-02 44-93 44-75 44-59 44-56 44-21 43-99 43-73 43-41 42-58 43-81 43-65 44-02 43-58 44-83 44-11 44-44 2-03 0-*57 0-*99 o"87 0-68 0-*63 0-53 Beech red . . white Birch Hornbeam Alder Ash Horse-chestnut Black poplar Lime Scotch fir, old wood . . . , , young wood Spruce fir Silver fir Larch . Apple ... Box Walnut . . . ]\lcan 49-21 6-27 44-52* * Oxygen, inclusive of nitrogen. Proportion of water m wood. All wood when recently felled contains a large quantity of water, which varies in amount with the nature of 3 Ueber den absoluten Werth der ge- brauchlichsten Holzarten als Brennma- teriul; von lYIrrscii mid Schodler. An- iiulrii der Pharmacie von J. Liebig u. and. 17, p. 139. Heidelberg, 1836. All the analyses are given in the Ann. d. Mines, 3, s. 11, p. 435. 9 Brix. Untersuch. d. Heizkraft d. wichtigen Brennstoffe d. PreOBsischeD Staates. Berlin, 1853. 66 PROPORTION OF WATER IN WOOD. the tree, the part of the tree, the season of the year at which it is felled, and, in trees of the same kind, with the place of their growth. When wood is exposed to the atmosphere during a sufficient length of time, under conditions favourable to desiccation, it loses the greater part of its water ; but all kinds of wood, however well air-dried, retain on an average from 18 to 20 per cent, of water. This is a point of great practical importance in reference to the direct application of wood as fuel. The amount of hygroscopic water in wood may be ascertained by exposing it to a temperature between 100 arid 150 C. until it ceases to lose weight. Violette has determined the proportion of water expelled from wood by desiccation at gradually increasing temperatures, and has given the results in the following table : 1 Temperature of Desiccation. Water expelled from 100 parts of Wood. REMARKS. Between 200 and 225 there is slight de- composition, and water alone is not evolved. The statements in works on chemistry that wood contains a given quantity of water can only be exact in so far as they indicate the degree of desiccation. Oak. Ash. Elm. Walnut. 125 C 15-26 17-93 32-13 35-80 44-31 14-78 16-19 21-22 27-51 33-38 15-32 17-02 36-94 33-38 40-56 15-55 17-43 21-00 41-77 36 56 150 175^ 200 225 The wood which Violette operated upon had been kept in store during two years. In each experiment the specimens were exposed during two hours to desiccation in a current of super-heated steam, of which the temperature was gradually raised from 125 to 225 C. When wood which has been strongly dried by means of artificial heat is left exposed to the atmosphere, it re-absorbs about as much water as it contains in its air-dried state. 2 Af Uhr proved that the desiccation of wood by exposure to the atmosphere is much promoted by the removal of the bark. Trees were felled at the same time in June after the sap had risen, and the wood was left to dry under cover during the four following months ; it was in pieces of unequal length and diameter, of which some were barked and others left with the bark entire. His results are recorded in the following table : Loss per cent, of the original Weight of the Wood. July. August. September. October. Barked stems 34-53 0-41 38-77 0-84 39-34 0-92 39-62 0-98 TJnbarked stems Thus, after the lapse of three months the barked wood was completely air-dried, whereas the unbarked wood had not even lost one per cent. 3 1 Ann. de Chim. et de Phys. 3, s. 39, p. 307. 2 Sclmbarth, Handb. 3, p. 217, 3 Anleitung zur vortheilhaften Ver- PROPORTION OF WATER IN WOOD. 67 The wood of the youngest branches of any individual tree contains about twice as much water as that of the trunk and older branches. The following table presents the specific gravity of the wood of various trees, and the proportion of water which it contains before and after atmospheric desiccation : 4 Specific Gravity. Per centage of Water. Xame of the Tree. Freshly felled Air-dried. Freshly felled. Air dried. Quercus Robur 1-0754 0-7075 34-7 16-64 var. pedunculata . . . Salix alba 1-0494 0*9859 0-6777 0*4873 35-4 50-6 Faois sylvatica 0-9822 0'5907 39-7 18-56 Ulniu.s canrpestris 0-9476 0-5474 44-5 18-20 Carpinus Betulus 0-9452 0-7695 18-6 Larix EuropSBa 0-9205 0-4735 48-6 Pinus sylvestris 0-9121 0-5502 39-7 Acer Pseudo-platanus .... Fraxinus excelsior 0-9036 0-9036 0-6592 0-6440 27-0 28-7 18-63 Betula alba 0-9012 0-6274 30-8 19-38 Py rus Aucuparia 1 (Mountain-ash.) J Abies excelsa 0-8993 0-8941 6440 0-5550 28-3 37-1 17 53 . pectinata 0-8699 0-4716 45-2 Pyrus torminalis ) (Wild service tree.) j ^Esculus Hippocastanum Alnus glutinosa ... 0-8633 0-8614 0-8571 0-5910 0-5794 0-5001 32-3 38-2 41-6 " Tilia Europaea 0-8170 0-4390 47-1 18-97 Populus niTa 0-7795 0-3656 51-8 treniula . 0-7654 0-4302 43-7 i'astigiata 0-7634 0-3931 48-2 19-55 Salix capraea ... . 0-7155 0-5289 60-0 The degree of dry ness will necessarily vary with the state of the atmo- sphere as to moisture. By air-drying wood may shrink ^ in volume, or perhaps more. 5 The proportion of water in wood is least at the fall and beginning of the year, as will appear from the following results : 6 Proportion of Water. Kinds of Wood. Jan. 27. April 2. Abies excelsa 52*7 61-0 Corylus Avelli>na 40*9 49-2 ^Esculus Hippocastanum Acer Pseudo-platanus 40*2 33*6 47-1 40-3 Fraxinus excelsior 28-8 38*6 koblung des Holzes in stehenden und j obtained by Schiibler and Neuffer, and liegenden Meilern von Carl David af j Kumford. Uhr. Aus dein Schwedischen iibersetst v. Dr. J. C. L. Blumhof. Gicssen, 1820, p. 13. 4 Extracted from Schubarth's Handb. tullur. Karsten, 3, p. 21. 3, p. 217. It is prepared from the data 5 Syst. d. Metal. Karsten, 3, p. 23. Schiibler and Neuffer, System d. Me- F 2 68 SPECIFIC GEAVITY OF WOOD. According to Werneck the wood of trees which have grown 011 mountains, under the same conditions, is more compact than that grown in plains ; the wood of closely-grown trees is more compact than that of isolated trees ; and the compactness appears to increase in proportion to the dryness of the soil. 7 Specific gravity of wood. Wood being extremely porous, its specific gravity must necessarily vary according to the amount of water which it contains : as the water evaporates, the spaces which it occupied become filled with air, except in so far as they are contracted by shrinking. From the fact of all woods having nearly the same ele- mentary composition, it might be anticipated that the specific gravity of all would be nearly the same ; and this is found to be the case when the determination is made with wood of which the pores have been completely deprived of air and afterwards filled with water. The variations in specific gravity, which appear in the preceding table, are in great measure due to the variable proportions of air and water con- tained in the pores of the wood. After complete expulsion of air, Rumford obtained the following results : Specific Gravity. , Specific Gravity. Oak 1-5344 Pine 1-4621 Elm 1-5186 Birch 1-4848 Beech . ,. 1-5284 Lime 1-4846 Sycamore 1-4599 Poplar 1-4854 By immersion in water, and subsequent drying, the specific gravity of wood is somewhat diminished. Thus, Werneck found that the specific gravity of beech-wood was 0*56, which, after the wood had been transported by floating in water, was reduced to 0*537 ; and that similarly the specific gravity of spruce fir was reduced from 0-493 to 0-464. 8 Proportion of ashes yielded by wood. This subject has been investigated by many observers, and by no one in a more complete and satisfactor} r manner than by Chevandier, to whom we owe the following average results of not less than 524 incinerations. 9 Number of Mean percentage Name of the Wood. Incinerations. of Ashes. Willow 17 2-00 Aspen 59 1-73 Oak 93 1-65 Hornbeam 73 3-62 Alder 26 1'38 Beech 93 1-06 Scotch fir (pin) 28 1-04 Silver fir (sapin) 46 1-02 Birch .. 89 .. .. 0'85 7 Karsteu, op. cit. p. 19. Ueber Ver- I * System d. Metall. Karsten, 3, p. 22. kohlimg des Holzes, etc. F. Klein, p. ' Comptes Eendus, 24, p. 269. 1847. 92. Gotha, 1836. COMPOSITION OF THE ASHES OF WOOD. 69 The general average percentage of ashes of all these woods taken together is as follows : Quality of the Wood. Percentage of Ashes. Entire wood 1 of young shoots 1'23 Wood split into billets 2 , 1-34 Entire wood of branches 1-54 Faggots of twigs 2'27 The geological character of the ground does not appear to exert a decided influence on the proportion of ashes, at least in the case of hard woods. The variation in the proportion of ashes found by Chevandier to occur in the different specimens of wood which he analysed indicates that this inorganic constituent of wood is affected by the nature of the ground and by that of the waters which moisten the roots. 3 Wood from different parts of an old tree contains a variable proportion of ashes : the trunk contains the least, and the small branches the most ; young shoots generally contain less than old trees. The inorganic matter may not be equally distributed in the same piece of wood : thus, one portion yielded 2 '64 per cent., and another only 0-69. The same result occurred in the incineration of 10 other specimens. 4 Composition of the ashes of wood. Although numerous analyses of the ashes of various kinds of herbaceous plants are recorded, yet I have found but few of the ashes of wood ; and of these scarcely one can be regarded as complete. The following will suffice for illustration : Composition of the Ashes of Wood. Composition of the Ground. L 2. 3. 4. 5. 6. 7. 8. Potash 12-81 1-60 26-72 2-22 1-38 0-78 trace 2-88 18-83 8-13 0-02 5-67 1-25 46-89 1-69 0-42 0-47 1-67 1-51 24-67 4-22 ;; 14-78 2-77 11-78 3-' 81 4 00 12-92 16-65 2-77 6-94 0-34 43-59 5-39 0-62 trace 2-13 28-29 7-54 0-62 10-91 1-23 13-55 12-03 0-05 3 : 47 6-36 26-24 5-64 1-04 0-03 0-03 50-95 2-11 0-78 6-02 7-07 10-20 trace 0-06 0-05 0-49 0-49 ; 49 0-02 3-24 0-03 trace 94-49 1-52 " 0-81 0-77 4-21 2-77 11-46 15-15 12-49 1-32 0-"6l 0-03 39-88 11-10 S(i(l;i Lime Mii"'iu'sjia Sesquioxide of iron Protoxide of manganese Silica Carbonic acid Phosphoric acid Sulphuric acid Chlorine Residue insoluble in acid Water 7-17 19-84 0-62 7-17 2-40 0-46 9-84 4-67 0-49 0-62 3-66 " 6-68 10-71 2-08 18-81 3 99 Charcoal In Liebig and Kopp's Jahrcsbericht, from which I have extracted these analyses, the sum of the con- stituents under each of the first five columns is given as exactly 100 -00, but the addition is incorrect. The proportion of water in the first and fifth analyses must surely be erroneous. The analyses were originally published in the Chemisch-Pharmaceutiscbes Centralblatt for 151. Unfortunately 1 have not been able to refer to tliis work and correct the errors. It is not in any of the following Libraries : Royal Society, Koyal Institution, College of Surgeons, British Museum, and not even in that of the Pharmaceutical Society '. 1 Rondinage de jeunes brins. Bois ! plied to wood cut in lengths, each of dc roni linage means wood sawn across in lengths, but not split into billets. -' 15 us dc i[(i;ulier. This term is ap- which is split into twjo or more billets. 3 Ann. du Chim., 3, s. 10, p. 150. 4 Comptes llendus, 24, p. 420. 70 COMPOSITION OF THE ASHES OF WOOD. These analyses were made by Witting, under the direction of Genth. Nos. 6-8 are analyses by the same chemist of the ground on which the trees grew. Nos. 1-3 are of the ashes of birch, and Nos. 4, 5 of the ashes of beech. No. 1 grew on ground No. 8 ; No. 2 on ground No. 6 ; No. 3 on ground No. 7. No. 4 grew on ground No. 6 ; and No. 5 on ground No. 7. No. 6 at Morschen, Hesse-Cassel, on the Muschel- kalk formation ; No. 7 at Marburg, in the same State, on Bunter- sandstone ; and No. 8 at Akureyri in America, on a formation of volcanic palagonite origin. 5 The presence of alumina is remarkable. Phosphoric acid is not men- tioned in the ground No. 8, and yet it must have been present, because it existed in the ashes of the tree which grew upon that ground. In No. 6 only part of the lime could have been combined with carbonic acid; and as to the state of combination of the other part, no evidence is advanced. These analyses are instructive as showing the influence of 'the nature of the ground upon the so-called inorganic constituents of trees. COMPOSITION OF THE ASHES OF WOOD. 1. 2. 3. 4. Potash 15-80 2-79 0-93 15-24 Soda 2 '76 15-99 14-59 7-27 Lime . 60-35 30-36 33 99 25-85 Magnesia 11-28 19-76 20-00 24-50 Oxide of manganese (Mn 3 O 4 ) Phosphate of sesquioxide of ironi (2F-O 3 3PO 5 ).... ./ 1-84 18-17 5-10 7-61 2-28 13-51 6-18 Sesouioxide of iron 7-73 Phosphate of lime (3CaO, PO 5 J Sulphate of lime 3-99 2-30 3 '31 5-05 2-91 Chloride of sodium o-^i 1-48 2-52 0-92 Silica 1*46 3. 04 5-27 3-60 Ash per cent, in the wood dried at) KXP C / 99-99 100-00 0-143 99-97 0-190 99-98 0-322 These analyses were made by Bottiiiger at Giessen, under the direc- tion of Will. No. 1. Fagus sylvatica, from Neufchatel, Switzerland. Nos. 2, 3. Firms sylvestris, from the vicinity of Giessen, not far from which are dolomite and mines of manganese. The ashes of Nos. 2, 3, 4 were brown-black, and evolved chlorine copiously when heated with hydrochloric acid. No. 2 was diseased. The ashes of No. 3 were ob- tained from a tree which had died. No. 4. Larix Europsea, from the same locality as Nos. 2, 3. In these analyses the calculation has been made after deduction of carbonic acid, charcoal (from imperfect incine- ration), and sand ; but, if proper precautions had been observed in the preparation of the ashes, it is difficult to understand how sand should be present in them. 6 5 Jahresbericht, Liebig u. Kopp, for 1S51, p. 70S. " 6 Aniialen der Chemie mid Pharniacic Wohler and Liebij* 50. p. 40(5. 1844. RAPIDITY OF GROWTH OF WOOD. 71 On the rapidity of growth of wood. On this subject there are many recorded observations, from which, on account of their importance to the metallurgist in some countries, I introduce a selection. Chevandier ascertained that on the western slope of the Vosges mountains and in the plain extending from their base, where the gres bigarre (lower trias) occurs, the average annual production of wood in forests of large beech-trees is about 9 steres (1 st. = 1 cubic metre = 317-849 cub. feet) per hectare (10,000 square metres, nearly 2J- acres). The average weight of dry wood annually produced in these forests amounts to 3650 kilogrammes (8047 Ibs. avoird.), which contain 1800 kil. (3968 Ibs.) of carbon and 26 (57 Ibs.) of hydrogen in excess of that required to form water with the oxygen of the w T ood. He has cal- culated that a bed of coal (containing 85 per cent, of carbon) cor- responding to the annual growth of these forests per hectare, would have an average thickness of O m 000165 (0-006496 inch). 7 The produce of coppice wood (taillis) is much influenced by -the nature of the ground; the more permeable and hygroscopic it is, the greater the produce. But in woods of large trees the geological influence of the ground disappears. The difference is explained by the fact that in coppices the ground is not so protected from the de- siccating action of the sun as in woods, so that it does not so long retain the necessary moisture, and the vegetative season is, conse- quently, shortened. Chevandier has shown that the influence of moisture is so considerable that silver firs (sapins) raised in boggy ground only grow at the rate of l kil> 80 (3-968 Ibs.) a year ; in dry ground at the rate of 3 WL 40 (7'496 Ibs.) a year; when watered by rain at gkn.2o (ig-078 Ibs.) a year ; and when watered by running streams at ll kil -60 (25-573 Ibs.) a year. The maximum of growth in different trees is stated by Chevandier to be as follows : oak at 77 years ; ash at 80 years ; silver fir (sapin), in very good ground, at 115 years ; in ground of medium quality at 76 years ; Scotch fir, in good ground, at 51 years; and in ground of medium quality 50 years. The difference in the last case is remark- ably small ! In the Black Forest in the Duchy of Baden trees of hornbeam pro- duced annually 2560 kil. (5644 Ibs.) of dry wood per hectare (2J acres), and trees of silver fir (sapin) 3903 kil. (8604 Ibs.). 8 These numbers include the wood obtained in thinning the forests, which is estimated at 15 per cent, in addition to the wood obtained in felling the trees. The actual data are recorded in the table on page 72. In Sweden and Norway the forests chiefly consist of Scotch fir ; spruce fir and birch occur in much less proportion, and even the silver fir is not common in several provinces. The annual yield varies con- siderably, ranging from 3 to 8 steres per hectare (105-950 to 282'533 c. f. per 2 acres). In those parts of Sweden which do not notably exceed 60 of latitude, 6 steres (211-900 c. f.) maybe taken as the average. ' Aim. de Chim., 3, s. 10, p. 156. s Chevandier. Comptes Jlemlus 24, p. 275. 422. 1847. 72 WEIGHT OF WOOD. FOKESTS OF THE GRAND DUCHY OP BADEN. 9 Kind of Wood. Mean Annual Growth per Hectare (2-4711, say 2 acres) in Cubic Metres (cub. feet 35-3166) in actual volume. Number of Cubic Metres (Steres) cor- responding in apparent volume. Dry Wood, correspond- ing in Kilogrammes and Pounds. Gneiss, granite, porphyry, j gres bigarre (lower trias), I Oak 5 221 j 1 84 388 cubic [ 7-57 f 2900-81 vieux calcaire jurassique, J rolled flints .... 1 feet, j { 6395 Ibs. Gneiss, granite, gres rouge, gres bigarre, vieux calcaire jurassique (lower Jurassic limestone), nouveau cal- !Ash (mountains j of medium al- I titude) | | 5-224 ) \ 184-594c.f. / 7-57 ( 2994-28 t 6601 Ibs. caire jurassique (upper ju rassic limestone), molass . . . Gneiss, granite, porphyry, j terrain de transition, nou-> (Ash(highmoun-j f 4-559 1 \ 161 008 c. f. J 6-61 ( 2574-62 ( 5676 Ibs. veau calcaire jurassique . . . ) Rolled flints (Hornbeam ) ( 4-008 ) 5-81 / 2226-04 Gneiss, granite, gres bigarre,) muscholkalk . f \ (charme) j Silver fir (sapin) \ 141 549 c. f. j / 8-304 \ \ 293-269 c f j 12-04 \ 4907 Ibs. / 3394-21 \ 7483 Ibs Granite, gres bigarre, muschel-1 kalk rolled flints . ) Scotch fir (pin) 7-330 1 ( 158-871 c f j 10-63 / 2798-71 \ 6170 Ibs In latitudes bordering on 60 the diameter of a pine (pin) 25 years old is only 11 centimetres (4-33 inches) : whereas on tolerable land in France and Germany the annual growth during this first stage of the life of the tree is half as much again. Wood intended for the smelting- works is cut at intervals, which vary with the locality, from 25 to 60 years ; and when it is to be converted into charcoal, it appears most advantageous to cut the trees at the age of 30 years. 1 Weight of wood. The following determinations have been made in Prussia of the weight of a cubic foot, in pounds (Prussian), 2 of various kinds of wood in different states of dryness : 3 Quite fresh. Half dry. Quite dry. Oak Beech Birch Hornbeam Scotch fir . .... 70 .... 65 .... 60 .... 62 .. 60 60 50 50 56 48 39 42 50 36-26 Practical directions for the cutting and storing of wood intended as fuel. The trees should be of mature growth, and should be felled when most 9 In this table no account is taken of the intermediate products resulting from thinnings (eclaircies), which would in- crease the numbers given above by about 15 per cent. 1 Sur 1'Exploitation des Mines et des Usines dans le Nord de 1'Europe. Par M. J. Dnrocher. Ann. des Mines, 5, s. 9, p. 356. 1856. 2 1 foot English = 11-65368 inches Prussian. 1 cubic do. - 1582-667 cubic inches do. 1 pound Prussian = 1 031236623 pound avoirdupois. 3 Bergwerksfreund, 3, 8. PEAT OR TURF. 73 free from sap. The wood should be cut to a suitable length, and should be so stacked as to allow the air freely to circulate through the pile, in order to promote desiccation as much as possible. If uncleaved in the direction of its length, it should be previously deprived of bark, either wholly or partially, by taking it off in strips all round. It should be protected from wet and rain ; if transported to a distance in rafts, it should not be left in the water longer than absolutely necessary, because experience has proved that its value as fuel deteriorates by long soaking. PEAT OR TURF. Peat is the product of the natural decay of various kinds of plants under special conditions of heat and moisture which occur in humid and temperate climates. Immense accumulations of peat constituting peat bogs exist in this country, especially Ireland. Peat is abundant in the northern and central parts of France, and in various other parts of Europe. It is found on mountainous declivities where it seldom exceeds four feet in thickness, and in low grounds where it may exceed even forty feet in thickness. It does not appear to exist within the tropics, yet the " Great Dismal " Swamp, in 36 8' lat., between Virginia and North Carolina, consists of black peat-like matter without any admixture of earthy particles, to the depth of fifteen feet. 4 The peat of Europe is chiefly derived from mosses belonging to the genus Sphagnum ; but that of South America contains no remains of mosses. In the sequel will be given the analysis of peat which my friend Dr. Falconer obtained from the bottom of a lake in Cashmere, and which is also free from the remains of mosses. It is certain that many peat bogs occupy the site of former forests ; and hence they occasion- ally contain imbedded trunks of trees, sometimes of very large dimen- sions. As peat is composed of the tangled remains of plants in different stages of decay, it may be more or less fibrous or earthy. It varies much in texture, but is generally so spongy as to retain a large quantity of water. The deeper the peat is taken from the bog, the more con- densed and changed is the vegetable structure, of which, however, suf- ficient always remains to be readily distinguishable. 5 It is composed of the same elements as woody tissue, with variable proportions of water and inorganic matter. This matter is in part derived from the inorganic constituents of the plants from which the peat has originated ; but it may occasionally be much increased by the presence of extraneous substances carried by streams or floods into the bog. XjH'cific gravity of peat. "It must obviously be subject to much varia- tion from several causes, such as difference in structure, degree of 4 Principles of Geology. Sir C. Lycll. 5 Dr. Bennett. Trans, of the Royal 1853. p. 724. Soc. of Edinburgh, 1854 v. 21. p. 183. 74 PEAT Oil TURF. desiccation, proportion of earthy matter, state of decomposition, and mode of preparation. According to Vogel, the specific gravity in some kinds is as low as 0*25, while in others which have not been compressed it ranges from 0-6 to O9. 6 COMPOSITION or PEAT. 1 1 Exclusive of Ash. O w d $ 1 g s i a > fl " No. Locality. C3 I 1 So I I I | 1 I . 1 S3 |f=| 1 2 Cappoge, Ireland Kilbeggan, do. . 51-05 61-04 6-85 6-67 39-55 30-46 2-55 1-83 26-37 24-50 52-38 7-03 62-18! 6-99 40-59 31-03 3 Kilhaha, do. 51-13 6-33 34-48 8-06 55-62 6-88 37-50 4 5 Phitlipstown, do. 0-405 Do. do. 0-669 58-69 60-48 6-97 6-10 32-88 32-55 1-45 0-88 1-99 3-30 6 Wood of Allen, do 335 59 92 6-61 32-21 1-26 2-74 );o-639 7 j to 61-02 5-77 32-40 0-81 7-90 : 0'672 8 Devonshire . . 0-85 54-02' 5-21 28-17 2-30 0-56 9-73 25-56 29-3059-74 5'77 31-32* 9 Abbeville, France 57-03 5-63 29-55 2-21 5-58 .. 60-40 5-96 33-64 10 Do. do. 58-09 5-93 31-37 .. 4-61 .. 60-89 6-21 32-90 , , TTmmrmt Hn 57 "'59 1 1 30 7*7 5-33 32-50 12 Thesy, do. '. ! 50-67! 5-76 :U-95 1-92 6-70 .. 36-95 54-31^ 6'17 39-52 13 Camon, do. . .. 46-11 5-99 35-87, 2-63 9-40 .. 38-1150-89 6-61 42-50 14 Cashmere . . : .. 33-28 3-66 21-03 1-81 ' 29-81 10-40! .. 55-66 6-12 38-22 * Exclusive of nitrogen and sulphur. 1 to 3. 7 Sir B. Kane states that it is very usual to find the turf of commerce containing one-fourth of its weight of water ; and that when it is dried in the air, under cover, it still retains one-tenth of its weight. 4. 8 Light surface peat, of a pale reddish-brown colour, containing small roots of Erica (heath) and leaves of grasses and Carex (sedges). Average thickness of the bog 18 feet; area 6582 acres. 5. Rather dense peat from the same bog as the last. Colour dark- reddish brown. Structure of moss still distinguishable, but species difficult to be determined. 6. Light surface peat, pale yellowish brown, moss very open- grained and fibrous. Principally composed of Sphagnum, Hypnum, etc., the species of which may readily be distinguished. 7. Lower layer of the same bog as the last. Compact and dense. Colour deep blackish-brown. Fracture earthy, appearing almost con- 6 Der Torf, seine Natur und Bedeu- tung. Eine Darstellung der Enstehung, Gewinnung, Verkohlung, Destination und Verwendung desselben als Brenn- material. Von Dr. August Vogel, Pro- fessor in Miinchen. Braunschweig, 1859, p. 24. This is a comprehensive treatise in 170 pages, well illustrated with wood- engravings : it might be advantageously translated into English. ' By Sir Robert Kane, ' The Industrial Resources of Ireland,' 1845, p. 37. 8 4-7. Report on the Nature and Pro- ducts of the process of the Destructive Distillation of Peat, considered specially with reference to its employment as a branch of manufacturing industry. By the Director of the Museum of Irish In- dustry (Sir R. Kane). 1851. COMPOSITION OF THE ASHES OF PEAT. 75 choidal, and exhibiting a resinous lustre when rubbed. All appearance of vegetable matter totally obliterated. 8. 9 Cut for fuel from the neighbourhood of the Military Prison, Princetown, near Tavistock, Dartmoor, Devonshire. Colour earth- brown; porous and largely intermingled with vegetable fibres of a light-brown colour, hard enough to be cut with a knife. When heated the gas gives a brilliant flame with muck smoke ; the residual coke is much shrunk in bulk, but retains the shape of the pieces employed ; it is without lustre and pulverulent, and is readily converted into a light bulky ash of a pure white colour, quite infusible before the blow- pipe. 9. 10 Was in a very advanced state of change ; it presented here and there some fragments of vegetables which had kept their form. Colour very dark brown ; powder, brown. 1 * 23 of ashes contained 90 of carbonate of lime and 0-33 of clay. The lime did not exist as carbonate in the peat, but was combined, Eegnault believes, with some organic acid, such as the so-called humic. The carbon, therefore, of the carbonic acid of the carbonate of lime is added to that obtained on the combustion. Before this correction the ash was 8 * 20 per cent. The peat was from Vulcaire, near Abbeville. 10. Similar to No. 4, from Long, near Abbeville. A similar cor- rection was made in this analysis for the carbonic acid in the ashes. 11. From Champ-du-Feu, near Framont (Vosges). This peat was in a little less advanced state of change than Nos. 4 and 5. 12. 1 Black peat of first quality. Previously to analysis, air dried during many months and in a dry vacuum during 24 hours. 13. First quality. Dried for analysis like No. 12. 14.* This peat was taken by my friend Dr. Hugh Falconer from the bottom of a lake in Cashmere. It contains the remains of the roots of aquatic plants free from those of mosses. Two analyses were made. Composition of the Ashes of Peat. A series of not less than 27 analyses of the ashes of peat from various localities in Ireland has been made in the laboratory of the Museum of Irish Industry, under the direction of Sir Eobert Kane ; and of these I insert the four following of the ashes of the varieties of peat, numbered 4, 5, 6, 7 in the preceding table of the composition of peat. 3 (See Table, p. 74.) It is stated that the quantity of carbonic acid found by experiment was nearly in every case much less than that required to saturate the lime, admitting the whole of the sulphuric acid to exist in sulphate of lime ; and it is inferred that the greater part of the carbonic acid, which is supposed to have been in combination with lime, was expelled during incineration. 4 According to Eegnault the lime in peat is certainly not present in the state of carbonate, but is in 9 ByVaux. Jo.Chem.Soc. vol.i. p.318. Mines. 5. s. 12. p. 406. ]857. 10 9-11. By Eegnault. Ann. des Mines, 2 By C. Tookey, in my laboratory. {. s . 12. p. 230. 1837. 3 Eeport ante cit. 1851. p. 72. 1 12, 13. By M. anchill against Messrs. Kussel and Son, 80 DIFFICULTY OF DEFINING COAL. Perhaps the nearest approach to a definition would be the following : Coal is a solid mineral substance, more or less easily combustible, varying in colour from dark brown to black, opaque except in ex- tremely thin slices brittle, not fusible without decomposition, not sensibly soluble in ether, benzole, chloroform, or turpentine not con- taining sufficient earthy matter to render it incapable of being applied with advantage as a source of heat in ordinary fire-places or in fur- naces. This would exclude amber and other resinous matters, bitu- men, and bituminous schists when they contain so large an amount of earthy matter as to be incapable of being employed as fuel ; and, I think, it would also exclude even the most compact varieties of peat, none of which can be properly considered as brittle. My friend Dr. Bennett has proposed a definition of coal founded on the appearance, under the microscope, of certain rings in a well-made transverse section ; and he maintains that by means of this appearance all kinds of coal, whether household or cannel, can at once be distin- guished from the Torbanehill mineral. 5 I do not, however, find that this distinction has been accepted by histologists. Dr. Bennett repu- diates the verdict of the Scotch jury. He writes, " at the trial it was very plausibly argued that, in a bargain between man and man, scien- tific truths were of no value, and that a whale among whalers was still a fish. But as no naturalist, conversant with the structure and func- tions of a whale, would for a moment suppose it to be a fish because it inhabits the water and resembles one, so, I contend, no histologist acquainted with the structure and properties of the Torbanehill mine- ral ought to maintain that it is coal because it is dug out of the earth and burns in the fire." Whatever difference of opinion there may be respecting the mode of accumulation of the great deposits of coal whether these deposits have, like peat, been produced in situ, or whether they have been derived from drift-wood there can be no doubt that coal, as has been previously stated, is the product of the decomposition of vegetable matter under special conditions. If we study the analyses of coals, we can readily select a series which illustrates the gradual passage of woody tissue into anthracite, or that variety of coal which consists al- most wholly of carbon. Such a series is presented in the following tabular form, the proportion of carbon being estimated at the constant amount of 100. 1. Wood (from mean of analyses 1 26) 2. Peat Carbon . 100 100 Hydrogen 12-18 9-85 Oxygen. 83-07 55 '67 3. Lignite, average of 15 varieties 4. Ten-Yard coal of the South Staffordshire basin 5. Steam coal from the Tyne 6. Pentrefelin coal of South Wales 7, Anthracite from Pennsylvania, U. S . 100 100 . 100 . 100 . 100 8-37 6-12 5-91 4-75 2-84 42-42 21-23 18-32 5-28 1-74 6 An investigation into the structure of the Torbanehill mineral and of va- rious kinds of coal, by John Hughes Bennett, M.D., F.R.S.E. Trans, of the Eoyal Soc. of Edinburgh, v. 21, part I, p. 173. 1854. DERIVATION OF COAL FROM WOODY TISSUE. 81 From this it will be perceived that through these various stages of conversion the proportion of carbon relatively increases while that of the hydrogen and oxygen decreases. If we examine the relation between the hydrogen and oxygen in these successive degrees of conversion, we shall find that, with the exception of anthracite, there is a steady increase in the proportion of hydrogen beyond that required to form water with the oxygen pre- sent. The following numbers express this gradual increase in the excess of hydrogen. Hydrogen. Hydrogen. 1 1-80 2 2-89 3 3-07 4. .. 3-47 5 3-62 6 4-09 7. 2-63 The formation of these products of the decomposition of woody tissue may be explained by the elimination of hydrogen in combina- tion with carbon as marsh-gas, of oxygen in combination with carbon as carbonic acid, and of hydrogen in combination with oxygen as water. 6 That the derivation of coal from woody tissue has been accompanied with the evolution of marsh-gas might be inferred from the nature of the gases generally occurring in coal-mines, of which that gas is the chief constituent. However, the hydrogen and oxygen are never completely separated, as these elements are always present in sensible proportion in anthracite, which is furthest removed from woody tissue in composition. Nitrogen is always found in coal, and its proportion seems to range pretty constantly between 1 and 2 jper cent. Although nitrogen does not appear to be an essential element of woody tissue, yet that tissue is always associated with nitrogenous matters, so that the source of nitrogen in coal is readily accounted for. Hence ammonia is gene- rated in the dry distillation of coal, and the product is, consequently, alkaline. In the dry distillation of woody tissue, on the contrary, the product is acid, owing to the presence of free acetic acid. Sulphur is always present in coal. It may exist in the state of sulphuric acid in combination with a base ; in combination with iron as iron pyrites or bisulphide of iron ; and, probably, also in combina- tion with the organic elements of coal as it exists in albumen. It chiefly occurs in the state of iron pyrites. In the dry distillation of coal a portion of it is evolved partly as sulphuretted hydrogen and partly as bisulphide of carbon. All coal contains a greater or less proportion of water, which may be expelled at or slightly above 100 C. A coal may appear perfectly dry and yet lose by desiccation a very large amount of water ; as much even as 20 per cent. Whether the water exists wholly as hygroscopic water, or whether, in some cases, it may not be partially combined, is doubtful ; yet facts will be advanced in the sequel which seem to support the latter view. The amount of water which may re- 6 Vid. Organic Chemistry in its ap- plications to Agriculture and Physi- ology. Liebig. Trans., 1840, 327. Also, Lchrlmch dcr Chemisch. u. Physikalisch- Geologie. Bischof, 2. p. 1780. 82 CONSTITUENTS OP COAL. main in some varieties of coal, even after desiccation by free exposure to the air, deserves more attention than it generally receives from me- tallurgists. All coal contains a sensible amount of inorganic matter, of which the chief constituents are silica, alumina, lime, and iron. It is this matter which appears as the ashes of coals. Now, admitting that alumina is not an ordinary constituent of plants, it follows that the matter of which the ashes are composed has been derived, not merely from the inorganic elements originally existing in the plants from which coal has been produced, but also from an extraneous source. When we consider the conditions under which coal must have been formed, we can be at no loss to understand how such extraneous matter should have been deposited in a coal-field by the agency of water. We have only to inspect a coal-seam to have ocular and positive proof of the interstratification of coal and earthy matter such as shale and sandstone. In illustration we may take the familiar but extreme case of blackband ironstone. This ironstone consists essentially of carbonate of iron intimately mixed with coal. If further proof were needed of the admixture of coal with such ex- traneous matter, it is furnished conclusively by the results of Taylor, who has analysed the matter forming the bed and the roof of a coal- seam at Newcastle (Buddie's Hartley and Blaydon Burn Colliery), as well as the ashes of the intervening coal. The analyses are as follow : 7 1. Silica 62-44 .. 2. 59-56 3. 4. 64-21 56-51 . 5. 58-99 Alumina . ... 31' 22 .. 12-19 28-78 . ... 31 '89 . 26-19 Seso uioxicte of iron 2 '26 15-96 2-27 5-14 Protoxide of iron . 7 04 . . 5-11 Lime 0*75 9-99 1-34 1-69 0-67 Melanesia * 85 . . 1-13 1-12 .. 0-85 .... 1-54 Potass 2-48 . 1-17 2-28 1-38 2-34 Soda .. . 0-61 . 100-00 100-00 100-00 99-97 99-98 1. Fire-clay, on which the coal rests, after subtraction of 10*5 per cent, of water and 0*44 per cent, of chloride of sodium and sulphate of soda. 2. Ashes of good coal (1'36 per cent.), after subtraction of 8*2 per cent, of sulphuric acid. The source of the sulphuric acid will be explained in the sequel. 3. Ashes of the coarse coal (16 '9 per cent.), after subtraction of the sulphuric acid. 4. Bituminous shale, after subtraction of 39-35 per cent, of organic matter. 5. Bluish shale, after subtraction of 11 per cent, of water. Ashes of coal In the choice of coal, attention should be directed to the nature as well as the proportion of the ashes which it may yield. The proportion may be so great as to lessen its value in a very mate- rial degree, or even to render it useless as a fuel for many purposes. It may be determined by incineration at a red heat in a platinum or porcelain crucible. The process may be conveniently conducted ? Biscliof, Lehrbuch der Chemiscli. u. I New Philos. Journ., 1851, v. 50, p. 140. Physikal. Geologie, v. 2, p. 1771. Edinb. | ASHES OF COAL. 83 either over an air-gas flame or in a muffle. In respect to the nature of the ashes, the amount of oxide of iron is a point of chief importance. Iron pyrites is present in all coal, in which it may occur either invi- sibly diffused through the mass in visible laminaB or in layers and nodules, sometimes of considerable size. The spontaneous com- bustion of coal in coal-mines appears to be due, in part at least, if not wholly, to the heat developed by the oxidation of intermixed pyrites by atmospheric air. When some kinds of coal are left exposed to the air, yellow spots will eventually appear here and there on the surface, which consist of basic sulphate of sesquioxide of iron (misy) derived from the oxidation of pyrites. At first sulphate of protoxide is formed, of which the quantity is occasionally so great as to cause complete disintegration of the coal. The water which flows from coal-mines frequently contains in solution sufficient sulphate of iron to communi- cate a strong chalybeate taste. Such a solution rapidly corrodes iron boilers, and by free exposure to the atmosphere will deposit a yellow- brown ochreous compound of basic sulphate of sesquioxide of iron. Shale, which consists essentially of silicate of alumina, is generally present in coal in a greater or less degree, as will appear from the analyses of the ashes of coal, and when it is in contact with iron pyrites in process of oxidation by exposure to the air it may give rise to the formation of iron-alum in white asbestiform fibres. The formula of this beautiful substance is FeO, S0 3 +A1 2 3 , 3SO* + 24H0. 8 The iron pyrites in coal will be represented in the ashes by an equi- valent amount of sesquioxide of iron, provided incineration be perfect ; and generally by far the greater portion of the oxide of iron in the ashes is derived from pyrites. The proportion of that oxide may be sufficient to give a red colour to the ashes, or it may not be sufficient to cause sensible coloration ; hence the distinction between red and white ash coals. The intensity of the red colour, taken in connexion with the amount of ashes in coal, may serve as an indication of the proportion of sulphur existing in the state of pyrites. When a coal contains pyrites in considerable quantity, especially in the state of nodules, it may speedily destroy the fire-bars of a furnace. In one instance in which a highly pyritic coal was employed in a reverbera- tory furnace I observed ferruginous masses depending like stalactites from the lower surface of the bars, which were rapidly corroded in consequence. When the ashes sinter together or melt, they form a more or less vitrified mass termed clinker, which may accumulate upon the bars so as greatly to impede the passage of air between them, and, as a consequence, lower the temperature of the furnace. Moreover, when the bars are covered with a firmly-adherent bed of clinker, and are no longer cooled by a continuous current of cold air, they may soon become heated and destroj^ed by oxidation. Now the presence of much oxide of iron in the ashes may, in a certain degree, increase their fusibility, and so tend to induce the evil arising from the forma- 8 Handworterbuch des Chemisch. ralogy, etc. Thomson, 1, p. 472. 1836. Tlieils der Mineralogie, Kammelsberg, Bertbicr, Ann. d. Mines, 5, p. 259. 1820. purt 1, p. 10, 1841. Outlines of Mine- G 2 84 ANALYSIS OF COAL. tion of an impervious bed of clinker. In this case a greater ex- penditure of manual labour would be required in removing clinkers, or, as it is termed, in dinkering the grate. But sometimes a bed of clinker is expressly formed, and made to serve an important purpose. Thus in furnaces in South Wales such a bed is ingeniously used as a substitute for a grate, on which small and inferior coal may be con- sumed which could not be otherwise profitably applied. The bed is allowed to become 12 or 20 inches thick, and requires to be supported only by a few bars, which, being far removed from the fuel, remain comparatively cool. It is kept broken up, so that a sufficient quantity of air may traverse it ; and, in proportion as it accumulates above, it is cut away underneath by means of a crowbar with a cutting-edge and a heavy hammer. This operation is termed breaking the grate. However, as a rule, the less a coal clinkers the better for most purposes, and for none more so than boiler -furnaces. Irrespective of the formation of clinker, a certain amount of inorganic matter in coal is sometimes beneficial in preventing the fire from too rapidly collapsing, as it were, in the furnace. On this account a kind of coal called bmsils, which occurs in the middle of the ten-yard coal in South Staffordshire, is preferred for reverberatory furnaces by some smelters in Birmingham. Errors in analysis of coal. When coal contains much inorganic matter, especially iron pyrites, the usual method of calculating its composi- tion from the data obtained in the process of organic analysis may be erroneous in a sensible degree. The ashes left by incineration are estimated as inorganic matter, and the proportion of oxygen is found by subtracting the sum of the carbon, hydrogen, nitrogen, and ashes from the amount of dry coal subjected to analysis. By incineration the iron of the pyrites is converted into sesquioxide, and the sulphur, in a greater or less degree, into sulphuric acid, which may remain in combination with any base in the ashes such as lime capable of forming a sulphate not decomposable at a red-heat. Supposing the whole of the sulphuric acid to be thus retained in the ashes, for 1 part of iron pyrites there would be an increase of 1 due to oxygen derived from the air during incineration. The whole amount of this error, provided no correction be made, would fall upon the oxygen. It is not asserted that the whole of the sulphur is actually converted into sulphuric acid and retained in the ashes, but that a considerable portion of a stable sulphate may be produced during incineration will appear from analyses of coal in the sequel. It is certain that the alu- mina in the ashes must, in great measure, exist in combination with silica as clay ; but clay holds water in combination which cannot be expelled except at a temperature far more than sufficient to decompose coal. Hence, during the process of organic analysis water may be evolved from the clay present in coal, and so occasion an error of excess in the determination of the hydrogen. This source of error has been pointed out by Eegnault. 9 Carbonate of lime is sometimes pre- sent in coal in very appreciable quantity, in which case carbonic acid would be evolved during the analysis, and so an error of excess would Ann. des Mines, 3 s. 12, 167. LIGNITES. 85 be caused in the determination of the carbon. M. de Marsilly has observed that, however pure a piece of coal may be, and however homogeneous it may appear to the eye, its different parts do not yield the same proportion of fixed residue by incineration ; and the same is true in respect to the proportion of coke obtained by the calcination of different fragments of the same lump of coal. Hence, in every case, the proportion of ash and coke should be determined by operating upon an average sample taken from the powder of the coal. 1 LIGNITES. The synonyms are brown coal, bituminous wood, and pitchcoal of Werner (Pechkohle) . Geologists apply the term lignite only to those carbonaceous minerals which occur in deposits of later date than the true coal-measures ; and yet these minerals may occa- sionally not be distinguishable, either in physical characters or che- mical composition, from undoubted bituminous coal. However, the cases in which lignites thus resemble bituminous coal are exceptional. Lignites seem generally to differ from other kinds of coal in the large proportion of water which they retain, even after free exposure to the atmosphere at the ordinary temperature. They appear to pass by insen- sible gradations into bituminous coal ; and at present it seems impossible to construct a definition, except on geological grounds, which shall be characteristic of lignites alone, to the exclusion of all other kinds of coal. For the sake of convenience some varieties of coal which occur in formations posterior to the coal-measures, and which approximate closely to, if they are not identical with, certain kinds of bituminous coal, will be classed among the lignites proper. Moreover an illustra- tion will thus be afforded that true coal, so far as relates to composition and external characters, is not confined to the coal-measures. Lignite is either wood-like in structure, earthy, or compact ; frac- ture conchoidal, wood-like, or uneven; colour of various shades of brown, brown-black, or black ; lustre dull, shining or fatty ; spe- cific gravity generally ranges from 1-2 to 1*4; in composition it approaches nearer to wood than bituminous coal, and is specially distinguished from other kinds of coal in retaining a large proportion from 15 to 20 per cent, or more of water, even after desiccation by long and free exposure to the atmosphere at the ordinary tempera- ture. Lignites are generally non-caking ; that is, the powder of lignite when heated to redness in a close vessel does not yield a coherent coke. The term lignite is restricted by some /persons to such varieties as manifestly present the appearance of woody tissue, the term brown coal being applied to all other varieties. Brown coal is the term at present used by the Germans as including all varieties, whether woody, earthy, or compact. Classification of lignites according to external characters : a. Pitchcoal (Pechkohle, Glanzkohle). Compact, occasionally cleav- ing into prismatic pieces ; fracture conchoidal ; pitch black ; lustre waxy or fatty. b. Common brown coal. Compact ; generally with slaty cleavage : Comptes Rcndus, 1848, 46, 882. 86 CLASSIFICATION AND COMPOSITION OF LIGNITES. wood-like structure indistinct; smooth conchoidal fracture; black- brown to pitch-black ; lustre more or less glistening, or slightly fatty ; of various shades of brown ; lustre dull or glistening. c. Woody brown coal (Bituminous wood). Massive, presenting the form and structure of wood. d. Schistose coal Distinct slaty cleavage, sometimes separating into very thin leaves as in paper coal (Dysodil) ; occasionally sensibly elastic ; brown. e. Earthy coal Compact, but easily rubbed to powder; fracture earthy ; dull ; of various shades of brown. 2 Composition of lignites. We are indebted to Kegnault for an excellent memoir on combustible minerals, published in 1837, from which the following instructive table is extracted. In this table is presented a series of analyses of lignites of tertiary origin, classified according to the degree in which their conversion from vegetable tissue has advanced. TERTIARY LIGNITES. 1 Composition Composition. ; after deduction ^ Specific of Ashes. | Character of the Lignite. Locality. Nature of the Coke. Gra- vity. ! . i i .i g . . Carbon a fit 1 | i K I'll i 1. Perfect lignite 1. Dax, south of France Pulverulent 1*272 70*49 5-59 18-93 4-9974-19 5-88 20-13 49-1 2. Bouches-du-RhOne do 1-254 63-884-58 18-11 13-43 73-795-2920-92 41-1 3. Hesse-Cassel do 1-351 71-71 4-85 21-67 1-77 73-004-93 22-07 48-5 4. Basses Aipes do 1-276 70-025-2027-77 3-0172-19 5-3622-4549-5 2. Imperfect lignite 5. Greece .... . -,\\ 1-185 6i-20'5-0024-78 9'02 67-28 5-4927-23 (LiKewood^ riQQ 63.29 4-98 26-24 5'49 66'96 5'27 27'77 38-9 36-1 6. Cologne 7. Usnach (fossil wood) I c Jai J 1-167 56-04<5-70 36-07 2-19 57-29 5-8336-88 3. Lignite passing ) into bitumen... J 8. Elbogen, Bohemia . I Puffed out } < (bom- }. ( souffle") f 1-157 73-79 7-46 13-79 4-9677-64 7-85 14-51 27-4 9. Cuba do 1-197 75-85 7-25 12-96 3-94 ] 78-96 7-55 13-4939-0 Asphaltnm 10. Mexico . , 1-063 79-18^-30 8-72 2-80 8T46 9-57 8-97 9-0 1. Fine black; powder brown; fracture uneven; but slight lustre; no ligneous texture ; the particles keep their form and do not stick together or cake when heated. 2. Occurs in limestone ; very schis- tose ; pure black ; powder brown ; very brilliant ; the texture of wood can no longer be recognised, except in less changed portions, which are brown ; particles keep their form when heated, and do not cake ; burns with a very brilliant and smoky flame. 3. Occurs in clay resting on Muschelkalk. The specimen analysed was extremely bril- liant; fracture conchoidal; it resembles the finest jet, but is more tender; powder black-brown; the particles when heated stick a little together, but without fritting, i. e. forming a sensibly aggluti- nated mass. 4. Occurs in limestone ; black ; powder clear brown ; lustre fatty ; compact ; coke slightly swelled out ; it may be used in a smith's fire. 5. This lignite is worked on the banks of the river 2 Vide Naumann, Elemcnte der Mine- ralogie, 18-16, p. 426. Also Classification u. Bescln-eibung der Felsarten von F. Senft, 1857, p. 403. COMPOSITION OF LIGNITES. 87 Alpheus, in Elis, Greece. Foliated ; laminae thick ; dull black ; powder brown; presents much indication of vegetable structure, which in some pieces is perfectly preserved ; the particles keep their form when heated. The portion analysed was treated with hydrochlo- ric acid in order to remove carbonate of lime, of which it contained a large quantity. 6. Occurs in thick beds in sand and clay on the banks of the Rhine from Cologne to Bonn. Reddish brown ; powder brown red ; friable; ligneous texture very decided ; gives a coke like wood charcoal. 7. Occurs at Usnach, on the borders of the Lake of Zurich, Switzerland. Brown, almost black ; powder clear brown ; veiy hard and cannot be cut, yet with much trouble it can be pounded in a mortar ; woody texture, still perfect. 8. Occurs as a thick bed in clay, and is used for firing porcelain. Blackish-brown; powder reddish-brown ; compact ; homogeneous like jet ; fracture conchoidal, dull ; yields a very light brilliant coke semi-metallic in lustre ; con- tains 1-77 per cent, of nitrogen. 9. Velvety black colour; lustre veiy fatty ; easily melts when heated, and leaves a very light puffed out coke. Regnault regards this mineral as forming the passage from lignites to bitumens or asphaltum. Geological position not certain, but presumed to be tertiary. 10. Black; powder black; very bril- liant ; emits a very strong and disagreeable odour ; melts below 100 C. Its geological position is not known. COMPOSITION OF LIGNITES DRIED AT 100 C. OR UPWARDS. Exclusive of Sulphur* . . and Ash. 1 Locality. ft | 1? a 1 | ^ 'I ^ Oxygen, inclusive 1 CO C5 1 1 I 1 1 <1 ts 1 Carbon. Hydrogen. of nitrogen. 1 Bovey, Devonshire 1-129 66-31 5-63 22-860-572-36 2-27 34'66 30'79 67-85 5-75 '23-39f 2 Oedenburg, Hungary 1-285 . . 0-9li 2-39 18-60 70-84 4-71 24-44 3 Do. do 1-334 1-63 4-6417-10 71-36 5-09 23-54 4 Bodoncspatak, do. 1-327 . . . 4-27 3-30,10-84 .. 59-88 4-55 35-56 5 Palojra do 1-256 2-59 1-4111-07! .. 70-40 5-73 23-87 6 /^inle.Comorn.do. 1-3*7 , , * tt 0-57 4-3512-6059-55 71-89 4-79 23-31 7 Wildshut, Upper} Austria J 1-306 53-79 4-26 25-39 0-98 15-58 26- 15 54-7 64-46 5-10 30'44 8 Thai lern, Austria ... 1-413 49-583-84 22-68 tt 4-5619-34 22-5363-7 65-15 5-05 29*80 9 Gloggnitz, do. 1-364 57-71 4-4922-14 '.'. 3-12 12-54 25-15 [ 54-4 68-42 5-33 26-25 10 Schonfeld, Bohemia 61-205-17 21-28 .. J12-3521-2 >< 69-82 5-90 24*28 11 Do. do. 8'65 70*80 5 "81 23*39 12 Meissen , Saxony ... 58-905-36 21-63 6-61 7-50 .. \\ C8-58 6-24 25-18 13 Do. do 62*185-47 1ft -05 . . a-sn R-ool .. 72'56 6*38 21-06 14 Riestedt, Prussia ... 1-218 61-13 5-09 31-95 1-8331-66 02-27 5-18 32-55 15 Loderburg, do 1-219 55-30 4-90 31-95 . 1 7-8549-50 60-01 5-31 34-68 16 Teuditz, do 1-263 54-025-2827-90 12-8048-60 61-95 6-06 31-99 17 Do. do 49-915-2032-42 12-47 .. \\ 57-02 5-94 37-04 18 Brumby, do 1-263 47-78;4*28;18-42 .. '. J<-52'40'60 67-79 6-07 26-14 19 Frankfort-on-the ) Oder, do. J 59-65 4-86 26-41 .. 9-0816-07 65-61 5-34 29-05 20 Wittenberg, do 64-075-0327-55 .. 3'35'l7-26 66-29 5-20 28.51 21 Zscherben 64-265-7617-44 .. 12-5445-37 73-47 6-59 19-94 22 Tiflis, Georgia 63-345-67 27-93 1 .. 3-04 t< 65-34 5-85 28-81 23 Irkutsk, Siberia 47-40 4-56 33-02J .. 14-95 55-81 5-36 38-83 24 Laubach, Hesse- ) Darmstadt } .. 57-28 6-0336-lo| .. 0-59 .. 57-62 6-07 36-31 25 Near Cassel, Hesse- ) Cassel 5 .. 62-60 5-02 26-52 .. 5-86 .. .. 66-49 5-33 28-28 26 Sipplingen, Lake ) Constance j .. .. 5-50 .. C4-96J 3-48 31-56 27 Island of Sardinia... 59 984-7529-42 . 5-85 50-00 63'7li 5-05 31-24 28 29 /Brit. N. America, "j \ Prairies, east of, I KockyMountains' .. 56-50 .. 50-60 3-65 3-24 18-91 8( 14-410-90 0-60 0-42 5-62 15-93 13-92 *.. 14-50 .. 70-75 73-17 4-57 4-68 24-68 22-15 1 * Kxcept when the sulphur is not given in the eighth column. f Exclusive of nitrogen. 88 COMPOSITION OF LIGNITES. 1. Brown; structure fibrous and lamellar; becomes rotten by im- mersion in water ; does not soil the fingers ; coke has a semi-metallic lustre ; does not swell, and cakes but slightly ; ash bulky and red ; copper and lead were detected in this coal ; it evolves an extremely offensive odour on burning. 3 2. Finely fibrous brown coal. 3. Not fibrous. 4. Pitch black ; powder brown ; lustre of fresh surface often vitreous ; structure here and there wood-like ; breaks in rhombic pieces. 5. Black brown to light brown ; powder brown ; woody structure dis- tinct ; hard and difficult to pulverise ; it contains a peculiar resin. 6. Black; powder brown; lustre imperfectly fatty; fracture uneven, schistose, often conchoidal or rhombic ; no trace of vegetable structure ; resists exposure to the air. 4 7. Wood-like ; the coke was obtained by slow coking ; by rapid coking the yield was from 2 to 3 per cent. less. The dry coal absorbed from the atmosphere 10-8 per cent, of water in 24 hours, that is, only 7 -3 per cent, less than the total amount expelled by desiccation at 100 C. 5 8. Black-brown ; wood-like (brown coal). The dry coal absorbed from the atmosphere 1 2'7 per cent, of water in 24 hours. 6 9. Wood-like ; much fissured (brown coal). The dry coal absorbed from the atmosphere 15*9 per cent, of water in 24 hours. 7 10. Brown coal. 8 11. Dark black-brown (brown coal). 9 12. Brown coal. 1 13. "Black coal" (Schwarz-kohle). Black variety of brown coal. 2 14. 15. 16. 18. Brown coal from the Prussian province of Saxony; 14. "Fossil wood," i. e. presenting wood-like structure; 15. Earthy; 16. and 18. Earthy. The specific gravity and water were determined with the coal fresh from the workings. Colour of ash 14. reddish- white ; 15. yellow-brown; 16. greyish- white ; 18. greyish white. 3 17. Earthy; from the same locality as 16, but the analysis is by another operator. 4 19. 20. and 21. Brown coal. 5 22. 23. Brown coal. 6 24. Wood-like. 25. Brown coal. 7 26. Brown coal. 8 27. Occurs, rather more than half a mile from the sea, at Goneza, province of Iglesias, to the west of Cagliari. An analysis was made at Turin ty Abbene and Rossi, and the mineral was also examined at the Ecole des Mines, Paris, and described as black coal, schistose and pyritic, yield- ing a pulverulent coke and very ferruginous ashes. Sir Eoderick Murchison informs me that this mineral probably belongs to the true coal-measures, in which case it presents an interesting illustration of 3 1. By F. Vaux, Journ. of Chem. Soc. London, v. 1. 318. 1849. 4 2 to 6. By Nendtwich, Chemisch- techn. Unters. d. vorziiglicheren Stein- kohlen-Lager Ungarns. Von Prof. Nendtwich, (aus dem October-Hefte des Jahrganges 1851 der Sitzungsberichte der matli-naturw. Classe der kaiserl. Akademie der Wissenschaften besonders abgedruckt.) 5 7. By Schrotter, Liebig u. Kopp, Jaliresb. 1849, p. 708 ; Kenngott, 1853, p. 149. 6 8. Id. 7 9. Id. 8 10. By Baer, Kenngott, 1853, p. 151. 9 11. By Kottig, Kenngott, 1852, p. 258. 1 12. By Gragcr, Berzelius, Jahresber. 1848, p. 261. 2 13. Id. 3 14 to 16 and 18. By F. Bischof, ma- nager of the salt-boiling works at Dtir- renberg (Ober-siedemeister). Berg. u. H. Zeit. 1850, p. 69, quoted from Bergwerksfr. vr. 13. 4 17. By Wagner, id., quoted from Polyt. Centralbl. 1847, p. 1496. 5 19 to 21. By Baer, Liebig u. Kopp, Jahresb., 1852, p. 733. 6 22, 23. By Woskressensky, Kenngott, 1852, p. 256. 7 24, 25. By J. von Liebig, Kenngott, 1852, p. 257. 8 26. By L. Gmelin, Kenngott, 1856, p. 117. COMPOSITION OF LIGNITES. 89 the fact that a coal of the coal-measures may remarkably resemble a true lignite in respect to composition. 9 28, 29. l Brought by Dr. Hector. 28. Dark-brown, compact, in part wood-like and in part resembling coal of the coal-measures ; fracture more or less conchoidal. 29. Cracked in small pieces during desiccation by exposure to the air. Much resembling coal of the coal-measures in appearance. Mr. G. P. Wall, formerly a student of the School of Mines, in the course of a recent official survey of the geology of the island of Trinidad collected specimens of lignites which present many points of interest. They have been analysed in my laboratory by C. Tookey. The combustion was effected in a current of oxygen. The tempera- ture at which desiccation was effected ranged from 100 to 110 C. The results are as follow : Composition inclusive of Hygroscopic Water. Hygroscopic water 1. 20*50 2. 5-90 3. 16*80 4. 17-65 Carbon ... 60-13 ... ... 69-53 . 57-38 56-19 Hydrogen .. ... 4-14 . . 5-36 3-74 4-14 Oxygen and nitrogen 10-77 15-22 17-50 17-39 Sulphur ... 2-36 ... ... 0-55 . 0-68 2-23 Ash.. , 2-10 .. . 3-44 , 3-90 2-40 Composition exclusive of Hygroscopic Water. 1. 2. 3. 4. Carbon .. .. 75'63 .. .. 73'11 , 71-58 . 68-23 Hydrogen 5*20 5'63 Oxygen and nitrogen 13 57 17 ' 08 Sulphur 2-96 0'57 Ash 2-64 3-61 4-66 18-09 0-84 4-86 5-02 21-14 2-70 2-91 No. 1. Black; fracture dull; powder brown; does not cake when heated in a close vessel ; yields 43' 1 5 per cent, of a non-coherent coke. L Congestion J Hydrogen per cent. { } Carton per cent. { %* No. 2. Black ; bright, like good bituminous coal ; friable ; fracture uneven ; powder dark brown. This variety scintillates much when held in a flame; when heated it evolves an odour like petroleum; cakes, and yields 54 per cent, of a firm, coherent coke ; ash red. 2. G m i[on } Hydrogen per cent. { *\ 5 } Carbon per cent. { '.* No. 3. Black ; compact ; fracture conchoidal and smooth ; powder brown; does not cake; yields 51'S per cent, of a non-coherent coke. 1. Comustion } Hydrogen per cent . { 01 } Carbon per cent. No. 4. Black ; compact ; fracture uneven and dull ; powder brown ; does not cake ; yields 44-95 per cent, of a non-coherent coke. 1. Combustion J Hydrogen per cent. { .' } Carbon per cent. { ' 9 27. Ann. des Mines, 4. s. 20. p. 680. Tertiary (?). Specimen taken Aug. 1857 ; 1851. analysis made June 1861. 29. From 6 ft. 1 28,29. By C. Tookey, in my 1-abora- seam, right bank of Saskatchewan, at Fort tory. 28. Saskatchewan Plains, La Roche I Edmonton, Lat. 53 J 33' N., Long. "" Perce'e, Lat. 49 ^ 7' N., Long. 115 W. | 20' W. Lower cretaceous (V). 90 COMPOSITION OF LIGNITES. On inspecting the preceding analyses of lignites it will be observed that, with the exception of No. 2 of the Trinidad specimens, all con- tain a large proportion of water, a proportion far exceeding that which is found in any coals of the coal-measures. No. 2 presents an example of a lignite which, on the ground of physical character and chemical com- position, should be classed as a bituminous coal of the coal-measures. Lignites from Auckland (New Zealand) and Tasmania have also been analysed in my laboratory by 0. Tookey. The results are as follow : 1. 2. Hygroscopic water 14-12 13-43 Carbon 55-57 59'90 Hydrogen 4 13 4-66 Oxygen 15'67 15-99 Nitrogen 1-15 1-08 Sulphur 0-36 0-30 Ash 9-00 4-64 No. 1. From Auckland; black; lustre dull; fracture uneven, more or less conchoidal ; distinct cleavage ; shining, more or less trans- parent; brown resin occurs diffused through this lignite in pieces varying in size from a pea to considerable masses. Two combustions were made. In the first the hydrogen was 4-07 and the carbon 55-65 per cent., and in the second the hydrogen was 4' 18 and the carbon 55-48 per cent. No. 2. From Tasmania. The specimen was sent by Governor Denison. In physical characters it was similar to the last described, and it also contained resin diffused in like manner through its substance. Accompanying the specimen of lignite was a piece of resin as large as the fist, which was more opaque, and less resembling ordinary varieties of amber in appearance, than that of No. 1. By the action of benzole a portion only dissolves ; a gum-like insoluble mass is left, which retains the form and bulk of the original resin. Composition of the ashes of lignite. The following analyses by Kremers will suffice for illustration : 2 1. 2. 3. 4. Silica 15-48 45-13 GO -23 31-30 Sesquioxide of iron Alumina 74-02 5-28 25-83 22-47 6-36 31-63 54-47 8-31 Lime 2-26 2-80 1-08 3-44 Magnesia 0-26 0-52 0-35 1-60 Potash 0-53 0-60 11 0-07 Soda 0-28 0-29 Sulphuric acid 2-17 2-37 0-24 0-52 101-53 101-60 98-37 98-95 Ash per cent 1-99 1-89 1-74 11-18 1. From Oberndorf. 2 and 3. From Zwickau, Saxony, burg, Silesia. 4. Walden- 2 De relatione inter carbones fuscos atque nigros. Dissert. Inaug. Auctor Petrus Kremers. Berolini. 1851. BITUMINOUS COALS. 91 JDaubree has detected arsenic in the lignite of the tertiary strata at Lobsann, Lower Khine. Ordinary specimens of this fuel contained from 0-002 to 0-0008 of their weight of arsenic. 3 BITUMINOUS COALS. This term is commonly applied to coals from the coal-measures which, under ordinary conditions, burn with a more or less smoky flame. In respect to the degree in which they are removed from woody tissue in chemical composition, they occupy a position between lignites on the one hand and anthracites on the other. The passage of lignite into bituminous coal is as gradual as the passage of bituminous coal into anthracite, so that there is no precise line of demarcation between coals of this class and those of the other two classes. Hence it may well be conceived that in the class of bitu- minous coals many varieties of coal must be included which in ex- ternal characters and composition present considerable diversity. The mineral bitumen burns with a smoky flame, and on this account coals which burn in a similar manner have been characterised as bituminous. This application of the term has occasionally led to the erroneous notion that bituminous coals necessarily contain a substance similar to bitumen. Natural bitumen readily dissolves in solvents such as ether or benzole, but no sensible amount of matter can be extracted from ordinary bituminous coals by these solvents. By some writers the term, bituminous is used to denote the matter which is volatilised when a coal is heated, at least to redness, in a close vessel, in which sense it is synonymous with volatile matter, both terms being employed indis- criminately. By other writers it has been used to express the so-called organic elements of coal other than carbon, namely, hydrogen, oxygen, and nitrogen. The characters of bituminous coals may be summed up as follows : solid, brittle, opaque ; lustre dull, shining, or fatty ; colour black or brown-black ; colour of the fine powder brown or brown-black ; some soil the fingers, and others do not ; hardness variable ; fracture even, conchoidal, or uneven; they frequently break into pieces more or less cubical or rhombic ; they present no definite signs of crystal- lization ; they consist of carbon, hydrogen, oxygen, nitrogen, and sulphur, and of fixed or inorganic matter ; they burn with a more or less smoky flame, and when heated in a close vessel leave a solid car- bonaceous residuum termed coke, which contains the fixed inorganic matter or ashes. Caking coal. -When some coals of this class are heated to a certain degree they swell, become pasty, and more or less fused, emitting bubbles of gas, which bum with a bright flame as they escape. If the powder of such coals is thus heated to the pasty state the particles stick together and form a coherent mass, or, in technical language, the coal is said to cake. When coal in the pasty state is taken out of a fire it remains for a short time soft and dough-like, but on cooling it becomes solid and brittle. In the quality of caking there may be Aim. (1. Minus, 5. s. 14. p. 472. 1858. 92 CAKING COAL. every degree, from slight fritting or sintering to almost complete fusion. Caking does not occur at a temperature below that at which the coal suffers decomposition, so that it is not due to the sold- ering of the particles together by mere fusion, as would occur with particles of wax by exposure to a gentle heat. The mode in which a coherent coke may be conceived to be produced may be illus- trated by the following experiment. When the powder of charcoal or anthracite is heated in a covered crucible, even after having been well rammed in, no degree of caking will take place under any con- ditions of temperature ; but if the powder be mixed with a little coal-tar or coal-tar pitch, and afterwards similarly heated, a firmly caked or coherent mass may be obtained, which may be thus made so solid and hard as to give a sonorous ring when struck. The tar or pitch is decomposed when sufficiently heated in a close vessel and resolved into volatile products and a fixed residuum of shining car- bonaceous matter; and it is this matter which acts the part of a solder and binds firmly together the particles of charcoal or anthra- cite. Now in the case of caking coals at a certain degree of heat, products appear to be formed which are subsequently resolved, at a higher temperature, into volatile matter and a carbonaceous residuum, and so a coherent coke is produced. Of the nature of these products I am not aware that we have at present any certain knowledge. ' It is interesting to inquire whether the caking quality depends simply on the elementary composition of a coal irrespective of its proxi- mate constitution. It is obvious that this question would be imme- diately solved if it could be shown that a caking and non-caking coal may have the same elementary or ultimate composition. Mr. Nicholas Wood, so well known in connexion with the coal -trade of the Tyne, obligingly supplied the author with samples of two kinds of coal from the vicinity of Newcastle, which, he believed, had the same elementary composition, but of which one was caking and the other non-caking. These coals were carefully examined and analysed by Mr. Dick in the metallurgical laboratory of the School of Mines, and the result was, that although their elementary composition was very similar, it was yet far from identical, and that both caked, though in a different degree. If we inspect the following table, in which is presented the compo- sition of a series of coals, caking and non-caking, we shall be dis- posed to suspect that there might be an essential connexion between the caking quality and elementary composition. In order to be able to compare these coals with each other, the hydrogen and oxygen, inclu- sive of nitrogen, are calculated in relation to the constant quantity of 100 of carbon. 1. 2. 3. 4. 5. 6. 7. 8. 9. Carbon ...... 100-00 lOO'OO 100-00 100-00 100-00 lOO'OO 100-00 100-00 100-00 Hydrogen ... 4-75 4-45 5-49 5'85 5-91 6'34 6-12 6'04 5'99 5-28 7-36 10-86 14-52 18-07 21-15 21-23 22-55 23-42 Non-caking. Caking. Non-caking. CAKING COAL. 93 The excess of hydrogen in these nine varieties of coal above that required to form water with the oxygen will be respectively 4-09 4-53 4-13 4'04 3'65 3'70 3-47 3-22 3'06 These numbers are erroneous to the extent of the nitrogen, which in the calculation has been included in the oxygen, but the amount of error arising from this cause may be disregarded. Hence it would seem that no essential relation exists between the excess of hydrogen and the property of caking, for that excess is nearly identical in Nos. 1 and 4, which are non-caking and caking respectively. If we now estimate the sum of the hydrogen and oxygen existing in each of these coals we obtain the following results : 10-03 11-81 16 35 20-37 23'98 27-49 27-35 28-59 29-41 We might, therefore, be led to infer from the preceding data that the property of caking is connected essentially with the proportion of oxygen, and that when the oxygen in a coal exceeds in round numbers 7 per cent., and does not exceed 18 per cent., the coal would have the property of caking. This inference, however, in respect to general application, is entirely opposed to the researches of Stein, Professor of Chemistry at the Polytechnic School of Dresden, to whom we are indebted for a monograph on the coals of Saxony. 4 I have selected from this monograph the following series of results, which would appear to establish the fact that caking and non-caking coals may have the same ultimate composition. STEIN'S TABLE. Composition > exclusive of Ash. 2 ' t^ P Result J Locality. C> 1 | i >? 1 i 4 .* -3 | ? a iU IH of Coking. / 02 6 w o 2 ti <1 d a W & fc I ?, OlxM-hohndorf Zwickau 1-265 1-300 82 '42 80-25 4'50 4'01 11-61 10-98 0-43 0-49 1-21 ?-99 0-74 1-57 4*75 5-91 83-28 83-82 4-55 4-19 11-73 11-47 0-44 0-51 47-70 69-59 Caking. Coke described ns like the coal itself, so that not even fritting could have occurred. 3 Do. 298 76-59 4-12 12-87 0-33 0-81 6'00 5-07 81-47 4-S8 13-71 0-35 54-64 Caking. 1 5 6 7 Niederwiirschnitz . Do Pianitz Niederwiirschnitz .. 378 311 280 454 72-35 80-49 81-23 77-42 4-17 11-99 4-10 10-62 4-43 9-86 4-65 11-73 0-62 0-20 0'21 0-23 2-65 1-10 0-55 1-6S 8-33 3 64 4-25 4-93 7-15 9-11 4-85 7-53 81-17 4'67 84-364-30 84-844-63 82-34I4-73 13-47 11-13 10-74 12-69 0-68 0-21 0-23 0-24 69-73 62 '40 63-89 66-43 Coke crumbly. Coke slightly fritted. Caking. Coke sandy. Zwickau.... 275 72-27 4-1610-73 0'34 0-88 12-50 5 'OS 2-59 4-76 12-260-39 77-29 Caking. 9 Do 291 75-26 4-08 16-07 0-20 1-71 3-07 6-30 78-71 4-27 16-81 0-21 77-44 Caking. 10 Niederwiirschnitz .. 331 76-03 4-35 16-05 0-13 0-13 3 31 8-48 78-74 4-51 16'62 0-13 60-81 Coke crumbly. If we admit the fact that a caking and non-caking coal may have the same ultimate composition, we should be necessarily led to the conclusion that the quality of caking depends upon proximate constitution, that is, upon the manner in which the elements are combined. But we must 4 Chcmische u. chemisch-technischo Untersuchun< W. Stein. 1857. der Stcinkohlcn Sachsens. 94 CAKING COAL. wait for future information on this interesting subject, which seems well worthy of investigation. As the passage of woody tissue into bituminous coal and of bitu- minous coal into anthracite is gradual, and as neither woody tissue nor anthracite is in the least degree caking, it follows that, beyond a cer- tain limit as a coal approximates in composition to woody tissue on the one hand and to anthracite on the other the presence or absence of the caking property may be certainly predicated from ultimate com- position alone. It is asserted that some coals speedily lose the property of caking after having been drawn from the pit ; and the author has it on good authority that a variety of coal which occurs at Penclawdd, near Swan- sea, loses its property of caking after exposure to the atmosphere not longer than one or two days. In this case the cause of the loss of the caking quality will probably be found to depend on the escape of some r thing from the coal, or on the oxidizing action of the atmosphere upon the coal. M. de Marsilly states that strongly caking coal, which yielded an excellent coke when fresh from the pit, yielded only an imperfectly formed coke in the same ovens after having been exposed to the air during six months. 5 To the same observer we owe some interesting and important obser- vations, which may throw light on the cause of the property of caking. He asserts that all caking coals (houilles grasses) from pits in which fire-damp occurs cease to swell up and cake when they have been pre- viously heated to 300 C. ; so that when they are calcined in the state of powder after having been thus heated they will be found in the state of powder after the calcination. I have confirmed the correct- ness of this assertion with respect to the strongly caking coal of New- castle-on-Tyne. The powder of this coal was heated in the hot-air bath at a temperature ranging between 300 C. and about 304 C. It may be thus heated for about a quarter of an hour or so without sensibly losing its property of swelling up and caking ; but when it is kept exposed to this temperature during one or two hours it does not swell up on subsequent calcination, and yields only a very slightly fritted coke. M. de Marsilly infers that the loss of the caking property by exposure to the air during a long time, and to the action of heat under 330 C. during a short time, is due to the same cause the volatilization of matter upon which, he believes, the property of caking depends. Whether the loss of this property by long exposure to the air is due solely to the cause in question does not appear to be established. The most important results of M. de Marsilly's investigation are the following : 1. The loss of weight which coal suffers by desicca- tion in vacuo is always less than that which is occasioned by exposure to 100 C. At 50 C. gas begins to be evolved from coal, but the evo- lution only becomes very sensible at 100 and upwards. It goes on in- creasing to 330, when, probably, the decomposition, properly so called, Comptes Reudus, 1858, 46. 882. CAKING COAL. 95 of coal commences. The amount of gas evolved varies from 1 to 2 litres per kilogramme of coal (61 to 122 cubic inches per 2J Ibs. avoird. nearly) ; and there is also distilled a liquid product, having the odour of benzine, of which the weight varies from 10 to 15 grammes per kilo- gramme of coal (154 to 231 grains per 2 Ibs.). The loss in weight which coal undergoes at 300, owing to the separation of gas and liquid com- bined, ranges from 1 to 2 per cent. 2. It is a remarkable fact that from the coal of pits subject to fire-damp, carburetted hydrogen is always and almost exclusively disengaged, whereas the gas evolved from the coal of pits free from fire-damp consists chiefly of nitrogen and carbonic acid without any trace of carburetted hydrogen. 3. Carburetted hydrogen is spontaneously evolved from coal newly won, even under a pressure five times as great as that of the atmosphere. After having been exposed to the air during six months or probably less coal yields no gas, not even at 300. Hence M. de Marsilly is disposed to conclude that the gases disengaged by the free exposure of coal to the atmosphere are the same as those obtained by its exposure to 300. This, however, is a conclusion which nothing short of analysis can justify. In some instances the caking of a coal depends on the manner in which it is heated and the degree of heat to which it is subjected. Thus the South Staffordshire coals (Nos. 27-30 in the Table, p. 102) are practically non-caking coals ; that is, the powder of these coals when heated in the usual way does not yield a coherent coke, but only a very slightly fritted and crumbling mass. Yet if it be rapidly exposed to a high temperature, such as a bright red heat, in a dose vessel, a good solid coke may be obtained. There is an enormous amount of the slack of such coals annually raised in South Staffordshire, and an enormous amount also left under ground, which might be turned to a profitable account if coking, under the special conditions just men- tioned, could be economically conducted. 6 The amount of water in coal is not without influence on the pro- perty of caking, as will appear in the sequel. \\hen the proportion of inorganic matter in coal is very large, its influence in diminishing, if not in destroying the caking quality may be readily understood. However, according to Stein, a coal may yield as much as 21-67 per cent, of ashes and yet be caking. 7 The property of caking may be very important in respect to the application of coal as a fuel, especially in metallurgical operations. Strongly caking coal may soon become agglomerated in the furnace into a mass so compact as to be, in a greater or less degree, impervious to air, in which case the fire, without stirring, would speedily be extinguished, and stirring, in many cases, would be quite impracti- cable. In the copper-furnaces at Swansea caking and non-caking in other words, binding and free-burning coal are used in admixture with 6 This is a subject affecting not merely to 160,000 tons of small coal in a year." South Staffordshire. According to Mr, Colliery Guardian, July 6, 1861. Nicholas Wood " the waste at the Hetton I ? Op. cit. p. 93, Nos. 5 and 6 in the and Black Boy Collieries alone amounted | Table. 96 FREE-BURNING AND CANNEL COAL. great advantage, the non-caking coal serving effectually to keep the fire open to the passage of air. Some kinds of coal possess the quality of caking in so high a degree as to stick firmly to the bars of a fur- nace, and cannot, therefore, be advantageously used even under steam- boilers. Free-burning Coal. This term is applied to coal which does not, in burning, sinter together, or cake, in a sensible degree. The fire remains open, so that the air can pass freely through it ; whereas the reverse is the case with caking coal, which cannot, without admix- ture with free-burning coal, be directly applied in various metallur- gical operations. Cannel Coal. This coal burns readily without melting, and emits a bright flame. A piece of good cannel coal when ignited will continue to burn for some time afterwards. The term cannel a corruption of candle is applied to coals of this kind because they burn like a candle. They are brown, brown-black, or black ; fracture uneven or largely conchoidal ; some are comparatively tough, others as brittle as other kinds of coal ; they do not soil the fingers ; some are susceptible of a fine polish, and may be wrought into articles of ornament. Thus jet is a variety of cannel coal. The term " Parrot " is applied to a cannel coal occurring near Edinburgh, on account of its burning with a crackling noise; and a variety of cannel coal from South Wales is termed " Horn Coal " because it emits while burning an odour like that of burning horn. Cannel coal is especially valuable as a gas- coal. ANTHRACITE. This coal is the ultimate product of the conversion of vegetable matter into coal. It generally contains upwards of 90 per cent, of carbon. The external characters by which anthracite is dis- tinguished are as follow : very compact ; deep black ; lustre bright, occasionally somewhat bronze-like or semi-metallic ; brittle ; fracture uneven or conchoidal ; does not soil the fingers ; bums with a feebly luminous, smokeless flame, and is much less combustible than other kinds of coal ; when heated does not in the least degree sinter, but frequently decrepitates considerably. Fibrous and granular matter in coals. In some kinds of coal may often be observed thin layers or patches of black, fibrous, soft matter, which soils the fingers, and is much like wood-charcoal in appearance. It is met with in cannel coal, but in less quantity than in other bituminous coals. Under the microscope it is found to possess the structure of woody tissue. 6 A granular and more or less pulverulent variety of this matter also occurs in coals. Dr. Eowney has examined both kinds, and finds them to differ sensibly in composition from the coal with which they are associated. Both kinds exist in coal of the carboniferous series, but only the fibrous kind is mentioned as present in coal of the oolitic and tertiary formations. Eowney describes the pulverulent kind as occurring sometimes as a light powder, and at others as a cindery sub- stance which peels off the coal in flakes easily reducible to powder. Dr. Bennett, Trans, of the Royal Soc. of Edinburgh, 1854, v. 2, p. 186. COMPOSITION OF COALS USED IN COPPER-SMELTING. 97 COMPOSITION OF FIBROUS AND GRANULAR MATTER IN COALS, DRIED AT 100 C. Character. Carbon. Hydrogen. Oxygen. Nitrogen. Ash. 1. Fibrous 82 '97 3-34 6-84 0-75 6-08 72-74 19-08 3 Fibrous 73-42 2-94 8-25 15-39 4 Do .. . 74-71 2-74 7-G7 14-86 5 Do 81-17 3-84 - ^_f 14 -9 -^ 1. From the common household coals of the Glasgow coal-fields. 2. From the Stonelaws coals. 3. From Ayrshire coal. 4. From the Elgin splint coal, Fifeshire. 5. From the 5-feet seam, Elgin coal, Fifeshire. 7 The composition of the coals from which Nos. 3, 4, and 5 were taken, as determined by Kowney, will be found in Table, p. 102, Nos. 37-39. Composition of the coals used in copper-smelting. It has been previously stated (p. 84) that in the smelting-works of Swansea and the neigh- bourhood the clinker is allowed to accumulate to such an extent as to form a bed of considerable thickness, and that upon this bed, which is kept sufficiently open to allow of the passage through it of the proper amount of air to sustain combustion, small coal may be effec- tually burned which would in great measure drop through an ordinary grate composed of bars arranged in the same plane. It is requisite, therefore, that the inorganic matter of the coals employed should, at the temperature of the furnace, sinter or fuse together so as to form a solid coherent clinker, but that it should not run into an easily fusible vitrified mass. Being desirous of ascertaining the composition of the ashes of the coals which produce a suitable clinker, I applied to Mr. William Morgan for average samples of the coals which are consumed at the Hafod Copper-Works, of which he directs the smelting department. The samples were promptly supplied, and I have much pleasure in acknowledging the willing assistance which I have at all times received from Mr. Morgan, especially during the lifetime of the late Mr. Vivian a man who was ever ready to aid the scientific inquirer, who made no pretension to the exclusive possession of mysteries, and whose memory is deservedly revered by the inhabitants of Swansea. Three kinds of coal are used in admixture at the Hafod Works : one binding, or caking; and two free-burning, or non-caking. The mix- ture consists of the same weight of each of these three coals, viz., 1 part of Mynydd Newydd, or binding coal ; 1 of Tyrcenol and 1 of Pentrefelin, or free-burning coals. The analyses have been made in my laboratory by my friend Mr. Dick. The ashes of each of these coals were exposed separately to a high temperature, and found to have about the same degree of fusibility. ~> Kdinb. Now Phil. Jour., 1855, v. 2, p. 141. 98 COMPOSITION OF COALS USED IN COPPER-SMELTING. COMPOSITION OF THE ASHES. 1. 2. 3. Silica 35-05 35-01 36'15 Alumina 2CrOO 28'01 28-12 Sesquioxide of iron 19-56 19-06 26-26 Lime 5-80 4-53 2'28 Magnesia 1'95 2-14 1-68 Potass 2-55 2'95 1-36 Soda 0-65 0-95 0-64 Sulphuric acid 8-45 7'14 3-17 100-01 99-82 99-66 1. Mynydd Newydd. 2. Tyrcenol. 3. Pentrefelin. The ash of each coal was reddish, and that of No. 3 most so. The total sulphur in each coal was determined by deflagration in a gold crucible, with a mixture of nitre and chloride of sodium ; and the sulphuric acid pre- existing in the coal in the state of sulphate was determined by digest- ing the powder of the coal in hydrochloric acid, and proceeding by the addition of a baryta-salt in the usual way. The greater part of the sulphuric acid in the ashes is the product of oxidation during incine- ration in the presence of strong bases. If we estimate the sulphur exclusive of that in the sulphuric acid pre-existing in the coal in combination with iron as iron-pyrites, the complete analyses of these coals will be as follow : 1. 2. 3. Carbon .'. 73'87 76- 81 78-49 Hydrogen 3'73 3'42 3*73 Oxygen and nitrogen 8*02 5*65 4*15 Silica 5-05 4'68 4-24 Alumina 3'75 3'74 3-29 Sesquioxide of iron 0-88 0-10 0-00 Lime 0'83 0-60 0-27 Magnesia 0-28 0-28 0-19 Potash 0-36 0-39 0-16 Soda 0-09 0-12 0-07 Sulphuric acid 0-23 0-54 0-69 i r o, pyritfis { g^EEE: J^ ::::::::: J^ :::::::: 1% 100-00 100-00 100-00 Ash obtained by incineration,! 1. 13-39 14-35 11-76 percent j 2. 13-34 14-32 11-71 Mean 13'37 14-34 11-73 The iron introduced in the foregoing analyses is estimated as sesqui- oxide, a portion of which only, probably, existed in combination with the shale or clay, which was evidently present in very appreciable quantity. Le Play has published the following analysis of clinker, reputed of good quality, which, if I mistake not, was procured at the Hafod Works. 1 1 Description des Precedes Metallur- I pour la Fabrication du Cuivre. Paris, giques employes dans le Pays de Galles | 1848, p. 122. BRITISH CAKING COALS. 99 Silica Sesquioxide of iron.... Protoxide of iron Alumina Lime Magnesia Sulphur Iron Carbon 52-0 5-2 22-0 14-2 2-4 0-7 1-3 1-0 1-2 100-0 Supposed mode of Combination. Silicate of sesquioxide of iron.... 11 , , protoxide of iron .... 40 4 , , of the earthy bases. .. 45 '1 Sulphide of iron (PeS) 2-3 Carbon 1 -2 100-0 These results do not closely accord with those obtained by Mr. Dick. Le Play makes no mention of potass and soda, which are certainly present in the ashes of the t3oals supplied by Mr. Morgan. Reasons will be given in the sequel which render doubtful the existence of silicate of sesquioxide of iron in the clinker. Composition of Bituminous Coals. In the following tables is presented a selection of analyses of British and foreign bituminous coals and anthracites. In a subsequent part of this work will be found analyses of certain coals of special interest in the smelting of .iron in particular localities. BRITISH CAKING COALS. 1 2 4 5 6 7 8 I^ocality. Specific Gravity. too \ I 5 78-65 82-42 81-41 78-69 77-40 82-56 83-44 83-00 Hydrogen. | O d 2-05 2-37 1-55 1-65 1-66 1-49 fcl 1 1 2-49 0-79 2-07 1-36 3-90 1-46 2-45 4-00 O 1 1 : 35 3 : 50 ^ Exclusive of Ash. a 1 80-541 4-76 83-73 4-90 83-26 5-65 81-01! 6-17 84-43; 5-41 84-42 5-48 86-25' 5-90 87-14 6-49 % O 2-87 2-44 1-69 1-70 1-72 1-56 Nortliuinlicrlaiid . Ditto Ditto Ditto Blaina,Soutb WaU-s Ditto Ditto 4-65 4-82 5 S3 6-00 4-96 5-36 5-71 6-18 14-21 11-97 7-90 10-07 7-77 8-22 5-9:5 4-58 0-55 0-86 0-74 1-51 0-92 0-75 0-81 0-75 66 '-7 63-18 14-70 11-37 8-22 10-38 8-47 8-40 6-13 4-81 * \VlKTC no figures arc given under tliis column, the Nitrogen is included in the Oxygen. I. 2 Sent by Mr. Nicholas Wood. The " Seaton Burn Steam-coal;" it is from the same bed as all the Hartleys, viz., West Hartley, Huddle's Hartley, &c. The powder swells up slightly when heated to redness, and forms a coke. In burning, this coal does not stick to the bars like No. 2, and it is therefore styled " open burning ;" but this term is only correct when used comparatively with a coal like. No. 2. An average sample for analysis was obtained from a large lump. In the table jio correction has been made, the number under the column of Oxygen being found by subtracting the sum of the carbon, hydrogen, and ash (as left by incineration) from 100. The 2 -49 of ash con- tained 49 of sulphuric acid ( = 19 sulphur), and 39 of sesquioxide of iron ( = 0-273 iron). But this quantity of iron requires 0-31 of By Dick in my laboratory. H 2 100 FOREIGN CAKING COALS. sulphur to form pyrites. The excess of sulphur, therefore, beyond what was thus required, is 0' 55 0*31=0' 24. 2. 3 Also from Mr. Wood. The " Peareth Gas-Coal." It is of the quality called gas-coals, and is the same as the Pelow Main, Felling, Peltree, and other coals of the Tyne. The powder swells up when heated to redness, and forms a coherent coke. 0'74^of ash contained 0-06 of sulphuric acid. The ash had a reddish grey but lighter colour than that of No. 1. But if the ash had consisted entirely of sesquioxide of iron, there would not have been sufficient iron to form pyrites with all the sulphur present. Neither No. 1 nor 2 contained any appreciable amount of carbonate. Mr. Wood remarks that "it is curious that both these coals 'are from the same bed ; but steam-coal exists north of Newcastle, and the gas coal south of it, the change taking place nearly in the line of the river, and within a zone of three miles on the north side." 3. 4 Name of the seam not stated. The coke caked together, swelling considerably. Pyrites was visibly present in this coal. 4. 5 Fracture conchoidal ; interspersed rather abundantly with iron pyrites ; chiefly used for steam purposes, for which it is largely ex- ported. From the " Low Main Seam," Buddie's Hartley Colliery. 5. 6 From Shireoak Colliery, belonging to the Duke of Newcastle. Ash bulky, and slightly pink in colour. 6. 7 Ell-vein coal. 7. Three-quarter-vein. 8. Big-vein. Mr. Adams, of the Ebbw Vale Iron- Works, which are situated in a valley adjoin- ing and parallel to that in which are the Blaina Iron-Works, informs me that the Ell-vein and Big- vein are steam-coals with a white ash ; and that the three-quarter coal is a furnace-coal. FOREIGN CAKING COALS. ^ Exclusive of Ash. 1 Locality. fie Gravi 1 fj J ^ o i | c 02 1 1 O I f 1 3f i I 1 I | 9 10 Epinac .... Alais, de'p. du Gard 1-353 1-322 81-12 89-27 5-10 4-85 11-25 4-47 2-53 1-41 63-6 78-0 83-22 90-55 5-23 4-92 11-55 4-53 11 Rive-de-Gier 1-298 87-45 5-14 3-93 T70 T78 68-0 89-04 5-23 5-73 12 Ditto 1-288 82-04 5-27 9-1? 3-57 72-0 85-08 5-46 9-46 13 ( Ce"ral, dep. de i \ 1'Aveyron 3 1-294 75-38 4-74 9-02 10-86 58-4 84-56 5-32 10-12 14 Saint-G irons 1-316 72-94 5-45 17-53 4-08 44-8 76-05 5-69 18-26 15 Mons 85-10 5-49 7-25 2-16 72-90 86-98J 5-61 7'41 16 Ditto SO -55 5-53 9-5? 4-40 69-15 84-26 5-78 q-96 17 Ditto 86-38 4-48 6-09 3-05 80-58 89-10 4-62 6'28 18 Charleroi 86-47 4-68i 5-30 3-55 84-43 89-65 4-85 5-50 19 Valenciennes 84-84 5-53 6-83 2-80 67-75 87-28 5-69 7-03 20 Pas-de-Calais 86-78 4-98 5-84 2-40 77-05 88-91 5-10 5-99 21 Hungary 295 0-86 0-89 20 78-85 88-72 4-66 6-61 22 Ditto 300 0-99 2-85 1483-14 88-85 4-23 6-92 23 24 Ditto Ditto 313 378 2-83 5-53 5 82 11'41 04j82-82 57I77-81 88-30 83*76 4-80 4 '97 6-90 11'26 ' 25 26 Ditto . . New Zealand 350 79 : 00 5-35 7 : 7l : 89 0-90 2-50 10-33 3-50 08 1-05 81-55 64-32 89-69 84-90 5-03 5-75 5-27 8-29 : 96 3 By Dick. 4 By Vaux, op. cit. 5 4. By Taylor, Ed. New Philos. Jour, v. 50. p. 142. 6 5. By C. Tookey, in my laboratory. 7 6-8. By Dr. Noad, and communicated by him to the author. FOREIGN CAKING COAlMjJ N I V E K SM" Y 9. 8 Powder brown. Falls to pieces by sequence of the pyrites which it contains. It volume by calcination; coke semi-metallic in lustre and agglutinated, but the different pieces from which it has been formed may be easily recognised. 10. 9 Powder black brown. Coke semi-metallic in lustre, slightly puffed up ; the pieces from which it has been derived can often be distinguished. The coal is regarded as very hard, i.e. difficult to burn, but capable of producing a very high temperature. The coke is ex- cellent for blast furnaces. 11. 1 Powder brown. Coke much puffed out. The small in much request for the making of coke. 12. 2 Powder brown. Coke puffed out. 13. 3 Powder black brown. Coke semi-metallic in lustre, fritted ; the particles stick well together. This coal occurs in lower marls of the inferior oolite. 14. 4 Powder brown. This coal is a very brilliant jet; very hard; fracture conchoidal ; used for ornamental purposes. Coke semi- me- tallic in lustre ; brilliant ; the particles become rounded, and stick together pretty firmly. It occurs in the upper series of the second- ary formation (Terrain cretace). 15 20. 5 In all the coke is described as well formed, and in No. 10 it is stated to be well formed and puffed out. 2 1. 6 Colour pitch black ; lustre fatty, here and there passing into vitreous ; fracture very uneven ; caking. From Eesicza mine, county of Krasso. 22. Colour of the coal and its powder pitch black ; lustre in the direction of the layers fatty, but on the cross fracture glistening. Is very easily pulverized, but does not fall to pieces when exposed to the air. Xtro/if//// ciikiiitj. From Fiinfkirchen, county of Baranya. "2-'>. Colour pure pitch black ; lustre vitreous ; compact, difficult to pulverize, yet falls to pieces by exposure to the air, though not to powder. Sfi'niii/lf/ raking. From the same locality as No. 14. 24. Colour of the coal and its powder pure pitch black. Lustre fatty. Easily rubbed to powder between the fingers, and falls to fine powder by exposure to the air. Strongly caking. From Szabolcs, county of Baranya. 25. Colour of the coal pitch black. Compact, difficult to pulverize, resists the action of the air, and does not fall to pieces after several years' exposure. '-'i;. 7 From the west coast of the Middle Island. Ash remarkably Avliite. Coke contains 2 -35 percent, of sulphur." No sulphuric acid was detected in the hydrochloric acid in which the powder of the coal Lad been boiled. It would appear therefore that the sulphur was '.). Renault, Ann. d. Mines, 3. s. 12, p. IS! . " K). Do. do. p. 185. 1 11. J>o. do. p. 196. - 1-J. Do. do. p. 200. 3 1:5. Do, do. p. 214. 4 14. Regnault, op. cit. 216. 5 1520. M. de Marsilly. Comptes Rendus, 46. 891. 6 21 25. By Nendtwieli, op. cit. pp. 15-22. ? 26. By C. Tookey, in my laboratory. 102 BRITISH NON-CAKING COALS. BRITISH NON-CAKING COALS. > Exclusive of Ash. Locality. 5 d j O o o ^ q 3 1 a o 3 g to % o o o 1 1 J &Q ^ % 1 o 5 ps I 3 a 1 3 2 3 CQ 1 A a 1 $ B 2 South Staffordshire 76-12 4-83'l6-72j 1-00 2-33 (78-46 4-9616-58 .. 77-01 4-71 16-72f 0-74 1-56 f 78-53 4-8016-66 . 76-40 4-62l7-43f* .. 0-55 1-55 ) 77-68 4-6917-62 72-13 4-3217-llJ 0-54 6-44 (77-32 4-67,17-99 Jf Nr. Wolverhampton St. Helen's, Lancashr 1-2 878-57 1-2 975-81 5-2912-88 5-2211-98 : 84 0-39 1-03 11-2957 21 930-905-17 3-2365 50 79-38 79-93 5-34 5-50 13-02 11-58 Dowlais, S. Wales 89-33 88-13 4-43' 3-25 4-51 2-94 24 41 0-55 1-01 1-201 0-79 2-00 0-68 90-93 90-86 4-51 4-65 3-30 3-03 1-26 1-46 87-62 4-34 2-52 131-07 3-32 0-68 91-64 4-54 2-64 1-18 " ! 82-60 4-88 3-44 281-22 7-18 0-78 90-18] 4-67 3-76 1-39 rtland" . 76-08 5-31 13-33 09 1-23 1-96 78-59 5-4913-77 2-15 80-63 5-16 10-61 330-84 1-43 82-50 5-2810-86 1-36 . 80-93 5-21 10-91 570-636-75 " 82-06 5-29 11-06 1-59 1 * Oxygen found by deducting the sum of the carbon, hydrogen, and ash from 100, exclusive of sulphur, f Calculated after corrections. See text. present in the same state of combination in the coal as it exists in albumen, fibrine, &c. It could not have been combined with iron, for in that case the ash would have had a decided red colour. Geologists are of opinion that this coal occurs in strata of miocene age, in- which case it presents an example of a coal similar in external characters and chemical constitution to varieties of coal occurring in the true coal measures. 27 30. 8 Thick or ten-yard coal from the vicinity of West Brom- wich, This seam consists of ten or more beds, to which special names are applied. No. 27 is "Rooves;" No. 28, "Top-slipper;" No. 21), "White-coal;" and No. 30, " Brasils." No. 30 contains much earthy matter, and is in request for certain reverberatory smelting furnaces in Birmingham. The powder of each of these coals is brownish black. The colour of their respective ashes" is as follows : No. 27, reddish- grey; No. 28, yellowish-red; No. 29, yellowish-red; No. 30, reddish- grey. The portion analysed of each coal was dried somewhat above 100 C. Combustion was effected in oxygen. The total amount of sulphur in each coal was determined, as was also the amount of sul- phuric acid in the ashes. No. 27 contained a trace of sulphuric acid. Sulphuric acid in the ashes of 100 parts of coal. Iron in the ashes of 100 parts of coal. Sulphur re- quired to form iron pyrites with the iron. Total sulphur in 100 parts of coal. The + and - signs indicate an excess of sulphur or otherwise. Pyrites in 100 parts of coal, estimating the whole of the iron as FeS 2 , except in No. 4, where the iron is in excess. No. 27 24 0'32 0-37 I'OO + 0-63 O'f>9 28 0-31 0-12 0-14 0-74 -f-0-60 0'26 29 0-33 0-03 0-034 0*55 + 0*52 0-064 30 1-31 0'94 1*08 0'54 0-54 1 01 2730. By A. Dick, in my laboratory. BRITISH NON-CAKING COALS. 103 In calculating the composition of these coals in the table, the fol- lowing corrections have been made : Deduction from the weight of ash of the oxygen corresponding to the sulphur present as sulphuric acid, the product of oxidation ; deduction of the oxygen of the sesqui- oxide of iron existing in the ash, except in No. 4, in which a deduction of oxygen from the sesquioxide of iron was made equivalent to the proportion of iron required to form bisulphide with the sulphur in the coal, the sulphur in this case being the reverse of excessive ; and in No. 30 a correction has been made in the ash by the addition of 2 04 per cent, of carbonic acid 0-43 found in the ash (= 0'72 of carbon) which was present in the coal as carbonate. Hence the amount of inorganic matter in this coal maybe estimated at 7 '67, i.e. by the addition of the carbonic acid expellee! by incineration to the actual amount of ash obtained, and with the correction for the sulphuric acid and sesquioxide of iron as above stated. The 7 67 would consist of iron pyrites I'Ol and 2-64 of carbonic acid in combination in the 4-02 of fixed residual inorganic matter. The first four analyses after these corrections will be : Carbon. Hydrogen. Oxygen and Nitrogen. Sulphur. Iron Pyrites. Residue. No. 27. . . . 76-12 4 83 16-10 0-63 0-69 1-63 28 77-01 4-71 16-35 0*60 0*25 1 -08 !>.) 76-40 4-62 17-23 0-52 0'064 1-17 :;. 72-13 4-32 16 88 1-01 6-66 It is no libel on South Staffordshire to assert that this magnificent bed of coal has been most barbarously treated. The pits have gene- rally been worked by contractors, called butties, under the superin- tendence of viewers, called ground-bailiffs. In consequence of the rapacity and rascality of many of the former, and the ignorance, inat- tention, and fraudulent connivance of many of the latter, an enormous amount of coal has been irremediably lost to the nation. Even at the present day the South Staffordshire colliery-viewers are frequently very imperfectly educated for their responsible duties, and the system of colliery mismanagement which still extensively prevails in this part of the country is a disgrace to the age. 'M. 9 From the ten yard coal. Coke very lustrous ; swells up very much into radiating cauliflower- shaped masses ; does not take the form of the crucible. Under ordinary conditions of coking, i.e. with the partial admission of air, this coal, which is described as caking, is practically non-caking ; the small cannot be -coked per se in ordinary ovens. 82. 1 Eushey Park seam. Slightly caking. Lustre of coke semi- metallic. 33. 2 Upper four-feet coal ; seam about 2 feet 9 inches thick ; the best s 31. By Vaux, Jo. of Chem. Soc. 1, 328. 1 32. By do. do. 321. 2 33-36. By E. Kiley. by him to the author. Communicated 104 FOREIGN NON-CAKING COALS. FOREIGN NON-CAKING COALS. & , Exclusive of Ash. t^ Locality. 5 d e O o o --~ | 1 ft eg 1 Carbon. I 1 | | g I 5 ^ 1 1 i Carbon. I 1 40 Blanzy, France 1-362 76-48 5-23 10-01 . 2-28 ~7T 57-0 78-26 5-35 16-39 41 Commentry 1-319 82-72 5-29 11-75 0-24 63-4 82-92 5-30 11-78 42 Noroy des Vosges . . 1-410 64-28 4-3513-17 . , 19-20 60-3 78-32 5-38 16-30 43 Mons 82-91 5-2210-13; . 1-741 . 66-96 84-38 5-31 10-31 44 Ditto 1 82-95 5-42|lO-93 0-70 . 63-58 83-53 5-46 11-01 45 Valenciennes 90-54 3'66 2-70 3-10 . 93-17 93-44 3-78 2-78 46 Pas-de-Calais 82-68 4-18 4-54 . 8'60 87-62 90-46 4-57 4-97 47 Charleroi ... ! 90-89 3-65 3-98 1-48 91-86 92-26 3-70 4-05 48 Ditto 88-69 4-251 5-26 1-80 85-57 90-32 4-32 5-36 49 Hungary ... 1 423 1 . . 1 0-58 10-53 3 : 0676-33 82-54 4-35 13-10 50 Ditto . 1 366 n-74 1*55! T'Sn Tn-ftn 78*37 3'Q9.i 1 7 -7(1 51 Ditto '1-317 0-20 1-60 2-6673-11 85-29 5-05 9-65 52 Ditto 1-319 0-87 2-26 3-21 69-98 81-57 4-41:14-01 53 i Near Aix-la-Chapelle 1-343 91 : 45 4 : Is 2-12 2-25 .. 89-4 93-56; 4-28 2-16 coal at Dowlais. 34. Eas Las: the so-called "brass" (clay iron ore mixed with coaly matter) and iron pyrites are visibly disseminated through this coal. 35.- Bargoed big-coal ; seam from 7 to 8 feet thick ; the cheapest coal. 36. Tomo yard coal ; bad coal for blast furnace. All these coals are used in the blast furnaces at Dowlais, and the Eas Las is also used for forges. 37. 3 Ayrshire. 38. Splint coal, Elgin, Fifeshire. 39. Ditto, Five- feet seam. 40. 4 During calcination the particles stick together a little, but separate under the slightest pressure ; they keep their form, being only a little rounded on their edges. This coal burns with a good flame, which only lasts a short time. Coke cannot be made with it, but it is esteemed for boilers. It is characterized as a dry coal with a long flame (houille seche a longue flamme). 4 1. 5 Powder black brown. Coke semi-metallic in lustre, grey, nearly white, very brilliant, and only fritted. Eegnault describes this as a true cannel coal. 42. 6 Powder brown. It contains much pyrites disseminated through the whole mass. This coal occurs in the lower series of the Jurassic beds (etage inferieur, mornes irisees). 43, 44. 7 Coke fritted. 45. Coke not formed. 46. Coke not formed, in powder. 47. Coke not formed, in powder. 48. Coke hardly formed. 49 . 8 Colour pitch black ; lustre fatty ; non-caking (Sand-kohle). 50. Grey-black ; lustre dull, somewhat fatty ; non-caking (Sand-kohle). 3 37-39. By Kowney. Edinb. New Phil. Journ. 1855. v. 2. p. 141. 4 40. By Regnault, Ann. des Mines, 3. s. 12. p. 190. 5 41. By Regnault, op. cit. 3. s. 12. p. 193. 6 42. By Regnault, op. cit. p. 210. ' 43-48. By M. de Marsilly. Comptes Rendus, 1858. v. 46. p. 891. 8 49-52. By Nendtwich, op. cit. p. 15. CANNEL COALS AND ANTHRACITE. 105 51. Coal and its powder black; lustre fatty. Coke fritted (Sinter- kohle). 52. Colour black ; but little lustre ; difficult to powder, and per- fectly resists exposure to the air. Coke fritted (Sinter-kohle). All the Nos. from 49 to 52 from the county of Krasso. 53. 9 Powder pure black ; it has the vitreous lustre of compact anthracites and the foliated structure of ordinary coals. It is very little changed in appearance by calcination. It occurs at Kolduc. Regnault regards it as a link between anthracites and bituminous coals. CANNEL COALS. | & Exclusive of Ash. 1 Locality. ca i I d S 1 s f I O o o o ? 1 f a trogen. X. k a n O 585 S < 6 a A O K 54 Wigan 1-317 84-07 5-71 7-82 2-40 59-0 85-81 5-85 8-34 55 Ditto 1-276 80-07 5-53 8-10 2-12 1-50 2'7f) 0'91 82-29 5- fiS 8 -si 56 Tyneside 1-319 78-06 5-80 3-12 1-85 2-22 8-94 87-86 6-53 2-53 2-09 54. 1 Colour black brown, without lustre ; powder black brown. Coke silvery metallic grey, very brilliant, only fritted. The particles keep their form, but stick together. 55. 2 Coke hard, semi-metallic lustre, swells up and takes the form of the crucible. 5 (').'* Black, homogeneous, hard, brittle. Fractiire conchoidal. Sus- ceptible of a fine polish. From " Blaydon Main " Colliery, Tyneside. It is often found in connection with the coal (Buddie's Hartley), as roof, base, or even interstratified with it. BRITISH AND FOREIGN ANTHRACITES. ^5 Locality. Specific Gravity. j H i : 83 : 92 1-22 ! 1 o'ii o o 1-58 1-61 4-67 2-25 10-20 1 2-00 Exclusive of Ash. 94-05 91-87 94-89 94-72 94-64 1 1 3-38 3-34 2-55 2-69 2-73 l_ 2-57 3-02 2-56 2-58* 2-64 58 59 60 61 South Wales, near Swansea. South Wales Pennsylvania l-.'ils 1-392 1-462 92-56 90-39 90-45 92-59 84*98 3-33 3-28 2-43 2-63 2-45 2-53 2-97 2-45 1-61 1-15 Ditto Ditto * Inclusive of Nitrogen. 57. 4 Anthracite used at. the Yniscedwin blast furnaces. Powder pure black. 58. 5 Bright metallic lustre. Does not soil the fingers. Burns with- out smoke. 9 53. By Kegnault, Ann. des Mines, p. 184. 1 54. By Kegnault, op. cit. p. 191. 2 55. ByVaux,Jo.ofCiicm.Soc.l.320. 3 5G. By Taylor, Ediub. New Philos. Journ. 1851. v. 50. p. 145. 4 57. By Regnault, op. cit. 183. 5 58. By Vaux, op. cit. 324. 106 COMPOSITION OF THE ASHES OF COALS. 59. 6 From Fittsville. Powder pure black. The sides of the fis- sures are beautifully iridescent. 60. 7 Lustre shining. From the Lehigh Summit Mine, Pennsyl- vania, U.S. I received this specimen from Sir C. Lyell. 6 1. 8 Lustre somewhat glistening; powder black. From Mauch- Chunk, the shipping place of the Lehigh coal, at the eastern end of the southern anthracite basin. In the following table is presented, for the sake of illustration, a selection of analyses of the ashes of various coals : COMPOSITION OF THE ASHES OF COALS. 1. 2. 3. 4. 5. 6. 7. 8. 9. Silica 35-73 24-18 37-61 39-64 40-00 53-00 37-60 43-68 53-60 41 '11 20-82 38-48 39 20 i 39-34 36 '69 Sesquioxide of iron Lime 11-15 2-75 -6-00 9'38 14-78 2-53 11-84/ 1-81 44' 78 12-00 35-01 3-94 52-00 < 3-73 8-22 5-76 5-59 2-86 Magnesia 2-65 9-74 2-71 2-58 Trace. 2-20 1-10 3'00 1-08 Oxide of manganese . . Sulphuric acid Phosphoric acid Sulphide of iron (FeS) 4-45 0-99 8-37 0-21 0-38 : 29 2-00 Traces. 3-01 2-22 0-75 4 : 89 0-88 4-1 1 0-88 0-19 98-83 99-08 98-40 98-08 99-75 99-92 99-45 100-00 100-01 Nos. 1-4 9 are analyses of the ashes of the Dowlais coals, , num- bered from 33 to 36 respectively. No. 5. Kock Vein coal, Polity pool. No. 6. Four-feet steam-coals, Ebbw Vale. No. 7. Fordel splint coal, Fifeshire, Scotland. Nos. 8, 9. Anthracites of the United States. On the occurrence of certain metals in coals. Daubree has detected traces of arsenic and antimony in the coal of Newcastle-on-Tyne ; in the coal of Sarrebriick he found 0-003 per cent, of arsenic ; and in a variety of coal occurring at Ville (Bas-Rhin), in France, as much as 0-0415 per cent., besides traces of antimony and copper. 1 The same observer, as previously stated, detected arsenic in certain lignites. The Nottinghamshire coal (No. 5, Table p. 99) contained decided traces of arsenic. I have seen galena in coal from Bedworth, War- wickshire. The anthracite of South Wales, which we have been accustomed constantly to employ in the metallurgical laboratory of the School of Mines, contains decided traces of copper. Fremy's chemiral researches on combustible minerals. 9 ' These researches have been published since the preceding article on coal was in type. The following is a resume of the "chief results: 1. Lignite. The varieties known as bituminous wood are, like peat, partially soluble in alkalies, but they dissolve almost completely in nitric acid and the 6 59. By Regnault, Ann. des Mines, 180. 7 60. By the Author. Proceedings of the Geolog. Soc. 1. 202. 8 61. By the Author, op. cit. 9 Nos. 1 4. By E. Riley. Communi- cated to the Author. Nos. 57. First Report on the Coals suited to the Steam Navy, hy Sir Henry De la Beche and Dr. Lyon Playfair. Museum of Practical Geology, London, 1848. Nos. 8, 9. By Walter R. Johnson : A Report to the Navy Department of the United States 011 American Coals applicable to Steam Navi- gation and to other purposes. Washing- ton, 1844. 1 Recherches sur la presence de 1'arse- nic et de rantimoine dans les combustibles mineraux, etc., par M. A. Daubre'e, iiige- nieur des mines, Ann. d. Mines, 4 s. 19, p. 669. 1851. 2 Coinpk-s Rend. 52, p. 1 14. Jan. 1861. CHARCOAL. 107 hypochlorites. The compact and black varieties resembling bituminous coal are in general not acted upon by alkalies, but dissolve completely in hypochlorites and in nitric acid. 2. Bituminous coals. They do not dissolve either in alkaline solutions or in hypochlorites. Both bitu- minous coals and anthracite dissolve completely in a mixture of mono- hydrated sulphuric acid and nitric acid; a dark brown solution is produced, containing an ulmic compound, which is entirely precipitated on the addition of water. CHARCOAL. When wood is heated without access of air to about 300 C 8 ., it is resolved into volatile products, and a fixed residue or charcoal. The volatile products consist of water, acetic acid, tar, and other matters which are condensable, and of the permanent gases, carbonic acid, carbonic oxide, hydrogen, and carburetted hydrogen. This process of decomposition of organic compounds by heat without access of air is known as that of dry distillation or carbonization. Charcoal is extremely porous, and retains the structure of the wood from which it is derived. It consists essentially of carbon and of the fixed or inorganic matter which exists in wood ; but if carbonization be imperfectly effected, it may contain a sensible amount of hydrogen. Wood may in a great de- gree be carbonized at 220 C., but not completely below a red heat. 4 Good charcoal is black, gives a sonorous ring when struck, breaks with a more or less conchoidal fracture, does not crumble under moderate pressure, although it is brittle under a blow, does not sensibly mark the finger when it is rubbed against a freshly fractured surface, or make a mark which cannot easily be rubbed off, swims on water, and does not burn with flame when ignited in separate pieces. It is a bad conductor of heat and electricity ; but its conducting power is in- creased after it has been exposed to a high temperature. One end of a short piece of common charcoal may be held in the hand without inconvenience while the other end is burning actively ; but this could not be done with impunity with a short piece of charcoal which has been strongly heated. Violette states that charcoal made at 260 C. bums most easily ; but that made between 1000 and 1 500 C. cannot even be ignited like ordi- nary charcoal. Charcoal made at a constant temperature of 300 C. takes fire in air when heated between 360 and 380 C., according to the nature of the wood from which it has been derived ; charcoal from light woods, ccet. paribus, burning most easily. Charcoal made between 260 C. and 280 C. ignites between 340 and 360 C. ; that made be- tween 290 C. and 350 C. between 360 and 370 C. ; that made at 432 C. burns at about 400 C. ; that made between 1000 and 1500 C. ' A Jaliresbericht iiber die Forts, der Chem. Technologic, Wagner, 1857, p. 474. 4 I have obtained the following results concerning the temperature at which car- Ixmi/ation may be effected. A narrow test-tube, containing a thin strip of deal, was depressed ill mercury to the depth of an inch or two with the closed end downwards, and the temperature of the mercury was gradually raised. At 220 C. the wood became perceptibly brown, at 240 it became sensibly darker, and at 255 very brown. It was kept for some time at 280, when it became deep brown-black. At 310 it became very brittle, and could easily be pulverized. 108 CHARCOAL. between 600 C. and 800 C. ; and that made at the melting-point of platinum only inflames at about 1250 C. Violette, it will be per- ceived, applies the term charcoal to very imperfectly carbonized, scarcely more than desiccated, wood. 5 Owing to its highly porous structure, charcoal has the power of absorbing and condensing gases in considerable quantity. The follow- ing Table indicates the amount in volume of various gases which a given volume of freshly burnt charcoal is capable of absorbing. Ammonia 90 Hydrochloric acid gas 85 Sulphurous acid 65 Sulphuretted hydrogen 55 Carbonic acid . ..35 Carbonic oxide 9'42 Oxygen 9'25 Nitrogen 7*5 Hydrogen 1'75 From these data it would appear that the volume of gas absorbed is great in proportion to the condensability of the gas by pressure. It readily absorbs from 10 to 12 per cent, of its weight of aqueous vapour, and may even absorb as much as 20 per cent. Charcoal made at a low temperature absorbs moisture more rapidly and in a greater degree than that made at a high temperature. The power of charcoal to absorb moisture is much affected by the temperature at which it was produced ; the lower the temperature the greater the power. Thus charcoal prepared at 150, 250, 350, 430, and 1500 degrees Centigrade, absorbs 21, 7, 6, 4, and 2 per cent, respectively of moisture by ex- posure to a moist atmosphere. Charcoal in powder absorbs about twice as much moisture as the same charcoal in pieces. 6 The propor- tion of water which it retains will necessarily vary with the state of the atmosphere in respect to moisture. Commercial charcoal may be estimated to contain not less than 9, generally about 12, per cent, of moisture (Karsten, Scheerer) ; and when heated to whiteness, it loses from 14 to 15 per cent, of combustible gases and aqueous vapour (Berthier). The specific heat of ordinary charcoal, according to Eegnault, is 0*2411. The specific gravity of charcoal varies with the nature of the wood from which it is made, with the age of the wood, with the state of the wood as to dryness, and with the process of carbonization. New and accurate determinations of the specific gravity of charcoal from different kinds of wood, with statements as to the age of the wood, mode of carbonization, &c., are needed. In the old Table of Hassen- fratz, the specific gravity of charcoal, inclusive of the pores, ranges from 0*203 in birchwood charcoal, to 0*106 in limewood charcoal, that of oak being 0*155 or intermediate. The specific gravity of charcoal according to Violette, exclusive of pores, that is after per- fect replacement of the air contained in the pores by water, ranges from 1*402 to 2*002, according as it has been prepared at tempera- tures between 150 C. and 1500 C. respectively. 7 A knowledge of the specific gravity of charcoal is of no practical value, because the amount of matter absorbed may vary considerably according to circum- 5 Ann. de Ch. et de Phys., 3, s. 39. 1853. e Violette, op. oil Do., 340. CHARCOAL. 109 stances ; and because, owing to difference of size and form in the pieces of charcoal, there may be likewise great variation in the amount of interstitial space (Karsten). It has been accurately determined at the Prussian iron-works, that 1 cubic foot of charcoal of Scotch fir weighs from 10-3 to 10'9 pounds avoirdupois ; and that 1 cubic foot of oak or beech charcoal weighs from 13 -2 to 14-1 pounds. 8 The proportion of ash in charcoal will obviously vary with the nature of the wood. On the average it may be estimated at about 3 per cent. 9 But on reference to the Table of the composition of various kinds of wood, it will be seen that the proportion of 3 per cent, of ash may be greatly exceeded. Knowing the percentage of ash in any given wood, and estimating the average yield of charcoal at 23 per cent. 1 by weight, the proportion of ash in the charcoal may be calculated. Commercial charcoal, even when well burnt, contains a sensible amount of hydrogen and oxygen, as Bunsen and Playfair have demon- strated. 2 They analysed the gases evolved from various specimens of charcoal heated in close vessels. Their results are as follows : 1. 23-65 15-96 43-39 11-00 2. 15-96 13-62 50-10 20-32 3. 19-58 20-57 39-10 20-75 4. 35-36 14-41 29-45 14-41 Carbonic acid Carbonic oxide Hydrogen Carburetted hydrogen .... 1. Very well burnt charcoal from beechwood. 2. Well burnt firwood charcoal. 3. A Veil burnt oak charcoal O65 gramme, gave of carbon 0*47, and 70 cubic centimetres of gas, at C. and 0'76 bar. 4. Imperfectly burnt beechwood charcoal, pulverulent and of a black ish-browri colour, 0-733, gave of carbon 0-443, and 250 cubic centimetres of gas, at C. and 0'76 bar. Ebelmen has determined the composition of the charcoal of poplar and young oak, obtained by charring in piles. 3 Both were dried be- tween 140 and 150 C. The oak charcoal had been long exposed to the air. The poplar charcoal lost 5-2 per cent., and the oak charcoal per cent, in weight by desiccation at the temperature above stated. The loss by exposure to a white heat in a platinum crucible was also ascertained. The results are given in the following Table. Carbon Poplar. Oak. Loss by exposure to a white heat. Poplar. Oak. 87-22 3-20 8-72 0-86 Ash deducted. 87-98 3-22 8-80 87-68 2 83 6-43 3-06 Ash deducted. 90-46 2-91 6 -63 17-07 per cent. 13 06 per cent. Hydrogen Oxven Ashes 8 Handworterb. der Chemie, 4, 443. 9 Richard, Etudes sur 1'Art d'extraire immediateinent le For de ses Minerals, p. 48. 1 Scheerer, Handb. d. Met. p. 254. 2 British Assoc. Rep. 1846, p. 145. :$ Recueil des Trav. Scient. 2. 228. 110 CHARCOAL. Faisst has found the composition of several kinds of charcoal to be as follows : Water Carbon Hydrogen Oxygen with nitrogen Ash . . , 1, 7-23 88-89 2-41 1-46 3-02 6-04 85-18 2-88 3-44 2-46 3. 8-21 87-43 2-26 0-54 1-56 1. Beech wood charcoal from piles. 2. Hard charcoal from wood vinegar works. 3. Light charcoal from wood-gas works. 4 The proportion of hydrogen and oxygen existing in dry charcoal will depend upon the temperature at which carbonization has been effected : the higher the temperature the greater will be the amount of carbon in charcoal. Violette has analysed charcoal prepared from the same wood at gradually increasing temperatures, from that at which carbonization commences to that at which platinum melts. His most important results are contained in the following Table. 5 Temperature in Composition of charcoal produced. No. Centigrade degrees at which carbonization was effected. Carbon. Hydro- gen. Oxygen, Nitrogen, and loss. Ashes. Observations. 1 150 47-51 6-12 46-29 o-osl {The products obtained at these temperatures cannot pro- 2 200 51-82 3-99 43-98 0-23J perly be termed charcoal, 3 270 70-45 4-64 24-19 0-85 but merely desiccated wood. 4 350 76-64 4-14 18-44 0-61 5 432 81-64 1-96 15-24 1-16 melting-point of antimony. 6 1023 81-97 2-30 14-15 1-60 do. silver. 7 1100 83-29 1-70 13-79 1-22 do. copper. 8 1250 88-14 1-41 9-26 1-20 do. gold. 9 1300 90-81 1-58 6-49 1-15 do. steel. 10 1500 94-57 0-74 3-84 0-66 do. iron. 11 Beyond 1500 96-52 0-62 0-94 1-94 do. platinum. No. 3 was what the French term " tres roux," from its yellowish- red or rusty colour ; it was beginning to be pulverisable. No. 4 was black- charcoal like that of all the succeeding numbers. The results of Nos. 6 and 7 are somewhat discordant ; but it must be difficult, if not impracticable, to insure perfectly harmonious results in experiments of this kind. The wood operated on was that of Ehamnus Frangula, which furnishes a charcoal suitable for gunpowder : it was dried at 150 C. The following Table will serve as a companion to the last, and is interesting as showing the amount of carbon and other matter ex- pelled during carbonization at different temperatures. 6 4 Wagner, Jahresber. 1856, p. 457. { 322. 1851. 5 Ann. de Ch. et de Phys. 3. s. 32, j 6 Violette, op. cit. p. 325. VARIOUS MODES OF CHARCOAL-BURNING. Ill Products of the decomposition of U T ood by carbonization. Temperature in Centigrade degrees at which carbonization Solid matter or charcoal in 100 parts. Matter volatilised and carried over in 100 parts. General Total of was effected. Carbon. Gaseous elements. Ashes. Carbon. Gaseous elements. 150 47-51 52 41 0-08 _ 100 200 39-95 36-97 0-18 7-56 15-34 do. 270 26-17 10-65 0-32 21-34 41-52 do. 350 22-73 6-75 0-18 24-78 45 56 do. 432 15-40 3-25 0-22 32-11 49-02 do. 1023 15-37 3 12 0-30 32-14 49-11 do. 1100 15-32 2-86 0-22 32-19 49-41 do. 1250 15-81 1 91 0-22 31-70 50-36 do. 1300 15-86 1-40 0-20 31-65 50-89 do. 1500 16-37 0-83 0-11 31-14 51-55 do. Beyond 1500 14-48 0-23 0-29 33-03 51-97 do. VARIOUS MODES OF CHARCOAL-BURNING. Charcoal-burning is effected in the open air in piles or stacks provided with a yielding cover, in pits, in closed chambers of brick or stone, and in iron retorts heated externally like common gas-retorts. I shall not present any de- scription of the latter method, which is practised by the manufac- turers of pyroligneous acid and gunpowder, and is not specially within the province of the metallurgist. Fig. '2. Vertical section through the centre of a pile, from Karsten's Atlas, No. 301. At the foot sn tln> left the cover is shown resting on stones, and on the right it is shown resting on branches supported by forked sticks. Cliarcoal-baniing in piles or stacks. Dry level ground, well sheltered from the wind, and near a water-supply, should be chosen as the site. When the pile is circular, the bed on which it rests should have a slight inclination upwards, from the circumference to the centre. In the centre three upright stakes are driven in about a foot equi-distant from each other, , fig. 2, and so secured that they may resist pressure from without. This may readily be done by means of pieces of wood placed crosswise from stake to stake, or by suspending a block of wood wi tli in ihe stakes. Pieces of wood of equal length are piled con- centrically around the stakes, placing those nearest the centre almost 112 VARIOUS MODES OF CHARCOAL-BURNING. vertical, and giving the surrounding pieces a slight, but gradually increasing inclination. A second row, and in the case of very large piles even a third, may be stacked in a similar manner, one above another. The wood should be packed as close as possible ; and all large interstices due to irregularity of shape in the pieces should be filled with the small wood of branches. The top of the pile is covered with a layer of the same kind of small wood, placed hori- zontally and radially, so that the whole pile may have the form of a truncated cone, rounded at its upper and smaller end. Close to and round the base of the pile a row of forked sticks is driven in, with the forked ends uppermost and about six inches out of the ground. The pile is then encircled with a band of branches, resting in the forks of these sticks. This band supports the cover of the pile which has next to be applied. A row of stones or pieces of wood may be used for the sam% purpose. The surface of the pile is made more or less even by packing in here and there bits of wood or small branches. The whole pile above the band of branches is now covered, except the space at the top of the three central stakes, with turf, placing the grassy side inwards ; and if turf cannot be got, leaves or moss may be used. The turf is plastered over with a layer some inches thick of the soil which may be at hand, or, when procurable, with a mixture of the residual charcoal-dust of previous burnings and soil, moistened sufficiently with water. As a rule the cover should be most solid and thickest at the top of the pile, where it is longest and most exposed to the action of heat. The pile is now ready for lighting. It is desirable that this should be done early in the morn- ing, and during fine weather, because at first much attention is re- quired on the part of the charcoal burner ; and because it is important that the pile should be well and regularly kindled, a condition which cannot be ensured in bad weather. The space within the three central stakes or chimney is filled with easily inflammable wood, which is then ignited, and the fire is kept up by a supply of fresh wood or charcoal, until the centre of the pile has become thoroughly kindled. Any sinkings-in which may occur at the top of the pile must be made good by taking oif the cover from that part and putting in fresh wood. The chimney is afterwards well filled with small dry wood or charcoal, and effectually stopped by extending the turf and soil covering over it and pressing it well down. In this, the first or sweating stage of the process, much water condenses in the cover, and especially appears round the base or foot of the pile, which is left uncovered below the band of branches ; during this stage, without proper attention on the part of the charcoal burner, explosions are apt to occur, occasioned by the ignition of explosive mixtures of atmospheric air and the inflammable gaseous products of the carbonization. That the explosions are due to this cause would appear from the fact that they are stated never to occur when much steam escapes from the cover, and that they very frequently occur when dry and resinous wood is used. 7 Karsten, however, attributes 7 Wehrle, Lehrbuch, 1. 809. VARIOUS MODES OF CHARCOAL-BURNING. 113 these explosions to the sudden escape of steam. When the sweating stage is over, the foot of the pile is covered, and any hollows which may be found by probing with a pole are filled up. The cover in every pail is made solid and impervious to air, and the pile is left to itself during three or four days, the heat existing in it being sufficient to effect the carbonization of much of the surrounding wood. If left too long in this state the fire would be extinguished, to prevent which holes or vents are made in the cover round the pile, on a level with the top of the first row of wood (fig. 2). Thick yellowish-grey smoke at first escapes from these vents, but after a time it becomes bluish and nearly transparent. The vents are then stopped, and another row below them is opened, when the same change in the appearance of the smoke ensues. The character of the smoke indicates exactly the degree of carbonization in that part of the pile from which it issues. If necessary, after the stopping of the second row of vents, a third may be opened about nine inches or so below. These vents serve for the escape of the volatile pro- ducts of the carbonization, and not for the admission of air, which enters chiefly at openings made at the foot of the pile. When only bluish transparent smoke proceeds from the lowest vents, every part of the surface must be well covered and rendered as impervious to the admission of air as practicable. The position of the vents may be varied according to circumstances. The object of the charcoal burner should be to conduct the combustion as uniformly as possible from the top towards the bottom, and from the centre towards the circumference of the pile. By making a vent in any part of the pile, he has the power of establishing a current of air through, and, consequently, of increasing the combustion in, that part. During the process of car- bonization wood decreases considerably in volume, so that the degree of regularity in the contraction of a pile during the progress of burning is a measure of the regularity with which the process has been conducted. The cover being yielding, adapts itself to the gradually decreasing size of the pile. The pile is left at rest during a few days, after which the charcoal may be drawn, beginning on one side, at the bottom, and from this point proceeding all round, care being taken to cover up the pile as the charcoal is drawn, and to quench the latter with water. If water cannot be had, the charcoal must be covered with the dust of previous burnings, or with dry soil. If the pile were left to itself, it would in time be perfectly extinguished ; but experience has shown that the charcoal in that case is less sei vice- able than that which is rapidly extinguished. 8 The method of charcoal - burning just described is extensively practised; and in respect to yield and quality of charcoal, it is not, when properly conducted, surpassed by any other. It has the ad- vantage of not requiring any permanent construction, so that the iiniLcr may be burned on the spot on which it is felled, and thus the expense of carriage to a distance may be greatly diminished, as the 8 Karsten, Sys. d. Metall. 3. 114 VARIOUS MODES OF CHARCOAL-BURNING. wood weighs about five times as much as the charcoal produced. It may be modified in details according to local circumstances, and the practice of the charcoal burner. a c, fig. 3, represents a vertical section of half a pile, similar to that shown in fig. 2, but with a different method of supporting the cover, Fig. 3. Copied from Nos. 302 and 300 of Karsten's Atlas. namely, by boards placed horizontally round the pile, resting on wooden props, a. c b is a vertical section of the half of another kind of pile described farther on. 1. The three central stakes may be replaced by one, c, fig. 3 ; but in this case it is necessary to construct a channel from the outside of the base of the pile, by means of which burning fuel may be introduced so as to ignite the wood at the bottom of the central stake, immediately around which easily combustible matter should be placed. For this purpose imperfectly charred wood from a previous burning may be used. When the pile is well kindled, the channel must be stopped. The channel may be made, either by leaving a space between the billets at the bottom, or by making a furrow in the bed of the pile. Pieces of wood may be placed upon the bed, radiating from the centre as in fig. 4, and the channel formed by two parallel pieces as at a ; pieces of wood are next arranged concentrically, as shown in one -half of the figure, and so a firm foundation of wood is made for the pile. The outer dark ring in the cut is the dust of char- coal or breeze 9 covering the. bed of the pile. Fig. 5 is a vertical section of a pile through the centre. Around the central Fig. 4. Copied ante 2 of Klein's inflammablo 9 I use this word breeze to denote the j the French word braise, which is applied small residual charcoal obtained in char- to the residual charcoal obtained in the coal-burning, just as the coke-burner ap- plies it to the small residual coke ob- tained in coke-burning. It is derived from heating of bakers' ovens. It is drawn out, extinguished with water, and sold under the name of braise. VARIOUS MODES OF CHARCOAL-BURNING. 115 wood, such as the imperfectly carbonized pieces termed brands from a previous burning. The lower part of the cover is supported by stakes c: in the middle of each of these stakes is fixed, at right angles, a piece of wood (L Resting upon the tops of the stakes c are boards n tin- ]il;ink is nigaged in carrying up the breeze with which to complete the cover at the top ; in the left foreground are various im- plements nse.l by the charcoal burner; the scene is Kuhpolding in Havana, and the artist is Mr. Justyne, who, with the exception of the pietmvs(|ue group of figures, lias derived the materials of his drawing from the plates in the work of Klein. 1 The channel may be omitted, 1 I'clii-r Ycrkohluni: de.s Holzes in stehenden Mcilern. Von Ferdinand Kit-in. (lutlm. : i 2 116 VARIOUS MODES OF CHARCOAL-BURNING. and a central stake fixed, extending upwards about one-third of the highest part of the pile, a hollow space being left above the stake for the purpose of igniting in the manner first described. 2. The wood may be piled horizontally and radially in concentric rows, 6, fig. 3. The spaces between the pieces will be wider towards the outside than towards the centre, and these must be well packed with small wood. By sawing the wood to suitable lengths, the pieces may be so piled as to form a series of steps round the outside of the pile, which will tend to prevent the cover from slipping off. By this means of supporting the cover, the pile may be made much steeper, a condition favourable to complete carbonization of the whole mass. The outer ends of the pieces upon which the cover rests are less cooled in steep piles than in flat ones. 2 3. The pieces may be piled at first vertically round the axis for some distance, and then horizontally as in fig. 3, b. In this arrange- ment Karsten remarks that the hollow spaces are reduced to a mini- mum ; and with pieces of equal length the outer ends of the last row naturally form a series of steps by which the cover may be firmly supported. It is especially necessary, in a pile of this kind, to cover the bed with a layer of waste pieces of wood, in order that the charring may extend well to the bottom (Karsten). There has been much discussion whether it is most advantageous to stack the wood vertically or horizontally, and practical charcoal-burners are still far from unanimous on the subject. Experimental results have been advanced in favour of each method of stacking. 4. A conical cavity, lined with brick, 1 33 (4 ft. 4J in.) in dia- meter at the top, O m 50 (1 ft. 7f in.) at the bottom, and O ra 50 (1 ft. 7f in.) deep, is made in the centre of the bed. Three brick flues, 12 (4f in.) on the side, proceed from the bottom of this cavity, and communicate with the external air beyond the base of the pile. The cavity is filled with small wood and imperfectly charred pieces, and then covered with sheet iron. The construction is similar to that described by Karsten, and which is represented in woodcut No. 7. The diameter of the pile at the base is 9 metres (29 ft. 6J in.) ; the wood is sawn in lengths of O m 67 (2 ft. 2J in.), and piled vertically in three rows one above another. In every part corresponding to the pro- jection of the cavity beneath, a thick layer of soil and small charcoal is put upon the first row of wood, but in other respects the pile is made in the usual way ; care being taken to diminish the interstices as much as possible, and to stack each piece in a diametral plane passing through the axis of the pile. The fuel in the cavity is then ignited. The upper part of the pile is uncovered, and holes are opened round the base. When it is well kindled, the three flues above-mentioned are closed, the top is covered, and the process conducted as usual. From 28 to 35 cubic metres (989 to 1236 cub. ft.) of wood may thus be burned in four or five days. At Audincourt, where this process is practised, Karsten, Sys. d. Metall. 3. 55. VARIOUS MODES OF CHARCOAL-BURNING. 117 it has been found better to operate upon this quantity of wood than upon 150 or 180 cubic metres (5298 to 6357 cub. ft.) at a time, as was formerly done. 3 The advantage of this method is, that it obviates the necessity of repeatedly charging the centre of the pile with fresh wood during some time after lighting ; but it is not adapted to char- coal-burning in forests, where the site of the pile is constantly changed ; nor can it well be employed in very moist soils, on ac- count of the difficulty of kindling the wood in the cavity. More- over, another advantage in these small piles is, that the charcoal burner can more easily manage them than large ones. 5. When carbonization can be economically conducted in one spot, a permanent bed of brickwork may be constructed, with an arrangement Fig. 7. Copied from No. 297 of Kai> ten's Atlas. for collecting tar and pyroligneous acid. Fig. 7 4 represents a vertical section through the centre of such a bed a. It slopes towards the centre, and not from it as in the ordinary pile. At the bottom in the centre, there is a cylindrical cavity, from which proceeds a channel 6 to a reservoir c, provided with a moveable cover g, such as a plate of iron ; d is a square plate of iron, of which the corners are rounded. The tar and other liquid products condense and trickle down between the sides of the plate d and the brickwork, and flow into the reser- voir c. The pile is stacked up in the usual way. Care must be taken to prevent access of air through the channel b. (5. Charcoal-burning in large rectangular piles is extensively prac- tised in some parts of Sweden. Af Uhr has published a detailed account of the process, of which I avail myself in the following description. 5 The ground on which the pile rests should be solid, dry, and even, free from roots and stones, and should slope gradually from one end to the other, about 18 inches in 24 feet. The base is usually a square of 20 feet on the side or upwards. Upon the ground, in the direction of the slope, three poles, a a a, fig. 8, are placed parallel to each other, one in the middle and the other two about 2 feet from the sides of the pile respectively. These poles may be i ' < >r 7 inches in diameter at one end, and 4 or 5 at the other ; the thickest ends should be placed at the upper part of the slope. At the lower end or foot of the pile are firmly fixed two posts, b b, inclining some- what towards the pile, and supported by props on the outside, b\ fig. 9. , Trav. Scieiit. 2. 10G. 1 Kar.stni's Atlas, No. 297, Sys. d. Met. 8. p. 51. 5 Aiilritun^ zur vortlu'illiafteu Ver- kohlung, p. 35. Sec also Notes by Du- rocher, Ann. des Mines, 5 ser. 9, p. 359. 1856. I have also received a MS. de- scription from my friend Andreas Grill, of Sweden. 118 VAKIOUS MODES OF CHARCOAL-BURNING. At this end and all along the bottom, the longest and thinnest pieces of wood are placed crosswise on the three poles forming the foundation. The largest pieces should be placed in the centre, and towards the high end or back of the pile, where the heat will be greatest. The large and small ends of the pieces should be placed alternately, so that Fig. 8. Plan of the Pile. Copied from Plate 3 of Af Uhr's Treatise. The inclined props on each side have been omitted. the pile may be level and compact ; the sides formed by the ends of the pieces of wood should be even and vertical. It is not necessary to split the wood, for Af Uhr found that pieces 23 feet long, 2 feet in diameter at one end, and rather more than \\ at the other, were as thoroughly carbonized in piles of this description as the smaller pieces near them. It is hardly necessary to remark that wood of these dimensions is not used for charcoal-burning when it can be more pro- fitably disposed of as timber. A hole, A, fig. 9, 6 inches square, should be left through the pile from side to side, near the foot, at about 9 inches from the top. At the back the lowermost piece of wood, /, should be let into the three poles underneath. Upon this piece at about 3 feet from each end is placed, at right angles, a wedge-shaped piece of wood, <7, about 3 feet long and 5 inches thick at the thick end, which is directed outwards. Upon the wedges a second piece of wood,/', is placed transversely, and so alternately wedges and wood until the pile is com- pleted, as shown in fig. 9, which is a vertical section on the line A B, fig. 8. By means of these wedges the transverse pieces of wood,/,/', &c., VARIOUS MODES OF CHARCOAL-BURNING. 119 are supported, and from the bottom to the top of the pile, at this end, is formed a series of parallel openings, which extend inwards as far as the thin ends of the wedges, and which are intended as draught open- ings. The upper surface of the pile is made level by covering it with Fig. 9. Vertical Section along the line A B, fig. 8. Copied from Plate 4 of Af Uhr's Treatise. a layer, 4 inches thick, of small pieces of wood. The larger openings in the sides of the pile must be packed with suitable pieces of wood, and the smaller ones with brushwood. The upper surface of the pile is next covered with fir branches or twigs of sufficient thickness to feel soft under foot. Over the foot and the upper transverse pieces of wood at the back the brushwood is bent down, and the depending Fig. 10. Side Elevation. Copied from Plate 4 of Af Uhr's Treatise. ends inserted in the interstices of the pile. The sides of the pile are also covered thinly with brushwood, by sticking the thick ends into the spaces between the pieces of wood, so that the other ends hang 120 VARIOUS MODES OF CHARCOAL-BURNING. down. The pile is thus prepared to receive its outer carbonaceous or black-coating of breeze from previous burnings, mixed in a greater or less degree with soil, and moistened with water to the consistency of thick paste. Two posts, c c, fig. 8, are firmly fixed in the ground in front, and three, d d d, on each side, supported by props. Chips and twigs are carefully swept away round the bottom of the pile ; for if they were left, they might cause irregularity in the admission of air into the pile. The black-coating is first applied along the top, over the brushwood, to the depth of 6 or 8 inches. At the back of the pile a piece of wood, A, fig. 8, is fixed across the projecting ends of the poles, a a a, &t a, distance of about 6 inches from the wood forming the back. The space thus left is filled with breeze, which is gently pressed down and continued to be applied until it forms a wall 7 or 8 inches above the piece. Upon this wall is placed a piece of wood, ', fig. 9, extending from one side of the pile to the other. A second wall of breeze is built on this wood, and so on until at length the whole of the back of the pile is well coated. These transverse pieces of wood, i, are supported by props, m, as shown in fig. 9. On the sides of the pile at the bottom, stones about 5 or 6 inches high are laid at intervals, and on these is built up a wall of wood, which is supported by the upright posts, d d d, which are firmly stayed by the inclined posts, d' d' d', fig. 1 0. 6 There should be a clear space of 5 or 6 inches between the inner side of the wall of wood and the sides of the pile formed by the ends of the wood intended for carbonization. This space is filled with breeze simultaneously with the building up of the outer wall of wood. The front of the pile is coated in a similar manner. By means of the stones on the sides and the crosspiece of wood, A, at the back, suitable vents may be opened round the bottom of the pile. Care must be taken to stop any open- ings in the angles at the back, formed by the ends of the transverse pieces of wood which support the cover and the outer side-walls of wood. The pile may now be lighted through the fire-hole, A, fig. 10, on the side which at the time may be least affected by the wind, the opposite end being stopped, and kept so for some time after %nition. When the sur- m Fig. 11. . the opposite side. The round spots rounding wood is well kindled, small holes between these lines indicate the a * P . openings through the roof or cover, are made trom the root down into the fire- hole, A, successively along it to draw the fire towards the opposite end, as indicated by the direction of the arrow in the annexed cut. When the fire has reached about one-third of 6 It will be borne in mind that fig. 9 is not a perspective drawing. The bed on which the pile rests is supposed to be raised above the surrounding ground. VARIOUS MODES OF CHARCOAL-BURNING. 121 the length of the wood as at s, fig. 11, all the vents on this side, in- cluding the fire-hole, are closed. Other similar vents are then made on the opposite side, and these are also closed as soon as the wood on that side is ignited. With dry wood it may require a day to light the pile ; but with wood which is not dry, the lighting may occupy a day and a half, or even two days. The object should now be to cause the fire to extend equably through the wood towards the caver above, and from thence forwards and downwards to the bottom of the pile. With this view, immediately after the closing of the end of the fire-hole opposite to that at which the lighting took place, five or six vents are made with a spade crosswise in the cover, near ??, fig. 10, between the two middle posts which support the walls of wood on the sides of the pile. As carbonization gradually advances towards the vents in the cover, it proceeds also towards the bottom of the low end of the pile. Wherever the wood feels loose under foot, on walking over the top, it is a sign that carbonization is completed, and the cover over that part of the pile should be beaten down. Owing to the contraction of the wood by carbonization, the cover gradually sinks to a certain extent ; and this sinking should take place uniformly across the pile, from one end to the other, without the formation of any steep ridge or irregular projections. If the wood is dry when stacked, carbonization always goes on better and more easily, and after the lapse of five or six days may have extended to the bottom of the foot or low end of the pile. As soon as the vents in the cover emit light smoke, they should be closed, and others made in the cover at the highest part of the pile ; and these must be stopped as soon as the smoke which issues from them presents the same light appearance. The whole cover should then be well stamped down, to check the passage of air through it. Other vents, about 12 or 18 inches apart, are next made along the upper surface of the uppermost transverse piece of wood at the back of the pile ; and as soon as light smoke escapes from them, they in turn must be closed, and others made at y, fig. 9, along the two or three next underlying transverse pieces of wood, so that two or three rows of vents may exist one above another. Vents are at last made round the bottom of the pile, as well on the sides as at the "back under h ; and when it appears that the lowest stratum of wood is carbonized, all the vents are closed and every part of the pile rendered as impervious to air as practicable. The vents are made with a shovel through the coating of breeze into the twig covering ; and sometimes they are made deeper by means of an iron bar. When the charring has extended over the whole pile, the cover is gradually taken off, beginning at the foot, and the charcoal lightly quenched with water. It is then removed to a store-house, where it remains until winter, when the roads become suitable for its transport to the furnaces. The following modifications of this method of charcoal-burning are described by Karsten and other authors. The wood, sawn into lengths of about 8 feet and uncleaved, is placed transversely so as to form a pile about 24 feet long, increasing in height from one end to the other. The bed may be horizontal or slightly inclining upwards 122 VARIOUS MODES OF CHARCOAL-BURNING. toward the high end. Fig. 12 represents a side view o.f a pile partly in section. Immediately upon the bed, which is covered as usual with a layer, several inches thick, of small charcoal, long cleaved pieces or straight branches are laid in the direction of the long axis of the pile, and upon these the wood is placed across at right angles. Thinner and dry pieces should be put at the low end where the pile is kindled. All round the pile, at a distance of about 6 inches, upright posts are driven iinnly into the ground, at intervals of 2 or 3 feet. Boards are then fixed on the inside of these posts, and the space between them and the sides and ends of the pile is rammed with moistened breeze. The low end may be about 2 feet high, and the high end about 7 or 9 feet. The pieces are piled without reference to uni- formity in size, but the largest should be put towards the high end. fe^gg! Fig. 12. Copied from Nos. 304 and 305 of Karsten's Atlas. Perspective view, partly in section. The interstices between the wood should be packed with the small wood of branches. The top of the pile is covered with leaves and faggots, and lastly with small charcoal in the usual way. The pile is ignited at the low end, where a hole c is left for the purpose. As soon as light blue smoke escapes from the cover at this end the hole is closed. Three or four holes, an inch in diameter, are then opened through the breeze-coating, about 15 inches from the ground, as seen at a a, fig. 12. When, after some time, only light blue smoke makes its way here and there through the cover above, it is a sign that carbonization is so far finished at the low end of the pile that air should be excluded from this part. The holes at a are then closed, and three or four fresh ones opened at 6, 6 or 7 feet farther on, about 2 feet from the ground, and 12 or 15 inches apart, and so the process is carried on to completion. The next holes should be made some inches higher, in proportion to the increasing height of the pile. After the closing of the holes at , two or three fresh ones of the same diameter should be made underneath, quite close to the ground. But these openings ought only to remain open until the appearance of a light scarcely perceptible smoke. W hen the process has proceeded so far that openings have been made at 8 feet from the high end, charcoal may be begun to be drawn from the low end, when it will rarely be found to be ignited. W 7 hen the high end of the pile VARIOUS MODES OF CHARCOAL-BURNING. 123 is actively burning, the charcoal from half of the pile towards the other end should be removed. 7 7. A pile, similar in form to the last described, is made by placing the wood lengthwise, and inclining upwards towards the high end. Steins which are more than 9 inches in diameter must be cleaved. Fig. 13. Longitudinal vertical section. Copied from No. 307 of Kursten's Atlas. The interstitial space is less than in the last method ; and owing to the direction of the wood, combustion is more rapidly propagated through the pile. In large piles of this form, the products of distilla- tion may be collected by inserting cast-iron pipes, p, in the high end, and connecting these with a barrel to act as a receiver. The pipes must be kept cool by water flowing over them. The process of charcoal-burning has been divided into three stages. In the first or sweating stage (Schwitzen) desiccation of the wood is effected ; one portion of the steam produced escapes through the heated part of the cover, and another condenses and appears especially round the open foot of the pile. The hygroscopic water thus evolved is increased by that derived from the wood in combustion. When this stage is completed, the heat existing in the central and upper part of the pile is sufficient to continue the carbonization of the surrounding wood to a considerable extent; and, accordingly, the supply of air should then be very much reduced by closing, in a greater or less degree, the foot or base of the pile. This constitutes the second stage (Treiben). Currents of air are next established successively in various parts of the pile, by making suitable openings at the base, and other openings or vents in the cover above ; and so, by a carefully regulated combus- tion, the wood at the circumference becomes at length converted into charcoal. This is the third stage (Zubrennen). Admission of air to every part is then stopped as completely as possible, and the pile is left to itself for a day or so, when the charcoal is withdrawn and ex- tinguished in the manner described. Chinese methods of charring in pits. 8 Two methods of making wood- charcoal are practised in China. When the soil is sandy, charring is effected in pits ; and when, on the contrary, it is clayey and the 7 Karsten, Sys. d. Metall. 3. 71 et seq. I have availed myself of Karsten's de- scription condensed as much as possible. 8 De la fabrication du Charbon de Bois en Chine. Par M. Kovanko, Major au Corps des Inge'nieurs des Mines. Annu- aire du Journal des Mines de Russie, annce 1838, St. Petersbourg, 1840, p. 375. I have given nearly a literal translation. 124 VARIOUS MODES OF CHARCOAL-BUBNING. Fig. a. Fig. b. locality is suitable, arched chambers are excavated, and in these the wood is carbonized. The last method is preferred by the Chinese, who have carried it to such a degree of perfection that all the small branches and twigs are carbonized without losing their form. Tfie first method. The pits are circular, and are never deeper than I m 8 (6 ft.), but they sometimes exceed 4 m 3 (14 ft.) in diameter. 9 In the annexed woodcuts the pit is shown in vertical section and in plan. B is a round chimney of which the base is from O m 2 (8 in.) to O m 35 (14 in.) below the bottom of the pit ; it rises about l m (3 ft. 3J- in.) above the ground, and is connected with the pit by an oblong opening, C, not ex- ceeding O m 35 (14 in.) in length, and from O m 05 (2 in.) to O m 10 (4 in.) in depth: the dimensions of this opening depend upon the quantity and the size of the wood to be charred. In pits 4 m 27 (14ft.) in diameter, the chimne} 7 at the base is O m 35 (14 in.) in width, and narrows upwards to O m 18 (7 in.) in width. In that part of the pit which is opposite the chimney is an inclined conical channel, D, from the lower end of which a vertical cylindrical chimney, O m 1 (4 in.) in diameter, rises to the surface. The axis of the conical channel ought to have such an inclination that its lower or narrow end is about equidistant from the bottom and upper edge of the pit. The bottom of the pit is covered with a bed of dry branches, upon which the wood is piled vertically, taking care, as usual, to leave as little interstitial space as practicable. When the pit is filled, the wood is covered first with small branches and then with a layer of soil suffi- ciently thick to be impervious to smoke. The wood is lighted through the opening opposite the chimney. Occasionally, for the sake of greater facility, a small opening is made above at about O m 25 (10 in.) from the chimney, but which is closed immediately when smoke begins to escape. If the smoke is copious, the pit is covered with stones, a small opening only being left to promote combustion. Five days after lighting the smoke begins to get purer, and when it has become quite transparent the pit and chimney must be hermetically closed. Five or six days suffice for the complete extinction of the charcoal, after which the pit may be opened. Experience in China has shown that the more freshly-cut the wood, the less is the loss : 100 pounds of freshly-cut wood are stated to yield from 30 to 35 of charcoal, which cannot Fig. 14. Chinese Method of Charcoal-burning. Copied from Kovanko's engraving. Fig. b, plan of the pit. Fig. a, vertical section on the line E F, fig. I 9 In reducing the French to English measures I have avoided fractions and given the nearest whole numbers. VARIOUS MODES OF CHARCOAL-BURNING. 125 possibly be correct. When a large quantity of wood is to be charred, the pits are made wider, but not deeper. The second method. The arched chamber excavated in the clayey ground is I m 4 (4 ft. 7 in.) high and 4 m 3 (14 ft.) wide (see the wood- cut, fig. 14). A lateral chimney is formed just as in the first method. In the side of the chamber opposite the chimney there is a conical channel of which the base is directed towards the chamber and extends nearly to the arch, while the narrow end is about intermediate between the bottom and the arch above. The chamber is entered by a low door, which is closed with stones as soon as the charging is completed. The wood is placed horizontally with the usual precautions. The kiln is lighted through the channel opposite the chimney; and when the smoke begins to issue from this channel, it is closed with stones, a very small opening only being left for the passage of the air. At the end of the charring the same course is followed as in the first method. When a judgment respecting the stage of the process cannot be formed from an examination of the smoke, one or two freshly-cut sticks of the size of the finger are placed across the chimney, and when these sticks, impregnated with oil, are dry and their fracture is black, it is a certain proof that carbonization is ended. Charcoal-burning in ovens or kilns at Dalfors Iron - Works, Sweden. I am indebted to my friend Andreas Grill for a description of this pro- cess, and the drawings from which the annexed woodcuts have been Fig. 15. Fig. 16. 12G VARIOUS MODES OF CHARCOAL-BURNING. Seal* cf Swedish Feet. 5 . . .10 of English Feet af> _ 30 Fig. 17. Plan and Horizontal Section on the Line E F. executed. The works belong to his brother-in-law, the Hon. W. Didron, who expresses himself well satisfied with the process, a, a, body of the oven, enclosed above by the arch b, b, which rests on the side- walls b' b', supported by buttresses b" b". The oven is filled as closely as possible with layers of wood, first through the lower openings, c c, and then through the higher ones, c' c' : these openings are afterwards closed as tight as possible by cast-iron doors, and all fissures are stopped up with clay. During the whole time of charring much care is required to prevent the admission of air through the sides of the kiln. A sunk passage, c7, leads to the fireplace, H to 'PC 'SI ^B U9J\[ ^ g 3 ^ PUB BATJa ? CO lO Observations. The horses were employed to draw the wood to the oven and the charcoal to the store-sheds. The bgtter yield was obtained when the wood was dry, and the worse when it was not sufficiently dry. These results were obtained in an oven exactly similar in construction to, Taking out. 11 'S9SJOJI c, U9H to to -a PUB SAXI M - ganooo pnB SJJFQ S o s i; Charring. II 8aS ao H : : U9H r ? < ?5 ea-a pro? BABCT ? ? c< to Charging. , M ,, a 89SJOH : "* a8 W O5 -- n,u i^qS^ PUB SjCV([ .g 1 II II 128 YIELD OF CHARCOAL. These ovens are not much used in Sweden, and can only be economically employed in peculiar localities, where the works are situated on the banks of a lake or river in the vicinity of a large tract of forest, of which the wood may be readily conveyed by floating. Yield of Charcoal. The yield will vary with the nature of the wood, with its age and condition especially as to dryness, and with the mode of conducting the process of carbonization. Of all these causes of variation, the last is the most influential. The yield of charcoal may be computed by measure or by weight. In practice, estimation by measure is usually preferred, because .the proportion of water in charcoal is far from constant ; and unless special precautions were taken to expel this water before weighing, a very erroneous result might be obtained. But such precautions on the large scale would be impracticable. In order to ascertain the yield by weight, the data required are simply the weight of the charcoal and the weight of the wood from which it is produced. But if by measure, several methods of computation may be adopted. One German writer enumerates not less than six different modes of estimating the yield of charcoal by measure. 1 One of the most obvious methods, and that which is in use in this country, is to measure the charcoal and the wood, inclusive of the interstices between the pieces, the wood being supposed to be suitably stacked for the purpose. Thus the wood may, according. to the usual practice, be estimated in cords, and the charcoal in bushels ; and from the data thus obtained, the relation, in cubic feet, between the charcoal and the wood may be readily found. A cord is a pile of wood 4 feet broad, 4 feet high, and 8 feet long ; and a bushel is 1-2836 cubic feet. When yield of char- coal by volume is mentioned in this work, the preceding method of computation is referred to, if not otherwise stated. A second method of estimating yield by volume is, to compare the amount of charcoal as determined by measure, inclusive of the interstices between the pieces, with the amount of wood, exclusive of the interstices between the pieces ; in other words, to compare the apparent volume of the charcoal with the actual volume of the wood. The amount of interstitial space in a closely packed pile of wood, consisting of uncleaved stems, may be taken at about 30 per cent, of the mass ; but in the case of cleaved wood or branch wood, this space may be estimated at from 40 to 50 per cent., or even 52 per cent. 2 The statements of different authors do not, as might be expected, exactly agree as to the amount of interstitial space. At Hiflau it is estimated at 25-5 per cent, in wood, and 36 per cent, in charcoal. 3 A third method is to compare the actual volume of the wood with the actual volume of the charcoal, that is, exclusive of in- terstitial space both in wood and charcoal. In reading German works on charcoal-burning, it is necessary to remember these differences in the estimation of the yield of charcoal by volume, otherwise much confusion and error may be the result. In some of these works elaborate formulae are given for the estimation of the cubic contents 1 Klein, op. cit. 155. 2 Scheerer, Lehrb. 1. 237. 3 Ann. d. Mines, 3. s. 7. G. YIELD OF CHARCOAL. . . f 120 of piles of different forms, but it is presumed t^at/ltlje reader is suffi- ciently acquainted with the mensuration of solids not to require any special formulae of this kind. Yield by Volume. In Sweden the average yield of charcoal is estimated at 63*2 per cent, of the volume of the wood of Pinus picea and Abies excelsa. 4 The extremes were 50-5 and 75 per cent. Comparative trials have been made in the same country to determine the difference of yield between piles in which the wood is stacked vertically, and those in which it is stacked horizontally. In the former the yield amounted to 67 per cent, and in the latter to 75 per cent, in volume; but these are considered as maxima, the ordinary yield being from 60 to 65 per cent. 5 In Upper Silesia the following yields have been obtained : 52*6 per cent, on the average from stem wood, and occasionally 60 per cent. ; 42 -7 from branch wood, and 39 -5 from, root wood. 6 Lampadius states, that in the Saxon Erzgebirge the yield from the stem wood of Pinus picea, in the best managed charcoal burning in round piles, is as much as 80 per cent, in volume. 7 At Hiflau, Styria, the yield from long piles of fir wood was found to be 76*8 of large charcoal, and 2 1 7 of small, making a total of 79 '5 per cent, in volume. 8 At the same place the yield from large round piles has reached 86-2 per cent. 9 In some cases the volume of charcoal is stated to have exceeded that of the original wood even to the extent of 28 per cent. The discordancy between these results is so great, that it cannot be attributed to the causes of variation previously mentioned ; it must be due either to error of observation, or to a variable or fallacious method of computation ; and that the latter may in some instances have been the true cause, will appear from the following remarks. The volume of the wood may have been deduced from the cubic contents of the pile as determined by admeasurement. Now the amount of inter- stitial space may vary greatly in different piles, not only with the degree of compactness with which the wood is stacked, but also the size and irregularity in shape of the pieces. The amount of the space, as Karsten remarks, could only be correctly determined by filling the interstices with sand, and afterwards measuring its volume ; but no such determination seems to have been made, at least on a large scale. In respect to the volume of charcoal, the amount of interstitial space may vary considerably, especially from the greater or less degree of splitting in the wood during carbonization. W r hen the wood is not sufficiently dry, splitting may take place to a very sensible extent ; and probably in this way the statement that the volume of charcoal has, in some cases, exceeded that of the original wood, may be explained. We know certainly that wood contracts sensibly in all dimensions by carbonization, and in length from 11 to 12 per cent. Cork, however, which is bark and not wood, increases in volume by 4 Af Ulir, Sclieerer, Met. 238. Mr. Grill of Sweden confirmed this state- ment to the author. 5 Durocher, Ann. d. Mines, 5. s. 9. 363. G Wittwc-r, Karsten's Archiv, 2. r. 24. 293. l Hiittenkunde, 1827, 48. 8 Ann. d. Mines, 3. s. 7. 6. The spe- cies of fir is not stated. Op. cit. 18. K 130 YIELD OF CHAECOAL. carbonization. 1 The degree will depend upon the kind of tree, its age, the part of the tree, and other circumstances. Young wood and branch wood contract more than old or mature wood and stem wood. In general, coniferous wood contracts less than other wood. Fir wood, when young and soft, loses from 46 to 47 per cent, in actual volume by carbonization, but when old and strong at least 44-5 per cent. In the case of non- coniferous wood, the hard kinds contract considerably more than the soft : the young wood, the branch wood of the softer kinds, and the harder kinds of wood, lose 52 per cent, in volume, whereas the mature wood of the softer kinds loses only 50 per cent. A yield of 55 per cent, in volume from coniferous wood, and 48 per cent, from non-coniferous wood, may be regarded in general as favourable. 2 Yield by Weight. The yield, by weight, of charcoal has been often determined ; but as might be expected, there is great discrepancy in the results which have been published on the subject. Various causes influence the yield in a marked degree, so that unless the precise con- ditions under which the charring has been eifected are stated, the ob- servations of one author cannot well be compared with those of another. The weight, however, may be considered as ranging between 15 and 28 per cent, of that of the wood. In different localities in the north- east of France, the yield from circular piles, containing from 60 to 90 cubic metres of wood, has been found to range from 17 to 21-33 per cent, in weight. 3 The kinds of wood employed were beech, oak, poplar, willow, and hornbeam. It is estimated in France, that with good charcoal burners an average yield of not more than 19 per cent, should be calculated on. 4 In Belgium, from wood of 15 to 20 years' growth, a yield of 15 to 17 per cent, is obtained ; and if the charring be slowly conducted, it may amount to 20 or 22 per cent, at the most. The wood employed consists of -f to of the hard kinds. 5 In Sweden, the yield from Finns sylvestris and picea by charring, in rectangular piles, has been found, according to Af Uhr, to range from 20 to 28 per cent. 6 A yield of 22 per cent, in weight may be regarded as very favourable. 7 The average yield of twenty-six recorded observations on charring in piles in different localities is, in round numbers, 23 per cent., the extremes being 19'5 and 28'0. 8 By carbonization on the large scale, in iron vessels, heated from without, Lampadius obtained a yield, from air-dried Pinus picea, of about 27 per cent, in weight. 9 The author is indebted to his friend and former pupil Mr. C. B. Hambly, who is engaged in the manufacture of pyroligneous acid, for the follow- ing information on this subject. The wood used consisted of three- fourths oak and one-fourth of mixed beech, ash, &c., and the distillation was eifected in retorts 3 feet 6 inches wide and 5 feet long. The yield 1 Karsten, Eisenhiittenkunde, 2, 267. 2 Karsten, op. cit. 2, 268. 3 Sauvage, Ann. d. Mines, 3. s. 11. 359. 4 Tr. de la Fabrication de la Fonte, &c. Par Flachat, &c. Part 1. 112. 5 Valerius, Tr. The'or. et Prat, de la Fabric, de la Fonte, 232. 1851. 6 Op. cit. 7 Karsten, Eisenhiittenkunde, 2 part. 286. 8 Scheerer, Lehrb. 1. 236. 9 Grundriss einer allgom. Flatten- kunde, 1827, p. 48. RESULTS OF CHARCOAL-BURNING IN PILES. 131 in weight varied from 25 to 27 per cent. ; it was deduced from the distillation of 328 tons 6 cwt. of wood. Influence of Temperature upon yield. Numerous experiments by Karsten have proved that the more slowly or what is equivalent, the lower the temperature at which the charring is effected, the greater will be the yield. The following Table contains his results. 100 parts by weight of the following kinds of wood gave of charcoal By rapid charring. By slow charring. Yoiiii (y oak 16-54 15-91 14-875 14-15 13-12 13-65 14-45 15-3 13-05 12-2 12-15 14-25 14-05 16-225 15-35 15-52 13-75 13-33 13-4 17-0 14-65 25-6 25-71 25-875 26-15 25-22 26-45 25-65 25-65 25-05 24-70 25-1 25-25 25-0 27-725 24-75 26-07 25-95 24-60 24-60 27-95 26-45 Old do Youii" 1 beech (Fagus sylvatica) Old do .......... Youi! * hornbeam (Carpinus Betulus) .... . . Old do Youu alder (Alnus orlutinosa) . Old do Youn- birch Old do Birch of a post which had stood over a grave above 100 Old do Old do Youn ' Scotch fir (Pinus sylvestris) Old d*o '. Lime (Tilia EuropsBa^ Rye straw Dried fern Reeds ]Mca.n 14-42 25-69 The wood operated on in these experiments had been previously well air-dried. 1 Karsten's results have been confirmed by Violette. 2 Ebelmen has determined the effect of rapid carbonization upon green and dry wood respectively. 3 His conclusion is, that the yield of char- coal from dry wood, as compared with green, is in excess proportionate to the degree of desiccation. Ebelmen also found, that when different weights of the same wood, in the same state of desiccation, are exposed to a constant temperature in vessels of the same nature and the same capacity, the proportion of charcoal obtained increases up to a certain limit with the weight of wood operated on. Illustrative results of charcoal-burning in piles. The following Table has been compiled by Beschoren from actual results. In the first column the yield is estimated by weight ; in the second column it is stated according to the method of computation by measure first described ; and in the third column it is stated according to the second method of computation. System. 3. 34. 2 Ann. do Ch. et do Phys. 3. s. 32. 315. 1851. 3 Recueil des Trav. Scient. 2. 178. K 2 132 RESULTS OF CHARCOAL-BURNING IN PILES. yield of charcoal. Weight from 100 of wood. From 100 cubic feet of wood in apparent volume. From 100 cubic feet of wood in actual volume. ] Oak 21 302 71 ' 842 98-672 9 Do 23 447 74-299 102-009 3 Beech 22 661 73 029 100-369 4 Birch 20 945 68-518 94-189 5 Hornbeam 20-575 57 197 78-584 o Scotch fir (Kief or) 25-029 63-561 87-157 Mean 22 355 68-195 93-645 The charring was effected in piles in which the wood was stacked vertically, in either two or three layers ; and all the observations seem to have been made with great care at Eisleben, under the immediate direction of Beschoren, who styles himself charcoal manufacturer. 4 The oak and beech were from trees 150 to 200 years old, the birch' and hornbeam from 50 to 60, and the fir from 70 to 80. The following summary of the actual results obtained by Beschoren may be interesting in a practical point of view, as showing the time required to effect the charring of given quantities of wood in piles. The numbers in the first column correspond to those in the first column of the last Table. Weight in pounds Prussian. 5 No. When lighted. When cooled. Dura- tion of charring process. Weight of Wood. Wood consumed in rilling the pile from time to time to be deducted. Actual weight of wood from which the charcoal was derived. Yield of charcoal. Days. 1 Sept. 4 Sept. 16 13 68035 120 67915 14467 2 7 20 14 67595 205 67390 15801 3 .. 1 12 12 47685 102 47583 10783 4 Aug. 30 12 14 47630 . 47630 9976 5 ., 7 Aug. 17 11 52470 80 52390 10780 6 Sept. 27 Oct. 15 19 55660 1983 6 53677 13435 The weather was unfavourable during the charring of No. 6. Too much air entered the pile, owing to the dryness of the ground. The large yield is to be explained by the fact that the wood was much drier than that charred in the preceding experiments. In No. 5 the wood was not nearly so dry as in the first four experiments. The apparent volume of wood operated on ranged from 1149 to 2025 cubic feet (Prussian), and the actual volume from 836*02 to 1476*77 cubic feet. 4 Versuche iiber das Ausbringen an I avoirdupois. Holzkohlen aus verschiedenen Holzsor- | 6 This includes a considerable amount ten. Bergwerksfreund, 3. 1. 1840. I of brands, i. e. imperfectly carbonized 5 1 pound Prussian 1 031 pound j pieces. PRACTICAL DIRECTIONS IN CHARCOAL-BURNING. 133 Summary of practical directions in charcoal-burning. The \vood should be of mature growth, neither too old nor too young. It should be felled when most free from sap, that is during winter. It should be partially or wholly barked and air-dried for some months before burning. Experience teaches that the best result is obtained when the wood is moderately dry. If too dry, the combustion is too quick, and not easily regulated by the charcoal-burner, carbonization proceeds irregu- larly, charcoal is uselessly consumed, and as a consequence its quality is injured, and the yield diminished. In this case the pile must be as flat as practicable, and the cover must be made thicker and more dense, so as sufficiently to reduce the supply of air to the pile. If, on the other hand, the wood is too moist, the process is considerably prolonged, more care and labour are required, but the charcoal pro- duced is sounder and of better quality than in the last case. The steam must be allowed freely to escape, by either partially removing the cover, or diminishing its thickness and solidity at the upper part during the sweating stage. 7 Rotten and worm-eaten wood should be avoided, as the charcoal from it is so bad as to be unfit for smelting. By long exposure to the action of water, whether by floating down livers or remaining unprotected from heavy rains, wood is sensibly deteriorated for charcoal-burning. Immersion, however, during ten days or a fortnight occasions no injurious effect. 8 If the ground on which the bed is made is too porous, too much air may find its way into the interior, and so cause unnecessary waste of charcoal. In this case the bed should be rendered less pervious to air by making it more solid ; and when practicable, by covering it with a layer of the residual charcoal dust from previous burnings. A clayey ground is bad, as it may become fissured by the heat evolved in the process; and the fissures may serve as channels for the admission of air, when the same evil would occur as in the last case. In moist ground, a foundation of wood covered with soil should be first laid, and on this the pile should be raised. If a sheltered situation cannot be found, the pile must be protected from wind by hurdles or other suitable expedients. Exposure to wind will obviously tend to prevent regularity in burning. Any hollow spaces caused by the burning away of the fuel used in lighting the pile should be replenished with wood or charcoal, or the imperfectly charred pieces from previous burnings, or a mixture of both, which should be pressed well in by means of a pole ; and the surface of the pile should be sounded from time to time, to ascertain if hollows exist. Any fissures which may appear in the cover should be stopped. There is nothing fixed in respect to the dimensions of piles ; they vary in diameter from 10 feet to 50 and upwards. Most frequently the diameter is from 20 to 30 feet. In height they vary from \ to of the diameter measured at the base. 9 Charcoal should not be used in blast furnaces or forges immediately after burning, as it has been found, to improve by keeping at least ~< Helmest, Erdinann's Journ. 4. 230. 8 Af Uhr, op. cit. M. 9 Schecrer, Lebrb. 1. 221. 134 BURNING IN CIRCULAR AND RECTANGULAR PILES. during several 'months. With the same charge of ore and flux in an iron smelting furnace, the same quantity of freshly burnt charcoal was found to be less effective than charcoal which had been kept during two years well under cover. 1 T/ieory of charcoal-burning in circular and rectangular piles. From the mode of conducting the process of charcoal-burning in piles, it might be inferred that the combustion is propagated from above downwards, and from the centre to the circumference. Ebelmen has given an experi- mental demonstration of this fact. 2 A pile containing 30 cubic metres (1059 cub. ft.) of oak, beech, and fir, in pieces O m 70 long (2 ft. 4 in.), was made in the usual way. In the centre was a chimney O m 25 (10 in.) in diameter, extending from the bottom to the top of the pile, and around it the wood was stacked in three rows, one above another; the large pieces being placed in the centre and the small outside. The diameter of the pile was 7 metres (22 ft. 11 J in.), and the height about 2 metres (6 ft. 7 in.). It was covered all over as usual with soil and breeze. It was lighted in the morning by putting ignited charcoal into the chimney, which was left open for some hours. Vents were made all round the bottom of the pile, and remained open during the whole process, to supply air for the combustion. When the pile was sufficiently ignited, the chimney was filled with small wood and then closed. In the evening the vacant space caused by the burning away of the wood in the chimney was filled with breeze. This was again done next morning. In the course of the day vents were made in the covering of the pile near the top. The smoke which escaped from them was white, thick, and copious ; but after some hours it became bluish, almost transparent, and much less abundant ; when the charcoal-burner made fresh vents, about 20 (8 in.) or 25 (10 in.) centimetres below those above. On the third day, when the vents were l m 20 (3 ft. 11 J in.) above the ground, half of the pile was removed, and the ignited wood and charcoal were extinguished with water. The annexed diagram shows Fig. 18. Copied from Ebelmen's figure. the condition of the pile at this time. The charcoal was all contained within the space produced by the revolution of the plane A R s P round 1 Karsten, Sys. 3, 45. 2 Recueil ties Travaux Seientifiques de M. Ebelmen, 2, p. 104 et seq. I have as far as practicable literally translated Ebelmen's language. BURNING IN CIRCULAR AND RECTANGULAR PILES. 135 the axis R s. This space represents nearly an inverted truncated cone, of which the radius of the small base next the ground is about O m 40 (1 ft. 4 in.). In the rest of the pile the wood was unchanged, the pieces being only blackened on the surface by tar, and exhaling an empyreumatic odour ; on sawing them across it was evident that they had not even begun to undergo desiccation. The greater part of the charcoal contained within the space A R s p was in pieces placed irregularly as in a heap of charcoal, and without any connection with the surrounding wood. It was only in that part of the pile corre- sponding to the triangle ABC, and the space included between the line A P and the parallel line E D, by their revolution round R s, that the charcoal remained attached to the wood. The distance be- tween D E and A P was from 10 (4 in.) to 15 (6 in.) centimetres. On each of the pieces of wood included within A p, the passage from perfect charcoal to unchanged wood might be traced, the two being separated by partially carbonized brown wood to the distance of 7 (2| in.) or 8 (3 in.) centimetres. The carbonized part of the wood had undergone very sensible contraction. If carbonization had been allowed to proceed unchecked, the angle H D E would have continued to decrease, until at length the line D E ^would have coincided with H D, and then all the wood would have been converted into charcoal. Hence it is clear that carbonization in piles is propagated from above downwards, and from the centre to the circumference. The air enters at the bottom of the pile, and finds its way to the space within A p, to which combustion is limited ; and the volatile pro- ducts of carbonization escape at the vents A B, round the upper part of the mound. It is in this space that the charcoal last formed remains attached to the wood. But as the volume of charcoal is considerably less than that of the wood from which it is produced, the spaces be- tween the carbonized parts of the pieces of wood must be considerably greater than between those which remain uncarbonized. Moreover, within the line E D the charcoal is detached, broken, and irregularly piled in a heap. Hence, the circulation of air should take place most readily where the least resistance is offered, that is, upwards through the space c E D P, with the upper part of which the vents are in communication. P D Fig. 19. 136 COMPOSITION OF PERMANENT GASES FROM PILES. In the rectangular pile the process of carbonization would appear to take place in much the same manner as in the circular pile, In the preceding diagram, fig. 19, let R H s represent the left half of the ver- tical section of a circular pile (see fig. 2, a), and let R K H' s represent the vertical section of the rectangular pile, fig. 8. Now in the circular pile carbonization commences along the line R s and proceeds outwards and downwards in the direction of the line E D. Suppose the wood within the space R E D s to be already converted into charcoal, and the process of carbonization to be active within the space E A p D, the air by which combustion is sustained circulates upwards in the direction of the arrow. But this is certainly the direction in which the air cir- culates through the rectangular pile R K H' s. In the circular pile the smoke escapes through vents all round on a level with A ; whereas in the rectangular pile it escapes through vents across the top at A'. According to Ebelmen, the heat by which carbonization is effected in piles is produced solely by the combustion of charcoal already formed, and not in any degree by the combustion of the volatile products evolved. He analysed the volatile products which issued from the vents of piles in different stages of combustion, and those which are produced by the carbonization of wood in dose vessels. In the case of the pile, the permanent gases will contain all the nitrogen of the air which has contributed to sustain combustion, to- gether with an amount of carbonic acid corresponding to the oxygen of that air ; that is, admitting that the oxygen is wholly converted into carbonic acid by contact with the ignited charcoal of the pile. Now, if we deduct from these permanent gases all the nitrogen, and an amount of carbonic acid containing oxygen, proportionate to the amount with which nitrogen is associated in atmospheric air, the residual gases will be found to approximate in composition to the per- manent gases produced by the carbonization of wood in close vessels. The following experimental results in proof of this were obtained by Ebelmen. COMPOSITION OF PERMANENT GASES FROM PILES. Carbonic acid Carbonic oxide Hydrogen 1. 2. 3. 4. 5. 6. 7. 8. 9. 25-57 8 68 9-13 56-62 26-68 9-25 10-67 53-40 27-23 7-67 11-64 53-46 25-89 9-33 9-28 55-50 28-34 15-17 8-87 47-62 21 26 5-18 8-84 64-72 23-51 5-00 4-89 66-60 23-28 5-88 13-53 57-31 23-08 6-04 14-11 55-77 Nitrogen 100-00 100-00 100-00 100-00 100-00 100-00 100-00 100-00 100-00 COMPOSITION OF THE GASES AFTER DEDUCTION OF THE NITROGEN AND CORRKSPONDING CARBONIC AdD. Carbonic acid Carbonic oxide Hydrogen 1. 2. 3. 4. 5. 6. 7. 8. 9. 37-5 30-4 32-1 38-8 28-4 32-8 40-3 23-6 36-1 37-8 31-2 31-0 39-7 38-0 22-3 23-5 28-2 48-3 37-8 31-4 30-2 29-8 21 2 49-0 29-6 21-1 49-3 100-00 100 00 100-00 100-00 100-00 100-00 100-00 100-00 100-00 COMPOSITION OF PERMANENT GASES. 137 (1.) Gas from the vent of a pile, at one-third of its height, two days after lighting. The pile contained 60 cubic metres (2118 cub. ft.). The vent had been opened 6 hours before, and gave much thick white smoke. A mercurial thermometer put in to the depth of O m 10 (4 in.), and taken out after 8 minutes, marked 260 C. (2.) Gas 24 hours after No. 1, from a vent in the same pile, opened an hour previously. There was much dense white smoke. (3.) Gas 24 hours after No. 2, from a vent in the same pile, at O m 60 (1 ft. 11^- in.) from the ground, opened an hour before. There was much white smoke. Temperature 230 Q. (4.) Gas from a vent of another pile containing 35 cubic metres (1235J cub. ft.), at O m 60 (1 ft. 11^ in.) from the ground, 4 days after lighting. Much dense white smoke issued from this vent. (5.) Gas from the same pile as No. 4, at O m 30 (llj in.) from the ground, and 18 hours before the termination of the burning. Dense white smoke. (6.) Gas from the vent of another pile, O m 60 (1 ft. 11 J in.) from the ground and 36 hours before the termination of the burning. Smoke bluish, transparent, and not copious. (7.) Gas from the same pile as No. 1, 18 hours afterwards. The vent was O m 30 (llf in.) below those from which gas No. 1 issued, and was made 4 hours pre- viously. White and not very dense smoke. (8.) Gas from the same vent as No. 7, but 5 hours afterwards. Smoke clear and slight. Tem- perature 250 0. (9.) Gas from the same pile as No. 8, 24 hours afterwards; the smoke which issued from the vent was slight, bluish, and almost transparent. The piles employed in these experiments m-ve constructed like that described under No. 4 (p. 115). COMPOSITION OF PERMANENT GASES BY CARBONIZATION IN CLOSE VESSELS. 10. 11. Carbonic acid 44 9 29-2 Carbonic oxide 36 -8 24-9 Hydro H 12 Chrysen C 12 H 4 (?) Pyren 0H U (?) Eupion (?) ACID COMPOUNDS. Acetic acid C 4 H 4 4 Phenylic acid or phenylic I C 12 H 6 2 alcohol I Cresylic acid or cresylic al-| cohol / Phlorylic acid or phlorylic I alcohol I Rosolicacid. 0< 6 HW(?) Brunolic acid (?) BASIC L 2 148 GENERAL CONSIDERATIONS ON PREPARATION OF COKE. Other products of the distillation of coal are similarly decomposed by exposure to a high temperature. In this manner the formation of the deposit in question, or gas-retort carbon, as it is termed, admits of easy explanation. Now, in the preparation of coke, carbon may be deposited upon the coke in precisely the same manner. Let us suppose coal piled to the thickness of two or three feet in a fire-brick chamber, entirely closed, with the exception of a hole in the top to act as a chimney, and a few small openings through which air from without may enter, above the surface of the coal ; and let us further suppose that the whole of the upper part of the coal is burning actively, air to sustain combustion entering through the small openings, and the products of combustion escaping through the hole in the top. The oven above the coal will speedily be heated to redness, and heat will be propagated downwards through the coal, of which every portion will be successively subjected to destructive distillation. The volatile products from below will ascend ; but in traversing the red-hot stratum of coked coal above, they will partially suffer decomposition : the coked coal will be coated with a deposit of lustrous carbon, while the more or less decarbonized residual gases will take fire as they escape from the incandescent mass. The coke which is formed above will continue to be enveloped in an ascending current of volatile pro- ducts, and will be thereby protected from the action of the air, and so from, burning to waste. Occasionally hair-like threads are observed on pieces of coke. They are solid, and under the microscope present somewhat of the appearance of a string of beads which have been soldered together. They consist of carbon, which seems to have been deposited on bubbles, of gas. One bubble being produced, a second is then formed in connection with the first, and so in succession until the completion of a continuous tube, constricted at the junction of every two bubbles. Gas would continue to flow through this tube, depositing carbon in its course on the inner surface, until at length the tube is converted into a nearly solid fibre. Such appears to me to be the mode of formation of this curious hair- like matter, though I am by no means certain of the correctness of this view. AY hen caking coals are thus heated, they first agglutinate into one mass, which, as the process proceeds, becomes fissured from top to bottom so as to form a series of columnar pieces, resembling the columnar structure which is induced in sandstone by the long-con- BASIC COMPOUNDS. I Chinoleine ' Ammonia ........................ H 3 N Aniline ........................... C 12 H?N Leucoline Lepidine C 20 H 9 N Cryptidine Pyridine C 10 H 5 N Picoline C 12 H'N ! Pyrrol C 8 H 5 N (?) Lutidine C 14 H 9 N Collidine C 16 HN Parvoline C 18 H la N Hydrogen H The equivalent numbers of CO, CO 2 , etc., are given as C 2 2 , C 2 O 4 , etc., the numbers now adopted by Dr. Hofmann. COKING IN CIRCULAR PILES. 149 tinned action of a high temperature, or the structure which starch acquires by desiccation. To this state of aggregation the term crys- tallization is frequently, though very erroneously, applied. The quality of coke is much affected by the degree of heat and duration of the coking process. It may be stated as a general rule that the higher the temperature, and the longer the exposure to that temperature, the harder, more dense, and less easily combustible will be the coke. M. de Marsilly tried the effect of coking during 96 and 1 20 hours, but found that no advantage was derived by prolonging the process beyond 48 hours. COKING ix PILES. The terms coke-hearth and coke-fires are employed synonj-mously with piles. The piles are either circular or in the form of heaps having a long narrow rectangular base. The ground on which they rest should be flat, dry, and solid ; and a plentiful, supply of water should be at hand. Circular Piles. They vary considerably in size according to cir- cumstances and the caprice of the coke-burner. The annexed woodcut represents a pile of this description which I saw at the Russell's Hall Furnaces, near Dudley. The diameter at the base was 30 feet. The bed was of earth, not brick. In the centre is a chimney, built of bricks without mortar ; it is shown in elevation at a, and in plan at b. LJ LJ Fig. 20. The heat of the pile is sufficient to vitrify the surface of the bricks, and so firmly to unite them together. Four bricks are first laid as in b ; upon each brick three others are placed, and thus four pillars of four bricks each are formed. Across the ends of these pillars two courses' of bricks are laid, and then the chimney is continued upwards by placing the bricks circularly, so as to leave regular spaces between their ends all round, as shown in the woodcut. The upper part of the chimney is built without spaces. A large flat brick may when necessary be placed on the top Fig. 21. Vertical section through to act as a damper ; or a suitable iron damper may be provided, as shown in the annexed woodcut, fig. 21. It is a cylinder of cast-iron, about 14 in. high and 10m. in internal diameter ; at the bottom is a flange, which rests on the top of tjie brick chimney. 150 COKING IN CIRCULAR PILES. The diameter of the opening at the bottom is 8 in. When necessary, the damper, cr, consisting of a disc of iron, provided with an upright handle, is dropped in. The damper is covered over with sand when it is desired to close the chimney perfectly. The coal employed was the non-caking thick or ten-yard coal. The large coal is stacked round and inclining against the chimney, and then follow the lumps (smaller coal) ; when the stacking of the coal is completed, the whole surface of the pile is covered with a layer of coke-dust from previous burnings, except round the bottom of the pile, e e, to the height of about a foot. The height of the pile from the centre to the highest part near the chimney was five feet. The dark shaded part, c, on the left of the chimney is a view in elevation of that part, the rest of the pile on the right being a vertical section through the centre of the bed, with the exception of the. chimney, which, as has been previously stated, is seen in elevation. In some piles the chimney was larger, and rested on six pillars of bricks instead of four. The diameter at the base of a chimney of this kind was 3 ft. 3 in. outside measure; and circular bricks were employed. In the pile described, about 20 tons of coal (1 ton = 2640 Ibs.) were stacked. Ignition is effected by putting live coals on one side of the chimney near the top. Combustion is thus conducted downwards through every part of the mass. Thick smoke speedily appears, and flame issues from the chimney and various parts of the surface. The progress of the coking is carefully watched ; where the combustion appears too vigorous, the coker damps it by applying coke-dust. After a time beautiful blue flames of carbonic oxide appear here and there over the surface. When coal smoke ceases to escape, and the process of burning is completed, which will occur ijL about 5 or 6 days, wet coke-dust is plastered over every part of the * surface of the pile by means of a spade, and the chimney is perfectly closed. In windy weather much attention is necessary on the part of the coke-burner to prevent as far as practicable the waste which is liable to occur from the increased combustion of the part exposed to the wind. Numerous precautions are necessary to ensure a successful result, but these can only be learnt by experience. Simple as the process may appear, yet the men who conduct it differ much in their degree of skill. On about the tenth day after lighting the coke may be drawn. Before drawing the pile is watered; that is, water is thrown upon it. The yield from such a pile was stated to be 3i barrows to the ton of coal (2640 Ibs.) or 13 cwt., that is 65 per cent, of the coal. This information was not based on exact data, and I should be inclined to consider the amount as too great. A consider- able quantity is burned to waste in this method of coking, as may be inferred from the layer of ashes covering the entire surface of the pile at the conclusion of the process. At the Coalbrook Vale Ironworks, South Wales, I observed coking in circular piles about 18 ft. in diameter at the base and 6 ft. high in the centre. Mr. James, the furnace manager, informed me that the pile was lighted at these works by putting live coals down the chimney, and that the coal round the bottom became first ignited. COKING IN CIRCULAR PILES. 151 From thence the fire creeps towards the outside of the pile round the base, and extends upwards ; and in proportion as it rises, the surface of the pile underneath is damped in the usual way. A column of flame 2 ft. high may continue to issue from the chimney during some time after lighting. The chimney is left open till the fire nearly reaches the top, when it is covered with an iron plate. In Mushet's description of this process, 3 it is stated that brick flues, or channels formed by pieces of coal, are laid to communicate with the lower tier of holes in the chimney, so as to conduct the air through the interior of the mass. In other respects the description is similar to that last given. Mushet, however, observes that as the fire proceeds upwards the ignited surface of the exposed coal (which forms a zone never more than 4 or 5 inches broad) is from time to time covered with coke- dust and a new surface exposed, the dust crumbling above and pro- tecting the ignited surface of the recently formed coke below. In this way the fire in 2 or 3 days reaches the upper surface, when the flame in the chimney becomes less, until it dies away altogether. Mushet further remarks that this process may be modified so as to yield coke of different qualities from the same kind of coal. \Vhen the opening round the bottom of the pile is much diminished the process is re- tarded, and the fire requires a longer time to reach the outer" surface. The yield thus obtained is said to be greater than when the coking- takes place more rapidly ; the coke is darker, less combustible, has a higher specific gravity, and is little changed in form (from the original coal ?). On the other hand, when the pile is left open round the base to the height of a foot, combustion proceeds more rapidly on account of the freer access of air ; a higher temperature is produced ; the coke is honeycombed ; is more grey in colour ; has a lower specific gravity ; is more combustible ; and less in quantity. * . According to Mushet this system of coking under wetted dust was introduced at the Muirkirk and Clyde Iron- Works about the year 1801, and in a few years became general at the Scotch furnaces. In 1805 it was introduced into Derbyshire, and subsequently into Staffordshire, Yorkshire, and Shropshire. He further states that it had been attempted at Merthyr, but without success, as the coal passed " almost unchanged in form into a ponderous coke resembling anthracite ; " and at other works in South Wales in which a caking coal was employed, it was tried and abandoned in consequence of the slow combustion occasioned by the welding of the coal, and the rents and cracks caused upon the surface of the pile by the great enlarge- ment of its volume. 4 Xow it has been shown that in Staffordshire coking in piles was commonly practised long anterior to 1805 ; but Mushet probably means to intimate that a covering of " wetted dust " had not been previously applied. In times anterior to that date the pile seems to have been constructed exactly like a charcoal pile, and to have been covered with straw, leaves, and soil in succession. 5 3 Papers on Iron and Steel, 1640, p. 304. 4 Op. cit. p. 304. 5 Traite" Me"thodique de la Fabrication du Coke, etc. Par M, Pelouze, pere. Paris, 1842, p. 9. 152 COKING IN LARGE OPEN HEOTANGULAR KILNS. Coking in long Piles or Midges. These piles, which are technically termed " pits," may be extended to any length, and may be con- veniently arranged in parallel rows. At the Coalbrook Yale Iron- Works, I measured one, of which the transverse section at the base was 12 ft., and the height in the centre 3 ft. 6 in. A pile of these dimensions will contain from 2 tons 10 cwt. to 3 tons per linear yard. Mr. Levick informs me that they vary from 4 to 5 ft. in height in the centre, and from 9 to 12 ft. in width at the base. There are no chimneys as in the circular/ piles. A layer of small coal, from 16 to 12 in. thick, is left at the bottom ; and upon this the large coal is stacked, inclining towards the middle of the ridge, and in such a manner as to leave air passages all through the inside of the pit ; the outside is covered with a layer of small coal. The pile is lighted at short intervals along the top, and the combustion is conducted down- wards. As the flame ascends up the outsides of the pile, the coker con- tinues to damp them with wet coke-dust until the coal is completely coked throughout; when this occurs the pile is well plastered over with wet coke-dust, and then left to itself. Before the fire is quite extinguished the pile is watered and the coke drawn as required. COKING IN LARGE OPEN RECTANGULAR KILNS. A description of this method, as practised at Gleiwitz, in Upper Silesia, was published in 1851 by Brand, 6 who states that it had been previously in operation in the principality of Schaumburg Lippe, where a pure but very tender and strongly caking coal is raised. According to Brand, who is manager of the iron-works at Gleiwitz, and who writes from personal experience on the subject, the advantages of this method are, that it recmires only a very moderate outlay, and produces a dense coke of exllllent quality. The kiln consists of two parallel walls of brick, a, a, fig. 23, and a flat bed, ft, of bri^s set edgewise, underneath which is a stratum of a glassy blast-furnace slag broken small, so that proper drainage may be secured. Fire-bricks are only used to form the bed and the inner surface of the walls ; the walls are five feet high, eight feet apart in the clear, and from rfto 60 feet long (Prussian measure). In each wall is a series of transverse openings, c, c, &c., at a distance of two feet from each other, and at the same height above the ground, so that the openings in one wall are respectively opposite to those in the other. From each of these openings, c, rises a vertical chimney, d. The charging of the kiln is effected as follows: One of the end openings, e, e, is bricked up, and through the opposite one coal slack is wheeled in, spread over the bottom, watered, and stamped down so as to form a solid stratum nine inches thick, or as high as the lower edges of the openings, c, c, &c. ; indeed, the height may be made two feet with advantage, if the coal be suitable. Pieces of wood, six inches in diameter at one end, and four at the other, and in length equal to the width of the kiln, are then passed through the openings in one wall so that their opposite ends may respectively lie in the Berg. u. Hiittenman. Zeitung, April 2, 1851, V. 10, p. 217. COKING IN LARGE OPEN RECTANGULAR KILNS. 153 corresponding openings in the other wall. Wetted coal slack is spread over the pieces of wood, and stamped carefully down. The kiln is then filled up with slack, which at every six inches of additional height should be watered and stamped down. % Brand well remarks that the mode of filling just described is very hard work when the kiln exceeds 40 feet in length. After the filling is completed, the top of the coal is covered with a layer, two or three inches thick, of coal- dust ; or, failing this, of loam. The end opening through which the kiln has been charged is at last bricked up. The pieces of wood are now carefully drawn out, and thus a series of channels are formed in the coal, upon the maintenance of which the success of the process essentially depends. Should an injury occur to any of the channels at the commencement, it can hardly be repaired afterwards. ' Before lighting the kiln, all the chimneys on one side are stopped by placing a brick on the top of each (see fig. 25, d), those on the opposite side being left open ; while on this side the openings or draught holes are stopped by bricks, fig. 25, c', the holes 09 the opposite side being left open, as in c. The kiln is now lighted by means of sticks of easily inflammable wood introduced into all the openings, c, on the left. A current of air is established through the transverse channels Fig. 22. Side elevation. n a p ^ bj ( ( d H * J2J j n i< 60' id H H H H |H Fig. 23. Plan. Fig. 24. Section on the line A B, fig. 23. Fig. 25. Section of kiln after filling (Mr. Rogers). 154 COKING IN LARGE OPEN RECTANGULAR KILNS. in the coal, in the direction indicated by the arrows. After the lapse of six or eight hours the fire will have reached the opposite ends of these channels, when the chimneys on the left, d, and the draught holes on the right, c'^mnst be opened, and the chimneys on the right, d', and the draught holes on the left, c, must be closed. This, how- ever, should only be -done when the fire has regularly spread through the entire extent of the channels. Special care in this respect at the commencement will prevent further trouble afterwards. According as the weather is stormy or settled, the direction of the currents of air through the kiln may be changed from every two to four hours. Should the coking be found to proceed irregularly, it may be necessary to keep open some of the chimneys on one side longer than others, and, consequently, not at once to change the direction of all the currents. Irregularity in the coking may depend either upon the quality of the coal or negligence in piling it in the kiln ; and in either case the yield may be diminished. In the management of the. process the work of the coke-burner is reduced to keeping open the transverse channels in the coal by raking out any pieces of coal which may fall into them and obstruct the passage of the air, and in preventing them from sintering together. For this purpose he uses a slender iron rod, somewhat bent at one end. The re-opening of a channel which has once become stopped is attended with much difficulty, and is generally impracticable ; and if several neighbouring channels are closed, the process is thereby much impeded. In windy weather the draught of air through the kiln must be carefully regulated by closing, in a greater or less degree, the chimneys. Any cracks which may occur during the process in the covering on the top of the coal must be well stopped, in order to prevent the ascent of currents through them. The proper regulation of the draughts through the kiln has an 'important influence upon the quality as well as the yield of coke. In about eight days the process will be completed, as may be known by the escape of white flame from the chimneys and the hardness which is perceived on plunging an iron rod through the cover on the top. All the openings must now be closed, and in the course of two days afterwards the fire will be gradually extinguished. One of the end walls is taken down and the coke removed. The coke at the height of the channels will be separated into two distinct layers ; that in the upper layer especially is remarkably beautiful (sic), dense, hard, and when carefully withdrawn is frequently in pieces 3 feet long and 1 foot in diameter. The yield per 7'768 cubic feet (English) of coal ranged from 241-25 to 261-87 Ibs. (avoirdupois). The loss in weight is 20 per cent, of the coal, an amount which, according to the quality of the coal, is often much reduced. For the measure of coke above mentioned the workmen received just over Ifd. The coke produced by this method is stated to have yielded most excellent results in the cupola, 1-36 cubic feet (English) being used to melt from 283 to 510 Ibs. (avoirdupois) of pig-iron, according as it is intended to be run into thin vessels (Potterie) or heavy castings. COKING IN -LARGE OPEN RECTANGULAR KILNS. 155 In the vicinity of Sarrebriick experiments have been made with kilns 10 feet high, having in the middle of each side a second row of draught holes ; but the result was unfavourable, and the kilns were abandoned. Several years after the publication of the preceding description, 7 the same process was patented in England. On the 28th January, 1857, my friend Mr. Eogers, of Abercarn, communicated to the Institution of Mechanical Engineers in Birmingham a paper on the manufacture of charcoal and coke, in which he introduced to their notice the process in question. In that paper the following passage occurs : 8 " A short time ago a plan was mentioned to the writer as having been used in Westphalia, by which wood was charred in small kilns; as the form of kiln described was quite new to him, it led him to some reflection as to the principles on which it acted, which were found to be so simple and effective, that he determined to apply them on a large scale for coking coal. The result has been that in the course of a few months the original idea has been so satisfac- torily matured and developed, that instead of coking six tons of coal in an oven costing 80/., 150 tons of coal are now being coked at once in a kiln costing less than the former single oven." The description and drawings of the kiln contained in the paper prove that in every essential point it is identical with that which I have just described. Mr. Rogers is mistaken in supposing that coal had not previously been coked in such kilns. The theory of coking by this method is perfectly intelligible. The coal surrounding the transverse channels is ignited, and through these currents of air are established. Heat is thus developed partly by the combustion of the coal in the vicinity of the channels, and partly by that of the volatile products arising from the destructive distillation of the coal. The coking will, therefore, proceed simulta- neously upwards and downwards. No currents, as has already been stated, can ascend through the coal aoove the channels, if the kiln be properly attended to. The air which sustains combustion can only enter the kiln through the lateral draught holes ; and, obviously, none can descend from above. At the conclusion of the process an accu- mulation of tarry matter always occurs immediately under the coal at the top of the kiln, which would further tend to prevent the descent of air from above as well as the ascent of currents from below ; and it is there that the most solid coke is produced. In South Wales, and I believe also in other districts, kilns of this kind have been erected of not less than 15 feet in width from wall to wall, measured within. The transverse channels may be made by suitably piling lumps of coal. When the coal is of different sizes, it is advantageous, according to Mr. Rogers, to let the size of the pieces diminish towards the top of the mass. In these larger kilns the mass 7 It was also published in England in 1855. Vide Chemical Technology, Bail- liere, London, v. 1, p. 117. 8 Proceedings of the Institute of Me- chanical Engineers, 1857, p. 31. 156 COKING IN LARGE OPEN RECTANGULAR KILNS. becomes well ignited in from twenty-four to thirty-six hours. During the process the workman walks on the top of the coal ; and from time to time at different parts of the surface he inserts an iron bar, which is easily pushed down until it reaches the mass of coke. In this way the height to which the coking process has reached is satisfactorily ascertained : if he finds it to have progressed higher at one part than another, he closes the chimney communicating with that part, and so retards the process there. When the mass has been coked up to the top, which takes place in about seven days, it is quenched with water, and the coke withdrawn in the manner already described. 9 "The new kilns," writes Mr. Rogers, "have proved entirely suc- cessful ; they are already in use at some of the largest iron-works in the kingdom, and are being erected at a number of other works. The great saving in the first cost of oven, economy in working and mainte- nance, increased yield, and improved quality of coke, will probably soon cause this mode of coking to supersede the others now in use. The kilns are most advantageously made about 14 feet in width, 90 feet in length, and 7 feet 6 inches in height : this size of kiln contains about 150 tons of coal." Mr. Eogers asserts that an outlay of only 41. is required to produce one ton of coke per day from the Welsh coals, and that the cost of working does not exceed 6d. per ton. In some places the coal has been actually tipped into the kiln from the colliery waggons, and the coke waggons were afterwards run into the kiln to be loaded direct from the mass of coke produced, thus reducing the labour to a mini- mum. The kilns need only be built of rough rubble work with a plain lining of fire-brick and without any iron work, so that the expense of repairs amounts only to a small sum. This exactly accords with Brand's experience of the German kilns. W hen interrogated at the meeting before which the paper was read as to the yield of coke by this method, Mr. Eogers replied that 18 cwt. per ton of coal (Welsh) had been obtained, an amount nearly equal to that of the carbon which existed in the original coal. It is hardly necessary to observe that this statement must be erroneous. It does not appear that the proportion of water retained by the coke after quenching had been determined ; and if this had been done, the result would probably have been widely different. Mr. Biley informs me that he found as much as 22 per cent, of water in coke prepared at the Dowlais Iron-Works by the method in question. In 1859 I visited several of the large iron-works in South Wales where these kilns had been tried, and I inquired particularly con- cerning the results. Opinions on this subject were certainly not concordant. At the Dowlais Iron- Works they have been erected, and after repeated trials abandoned. The Ebbw Vale Iron Company has also made trial of them, and my friend Mr. Adams, the manager, informs me that they appear to be adapted to one kind of coal, but that for their usual good coal they are Vide paper above referred to, p. 32, from which I extract these details. COKE-OVENS. 157 wasteful and expensive ; much of the large coal which is used to form the transverse channels is burned away, and as he quaintly observed, "you might hunt badgers through the coke." At Pontypool Iron- Works I inspected one of these kilns from, which the coke had been partially drawn, and I remarked that a good deal of the coal opposite the draught holes appeared to have burned away : some of these kilns were much higher than I had seen elsewhere. Experiments have been made at these works with kilns having double rows of draught holes on each side; but I am informed the result was not satis- factory. COKE-OVENS. In its simplest form a coke-oven is a chamber of fire- brick or some other refractory material, having an arched roof in which is a hole and an entrance below. Many years ago Parkes described ovens of this kind, which were in use at the Duke of Norfolk's Colliery near Sheffield. 1 Each oven was a circular building 10 ft. in diameter within, having a floor laid with common bricks set edgewise. The wall of the oven rose perpendicularly 19 inches above the floor, and was covered with a brick arch rising 3 ft. 5 in. in the centre and forming a cone, which measured within was 10 ft. at the base and 2 ft. high at the apex. The opening at the top acted as a chimney during the process of coking, and through it the coal was introduced. The entire height of the oven from the floor to the top of the arch, outside measure, was 5 ft. ; the surrounding wall, 18 in. in thickness, was built with good bricks closely laid, so that no air might get in through any part of the work. The floor was raised 3 ft. above the ground for the convenience of placing a low carriage under the doorway, to receive the coke as it was raked from the oven. The oven was enclosed up to the top in a square, formed by four vertical walls, 20 in. in thickness, and built of rough unhewn stone. The four corners between the circular oven and the surrounding walls were filled with soil or rubbish, which was well rammed to give greater firmness to the work, and the more effectually to exclude the air. Parkes has given the following excellent description of the mode of conducting the process, which cannot well be improved, and which is applicable to coking in ovens as practised at the present time. When once the ovens are heated, the work goes on night and day without in- terruption, and without any further expense of fuel. Small refuse coal is thrown in through the hole in the top, sufficient to fill the oven up to the springing of the arch ; it is then levelled with an iron rake, and the doorway built up with loose bricks. The heat which the oven acquired in the former operation is always sufficient of itself to light up the new charge, the combustion being sustained by the entrance of atmospheric air through the joints of the loose bricks in the doorway. In two or three hours it is necessary to check the influx of atmospheric air by plastering up the doorway with a mixture of wet soil and sand, except the top row of bricks, which is left unplastered all night. Next 1 Chemical Catechism, 12th eel., 1826, p. 453. 158 COKE-OVENS. morning, after the charge has been in the oven twenty -four hours, this row is also completely closed ; but the chimney remains open until the flame is gone, which is generally quite off in twelve hours 'more; a few loose stones are then laid on the top of the chimney, and closely covered with a thick bed of sand or earth. All connexion with the atmosphere is now cut off, and in this state the whole remains for twelve hours more to complete the operation. The doorway is then opened, and the coke is raked out into wheelbarrows, or low wag- gons, to be carted away. About two tons of coal are put in for each charge. The coke thus produced is ponderous, extremely hard, of a light grey colour and shining metallic lustre. It is used in those manufactures that require an intense and long-continued heat. The coal is thus subjected to destructive distillation ; the volatile products, as they escape from the surface of the mass, meet with atmospheric air, take fire, and burn, with the evolution of much heat. Combustion is thus sustained along and above the surface of the coal, and the process of coking is gradually propagated from above down- wards. It is evident that the coal underlying the incandescent stratum must be exactly in the condition of coal undergoing distillation in a close vessel ; for air is supposed only to enter the oven above the level of the coal. A current, therefore, of inflammable gases and vapours will continue to ascend until the lowermost stratum of coal is converted into coke. Hence it might be inferred that the process of coking in ovens would be effected, at least to a considerable extent, by the heat developed by the combustion of the volatile products evolved. Ovens have been made circular, more or less oval, or rectangular ; and in dimensions they have varied considerably. Contrivances have from time to time been adopted with the following objects : First, to prevent as far as practicable the escape of heat from the oven ; secondly, to effect the admission of air so as most completely to burn the volatile matters evolved from the coal ; thirdly, to economize the waste heat so as to cause the process of coking to proceed from below upwards as well as from above downwards ; and fourthly, to facilitate the removal or drawing of the coke from the oven, with the view, not merely of diminishing labour, but also of reducing as much as possible the amount of heat which, in a greater or less degree, must necessarily be lost during this part of the operation. The first object has been accomplished by making the walls thick, and covering the roof with sand, or some other bad conductor of heat ; by building a second arch at some distance over that which forms the roof of the oven proper, and causing the products of combustion to pass between the two arches in their way to the stack ; and by constructing in one rectangular pile of building two parallel rows of ovens back to back. The second object has been effected by allowing the air to enter the oven in several places, and accordingly passages have been formed in the brickwork by which air may pass through the sides or back, above the surface of the coal ; the supply of air through the front being either wholly or partially stopped. The third object has been achieved COX'S COKE-OVEN. 159 by causing the products of the combustion within the oven to circulate through flues immediately under the bottom, and sometimes also round or down the sides. The fourth object has been attained by building the oven rectangular below the arched roof, and a little wider in front than at the back ; then on the floor at the back is placed an instrument called a drag, which consists of a strong piece of flat iron, having attached to it at right angles a rod of iron sufficiently long to protrude beyond the front of the oven. This drag is left in the oven during the process of coking, after the comple- tion of which the whole mass of coke may be drawn out at once by means of a windlass in front, with which the protruding end of the drag has been connected. In some ovens a small gutter is made from front to back in the floor of the oven ; and only the transverse piece of the drag is left in the oven. The gutter is covered over with little pieces of bar-iron preparatory to charging it with coal, in order that it may not become obstructed. Just before the coke is drawn, a long rod of iron, called a needle, is pushed along this gutter, and by a simple contrivance catches hold firmly of the centre of the transverse piece, after which the operation may be proceeded with in the manner described. By this arrangement the destruction of iron which occurs when the entire drag remains in the oven is considerably diminished. As soon as the coke is drawn, it is extinguished with water, of which a copious supply should be always at hand. The mouth of the oven may be stopped either by building it up with bricks, or by a door con- sisting of a framework of cast-iron, in which fire-bricks are set. The door may be made in two parts, and hinged like an ordinary folding- door, or it may consist only of one piece, and may be raised or lowered by means of a chain passing over a pulley above, and having a coun- terpoise weight at the end. The charging is generally effected through a hole in the roof, to which the coal is conveyed on a railway along the top of a pile of ovens. When the coke is not drawn out in mass, it must be removed in pieces, which will obviously occasion a greater expen- diture of time and labour, and the oven will in consequence lose much heat, especially if water be injected into it to extinguish the coke. Instead of drawing out the coke at the front, it may be pushed out from the back by a suitable apparatus, capable of being moved along a railway from one oven to the other in succession. 2 This method of drawing is employed at Cyfartha and at Beaufort. I have selected for special description the following series of ovens, as sufficient to illustrate the essential modifications which have been proposed and adopted. Cox's Cohi-oven. This oven was patented in 1840. 3 I am indebted to the Ebbw Vale Iron Company for the drawings from which the an- nexed engravings were made. The oven consists essentially of a nearly-rectangular chamber of fire-brick, arched over from side to side, 2 Engravings of such apparatus are given I bach. Ann. d. Mines, 5. s. 15, p. 489. 1859. by Dieudonne in his Memoire sur la Fa- 3 Vide published Specifications, No. brication du Coke h Forbach et Hirsch- 8709, A.D. 1840. 160 COX'S COKE-OVEN. and open in front. The bottom is flat, and inclines slightly down- wards and forwards : see fig. 27 t. The width between the side walls increases gradually, but only in a small degree from back to front, as shown by the dotted lines, fig. 28. At a distance above the arch forming the roof is a second arch, fig. 26 ft and fig. 27 e. The side walls and the front wall above the arched mouth are carried up to a considerable Fig. 26. a Front elevation. Section on the line G H, fig. 28. y Section of the stack on the line i J, fig. 28, above the top of the lower arch. Section on the line A. B, fig. 28. e Section on the line c D E F, fig. 28. x* COX'S COKE-OVEN./"' UNIVERSITY 1 J! Fl in, showing section of the stack on the line K L, fig. 27 height, and the space enclosed by these walls and the stack at the back is filled with sand, so that a considerable amount of matter is thus accumulated, which will have the effect of retaining much of the heat generated during the process of coking. In the front and on the right of each oven is a flue, a, fig. 27 and fig. 28, which passes upwards, back- wards, and downwards in succession, and then along the upper and back part of the oven (see the dotted lines in fig. 28), with the interior of which it communicates by three openings, 666, fig. 26, /3. Between the front wall of the oven and the lower arch a space is left, as shown by the arrow in this part, fig. 27, e. The space between the two arches communicates at the back with the stack. From the engravings it will be observed that the ovens are built back to back in double rows. There is but one stack to every two ovens ; at a certain distance from the top it is divided by a partition wall (see fig. 26, y), two flues being thus formed, one for each oven. Each of these flues is provided with an opening, and also with a sliding damper for regu- lating the draught : see fig. 27, L. The mouth of the oven is closed by a door, consisting of a frame of cast-iron filled with fire-bricks. The door may be conveniently raised or lowered by means of a chain passing over a pulley, and having a counterpoise weight at- tached to it. When lowered, it is securely fixed against the mouth by an iron bar placed transversely, as shown in fig. 26, y. All these details will be sufficiently evident on an inspection of the engravings. The charge is introduced through the mouth, and, as usual, piled of the uniform depth throughout of 3 ft. 6 in. The door is then closed and luted. The air which supplies combustion enters the flue, a, and passes into the oven at the back, through the three openings, 666. The pro- ducts of combustion rise through the space in front of the lower arch, and then pass backwards into the stack through the space between the two arches, as may be clearly seen from the direction of the arrows in fig. 27, . These ovens are necessarily expensive in construction ; but I am informed on good authority that they have been found in practice to be sufficiently advantageous to justify the outlay. The admission 1G2 JONES'S COKE-OVEN. of air can be surely regulated, and the amount of waste heat retained by means of the arrangement of the two arches with the superin- cumbent mass of matter effectually contributes to the coking of the next charge of coal. Under the impression that more sulphur is re- moved, the coke is quenched with water before drawing, the effect of which is to injure the bottom and increase the expense of repairs. Mr. Parry, of the Ebbw Vale Iron-Works, informs me that he has recently (April, 18G1) effected an important improvement in these ovens which is especially adapted to the coking of small coal. In the wall at the back are four vertical flues, in connexion with four flues immediately under the bottom, in the direction of its length. The air which sustains combustion is admitted, as in Jones's oven (fig; 29), only at the front into the space between the two arches (fig. 27, e), and not below the lower arch into the chamber containing the coal. The gases evolved are thus burned on the outside of this chamber, and the gaseous products of combustion pass backwards and then downwards through two of the flues at the back, and from thence forwards through two of the flues under the bottom, returning by the other two flues under the bottom, and ascending through the other two flues at the back into the stack. No flame appears at the top of the stack, which is the case in the old ovens ; so that all the heat developed by the combustion of the gaseous products is applied advantageously in the coking process. By this arrangement 50 per cent, more small coal can be coked at a time and with much greater economy than in the old ovens, as less coal is burned to waste than in those ovens. A stratum of small coal 4 feet thick may bo uniformly coked to the bottom in the new ovens, whereas in the old ovens, in a stratum 3 feet thick, there was always a layer at the bottom, not less than 6 inches thick, of soft coke from imperfect carboni- zation. Jones's Coke-oven. This oven was patented by the late Mr. Edward Jones, Manager of the Russell's Hall Furnaces, near Dudley, belonging to Mr. S. H. Blackwell. The annexed en- gravings were made from drawings prepared under ihe direction of my friend Mr. George Shaw, of Birmingham. The oven is built wholly of b brick. The bottom is rectan- gular and flat, and inclines forwards and downwards from a to b, fig. 30. The sides and back are vertical ; and above is an arch enclosing the top ; at the front, b t is an arched opening, or mouth, which III o Fig. 29. Front elevation. JONES'S COKE-OVEN. 163 may be closed by movable brickwork. In the centre of the arch forming the top of the oven is a circular opening, s ; and at the back, r, is a narrow opening extending across : both these openings are closed with movable covers, around the edges of which sand is piled. In the back wall Fig. 30. Longitudinal vertical section. there are two arched openings, c, d, fig. 32, from which flues pass down- wards and forwards under the bottom, as shown by the arrows in fig. 33 ; and then, as shown by the arrows, e, f, in the same figure, to the back, where they communi- cate with the flues, , ?>, &c., extending across the spaces, i, i, &c., and placed quincuncially in the partition walls and the sur- rounding vertical walls : there are 60 of these bricks in each compartment. At the top of each compartment is an opening, o, fig. 36, formed by o o the walls on the short sides rising vertically, and the walls on the long side approximating upwards in a series of steps ; and at the bottom is an opening, p, provided with a cast- iron door. The partition walls of the compartments rest upon frames of cast-iron, fig. 36, M, O m 03 (about an inch) in thickness, which are supported in the direction of their length by brick arches O m 24 (about 9J inches) wide. These arches are so arranged Annales cles Mines, 5. s. T. 13, p. 417. 168 COKE-OVEN OF THE BROTHERS APPOLT. that an open space is left under each compartment. Instead of arches cast-iron bearers may be employed. The bottom of the free spaces is covered with fire-brick to the height of O m 27 (about 10J inches) above the frame-work of cast-iron. Under the compartments and in the direction of their short axes two parallel passages arched at the ends extend through the building. The outer walls of the oven are vertical up to the frame-work of cast-iron, from which they incline inwards to the top. The cast-iron doors which close the compartments at the bottom are O m 02 (about f inch) thick. Each door is strength- ened by three transverse bars of wrought-iron, by means of which it is firmly fixed to an iron rod supported at each end by a staple let into the under side of the cast-iron frame. A hinge is thus formed. The end of the suspending rod which is directly towards the long side of the oven is prolonged, and the projecting part is squared so that it may be turned by a key like that of a piano. In the centre and passing through the middle transverse bar of the door is a pivot, and on this turns a flat bar, of which the ends may slide into grooves in pieces projecting from the under and short sides of the cast-iron frame : the door may thus be securely fastened. It may be shut by a key, which passes through an iron pipe built in the brick-work and Fig. 36. Front elevation. Vertical section on the line G H, fig. 37. COKE-OVEN OF THE BROTHERS APPOLT. 169 firmly fastened to the iron frame, that it may not turn round with the key. In a space included between the vertical heights of O m 42 (about 16 inches) and O m 57 (2 feet) from the bottom, the partition walls of each chamber are traversed by two rows of small horizontal openings O m 14 (oj inches) long and O m 02 (about f inch) high, fig. 36, /: there are nine such openings on each long side and three on each small side. At the upper part there are three similar openings on each long side only, /'. Through these openings the volatile products evolved during the coking of the coal pass into the surrounding open surfaces, in which they are burned by atmospheric air admitted through holes in the long sides of the oven. It is asserted that the heat thus deve- loped more than suffices to coke the whole coal from which these volatile products have been derived. Experience has proved the utility of the small openings,/', at the upper part or somewhat lower, as at/" ; for in operating upon certain coking coals, if the whole of the tar evolved is obliged to traverse the lower part of the cake of coke, too much carbon may be deposited, and the descent obstructed. The height of all the small openings,/,/', is purposely restricted to O m 02 (about J inch) in order that the fine coal may only produce a small talus ; and in the event of their becoming choked with little stoppers of coke, the latter may be withdrawn of themselves in following the mass of coke as it shrinks. In the long Fig. 37. Horizontal section on the line K F, figs. 36,33. 170 COKE-OVEN OF THE BROTHERS APPOLT. Fig. 38. Longitudinal section on the line A B, fig. 37. side walls of the oven are the flues, g, g r , which receive the products of combustion from the spaces surrounding the chambers and convey them to the stacks. There are twelve in all three below and three above in each of these walls. Those below are square in section, O m 25 (about 7 J inches) on the side ; at first they pass horizontally to the middle of the outer brick- work, then ascend vertically and open into a horizontal flue, h. Those above are O m 20 (about 7f- inches) long by O m 17 (about 6J inches) broad; they also pass horizontally into the outer mass of brick-work, ascend vertically, and open into a second horizontal flue, A', parallel to the first. All these vertical flues have dampers of fire-brick at the top, so that the draught may be regulated at will. The four horizontal flues, of which there are two on each side, #, A, have all the same height of from O m 54 (about 1 foot 9 inches) to O ra 67 (about 2 feet 2 inches). Into the two outer flues of which the width is from O m 25 (nearly 10 inches) to O m 29 (nearly 11^- inches) open the six vertical flues from below ; and into the two inner flues of which the width is O m 17 (about 6f inches) open the six flues leading from the compartments at the top. The two horizontal flues on each side are separated by a wall of the width of a single brick, and open into a stack of which the internal sectional area is a square, 48 (about 1 foot 7 inches) on the side, and the height COKE-OVEN OF THE BROTHERS APFOLT. 171 Fig. 39. Horizontal section on the line C D, tig. 36. above the top of the oven 5 m OO (about 16 feet 6 inches). There are two stacks, one on each side, fig. 36 ; in the section the position of one of the stacks is indicated by dotted lines. The lower part of the stacks within, to the height of l m 00 (about 39 inches), is divided by a partition of single brick into two parts, corresponding to the hori- zontal flues which open into them. At j, j\ are openings with iron frames, by means of which the horizontal flues may be cleaned out. It is hardly necessary to remark that every part of the oven exposed to a great heat should be made of fire-brick. The free spaces surrounding the compartments are closed at the top with two courses of fire-brick, upon which is ordinary brick-work of sufficient thickness to prevent the loss of too much heat. A few hollows may be left in the- brick- work to allow for the effect of expan- sion by heat, as shown at .9, fig. 36. The platform on the top of the oven is slightly inclined towards the two long sides, and is covered with cast- iron plates. The bottom of the free spaces may be cleaned out from the outside of the oven through openings d. There are small open- 172 COKE-OVEN OF THE BROTHERS APPOLT. ings provided with sliding registers through the brick-work, a, which answer the double purpose of supplying air to effect combustion, and of enabling an. inspection to be made of the interior of the oven. There are also other small openings, n, for the admission of air, through the arches upon which the long sides of the compartments are built. The brick-work is held firmly together by suitable iron tie-rods, as shown in fig. 36. The oven is charged at the top, and the coke is withdrawn at the bottom and removed in iron waggons. In order to prevent the coke from falling with too much force into the waggon, inclined and projecting plates of cast-iron are fixed in the walls underneath each compartment, fig. 36, A. The mass of coke lodges on these plates, from which it may be detached in pieces and carried away. Mode of conducting the process. The oven is supposed to be new and ready for lighting. At the bottom of each compartment a tempo- rary grate of iron bars is adjusted; and the sides, to the height of O m 3 (about 1 foot) above the grate, are lined with fire-bricks placed slanting, in order to prevent the adhesion of the clinker produced in heating the furnace in the first instance to the walls of the compart- ment : a moderate fire is made, and kept up by throwing in coal at the top, which is allowed to remain open until the walls of the com- partments have become red-hot. When this occurs, it is generally kept closed ; so that the flame from the temporary grate is compelled to escape through the small openings leading into the surrounding free spaces, and all the interior of the oven is thereby heated. By only partially opening the registers of the air-flues a portion of the products of combustion will escape through the outer walls of the oven, and promote their desiccation. After eight or ten days' firing, gradually increased, the oven will be found to have attained the temperature of from 1200 to 1400 C., which is necessary for the commencement of the charging. In order always to insure an equable degree of heat through the oven, and to simplify the management of the latter by the registers and air-flues, it is expedient to charge the two series of compartments alternately. The temporary grate and brick-lining at the bottom is removed from the compartment which it is proposed to charge. The door is closed and securely fixed in the manner pre- viously described, and then covered with a layer of coke-dust about O m 30 (about 1 foot) thick : this is done to protect the door from heat, to close effectually the bottom of the compartment, and to prevent loss of heat. .The charge of coal is now introduced, and a cover placed over the top luted with coke-dust or clay. The gases which are immediately evolved when the coal comes in contact with the red-hot sides of the compartment pass into the surrounding free spaces, where they are burnt, and so sustain the heat of the oven. An hour afterwards the second compartment is charged in like manner, and so in succession until all have been charged. As the amount of gas produced increases din-ing the day with the number of charges, it is necessary to open the registers, and all that is required to be done during the night is gradually to shut them again COKE-OVEN OF THE BROTHERS APPOLT. 173 in proportion as the evolution of gas decreases. Carbon'zation being completed at the end of twenty-four hours, on the following day the coke is drawn from the first compartment at the same time as the charging took place on the previous day. Immediately afterwards the compartment is charged again. The process is thus continued without interruption, the Jcoke being drawn from each compartment twenty-four hours after it has received its charge of coal. No inconvenience arises from the use of washed coal which still retains moisture. By suitably decreasing the admis- sion of air and the exit of gases from the oven, the charging may be omitted on particular days ; and yet the heat will continue sufficiently high to enable the charging to be effected on the following day. Principles on which the oven is constructed. During the process of coking each compartment is in reality a closed vessel, with the excep- tion of the apertures through which the volatile products escape into the surrounding space ; and in so far it resembles a gas-retort. No air from without can reach the interior of a compartment, even though cracks may be produced in its walls. In this respect it differs essen- tially from ordinary coke-ovens, into which air is allowed to enter above the coal undergoing carbonization. If from neglect of the workmen or other circumstances too much air passes into these ovens, a considerable amount of coke may be needlessly consumed ; whereas in the Appolt oven this evil cannot occur. By the subdivision of the oven into a series of compartments, each of which is surrounded by burning gas, a very great extent of heating surface is obtained, which in the oven described is nearly 190 square metres for a charge of 17,000 kilogrammes of coal, a surface two or three times greater in proportion than that of the most improved kinds of other ovens. As the coke is divided into masses of but little width, it can be readily penetrated by heat, and, therefore, subjected to rapid carbonization. The combustion of the gases in these ovens is stated to be more perfect and active than in ordinary ovens, because air enters through numerous openings in the outer walls of the oven, and the mixture of air and gases freely circulates through a large extent of space. This result is further promoted by the exit of the jets of gas through numerous small openings, and, in consequence, its more rapid and complete admixture with air. The partial exit of gas from the lower part of each compartment obviously tends to produce an equable dif- fusion of temperature through every part of the oven. ^The changes of temperature which occur in other ovens, from the charging of the coal to the drawing of the coke, are avoided ; for, as the charging of the compartments takes place at successive intervals in a well- arranged order, the heat of the oven is maintained at nearly the same degree during the whole course of the operation. The heating surface of this oven, compared with its external surface of brick-work, is greater than in other ovens, and, consequently, much less heat is lost by cooling from without. The vertical position of the compartments is said to be important, as presenting the following advantages : it is possible only by this means to secure the advantageous relation between the heating and cooling surfaces, so that a large quantity of 174 COKE-OVEN OF THE BROTHERS APPOLT. coal may be coked in a proportionately limited space in respect of its relative power of production it occupies much less space than other ovens as there is no arch exposed to the action of heat, the oven is more solid, and the coke in dropping down exerts no injurious amount of wear of the sides as the charging and drawing may be veiy quickly effected, the walls of the compartments are less liable to be cooled during these operations the pressure of the column of coal produces a coke of much greater density than that obtained in other ovens. Actual results obtained in coking by this method. At the date of the publication of Messrs. Appolt's description, June, 1858, the oven at Marquise, having worked regularly and without the least interruption since the time it was lighted on the 1st of September, 1857, yielded the following results. Each compartment contained from 1350 to 1400 kilogrammes, that is, somewhat less than 1 \ tons of coal (1 ton = 20 cwt. of 112 Ibs.). The coking was completely effected in the course of twenty-four hours. The workmen suffered not the slightest inconve- nience in the operations of charging and drawing, which took place in the day- time. I should have expected otherwise. The service of four men was required. Belgian caking coal gave from 80 to 82 per cent, of coke, and English caking coal from 72 to 73 per cent. This yield is stated to be from 10 to 12 per cent, greater than that of ordinary ovens. Mixtures of non-caking and caking coals in different proportions also gave good results. The first experimental oven was erected at St. Avoid, department Moselle, France, by which the correctness of the principles of the method was established. Another larger experimental oven was subsequently built in the centre of Sarrebriick coal-field in Prussia ; and although it did not possess several of the contrivances which have since been adopted, yet the result was satisfactory. An oven had been previously erected at Bive- de-Gier in 1856, and continued at work regularly during several months, by which the maximum yield and solidity of construction were established ; but it was afterwards discontinued, as the labour was found too expensive, from there being only six compartments. The coke produced at Marquise was used in the iron-smelting furnaces of that establishment, and was admitted to be of very good quality hard, dense, close-grained and possessing all the characters of good coke for metallurgical purposes. It was demonstrated that the quan- tity of gas evolved during coking was far more than sufficient to effect carbonization and to maintain the heat of the oven at the proper degree, so that the excess of gas might be advantageously applied to raise steam, &c. Under ordinary circumstances the cost of erection of an oven like that of Marquise may be estimated at from 14,000 to 15,000 francs, that is, from 560/. to 600/. Remarks. This oven differs much in construction from all other coke-ovens, and appears completely to fulfil the conditions of a close vessel or retort. Now it has been previously stated that the non- caking thick coal of South Staffordshire will cake and produce a solid coherent coke, provided it be rapidly exposed to a high temperature in a perfectly close vessel. My friend Mr. Samuel Blackwell, of Dudley, some years ago informed me of this fact, about which, I COMPOSITION OF WASTE GASES OF COKE-OVENS. 175 confess, I was at first somewhat sceptical. However, I visited Mr. Blackwell's works, and saw the experiment made to my entire satisfaction. A Hessian crucible five inches deep was nearly filled with powdered thick coal slack, which was pressed well down and then plastered over with wet clay. In this state the crucible was put into Jones's coke-oven, described p. 162, through one of the holes at the back, and there exposed to a bright red heat during twenty minutes or half an hour, when it was withdrawn and allowed to cool. A per- fectly solid coke was produced. The exclusion of the .air by clay is essential ; and it was this which suggested the experiment of covering the surface of the coal in the oven, as recorded at p. 166. To the coal-masters of South Staffordshire an economical solution of the problem of coking the thick-coal slack would be of immense value. A prodigious amount of the fine slack has been and still continues to be left in the pits, because it cannot be raised with profit. I have no doubt that should any person be so fortunate as to succeed in convert- ing this at present worthless material into good coke at a moderate cost, he would realize a large fortune ; and he would, moreover, have the satisfaction of prolonging the industrial life of South Staffordshire, which has begun to suffer from the exhaustion of its fuel. In 1854 a patent was taken out by Mr. Dawson for converting coal-dust and coke into solid blocks of fuel. 5 The dust, moistened with water, is pressed into a cast-iron box having a tightly-fitting cover, and the whole is exposed to a temperature ranging from 300 to 700 Fah. An experi- mental oven was erected near Whitechapel, which I visited. At my request, Mr. Dawson operated upon some of the thick-coal slack from West Bromwich, with which I supplied him; but the result was unsuccessful. Whether the Appolt oven will satisfactorily realize the conditions essential to the coking of this slack, experiment alone can determine. It is a costly structure, it must be admitted : nevertheless, according to the statement of the inventors, the cost is less in propor- tion to yield than in other coke-ovens. Inventors are apt to be sanguine, arid often in good faith ascribe merits to their inventions which experience proves either to have had no existence, or at least to have been much exaggerated. We must, therefore, be cautious in admitting without further confirmation all the advantages enumerated by the Messrs. Appolt in favour of their oven. Composition of ihz waste Gases of Coke-ovens. We are indebted to Ebel- men for analyses of the gases of coke ovens at Seraing in Belgium. 6 The floor of the oven is a rectangle terminated at each end by a trape- zium. 7 The roof is cylindrical above the rectangle, and conical above the trapeziums. There are three chimneys in a line : one in the centre of the cylindrical part of the roof, and one on each side at the junction of the cylindrical with the conical part of the roof. The area of the central chimney is double that of each of the others. The three chimneys are never in use at the same time : the two lateral chimneys 1 Specification, A.D. 1854. No. 3. r> Ilecucil (lew Tntv. Srient. 2, j>. 11 '2. Vid. Traitc Theor. et Prat, tie la Fabrication de la Fonte. Par B. Valerius, 1851, p. 255. 7 This oven is described in detail at p. 1H2. 176 COMPOSITION OF WASTE GASES OF COKE-OVENS. are closed when the central one is open, and conversely. The central chimney conducts the gas'es which escape from the oven under the boiler of a steam-engine. Caking coal was coked in these ovens, which yielded 80 per cent, of coke, consisting of 78 parts of carbon and 2 of ashes, and 20 per cent, of volatile matters. Ebelmen gives no elementary analysis of this coal, but inferred its composition from that of a coal at Eochebelle, near Alais, which yielded very nearly the same per centage of coke, namely, 78, and of which the following analysis was made by Eegnault : Carbon 89'27 Hydrogen 4-85 Oxygen and nitrogen 4-47 Ashes... 1-41 100-00 The charge for each oven is 3 cubic metres ( = 2750 kil., or 2 tons 14 cwt. 16 Ibs.) of small coal, which is spread as evenly as possible over the floor, forming a stratum about O tn 33 (= 12 '99 inches) in thickness. All the chimneys are open at a time for the comfort of the workmen. When the charging is over, the lateral chimneys remain open while the central one is closed, and continue so for two or three hours ; the doorways are closed, but not luted, and carboni- zation commences. It may be divided into three stages, as follows : in the first stage, which lasts about three-quarters of an hour, there is only disengagement of water ; in the second stage, which lasts about an hour and a half, the gases take fire and partially burn with a very smoky red flame, the chimneys remain wide open, and the doorways are closed, but not luted ; in the third stage, the gases burn well with a smokeless white flame. The coal appears incandescent to the depth of 8 or 10 centimetres (3 to 4 in.) from the surface ; the doors are then luted, and only a small fissure is made in the clay luting at the upper part. The lateral chimneys may now be closed, and the central one entirely opened. AY hen the flame begins to decrease, the fissure in the clay luting is gradually contracted and at last completely stopped, and when flame ceases the chimney is closed. The period of coking, inclusive of charging and drawing, lasts from 22 to 24 hours. The average yield of coke is 160 per cent, in volume, and 67 per cent, in weight. Ebelmen analysed the gases collected from these ovens at three different stages of the process, and obtained the following volu- metrical results : 2. Mean. Carbonic acid 10-13 Carbonic oxide 4'17 Marsh gas (CH 2 ) 1-44 Hydrogen 6-28 Nitrogen 77'98 9-60 13-06 10-93 3-91 2-19 3-42 1-66 0-40 1-17 . 3-67 1-10 3-68 81-16 83-25 80-80 Katio between the volume } of oxygen combined with f , ,- carbon and 100 vols. of j nitrogen J 100-00 100-00 100-00 100-00 14-2 , 17-0 . 15-6 COMPOSITION OF WASTE GASES OF COKE-OVENS. 177 1. Gas collected two hours after the kindling of a charge from one of the lateral chimneys of an oven ; smoke black and dense ; reddish flame appearing at intervals. 2. Gas collected 7-J hours after charging ; bright, but still somewhat "reddish, flame, without smoke. 3. Gas collected after 14 hours' coking ; flame clear, but of little volume ; carbonization being apparently near completion. The relation in weight between the elements contained in the gas of mean composition is as follows : 8 I In carbonic acid ................... 1-408) In carbonic oxide .................. 0-443 2-004 In marsh gas ........................ 0-153) Nitrogen ................................................................. 24- 353 30-835 Oxygen which has disappeared, as deduced from thei ratio existing between nitrogen and oxygen in atmos-> 2 '925 pheric air .............................................................. I 24*353 of nitrogen are associated with 7 -273 of oxygen in atmospheric acid : hence, 7'273 - 4 '348 = 2 '925. The per centage of coke obtained in these ovens was 67, so that 33 per cent, of the coal had been removed either by volatilization or combustion. Now, supposing that the 67 per cent, of coke consisted only of 65 parts of carbon and 2 of ashes, it follows that the 33 per cent, of matter lost by coking contained the following elements : Carbon ....................................... 23-68 Hydrogen .................................... 4*85 Oxygen and nitrogen ..................... 4-47 33-00 The relation in weight between the carbon and hydrogen is 1 : 0'205 ; but the relation in weight between the carbon and hydrogen deduced from the mean composition of the gases above stated is 1 : 0-065. Hence, Ebelmen draws the conclusion that more than two-thirds of the hydrogen contained in the coal are burned during the process of coking. He remarks, however, that in this computation no account is taken of the amount of tar and other condensable matters evolved during the process ; but he considers that, in consequence of the very high temperature of the oven during nearly the whole period of coking, the proportion of condensable products is of little account, and that it is only at the beginning of the operation that they are disen- 8 Data from which these calculations have been made at O m 760 (= 30 in.) and 15 5 C. (60 F.) : Grains. 100 cubic inches of carbonic acid weigh 47-26 ,, carbonic oxide ,, 30-21 ,, oxygen , , 34-29 ,, hydrogen ,, 2-14 ,, marsh gas ,, 17' 41 nitrogen ,, 30'14 X 178 ECONOMIC APPLICATION OF THE gaged in appreciable quantity. It must be borne in mind that the data from which Ebelmen draws his conclusion involve an assumption as to the composition of the coal of which no analysis was made. The volume of oxygen existing in combination in the gas of mean composition is to that of the nitrogen as 15-63 : 100 ; whereas the relation in volume between the oxygen and nitrogen which entered the oven in the state of atmospheric air is as 26-26 : 100. The difference, 10-63, represents the volume of oxygen which has served to burn the hydrogen. Hence two-fifths of the oxygen of the air introduced into the coke-oven have been converted into water. In this calculation Ebelmen remarks that the small amount of oxygen contained in the coal has been neglected, but that the correction required to be made in consequence will in no way affect the conclusions enunciated. The quantity of atmospheric air which the process of coking will require may also be deduced from the composition of the gases. The relation in weight between the nitrogen and carbon in the gas of mean composition is 12-2 : 1. Atmospheric air contains 77 per cent, of nitrogen by weight ; consequently, for 1 part by weight of carbon in i he gases, 15*8 of air will enter the oven. Now, it has previously been stated that the quantity of carbon carried off in the gases is 23'68 per cent, of the weight of the coal ; the weight of air, therefore, intro- duced during the process of coking is to that of the coal as 3'74 : 1. Hence, in coking 2 tons 14 cwt. 16 Ibs. ( = 2750 kil.) of coal, not less than 10 tons 2 cwt. 55 Ibs. of air will be required, or in volume 296188 cubic feet (100 cubic inches of air weighing 31-0117 grains). Sup- posing the duration of the coking process to be 24 hours, 12341 cubic feet of air (1 Ib. of air = 13-06 cubic feet) will enter the oven per hour, very nearly 205-7 cubic feet per minute, and 3*43 cubic feet per second. Ebelmen estimated that this amounted to about two-thirds of the air blown into a charcoal iron-smelting furnace yielding 2 tons of pig-iron in the 24 hours. Economic application of the waste Gases of Coke-ovens. From the data which Ebelmen obtained in his investigation concerning the composi- tion of the gases of the coke-ovens at Seraing may be approximately calculated the amount of heat produced during the process of coking, as well as the amount which may be further developed by the com- plete combustion of the carbonic oxide, hydrogen, and marsh gas existing in those gases. Ebelmen made such a calculation, and came to the conclusion that two-thirds of the heat capable of being evolved by the complete oxidation of the volatile products were rendered sensible in the ovens, and only one-third remained to be generated by the subsequent oxidation of the combustible constituents of the gases which escaped from the ovens. Suppose that the coal employed yielded 67 per cent, of coke, and 33 per cent, of volatile products, consisting of 4-85 of hydrogen, 23'68 of carbon, 2-97 of oxygen, and 1-50 of nitrogen ; and grant that only the amount of hydrogen in excess of what is required to form water with the oxygen in the coal is available as a source of heat, namely 4-479. In 1 part by weight of coal, 0*04479 of hydrogen and 0-2368 of WASTE GASES OF COKE-OVENS. 179 carbon will suffer combustion ; and the number of units of heat which will be evolved by the conversion of the hydrogen into water and the carbon into carbonic acid will be respectively 0-04479 x 34000 (calo- rific power of hydrogen) = 1522-860, and 0-2368 X 8080 (calorific power of carbon) = 1913-344. But as the water formed from the hydrogen will pass off as steam at the temperature at least of 100 C., it is neces- sary to deduct the latent heat of that steam ; for it will be remem- bered that in the determination of the calorific power of hydrogen in the calorimeter, the vapour of water produced is condensed, and its latent heat rendered sensible. Not only must the latent heat of the steam, resulting from the combustion of hydrogen in the oven, be deducted, but also that of the water which is supposed to exist in the coal. Hence the total amount of latent heat to be deducted is 0-0485 X 9 x 537 = 234-400. The greatest number of available units of heat from the hydrogen will, consequently, be 1522-860 - 234-400 = 1288-460; and the sum of the units from the hydrogen and carbon in 1 part by weight of coal will be 1288-460 -f- 1913-344 = 3201-804. The number of units of heat resulting from the combustion effected within the oven during the coking process is easily found. The total quantity of carbon in the volatile products of 1 part by weight of coal is 0-2368 ; for 100 parts of coal yielded 33 per cent, of such pro- ducts, of which 23-68 consisted of carbon. Now, in a volume of the gases of the coke-ovens at Seraing which contained 2-004 parts of carbon by weight, 1-408 were present in carbonic acid, 0*443 in car- bonic oxide, and 0-153 in marsh gas. Hence, of the total 0-2368 of carbon in the volatile products from 1 part by weight of coal, 0-1664 existed in carbonic acid, 0*0523 in carbonic oxide, and 0-0181 in marsh gas. The carbon and hydrogen in this gas may be regarded as in a free state ; as its calorific power is nearly the mean of the calorific powers of its components. The number of units of heat resulting from the combustion of carbon in the oven is 0-1664 x 8080+0-0523 x 2473 (the calorific power of carbon when the product of its combustion is carbonic oxide) = 1473-850. The number of units of heat which will be evolved by the perfect combustion of the carbonic oxide, hydrogen, and marsh gas in the gases escaping from the ovens, may also be found in a similar manner. The relation in weight between the carbon and hydrogen in those gases is 2-004 : 0-130 ; and the relation in weight between the hydrogen existing in a free state and combined with carbon is 0*079 : 0-051. Hence, in the volatile products from 1 part by weight of coal, 0-0933 of hydrogen will be free, and 0-0603 combined ; and as the combined hydrogen may, for the reason assigned above, be regarded as free, the number of units of heat which will result from the combustion of the hydrogen is 0-1536 x 34000 0*1536 x 9 x 537 (latent heat of the steam produced) = 448-0051. The number of units of heat which will be developed by the combustion of carbonic oxide is 0-0523 x 5607 = 293*246 ; and by the combustion of the carbon combined with hydro- gen, 0-0181x8080 = 146-248. From the preceding data it appears that the total maximum number N 2 180 ECONOMIC APPLICATION OF THE of 'units of heat, capable of being evolved from the perfect combustion of the carbon and hydrogen, separated from 1 part by weight of coal during the process of coking, is 3201-804. Now, by deducting from this number the sum of the number of units of heat resulting from the combustion of carbon in the oven (1473-850), and of the number yet to be evolved on the complete combustion of the gases as they escape from the oven (887*4992), we obtain the number of units of heat which is evolved by the combustion of hydrogen in the oven, namely, 840-455. The number of units of heat, therefore, actually evolved during the process of coking 1 part by weight of coal, is 1473-850 + 840-455 = 2314-305. Hence the ratio between the number of units of heat evolved and that which remains to be evolved is nearly as 2 : 1. The maximum number of units of heat capable of being developed during the coking of 2750 kil. of coal is 8804961, of which 2440622 may be made available by the combustion of the gases escaping from the oven. The preceding approximate calculations must be correct, if the data upon which they are founded be true. But let us now inquire whe- ther we have reason to doubt the truth of these data. In the gas of mean composition, 24*353 parts by weight of nitrogen were associated with 4*338 of oxygen, which existed in combination with carbon. But in atmospheric air the same weight of nitrogen is associated with 7*273 of oxygen ; and as the nitrogen present in the gas was derived from atmospheric air, it follows that 7-273-4-338 = 2-925 parts by weight of oxygen have disappeared in consequence of its combination with hydrogen and the formation of an amount of water containing 0-3656 parts by weight of hydrogen. The relation in weight between the oxygen which has disappeared and the carbon in the gas of mean composition is 2-925 : 2*004. Hence the relation in weight between the oxygen which disappears in the coking of 1 part by weight of coal and the total weight of carbon separated is 0-3456 : 0-2368. But 0-3456 of oxygen requires 0-0432 of hydrogen, which would evolve 1468-800-208-786 (latent heat of steam produced) = 1260-014 units of heat. Now, according to the previous calculation, the number of units of heat evolved by the combustion of hydrogen in the oven is 840*455. It may, therefore, be inferred that the data of Ebelmen cannot all be true. The composition of the coal, and the weight and elementary composition of the volatile products separated during the process of coking, may be exactly determined ; but it must, obviously, be very difficult to arrive at the correct average composition of the gases evolved during the whole process of coking. The three specimens of gas which Ebelmen collected at intervals of 2 hours, 7j hours, and 14 hours after the commencement of the process, are assuredly quite insufficient for the purpose. Thick yellow smoke is copiously evolved at an early stage of the process ; but the matter which produces this smoke, and which cannot be inconsiderable in amount, is not repre- sented in the gas of mean composition. It seems, therefore, most probable that the opposite conclusions concerning the amount of hydrogen consumed during the process of coking are due to an error in the average composition of the* gases/ WASTE GASES OF COKE-OVENS. 181 1 1 1 * r 1 ^^ CO CO CO (M II II + "+ **. & C ,0 c 11 . ' 'S s or S> gj I* ! US j 1 rC f b the combustion i in the oven ... obtainable by pnmVme'hinn r\f 3 oxide, marsh hydrogen after ;ft the oven / = Difference ... ^ 1 ^ E< I r2 CS t J '3 !i A3 5: f :^^^ ^H "-3 g ?! ili-3 V C g-s o ^ ! o S? "s P ^ b .S o -2 'S ^ fl i^ >5 II t "1 -^/ il Available units of heat. 1 f CO CO rH CO Ci (M 73 rH rH j^ 1 CO -h C5 rA o IO CO CO 5 1 1 COCO Tfl Tfl O a I? co co (M CO CO CO O ^ CM (M c -^ ' .cso.floj g)[5pH as " ! 1 ^ 1. f s '' 1 J I * "1 1 S If 1 f S | 1 1 I 1 t 1 1 f | | 11 1 1 Si ! I.I * i 1 '*.! * " 1"! * S| 1^1 ^ SI 1 a 1 * 1 I 1 5 j < rll.l if Hi ii Hi if iitii^.ia S?A "S^ "^ S 'P t. a -11 8 ^SJ'ifirf^e ^^ 1.1 o^d -S 0 .fOINlOXlOClOMlOlogoCM foS OOrHd >4fO-*Odi-i3o2cj? odo oooooooofli'C I 9 *S saCG BQ..QOQQQQ|-!3j'li c v S a 5 _2 J5 Li, K PM O5 W <1 P o rHCMCO-* >0DtOOO30i-l(NCO't>03-t-00 214 FIRE-CLAYS. 1 c 8 ji s tl is* ill it! gritty sand being separated by washing. No sulphur was detected by boiling with nitric acid. 2. Best Stourbridge clay, such as is used at Messrs. Chance's works. 3. Darker than No. 1. Silica partly free. No. 1 and this clay were fashioned into small prisms with sharp edges, and ex- posed in the same covered crucible to a very high temperature. No. 1 becomes pale brown by buining, and No. 3 grey. No. 1 is decidedly mure refractory than No. 3, which is more glazed on the sur- face. The two trial pieces adhered finnly together where they had been in contact. The alkali was determined with great care. The results were as follow : 5^1gf 1^ Jslfl-a --II Ii*lfe1 2 IT ^lllil'lllil"! ^1* Ulil'!:l 2" S2" 'S'E'f H -2 S^^'l |-| " ^ -.it '^ t S| > |oc" < . ~ fe" fl 3 o : : : IM a 1 IM i co IM I-H in ' o CN co IM 35 | . : S t- CO ^H o S^S^aS s : | : : :::::: : 1 : : : : : : | | 1 : : : : c* r- 10 CD I-H S CO CD jj^HCD 1O * CO "J CO "S^5 ^ fl ^0_0 i : : | : Ifjs : M 00 o o Cl "^ ^ CO Oi Oi 00 d Oi GO * I CS| 1 ? S3 S S S: S f ^ 8 S g S ?? S ^ S s 8 s M *- ^ ^H IM IM * C O lO 03 00 00 10 N if : : : 03 C Stourbridge, Wor- ) t cestershire 3 4 i 4' * 1 ! o 1 SZ o fl o l| | | |1 N CO * IO CO t- 00 O3 O i-H ^ FIRE-CLAYS. 215 r - >, -s-Z .2 3 S S -2 if s "S .SaSarc.SAg'SjScd s " s g 5': -s'p>. -^.^jsl^ssl ||l|li^l|| s 1 IS li Ull if 3 = '=_-'*- ^ ?= % -Vtr as "a "8* rrS'SiB ^aT Eiii.3 rf-ar a a. cs =2 -J3 i O 00000 OOrH Or-lOOO CM CM S S 2 2 % a . a a 216 EARTHEN OR CLAY CRUCIBLES. CRUCIBLES. The term crucible is generally applied to earthenware vessels in which metals or other substances are heated in furnaces ; but the term is not restricted to 'vessels of earthenware : it is extended to all open- mouthed movable vessels, of whatsoever material composed, in which substances are in any way exposed to high temperatures. Thus, there are crucibles consisting of graphite, iron, platinum, and lime. In English it is usual to confine the use of the term crucible to vessels which can be conveniently handled by means of tongs or forceps, and which are not intended to remain as fixtures in a furnace during any lengthof time. In French the term crucible (creuset) is applied not only to large movable vessels, but even to the hearth of a blast- furnace. It has been asserted that the origin of the word crucible is due to the superstitious practice of the alchemists of marking these vessels with the sign of the cross ; and, as Latin was the language in which they wrote, the word was derived from crux, crucis. 6 Hence the expression crucial experiment. The use of superstitious signs is not yet extinct with physicians, for in their prescriptions the pro- longed tail of the letter !, which is intended as an abbreviation for Recipe, is a relic of the astrological symbol of Jupiter. 7 The qualities which may be required in crucibles, according to the special purposes to which it is proposed to apply them, may be thus enumerated : a. They should resist a high temperature without fusing or softening in a sensible degree. b. They should not be tender while hot, so as to be liable to crumble or break when grasped with tongs. c. In some cases they should resist very sudden and great alternations of temperature, so that they may be plunged while cold into a nearly white-hot furnace without cracking. In other cases it is only n'eces- sary that they should resist a high temperature after having been gra- dually heated. d. In some cases they should withstand the corrosive action and permeation of such matters as melted oxide of lead. e. They should all resist sufficiently well the corrosive action of the ashes of the fuel with which they may be surrounded, I am not aware that any crucibles have been made which combine in the highest degree all the qualities above mentioned. Crucibles must, therefore, be selected with special reference to the conditions to which it is intended to expose them. EARTHEN OR CLAY CRUCIBLES. These crucibles are made of fire-clay in admixture with silica, burnt clay, or other infusible matter. When clay is dehydrated by heat, it contracts considerably ; and even after 6 Johnson's Dictionary, London, 1805, ; marked with a cross." 9th ed. Longman and Co. " Crucible 7 Pharmacologia. By J. A. Paris, M.D., a chymist's melting-pot, made of earth : j F.R.S., 8th ed., 1833, p. 13. so called because they were formerly EARTHEN OR CLAY CRUCIBLES. 1M 7 the complete expulsion of its water of combination it may suffer much further contraction by exposure to a higher degree of heat. 'Wedgwood in the construction of his pyrometer availed himself of the property which clays thus possess of shrinking or contracting in a degree corresponding to the degree of heat to which they are subjected. In proportion to the amount of contraction which a clay crucible suffers when heated will be its liability to crack. In order to counteract as far as practicable the evil due to contraction, clay crucibles are made of a mixture of clay and of some other substance, which either expands somewhat by heat, or at least does not contract in a sensible degree. The other qualities required in the substance added are infusibility at high temperatures, and freedom from any tendency to fuse or soften the clay, conditions well fulfilled by highly burnt fire-clay, silica, and carbon in the state of graphite or coke. The unburnt clay must be reduced to powder by grinding ; and complete disintegration may be effected by allowing the clay to be freely exposed to the action of the weather during a considerable time, especially the winter season. The burnt clay is generally prepared by grinding crucibles which have been used, or, still better, glass-pots which have been exposed to a high temperature during a long time. Failing these materials, fire-clay may be burnt expressly for the purpose. Before grinding up pieces of old pots, their surfaces, which are generally more or less vitrified, and have extraneous matter adhering to them, should be carefully chipped off. The size of the particles of burnt pot is a matter of importance. Brongniart insists that variation in the size of the particles is an advantage. The pro- portion of burnt to raw clay may be altered according to circumstances. A mixture of about -f by measure of raw clay to -^ of burnt clay is suitable for man)'- crucibles. But on purely practical points of this kind experience alone will enable the operator to decide. The addition of an infusible substance like silica, graphite, or coke to raw clay may not only tend to counteract the evil due to contraction of the clay, but may also fulfil another important purpose. Crucibles composed of good fire-clay may be softened and lose their shape at the high temperatures occasionally produced in furnaces. In such a case the presence of an infusible substance may act, as it were, the part of a rigid skeleton to the crucible, and prevent its collapsing in the furnace. According to Berthier, if silica be employed, it may, by long exposure to a high temperature, gradually enter into combination with the clay, and form a more or less homogeneous and pasty mass : and this evil will be specially likely to occur if the particles of silica are too fine. 8 In all mixtures intended for crucibles, it is requisite that there should be a sufficient quantity of raw clay to produce the proper degree of plasticity for working. It is important to note that the fusibility of a crucible or fire-brick may not be altogether dependent on its ultimate composition, but that it may to a certain extent depend upon the manner in which its proximate Traitcdcs Essais, 1. 66. 218 EARTHEN OR CLAY CRUCIBLES. constituents arc mixed. Granite affords a fitting illustration of this fact. This rock consists, as is well known, of an agglomeration of quartz, felspar, and mica, in particles of considerable size, compara- tively speaking. Now, \vhen a piece of granite is exposed even to a very high temperature, it will not melt ; but if it is reduced to powder and then similarly heated, it will melt with facility. The selection of a fire-clay for the manufacture of crucibles should be carefully made, especially in the case of those of large size in which a valuable metal is to be melted. Attention should be particu- larly directed to the presence of iron pyrites, which in clays from the coal-measures may exist irregularly disseminated in particles. W hen a crucible made of such clay is heated strongly for any length of time, little cavities, or even perforations, will be produced in the substance of the crucible wherever particles of pyrites may have been imbedded. The pyrites is converted into oxide of iron by oxygen derived from the gases of the furnace, which readily permeate most crucibles. It might be supposed that at the high temperature of our furnaces filled with ignited fuel oxidation could not at any time occur ; yet that it may occur I have ascertained experimentally by thus heating crucibles in the substance of which particles of iron pyrites had been expressly imbedded. When the experiment is made by substituting small pieces of chalk for iron pyrites the result is the same. The presence, how- ever, of a small quantity of lime uniformly diffused through a fire-clay may not be injurious, and may possibly tend to render the substance of a crucible made of such clay compact and close in grain. The presence of potass or soda in sensible proportion in a fire-clay would certainly make it less refractory ; but in the proportion in which they appear to exist in some of the best fire-clays their effect may be bene- ficial rather than otherwise, by soldering, as it were, the particles firmly together. I believe that the presence of fixed alkali may be detected in all clays ; and no analysis of a clay should be considered of much practical value in which its proportion has not been determined. Potass appears to be present more frequently and in greater quantity than soda ; but the late Mr. Henry informed me that in certain Welsh fire-clays he found from l to 3 per cent, of soda. As all clays appear to be the result of the slow decomposition of felspar or some other similar mineral, which contains either potass or soda, by the action of water, the retention of some of the alkali in the residual clay might be expected. The quality of a fire-clay in reference to its fitness for the manu- facture of crucibles may be satisfactorily tested on a small scale, either in an air or small blast-furnace. For this purpose it should be kneaded with water and fashioned into small prismatic pieces with sharp edges ; and these when dry should be enclosed in a covered crucible and sub- jected to a very high temperature. The edges of the pieces should afterwards be examined : if they continue sharp, the clay may be regarded as very refractory ; but if they are much rounded, it is an evidence at least of incipient fusion ; if the pieces are melted down, the clay is worthless. With a little practice in experiments conducted in EARTHEN OR CLAY CRUCIBLES. 219 this way very satisfactory results may be obtained. It is necessary to enclose the trial pieces in a crucible in order to eliminate the effect of the ashes of the fuel. The quality of a clay crucible may be readily determined by an actual trial under the special conditions to which it is intended to be subjected. To test the property of resisting corrosion, protoxide of lead, or, still better, a mixture of that oxide and dioxide of copper, may be melted in the crucible. If a clay crucible will stand this test for any length of time without being permeated or corroded in a very sensible degree, no other need be applied. Generally, after such a trial of a few minutes only, ordinary crucibles will be very much corroded and perforated. The substance of the crucible will not be uniformly removed, but will be found to be eaten away irregularly into cavities ; and perforation will, probably, only have taken place in one or two spots. The grain of ordinaiy clay crucibles is very irre- gular, owing to the nature of the mixtures of which they are made ; and hence the great irregularity of action of the oxide of lead. It may now be conceived how regularity of structure and fineness of grain should tend to insure uniformity in the corrosive action of the oxide of lead, and so to prevent perforation. It is necessary to distinguish between simple infiltration and perme- ation due to corrosion. The body of some crucibles may be so porous as to admit of their being readily infiltrated by liquids which exert no corrosive effect. As a general rule, clay crucibles will resist permea- tion and corrosion in proportion to their fineness and regularity of grain ; but, unfortunately, in the same proportion is their liability to crack increased. The property of resisting corrosion is not neces- sarily connected with that of infusibility or softening at very high temperatures ; but rather the reverse. Crucibles are used both in the unburnt and burnt state. Small cru- cibles are generally kiln -burnt before they are used. The large Stour- bridge clay crucibles or " casting-pots," which are extensively employed in the brass foundries of Birmingham, are never previously burnt. They are gradually and thoroughly dried by the maker in chambers artificially warmed ; and are' generally kept by the founder in a dry warm situation on shelves in the casting-shop. They will hold forty pounds or more of melted brass. In Birmingham these crucibles are heated in rectangular air-furnaces, about 10 inches square and 2 feet deep. The fuel used is coke. When the furnace is cold, the fire is made and covered to the depth of a few inches with coke, upon which the crucible is placed inverted. The furnace is then filled with coke, and the crucible is gradually heated to redness. When red-hot it is taken out of the furnace, and immediately put in again with the mouth upwards. When the furnace is hot, and a fresh pot is needed, the fire must be allowed to go down, and the second pot heated with the same precautions as the first. If the pots were put into the furnace at first with the mouth upwards, they would almost certainly crack, notwith- standing every precaution. I have made numerous attempts to use crucibles of the same size and make, after having been kiln burnt, but 220 STOURBRIDGE CLAY CRUCIBLES. I could not succeed in heating them to redness without cracking ; although for special reasons I was very desirous of doing so. It may be stated as a general fact that the tendency of an earthen- ware crucible to crack increases with its size and thickness. In the case of crucibles which are used in the burnt state, it may also be observed that their liability to crack is increased by over-firing during the process of burning them in kilns or otherwise. Stourbridge day Crucibles. The materials used in admixture with the clay are burnt clay and coke. The burnt clay is obtained by pounding and grinding old glass-pots, from the surface of which any adherent glass has been carefully chipped off: only the best Stourbridge clay is used in the making of these pots. The clay is ground under edge- stones, which revolve on edge in the usual way round a circular bed, and the ground clay is passed through a sieve. Thus prepared, it is next mixed with the proper amount of burnt clay and water, and kneaded by men who tread barefoot upon it until it has acquired the right degree of consistency. These crucibles are made by hand in the following way. The workman sits before a bench, in which is a wooden block of the shape of the cavity of the crucible. At the widest end of the block is a flange or projecting border of the same width as the wet crucible is to be made thick at the mouth. In the middle of the same end an iron spindle is inserted, which fits into a socket in the bench. The block may thus be made to revolve in the position indi- cated in the annexed woodcut ; it is not fixed, but may be taken out or dropped into the socket at pleasure. On the narrow or upper end of the block is placed a lump of tem- pered clay, which the workman then moulds round the block by first striking it with a flat piece of wood, and then slapping with both hands, so as to turn the block more or less each time, as occasion may require. The clay is thus rapidly extended over the whole block down to the flange. A sliding vertical gauge is fixed in the bench near the block, by means of which the thickness of the sides and bottom of the crucible tig. 57. mav ^ regelated. As soon as the moulding is finished, the block is lifted out of its socket and inverted, when the crucible, with a little easing, will gently drop off. The spout for pouring out metal is then fashioned by the finger. The clay may likewise be moulded upon a linen cap wetted and slipped over the block, so that, on inverting the block, the crucible and cap slide off together, after which the cap may be easily pulled out. The wet crucible must be very gradually dried. These crucibles are very extensively employed by brass-founders in Birmingham. After they have been once used in the furnace and allowed to become cold, they are worthless. They are largely manufactured by Messrs. King of Lichfield Street, Birmingham ; and I can testify to the excellence of the articles of that firm from an experience of 20 years. The muffles which they have made for me of the same materials as their crucibles are superior in every respect to any others which I have tried. They make 25 sizes of crucibles ; the smallest contain 10 Ibs. of metal, and CORNISH CRUCIBLES. 221 sell at 2s. per dozen, and the largest 140 Ibs., and sell at 11s. 3rf. per dozen in the unburnt state. A convenient mould for making crucibles is represented in the annexed cut, which is a vertical section through the centre. a is a vessel of cast-iron, exactly similar in shape to a com- mon flower-pot, open at the top, and perforated at the bottom ; b is a disc of wrought-iron, having a spindle fixed in the centre ; c is a flat ring of cast or wrought iron, of which the internal diameter is less than that of the upper end of a, it is made with a rim on the under side, which loosely clips the top of a ; d is a plug of cast-iron, which passes through and rests on c, it should be turned in a lathe ; e is empty space round the depending part of the plug, in which the crucible is to be moulded. To this end the mould must be placed on a solid block of wood, in the centre of which is a hole to receive the spindle of b ; the plug is then Fig- 5S- withdrawn, and a lump of clay, somewhat more than sufficient for one crucible, is put into the vessel a, on the top of the disc or piston, b ; the plug is then rammed home by means of a suit- able hammer until it comes to the position in which it is shown in the woodcut ; the superfluous clay escapes through the ring or cover, c, around the plug. The space e is thus solidly filled with clay, and a crucible formed. The plug is then taken out, the disc, 6, forced up by means of the spindle, and the crucible raised out of the mould. A spout may then be fashioned by hand in the usual manner, and the crucible removed to a suitable place to dry. In 1762 a patent was granted to William White for a "new invented manufacture of cruci- bles for the melting metals and salts," &c. The specification directs that Stourbridge clay and Dorsetshire clay are to be mixed with Wool- wich sand and water ; and that the mixture is to be trodden with the feet. 9 The use of coke in admixture with Stourbridge clay in the manufacture of crucibles was patented by Anstey : he directed that two parts of fine ground raw Stourbridge clay should be mixed with one pint of the hardest gas coke, previously pounded and passed through a sieve of half-inch mesh. 1 ( 'firnish Crucibles. These crucibles are manufactured on a large scale in Cornwall for the use of copper- assay ers. According to Price the first manufactory of these crucibles was established at Truro a few years before the publication of his w^ell-known work ; 2 and a premium was awarded to the inventor of them by the Society of Arts. They are generally made round, and of two sizes, of which one fits into the other. Those of the larger size are 3 inches in diameter at the top and 3 inches high, outside measure. They are coarse in grain, and their surface has a grayish-white colour. They are spotted both within ;m oualitv . . ( Hafnerluden, Mo- 1 I ravia J Kaisersberg ( India, from the"! 1 Himalaya / Ceylon, unpurified f Do. coarsely 1 \ purified I Do. crystallized 1. By Karsten, op. cit. It contained also traces of chromium and magnesia. Surely so large a pro- portion of titanic acid cannot be present in this graphite. 2, 4, 5. By Ragsky, Kenngott, Uebersicht, 1856, p. 119. 3. By Berthier, Tr. des Essais, 1, p. 50. 6. By Ferstl, Jahrb. d. k-k. Geolog. Reichsanstalt, 1854, p. 869. 7-10. By Prinsep. No. 3 contained a trace of pyrites. Graphite intended for crucibles is ground and sifted ; and the powder is mixed with a sufficient amount of refractory clay to render it plastic, as it possesses no plasticity of itself. The well-known cru- cibles of Passau are said to be made of a mixture of 1 part of clay from 5 Archiv, 1st Ser. 12, p. 93. 6 Dumas and Stas. Ann. de Chimie et de Phys. 3rd ser. 1, p. 26. GEAPHITE, BLACK-LEAD, OR PLUMBAGO CRUCIBLES. 227 Schildorf and from 2 to 3 parts of an impure graphite which occurs in gneiss in that locality. 7 Good black-lead crucibles may be characterised by the following properties : they support, even when of the largest size, the greatest and most sudden alternations of temperature without cracking ; they may be used after repeated heating and cooling, so long as they are not too much reduced in thickness by the burning away of the graphite to bear the weight of the metal which may be melted in them and to admit of being held by tongs without breaking ; their surface within as well as without may be made very smooth, so that particles of melted metal will not hang about the sides ; an advantage not possessed by any other crucibles in the same degree, and much valued in the casting of metal for coining, as it may be poured out perfectly clean or free from particles derived from the crucible : 8 they have, however, the disadvanatge of being expensive. Although the graphite may burn away in a greater or less degree on the surface of a crucible, yet it is pretty well protected by the clay with which it is mixed from the action of any free oxygen which may exist in the gases of the furnace. When these crucibles are kept some hours at very high temperatures, much of the graphite may burn away and the exposed clay be melted into slag. It is usual with some persons, before using black-lead crucibles, to coat them externally by dipping them in a mixture of the consistency of cream, prepared with clay and water, containing borax in solution. Some years ago I had occasion to employ black-lead crucibles of large size, capable of holding from 40 to 60 Ibs. of metal, and to subject them during many hours continuously to the highest tempera- ture of air-furnaces, in which nickel could be easily melted by pounds at a time. I found that while some stood perfectly well under these conditions, many were quite worthless. I did not previously coat them with clay, and notwithstanding they resisted well. This differ- ence in quality must be due either to the amount or quality, or both, of the matter associated or mixed with the graphite. Although graphite is of itself infusible, yet the foreign matter with which it is mixed, either naturally or artificially, may become soft and pasty at very high temperatures, and so render the crucible more or less yielding to the tongs. We have found the black-lead crucibles manu- factured by Mr. Euel, of High Holborn, to be of excellent quality, and capable of resisting the highest temperature of our air-furnaces in which we can melt manganese or wrought-iron with facility. I used them with perfect success in numerous experiments on the artificial formation of silicates, which, in many cases, required a very high and long-continued heat. It should be remembered that natural graphite frequently contains a sensible amount of oxide of iron, which may be reduced to the metallic state by heat, so that matters melted or heated in such crucibles may become contaminated with iron. The black- lead crucibles which Mr. Ruel placed in the Great Exhibition of 1851 were reported on very favourably by Mr. Henry, who subjected them to severe tests. "Mr. Ruel," writes Mr. Henry, "has so improved the Knapp, Lehrbuch d. chomisch. Technologic, 1847, 1, p. 598. 8 Ibid. Q 2 228 MOULD FOR MAKING SMALL CRUCIBLES. black-lead or plumbago crucibles as to drive the foreigner out of the Eng- lish market. I have repeatedly tried them, and found them excellent. Messrs. Brown and Wingrove, the gold-melters, use them exclusively." 9 That firm furnished the Jury with the following written opinion of their experience of these crucibles : They are " the best quality for all the most important purposes for which such utensils are required." Of late we have employed black-lead crucibles manufactured by the Patent Plumbago Crucible Company at Battersea, and have found them in quality equal to those of Mr. Euel. Of the black- lead crucibles manufactured on the Continent, those of Passau have long been highly esteemed. At the Great Exhibition of 1851 the firm of Messrs. Lorenz, Kapeller, and Son, of Hafnerzell, near Passau, exhibited a series of these crucibles, which Fig. 62. Section of juieffs were reported to-be highly refractory; they were crucibles, to the scale of J. .. & J ,-,-, , well made, and one was unusually large, being 2 feet high and 20 inches wide. Annexed is the section of a plumbago crucible manufactured for tin-assay ers by Juleff of Eedruth. Mould for making very small Crucibles. More than ten years ago Mr. ^0%%^ Einmann, of Sweden, presented me with a mould which he informs me was employed in Sweden for making small crucibles used in assaying iron-ores. 1 Ever since we have employed crucibles of this kind with great advantage, not only in assaying iron-ores, but also in metallurgical experiments of various kinds. A description of the mould has appeared in Swedish, and more -recently in German. The following descrip- tion of it includes certain altera- tions in details, which Mr. Smith, in the course of long practice in our assay-laboratory, has found it expedient to adopt. Fig. 63 is a vertical section through the cen- tre ; fig. 64 is a horizontal section on the line A B, fig. 63 ; and fig. 65 is an elevation of the part/, seen in section in figs. 63 and 64. All the figures are drawn to a scale of \. a is a short hollow slightly conical piece, open at both ends, made of gun-metal, b b are short Fig. 63. Fig 64. Jury Reports, p. 585. 1 It is described and fi figured in the Jern-kontorets Annaler, 1852, p. 56. LINING CRUCIBLES WITH CARBON. 229 pins of iron inserted one in each side of a, near the bottom or narrower end. c is a round block of wood, in the centre of which, on the upper surface, is a circular cavity large enough to receive the lower end of a, including the projecting pins, b b ; through the middle of this cavity is a hole, A, and upon the bottom lies a disc of gun- nietal, as seen in fig. 63, which also has a hole in the centre of the same size as h. Around the edge of this cavity is screwed a flat ring of brass, d d, which projects inwards to the extent shown by the dotted line in fig. 64. c c are notches to allow the pins, b 5, to pass through, so that a may be turned in the position seen in fig. 64, when the pins, 66, will hold it firmly under the projecting edge of the ring of brass, /is a circular wooden plug, in the bottom of which is fixed the iron pin, g ; by means of this pin the plug is kept upright exactly in the centre of the mould. Fig. 65 is another wooden plug, without a pin, but having an oblique groove, i. The inner surface of a is very slightly oiled, and also the outer surface of /, where it passes into a. a being adjusted, as shown in fig. 64, a small lump of well-tempered clay is put into a, when the plug,/, is forced down and turned round ; the excess of clay escapes from the upper edge of a ; /is now withdrawn, and a small bit of clay is dropped in, when/', fig. 65, is forced down and turned round and round ; the excess of clay escapes by the groove, i. f is then taken out, and a, with the con- tained crucible, detached. The crucible may be removed by being pushed gently upwards with a circular piece of metal or wood applied to the bottom. The mixture we usually employ consists of raw and burnt clay (the latter Fi e- 65 - being obtained by simply burning the raw clay), in the proportion of about 2 measures of unburnt to 1 measure of burnt clay. The crucibles are burnt in a muffle before they are used. Lining Crucibles with Carbon. This is done when it is necessary to protect the crucible from the corrosive action of matter which may be heated in it, or when a small quantity of a metallic compound is reduced, of which every particle of the metal must be collected, as in the assaying of iron-ores in the little crucibles just described, when not more than 10 grains of ore are operated upon at a time. The minutest portions, which may. become entangled in the lining, may be separated by gently crumbling it to powder and applying the magnet. In the case of a difficultly-reducible oxide, like protoxide of manganese, the oxide would powerfully corrode an earthen crucible at a tempera- ture far below that at which reduction occurs; but a carbonaceous lining would effectually prevent contact between the oxide and the sides of the crucible. The atmosphere of a crucible so lined will always be reducing at a high temperature, owing to the carbonic oxide which must always be present. When the lining is sufficiently thick, it may not even be necessary, in certain cases, to mix the oxide intended to be reduced with carbon. Such a lining may tend to prevent the crucible from sinking down upon itself when softened by an intense heat and exposed to the weight of the fuel. Various 230 COVERS OF CRUCIBLES. kinds of carbonaceous matter may be employed, according to circum- stances. In the case of very small crucibles, charcoal powder, mixed with sufficient gum-water, starch, paste, or treacle, just to make it adhere together by pressure, should be gently rammed in so as entirely to fill the crucible. A cavity may then be made in the charcoal by boring with an instrument like fig. 66, and the surface of the cavity so made may be afterwards rendered per- fectly even and smooth by pressing down the instrument, fig. 67, and turning it round. A section of a crucible thus prepared is represented in fig. 68. The diameter of the cavity should gradually diminish from top to bottom. For some time we used lamp-black, which has the advantage of being easily and solidly compres- sible without any addition; but we abandoned it on account of the impurity of the commercial article, which contains sulphates. The charcoal lining answers perfectly : we have occasionally replaced it by little crucibles of charcoal ffom a solid wood like box or ebony. Large cru- cibles may be coated internally with the same mixture of charcoal as is used in the small ones ; but a mixture of anthracite powder, or fa Q p OW( j er of gas-retort carbon, and gas-tar answers still better. The carbon crucible, for such is the lining, may be conveniently made in a separate mould, and afterwards carefully dried, then imbedded in anthracite powder or coke- dust in a closed iron box, and exposed to a red heat. The box should have an overlapping lid, and, after heating, should not be opened until the contents are cold. When gas-retort carbon is used, a vessel may be obtained in this way which, when struck, yields a sonorous ring like pot or metal. One of these vessels may be dropped into a clay crucible just large enough to receive it, and, when properly treated, it may be used more than once. A carbonaceous lining is, as has been previously mentioned, termed brasque by the French, who employ the verb brasquer to express the lining of a crucible with carbon. As the words are short, and often used in English, they will be adopted in this work. Covers of Crucibles. When necessary, the mouth of the crucible may be fitted with a cover, made of the same materials as the crucible. Such covers may be easily made by cutting or stamping them out of the clay mixture rolled out upon a flat surface. Pieces of old crucible or thin fire-bricks, termed split-bricks, may also be conveniently used for the same purpose. Sometimes it is desirable to lute on the cover with clay. In the case of small crucibles a bit of old crucible may be stuck into the mouth and plastered over with clay. For the small crucibles prepared in the mould, of which a description has been given, we are accustomed to make nicely-fitting covers in a mould of the following construction, and which is shown in the annexed woodcut in vertical Fig. 63. To the scale of . The boring parts are made of gun-metal. SEFSTROM'S BLAST FURNACE. 231 section through the centre, a is a cylindrical piece of wood, upon which is placed a cylinder of brass, b b, perforated with several holes, c c, &c. ; c/, a cylindrical plug of wood hollowed out at the bottom, as repre- sented in the woodcut. This plug fits into the brass cylinder, b 6, upon the top of which it is supported by a shoulder, so as to leave a space, e. A small lump of clay being put into the cylinder, the plug d is pressed down and turned round, when a cover is moulded in the space e, the excess of clay being expelled through the holes, c c, &c. The plug is first withdrawn, and then the cylinder, after which the cover, which is formed with the top or flat side downwards, may be detached and left in a warm place to dry. Crucible stands. When crucibles are heated in a common air-furnace, it is desirable to Fi s- 69 - Moi i ld for making covers, to the scale of . support them on a stand about 2 or 3 inches above the bars. Supports of refractory clay are sometimes expressly made, but bits of fire-brick, chipped to a convenient shape and size, answer perfectly. Tongs for Crucibles. The an- nexed engravings are taken from photographs of various kinds of tongs for manipulating with crucibles. They are made of iron. Fig. 70, adapted for small crucibles. Fig. 71, used by the Cornish copper - assay er : the ends consist of rectangular pieces of iron. Fig. 72, of general uti- lity; but as the leverage is great, they must be used with care. Fig. 73, for large crucibles, such as are used by brass-founders. When these tongs are employed to take out of the furnace a large crucible containing a heavy weight of metal, it is customary to slide an iron ring over the handles, in order to prevent the crucible from slipping through. Sefstroms blast furnace. For this useful furnace we are in- debted to Sefstrom, the well- known Swedish metallurgist. It is extremely convenient for F i g . o. Fig. n. Fig.w, Fig. 73, Scale of ll inch to 20 inches. 232 DEVILLE'S BLAST FUENACE. o JIN? easily and rapidly producing high temperatures. It may be made very portable, and at a cost only of a few shillings. The annexed woodcut represents a vertical section through the centre of a small furnace of this kind, in which the crucibles described at page 228 are heated. During many years I have employed this furnace with great advantage in various metallurgical experiments, and I can confidently recommend it to persons who may travel in distant parts and desire to take with them a furnace not exceeding the size of a hat. It is cylindrical, and is made of sheet-iron ; and consists of an outer cylinder a a, closed at the bottom except at /, where a short pipe is inserted for the en- trance of the blast ; and of an inner cylinder 66, completely closed at the bottom, and fixed in the posi- tion shown in the woodcut by a rim g g. There is thus formed a hollow space or chamber c c -c, into which air can be blown through the orifice /. In the inner cylinder b b, are eight small holes e e e, at the same height from the bottom, and at equidistant spaces. Into Fig. 74. vertical section through the middle, each of these holes is inserted a small nozzle of sheet-iron tapering inwards. The interior of the inner cylinder is lined with a suitable mixture of fire clay d d, which is represented by the shading of sectional lines. Any cracks which may occur during the drying of the clay must be filled up with the clay mixture, so that the lining may be rendered compact. It will be observed that the inner cylinder projects somewhat above the rimgg. Eound this projecting part a hoop of sheet-iron, h h, is dropped so as to rest on the rim g g, by which means the furnace may be heightened, and space provided in which fuel may be piled. The hoop h is formed by bending the sheet-iron and overlapping the ends, in one of which there are two small holes, i i, while in the other is fixed a button, which may be fitted at pleasure into either of the holes. The diameter of the hoop may thus be made greater or less, so as to be* placed either round the top of the inner cylinder above gg, or, when the furnace is not in use, round the outer cylinder a a. Charcoal is the fuel usually employed, and it should be reduced to pieces about as large as walnuts. A small pair of double bellows is required to produce the blast. Four of the small iron-assay crucibles of Ekman, or even more, may be heated at a time in this furnace. Sefstrb'm's furnaces may be made of much larger dimensions when portability is not an object. Devilh's blastfurnace* In the metallurgical laboratory of the School of Mines we have used this furnace during several years, and find it 2 Ann. de Chimie et de Phys. 3. s. 46, p. 190. 1856. DEVILLE'S BLAST FUENACE. 233 Fig. 75. very convenient for obtaining high temperatures. Deville has melted platinum in it. Fig. 75 is the plan, and fig. 76 is a vertical section through the centre of fig. 75. A is a vessel of cast-iron, having a circular hole b ; it is supported by a tripod, of which d d are two legs ; a a circular cast-iron plate, which forms a cover to A ; in a is a series of holes, c c, equidistant from the centre. B is a cylinder of sheet-iron, strengthened at the top and bottom by two rings of iron, ef; within it is lined with a mixture of burnt and raw fire-clay to the thickness of the rings, ef ; any cracks which may be formed during the drying of the clay lining must be filled up. The blast enters at b, fig. 76, and rises through the 10 holes in the plate c. The crucible is placed on a stand in the cen- tre. The construction of this furnace is such as to produce an intense heat over a wide surface, but within a very limited height from the bot- tom. Coke, not too dense, of about the size of walnuts, is an excellent fuel for this fur- nace. Deville uses cinders or "breezes" from the ash-pit of a furnace in which the non- caking coal of Charleroi is burnt; the "breezes" are sift- ed and thus obtained, varying in size from a pea to a hazel- nut. Ignited charcoal is first introduced to the height of 2 or 3 inches, then pieces of coke of the size of walnuts, and lastly " breezes." According to Deville's experience, at the high temperatures which he obtains, and which he charac- terises as blue heat, the best earthen crucibles become as liquid as glass. He accord- ingly prepares crucibles of lime, carbon, or alumina. A piece of well-burnt, slightly hydraulic lime, is cut by means of a saw or knife into a rectangular prism with a square base, 3 or 4 inches on the side, and from 5 to 6 inches high. The edges of the prism are roughly rounded off, and a cavity of variable dimensions is bored in the centre. Often in experiments requiring a very high temperature, a small crucible of well-selected lime is enclosed in another lime Fig. 76. 234 DEVILLE'S BLAST FUKNACE. crucible. Lime crucibles may be easily and rapidly made. Covers of lime are also employed. When cut in the manner described, no stands are necessary for these crucibles. When the substance to be heated is very refractory, only one crucible is used, of the dimensions above prescribed for the outer one ; and the diameter of the cavity should at the most not exceed | or 1 inch, so that the thickness of the walls may remain from about 1 to 1 inch. The thickness of the bottom should be from 2 to 2% inches. The temperature should be gradually raised ; and before the " breezes " are put in, care should be taken to ascertain that the lime crucible is not cracked. The carbon crucibles which Deville employs are turned in a lathe out of gas-retort carbon. They should be perfectly cylindrical, and never exceed the height of 4 inches, inclusive of the thickness of the bottom, whatever may be their width. The reason assigned for this is that the zone of maximum temperature in Deville's furnace hardly extends to the height of 3 inches. These crucibles may be freed, at least to a certain depth below the surface, from impurities, such as sulphur, iron, silica, and alumina, by exposing them, &long with their covers, in earthen crucibles, through the bottom of which a current of chlorine is introduced. The carbon often loses sensibly in weight by this process, but keeps its solidity. When used, a carbon crucible is placed in another crucible of lime or refractory clay, and the space between the two is filled with powder of alumina, which has been previously exposed to a white heat. A cover of gas-retort carbon is placed on the carbon crucible, and over the cover powder of alumina is spread and pressed strongly down. Upon the whole is placed a clay cover. During the process of heating the outer covering may melt completely, yet the carbon crucible within will remain pro- tected by the alumina, which the clinker of the fuel will scarcely attack. Alumina crucibles are made of a mixture of gelatinous alumina mixed with a proper proportion of alumina which has been previously heated very strongly during a long time. Deville prefers alumina made at a very low temperature from ammoniacal alum, as it forms a plastic mixture with water, and it is very difficult to triturate the lumps in ordinary gelatinous alumina. Instead of calcined alumina, Deville uses the powder of the product obtained by exposing an intimate mixture of equal parts of alumina and pounded marble to the highest tempera- ture of a good air-furnace. It is often scoriaceous and translucent, and resembles dried flour paste. With equal parts of plastic alumina, calcined alumina, and the product last described, or the so-called alu- minate of lime, crucibles may be prepared which soften a little at the temperature of melted platinum, but which, by being strongly heated, acquire a remarkable degree of solidity. For very high temperatures less aluminate should be added ; but it is desirable that the mixture should contain from 5 to 10 per cent, of lime. Deville remarks that however these crucibles may be made, when once baked they will stand every test. They resist sudden and great changes of tempera- ture, and almost every kind of matter, even sodium, which may be heated in them. FIKE-BRICKS. 235 Fire-bricks. The term fire-brick is applied to bricks capable of resisting high temperatures, whether made of natural fire-clay or other refrac- tory matters. They are only used in those parts of furnaces where the heat would be sufficient speedily to destroy ordinary bricks, the use of which is therefore restricted to the external or cooler parts. They are made of varying shapes and sizes to suit the manifold requirements of the furnace-builder. In large establishments, such as the iron- works of South Wales, they are made on the spot. Much of what has been previously stated concerning crucibles will equally apply to fire- bricks. They are made of raw clay ground between rolls or under edge stones, and suitably kneaded by treading after the addition of water. They are fashioned by hand in moulds similar to those used in the manufacture of common bricks. They are carefully dried in stoves, and burnt at a high temperature in closed kilns. In some establishments the powder of burnt clay is used in admixture with raw clay. The tenacity of a clay must be much affected by the amount' of free silica which it contains : when it is naturally too tenacious to admit of being directly applied, the right temper may be readily produced by the addition of a proper amount of burnt clay in coarse powder. In setting fire-bricks, fire-clay should be used instead of lime-mortar. The qualities which may be required in fire-bricks are as follow : a. They should not melt or soften in a sensible degree by exposure to intense heat long and uninterruptedly continued. b. They should resist sudden and great extremes of heat. c. They should support considerable pressure at high temperatures without crumbling. d. They may be required to withstand, as far as practicable, the cor- rosive action of slags rich in protoxide of iron. Experiment on the large scale is essential to the formation of a correct judgment as to some of the qualities which may be required. External characters alone will not suffice. Stourbridge bricks have long been highly valued for their refractory qualities, and they have been exported to various parts of the world, even far remote. Excellent fire-bricks are also manufactured in many parts of the United Kingdom with clay from the coal-measures. All these bricks have a pale-brownish colour. Sometimes they are copiously mottled with dark spots, due, I believe, to the existence of particles of iron-pyrites diffused through the clay. Brongniart mentions several kinds of French fire-bricks as being in the highest degree refractory, namely, those from Mouchy, Saint Eloy (Oise), manufactured by Deyeux; those from Septveitte, near Provins; and those from Saint Vdlier, near Oriol. He states that all these bricks were exposed to the highest temperature of a porcelain-furnace, but protected from the direct action of the potash contained in the ashes, wood being the fuel, and that they suffered no change ; whereas, Stourbridge bricks exposed to exactly the same conditions became red- brown, and were completely softened. 3 Traite des Arts Ceramique*, 1, p. 341. 1844. 236 DINAS FIRE-BRICK. In the following short table is presented the composition of fire- bricks from different localities : Composition of various Fire-Bricks. i. 2. 3. 4. 5. 6. 7. Silica . . 63-09 84-65 88-1 84-0 88-43 69-3 77-6 Alumina 29-09 8-85 4-5 14-1 6-90 28-5 19-0 Lime 0-42 1-90 1-2 0-7 3-40 Magnesia 0-66 0-35 trace. 2-8 Sesquioxide of) iron 1 2-88 4-25 6-1 0-5 1-50 2-0 0-3 Potash 1-92 Soda 0-31 Titanic acid 2-21 100-58 100-00 99-9 99-3 100-23 99-8 99-7 1. Communicated to the Author. 2. Richardson, Knapp's Technology, Trans. 2, p. 481. 3. 4. 5. Napier, Phil. Magazine, 4 s. 4, p. 348. 6. 7. By Berthier, Traite" des Essais, 1. p. 67. 1. By Riley. This analysis, I am informed, was made with extreme care. The clay from which the brick was made is known as the little vein west clay, Dowlais. 2. From Windsor clay, which is stated to be a mixture of 70 per cent, of sand and 30 of clay. 3. From Flintshire. This brick is stated to be used in the construction of furnaces and chimneys in the copper-works of Wales in parts exposed to great heat and currents of air ; but not where melted matter can come in contact with it. 4. From Lysnewydd, S. Wales : it is used for fireplaces and hearths. 5. From Pembroke : it is used in copper-works. The analysis is by Cameron, late of the Spitty Works. 6. From Creusot, France : it is used for blast-furnaces. 7. From Provins, France : it is used in reverberatory furnaces. Dinas fire-brick. This brick consists almost entirely of silica. It was invented by the late Mr. W. Weston Young, a land-surveyor, of Newton Nottage, Glamorganshire, whose original documents on the subject have been placed at my disposal by his great-nephew, Mr. Edward Young. A company was established for the manufacture of these bricks in 1822 by Mr. Young, who thus describes the history of his invention : " The material at the Dinas (the well-known rock of that name in the Vale of Neath), from which we procure it, is nearly pure silex ; but, from its lying on the limestone and occasionally inter- mixing with it, there is, taking the average of the general working, perhaps, about 5 per cent, of calcareous matter and 1 per cent, of metallic, either iron or copper. Its use as a sand was discovered about 40 years ago, 4 when the fine of it was taken to one of the copper-works and used as a cement, and for mending their furnaces while at work, by placing it with a long iron-handled ladle or spade where the wash of the metal had destroyed the brick ; and, from its remarkable pro- perty of swelling in high heats, it fixed itself firmly. It gradually 4 The description is not dated, but it was written 12 years after the bricks had come into use. DINAS FIRE-BRICK. 237 gained from one copper- work to another till its use became general : in fact, they are not able to find any other sand that will answer the purpose so well. Its fire-proof qualities being known, many attempts were made to produce a brick from it ; but all the common combina- tions of different clays, &c., failed. About 14 years ago I became acquainted with it, and soon after devised a method of producing a brick from it of very extraordinary fire-proof qualities. AVhen set in its own cement, for very high and long-continued heats it certainly will exceed in duration any other known brick. It does not suit every situation, as, in fact, no fire-brick will : the nature of it at once tells you it must not be placed near alkaline substances ; neither will the effluvia from some lead ores suit it. Perhaps it does not exceed Stourbridge (brick) for grates ; but for the body of furnaces of most kinds it exceeds, as said before, that and every other known brick in duration. The manner in which the brick is made gives it a rough coat compared with most others ; indeed, it is peculiar in this respect ; but, as it is made in machines perfectly square, all the works here prefer it with its rough coat ; they say it sits better in the work, and they have now had more than 12 years' experience. This brick ought to be kept dry if possible, for, being open in its texture, it imbibes moisture freely." " The fire-place, roofs, sides, and bridge of the furnace, also the lower part of the stack, should be built of Dinas ; the back part and the remainder of the stack will do best of the other kind (similar to Stourbridge in quality) ; slabs for leaving the flues and doors are also best made of this material. The appearance of the Dinas brick is peculiar in colour and the roughness of its surface." The mode of making the Dinas brick was long kept rigidly secret, and even now it is not generally known. The material, which is called " clay," is found at several places in the Vale of Neath, some of which I visited with Mr. E. Young (1859). It occurs in the state of rock and disintegrated like sand. Its colour, when dry, is pale grey. The rock, when not too hard, is crushed to coarse powder between iron rolls. By exposure to the air the hard rock becomes somewhat softer, but some of it is so hard that it cannot be employed. The composition of Dinas " clay," from two localities in the Vale of Neath, is stated in the following table. The analyses were made in my laboratory by Mr. W. Weston. No. 1 was rock of medium hardness, which I obtained near Pont Neath Vaughan, on the occasion of my visit with Mr. E. Young ; and No. 2 was sent to me from the same locality, though not from the same mine. 1. 2. Silica 98-31 96*73 Alumina 0-72 1-39 Protoxide of iron '.... 0-18 0-48 Lime 22 0-19 Potass and soda 0-14 0'20 Water combined . , 0-35 .. .0-50 99-92 99-49 The powder of the rock is mixed with about 1 per cent, of lime and sufficient water to make it cohere slightly by pressure. This mixture 238 SAND AND SANDSTONES. is pressed into iron moulds, of which two are fixed under one press, side by side. The mould, which is open at the top and bottom, like ordinary brick-moulds, is closed below by a moveable iron plate, and above by another plate of iron, which fits in like a piston, and is con- nected with a lever. The machine being adjusted, the coarse mixture is put into the moulds by a workman, whose hands are protected by stout gloves, as the sharp edges of the fragments would otherwise wound them : the piston is then pressed down, after which the move- able bed of iron on which the brick is formed is lowered and taken away with the brick upon it, as it is not sufficiently solid to admit of being carried in the usual manner. The bricks are dried on these plates upon floors warmed by flues passing underneath ; and when dry they are piled in a circular closed kiln covered with a dome, similar to kilns in which common fire-bricks are burned. About 7 days of hard firing are required for these bricks, and about the same time for the cooling of the kiln. One kiln contains 32,000 bricks, and consumes 40 tons of coal, half free-burning and half binding. The price (1859) is 60s. the thousand. They are manufactured of various shapes and sizes, to suit the furnace-builder. The fractured surface of one of these bricks presents the appearance of coarse irregular white particles of quartz, surrounded by a small quantity of light-brownish yellow matter. The lime which is added exerts a fluxing action on the surface of the fragments of quartz, and so causes them to agglutinate. These bricks expand by heat, whereas bricks made of fire-clay contract. On this account they are stated to be advantageous for the roofs of reverberatory furnaces, and in all parts where a solid and compact lining is needed. From their siliceous nature it is obvious that they should not be exposed to the action of slags rich in metallic oxides. Sand and sandstones. Silica is extensively used by the metallurgist, both in the state of sand and sandstone. The beds of the reverberatory furnaces in which the operations of smelting and refining are effected in the copper- works at Swansea are made of sand. Great accumula- tions of blown sand occur on various parts of the neighbouring coast suitable for this purpose ; and some of the best quality, I am informed, is met with at Briton-Ferry, near Neath. I received from Mr. F. F. Bankart, late of the Briton-Ferry Copper Works, a sample of this sand, which has been analysed in my laboratory by Mr. W. Weston. It has a brownish-yellow colour, and contains minute fragments of shells. Le Play has also determined the composition of similar sand from Swansea. The analyses are as follow : Weston. Le Play. Silica. 87-87 86'0 Alumina 2*13 1-0 Sesquioxide of iron 2'72 1-2 Lime 3-79 ............ 5-7 Magnesia 0-21 0-8 Carbonic acid and a little water .... 2-60 . 45 99-32 99-8 SAND AND SANDSTONES. 239 These results show that the composition of the sand is more uniform than might have been expected, as the sample on which AYeston operated was obtained (1859) more than ten years after Le Play's. No sensible amount of chloride was detected. The presence of lime is, doubtless, important, by tending to cement the particles of sand together into a more or less compact mass. M. Kampmann has given the following analyses of sands employed for moulds in various foundries : 5 1. 2. 3. 4. Silica ....... 92-083 91-907 92-913 90 625 Oxide of iron 2 '498 2'177 1*249 2-708 Alumina 5-415 5-683 5-850 6-667 Lime traces 0-415 traces traces 99-996 100-182 100-012 100-000 1 . Sand from the foundry of M. Freund at Charlottenburg. 2. Sand employed at Paris for bronzes. 3. Sand from Manchester. 4. Sand from the establishment of Lagua near Stromberg. According to M. Kampmann, a good sand for moulds may be artifi- cially made from the following mixture : Fine quartzose sand 93 Red English ochre 2 Aluminous earth the least possible calcareous 5 In the Museum of Practical Geology is a very fine iron casting which was exhibited at the Paris Exhibition in 1855. It is a circular disc, 40 inches in diameter, and about -fa of an inch in thickness, pre- senting a pattern of elegant perforated tracery- work ; its surface is re- markably smooth, and the casting is sharp and even : it was produced at the works of Count Stolberg-Wernigerode, at Ilsenberg, in the Harz Mountains. The sand which adhered to the surface of the casting as it came from the mould was purposely left attached, and of this a por - tion was taken for the analysis, which was made in my laboratory by Mr. J. Spiller. Silica ^ 79-02 Alumina 13'72 Protoxide of iron 2-40 Oxide of copper (CuO) trace. Magnesia 0-71 Potass 4-58 100-43 This sand is stated to consist of three different kinds of material, namely, common argillaceous sand, sand found in diluvial deposits, and sand from solid sandstone. As the first two contain clay, they are carefully heated to dehydrate the clay. The sandstone is pounded under a hammer, and mixed with an equal weight of each of the other 5 Ann. d. Mines, 4, s. 8, p. 689. 240 SAND AND SANDSTONES. two kinds of sand. The mixture is ground by iron balls in a revolving drum, and afterwards passed through a woollen cylinder, which moves up and down ; it is thus obtained in the state of the finest flour, which in moulding may be made to receive the most delicate impress. The moulds used in making the so-called " lace-castings" of cast-iron are also prepared with this flour of sand, the patterns being formed of stamped and perforated paper. A valuable casting- sand is obtained from the New Red Sandstone at Birmingham. There is a quarry of this sand at the old Cemetery, the value of which one of the Directors of the Company some years ago informed me was estimated at not less than 20,000?. Blue bricks. These bricks, which receive their name from the dark grey glaze on their surface, are made of brick-earth in moulds of the usual kind. They are fired in closed kilns, having been previously dusted over with " iron-scurf," the substance produced by the wear of the siliceous grind-stones employed in grinding gun-barrels, etc., and which consists of an intimate mixture of fine particles of stone and iron. In firing the iron becomes oxidized, and combines with the silica to form silicate of protoxide of iron ; and it is this compound which forms the glaze. They are much prized by engineers for special purposes. They are not refractory ; but I mention them as the mode of glazing may be interesting to the metallurgist. There are many natural mineral substances, such as those consti- tuting igneous and metarnorphic rocks, which are used in the con- struction of furnaces, and which I shall not particularly describe in this place, though I shall be careful to specify them in the sequel. COPPER. UNIVERSITY History. COPPER was in use -in the earliest times of which any record exists. The ancients obtained it from various localities, amongst which was the island of Cyprus, where, according to Pliny, the metal was first discovered. The copper from this island was known in the Roman market as ces Cyprium or Cyprian copper. 1 The adjective Cyprium, at first only used to express locality, became cor- rupted into the substantive cuprum, which replaced the original name, ees ; and from cuprum the English word copper is derived. Colour. It is distinguished by its red colour from all other metals, or metallic compounds, except that of titanium, which frequently occurs in iron-smelting furnaces. Lustre. It is capable of receiving a very brilliant polish. Crystalline system. It crystallizes in the regular system ; and after fusion it may occasionally be obtained in imperfect skeleton or solid octahedrons of considerable size. Mallea- bility and ductility. It possesses these qualities in a high degree. It may be rolled into very thin sheets, beaten out into leaves, and drawn into fine wire. By cold rolling or hammering it becomes hard ; but its malleability is restored by annealing at a red heat. It is immaterial whether the heated copper be cooled rapidly or slowly in the anneal- ing process. Tenacity. According to Sickingen, a wire O m - 00216 (0-0864 in.) in diameter supports a weight of 151 kiu (332-9 Ibs.) with- out breaking; 2 whereas it is stated by Berthier that a wire O m - 002 (0-08 in.) in diameter breaks under a weight of 137 kiL 4 (302-9 Ibs.). 3 Results of this kind are not of much value unless accompanied with precise information as to the degree of purity of the copper, the parti- cular method of its preparation, and the exact conditions, as to tempera- ture, &c., under which they were obtained. Specific heat 0' 0951 5, between C. and 100 C. (Regnault). Linear dilatation by heat. The published results do not closely agree. According to Troughton, the co-efficient of linear dilatation* for copper wire is 0*000019188, that is, the degree of dilatation of a unit of length for 1 C. Action of heat. It melts at a lower temperature than gold, and at a higher temperature than silver ; or, according to Pouillet's estimation of the melting-points of those metals, between 1200 and 1000 C. Before the oxyhydrogen blow-pipe it may be volatilized with facility. It is not sensibly vola- tilized in close vessels at the high temperature of a porcelain furnace : thus Berthier exposed a known weight of copper in a brasqued cru- cible to the heat of the furnace at Sevres during the whole period of one firing, and found that at the most the .loss did not exceed ^ per cent. At a temperature below, } T et bordering on, its melting- 1 Nat. Hist., L. xxxiv. cap. i. ii. 18,")1. Sillig. 2 Bcrzelius, Tr.de Chimie, 1846, 2, p. 519. 3 TF. dcs Essais, 2, p. 395. 242 ACTION OF OXYGEN DIOXIDE OF COPPER. point, it becomes so brittle that it may be readily reduced to powder by trituration. It is usual in foundries to break ingots 'of copper in pieces while thus heated. The fractured surface of an ingot broken while strongly heated is coarsely fibrous or columnar ; and amongst the fibres I think I have perceived imperfect octahedrons. When comparatively pure copper is melted and poured into a mould without exposure to oxygen, the upper surface sinks in considerably on cooling, and a sound ingot is obtained. However, according to Karsten, 4 the surface of pure copper rises in the mould during solidi- fication after fusion : this point will be fully discussed hereafter. Atomic weight. 31*648. Action of oxygen. There are two oxides of copper, a knowledge of which is important to the metallurgist, namely, the red, or dioxide, and the black, or protoxide. At the ordinary temperature of the air copper is not acted upon either by dry or moist oxygen ; but when ex- posed to moist oxygen and carbonic acid, its surface becomes coated with a green rust of carbonate of copper, commonly, but erroneously, called verdigris, which is acetate of copper. When copper is heated to redness with access of air, its surface is converted into dark-coloured oxide, or copper scale, which may be more or less perfectly detached by plunging the copper while hot into cold water, or by bending it backwards and forwards after cooling. Sheet copper may soon be com- pletely converted into oxide by alternately removing the scale and reheating. A large quantity of this scale is produced in the process of annealing sheet copper in rolling mills. Although the scale has a superficial dark grey, nearly black colour, yet it consists almost wholly of dioxide of copper ; it breaks with a crystalline fracture, and, when in thin laminae, transmits a beautiful ruby-coloured light. The scale, in the state in which it is detached from the surface of the copper, may be exposed during a long time to a strong red heat in a muffle without passing in sensible proportion to a higher degree of oxidation ; but when it is thus heated after having been reduced to powder, it swells up considerably, and is speedily and completely changed into protoxide. When copper in a finely divided state such as is produced by precipitation from sulphate of copper by iron, or by the reduction of the oxides of copper by hydrogen at a low temperature is heated with access of air to a degree far below redness, it is rapidly oxidized. Dioxide of copper. Formula, Cu 2 0. It crystallizes in the regular system. It melts between a bright red and a white heat. It is always formed when copper is heated to redness with access of air, or in contact with protoxide of copper. It is easily reduced at a red heat by hydrogen, carbonic oxide, charcoal, or other carbonaceous matters, and metals having a strong affinity for oxygen, such as iron or zinc. When heated in the state of powder to redness with access 'of air, it is rapidly converted into protoxide. It is resolved by the action of dilute sulphuric acid into protoxide of copper, which dis- Sys. 5, p. 25 PROTOXIDE OF COPPER. 24., solves, and into finely divided metallic copper, which remains. By the action of nitric acid upon dioxide of copper, except when very dilute and cold, nitrate of protoxide is formed. It dissolves in hydro- chloric acid and ammonia, and the latter solution, which is colourless, becomes blue by exposure to air. It is used in the arts to communi- cate a fine ruby colour to glass. Protoxide of copper. Formula, CuO. It is this oxide which forms the base in ordinary salts of copper. According to Berthier, it melts at a white heat. It is as easily reduced as the dioxide, and by the. same reducing agents. Favre and Maumine state that when pro- toxide of copper is exposed to about the melting-point of copper, oxygen is given off in a regular stream, which, having once ceased, will not again occur, though the heat may be increased. 5 In four experiments the loss of oxygen varied from 8 to 8 2 per cent. The product, which was melted, was black, and consisted of 2Cu 2 -f- CuO. I had long previously found that when protoxide of copper was exposed in a clay crucible to a high temperature in a common assay furnace, it became brown and formed a sintered mass ; but 1 was not sure that the reduction was the simple effect of heat, and thought that probably the partial reduction might have been caused by the gases of the furnace. It dissolves in ammonia, forming a deep blue solution : when ammonia is poured upon it, scarcely any colora- tion will take place ; but the addition of a little carbonate or other salt of ammonia will instantly cause the blue colour to appear. Dioxide of copper heated with silica. Berthier prepared silicates of this oxide by heating mixtures of fine quartzose sand and protoxide of copper with sufficient metallic copper to form dioxide. Dioxide of Copper per cent. Silica per cent. 1. 3Cu 2 0, SiO 3 82^3 17-7 2. 8Cu 2 O, 2SiO 3 69'9 30-1 3. Cu 2 0, SiO 3 60-7 39-3 1. The product was a homogeneous button, easily detached from the crucible, and had only undergone the commencement of pasty fusion ; it was compact, tenacious, red-brown, with a somewhat metallic aspect; its powder was bright red. 2. The product was melted into a button filled with small bubbles ; its fracture was uneven, shining, and of a fine deep red violet colour. It must have been very liquid, and had partially traversed the substance of the crucible. 3. The product had the same form as the mixture ; it was tenacious and cellular ; its fracture was partly dull and partly shining ; it must have been strongly softened, but yet not perfectly liquid. A portion of the grains of quartz had risen to the surface. In making experi- ments of this kind, it is desirable in the first part of the process to employ a comparatively low and long-continued heat. Experiment by E. Smith : 3Cu*O = 1400 grains. 2SI0 3 = 600 do. Jalires-Bericht. Beiv.elius. 1F46, p. 184. l! 2 244 PROTOXIDE OF COPPER HEATED WITH SILICA. The mixture was heated strongly in a plumbago crucible. The product was fritted, but not melted, and red like dioxide of copper. A few small particles of copper were found on the exterior. Protoxide of copper heated with silica. Berthier heated the following mixture of this oxide and silica : 3CuO = 0-421 gramme. 4SiO 3 = 0-579 do. The product was only semi-fused and blood-red in colour, which proves that the protoxide had been reduced to dioxide. This experiment has been repeated by K. Smith. The proportions employed were 1160 grains of protoxide of copper and 840 of silica. The mixture was exposed in a fine-grained crucible to a high tempera- ture during two hours. The product was fritted, and very similar in appearance to that obtained by heating a mixture of dioxide of copper and silica. Its upper surface was black, and where it was in contact with the crucible it was orange-coloured. In order to be certain that the gases of the furnace in which the crucible was heated had not contributed to effect the reduction of the protoxide to dioxide, the following experiment was made in a muffle, in which the atmosphere was oxidizing : 3CuO = 580 grains. 2SiO 3 = 440 do. The materials were intimately mixed, and the mixture was exposed in an uncovered platinum dish to a strong red heat in a muffle during 3 j hours. A sound, such as is produced by the evolution of bubbles of gas from a thick liquid, was perceived during the process. The dish with its contents was left to cool in the muffle. The product was detached in one piece from the platinum ; it was somewhat com- pact, semi-fused, opaque, and brown-red ; the upper surface was black and porous. The platinum dish was attacked where it had been in contact with the mass. About half of the product was again exposed during 5J hours in the same platinum dish in a muffle to a degree of heat approaching whiteness. No perceptible change occurred, except the blackening of the surface of the mass. From the preceding experiments it may be concluded that, under the influence of silica, protoxide of copper is reduced at a high temperature to dioxide. Dioxide of copper heated with silica and alumina. The following experi- ments were made by E. Smith : 1 . 3Cu 2 O = 1200 grains. APO 3 = 280 do. 2SiO 3 = 520 do. The mixture was exposed to a high temperature during an hour and a half in a plumbago crucible. The product consisted of a vitreous, porous, dirty orange-red slag, and a button of porous copper weighing 550 grains. 2. The same quantities were used. The mixture was exposed in a fine-grained crucible during more than five hours to the highest tem- perature attainable in a muffle. The product was fritted ; it was sepa- BORATES OF COPPEK. 245 rated as completely as possible from the adherent substance of the crucible, and exposed in a Cornish crucible placed within another during about an hour to a white heat. The product was perfectly melted except on the upper surface, which was porous ; its colour was greenish orange. 3. (By myself) Cu 2 O = 240 grains. Al-'O 3 (containing 43 per cent, of dry alumina) = 133 do. SiO 3 (ground flints) = 102 do. The mixture was gradually heated to bright redness in a crucible closed by pieces of earthenware biscuit luted over with clay. The product was well melted, free from bubbles, opaque, and of a red- orange colour. Berthier made the following experiment : 3Cu 2 O = 60-0 A1 2 O 3 = 14-4 2SiO 3 = 25-6 The product was compact, free from bubbles,- with a slightly con- choidal fracture, very bright (tres eclatant), of a fine sealing-wax red colour, opaque, even in the thinnest fragments. ' Protoxide of copper heated with silica and alumina. The experiment was ' made by K. Smith. 3 CuO = 120 grains. A1 2 3 = 52 do. 2SiO 3 = 92 do. The mixture was heated in a fine-grained crucible in the same furnace, and during the same time, with the crucible in experiment 2, p. 244. The product was melted, compact, and of a greenish-orange colour, like that obtained in experiment 2. Borates of copper. The following experiments were made by I jerthier : 1. Cu 2 O -- 17 -83 grammes. 4B0 3 = 29-44 do. The mixture melted easily and became veiy liquid. The product \\ as compact, very hard and tenacious, opaque, cinnabar-red in colour, and had an uneven, slightly shining fracture. It should consist of 50-7 per cent, of dioxide of copper and 49'3 of boracic acid. 2. CuO = 9-91 grammes. 2BO 3 = 14-72 do. The mixture melted easily without any intumescence. The product was tenacious, opaque, and red-brown in colour, spotted with blue. It contained cavities, in which were brilliant prismatic crystals, some red and others of the finest blue colour. Part of the protoxide of copper must have been reduced to dioxide. The following experiment was made by E. Smith : 3. 3CuO = 240 grains. 2BO 3 = 140 do. 246 BISULPHIDE OF COPPER HEATED WITH OTHER SULPHIDES. The mixture was heated in an open Cornish crucible in a muffle, and in about 20 minutes fused easily at a red-heat ; the product was a dark greenish coloured glass when seen by transmitted light, but it was blue and iridescent on the surface. Bisulphide of copper. Formula, CVS. Copper has a strong affinity for sulphur. ' When a mixture of sulphur in powder and copper-turn- ings is exposed to a red heat, combination takes place with incan- descence, and disulphide of copper is formed. It may easily be made on a large scale by dropping pieces of sulphur upon copper heated to bright redness in crucibles, or by heating a mixture of copper scales and sulphur. When thus artificially prepared it is compact, and breaks with a granular, columnar, or more or less conchoidal fracture ; it is black, with a bluish-grey tinge, and has a feebly -metallic lustre ; it melts at a lower temperature than copper, and when melted does not permeate the crucible like galena ; it may easily be reduced to powder by trituration. According to Karsten its specific gravity is 5'9775. It undergoes no change when strongly heated without access of air. Disulphide of copper heated with other sulphides. According to Berthier it has a strong tendency to combine with all metallic sulphides. Double sulphides of copper and the alkaline metals are easily obtained by heating mixtures of sulphate of copper and alkaline sulphate in brasqued crucibles. Berthier prepared in this way the two following double sulphides : 6 1. Disulphide of copper 55 parts. Sulphide of barium 45 do. It was compact, fragile, lamellar, of a bright lead-grey colour, and resembled galena. 2. Disulphide of copper 67 parts. Sulphide of calcium 33 do, It was bubbly ; its fracture was granular and crystalline ; it had a bluish, metallic, grey colour, and somewhat resembled sulphide of antimony. The following experiments on the combination of disulphide of copper with sulphide of iron have been made in my laboratory. Nos. 3 and 5 by W. Baker, and No. 4 by E. Smith. 3. Cu 2 S = 1000 grains. 2FeS = 1109 do. The mixture was exposed in a covered crucible to a strong red heat, and pieces of sulphur were dropped into the fused mass. When cold the product weighed 2176 grains, showing an increase of 67 grains. It consisted of two distinct layers : an upper one, which had a yellow colour and metallic lustre, resembling native iron pyrites ; and a lower Tr. dcs Essais, 2, p. 40(5. HEATED WITH ACCESS OF AIR. 247 one, resembling protosulphide of iron. The whole was remelted, when sulphur was again added, and mixed with the melted mass by stirring with a stick. The product, which was brittle, was broken when cold. It appeared quite homogeneous ; its colour was between brass and bronze-yellow; it contained 29-6 per cent, of sulphur. The formula Cu 8 S + 2FeS corresponds to 28-69 per cent, of sulphur. 4. Cu 2 S = 320 grains. FeS = 176 do. The mixture was melted in a covered crucible. The product weighed 450 grains, the loss being 46 grains ; it had a granular fracture, and a dark bluish grey colour ; copper in minute particles was diffused through the mass, and moss or filamentous copper occurred in cavities in the interior ; it resembled the kind of regulus called blue-metal by the copper-smelters. 5. Cu 2 S = 1000 grains. 2FeS 2 (native) = 1513 do. The mixture was exposed in a crucible, having a luted cover, to a strong red heat during twenty minutes. The product, which resem- bled that obtained in experiment No. 3, weighed 2160 grains, the loss being 353 grains ; it contained 3O39 per cent, of sulphur. If the materials had been perfectly pure, and the bisulphide of iron had been completely reduced to protosulphide, the loss would have amounted to 403 grains. Disulphide of copper heated with access of air. When disulphide of copper in the state of fine powder is gradually heated with free access of air to incipient redness, and stirred continually, both elements are oxidized. The sulphur is partly converted into sulphurous acid, which escapes, and partly into sulphuric acid, which remains in combination with protoxide of copper, forming sulphate of that oxide. If the roasting be continued until sulphurous acid ceases to be evolved, and at a temperature insufficient to decompose sulphate of copper, the product will consist of a mixture of that salt and protoxide of copper. If, on the other hand, the temperature be raised to a pretty strong red heat, the sulphate of copper will be decomposed, and the product will consist entirely of protoxide of copper, the sulphuric acid of the sulphate being partly volatilized, and partly resolved into sulphurous acid and ox}^gen. Theory of the process of heating disulphide of copper with free access of air, or wasting. We are indebted to Plattner for valuable and interesting researches on the chemical changes which occur in various roasting processes. He states that he has obtained by experiment the follow- ing results concerning the roasting of disulphide of copper. 1. A mixture of sulphurous acid and atmospheric air in the ratio of two volumes to five, dry as well as moist, was passed through a glass tube heated to moderate redness. No sulphuric acid was formed. 2. The experiment was repeated with a spiral coil of fine platinum wire placed in the tube. Sulphuric acid was formed, whether the gaseous mixture were dry or wet. The same result was obtained 248 PLATTNER'S EXPERIMENTS ON ROASTING. when the platinum was replaced by gold or silver in a finely divided state, the metals themselves undergoing no change. 3. Metallic copper. Experimental) was repeated with finely divided copper, precipitated by iron, from the solution of a salt of copper. At incipient redness dioxide of copper and sulphate of protoxide were formed. The experiment was not continued till the copper had become completely oxidized. During the process only a slight odour of sulphurous acid could be perceived at the open end of the tube. 4. Dioxide of copper. Experiment (1) was repeated with native red oxide. At incipient redness it was very soon changed into protoxide, and this into sulphate of protoxide. During the process only very little free sulphurous acid passed over. 5. Protoxide of copper. Experiment (1) was repeated with this oxide. At incipient redness it was changed into sulphate of protoxide, and neither sulphurous nor sulphuric acid was perceived at the open end of the tube. 6. Dry sulphurous acid gas was passed over protoxide of copper in a glass tube heated to incipient redness, atmospheric air being com- pletely excluded. When cold the matter in the tube was dirty red. It contained a sensible quantity of sulphate of protoxide of copper, which was dissolved out by water. During the passage of the dry sulphurous acid a sublimate of sulphur appeared near the oxide ; and at the open end of the tube neither sulphurous nor sulphuric acid was observed. Plattner obtained this deposit of sulphur by exposing oxide of zinc, protoxide of lead, and sesquioxide of iron to precisely the same conditions. Hence he draws the conclusion that, when sul- phurous acid gas is passed over easily reducible metallic oxides at a feeble red heat atmospheric air being excluded it is not only con- verted into sulphuric acid at the expense of the oxygen of the oxides with which it comes in contact, but at the same time it may also by simple contact with red-hot solid bodies be resolved into sulphuric acid and sulphur. 7. Dry sulphurous acid gas, without admixture of atmospheric air, was passed over finely divided copper (prepared by precipitation with iron), heated in a glass tube to incipient redness. The copper glowed more brightly, but without causing any manifest decomposition of the sulphurous acid. After cooling, the copper throughout appeared to retain its characteristic colour ; nevertheless, on boiling with distilled water, a small quantity of sulphate of protoxide of copper was dissolved out. 8. Experiment (7) was repeated at a red heat with silica in the state of sand and quite free from iron. At first sulphuric acid was formed, but afterwards sulphurous acid cjiiefly escaped ; sulphur was deposited in the tube. This result is remarkable and important, as showing that, at a red heat, by the mere contact of inert bodies in a state of fine division, dry sulphurous acid gas is resolved into sulphuric acid and sulphur. The following conclusions are drawn from the preceding data : 1. According to Plattner, when finely divided disulphide of copper is BISULPHIDE HEATED WITH OXIDES OF COPPER, ETC. 249 exposed to the action of the air at nearly a red heat, with frequent stirring, so as to change the position of the particles, sulphurous acid and dioxide of copper are at first produced. 2. The sulphurous acid by contact-action (Exp. 8) is partially converted into sulphuric acid at the expense of the oxygen of the current of atmospheric air ; and this sulphuric acid, together with the oxygen of another portion of atmospheric air, immediately exerts an oxidizing action upon the dioxide of copper, and probably also upon any unchanged disulphide present ; part of the acid combining with the resulting protoxide of copper to form sulphate of copper, and part being decomposed with the evolution of sulphurous acid. The sulphurous acid thus set free may either be driven off by freshly-formed sulphurous acid, or it may again be converted by contact-action into sulphuric acid in the manner described, at the expense of the oxygen of the current of air, which is continually flowing over the matter in process of roasting. 3. So long as sulphurous acid is formed in sensible quantity the whole of the dioxide of copper cannot be converted into protoxide (Experiment 6). Hence, after the oxidation of the disulphide of copper, the product contains from 20 to 30 per cent, of dioxide of copper, mixed with protoxide and sulphate of protoxide. 4. At a higher temperature the sulphate of protoxide of copper will be decomposed, the sulphuric acid, as has been previously stated, being partial^ volatilized and partially resolved into sulphurous acid and oxygen : this oxygen, Plattner remarks, may, together with that of the atmospheric air present, con- vert the remaining dioxide of copper into protoxide. But, it may be asked, will not the sulphuric acid volatilized be reduced by dioxide of copper at this increased temperature with the formation of sulphurous acid and protoxide of copper ? If the vapour of water be present, as must always be the case in ordinary roasting operations, it will com- bine with the anhydrous sulphuric acid which may escape, and produce a white vapour. Disulphide of copper heated in admixture with dioxide, protoxide, or sulphate of copper. It may be stated as a general fact, that when disulphide of copper is intimately mixed with one or more of these oxidized com- pounds of copper, in such proportion that the sulphur and oxygen exist in the ratio in which they are combined in sulphurous acid, and the mixture is heated to the melting-point of copper, the whole of the copper will be reduced to the metallic state, and the whole of the sulphur will be evolved as sulphurous acid. Hence it is easy to conceive how, by the action of air and heat alone, disulphide of copper may be completely reduced. It must first be roasted until the product contains sulphur and oxygen in the ratio above-mentioned, and then the heat must be raised to the melting- point of copper. When the oxygen exceeds this ratio, a propor- tionate amount of dioxide or protoxide of copper, as the case may be, will remain in the product. Conversely, when the sulphur exceeds this ratio, a proportionate amount of disulphide of copper will remain in the product. The following formulae express the 250 COPPER HEATED WITH PROTOXIDE OF LEAD. results, which will occur under varying relations between the sulphur and oxygen : 1. Cu 2 S + 2Cu 2 O = Cu 6 + SO 2 2. Cu 2 S + 2CuO = Cu 4 -f SO 2 3. Cu 2 S + 3CuO = Cu 3 + Cu 2 O -f SO 2 4. Cu 2 S + CCuO = 4Cu 2 O + SO 2 5. Cu-S + CuO, SO 3 = Cu 3 + 2SO 2 6. Cu-'S + 2CuO,SO 3 = 2Cu 2 + 3SO 2 7. Cu 2 S + 4CuO, SO 3 = 6CuO + 5SO 2 Most of these reactions have been confirmed by experiment in the metallurgical laboratory by my former pupil, Mr. W. Baker, of Sheffield. It must not, however, be supposed that in experiments of this nature theoretical accuracy can be attained, especially on account of the corrosive action exerted at a high temperature by oxides of copper on the substance of the crucibles employed, and the consequent formation of compounds consisting of silica, alumina, and oxide of copper. Nevertheless, with proper care and experience, results may be obtained which sufficiently approximate to those indi- cated by theory to demonstrate the practical correctness of the formulae. It is especially desirable to conduct such experiments at the lowest temperature at which the reactions take place, in order to lessen, as far as practicable, the error arising from the corrosion of the crucibles. Care must also be taken to prevent reduction of the oxidized com- pounds of copper by the gases of the furnace in which the crucibles are heated. By way of illustration, the following results of actual experiments are presented : 1. Cu 2 S = 1000 grains. 2Cu a O = 1796 do. The button of copper weighed 2301 grains. The theoretical quantity is 2391 grains : difference, 90 grains. 2. Cu 2 S = 1000 grains. 2CuO = 1000 do. The button of copper weighed 1295 grains. The theoretical quantity is 1596 grains: difference, 300 grains. A thin layer of regulus, like disulphide of copper, adhered to the button. 3. Cu 2 S = 500 grains. CuO, SO 3 = 500 do. The button of copper weighed 516 grains. The theoretical quantity is 596 grains : difference, 80 grains. Copper Jieated with protoxide of lead.- The following experiments were made in the metallurgical laboratory by Mr. R. Smith : 1. Cu 4 (granulated) = 320 grains. PbO (commercial litharge) = 280 do. The product was melted, and consisted of slag, 7 and a button of ' In the description of all these expe- I denote that part of the product which riments I have used the word sine to was neither rfimdii* nm- mP tni COPPER HEATED WITH PROTOXIDE OF LEAD. 251 metal weighing 340 grains. The slag was compact, opaque, and red- brown. The button was copper-coloured on the external surface and brittle; its fractured surface was fibrous and greyish red; it contained 308-4 grains, or 90-7 per cent., of copper. 2. Cu 2 (thin turnings) = 320 grains. PbO (commercial litharge) = 560 do. The litharge was placed upon the copper, and the mixture was gradually heated to bright redness. The product was melted, and consisted of slag, and a button of metal weighing 340 grains. The slag on its upper surface was dark grey, and presented acicular crys- tals confusedly grouped ; on fracture it was vitreous, opaque, and red- brown. The button was copper-coloured externally, and cracked under the hammer; its fractured surface was dull lead-grey; it contained 288-8 grains, or 83*18 per cent., of copper. 3. Cu = 320 grains. PbO = 1120 do. The product was melted, and consisted of slag, and a button of metal weighing 321 grains. The slag on its upper surface was dark grey, somewhat metallic in lustre, and presented groups of interlacing acicular crystals ; on fracture its colour was red-brown, approaching black, near the upper surface. The button was copper-coloured ex- ternally ; its fractured surface was granular and dull lead-grey ; it contained 233-7 grains, or 72-8 per cent., of copper. , 4. Cu = 320 grains. 2PbO = 2240 do. The product was melted, and consisted of slag, and a button of metal weighing 392 grains. The slag was similar in appearance to that of No. 3, except that it was somewhat mottled. The button was copper- coloured externally, and more malleable than that of No. 3 ; its frac- tured surface was finely fibrous, or silky, and lead-grey. The experi- ment was repeated with granulated copper, and with similar results, except that the button weighed 373 grains ; it contained 203-9 grains, or 54- 66 per cent., of copper. 5. Cu = 160 grains. 3PbO = 1080 do. The product was melted, and consisted of slag and a button of metal weighing 195 grains. The slag on its upper surface was dark grey and somewhat metallic in lustre ; the lower part was vitreous and dark brown. The upper part, or about three-fourths of the whole, was granular, and in places presented crystalline fibres ; without lustre ; varying in colour from red-brown to greenish yellow-brown, and even black ; where in contact with the button it was red ; it was translu- cent in thin slices. The button was composed of two layers ; an upper one, hard, fibrous on fracture, and, from its appearance, evidently rich in copper, and a lower one, soft, and resembling lead; it contained 56-9 grains, or 29-2 per cent., of copper. 6. Cu = '64 grains. - 1344 do. 252 COPPER HEATED WITH SULPHATE OP LEAD. The product was melted, and consisted of slag and a button of metal weighing 86 grains. The slag was vitreous, semi- opaque, and dark brown. The button resembled lead, and presented no appearance of separation into two layers; it contained 1O6 grains, or 12-3 per cent., of copper. These results agree pretty well with those of Berthier, 8 and show that when copper is heated with litharge it is only oxidized to the degree of dioxide ; the same fact is also proved by heating a mixture of dioxide of copper and litharge (vide p. 253). In the six experi- ments preceding, the differences between the weights of the buttons actually found and those which theory would indicate are respectively as follow (the atomic weight of Cu being taken as 32 and that of lead as 104) : 12-75, 0-5, 52'9, 0-5, 29-4, and 0-77 grains. These numbers, with the exception of the third, are not greater than may be expected in experiments of this kind. From the last experiment it appears that when copper is heated, with 21 times its own weight of litharge not less than one -sixth of it remains unoxidized. Copper heated with sulphate of lead. Experiments by K. Smith. 1. Cu 2 = 320 grains. PbO,S0 3 = 760 do. The reaction is very energetic, but it does not occur below a strong red heat. Sulphurous acid is evolved. The product was a compact, opaque, red-brown slag, black and scoriaceous on the upper surface and orange-coloured at the lower part. Berthier describes the product which he obtained from this mixture as compact, with a shining frac- ture, opaque, and of a very fine sealing-wax-red colour. 2. Cu = 160 grains. PbO,S0 3 = 760 do. The result was similar to that of the last experiment. The slag was brownish-red, -and darker near the upper surface, which was coated with a black film of a somewhat metallic lustre. There were no globules of copper, so that the copper was completely oxidized. Copper heated with sesquioxide of iron. Experiments by K. Smith. Finely pounded haematite and copper in thin shavings were employed. 1. Cu 2 = 64 grains. Fe 2 O 3 = 80 do. The mixture was made as intimate as possible. It was subjected, in a small covered clay crucible, to a high temperature in a muffle during an hour. The contents of the crucible adhered together in a black mass. The copper was oxidized on the surface and very brittle. 2. Cu = 32 grains. Fe 2 O 3 = 80 do. 3. The result was similar to that of the last experiment. Copper heated with peroxide of manganese. Experiments by E. Smith. 1. Cu = 64 grains. 2. Cu 2 - 64 grains. MnO 2 = 200 do. MnO 2 = 50 do. Tr. des Essais, 1, p. 385. PROTOXIDE OF COPPER HEATED WITH METALLIC LEAD. 253 The results were similar to those obtained with sesquioxide of iron. According to Berthier the peroxide of manganese is reduced to prot- oxide, and, by the addition of a little glass, a very fusible slag is funned, which contains dioxide of copper and protoxide of manganese. Protoxide of copper heated with metallic lead. Experiments by E. Smith. 1, 2CuO = 800 grains. Pb (granulated) = 1040 do. The mixture melted readily into an opaque, crystalline, black slag, having a semi-metallic lustre. The product should have the formula Cu'O + PbO. 2. 3CuO = 480 grains. Pb 2 = 832 do. The product was melted, and consisted of slag, and a button of metal weighing 460 grains. The slag, which was only partly vitreous, was opuquc and reddish-brown. The button was copper-coloured exter- nally, and its fractured surface was dull lead-grey ; it contained 330-7 grains, or 71 '9 per cent., of copper. 3. CuO = 400 grains. Pb = 1040 do. The product was melted, and consisted of a red-brown slag, and a button of metal weighing 343 grains. The experiment was repeated with half the preceding quantities. The slag was vitreous, and redder in colour than in the first experiment. The button resembled lead, and weighed 180 grains ; its fracture was fibrous; it contained 119'7 grains, or 66-5 per cent., of copper. Dioxide of copper heated with protoxide of lead. Experiments by E. Smith. 1. Cu 2 = 720 grains. PbO = 1120 do. The mixture melted at a low red heat ; it attacked and traversed the substance of the crucible with great rapidity. The product is crys- talline and reddish brown-black. 2. Cu 2 O = 720 grains. 2PbO = 2240 do. The mixture melted as in the last experiment, and the product had nearly the same characters ; its upper surface was coated with a black film having a semi-metallic lustre. It follows from these results that dioxide of copper is not in any degree oxidized when heated with protoxide of lead ; for, otherwise, metallic lead would have been separated. Protoxide of copper lieated with protoxide of lead. Experiments by E, Smith. 1. CuO = 400 grains. PbO =1120 do. The mixture melted into a compact, hard, dull slag ; its upper surface was black, crystalline, and metallic in lustre ; the colour of its fractured surface varied from brown to black from below upwards. 2. CuO = 200 grains. 2PbO =1120 do. 254 PROTOXIDE OF COPPER HEATED WITH SULPHIDE OF LEAD. The mixture melted into a crystalline, shining, dark-green slag, much softer than the last, and more resembling fused protoxide of lead in appearance; its upper surface was smooth, black, and semi-metallic in lustre. Protoxide of copper heated with sulphide of lead. Experiments by E. Smith. The purest galena was employed. 1. CuO = 400 grains. PbS = 1200 do. The mixture melted with considerable effervescence. The product consisted of a vitreous, opaque, black slag, and a regulus weighing 705 grains ; its fractured surface was dark grey and crystalline, resem- bling fused galena in appearance. The experiment was repeated with half the preceding quantities ; the regulus weighed 302 grains, and contained 142-3 grains, or 47-13 per cent., of copper. Berthier describes the slag obtained from this mixture as of a fine red colour. 2. 30uO = 600 grains. 2PbS = 1200 do. The mixture melted with much effervescence. The product consisted of a vitreous, opaque, brownish-red slag, and a purplish lead-grey regulus, weighing 710 grains; a small button of soft lead was attached to the lower part; it contained 504-4 grains, or 71-04 per cent., of copper. 3. 2CuO = 800 grains. PbS = 1200 do. The mixture melted with considerable effervescence. The product consisted of a compact, hard, brittle, opaque, sealing-wax-red coloured slag, and a purplish grey regulus, to the lower part of which a copper- coloured button of metal adhered with great tenacity ; the fractured surface of this button was close-grained, dull, and reddish grey; it cracked under the hammer. The regulus and button weighed together 565 grains ; the button weighed 307 grains, and contained 299 grains, or 97-4 per cent., of copper. 4. 30 uO = 1000 grains. PbS = 1000 do. The mixture melted, but not with quite so much effervescence as those preceding. The product consisted of a brownish red slag, darker on the upper surface, and a copper-like button of metal weighing 140 grains ; it contained 138-6 grains, or 99 per cent., of copper. 5. 4CuO = 800 grains. PbS = 600 do. The mixture melted with slight effervescence. The product consisted of a slag like that of the last experiment, and a copper-like button of metal weighing 162 grains; its fracture was fibrous; it contained 156-7 grains, or 96 '1 per cent., of copper. Dioxide of copper heated with protosulphide of iron and silica. Experiment by E. Smith. 3Cu 2 O = 1296 grains. 3FeS = 786 do. SiO ! = . 276 do. BISULPHIDE OF COPPER HEATED IN CONTACT WITH STEAM. 255 These substances were intimately mixed, and the mixture was ex- posed to a high temperature in a close-grained crucible. The product consisted of a vitreous, opaque, black slag, and a regulus which resem- bled disulphide of copper in appearance, and weighed 1330 grains. The proportions of the mixture are such, that, supposing the copper to be wholly converted into disulphide, tribasic silicate of protoxide of iron would be formed. Admitting this reaction to have occurred, the regulus should have weighed 1440 grains ; and as the difference between this number and that actually obtained is only 110 grains a difference by no means great, considering the nature of the experiment it may be concluded that the reaction supposed does take place, at least in a very great degree. This reaction is of much importance in copper-smelting, and may be expressed by the following formula : 3Cu 2 O + 3FeS -f SiO 3 = 3FeO, SiO 3 + 3Cu 2 S. The same experiment with the same proportions was repeated in a plumbago crucible. A similar slag was produced, and a regulus like disulphide of copper, which contained " moss copper "'in cavities near the surface. At the bottom of the crucible there was also some metallic copper, through which disulphide of copper wag diffused. Disulphide of copper exposed to the action of hydrogen at high temperatures. According to Berthier 9 and H. Eose 1 disulphide of copper is not changed when heated to redness in a stream of hydrogen. The fol- lowing experiments were made by A. Diclpm my laboratory. Perfectly dry hydrogen was passed over disuj^iide of copper in a tube of German glass heated until the glas&a^oftened. Sulphuretted hydrogen was evolved in small quantity ; ,md the residue, where it had been exposed to the greatest heat, Jgal a coppery aspect. Disulphide of copper exposyj&o the vapour of water at a high temperature. Disulphide of copper ytas prepared by heating together electrotype copper and sulphur. The product was reduced to powder and again heated with nujpliur in order that no metallic copper might remain. A porcelain>tube, coated externally with a mixture of fire-clay and asbestos, was fbsfed in a furnace capable of producing a white heat. The disulphide in powder was placed in a little porcelain boat, to the bottom of which asbestos was attached by platinum wire so as to prevent its sticking to the tube. Steam was passed through the tube, and the boat gradually pushect into the hottest part. At first sulphuretted hydrogen was evolved, as detected by its smell, and free sulphur was deposited in the condensed steam. The odour of sulphurous acid in admixture with sulphuretted hydrogen was also clearly detected by Dick and myself, notwithstanding these gases mutually decompose each other, with the formation of water and the separation of sulphur. After a short time these gases ceased to escape with the steam, or, if they did, it was only in very small quantity. During three-quarters of an hour the tube was heated nearly to whiteness, and a good current of steam was kept up all the time. The boat was then gradually withdrawn Tr. dcs Essais, 2, p. 402. ' Jahrcs-Bericht. Berzelius, 1827, p. 110. 256 METALLIC COPPER HEATED IN CONTACT WITH STEAM. from the tube by means of a wire, and, when the temperature was reduced to about 100 C., it was taken out of the steam. The product consisted of several globules of melted disulphide, the surfaces of which were covered with little excrescences of metallic copper. In a second experiment a considerable quantity of globules of copper were obtained. The experiment was repeated a third time with the same results. In 1837 Eegnault experimented on this subject. He found that, at a red heat, the vapour of water only exerts a very feeble action upon disulphide of copper, metallic copper and sulphuretted hydrogen being separated in small quantity. At a strong white heat, however, the disulphide is energetically decomposed, hydrogen being copiously evolved along with sulphuretted hydrogen, and drops of sulphur being condensed in the tube. After seven hours the disengagement of gas ceased, when it was ascertained that the disulphide of copper had been completely reduced to the metallic state. The button of copper was very brilliant. Eegnault explains the production of free hydrogen by the decomposition which sulphuretted hydrogen undergoes at a high temperature. He found that when this gas is exposed per se to a white heat, it is only very partially decomposed; and, as the presence of the vapour of water cannot favour the decomposition, he supposes that, in the foregoing experiment, the copious evolution of hydrogen may be due to the fact that sulphuretted hydrogen in the nascent state is much more easily decomposed than when it has acquired the gase- ous state. 2 In reference to these results Gmelin inquires what be- comes of the water ? 3 This question may be readily answered. If the hydrogen of the water combine with sulphur, its oxygen must combine with copper to form oxide of copper ; but the oxide of copper thus formed would, at the high temperature of the experiment, imme- diately act upon the unchanged disulphide with which it may be in contact, forming sulphurous acid and metallic copper. Part of the sulphurous acid may escape from the tube, as was the case in Dick's experiment, while the remainder may be decomposed by the sul- phuretted hydrogen, with the formation of water and free sulphur. Metallic copper exposed to the action of the vapour of water at high tempera- tures. According to Eegnault, when the vapour of water is passed over metallic copper in a porcelain tube heated nearly to whiteness, hydrogen gas is evolved in very sensible quantity. On continuing the operation during three or four hours, from 80 to 90 cubic centi- metres of inflammable gas were collected. After the conclusion of the experiment the tube was broken, when the anterior part was found to be coated with a very thin layer of black matter which dissolved in acids and presented the characters of oxide of copper. Eegnault sup- poses that this layer of oxide may have been formed by the volatiliza- tion of a little copper and its oxidation in that very finely divided state by the vapour of water. The copper in the tube was melted, and had a fine metallic lustre ; at some points of its surface it presented an 2 Ann. des Mines, 3. s. 11, p. 44. 3 Handbook of Chem. Cavend. Soc. 5, p. 421. BISULPHIDE OF COPPEE HEATED WITH IRON. 257 excessively thin and deeper coloured film of oxide. This experiment proves that metallic copper is only capable of decomposing the vapour of water at a very high temperature, and then only so feebly that it would be impossible to produce complete oxidation. It might be objected, as Kegnault remarks, that copper only decomposed the vapour of water under the influence of the silica of the tube. In order to ascertain the force of this objection he exposed a mixture of copper turnings and finely divided silica (as obtained by the action of water on hydrofluosilicic acid) to the action of the vapour of water in a por- celain tube heated to whiteness. The operation was continued during four hours, but hydrogen was not disengaged in greater quantity than in the first experiment. The metal was only melted in drops amongst the silica in certain places ; in others it had preserved its form ; its surface was not more brilliant than before the experiment ; its colour had become very rosy, like that of copper containing oxygen. No trace of silicate of either of the oxides of copper could be detected. 4 Disidphide of copper heated with carbon. According to Berthier carbon slowly reduces disulphide, but only at a very high temperature. 5 The following experiment was made by R. Smith : 200 grains of disulphide were exposed to a high temperature in a brasqued crucible ; the pro- duct consisted of 189*5 grains of unchanged disulphide and 8'5 of metallic copper. Sulphur must be evolved in combination with carbon. Disulphide of copper heated with iron. The experiments were made by W. Baker. 1. Cu-S = 1000 grains. Fe (in filings) = 353 do. The mixture was exposed during twenty minutes to a strong red heat in a crucible with a luted cover. The product was melted, and consisted of regulus, and a button of metal weighing 538 grains. The metal was analysed by Mr. Tween, and found to have the following composition : Copper 62-45 Iron 31-70 * Sulphur 3*85 98-00 2. Cu 2 S = 1000 grains. Fe 2 = 700 do. The product was melted, and consisted of regulus and metal ; it was brittle and slightly crystalline on the lower and external parts ; its fractured surface was uneven, and had a dark iron-grey colour with a reddish tinge. 3. ( 'u 2 S = 1000 grains. Fe 2 = 910 do. The mixture was exposed in a crucible with a luted cover during forty -five minutes to the high temperature of a furnace heated with Aim. des Mines, 3. s. 11, p. 26. 5 Tr. des Essnis, 2, p. 402. S 258 BISULPHIDE OF COPPER HEATED WITH ZINC. anthracite. There was no separation into regulus and metal. The product was brittle and purple coloured ; its fractured surface was uneven, crystalline, especially the lower portion, and reddish dark grey. 4. Cu 2 S = 1000 grains. Fe 4 = 1410 do. The product was melted and dark reddish grey. There was no distinct separation into regulus and metal. On its upper surface was a thin layer of compact, non-crystalline, dark grey regulus, below which the mass was crystalline. It contained moss-copper, especially in the centre, and at the depth of about one-third from the top, where there appeared to have been a cavity. Mixed with the moss-copper I observed distinct dark grey crystals, resembling those of disulphide of copper, which are sometimes produced during the process of copper- smelting. From the preceding experiments it appears that disulphide of copper is only partially reduced to the metallic state when heated with iron. A double sulphide of copper and iron is formed, upon which iron exerts no reducing effect. Disulphide of copper heated with zinc. The experiments were made by E. Smith. 1. Cu 2 S = 320 grains. Zn (in fine p.owder) = 128 do. An intimate mixture was made and exposed to a bright red heat during twenty minutes in a Cornish crucible with a luted cover. The product weighed 246 grains, the loss in weight being 202 grains ; it consisted of a thin layer of a dark bluish-grey regulus, and a button of metal of the colour of brass ; they adhered together tenaciously. The regulus contained 58*4 per cent, of copper, and the button 81'8 per cent. 2. Cu 2 S = 160 grains. Zu 2 = 128 do. The product weighed 182 grains, the loss in weight being 106 grains. It consisted of a layer of regulus weighing 119 grains, and a button of metal weighing 63 grains. The regulus was compact, brittle, finely granular on fracture, and crystalline on the exterior ; its upper surface was partially covered with moss-copper ; it contained 62-37 per cent, of copper. The button was well melted, of the colour of brass, and contained 80-9 per cent, of copper. The loss of copper amounts to 2-82 grains. About two-sevenths of the copper of the disulphide were reduced to the metallic state. Disulphide of copper heated with lead. The experiments were made by R. Smith. The lead was prepared by constantly shaking it during solidification after fusion, and finely sifting the powder thus obtained. The materials were intimately mixed and heated in Cornish crucibles, covered, but not luted. 1. Cu 2 S = 400 grains Pb (in fine powder) = 520 do. The product was melted, and consisted of regnlus, and a button of BISULPHIDE OF COPPER HEATED WITH TIN. 259 metal weighing 495 grains. They adhered tenaciously to each other. The regulus was compact, hard, purplish-grey, and semi-vitreous in lustre. The metal resembled lead, was malleable, but could be easily broken in two ; it contained 10'3 per cent, of copper. 2. 2Cu 2 S = 400 grains. Pb 3 = 780 do. The description of the product of the last experiment applies equally to that obtained in this. The button of metal weighed 753 grains, and contained 8-84 per cent, of copper. 3. Cu 2 S = 400 grains. Pb 2 = 1040 do. The product was similar in characters to that of the first experiment. The button of metal weighed 1019 grains, and contained 6*76 percent, of copper. In the three preceding experiments, the total amounts of copper reduced to the metallic state were respectively as follow : 50-99, 66-57, and 68*88 grains ; the amount of copper in the disulphide employed being the same in each experiment, namely, 320 grains. Hence it may be concluded that when disulphide of copper is heated, either with a small or a large quantity of lead, between one-fifth and one-sixth of the copper is reduced to the metallic state. Disulphide of copper heated with tin. Experiment by Mr. W. Baker in the metallurgical laboratory. Cu 2 S = 1000 grains. Sn (in powder) = 741 7 do. An intimate mixture was made and heated to bright redness during ten minutes in a Cornish crucible, enclosed in another crucible with a luted cover. The inner crucible with its contents was weighed before and after the experiment: the loss of weight amounted only to 18 grains. The product consisted of regulus, and a button of metal, which were easily separated from each other. The regulus which contained tin was compact, crystalline, and more grey than disulphide of copper. The metal was crystalline, with a largely lamellar fracture, very brittle, and white. It had the following composition : Tin 65-17 Copper 33-25 Sulphur 0-37 98-79 It may, thereefor, be regarded as a definite alloy of the formula CuSn. The reaction which occurs may be expressed by the following equation : 2Cu 2 S + Sn 2 = CuSn + (Cu 2 S + Cu, Sii, S.) Thus in two equivalents of disulphide of copper, one equivalent -of copper is replaced by one of tin ; and the equivalent of copper so replaced forms a definite alloy with one equivalent of tin. The regulus s 2 260 BISULPHIDE OF COPPER HEATED WITH ANTIMONY. raay.be considered as resulting from the substitution of one equivalent of tin for one of copper in two equivalents of the disulphide. sulphide of copper Iieated with antimony. The experiments were made by E. Smith. 1. Cu 2 S = 160 grains. Sb = 258 do. 6 An intimate mixture was made. The product weighed 409 grains, the loss being 9 grains ; it consisted of regulus and a button of metal, which adhered tenaciously together. The regulus amounted to about one -fourth of the bulk of the whole mass ; it was compact, fine-grained, and purplish grey ; it contained 5 7 -48 per cent, of copper. The button of metal resembled antimony, and had a largely lamellar fracture ; it contained 19 -2 per cent, of copper. " 2. 3Cu2S = 240 grains. Sb = 129 do. The product weighed 365 grains, the loss being 4 grains ; it consisted of regulus and metal similar in appearance to those of the last experi- ment. The regulus and metal adhered so firmly together that it was impossible to separate them so as to obtain the weight of each with any degree of accuracy. The regulus contained 57 '9 per cent, of copper. The metal had the following composition : Copper 33-40 Antimony (by loss) 60'56 Sulphur 6-04 100-00 Copper heated with tersulphide of antimony. The experiments were made in the metallurgical laboratory by my former pupil, Mr. Ambrose Tween. The old atomic weight of antimony, 129, was taken. 1. Cu 3 = 190 grains. SbS 3 * = 354 do. The copper was employed in the state of wire cut into small pieces. An intimate mixture was made, and exposed during twenty minutes to a bright red heat, in a Cornish crucible with a luted cover. The product weighed 517 grains, the loss being 27 grains; it con- sisted of regulus and a button of metal, which adhered very tenaciously together. The regulus and metal were composed as follows : Regulus. Metal. Copper 47-53 3-86 Antimony (by loss) 29 '32 95-97 Sulphur 23-15 0-17 100-00 100-00 By adopting the recent atomic weight of 120-3, as established by The old equivalent of 129 was taken. BISULPHIDE OF COPPER HEATED WITH PROTOXIDE OF LEAD. 261 Schneider and H. Rose, the composition of the regultis may be nearly expressed by the formula 3Cu 2 S + SbS 3 2. Cu 6 = 380 grains. SbS 3 = 354 do. The product weighed 689 grains, the loss being 45 grains ; it con- sisted of regulus and metal, which adhered firmly together. The regulus and button of metal were composed as follows : Regulus. Metal. Copper 66-44 42-54 Antimony (by loss) 16-91 57'06 Sulphur 16-65 0-40 100-00 100-00 3. Cu 12 = 761 grains. SbS 3 = 354 do. The product weighed 1087 grains, the loss being 28 grains; it con- sisted of regulus and metal, which could not be separated, and were composed as follows : Regulus. Metal. Copper 75-9 66-72 Antimony (by loss)... 22-8 32-98 Sulphur 1-3 0-30 100-0 100-00 4. Cu 18 = 1141 grains. SbS 3 = 354 do. The product weighed 1474 grains, the loss being 21 grains ; it con- sisted of regulus and metal, which adhered firmly together, and were composed as follows : Regulus. Metal. Copper 77-36 75-90 Antimony (by loss) 21-31 24-03 Sulphur 1-33 0-07 100-00 100-00 Bisulphide of copper heated with protoxide of lead. The experiments were made by W. Baker, and the quantitative determinations by A. Tween. The materials were intimately mixed, and heated in covered Cornish crucibles during about ten minutes, at a temperature just sufficient to effect perfect fusion. 1. Disulphide of copper = 400 grains. Litharge = 2000 do. The product consisted of slag, regulus, and a button of metal ; the button weighed 410 grains, and resembled lead; it contained 22-55 grains of copper, or 5-5 per cent* The slag appeared to be composed of protoxide of lead and dioxide of copper. 2. Disulphide of copper = 100 grains. Litharge = 2000 do. 262 BISULPHIDE OF COPPER HEATED WITH NITRE. The product consisted of a crystalline, opaque, reddish-brown slag, and a button of metal weighing 394 grains ; it resembled lead, and contained 22-85 grains of copper, or 5*8 per cent. 3. Bisulphide of copper = 100 grains. Litharge = 2500 do. The product consisted of a slag like that of the last experiment, and a button of metal weighing 422 grains : it resembled lead, and con- tained 21-67 grains of copper, or 5-16 per cent. Hence it appears that when disulphide of copper is heated with twenty times its weight of protoxide of lead, the whole of its sulphur is oxidized; for otherwise some regulus would have been obtained in the second experiment. The sulphur is converted into sulphurous acid, which, on escaping, causes the effervescence observed. About four parts by weight of protoxide of lead contain oxygen sufficient to convert the sulphur of one part of disulphide of copper into sul- phurous acid, and the copper into dioxide ; but, practically, a much larger quantity of protoxide of lead is required to produce this effect. The results of the preceding experiments tend to confirm the state- ment of Berthier, that, when litharge is combined with a certain pro- portion of dioxide of copper, it ceases to exert any action upon disulphide of copper, although each oxide separately has the power of decomposing this sulphide. Disulphide of copper heated with sulphate of lead. According to Berthier, sulphate of lead attacks simultaneously both elements of the disulphide; S3 that neither copper, nor an alloy of copper and lead, ever results from the reaction between these two substances. The product consists of a regulus, which appears to contain sulphide of lead, and a red slag composed of oxide of lead and dioxide of copper, sulphurous acid being evolved. In order to decompose the whole of the disulphide of copper, it would be necessary to mix it with at least seven times its weight of sulphate of lead ; and the product resulting from such a mixture would be a slag formed of oxide of lead and dioxide of copper. 7 Disulphide of copper heated with nitre. The action is energetic at a nascent red heat. When the quantity of nitre is sufficient to convert the whole of the sulphur into sulphuric acid, and access of air is pre- vented, the whole of the copper is reduced, and the slag consists entirely of sulphate of potash. In order to moderate the deflagration, it is necessary to add to the mixture a considerable quantity of alkaline carbonate (Berthier). The following equation should express the reaction : Cu 2 S + KO, NO 5 = Cu 2 + KO, SO 3 + NO 2 . When the nitre is in excess, the slag will contain dioxide of copper. Disulphide of copper heated with caustic soda. According to Berthier, the disulphide is partially decomposed, with the separation of metallic copper, and the formation of sulphate of soda, and a double sulphide of sodium and copper. The presence of charcoal much promotes the " Tr. (k-s Essais, 2, p. 406. BISULPHIDE OF COPPER HEATED WITH CARBONATE OF SODA. 263 desulphurization. From a mixture of 1 part of disulphide and 2 parts of caustic soda, Berthier obtained O32 of copper; and from a mixture of 1 part of disulphide, 1 of caustic soda, and 0-4 of charcoal, he obtained 0*54 of copper. The first amount is nearly the same as that we obtained by heating a mixture of disulphide, carbonate of soda, and charcoal. The following formula expresses the reaction : 4Cu 2 S + zWS + 4NaO = NaO, SO 3 + 3NaS, zCu 2 S . + Cu 8 . Disulphide of copper heated with carbonate of soda. The experiments were made by K. Smith. 1. A mixture of 200 grains of disulphide of copper, and a consider- able quantity of carbonate of soda, was heated in a covered Cornish crucible. No metallic copper was separated. The product consisted of a brownish black slag, and a button of regulus like disulphide of copper. The experiment was repeated with precisely the same result. 2. A mixture consisting of 200 grains of disulphide of copper, about 400 of dry carbonate of soda, and 50 of charcoal, was heated in a covered Cornish crucible. The product consisted of a somewhat crys- talline black slag, and a ^brittle metallic button resembling copper, and weighing 61 grains. The experiment was repeated with the same quantities, and a button of metal was obtained weighing 64 grains. Taking the mean of the weights of copper obtained in both experi- ments, the amount of copper reduced is somewhat more than 39 per cent, of the total copper. Berthier states that when a mixture of disulphide of copper and car- bonate of soda in the ratio of Cu' 2 S : 3 NaO, CO 2 is exposed to a high temperature in a brasqued crucible, the reduction of the copper is nearly complete. According to the same authority, elevation of tem- perature singularly favours the reaction. According to Berthier, disulphide of copper is partially reduced when heated with the pearlash of commerce ; because the latter always contains some caustic potash : with 6 parts of pearlash to 1 of disulphide, only 0-4 of copper is obtained, and no more is separated by increasing the proportion of flux. Disulphide of copper heated with laryta or lime. Baryta and caustic lime partially decompose disulphide of copper like the alkalies, at least in the presence of carbon ; but the copper separated remains diffused in the form of shots through the double sulphide of copper and barium, or calcium, which is formed and does not thoroughly melt even at a very high temperature, on account of admixture with a certain amount of baryta or lime (Berthier). Disulphide of copper heated with cyanide of potassium. The experiments were made by R. Smith. The mixtures were heated in covered Cornish crucibles. The disulphide is partially reduced. 264 COPPER AND DIOXIDE OF COPPER. 1. Grains. 2. Grains. 3. Grains. 4. Grains. 5. Grains. Disulpliidc of copper 100 400 20 46 100 800 40 43 100 1000 50 41 50 550 50 50 300 23 21 20 21 Do. do. on repeating the experiment Observations. 1. The copper was very fine and tough. 2 and 3. The slag was black. 4. The mixture was covered with common salt. 5. Plumbago crucibles were employed. Copper and dioxide of copper. Copper in the state of fusion has the property of dissolving dioxide of copper to a considerable extent. When it contains this oxide to the maximum, it is technically termed " dry copper 1 ' by English copper-smelters, or " ueler-gaar " by German smelters. Dry copper is distinguished by the following characters : it is brittle either when cold or hot, so that an ordinary ingot may be easily broken in two ; the fractured surface is uneven, minutely granular, and without any appearance of fibre ; it presents here and there the appearance of films of dioxide and globular cavities ; it is dull, and of a comparatively deep purplish red colour ; and, when cast in an open narrow mould of copper of the usual well-known form, the upper surface of the ingot is marked by a shallow longitudinal depression or furrow, extending nearly from end to end along the median line. Copper which contains much less than the maximum of dioxide is said to be more or less " dry ; " and the degree of dry ness is proportionate to the amount of this oxide present. An ingot of copper in the driest state, which, through the kindness of the late Mr. Vivian, I saw laded from the furnace at the Hafod Works, has been examined by Dick in my laboratory, with a view to determine the proportion of dioxide contained in it. 8 1. A portion of this ingot was rolled out as thin as possible, and the rolled metal was cut into small pieces, of which 132'34 grains were exposed to a current of dry hydrogen in a combustion tube heated to redness, and connected with a weighed tube containing chloride of calcium ; the gas was previously passed through the tube in order completely to expel atmospheric air. When the temperature rose to redness, the gas which escaped had a distinct odour of sulphuretted hydrogen, and instantly blackened paper impregnated with a salt of lead. This result is remarkable, as showing that copper saturated with dioxide may yet retain a very sensible amount of sulphur. During the course of the experiment a slight metallic sublimate appeared in the cooler part of the tube, which, however, was much 8 All the experiments by Mr. Dick on the metallurgy of copper, to be subse- quently detailed, were made in the me- tallurgical laboratory, expressly for this work: and I not only consented to but urged their publication in extenso in the Philosophical Magazine for June, 1856. COPPER AND DIOXIDE OF COPPER. 205 too hot for the condensation of arsenic : it was found to contain lead, but its quantity was too small to a limit of satisfactory examination. The chloride of calcium tube increased in weight 1*93 grain, which is equivalent to 10-21 per cent, of dioxide. In a second similar experiment upon another portion of the same dry copper, the increase was 1-82 grain, which is equivalent to 9-34 per cent, of dioxide. Sul- phuretted hydrogen was again detected at the beginning of the expe- riment. Precautions were taken to make it certain that the hydrogen employed contained neither water nor sulphur. It is scarcely to be expected that this method would yield uniform and exact results, unless the copper were in a very much finer state of division than that operated upon. 2. An attempt was made to deduce the oxygen of the dioxide from the loss occasioned by melting the copper in hydrogen. The apparatus employed for the purpose was a small Stourbridge clay crucible fitted with a perforated cover. It was filled with hydrogen by means of a small porcelain tube passing through the hole in the cover, and was heated by charcoal to the melting-point of copper. Great spirting occurred, and it was not possible to collect all the pro- jected globules of copper, which adhered to the inner surface of the cover and sides of the crucible ; so that the loss of weight, due to the reducing action of the hydrogen, could not be ascertained. Spilling took place even when the crucible was very gradually heated. 3. A weighed quantity of the dry copper was- dissolved in nitric acid, and the solution was saturated with caustic potass, and boiled. The precipitate was collected on a filter, washed, ignited, and weighed: it was then, evaporated with nitric acid, and exposed to a red heat until its weight became constant. From the protoxide of copper thus obtained, the proportion of copper which it should have contained, supposing it to have been pure, was calculated ; and the difference between the weight deduced and that of the original copper was estimated as oxygen. As commercial copper is never perfectly free from certain other metals, and probably also other matters, this method cannot yield absolutely exact results; yet, as the proportion of these foreign matters is comparatively very small, the error cannot be very great. In one experiment, 1O73 grains of dry copper gave 13-18 of protoxide, which is equivalent to 98-09 per cent, of copper ; and in another, 9-17 grains gave 12-26 of protoxide, which is equiva- lent to 98-01 per cent, of copper. The difference estimated as oxygen corresponds, in the first experiment, to 17-04 per cent, of dioxide of copper ; and, in the second, to 17*74 per cent. Karsten ftmnd that copper which was purposely made very dry, and could not be forged at any temperature without crumbling to pieces, contained 13-47 per cent, of dioxide, and had only a specific gravity of 8-0552 ; whereas it had previously, while in the state of maximum malleability, a specific gravity of 8-7574. According to Karsten, the effect of the presence of dioxide of copper in copper is to diminish the tenacity of the metal less at high than at ordinary temperatures ; and when y;///v copper contains 1/1 percent. 266 COPPER AND DIOXIDE OF COPPER. of dioxide, it is no longer sufficiently malleable to admit of being worked at ordinary temperatures without splitting into laminse and cracking at the edges. When the proportion of dioxide amounts to 1^ per cent, the decrease in tenacity is very perceptible at high tem- peratures, and the copper is brittle whether cold or hot, or, in technical language, it is both "cold-short" and " red-short." When copper which is smelted on the large scale is in the highest degree malleable at all temperatures, it is technically said to be "at tough pitch" by English smelters, or " hammer-gaar " by German smelters. In this state it is usually cast into flat rectangular cakes called " tough cake," a form suitable for rolling or hammering. No commercial copper is pure, but in the preparation of tough cake a notable quantity of lead is always expressly mixed with the copper just before it is laded from the furnace into ingot moulds. The reason of this addition of lead will be stated hereafter. When copper at tough pitch is cast into a narrow open ingot mould, the upper surface of the ingot is flat, presenting neither furrow nor ridge. When such an ingot is broken in two cold, its fractured surface is even, close- grained, free from fibres or cavities, presenting, especially towards the centre, numerous shining grains of a bright metallic lustre ; its colour is fine salmon-red, neither purplish-red nor orange. It is very interesting to watch the successive changes in external characters which copper undergoes in its passage from dry copper to tough pitch by the gradual, though not complete, reduction of the dioxide of copper. As the reduction proceeds, the fractured surface of the ingot becomes more and more even and acquires a paler and purer red colour, while the furrow on the upper surface becomes less and less distinct until it at last disappears altogether. When an ounce or so of tough-pitch copper is laded from the furnace and, after it has become cold, broken in two, its fractured surface is pale red and very finely fibrous, and has a characteristic silky lustre. The fracture may be conveniently obtained by nicking the piece of copper with a chisel, fastening it in a vice on one side of the nick, and bending the other with pliers backwards and forwards until it breaks. The toughness of the copper is proportionate to the number of times which it may thus be bent without breaking. An ingot of tough-pitch copper which was laded from the furnace at the Hafod Works in my presence has, together with other specimens of commercial copper in this state, been examined by Mr. Dick in my laboratory with the following results : 1. Some of the metal was heated to redness in a current of dry hydrogen, when water was pro- duced. Various experiments were made by this method with a view to determine the proportion of dioxide, just as in the case of dry copper, and the results obtained were equally discordant and unsatis- factory. The highest amount of dioxide indicated was 2*95 per cent. Towards the beginning of each experiment a trace of sulphuretted hydrogen was evolved, and, on testing the copper operated on, the presence of sulphur was detected. A slight metallic sublimate con- taining lead was formed, just as in the similar experiments upon dry COPPER AND DIOXIDE OF COPPER. 267 copper previously recorded. Tough-pitch copper which had thus been exposed in the state of wire or foil to the action of dry hydrogen at a red heat was found to have become extremely brittle, so that it could not be bent even once without breaking, and it had, moreover, lost the lustre of its surface. Flexibility could not be restored to the wire or foil by annealing at a red heat in a current of steam, which was employed because it exerts neither an oxidizing nor a reducing action. The same loss of flexibility occurred when hydrogen was replaced by car- bonic oxide or coal-gas. The brittleness thus induced seems to be due to the porosity occasioned by reduction of the dioxide of copper dif- fused through the copper, and must be distinguished from the brittle- ness which the copper acquires by being melted in any of these reducing gases. If tough-pitch copper so melted be rolled out into foil which may be done, though the metal cracks somewhat at the edges and afterwards heated to redness in any of these gases, it will not become in the least degree brittle. Malleable electro-deposited copper, which it was certain contained no dioxide, either before or after fusion under charcoal, acquired not the slightest brittleness by exposure to the action of those gases at a red heat. These results have an important practical bearing, and are worthy of particular attention. 2. An attempt was made to deduce the amount of dioxide of copper in commercial copper-wire which is always made from tough-pitch copper from the loss occasioned by melting in hydrogen. Spirting took place, and though in a less degree than in former experiments of the same nature, yet sufficiently to render the method inaccurate. 3. A known weight of copper-wire was melted under charcoal purified, as in all these experiments with a view to deduce the amount of dioxide from the loss consequent on its reduction. Even when heat was very slowly applied the same kind of spirting some- times took place, so that after each experiment the charcoal in the crucible was always washed by decantation to separate and collect any globules of copper which might have been projected into it. It will be shown hereafter that when copper is melted under charcoal, it certainly does not take up more than a very minute quantity of, if any, carbon. Two different samples of copper wire, A and B, were subjected to experiment. A was thicker than B. A. 218-24 grains lost by melting under charcoal 0-76 grain, which, estimated as oxygen in combination with copper, is equivalent to 3-10 per cent, of dioxide. B. 176-48 grains lost by similar treatment 0-635 grain, which is equivalent to 3*37 per cent, of dioxide. A contained 0-17 per cent, of lead, but no antimony could be de- tected in it. Its specific gravity was 8-853. B contained 0-29 per cent, of lead and 0*31 per cent, of antimony. Its specific gravity was 8-733. In a sample of commercial sheet-copper 0-27 per cent, of lead was found, but no antimony. Not one of these varieties of copper, -after fusion under charcoal, could be hammered out at a dull red heat without cracking at the edges, although each could be hammered out 268 COPPER AND DIOXIDE OF COPPER, cold without cracking in the least degree. The pieces employed to test malleability weighed from 1 50 to 200 grains each. Similar pieces cut from the tough-pitch ingot which I obtained at the Hafod Works could not, after fusion under charcoal, even be hammered out cold without cracking at the edges. The amount of lead in this copper was not ascertained. Eussian copper coin, which is said to be esteemed on account of the purity of the metal, was found to contain oxygen, and yet, after fusion under charcoal, it could be hammered out, even at a red heat, without cracking much. This seems to indicate a con- siderable degree of purity, though inferior to that of the electrotype copper employed in these experiments. Several experiments were made with the view of finding some means of remelting tough-pitch copper in crucibles without causing any variation in the amount of dioxide contained in it. Commercial copper wire was plunged under common salt in a state of fusion and melted, but the copper could not afterwards be hammered out at a dull red heat without cracking, though electrotype copper which had been treated in the same way could be hammered out either cold or at a dull red heat without in the least cracking. Similar results were obtained when chloride of calcium was substituted for common salt. Copper suffers a sensible loss of weight by fusion under these salts. Thus, in one experiment, the wire B lost by fusion under common salt 2-44 per cent, in weight; in a second experiment 2'05 per cent. ; and in a third, at a temperature just sufficient to melt copper, 1-35 per cent. The loss of weight was much greater when chloride of calcium was employed, amounting in one instance to not less than 7 '17 per cent. No similar experiments were made with electrotype copper. When the common salt in which the metal had been melted was dis- solved in water, insoluble matter was left, which contained both copper and chlorine. This deserves investigation. It may, possibly, be a compound of dioxide and dichloride of copper. When tough-pitch copper is kept melted under charcoal during a sufficient time and is then laded into a narrow open mould, the upper surface of the ingot when cold presents a distinct ridge along the median line. This ridge is sometimes of considerable height. In the act of solidification globules (Streu or Sprutz-kupfer of the Germans) of melted copper may be projected with force along the line on which the ridge will be formed. The copper is more or less brittle, so that the ingot may be easily broken in two. Its fractured surface is more uneven than that of tough-pitch copper, and appears fibrous through- out; it frequently presents small, irregular, tube -like cavities, of which the direction is from the sides and bottom of the ingot towards the median line along the upper surface, and through which gas appears to have escaped; its colour is paler and less red than the fractured surface of tough-pitch, and strongly inclines to orange. The appearance of the surface of the ingot suggests the notion that, after the upper surface had solidified in a certain degree, the still liquid, but more or less viscous, copper within had been squeezed through the upper surface along the line of least resistance. Copper in this state COPPER AND CARBON. 209 is said by English copper-smelters to be " overpoled" the reason of which appellation will be given hereafter and by the German smelters it is called " zu junges Kupfer" The cause or causes of this remarkable change in the external characters of copper will be pre- sently considered. Copper and carbon. According to Karsten, when pure copper is imbedded in lamp-black and exposed first to a strong red heat during several hours, and afterwards to the melting-point of copper, it may combine with as much as, but not more than, O2 per cent, of carbon. Copper containing this maximum proportion of carbon is distinguished by its pale yellowish-red colour and very strong metallic lustre, and at a low red heat it crumbles under the hammer. \Vhen the proportion of carbon does not even exceed O05 per cent, copper cannot be forged or rolled at a high temperature without split- ting into laminae and readily cracking at the edges. On the contrary, the presence of carbon scarcely affects the tenacity of copper at ordi- nary temperatures. The rigidity and brittleness which copper con- taining carbon acquires when hammered out cannot be removed by annealing at low temperatures, and, on this account, such copper is unsuited to fine plated work on the old system of plating copper with gold or silver. Such are the characters which Karsten ascribes to copper containing carbon. As this subject is one of much importance in the metallurgy of copper, the evidence in support of the existence of a combination of copper and carbon must be examined with care. Karsten properly remarks that it is very difficult to determine with certainty the highest amount of carbon which may be taken up by copper. He adopted the following method of analysis. The copper was left in an aqueous solution of nitrate of silver, when the carbon supposing it to be the only foreign matter present was left undissolved in admixture with the precipitated metallic silver while the copper was completely dissolved in the state of nitrate. A button of copper weighing from 8 to 10 grammes was found to be wholly converted into nitrate of copper by being left in an aqueous solution of nitrate of silver during 6 or 8 days. The precipitated silver is dissolved by dilute nitric acid, and the quantity of the insoluble residue of carbon is so small that it only becomes visible after diluting the solution of nitrate of silver consider- ably and then gently warming it. Karsten anticipated the obvious objection which may be made to this process, namely, that the carbon, especially in this finely divided state, would probably be acted upon in a greater or less degree by the nitric acid employed to dissolve the precipitated silver. Moreover, Karsten presents us with no evidence to prove that any dark-coloured residue which may have been observed after dissolving the silver by nitric acid was carbon, either wholly or in part. 9 It has long been known, that when commercial copper, even of the purest kind, is kept melted in contact with carbonaceous matter, so as 9 Sys. der Metall. V. p. 231 et seq., 1832. 270 COPPER AND CARBON. completely to prevent the action of the oxygen of the atmosphere, it acquires the properties which Karsten ascribes to copper supposed to contain carbon. Some years ago a specimen of " best selected " copper, which I knew had been prepared with particular care by Messrs. Newton, Keates, and Co., was subjected by myself to the following treatment : It was reduced to powder by pounding it while hot in a mortar. The finest of this powder, as obtained by sifting, was mixed with a large quantity of powdered wood- charcoal, and heated in a covered crucible during many hours, at about the temperature of melted copper. "When cold the crucible was opened. The copper was melted into small round shots, which were diffused through the char- coal. By blowing with bellows the charcoal was removed. The residual copper was re-melted in a crucible under charcoal powder, and cast, by an experienced caster of Birmingham, into a flat ingot in a closed iron mould, which had been duly coated with oil and charcoal, and then gently heated. The ingot was carefully rolled at Mr. Clif- ford's mill in the same town, without cracking at the edges ; and a piece of the rolled metal was drawn out into wire ; and yet I- had many times tried in vain to roll ingots of the same kind of copper, which had not been previously heated in the manner described with charcoal powder. Although the copper could be rolled, yet it could not be forged at a red heat without crumbling to pieces. The metal was melted under varying conditions of temperature under charcoal, as well as with the surface exposed to the air ; and yet in no instance could it afterwards be rolled without cracking at the edges, notwith- standing the rolling was conducted with great care and under different conditions of temperature. A piece of the sheet obtained from the copper heated with charcoal was placed on platinum foil and immersed horizontally in a solution of sulphate of copper : in the same solution, at some distance above, and parallel to, the foil, was fixed a plate of copper, which was connected with the negative pole of a constant battery, the platinum foil being connected with the positive pole. The whole was thus left until everything capable of being removed by the voltaic current from the sheet of copper on the platinum foil had been transferred to the plate of copper above. A very small quantity of dark-coloured matter was left on the platinum, in which any carbon which might have existed in the copper ought to have been present ; but I did not succeed in satisfying myself of the presence of carbon, which, however, I am sure, could only have formed a small proportion of the residue in question. Some years afterwards, Mr. Dick, at my request, investigated in the metallurgical laboratory this, amongst other subjects relating to the metallurgy of copper. The copper in which carbon was sought in the foregoing experiment was submitted to very careful examination, and was found to contain a notable quantity of silicon, and a small quantity of phosphorus and iron. Now it will be shown that when copper is exposed to a long- continued high temperature in contact with carbon in admixture with silica or phosphate of lime, it will take up silicon or phosphorus ; and it will be further shown that when COPPER AND CARBON. 271 ''best selected" copper contains only a very minute proportion of various foreign matters amongst which may be mentioned silicon and phosphorus it may be re-melted in a crucible, and cast into ingots capable of being rolled without cracking at the edges. Hence it may be inferred that the silicon and phosphorus in the copper in question were derived, at least chiefly, from the ashes of the charcoal with which it was heated, and in sufficient proportion to affect in a striking degree the working qualities of the metal. In the following experiments Mr. Dick employed electrotype copper, which was prepared from sulphate of copper at Messrs. Elkington and Co.'s Electro-plate Works in Birmingham ; and although not neces- sarily pure, yet it is likely to be free from lead, and some other foreign matters, which may occur in smelted commercial copper. The charcoal powder was digested in hydrochloric acid, and afterwards thoroughly washed with water. 1. Electrotype copper, melted in small pieces under charcoal, could be hammered out without cracking, whether cold or hot. 2. Some rather larger pieces of electrotype copper were imbedded in charcoal powder and heated for about half an hour at a temperature approaching whiteness. The contents of the crucible were then stirred with a piece of wood, so as to cause the metal to sink to the bottom ; after which it was cast into an ingot mould. 3. Several pieces of electrotype copper were placed in a brasqued crucible, which was then completely filled with charcoal powder, covered, and exposed to a temperature approaching whiteness for about an hour. The crucible was left to cool gradually in the furnace. The copper thus treated was re-melted under similar conditions and sent to Birmingham, together with that obtained in the last experi- ment, to be rolled into sheet and drawn into wire, with the direction that they might be treated exactly like ordinary copper when sub- jected to these processes. No mention was made of the treatment which the copper had received. They rolled well into thin sheet, and were drawn into tolerably fine wire. The report received from the rollers was that, although the casting was not good, the metal was fit for any work. 4. A piece of the sheet obtained in the last experiment was boiled in a solution of caustic potass, to remove any oil which might be present; it was then washed and dried, when it weighed 221-10 grains. It was laid in a platinum basin, immersed in a solution of sulphate of copper. The basin was connected with the positive pole of a voltaic battery, and a plate of copper connected with the nega- tive pole was fixed in the solution above the basin. The whole was protected from dust, and left until the residue from the copper in the basin was very small. It still contained a little metallic copper, which was dissolved out by a solution of sesquichloride of iron mixed with a little hydrochloric acid. The insoluble residue was then washed by decantation and dried, when it weighed - 08 grain ; its colour was very dark grey, nearly black. When a portion of it was heated on platinum foil, it evolved a slight and peculiar odour, glowed for an instant, and left a small amount of fixed matter. 272 COPPER AND CARBON. Another portion, weighing 0-012 grain, was introduced on a very small piece of platinum foil into a small glass tube, of which one end dipped into baryta water protected from the air, while the other was connected with an apparatus from which a very feeble current of air, quite free from carbonic acid, could be sent. A current of air was passed over the platinum, and produced no cloudiness in the baryta water. That part of the tube containing the platinum foil was then heated by means of a spirit lamp, when a very slight deposit of the colour of sulphur appeared in the cold part of the tube ; but when the platinum became red hot, every bubble of air, as it traversed the baryta water, caused a white precipitate, which dissolved with effervescence on the addition of excess of hydrochloric acid. The residue upon the plati- num foil weighed 0'003 grain ; it had a light reddish colour, and dissolved almost entirely in hydrochloric acid ; the insoluble portion was, probably, a minute quantity of silica derived from silicon, which the copper may have extracted from the charcoal. The solution contained a trace of iron, but was not turned blue by ammonia ; the quantity was too minute for any further experiments. The weak point in this evidence as to the presence of carbon is, that the preci- pitate in the baryta water might have been sulphite, and not carbonate, of baryta ; and that sulphurous acid might, not to say must, have been formed, is evident from the sulphur which was evolved and condensed at the beginning of the experiment. The copper employed in the foregoing experiment was tested for sulphur, and found to contain 0'05 per cent, of that element. 1 We are, therefore, not justified in con- cluding that this copper certainly contained carbon. It is hardly probable that the sulphur was derived from the charcoal, which had been purified by washing, first with hydrochloric acid, and afterwards with water, or from the gases of the furnace ; but it seems not impro- bable that the electrotype copper may have retained in pores some of the solution of sulphate of copper from which it was deposited : any sulphate so imprisoned would, in contact with copper at a high temperature, be reduced to disulphide. 5. Across the mouth of an ingot mould a copious flame of coal gas was maintained, and through the burning gas 33 6 '9 grains of electro- type copper, melted under charcoal, were poured. The copper, which contracted sensibly in cooling, was rolled without annealing, during which it cracked a little at the edges. ' To free it from any adherent oil, it was boiled in a solution of caustic potash and washed ; it was then acted on for a short time by dilute nitric acid, washed, digested in ammonia, washed well with dilute ammonia, and, lastly, with hot distilled water. In this state it was placed in a stoppered bottle filled with a filtered solution of sesquichloride of iron, which contained free hydrochloric acid and chloride of sodium. The bottle was digested in a water-bath during several days, and then allowed to stand at rest 1 Other specimens may have contained less, or more, or none. The presence of this compound in electrotype copper was unfortunately not suspected till nearly the close of the investigation. * The sulphur in this copper has been re-determined with the greatest possible care by Mr. Tookey, and found to amount only to 0-00259 ( = Cu 2 S 0-01279) in 100 parts of copper. OVERPOLED COPPER. 273 during one night. The clear supernatant solution was removed by a syphon to within a quarter of an inch from the bottom. The residual liquid was diluted and left at rest. Black matter subsided, which was thoroughly washed, and weighed after drying ; the weight was 0-16 grain. A minute portion of this sediment was cautiously heated on platinum foil, when it glowed for an instant, yielding a black residue, which at a higher temperature melted. The whole of the re- mainder was mixed with a slight excess of litharge, which had been previously melted with free access of air, and put into a very small tube drawn out to a point and closed at one end, and connected at the other with a small weighed tube containing a few grains of fused hydrate of potash. Heat was then applied to the first tube, when there was instant partial reduction, with the separation of distinct globules of soft lead. The point of this tube was broken off, and a little drawn through the potash tube, just as in an ordinary organic analysis ; the potash tube increased in weight 0'45 grain. It must be remembered that litharge is reduced when heated with disulphide of copper. 6. About 672 grains of filings of electrotype copper, which had been melted under charcoal, were mixed with freshly-ignited chromate of lead in a combustion tube connected with a Liebig's potash-appa- ratus. No gas was perceived to bubble through the potash solution. At the end of the process, some air having been drawn through the combustion tube, the potash apparatus was found to have increased 0-115 grain in weight. Assuming this to be carbonic acid, and that the whole of the carbon supposed to exist in the copper had been converted into carbonic acid, this increase in weight corresponds to 0*031 grain of carbon in 672 grains of copper, or only 0-0046 per cent. The preceding data, it must be admitted, do not surely establish the fact that copper melted under, or exposed to a high temperature in contact with, charcoal, does combine with carbon ; though they .seem to indicate the probability that copper may under these con- ditions take up a very minute quantity. It may, however, be re- garded as pretty certain that if copper have this power, it can only combine with a very minute proportion of carbon. But there is one point of practical importance which is proved by the foregoing expe- riments, namely, that comparatively pure copper is not rendered brittle by being heated, or melted in contact with comparatively pure charcoal. This conclusion will be further supported by the results of other experi- ments yet to be described. I much regret that the investigation, which had proceeded so far, should, from one cause or other, not have been resumed and carried to completion. Overpoled copper. As this variety of commercial copper by which is meant such as is ordinarily produced in this country is the result of the action of charcoal or other carbonaceous matter on copper con- taining dioxide while in a state of fusion, it is necessary to enquire whether the remarkable change which the metal undergoes during this process is due simply to the total reduction of the dioxide with which it was previously impregnated, or, as has been commonly 274 OVPmPOLED COPPER. supposed, to the combination of carbon with the copper after total reduction of the dioxide ; or, whether it may not depend on the joint operation of both these causes. One of the most characteristic pro- perties of commercial overpoled copper is brittleness ; but compara- tively pure copper such as the electrotype copper employed in the preceding experiments may, after having- been melted under char- coal, be hammered out hot or cold without cracking : it is also equally malleable after having been melted and cooled in hydrogen. Some of the copper thus treated was rolled out with smooth, or tl wire " edges, to the thinnest foil. Here, then, is a wide distinction between overpoled commercial copper, and pure copper melted in contact with charcoal. Both kinds of copper may be perfectly deprived of dioxide, and both may equally be exposed in a melted state to the action of charcoal, and yet one shall be excessively brittle, and the other malleable at all temperatures. From these data the conclusion is inevitable, that the brittleness of overpoled commercial copper must depend upon the presence of some foreign matter. It cannot be due to the presence of carbon ; for pure copper may be kept melted any length of time under pure charcoal without becoming brittle. Now, commercial copper is never perfectly free from certain other metals ; and amongst those which very frequently occur may be mentioned lead and antimony. Although only a minute quantity of these metals may exist in copper, yet it may be quite sufficient to determine the brittleness of overpoled commercial copper. But tough-pitch com- mercial copper must equally contain these metals, and, nevertheless, it is malleable, hot as well as cold. However, in tough-pitch copper there is always present a certain proportion of dioxide of copper, which cannot be removed without brittleness resulting. The exist- ence, therefore, of a certain proportion of dioxide of copper in com- mercial copper would seem to be essential to its malleability. This is precisely the conclusion at which Karsten had previously arrived. He states that while dioxide of copper, even in small proportion, decreases the malleability of pure copper while hot as well as cold, the presence of a certain amount of dioxide is essential in impure copper, in order that the brittleness of the metal at high temperatures may be as much as possible counteracted. He further remarks that the effect of all foreign matter in copper is specially to diminish its malleability in a greater degree at high than at low temperatures in other words, to render it more "red-short" than "cold-short" and that the dioxide diminishes the " red-shortness" of impure copper, so long as it does not exceed from If to 2 per cent. 2 We have previously seen that tough-pitch copper in the form of wire or foil is rendered excessively brittle by exposure to the action of ^ hydrogen or carbonic oxide at a red heat far below the melting- point of copper ; and as there is no reason to suppose that these gases act in any other way than by reducing the dioxide contained in the copper, we may infer that the oxygen of the dioxide cannot be abstracted 2 Sys. der Metall. 5, p. 248. OVERPOLED COPPER, 275 even from solid tough-pitch copper without destroying its mallea- bility. Admitting the preceding conclusion to be correct, it seems probable that the proportion of dioxide of copper required to produce the highest degree of malleabilit}^ in tough-pitch copper will vary, to a certain extent, with the nature of the foreign metals, or other matters, which may be present in commercial copper. But this is a point which can only be determined by very careful experiments. It need scarcely be remarked that, as copper passes from the state of tough-pitch to that of overpoled to the maximum, the proportion of dioxide is gradually reduced, until at last not a trace remains; so that copper much over- poled, but not to the maximum, may yet retain a sensible amount of dioxide. Concerning the mode of action of dioxide of copper we have not at present, so far as I am aware, any certain knowledge. \\ hether its action is simply mechanical, or whether the foreign metals present in commercial copper exist in all cases and in any degree as oxides in combination with dioxide of copper, future experiments must decide. It is well to repeat in this place that, although the proportion of dioxide in tough-pitch copper has been assumed to be equivalent to the oxygen actually found by experiment, yet, when other metals are present, it by no means follows that the whole of this oxygen is com- bined with copper to the exclusion of those metals. Indeed, an oxi- dized compound of copper, nickel, and antimony, called copper-mica, and having the formula 12(Cu,Ni)0+Sb0 3 , is occasionally found in dry copper smelted from ores containing nickel and antimony. 1 In this formula it will be perceived that the copper is said to exist as protoxide, and not as dioxide. It has been stated that when overpoled copper is laded under ordi- nary conditions into an open mould the surface of the ingot mes, and that in the case of a narrow ingot, it is thrown up so as to form a longitudinal ridge which is sometimes very distinct, and not less than a quarter of an inch in height. The cause of this has now to be con- -sidered. 1. AVhen electrotype copper of the degree of purity used in the fore- going experiments was melted under charcoal in a crucible and left to cool therein, not only was there no rising, but, on the contrary, a con- siderable depression in the centre of the upper surface of the button. In a specimen in my collection which was prepared in this way by Mr. Dick is a cavity of considerable size near the centre of the upper surface of the button, and projecting into it are crystals of copper grouped branchwise, much like certain specimens of the native metal, or the so-called fern-leaves on the surface of the large round cakes of commercial antimony. The depression in the surface and the crystal- line markings were uniformly observed in numerous repetitions of this experiment. It is manifest, therefore, that considerable contraction occurs in the centre of the mass during the Solidification of the metal, 3 Rnnimclsherg, Lchrk 8 '940 the air ) Another piece of the same ingot 8-952 Piece of another ingot cast in the same manner 8-922 Unannealed wire made from the last ingot 8*952 Unannealed wire made from copper which had been melted under I 0.037 charcoal and allowed to cool in the crucible I The same wire annealed 8*930 Piece of an ingot cast iu coal-gas 8 -948 Another piece of the same ingot 8-958 It will be remarked that the specific gravity of copper which has been melted under charcoal and cast with a porous structure is in- creased by the process of wire-drawing, so that it is nearly as high as that of wire made from copper which had been cast with a dense struc- ture. The specific gravity of copper which has been melted under charcoal and cast with a dense structure is not increased by the pro- cess of wire-drawing, and the specific gravity of the wire drawn from such copper is nearly the same before as after annealing. The specific gravity 6f one specimen of copper which had been cast in coal-gas, namely 8-958, is sensibly higher than the highest obtained by Mar- chand and Scheerer, namely 8'952, which occurred in a specimen of copper-wire hammered out to the thickness of O m ,00015 (O'OOG in.). The specific gravities of four specimens of crystallized native copper which had been carefully rolled out were found by Marchand and Scheerer to be 8-940, 8-935, 8-933, and 8-962. The last specimen was from Brazil, and contained a sensible quantity of foreign matter. It had the following composition : Copper 99-56 Silver 0'30 Gold 0-08 Iron 0-10 100 04 ELECTRIC CONDUCTIVITY OF COPPER. 287 The facts which have now been advanced seem to render it probable that copper does not suffer a permanent decrease in volume by com- pression, and that the observed variations in specific gravity were caused by the varying porosity of the metallic mass. Were it even possible to cast pure copper absolutely free from cavities, yet, if it con- tract during solidification after fusion, spaces must be formed in the interior of the metal, because the exterior first becomes solid, and in that state is able to resist the pressure of the atmosphere as the still liquid metal within gradually contracts and solidifies. Since the above was in type Mr. Charles O'Neill has communicated a paper to the Manchester Philosophical Society (March 5, 1861) on the Changes in Specific Gravity which take place in Rolled Copper by Hammering and Annealing. His results are embodied in the following extract from the published Proceedings of the Societ t y : " In the first series of experiments ten pieces of copper were cut from a sheet of the thickness of ^ inch ; the pieces weighed from 250 to 320 grains each ; their mean density was 8*879. The pieces were then separately subjected to the action of a powerful compressing machine, acting on the principle of the genou, about fifty blows being given. The density of these hammered pieces showed a mean of 8*855, being a loss of 0-024. The same pieces were annealed by being placed in red-hot sand, and cooled slowly ; when cleared from adhering oxide, the mean density was found to be 8-884, being an increase of 0-029 on the hammered pieces, and 0-005 on the original pieces. A second series of experiments, made with very great care, corroborated the first in the main points. The pieces were from another and better sheet of copper : ten pieces, weighing each from 420 to 520 grains, showed a mean density of 8-898, being hammered by the same machine ; their mean density became 8-878, showing a loss of 0-020 by hammering; upon annealing in a charcoal fire, the mean density of five out of the ten pieces was 8 - 896, showing a gain of 0-018 upon the hammered pieces, and a loss of 0'002 upon the original. A third series of experi- ments upon the change of density in a bar of copper by successive hammerings showed a loss of density from 8-885 to 8*867. " The conclusion drawn by the author from these experiments is, that the specific gravity of best commercial rolled copper is, contrary to what might have been anticipated, sensibly diminished by ham- mering. Electric conductivity of copper. We are indebted to Matthiessen and Holzmann for the best investigation of this subject. 8 Their results will be found in the table subjoined : the conducting power of a hard- drawn silver wire is taken as the standard and as equal to 100. 8 On the Effect of the Presence of Metals and Metalloids upon the Electric Conducting Power of Pure Copper. Trans. Roy. Soc. April 26, 18GO. On the Electric Conducting Power of Copper and its Alloys. By A. Matthiessen, Ph. D. Proceedings of Roy. Soc. Feb. 28, 1801. 288 ELECTRIC CONDUCTIVITY OF COPPER. Wires hard-drawn. 1. Pure copper 2. do. 3. do. 4. do. K -, do - How prepared. Temperature Conducting at which the ob- Power. servation was made. . Protoxide reduced by hydrogen ...... 93*00 . Electrotype copper, not melted ...... 93*46 . Do. commercial, do ................ 93'02 . No. 3, after fusion in hydrogen ...... 92*76 /No. 3, hydrogen passed through the\ oo.nn metal while melted .. .../ y Mean of the 12 determinations from which the) no no > 90 UO preceding numbers were deduced 186 C. 20 C 2 184 193 175 18 C 9 2(5. 27. Extremes observed 92-22 at 19 C 3 and 93-81 at 19 "7 Tlic conducting power was found to be increased about 2*5 per cent, by annealing the wires. 19f> 17 C 5 221 2oo 16 C 8 191 197 103 Copper containing dioxide, of which the proportion could not be determined with accuracy 73-32 Electrotype copper melted in contact with air Copper containing 2 do. do. do. do. do. do. 5 50 per cent, of phosphorus 95 do. do 13 do. do. 40 do. arsenic . 7-24 . 23-24 . 67-67 . 6-18 do. do. 2-80 do. do . 13-14 do. do. traces do. do . 57-80 do. alloyed with 3 20 do. zinc . 56-98 do. do. 1 60 do. do. . 76-35 do. do. traces do. do. . 85-05 do. do. 1 06 do. iron . 26-95 do. do. 48 do. do. . 34-56 do. do. 4 90 do. tin * . 19-47 do. do. 2 52 do. do. . 32-64 do. do. 1 33 do. do. . 48-52 do. do. 2 45 do. silver . 79*38 do. do. 1 22 do. do. . 86*91 do. do. 3 50 do. gold . 65-36 do. do. and 31 29 do. do. antimony \ lead } . 64-5 190 144 168 197 20 C 7 Electrotype copper, from a dense ingot melted under) ^o . charcoal, and cast in coal-gas by Dick (see p. 276). / do. from a porous ingot of the same] copper as No. 25, but poured into a mould under) 94-8 ordinary circumstances I According to Matthiesseii the conducting power of pure copper would be 96*4 at 13 C C. Electrotype copper cemented with charcoal, and con-) taming silicon and traces of phosphorus and iron.! 62*8 (See Dick's experiments, p. 270) ) do. do. , 63-2 12 12 C 8 13 13 14 C 2 In the following table Mr. Matthiessen has given the relative con- ducting powers of various kinds of commercial copper, as compared with pure unmelted electrotype copper, which is taken as the standard and as equal to 100 at 15'5C. 9 3 Report to the Submarine Cable Com- mittee of an Investigation relating to the causes of the different Electric Conduct- ing Powers of Commercial Copper, by A. Matthiessen. COPPER-SMELTING IN GREAT BRITAIN. 289 Temperature at Wires Conducting which the annealed. Power. observation was made. 1. Spanish, Rio Tinto. It contained 2 per cent, of arsenic, ) besides traces of lead, iron, nickel, dioxide of> 14' 24 148 C. copper, &c ) 2. Russian, Demidoffs make. It contained traces ofi 59.34 1007 arsenic, iron, nickel, dioxide of copper, &c ) 3. Tough-cake, make not specified. It contained traces) 71-03 5703 of lead, iron, nickel, antimony, dioxide of copper, &c. ) 4. Best selected, make not specified. It contained traces ) 01. Q* 142 of iron, nickel, antimony, dioxide of copper, &c. ... I 5. Australian, Burra-Burra. Traces of iron and dioxide I CQ Qfi -ijon of copper only were found f b 6. American, Lake Superior. It contained traces of iron ) Q9 ,- 7 -.CCA and dioxide of copper, and 0-03 per cent, of silver ... I HISTORICAL NOTICES ON COPPER-SMELTING IN GREAT BRITAIN. At present I am not in possession of sufficient materials to enable me to present a complete history of copper-smelting in this country. How- ever, such information on the subject as I have been enabled to collect from various sources, I shall now introduce. I am informed by Mr. Albert Way and Mr. Franks, the eminent archaeologists, that lumps of metallic copper, more or less rounded, have been discovered in different parts of the country, but always in association with articles of bronze. Mr. Franks showed me one of these lumps, which evidently had been melted, and which, on examination in the metallurgical laboratory, proved to be practically pure copper. Pennant describes a cake of copper found at Caerhun (also spelled Caer-hen), the ancient Cono- vium, near Con way : it weighed 42 Ibs., and in form resembled a cake of bees' wax; the diameter of the widest part was 11 inches, and the thickness in the middle 2| ; on the upper surface was a deep concave impression, with the words Socio Romce (to my partner at Rome), and obliquely across these was impressed in smaller letters the inscription, Natsol. Caerhen is only a few miles from Llandudno, where ores of carbonate of copper continue to be raised to this day. The first in- scription, and the occurrence of the lump in close proximity to mines yielding copper-ore of the most easily reducible kind, would lead us to conclude that this ore was melted in situ by the Romans. In the time of Elizabeth there was a rich copper-mine at Keswick, in Cumberland, of which that Queen deprived the Earl of Northum- berland on the ground that it was a mine-royal. 1 It is reported that not less than 4000 men were employed at this mine ; but this is pro- bably a great exaggeration. The ore appears to have been a sulphide ; for Webster, the author of the * Metallographia,' describes it as an ore " that must be often melted in the fire ere it be brought into the form of good copper." According to Camden, much good copper con- tinued during a long time to be made at Keswick and Newland ; but Webster, in 1671, wrote that " now the Work is quite left and decayed : 1 Some Account of Mines, and the I Thomas Heton, M.A., Vicar of Lctyston Advantages of them to the Kingdom. By | in Hertfordshire, etc. London, 1707, p. 15. U 290 HISTORICAL NOTICES ON yet I am informed that some do now melt forth as much- very good copper as serveth them to make half-pennies and farthings." a More ancient records of copper-mines exist : thus Edward III., in the fifteenth year of his reign, granted the right of working " the copper-mines of Skildane in Northumberland, and the copper mine of Alston-Moor in Cumberland, and the copper- mine near liichmond in Yorkshire, during a term of fifteen years, and on payment of a royalty to himself of one-eighth, and one-ninth to the Lord of the Soil," to a company of adventurers, amongst whom his brother Richard, Duke of Gloucester, and Henry, Earl of Northumberland, are mentioned. 3 That the copper-ore which was raised in these earlier times was smelted at or near the mines, I think there is reason to suppose, notwithstanding the absence of any positive historical record of the fact. The Hindoos have smelted copper from time immemorial ; and to this day conduct the operation in small blast furnaces about 3 feet high, with charcoal and cow-dung as the fuel. The ores which they employ are not those of the easily reducible class, such as car- bonates, but sulphuretted ores, like copper-pyrites. But, if these rude tribes of mankind are able to smelt copper-ores with success, it is not difficult to believe that our ancestors, at least those of the fourteenth century, possessed an equal degree of metallurgical skill. Moreover, it appears certain that copper-ore was raised in this country many hundred years ago, and it must either have been smelted at home or exported ; but I am not aware whether there is any historical evidence of the fact of such exportation : if not, we have an additional though negative argument in favour of the supposition which I have above ventured to express concerning the early history of copper-smelting in England. On the other hand it should be stated that our ancestors imported copper from Hungary 4 and Sweden, and allowed calamine to be exported as ballast. 5 Copper-works were in operation in Yorkshire during the last cen- tury. Mr. Keates has communicated to me the following particulars on this point : " Copper-smelting, I believe, was carried on in York- shire to a limited extent ; but all that I know of it was told me by old Samuel Burgoyno in 1822, who at that time was eighty-four years of age, and was consequently born in 1738. His father worked at the copper-works at Middleton Tyas, in Yorkshire. He said : ' The ore was green and red, and melted by blast ; the work stopped when I was about twelve years, and we came to live at Ecton.' " Mr. Keates has furnished me with a copy of a memorandum which confirms the preceding statement: "April 17th, 1752. Essayed the sample of Middleton Tyas round ore brought me by Mr. Rotton's son. Quantity 13 T - 4 C - 2 13 20 dwts. produce 9 dwts. of fine." This shows that the ore yielded 45 per cent, of fine copper. Jars states that in 1765 copper-smelting in this locality was effected in re- 2 Metallographia, etc. By Jolm Web- ster, etc. London, 1671, p. 244. I have cited Camden on the authority of Web- 3 Heton, op. cii, p. 9. 4 Seo Specification of Patent to George Danby, A..D. 1636, Jan. 21. ster. s Heton, op. cit. pp. 1")3, Io4. COPPER-SMELTING IN GREAT BRITAIN. 291 verberatory furnaces, and that various kinds of ore were raised from the neighbouring mines, amongst which he mentions green carbonate of copper, vitreous-copper, and rarely yellow ore, or copper-pyrites. 6 In Staffordshire copper-smelting was carried on at the village of Ellaston, near Ashbourne. The ore was obtained from the well-known Ecton mine in the vicinity. Specimens of this ore, which I have seen, consist of copper-pyrites in association with calc-spar. Plot, writing in 1686, informs us that when he visited Ecton, the mine had ceased to be worked, and that at the mills at Ellaston, where they smelted three kinds of ore, " all was out of order," the famous wooden-bellows that had no leather about them "having been earned away to Snelston, in I hirbyshire," whither he went to see them. From this it is clear that the smelting was conducted in blast-furnaces. 7 According to Plot, the stoppage of the mine and smelting- works was on account of "Copper comeing cheaper from Sweden than they could make it here." The working of the Ecton mine was resumed ; and Mr. Keates informs me that about 1750 the ores raised from this mine were smelted at Whiston, and some of the copper was carried to a forge at IJosley, on the river Dane, near Macclesfield, where it was hammered out into pans, &c. Other Staffordshire copper-ores were smelted at Cheadle about 1780. Mr. Keates has also communicated to me the fact that copper-ore was raised at the Eibden mine, distant a few miles from Alton Towers, and smelted at a place in the vicinity named " Blazing Star," on account of the light emitted at night; so that a blast furnace was probably employed. The ore consisted chiefly of carbonate and oxide of copper. Webster states, on the authority of one I )r. Merrett, that a copper-mine existed at Wenlock, in Staffordshire. 8 The following historical notice of copper-smelting in Lancashire has been kindly supplied by Mr. Keates : " The first introduction of copper-smelting into Lancashire was by the ancestor of the present Colonel Patten ; the works were at Bank <,)uay, on the banks of the Mersey, near Warrington. The building of these works commenced in 1717 or 1718. The ores were princi- pally ( 'ornish and Irish, with small importations from the West Indies, and the British Colonies in North America ; some also were got from Vldcrl y 1'Mge, Coniston, &c, These works were dismantled, I believe, about 1780. The next works in Lancashire were built very near I nverpool : the present Mersey Iron and Steel .Works stand on their site. They were carried on by Roe and Co., who had a brass manu- factory at Macclesfield. Cornish and Irish ores were smelted at these copper- works, which were discontinued about the year 1800. Next in succession were the works at St. Helens, and at Stanley, a few miles distant. These works were of considerable magnitude, and were established by the father of the late Lord Dinorben and his partners for smelting the ores raised at the Parys and Mona mines in Anglesea : [ have not the exact dates ; but I believe they were begun about 1780, 6 Voyages Metal lurgiqnes, 3. p. 72. I By Robort Plot, LL.D. Oxford, 1G8G. 7 The Natural History of Staffordshire. | ' Op. cit, p. 244. u 2 292 HISTORICAL NOTICES ON and discontinued between 1812 and 1815. Copper-smelting then ceased entirely in Lancashire, but was resumed in 1830, when the writer built works at Ravenhead, near the site of the old St. Helens' Works, primarily with the object of smelting the ore raised at the mines of General Bolivar in Columbia ; the legislature having granted permission to import and smelt foreign ores in bond, on condition that the produce should be exported in the state of cake or ingot copper. The works at Sutton, near St. Helens, were also built by the writer shortly after those at Eavenhead ; and these have been followed by others, so that at present the quantity of fine copper smelted from ore in Lancashire is probably not less than 6000 tons per annum. The principal ores smelted are from the West Coast of South America, Canada, Cornwall, Ireland, and Wales, together with the sulphides of low produce imported by the chemical manufacturers from Spain, Por- tugal, &c., who first extract the sulphur from them, and then turn them over to the copper-smelters." In the last century copper-smelting was carried on in Gloucester- shire, at Bristol, and other neighbouring localities; but I have not been able to ascertain when it was first established in this county, or when it was discontinued. Through the aid of Mr. George Grant Francis, of Swansea, I have had access to MSS. in which I have found some precise evidence on the subject : it is contained in a formal depo- sition, which I insert verbatim, and of which the object was to prove that the smelting of copper was commonly effected in blast furnaces, and coke used as the fuel. The original is not punctuated : " Copy. Edward James of Lower Forrest in the County of Glamor- gan coppermaii aged seventy-two years and upwards maketh oath that sixty years ago or thereabout he was employed in a copper-smelting work at Redbrook in the county of Gloucester under one Mr. Thomas Coster and that it was then in constant practice in the said copper work to make use of a blast-furnace for melting calcined copper ore with pit coal by which operation coarse copper was produced at the first smelting ready for the refinery and that his father Edward James now deceased was employed by the said Mr. Thomas Coster at Red- brook aforesaid upwards of thirty years at the said blast-furnace in making copper in the manner aforesaid as he believes and has often heard him declare and that in or about the same time at another copper work called Lower Redbrook situate near Redbrook aforesaid there was another copper work under one Mr. Chambers at which work it was also constantly practised to make copper in a blast-furnace with coked coal which blast-furnace this deponent hath so seen at work And this deponent on his oath also saith that upwards of fifty years ago he himself has worked at a blast-furnace in which copper was made and wherein coked coal was used at Landore in the said County of Glamorgan under Dr. Lane of the City of Bristol now deceased and under the late Robert Morris Esqre of Swansea now also deceased And this deponent on his oath further saith that so far from its being any new inventation to make copper by a blast-furnace and the use of cocked (sic) coal or charcoal it was the ancient method of making COPPER-SMELTING IN GREAT BRITAIN. 293 copper and formerly most in practice And this deponent has himself also used some years ago peat both raw and charred in a blast-furnace for the purpose of making copper. (Signed) " EDWARD JAMES. "Sworn at Swansea this 23 day of January 178 (sic) the words upwards of fifty years ago being then interlined before me 11 JN BE VAN, " One of his Majesty's Justices of the Peace in and for the said County of Glamorgan." - Jars published, in 1781, a description of the smelting of copper in the vicinity of Bristol. There were two works, to which the greater part of the ores raised in Cornwall were conveyed by sea. Eeverbe- ratory furnaces were used, of which there were not less than fifty in one of these works. The regulus preparatory to calcination was broken and ground under edge-stones by horses. 9 Aikin, writing in 1797, states that at Anilwch port in North Wales, the poorest ores of the Parys mine, which yielded only from li to 2 per cent, of copper, were partially smelted, so as to produce a regulus con- taining 50 per cent, of copper, which, together with the rich ores, was exported to Swansea. There were two companies, each of which had a smelting-house, in which were thirty-one reverberatory furnaces. 1 Copper- works were established by the Union Company at Eisca, i\ear Newport, Monmouthshire, in 1807, and continued in work till 1817, when, the copper trade being much depressed, the smelters deter- mined to reduce the number of works, and they accordingly drew lots to decide which works should be given up. The lot fell upon the Eisca works, which were consequently abandoned, and the buildings have since been used as chemical works. 2 We now arrive at the history of copper- smelting in South Wales. In Carew's ' Survey of Cornwall,' of which the first edition was pub- lished in 1002, is the following passage : " Touching metals : Copper is found in sundry places, but with what gain to the searchers I have not been curious to enquire, nor they hasty to reveal ; for at one mine (of which I took a view) the ore was shipped to be refined in Wales, either to save cost in fuel, or to conceal the profit." 3 From the evidence which I shall adduce, and for which I am indebted to Mr. G. F. Francis, it may be certainly concluded that the first copper-smelting works at Swansea were not erected until after 1720 ; and that anterior to this date copper-smelting works existed at Neath. The following interesting letter was written by Mr. Gabriel Powell (the father of the locally celebrated Gabriel who died about 1787) to Mr. Burgh, who, 9 Voyages Metallurgiques, 3. p. 222. i are added notes illustrative of its history 1 Journal of a Tour through North , and antiquities. By the late Thomas Wales, etc. By Arthur Aikin, London, Tonkin, Esq., and now first published 1797, p. 133 et seq. IV. na the original manuscripts by Francis 2 I am indebted to Mr. Octavius Mor- j Lord de Dunstanville, London, 1811, p. 21. gan for this information concerning the See also the note p. xii. as to the date of Kisca works. the first edition. 3 Carew's Survey of Cornwall ; to which 294 HISTORICAL NOTICES ON according to Mr. Hooper, the agent of the present Duke of Beaufort, was the then trustee and head manager of the Beaufort family : " Dear Sir. I have been favoured with 3*0 togeather with one inclosed to James Griffith well I delivered f he has returned an Answear in the same words as the Copy thereof hereinclosed wch will save me the trouble of writeing a great deal of what I thought it my duty to mencon to you in that affayre. I shall therefore only observe that it must be granted that the Duke by obliedgeing us in a pticular (Taken by others without makeing the least acknowledgem 1 to his Grace) will gain considerably if so I hope Mr. Beavans myselfe will not be thought the worse of encourageing the undertakeing. It may be imagined from the trouble wee take in solliciting this affayre that our measures are entirely broak if wee are disappointed therein but if I may creditt our managers It will not be 10 a yeare losse or advantage to us. They give instances of several workes carryed on without watter but were it absolutely necessary there are other wayes to supply that defect with ab* 40 fi mor expence wch were it to fall to my own share I should not value very much but what gives me the greatest uneasinesse is that a cunning crafty pson (I mean Mr. Popkins for I dare name him tho' he dare not avow his false as wel as sly insinuations now [sic: nor?] the 10 th pte of his incroachm 18 upon my L d Duke his. other neighbours) who has from the beginning opposed my L d Duke his interest should prevayle with that family (wch wee have served to the utmost of our power) to Deny us a favo r wch tends to his Graces interest. I cannot imagine how he comes to be soe careful of the health of the inhabitants of Swanzey all of a suddairi he has opposed its welfare all that lay in him p'ticularly in the contest between us the out Burgesses of Loughor for wch reason (among others) wee refused to bring him in Alderman Sil : Bevans myselfe are much more concerned for the health of this place, our- selves f family (wch are numerous) live in it wee are not soe necessitous or soe covetous that wee would endanger our healths for any considerations whatsoever. There are several copper workes near Neath, several, inhabitants ab* those workes yet we doe not hear the least complaint of any unhealthinesse thereby, nay, I am told there is a copper worke in y e the middle of South wark must ours be more unhealthy than those or Docto r Lanes wch is surrounded with inha- bitants, verry strange indeed what wee never heard untill y last Letter. It was insinuated unto me (who has been at 150 e ab l my Garden) in order to divert me from this affayre that it would spoyle my Garden but that I look upon as idle as the least of their ptences. There is but one wind in the 24 that wch blows y e the most seldom viz' N.N.E. that will blow upon any p,te of y e Towne y e work being three fields distance from y e uppermost house in Towne I am wel satisfy ed It will not affect us But upon y c whole let y e conse- quence be what it will wee are determined to Goe on think it very hard that the Inhabitants of this place should be debarred of seekeing those advantages wch their situation in titles them to without being opposed by strangers whose Avarice would ingrosse all advantages to COPPER-SMELTING IN GREAT BRITAIN. 295 themselves had wee the happinesse to see you here I should hope to have this affayre Terminated in our favo r but since that is denyed us I do insist on y e p,mise to visit our friends in Breconshire Col 11 Yaughan, Major W ms f honest M r Will : Aubrey expect you my house Pennant (wch I desire you will make y 8 ) is within halfe a mile of Brecon. If you will lett me know y e time I will be sure to be there to Receive you I doe not doubt but Ned Gatchmayd (m or w ?) f Will : Edwards will attend you j ani) worthy S r y r most obliedged humble For Servant John Burgh, Esq att GAB POWELL Foy near Monmouth 29 th Sept. 1720." These. . Three days before the date of the preceding letter, the municipal authorities of Swansea had formally sanctioned the erection of copper- smelting works within the precincts of the borough. This sanction is contained in the following document, which cannot now be found, and to which, according to Mr. Francis, there is no reference in the Hall Book : " Wee the Portreeve, Aldermen & Burgesses of the Burrough of Swansey in the County of Glamorgan doe hereby certifye all whom it may concerne That wee doe approve of the erecting of a Copper \Vorke upon the Bank commonly caled (sic) M r Ley's Alias Thomas Evans Coale Bank and wee doe further Declare that it is our unanimous opinion that the Carrying on the said Worke will prove very much to the advantage and not in the least prejudicial or Hurtful to our said BuiTOugh or the Inhabitants thereof. In Wittnesse whereof wee have hereunto sett our hands & the Common Seal of the said Burrough this 26 th Day of September. Anno : Dom : 1720." On this document is the following remark by the before-mentioned Gabriel Powell : "Since y e last Lre (letter?) y e Portreeve Aldermen (except Al n Ayres who is steward to Mr. Popkins, Jenkin Jones whose sonnes in Law are imployed in bringing D r Lane's Oars f M r Da : Tho 8 who is p,missed to succeed in y e Dukes affayres) Burgeses have putt their names f y e Comon (sic) Seal of the Burrough to an instrum* a Copy whrof is above written. Q p " Tho site referred to in the preceding document is, according to the late Mr. Dillwyn, that on which the Cambrian Pottery- Works are now situated. In George the Third's collection of topographical engravings in the British Museum I have found a curious old Indian ink drawing of copper-works at Llangefelach, the parish adjoining Swansea; and though I do not know when they were erected, yet it will be shown in the sequel that they were in operation in 1745. 296 HISTORICAL NOTICES ON From the evidence which has now been advanced we may, I think, conclude with certainty that copper-smelting had been extensively carried on at or near Neath for a considerable period before it was established at Swansea ; but I have not yet succeeded in obtaining more precise information on this subject. Carew, however, it will be remembered, states that copper "was refined in Wales ;" and as this statement was published in 1602, there can be no doubt that copper-smelting was in operation in the Principality before that date. The term refined, in the passage quoted from the ' Survey of Cornwall,' is evidently used as synonymous with our present word smelted. Hence, unless it can be shown that when Carew wrote, copper-smelt- ing was conducted in other parts of Wales, we may reasonably infer that the art had attained a considerable degree of development at or near Neath at least 120 years prior to its introduction into Swansea. It must be left to future antiquarian researches to elicit more precise evidence on this subject than we at present possess. In Cornwall during the last century several unsuccessful attempts were made to smelt copper, of which a record has been preserved by Tonkin ; and as the history of these failures may convey an important lesson to persons engaged in mining adventures, I insert this record without abridgment : it is contained in Lord de Dunstanville's edition of Carew's 'Survey of Cornwall,' and was evidently prepared in 1739 with a view to publication : 4 " This variety of ores and great increase in the mines has occasioned the setting up of six several companies for the buying of the ore, but who take care to keep us as much in the dark as they can, by shipping off all the ore to be smelted in their houses near Bristol, in Wales, &c., under a pretence of saving cost in fuel, but in reality to conceal the profit, as Mr. Carew very justly observes ; so that we must be entirely at their mercy, as not understanding the true value of the commodity ourselves ; or, if we did, they know that it would require a greater purse than any one private gentleman can be supposed to be enabled to lay out. It was, however, attempted about thirty years since by the late John Pollard, Esq. and Mr. Thomas Worth, jun. at St. Ives ; and before them by Mr. Scobell, at Pol Kuddan, in St. Austell, with whom the late Sir Talbot Clarke and the old Mr. Vincent joined, and where the first piece of copper that ever was so (sic) in this country was smelted, refined, and brought to perfection. But both these attempts failed of success, more through ill-management, roguery of the workmen, and the ill-situation of the said smelting-houses, than any defect in the ore, or charge of the fuel. Since this, one Gideon Collier, of St. Piran in the Sands, erected an house for the like purpose at Penpol, in the parish of Phillack ; but being soon taken off by a fever, in the best of his time, when he had made a fair progress in it, the same was carried on by the late Sir William Pendarves and Robert Corker, Esq., who have (particularly the last, with .whom I have often discoursed about it) assured me that they could smelt their ore as 4 P. 22. COPPER-SMELTING IN GREAT BRITAIN. 297 cheap there, all hazards considered, as the companies could pretend to do at their houses in Wales, &c., and did so accordingly for some years. But they being" both since dead, and their affairs falling into such hands as had other interests to mind, this project too sunk with them. A small beginning was also made to the same purpose at Lenobrey in St. Agnes, where they smelted some pieces of copper with good success ; but were forced to give it over for want of a sufficient stock to go on with it. From all which essays, and some observations I have made of my own and gathered from some workmen abroad (but chiefly from the late Mr. John Coster, who owned to me that most of our ores might be smelted rough here as cheap as abroad, but not brought to the true fineness (for what reasons you may easily guess), and there- fore must be shipped off to be refined), I am fully convinced that the ore may be smelted here, and refined too (that pretence being a mere cant, to conceal the real value), all things considered, at as small an expense as it can be done in Wales, &c. And if we allow for the great salaries the said companies are obliged to give' to their agents here and elsewhere; the great charges they are at in working the mines (which they covet at any rate to get into their own hands) ; the hazard of the ore on shipboard, especially in time of war ; and many other incidents, which would be saved if the ore was smelted here : I believe it would amount to a demonstration, that it would even be done much cheaper, in some convenient places in this county, than in Wales, &c. What advantage from this would accrue to our country in general, is too obvious to need any more words : and this the copper companies know but too well ; and therefore keep us as much . . . . . . [left unfinished by the author.] " In 1754 copper- works were erected at Eiitral, in the parish of Cam- borne, and afterwards removed to Hayle, where coal could be procured at a less cost. According to Price " the [copper] companies left no method unsought to traduce the credit and stab the vitals of this undertaking. Threats and remonstrances were equally used to oblige or cajole the owners of the mines to abandon or suppress the new com- pany at Hayle. The opponents of this association, using every expe- dient to mortify the spirit of this arduous undertaking, alternately raised the price of copper- ores and lowered the value of fine copper, to the great loss of the contending parties ; which will ever be the case where monopolies are disturbed and the almighty power of opulence can prevail. But happening to have men of fortune and capacity at their head, they were founded in prudence, and withstood the shocks ol power and artifice." 5 The same author informs us that copper- works were subsequently erected at North Downs, in Kedruth ; but the locality proving unsuitable, they were removed to Tregew, on a branch of Falmouth harbour, where they were carried on with advantage. From the language of these writers, it is evident that the Cornish mine-adventurers considered themselves the victims of a conspiracy on the part of the Welsh copper-smelters. But it is difficult to under- Mineralogia Cornubiensi*. London, 1778, p. 279. 298 HISTORICAL NOTICES ON stand why copper -smelting should have ceased in Cornwall if it had really been so profitable as Price declares. In one instance, at least, failure was not due either to deficiency of capital or incapacity in the management. As the adventurers felt themselves so much aggrieved by the smelters, they might have entered into a combination to keep up the price of copper-ore. Of all facts, none are more stubborn than those of political economy; and the truth of the matter appears to be, that copper-smelting can be conducted with greater profit in Wales than in Cornwall, and, therefore, it has become extinct in the latter county. When a man has an article for sale, he ought to know how much it has cost to produce it, and to fix such a price upon it as he considers remunerative. So, the miner should determine the value of the ore which he raises irrespective of the profit which it may subsequently yield to the smelter ; and he has no right to impute injus- tice to the smelter who declines to inform him of the gains arising from the metallurgical treatment of the ore and to allow him to parti- cipate in those gains, which often entirely depend upon the exercise of individual skill and the possession of sound commercial knowledge. Whatever the profits of copper-smelting may have been in former times, it is certain that the smelters of the present day do not, in general, realize more than they are fully entitled to expect. The last county to be mentioned in which copper-smelting has been conducted is Middlesex. About fifteen years ago works were erected on Bow Common for the purpose of smelting copper by a process de- vised and patented by Mr. James Napier, which will be described in the sequel. The locality was not suitable, and, as might have been anticipated, the works were speedily adandoned. The chief promoter of the undertaking was, I believe, the late Mr. Benjamin Smith, the silversmith, of Duke Street, Lincoln's Inn Fields. Towards the end of the last century, probably between 1780 and 1790, copper-smelting was carried on at Ballymurtagh, Wicklow, Ire- land. Through the kindness of Mr. Moyle, of Chatham Dockyard, I have received the following information on this subject from Mr. Edward Barnes, the present resident director of the Ballymurtagh Mines, now worked by the Hibernian Mining Company. Mr. Barnes writes that " when we first commenced the mine, none of the persons employed at the works were living, or at least remaining in the neigh- bourhood, and no records are to be found in the office of the Hibernian Mining Company on the subject. I think I have heard it stated that the smelting-works were erected by English parties, the Mining Com- pany selling them the ore as raised. At the period referred to I rather think a considerable export duty was levied upon copper-ores, which, added to the low produce of the ore of Ballymurtagh and its high percentage of sulphur, were the chief causes of erecting smelting- works near the mine. An attempt was also made to save the sulphur by calcining the ores in rude kilns in the open ay?, the sulphur-fumes being received into long horizontal flues. This process was very slow and unsatisfactory, and there is reason to believe the Company were losers by it. Judging from the cleanness of the slag at Ballymurtagh COPPER-SMELTING IN GREAT BRITAIN. 299 and at Arklow, it would appear that the process was well carried out and no copper left in it ; but no doubt there must have been great disadvantage in operating upon one stubborn class of ore. The Com- pany had a patent for coining their own copper tokens, as had also the Associated Irish Mining Company at Cronebane, who tried smelting on a smaller scale. This, I believe, was a general medium of payment with similar companies at the period." 6 It would be difficult to select in this country a more eligible site for copper-smelting works than Swansea, and this for two reasons. The first is, that it is a good seaport, which is only at a short distance from Cornwall and Devonshire, the two counties in which the greatest amount of copper-ore is raised, and it is also easily accessible to vessels conveying ore, or products containing copper, from South America, Australia, and other parts of the world. The second is, that extensive collieries exist in the immediate vicinity, from which an abundant supply of coal can be obtained at a low price. Many of the smelters are themselves engaged in the working of collieries, and are thereby enabled to dispose of their coal to the greatest advantage, the large being sold at a good profit, either for home consumption or exportation, and the small, which is often very dirty from an admixture with shale, being reserved for the copper- furnaces. It is advantageous, both for the mine-adventurers and the smelters, that the process of smelting should be carried on in a locality where copper-ores of various kinds may be procured, for it is well known that frequently copper can be extracted at a less cost by smelting several ores in admixture than by smelting any one ore by itself. An illustration will make this point plain. Suppose we have two kinds of very poor ore, one consisting almost wholly of oxide of iron and the other almost wholly of quartz. Et might not be profitable to smelt either separately ; for, in the case of that of oxide of iron, it would be requisite to add quartz as a flux, and, in the case of that of quartz, it would be requisite to add oxide of iron as a flux. But it might be profitable to smelt the two ores together, as one would then serve as a flux to the other, and each would yield copper. This is not an imaginary case. The smelter, by having at command a variety of ores, may render an ore profitable which otherwise would have no value. Adventurers in copper-mines would do well to consider this matter, and to be cautious how they embark capital in the erection of smelting-works which can only derive a supply of ore from some one particular mine. However, I do not mean to assert that particular copper-ores cannot be smelted with profit. The ad- vantages which Swansea possesses as a site for copper-smelting are shared in a greater or less degree by other localities in the vicinity, such as Tseath and Llanelly. The copper-works near the Lancashire roast may be well situated for the importation of ores, for the exporta- tion of copper from Liverpool, and for supplying the great local demand in Lancashire and the West Biding of Yorkshire ; but they cannot obtain coal at the same price as Swansea and its neighbours. The 6 MS... Avoca Lodge, Avoca, May 9th, 1861. 300 DRESSING AND SAMPLING COPPER-ORES IN CORNWALL. Swansea smelters enjoy the privilege of pouring dense volumes of thick sulphureous and arsenical smoke from comparatively low chimneys into the atmosphere, and destroying vegetation with impunity, for a con- siderable distance round. This privilege has now in the lapse of time become an established right, which would not readily be conceded in many other parts of the kingdom. The inhabitants of Swansea gen- erally seem to be habituated to the inhalation of the smoke, and to sub- mit to the evil, if evil it be regarded, with unmurmuring resignation. Description of the mode of dressing and sampling copper-ores in Cornwall. I am indebted for this description to rny former pupil, Mr. Pearce, of the Royal Institution, Truro. The ores consist chiefly of copper-pyrites, intermixed with small quantities of other cupriferous minerals, iron-pyrites, and the consti- tuents of the vein in which the ore is found. The ore is drawn from the shaft to about 30 feet above the surface and taken in waggons to an adjoining dressing-floor, where it is allowed to fall from this height on two rows of inclined bars, by which means it is separated into three divisions or sizes, called rocks, roughs, and smalls, for convenience of dressing. In the first, or top row of bars, which separates the rocks, the bars are 4 inches apart ; and in the second row, which divides the roughs from the smalls, they are 1^ inch apart. The rocks are broken, by a process called spalling, into pieces of about one pound weight each. A small hammer, called sledge, is used in this operation, which is conducted generally by girls, who carefully separate the ore from the refuse. The richest part of the selected ore, termed prills, is taken, without any further preparation, to the crusher, which consists of two iron cylinders, or rolls, revolving against each other. They are worked by steam or water power, and between them the ore is passed, and reduced to a size small enough to pass through a sieve divided into holes of \ inch square. The preparation of the ore is now completed. The poorest part of the ore selected from the rocks, called dredge, containing from 10 to 20 per cent, of copper disseminated through a large bulk of matrix, is also crushed and passed through a sieve of holes about J inch square. This prepares it for what is called jigging, of which the object is the separation of the ore from its matrix. The ore is put into a sieve of holes inch square, suspended in a large cistern of water. An up and down motion is then communicated to it by a break-staff, or lever, worked either by girls or by steam or water power. By this operation the lighter particles or refuse, on account of their less specific gravity, rise on the top, and are carefully skimmed off with a small wooden scraper. The ore that passes through, together with what stands in the bottom of the sieve, is then ready for the market. The roughs from the second row of bars are placed under a stream of water to clean the ore from all extraneous matter, so that the copper- ore may, from its colour, be readily distinguished from the refuse, which is easily separated by what is termed hand picking, an operation conducted by girls who sit on tables near the stream. The picked ore, SALE OF COPPER-ORES. 301 by this operation, is divided into prills and dredge, and treated by the methods before described. The smalls from the first separation are thrown on a griddle or sieve of holes J inch square ; the coarse or larger particles which do not pass through are subjected to precisely the same treatment as the roughs, and the fine or smaller particles, if rich, are ready for the market, but if poor they are jigged through a sieve of holes \ inch square, preparatory to sampling. The prills and smalls are then carefully mixed together into one pile or parcel, the dredge usually forming a separate parcel. The pile is then equally divided into 4 or 6 sub-piles, or doles, the number being regulated by the weight of the parcel ; if above 20 tons it is divided into 6 doles, and if below that weight into 4 only. This division into doles is effected by means of a measure called barrow, which when exactly filled with the ore weighs about Ij cwt. The number of bar- rows is carefully noticed, and, the exact weight of one being ascer- tained, a rough estimate of the total weight of the parcel can then be arrived at for the ticketing or sale. The parcel is now ready for sam- pling, which is conducted in the following manner : The Company's agent or sampler goes to the mine about a fort- night before the time appointed for holding the ticketing or sale, and selects two of the doles of the parcel ; these, called the sampling doles, are then cut through or divided each into two parts, the division being about 1 2 inches wide. The sampler then takes down from the sides of the divisions about 1 cwt. of each dole, which is taken to the sampling-house, and there carefully mixed by him. A portion of it, about 30 Ibs. weight, is then passed through a fine sieve, and again mixed by hand and put into as many small bags, holding about Ij lb., as there are smelting companies of which the number at present is 13 and one bag is also filled for the mine. The samples are then sent by the sampler to the companies' assayers, whose business it is to ascertain the quantity of fine copper in the sample according to the Cornish method of dry-assaying, and, according to the produces of copper found, the different companies fix the prices per ton they will offer for the parcel at the approaching sale or ticketing. The sale of copper-ores. Originally the Cornish miners disposed of their copper-ore by private contract ; and, according to Price, at the end of the 17th century gentlemen from Bristol purchased it for 21. 10s. to 41. per ton. In about 1720 other gentlemen of Bristol entered into " a covenant with some of the principal miners of Cornwall to buy all their copper-ores for a term of years at a stated low price, particularly with Mr. Beauchamp to buy all the copper-ore which should rise out of a mine stocked for 20 years at 51. per ton, and ore from other mines at 21. 10s. per ton." 7 In about 1730 a gentleman from Wales visited Cornwall and purchased, at the rate of 6?. 5s. per ton, ready money, 1400 tons of copper-ore which had been lying unsold during some 7 Mineralogia, p. 286. I derive my I copper-ore trade of Cornwall chiefly from information of the early history of the | this work. 302 SALE OF COPPER-OEES. years, and for which previously only 41. Ws. had been offered. The Welsh visitor bought 900 tons more at 71. per ton ; and in the course of six months, before he left Cornwall, he purchased 3000 tons, upon which, Price informs us, he deservedly made very little, if at all, short of 40 per cent, profit. Soon afterwards it was agreed between the miners and smelters that the latter should, at stated periods, tender for the ores which might be ready for sale. At the appointed time and place of sale the agents of the smelters assembled, and a person 'was appointed on behalf of the miners to conduct the proceedings. Each agent delivered a paper, or ticket, upon which were written the name of the smelter, or company of smelters, and the sum tendered. The papers were then read aloud by the president, who declared the ore sold to the highest bidder. With the exception that in Cornwall the results of the sale are read by a clerk, the same mode of proceeding is practised to this day, and the places of sale now are Swansea, Truro, Bed ruth, Pool, and Camborne. The first sale of ore by ticketing at Swansea took place May 14th, 1804, when 52 tons of ores from North Wales were disposed of. In former times Price states, that " on this ticketing-day a dinner almost equal to a City feast was provided at the expense of the mines in proportion to the quantities of ores each mine had to sell." Until quite recently the ticketing-day at Swansea was celebrated by a dinner, but now the sale takes place early in the day and no refreshment of any kind is supplied. However, the good old custom is still maintained as a sacred institution in Cornwall ; not, it is reported, with the willing concurrence of the mine-adventurers, who defray all the expenses of the treat, but for the special gratification of the smelters' agents, who, doubtless, like some other philosophers, con- sider it wise to season business with pleasure. When the same sum is offered by two or more smelters for a parcel of ore, it is equally divided between them. Each smelter has printed ticketing-papers with his name thereon. I subjoin a specimen of one in use at Swansea. Ores for Sale, 186 . Mine. No. 21 Cwts. Price. Per 21 Cwts. Dry Weight. FOE COPPEE COMPANY. Immediately after the sale the results are printed in a tabular form and distributed. When at Swansea in 1859 I attended a sale, after which the following table appeared : SWANSEA COPPER-ORE CIRCULAR, 303 ??s j ^2 r- 333? ss. 82 . 77 S tcoa S ?S a 2St & ^-? S 7". H j 1 1 i | 1 . l|l 1 :fl 2. ?. 1. p* c ri 5 S l ffi K S 3 *" C" 1 ^fl!^ O "^ q I | Copper <2 FOR SAL SEPTEMBER 1 Cobre August aS 1 - 1 I 1 c c 11 SJSS s ^C IS si n TnVlT'l? > I'J , !? 8. ~- I r. 1 ^;~ S; | ;w j!,ngiisn ana oa 39 Sundries 1 ( s | i t 3 < o O 1- a o O X < i 3 8 o o X c X to CO lO o r-. -" o CO O 500 r o 1 H o s 6 s ill* ' B o e - o o c D J g s 3 o X X o O X i ,si y v 7! 7! 7> 7! SI ~ -i O O S 8 i CO . X I- 1 2 1st: ^0 o co to M .- "22 O o o o S J to i p o> lO |3J ^2 71 7! 7j 7' 2 Z^ \ :: 2 2 2 ? 1 1 ^0 CO X X J 00 I 1 = *1* *7 71 7! 71 7! W N - H _ a 5 >0 CO 01 X X C c to o 5 c: ^ -. r vr s s 1 ^ 02 2 W r?) M 7! O 7! CO t - O (M -. r-. ae o g o o o ^O CO ai ifS ii 'X X 1 2 x o o i - o 1 S 77 tf s I m s I - 71 O 7! ^ C ot - - 5 to O - - 71 o s o 5 D O V5 CO X 2 " X X I 22'**' - r g E_, ^ , -I*" 6 W- 7! r -' 7) W C O CO in a - * * c o z - * e e o - : 5 3 * S GQ Mil Sa 7, 7' 71 rH "JfilJi *0*~ ~ ,, _ 71 Ot - >i o ^ - o o e 01 _, 1 L- ^ n en 01 X t I to x 71 _ 7^ ?-' V s s & o I & ^ O 7J 71 i- .-; 71 O C N - 3 t^ O o S o s O ^* N N X 2 "* X X C 5 CH l^ C to O jj X X II [i o a o 3 " 'I 1 *' r]>9 7! 7! 7! 7) IM i - -, -H _ | | 1 :- * TC - x x a o ! 7^ 1 H tf i g3 sis tt- 00 2 C 3 2 * I- O >c 3 8 o o 2 2 o X O . 2 2 c / 30 15 m n ; ~ s & g QQ -fjd o a 7! a ~ 71 IM 'i - f O X g i o co - o> t^ o 1 s c E tfi i j w I c^ | g CO | S? S? e 1 IM CO i ; ^ 5 -q. - o S? S 1 ~- t~ r~ "- -r 7- " r r ?.' M O ' . . . ' O ^ ^ 5s S P " !: ? P S ? 3 Si 5 -i * -r 3 S H w ' a 22" CM S *" CO 7 7! 77 S I a 02 <- =*- 1 1 i exR.E. i . ?: :' /: w H --. 5 I .1 $ a II- . . I : ! ? r 9 * I 1 i 1 Cobre Augustc 6 7 Leon RaymtuH j j c c> s M .^ =. I < !* 2. 2 2 2 17 Knockmahon ,. 2 121 Spanish Ore. 1 22 1 - 8S Wildberg.... 126 English & Can 97 e 29 AustralianReg EACH COM 304 DEFINITION OF " STANDARD." Under the heading " Produces " there are, it will be noticed, two columns : in the first are stated the results of the assays, which have been made on behalf of the sellers, and about which no secrecy is maintained ; the second is left blank, and is intended for the insertion of the results of the smelters' assays, which are kept rigidly secret. The sums immediately below which are the transverse lines in the columns of tenders are the highest bids, and, accordingly, those at which the ores were sold. I now insert (see p. 305) a printed form of the results of one of the copper-ore sales at Truro, introducing only a few of the biddings by way of example, as the table is too long to be given in extenso. Standard. It will be remarked, that at the head of the last column but one on the right the term standard occurs; a term which is in common use, and is generally quite unintelligible to persons not actu- ally engaged in copper-smelting. For the following history of its origin, and explanation of its present meaning, I am indebted to Mr. Keates, who designates it as an "everlasting stumbling-block of copper- trade technicality." " Originally there were few copper-smelters and few miners, and it was customary for the former to contract with the latter to buy their ores for periods varying from a quarter to one year, agreeing to pay them for the same according to a standard price of copper determined on ; and which price was usually the selling price of tough-cake copper at the time. Thus, if copper were selling at 120/. per ton, the standard was fixed at that rate, or thereabouts. Out of this standard price the miner returned to the smelter a certain sum on every ton of ore sold, 8 which for many years was fifty-five shillings per ton, though originally it was more. This was called the returning charge. The miner also gave the smelter 1 cwt. of ore upon every ton, to cover waste on removing the ore from the mine to the smelting- works : the smelter was also allowed a varying number of pounds of ore in each ton as compensation for moisture in the ore, the bulk being weighed wet, while the assay er's sample was weighed dry. Examples. Suppose a bargain made, and the ore weighed by the miner to consist of two par- cels, one of 1004 cwt., which is guessed to contain -| cwt. of moisture per ton ; the other of 800 cwt., which is guessed to contain cwt. of moisture per ton ; so that in the first case the quantity of ore paid for is only 46 tons 14 cwt. 1 qr., and in the second it is only 36 tons 15 cwt. 2 qrs. Let the first parcel produce by assay 1 Oj- per cent, and the second 5J per cent, of copper. In the first parcel the gross value of the copper in the ton of ore will be 12/. 12s., the standard price of copper being 120Z. For 100 parts of copper (say 1 ton) : 120?. : : 10J- parts : 12/. 12s. But from this gross value of the copper in the ton of ore the returning charge must be deducted. Thus, gross value of the copper 12/. 12s. ; returning charge, 21 15s. ; price per ton of ore, 9Z. 17s. In like manner the price of the second parcel of ore will be found to be 31 1 Is. Thus far the term standard is simple and intel- ligible, meaning neither more nor less than the price of copper. 8 The ton of copper ore is always 2352 Ibs. or 21 cwts. if not otherwise expressed. SALE OF COPPER-ORES. 305 S e a -? Epipls^^a^^cSii^ FEINTED BY MES TREOASKIS, he Ticketing Paper Office, EEDRUTII. AT TABB'S HOTEL, RED ORES SAMPLED Average Stand- ard, Produce, and Price. Quantity of Ore, Copper, and Money, &c. &c. is O "* o II it tandard and Produce of last Sale. 138 1 0. i --' > 000 tC tOtO 2 = 22^'2 ; T> =1 O C< O 00 O to to O .-~ = <3o O O o o to o o o to o o to 2 2 * 2 2 to to to to to to -, _ 2 o o -H O O 00 O O O * wU tc to o to to tO r- !C OS to c o o o o o 2 2 " S -* S * o HJ'-O S2 m '*3 ow S3 * e a to o o o o'W O -H 0> o to o to to o o to o -otooooo-^ooooooto ^-^; i WIIEAL BASSET ____ iow c/ the detuU of thi the following Mines i ted to save space.] W [X :::=:::::::: ; ;= = ! ja 5 : : - ^ li : : : g o : : 306 SALE OF COPPER-OPiES. The sources of the smelter's profits were the care with which he got his ores transported from the mine to his works, so as to save as much as possible of the 1 cwt. of ore in each ton which he did not pay for, the portion of the 21. 15s. returning charge which was not actually expended in smelting, and the surplus, or quantity of copper which his furnaces yielded in excess over the crucible of the assay er, and this of course would vary with the skill as well of the assayer as of the smelter. Originally copper-ores were dressed to a pretty uniform rate of produce, perhaps from 9 to 12 per cent., and, whatever the produce, the standard did not vary. By and by the ores were not dressed so uniformly; some came to market, say of 15 per cent, produce, others of 5 per cent. But the smelter, more acute than his neighbours, saw that he had better buy those of 5 per cent, and leave the others, because it took a much less portion of the 2L 15s. to smelt a ton of ore of 5 per cent, than a ton of ore of 15 per cent, produce. 9 Now the simple mode of meeting this was to have had a varying scale of returning charges, instead of which these charges remained the same, while the standard was varied with the varying produce of the ores, so that with copper at 120?. there might be a standard of 115?. or 130?., and thus the word standard lost its former simple and correct meaning. Competition went on increasing, processes were improved, carriage, freights, coals, &c., were lowered, but the returning charge continued the same, with, of course, less applicability than ever to the varying produce of the ores. An illustration of what actually occurs at a modern sale will make the matter plain. Out of a modern sale of 3000 or 4000 tons of ore, varying in produce from 4 to 20 per cent., let us select the following lots, with the prices at which they were sold : . . d. 100 tons of 5 per cent, produce 4 12 id. 8 id. 810 id. 12 id 12 18 id. 16 id. 17 8 id. 20 id 21 15 The smelter has no longer got his standard price of copper arranged with the miner as of old, but he opens his eyes to all the circumstances, or ought to do ; he sees what sort of ore he wants ; he knows the rale of carriage and freight which he will have to incur on each parcel ; he knows that one lot melts easily, another with difficulty, a third makes good copper, a fourth bad, and so on ; and, in the end, he finds he has bought the five lots of ore above mentioned at the prices affixed. Immediately these prices are disclosed in the sale-room the miners' and smelters' clerks proceed to calculate the standard in the folio wing- manner : . 6!. d. Price of the ore of 5 per cent, produce 4 12 Add returning charge 2 15 .770 But this sum refers to the ton of ore, or 5 per cent, of the ton of copper, so that the standard of the ton of copper will be 7?. 7s. x 20 = 147?. 9 The father of the late Mr. Vivian, it I to apprehend this important commercial is reported, was the first person clearly | truth. COPPER-SMELTERS IN ENGLAND AN UNIVERSITY Again: Price of the ore of 20 per cent, produce 2 Add returning charge 2 1 This multiplied by 5 gives the standard of 122?. 105. Hence the standard is now deduced from the price, and not the price from the standard as formerly. The buyer makes his offer without thinking of the standard. "When the sale is over, the average produce of all the parcels of ore is deter- mined, and also the average standard. Taking the 5 lots enumerated, the average produce is 12^6- nearty, and the average standard 132/. 45. nearly. The only purpose which this modern standard serves is a ready mode of comparison of prices or of rates at which copper in the ore has been sold. For instance, instead of saying last week ores of 5 per cent, produce sold for such a sum and this week they sold for such a sum, the phrase is, the standard is down a couple of pounds, or up 5?., as the case may be. Copper-smelters in England and Wales. I am indebted to Mr. Keates for the subjoined list of smelters in July, 1861. Proprietors. Pascoe Grenfell and Sons Do. Vivian and Sons Do. Williams, Foster, and Co Do. Do. Do. Sims, Willyams, and Co , Copper Miners' Company Mona Mining Company Keys and Son British and Foreign Company Newton, Keates, and Co Xewton, Keates, and Co Bibby, Sons, and Co Mason and Elkington Charles Lambert Do. Frederick Bankart Sweetland, Tuttle, nnd Co Vivian and Williams , Williams and Vivians and others James Radlcy Bold Copper Company , Name of Works. Locality. Middle Bunk Swansea. Upper Bank Do. Hafod Do. Taibach Aberavon Morfu Swansea. Landore . . Do. Rose Do. Crown Neath. Llanelly ijlanelly. C win ft von Aberavon. Mona Ami wch Winston Cheadle Parr St. Helen's, Liverpool Sutton Do. Ravonhead .... Do. Pembrey Near Llanelly Port Tennant Swansea. Widnes Dock. Liverpool. Red Jacket Neath Britonferry Do. White Rock Swansea Mines Royal * Neath Pocket Nook St. Helen's, Liverpool Bold ... Do. Associated copper-smelters. There are certain of the smelting com- panies about half whose assa} r ers act in concert, and assemble weekly, when each presents the results of his assa} T s of the samples of ores announced for sale on a given day. The assayers compare their results and agree upon a uniform list of produces, which is called the Incorporated by Royal Charter, James I., A.D. 1504. 308 ASSOCIATED COPPER-SMELTERS. OBES OF COPPER. " settled list;" and by this the associated smelters are supposed to be guided in their biddings for the ores. But this may not always be the case. Thus, suppose the produce of a particular lot of ore to be returned as 9J per cent, by the private assayer of a company while it is only fixed at 9 per cent, in the "settled list," the company would probably bid on a produce of 9^, and vice versa. Admission to this conclave of assayers is believed to be of great advantage, because the error of any individual assayer is sure to be found out and corrected. The strictest secrecy is attempted to be maintained with respect to the " settled list," both the smelters and the assayers of the association in question being under a promise not to impart information concern- ing their proceedings to any " outsider." The companies have of course a perfect right to enter into a combination of this kind, but it is questionable whether it be wise on their part to affect so much mystery, and forbid the publication of the " settled list " after the sale. Secrecy engenders suspicion, and people are apt to conclude that deeds are kept in the dark because they will not bear the light. Thus many mine-adventurers are under the impression, which may be very' erro- neous, that all these strict injunctions as to privacy on the part of the associated smelters can have no other object than that of keeping down the price of ores. My own conviction is, that if the "settled list" were published in due course after the sale, all cause of suspicion would be removed, and the association would benefit rather than suffer. It has been reported that in making out the "settled list" only the lowest produces are selected, and that the average of all the produces is not taken ; but, from what I have seen, I believe this report to be without sufficient foundation. Evidence on this subject will be presented under the head of Assaying in a subsequent part of this work. However, it is confidently asserted, that on one occasion good reason existed for disputing one of the produces set down in the " settled list." The following anecdote, which I received from an excellent authority, may be adduced in confirmation of this statement. Some years ago a rich copper-ore was assayed by a professional assayer of great experience, and reported by him to contain from 30 to 40 per cent, of copper. The produce, whatever it might be, was 5 per cent, higher than that in the " settled list." On the ticketing-day the par- ticular lot of ore was presented .for sale, when, on account of the dis- cordance above mentioned, and which came to be known, a proposal was made to withdraw it. Some of the smelters present objected to this proposal on the ground that it was not likely that the assays of their united assayers should all be wrong and the assay of a single assayer correct. At this period the manager of certain copper-works rose, and with a degree of moral courage which did him honour, boldly avowed that his private assay was 4^ per cent, higher than that of the settled list. This decided the point, and its sale was postponed. It was subsequently sold at a price corresponding to the higher produce of the single assayer. I have recorded this anecdote simply to show that the "settled list" may not, in every case, be quite infallible. ORES OF COPPER. 309 ORES OF COPPER. Under this head an enumeration of the various mineral species which are subjected to metallurgical treatment will be given, and their com- position stated ; but for information respecting their physical characters the reader is referred to any of the excellent standard works on Mineralogy. 1 1. Native copper. Tt is not an unfrequent constituent of certain copper-ores. It occurs diffused in isolated particles, in the form of thin laminae, in dendritic pieces, and in solid blocks, occasionally of large dimensions. On breaking pieces of ore, a nucleus of metallic copper may sometimes be found, coated successively with red oxide and carbonate of copper. The richest deposits of native copper which have been discovered are those of Lake Superior, in North America. My friend Professor Brush, of Yale College, U.S., has communicated to me the fact that in 1858 6000 tons of copper were procured from the native copper of Lake Superior alone. Mr. Petherick, the well- known mining engineer, informs me that at Minnesota in 1854 not fewer than forty men were engaged during twelve months in cutting up a single mass of native copper weighing about 500 tons ! Native copper is generally believed to be extremely pure ; but I do not find that many analytical examinations of it have been made. That it is not necessarily pure, is shown by the analysis recorded at p. 286. The native copper at Lake Superior in some places occurs curiously intermingled, but generally not alloyed with, native silver. Haute- feuille analysed a specimen from this locality which contained both silver and mercury. The results of his analysis are as follow 2 : Copper 69-280, Silver 5-453, Mercury 0-019, Matrix 25*248. A considerable quantity of ore is imported into Swansea from Chili under the name of " copper sand" or " copper barilla;" it consists of from 60 to 85 per cent, of metallic copper intermixed with quartz. Native copper is generally remarkable for its toughness. Mr. Morgan, of the Hafod Works, informed me that the toughest copper he had ever seen was a piece of native copper from Chili, about three-eighths of an inch in thickness : it was bent backwards and forwards forty- eight times before breaking. 2. Red oxide of copper, Cu 2 0. When pure it contains 88*78 per cent, of copper. It occurs in Cornish, South American, and especially Australian ores. 3. Black oxide of copper, CuO. When pure, it contains 79'82 percent, of copper. As obtained from the mines it is frequently very impure. It occurs at Lake Superior. 4. Green carbonate of copper, or malachite, CuO, CO 2 -}- CuO, HO. When pure it contains 57*33 per cent, of copper. It is a frequent consti- 1 Amongst these may be mentioned 8vo., New York and London, 1854 ; Ele- tlie following : Phillips's Mineralogy, by mente der Mineralogie, von Dr. C. F. Brooke and W. H. Miller, 8vo , Longman and Co., London, 1852 ; A System of Mineralogy, by James D. Dana, 2 vols. Xuiimunn, Svo. Leipzig, 1846. 2 Ueb/ersicht, Kenngott, 1800, p. 108. 310 ORES OF COPPER. tuent of copper-ores. It lias been largely imported from South Australia. In analyses of this mineral from various localities, no other constituents are recorded than oxide of copper, carbonic acid, and water. 5. Blue carbonate of copper, 2 CuO, C0 2 +CuO, HO. When pure it contains 55*16 per cent, of copper. A considerable quantity of this mineral was formerly obtained at Chessy, near L^ons, in France. It has been imported from South Australia in admixture with green carbonate. From the analyses of this mineral which have been published, it would appear to be equally free from foreign matter as the green carbonate. It must, however, be borne in mind that mineralogists would only select for analysis specimens of the greatest purity. Although the mineral species themselves may be pure, yet they may occur in association with other substances, which would cause them to yield an inferior quality of copper. Thus I was assured that the carbonate ores from Kanmantoo, in Australia, contained anti- mony and bismuth. Generally the copper produced in the smelting carbonates is of the best quality. 6. Vitreous, or grey sulphide of copper, Cu -2 S. When pure it contains 79-79 per cent., or nearly four-fifths of its weight, of copper. It is of frequent occurrence in Cornwall. In nine analyses of this mineral recorded by Eammelsberg, 3 iron appears as a constituent, varying in amount from 0'5 to 3'33 per cent. 7. Purple copper, 3 Cu 2 S-j-F 2 S 3 . When pure it contains 55-54 per cent, of copper. It generally occurs massive and disseminated, and but very seldom crystallized. Eammelsberg has classified the varieties of purple copper according to the proportion of copper which they contain : in the 1st class, the copper ranges from 56 to 58 per cent. ; in the 2nd, from 60 to 64 per cent. ; and in the 3rd, it is fixed at 70 per cent. The following selected analyses furnish examples of each of these classes 4 : Copper .. 1st ( 56-76 11-84 28-24 Jlass. 2. 2ndC 3. Mass. 4 3rdC 5. lass. \ 6. 56-10 17-36 25-80 99-26 60-80 13-67 25 46 61-07 14-00 23-75 69-72 7-54 22-65 70-0 7-0 22-3 Iron Sulphur 99-84 99-93 98-82 99-91 99-3 No. 1. From the Condurrow mine, Camborne in Cornwall, by Platt- ner. No. 2. From Sweden, by Plattner. No. 3. From Coquimbo, Chili, by Booking. No. 4. From Killarney, Ireland, by Phillips. No. 5. From Eisleben, Prussia, by Plattner. No. 6. From Tuscany, by Berthier. The rational constitution of some of these minerals at least has not been satisfactorily established. 8. Copper-pyrites, or yellow copper-ore, Cu 2 S + Fe 2 S, or, as it was for Handbuch der Mineralchemie, 1860, p. 50. 4 Op. cit., p. 113. ORES OF COPPER. 311 merly expressed, CuS + FeS. When pure, it contains 34-81 per cent, of copper. It is the most abundant ore of copper. It is largely imported from Cornwall, Devonshire, Cuba, and South America. Mag- nificent specimens of it are yielded by a rich mine in Tuscany. In eight analyses of this mineral recorded by Eammelsberg, no mention is made of the presence of any foreign matter, except a little quartz. In the ore raised at the Fowey Consols Mines in Cornwall, and in which the copper exists in the state of copper-pyrites, both nickel and silver occur in small quantity ; but whether they actually exist in the copper-pyrites itself I am not aware. In ores containing this mineral iron-pyrites is frequently present in large quantity. 9. True grey-copper ore, or fahlerz. This ore is found in numerous localities, and is very complex and variable in composition. It may be divided into three principal classes, of which the following analyses are selected as examples 5 : 1. ANTIMONIAL GREY-COPPER ORE. 1. 2. 3. 4. Copper 30-47 34-48 37-11 37-95 Antimony 26'56 28-24 25'97 28-78 Silver 10-48 4-97 1-09 0-67 Iron . . 3-52 2-27 4-42 2-24 Zinc 3-39 5 '55 5-02 2-52 Lead 0-78 0-54 Sulphur 24-80 24-73 23 76 25-82 100-00 100-24 97-91 97-98 Xo. 1. From Neudorf, Harz, by Rammelsberg. No. 2. From Claus- thal, by H. Rose. No. 3. From Durango, Mexico, by C. Bromeis ; 0-47 of matter was left undecomposed. Xo. 4. From Goslar, Harz, by Rammelsberg. 2. ARSENICAL GREY-COPPER ORE. 1. 2. 3. 4. Copper 47-70 51-62 41-07 42-60 Arsenic 12-46 19-03 18-87 19-01 Iron 9-75 1-95 2-22 9-21 Zinc 8 89 Lead 0-34 Sulphur 30-25 26-61 28-11 29*18 100-16 99-21 99-50 100-00 Xo. 1. Tennantite, from the Trezavean mine, Kedruth, Cornwall, by Phillips. Xo. 2. Tennantite, from the same locality, Eammels- berg. Xo. 3. From the Prophet Jonas mine, Freiberg, by Plattner. X'o. 4. Modurn, Xorway, by Fearnley. Rammelsberg, op. cit., p. 86 et seq. 312 ORES OF COPPER. 3. GREY-COPPER ORE CONTAINING BOTH ANTIMONY AND ARSENIC. 1. 2. 3. 4. 5. 6. Copper . . . 38-42 37-98 39-18 40-57 38-63 30-73 25-27 23-94 23-66 21-47 16-52 17-76 Arsenic 2-26 2-88 4-40 2-42 7-21 11-55 Silver 0-83 0-62 0-56 2-37 10-53 Iron . 1-52 0-86 6-99 2-92 4-89 1-42 Zinc 6-85 7-29 5-07 2-76 2-53 Sulphur 25-03 25-77 25-64 26-10 26-33 25-48 100-18 99-34 99-87 99-11 98-71 100-00 No. 1. From the Aurora mine, Dillenburg, by H. Rose : this mineral gave a red streak. No. 2. From Kapnik, Hungary, by H. Eose ; streak red. No. 3. From Cornwall, by Wittstein. No. 4. From Beresow, Siberia, by Lowe ; the 0-56 of silver was inclusive of some vein-stuff. No. 5. From Gersdorf, Freiberg, by H. Eose ; streak black. No. 6. From North Carolina, U.S., by Genth ; streak brown-red. Grey-copper ores containing mercury has been found in Hungary, the Tyrol, and Tuscany. 6 In eleven analyses recorded by Eammels- berg, the mercury ranged from 0'52 to 17*27 per cent. Much has been written on the subject of the rational constitution of grey-copper ores, and numerous and complicated formulae have been proposed; but not until we know more concerning the so-called foreign matter, or, as some mineralogists designate it, " dirt," is it probable that satis- factory formula will be established. 10. Chrysocolla. It is essentially a hydrated silicate of protoxide of copper. Two species are accepted of the formulae 3 CuO, 2 Si0 3 -f 6 HO and 2 CuO, Si0 3 -f 3 HO. 7 Both, I presume, are confounded under the general term of Chrysocolla. The following analyses will suffice for illustration. l t 2. Silica .. .... 32-55 40-09 42-32 27-97 1 63 Protoxide of iron 4 94 1-76 1-49 1-06 0-78 Water... . 20-68 .. 24-73 Protoxide of copper Sesquioxide of iron Lime .. 100-00 100-00 No. 1. From Lake Superior, by Eammelsberg. No. 2. From Chili, by Kittredge. 11. Atacamite. It is esentially a hydrated oxy chloride of copper, of the formula CuCl-j-3CuO, combined with different proportions of water. It occurs in Chili, and other parts of the West Coast of South America. The following analyses of crystallized mineral from Copiapo were made by Field : 1 2. Chlorine 11-94 15-01 Protoxide of copper 70'74 70'48 Water .. 17-79 18-00 Rammelsberg, op. cit., p. 89. ' Ibid., op. cit., p. 551 COPPER-ORES OF CORNWALL AND DEVON. 313 These results lead to the formula 2 (CuCl+3 CuO) + 9 HO. A con- siderable quantity of this interesting mineral has been imported into Swansea, where I have seen it in the ore-yards. I am indebted to my valued friend and colleague, Mr. Warington W. Smyth, Lecturer on Mining at the Government School of Mines, for the following notice of the copper-mines of Devon and Cornwall. Mr. Smyth's position as Mining Inspector on behalf of the Duchy of Cornwall gives him peculiar facilities for obtaining accurate informa- tion on this subject. Copper-ores of Cornwall and Devon. Although a certain character may distinguish the ores of particular mines, it is subject to vary with the diiferent depths at which productive courses of ore may occur ; and at most of the mines considerable discrepancy will be found among the various " parcels," according to their being produced from various lodes within the same sett, or at various depths from the surface. Among the most persistent characters may be cited the association of the copper glance (disulphide of copper) in the mines of the St. Just and St. Ives district (Botallack, Levant, Pendean, St. Tves Consols) with haematitic iron-ores, occasionally crystallized as specular iron, with jasper, and more rarely with bismuth, &c. The same mixture of sesquioxide of iron with this rich ore of copper is very noticeable again in the valuable group of mines (equally important for tin) which lie on the flanks of the granite between Camborne and Kedruth, Dolcoath, Tincroft, Cook's Kitchen, and Carn Brea. From some of the lodes producing the copper-glance, copper-pyrites is altogether absent ; in some cases the latter ore occurs at greater depth than the former. Sometimes at the same depth one lode will contain only the one ore, a neighbouring vein only the other, as in the neighbour lodes of Botallack Crowns lode and \Vheal Cock lode, of the old Levant lode and the North lode. Iron-pyrites and quartz are the most abundant concomitants of the copper-ores generally, the former especially so in some of the coarse and often large lodes, the character of which can often be inferred from the price obtained at the ticketings. Carbonate of lime is rare, and generally confined to crystals lining a few drusy cavities : the same may be said of barytes, hitherto found only in the United Mines and a few mines in the Liskeard district. Chlorite (peacli) is a constituent of most of the lodes in one part or other ; but in some instances, as in the Par Consols district, appears in great abundance, and very commonly as the accompaniment of a copper- pyrites of unusually rich colour and high per centage. With the chlorite, as well as with some of the previously cited asso- ciates of copper ore, tin-stone very often occurs, and sometimes so mingled with the copper-ores, both pyrites and sulphide, as to render the separation by picking and dressing very difficult. Fluor-spar is an important constituent of the matrix in several of the groups of mines, especially some of those placed in or very near the granite, as in the South Frances, Bassett and Buller group, Kelly Bray, Gunnislake and Bedford United, &c. 314 FURNACES EMPLOYED. CALCINER. Carbonate of iron is a frequent companion of copper-pyrites. It may be specially noticed in the Devon Consols and Tavistock group of mines, and in a variety of crystalline forms at Fowey Consols. Among the rarer accompaniments may be mentioned wolfram, isolated crystalline portions of which may be seen in the ores of St. Day, United, Holmbush, Kingston Down, and Gunnislake. Galena and zinc-blende, although not of unfrequent occurrence in moderate quantity in the copper lodes, are not very characteristic of the ores sold from particular mines. The grouping of the ores is often remarkably similar in lodes placed under the same conditions, as an example of which we may take a beautiful association of rich and finely-coloured ores in the lodes of Phoenix and of South Frances, bolh working at considerable depth in granite, and yielding in some parts good copper-pyrites ; in others, and the more gossany portions malachite, chrysocolla, cuprite, and the delicate crimson tufts of chalcotrichite which relieve the green and blue tints of the former species. THE WELSH PROCESS OF COPPER-SMELTIXG. This method consists of not less, and generally more, than six dis- tinct operations. Eeverberatory furnaces are exclusively employed ; and of these there are only two kinds calciners and melting furnaces. The fuel consists of a mixture of binding and free-burning coal, which is burned on a bed of clinker in the manner previously described. The process is varied somewhat in different establishments ; but the modifications are not considerable. I shall first describe the process, as it may be conducted with the smallest number of operations, and as I saw it practised in 1848 at the Hafod Works, near Swansea. I have much pleasure in publicly acknowledging my obligation to the late Mr. Vivian for granting me free access to these works ; and to Mr. William Morgan, the Manager, for much valuable information which, with the consent of Mr. Vivian, he so willingly afforded me. FURNACES EMPLOYED. Cakiner. I am indebted to a firm near Swansea for the drawings from which the annexed engravings are taken. Some Side elevation. details of minor importance have been suppressed ; and in the vertical sections the parts in view beyond the lines of section are not indicated. FURNACES EMPLOYED. CALCINER. IJIT. But all the details which furnace builders usually consider necessary are represented. The interior of the furnace, which is exposed to the highest temperature, is built of fire-brick, while the exterior and parts under the bed are built of common brick. The fire-place is at one end ; a is the ash-pit, above which are three transverse wrought-iron bars, as shown in section, fig. 78 : on the two uppermost the bars of Fig. 7* Vertical section on the line A B, fig. 81 (see p. 316). the grate rest, and the lower one serves as a support for the long crowbar employed in " breaking the grate ;" /, the fire-bridge; g, an arch extending across the furnace from one side to the other, and intended to protect the ore on the bed underneath nearest the fire- bridge from exposure to too high a temperature. There are four rec- tangular openings, A, A, &c. in the roof, through which the ore is allowed to fall into the furnace from the hoppers or bins, ra, m, &c., supported by a framework of wrought iron ; each opening is closed by placing a large fire-brick over it. Each hopper may be made of four cast-iron plates tapering downwards, and firmly braced together. The bed is flat ; on each side are square openings, , i, &c., through which the ore after calcination is transferred to chambers or vaults underneath ; during the process of calcination these openings are kept covered with fire-bricks. There are four doors n, n, &c. ; on each side of the furnace in front of the openings i, i, &c., respectively. In fig. 79 are shown tri- Fig. 79. Horizontal section, showing plan of the bed. angular projections of brick-work, of which the object is to prevent the accumulation of ore on the bed of the furnace between the side doors, where it could not be reached by a rake, or " rabble," except at great inconvenience from the opposite side. At the end opposite the fire- place is a flue z, communicating with a large flue k, which is supported on iron plates and communicates with a high stack : several furnaces 316 FURNACES EMPLOYED. CALCINER. are connected with the flue k. Below the bed are four arched brick- chambers, 6, 6, c. ; which, above, communicate with the openings , t, &c., by the channels ?, ?, as shown in fig. 82 ; and, below, with the horizontal flues d, d, &c. by means of the openings c, c, &c. These flues are connected with a large flue, e, e, leading into the stack : by this arrangement any sulphureous vapour which may continue to escape from the ore, after it has been allowed to fall into the chambers 5, b, &c., passes into the flue e, e, and is thereby prevented from incom- moding the workmen. Fig. 80 shows the manner of placing the M- * ^mi U4 i-W Fig. 80. Horizontal section on the line E F, fig. 11 (p. 314). wrought-iron cramps, into which fit the lower ends of the cast-iron standards (see fig. 77) ; the upper ends of these standards are connected ...JS&.....B. Fig. 81. Plan of foundation. by wrought-iron tie-bars extending above the roof of the furnace, from side to side and from end to end. Some of the cramps pass quite Fig. 82. Vertical section on the line GH, fig. rs. Fig. 83. End elevation of the fire-place. Fig. 84. End elevation near the stack. through the furnace, and others which are short only a little way into the brick-work : the latter should be curved at the ends, directed inwards. Each opening n, n, &c., is formed of a cast-iron door frame, FURNACES EMPLOYED. CALCINER. 317 having a sheet-iron door fixed on hinges at the top, so that it may open outwards and upwards. The clinker grate is used in this furnace. This calciner is much larger than usual, and is charged with about seven tons of ore at a time. The annexed engravings represent a calciner of ordinary dimensions, and, I believe, of the best construction. The drawings of this furnace Fig. 35. Side elevation. were kindly prepared by Mr. John Keates, expressly for this work. The description of the large calciner which has just been given will Fig. se. Vertical section on the line A B, fig. 87. in great measure apply to this, so that only a short additional explana- tion is necessary. There is a channel o, o, extending across the furnace Fig. 87. Horizontal section on the line G H, fig. 85. 318 FURNACES EMPLOYED. MELTING FURNACE. through the lire-bridge and open at each end ; from this channel proceed three passages, i, through which the external air may enter the furnace. Above these passages is the " curtain-arch," h, of which the object has been stated in the descrip- tion of the large calciner. By this construction, which was patented in 1812 by William Evetts Sheffield, 8 the fire-bridge is not only cooled, but air is admitted where it is re- Fig, ss. Vertical section on the line E F, fig. 86. quired in the process of calcination. The bosses, or projections of brick- work, between the side-doors are rounded, and not triangular. There are two openings in the roof, which communicate with one large hopper, or bin, g. The ligbt sectional shading of the internal brick. Fig. 89. Vertical section on the line C D, fig. 85. Fig. 90. End elevation of fire-place. work represents fire-brick. The mode of bracing the furnace together is clearly shown- in figs. 77 and 78. This calciner is charged with about three tons of ore at a time. Melting furnace. For the drawings from which the annexed engravings were made, I am indebted to the same firm who supplied me with those of the large calciner : a, a, side-walls of the fire-place ; ?>, 6, walls, which are carried up vertically, and enclosed above by the arch o ; an arched chamber, ??, is thus formed, which extends from the back of the ash-pit, m, to the opposite end of the furnace, where it is closed by a vertical wall. Upon the top of the arch is a flat plat- form of brick-work, #, . Metal calciners. The raw or granulated naetal from No. 2 is cal- cined during from 24 to 30 hours, stirring every second hour. 4. Coarse-metal furnaces. The charge is, calcined metal from No. 3 20 cwts., fine-metal and roaster-slags 4 cwts., and carbonates 3 cwts. It is melted in 6 hours, when, after skimming the slag, the melted metal is tapped into sand. 5. Fine-metal furnaces. The charge is, coarse metal from No. 4 24 cwts. and refinery slag 2 cwts. It is melted in 6 hours and tapped into sand. 6. Roasting. The charge of fine-metal from No. 5 should be suffi- cient to yield 2 tons of copper. The heat is raised to the melting- point of the metal ; the air (teasing) holes are now opened and the metal gradually melted down in from 6 to 8 hours ; it is kept in a melted state during about 1 2 hours and then skimmed. If the fine- metal, now called for the first time regule, is all reduced to copper, the heat is raised, the metal well melted, and tapped out into sand as pimple or blister-copper as required. The whole operation of roasting requires 24 hours. 7. Refining. The charge is from 5 to 7 tons of coarse copper (i. e. impure metallic copper). It is melted and air admitted at the side door. Black blisters rise to the surface of the metal and burst. The copper is well rabbled all the time. If the charge is of pimple-metal, a pole is put in to agitate and expose it to the action of the air ; when the copper ceases to " work," it is allowed to set ; it is afterwards again melted and skimmed. The refiner now takes the charge, and the poling commences. AVhen necessary as he judges from the assay which he takes out from 16 to 20 Ibs. of lead are thrown in and well stirred with the copper, and the door is put up, after which the charge is skimmed and covered with charcoal or stone-coal (an- thracite). Some of the copper or a " trial " is now beaten out, and, if the result is satisfactory, the lading begins, during which the re- finer constantly takes out " assays," and regulates the pitch-copper accordingly by throwing in a piece of pole or admitting air till the charge is all laded out. The whole process of smelting lasts from 70 to 96 hours. At these works from 13 to 18 tons of coal, which now (1859) costs five shillings a ton, are required to make one ton of copper, and about the half of this quantity is consumed in the first and second operations of calcining and melting. A mixture of three parts by weight of free- burning and one of binding coal is employed. Second modification, as practised in 1859. The mixture of ores em- ployed contains, on an average, 9 per cent, of copper, as determined by the Cornish method of assaying. 1. Calcination. The large calciner, of which engravings are given (pp. 314-16), is employed; it holds 7 tons of ore; the process lasts from 12 to 24 hours. 2. Ore furnace. The charge is calcined ore from No. 1, 22 cwts. 328 MODIFICATIONS OF WELSH PROCESS OF COPPER-SMELTING. (1 cwt. = 112 Ibs.) and 6 of metal or sharp slag. The products are metal, which is granulated, and ore-furnace slag, which is thrown away. 3. The granulated coarse-metal from No. 2 is calcined during from 15 to 18 hours. 4. The charge is calcined granulated coarse-metal from No. 3,45 cwts., with 6, 9, or 12 cwts. of slag from No. 6, and roaster-slag or native carbonates of copper, according to circumstances. The products are fine-metal, which is granulated, and slag. 5. The granulated fine-metal from No. 4 is calcined during 18 hours. 6. The charge is calcined granulated fine-metal from No. 5, 50 cwts. and refinery slag only from 3 to 6 cwts. The products are -fine-metal (still called only fine-metal) and slag. The metal obtained in this operation may occur in the state of blue-metal or pimple-metal. The Welsh name for the last metal is crych, which means rough ; the surface of the metal being rough from pimple-like excrescences. When the metal is still further advanced beyond the state of pimple-metal, it is called close regule, which almost always contains metallic copper. 7. Roasting. The charge is -fine-metal from No. 6, 2 tons. Tt is heated at first rapidly and then melted down, so as to require 4 hours for complete fusion, after which the slag is skimmed off through the front door the opening near the stack being commonly called the front of the furnace. The side door is then opened in order to admit the air freely ; this opening into the furnace is near the fire-bridge on one side. The metal is kept in a melted state during 18 hours. It is skimmed twice, or oftener, according to circumstances, during the process of roasting. Towards the latter part of the operation, the heat of the furnace remaining constant, the metal becomes what is termed " dead," i. e. more or less set or solidified at the surface. In propor- tion as the impurities are removed, a higher temperature is required to maintain the state of fusion. The heat is finally increased by closing the side door so as perfectly to fuse the contents of the furnace. The product is blister-copper, which is tapped oft'. 8. Refining. The charge is from 6 to 7 tons of blister-copper. The copper is kept melted during 18 hours, and is skimmed about three times at intervals ; it receives one good skimming after it is thoroughly melted down, and another before throwing on the " fluxing" coal pre- paratory to poling. Free-burning coal is used in this process, and is left on the surface of the metal during the whole time, including that of lading. Generally at these works the fuel employed consists of two parts by weight of free-burning and one of binding coal. With some kinds of coal equal weights are used. Third modification. 1. Melting. The mixture of ores employed contains, on an average, 9 per cent, of copper. It is melted without previous calcination. The charge is, mixed ores 22 cwts. and sharp slag 4 cwts. The products are, ore-furnace slag, which is thrown away, and metal, which is first granulated and then all passed between rolls, by which means it is reduced to small particles. It is said to contain as much as 38 per cent, of copper. PROCESS OF MAKING "BEST SELECTED" COPPER. 329 2. Calcination. The charge is 5 tons of crushed granulated coarse- metal from Xo. 1, and the calcination is continued during 24 hours. 3. Second melting. The charge is product of No. 2, 36 cwts., roaster and refinery slags, and native carbonates of copper, together 3 cwts. The products are blue- metal and sharp-slag. 4. The blue-metal, from No. 3 is roasted to pimple-metal ; the charge is 4 tons. The products are, this metal and coarse roaster-slag. 5. The pimple-metal from No. 4 is roasted to pimple-copper ; the charge is 5 tons. The products are, this copper and roaster-slag. 6. Refining. The charge is about 8 tons. In the preparation of tough-cake copper about 35 Ibs. of lead are added. The ore is raised by a hydraulic lift and conveyed in waggons direct to the ore-furnaces on a railway, which is constructed on the outside round these furnaces and on a level with their tops. By this arrangement labour is much economised. % The process of making " best selected" copper. I am indebted for the fol- lowing historical notice of the introduction of this process to my friend Mr. Keates. " Best selected " is a comparatively modern term, Best being the old term. The introduction of the manufacture of brass on a large scale into this country does not date much further back than the year 1680, and the manufacturers were not long in discovering that copper taken indiscriminately as it occurred in the market frequently produced brass quite unfit for manufacturing into battery, sheets, and wire, and they rightly attributed this to its impurity. The English copper generally in use at the beginning of the 18th century was derived from Cornish ores, which were then, to a greater extent than at pre- sent, mixed with tin ; and it is most creditable to the sagacity and practical skill of the smelters of that day that they devised a mode of remedying the evil which, in effect, has not been improved upon by their successors. The following " Directions for taking out the second or common copper so as to make best copper for brass " date about the year 1743, and were in manuscript, but no doubt had been practised years before : " In the first place you must calcine the ore for a certain time, so that, when smelted, the metal from it when broken appears of a brownish colour tinged with blue ; in general ten or twelve hours will be sufficient. When you smelt the ore if you find it melts very liquid I would advise to tap out the contents instead of skimming, and you will find the metal in one, two, or three of the first pigs, according to the quantity of it ; which metal must be drawn aside and a quantity of cold water thrown upon it until it begins to crack and fall to .pieces, and in this manner you must proceed until you have got metal enough for the next operations. But if the ore does not melt very thin, then you must skim off the slag and tap out the metal, and quench it with water as before directed. You must always endeavour to keep your metal in the afore-mentioned pitch, so that, when the water is thrown upon it while hot, it will fall to pieces like a lump of quick-lime, and not require to be buckered to pieces. You must now take your metal 330 PROCESS OF MAKING "BEST SELECTED" COPPER. to the calciner and calcine it for such a length of time as, when it is melted, will give you a little bell-metal in the runner-pig. Twelve or fourteen hours will generally do to calcine. A\ hen the metal is cal- cined, take enough of it for a charge for the metal, or regule, furnace, to melt in four or five hours, and add to it a box or two of refinery-slag pounded down small, and also two or three boxes of good thin slag and two or three shovels of cokes or cinders. "When this mixture is pro- perly melted, tap it all out of the fui'nace together and the metal will be found in five or six of the first pigs, and at the bottom of the runner- pig will be found a quantity of whitish-looking metal called bell- metal, and this must be separated from the metal, or regule. The regule is then taken to the roaster-furnace and roasted such a length of time that, when tapped out, there shall be found at the bottom of the pigs four or five hundred- weight of coarse and common copper, which is separated and kept for inferior purposes. The regule, which now begins to get hollow and spongy, is kept to make the best copper, which it does when roasted." The term "bucker" means to pound or bruise with hammers into small pieces. " Runner "-pig is the pig or mould in the bed of sand into which the metal first runs from the tap-hole. Modifications of the process of making "best selected" copper at different works in 1859. 1. This method is practised at the works where the first modification (p. 326) is carried on. About two tons of fine-metal (from the third fusion) are melted down, and roasted during a certain time according to the judgment of the furnace-man. When ready the contents are tapped into sand-beds, or moulds (made as described at p. 324) ; from five to seven pigs of reduced impure copper will be found in the moulds nearest the tap-hole; and from the surface of these pigs the regule is stripped off immediately on its setting. The total number of pigs may be about eighteen or twenty, according to the size of the sand beds. About one-fourth of the copper is reduced in this first melting in making best-selected copper. The regule thus obtained is again treated in the manner just described, when about the same pro- portion of copper will be abstracted as impure, so that in the two meltings about half of the total quantity of copper is reduced. In this reduced copper, termed ** bottoms," certain impurities, especially tin, will be found concentrated. That produced in the second melting is roasted to blister-copper, which is refined in the usual way. The usual proportions which should be obtained when common ores are treated are about eleven tons of best-selected copper and nine tons of bottoms. These bottoms are cast into small rectangular tile-shaped pieces, which are known in the market as "tile-copper;" or, accord- ing to the necessities and opportunities of the smelter, they may be cast into cake-copper. 2. At the works where the large calciner was used (second modifi- cation, p. 327), I obtained the following information : The selecting process follows No. 6. The product of No. 6 is roasted, so that about half the copper which it contains may be reduced. The residual COPPER-SMELTING IN CHILI. 331 metal is termed regule, lest regule, or spongy regule. It is afterwards roasted and refined by itself. 3. This method is adopted at the works when the third modification (p. 328) is carried on. Best-selected copper is prepared from pimple- metal. \Vhen this metal is tapped into sand-beds in the usual way, eight or ten pigs nearest the tap-hole are put aside for tough-cake copper, and the remaining pigs, about fourteen in number, are reserved for the selecting process. The pigs put aside contain metallic copper in strings ; and at these works the roasting is not carried so far as to cause a distinct separation between the bottoms and regains. The pigs used for selecting are in the state of dose, not spongy, regulus. The furnace is charged with about five tons of pimple-metal pigs ; the furnace-doors are then closed until the metal becomes red-hot, but not melted. The air is now freely admitted into the furnace through holes, termed "port-holes," of which there is one on each side near the fire-bridge. The temperature is so regulated that the pigs may bo sweated down in six hours ; the port-holes are now closed, and the metal becomes thoroughly melted in the course of three hours, when it is skimmed. One of the port-holes is opened, and the charging-door at the side partially. The regulus is thus left until it is converted into rough-copper. The side-door and port-hole are now closed until the contents of the furnace are well melted ; after which the tap-hole is opened, and the metal obtained sent to the refinery. As large quantities of concentrated regulus as well as cake copper have been imported from South America to this country, the following description of the method of smelting which was practised by the Mexican and South American Company may be interesting. I received it from the furnace-manager of the company, Mr. M'Auliffe, who was for some time a student in the Metallurgical Laboratory of the School of Mines. Copper-smelting in Chili. The smelting was effected in reverberatory furnaces with coal from this country. 1. Fusion for regulus. One charge consists of 70 quintals (over 3 tons English), and is composed as follows : Average Ore. Locality. Quintals. per centage of copper. Carbonates and oxychloride (green and dark\ n , , brown) / ^ aldera Silicate Tougoy 8 10 Iron fluxes from various parts of Coquimbo 14 8 Limestone 4 3 Carbonates and oxychloride ; hard to be) fused; containing a good deal of " Tofo ''' 2 8 (chiefly carbonate of lime) ) Sulphides (blue) Tougoy 6 20 (yellow) Various places 6 8 (yellow) Totorallillo 16 8 lloasterslag 2 9 Total quintals 70 Each furnace smelts four charges in 24 hours. The regulus con- tains about 60 per cent, of copper ; and the slag, which is " sharp " and brittle, is said rarely to contain more than 1 per cent, of copper. 332 ON THE RE-ACTIONS WHICH OCCUR IN THE 2. Roasting for spongy regulus. A charge of 4 tons is roasted during about 8 hours ; the time from charge to charge, inclusive of charging, tapping, &c., is 10 hours. Out of 20 pigs of the spongy regulus (metal) about 6 or 8 have "bottoms." The metal is not allowed to become too spongy, as in that state it would become mixed with too large a quantity of sand, which it is stated would retard the next roasting. The slag contains 9 per cent, of copper. 3. Roasting for blister-copper. The charge consists of sufficient spongy regulus and " bottoms " to yield from 4 to 5 tons of blister-copper. The time required to work off the charge is from 16 to 18 hours. This operation of roasting is conducted as follows : The charge is first allowed to sweat down, which requires about 6 hours, when the air- holes are luted and the temperature raised, so that the whole charge may be perfectly melted in 1 hour. The front door is now taken down and any slag that may have accumulated is skimmed off, after which the charge is allowed to set by opening the side door and air- holes. As soon as it is quite set or before, if the copper can easily be seen by striking back the " rigole " (regulus) floating on the surface of the bath, with a skimming rabble the side door is luted and the temperature gradually increased until the whole charge is in a state of fusion. During this period the two air-holes are left open, but too much air must not be admitted, as the charge would be thereby pre- vented from melting. If the furnace has had proper attention, and the charge is " working " (7. e. appears to boil), it continues in this state from 30 to 40 minutes. The "working" ceases first in those parts near the air-holes, and soon afterwards in every part of the furnace, the surface becoming covered with a thick yellow coat called " cream," from which small blisters about the size of a pea are thrown up. The Ulster-copper is now tapped into sand moulds. After the second fusion, or before the charge begins to work, any slag covering the face of the bath should be skimmed off. Three furnaces are used. In an establishment of 9 furnaces, smelting-furnaces will keep the 2 "roasters" continually going ; but this depends on the per centage of copper in ores used. ON THE RE-ACTIONS WHICH OCCUR IN THE WELSH PROCESS OF COPPER-SMELTING. Le Play, formerly Professor of Metallurgy at the Ecole des Mines at Paris, has published an elaborate description of the process, and the results of an analytical examination of the products which are formed in each operation. 9 His information was chiefly obtained at the Hafod Works, where, I am informed, he would frequently remain at the furnaces from morn- ing till night. Mr. James Napier subsequently communicated to the * Philosophical Magazine ' a series of papers on this process of copper- smelting. 1 He resided at Swansea, and was engaged at the Loughor 9 Description des Precedes Metallur- giques employes dans lePays de Galles pour la Fabrication du Cuivre. Paris, 1848, pp. 496. 1 Fourth Series, 1852, vol. 4, pp. 45. 192, 2G2, 345, 453. Vol. 5, 1853, pp. 30, 175, 345, 480. WELSH PROCESS OF COPPER-SMELTING. 333 or Spitty Copper-works, when I visited them in 1848. In the follow- ing pages I shall freely avail myself of the writings of both the authors above-mentioned. Calcination. Atmospheric oxygen, aided by heat, is the essential agent in this operation. In order to understand the re-actions which occur, the composition of the ore before, and after, calcination should be ascertained ; but, according to Le Play, the changes effected in the chemical composition of the ore by calcination cannot be completely revealed by comparative chemical analyses of the raw and calcined ores ; because, owing to the mechanical mixtures of several sulphides and oxides in different degrees of sulphuration and oxidation, the data obtained by such analyses do not suffice for the calculation of the exact proximate composition of the ore. That this problem maybe difficult of solution from the cause assigned by Le Play, is probable ; but that it should be incapable of solution, even by an experienced analyst, may fairly be questioned. However, Le Play believes he is able to explain the re-actions which occur in the process in question, so as very nearly to approximate to the truth. Some of Le Play's results are presented in the following table, which is extracted from his work : Composition of the Raw Ore. Composition of the Calcined Ore. Dioxide of copper 0*4 5'4 Copper-pyrites 22'7 11-2 Iron-pyrites (FeS 2 ) 22 4 Sesquisulphide of iron (Fe 2 S 3 )... 11'2 Various sulphides 1-0 0'6 Sesquioxide of iron O'G 11*7 Various oxides 0'3 0-6 Silica 34 3 84-3 Earthy bases 2-0 2-0 Water and carbonic acid in the state) n K of combination . . , . . f * 5 Sulphuric acid combined 1 1 84-2 78-1 Atmospheric oxygen consumed by) this amount of ore in the process) 15 '8 of calcination . Evolved . 5 as gus ( Sulphurous acid'.'.'.'.'. . -21-4 100-0 100-0 The loss of weight which occurred during calcination was found to be 7 '2 per cent., and the loss of sulphur 51-9 per cent, of the total in the ore. The chemical changes which occur when di sulphide of copper is heated with free access of air have been previously considered. When copper-pyrites (Cu 9 S+Fe 2 S 3 ) in a finely-divided state is heated under the same conditions until no further evolution of sulphur in any state takes place, even when the temperature is raised to bright red- ness, the product will consist of a mixture of protoxide of copper and ses(|uioxide of iron. Sulphate of copper is formed and afterwards decomposed, just as in the calcination of disulphide of copper. A sulphide of iron may, in like manner, be converted into sesquioxide : sulphate of protoxide is formed and decomposed with the formation of sulphate of sesquioxide, from which the sulphuric acid may afterwards be wholly expelled. Le Play estimates the whole of the oxide of iron 334 ON THE RE-ACTIONS WHICH OCCUR IN THE in the calcined ore as sesquioxide ; but if I mistake not, some mag- netic oxide of iron at least will always exist in the calcined ore, when any sensible amount of the sulphides of copper and iron is left, as is invariably the case in the calcined ore of copper-works. The calcined ore which I procured from the Hafod Works contained a considerable quantity of black matter, capable of being separated by the magnet ; small lumps were picked out of this ore, consisting of minute, loosely-aggregated, brilliant black crystals, which under a lens were found to present triangular faces, were attractable by the magnet, and dissolved without effervescence in hydrochloric acid, pro- ducing a solution containing proto- and sesqui- chloride of iron : so that the crystals were evidently magnetic oxide of iron. The presence of a sulphide of iron promotes the oxidation of disulphide of copper during calcination ; because the former sulphide is more easily oxidized than the latter, and the resulting sulphate of protoxide of iron may, by its subsequent decomposition, produce the same kind of oxidizing action as is caused by the decomposition of sulphate of copper in the manner described at p. 249. When iron-pyrites is present the same re-actions take place, the only difference being that part of the sulphur burns with a blue flame, like free sulphur, and evolves a considerable amount of heat. When iron-pyrites containing copper-pyrites is roasted under certain conditions, not in a state of fine division, but in lumps, curious phenomena are observed, which will be particularly explained here- after. It will be remarked, sulphur is only partially expelled : it escapes chiefly in the state of sulphurous acid, which, in certain directions ol the wind, may be tasted in every house in Swansea; and it is also evolved to a certain extent in the state of sulphuric acid. Le Play detected the presence of this acid in the smoke of the ore-calciner at all periods of the calcination, by keeping moistened tow, which had been carefully washed with distilled water, in the smoke for some time ; the tow was afterwards digested with water. When chloride of barium was added to the solution thus obtained, copious turbidity followed, which was .not removed by hydrochloric acid. In 1822 Mr. Faraday and the late Mr. Kichard Phillips found sul- phuric as well as sulphurous acid in water which had been well ex- posed to contact with the smoke from the ore-calciners 'at the Hafod Works ; and yet in their analysis of this smoke, before it came in contact with water, there is no mention of the presence of sulphuric acid. 2 Le Play makes the singular remark " that the prevailing taste of the gas in the Welsh smelting- works is by no means (nullement) that of sulphurous acid ; but is precisely that of those thick vapours which disturb the transparency of the air in laboratories where sul- 2 Proceedings of the Subscribers to the Fund for obviating the Inconvenience arising from the Smoke produced by Smelting Copper Ores; Report of the Judges who decided on the Merits of the Trials submitted to their consider- ation; and Statement and Plan explana- tory of the Experiments made at the Hafod Works; with an Account of the Process of Smelting Copper, etc., Svo. pp. 95. Swansea, 182H. WELSH PROCESS OF COPPER-SMELTING. 335 phuric acid is evaporated, or sulphates are decomposed." 3 However, according to my experience, the prevailing taste of the gas of these works is certainly that of sulphurous acid ; but de gustibus, &c. It is, I believe, a common observation at Swansea and in the neighbourhood that the copper- smoke is rendered much more opaque by rain ; and this may, probably, be due to the fact that the sulphuric acid, which is produced by the decomposition of the sulphates of iron and copper, passes into the atmosphere, partially at least, in the state of anhydrous acid, which, as is well known, causes dense white fumes when exposed to moist air. It is true that the gaseous products of the combustion of the coal and the air which finds its way through the sides or bridge of the furnace, pass over the ore, and both contain aqueous vapour ; yet, from one cause or other, some of the anhydrous sulphuric acid evolved may issue from the top of the stack into the atmospjiere. In special cases there may be other conditions which tend to increase the opacity of the copper-smoke. A dense cloud of this white smoke per- petually hangs over the copper-works of Swansea and the vicinity, and Swansea, ;is seen at Sea near the Miimblos. occasionally beautiful eifects are produced in the landscape when the rays of the sun fall upon it, especially towards evening. In favour- able states of the atmosphere, I have frequently seen it with the utmost distinctness at Lynton, which is situate on the south side of the 1 Bristol Channel, at a distance of twenty-seven miles in a direct line from Swansea. Le Play remarks that the opacity of the sulphureous vapours evolved during calcination of the ore is never greater than at the moment when it is taken out of the furnace ; and at first he concluded that the proportion of sulphates contained in the calcined ore would be very 3 Op. cit., p. 336 ON THE RE-ACTIONS WHICH OCCUR IN THE considerable. However, he found that while the proportion of sul- phuric acid never exceeded 2-2 per cent, in any ore, many ores con- tained not a trace ; and this trace, he believed, existed in combination with lime and magnesia, rather than the metallic bases. The amount of sulphuric acid which may exist in combination with these bases in the calcined ore depends entirely on the degree of heat to which it has been exposed : of this ample proof, if any were required, will be found in the sequel. According to Favre and Silbermann, some anhydrous sulphuric acid is produced by the direct combustion of sulphur in oxygen gas, and an almost imponderable quantity of this acid suffices to render a very large volume of air opaque. 4 But I am not aware that the transparency of the atmosphere is affected by the burning of sulphur either in large or small quantity, which would be the case if anhydrous sulphuric acid were formed, even in minute proportion. Mr. Napier has published the following statements relating to the pro- cess of calcination, which, in my judgment, cannot be accepted. " "We took," writes Mr. Napier, "a charge of Cuba ore and calcined during twelve hours, and tried every hour, gave the following results " (sic). COMPOSITION or THE ORE. Copper j 12- 3 13- 012- 2 12- 2 Iron \ 32-730-024-432' 731-333- Sulphur 3M28-3|23 6'l8-629-224-412-218-120-0 Silica | 24-028-032-028-02B-028-034 100- 199 -3 92-2 91-6 96-9 95-9 94-4 a 5 13-012-213-8 828- 12-612-6'12-513-213-8 12-2 6 30 '6 30- 027- 6 24-3 40- 3 27-0 15-918-817-516-2 021-040-0 832 030-030-833 93-392-6 889-392-6 95-4 Mr. Napier remarks : " when we take into consideration the several amounts of sulphur, we observe what appears very anomalous that there is less sulphur at the end of six hours than after twelve. It may be asked, where the sulphur is gone, whence comes it again? In all our experiments this intermitting action of the sulphur is ex- hibited." 5 He then presents, in rather unintelligible language, an explanation of this alleged action. Now, in another experiment, the results of which are given in the very same paper as that from which the preceding extracts were taken, Mr. Napier found that the sulphur gradually diminished from the commencement to the end of the cal- cination, which was continued during forty-four hours. But we hardly seem to require the aid of experiment to demonstrate the fallacy of the "intermitting action." It is certain that during the entire period of calcination sulphur, especially in the state of sul- 4 Ann. de Chimie et de Phys. 3 s. 1852, 34. p. 445. 5 Phil. Mag. 4 s. 1852, 4. p. 459. GASEOUS PRODUCTS FROM ORE-CALCINER. 337 phurous acid, issues in a continuous current from the furnace ; and as this sulphur must be derived from the ore, except the compara- tively minute and quite insignificant amount which may be evolved from the fuel, it follows necessarily that its proportion in the ore must continually decrease from the beginning to the end of the process. In order that Mr. Napier's results should be of any value, it is essential that the ore operated upon should be absolute^ homo- geneous throughout, and that every portion withdrawn from the furnace for the purpose of analysis should be a perfect sample : in other words, a specimen which, for the time being, correctly represents the average composition of the ore ; but it must obviously be extremely difficult, especially when operating upon large quantities of ore in furnaces, to ensure this indispensable condition ; and it may be proved from the very data which led Mr. Napier to admit " the intermitting action " in question that the successive portions of ore which he extracted from the calcining furnace, and afterwards analysed, could not have been samples. If the specimens successively taken out of the furnace had been samples, the ratio between the fixed constituents, namely, silica, copper, and iron, should obviously be the same in each. But on referring to the table of results above inserted, we find that such was not the case. Let us compare the ratios between the silica and copper: in the ore before calcination it was 100 : 51 ; in 1 hour, 100 : 46 ; in 2 hours, 100 : 38 ; in 4 hours, when the sulphur increased from 18 '6 to 29*2 per cent., 100 : 50; in 6 hours, when the sulphur decreased from 24- 4 to 12 '2 per cent., 100 : 39 ; in 11 hours, 100 : 65 ; and in 12 hours, 100:30. When the mixture of ores smelted contains iron-pyrites, as is gene- rally the case, free sulphur may be volatilized during calcination ; for iron-pyrites, when heated to redness, loses about half its sulphur. Faraday and Phillips found a little sulphur deposited from the smoke of the ore-calciner at Ihe Hafod Works. 6 Mr. Napier remarks that " sulphur will not sublime freely from ores in an atmosphere of sulphur," and alludes to a "law" which "must be attended to in all subliming operations." 7 I have never found any difficulty in subliming free sulphur in a flask, or volatilizing sulphur from iron- pyrites when heated in covered crucibles. Composition of the gaseous products which escape from the ore-calciner. Faraday and Phillips made two analyses of the air from the ore- calcincr flue at the Hafod Works : their results are as follow . 8 1. 2. Sulphurous and carbonic acid gases,! in .c^ absorbable by water I ! Oxygen 8-94 9 66 Nitrogen 8042 81-06 100 00 100-00 These results are interesting, as showing that a considerable amount of free oxygen exists in the gases which issue from the calciner. I am 6 Op. cit, p. 61. 7 Phil. Mag. 4 s. 4. p. 4f2. 8 Op. cit., p. 64. 7; 338 AMOUNT OF SULPHUR ANNUALLY EVOLVED FROM not aware whether the composition of these gases has again been inves- tigated by any competent chemist. Total amount of sulphur annually evolved from the copper-works of Swansea and its vicinity. We are indebted to Le Play for the following calcu- lations : Sulphurous acid forms 21 per cent, of the total weight of the sum of the fixed and volatile products of calcination, and 25 per cent, of the weight of the ore subjected to calcination : the weight of sulphur contained in this gas amounts to 13 per cent, of that of the ore. During the entire process of smelting, the sulphurous acid and sulphur expelled amount, respectively, to 56 and 23 per cent, of the weight of the ore. The total weight of copper-ore smelted in South Wales (some time before 1848) being about 200,000 tons, about 46,000 tons of sulphur were annually volatilized, producing 92,000 tons of sulphurous acid. 9 In the works situate near Swansea, nearly two-thirds of the ore im- ported into South Wales are smelted, so that daily 65,900 cubic metres of sulphurous acid are projected from these works into the atmosphere. This acid being very hurtful to vegetation, not a blade of grass will grow on the neighbouring hills, which are particularly exposed to its influence. The sulphuric acid contained in the copper-smoke is, pro- bably, more injurious than the sulphurous acid, as every drop of rain, in falling through the smoke, becomes a solution of oil of vitriol, which, alighting upon foliage, is rendered more corrosive by subsequent eva- poration of a portion of the water. The value of the sulphur annually dissipated in the atmosphere was estimated by Le Play at 200,000?. Various attempts have been made to lessen the nuisance caused by the copper-smoke, and, at the same time, to turn to account the sul- phurous acid which it contains, by applying it to the manufacture of sulphuric acid ; and I believe large sums of money have been from time to time expended in the execution of schemes proposed with this object. But all such projects have hitherto proved unsuccessful. The proprietors of the Hafod Works forty years ago incurred a direct expenditure exceeding 6,000?., exclusive of other indirect expenses, in attempting to abate the evil of the copper-smoke. The smoke was made to take a tortuous course through long flues and chambers, into which water was injected in numerous fine streams ; and only the uncondensable portion was afterwards allowed to escape into the air from the top of a high stack. That the late Mr. Vivian believed these contrivances would prove successful is shown by the following extract from his pamphlet : " And if we cannot flatter ourselves that we have absolutely, entirely, and effectually got rid of every particle of the matter which has been considered as producing inconvenience, we may at least affirm that we have abated it to a degree beyond the possibility of its producing, as far as our Works are concerned, future cause for complaint ; and we feel confident that the liberality of the inhabitants of Swansea and the immediate vicinitv of the Works 9 Le Play uses the word tonneau, which is equal to 1000 kiL , or about 1 English ton. THE COPPER-WORKS OF SWANSEA AND VICINITY. 339 will do us the justice to believe that we have been stimulated in our endeavours to effect this, much more by a sincere desire on our parts to meet their- wishes, and by a sense of the advantages that would arise to the town and neighbourhood, 1 than by the hope of benefiting ourselves or the fear of any proceedings against us in a court of law." This method, however, of condensing the smoke does not appear to have been attended with the success at first anticipated, if we may judge from the fact that it was soon abandoned, and the Hafod Works still contribute their full share of copper-smoke to the general stock. Quite recently an energetic inhabitant of Swansea has endeavoured to apply the existing law concerning the suppression of smoke to the particular case of copper-smoke ; and I am informed that legal autho- rities consider this kind of smoke as exceptional, possibly because it is trhite, and not black, like ordinary smoke. However this may be, the proprietor of the Patent Fuel Works at Swansea has been compelled at great expense to construct a long flue to the top of the Kilvey Hill, at the base of which the works are situated, in order to carry farther away some dark-coloured, foul-smelling smoke, which is intolerable to the inhabitants of a town who can submit without a murmur to the sulphureous and choking exhalations of the copper-works. 2 Nay, it has, I understand, even been gravely maintained by some persons that copper -smoke is beneficial, if not agreeable, rather than other- wise. This, however, was assuredly not the opinion of the late Member for Swansea, Mr. Vivian, who, in opposition to his own interest, had the honesty to declare that the suppression of the smoke would be advantageous to the town and neighbourhood ; nor does it appear to be the opinion of the existing smelters, if we may judge from the fact that, with scarcely an exception, they have selected residences which the smoke cannot reach. The smoke is an unmis- takeable nuisance ; and the man who pretends that it is not, must either have a peculiar constitution or lie under some strange delusion. 8 That the smoke may be conveyed to a considerable distance with- out injuriously affecting the working of the furnaces, may, I think, be reasonably inferred from the following facts : 1. The experience of the late Mr. Vivian proved that at least several calciners might 1 Op. cit, p. 49. The italics are not in the original. 2 The flue is egg-shaped, 4 ft. 3 in. wide by 4 ft. 9 in. high. It is about 850 yards in length, and terminates in a stack 50 ft. high, of which the top is 400 ft. above the furnaces at the works below. 3 In 1854 Dr. Thomas Williams, of Swansea, published a " Eeport on the Copper-Smoke, its Influence on the Pub- lic Health, and the Industrial Diseases of Copper-men." The following extract is a fair specimen of the style of the author: "The furnace-chimneys of the copper-works, thousands ^?) by number, emit gracefully-gyrating, white, smoky, and fleecy columns, which circlingly and wideningly ascend to the upper regions of the atmosphere, thereat to be lost in the purity of invisible air, or, marrying art to nature, to be mingled with the clouds.." p. 10. The author intimates at p. 4, that in consequence of the direct action of the copper-smoke during twenty years ague has ceased to prevail over certain marshes near Swansea; whereas at p. 12 he intimates that the suppression of malaria is due to the rilling up of a tidal morass by enormous accumulations of slag. Z 2 340 AMOUNT OF SULPHUR ANNUALLY EVOLVED FROM advantageously communicate with a single stack. 2. At the Cwm Avon Copper-Works, 42 furnaces (I believe all, except those of the refinery) are connected with only one large chimney. This chimney runs up the side of the adjoining hill, where, on the top, it rises vertically 40 feet from the ground, forming a stack : its length, from the works to the top, inclusive of the stack, is 1100 yards; and its height, above the level of the works to the top of the stack, is 1100 feet; its internal diameter is 13 feet. The stack is a most conspicuous object in the landscape on all sides, and may be seen at a great distance. The hill on which it is erected is more or less conical ; and the stack not being visible from the low ground of the surrounding country, the smoke appears to issue from the top of the hill, which in certain aspects might be mistaken for an active volcano. In 1859 I ascended the hill, and found the summit for a considerable distance round the stack utterly devoid of vegetation : the plant which sur- vived at the highest elevation was a small eriophorum. 3. At the Llanelly Copper-Works one stack is common to all the calciners. 4. At the Pernbrey Copper- Works all the furnaces, with the exception of two refining furnaces, communicate with a single stack 270 feet high and 31 feet square on the outside at the base. It is built square to the top. Up to the height of 100 feet there are double walls, with a free space between them, into which air is allowed to enter through holes at the bottom : by this means the outer casing of brick-work is kept cool, and preserved completely from the action of the copper- smoke. I was assured at the works that the furnaces worked per- fectly with this arrangement. 5. At the Spitty Works, where I am informed on good authority the process of copper-smelting was suc- cessfully carried on, so far as the furnaces were concerned, there was one large circular stack with which nearly all the furnaces were connected. Now, although high stacks at the copper- works at Swansea might relieve the town from much of the smoke, yet they would doubtless be detrimental to the interests of the neighbouring landed proprietors, whose crops and trees would then be seriously damaged ; the effect of a high stack being merely to deliver the smoke into the atmosphere at a great elevation, and so cause it to travel to a greater distance before descending to the ground. I fear that unless some economical system can be put in practice of extracting the noxious constituents of the smoke without in any degree interfering with the working of the furnaces, Swansea must continue to tolerate the nuisance ; and she will probably console herself with the reflection that it is inseparably con- nected with a trade which has so largely contributed to her develop- ment and prosperity. Under any circumstances, should a method hereafter be discovered of suppressing the nuisance without incon- venience to the smelters, Swansea should in justice be required to contribute a fair share of any expenditure which it may be necessary to incur in putting that method in practice. Mr. Grenfell informed me in 1848 that apparatus had been erected at his works at a cost of 2,000/., under the direction of Schaf heutl, of THE COPPER-WORKS OF SWANSEA AND VICINITY. 341 Munich, with a view to employ the smoke as a source of sulphurous acid in the manufacture of sulphuric acid. But the result was a failure. Yet I am inclined to believe that the nuisance will not eventually prove irremediable ; and that the desired end is most likely to be attained by following up the experiments made at Mr. Grenfell's works. In some localities significant improvements in reverberatory furnaces are already in operation, which consist essen- tially in rendering the draught of the furnace less dependent on the .stack, by causing the air to enter by the direct application of mechanical power; and, if I mistake not, considerable progress has yet to be made in this direction. Whether the principle of thus supplying air is capable of being applied with advantage to copper furnaces can only be decided by experiments. Should it be found practicable, it will then be possible to expose the smoke to conditions much more favourable to the separation of its sulphureous constituents than when the draught of the furnaces is determined by the exhausting power of the stack alone. In the calciners especially, from which the greater part of the smoke is evolved, the velocity of the current of air through them is small as compared with that which passes through the melting furnaces ; so that there would probably be much less difficulty in applying the principle in question to the former than the latter de- scription of furnaces. At the copper-works in the Isle of Anglesea, where pyrites ores are smelted, it has long been the practice to collect a considerable quan- tity of the sulphur evolved during the first process of calcination. Of late soda-makers have largely employed Irish and Spanish iron-pyrites containing a small proportion of copper in the manufac- ture of sulphuric acid ; and they have been accustomed to dispose of the residues of cupriferous oxide of iron to copper smelters. But I have recently been informed by a large firm in the North of England that they propose forthwith to erect furnaces for the extraction of the copper from these residues. It is not improbable that, should this project be found successful in a pecuniary point of view, the combination of copper-smelting with the manufacture of sulphuric acid may become pretty general. The utilization of the sulphur for this purpose would render copper-smelting comparatively innocuous to vegetation, and allow of its establishment in districts where at present it would not be tolerated. Moreover, the profit derived from the sulphur of the ores would tend to counterbalance in some degree the disadvantages arising from the additional cost of freight or inland carriage in the case of works situate at a distance from Swansea or Liverpool. 1'ho Spanish and Irish (Wicklow county) pyrites contain from 1 to 4 per cent, of copper. The calcined residues are smelted with the addition of * th of their weight of sand and J th of poor uncalcined copper-ores. The regulus contains about 40 per cent, of copper. The slag is re- markably crystalline. As the regulus and slag are liable to be much intermixed, the assay of the bulk of the regulus seldom gives more than 12 per cent, of copper. (See note at the end of the article on copper). 342 PRACTICAL CONCLUSION CONCERNING CALCINATION. Important practical conclusion concerning calcination. The sulphur is only very imperfectly expelled, and more is left in the calcined ore than is sufficient to form disulphide of copper with the whole of the copper, and protosulphide of iron with an amount of iron equal in weight to that of the copper. The reader is, for the present, requested to fix his attention upon the four essential constituents of the ore, namely, copper, iron, sulphur, and silica : the changes which any other sulphides or other matters present in the ore may undergo will be separately considered hereafter. 2. Melting of the calcined ore. The products are coarse-metal and ore- furnace slag. External characters of coarse-metal. It is brittle, and may easily be reduced to powder by trituration. The fractured surface of a piece of this metal is n on- crystalline, uneven, more or less granular, gene- rally vesicular, and bronze-like in colour. Composition of coarse-metal. Le Play made the following analysis of a sample, which was carefully prepared from a mixture of 58 specimens of coarse-metal taken at intervals during a fortnight from a furnace in which the nature of the charge was constant. Copper 33- 7 Iron 33'6 i Nickel, cobalt, manganese 1-0 Till : 0-7 Arsenic 0-3 Sulphur 29-2 Slag, mechanically mixed 1-1 99-6 Mr. Napier has published analyses of six specimens of coarse-metal, produced at different times in the ordinary process of smelting : they indicate considerable variation in composition. I have selected the two following, which contained respectively the minimum and maxi- mum of copper. 1. 2. Copper 21-1 39-5 Iron 33-2 36-4 Sulphur 45*5 25-0 99-8 100-9 Le Play proposes the formula 3 Cu 2 S-hFe 2 S 3 +4FeS as representing the composition of the coarse-metal produced in the Welsh method of smelting: the iron in the protosulphide is supposed to be partially replaced by other metals. But copper, iron, and sulphur may exist in very variable proportions in a regulus, which may, notwithstand- ing, be quite homogeneous throughout; and no evidence has been advanced to prove that coarse-metal should be regarded as a definite compound, and riot as a mere mixture of two or more definite com- pounds. It is better that we should feel our ignorance than hastily apply formulae on insufficient grounds to matters so variable in coin- position as a regulus of this kind. External characters of ore-furnace slag. It generally consists of a hard, COMPOSITION OF ORE-FUKNACE SLAG. 343 brittle, compact, opaque, black matrix, in which are imbedded sharply- defined angular pieces of white quartz, and hence it presents a por- phyritic appearance. Its fracture is uneven, and may be here and there more or less vesicular. It frequently contains small shots of coarse-metal, which, on a freshly-broken surface, may be readily dis- tinguished with the naked eye. Composition of the ore-furnace slag. Le Play determined the composi- tion of a sample of this slag prepared from 58 specimens taken under exactly the same conditions as those from which the sample of coarse- metal was obtained. The results of his analysis are as follow : Quartz in admixture 30'5 Silica in combination 30'0 Alumina 2-9 Protoxide of iron 28' 5 Lime 2-0 Magnesia 0*6 Various oxides (of tin, manganese, nickel, and cobalt; 1*4 Fluor 1-0) , Calcium 1-1 Copper 05 Iron 0-9 Sulphur 0-6 2-0 100-0 The oxygen of the combined silica is nearly double that of the bases, so that in constitution this slag approximates to a sesquibasic silicate. According to Le Play this slag never contains the slightest trace either of oxide or dioxide of copper ; and the copper which it contains exists only in the mechanically-diffused coarse-metal. The same author in- forms us, " that his numerous researches on this point led him to dis- cover a fact which had previously escaped the attention of metallur- gists, and which, in his view, constitutes the most general and essen- tial characteristic of the theory of the smelting of sulphuretted ores of copper. This fact, which, without a single exception, is true of all the slags of the seven principal groups of copper-smelting works of Europe, may be enunciated as follows. If we analyze simultaneously the regulus and the slag produced at the same time in smelting sulphur- etted ores of copper, we find that for a given quantity of copper the slag always contains more sulphur than the regulus. With the Welsh slag in particular I have established the following facts. The regulus pro- duced during a fortnight in one of the furnaces (No. 2) contained on the average, copper 34-6 and sulphur 29'8. The slag produced at the same time, in the same furnace, contained on the average, copper 0'5 and sulphur 0'6. The proportion, therefore, of sulphur (for 1 of copper) is In the regulus 0-86 or 100 In the slag 1-20 ,, 140" 4 The proportion of copper has been determined in our metallurgical laboratory in two specimens of ore-furnace slag from the Hafod Works : one contained 0-45 per cent., and the other O61 per cent, of copper. 4 Op. cit, p. 212. 344 COMPOSITION OF ORE-FURNACE SLAG. Le Play supposes that the excess of sulphur exists in combination with iron as protosulphide, which is " dissolved in the slag by the affinity of the silicate of protoxide of iron, forming a compound which might be called sulpho-silicate of iron. . . . The constant presence of the sulpho-silicate of iron in slags from the fusion of sulphuretted ores explains perfectly the absence of the oxides of copper, because the existence of these is incompatible with that of sulphide of iron. The oxide of copper, which at first tends to dissolve by the action of the silica, is reduced by the two elements of the sulphide. The metallic copper formed separates immediately from the slag, either to deposit itself or dissolve in the regulus." 5 Simple inspection of ore-furnace slag suffices to prove that a sensible portion of the small quantity of copper which it contains exists in the state of coarse-metal, and, however probable it may be that the whole exists in that state, yet I do not find that Le Play has proved this to be the case. Moreover, he has neither demonstrated that the portion of sulphur which, in relation to the copper present, exceeds that con- tained in the coarse-metal, is in combination with iron, nor established the fact of the existence of such a compound as sulpho-silicate of iron. Admitting, for the sake of argument, that ore-furnace slag contains no oxide of copper, there would appear to be no difficulty in explaining the fact if we refer to the reaction between oxide of copper and sulphide of iron in the presence of silica, which has been previously described (see p. 254) ; for, supposing that some of the oxide of copper which was present in the calcined ore had combined with silica and passed into the slag, yet, as every part of this slag must have been thoroughly in contact with a large excess of sulphide of iron, as may be certainly inferred from the composition of the coarse-metal, any silicate of copper which it might have contained would have been decomposed with the formation of disulphide of copper and silicate of protoxide of iron. The necessity, therefore, of Le Play's hypothesis concerning sulpho-silicate of iron, under the circumstances, does not seem to be required. Le Play intimates that such a compound as sulpho-silicate of iron would afford but little interest to chemists ; 6 but in this, I am inclined to think, he is mistaken. There is one very interesting mineral helvin, or tetrahedral garnet which has attracted much attention, because it consists of a silicate apparently in combination with a sulphide. It occurs well crystallized, and, according to Rain- melsberg, has the following composition : 7 Sulphur 5 71 Silica 33-13 Glucma 11-46 Protoxide of manganese 49 12 Protoxide of iron . . 4-00 The sulphur is believed -to be in combination, partly with manga- nese and partly with iron, but whether in this mineral there is an 5 Op. cit, p. 213. Op. cit., p. 213. 7 Handbuch der Miueralcliemie, 1860, p. 700. CONCLUDING OBSERVATIONS. 345 actual definite combination of a sulphide with a silicate is not yet proved on indisputable evidence. It is certain that many well-crys- tallized minerals do contain a considerable amount of matter or "dirt," as it has been called which is merely accidentally present, and is not an essential part of their constitution. Beautifully-crystal- lized slags of iron-smelting furnaces, as has been already remarked (p. 23), contain a small quantity of sulphide, which is certainly only mechanically diffused, because minerals identical in crystalline form and chemical composition, except that they contain no sulphide, occur in nature. In the sequel other instances will be given in which sul- phides exist in small quantity in well-crystallized slags, which consist of definite silicates. It may be that it is an error to suppose that in these and other cases the sulphide is present in an uncoinbined state, but at least additional data are necessary to justify a positive con- clusion on the subject of the existence of sulpho-silicates, notwith- standing the opinion of Le Play, that "the important function of the sulpho-silicate of iron throws new light on the metallurgy of copper." Specific yravity of the coarse-metal and ore-furnace slag. Le Play deter- mined the specific gravity of these substances when in the state of fine powder. His results are as follow : Specific gravity of coarse-metal, containing 33-7 per cent, of copper 4-56 Do. slag, containing 1*5 of coarse-metal and 30-5 of quartz 3 '21 Concluding observations. From the preceding analytical data we learn that in the ore-furnace much of the iron and the whole of the so-called earthy matter of the ore are separated as slag, and that the whole of the copper, except the small quantity which escapes in the slag, is concentrated in a regulus, which contains, on an average, about as much copper as pure copper-pyrites. One portion of the quartz which existed in the ore combines with protoxide of iron, derived partly from the ore and partly from the metal slag, which is a silicate of prot- oxide of iron, containing a much larger quantity of this oxide than ore-furnace slay, while the remainder of the quartz is diffused through, arid remains suspended in, the slag, because it has a lower specific gravity than either the regulus or the slag. The oxide of copper which is present in the calcined ore is entirely converted into disul- phide, which passes into the regulus. This conversion is effected through the joint agency of sulphide of iron and silica, as has been previously explained. The oxide of iron in the calcined ore is in the two states of oxidation of protoxide and sesquioxide, and a considerable quantity of magnetic oxide may also be present. The metallic sul- phides in the raw ore contain more sulphur than the coarse-metal Sul- phur must, therefore, be evolved during fusion in the ore-furnace. \\hen any of the oxides of iron are heated in contact with excess of sulphur, sulphurous acid and protosulphide of iron are generated ; but when either magnetic oxide or sesquioxide of iron is heated in contact with silica and a proportion of sulphur only just sufficient to combine with the oxygen in excess beyond what is required to form protoxide of iron, silicate of this oxide is obtained. Hence 346 CALCINATION OF GRANULATED COARSE-METAL. it is easy to understand how the sesquioxide of iron in the calcined ore should be reduced to protoxide. It is observed, that during the fusion sulphurous acid escapes with effervescence from the melted mass. Any copper which may exist in the state of silicate in the metal slag forming part of the charge of the ore-furnace is converted into disulphide and passes into the coarse-metal ; for when silicate of copper is heated with an excess of sulphide of iron, such as is always present in this furnace, silicate of protoxide of iron and disulphide of copper are formed. In like manner may be explained the extraction of copper from old fire-bricks, or other furnace residua impregnated with oxide of copper, when these are introduced into the ore-furnace. 3. Calcination of the granulated coarse-metal. I have not met with a satisfactory analysis of this product. Le Play determined that the sulphur was reduced by calcination from 29'5 to 16'4 per cent, in coarse-metal, consisting of Copper 33-7 Iron 34-2 Various metals 1 '5 Sulphur : 29-5 Slag, mechanically mixed 1 ! 100-0 From the proportion of sulphur above stated he infers that the ap- proximate composition of the product of the calcination of such coarse- metal is as follows : Copper 34-6 Coarse-metal, unchanged 54-9 Iron 35*1 Protoxide of copper 19'5 Various metals 1*5 Sesquioxide of iron 22-5 Sulphur 16*4 Various oxides 2-0 Oxygen 11-3 Slag, mechanically mixed 1 1 Slag, mechanically mixed 1 1 100-0 100-0 Le Play does not appear to have analysed this product, and it is not possible, from the determination of the sulphur alone, to deduce its composition. It may, indeed, be regarded as certain that the composi- tion assigned cannot be correct ; for, so long as a considerable amount of the coarse-metal remains unchanged, the oxides of copper and iron could not have wholly existed in the high degrees of oxidation supposed. According to Mr. Xapier " the following analyses give a fair ave- rage of the results of calcining coarse-metal" Before calcination. After calcination. Copper 32 33 Iron 36 36 Sulphur 25 13 Oxygen 11 Insoluble matter 7 7 100 100 Eough results of this kind will hardly satisfy an analyst of the pre- sent day, any more than the preceding deductions of Le Play. They MELTING OF CALCINED GKANULATED COARSE-METAL. 347 may be correct, or, as is probable, they may not. It is, however, certain, that during calcination a considerable amount of sulphur is evolved in the states both of sulphurous and sulphuric acids. -I. Melting of calcined granulated coarse-metal. The composition of the regains produced in this operation will vary with the proportion of oxide of copper in the matters which are fused in conjunction with the calcined granulated coarse-metal. It is only when a sufficient quantity of ore containing oxide of copper, such as the Australian carbonates and red oxide, is added along with the roaster and refinery slags, that the regulus occurs in the state of white-metal. \Vldte-metal. It is compact and brittle. Its fracture is uneven, -i -similar, and more or less crystalline ; it has a feebly metallic lustre and dark bluish-grey colour. Le Play has given the following analysis of a characteristic speci- men of this metal: Copper ....................................... 77 4 Iron ........................................... 07 Nickel, cobalt, manganese .............. traces Tin, arsenic ................................. O'l Sulphur ...................................... 21-0 Slag and sand, mechanically mixed. . . 3 99-5 It was selected from a part most free from cavities, and had a specific gravity of 5*70. Le Play found that an average sample, pre- pared from numerous varieties of white-metal collected during a fort- night, contained 73*2 per cent, of copper. From the preceding analysis it appears that white-metal approximates closely in composition to disul- phide of copper. The proportion of sulphur, however, after deducting the maximum amount which may be combined with the iron and other metals, is sensibly greater than that in disulphide, but possibly to this extent the analysis may be erroneous. Slag. This slag resembles some of the slags produced in the manu- facture of iron. It is brittle, compact, and occasionally very crystal- line. Its fracture has the following characters : uneven, more or less conchoidal, granular, or distinctly crystalline; here and there, espe- cially towards the upper surface, it presents small rounded cavities, caused evidently by gas; its colour, when freshly fractured, is dark bluish-grey ; in parts it appears somewhat iridescent, the tints ranging from blue-grey to bronze-yellow ; its lustre inclines to metallic ; in places, especially near the button, small round shots of regulus may be observed, which may be readily distinguished from the surrounding slag by their lighter bluish-grey colour and more highly metallic lustre. The following analysis is by Le Play: Silica ................................. 33-8 I Magnesia .............................. 0'3 . -" Dioxide of copper .................. <)! : nixed.... ........... ( Sulphur ... 0-8 Various oxides ..................... 2'] I - Lime ................................. ]4 I 100-0 348 WHITE-METAL. Le Play states, that in what he terms rich slag the dioxide of copper amounts to 2*7 per cent. A specimen of metal-slag which I procured from the Hafod Works contained not more than 1-83 per cent, of copper. White-metal may be regarded as coarse-metal deprived of nearly the whole of its sulphide of iron. The same result would have been directly attained by heating together coarse-metal, oxide of copper, and silica, intimately mixed in such proportions that the oxygen of the oxide of copper should just be sufficient to convert the iron of the coarse-metal into protoxide, and the silica sufficient to combine with the protoxide thus formed and produce an easily-fusible slag. The fol- lowing formulae exactly express the reactions : FeS + Cu 2 + a-SiO 3 = Cu 2 S + FeO,o:SiO 3 2FeS + 2CuO + zSiO 3 = Cu 2 S + S + 2FeO,zSiO 3 . By the calcination of the granulated coarse-metal a certain amount of sulphur has been evolved and replaced by an equivalent proportion of oxygen, or, in other words, a product has been obtained consisting essen- tially of copper, iron, sulphur, and oxygen. Oxide of copper (Cu' 2 and CuO) has also been added by introducing into the furnace roaster and refinery slags and ores containing carbonate of copper. But the calcined granulated coarse-metal, the two slags above-mentioned, and, probably, the ore, contained oxide of iron. The presence of this oxide does not affect the result. The oxide of iron in the slags is already in combina- tion with silica, and continues so ; and if we refer to the analysis of the calcined granulated coarse-metal, we shall find that there remains more sulphur than is sufficient to form disulphide of copper with the copper present and to combine with the excess of oxygen beyond what is required to form protoxide of iron with the iron present, so that there is no difficulty in understanding how any sesquioxide of iron should be reduced to protoxide. We have only, therefore, to fix our attention on the proportions of copper, iron, sulphur, and oxygen, which may exist in the various matters introduced into the furnace, quite irrespective of the manner in which these elements may be combined with each other, in order to have a clear comprehension of the metallurgical reactions which concur in the production of white-metal. The silica required to combine with the protoxide of iron may be derived, not only from the slags added, but also from the ore and materials of which the furnace is made. On the large scale it would not be desirable, even were it practi- cable, to effect an intimate mixture, in the manner supposed, of the various matters composing the charge. Such a course would, on the ground of expense alone, be inexpedient, and would, it is alleged, be disadvantageous in other respects. In the first place it is stated that the bottom of the furnace would be exposed for a considerable time to contact with the oxides of copper and iron, and would, consequently, be much corroded, whereas, by the present system, the unchanged regulus remaining in the calcined granulated coarse-metal fuses rapidly and trickles down on the sand bottom of the furnace, which is thereby protected from such corrosive action ; and, in the second place, me- BLUE-METAL. 349 tallic copper would be liable to be separated, which, in the present stage of the process of smelting, should be avoided, because copper thus reduced would, for reasons to be hereafter explained (see best selected process), be very impure. 8 But this result would also occur in the usual method of conducting the fusion if a considerable excess of oxide of copper should happen to be put into the furnace. Le Tlay defers to the judgment of workmen on questions of this nature to a much greater extent than many experienced smelters in this country would, I think, be disposed to do. In the first stage of this operation a regulus is formed which con- tains much less copper than white-metal, and a considerable time is required completely to liquefy the matter with which its surface is, in a greater or less degree, covered. This matter is, as we have seen, rich in oxide of copper, much of which is combined with silica. A gradual interchange takes place between the elements of this silicate of copper and those of the sulphide of iron in the regulus, whereby the latter becomes enriched with the addition of disulphide of copper, of which the copper is derived from the superincumbent slag and ore, and in a corresponding degree deprived of sulphide of iron. Le Play makes the following remarks : " The oxide of copper especially acts upon the sulphides of iron and copper, producing metallic copper, protoxide of iron, and sulphurous acid. The undecomposed sulphides not being completely saturated with metals dissolve the copper." 9 Now the only sulphides to which he can refer are disulphide of copper and protosulphide of iron, but I am not aware that either separately has the power of dissolving metallic copper, though when combined they may possess that power in a limited degree. It is not to be supposed that the proportions of the various matters which form the charge of the furnace in this operation can always be so nicely adjusted as to produce a regulus so nearly approximating to the composition of disulphide of copper as the specimen of white-metal of which the analysis has been given. Should the oxidized compounds of copper be deficient, as from the exhaustion of the stock of ores con- taining carbonates and oxide of copper, the regulus would contain a considerable quantity of iron, and constitute what, in consequence of its bluish colour, is termed blue-metal. On the other hand, should these compounds be in excess, some copper would be reduced, and a regulus obtained similar in composition, colour, and fracture to white-metal, but when cold presenting on its upper surface numerous small, round, pimple-like excrescences, which have caused it to be termed pimple-metal. White-metal passes insensibly into blue-metal as the pro- portion of iron increases, and between the two there is no line of demarcation. Blue-metal. It is brittle and breaks with an uneven fracture. The colour of its fractured surface is modified by the temperature at which it is produced. When broken hot, but much below a red-heat, it has a fine, deep, purplish-blue colour ; but when broken cold, it has a pur- Le Play, Op. cit., p. 2GO. 9 Op. cit, p. 256. 350 BLUE-METAL SLAG. plish-red bronzy tint, of which it is difficult to convey an accurate notion by description : its lustre strongly inclines to metallic. On examining the freshly-fractured surface of a characteristic specimen, from which the preceding description was taken, under a good lens, it is seen to be studded throughout with bright, metallic, copper-like particles, and the peculiar colour and lustre which it possesses seem to result from the combined effect of these particles and the purplish- blue matrix of regulus in which they are embedded. Le Play has given the following analysis of a sample prepared from a mixture of all the varieties of blue-metal produced during the course of a week in the same furnace : Copper 56'7 Iron 16-3 Nickel, with traces of manganese 1*6 Tin, with traces of arsenic 1 2 Sulphur 23-0 Slag, mechanically mixed 0'5 99-3 Of several analyses of blue-metal published by Mr. Napier, one is nearly the same as this of Le Play. Slag. There is nothing characteristic in the slag which may be formed in conjunction with blue-metal ; it may be regarded as in all respects similar to that of white-metal, of which an analysis has been already presented. It may be expected to contain less copper in the state of oxide in proportion as the regulus is rich in iron. In the preceding analysis the amount of sulphur required to form disulphide of copper and protosulphide of iron with the copper and iron present is 23-64, so that there is a deficiency of 0*64 in addition to that needed for the various metals which must have existed as sul- phides. This deficiency would be readily explained by the fact that metallic copper is generally present in blue-metal. The copper is dif- fused through the mass in minute angular particles, not in the least globular like shot, and, for the most part, invisible without the aid of a good lens ; and where cavities occur, it may be seen protruding into the interior in the form of teeth and delicate hair-like filaments. Le Play has well described these appearances. 1 Now, as Le Play has remarked, it appears singular that metallic copper should occur as the rule in blue-metal and only as the exception in white-metal, notwithstanding blue-metal is necessarily the intermediate stage between coarse-metal and white-metal, and no metallic copper is separated in the furnace until the operation following the fusion in which white-metal is produced. One might be disposed to infer that, if copper existed in a free state in blue-metal, a fortiori it should exist, even in greater proportion, in white-metal. If free copper had been present in the liquid regulus in the furnace, so far as my knowledge extends, there is every reason to suppose that it would have assumed the form of globules, and subsided, in a greater or less degree, on the bottom of 1 Op. cit., p. 272. BLUE-METAL. 351 the furnace. On the other hand it is difficult to understand how the copper, which is seen to be free in a piece of solid regulus, should have existed in a state of combination in the liquid regulus, because the amount of sulphur found by analysis does not suffice to convert the copper into disulphide and the iron into protosulphide. But the follow- ing fact, recorded by Plattner, tends to remove this difficulty. Copper regulus of a bluish-black colour, and consisting of #Cu 2 S+FeS, may, during fusion, take up a small additional quantity of metallic copper, which, 011 the subsequent rapid solidification of the regulus, does not separate ; a grey colour is thereby communicated to the finely-granular fractured surface of the regulus. "When the regulus thus enriched with copper is melted and allowed to cool slowty in a crucible, its frac- tured surface will regain its original bluish-black colour, and present, here and there, small cavities lined with little teeth of xfltetallic copper. According to Plattner this effect is due to the action of protosulphide of iron (FeS), which, during the melting of the regulus, yields a por-- tion of its sulphur to the metallic copper, forming disulphide of copper and disulphide of iron (Fe 2 S). When this disulphide of iron is kept fused at a certain temperature, and afterwards rapidly solidified, it undergoes no change ; but when, on the contrary, it is allowed to cool slowly after fusion, owing to the affinity of disulphide of copper for protosulphide of iron, the disulphide of iron is again converted into protosulphide at the expense of the sulphur of some of the disulphide of copper, with the separation of an equivalent proportion of metallic copper. During the process of cooling the regulus contracts consider- ably, and, as solidification proceeds from without inwards, cavities are formed, into which the liberated copper protrudes. These internal cavities are due less to the escape of gases or vapours than to the cause just assigned. 2 A somewhat analogous play of affinity at different temperatures seems to occur between the oxides of copper and iron when in com- bination with silica. I have a specimen of glass (silicate of soda and lime) which contains both copper and iron, and has a pale green colour. \Yhen this glass is heated to the degree at which it softens or melts, and is afterwards rapidly cooled, its colour is not changed ; when, on the contrary, after having been thus heated it is slowly cooled, or, after having become cold, it is gently reheated, it acquires an intense red colour like that communicated to glass by dioxide of copper. The glass which has become red regains its original green colour by being strongly reheated and afterwards rapidly cooled. This alternation of colour may be effected any number of times. The experiment may be easily made by heating the glass before the blowpipe beyond the point of the flame ; and, as far as 1 have observed, the phenomenon is inde- pendent of any oxidizing or reducing action from without. Dioxide of copper has a much greater colouring power than an equivalent propor- tion of protoxide, for glass of which the surface contains protoxide of copper sufficient merely to communicate a just perceptible greenish- 2 Berg. u. Huttenm. Zeitung, 1855, p. 143. 352 BLUE-METAL. blue tinge becomes intensely red by being heated in a reducing gas. The alternation of colour above described may be susceptible of the following explanation. At a high temperature only protoxide of copper, in association with some protoxide of iron, exists ; but at a low temperature it is reduced to dioxide, with the formation of an equiva- lent proportion of sesquioxide of iron, which has but a feebly colouring power as compared with protoxide of iron. I have made the following experiments -with a view to ascertain the conditions under which the separation of metallic copper occurs in Hue-metal. 1. A characteristic specimen of blue-metal, through the substance of which metallic copper was diffused in fine particles, was melted under charcoal in a very small covered clay crucible at a bright red-heat, aiid the crucible, immediately after its removal from the furnace, was plunged into cold water, so that the melted regulus might be cooled with the utmost rapidity. By the action of the water some sulphu- retted hydrogen was evolved. The regulus had been thoroughly fused. On the surface of the button which had been in contact with the cru- cible numerous circular cavities existed, due clearly to bubbles of gas. The prevailing colour of this surface was dull copper-red, which was produced by extremely minute particles, or, as it were, dust of copper, for the characteristic metallic lustre and colour of copper were instantly developed by drawing the end of a penknife over a portion of the red surface. On holding the regulus obliquely towards the eye, the red deposit of copper presented a velvety aspect. The button was friable, breaking easily in the direction of cracks, which appear to have been caused by the rapid cooling in water. The fracture was uneven, and more or less conchoidal. There were a few tolerably large globular cavities towards the exterior, but the mass was generally compact, except in the central portion, where the fracture was hackly, as though small spaces had been produced by contraction during solidification. The red copper dust was visible to the naked eye to a greater or less extent over the entire fractured surface, but it was most abundant towards the exterior. On looking at the surfaces of some fragments placed at right angles to the line of vision, the colour was dark bluish- grey, with a distinct vitreous lustre ; but on looking at the surfaces of the same fragments placed obliquely before the eye, the copper-red colour became more or less evident. I examined various fragments under a simple microscope, and observed copper-like particles where I could detect none with the naked eye. 2. Another piece of the same blue-metal was melted under charcoal in a very small covered clay crucible at a bright red-heat, and the crucible was left to cool in the air. The regulus was perfectly melted. On the external surface, which had been in contact with the crucible, were numerous small globular cavities, which had evidently been caused by bubbles of gas : many of these, which were disclosed by detaching the adherent particles of crucible, were completely coated internally with filaments of copper, directed towards the centre ; others contained only a few minute projecting teeth of copper ; and P.LUE-METAL. 353 others, again, were free from copper. The outer portion of some of these ca\ 7 ities was formed by the substance of the crucible. The cavities lined with copper were beautiful objects under a simple microscope. The prevailing colour of the whole surface, except the top of the button, was dark bluish-grey, intermixed with red. The red was less intense, and was due to particles of metallic copper, larger and more distinct than in the first experiment. The top of the button was dull, and was covered with small, angular, apparently- crystalline, isolated, and slightly projecting particles of copper. The regulus was very much less friable than that which had been cooled in water. The fracture was tolerably even ; less conchoidal and vitreous than that of the button in Exp. 1 ; its colour was dark purplish-grey with a reddish hue. The tint changes somewhat with the direction in which the surface is seen ; under a lens, or simple microscope, minute angular particles like metallic copper were observed, apparently equally distributed over the whole surface. 3. Another portion of the same blue-meted was meltedMn a covered clay crucible, placed in another crucible and surrounded with anthra- cite powder. Fusion was effected at a bright red-heat, and the whole left to cool in the furnace until the following morning. The button was well melted. The top was concave and dull, and, under a good lens, 110 particles of metallic copper could be detected upon it. The remaining surface of the button, which had been in contact with the crucible, presented nearly the same appearances as in Exp. 2. There were several small globular cavities lined with protruding fibres of copper. The fracture was similar to that of the button in Exp. 2, except that it had a dark bluish-grey colour without the decided red tinge which the other possessed. When the two were compared side by side, the difference of tint was very decided. On examining the fracture under a simple microscope, minute angular particles, like copper, were observed in every part. 4. Bisulphide of copper, prepared by heating best-selected copper and sulphur together and fusing, was intimately mixed by trituratioii with a sulphide of iron 3 containing 29-9 per cent, of sulphur, and prepared by heating thin sheet-iron and sulphur together. The proportions were as 7 to 3. Of this mixture 600 grains were triturated with 120 of the powder of copper (obtained by reducing oxide of copper in a current of hydrogen), and heated to bright redness in a covered clay crucible contained within another covered crucible, the space between the two being entirely filled with anthracite powder so as to cover the top of the inner crucible. The button was well melted. The external surface was studded with globular depressions or cavities, some very small and others as large as an ordinary pin's head, most of which were filled with converging filaments of copper. On the fractured surface small angular particles, or small laminse of metallic copper, appeared throughout ; but the central portion, in which spaces seemed to have been caused by contraction during solidification, consisted 3 It contained 6-46 per cent, less sulphur than true protosulphide of iron. 2 A 354 BLUE-METAL. apparently of a crystallized regulus mixed with so much fibrous copper as to communicate a coppery-red colour to the whole of this portion. Under a simple microscope this mixture of crystals of regulus and fibres of copper was extremely interesting. In no part did 1 detect any shots of copper, not even at the bottom of the button. 5. Exp. 4 was repeated with 500 grains of the mixture tritu- rated with 50 of the powder of copper. The regulus was well melted. The fractured surface was beautifully mottled with coppery-red and bluish-grey. Under a lens or simple microscope, minute, angular, copper-like particles were everywhere observed on the fractured sur- face. The external surface of the button presented similar characters to that of Exp. 4. I did not detect any shots of copper in any part of the button, not even at the bottom. 6. Exp. 4 was repeated with 500 grains of the mixture with- out the addition of metallic copper. The product was well melted. The characters of its outer surface and fracture were in all respects similar to those of the last, and, as far as the eye could enable me to judge, the amount of metallic copper diffused seemed to be quite as great as that in Exp. 5. 7. A mixture consisting of 250 grains of the mixture employed in Exp. 4 and 250 grains of disulphide of copper was heated in a clay crucible enclosed in another containing anthracite powder. The button was well melted. The outer surface of the button, except the top, was studded with small globular concretions of fibres of copper, which appear to have been deposited in the interior of cavities. These copper-lined cavities, or concretions, as I observed in buttons which have been previously described, are much more numerous round the sides of the button. At the bottom of the button were numerous small globular cavities, which were either empty or contained only a few protruding teeth of copper. The fracture was uneven, and in places conchoidal. It was beautifully mottled with fine-bluish purple and rich coppery-bronze, and was everywhere studded with minute, angu- lar, copper-like particles. No shots of metallic copper were found. 8. The mixture used in Exp. 4 was triturated with a large excess of sulphur, and heated simply in a covered clay crucible in a muffle to bright redness, without being protected by charcoal or anthracite powder. The product was well melted. On the surface were nume- rous globular cavities, but in none did I observe any metallic copper. The fracture was uneven ; it had a reddish bronze-like colour, mottled with dark-purplish grey, with metallic lustre ; there was one consi- derable cavity near the centre, which was lined with distinctly crys- tallized dark-grey regulus. The mass seemed to consist of a mixture of two kinds of regulus, distinguished by the colours above mentioned. With the naked eye or even under a common lens I could not detect the presence of metallic copper;- but under a simple microscope the whole mass seemed, as it were, infiltrated with copper in very fine particles. The bronze-like colour is evidently due to copper thus finely disseminated. 9. A mixture consisting of 300 grains of disulphide of copper and BLUE-METAL. 355 136 of iron-pyrites, from South Wlieal Frances, was heated in the same manner as the mixture in Exp. 4. The product was well melted. No metallic copper could be observed on the surface. The fracture was uneven ; its colour was coppery-bronze, resembling that of the button in Exp. 8, but richer, and with a more highly metallic lustre ; in places this fine colour was uniform, but for the most part the surface was mottled with a mixture of this colour and dark-bluish grey. Under a simple microscope minute particles of copper were seen to be diffused everywhere through the mass, and thus examined it was evident that the rich bronze colour was entirely caused by particles of metallic copper, just as in the last experiment. 10. Some of the disulphide employed in the foregoing experiments was heated per se in a covered clay crucible, enclosed in another crucible containing anthracite powder, just as in Exp. 4. The product was well melted. The fracture was somewhat uneven and conchoidal. There were cavities on the outer surface, as well as a few in the interior, in all of which were observed angular, apparently crystallized, isolated copper-like particles. 1 1 . Some of the same disulphide was mixed with a large excess of sulphur, and heated just as in the last experiment. The button was perfectly melted. Angular metallic particles were observed on the top, as well as the sides of the button, which had been in contact with the crucible. The fracture was even, smooth, almost vitreous in lustre, and somewhat conchoidal ; the colour was dark grey, like that of disulphide of copper ; a single cavity existed on the fractured sur- face, in which angular particles of metallic copper were observed, but on no other part of the surface could a speck of copper be detected. 12. A mixture of 500 grains of disulphide of copper in powder and 100 of copper turnings was exposed in a covered clay crucible to a strong heat in a muffle during about two hours. The crucible was then taken out, cooled as rapidly as possible, and broken. At the bottom was a button of copper weighing 68 grains. Ko metallic copper was detected in the disulphide, which had the usual appearance, and was not in the least crystallized. 13. The last experiment was repeated, and the crucible left in the muffle during the night to cool. A button of copper, weighing 77 grains, was found at the bottom. Here and there on the surface of the button particles of copper were observed, but not in the form of moss; and the top was covered with small crater-like prominences, as though produced by the evolution of gaseous matter ; the fracture of the disul- phide was crystalline. 14. Feathered shot copper, best-selected, was melted with a large excess of roll-sulphur in a Cornish crucible. The whole was well stirred with a piece of wood, then covered with charcoal, and left to cool slowly in the crucible out of the furnace : particular care was taken to exclude the presence of iron. When perfectly cold, a well- melted lump of disulphide of copper was detached from the crucible, 3 inches in diameter at the top and 2| deep in the centre ; it weighed .12,710 grains, inclusive of a button of copper at the bottom weighing 2 A 2 356 BLUE-METAL. 775 grains. The lump was broken nearly vertically through the middle, when several cavities were disclosed, which measured from -i to |- inch in diameter, and were mostly in the upper half of the frac- tured surface ; there were other and smaller cavities. In several of these cavities, large as well as small, were projecting teeth and fila- ments of copper ; but I did not observe particles of copper anywhere imbedded in the mass. The surface produced by a fresh fracture had the characteristic dark bluish grey colour of disulphide of copper ; but, after exposure to the air, it acquired here and there a rich blue tarnish. The copper in the button at the bottom was brittle, some- what vesicular, and of a greyish red colour, very similar to that which appeared on the fractured surfaces of some of the specimens of regulus obtained in the preceding experiments. The last three experiments would seem to indicate that disulphide of copper alone may possess the property of taking up during fusion a small quantity of copper, which is again set free during solidification ; but, inasmuch as there was no uniform diffusion of copper in minute particles through the mass, and as the copper which was disclosed on fracture was very small in amount and confined entirely to the cavities which existed, the evidence is not sufficient to justify a posi- tive conclusion on the subject. The presence of copper in these cavities may have depended upon the action of the gaseous matter which caused their formation. What this matter was, and how it came to be present, I have as yet no certain knowledge. To account for the presence of metallic copper in blue-metal, Le Play has propounded a theory which, he asserts, " explains all the special facts of the Welsh method (of smelting), and which, moreover, throws great light upon a multitude of operations peculiar to the copper- smelting works of the continent of Europe." 4 It is obvious from this language that he regards this theory as highly original and important. That there may be no mis-statement on the subject, I subjoin a literal translation of Le Play's description of it, and the evidence on which it is founded : " Metallic copper, the presence of which characterizes blae and red- metal, 5 is not actually deposited during fusion ; it is a product which is only formed after the matters have left the furnace. AVhen, after tapping, the regulus and slag are superposed in the mould destined to receive them, these two substances continue during some time to act upon each other as they did in the interior of the furnace, and to make an exchange of the two metals (copper and iron). But affinities are gradually modified by the progressive cooling of the two reacting substances. A moment arrives when the temperature of the regulus, still perfectly fluid, falls to that at which copper tends to assume the solid state ; from this moment the reaction, which before was simple, 4 Op. cit., p. 273. latter, and less than coarse-metal. The 5 This term is applied by some smelters red colour is due to dispersed minute par- to a regulus less advanced than blue-metal, tides of metallic copper. that is, one containing more iron than the BLUE-METAL. 357 becomes double ; 6 the iron of the reguhis combines with the oxygen of the dioxide of copper ; but the copper set at liberty, instead of combi- ning with the sulphur previously combined with the iron, is in some way deposited in the molecular state, the temperature not being sufficiently high either to compel the copper to combine with the regains, or to cause it to collect in distinctly melted globules. The phenomenon takes place, then, during that period of the cooling of the matters tapped out when the regulus and slag still preserve the fluid state at their surface of contact, and while the temperature of the regulus already borders on that at which copper solidifies." The arguments adduced by Le Play in support of this theory are the following : 1. If metallic copper had existed in Hue or red metal before tapping, it is not conceivable that at the white heat which prevails in the furnace the metal should preserve the pulverulent or filiform state in which it is found in the cooled regulus, and not subside to the bottom. 2. When, according to Le Play, a furnace containing blue-metal is tapped, so that one portion may flow into a mould apart and free from admixture of slag, while the remainder is allowed to take its usual course and to solidify under the slag, copper is always entirely absent from the former, but is present in large quantity in the latter. 3. Le Play having remarked that blue-metals, identical in external characters, became charged with very unequal proportions of metallic copper, suspected that this might be due to a variation in the propor- tion of dioxide of copper dissolved in the silicate. Admitting the correctness of his theory, it should follow that in a regulus containing the same amount of sulphide of iron, the metallic copper ought to be proportionate to the dioxide of copper in the slag. This conjecture was subsequently confirmed by analysis. 4. It occurred to Le Play that he might be able to determine at will the production of metallic copper in any particular part of the regulus after tapping, either by introducing sulphide of iron into a very white metal, or dioxide of copper into a slag which contained but little of that oxide. He often made experiments of this kind, which always responded to his expectation. 5. In the middle of the bed destined to receive the contents of the furnace, Le Play made a circular cavity O m 60 (23'6 in.) in diameter, and O m 25 (9'8 in.) deep in the centre, towards which the sides sloped gradually down. This cavity, which W 7 as in communication with the rest of the bed, became completely filled with regulus, and then covered with a mass of slag O m 25 (9-8 in.) thick. As soon as the tapping was over, he introduced at one spot into the regulus, according to the nature of the matters upon which he was operating, a more or less strong dose of very ferruginous coarse-metal, and an equivalent dose of silicate of dioxide of copper into the corresponding part of the slag. Immediately after the pellicle of slag on the upper surface had solidified, charcoal powder : Des ce moment la reaction devient simple de double qu'elle etait." 358 BLUE-METAL. and a considerable mass of sand were projected upon it, so as to prevent sudden cooling. In one experiment of this kind, at the junc- tion of the regulus and slag, was found a large geode, of the capacity at least of 200 cubic centimetres (30-4 cub. in.), and entirely filled with extremely delicate threads of metallic copper, iridescent with the most vivid colours, and possessing nearly the same degree of flexibility as threads of organic matter. The whole of this deposit had about the same consistency as a large mass of tow. The copper thus separated had the following composition : Copper (by dry assay) 98*2 Iron -. 0-4 Nickel 0-6 Intermixed sand and carbon 0'2 99-4 I shall now proceed to consider the proofs of the theory in question seriatim. 1. It is a fact that blue-metal, richly impregnated with metallic copper, may be re-melted at a high temperature, without causing the copper to subside and collect in mass at the bottom. The experiments recorded at pp. 352, 3, establish this fact beyond question. 2. Issue must here be joined with Le Play on a point of observation. Metallic copper is not absent from blue-metal which has solidified without a covering of slag. I have it on the authority of an acute and expe- rienced smelter, that blue-metal may be tapped out of the furnace with- out any admixture of slag, and yet contain metallic copper disseminated through its mass. Again : if a large mass of blue-metal be carefully examined under a microscope, metallic copper may be seen dis- seminated, apparently in isolated particles, and in equal quantit}^ throughout. Now, if this separation of metallic copper were effected at the junction of the regulus and slag, it might certainly be expected to exist to a much greater amount in the vicinity of this junction ; and it might be supposed that fibres of copper more or less continuous would be traceable in the regulus from the upper surface, descending in a vertical direction ; which is not the case. 3. The argument con- cerning the connexion between the varying proportion of dioxide of copper in the slag and the proportion of metallic copper in the regulus is not of much weight ; because this proportion may, nay, doubtless does, depend on that of the sulphide of iron in the regulus ; and a relation must necessarily exist between this proportion of sulphide of iron and that of the dioxide of copper in the slag. Hence an effect which Le Play directly ascribes to the proportion of dioxide of copper in the slag, may be entirely due to the proportion of sulphide of iron in the regulus, of which that of the dioxide in the slag is the measure. 4. The experiment concerning the addition of sulphide of iron to very white metal furnishes no proof in favour of Le Play's theory of the supposed action of the oxide of copper in the slag ; because metallic copper is separated by the addition of sulphide of iron to disulphide of copper when no slag is present. It seems rather difficult to understand how a satisfactory result should be obtained by adding dioxide of copper, MOSS-COPPER. 359 or matters containing it, to the slag after its removal from the furnace, unless it be at a very high temperature and extremely liquid. But if this had been the case in Le Play's experiment, the copper separated ought certainly, according to his own views, to have been well melted and deposited in the state of shots. 5. That filamentous copper may be formed without the agency of silicate of dioxide of copper, may be inferred from the fact that cavities in every part of pigs of pimple-metal may be full of it ; and I am informed by Mr. William Edmond, that it may even be seen exuding from under the pigs into the sand beds. From a review of the evidence which has now been presented con- cerning the presence of metallic copper in Uue-metal, it appears to me that while the theory of Le Play cannot be maintained, the theory of Plattner is supported by strong evidence, both of observation and expe- riment. Indeed, when the fractured surfaces of Uue-metal and of some of the specimens of regulus obtained in the preceding experiment are carefully examined under a microscope, it seems difficult to resist the conclusion that the liberation of metallic copper must have been due to a cause operating equally through the mass. In some cases the copper is dif- fused in particles so numerous and minute as to be quite imperceptible to the naked eye, when it produces a reddish-brown uniform colour. Admitting that such an equally acting cause does operate through the mass of the regulus during the process of cooling, as the theory of Plattner requires, all the varied appearances which have been described may be perfectly explained. The copper is liberated during solidification, according to my experience, whether slowly or rapidly effected, and may be re-dissolved on subsequent fusion The sepa- ration of carbon in the form of graphite during the solidification of grey-pig iron may be adduced in illustration of the separation of copper in Uue-metal. That graphite exists diffused through solid grey- pig iron will be shown hereafter. When the iron re-melted, the graphite must dissolve in the liquid metal ; for otherwise, on account of the great difference in specific gravity between it and the iron, it would necessarily rise to the top, which is not the case. Moss-copper. In copper-works this term is commonly used to desig- nate those accumulations of filamentous, or moss-like copper, which are formed in cavities in pigs of certain kinds of regulus. Mr. Edmond informs me that, in making copper from Cornish ores, moss-copper seldom appears ; but more of it is produced when these ores are melted in admixture with a little Irish ore (copper-pyrites mixed with much iron-pyrites ?) : it occurs most abundantly when foreign ores are much used. It is chiefly observed and in the finest state in pimple-metal, when all the cavities are filled with it, and it is found protruding from the bottom of the pigs into the sand underneath ; sometimes a little of it, strong and wiry to the touch, appears on the upper surface of the pigs. According to Mr. Edmond, it may be seen in the little prills, or shots of metal, in the ore-slag ; and the surfaces of the pigs of metal from the caldned-metal furnaces are covered with a coating of it, gene- rally of a dark colour, and as thick as the nap, or pile, on velvet. In specimens in my collection the filaments of copper vary in size 360 MOSS-COPPER. from the finest thread to fibres VV of an inch in diameter ; and from one of these specimens, obtained by Mr. Edmond from a fine-metal furnace bottom, I have taken separate filaments perfectly continuous, and exceeding five inches in length. Under the microscope, the filaments present numerous minute, parallel, and longitudinal lines, or grooves, as though they consisted of bundles of extremely delicate fibres. Sometimes a continuous stratum of copper is found, entirely composed of densely-packed, delicate, parallel fibres. A remarkable specimen of this kind has been presented to me by Mr. Ednaond, who obtained it from a roaster-furnace bottom. The stratum is of an inch thick in the thickest part, and appears to have been formed in a narrow flat cavity in dark grey regulus, from the opposite walls of which the fibres seem to have been protruded until they met: the supposed plane of junction of the ends of these fibres is indicated by a layer about T ! ff of an inch thick, of a rich ruby colour, like that of the surface of Japan copper ; and along this plane the fibres are more or less contorted and convoluted. The colour of the fibres, even in the same mass, varies from brass-yellow to ruby-red; and occasionally it is pale grey, like that of slightly-tarnished silver : it is due to superficial tarnish by oxidation, or otherwise. The mode in which these fibres are produced is an interesting subject of inquiry. Each fibre seems to have been pushed, as it were, through a draw-plate, and at a temperature when the metal was soft, but certainly not exceeding that of well-melted copper ; for otherwise the fibres, immediately after their protrusion, would have been re-melted into globules. Filaments of silver, which, examined under the microscope, appear to possess identically the same structure as those of moss-copper, may be formed by heating finely-divided sulphide of silver in a current of hydrogen at a temperature sufficient to agglu- tinate the sulphide, but below the actual melting-point of silver. This beautiful experiment may be made in a glass-tube, through which a current of the gas is passed. Long delicate fibres of silver may be seen protruding from minute rounded masses of the sulphide ; and as they are produced while these masses are in a soft state and lying free in the tube, the idea that they result from the application of external mechanical pressure, in a similar manner to maccaroni, can hardly be entertained. There seems to be a force in operation at the base of each filament, which causes the particles of silver at the moment of liberation successively to arrange themselves in one continuous fibre, or series of fibres ; or in other words, each filament grows, as it were, from a root imbedded in sulphide of silver. Moss-copper is said to be remarkable for its purity, and to be equally pure whether the blue-metal on which it occurs may have been produced from the purest or most impure ores. 7 At p. 358 an analysis by Le Play of moss-copper has been recorded. The following analyses of this substance are given by Mr. Napier : 7 Napier, op. cit,, v. p. 346. 110ASTING BLISTER-COFFEE. 361 1. 2. Copper ! 98*5 99'0 Sulphur 0-4 0-4 Tin, Antimony, etc 1-0 0-5 Iron trace trace 99-9 99-9 No. 1 was brass-yellow in colour, and No. 2 red. Now, if the copper forming rnoss-copper had been set free at a temperature above the melting-point of copper, it would, in subsiding through the bath of regulus, have reduced, more or less completely, any sulphide of tin or antimony with which it might have come in contact ; and copper-bottoms, consisting of very impure copper, would have been obtained. Blue-metal is converted into white-metal by melting it in contact with slags rich in silicate of dioxide of copper, or ores containing car- bonate or oxide of copper and silica ; and at first it might seem difficult to understand how this conversion should be effected by the agency of oxidized compounds of copper, as blue-metal contains disulphide of copper and metallic copper, whilst white-metal may be practically regarded as pure disulphide. Whether Plattner's theory of the existence and modus operandi of disulphide of iron be true or not, if we admit his statement that a certain regulus, consisting of copper, iron, and sulphur in the proportions requisite to form disulphide of copper and protosulphide of iron, has the power of dissolving some metallic copper, the difficulty just alluded to may easily be removed. Thus, suppose we have a regulus, of which the composition may be repre- sented by the empirical formula #Cu 2 S+?/FeS-|-2:Cu, then the follow- ing equations will enable us to understand the conversion of blue-metal into white-metal by the action of oxidized compounds of copper and silica : *Cu 2 S+2 FeS+Cu 2 +2 CuO+*'SiO 3 =*Cu 2 S+2 Cu 2 S+2 FeO, eS + 3 Cu 2 0, *'SiO 3 =.rCu 2 S+3 Cu 2 S+3 FeO, Hence it will appear that the sulphide of iron may be completely separated from blue-metal without the evolution of any sulphurous acid. 5. Roasting. In this operation white-metal is melted with free access of air, and without the admixture of any other substance except silica, of which there is always a sensible amount in the state of sand adhering to the pigs of a regulus cast in sand-moulds. From the descriptions which have been previously given, it might be concluded that roasting is essentially a process of oxidation, due to the action of atmospheric air at a high temperature ; and that this conclusion is correct will be clearly established by the composition of the two pro- ducts obtained, namely blister-copper and roaster-slag. Blister-copper. This term indicates the appearance of the surface of the copper, which presents numerous blister-like elevations; caused by corresponding cavities in the substance of the metal underneath. The following analysis of blister-copper, obtained by the roasting of white-metal^ is by Le Play : 362 ROASTEB-SLAG. Copper 98-4 Iron * 0-7 Nickel, cobalt, manganese 0'3 Tin and arsenic 0'4 Sulphur 0-2 100-0 Mr. Napier has published three analyses of blister-copper, which, in essential points, agree with that of Le Play. They are as follow : 1. 2. 3. Copper 97-5 98-0 98'5 Iron 0-7 5 OS Tin and antimony 1-0 0-7 0-0 Sulphur 0-2 0-3 0-1 Oxygen and loss 0'6 0-5 O'G 100-0 100-0 100-0 Mr. Napier remarks that the oxygen existed in the state of dioxide of copper, dissolved in the metallic copper : a result which a priori would hardly have been anticipated ; but that sulphur and dioxide of copper may co-exist in metallic copper has already been demonstrated (see p. 264). Before the copper reaches the state of blister-copper, it passes through that of pimple-copper, so called on account of the pimple-like excres- cences with which its surface is studded. Mr. Napier gives analyses of six varieties of copper produced in the operation of roasting, and in a less advanced state than blister-copper. The extremes of variation per cent, in the copper, iron, and sulphur, are as follow : copper, 89-4 95-6; iron, 0-3 2*4; sulphur, 0-4 2-5. To all such varieties the term coarse-copper is applied in contra-distinction to blister-copper. Roaster-slag. This slag presents a characteristic appearance : it is vesicular, and more or less pumice-like, or scoriaceous ; it is without metallic lustre, and its prevailing colour is dark -reddish brown, but here and there it is grey-black. This slag contains shots of metallic copper, which may be separated by trituration and levigation. Le Play has published the following analysis of roaster-slag, which contained but little intermingled metallic copper : Silica 47-5 Alumina 3'0 Dioxide of copper 16 '9 Protoxide of iron 28-0 Oxides of nickel, cobalt, manganese 0'9 Protoxide of tin 0-3 Lime and magnesia traces Metallic copper 2-0 98-6 Le Play estimates that on an average roaster-slag contains 20 per cent, of copper, inclusive of that present in the metallic state. A piece of roaster-slag which Mr. Morgan gave me at the Hafod Works, and which he considered to be a characteristic specimen, contained as much THEORY OF THE ROASTING PROCESS. 363 as 43*73 per cent, of copper existing as oxide, and O85 as metallic copper : these results were confirmed on repetition. In the following analysis of roaster-slag recorded by Mr. Napier, the per centage of copper is 39-95: Silica 45 Oxide of copper (Cu 2 O?) 25 Oxide of iron (FeO?) 28 Sulphur 2 100 From the fact that sulphur is stated to be a constituent, it is to be inferred that some metallic sulphide was present ; and as the slag is scoriaceous and only imperfectly melted, it is not difficult to understand how it should retain intermingled sulphides, notwithstanding the presence of so large an amount of oxide of copper. In the sequel it will be demonstrated that well-melted slags essentially composed of silica and protoxide of iron may contain a sensible quantity of sulphide of iron. From the preceding analytical data, the reactions which occur during the process of roasting may be deduced, and a reason assigned for the particular method of conducting that process. The manner in which copper may be completely reduced from disulphide by the joint action of heat and atmospheric air has been fully explained ; and it is precisely in this way that blister-copper is reduced from white-metal in the operation of roasting. The chief object, therefore, of the operation is to effect the reduction of the disulphide of copper constituting white-metal by the generation of oxide of copper. Now, if, in roasting, the regulus were rapidly melted, the action of the oxygen of the air which enters the furnace would be limited to the surface of the bath of melted regulus ; and although reduction would in this case proceed, yet the time required to effect the removal of the sulphur, as sulphurous acid, would be greatly prolonged. But, in the ordinary method of roasting, the regulus is, as we have seen, very slowly melted, so that eveiy drop as it trickles down is exposed to the action of the free oxygen in the gaseous current of the furnace. A consider- able quantity of oxide of copper is thus formed, which decomposes the disulphide in contact with it ; and so reduction takes place pari passu with the fusion of the regulus. Still, after complete fusion, desulphu- rization of the product which consists of regulus and metallic copper is far from complete. In the next, or second stage of the operation of roasting, the temperature of the furnace 'being much lowered, the rising of the regulus occurs. Owing to the continuous formation and evolution of sulphurous acid during solidification, the undecomposed regulus is thrown up into porous crater-like prominences, whereby the surface exposed to oxidation is greatly increased, and a consider- able quantity of oxide of copper is accordingly formed. As the mass is now in a pasty, almost solid state, the superficial oxide of copper does not become mixed with the undecomposed disulphide, and is, therefore, persistent ; nor, indeed, if the oxide and disulphide were in contact, would reciprocal decomposition take place, because the 364 BEST-SELECTED PROCESS. temperature would not be sufficient to produce that effect. In the third stage, the heat of the furnace is increased so as to completely melt the solidified product; and during fusion the oxide of copper formed on the crater- like surface becomes thoroughly intermingled with the remaining" regulus, which is thereby more or less completely reduced with the evolution of sulphurous acid. By this mode of proceeding the object of the roasting process is attained without the necessity of frequent rabbling, which would be required if the fusion in the first instance were rapidly effected and the product kept in a state of perfect liquidity until the complete expulsion of the sulphur. In the last stage the product is kept well melted with free access of air during the remainder of the operation. Sulphurous acid continues to escape from the surface of the melted mass, producing an appearance of ebullition and a sound like that of frizzling ; and, as we learn from the preceding analyses, a sensible amount of sulphur remains in the final metallic product of the roasting process. The gas seems to be formed at some depth below the surface of the liquid mass, as every bubble on escaping occasions some degree of spirting. AVhen a' pig of blister-copper is broken vertically across, numerous long, narrow, bright tube-like cavities may be seen occasionally on the fractured sur- face extending from the bottom to the top. A beautiful specimen of this kind is in the collection at the Museum of Practical Geology, presented by Mr. Hussey Vivian. Oxide of copper is formed on the surface, which, by the action of currents in the liquid, may be submerged, and by coming in contact with still remaining disulphide may determine the evolution of sulphurous acid, and give rise to the spirting in ques- tion ; or possibly oxidation on the surface may be carried to a degree sufficient to produce a little dry copper, which, on coining in contact with the subjacent metal containing disulphide (mechanically mixed or dissolved '?), would produce the same effect. The silica of the slag is, as has been slated, derived partly from the sand adherent to the pigs of white-metal, and partly from the sand- bottom and other materials forming the interior of the furnace. The presence of this acid may be useful in combining with the oxides and forming a light, pasty slag, which admits of being easily skimmed off. Moreover, it would tend in a material degree to promote the oxidation of the iron. Best-selected process. The object of this process is to produce copper of the highest degree of purity ; by which is meant, as free as possible from foreign metals. Best-selected copper commands the highest price in the market, and is in special request for the manufacture of certain alloys, such as the best qualities of brass, and the white alloy known as German silver ; it is prepared, as we have seen, by a special process of purification ; and it is generally pretended that it is exclusively produced from the purest ores. That some smelters do employ the purest ores in making this variety of copper may be true ; but that others do not, seems very probable from the fact, which is noto- rious in Birmingham, that copper which is occasionally (I might say, frequently) sold as best-selected, is extremely bad. Some of the smelters BEST-SELECTED PROCESS. 365 know perfectly well that what I now state is correct, as they have suffered pecuniary loss from having had no small quantity of best- selected copper returned on their hands at different times. That the selecting process would in a certain degree effect the separation of certain metals, such as tin, might be anticipated from various experiments previously detailed, and that it actually does so will be established by analytical data. The process, it will be remembered, essentially con- sists in the partial reduction of the copper contained in a regulus, having the composition of white-metal. After the principles which have now been expounded and proved by analysis, the mode in which this partial reduction takes place will be clearly understood without any further explanation. The metallic copper, which is separated in the selecting process, will be found to contain a considerable quantity of tin, if stanniferous copper-ores had been employed to produce the white-metal. Metallic copper decomposes sulphide of tin, in a greater or less degree ; but whether this sulphide is capable of being completely decomposed by any proportion of copper, does not appear. Supposing the regulus, then, to contain tin in the state of sulphide the state in which it would exist it would ,be in a greater or less degree reduced by the metallic copper in the course of its precipitation through the regulus. So also, if ihe regulus contain sulphide of antimony, this metal would be set free by the reduced copper, with which it would alloy and sub- side (see p. 260). The copper which is thus reduced and alloyed is termed bottoms. The bottoms obtained in the selecting process generally, not to say always, contain a considerable proportion of foreign metals, namely, tin, antimony, and others ; from which we may infer either that best-selected copper is not made from the purest ores, or that what are called the purest ores are in reality impure. A specimen of the regulus (called regule) produced in the first run- ning down, as described at p. 330, had the following characters : its upper external surface was pimpled ; the colour of a fresh fracture was dark grey, like that of disulphide of copper ; it was porous and highly vesicular ; it contained large irregular cavities, having a red- brown coating, of a metallic lustre, confined almost entirely to their floors ; angular particles of metallic copper protruded everywhere into the interior of these cavities, and in some there was filamentous copper ; on the bottom of the pig, which had been in contact with sand, there was also filamentous copper. Another specimen from the Hafod copper-works, and produced many years previously, had precisely the same assemblage of characters. Le Play has given the following analysis of a specimen of bottoms obtained in the selecting process : Copper 92-5 Iron, nickel, manganese ]-6 Tin 0-2 Arsenic 0-4 Sulphur 4-8 99-5 366 REFINING. Mr. Napier has given the following analysis of what he considers a fair sample of copper bottoms : Copper 74-0 Tin 13-8 Antimony 4'5 Lead 0-8 Iron 2-5 Sulphur 3-9 99-5 6. Refining. The physical and chemical changes which the copper undergoes during this operation have been so fully described in a former part of this work, that nothing remains to be added on this subject. The refinery-slag is very heavy, and black externally ; its fractured surface is vesicular and porous, dull and not vitreous, and deep brown- red, with a purplish hue ; where free from cavities, the substance of the slag is well melted and compact. This slag contains numerous shots of metallic copper. Le Play has given the following analysis of refinery-slag : Silica 47-4 Alumina 2-0 Dioxide of copper 36-2 Protoxide of iron 3*1 Oxides of nickel, manganese, etc 0-4 Protoxide of tin 0-2 Lime 1*0 Magnesia 0-2 Shots of metallic copper 9'0 99-5 The proportion of metallic copper was determined by stamping and washing 10 kilogrammes (22 Ibs.) of a mixture of different specimens of refinery-slag. A specimen of this slag, which I obtained at the Hafod Works, contained 57*9 per cent, of oxide, existing as dioxide, and 2*65 of metallic copper, making a total of 6O55 per cent, of copper. In refining best-selected copper, no lead should be added as in the case of tough-cake copper, which is intended for rolling. When tough-cake, which is made of the ordinary description of copper, is laded without addition of lead, it is apt to rise on the face of the cake ; notwithstand- ing which it will roll extremely well, but not so well as after the addition of a small quantity of lead. The proportion of lead may be varied considerably without causing an appreciable difference in the rolling quality of the copper, but not without affecting in a very appreciable degree the appearance of the cake on fracture. Some persons profess to be able to decide with certainty concerning the quality of copper from the appearance of its fracture ; but a decision on such grounds may undoubtedly prove erroneous. I have often been unable on the most careful examination to detect the slightest differ- ence in the appearance of the fractures of ingots of copper, which I knew differed much in regard to working qualities and degree of purity. Ample proof has already been advanced that the same copper REFINING. 367 may be made to present widely different appearances on fracture, according to the manner in which it is cast. Temperature alone will suffice to modify the characters of the fracture in a very perceptible degree. One of the best smelters at Swansea, a man of great experi- ence and an acute observer, in writing to me remarks, " you would be surprised how at different temperatures in the lading the appearance of the fracture is altered. If laded very hot that is, what a copper-man would call very hot, for, what other people would call very hot, he says is as cold as ice the structure of the copper seems quite altered ; and it presents on fracture an assemblage of crystals, larger or smaller, and more or less perfect, which do not in the slightest degree make their appearance when laded at a lower temperature, and which do not at all affect its malleability." That variation in the temperature of a metal at the time of lading, and the consequent variation in the rate of cooling, should thus modify the fracture is very intelligible, for reasons which have been given in an early part of this work. Experiments on the small scale will not, except in particular cases, afford satisfactory results on this subject. In the small assay no such differences are observable as are indicated by the fracture of a cake or ingot, when obtained at different temperatures. An assay at the tough- pitch, when the copper is too cold to lade well, and one taken when the copper is as hot as it is usually laded, will present no perceptible difference on fracture. The actual amount of copper in both cases is so small that the metal sets immediately, and the rate of cooling in both may be regarded' as practically the same. At my request, my friend Mr. John Keates was so obliging as to pre- pare for my examination ingots of best-selected and tough-cake copper, laded at different temperatures into moulds exactly similar in all respects. There were three of each kind of copper : the three of best-selected will be indicated, respectively, by the letters b, &', 6", and the three of tough-cake by the letters c, c', c". The ingots &, c, were laded at the highest temperature which could be conveniently attained in the refining-fumace ; I', c', were laded after the greater part of the charge was out, and the metal had cooled down; and 6", c", were laded after the metal had cooled down still more, and was not far from solidification. The characters of the fractured sur- faces of these ingots were examined and recorded immediately after fracture, which was obtained in the usual way after first nicking the ingots across the under surface. The upper surfaces of all the ingots were flat. The colour of the best-selected copper was perceptibly paler than that of the tough-cake. All the ingots presented on their fractured surfaces numerous shining grains, which, on close inspection, especially under a lens, are clearly seen to be caused by globular cavities, b un- even, more or less columnar, studded with minute globular cavities in every part ; &', somewhat less columnar than b, but in other respects similar : ?>", much less columnar and more even than in b. c, columnar structure distinct, but less so than in Z>, studded with numerous globular cavities : c', not perceptibly different from c : c", very similar to c', ex- cept along a band across the top, about ^ inch deep, where it was silky, 368 .REFINING. like that of the assay, at tough-pitch, tolerably even, with scarcely a perceptible trace of columnar structure. The existence of these globular cavities merits special attention ; they have frequently been mistaken for crystalline faces. I have never seen the fracture of an ingot of copper free from this porous structure, except when cast in a reducing atmosphere, as described at p. 276. It was formerly, and for aught I know to the contrary may still be, the practice of persons entrusted with the examination of the copper supplied to the Navy, to rely entirely on the appearances which it presents on fracture in forming a judgment respecting its quality. How far such a method of judging should be considered satisfactory, may be inferred from the following fact, of the truth of which I have received unquestionable evidence : In execution of a large contract, a quantity of copper was sent to one of H.M. dockyards, and rejected as inferior on account of the appearance of its fracture. The same copper was simply remelted by the smelter who supplied it, and laded at a temperature different from that at which it had been cast in the first instance, and then returned ; when it was accepted as good. I -have it on good authority, that the proportion of lead and the temperature at which the copper is laded are the two conditions to be attended to in order to produce tough-cake copper which shall present a fracture of which the dockyard authorities will approve. I do not for a moment mean to assert that in no case will the fracture afford any certain indication of quality, although T affirm, without fear of contradiction, that very different qualities of copper may present precisely similar fractures. In the casting of copper, ingot-moulds of copper or iron are employed. Some smelters have assured me that, according to their experience, the ingots are invariably full of pin-holes when iron moulds are used, while others of great experience maintain that it is quite immaterial of which metal the mould may be made ; and this accords with my own observation. I have examined ingots which have been cast in both kinds of moulds, and I have found no difference, either in their outward appearance or internal structure. However, moulds of copper are, I believe, in most frequent use ; they may be readily made by casting the copper in suitable moulds of cast-iron. The colour of the surface of an ingot of copper is modified by the temperature of the water in which it is cooled. In casting copper in small ingots of the usual form, the moulds are arranged in a frame on the edge of a long trough filled with water, so that they may be easily inverted without being detached, and allow the copper to fall into the water. As soon as the ingots of copper have solidified sufficiently, which occurs in the time required for the filling of three or four moulds, they are thrown into the water. At first, while the water is cold, the surface of the copper has an orange colour ; but when the water has become warm, it acquires a rosy tint, similar in kind to that of Japan copper, though much less intense. In the Museum of Practical Geology is a very fine and characteristic series of specimens, which Mr. Hussey Vivian has presented in illustra- ELIMINATION OF FOREIGN METALS IN COPPER-SMELTING. 369 tion of the operations of copper-smelting as conducted some years ago at the Hafod Works ; and of these I subjoin a list, with the names and statements of the amounts of copper which were attached to them at the time they were received : DESCRIPTIONS OF KEGULUS. copper percent, Raw-metal (coarse-metal) 39'1 Red-metal 48-1 Blue-metal (tliis is decomposing, the surface being covered here and I KQ . g there with a blue efflorescence) j> Metal between blue and sparkle-metal : it contains much moss copper 69 1 Sparkle-metal 74-3 White-metal 76-6 Pimple-metal 78-8 Close-regulus (best selecting process) 79'6 Open-regulus ( ,, id.: both this and the preceding contain ) OQ.Q much metallic copper) / DESCRIPTIONS OF COPPER. Pimple-copper 99 ! Blister-copper (both this and the preceding must be unusually pure)... 99-3 DESCRIPTIONS OF SLAGS. Quartz-ore slag 47 Vitreous do., containing alumina and zinc (is not alumina always I ft . 2 present?) j Ore-slag, very sharp (rich in protoxide of iron) 0-5 Ore-slag, of ordinary quality 0'3 Metal-slag 2-1 Roaster-slag 23 9 Refinery-slag 71-0 ON THE ELIMINATION OF CERTAIN FOREIGN METALS DURING THE WELSH PROCESS OF COPPER-SMELTING. This is a subject of great interest in a scientific as well as practical point of view ; but it has not, I believe, been so thoroughly investi- gated as it deserves. Such an investigation could only be properly conducted by a skilful analytical chemist, and would be very laborious ; and it would be necessary that he should have constant access to copper-works, in order to be able to select with due care every specimen intended for analysis. If one of our great Welsh smelters could be induced to take an interest in the science of his art for its own sake, and quite irrespective of any consideration of pecuniary gain, many important questions concerning the metallurgy of copper would soon be decided. Some of these questions, perhaps, may have been solved long ago within the precincts of copper-works, and the results regarded of too much commercial value to be divulged. The Science of Metallurgy in this country has hitherto not only not received much encouragement from the men who have amassed great wealth by the practice of metallurgic arts, but has frequently had to encounter their strenuous opposition ; and the opposition, I have generally remarked, has been proportionate to their ignorance of the principles of the pro- cesses which they were engaged in carrying on, and their dependence 2 B 370 ELIMINATION OF AKSENIC. upon the brains of others for the successful management of their works. It is this class of men, moreover, who affect superior knowledge, and delude themselves with the notion of their being the exclusive pos- sessors of mysteries a class which, it is to be hoped, is doomed to speedy extinction. Within the last twenty years I have witnessed great changes in the views of many of our smelters, especially the iron- masters, concerning the value of the Science of Metallurgy. In several of the great iron- works skilful chemists have been successfully engaged in investigating problems of interest, and, with but few exceptions, the results have been freely disclosed. The elimination of arsenic. Arsenic is generally, if not always, present in the mixtures of ores smelted in this country : it may occur in combination with copper in true grey copper-ore (Fahlerz), in small proportion as arseniate of copper, in mispickel (FeS 2 + FeAs), in arsenical iron (Fe 4 As 3 and FeA), and occasionally in other minerals. Although arsenic is separated in a greater or less degree during the smelting process, yet it appears to be rarely, if ever, completely eliminated ; for, according to recent experiments by Dr. Alfred Taylor, distinct traces of the metal may be detected in all commercial copper. 1 Dr. Taylor states that he found arsenic "in not fewer than forty samples of copper, as it is employed by chemists in the form of wire of various sizes, of foil of various thicknesses, and of gauze, coarse and fine." He further states that he detected arsenic in two out of five specimens of electrotype copper. The copper in the forty speci- mens above mentioned was, doubtless, of the description known in commerce as " tough cake," as it is only this description which is employed in rolling and wire-drawing. He does not appear to have examined any specimens of " best selected copper," so that his experi- ments do not justify the conclusion to which he has come, "that all the copper used in commerce, the arts, and chemistry, refined or unre- fined, contains arsenic." It needs scarcely be remarked that all copper used in the arts has undergone the operation of " refining." I regret that at present I am unable to offer satisfactory information concerning the degree in which arsenic is expelled at each successive stage of the process of smelting, and the precise reactions accom- panying its expulsion. In the analyses of Le Play, previously inserted, arsenic appears as a constituent in coarse-metal, blue^netal, white-metal, blister-copper, and the bottoms obtained in the 'selecting process; but the proportion was only determined in coarse-metal and bottoms. I have perceived a strong odour of arsenic in the steam produced on damping the calcined granulated metal with water after its removal from the furnace. In cavities in the spongy regulus obtained in the selecting process arse- nious acid is found condensed occasionally in beautiful crystals and in considerable quantity. I have received the following observations on 1 Facts and Fallacies connected with the Research for Arsenic and Antimony, &c. By Alfred S. Taylor, M.D., F.R.S., p. 28. ELIMINATION OF ARSENIC. 371 this point from Mr. Morgan, of the Hafod Works. " In the cooling of the pigs of selected metal, large cavities are produced near the top of the pig, either by the disengagement of gas from the metal itself, or by means of a portion of steam, which is frequently formed from moisture in the sand into which the metal is tapped. It is on the upper surface, or rather depending from the roof, of these cavities, that they occur ; and the thickness of the scale or plate of metal which forms the upper boundary or cover of the cavity I have scarcely ever, I think, found to exceed -^ of an inch." When mispickel (FeS 2 -f FeAs) is heated to incipient redness, sulphide of arsenic is evolved, which, in contact with atmospheric air, is converted into sulphurous and arsenious acids. Basic arseniate of sesquioxide of iron exists in the product obtained by roasting this mineral at a sufficiently high temperature. When arsenical iron (Fe 4 As 3 and FeAs) is heated to incipient redness, metallic arsenic is volatilized ; and when roasted with access of air, much arsenious acid is evolved, with the formation of a sensible amount of basic arseniate of sesquioxide of iron. When arsenious acid and atmospheric air are passed over sesqui- oxide of iron or protoxide of copper at a red heat, basic arseniates of these oxides are formed. 2 When arsenious acid is passed over protoxide of copper at a red heat without access of air, dioxide of copper and basic arseniate of protoxide of copper are formed. 3 When arsenious acid is heated with sulphate of copper, arseniate of protoxide is formed. When arseniates of copper and iron are strongly heated with silica, the arsenic acid is displaced and resolved into arsenious acid and oxygen, and silicates of the metals are formed. These reactions must occur during the fusion of products, such as calcined ore, calcined granu- lated coarse-metal, &c., in which arseniates may have been generated during the process of calcination. When fluor-spar is present in the ores, fluoride of silicon may be evolved (see p. 44) during calcination as well as in the fusion of the calcined ore. The sulphuric acid liberated during the former process may act upon the fluor-spar, causing the production of hydro-fluoric acid, and, owing to the constant presence of quartz, of fluoride of silicon. These compounds of fluor may, possibly, tend to eliminate arsenic in the state of fluoride, which, on subsequent exposure to moisture, would be converted into hydro-fluoric and arsenious acids. Faraday states that in his examination of the condensing apparatus erected by the late Mr. Vivian at the Hafod Works he found " most decided indications of fluoric acid " in water from the first shower-chamber connected with the ore-calciner flue. 4 In a refinery at Birmingham, where the sweep- ings of silversmiths' and jewellers' workshops technically called "sweep" are melted in reverberatory furnaces with the addition of 2 Plattner, Die metallurgischen Rost- I 3 Plattner, op. cit, p. 15Q. prozesse, p. 89. 4 Proceedings, &c., ante cit., p. 67. 2 B 2 372 ELIMINATION OF ANTIMONY. lead-slags rich in flnor, so great has been the amount of fluorine com- pounds disengaged that the glass of all the windows in the immediate vicinity which have been exposed to the smoke is corroded and ren- dered more or less opaque. The elimination of antimony. When this metal is present in copper- ores, it, probably, almost always exists in the state of antimonial grey copper-ore. It may sometimes occur in ores in which its presence might not be anticipated : thus, I know that it has been detected in copper alleged to have been made exclusively from Burra-Burra ores; and, if the allegation be true, the presence of a sensible quantity of antimony in these fine ores is, I should think, quite exceptional. In Le Play's analyses of the products of copper smelting, no mention is made of antimony. Napier found more than 4 per cent, of it in bottoms pro- duced in the selecting process. It may, certainly, be frequently detected in commercial varieties of copper, even in those reputed to be of the best quality. During calcination, it is possible that some of it may be volatilized in the state of oxide (SbO 3 ) ; but it passes, undoubtedly, in considerable quantity into the coarse-metal. That it is removed in great measure in the selecting process, is proved by the analysis of Napier above referred to, and may also be inferred from the reactions which occur when metallic copper is heated with sulphide of antimony (see p. 260). In 1856 a patent was granted to MM. Beudant and Benoit, French engineers, for "separating antimony and arsenic from copper-ores." 5 The ores are smelted in the usual manner, so as to yield a regulus similar to coarse-metal, and from this regulus the antimony and arsenic are separated by one or other of the three following methods : 1st. Metallic iron, wrought or cast, is introduced into the melted regulus, whereby, it is stated, nearly all the antimony and arsenic are precipi- tated in combination with iron and copper, according to the tempera- ture and state of the regulus. But a small quantity of antimony and arsenic generally remains in the reguhis, and this is precipitated by adding 1 or 2 per cent., more or less, of lead or galena, and continuing the action of the iron. By this treatment the patentees affirm that the regulus is almost entirely deprived of antimony and arsenic, a small quantity of antimony and a considerable proportion of arsenic being volatilized during the operation. If the regulus contains much sulphur, some roasted ore is added in order to lessen the amount of iron required to effect precipitation. The precipitated metallic mass of antimony and arsenic is melted with a mixture of ore and iron- pyrites, by which means the copper and iron contained in it are ex- tracted, and " the antimony and arsenic are left nearly pure ;" the cupreous regulus thus obtained is treated again in a subsequent opera- tion. 2nd. Lime or roasted ore, or a mixture of the two, is added to the melted regulus, and the surface is covered with charcoal or carbo- naceous matter, by which process is obtained " a metallic button, or mass of antimony and arsenic, either pure, or mixed with iron and 5 A.D. 1856. No. 1061. ELIMINATION OF ANTIMONY. 373 copper, in quantity varying with the degree of sulphuration of the matt (regulus) and the quantity of lime and roasted ore employed : " the antimony and arsenic remaining in the regulus after this treatment are precipitated by the addition of a little lead or galena, assisted by the action of metallic iron or by that of the charcoal on the surface. As an example of this mode of treatment, the patentees state " that a grey copper or coarse metal, consisting of 68 parts by weight of proto- sulphuret of iron, 20 parts of sulphuret of copper, and 20 parts of sulphuret of antimony, was melted, with the addition of 16 parts of lime and 6 parts of roasted ore. A button of antimony weighing 13 parts was obtained, and the operation was completed by the addition of 2 parts of galena and stirring the mixture with a piece of iron." 3rd. A sufficient quantity of roasted ore to saturate the excess of sulphur is added to the regulus, and then metallic lead ; antimony and arsenic are precipitated, and carry down with them a quantity of lead, varying with the time which is allowed to elapse after the addi- tion of lead, the state of sulphuration of the regulus, and the quantity of lead added. According to the patentees the operation may be so conducted as to cause a complete precipitation of the antimony with only a very small quantity of lead. In lieu of metallic lead galena may be added to the regulus, together with roasted ore or lime, or both, and the whole is then covered with charcoal or carbonaceous matter. The precipitated button of antimony and arsenic contains lead, copper, and iron. The regulus, after having been thus deprived of antimony and arsenic, is smelted in the usual manner. The operations above mentioned are to be effected in a reverberatory furnace, having at the end furthest from the fire a cavity in the bed for the reception of the precipitated metals, so that they may be readily tapped off. AVhen the regulus is well melted a mixture of lime and roasted ore is added, in the proportion of about 80 parts of lime and 30 of roasted ore for each 100 of sulphide of antimony in the mass. These proportions have been found to produce a good result. After the whole mass, with these additions, is completely fused, its surface is covered with carbonaceous matter and the heat continued ; the anti- mony and arsenic collect in the cavity in the bed, from which they are tapped out ; about 2 per cent, of galena is then added to the remaining regulus, which is stirred with an iron tool. A fresh portion of antimony, containing some lead, collects in the cavity, and is tapped out; after which the regulus is tapped out. If the last portion of antimony has carried down a little lead with it, the regulus may be regarded as pure ; but if the antimony is free from lead, a small portion of lead should be added to the regulus before it is tapped out of the furnace. The reactions which occur in this process are generally quite intelli- gible, which is more than can be declared of many patented "improve- ments" in copper-smelting. The sulphides of antimony and arsenic are reduced by iron with the formation of sulphide of iron. Iron is the reducing agent commonly employed on the large scale in the preparation of antimony from the native sulphide. The patentees do 374 ELIMINATION OF TIN. not affirm that the regulus is totally deprived of antimony and arsenic by the use of iron alone ; but they do affirm that by the conjoint use of iron and lead, or galena, a complete precipitation of these metals may be effected. If galena be added, it would be taken up by the regulus, sulphide of lead combining readily with various metallic sulphides : then, by the action of iron, it would be reduced with the evolution of metallic lead, which, on coming in contact with the sulphides of antimony and arsenic, would effect their reduction ; and the excess of metallic lead in subsiding would tend to collect and carry down the antimony and arsenic. The principle of this process is similar to that of making best selected copper ; but whether it is more econo- mical or more effective copper-smelters must determine. Iron, it may be stated, will not, as some persons suppose, precipitate copper from every description of copper regulus. When the regulus contains a certain proportion of sulphide of iron, metallic iron will not precipi- tate a trace of copper from it. Thus, thin hoop -iron in large excess may be kept immersed in melted copper-pyrites without causing the separation of any metallic copper. The regulus resulting from this action is reddish-brown, has a bronzy tint, and contains, it is true, minute angular, not globular, particles of metallic copper disseminated through its substance : this copper is not precipitated by the direct action of the iron, but is separated in the same manner as that occurring in blue-metal. The elimination of tin. Ores such as " burnt leavings " (see p. 322) may contain a sensible quantity of peroxide of tin (SnO 2 ). During the first operation of calcination, sulphide of tin may be formed by the action of copper-pyrites or any bisulphide of iron present. When peroxide of tin is heated with sulphur or either of these sulphides, it is converted into sulphide of tin with the evolution of sulphurous acid. During the selecting process, the tin is in great measure separated from the regulus, and passes into the bottoms, which sometimes contain sufficient tin to form a white alloy ; and this separation is explained by the reaction which takes place when copper is heated with sulphide of tin. In 1851 Mr. John Cameron, chemist at the Spitty Copper- Works at Loughor, published a description of crystals of oxide of tin which he had observed in copper slags at certain stages of the process. They were insoluble in nitric acid, and when heated -with black flux gave a button of tin. Heated in an atmosphere of hydrogen, they were reduced, and water was formed. Mr. Cameron inferred that they were a quadroxide (SnO 4 ), but this inference was, no doubt, erroneous. 6 I' have received from my former student, Mr. Eidsdale, who was for some years engaged at copper-works in South America, but who now holds an appointment at H. M. Mint, very beautiful colourless, trans- parent, acicular crystals of peroxide of tin, some of which are I inch in length. They were discovered in the chimney of a copper-roasting furnace, which was undergoing repair, and formed part of an incrusta- tion at the base of the chimney flue, close to the bed. "When found 6 Chemical Gazette, 1851, p. 125. ELIMINATION OF NICKEL AND COBALT. 375 they were nearly black from the presence of protoxide of copper, with which they were coated. The crystals which I received had been pre- viously boiled repeatedly in nitro-hydrochloric acid to free them from associated matter, and afterwards sifted from some amorphous oxide of tin occurring with them. 7 As the furnace from which the crystals were obtained had been in use a long time, and had been employed to roast regulus produced from various descriptions of ore, it was not possible to ascertain from which ore they had been derived. In an experiment by Smith, in which copper-pyrites was heated in admixture with peroxide of tin, crystals similar in appearance to those above described were obtained. The details are as follow : 368 grains of pure copper- pyrites were intimately mixed by trituration with 148 of dressed Cornish tin-ore, yielding 72 per cent, of metallic tin by dry assay ; the mixture was put into a covered clay crucible, inclosed within another, the space between the two being filled with the coarse powder of burnt fire-clay, and the whole was exposed to a strong red-heat during f of an hour. A well- melted regulus weighing 412 grains was obtained. Very delicate acicular crystals adhered to the under- surface of the lid, and a considerable amount of similar, but less delicate, crystals were found lining a cavity in the upper part of the regulus. The fractured surface of this regulus was granular and yellowish iron-grey, except near the bottom, which consisted of brilliant, well-defined, interlacing, lamellar crystals, having a super- ficial bluish-black colour and some degree of iridescence. The inner surface of the crucible surrounding the regulus was coated with a thin layer of black, opaque, vitreous slag. Sulphurous acid had, doubtless, been formed in this experiment at the expense of the oxygen of the oxide of tin and the sulphur of the sulphides ; but, as the regulus has not yet been analysed, it is impossible to state with certainty the reaction. The elimination of nickel and cobalt. According to Mr. Hussey Vivian, " nickel and cobalt in quantities of considerable value are contained in copper-ores, either the produce of foreign countries or of England;" and in the smelling of such ores these metals pass in part into the 7 Professor Miller of Cambridge, at my request, has, with his usual kindness and promptness, examined these crystals, and communicated to me the following re- sults : " The supposed crystals of SnO 2 are right-angled prisms, without any other faces, not even faces truncating the edges. The ends are imperfect, being either undeveloped or broken. The light reflected from a face of the crystals is most nearly extinguished by a Nicol's prism, held in a proper position, when the angle of incidence is about 64 34' or 65. If the crystals were not doubly refractive, this would indicate an index of refraction between 2'1028 and 2-1445. In the case of a crystal having double refraction, such a number only approxi- mately indicates the refractive energy of the substance. I was unable to find the optical constants of stannic acid in any of the books upon such subjects. I presume it is difficult to find crystals sufficiently transparent for observation. One frag- ment bounded by cleavage-planes in the University Collection gave, by a very rough observation, Index of ordinary ray 2*010 Greatest index of extraordi-\ , 1P7 nary ray / 2 ' 117 These observations confirm the sup- position that the furnace crystals are >tanui(- acid. The lustre also favours the supposition.'' Jan. 30, 1861. 376 ELIMINATION OP NICKEL AND COBALT. refined copper, in part into the slags, and in part into " a product called white or hard metal (bottoms ?), commonly used in the manu- facture of nails or sold at an inferior price." The Fowey Consols copper-ore is stated to contain, in addition to silver, a sensible amount of nickel, to which the alleged inferior quality of the copper produced from this ore has been attributed. In 1851 a patent was granted to Mr. Hussey Vivian for certain methods of separating nickel and cobalt from copper-ores in the process of smelting. 8 These methods are founded on facts previously well known to persons who had any practical acquaintance with the metallurgical treatment of nickel and cobalt ores in this or other countries, namely, the affinity of nickel and cobalt for arsenic, and the affinity of copper for sulphur. Thus, Lampadius, in 1827, published the following statement. " In lead- smelting at Freiberg (bei der Freiberger Blei-und Bleisteinarbeit) speise is formed ; whilst the sulphur betakes itself to the lead, copper, and iron, the arsenic combines with the nickel, cobalt, and bismuth." 9 It will be seen hereafter that in the smelting of nickel or cobalt ores, arsenic acts a part exactly analogous to sulphur in the smelting of copper-ores. Mr. Vivian claims as his invention " the separation of nickel and cobalt, or either of them, in the form of arsenurets from ores, slags, regulus, and other combinations or alloys of copper ... by means of the affinity of nickel and cobalt for arsenic and of copper for sulphur." In the following details I shall adhere as closely as possible to the language of Mr. Vivian. The process is for the sake of convenience divided into two steps : in the first step, an impure arsenical compound containing nickel and cobalt is separated from the mass of matter with which these metals were first associated ; in the second step, the arsenides of nickel and cobalt are obtained in a marketable state. Ores, slags, and other compounds containing copper and other metals, the whole or chief parts of which are in an oxidized state, are melted with arsenical pyrites containing sufficient arsenic to absorb the whole, or nearly the whole, of the nickel and cobalt, together with a portion of the copper present in the material ; and at the same time iron-pyrites or raw ore furnace mctalis added in suffi- cient quantity to yield sulphur to combine with the remainder of the copper, and, lastly, coal or carbonaceous matter for the more complete reduction of the oxides. The melted mass, after skimming off the slag in the usual manner, is tapped into a bed, when the arsenical portion sinks to the bottom of the pigs, and is readily separated on cooling from the superincumbent copper regulus. The arsenical com- pound thu8 produced contains nearly the whole of the nickel and cobalt present in the material treated ; but should the whole of this compound not have been separated from the regulus, the latter must be re-melted once or twice, and tapped off as above described until the arsenical deposit is exhausted. Mr. Vivian found that in treating an oxide of copper, produced by calcining regulus of copper of 70/ , 8 Obtaining nickel and cobalt. Speci- fication A.D. Nov. 4, 1851, No. 13,800. 9 Grundriss einer allgemeinen Hiitten- kimde. Gottingen, 1827. p. 35. ELIMINATION OF NICKEL AND COBALT. 377 the addition of 8 cwt. of arsenical pyrites, 12 cwt. of raw ore furnace metal containing about 30/ of sulphur, and 2 cwt. of coal to 20 cwt. of the oxide, extracted nearly the whole nickel and cobalt from the oxide and concentrated them in the arsenical product as required. In treating other products of this class, sufficient sulphur should be present to yield a copper regulus of between 50 and G5/ , and suffi- cient arsenic to prevent the bottom from assuming a coppery appear- ance and to convert it entirely into a compound, which, when broken, will have a sharp, bright, white fracture. Copper-ores and regulus containing small quantities of nickel and cobalt, and in which the metals are not in an oxidized state, are smelted in the usual manner of smelting copper-ores until they furnish a regulus of white-metal of about 70/ . This regulus is then roasted so as to produce a metallic bottom. On the completion of the first portion of the roasting process, when the plug-holes are to be closed and prior to melting for tapping, from 3 to 5 cwt. of arsenical pyrites are added. The contents of the furnace are tapped into an ordinary sand-bed, and, when the pigs are cold, under the first three or four will be found a metallic bottom, in which the largest portion of the nickel and cobalt contained in the regulus originally treated will be concentrated. This metallic bottom consists of a low arsenical compound, easily broken, and presenting a greyish fracture approaching that of cast-iron. Mr. Vivian states that, as arsenic is generally associated with nickel and cobalt in copper-ores, a certain degree of concentration takes place in any metallic bottoms formed by roasting the regulus of such ores, but only to an inconsiderable extent. Copper-ores or regulus rich in nickel and cobalt, and in which the metals are not in an oxidized state, generally contain a large quantity of arsenic, and are melted in the usual manner of melting copper-ore, except that the metallic products are not granulated, but run into pigs. At the bottom of one or more of the first pigs a product will be found different from the superjacent regulus ; it has a higher specific gravity than the regulus, breaks with a light, white fracture, and consists of nickel, cobalt, and other matters ; it separates easily from the regulus, and contains the greater portion of the nickel and cobalt originally present in the ore. Nickel and cobalt are separated from cupreous alloys on the same principle namely, that of the production of two compounds ; one a regulus of copper and other metals, with little or no nickel and cobalt ; thu other an arsenical compound or speise, containing nearly all the nickel and cobalt originally present. These alloys are treated in any of the above described processes by being placed at the top of the charge of materials when they are first put into the furnace ; or by melting the granulated metal with materials containing sulphur and arsenic, and proceeding in the manner detailed, but without the admix- ture of coal. The arsenical compounds resulting from these processes usually con- tain from 1 to 12/ of nickel and cobalt, and from 15 to 50/ of copper, together with iron, antimony, and other matters. These 378 ELIMINATION OF GOLD AND SILVER. impure arsenical compounds are pulverized or granulated, and then calcined in the ordinary manner of calcining copper-ores. The calcined product is melted with the addition of sulphur and silica, and with arsenic in amply sufficient quantity to combine with the nickel and cobalt. Sulphate of baryta is recommended as the source of sulphur, arsenical pyrites as the source of arsenic, and common sand as the source of silica. A mixture actually used consisted of 12/ of silica, 10/ of arsenical pyrites, 8% of sulphate of baryta, to 70/ of arse- nical product, and 1 part of coal to 4 of sulphate of baryta. The products of this melting were, first, regulus containing copper, anti- mony, zinc, iron, lead, silver, and other such metals as may have been present, and also a certain small portion of nickel and cobalt ; secondly, nickel and cobalt speise, together with portions of the other metals, which, it is preferred, should thus pass into the speise, rather than the risk should be incurred of any notable portions of nickel and cobalt remaining in the regulus for want of arsenic; thirdly ', slag consisting chiefly of iron originally present in the material : the slag is skimmed off in the usual manner, and the regulus and speise are tapped off, when, on cooling, they separate according to their respective specific gravities. The regulus is melted slowly and tapped into a bed, when an arsenical product containing nickel and cobalt will again be found under the first few pigs on the first two or three meltings, after which the regulus will be generally free from nickel and cobalt, and may be treated for copper in the usual manner ; and in case it should not be quite free, a semi-arsenical bottom may be taken from it, as in the case of ordinary copper regulus before described, which will free it from the last portions of nickel and cobalt, or nearly so. The arsenical product obtained by this second step will be found to be freed from the greatest portion of its impurities, and will contain a largely increased percentage of nickel and cobalt ; but in case it should not be sufficiently pure to be marketable, which will commonly be the case, it is to be again pulverized or granulated, calcined and melted, and otherwise treated in the manner described until a speise of the required purity is produced. The elimination of gold and silver. In smelting auriferous or argenti- ferous copper-ores, the gold and silver may be more or less completely concentrated in the bottoms obtained in the selecting process ; but in the absence of this process they will be found diffused through the whole mass of the copper produced. In 1856 a patent was granted to Mr. H. Hussey Vivian and others for the " manufacture of copper, and obtaining gold and silver from cupreous ores." l The bottoms impreg- nated with the precious metals are granulated in cold water and then converted by calcination into oxide, which is susceptible of being pounded to dust in a mortar. An ordinary copper- works calciner is capable of bringing 3 tons of copper into this state in 72 hours. This 1 Date of the specification Dec. 22, A.D. 1850. The persons associated with Mr. Vivian in this patent were Mr. B. Gustav Herrman, manager of his Silver- Works, and Mr. W. Morgan, manager of the Hafod Copper- Works. ELIMINATION OF GOLD AND SILVER. 379 method of oxidizing copper by granulation and calcination is not original, as will be shown hereafter. A mixture consisting of 16 cwt. of the pulverized oxide and 26 cwt. of "sulphurous copper-ore" (con- taining 30 per cent, of sulphur) is melted, when a regulus (containing about 40 per cent, of copper) is formed. The same result may be arrived at by melting the oxide in admixture with raw ore furnace metal and siliceous matter. A small metallic bottom rich in gold is usually produced in this process. It is recommended to add sufficient "sulphurous material" to change nearly the whole of the oxide of copper into regulus. By calcining and smelting in the usual manner, the regulus is advanced to white-metal (of about 70 per cent, of copper), which is subjected to the selecting process, so as to furnish light regule and metallic bottoms. From the ordinary furnace charge of 2 tons, 30 cwt. of regule and from 5 to 6 of bottoms have been obtained with a good result. In these bottoms will be found nearly the whole of the gold originally existing in the ore ; but if any sensible amount of it remains in the regule, a second selecting will remove it and concentrate it in the bottoms. The patentees state that lead, arsenic, and antimony are generally (collectively or separately) present in the auriferous copper bottoms, and that the presence of these metals materially facilitates the concentration of the gold in the bottoms ; and they advise that in the absence of any or all of these metals, lead, in the shape of litharge or ore, should be added in the conversion of the oxide into regulus. The metallic bottoms, thus formed, are again and again submitted to the processes of granulation, oxidation, conversion into regulus, and concentration by selecting, until the gold exists in such a proportion to the copper as to render its separation* by any of the well-known methods economical. When silver is present in the bottoms, it is extracted from the regulus prepared from them in the manner above described by a special process which will be hereafter described. In 1846 I made experiments on a considerable scale on copper containing a large quantity of silver and a small quantity of gold. The metal was in ingots, and was converted into sulphide by heating it in large crucibles, with the addition from time to time of sulphur. When the conversion was only partial, I found that the unchanged metal at the bottom was very much richer in silver and gold than the original ingots, but that the regulus with which it was covered retained a considerable proportion of silver and also a sensible amount of gold. In 1781 Jars, in his well-known work, 2 proposed a process founded on essentially the same principle as that patented by Mr. Vivian, the separation of copper from gold by converting the former into a sul- phide. He writes as follows: "It is more than proved that if we had auriferous copper, this process would be still more advantageous than for silver." The process referred to is that of melting auriferous silver in conjunction with iron pyrites in a small blast-furnace, so as to " mineralize " the silver, t. e. convert it into sulphide and allow 2 Voyages Metallurgiqucs, 3. p. 285. 380 COPPER MADE BY OLD WELSH PROCESS. the gold to subside. " In case we should have much of these auriferous coppers, as may happen at a mine, the separation of gold from the copper after it (the latter) had been mineralized might be very well effected by means of litharge in a reverberatory furnace ; for although gold may have a greater affinity for copper than for lead, yet on being disengaged from the copper through the medium of sulphur, it would unite to the lead which in roasting first acquires the metallic state." The regulus or the plachmall freed from, gold is directed to be roasted and smelted for copper in the usual manner. Jars thus concludes : " How many argentiferous and cupriferous matters containing gold occur every day in commerce which it is not considered worth while to submit to any process of separation, and which might be most advantageously treated by the method which I have just de- scribed ! " It is remarkable how much valuable matter lies entombed in metal- lurgical treatises now regarded as old and worthless. In the course of this work it will be shown that many a modern patentee has claimed inventions as his own which had been known and practised long before he was born. I have frequently detected traces, and more than traces, of gold in copper, especially, I think but I am not sure in that derived from South American ores. Some years ago Dr. Lyon Playfair communicated to me the following incident relating to the presence of gold in copper. At large chemical works, where sulphate of copper was prepared by dissolving copper in sulphuric acid, an insoluble residue was produced in the process which had been put aside from time to time and had not, fortunately, been thrown away. A small sum was offered by certain persons for this residue, which had not previously been regarded as of much value. Suspicion was excited, especially by the quarter from which the offer proceeded, and it was declined. The residue was examined, and found to contain 700?. worth of gold ! Alleged superiority of the copper made by the Welsh process as formerly prac- tised. A practical smelter of great experience assures me that better copper was made by what he terms the old method of dry roasting. The pigs of regulus were exposed, during the greater part of the pro- cess of roasting, to a less degree of heat than sufficed to melt them, and the roastings were repeated often three times, until blister-copper was obtained. The tin, antimony, arsenic, &c., are stated to have been more completely removed than is the case in the modern practice of wet roasting, in which the regulus is melted down rapidly (not always) and kept boiling until the sulphur is expelled and blister-copper formed. This method is more economical than the old one, but the quality of the copper is deteriorated. NOTICES OF IMPROVEMENTS IN COPPER-SMELTING. 381 MISCELLANEOUS NOTICES OF VARIOUS IMPROVEMENTS IN COPPER-SMELTING. Granulation of regulus. A patent was granted for this process to Thomas Williams in 1778. 3 Furnaces. In 1815 William Bevan the younger and Martin Bevan obtained a patent for heating calciners with the waste heat of melting- fumaces. 4 Melting coarse-metal with non-sulphuretted ores of copper. A patent was granted to Joseph James for this process in 1828. 5 The proportions should be such as to yield a regulus having as nearly as possible the composition of pure disulphide of copper. A reverberatory furnace, about 11 feet long and 7 or 8 feet broad, is recommended for operating upon about a ton of regulus in admixture with the proper quantity of ore. The advantages ascribed to the process are, that "the separation of iron, sulphur, and other impurities, from copper ores " is facilitated, and that fuel is economised which would otherwise be required for cal- cination. It is difficult to understand on what principle such a patent could be granted. The process essentially consists in melting regulus with matters containing oxide of copper, but roaster and refinery-slag are matters of this kind, and they had been used with the same object long before. However, in 1848 it was again introduced into a patent obtained by Alexander and Henry Parkes. 6 Furnaces. In 1838 Nicholas Troughton procured a patent for " improvements in obtaining copper from ores," amongst which is the introduction of a blast, either cold, or, by preference, heated to about 500 Fah., into the closed ashpit of reverberatory furnaces. 7 This con- trivance was subsequently claimed in a patent granted to Julius Adolph Detmold in 1843, 8 and again in 1849 in a patent granted to Thomas Symes Prideaux, with the title of "improvements in puddling and other furnaces." 9 The Ebbw Vale Iron Company, in Monmouthshire, purchased Detmold's (?) invention, and, after having for some years applied it to puddling-fumaces, have recently been attacked by Pri- deaux on the ground of infringement. Costly litigation has ensued, which would have been impossible if the officers of the Crown had been compelled to exercise the least discrimination in the concession of patent rights. It is not consistent with common honesty that the Crown, any more than an individual, should sell the same article to two or more persons and coolly leave them to fight for the possession of it. Grievous injustice of this kind has been, and is, so frequently per- petrated in the name of law, that it is high time the reckless system of issuing letters-patent should be amended in accordance with recognised principles of justice. 1 Furnaces. In 1843 Edward Budd and William Morgan, of the Hafod Works, obtained a patent for "improvements in the treating or reducing 3 A.D. 1778, May 7. No. 1191. 4 A.D. 1815, July 12. No. 3938. 5 A.D. 1828, July 17. No. 5676. 6 A.D. 1848, Nov. 11. No. 12.325. s A.D. 1843, Oct. 18. No. 9911. 9 A.D. 1849, Aug. 30. No. 12,750. 1 Since the above was in type a ver- dict has been given, and in my opinion " A.D. 1838, May 22. No. 8075. most justly, in favour of the Defendants. 382 NOTICES OF IMPROVEMENTS IN COPPER-SMELTING. of copper-ores, and in the construction of furnaces for heating such ores, part of which improvements are applicable to other ores." 2 One of the claims is for a "mode of constructing the bottoms of copper- melting and roasting-furnaces known as ore-furnaces, metal-furnaces, and roasters in such manner that the under surfaces may be kept cool by air or other fluid." The sand bottom of the furnace was sup- ported by a bed formed of cast-iron plates. Mr. Morgan informs me that bottoms thus constructed were not found successful, and that they have been abandoned long ago. It would be sheer waste of time even to notice many of the mis- called improvements in copper-smelting for which patents have been granted in this country during the last 20 years. Some of the patentees display such deplorable ignorance of the first principles of chemistry, and such utter want of practical knowledge, as would seem hardly possible with the present facilities of acquiring information. NAPIER'S PROCESS. This was patented in 1846, 3 and was for some time carried on at the Spitty Works, at Loughor, near Swansea, where I saw it in operation in 1848. I then wrote the following description of the process, with Mr. Napier's assistance. It consists of four operations, corresponding to the four operations of the Cornish method of assaying copper- ores : Sulphuretted ores alone are mixed, as well as practicable, according to the relation between the iron and earthy matters and the silica. n Fig. 98. Side elevation of Great Calciner. They are then calcined during nine hours, and the calcined ore is melted in the usual way for coarse-metal. The slag is skimmed off, and, previously to tapping, for every ton of metal in the furnace are added 120 Ibs. of salt-cake (impure sulphate of soda), 40 Ibs. of slaked lime, and 60 Ibs. of coal. The whole is well mixed by rabbling, after which the furnace door is kept closed for about 15 minutes, when, after again well rabbling, the metal is tapped into sand. When the 2 A.D. 1843, Dec. 28. No. 9999. I Smelting Copper Ores," A.D. 1846, July 3 James Napier, " Improvements in | 20th, No. 11,301. NAPIER'S PROCESS. 383 pigs are set they are put into tanks of water, in which they imme- diately begin to disintegrate. An alkaline solution is formed, in K... Fig. 99. Vortical section on the line A B, fig. 100. which, if the ores contained tin and antimony, those metals should be dissolved. The ley, after three or four hours, is drained off and Fig. 100. Horizontal section on the line E F, fig. 99. thrown away, and the disintegrated powder well washed with water. The washed powder is calcined sweet, in large calciners consisting of Fig. 101. Horizontal section on the line G H, fig. 99. three tiers. Engravings of this furnace are annexed. They have been made from drawings which T obtained at the time of my visit to 384 NAPIER'S PROCESS. the works. After the descriptions which have been already given of copper-calcining furnaces, any written explanation would be super- fluous. There is only one point to which attention need be called, namely, that the fire-place may, when necessary, be put in connexion with the uppermost bed by drawing out the damper (a, figs. 98, 99). Four tons of the powder are cal- cined during nine hours on each bed in succession, beginning with the Fig. 102. Vertical section on the line D, fig. 100. -, , uppermost, so that the ore is sub- jected ty a graduated calcination, the temperature of the calciner increasing from the lowest to the uppermost bed. The calcination is continued during 27 or 30 hours for 12 tons of powder. The compo- sition of the washed powder before calcination is nearly as follows, including some adherent silica : Copper 33 Iron 36 Sulphur 27 The calcined powder contains about 45 per cent, of oxide of copper and 52 of the mixed oxides of iron. It is then fused with as much fine anthracite as will reduce all the copper; and some silica is also added in order that it may combine with the oxide of iron and protect the furnace. By this fusion metallic copper is obtained with about Ij per cent, of metallic iron. If there is a stock of non- sulphuretted ores, they are mixed with the calcined washed powder and anthracite. The former act as a flux to the latter, supplying the silica necessary for the oxide of iron produced during the calcination. None of the slags are, in this case, remelted, as on testing, which is never omitted, they are not found to contain a sensible amount of copper. Copper is thus reduced and obtained in a fit state for the refinery. It contains only iron, and no sulphur. It is melted as quickly as possible, care being taken to avoid the presence of carbonaceous matter. The whole is well rabbled from time to time. In about eight or ten hours the iron is found to be oxidized, and upon the surface of the metal is a slag, which is skimmed off, when, if necessary, poling may commence. Occasionally, however, the iron itself in oxidizing is stated to produce the effect of poling by reducing in a suitable degree the dioxide of copper which may be formed in the copper ; for the latter is sometimes found to be at the tough-pitch without poling. The sulphate of soda being put into the furnace in admixture with coal, becomes reduced to sulphide of sodium, which is uniformly diffused through the coarse-metal, pro- bably in a state of chemical combination. "When the pigs are thrown into water the sulphide of sodium slowly dissolves, and the whole mass becomes disintegrated and reduced to impalpable powder. Now, METHOD OF SMELTING PROPOSED BY RIVOT AND PHILLIPS. 385 as the sulphides of tin, antimony, and arsenic are strong sulphur- acids, and as sulphide of sodium is a strong sulphur-base, it was con- ceived that, in the event of any of these metals being present in the ores, they would be dissolved out during the disintegration and sub- sequent washing in the form of soluble sulpho-salts. I examined a solution which had been obtained in this manner, and it certainly contained antimony, as might have been anticipated if that metal had been present in the ore. In the course of practice, however, it was, I believe, found that the separation of these metals was far from com- plete, and, from one cause or other, this new process of copper-smelting was speedily abandoned and the ancient method resumed at the same works. The works were purchased a few years ago by Messrs. Williams and Vivian, who, it is reported, extracted from the furnace-bottoms and other cupriferous materials on the premises a large quantity of copper in excess of the estimated amount ; but whether this report be correct or not, I am unable to state. It certainly is not an improbable one. The works have remained closed since they became the property of the smelters above-mentioned. METHOD OF SMELTING PROPOSED BY MM. EIVOT AND PHILLIPS. The method was suggested to these gentlemen one of whom, M. Rivot, is a professor at the Ecole des Mines, Paris in 1845, in the course of experiments concerning the proposed application at that time, in Eng- land, of the voltaic current as an agent for eifecting the extraction of copper from its ores. The following abstract is from a MS. description. Sulphuretted ores of copper, free from tin, antimony, and arsenic, are to be reduced to powder and roasted sweet. The roasted ore is to be fused in a reverberatory furnace in admixture with lime or sand and slags from a preceding operation, so that the whole of the copper may exist as silicate in the melted mass. Flat bars of iron are then to be introduced into this mass through suitable apertures in one side of the furnace. The copper is displaced from the silicate by the iron, and subsides to the bottom of the furnace, from which it is tapped, an equivalent proportion of silicate of protoxide of iron being formed. Care is to be taken to keep the bars of iron suspended in the melted mass above the surface of the copper, which accumulates underneath, and which, by contact with these bars, would become very ferriferous. It is necessary to throw some small coal upon the melted mass in order to prevent the protoxide of iron in combination with silica from pass- ing to a higher degree of oxidation, of which the effect would be to diminish the liquidity of the mass and to prevent the complete separa- tion of the copper. When copper ores containing tin, antimony, and arsenic are subjected to this treatment, these metals pass more or less completely into the reduced copper. The method was put to the test of experiment in the vicinity of Paris, and the charge of roasted ore operated on at a time amounted to 150 or 170 kilogrammes. It is stated " that the action of bars of iron upon a melted metallic silicate 2 c 386 SMELTING RICH COPPER-SLAGS IN A BLAST-FURNACE. containing from 2 to 3 per cent, of copper is energetic and rapid, and that 3 hours suffice to reduce the copper in the slag to 0*004 or 0-006 per cent., the copper obtained being free from iron." It was estimated that with ores of 8 per cent, produce the cost of reducing 1000 kil. (about 1 ton) of copper would be 350 francs, and that with ores of 25 per cent, produce the cost would not exceed 112 francs. It was further estimated that, with ores of 8 per cent, produce, the balance in favour of this method, as compared with the Welsh process, would be not less than 175 francs for 1000 kil. of copper ! Yet the method has never been adopted, from which it is reasonable to infer that the estimate was erroneous. SMELTING RICH COPPER-SLAGS IN A BLAST-FURNACE. In 1859 a patent was granted Mr. Hussey Vivian for improvements in copper-smelting, which he describes in the subjoined Provisional Specification. 4 The patent became void by reason of the patentee having neglected to file the complete specification. Heretofore the ordinary process of smelting sulphurous copper-ores has been to calcine the ore in the first instance to such a degree only as to produce, when smelted, a regulus containing about 33 per cent, of copper, and afterwards again to calcine this regulus, so that when re-smelted it may produce a regulus much richer in copper; from, this rich regulus the copper is obtained by other subsequent opera- tions. It is found that if the first calcination is carried so far as to produce a regulus much richer than 33 per cent., the accompanying slag contains a considerable proportion of copper ; and as the object of the copper smelter has been to produce slags which may be thrown away without involving much loss of copper, the carrying the first calcining beyond the point above-mentioned has been as far as possible avoided. Now, according to Mr. Vivian's invention, the ore is calcined in the first instance so far that when smelted it will produce rich regulus, similar to that usually produced after two calcinings and two smeltings, and containing, say, 70 per cent, of copper (unless a somewhat lower regulus for the purpose of making best selected copper be required) ; and the accompanying slag, which will be rich in copper, is smelted in a blast-furnace, by which operation is obtained from it a further quantity of regulus, and a clean and more uniform slag; and much loss which now occurs from carelessness and me- chanical imperfection in slag trying is avoided. The fuel employed in the blast-furnace consists either wholly or in part of the coal and cinders which fall through the bars of the calcining and smelting- furnaces, and which cannot afterwards be conveniently burnt in such furnaces, and are consequently at the present time wasted. The blast-furnace in also employed to reduce the red or coppery slags re- sulting from processes of copper smelting which are subsequent to the 4 A.D. 1859. No. 962. COPPER-SMELTING IN BLAST-FURNACES. 387 production of regulus, whereby metallic copper is obtained from these slags instead of regulus, as is now the practice. It is stated that the impurities contained in these slags are not by this process re-intro- duced into fresh lots of copper, as is now the case ; and by the re- ducing action of the blast-furnace many impurities not easily separable in the reverberatory furnace are driven off, and the quality of the copper produced from the slags is improved. Ores of copper which do not contain sulphur are also treated in a similar manner, that is to say, by smelting in a blast-furnace ; and thus metallic copper is at once ob- tained from such ores, in place of, as at present, reducing the copper contained in such ores to the state of regulus, by smelting them with regulus containing an excess of sulphur, which process necessitates the use of a large proportion of sulphurous ores. In 1859 a blast-furnace was in vigorous operation at the Hafod Works, which at night lighted up the neighbourhood with the greenish flame issuing from its mouth. It is to be regretted that Mr. Vivian has not informed us what those many impurities are, which, he alleges, are driven off in the blast-furnace. In a subsequent part of this volume evidence will be adduced to show that certain impurities, well known to affect the quality of copper, are not absent from copper smelted in blast-furnaces. Mr. Vivian may, possibly, possess novel and important information on this subject, which he has not deemed it prudent to disclose. Supposing Mr. Vivian's patent not to have become void, it is difficult to understand how he could have established it on the ground of novelty of invention. But the word invention is not unfrequently used in the present day in what may be termed a non-natural sense. COPPER-SMELTING IN BLAST-FURNACES. In former times all copper was smelted in furnaces of this descrip- tion, and they are still employed for the purpose in Europe and other parts of the world. In tracing the history of a metallurgic art, nothing is more striking than the gigantic scale of operation at the present day as compared with that of ancient times. But in some countries no progress has been made, and smelting processes are still carried on just as they appear to have been at their commencement. The principles, however, upon which many of these processes are founded and the manipulations practised have remained substantially the same in all ages. In illustration of the truth of these statements the method of copper-smelting as at present conducted by the natives of various parts of India may be adduced. Nothing can be more insignificant than the scale of their operations, and yet, in principle, nothing can be more correct. The Hindoo smelters belong to inferior grades of society, and the nature of their occupation is regarded as incompatible with respect- ability. I have much pleasure in now presenting an admirable descrip- tion of Hindoo copper-smelting which I have received from my friend and former pupil, Mr. H. F. Blanford, of the Geological Survey of India, who, being well informed on the subject of metallurgy, under- 2 c 2 388 COPPER-MINE AND SMELTING- WORKS IN SIKKIM. stood what he saw, and has, therefore, given an intelligible account of the process ; and I regret that I cannot say as much for many of the authors of Indian Eeports on similar subjects, so true is it that men without special knowledge are unable to observe what they see. DESCRIPTION OF A NATIVE COPPER MINE AND SMELTING WORKS IN THE MAHANUDDI VALLEY, SIKKIM HIMALYA. I witnessed the process here described in one of the southern valleys of the Sikkim Himalya, a few miles from the Terai. The workmen were Nepaulese, by one of whom the little mine from which the ore was obtained was rented of government. The rock of this part of Sikkim, to the north of the great fault which runs along the base of the hills, is a highly foliated quartz and hornblende schist, the folia of which dip at an angle of 30 or 40 towards the north. The copper vein was small, and, apparently, almost coincident with the foliation, dipping evi- dently at a very low angle. The ore was copper pyrites with a large admixture of nrimdic (iron pyrites). I was unable to visit the workings, which appeared to be carried on in the rudest and most irregular manner, owing to the fires being lighted at the time of my visit for the purpose of loosening the vein-stone, a mode of " winning " at one time extensively practised in Europe, and still to be seen in some important mines. The vein-stone loosened by the fire was after- wards detached and broken up by means of the hammer and gad, sketches of which I here append (Fig. 103). A small pick of the annexed form (Fig. 104) was also used. The ore as brought from the mine, and which appeared to be very poor, was sepa- rated as much as possible with the hammer from the adhering rock, and was then pounded with a heavy stone mallet, another stone, with a slightly hollowed surface, serving as a " knockstone," on the centre of which, after each blow, the ore was swept together by a woman. The ore thus pounded was next washed by women in small tyes, which, in their general form, much resembled those employed by the tin-streamers of Cornwall, but were smaller and of more simple construction. This tye consisted of six planks about one foot in width fixed on edge in the ground, so as to form a partitioned trough of the form here shown (Fig. 105). The cavity above the head-board was nearly filled with clay, over which a stream of water, easily regulated in amount by a little clay placed in the feed- Fig. 103. Fig. 104. COPPER-MINE AND SMELTING-WORKS IN SIKK1M. 389 ing-channel, was allowed to flow, and enter the lower trough through a notch (a) in the head-board. The woman who sat, or rather squatted, by the tye, with one hand divided and directed the stream while with the other she held to the aperture (a) a handful of the pounded ore, which was thus washed down into the tye. \Vhen a consider- able quantity of ore had thus accumulated, it was further washed by being raked up with the fingers towards the head-board, while a good stream of water was allowed to flow over the mass. This was con- tinued until the ore was considered sufficiently " clean." The mass accumulated in the lower part of the tye was thrown away, and that in the upper part, occupying about one-third of its length, was removed to the furnace without being subjected to any further process. It consisted of a small quantity of copper pyrites mixed with a large proportion of iimndic, and also much gangue, principally quartz and hornblende. The furnace, formed of a sandy clay, was of the form and dimensions shown in the accompanying engravings (Figs. 106 and 107). It was built in a bank about two feet high, and consisted of a shallow square cavity, the bottom of which was slightly concave. The back wall of the furnace which, as a whole, may be well compared to an arm-chair was carried up to the height of about eighteen inches, in the form Fi s- 105 - of a truncated pyramid, the top being hollowed out for the purpose of receiving the slags as they were removed with pincers from the furnace. The front wall was very low, not more than about six inches above the bed, while the side walls were of interme- diate height, being about one foot above the furnace-bed and six inches above the earthen platform on which rested the bellows. Fig. KG. Bird's eye view. These bellows were of simple construction, two in number, one being placed on each side of the furnace. They consisted of a seam- 390 COPPER-MINE AND SMELTING- WORKS IN SIKKIM. less bag of goat-skin, formed of the skin of the body and fore-limbs of the animal. The bottom, formerly covering the neck of the animal, now embraced in like manner the earthen nozzle of the bellows. The mouth of the bag was gathered in, so as to leave a small opening only, and was grasped by a boy who squatted beside it and worked the bellows, alternately loosening and tightening his grasp as he raised or depressed the bag, thus producing an effectual, though intermittent, blast. The nozzle, moulded by hand, resembled in form the common mouth blowpipe, the end being bent at a right angle, so that, while the stem rested on the side wall of the furnace, the entrance aperture reached to within three or four inches of the furnace -bed, on which the blast impinged at a rather obtuse angle. Charcoal was the only fuel employed in the furnace. FT F 'g- l 1- Front elevation. The first smelting operation was a simple fusion. The furnace being heated with charcoal, a few handfuls of the washed ore, pre- viously dried and mixed with charcoal, were thrown on, and the bellows worked by boys, as above described. More charcoal was added as required until a perfect fusion of the ore was effected. The fused " metal " (regulus) then formed a small pool at the bottom of the furnace, covered with a layer of fused slag, while the burning charcoal floated on the surface. The fusion being complete, the charcoal was removed, water was sprinkled on the slag to solidify it, and it was then lifted off in successive cakes with a pair of pincers and placed to cool in the shallow basin at the back of the furnace. A fresh charge was then thrown on, and the same series of operations repeated, until a cake of " crude metal " weighing eight or ten pounds had accumulated in the bottom of the furnace. This was removed when cold, pounded, and kneaded with cow-dung into small balls, which were dried in the sun and then roasted in a shallow furnace formed of a ring of slag- cakes placed on edge. The roasted " metal " was afterwards refined in the same furnace in which the fusion of the ore had been effected, and in a precisely similar manner, the result being 1st. refined copper, which collected in the bottom in a cake, weighing four or five pounds ; COPPER-SMELTING AT S1NGHANA, IN INDIA. 391 and 2nd. slag, which was not, so far as I could learn, subjected to any further process, the probability being that the ash yielded by the very large amount of charcoal consumed in the process is sufficient to form a highly basic slag, and thus allow the whole of the copper to be reduced to the metallic state. Three pounds of " crude metal" were said to yield one pound of refined copper. An excellent account of the native method of copper-smelting at Singhana in India (lat. 28 5' N. and long. 75 53' E.) was published at Calcutta in 183 1. 5 The ore was copper-pyrites with a matrix of quartz. It was reduced to powder, mixed with cow dung, and kneaded by hand into sausage-shaped pieces 5 inches long and 1 J in diameter. These were sun-dried and roasted in circular heaps 4 feet broad and 1 J high. The fire was lighted in the evening, and on the following morning the roasted ore, which had a red colour, was smelted with charcoal in a small blast-furnace of the following construction. A quantity of common sand was spread upon the floor of a circular hut, and in the centre a small hollow was made from 12 to 15 inches in diameter and from 2 to 3 deep. In this was laid a stratum of fine yellow sand and then another of ashes. A sand bottom, or hearth, was thus formed, which, by the action of heat and the alkali in the ashes, would become firmly consolidated. Two clay nozzles were placed on opposite sides of the hearth and a third one midway between them, the fourth side being left for the escape of the melted slag. The nozzles were connected together with moist clay, so as to form a little circular wall a few inches high to serve as a basement for the upper part of the furnace, which consisted of three annular vessels of fire- clay placed one above another. Each of these vessels was about 15 inches in external diameter, 9 or 10 inches high, and about 3 inches thick. They were used over and over again, but the bottom of the furnace re'quired to be reconstructed daily in the manner above described. Holes were made round the basement, through which a poker might be introduced into the furnace, and there was one such hole obliquely directed through each nozzle near its junction with the furnace, so that a clear passage for the blast might always be main- tained. These holes were stopped with clay,. which could easily be removed when necessary. The blast was produced by three ordinary goat-skin bellows, of which one was attached to each nozzle. They were worked by men, women, and even children, all Mussulmans. Before the furnace was charged, a quantity of charcoal was burned in the hearth in order to dry it. It is reported that in the day (9 to 1 hours) a single furnace would consume 3 maunds of charcoal (1 mun or maund = 80 Ibs. avoirdupois), during which time were added 2J niaunds of the sausage-shaped pieces of ore and cow-dung and 2 or 3 maunds of iron- slag, which was required as a flux, and was brought from a distance. Four persons were emplo3 T ed at each furnace per- haps a man with his wife and two children who received for their iiiga in Science, N r o. 30, Dee. 1831. Calcutta, p. 380 et scq. 392 COPPER-SMELTING IN JAPAN. united services 10 rupees per month. The head man prepared the furnace and occasionally relieved one of the party working the bellows, of which all three were kept constantly in action. On the morning after the first melting, the mass of copper which had collected in the hearth was taken out and sent to the refining furnace. This is described as a small vessel which received the blast of a single bellows. The refined copper was cast into narrow, shallow, clay moulds, each about 1 foot in length. The ingots thus obtained weighed 2 or 3 seers each (1 seer = 2 Ibs. avoirdupois). The copper was lilac-coloured and brittle. Copper- smelting must have been carried on in this locality during a very long period, as the slags had accumulated to such an extent as to form a line of small hills several hundred feet in length and from 30 to 60 feet in height. There were four insulated stone bastions built on one of these artificial mounds. COPPER-SMELTING IN JAPAN. There exists a volume containing descriptions of copper-smelting in Japan written both in Japanese and Chinese, illustrated with numerous plates, representing the manner of working, ventilating, and draining the mines, the dressing of the ores, and the process of smelting, together with other metallurgic operations employed in extracting silver from copper by means of lead. An account and partial translation of this work have been given in the ' Chinese Repository,' 6 from which I have derived the following information on the subject. I much regret that I have not had the opportunity of seeing the original volume, which appears to be most complete, as it is stated there are not less than a hundred drawings merely of the implements employed, such as iron ladles, rods, forks, skimmers, pincers, sieves, brooms, tubs, crucibles, moulds, mortars, weights, bellows, &c. Various kinds of ore are obtained, which are described as yellow, black, reddish,' and grey, brilliant and dull, rich and poor. After leaving the mine the ore is broken in pieces, from which the barren vein- stuff is selected and thrown away. Generally the best ore yields 10, and the poorest 5, per cent, of copper. In the first operation the ore is roasted in solidly constructed and permanent kilns, in which are holes for the admission of air. They are built under a shed, and are filled with alternate layers of faggots and ore, a bed of faggots being first spread over the bottom, where a vent-hole is made for the free introduction of air. The smoke is so sul- phureous as to suffocate one, and the fire cannot be approached. The roasting continues during 10 days. In the Japanese description the time is stated to be 30 days. In the second operation the roasted ore is melted. This is effected in a large furnace erected on one side of a wall, on the other side of which two large bellows are fixed. A trough is described as leading from the furnace, and the whole is contained within a building. In the plate the furnace is represented as sunk in From May to December, 1840. Canton, v. 19. COPPER-SMELTING IN JAPAN. 393 the ground. The furnace being charged with coal (charcoal ?) and ore, two tall, powerful men pull the bellows, while a third man stands before the furnace to separate and level the mass. When the fire has reached its strength (sic), and the liquid metal has risen and filled the furnace, the earthy scoria floats upon the surface, and, little by little, flows off into the trough. As it flows out it is allowed to cool, or else water is sprinkled upon it, and it is taken out and thrown aside. When the ore is all melted, fresh charges of ore and fuel are added until the furnace becomes filled with good metal. All the scoria and fuel are then pushed off, and water is sprinkled upon the top of the fur- nace, when a solid crust (of regulus) is formed, which is peeled off by means of an iron rod and taken away. A second crust is produced and removed in like manner, and this process is repeated until the whole of the regulus has been taken out, when a mass of copper will be found at the bottom ; but if the ore is poor, there may be none. The coarse-metal thus obtained is roasted and melted. These operations are, for the most part, conducted like the roasting of the ore and first fusion. AVhen the furnace is full of liquid metal the top is luted with clay, " leaving a small hole in it in which to put the coal and blast the charge. If there is any scum, take it out immediately and wait till the whole mass is thoroughly fused, then open the furnace and entirely remove the ignited coal and earthy slag, after which wait till the heat has abated a little, and then, sprinkling the surface, take it out in the same manner as when taking out the coarse-metal.'* [This part of the description is not very clear.] All the preceding operations are carried on at the mine. The copper is next sent to a foundry, where, under the direction of (government ?) officers, it is melted, cast, assorted, and has a price affixed to it according to its quality. No copper is allowed to be pri- vately sold. That which is delivered at Nagasaki and Kw^ashi is from Besh-shi, Akita, and Nanibu, while that which is brought to market for ordinary purposes is from other localities. The number of founders is fixed; "they cannot be lightly increased or diminished lest mal- practices should arise." In the foundry the copper is melted under a blast, after which the slag and fuel are removed from the surface of the melted metal, which is taken off in successive crusts by sprinkling with water, &c. The product is fine-metal, which forms a mass a little smaller than the bottom of the furnace, about a cubit broad and half a cubit thick. About 250 cattis (100 cattis = 130 Ibs. avoird.) can be melted in the furnace at once, and there are three meltings in a day. The refined copper is again melted in an earthen crucible and cast into ingots. This operation is described as follows : " A tub of hot water is set near at hand and a square wooden pool made, into which the moulds are placed, and over them a thick hempen cloth is spread. ^ hen the copper is melted, the scoria taken off, and the fire reduced, hot water is poured into the pool (not very hot) until it is almost level with the moulds ; then the smelter, firmly grasping the crucible with a pair of large iron pincers, pours (the metal) into the moulds, which are pre- viously sprinkled with warm water lest they should crack. After- 394 COPPER-SMELTING IN JAPAN. wards water is sprinkled upon the bars to cool them, and they are taken out with a pair of iron nippers. Each casting produces 10 or more bars ; they are 7 or 8 inches long, and weigh about 10 taels (i. e. nearly 1 Ib. avoird.) each." The following note occurs in the original : "If cold water is indiscreetly sprinkled (upon the mould), or if the crucible is cracked, in both cases an explosion will take place ; and, because the lives of persons are endangered by such an accident, great care should be used to guard against it." All the copper bars which are sent to Nagasaki and Kwashi are made in this manner. It is stated that some copper contains both silver and lead, and that this is softer, and is hammered into sheets or drawn out into wire. The use of the hempen cloth is explained by Thimberg, who accom- panied the Dutch embassy to Yedo in 1771. After importunate en- treaties he was permitted to witness the process of casting copper, which he thus describes : 7 The copper was melted in a small hearth level with the floor by means of hand-bellows, and directly opposite, in the ground which was not floored, was dug a hole of an oblong form and about 12 inches deep. Across this were laid 10 square" iron bars, barely the breadth of a finger asunder, and all of them, with one of their edges upwards. Over these was expanded a piece of sail- cloth, which was pressed down between the bars, and cold water was afterwards poured in until it stood about two inches above the cloth. A series of long narrow depressions were thus formed in the cloth between the bars, and into these the melted copper was poured from iron ladles, so that it was actually cast in water. I find there is no difficulty in casting small sound ingots of copper under water in the manner described ; and the ingots thus obtained have a very clean, bright surface. The form of the well-known Japan copper ingots cor- responds to the above description. The transverse section is nearly an equilateral triangle, the edges being rounded and the upper surface somewhat convex. Two which I have measured are from 8 to 9 inches long and about half an inch on the side. One of these was supplied to me by my friend Harry S. Parkes, H.M. Consul in China. The metal was also cast into ingots of various shapes and sizes according to the purposes for which it was intended. The colour of the surface of Japan copper is rich crimson, due, apparently, to a very thin and tenaciously adherent film of dioxide of copper. In 1757 a patent was taken out by Eobert Morris for a " new invented method of fashioning and colouring- copper in imitation of Japan copper." 8 The copper made from British ore and purified in the usual way is melted and "run into small moulds, which are fixed in a machine which keeps them moving in a horizontal circle under water," of which the temperature is regulated to the proper degree. When the copper is set, it acquires the rich colour of Japan copper. Ingots of the colour of Japan copper have been made in this country and exported to India. According to the late Mr. Vivian they were about 6 inches long, and weighed about 8 ounces each. The rich red colour was produced by dropping the Op. cit. 8 A.D. 1757, Fob. 9th, No. 711. COPPER-SMELTING IN SWEDEN. 395 copper from the moulds immediately on its becoming solid into a cistern of cold water. 9 COPPER-SMELTING IN SWEDEN. The process selected for description is that conducted at the copper works at Atvidaberg, which are the largest in the kingdom. I have pleasure in stating that Mr. P. D. Malmqvist, director of the smelting department at these works, has revised the following description, and has communicated many important details. Mr. Malmqvist has the reputation of being the best copper-smelter in Sweden. I am also indebted to Mr. Andreas Grill and to the chemist of the works for information on the subject. In 1850 the Swedish copper-smelter Bred- berg published an account of improvements introduced at these works between 1844 and 1848. 1 In 1859 Julius Ahrend, director of the smelting- works at Oker, published a description, from personal inspec- tion, of the establishment at Atvidaberg. 2 I have largely availed myself of both these publications. The copper ores smelted in Sweden consist exclusively of copper pyrites, mixed with a large amount of iron pyrites and siliceous mine- rals. The principal mines are at Fahlun and Atvidaberg, in the old provinces of Dalecarlia and Ostrogothia respectively. The mines of Fahlun are very ancient, and of world-wide renown. The Atvidaberg Works are situate a few miles distant from Linkoping. The Swedish ores are very poor, the average yield of those of Fahlun, when dressed, being 4 per cent, of black copper, and of those of Atvidaberg 5 per cent. At the last locality there are several mines, of which the most im- portant is the Bersbo mine. The ores consist of copper pyrites, iron pyrites, magnetic pyrites, zinc-blende which in that of the Bersbo mine forms, on an average, one-third of the entire mass magnetic oxide of iron, quartz, felspar, mica, garnet, rarely calc-spar, and, very seldom, traces of galena. In the ore from one of the mines crystals of bright white cobalt occur together with the copper pyrites ; and in the same mine fluor-spar is found, and occasionally arsenical pyrites. The proportions in which these constituents occur in the ores vary consider- ably. The copper and iron pyrites and blende are sometimes so inti- mately mixed that they cannot be detected by an inexperienced eye. The ores are divided into hard and soft according as the quartz and hard silicates or the sulphuretted constituents preponderate. 3 The smallest pieces of ore are washed in a cylinder of thin iron bars, moving in water, whereby a separation into two kinds is effected, one of the size of gravel, and the other of the size of walnuts, both of which are used to cover the roast-heaps. Furnaces. Three kinds of furnaces are employed, namely, the ore- f urn ace, the black-copper furnace, and the refining-hearth. Ore-furnace. Figs. 108, 109, 110 are copied from Bredberg's engrav- 9 Proceeding's, &c., ante cit. p. s... 2 Berg.-u.-hutteiim. Zeitung, Feb. 28, 5i), p. G9et wq. 3 Ahrend, op. cit. 1 Bergwerksfreund, 1850, v. 13, p. 110 I 185i), p. G9et wq. h 39(> COPPER-SMELTING IN SWEDEN. ings. Fig. 108, horizontal section on the line AB, figs. 109 and 110. Fig. 109, vertical section on the line C D, fig. 108. Fig. 110, vertical section 011 the line EF, fig. 108. This furnace consists of a quadrangular shaft, open above, and below terminating in a shallow cavity called the hearth, 6, ft, 5, from the bottom of which, at one corner, proceeds a channel or tap-hole, c, fig. 108. The sides and back of the shaft are formed by walls built up from the ground, while the front is enclosed by a wall supported f on a bar of iron, d, fig. 110, called the tymp- iron. From this bar, to a certain extent upwards, the wall is narrow and vertical ; it is called the fore-wall, and may be taken down and replaced with facility. Higher up the front wall is much thicker, the inner side inclining inwards and the outer side rising vertically to the top. Where the thin wall ends (see fig. 110) there is an iron girder, /, to support the front of the furnace above. At the back, and some distance above the tymp-iron, is an open arched space (a, fig. 110), in which four twyers are fixed (a, a, , a, fig. 109). That part of the hearth which extends beyond and to the left of the tymp-iron (V, fig. 110) is called the fore-hearth, which is open above. In front of the furnace, above the fore- hearth, is fixed a hood to take oif the sulphureous gases which may escape, and which would otherwise greatly incommode the Fig. 108. Horizontal section. Fig. 109. Vertical section on the line C D, fig. 108. Fig. 110. Vertical section on the line E F, fig. 108. EE - FUR KA CE . furnace-men. The interior of the furnace is built of mica-schist, in which the mica is sometimes replaced by talc. The foundation is solidly constructed, and kept dry by means of a drain, e (figs. 109 and 110). For a considerable distance downwards from the top the walls are built double, a narrow space being left between the outer and inner walls, which is filled with sand. By this construction the inner walls can be removed when repairs become necessary without disturbing the outer walls, and, space being allowed for the expansion of the furnace by heat, cracks on the exterior are, to a certain degree, prevented. This space is indicated in figs. 109 .and 110 by the vertical strips of dotted shading. The dotted shading in the vicinity of the hearth also repre- sents sand. The lower part of the furnace is firmly braced by means of tie-rods and cast-iron plates (see fig. 108), of which the edges are left white in the woodcuts. On the bottom of the hearth a mixture of equal parts of sand and clay is stamped well down in a moist state, so as to form a layer 4 inches thick. The sides of the hearth are also coated with the same mixture in a somewhat wetter state. This lining is carefully dried by a gentle charcoal fire, and then upon it is applied, in the manner described, a second lining, composed of equal measures of clay, sand, and charcoal-dust. The clay mixture is also rammed into the tap-hole, and the pointed end of a stick 2 or 3 inches thick is driven in until the point is within 3 or 4 inches of the inside of the hearth ; the stick is afterwards loosened. The twyers lie nearly horizontal. They are made by bending iron plate, 8 or 9 inches broad and of an inch thick, and fashion- ing the ends into nozzles ; and, as they are at first several feet in length, when the nozzles are burnt away fresh ones can easily be formed by the smith. The blast-pipes are of copper, and are 1 \ inch in diameter at the ends from which the air is injected through the twyer into the furnace. The blast is produced by double-acting blowing-cylinders (see IRON SMELTING). By proper management an ore-furnace may be kept continuously in work during 3 or 4 months. The engravings accompanying the preceding description represent the improved construction of furnace described and figured by Bred- berg, and which was built some time before 1848. It differs consider- ably in dimensions from that which Ahrend incorrectly states to consume the least fuel and to give the best results. The measure- ments of this furnace in Swedish feet and inches, which are very nearly the same as English, are as follow. From the bottom of the hearth to the top of the shaft 1 8 ft. Width of the bottom of the hearth, at the back as well as in front, 3 ft. 8 in. Length from back to front 5 ft, 4 in. Width of the shaft'on a level with the twyers 3 ft. 10 in., and at 8 ft. from the bottom of the hearth 4 ft. ; from this point upwards it gradually contracts to 3 ft. 6 in. at the top. Width from back to front 2 ft. 8 in. at the tymp-iron, above which it continues the same to the height of 5 ft. 6 in., whence it gradually contracts until it is only 1 ft. 10 in. at the mouth. There are three twyers, which are nearly hori- zontal, or at most, only slightly inclined downwards and inwards. They are 4 ft. above the hearth bottom ; 1 ft. 6 in. above the lower edge of the 398 COPPER-SMELTING IN SWEDEN. tymp-iron ; and 1 ft. 9 in. above the top of the fore-hearth. Measured from the centre of the nozzles they are 1 2 in. apart. The opening of the nozzles rather exceeds If in. in diameter, and that of the blast-pipes is If in. There are three ore-furnaces of the kind described above by Ahrend, and four on Bredberg's plan, except that they are furnished with a partition wall in the shaft from the back to the front. This wall, which is 1 ft. thick, commences 10 ft. above the bottom arid extends to the top : it causes a more even descent of the charges, and was adopted when coke was first used as the fuel. The present dimen- sions of an ore-furnace are as follow : ft. in. From the bottom of the hearth to the top of the shaft 24 do. do. to the partition wall 10 do. do. to the twyers 4 do. do. to the lower edge of the tymp-iron 2 do. do. to the surface of the fore-hearth ... 2 3 Width of the furnace at the bottom of the hearth 3 8 do. on a line with the twyers .... 4 do. 8 feet from the bottom of the hearth 6 do. at the top (inclusive of the partition wall) ... 6 .0 From the back to the front, at the bottom of the hearth 5 6 do. at the surface of the fore-hearth 6 do. on a level with the tymp-iron 2 6 do. 8 feet from the bottom 2 6 do. at the top 1 9 F"" Mr. Malmqvist informs me that generally in Bredberg's furnaces there is a larger daily yield and less consumption of fuel than in those described by Ahrend. This year (1861) an ore-furnace of a new con- struction has been built for the use of hot blast ; and one of the larger furnaces has also been adapted to the use of hot blast. It is expected that by this means fuel will be economized. The blast is heated by the waste gases of the furnaces* The poorer ores of this year have ren- dered it necessary to try every means of lessening the consumption of fuel. Black-copper furnace. Fig. Ill, section on the line AB, fig. 109. Fig. 112, section on the line CD,fig.l08. This furnace is much smaller than the ore-furnace. In the an- nexed engravings -both furnaces are drawn to the same scale as the ore- furnace. The description which has been given of the ore-furnace applies in great measure to this, so that only a few additional words of explanation are required. Hereafter the reader will find more ample details and more elaborate illustrations of blast-furnaces. Di Fig. ill. Black-copper Furnace. Fig. 112. (Copied from Bredberg's engravings.) BLACK-COPPER FURNACE REFINING-HEARTH. 399 The twyers are 10 inches apart, and are inclined towards the interior of the furnace at. an angle of 3-| degrees with the horizon. The opening of the blast-pipe is ] inch in diameter. The blast is turned full on during the working of the furnace at a pressure equal to a column of mercury of inch (English). From the tap-hole proceeds a channel 36 feet long made of iron plates and divided by partitions into 10 compartments, of which that furthest from the tap-hole is about two inches deeper than that nearest it. In 1859 one of the black-copper furnaces was enlarged and furnished with a partition-wall and 3 twyers. The dimensions are as follow : ft. in. From the bottom of the hearth to the top of the shaft 17 3 do. do. to the partition wall 6 6 do. do. to the twyers 2 ;t. in. ft. in. Dimensions at the bottom of the hearth 24x20 do. on a level with the twyers 2 10 X 2 6 do. 6 feet above the bottom 4 6 X 2 G do. at the top 26x16 Thickness of the partition wall 6 Since this alteration the consumption of fuel has been diminished 19/ > an( l the daily yield increased 23/ . In order in some degree to prevent the reduction of oxide of iron to the metallic state, the gases are partially taken out of the furnace 8 ft. from the top of the hearth and are applied to the heating of steam boilers. Refining-hearth. The annexed engraving has been executed from a drawing, supplied by Mr. Grill, of a hearth at A vesta in Sweden, where the ores of the Fahlun copper-mines are smelted. Fig. 113, vertical section on the line G H, fig. 114, which shows the hearth in plan. The figures on the right of the plan are vertical sections of the hearth on the lines A B, CD, and E F, of fig. 114 respectively. The re fining-hearth consists essentially of a shallow, generally semicircular cavity, constructed in a platform of brick or stone-work, and solidly lined with a refractory substance, such as a mixture of finely-pounded clay and charcoal dust, suitably moistened with water, or of fire-clay and sand. Ahrend has given the following description of the refining-hearth at Atvidaberg : It is from 2 to 2J feet in diame- ter, and from 15 to 18 inches deep. It is made of English fire-clay mixed with sand, well beaten down. The platform of stone-work in which it is contained is 16 inches above the ground in front, but it rises somewhat towards the wall at the back ; it is entirely covered at the top with plates of iron. The twyer, which is of copper, projects 4 inches over the hearth, and is inclined downwards at an angle of 45 } with the horizon : its diameter at the nozzle is 1 J inch, or the same as that of the nozzle of the blast-pipe. The pressure of the blast varies from If to 2f inches of mercury. Beckmann has recorded the following singular feat, which he saw performed by the workmen at the Avesta smelting- works in September, 1765: " One of the workmen, for a little drink-money, took some of the melted copper in his hand, and after showing it to us, threw it against 400 COPPER-SMELTING IN SWEDEN. a wall. He then squeezed the fingers of his horny hand close to each other ; put it (the hand) a few minutes under his arm-pit, to make it sweat, as he said ; and, taking it again out, drew it over a ladle filled with melted copper, some of which he skimmed off, and moved his hand backwards and forwards very quickly, by way of ostentation. While I was viewing this performance, I remarked a Fig. 114. smell like that of singed horn or leather, though his hand was not burnt. The workmen at the Swedish smeltiiig-houses showed the same thing to some travellers in the 17th century ; for Eegnard saw it in 1681, at the copper-works in Lapland." 4 4 A History of Inventions and Discoveries. Translation. London, 1814, v. 3, p. 277. ROASTING, OR CALCINATION. 401 Boutigny, who has of late performed feats of a similar kind, has, as is well known, explained them by the action of water in the spheroidal state. 1. Roasting, or calcination. Ore which is only to be once roasted should be broken in pieces of the size of the fist. Ore which is to be roasted twice may be in larger pieces during the first roasting ; but in the second it must be broken in pieces not larger than the fist. Eoasting is effected in kilns and heaps, but chiefly in the latter. There are only two kilns, which are constructed as follows : From each end and the centre of a vertical wall 50 ft. long and 10 ft. high, proceed at light angles three walls 28 feet long, 10 feet high, and 7 feet thick: two rectangular spaces are thus formed, of which each is enclosed on three sides by walls, while the fourth side or front remains open ; the upper projecting angles of these walls are rounded oif in front. The ground on which the kilns are built is level and horizontal. On the bottom of the kilns wood is piled to the height of 1 foot, and then the largest pieces of ore to the height of 4 feet. On the top of the ore a layer, 4 inches thick, of small charcoal is spread, upon which ore is again piled to the height of 3 or 4 feet, then a second layer of small charcoal, and lastly ore until the kilns are filled. The whole pile is finally covered in front and on the top with a layer 1 or 2 inches thick of the ore dust or schlich. The wood at the bottom of the kilns may now be lighted. Combustion continues from four to six weeks, and at first, owing to the large quantity of zinc-blende present and the free access of air, much smoke is emitted : on the outer surface, as well as in the interior, the ore acquires a white coating of oxide of zinc. The other kind of roasting, which is the most general, is effected in free, or unwalled pyramidal heaps, from 28 to 30 feet square at the base, and built on level ground. A bed of wood, from 8 to 12 inches thick, is first formed, upon which ore is piled in alternation with layers of small charcoal, just as in charging the kilns, until the heap is raised to the vertical height of from 10 to 12 feet. The sides and top are afterwards covered with a layer of small pieces of ore from 3 to 6 inches thick, and over the whole ore dust is thrown. 5 Roasting by these methods must necessarily be more or less irre- gular : in some parts the ore is but little changed, in others it is pro- perly roasted, while in others again it may be sintered or even run together. The roasted ore is broken in pieces not exceeding the fist in size ; and the soft ore is roasted a second time, which requires three or four weeks. The heaps are prepared in the same manner as in the first operation. The twice roasted ore is broken in pieces of the size of hens' eggs, and is then ready for the next operation : it should retain sufficient sulphur to furnish, when smelted, a regulus yielding from 20 to 30 per cent, of copper. Roasting is a nice operation, and requires considerable attention, which, as at some old works at Fahlun, it does not always receive ; but in other localities, where it is properly 5 Ahrend, op. cit. 2 D 402 COPPER-SMELTING IN SWEDEN. conducted, it has been found that the care bestowed upon it is repaid by a better yield of copper. In former years, according to Bredberg, the roasting of the ore was very carelessly conducted : the large lumps were not properly broken up, and the heat was so irregular, that frequently at the bottom of the pile the ore was found melted into a thick, solid stratum, which required to be blasted with gunpowder, while that at the upper part of the pile had scarcely undergone any change. The consequence was that much of the zinc-blende remained unoxidized, and a refrac- tory substance, termed skumnas, rich in sulphide of zinc, was produced in the ore-furnace upon the surface of the melted regulus. It was partly removed along with the slag when the latter was lifted off. It was, in fact, a difficultly fusible regulus, in which zinc in great measure replaced iron ; and its specific gravity was less than that of the usual copper regulus. This skumnas, which frequently contained from 10 to 12 per cent, of copper, had during 50 years been thrown away as worthless, so that very large accumulations of it existed in the vicinity of the smelting-works. Within the last 20 years, owing to the improved method of roasting, the skumnas has ceased to be formed, and the copper has been extracted with profit from the old heaps of it which were broken up and assorted, when about one-third was thrown away as valueless, while the remainder, which contained on an average as much as 2'4 per cent, of copper, was retained. Bred- berg estimated that about half the copper in the ore had been allowed to escape in the skumnas. The picked skumnas was first well roasted in large heaps, and then smelted in the ore-furnace with the usual constituents of the charge. The formation of skumnas could not be prevented by merely increasing the height of the furnace, though, as has been stated, it was immediately checked by changing the system of roasting, so as to effect more perfect oxidation. 6 Formerly when badly roasted ore was introduced into the furnace, crystalline sulphide of zinc was found sublimed about 4 ft. above the twyers. The forma- tion of skumnas was entirely prevented by roasting the ore rich in zinc blende twice, and the second time with a larger proportion of charcoal. 2. Fusion of the roasted ore. The roasted ore is smelted in admixture with black-copper slags, which contain a large amount of protoxide of iron, and are produced in the fourth operation. Although the ore contains much magnetic pyrites and iron-pyrites, the present propor- tion of silica is so large that sometimes it is necessary to add carbo- nate of lime as a flux. Formerly, in the time of Bredberg, before coke was used, they could not procure so much fuel as at present, and only the richest ore was smelted ; and it was then necessary to add quartz in order to saturate the large quantity of protoxide of iron. The proportions of roasted ore and slags are so adjusted, that a regulus may be formed in which the copper shall not be under 20, nor above SO, per cent. ; and a slag, in which the oxygen of the silica shall as nearly as Bredberg, op. cit. FUSION OF THE ROASTED OKE. 403 practicable be double that of the bases. The slag consists essentially of silicate of protoxide of iron ; and experience has shown that, when it has the formula 3 FeO, 2 SiO 3 , it has the proper degree of liquidity to ensure the most complete separation of the regulus, and that it neither solidifies too rapidly in the fore-hearth, nor corrodes too powerfully the lining of the furnace, nor yet favours the accumulation of ferrugi- nous masses in the hearth to such an extent as to prevent the conti- nuance of the smelting- process during a long period without interrup- tion. Owing to the ascending current of gas containing carbonic oxide (see p. 52), and the presence of incandescent carbon, a portion of the oxide of iron in the roasted ore becomes reduced in its descent through the furnace, and consequently a ferruginous mass, sometimes of consi- derable size, is formed on the bottom of the hearth. A mass of this kind is termed " Eisensau " by German smelters. Similar masses occur in the hearths of iron-smelting furnaces, and have received different names in different localities : thus, in Wales they are called " horses," and in Staffordshire " bears." Should the roasting of the ore be carried too far, these masses would accumulate rapidly ; and in that case it would be necessary to add some raw ore to the charge of the furnace. Conversely, an increase in the quantity of sulphur in the ore, on account of imperfect roasting, will tend to prevent the formation of " bear." Other refractory lumps occur a little higher up, or follow the slag ; and these, which are occasioned by an excess of silica, may be removed by increasing the proportion of protoxide of iron in the charge by the addition of a larger quantity of black-copper slag. The ore and slags are mixed and weighed on a steel- yard at each charge ; and the number of charges is recorded by means of a peg and a board with rows of holes in it, the peg being advanced one hole at every weighing. The furnace is worked with a " nose " or slag prolongation of the twyer, from 4 to 6 inches long. When the " eye," or point of light seen through the twyer, is " dark," the nose is cleared by poking through the twyer ; and if the nose grows too fast, fuel is occasionally added without the usual charge of ore. The appearance of the " eye " affords important indications to the furnace- men. By varying the charge and fuel, the nose may be made longer or shorter at will ; and accordingly the blast may be made to enter the furnace at a greater or less distance from the twyer. It is easy to conceive how the working of the furnace should be affected by thus changing the position at which the air impinges on the fuel. By increasing the amount of fuel, the nose may be diminished or even melted off; and, conversely, it follows that an increase in the propor- tion of slag-producing materials will, cceteris paribus, tend to lengthen the nose. The regulus and slag accumulate in the hearth; and the former, being specifically heavier than the latter, constitutes the lowest stratum. About 60/ of the slag is allowed to run out into a sand-bed on the side of the fore-hearth ; this slag is considered as clean, and is thrown away : it contains from i to ^ per cent, of copper. The 2 D 2 404 COPPER-SMELTING IN SWEDEN. remaining slag, as it collects and solidifies in the fore -hearth, is taken off from time to time by means of a two-pronged iron fork, with a long iron handle suspended from a crane : the prongs are about 2J feet long and 2 feet apart. These crusts of slag must be put aside and remelted in the ore-furnace. They contain from f to li per cent, of copper. The furnace is not tapped until the hearth has become filled with regulus, which occurs at intervals of from 48 to 72 hours. The amount of regulus which runs out at a time upon a sand-bed varies from 4 to 6 tons. After two days, when the regulus has become cold, it is broken in pieces of about the size of the fist. It is purposely allowed to run over a large surface, so that it may be obtained suffi- ciently thin to be easily broken in pieces suitable for roasting. The blast being shut off, the tapping is effected by driving with a hammer the point of an iron bar into the tap-hole. The tap hole is stopped by driving in a piece of wood from 6 to 8 feet long and from 2 to 3 inches thick, and ramming sand round it : after the lapse of an hour the wood may be removed with safet} r , when the carbonized end will remain in the hole and act as a plug. After the tapping, before the blast has again been turned on, the fore-hearth, breast, and adjacent internal parts of the furnace are cleared from incrusting matter. If the fore- hearth be much injured, it must be repaired with talcose schist, cla}^, and brasque. The oxide of zinc in the roasted ore is reduced, and the vapour of zinc, as it rises through the upper part of the furnace, becomes more or less oxidized, and forms an incrustation of furnace-calamine round the interior. Every month this must be detached and removed to the depth of 5 or 6 feet from the mouth. In the course of working, the hearth becomes gradually contracted by the formation of blende-like stony masses, and at length, after the lapse of from 3 to 4 months, the lower part of the shaft becomes so much injured as to make it advisable to blow out the furnace and effect the necessary repairs. The upper part may last several years. The " bear " may be detached without difficulty, and consists for the most part of sulphide of iron and zinc, with very little sulphur, and from 6 to 12 per cent, of copper. The weight of the regulus varies from 17 to 20 per ce.nt., or even more, of that of the charge. Of late the copper in the regulus has ranged between 20 and 22 per cent. ; and in 1859 it did not exceed 18 per cent. In 1860 the ores became still poorer; but there appears reason to believe that this will be only temporary. The slags are stated to contain on an average not more than about J per cent, of copper. The fuel hitherto employed was charcoal, which of late has been replaced with great advantage by English coke. It has been ascer- tained by exact trials that in effect 1 cwt. of coke is equal to 2 cwt. of charcoal. This is attributed to the fact that, as coke has a higher specific gravity than charcoal, it yields bulk for bulk more heat than charcoal. To smelt 4 tons of the materials mixed in the pro- portions previously specified, 1 ton of charcoal is required ; but 1 ton ROASTING OF REGULUS FUSION FOR BLACK-COPPER. 405 of a mixture of charcoal and coke will smelt 5 tons of the same materials. 3. Roasting of the regulus from the last operation. This is effected in small kilns contained within what is called the Roast-house. In the middle of this house, and extending from one end to the other, is a wall 5 feet high and 3 feet thick, adjoining which on each side are 1 3 walls, built at right angles, of the same height as the median wall, 11 feet long, arid from 2 to 2J feet thick: there are thus formed 24 compartments, or kilns, 12 on each side, open in front, each of which in the clear is 11 feet long, from 4 to 5 feet broad, and 5 feet high. The bottom is made of a mixture of ore dust and clay, 6 inches thick, so that the height above this bottom is 4 feet 6 inches. In charging a kiln, wood is spread to the height of 8 inches over the whole of the bottom, which should be level : the broken-up regulus is piled on the wood without any charcoal in the first instance. Each kiln contains about 100 Swedish cwt. (between 4 and 5 tons) of coarse regulus. The roasting may now proceed. The process is repeated four, five, or even six times. When the 1st "firing" is finished, the regulus is " turned over," or transferred to an adjoining kiln, in which it receives a 2nd firing, and so in succession until it is roasted as "dead" or "sweet" as practicable, which may not occur until after the 6th firing. In the 1st firing no charcoal is put in; in the 2nd firing li measure of charcoal is spread over the wood ; in the 3rd firing 5 measures are introduced, partly over the wood and partly forming a layer in the midst of the heap ; in the 4th firing 8 measures and in the 5th firing 10 measures are applied in a similar manner ; and in the 6th firing 12 measures of charcoal are used, forming 3 layers, including that upon the wood. The entire roasting generally lasts from 7 weeks to 2 months. At each " turning over," the regulus is somewhat broken up, and the kiln charged as in the first instance. As the regulus contains sulphide of zinc, oxide of zinc is formed during roasting, and deposited on the pieces of regulus in the kiln. In the roasting both of copper ore and regulus, it is a remarkable fact that the copper becomes concentrated in the interior of the pieces, forming, as it were, kernels, which, as well in appearance as in the proportion of copper which they contain, frequently resemble purple copper-ore (3 Cu 2 S-f Fe 2 S 3 ). A special process will hereafter be described, which is founded on this principle of concentration. 4. Fusion for black-copper. The roasted regulus from the last opera- tion is smelted in the black-copper furnace in admixture with a roasted regulus, accompanying the formation of black-copper, with refinery slags, furnace residua containing copper, and ore-furnace slags ; and, when more silica is necessary, quartz is also added. The proportions of these matters must necessarily vary with the nature of the ore at the time. An actual charge was composed as follows : Roasted regulus, 200 Ibs. ; roasted thin regulus (see p. 406), which is not always added, 40 Ibs. ; ore-furnace slags, 20 Ibs.; cupriferous products, 20 Ibs. ; quartz, from 10 to 20 Ibs. The furnace should be worked with a nose from 4 to 6 406 COPPER-SMELTING IN SWEDEN. inches long. The breast of the furnace is stopped with sand, only a slight opening being left through which the flame blows out. Three products collect in the hearth in the following order of super-position from the bottom, namely, black-copper, thin regulus containing from 55 to 72 per cent, of copper, and black-copper slag. With 55 per cent, of copper this regulus is described as red-blue, and with 72 per cent, as steel-grey : it plays an important part in protecting the subjacent copper from waste by oxidation, and tends to effect the separation of any copper which may exist in an oxidized state in the slag. It is this rich regulus which, after roasting, is added to the charge in the present operation. The yield of black-copper may vary from 20 to 30 per cent, of the weight of the charge composed as above stated, but of late it has not reached 20 per cent. Slack-copper receives its name on account of its superficial black coating of oxide. W 7 hen the melted matter in the hearth has accumulated to such an extent as to reach the opening from which the flame proceeds, and to rise nearly up to the twyers, the slag is tapped off. This is done by removing the sand with an iron shovel from the breast or fore-part of the furnace, to the depth of 6 or 8 inches below the opening from which the flame issues, when a copious stream of slag flows into cavities in sand pre- viously moulded to receive it. The breast is then closed with sand as before, and the further escape of slag prevented. The slag is subse- quently allowed to escape from time to time as may be required. Meanwhile, black-copper and thin-regulus continue to accumulate in the hearth, until at length they must be removed. Before the final tap- ping, the slag and regulus are drawn off quite close to the surface of the black-copper in the manner just described. After this the tap-hole is opened, when the black-copper runs out into moulds of iron, by which means it is obtained in blocks : the tapping is repeated at intervals of two or three days. The black-copper is followed by regulus and slag, which run into sand beds. After tapping the blast is stopped in order to enable the smelter to cleanse the hearth. The slag, under which regulus may have collected, is tilted over and broken as soon as it is set, and while still red-hot, so as to separate the regulus as com- pletely as practicable. The slag is rich in oxide of iron, and therefore strongly attacks the lining of the furnace : it contains some copper, and is, as has been stated, remelted in the ore-furnace, in order that this copper may be extracted, and that the necessary amount of oxide of iron may be presented to the silica in the ore. The bear, which forms on the hearth-bottom, has a red colour, and consists chiefly of iron, but is rich in copper. The thin-regulus is broken up, and mixed with the ore-furnace regulus in the third firing, so that it receives only four firings instead of six : but Bredberg states that it is advisable to roast this regulus by itself. The bear is heated before the blast of a single twyer in a small hearth open on three sides, when the copper is liquated and the iron oxidized. 5. Refining. The hearth being filled up level with charcoal, thin pigs of black-copper are first placed on each side, with their ends pro- REFINING. 407 jecting several inches into the cavity, and then three pigs are piled one upon another across the hearth, not too near the twyer. From 13 to 15 cwt. of pigs are refined at a time. Copper residua from previous operations are afterwards added. Ignited charcoal is put in front of the twyer, and cold charcoal thrown on until the black-copper is covered with it. The blast is then gradually turned on, so that the metal may slowly melt down. Owing to the great inclination of the blast-pipe, the melted metal is exposed to oxidation, and continues in lively motion. At first the slag has a brownish-black 'colour, and often a blue tinge from the presence of cobalt ; but by degrees it becomes red, in consequence of the formation of a large quantity of dioxide of copper, just as in refining in the reverberatory furnace. The flame acquires a pure and deep green colour ; but it is not so variegated at the com- mencement as in the Lower Harz, and metallic vapour is scarcely ever perceived. The surface of the melted copper must be kept covered with charcoal. The addition of a little broken quartz is occa- sionally necessary, in order that the slag may have the proper consis- tency. As no lead is present, a pasty slag mixed with pieces rich in protoxide of iron is formed, which is not sufficiently thin to flow away, as is the case in some Continental refineries, where impure copper containing lead is operated on: the slag is, therefore, removed by skimming twice, and occasionally, when the hearth is large and deep, three times in the course of one refining operation. The skimming is effected as follows : The blast is stopped, all the charcoal is taken out of the hearth, and water is sprinkled over the slag, which being thus rendered solid, may be readily lifted off in the form of a crust. The condition of the melted copper is ascertained by taking out small portions at intervals during the progress of the operation in the same manner as in various copper- works on the Continent. For this pur- pose a cylindrical piece of iron, with a clean surface and rounded end, is employed : it is plunged through the twyer into the melted copper, quickly withdrawn, and immersed in cold water ; the coating of copper which it may thus have received is knocked oif and examined. This coating, or trial-piece of copper, is in the form of a hollow cylinder closed at one end, about 3 inches long and f inch in dia- meter : it is called " Gaarspahn " by the German smelters. If this trial-piece is thick, smooth on the outer surface, and yellowish-red in the interior, the copper is " too young," and must be further exposed to the action of the blast. When it becomes thin, brownish-red, and crinkled on the outer surface, of a pure copper-red in the interior with metallic lustre, and may be bent several times without breaking, the copper may be regarded as refined, or very nearly so. \\hen it becomes so thin as no longer to form a continuous coating, but merely to sur- round the iron in some places like net-work, and in others to present the appearance of small pointed or bearded excrescences, the blast should be immediately stopped, as the copper is now refined, or " gaar." If the process is continued beyond this point, the outer surface of the trial-piece becomes dull, and acquires a brown-red or reddish brown colour, and the copper cannot be bent without breaking ; it is then said 408 COPPER-SMELTING IN SWEDEN. to be " uebergaar," or dry. When copper in the hearth is "too young," its surface, freed from slag, appears perfectly tranquil ; but in the "dry" state, its surface after removal of the slag presents the appearance of ebullition. At Atvidaberg, owing to the comparative purity of the copper, the process is sooner concluded than in other localities, where copper containing lead, and it may be other foreign metals, is operated upon ; for although the trial-piece may still in a certain degree present the appearances indicative of "too young" copper, namely, a brass-yellow tint and a strong metallic lustre in the interior, the cast copper may, nevertheless, be pure (practically ?), and indeed, somewhat over-refined (uebergaar). At Atvidaberg the black- copper is sufficiently pure to admit of being refined and toughened at one operation. When the pitch is right, the metal is laded into iron ingot-moulds in the same manner as at Swansea. The following analyses of the black and refined copper (Gaarkupfer) of Atvidaberg were made at the Mining School at Fahlun : Black-copper. Gaarkupfer. Copper 94-39 99-460 Iron 2-04 O'Oll Zinc 1-55 Cobalt and nickel 0'63 O'llO Tin 0-07 scarcely a trace. Lead 0-19 do. Silver ' 0-11 0-065 Gold not looked for...., 0-0015 Sulphur 0-80 0-017 Arsenic trace 99-78 99-6645 Oxygen not determined. The greater part of the cobalt is separated during the process of refining, and becomes concentrated in the refinery slag. Generally at copper-works on the Continent where the refining- hearth (Gaarheerd) continues to be used, the copper is first refined and afterwards remelted in the same kind of hearth in order that it may be rendered malleable (hammergaargemacht), and suitable for rolling or hammering; but in this hearth the twyer is not inclined, whereby oxidation is prevented. The hearths are generally much smaller than those of Sweden, and the black-copper is often very impure as com- pared with the Swedish. The process of refining is conducted pre- cisely in the manner described. When much lead is present the slag is liquid, and may be allowed to run off by a channel cut in the side of the hearth. W 7 hen the trial-piece indicates that the copper is refined, or " gaar," the charcoal is immediately removed, the metal skimmed as clean as practicable, and allowed to cool down to a certain degree, when water is sprinkled over its surface. A superficial solid crust of copper is thus formed, which is lifted off and plunged into cold water. A second crust is produced in the same way and removed ; and this operation is repeated until nearly the whole of the copper is withdrawn from the hearth. The first crusts being generally some- REFINING TOUGHENING. 409 what dirty on the upper surface from adherent matter, they are remelted in the next refining operation. These round crusts or discs of copper are known in commerce as " rosette-copper." They are smooth on the upper surface, with which the water came directly in contact ; but underneath, where they were in contact with the liquid metal, they are rough and covered all over with excrescences. Around the circumference is a border directed downwards when the disc is held in the position in which it was formed in the hearth. It is found by experience that when the trial-piece presents the indications cha- racteristic of refined, or " gaar," copper, discs may be obtained having the greatest degree of thinness and presenting a rich red coloiir, cha- racters which are regarded by purchasers as unerring signs of the purity of the copper. The process of refining is not continued so long as to render the copper very dry, or " uebergaar," but is arrested at a point when the metal may be taken off in the thinnest discs, although it may, as Karsten observes, still retain a sensible amount of foreign metals, which, by a more prolonged exposure to oxidation in the hearth, might, in great measure, be separated. In the dry, or " ueber- gaar," state thin discs cannot be obtained, and yet this may be solely due to the presence of a large quantity of dioxide of copper, and not in any degree to the presence of foreign metals, or impurities, properly so called ; whereas thinner discs, containing only a small proportion of dioxide of copper, may, nevertheless, consist of copper in which a very sensible amount of foreign metals may exist. It is only when the trial-piece indicates that the copper possesses the qualities character- istic of what is termed "gaar" that discs of approved thinness and colour can be obtained. These qualities afford unequivocal signs that the copper has undergone a considerable degree of refining, and, in so far, they may be valuable, but they become fallacious when regarded as sure tests of the degree of purity of the metal. Thin discs cannot be obtained with the purest copper in any state except at a proper degree of heat. Karsten remarks, that " in most European states refined copper (Gaarkupfer) is not cast into bars, but is sent to market in the form of discs ; and this practice, once introduced, has led persons to pay attention to the external appearance of the copper rather than its intrinsic excellence. It is required that the discs of copper should be as thin as possible, and possess a fine deep red colour. These qualities are only possessed by absolutely pure copper. The more they are sought for in impure copper the less perfectly will it be freed from foreign matters, and the less will be the tenacity of the metal produced in the process of toughening (hammergaarmachen)." 7 In the toughening process the discs are slowly melted down under a very gentle blast from a twyer a little less inclined than in the refining process. The form of disc, as compared with that of ingot, is advantageous, as the copper does not press heavily on the charcoal and drop through unmelted, but fuses easily and gradually in every part. When melted it is in the " too young/' or overpoled, state (see Sys. 5. p. 38G. 410 COPPER-SMELTING IN SWEDEN. p. 269), but will afterwards pass to the state of tough-pitch by the effect of oxidation, just as, in the Welsh process, in the case of over- poling. Trial-pieces are taken out from time to time in the manner before described, and when the metal has acquired the maximum of toughness it is immediately laded into ingot-moulds. According to Karsten there is a temperature at which it may be cast without " rising" in the mould, but this is so difficult to be attained that it is usual, before lading, to add from a quarter to one-third per cent, of lead, which prevents the rising. The same author also states that the addition of lead renders the copper unsuitable for drawing into fine wire and for plating with gold. When required for these purposes the copper must be allowed to sink in temperature to precisely the right degree, and must then be instantly cast into ingot-moulds. 8 At Atvi- daberg lead is never used in the toughening process, either in the hearth or reverberatory furnace. Copper rain. During the time the temperature of the melted copper is allowed to sink to the proper degree, before the commencement of the operation of solidifying the surface with water and taking off the discs, the singular phenomenon called copper rain occurs. Minute spherical particles or shots of copper are projected from the surface of the melted metal on all sides, occasionally, according to Karsten, to the height of 4 feet, and this continues until the surface has become solid. To prevent loss of copper from this cause it is necessary to cover the hearth with sheet-iron in order that the shots may fall back into the melted metal. The same kind of copper rain is produced in the Welsh refining-furnace. In a specimen of it which I have received from Mr. Edmond the particles are pretty uniform in size, and are much less than the smallest pin's head. Karsten gives the following account of the rain and the conditions under which it appears. " The particles vary much in size, and sometimes exceed that of a pin's head, and the higher they are thrown up the larger they are. In other cases a dense red-coloured vapour appears at a short distance above the surface of the melted copper : it consists of innumerable small rotat- ing shots of copper, having a nucleus of metallic copper. The slight coating of dioxide with which the small shots are surrounded is pro- duced by the action of atmospheric oxygen at the moment of their pro- jection. Copper which produces strong copper rain contains but little dioxide ; often so little as not to admit of accurate determination. The red dust appears when the copper contains a larger proportion of di- oxide. Copper containing from 0'7 to O8 per cent, of dioxide still exhibits this appearance in a very perceptible degree. When the pro- portion of dioxide is still further increased the dust vanishes and the surface of the copper sets quite tranquilly. It seems that 1-25 per cent, of dioxide is necessary to prevent the formation of the rain. All impure copper does not present this phenomenon, especially when it has not taken up any dioxide the condition in which the strongest rain is evolved from pure copper. All copper which produces rain in Sys. 5. p. 405. COPPER-SMELTING AT RORAAS IN NORWAY. 411 a remarkable degree is in the malleable state, and might, therefore, be directly cast into ingot-moulds. But copper in this state rises in the mould if it be not poured at exactly the right temperature." 9 This spitting of copper has been discussed in a former part of this work. It occurs when commercial copper is overpoled. Mr. Grill informs me (1861) that a reverberatory furnace has just been introduced at Atvidaberg in order to conduct the operation of refining on the Welsh system, and that it is reported to answer ex- tremely well. It is the invention of Dr. C. Th. Boettger, of the Mans- feld Copper- works ; it is constructed for the use of charcoal with a view to economize fuel, and has a very high arch. Consumption of fuel. In making one ton of copper at Atvidaberg in 1858 were consumed on an average 240 cubic feet of pine-wood, 7 tons of charcoal, and 3*8 tons of coke. In the table (p. 412) will be found the results of copper-smelting at Atvidaberg in 1859. It is an exact copy and literal translation of that which was prepared at the works, with the addition of the English weights. Loss in smelting. Excepting what may be volatilized, the only loss of copper is in the ore-furnace slags, which are thrown away. Mr. Malmqvist estimates the quantity of these slags to be about 50 per cent, of the raw materials, and that, inclusive of the admixture of foul slags, they may be regarded as containing -| per cent, of copper, in which case the total loss would be about j per cent, of the weight of the raw materials. ' Cost of smelting. This will be subsequently given. Copper-smelting at Roraas in Norway. The process is conducted on the same system as in Sweden, and although the ores are richer, yet the yield is not so good as in Sweden. Professor Eggertz, of the Mining School at Fahlun, has published a complete description of the Roraas Works, together with analyses of the products, of which the following is a selection : x Ore-furnace regulus. 1. 2. Copper 22-03 20-11 Iron 52-14 52-40 Sulplmr 25-15 24-72 Insoluble residue... 3-00 320 102-32 100-43 1 . Smelted with hot blast. 2. Smelted with cold blast. 9 Sys. 5. p. 391. 1 -I( ru-Kontorets Annaler for 1849, , 275. Duchanoy has also published a description of these works, which he visited in 1852. Vid. Ann. dcs Mines, 5* me s. 5. p. ]81. 412 RESULTS OF COPPER-SMELTING AT ATVIDABERG. Q pH ^ .P E-j '3 6 J o 5 J m X *; it SJ5 :^s 3 il Ull i ;-- - ~ ic^? '- a II a S g S 8 : S ! I 8 !I I 1 Jf.H i HI ill C C.3 II Is ll "S3: 111 IW ll 68 ill !!!! sj i 31% &^- II I ? 2 SoS leW 5 H i :(i SMELTING OF COPPER-SCHIST IN PRUSSIAN SAXONY. 413 Ore-furnace slag. Black -copper slag. Silica r 1. 28-48 .. 2. ... 28-18 i 3. 30-85 -s 4. 23*95 Alumina 9-58 .. 9-46 4-00 ... . . 4-98 Protoxide of iron . 50-99 .. ... 51-17 66-25 70-74 1-18 1-41 0-47 0-42 Ma^nesiu .... 11-55 .. ... 11-35 Copper 0-38 .. 0*60 0'63 1-06 102-16 102 17 102-20 101-15 Nos. 1 and 3 smelted with cold blast. Nos. 2 and 4 smelted with hot blast. In Nos. 3 and 4 the magnesia was estimated with the iron. The excess is attributed to the use of water which had not been distilled. SMELTING OF COPPER-SCHIST IN THE DISTRICT OF M ANSFELD, PRUSSIAN SAXONY. This schist is the well-known Kupferschiefer of the Germans, and appears to be the equivalent of the marl-slate of English geologists. It occurs in the Permian series, formerly known as the Lower New Eecl Sandstone. It is a thin dark-coloured bed of a schistose or slaty structure, lying upon quartzose conglomerate (Rothliegende), and over- laid by limestone. By the miners it is usually divided into four strata, which have received different names in different localities. Thus at Eisleben, in the order of their downward succession, they are designated Noberge, Schieferkopf, Kammscliale, and Letten. Overlying the uppermost of these strata, the Koberge, is a bed called the Dach, which, though consisting chiefly of limestone, nevertheless forms part of this schistose formation, and is sometimes included, as at Sanger- hausen, in the term Kupferschiefer. The bed of copper -schist proper varies in thickness from 10 to 20 inches, of which only from 3 to 5 inches are worth smelting. It con- sists chiefly of clay, silica, and limestone, and contains oxide of iron, black bituminous matter, and water. Copper exists in the schist chiefly in the state of vitreous copper (Cu 2 S) and purple copper-ore (3Cu' 2 S-fFe 2 S 3 ), but it also occurs as copper pyrites, grey copper-ore, black copper-ore a mixture of protoxide of copper with oxide of iron and manganese red copper-ore, and native copper. The schist is not only cupriferous, but argentiferous, and generally in a sufficient de- gree to allow the silver to be extracted with profit. The silver is rarely met with in the metallic state. The following minerals have also been found in the schist, either as constant or as occasional con- stituents : iron-pyrites, blende, rarely galena, kupfernickel, iron- ochre (a mixture containing sesquioxide of iron), nickel-ochre, smal- tine or tin- white cobalt (Glanz-cobalt, consisting essentially of cobalt, arsenic, and sulphur), red earthy cobalt (a mixture containing oxide of cobalt), sulphide of molybdenum, and, very rarely, native antimony, bismuth, and arsenic, and, lastly, according to Kersten, vanadium. The schist of any one particular locality may contain only a portion of 414 SMELTING OF COPPER-SCHIST IN PRUSSIAN SAXONY. the minerals above mentioned. 2 The metallic minerals are sometimes so finely disseminated as to be imperceptible ; or they occur in little veins, in thin laminse, in nests, or in nodules. Fossil remains of fishes and plants are met with in copper-schist, and, less frequently, those of mollusca. The sandstone upon which the copper-schist immediately rests is sufficiently cupriferous to the depth of two or three inches to be extracted with advantage. It is called sand-ore (Sanderz), and con- tains vitreous-copper, purple-copper, copper-pyrites, blue and green carbonate of copper in small rounded nodules, rarely native copper, iron-p}a'ites, blende, galena, kupfernickel, and native bismuth ; but of these minerals the most frequent are vitreous-copper and copper- pyrites. The particles of quartz composing the sand-ore are cemented together either by argillaceous or calcareous matter. The proportion of copper in the sand-ore decreases downwards so rapidly that at San- gerhausen, whilst the uppermost layer of about the thickness of half an inch yields 12 Ibs. of copper per centner, that at about 2 or 3 inches below yields only 2 Ibs. per centner. So long as the sand-ore contains 4 Ibs. of copper per centner it is broken into pieces' of from one to half a cubic inch in size and smelted. Poorer sand-ore is dressed preparatory to smelting. The general average of copper in all the sand-ore smelted is about 6 Ibs. of copper per centner. 3 A speci- men from Sangerhausen in my collection, labelled " Sanderz," consists of a layer of dark grey earthy matter to which is attached one of iron-pyrites. By digestion in nitric acid it was partially dissolved : the washed and ignited insoluble residue amounted to 43*92 per cent. ; it was nearly white, very gritty to the touch, presented minute par- ticles of mica, and evidently consisted almost entirely of siliceous 'sand. Copper and iron in considerable quantity were found in the solution. Another specimen -was digested in nitro-hydrochloric acid without having been previously reduced to powder, when rounded particles, apparently of colourless quartz, were observed in the insoluble residue. According to Heine the Noberge at Sangerhausen consists chiefly of carbonate of lime, with scarcely any silica, and only very little clay ; and the copper which it contains exists for the most part as grey copper-ore (Speise), but vitreous copper in grains is also present. A specimen of Noberge from Sangerhausen in my collection is very dark grey in colour, and distinctly stratified. It effervesced on the addition of hydrochloric acid and partially dissolved. The washed residue, after having been heated to redness in a covered crucible, amounted to 55-42 per cent., which was reduced to 53-42 per cent, by subsequent ignition in an open crucible. The solution contained chiefly lime, with a little copper and iron. The ignited residue was pale brown, gritty to the touch, and contained minute scales of mica. Copper-schist is worth smelting when it does not contain less than 2 Ibs. of copper per centner. 4 The Dach at Sangerhausen, according to Heine, is almost entirely 2 Vide Die Lelire von den Erzlager- j 3 Heine, Ann. derPhys. u. Ohem. Pog- statten, Bernhard Cotta. Freiberg, 1855, I gendorff, 1835, 34, p. 531. Cotta, op. cit. p. 234 et seq. * Heine, op. cit. SMELTING OF COPPER-SCHIST IN PRUSSIAN SAXONY. 415 composed of carbonate of lime, and, as a general rule, contains only grains of vitreous-copper; but it is also stated to contain copper- pyrites, purple copper-ore, red copper-ore, malachite, rarely blue car- bonate of copper, iron-pyrites, and some galena. 5 A specimen in my collection from Sangerhausen, labelled " Oberberge," 6 is com- pact and brownish grey, like an ordinary clay iron-ore. It contains irregularly diffused small patches and minute particles resembling vitreous copper ; it effervesced considerably on the addition of hydro- chloric acid, and partially dissolved : the washed and ignited insoluble residue amounted to 35'32 per cent. Lime, magnesia (in considerable quantity), iron, and copper were found in the filtrate. The smelting of this cupriferous schist has been carried on at least during several centuries. Agricola, writing about the middle of the 16th century, minutely described the manner in which, at Eisleben and in the neighbourhood, it was burned in heaps preparatory to smelting. 7 The existing smelting-works are situate in the vicinity of the towns of Mansfeld, Eisleben, and Sangerhausen. Owing to the recent introduction of new methods of extracting silver, this metal, as well as copper, is now obtained from all the copper-schist raised in the district ; whereas formerly it could not be extracted with profit from the schist near Sangerhausen, which contains less than that of the other localities mentioned, and which, consequently, was smelted for copper alone. The smelting of sulphuretted ores of copper in the blast- furnace has always been conducted in essentially the same manner as at Atvidaberg, so that it will not be necessary to enter into much further descriptive detail upon the subject. The process formerly practised at Sangerhausen and some other localities consisted of the following operations : 8 L Eoasting of the copper-schist. The ores which did not contain much bituminous matter were not roasted. Some of the schists were sufficiently bitu- minous to continue to burn of themselves after having been once ignited. 2. Fusion of the roasted ore with slags and fluor-spar. The products were a regulus, containing from about 30 to 35 per cent, of copper and poor slag. 3. Roasting of the regulus of No. 2 in three successive fires. 4. Fusion of the roasted regulus with slags. The products were a concentrated regulus (Spurstein), containing from about 50 to 60 per cent, of copper, and a slag (Spurschlacke) contain- ing from about If to 2 per cent, of copper, which was resmelted. 5. The regulus of No. 4 was roasted in seven successive fires. 6. Fusion of the roasted concentrated regulus of No. 5 with slags. The products were black copper and a rich thin regulus (Diinnstein) containing about 63 per cent, of copper, which was roasted along with the regulus of No. 5 in the last three fires. 7. Refining of the black copper. The products were rosette copper, and slags or skimmings rich in 5 Cotta, op. cit. 6 A distinct bed forming the upper part of the Dach at Saugerhausen, but first edition is 1555. 8 Vide Notice sur les Mines de Schiste Cuivreux et sur les Usines du Pays de not appearing in other districts. Mansfeld. Par M. Manes. Ann. d. Mines, 7 Georgii Agricolse De Re Metallica. j 1824, 1. ser. 0, p. 1 et seq. Basilese, 1561, p. 218. The date of the 416 SMELTING OF COPPER-SCHIST IN PKUSSIAN SAXONY. dioxide of copper. 8. Kemelting of the rosette copper and toughening (Hammergaarmachen). The products were malleable or tough copper, and slags or skimmings rich in dioxide of copper. At Mansfeld and Eisleben the regulus produced in operation No. 2 of the Sangerhausen process was roasted six times successively ; and the roasted product (Gaarrost) was smelted with slags, when black- copper, accompanied as usual with thin regulus, was obtained. At some establishments the regulus was washed with water after each roasting, except the first and last, so that any sulphate of copper present in it might be dissolved out and crystallized. The silver became concentrated in the black-copper, from which it was separated by the ancient process of liquation with lead hereafter to be described. 9 In 1831 this process was discontinued, and the silver was extracted from a concentrated regulus by means of mercury. More than twenty years afterwards the use of mercury was abandoned, and the regulus was desilverized by the wet method of Augustin, which, after three years, was replaced by that of Ziervogel. At the present time (1861) the process of desilverization is only conducted at the Gottesbelohmmghiitte, near Hettstadt ; so that at the various smelting-works in the vicinity the smelting of copper-schist is not carried further than the production of a regulus, which is sent to the establishment above-mentioned. In illustration of the treatment to which the copper-schist is now subjected, I insert the following description of the practice at the Kreuz- hiitte, near Hettstadt, in 1860 : 1. The schist contains about 200 Ibs. (Prussian) of copper per Fuder, i. e. about 3 per cent. : it is roasted in pyramidal heaps containing from 6000 to 21000 ctrs. each ; the time required for this process varies from six weeks to three months. 2. The roasted schist is smelted in blast-furnaces about 20 feet high in admixture with fluor-spar, slags accompanying the formation of con- centrated regulus and black-copper, and desilverized residua. The furnace is cylindrical from the mouth down to the top of the hearth, of which the section is square. The dimensions (Prussian measure) are as follow : diameter at the top, 3 ft. ; below the top, at 10 ft, 6 in., 5 ft. ; at 11 ft. 6 in., 5 ft. 8 in. ; at 14 ft., 3 ft. 9 in. ; at the bottom, 2 ft. 6 in. total height, 19 ft. There are two twyers, opposite each other, one on each side, at 2 ft, 2 in. from the bottom ; the diameter of the nozzles is from 1 J to 2 in. The bottom of the hearth inclines forwards ; and in front are two pits, of which the vertical section is that of an inverted cone ; they are about 18 in. in diameter at the top, and 2 ft. deep. The melted matter is not left to accumulate in the hearth of the furnace, but flows out continuously, first into one of these conical pits and then into the other, the two pits being alternately filled and emptied. There are, consequently, two aper- tures at the bottom of the furnace, on a level with the lowest part of the hearth, one on each side corresponding to each pit ; they are called "eyes "by the Germans, and are kept alternately closed and open. The pits in front have been fancifully compared to spectacles, whence the name " Brillenofen," or spectacle-furnace. 'The pressure of the De la Kichesse Minerale, Heron de Villefosse. Paris, 1819, 3. p. 345 et scq. SMELTING OF COPPER-SCHIST IN PRUSSIAN SAXONY. 417 blast is equal to that of a column of water from 8 to 10 inches in height. The charge of the furnace is composed as follows : Roasted schist, 2 Fuders (1 Fuder = 60 ctrs. = about 3 tons English), fluor-spar 8 ctrs., concentration-slag 6 ctrs., and any residua (Gekratz) which may be at hand. The fuel is English coke at the rate of 5 tons measure (1 ton Prussian = 7 cub. feet Prussian) per Fuder : light coke weighs 150 Ibs. and heavy coke 200 Ibs. (Prussian) per ton. The products are ore- furnace regulus (Eohstein), containing from 30 to 35 per cent, of copper, and ore-furnace slag (Rohschlacke), which is stated not to contain more than from 2 to 5 loths of copper per ctr., i.e., from 0'06 to 0-15 per cent. 3. The ore-furnace regulus is roasted in three- walled kilns, similar to those at Atvidaberg previously described : after the first roasting, the contents of the kiln are > turned over, broken up, and such pieces as require to be roasted a second time are put aside for that purpose ; the greater part of the ore is usually roasted twice. It is hardly necessary to remark that care must be taken to leave suffi- cient sulphur in the roasted regulus to prevent the separation of metallic copper in the subsequent fusion. 4. The roasted regulus of No. 3 is smelted in a reverberatory furnace. The products are con- centrated regulus (Spurstein), which is tapped into water and granu- lated as in the Welsh process, and slag which is skimmed off at once into small waggons and wheeled away. The charge is, 42 ctrs. of roasted regulus, of which from i to \ has been only once roasted (one- fire Rohstein), and about 4 ctrs. of sand. It is introduced into the furnace in ladles through two doors, one at the end opposite the fire- bridge, and the other in the side opposite the tap-hole. The operation is conducted like the No. 4 fusion at Swansea. The slag is drawn out in from seven to eight hours, when a second charge of 42 ctrs. is introduced. The products are concentrated regulus (Spurstein) and concentration-slag : the regulus contains from 60 to 75 per cent, of copper, and amounts to about half the weight of the ore-furnace regulus. The average time required for the fusion of two charges of 42 ctrs. each, including charging, rabbling, tapping, &c., is about 19 hours. The fuel employed consists of equal parts of brown coal, which occurs in the neighbourhood, and English coal imported into Hamburgh. The regulus is stamped and ground, and in that state sent to the Gottesbelohnunghiitte to be desilverized. At the Kreuz- hiitte there were four blast furnaces, and one reverberatory furnace. 9 After the extraction of the silver from the regulus at the Gottes- belohnunghiitte, the pulverulent residue in which the copper exists in the state of oxide is smelted. It is kneaded with from 5 to 10 per 9 The ore-furnace slag is employed for pot. The lump of slag thus prepared domestic purposes as a source of heat, is placed on two iron bars in a wheel- A lump of red-hot slag is taken off and barrow, and taken away by women to the a stick is thrust into it, when, owing to neighbouring huts. By this means veget- the evolution of gases from the wood, it ! ables are cooked, coffee is boiled, &c. ; swells out, forming a hollow ball. The and in winter the rooms are heated by stick is then pulled out, and the hole left j hot slag, is made large enough to receive a culinary 2 E 418 ANALYSES OF THE MANSFELl} SCHIST. cent, of clay, and sufficient water to form a coherent mass, which is moulded by hand into balls from 3 to 4 inches in diameter. These are dried in hot chambers, and then smelted in a blast-furnace in admixture with ore-furnace slags, stamped quartz, and some thin- regulus ; or failing this, some iron-pyrites free from silver may be added. It is essential that a certain amount of regulus should be present, in order that the slag may be as free from oxide of copper as possible. If a reverberatory furnace be employed, the cupriferous residue may, in admixture with carbonaceous matter, be smelted without the addition either of flux or regulus. 1 Close to the Kreuzhiitte are the new smelting works called Eck- hardtshiitte, which were on the point of completion in September, 1860. They are stated to be well lighted, more roomy, better venti- lated, and in style much superior to the old establishments. There were four large blast-furnaces, of which two only were in blast, and two small ones; and a reverberatory furnace was in course of construction. The blast was produced by a fan worked by steam- power. The pressure was equal to a column of water of 4 inches in height, and the internal diameter of the nozzles of the blast-pipe was from 2j- to 3 inches. At the Oberhiitte the blast is obtained from a Cagniardelle, or machine constructed on the principle of the Archi- medes screw : it is stated to produce a blast equal to a column of water 14 inches in height. Analytical data in elucidation of the smelting of copper-schist in the district of Mansfeld : ANALYSES OF THE MANSFELD SCHIST BY 1. Unburnt. Silica 40-0 Alumina 10*7 Oxide of iron ( Fe 2 O 3 ) . . . . 5 Carbonate of lime 19'5 ,, magnesia... 6 '5 Potash 2*0 Water and bitumen . . 10-3 Roasted. 50-6 Silica Alumina ) 23 . 4 Magnesia } " Lime ..................... 7'8 Oxide of copper (CuO) 2-8 ,, iron(Fe 2 O 3 ) 9-0 Sulphur .................. 4*0 Loss by calcination. . . 0-8 3. Roasted. 43-8 17-2 18-0 2-5 7-2 2-4 6-0 100-0 98-4 97-1 These analyses of three different specimens of the schist show that, as might be anticipated, it varies considerably in composition. Determinations of the silica, alumina, lime, magnesia, and oxide of iron in the schist from Sangerhausen and other localities in the district of Mansfeld have been made by Grunow, 3 and the results 1 I am indebted for much of the pre- ceding information to my former pupil, Mr. Foster, who visited the works Sept. 1860. I have also derived information from the MS. accompanying the collection of specimens from the Mansfeld Copper Works in the Great Exhibition of 1851-, signed C. Th. Boettger, Eisleben, and from the work entitled Die Augustin'sche Silberextraction, by A. Grutzner. Braun- schweig, 1851, p. 91 et seq. 2 Ann. des Mines, 1. s. 1824, v. 9, p. 63. 3 Handb. der Metallurg. Huttenkunde. Kerl, 1855, 2. p. 256. ANALYSES OF ORE-FURNACE REGULUS. 410 agree pretty well with those inserted above. According to Berthier, the roasted schist melts very well without any addition in a brasqued crucible, forming a slag which is compact, vitreous, free from cavities, blackish, translucent, and very tenacious : shots of regulus are produced which are magnetic, and contain both copper and iron. The same metallurgist remarks that it is evident, from the proportion of sulphur in the 2nd and 3rd analyses, that the iron and copper are present in the roasted schist chiefly in the state of sulphides, and that, consequently, it is difficult to conceive the utility of the roasting pro- cess, of which the effect is the expulsion of the bituminous matter and a portion of the carbonic acid. Fresh analyses of this product are desirable. Silica.... ANALYSES OF ORE-FURNACE REGULTJS (ROHSTEIN). 5. 47-27 19-69 26-76 Zn, Ni, 1 Co, Mnj Copper Iron L< 52-44 20-49 2. 48-25 17-35 3. 42-10 19-25 4. 31-70 28-75 Sulphur Zinc 26-44 24-58 2-90 25-50 5 20 27-80 4-35 Nickel Cobalt Lend 0-41 0-80 1-05 1 05 1-50 1-25 0-65 Silver . . . 0-18 30 0-27 0-16 6. 43 62 23 35 28-70 3-45 99-91 1-55 1-15 96-78 96-02 1-65 96-31 Carbon, earthy matter, and loss 0-88 97-81 100-00 Observations. 1-4. By Heine. 1. The Rohstein was produced at San- gerhausen in 1 831 : the amount of copper was occasionally as low as 40 per cent., when the iron was proportionately increased. In addi- tion to the constituents mentioned, traces of manganese, zinc, cobalt, nickel, antimony, and arsenic were found. The charge of the furnace consisted of 3 parts of a mixture of dach, noberge, and sand-ore to 5 parts of copper-schist, with the addition of from 10 to 20 per cent, of fluor- spar and slags from the concentration and black-copper furnaces. The fuel employed in smelting was charcoal. 2. From the Oberhiitte, Eisleben. 3. From (the Katharinenhutte ?) Mansfeld. 4. From the Kupferkammerhutte, Hettstadt. The last three analyses were made by Heine in 1844. The loss consisted of small quantities of alumina, lime, magnesia, molybdenum, phosphorus, &c. The silica, I presume, was simply in mechanical mixture. 5. By Rammelsberg. This Rohstein, of which the locality is not stated, was partially crystallized in octahedra of the cubical system, built one upon another ; its sp. gr. was 4-73. 6. By Rammelsberg. This specimen was obtained in 1833 at the Katharinenhutte, Leimbach, from the lining of the hearth (Gestiibemasse) into which the regulus had infiltrated in small 4 Ann. der Phys. u. Ohem. 1835, 34. p. 533. The other analyses are extracted from Rammelsberg's Lehrbuch der Che- mischeii Metallurgie. 224-226. Berlin, 1850, pp. 2 K 2 420 ANALYSES OF ORE-FURNACE SLAGS. particles, and had afterwards crystallized during slow cooling. Crystals thus produced are more distinct than the last: they are combinations of the octahedron with the cube, having very smooth and brilliant faces, and sharply denned edges and angles ; in colour they resemble kupfemickel, but sometimes are coated with a steel- grey tarnish. A specimen of ore-furnace regulus in my possession produced at Mansfeld many years ago is coated with a blue and green efflorescence ; its fracture is granular, and has a reddish iron-grey colour : it contains cavities near the upper surface, in some of which is moss-copper. ANALYSES OF OEE-FUENACE SLAGS. 1. 2. 3. 4. 5. 6. Silica 57-43 53-83 49-8 48-22 50-00 54-13 Alumina 7-83 4-43 12-2 16-35 15-67 10-53 Lime 23-40 33-10 19-2 19-29 20-29 19-41 Magnesia 0-87 1-67 2-4 3-23 4-37 1-79 Protoxide of iron 7-47 4-37 13-2 10-75 8-73 10-83 Oxideofzinc .. .. 1-26 1-11 .. Dioxide of copper 0'30 0-24 .. 0-75 0'67 2-03 Fluor 1-97 2'09 1-1 Alkali (KO) and loss .. 2-1 99-27 99-73 100-0 99-85 100-84 98-72 Observations. I and 2 from Sangerhausen, made by Heine in 1831. 5 No. 1 was pearl-grey, and so light and porous that it floated upon water like pumice. No. 2 was perfectly melted, glassy, and leek- green in colour. The difference between these two slags was attributed to the fact that in the production of the second a larger quantity of fluor-spar had been used than in the case of the first. 3. From Mansfeld, by Berthier, 6 who describes these slags as vitreous, translucent, and of a deep green, almost black, colour, occasionally tinted with blue. 4, 5, 6. From the Kupferkammerhiitte. The analyses were made in Eammelsberg's laboratory, the first two by Hoffmann, apparently with the same specimen, and the third by Ebbinghaus. 7 The slags were of the usual description, vitreous and dark-coloured. Ore-furnace slags in my possession from Mansfeld and Sangerhausen are vitreous and olive-black when seen in mass ; but viewed in thin pieces by transmitted light the colour is dark olive. Other speci- mens are partly vitreous and partly opaque and brown-grey ; and others from Eisleben, which are reported to contain vanadium, are vitreous, and more or less blue, like slag occasionally produced in iron-smelting furnaces. 5 Op. cit, p. 534. 6 Ann a Mines, 1 . s. 1824, 9, p. 60. 7 Op. cit., p. 226. ANALYSES OF CONCENTRATED AND THIN REGULUS. 421 ANALYSES OF CONCENTRATED-REGULUS (SPURSTEIN) AND THIN-REGULUS (DiJNNSTEIN) FROM THE BLACK-COPPER FURNACE. Only the first analysis is of concentrated-regulus. The supposed rational constitution of these specimens of regulus, as calculated by Eammelsberg, is inserted below the ultimate analyses : 1. 2. 3. 4. 5. Copper 51-37 59'8 59-18 57-27 61*23 Iron 18-67 15'8 16-07 16-32 15-19* Sulphur 24-35 22-6 20-01 22-17 24-38 Zinc, nickel, &c 6-54 .. 297 2-55 100-93 98-2 98-23 98-31 100-80 Bisulphide of copper 67-47 46 '66 57-69 77-95 Protosulphide of iron 24-75 25-11 25-56 23-80 Sulphides of zinc and nickel .. 4-44 3-81 Metallic copper 5-98 21-96 10-08 traces. * With other metals. Observations. 1. From the Kupferkammerhiitte, by Ebbinghaus in Eammelsberg's laboratory. There is a deficiency of 2-48 of sulphur below what is required to form disulphide of copper, and protosul- phides of iron and the other metals, so that Eammelsberg suggests that part of the iron exists as disulphide (Fe 2 S). 2. From Mansfeld, by Berthier. 8 3, 4, 5. These specimens of thin-regulus were analysed in Rammelsberg's laboratory by De la Trobe, Schliesser, and Boujoukas respectively. Eammelsberg states that this regulus always contains moss-copper, not only in cavities, but also in its compact substance. A specimen of Spurstein in my collection from Mansfeld is inch thick : it contains numerous elongated tube-like cavities, perpen- dicular to the unfractured surfaces ; many of these cavities contain teeth-like projections of metallic copper ; the colour of a fresh fracture is dark grey, like disulphide of copper, with a distinct reddish tinge. A specimen of Dunnstein from Mansfeld is J- inch thick : its surfaces are parallel ; it is full of tube-like cavities, many of which contain projecting teeth of metallic copper ; the colour of a fresh fracture is grey, like disulphide of copper ; there is no efflorescence upon the surface. A specimen from Sangerhausen is pimpled on the surface exactly like pimple-metal ; it is compact, and not porous like the last ; the colour of its fracture is grey, like disulphide of copper ; particles of metallic copper are thinly disseminated through the mass. ANALYSES OF SLAGS ACCOMPANYING CONOENTRATED-REGULUS (SPURSCHLACKE) AND BLACK-COPPER. 1. 2. 3. 4. 5. Silica 33-18 34'11 33-6 38-15 37-90 Alumina 11-22 8-46 5'6 Protoxide of iron 32-03 37-68 51-5 47-22 49-23 Lime 17-14 13'38 5-0 11-56 9-07 Magnesia 2-96 4-57 .. 0-03 1-47 Copper existing ) .. .. Cu 2 O 3-0 2-86 1-^9 partially as disulphide ] 1 90 0-68 Sulphur not deterni. 0-46 98-43 99-34 98-7 99-82 99-26 Ann. d. Mines, 1. s. 9. p. 68. 422 ANALYSES OF ROASTED REGULUS AND BLACK-COPPER. Observations. 1, 2. Slags accompanying the formation of concen- trated-regulus ; the analyses were made in Eammelsberg's laboratory by Wornurn and Hoffmann respectively. The slags are described as more stony than glassy, bluish black, slightly shining or dull. They are decomposable by hydrochloric acid. In each analysis the oxygen of the silica is equal to that of all the bases (17-25 : 18-96 in the first, and 17*72 : 17*90 in the second); these slags are, therefore, tribasic or singulo-silicates. 3. Black-copper slag from Mansfeld, by Berthier, who describes it as compact, heavy, black, and magnetic, resembling certain slags produced in the conversion of pig-iron into malleable iron. 9 4, 5. Black-copper slags. The analyses were made in Itam- melsberg's laboratory by Lade and Gehrenbeck respectively. They are described as stony, black, and decomposable by hydrochloric acid. The oxygen of the silica is equal to that of the bases ; so that, like the slags accompanying concentrated-regulus, they are tribasic or singulo- silicates. 1 ANALYSES OF THE COMPLETELY-ROASTED REGULUS (GAAEROST). 1. 2. Copper 51-97 67'59 Iron 20-39 10'56 Zinc and nlck'J .. 0-67 Oxygen 13'61 8-67 Sulphur 2-11 1-6-1 Matter insoluble in acids . . 11-92 9 49 100-00 98-62 Observations. These analyses were made in Eammelsberg's labora- tory. The insoluble matter contained silica and protoxide of copper. From the preceding results it may be concluded that the roasted regulus consists chiefly of dioxide of copper and magnetic oxide of iron (Fe 3 O). A specimen of six times roasted regulus in my collection (Spur- gaarrost) from Sangerhausen is a heavy, imperfectly melted mass, containing lumps of metallic copper; it is brown-red, and appears to contain a large quantity of dioxide of copper ; when seen under certain conditions of incident light, the characteristic ruby colour of laminae of fiised dioxide of copper may be distinctly perceived on its surface. ANALYSES OF BLACK-COPPER. 1. 2. 3. Copper 95-45 89-13 92-83 Iron 3-50 4-23 1-38 Lead 0-97 2-79 Silver 0-49 not determ. 0-26 Zinc, nickel, and cobalt 3-98 1 05 Sulphur 056 1-07 1-07 100-00 99-38 99-38 Observations. I. By Berthier. No trace either of nickel or cobalt was found. 2 2 and 3. By Hoffmann and Ebbinghaus respectively, in '' Ami. d. Mines, loc. cit. J Rammelsberg, op. cit. 2 Op. cit., p. 68. ANALYSES OF THE "BEAR" (EISENSAU). 423 Rammelsberg's laboratory. The percentage of silver in No. 3 corre- sponds to 84 ozs. 18 dwts. 15 gr. per ton of 2240 Ibs. ANALYSES OF THE " BEAR " (EISENSAU). In the vicinity of Magdeburg in 1831 several lumps of iron-like metal were discovered in the ground, about 4 feet below the surface ; and as no iron-works had existed in the locality, it was suspected that they were of meteoric origin. They were analysed by Stromeyer, who found them to consist chiefly of iron, and to contain a consi- derable quantity of molybdenum, a metal of very unusual occurrence, and not previously known as a constituent either of meteoric iron, or of any furnace product. The metal was hard, and when in small pieces could be broken as easily as white cast-iron, and reduced to coarse powder in a mortar. A freshly-fractured surface was scaly- granular (schuppig-komiges), had a tolerably bright lustre, and a tin-white colour passing into grey. On examining different pieces, two varieties of metal were observed : one was distinctly scaly, more coarsely granular on fracture, and had a greyer colour ; the other was indistinctly scaly and more finely granular on fracture, somewhat lighter in colour, and more brittle. The sp. gr. of the coarse-grained variety was 7-2182, and that of the fine-grained 7*3894. In intimate admixture with the ferruginous mass, especially in the coarse-grained variety, was a large quantity of a metallic sulphide, which in appear- ance and other respects resembled purple copper ore (3Cu 2 S-|-Fe 2 S 3 ). This sulphide occurred chiefly towards the exterior, and in places constituted almost entirely the outer layer of the mass. In the interior of some of the pieces extremely small quantities of moss- copper were found. 3 Heine subsequently analysed " bears" from the Oberhiitte at Eisleben, and ascertained that their composition was identical with that of the metallic lumps discovered at Magdeburg. He published an elaborate paper on the subject in 1836. 4 Both varieties of metal are soluble in hydrochloric and nitric acids ; and at first metallic scales are separated, which dissolve more slowly than the rest of the mass. 1. 2. 3. 4. 5. Iron..., . 76-77 74-60 73-26 57-68 57-91 Molybdenum 9'97 10-19 9-13 27-33 28-49 Copper 3-40 4-32 1-79 2-49 2-45 Cobalt 3-25 3-07 0-77 ( 5<5() / 0-67 Nickel 1-15 1'28 4-63 ( } 3-42 Manganese 0-02 0-01 Arsenic 1'40 2-47 Phosphorus 1-25 2-27 6'04 4-58 3'51 Sulphur 2-06 0-92 0-09 0-46 0-60 Silicon 0-35 0'39 Carbon 0'38 0-48 1'42 1-31 0-87 100-00 100-00 97 13 99'35 97-92 Observations. I, 2. From Magdeburg, by Stromeyer: No. 1 coarse- 3 Ann. d. Phys. u. Chein. Poggendorff, I 4 Jour, fur prakt Chem. Erdmann. 1833, 28. p. 551. | 1836, 9. p. 177. 424 BLACK-V1TEIOL. grained, No. 2 fine-grained variety. 3, 4, 5. From Eisleben, by Heine : No. 3 coarse-grained, sp. gr. 7-578 ; Nos. 2 and 3 fine-grained, sp. gr. 7-578. According to Heine, Augustin was the first to suggest that in this interesting product the function of molybdenum might be regarded as analogous to that of arsenic in speise. The processes employed in all these analyses appear to be defective in certain respects. Eammelsberg has published the following analysis of a remarkable product from the Mansfeld smelting-works : its mode of formation is not known ; it is in imperfectly defined prismatic crystals, which are partially coated on the surface with sulphate of copper; they are magnetic, and their fracture is silver-white passing into grey, gran- ular and shining. 5 Iron 63-23 Copper 12-69 Cobalt with traces of nickel 8-22 Molybdenum 8'40 Sulphur 9-43 101-97 Black-vitriol. The rich regulus was washed with water several times during the process of roasting, and the solution of sulphate of copper was evaporated, in order to obtain the salt in crystals. The mother- liquor was added to the solution obtained in a subsequent washing of regulus, and the mixed liquors were evaporated, and set to crys- tallize. The second mother-liquor was again evaporated in admixture with fresh solution ; and this was again repeated. At last there remained a dark-coloured mother-liquor (Schwarzlauge) , from which crystals of black-vitriol were produced. They have the form of green- copperas or sulphate of protoxide of iron, are bluish-black in colour, and are often of considerable size. The salt, according to Eammels- berg, contains magnesia and protoxides of copper, iron, manganese, cobalt and nickel; 6 and, as its form would indicate, 7 equivalents of water. A specimen in my collection contains so much cobalt, that after the precipitation of the copper by sulphuretted hydrogen, the solution has a tolerably deep red colour. Intermixed with the crystals of black-vitriol, others of a pale bluish- green colour are occasionally found, which, according to Rammelsberg, have the following composition : Oxygen. Sulphuric acid 35-56 21-30 Protoxide of copper 4'47| Iron 0-52 Oxide of zinc, protoxide of nickel, with some) 1 K. O i 7 > 4 -31 cobalt and manganese I 10 '^' I Magnesia 0-63| Potass 18-39 3'12 Water 25-16 22-37 100-00 The composition of the salt may be expressed by the formula 5 Lehrb, p. 220. 6 I!)M . 2 31. CRYSTALS OF ARTIFICIAL FELSPAR. 425 (RO,S0 3 -fKO,S0 3 )4-6HO: and it is similar in form to the analogous double salts of potass and ammonia. 7 It is remarkable that the potass derived from .the ore and the ashes of the fuel should thus become concentrated in this salt. Crystals of artificial felspar. Heine was the first to analyse and describe this interesting product, which was met with for the first time at Sangerhausen in 1834, during the reparation of one of the blast (ore) furnaces. 8 It occurred in well-defined white and pale- violet crystals, which were attached to the back wall in a space included between 1 and 2-^ feet above the twyers, and were disclosed after the removal of an incrustation consisting chiefly of blende, from this part of the interior of the furnace. Similar crystals existed in fissures and holes in the stone-work, and also upon a carbonaceous coating which firmly adhered in thin layers to the sides of the hearth, and was not unlike graphite in appearance, except that it was deeper in colour : the stones used in the construction of the shaft and hearth of the furnace were quartz conglomerate. Besides the crystals, there was found in the furnace incrustation a compact spar-like sub- stance, having a conchoidal fracture, but in colour and other characters resembling the crystals. I have several specimens of these crystals from the Sangerhausen furnace. Some are perfectly colourless, others have a delicate amethystine tint, and others again appear almost black, possibly from the presence of intermixed carbon or dark blende : they vary considerably in size, one of the largest rhombic faces being T ! Vths of an inch in width ; they scratch glass, and, according to Heine, their sp. gr. is 2-56 ; they have generally the form of adularia. Crystals of the same kind of felspar have also been found in furnaces at Leimbach. ANALYSES. 1. 2. 3. 4. Silica 64-53 65-95 65-03 63-96 Alumina 19-20 18-50 16"84\ on , u Sesquioxide of iron 1-20 0'69 0-88/ ZU ' Lime 1-33 4'28 0-34 0-43 Magnesia .. 0-34 0'54 Protoxide of copper 0-27 0-13 030 Sesquioxide of manganese (Mn 2 O 3 )... ) . . . . 0-36 Oxide of zinc f traces traces ,, cobalt I Potass 13-47 10-47 15-26 12-49 Soda .. 0-65 0-65 100-00 100 02 100-00 98-11 Observations. 1 and 2. By Heine : 1, by fusion with carbonate of soda ; 2, by fusion with carbonate of baryta : in both analyses the potass was estimated by loss, and the presence of a little soda was considered probable. 3. From Sangerhausen, by Abich. 9 Sp. gr. ' Lehrb., p. 231. 8 Annal. d. Pliys. u. Chora. Poggen- dorff, 1835, 34. p. 531. Ueber kiinst- liche Feldspathbildung. Von Bergprobirer Heine in Eisleben. 9 Geolog. Beobacht. 1841, p. 10. , Braunschweig, 426 SMELTING OF COPPER-SCHIST IN HESSE. 2-56. The silica contained a trace of titanic acid. The analysis was made by means of hydrofluoric acid. Abich suggests that the amount of lime in No. 3 is erroneous, and that the error was, probably, occa- sioned by the presence of baryta. The crystals selected for analysis by Abich were perfectly pure, and of a pale amethyst colour. 4. By Kammelsberg. These crystals were obtained from smelting-works at Leimbach : they were vitreous in lustre, grey, and had the sp. gr. of 2-665. 1 From the preceding analyses may be deduced the formula KO,Si0 3 +Al 2 3 ,3Si0 3 , which is that of orthoclase. SMELTING OF COPPER-SCHIST AT KIECHELSDORF IN HESSE. Copper-schist occurs here in the same geological position as in the province of Mansfeld, and has long been smelted : it is not sufficiently argentiferous for the profitable extraction of the silver. The products of smelting at the Friedrichshiitte, Eiechelsdorf, have been analysed by Genth, under the direction of Bunsen ; and of all the analyses which have been published in elucidation of the chemistry of copper-smelt- ing, few appear to be so complete, or more worthy of confidence. 2 The Noberge and Unterschiefer (Letten) of the bed of schist and the underlying sand-ore were smelted at these works : the schist generally contains not more than from 2 to 3 per cent, of copper, and the sand-ore from 3 to 4 per cent., rarely from 6 to 7 per cent. Purple copper-ore and iron-pyrites are the chief metallic constituents ; in addition to these and disseminated through the ore occur copper- pyrites, red copper-ore, vitreous-copper-ore, blue and green carbonate of copper, native copper, kupfernickel, smaltine (tin-white cobalt, Speis-kobalt), zinc-blende, galena, molybdenite (sulphide of molybde- num), grey-copper (Fahlerz), and others. The process of smelting consisted of the following operations: 1. Boasting the schist in large open heaps. 2. Fusion of the roasted schist with black-copper slags in a blast-furnace with two receivers in front (Brill enofen). On the bottom of the hearth a ferriferous " bear" (Eisensau) was formed as usual ; on the top of this was regulus, covered with slag, which, according to its different modifications, received the names of raw-slag, Schwiel, and Schwiel-slag. In addi- tion to these ordinary products of ore-smelting, others of much interest in a scientific point of view were occasionally formed. Thus on the front wall of the blast-furnace and in the cooler parts sulphur, realgar, arsenious acid, blende, galena, &c., have been found sublimed. 3. The regulus from the last operation was roasted nine or ten times succes- sively ; the product was mixed with charcoal and ore-furnace slag, and smelted in a low blast-furnace (Krummofen), when black-copper and thin-regulus (Diinnstein) were obtained. This regulus was added to the ore-furnace regulus in the 4th or 5th fire, and roasted. 4. The 1 Lehrb. d. Cliem. Met. 238. 2 Chemische Untersuclmng der beim Kupferschiefer-hiittenprocess fallendeii Producte. Jour. f. prakt. Chem. 1846, 37. p. 193 et seq. Genth was assistant of Bunsen at Marburg. This investigation has also been published in the Berg. u. hiittenm. Zeit. 1846, p. 617 et seq. ANALYSES OF REGULUS. 427 black-copper was refined in the common refining-hearth (Gaarheerd) ; the products were as usual rosette -copper and refinery-slag, or skim- mings (Gaarkratze). The following analyses of the products were all by Genth : ANALYSES OF REGULUS. 1. 2. 3. Ore-furnace Ore-furnace Black-copper regulus. regulus. or thin-regulus. Copper 42-95 43-81 61-26 Silver 0'09 Lead 1-21 0-87 Iron 27-08 24-96 13-70 Nickel 0-57 1-14 traces Cobalt trace trace 4-11 Manganese 2*33 traces Calcium 0-44 0-96 do. Sulphur 28-29 26-57 22'51 100-54 100-73 101-58 Observations. 1. From a disc taken at the upper part of one of the receiving cavities. Crystalline-granular, fracture uneven, somewhat inclined to splintery, opaque, lustre metallic, colour of a fresh fracture pale bronze-yellow (Speisgelb), but after a few seconds acquiring a tint of copper-red and indigo-blue, sp. gr. 5*223 ; soluble in nitric acid ; full of cavities containing fine capillary metallic copper ; the magnet extracts from the finely-pounded regulus a sulphide of iron of the colour of magnetic pyrites. According to Genth the constitution of this regulus may be expressed as follows : Sulphide of lead 1-40 containing of sulphur 0-19 Magnetic pyrites (Fe^S 8 ) .... 44-84 ,, ,, 17-76 Sulphide of nickel (NiS) .... 0-88 ,, ,, 0'31 Bisulphide of copper 49'57 ,, ',, 10-03 Metallic copper 3'41 ,, ,, 100 10 28-29 The composition of magnetic pyrites may be equally well expressed by either of the two formula) 5FeS,Fe 2 S 3 and 6FeS,FeS 2 . 2. The regulus was from the lowest part of one of the receiving cavities (Kupfersteinkonig). In external characters it was precisely similar to the last ; sp. gr. 5-147. According to Genth its constitution may be expressed as follows : Silver 0-09 Sulphide of lead 1-00 containing of sulphur 0-13 Magnetic pyrites 41-33 ,, ,, 16-37 Sulphide of nickel 1-76 ,, ,, 0-62 ,, manganese (MnS) 3-68 ,, ,, 1-35 Bisulphide of copper 40-03 ,, , 8-10 Metallic copper 11-88 99-77 26-57 3. In thin crystalline plates. It is stated that in " physical charac- ters " it was similar to ore-furnace regulus ; but its colour must 428 ANALYSES OF SLAGS. certainly have been different; sp. gr. 5*004; its surface consisted of a thin coating of metallic copper. Its constitution may be expressed by the formula K 2 S,3Cu 2 S, in which E 2 = Fef + Cof. Calculated from this formula, its composition is as follows : Iron 14-40 Cobalt 3-80 Copper 61-13 Sulphur 20-67 100-00 The sulphur found is 1'84 in excess of that deduced by calculation ; and this Genth attributes to his neglect in the analysis to treat the sulphate of baryta after ignition with hydrochloric acid. ANALYSES OF SLAGS. 1. Ore-fur- nace slag. 2. Red- brown ore-fur- nace slag. 3. Ore-fur- nace slag, by the ad- dition of sand- ore. 4. Schwiel. 5. Black- copper slag. 6. Refinery Slag. 7. - Slag from refined copper. Silica 48-23 44-47 51 44 45-41 31-72 7-88 32-23 Alumina 6*51 12-96 19-32 18-11 2-83 0-81 5-60 Protoxide of copper 0-58 1-23* 0-30 1-07 1-26 4-79 , , molybdenum) (MO) / trace 0-38 0-25 0-23 2-36 0-87 , , iron 14-13 7-85 5-88 6-31 47-80 82-49 20-72 , , cobalt , , nickel , , manganese Magnesia trace do. 0-65 3*35 trace do. 0-30 7-00 0-89 1-40 0-84 7-15 0-25 trace do. 3-86 trace 3 59 trace 34-16 Lime 23-06 21-20 17-80 18-49 8-06 1-70 do Potass . 3-75 2-90 1-78 3-09 3-68 0-31 1-23 Soda 0-88 0*87 0-65 0-70 1-26 0-25 0-43 Protosulpliide of copper) (CuS) 0-67 Tersulphide of molybde- 1 num (MS 3 ) ... f 0-20 '.. Protosulpliide of iron 1 (FeS) . .) .. 1-40 .. .. .. .. 101-14 99-16 101-43 100-65 100-76 100-65 100-03 * Cu 2 0. Observations on the characters of the preceding slags. 1. Amorphous, fracture conchoidal and splintery, lustre vitreous or waxy, colour between pitch- and velvet-black, streak greyish-white, opaque, in thin splinters translucent, and greenish-grey by transmitted light, sp. gr. 2-834 ; readily soluble in hydrochloric acid with the separation of gelati- nous silica. The composition of this slag may be expressed by the for- mula 24 (3RO,2SiOO+2Al 2 3 ,Si0 3 . 2. Streaked with red-brown veins, sp. gr. 2-683, in other respects resembling the last. The composition of this slag may be expressed by the formula 3(2RO,Si0 3 )-r-Al 2 3 ,Si0 3 . 3. Similar to the ordinary ore-furnace slag, sp. gr. 2-731, more or less vesicular. The composition of this slag may be expressed by the ANALYSES OF COPPER. 429 (very improbable ?) formula 10(3RO,2Si0 3 )+3(4Al 2 3 ,5Si0 3 ). 4. Crys- talline-granular, fracture uneven and splintery, opaque, translucent at the edges, pearly vitreous lustre, colour between ash-grey and greenish-^rey, streak greyish white, sp. gr. 3-023; not completely decomposable by hydrochloric acid. Schwiel is an impure slaggy mass, which collects x round the edges of the receiving vessels in front of the furnace ; it contains much intermixed regulus, and is melted over again. Its composition may be expressed by the formula 9(2RO,Si0 3 ) + 5Al 2 3 ,Si0 3 , or by the (not less improbable?) fornmla 6(3RO,2Si0 3 ) + 5AP0 5 ,2Si0 3 . Partial analyses of three varieties of Schwiel slag were made, but the composition was found to be so variable that it was not considered worth while to complete them. The proportion of silica varied from 48-44 to 58-28 per cent., and that of the lime from 18*50 to 22-27 per cent. 5. Crystalline, in the form of sheets, with a wrinkled surface, structure granular and in part finely radiating, opaque, lustre vitreous inclining to metallic, colour between bluish- and velvet-black, streak grey, sp. gr. 3'512, magnetic; decom- posable by hydrochloric acid. The composition of this slag may be expressed by the formula 36(3RO,Si0 3 )+3AP0 3 ,2Si0 3 . 6. Knobby, vesicular, fracture uneven and crystalline, opaque, lustre between vitreous and metallic, colour iron-black, streak grey, sp. gr. 4*609, strongly magnetic ; not soluble in hydrochloric acid ; the cavities are in part filled with shots of metallic copper and an almost silver-white alloy of copper and nickel. 7. Small black slaggy particles, which in appearance exactly resemble the Lapilli from Vesuvius, sp. gr. 4-135, magnetic. It is not stated how this slag was produced : was it obtained in the remelting of refined copper in the process of tough- ening (Hammergaarmachen) ? Its composition may be expressed by the formula 5(3EO,Si0 3 )+Al 2 3 ,Si0 3 . ANALYSES OF COPPEK. 1. Black copper. 2. Black copper. 3. Refined copper. 4. Tough copper. 5. Refined copper residue. Oopper 83-29 92-24 83-90 99-31 98-97 Silver 0-051} 0-10 do. 0-10 0-13 Lead. . 0-31 0-89 0-60 0-21 0-07 Jron . 1-66 1-41 0-02 0-23 Cobalt traces traces Nickel 3-28 4-15 i-io 0-28 0-27 Protoxide of nickel 13-86 traces Calcium 0-05 0-13 0-10 0-03 0-04 Magnesium . ... o-oi traces 0-12 0-01 traces Potassium 0-03 0-10 32 0-04 0-07 Aluminium traces Sulphur 11-31 0-98 traces traces traces Slae- do. do. 0-22 100-00 100-00 100-00 100-00 100-00 Observations. In all these analyses, not less than from 15 to 20 430 SMELTING 'OF COPPER-SCHIST IN HESSE. * grammes were operated on in each analysis : the weight of the copper was deduced indirectly, by subtracting the sum of the weights of all the other ingredients from the weight of metal taken for analysis. 1. Upper disc. Crystalline with indented surface (zahniger Ober- flache), very finely granular, fracture hackly, opaque, lustre perfectly metallic, copper-red and silver-white, tarnished black; sp. gr. of a disc from the middle, 7 '305. 2. From the lump of copper remaining after the discs have been all taken off (Kupferkonig). Knobby, vesi- cular, in other respects similar in external characters to the last. 3. Uppermost disc. A plate- like mass, consisting of numerous thin layers of metallic copper cohering together, over which are dissemi- nated small black octahedra of protoxide of nickel ; crystalline, pulverisable, fracture hackly, colour between black-grey and copper- red. By the action of nitric acid is obtained an insoluble residue, which consists of minute crystals of pure protoxide of nickel mixed with some slag. These crystals are octahedra, belonging to the cubical system ; they appear black by reflected, and ruby-red by transmitted, light ; they are completely insoluble in concentrated sulphuric, nitric, hydrochloric, and nitro-hydrochloric acids. They undergo no change whatever in melted carbonate of soda ; but when heated to redness with bisulphate of potash, a double sulphate of potash and protoxide of nickel is formed. When heated in hydrogen they are reduced to the metallic state. 3 The production of these crystals in the refining process is a very interesting fact. I have had an opportunity of examining a specimen of them, with which I was favoured by Genth many years ago. It seems rather remarkable that potassium, calcium, and magnesium should exist in melted copper, which must certainly have contained dioxide of copper, and through which oxide of nickel is copiously disseminated. Moreover, the proportion of these metals in refined copper is much greater than in black-copper, notwithstanding the powerfully oxidizing influence to which the latter is subjected during the process of refining. It would even appear from the analyses that the whole of the iron and the greater part of the nickel may be removed from black-copper by oxidation, without any sensible reduction in the proportion of such extremely oxidizable metals as those of the potassium and calcium groups. 4. Thick disc of tough copper (Hammerkupfer), tough, but not " Hammergaarkupfer." Colour pure copper-red ; innumerable small channels, through which gas had escaped, traversed the metal from the lower to the upper surface, and communicated especially to this disc " a decidedly radiated, crystalline-granular fracture." 5. From the lump of copper (Gaarkupferkonig) after the discs of copper had been all taken off. Vesicular, crystalline-granular, radiated fracture. It was ascertained by Genth that in the melting and refining of black-copper the silver did not become concentrated in any particular part of the copper, so that it could be profitably extracted. In the 3 Jahres-Bericht. Berzelius, 1846, v. 25, p. 170. ANALYSES OF THE "BEAR" (EISENSAU). 431 black and refined copper of the Friedrichshiitte the proportion of silver in the centner varies from 2 to 3 loths, that is, from about 20 to 30 ozs. per ton : discs were successively taken off during the course of one operation of refining, and were found to contain the following proportions of silver : Upper discs of black-copper Middle do. do. Konig do. do. Uppermost disc of refined copper Second do. do. Contained of silver per cent. . 0-059 ,. 0-049 . 0-104 traces 0-058 When tough (Hamuierkupfer) do. 0-101 Gaarkupferkoni First thin disc of refined copper Second do. do. Third do. do. Fourth do. do. Fifth do. do. do. Contained of silver per cent. 0-059 0-066 0-105 0-067 0-066 0-132 Genth made the three following analyses of " Hammergaar" copper from Dillenburg, in Nassau : 1. 2. 3. Uppermost disc. Silver 0-056 Lead Iron traces Copper 99-944 Eesiduc.... 100-000 Middle disc. 0-056 0-038 0-107 99-799 100-000 Residue, or Konig. trace 0-069 0-015 99-916 traces 100-000 ANALYSES OF THE BEAU " (EiSENSAti). 1. 1-04 Phosphorus Arsenic Carbon .-.. 0'73 Sulphur 0-59 Silicon 2-98 Aluminium traces Iron 86-64 Manganese Nickel traces Cobalt 3-61 Molybdenum Copper 5'19 100-78 2. 0-04 traces 1-12 0-31 1 28 traces 84 24 traces 2-85 6-98 4-52 101-34 Observations. 1. From the hearth-bottom of the furnace. Crystal- line, coarsely-granular, fracture uneven, opaque, lustre metallic, colour steel-grey, brittle, sp. gr. 7-466, hardness = 5-5 (between apatite and felspar) ; it contains a few cavities, in which are small octahedra of metallic iron "united in toothed knitted masses," but very little moss- copper; here and there were small round grains of a crystallized black-blue compound, of which the quantity was too small for inves- tigation. 2. This " bear " was from one of the receiving cavities (Brillheerd). Crystalline, finely granular, fracture uneven, opaque, lustre metallic, colour between steel-grey and tin-white, brittle, sp. gr. 7-549, hardness the same as that of the last; soluble in hydro- chloric acid. 432 TOPPER-SMELTING. OTHER ACCESSORY PRODUCTS. Zinc-blende. It occurs in crystalline masses, foliated, occasionally composed of radiating fibres ; opaque, fracture uneven or jsplintery, lustre adamantine and metallic, colour between smoke-grey and greyish-black, streak greenish-grey, sp.gr. 3'784, hardness = 3 (calc- spar) ; soluble in nitric acid ; it is occasionally tarnished of a golden- yellow colour; it is not abundant. Genth has given the following analysis of a specimen : Analysis. ZiiM* 57-51 Supposed constitution. Sulphide of zinc 85-91 Containing of sulphur. 28-40 0-55 manganese 0-86 0-31 4. 08 iron (FeS) 6-42 2-34 Coppsr 1-06 Disulphide of copper 1-33 0-27 2-79 Sulpliide of lead .... 3-22 0-43 Molybdenum . Calcium .. .. 0-15 1-06 , , molybdenum ( M S :i ) . . , , 0-30 1-06 0-15 Sulphur 31 -J 99-09 31:90 The state of combination of the calcium is regarded as doubtful ; but there is no difficulty in understanding how it should be present as sulphide, and how this sulphide should be combined with other sulphides. Sulphur. It has been found in crystals, which in colour and lustre were precisely similar to those of native sulphur. Arsenious acid. It occurs in octahedrons and tetrahedrons up to Qmmg ( near ly j. i n .) i n diameter, generally with step-like depressions, opaque, and having an adamantine lustre. Realgar.- It had only once been met with at these smelting-works (1846), in pieces of a magnificent aurora red colour, and having a crystalline, foliated structure. I have a beautiful specimen of this substance from a roast-heap at Freiberg. Galena. It exactly resembles native galena, and presents distinct laminations parallel to the faces of the cube. Fume. It was in the state of dry powder, of a yellowish- white colour inclining to grey. It was partly soluble in water : the aqueous solution contained sulphates of protoxide of copper, sesquioxide of iron, protoxide of manganese, oxide of zinc, alumina, lime, magnesia, and potash, chloride of sodium and arsenious acid. The insoluble residue was boiled with hydrochloric acid : the solution contained the following substances sulphuric acid, oxide of lead, oxide of zinc, sesquioxide of iron, arsenious acid, oxide of antimony, lime, protoxide of nickel, magnesia, potass, soda, molybdic acid, oxide of bismuth, and organic matter. The residue insoluble in hydrochloric acid consisted of sulphate of lead and combinations of silica : the following bases were also present, sesquioxide of iron, alumina, lime, magnesia, potass, soda, oxide of zinc, protoxide of copper, and traces of oxide of bismuth and protoxides of manganese and nickel. ANALYSES OF CUPRIFEROUS SANDSTONE. 433 COPPER-SMELTING IN PERM, IN EussiA. 4 According to Gustavo Eose the ores of copper occur in the " Weiss- liegende " of German geologists. The sandstone of the country, which consists chiefly of grains of quartz and fragments of other hard rocks cemented together by a marly substance, is more or less impregnated with chrysocolla and blue carbonate of copper. In the lines of strati- fication are found malachite, small crystals of blue carbonate of copper, vanadiate of copper, and, rarely, vitreous-copper, copper-pyrites and iron-pyrites. Native copper is also present. Upon the sandstone is a dark-coloured bituminous massive or schistose clay, which contains dis- seminated globules of vitreous-copper and subordinate beds of copper- pyrites (strates subordonnees de pyrite cuivreuse). Blue and green carbonates of copper are generally found in the lines of stratification. The ores were smelted with the addition of 30 per cent, of dolomite. The average produce of copper might be estimated at 2-J per cent. The fuel was charcoal. The smelting was effected in blast-furnaces having one twyer each and a fore-hearth like the Atvidaberg furnaces. The chief dimensions were as follow: height of the shaft 4 m 44 (14 ft. 7 in.) ; diameter of the mouth O m 62 (2 ft.) ; the belly was l m 07 (3 ft. 6 in.) wide by l m 24 (3 ft 13 in.) ; the hearth l m 07 (3 ft. 6 in.) broad by O m 62 (2 ft.), and 71 (2 ft. 4 in..) deep; the fore-hearth was serai-elliptical, the long axia being O m 31 (1 ft.), fhe short axis O m 20 (10 in.), and the depth O m 33 (13 in.). The quantity of ore smelted daily varied from 3684 to 4502 kilogrammes ('. e. in round numbers from 3 to 4 tons). The products were slag, cupriferous piy-iron, and black-copper. The metallic products were tapped out once in 24 hours. The pig-iron formed the upper stratum, and was first removed in discs or round cakes, after which the black-copper was removed in a similar way. Choubine has given the following analyses of two varieties of cupri- ferous sandstone. 1. 2. Protoxide of copper 23*08 2*50 Vanaclic acid (VO 3 ) 0'53 Silica 34-91 53-15 Alumina 3-46 5*24 Sesquioxicle of iron % 1-23 4'19 Protoxide of iron 0-23 0-39 Oxide of manganese traces traces Lime 5-87 8'94 Magnesia 4-09 5'77 Potass 0-38 0-84 Sulphur 0-13 0-59 Bituminous matter 4-47 3-09 Carbonic acid 12-66 9-33 Water 7-08 2-90 97-59 97-46 4 Coup d'ceil sur le travail de Cuivre aux Usines de Perm (called Jougovsk and Mines de Russie. Annee 1842. S. Pe- tersbourg, 1845, p. 184 et seq. See also Motovilikhinsk) ; par M. le Lieutenant | the same work, Annee 1840, p. 250. Choubine. Annuaire du Journal des j Notice sur les produits de la fonte des 2 F 434 COPPER-SMELTING IN RUSSIA. It is stated that, in general, the slags were very good, and that the oxygen of the silica was double that of the bases. They varied in sp. gr. from 2*31 to 2-78. The black copper contained on an average 90 per cent, of copper; its sp. gr. varied, according to the propor- tion of iron present, from 7 '8 11 to 8-097. One specimen analysed by Choubine is reported to have had the following composition : Copper 90-52 Iron 6-17 Vanadium 1-21 Carbon 0-94 98-84 There was, probably, a considerable error in the determination of the carbon in this analysis, for, admitting the iron to have been com- bined with the maximum of carbon (about 5 per cent.), it would follow that not less than 0-615 of carbon must have been combined with the copper and vanadium, a conclusion which is entirely opposed to the results of other observers. The cupriferous pig-iron is a remarkable product. A specimen, of which the analysis by Choubine is inserted below, was white, graiiular in struc- ture, and even in fracture ; it was extremely hard, and easily scratched glass ; its sp. gr. was 7*432 ; particles, or small grains, of metallic copper were seen disseminated on its fractured surface. It was conjectured that its great hardness might be due to the presence of vanadium. ClTPKIFEBOrS PlG-lRON. Iron 75-97 Copper 12-64 Vanadium 1'99 Aluminium 0'89 Calcium 0-95 Magnesium 0*78 Silicon 2-51 Carbon 3 03 98-76 The discs of cupriferous pig-iron were inter stratified with charcoal, and melted in a refining-hearth (Gaarheerd), lined with a mixture of clay and sand. The products were black-copper and a slag of silicate of protoxide of iron, which, being very fusible, ran off; the silica was sup- plied by the coating of the hearth and the silicon contained in the metal. The sp. gr. of this slag was 4-070, and its composition was as follows :.-. Silica 18-15 Vanadic acid 1'57 Alumina 0'36 Protoxide of iron 75 '50 ,, copper ( Cu 2 O ?) 0-40 Lime 1 97 Magnesia 1'03 98-98 Minerals de Cuivre aux Usines de Perm, ! munication above referred to. There is by the same author. Some of the ana- also another paper by the same author, lyses diifer considerably from tliose of Aimee 1841, p. 319, containing the details similar products in the subsequent com- < of the analyses which I have inserted. ANALYSES OF CUPRIFEROUS PIG-IRON. 435 The formula approximates to 6FeO,SiO 3 . The copper present in this slag was in the form of metallic shots. In one refin ing-hearth 491 kil. (1082-6 Ibs.) of cupriferous pig-iron were worked off in the course of twelve hours, by which 73 kil - 5 (162 Ibs.) of black-copper and 564 kll> 9 (1245*6 Ibs.) of slag were produced. If this slag contained -| per cent, of copper, it was melted in the ore-furnace in conjunction with unclean slags, ore-furnace slag, and refinery-slag, when two products were obtained slag, and a cupri- ferous "bear " (loupe cuivreuse). The slag had the sp. gr.. 3*271, and was composed as follows : Silica 31-61 Vanadic acid 1-30 Alumina 1*48 Protoxide of iron 57"00 ,, copper (Cu 2 ?) 0-91 Lime 4*24 Magnesia 1-58 98-12 The formula approximates closely to 3RO,Si0 3 . It will be observed that this slag contains more than \ per cent, of copper ; it could not be considered as dean, for we have seen that the slag produced in melting the cupriferous pig-iron was remelted when it contained only -J per cent, of copper. The copper was mechanically mixed in the "loupes cuivreuses,', and on an average did not exceed 30 per cent. The sp. gr. of these " loupes " varied from 6'438 to 6'672. One of them was found to have the following composition : Iron 76'30 Copper 19'90 Vanadium 0'12 Aluminium 0-43 Calcium ) Magnesium j ' Silicon 0-83 Carbon 0*73 Slag imbedded 3-33 101-64 This method of treating the cupriferous pig-iron was necessitated by the impossibility of obtaining sufficient copper- or iron-pyrites. The copper is extracted from the " loupes cuivreuses " in the same manner as from the cupriferous pig-iron. According to Choubine, the cupriferous iron-slag resulting from the treatment of the cupriferous pig-iron in the refining hearth could not, for two reasons, be added to the charge of the ore-furnace : the first is, that on fusion half of the iron is disengaged from a silicate of protoxide of iron of the formula of this slag ; and the second is, that the lime and magnesia existing in the copper-ores treated would tend to displace the oxide of iron from the slag, and so favour the formation of ferruginous "bears." In regard to the first reason, I have no knowledge of any expe- rimental evidence in support of Choubine's statement concerning the separation of iron by the simple fusion of a highly basic silicate of prot- 2 F 2 436 COPPER-SMELTING IN RUSSIA. oxide of iron ; and the second reason cannot be reconciled with the pre- vious statements that the ore is essentially a cupriferous sandstone, and that not less than 30 per cent, of dolomite was added in order to flux the silica. A portion of this dolomite might surely have been advan- tageously replaced by slag rich in protoxide of iron. The black- copper is refined in the Spleissofen of the Germans, which will hereafter be described. It is a kind of reverberatory furnace, having a concave bed of brasque to receive the metal. The surface of the melted copper is exposed to the action of a blast from a twyer, which passes through one side of the furnace. The charge of black- copper was 1965 kil. (about 2 tons). Silicious sand is thrown upon the surface of the melted metal during the whole course- of the refining process. The diameter of the nozzle of the blast-pipe was O m 04 (1-58 in.), and the pressure of the blast was equal to a column of mercury of O ra 03 (1-18 in.) : the average quantity of air injected per minute was 6'3436 cubic metres (224 cub. ft.). The slag flows out of the furnace through a channel in one side, like litharge in the cupellation of lead. The refined copper is tapped out into circular cavities, from which it may be taken off in discs. These discs are melted in a small refining hearth (Gaarheerd), and when the copper has acquired the proper pitch, it is laded into ingot moulds. Choubine gives the following analysis of the refined copper : Copper 96-54 Dioxide of copper 1*41 Vanadium 0-21 Iron .. . 0-78 98-94 The proportion of. iron is considerable, and must surely be excep- tional. Tin, antimony, and arsenic are stated to be entirely absent from this copper. Rivot has recently published a description of the smelting of cupri- ferous sandstone in Perm, from information derived from Le Play ; 5 and in the following respects this description differs from that of Choubine. The cupriferous pig-iron is stated to be remelted in a cupola, such as is commonly employed in iron-foundries : the metal is allowed to remain in tranquil fusion during an hour, when it separates into two strata, one of highly ferriferous black-copper, and the other of pig-iron which retains a little copper. The copper is first tapped out, and afterwards the pig-iron, which may be applied to castings where strength is not required. Besides dolomite, from 25 to 30 per cent, of ore-furnace slag is mixed with the ore in the first fusion. The ore-furnace slag which flows from the fore-hearth does not contain more than 0-003 per cent, of copper ; and it is stated that this is the cleanest slag of all known copper-works. The slag which is tapped out along with the metal contains copper in shots, which is not lost, as the b Principes ge'neraux du traitement des Minerals Metalliques. Paris, 1859, 1. p. 86. THEORY OF THE PROCESS. 437 slag is remelted. \Yhen the furnace is in good order, the cupriferous pig-iron does not contain more than 3 per cent, of copper. The theory of the process of copper- smelting in Perm is very simple. The copper exists in the ore in the state of carbonate and silicate, salts in which protoxide of copper is the base. \Vhen car- bonate of copper is heated to low redness in an atmosphere containing carbonic oxide or in contact with carbon, the copper is easily and completely reduced to the metallic state ; and when silicate of copper is exposed to the action of the same reducing agents and lime at a strong red heat, the whole of the copper is also reduced to the metallic state. Now the ore in its descent through the blast furnace is exposed to these conditions of reduction ; and not only is the whole of the copper reduced, but a considerable quantity of iron is also reduced and converted into pig-iron. And from the formation of pig-iron it may be inferred, that the iron after reduction must have been during some time in contact with carbon at a much higher temperature than would suffice to effect the reduction of the whole of the copper. The presence of metallic iron would tend to ensure the separation of copper from any silicate of copper which might otherwise escape reduction. In Rivot's description of the refining of the black- copper, it is stated that "when the fusion is nearly complete, the oxidizing action is diminished, as far as the nature of the fuel (wood) will permit, by ceasing to inject air into the furnace ; and a charge of pyritic ores i. e., sulphides of iron and copper, and silicious fluxes is then intro- duced. The matrix of the ore serves to scorify the oxides of iron and copper produced during the fusion ; and the excess of sulphur of the pyrites forms with the metals protosulphide of iron and disulphide (protosulfure) of copper." 6 According to Choubine the regulus which is produced in the smelting of certain ores, and which contains 50 per cent, of copper, is not roasted, on account of its small quantity, but is treated in the refining furnace. By the action of sulphide of iron on any silicate of copper in the slag, silicate of protoxide of iron and disulphide of copper would be formed. In about three hours after the addition of the charge of pyritic ore the slag is removed, the blast again let on, and the process concluded in the usual manner. Rivot informs us that " the black-copper subjected to refining contains from 12 to 15 per cent, of carbon and iron." 7 It might be inferred from this language that the black- copper contained a considerable quantity of carbon ; and if Rivot intended to convey this impression, he is probably in possession of experimental evidence concerning the com- bination of carbon and copper, which he has not yet disclosed. As the copper-ores of Perm are extremely poor, and yet can be smelted with advantage, it will be instructive to inquire concerning the precise conditions under which the smelting is conducted. Accor- dingly, the following details on this subject from Choubine's description are presented : fi Op. cit., p. 106. " ibid. Pouds (1 Poud = 16 k. 372 = 36-1 IDs.). Percent- age of copper. Observations. Smelted in six blast-furnaces in the ) course of 40 days 5 Ores 60000 18000 2-5 90-0 15-0 50-0 0-3125 4-556 The slags are very liquid, and by dry assay yield only inappreciable traces of copper. The method of testing copper-slags by dry assny is quite unsatisfactory, and may lead to very erroneous conclusions. A slag may contain a considerable quantity of copper, and yet not yield an appreci- able trace of copper by this method. The quantity of air injected into a furnace was 9 '61 cub. metres per minute. Dolomite Products obtained < Black-copper . . 1470 615 30 450 Cupriferous pig-iron Loss of copper on the total copper in the ) ore and cupriferous products 5 Ore smelted daily in each furnace Charcoal consumed for all the ore smelted, 2350 korobs, or 3-9165 for 100 pouds. 1 korob (korb) of charcoal contains 20 pouds, or 800 pounds Russian. Unclean slags Total 2565 .... 250 Smelting of slag from cupriferous pig- 1 iron (laitier de fer, iron slag) in six \ blast-furnaces during 36 hours before ( they were blown-out for reparation ... J Products obtained Iron slag 1800 1800 1-5 0-3 27-0 0-689 The cost of smelting 100 pouds of ore amounts to about 3 r. 34| cop. arg. (1 rouble = 4 francs = 3s. 2tf. accord- ing to average rate of exchange), and 1 cop. arg. (silver copeck = 0,04 franc). Cost of winning 100 pouds of ore and carriage to the smeltiiig-works was 7 r. 22f cop. arg. Unclean ore-fur- ~) nace slags 3 Total 3600 Loupes cuivreuses... 120 Charcoal consumed, 81 korobs, or 2 for 100 pouds. Smelting of refinery-slag in a blast-fur- I nave during 42 days | Products obtained -J Refinery- slag Unclean ore-fur- 7 nace slags 5 Total Black-copper Cupriferous pig-iron Total 9000 9000 5-0 0-3125 90-0 15-0 2-277 18000 450 250 Loss of copper 700 Charcoal consumed, 385 korobs, or 2-139 for 100 pouds. Smelting of iron-slag in one blast-fur- I nace during 42 days | Iron-slag 7500 7500 1-5 0-3125 30-00 0-708 Unclean ore-fur- ) nace slag j Total Loupes cuivreuses... Product obtained 15000 4500 Loss of copper Charcoal consumed, 300 korobs, or 2 for 100 pouds of mixture. Cupriferous pig-iron treated in a re- \ fining hearth during 10 days 3 Products obtained ... . < Black-copper < Iron-slag..... 300 43 n> 30 (lib. = Wl 4093) 345 15-00 90-00 1'50 1-00 The cost of treating 100 pouds of cupri- ferous pig-iron was 2 r. 95 cop. arg. 1 Loss of copper Total 388-30 Charcoal consumed, 6| korobs, or 2| for 100 pouds. Loupes cuivreuses melted in a refining- \ hearth during 10 days . ' J Black-copper 300 ' '95 260 30-00 90-0 1-5 0-666 Products obtained J Loss of copper Total 355 . . . fiefining black-copper in one furnace 1 during 13 days Refined copper Refinery -slag Residua Brasque impreg- ) nated with copper 5 Total 2625 2360 550 2 -20 IDS. 13 90-00 97-5 5-0 95-0 3-8 1-30 According to Choubine, it is more eco- nomical to refine the black-copper in the spleissofen with wood as the fuel than in the Gaarheerd with charcoal ; in proof of which he adduces the fol- lowing facts. In refining 100 pouds of black-copper in the spleissofen % sag. cub. of wood is required ; and in re- melting 100 pouds of refined copper, which is much more fusible than black- copper, 3J korobs of charcoal are con- sumed. But J sag. of wood yields on an average 1| korob of charcoal. The charcoal is obtained from the sapin, pin, and lime-tree. Products obtained J Loss of copper 2925-20 Wood consumed, 9 cubic sagenes (1 cub. sag. = 2-1336 cub. metres), or for ] 00 pouds. He-melting and casting refined copper ) in a Gaarheerd during 10 days 5 Products obtained ! Copper in ingots ... Scrap copper 1250 1216'lOlbs. 0-5 41-10 97-5 99-5 99-5 5-0 0-126 Cost of remelting 100 pouds of refined copper was 4 r. 9 cop. arg. Slag Total 1257-25 Charcoal consumed, 43f korobs, or 3i for 100 pouds. KERNEL-ROASTING AT AGOKDO. 439 KERNEL-ROASTING, OR KERNROSTEN (German). This curious and singularly interesting process has already been alluded to (p. 405). When cupriferous iron-pyrites, containing, say from 1 to 2 per cent, of copper, and in lumps about as large as the fist, is subjected to a very gradual roasting with access of air, it is found that a large portion of the copper becomes concentrated in the centre of each lump, forming a nucleus composed essentially of copper, iron, and sulphur. This nucleus is termed the " kernel," whence the appropriate name of " Kernel-roasting." It is enclosed in a more or less porous " shell," or rind, consisting chiefly of sesquioxide of iron, from which it may readily be detached by a few gentle blows with a hammer. The kernels are separated by hand in this manner, and smelted for copper ; and the shells are lixiviated with water in order to dissolve out a small quantity of sulphate of copper which they contain. The washed residues, however, still retain copper in the state of oxide, and, probably, also of basic sulphate ; they are used to cover the ore during the roasting, and, when the process is conducted in open heaps, to form the bed on which the ore is piled. After having served these purposes they are again lixiviated, in order to extract from them any copper which may have been rendered soluble by the action of sulphuric acid, either directly produced by the decomposition of the sulphate of iron formed in roasting, or generated by the conjoint action of stilphurous acid and atmospheric air (see p. 248, ante). The process appears to be of comparatively ancient date ; but I have not been able to trace its history with certainty. At the present time it appears to be carried on with great skill at Agordo, in the Venetian Alps, where it is stated to have been first introduced in 1692 by a Prussian of the name of Weyberg : 8 it is also practised at Miilbach, in the Tyrol, but not on so great a scale as at Agordo. Swedenborg, in 1 734, published an account of the process as then conducted at Agordo, which appears to be precisely similar in principle and results to that now in operation. 9 I am informed by Mr. David Forbes, who was for some years engaged at the Espedal Nickel Works in Norway (now defunct), that kernel-roasting has long been, and still is, practised in Norway. I shall describe with some detail this process as it is now carried on at Agordo under the direction of Liirzer, to whom I have much pleasure in acknowledging my obligation for a complete and charac- teristic series of coloured drawings illustrating the successive stages 8 Meinoire sur les Etablissements d' Agordo (Haute-Ve'netie). ParN.Haton, ingenieur des mines. Ann. des Mines, 5. s. 1855. 8. p. 416. none are, in my judgment, more worthy the attention of those who are interested in the history of metallurgy. They form two tolerably thick folio volumes, copi- 9 Regnum Subterraneum, sive Minerale ously illustrated with copperplate engrav- de Cupro, 1734:, p. 144. The metallur- ings, and magnificently printed. I shall gical works of this remarkable man seem j have occasion to refer to them especially to be very imperfectly known at least in connexion with the subject of iron, they are rarely if ever quoted ; and yrt 440 KERNEL-ROASTING AT AGORDO. through which the ore passes from the commencement of the roast- ing to the development of the perfect nucleus, and also for drawings of the Styrian kiln, which -he has introduced, it is affirmed, with much success. 1 Liirzer promised to furnish me with a detailed description of the process ; but, unfortunately, I have not yet received it, and am, therefore, compelled to have recourse to other sources of information. This I much regret, as there are many important points of detail concerning which Liirzer would, doubtless, have given instruction, but which I cannot find in any of the published descriptions of the process. An account of the mode of occurrence, method of working, and metallurgical treatment of the cupriferous ores at Agordo, was published by Baton in 1855; 2 and of this I shall freely avail myself, as I shall also of the short and incomplete notices of the process previously published by Liirzer himself. 3 Composition of the ore. It occurs in enormous masses or pockets in a schist of ancient geological date. It is compact, has a brass-yellow colour, and a granular, steel-like fracture ; and to the eye it appears free from gangne, though, in reality, it contains quartz finely dis- seminated through the mass. Its mean composition is stated to be as follows : Copper 1-60 Iron 43-15 Sulphur 50-25 Quartz 5-00 100-00 Liirzer gives the average percentage of copper as 2. The ore, however, varies considerably in respect to the proportion of copper, and may either be entirely free from this metal, or, in exceptional cases, may contain as much as 25 per cent, of it. Liirzer is of opinion that the ore probably consists of an intimate mixture of iron- pyrites (FeS 2 ), magnetic -pyrites (Fe 2 S 3 -}-5Fe 2 S 3 ), and copper-pyrites (Cu 3 S-|-F 2 S 3 ). Blende occurs finely disseminated in the ore, and occasionally also galena ; cobalt is found in the solution of metallic sulphates extracted by washing the roasted shells ; tin is present in the copper produced at Agordo, and the ore contains arsenic and antimony, of which the amount may be not less than 2-5 per cent. The ore is broken in pieces about as large as the fist, and assorted by hand into the five following classes : 1 , non-cupriferous pyrites, which is thrown away ; 2, poor, containing less than 1 '5 per cent, of copper; 3, good, containing from 1-5 to 4 per cent. ; 4, best, or rich, 1 I must not omit also to acknowledge my obligation to Professor Miller, of Cam- bridge, through whose intervention I re- ceived these drawings. 2 Ann. des Mines, ante cit. 3 Berg. u. hiittenmannisches Jahrbuch. Tunner, 1853, 3. p. 339; and 1854, 4. p. 242. Rivot has a long article on the subject in his Treatise on the Metallurgy of Copper (1859) ; and though he seems chiefly to have derived his information from Haton's Memoir, yet on many points he differs from this author. His engrav- ings of the Styrian kiln seem to be copied from those of Haton, but with certain alterations ; and if the latter be correct, as they would appear to be from the fact of their agreement with the drawings which I have received from Liirzer, Rivot's engravings are not quite accurate. STYRIAN KILNS. 441 generally containing from 4 to 8 per cent, of copper and upwards ; 5, con- taining galena. The proportions raised of poor, good, and rich ores are respectively as the numbers 51-63, 46-16, and 2-22. The rich ores are smelted together with the kernels, &c., in the blast-furnace in the usual way, and are not subjected to the process of kernel-roasting. Two methods of roasting are practised at Agordo the old method, in heaps or piles, and the new method, in kilns. la piles. They are in the shape of long, truncated, pyramidal heaps, of which the base is rectangular and the transverse section trape- zoidal. Their length is variable; they are 6 m (about 19 ft. 8 in.) wide at the base, and 2 m 50 (about 8 ft.) at the top. On an average they contain 209 tonnes (1 tonne = 1000 kil., and is therefore nearly the same as the English ton) of fresh ore. The ground on which the pile rests is excavated to the depth of about l m 30 (about 4 ft. 4 in.), and the pit so formed is filled up with the washed ferruginous residues of a previous roasting. Upon the bed thus prepared the pyramidal heap of ore in lumps is raised, care being taken to divide it into sections, in the direction of its length, by four or five inclined beds of ore-dust (schlich), with the view of preventing too free a communi- cation between the diiferent parts of the pile and of intercepting the transverse currents of air. The whole is covered with washed ferru- ginous residues to the depth of O m 12 (about 4f in.) or O m 15 (6 in.). At the angles of the pile, near the bottom, small cavities are fashioned by means of blocks of wood, and in these small pieces of wood and chips are placed for lighting the pile, which is now done. In the course of the day the wood is entirely consumed, leaving holes which are stopped up, when combustion continues to be pro- pagated through the mass at the expense of the sulphur in the ore, and the process of roasting proceeds uninterruptedly during eight or ten months. After the lapse of five or six weeks sulphur begins to appear, and is collected. When sulphurous acid has ceased to be evolved, the pile is left to cool for a month, after which the cover is removed and the contents taken to the place where they are broken up. Sulphur is sublimed from the iron-pyrites, and at first condenses near the surface of the pile ; but as the temperature increases it melts, and at this period hemispherical cavities, O m 25 (about 10 in.) in diameter, are made in the cover by plunging into it a mould of that shape. These cavities are placed where the disengagement of sulphur appears to be most copious. The sulphur is daily removed from them by means of an iron spoon and poured into a little wooden tub. After some time the cavities become unproductive, when the cover around them must be removed and replaced by fresh matter. The yield of sulphur is only 0-2 per cent, of the ore, or 0'4 per cent, of the total sulphur in the ore. Styrian kilns. These kilns were first adopted at Agordo by Lurzer about 1853, and have been found to yield decidedly .better results than the old method of roasting in piles or open heaps. The con- struction of them is represented in the accompanying woodcuts: in fig. 115 a kiln is shown partly in side elevation and partly in section 442 KERNEL-ROASTING AT AGORDO. on, the line A, B, C, D, E, F, fig. 116, which is a horizontal section on the line G, H, fig. 115 : both these figures are copied from the engrav- ings accompanying Haton's memoir in the Annales des Mines; fig. 117 is a transverse section on the line K, L, fig. 116, and is executed from a drawing supplied by Liirzer to myself. I cannot improve on Haton's description, of which I subjoin nearly a literal translation. Fig. 115. Side elevation and section on the line A BCD E F, fig. 116. Horizontal section on the line G H, fig. 115 Fig. 117. Vertical section on the line K L, fig. 116. The kiln consists essentially of a long rectangular enclosure within four walls. It is divided into sections, and the bed of each of these consists of four inclined planes, sloping towards the four angles respectively, thus forming a low pyramid. Along the lines of inter- section of these planes with each other are gutters ; and other gutters, a, a, a, a, are also traced upon the faces of the pyramid, of which b is the apex, or the point of junction of the four planes. In the sides of the kiln and on a level with the ground are channels, c, c, figs. 116, 117, which terminate on the outside in vessels of the shape of a quarter of a sphere : the object of this arrangement is twofold, namely, to regu- late the admission of atmospheric air into the interior of the mass, and to allow the sulphur deposited on the bed to flow outwards. There is a channel at each angle for the supply of air required on the ignition of the kiln. In the walls are cavities or chambers, e, e, intended for the collection of the sulphur ; they communicate with the interior by means of nine channels, as shown in fig. 115, which incline and con- MODE OF CHARGING. 443 verge somewhat towards the exterior. The length of the kiln is arbitrary. They are protected from rain by roofing of planks. Mode of charging. This is begun by placing at the four angles the wood necessary for lighting the kiln, and covering the gutters with flat pebbles of flint. At the top of each pyramid, 6, chimneys are built up of superimposed courses of ore, which are separated from each other by cubical pieces, so as to form openings into the chimneys. A mixture of ore-dust and water is used in this operation of build- ing. By this arrangement the distribution of air through the whole mass is effected, and kernels are produced from the ore-dust or schlich, which could not be obtained by the old process. The whole of the space between the walls and chimneys is filled with ore, taking care to add alternate layers of large and small ore, and to place at intervals some beds of chips in order to accelerate the ignition of the kiln. This done, it only remains to rebuild that portion of the wall which it was necessary to pull down in order to remove the former charge of roasted ore, and then to set fire to the wood. When ignition has once been properly effected by means of the wood, combustion is maintained at the expense of the sulphur in the ore, just as in the case of the piles. The roasting is completed in five or six months, and about 288 tonnes of ore are treated at a time. After emptying the kiln the lumps of ore are separated from the ore-dust or schlich and broken with hammers by children, who then detach the kernels from the earthy shells of oxide of iron in which they are enclosed. It is important that none of the kernels should be passed over, as in that case the copper contained in them would not be extracted in the subsequent process of washing the shells to extract the metallic sulphates, and would, consequently, be wholly lost; 4 and it is also desirable that the oxide of iron coating should be separated as com- pletely as practicable. Any imperfectly roasted lumps are put aside to be again roasted. The average proportions of these various products are stated by Haton to be as under : 13-26 per cent, of kernels and 86-74 of ferrugi- nous shells; of the total weight of kernels 11 per cent, are poor, and 89 per cent. good. The average produces are, for Good kernels 5*02 per cent, of copper. Bad do 3-45 do. Ferruginous shells, deduced by calculation ...0-7041 do. The consumption of wood required per tonne of raw ore is Of cord wood 8t 025 (0-883 cub. ft.) Of chips 0*001 (0 03532 cub. ft.) The roasting requires 7 men, who are employed piece-work, and the 4 Mr. Petherick assures me that in a wards thrown away ! Mr. Petherick spoke most confidently to me on this point, and presented me with specimens of the recent journey to Spain he visited a loca- lity where a large quantity of cuprife- rous iron-pyrites is raised, from which the copper is extracted as sulphate after the roasting of the ore ; and that all the kernels formed during the roasting were washed along icith the shells, and after- \ able to be thrown away. roasted ore which he had himself col- lected, and which contained distinct and characteristic kernels, small, it must be admitted, yet, one should think, too valu- 444 KERNEL-ROASTING AT AGORDO. breaking of the ore 121, who ought each to furnish daily 54 kilo- grammes (119 Ibs.). The quantity of ore treated annually is 15302 tonnes. Mr. David Forbes informed me (1855) that in Norway, where poor ores of from only to 1 per cent, produce of copper are roasted, kernels are produced which contain generally not more than 7 and never more than 15 per cent, of copper, the former percentage being found the most profitable. Sulphur is first driven off from the iron-pyrites at the bottom of the kiln, which is the first part to be heated : it runs along the bed into the cavities on the outside prepared for the purpose. As com- bustion gradually extends upwards through the mass, it escapes suc- cessively through the different rows of inclined channels, from which it falls on the bottom of the chambers in the walls of the kiln. Towards the end of the process some cavities are made along the top of the kiln for the collection of sulplrur. The crude sulphur thus obtained being necessarily very impure, is refined by fusion in cast- iron vessels. When the sulphur is melted, it is left for some time at rest to allow the foreign matter either to rise to the surface or fall to the bottom. That which rises is skimmed off, and the sulphur is then laded out into wooden moulds formed of two symmetrical parts, which may be clamped together. The charge of sulphur refined at one time is 150kilogramm.es ( = about 330 Ibs.), and requires from four to five hours. The marketable sulphur obtained amounts to about I'l per cent, of the weight of the ore, or 2-2 per cent, of the total sulphur in the ore. Loss of copper. It is estimated that in the Agordo process, inclusive of all the operations practised, the loss upon 100 parts of copper amounts to 7 '2 14 or T J T th of the whole. The changes which the ore undergoes in kernel-roasting. These changes have been carefully investigated by Liirzer, who has described and delineated the appearances presented by the ore during the successive stages of the process, which he divides into four, as follows : 1st stage. When a lump of ore in this stage is broken across, it is seen to consist of a central mass of unchanged ore, enclosed, as it were, in a rind or shell of a reddish-brown substance like sesquioxide of iron ; and between the two is interposed a thin more or less continuous layer, which differs in lustre from, and contains more copper than, the original ore, and in appearance resembles copper-pyrites. Fig. 118, a, I, represents the structure of lumps of ore in this stage, which were taken out of the kiln 011 the 8th and 10th days respectively after lighting. They have been executed from the drawings of Liirzer to which I have previously referred. 2nd stage. This stage occurs at about the middle of the roasting process. The external appearance of the ore is the same as in the 1st stage, but the weight is much diminished. On breaking a lump across, several concentric layers may be observed. In the centre is a nucleus of unchanged ore, surrounded first with a layer similar in appearance to copper-pyrites ; secondly, with a layer having a greater lustre and of a ANALYSES OF KERNEL AND SHELL. 445 reddish colour, similar in appearance to purple copper-ore (3Cu 2 S-f- Fe 2 S 3 ) ; thirdly, here and there, with a layer having a metallic lustre and varying from the colour of indigo copper-ore (CuS) to that of vitreous copper (Cu 2 S) ; lastly, with a thick red-brown crust, forming the outer shell. It will thus be observed that in these concentric layers within the outer shell the proportion of copper successively increased towards the unchanged nucleus of ore ; and it has been found ihat the copper retained in the outer shell is not equably distributed through the mass, but likewise increases towards the interior. When a lump of ore in this stage is broken across, while the interior is still hot, the nucleus on the freshly fractured surface appears surrounded with a radiated ring of a bright red colour, which, however, soon afterwards disap- pears. In the opinion of Liirzer these rays, without doubt, indicate the course of the gaseous products evolved from the interior. Fig. 118. 3rd stage. On breaking across a lump of ore in this stage, which occurs when the roasting is nearly completed, a nucleus of unchanged ore can no longer be seen ; but within the now greatly increased outer red-brown crust some yellow, reddish, and bluish particles may yet be perceived ; and in the crust itself concentric stratification still con- tinues visible. 4th stage. In this, the final stage, on breaking across a lump of ore it is found to consist only of a central nucleus having the appearance of vitreous copper (Cu 2 S), or rather of rich copper regulus, and an outer red-brown shell not usually presenting any indication of concen- tric arrangement (fig. 118, c). Not unfrequently one, two, or more kernels may be formed in a single lump of ore, and the larger the lump the more likely is this to occur. The following analyses have been made in the laboratory of the Vienna Mint of a kernel, and of the richest or innermost portion of the shell, produced in roasting a rich copper-ore : Pure kernel. Innermost portion of the shell. Copper 41-64 Copper 3'31 Iron .- 28-70 Protoxide of copper (CuO) 1'58 Sulphur 29-28 ,, iron 0-10 Gangue 0-08 Sesquioxide of iron 85'70 Loss 0-24 Sulphur 0-92 Sulphuric acid 2-50 100-00 Gangue 2*85 Loss by heat estimated as water 3 04 100-00 This degree of richness in kernels is, of course, quite exceptional, and is due to the fact that they were derived from a copper-ore much 446 KERNEL-ROASTING THEORY OF THE PROCESS. richer than those which in the ordinary course are subjected to the process of kernel-roasting. The average produce of the kernels usually obtained has been previously given. By prolonging the roasting beyond the complete state, metallic copper will be found in the kernel. It should be borne in mind that the existence of the various com- pounds of copper described by Liirzer as successively formed in the lump of ore is inferred from external appearances, and not, if I am correct, from the results of analysis. Theory of the process. Karsten, Liirzer, Werther, 5 and Plattner have attempted to explain this remarkable process ; but their explanations, so far as I can understand them, appear to amount to little more than a detail of certain reactions, which, while they tend to explain the formation of regulus of copper, yet fail to render a rational account of the cause of the actual transference of the metal from every part of a lump of ore and its concentration in a small space in the centre. The phenomenon has been regarded as somewhat, if not strictly, analogous to what takes place in the formation of steel by the cementation process, in which particles of solid carbon are supposed to travel slowly into the interior of a solid bar of wrought iron while heated to strong redness. But this process of cementation is still very obscure, and as much needs explanation as that of kernel-roasting itself. The results of kernel-roasting clearly establish the fact that when copper-pyrites intermixed with a large excess of iron-pyrites is gra- dually roasted, the latter may be to a very great extent converted into sesquioxide of iron, while a large portion of the copper remains in combination with sulphur. When iron-pyrites is roasted with access of air at a gradually increasing temperature, sulphate of protoxide of iron is formed, which is subsequently converted into basic sulphate of sesquioxide at the expense of the oxygen of part of the sulphuric acid ; and, finally, this basic sulphate, by exposure to a higher temperature, is decomposed, sulphuric acid being evolved in the anhydrous state, which, if the temperature be sufficiently high, may be resolved into sulphurous acid and oxygen. Plattner admits that in roasting iron- pyrites and disulphide of copper analogous reactions occur; and that his contact-action plays a prominent part (see pp. 247, 248, ante). In kernel-roasting reactions appear to take place in some respects resembling those which occur in copper-smelting, except that silica takes no part, and the iron is separated in the free state as sesquioxide. There is probably the same play of affinities between oxide, sulphate and disulphide of copper, and between oxide of copper and sulphide of iron, &c., as are observed in smelting ; and this, I believe, is about all that can in the present state of our knowledge be stated concerning the theory of the process. I subjoin an extract, nearly literally translated, from Plattner's description of the theory of the process, from which, I think, it will be admitted that he had not the clearest possible notions on the Berg. u. hiittenm. Zeit. 1853, 12. p. 439. WET METHODS OF EXTRACTING COPPER. 447 subject. " The disulphide of copper remaining behind 6 which comes in contact with the sulphur vapours uninterruptedly escaping in small quantity from the interior towards the surface, where the heat operates, and which, therefore, is not only protected from oxidation, but is also exposed to a higher temperature, produced by the con- tinuous oxidation of the sulphide of iron and of the sulphur vapours with the fonnation of sulphurous acid and protoxide of iron, as well as of the chief part of the sulphurous acid with the formation of sulphuric acid passes over in a liquid state, and, in consequence of its affinity for sulphide of iron, combines with the sulphide of iron and sulphide of copper in immediate contact with it, so that an increase is there effected in the proportion of copper." When silver exists in the ores subjected to the process of kernel- roasting it is stated to become concentrated on the exterior. Mr. David Forbes thus writes to me on the subject (Jan. 15, 1855): " When silver is present in the ores, it appears to travel outwards, and, in some specimens, I have seen the outer surface of the piece of roasted ore covered by a most beautiful thin shell of metallic silver, as if electro deposited." WET METHODS OF EXTRACTING COPPER. PRECIPITATION OF COPPER FROM SOLUTION BY IRON. The water of copper-mines may contain sulphate of copper in solu- tion. This is the case in the mines of the Isle of Anglesea, where the water is conveyed into shallow pits or tanks containing plates of cast- iron. 7 The copper is precipitated in the metallic state by the iron, and sulphate of protoxide of iron is formed, which, by prolonged ex- posure to the atmosphere, produces a yellowish-brown mud, consisting of basic sulphate of sesquioxide of iron. The copper is detached at intervals from the surface of the iron, and the solution renewed. The precipitated copper, which is termed cement-copper by the Germans, is melted and refined. The Anglesea or Mona copper has long been highly prized in Birmingham on account of its malleability and ductility. BANKART'S PROCESS." The process was founded on the principle of converting the copper of sulphuretted ores into sulphate by roasting with access of air, and was carried on for some time in the vicinity of Neath and then discontinued. The pyritic ores of Cuba were employed, which con- tained, on an average, 16 per cent, of copper. The ore was ground under mill-stones. The calciner was constructed in a peculiar 6 That is, after oxidation of the iron. 7 Vid. Briefe iiber die Insel-Anglesea vorziiglich iiber das dasige Kupfer-Berg- werk u. die dazu gehorigen Schmelzwerke u . Fabriken A. G. L. Lentin. Leipzig, 1800, p. 68 ; also Aikin's Tour tbrough North Wales, etc., 1797. 8 A.D. 1845, Aug. 7. No. 10,805. Fre- derick Bankart, " Improvements in treat- ing certain metallic ores, and refining the products therefrom." I am indebted to Mr. F. Bankart, junr., for the following de- scription. 448 BANKAKT'S PROCESS. manner. The floor consisted of tiles, covering long fluos, which returned through a double roof. The calciner was charged with 2 tons ( = 20 barrows) of ground ore, which was roasted during 24 hours, and stirred at intervals, atmospheric air entering freely through holes at the fire-bridge end and the side-doors, of which, except during the stirring, the lower part only was left open, so that the air might impinge more directly upon the ore. The products of combustion in the fire-place were not allowed to pass over the ore. The calcined ore was drawn out into iron trams, which ran into archways under the bed of the calciner, and which were raised by means of a crane to the wash- ing floors. These floors were filled with wooden dissolving vats, 4 ft, 6 in. wide and 3 ft. deep; they were arranged in three tiers, one above another, so that the solution might flow successively from those on the highest to those on the lowest tier. Each vat, at the height of 3 in. above the bottom, had a false bottom, pierced with holes and covered with canvas, tow being stuffed tightly in round the circum- ference. A wooden spigot was inserted, so that any liquid contained in the space between the two bottoms might be drawn off". Each vat was provided with a steam-pipe and tap. Over each tier of vats was a tramway. Water was supplied to the highest tier of vats by means of a pump and troughs. The canvas on the false bottom having been well wetted, the calcined ore was dropped into the vat from the bottom of the tram, evenly spread, and then covered with as much water as could be safely introduced. Steam was injected into the water through a vulcanised india-rubber pipe, screwed on to the steam-pipe, until complete ebullition occurred. The solution was now drawn off into the next vats below, which had been previously charged with ore, when the heat developed by the action of the anhydrous sulphates in the ore was generally sufficient to cause the hot solution to boil. In like manner this solution was drawn off into the vats on the lowest or third tier. At length a clear saturated solution of sulphate of copper, containing some sulphate of iron, was obtained. The residual ore in the vats was washed with fresh water so long as any sensible amount of sulphate of copper remained undissolved ; practically three " waters" were almost always found sufficient for this purpose. On the ground floor were several rows (about 12) of similar precipi- tating vats, supported on blocks and filled with broken cast-iron or iron plates, each row being 6 inches higher than the next in succes- sion. The solution of sulphates from the lowest tier of the dissolving- vats was allowed to run into the highest precipitating-vats by means of a trough placed under the spigots of the former. A piece of board fixed against the point of influx directed the course of the solution to the bottom of the precipitating -vat. When the vat became filled, the solution, deprived of a greater or less amount of copper, overflowed through a short leaden flap into the vat immediately below, in which also the stream was directed to the bottom. When the solution had thus passed through the series of about 12 vats the whole of the copper was precipitated. It was necessary constantly to brush the broken iron and scrape the plates in order to detach the deposited copper and UANKAUT'S PROCESS. 449 expose a fresh surface of iron to the solution. The deposit of copper on the iron plates was so copious and dense that it could be separated in the form of a sheet. Before removing any of the copper in a vat, the whole of the liquor was drawn off into the vat immediately below, and fresh water thrown on the iron and deposited copper in order to wash off any adherent sulphate. The solution left after complete pre- cipitation of the copper contained sulphate of iron : it was evaporated in leaden tanks fixed over flues covered with tiles and then drawn off into vats, in which the salt crystallized. The lixiviated ore was transferred from the washing vats by copper shovels into wooden trams and conveyed to the roofs of the calciners, where it was spread over the surface and left to dry. When dry it was mixed with i-th part, or 4 barrows, of fresh ore, and the whole was calcined during 1 2 hours, the oxides of iron and copper in the lixiviated ore greatly facilitating the conversion of the sulphur in the fresh ore into sulphuric acid. The calcined ore was lixiviated as above described, and generally a larger quantity of copper was extracted from the product of the second calcination than from that of the first. The process was repeated three times, and on the last occasion the fresh ore consisted exclusively of iron pyrites. There were thus three calcinations in all. The several tiers of vats were reserved for ore in the successive stages of calcination. The vats on the uppermost tier were kept for fresh ore, those on the second for the product of the second calcination, and so on. In like manner a similar arrangement was adopted in regard to the calciners. The ore from the first tier of vats was conveyed to the top of the second pair of calciners, and so on. . The copper was melted and refined in an ordinary refining furnace. According to Mr. Bankart this method is liable to the following objections : 1 . The necessity of grinding the ore ; 2. The great bulk of matter which had to be removed several times during the entire process ; 3. Especially the loss of copper arising from leakage ; 4. The residual ore contained more copper than ordinary ore-furnace slag; 5. The value of the sulphate of iron obtained as an accessary product was far less than the cost of the iron and the crystallization ; 6. The refinery slag produced cannot be economised in the same manner as in the process of smelting. The second objection might, in great mea- sure, have been removed by a preliminary fusion and the concentration of the copper in a regulus. The great bulk of the ore would thus have been removed at once, and the expense of grinding the regulus would have been small compared with that of grinding the entire ore. The special advantage of the method is stated to be the production of copper remarkable for purity and malleability ; and, at first, sanguine hopes were entertained that it would successfully compete with the process of smelting. Experiments were made under the observation of the late Mr. Kichard Phillips with calciners only large enough to re- ceive each about a ton of ore at a time ; and it was considered that the experiments were made under very unfavourable conditions com- pared with such as might be conducted on a larger scale with every appliance suitable. The ore contained 11-62 per cent, of copper, and 2 G 450 W&T PROCESS BY M. ESCALLE. was passed through a sieve of 81 holes to the square inch. Mr. Phil- lips estimated that "the probable expense in coals and wages of ex- tracting a ton of copper precipitate by Mr. Bankart's process, under the proposed improved arrangements, would be from IL 15s. Id. to 1 5s. respec- tively, according to the per centage of the ore." 9 It is wise never to place too much reliance on the estimated cost of production in comparatively untried metallurgical processes ; for metallurgical, like engineering, estimates, have often proved sadly fallacious, and entailed ruin upon too confiding adventurers. The correctness of the preceding estimate of Mr. Phillips, which was, doubtless, most conscientiously given, was not subsequently confirmed. Mr. Bankart has, during many years, abandoned the process, in which he was once a stanch believer, and has ever since been engaged in. the ordinary process of smelting. The stern logic of facts is inexorable, and it is the only safe guide in deal- ing with the schemes of inventors. In Mr. Phillips's report it is stated that the quantity of copper remaining in the residue was not less than OG2 per cent., a quantity exceeding that which" should be allowed to escape in ore-furnace slag. From what has been previously advanced in this work 'the theory of Bankart's process should be quite intelligible. There is, however, one point which requires explanation, namely, the use of iron pyrites alone in the last calcination. When this sulphide is roasted under suitable conditions, sulphate of protoxide of iron is produced in the first instance, and is afterwards entirely decomposed, sulphuric acid being evolved and sesquioxide of iron left as a fixed residue. Sulphate of protoxide of iron is decomposed at a lower temperature than sulphate of prot- oxide of copper. Hence, in the last calcination, iron pyrites is em- ployed in order to furnish sulphuric acid to any oxide of copper which may be present in the calcined ore, and which otherwise might not be extracted. WET PROCESS BY M. ESCALLE. l This method was practised some years in the vicinity of Marseilles, but the result was not successful, at least in a pecuniary point of view. The ore (containing copper in combination with sulphur), in very fine powder, was calcined in a double-bedded reA 7 erberatory furnace. The calcination took place on the bed furthest from the fire. 1000 kil. were thus treated in 12 hours, with a loss in weight of about 17/ . The calcination finished, the ore was transferred to the bed nearest the fire, upon which hydrochloric acid at 16 (sp. gr. 1-124) was introduced. The " chloruration " lasted three hours, and the charge consisted of only about one-fourth of that drawn from the 9 Printed Copy of Report of Richard Vide Bulletin de la Societe de 1'In- Phillips, Esq., F.R.S., F.G.S., etc., on dustrie Minerale, T. III. 4 Wme Livraison, Mr. Frederick Bankart's Patent Process ! 1858. Notice sur les Usines a Cuivre et for reducing Copper Ores. Craig's Court, | les Usines a Antimoine dcs Bouches-du- Nov. 26, 1847. This estimate is exclusive | Rhone. Par M. L. Simonin, ingenicur of the iron required for precipitation. civil a Marseille, p. 559. HAHNER'S PATENT. 451 calcining bed ; the whole was rabbled. The chlorides thus produced were dissolved out by water, and the solution was decanted into wooden vessels, polysulphide of calcium, obtained from residues of soda and soapworks by what is called " fermentation," being introduced at the same time. The copper was first precipitated, the other metals only being subsequently thrown down. As soon as the liquid ceased to turn blue on the addition of ammonia, the supply of polysulphide was shut off. The supernatant liquor was decanted, and the precipitated sulphide of copper collected. It was drained, dried, compressed, and then moulded into cakes (pains). Sulphide of copper (CuS), with ex- cess of sulphur, nearly chemically pure, was thus obtained. These cakes were heated in a sort of blast-furnace in alternation with layers of charcoal. The product obtained was a blackish scoriaceous mass, or copper-sponge, mixed with metallic globules ; the sulphurous acid generated in this operation was conveyed into leaden chambers to be converted into sulphuric acid. The copper-sponge was remelted, with- out any addition, in a reverberatory furnace, and yielded in a short time very pure metallic copper. HAHNER'S PATENT. The specification of this patent 2 is in some respects unintelligible to me ; but as far as I comprehend it, the process consists essentially in exposing oxide of copper in admixture with chloride of sodium and silica to a red -heat in a reverberatory furnace, by which means chloride or oxy chloride of copper and silicate of soda are stated to be produced. Ores in which the copper exists in the state of oxide only require to be pulverized, but sulphuretted ores of copper are first to be roasted sweet. If the ore contains no silica, about 10 per cent, of this substance must be added. The chloride of sodium, or other chloride, is inti- mately mixed with an equal weight of roasted ore, and, if dry, the mixture is moistened. " The moistened chloride, or mixture of chloride and roasted ore, ought then to be. incorporated as intimately as pos- sible with the red-hot ore in the furnace, and kept in a continual movement and at a red heat until the smell of muriatic acid becomes less perceptible, and the ore commences to adhere to the workmen's tools," when it is withdrawn from the furnace and a fresh charge added. The roasted product is then lixiviated, if practicable, while still hot, with water acidulated with hydrochloric, sulphuric, or other acid. The copper is thus obtained in solution, and may be precipi- tated therefrom by any of the ordinary methods. Mr. Henderson has obtained a patent for the extraction of copper from ores and certain cupriferous products, by heating them in admix- ture with chloride of sodium, or certain other chlorides, so as to volatilize the copper in combination with chlorine, and condense it in a suitable apparatus.* It remains to be seen whether Mr. Henderson 2 A.D. 1856, No. 571. 3 A.D. 1859, No. 2900; and A.D. 1860, No. 2525. 2 G 2 452. KEMARKS ON THE PATENT LAWS. will succeed in effectually collecting the gaseous compound of copper : if he should, his condensing apparatus would be a valuable acquisition to the patentee, who proposed that in refining copper a blast of chlorim should be projected upon the surface of the melted metal ! < KEMAIIKS ON THE PATENT LAWS. Various patents have been granted for alleged improvements in the treatment of copper- ores, of certain products obtained in the smelting of copper-ores, &c., which are only worthy of notice as affording, as I conceive, satisfactory illustrations of the defective state of our existing Patent Laws. In one of these patents, the exclusive right is granted of extracting copper by the solvent action of acids from highly silicious ores con- taining oxides and salts of copper insoluble in water ; and also of ex- tracting copper by the same means from sulphuretted ores after they have been subjected to calcination. In another patent Her Majesty confers upon the patentee the monopoly of using sulphuric acid as a solvent for the extraction of copper from ores containing substances insoluble in acids, such as silica ! 5 That a man, who has worked out an original and valuable process from his own brain, and who may have incurred great expense in bringing it to a practical issue it may be after years of protracted toil and anxiety should have secured to him by law during a moderate term the exclusive privilege of reaping the substantial reward of his own invention appears to me as just and reasonable as that an author should be protected against piratical and unprincipled publishers. But that the law should confer upon a man the exclusive right of appropriating to his own benefit facts which are perfectly familiar to every tyro in chemistry, and of practising operations which are of daily occurrence in the laboratories of chemists, is as impolitic as it is unjust. And, surely, the particular "inventions" above referred to belong to this category. I cordially subscribe to the opinion ex- pressed by Mr. Grove, Q.C., namely, that the real object of patent law was " to reward, not trivial inventions, which stop the way to greater improvements, bat substantial boons to the public not changes such as any experimentalist makes a score a day in his laboratory, but substantial practical discoveries, developed into an available form." 6 Believing as I do that the existing Patent Laws frequently operate injuriously to the interests of the real inventor, while they afford 4 A.D. 1849, No. 12,534, dissolved out, is treated by iron, which 5 Vid. " Description du traitement precipitates the copper ; the cement- du Cuivre par cementation, pratique a copper is melted in a reverberatory fur- 1'Usine de Stadtberge, dans la Westphalie ; nace, and afterwards refined in the small par M. Achille Delesse, Eleve-Ingenieur hearth." The ores operated on contain des Mines." Ann. d. Mines, 4 s. 1842, 1. copper in the state of carbonate, and are p. 477. The process " consists in passing free from any sensible amount of lime, over the ores acid vapours which decom- I 6 Suggestions for Improvements in the pose the carbonate of copper and convert ! Administration of the Patent Law. it into sulphate ; the sulphate, after being REMARKS ON THE PATENT LAWS. 453 encouragement to the mere schemer and pilferer of other men's brains, I cannot forbear from directing public attention to the subject through the medium of these pages ; and I do so with the less hesitation as there is no manufacturing art which is more likely to suffer from the defective state or administration of these Laws than Metallurgy. At a recent meeting of the Institution of Mechanical Engineers held at Sheffield, Sir William Armstrong, who presided on the occasion, delivered an address on projectiles, in which he expressed an opinion strongly condemnatory of the present system of Patent Law, and even suggested that an entire abrogation of the law might be beneficial to the community, and not disadvantageous to inventors. Any person who has the pleasure of personally knowing Sir AVilliam Armstrong will be convinced that this opinion was as honestly formed as it was candidly stated ; and, although I am not prepared to admit that the protection of law should cease to be afforded to really meritorious inventors, yet I entirely concur in many of the observations contained in the following extract from Sir William's address 7 : " I am tempted to advert, before I conclude, to a subject intimately connected with mechanical progress, but upon which much difference of opinion may exist. That dauntless spirit which in matters of com- merce has led this country to cast off the trammels of protection has resulted in augmented prosperity to the nation, showing the injurious tendencies of class legislation when opposed to general freedom of action. Would that the same bold and enlightened policy were extended, in some degree at least, to matters of invention. Under our present Patent Law we are borne down with an excess of protec- tion. We are obstructed in every direction by patented inventions, which will never be reduced to practice by those who hold them, but which embrace ideas capable of useful application if freed from mono- poly. The merit of invention seldom lies in the fundamental concep- tion, but is to be found in the subsequent elaboration, and in the successful struggle with difficulties, unknown to the mere theorist, and often requiring years of labour, blended with disappointment, for their removal. Nothing can be more irrational, therefore, than to give equal privileges to the mere schemer and to the man who gives actual effect to an invention. Primary ideas ought to be the common property of all inventors, and protection, if we are to have it at all, should be sparingly awarded to those persons alone who, by their labour and intellect, give available reality to ideas. Apart from the impolicy of our present indiscriminate system, its operation is unjust. Philosophers who furnish the light of science to guide to useful dis- covery go altogether unrewarded and unrecognized. Practical men, who, like Watt and George Stephenson, devote the best part of their lives to perfecting inventions of immense importance to the world, seldom derive from patents any greater emolument than would flow to them without the aid of a restrictive system, while they are fre- quently involved in tormenting litigation about priority of idea. On From the Report of the Address in the Times, Aug. 2, 1861. 454 ASSAYING OF COPPER-ORES BY THE CORNISH METHOD. the other hand, we see numerous cases of disproportionate wealth realized by persons whose only merit has been promptitude in seizing upon and monopolising some expedient which lay upon the very surface of things, and required no forcing atmosphere of protection for its discovery. Finally, injustice is done by the existing law to those men who have no desire for monopoly, but who are compelled to become patentees for no other purpose than to prevent their being excluded from carrying their own ideas into practice. For my part, I incline to think that the prestige of successful invention would, as a rule, bring with it sufficient reward, and that protection might be entirely dispensed with. On this point, however, I speak with hesi- tation ; but it is, at all events, certain that extensive reform is urgently reqivired in this branch of legislation, and that the advance of prac- tical science is now grievously obstructed by those very laws which were intended to encourage its progress." 8 ASSAYING OF COPPER-ORES BY THE CORNISH METHOD. The successful performance of this process requires great skill and an education of the eye which can only be attained by long and con- stant practice. In the hands of an expert assayer it is capable of yielding results upon which the smelter may rely with implicit con- fidence, inasmuch as they are always below the furnace yields ; and, considering the nature of the process, it is surprising how closely diiferent assayers will approximate in their produces, except with par- ticular descriptions of ore. FURNACE AND IMPLEMENTS. Furnace. Fig. 119 is a vertical section through the middle of the grate of one of the furnaces used for assaying in the Metallur- gical Laboratory, a is the fireplace, lined with firebrick, 8 inches square by 12 inches deep ; b the ash-pit provided with a register door for the regulation of the draught ; the door may be entirely opened when the largest supply of air will be obtained, or it may be closed and the register f opened in a greater or less degree according to circumstances. This register is formed by a revolving disc of sheet-iron having a semicircular opening, the centre of which coincides with the centre of the door ; behind it is another similar semicircular opening in a fixed piece of sheet-iron, e, forming the door : the opening of the revolving disc may be made to coincide with the fixed opening behind, or it may be adjusted so as to regulate the 8 Since these observations were in type Sir W. Armstrong has read a paper at the recent meeting of the British Asso- ciation at Manchester, on the present system of Patent Law, and has strongly contended for its complete abolition. Mr. Fairbairn, the President, in his inaugural address, advocated the expediency and justice of protection, by patents, to really meritorious inventors ; but condemned the present indiscriminate system of issuing letters-patent, and animadverted strongly on that class of patentees who are chiefly distinguished by their magpie- like propensity of running away with the silver spoons of real inventors. FUftNACE AND IMPLEMENTS. 455 Fig. 119. admission of air with the greatest nicety. The flue, d, communicates with an upright shaft, about 60 feet high, with which five other similar furnaces are con- nected. \\hen practicable each furnace should have a stack to itself, or, what is equivalent, one stack may be divided into a series of fines, of which there should be one for each furnace. The exterior and the part below the grate are built of com- mon brick. The whole of the brickwork is kept firmly bound together by means of cast-iron plates and wrought- iron tie rods. The furnace mouth can be closed by means of the firebricks, -/, A, each of which is clamped with a piece of flat bar-iron firmly wedged at one end; sometimes a small screw is used instead of wedging : these firebrick covers are of two sizes, and the larger one only need be removed when the crucible is taken out of the furnace. The draught may be regulated not only by the register door of the ash- pit, but also by opening the furnace top in a greater or less degree, or by placing a piece of firebrick in the opening into the flue c, or by means of a damper at the top of the stack. With such arrange- ments perfect control may be obtained over the temperature of the furnace ; it can be kept below a dull red or increased sufficiently to melt nickel or manganese. In copper assaying in a furnace of the dimensions stated, 2 or 3 fusions for regulus or 4 calcinations can be made at one time. Furnaces specially constructed for copper assaying are of larger dimensions, so that 4 or more fusions for regulus, or from 6 to 8 calcinations, may be carried on at once ; but the size and number of furnaces vary according to the requirements of the assayer. The fuel employed is coke. Crucibles. Those known as " Cornish Crucibles" are the best: they are sold in nests of two and sometimes three each. The large size (No. 1, fig. 59, p. 222) is used for calcining the ore and fusion for regulus. The smaller and middle sizes are used for calcining the regulus, fusion for coarse copper and refining, according to the quantity operated on. They are sold without covers, a substitute for a cover being made, if required, from a piece of broken crucible. When several assays are carried on simultaneously, the crucible before use should be marked with red chalk, " reddle," or a mixture of haematite and water. 45G ASSAYING OF COPPER-ORES BY THE CORNISH METHOD. Crucible rests or stands. See fig. 1 20, A. They consist of a strong iron ring with a slot on each side to receive the ends of the tongs. They are of two sizes, and are used by some assayers for supporting crucibles when out of the furnace. Crucible tongs. Fig. 71, p. 231, represents those generally employed by copper assayers : they vary somewhat in size. Scarifiers or roasting dishes. They are. small, flat, shallow, cup-shaped vessels, made of fireclay, and are employed in roasting or calcining ore or regulus in a muffle. Those used for assaying silver-ores l will answer, but if they are made specially for the copper assayer they should be somewhat wider and shallower. The following dimensions are recommended : the cavity about 2 inches in diameter at the top and f of an inch deep in the centre ; th of an inch in thickness at the top; the height, including the foot, li inch. In calcining ore one of somewhat larger dimensions is necessary, viz., 3 inches in diameter and ths of an inch deep. Flux-spoon or ladle. It is made of copper, and is used for measuring out the fluxes. Thickness of bowl T ^th of an inch, width at the top 11 inch, depth i inch. The handle is upright, and made of stout copper wire 7 inches in length. Ladle. Fig. 120, G. It is made of copper, and is used for drying the sample or for washing the ore to ascertain the nature of the minerals or gangues present. Regulus bowl or pan. It may be made of copper, zinc, wood, or earthenware, but the first is to be preferred : it is circular, about 9 or 10 inches diameter at the top and 5 or 6 deep, and is kept partly filled with water for quenching the regulus, slags, &c. About 1 inch below the upper rim is fixed a flat metal ring of sheet-iron, copper, or zinc, about IT or 2 inches wide, so as to form a shelf for receiving temporarily the products of fusion. Instead of a shelf, a number of small, finely-perforated, flat copper ladles can be used, made of thin sheet-copper 2J to 3 inches diameter at the top and \ inch deep in the centre, with an upright handle of copper wire from 3% to 4 inches high, and bent in the form of a hook at the top, so that they can be readily suspended from the inside of the bowl and removed when required, the water draining off in the removal ; the bottom should be made somewhat flat, that it may stand upright on the table. Forceps. Two are required, one of copper, fig. 120, B, and one of steel, fig. 120, C ; they are used for manipulating with the slags, regu- lus, &c. Copper scoop. Fig. 120, F. It is used for transferring the fluxes, &c., to the crucible either when out of or in the furnace. Mould of iron. Fig. 120, H. It is of the shape generally used by cop- per assayers; one with somewhat deeper and more hemispherical cavities is preferable : it is employed to receive the slag and regulus poured out after fusion. To prevent the fused products from adhering to it, the cavities from time to time should be rubbed with an oiled rag and They will be described hereafter. IMPLEMENTS. 457 dusted over with charcoal, or smoked over a gas-flame, or the}' may he ruhhed over with graphite, or a mixture of tar and tallow. Fig. 120. Calcining rods. Fig. 1 20, 1. The blade is usually of wrought-iron, hut it is better of steel ; it is kept clean by the aid of a rasp or file to pre- vent adhesion of regulus during roasting : a separate rod is employed for each crucible. When the roasting is conducted in a muffle, a roasting spatula is required ; it is made of steel, and is 6 inches in length ; it is flattened out at one end into a blade about 2 inches long and from |-ths to fths of an inch in width, the other end heing fashioned into a point. The blade end is useful for stirring, and the point for breaking up any agglomerated particles during temporary removal from the muffle. A stout iron rod, from 2? to 3 feet in length, flattened out at one end and bent at right angles at the other, is used for occasional stirring within the muffle. Steel anvil It is used for flattening the copper button upon, &c. ; its face should be about 3 inches square. Hammer. Chisel. It is required for cutting the copper buttons, &c. 458 ASSAYING OF COPPER-ORES BY THE CORNISH METHOD. Iron slab, or flat plate of cast-iron, about 15 inches square and I inch thick, used for breaking down the slags upon, &c. The iron ring, fig. 120, E, serves to prevent particles from being lost by projec- tion during manipulation. 2 Mortar and pestle. Fig. 120, D, is the mortar made of bronze, and the pestle of iron or steel ; a cover is required with a depending rim and having a hole in the centre to admit the handle of the pestle ; by this means any particles of the regulus are prevented from being lost by projection out of the mortar during trituration. An iron or steel mortar not quite so deep is to be preferred ; a camel-hair pencil or hare's foot is employed in dusting out the mortar. Sieves. For sifting samples ; they are from 9 to 1 2 inches in diameter, and are made either of hair or wire, with from 40 to 60 holes to the linear inch. Slab of cast-iron. About 22 inches long, 11 wide, and 1 thick, with a border about 1^ inch high, and a bruiser with a face about 11 inches wide, 7 long, and If thick, for grinding down the ore ; or a large cast-iron mortar and pestle of the following dimensions may be used as a substitute internal diameter, near the top, 9 inches, and depth 9 inches. The pestle may be attached by means of a rope to a spring fastened in the wall or ceiling to relieve the arm during trituration. Balance. It should carry about 500 grains, and turn with from Vth to -jLth of a grain. Weights. Special weights, of which the unit is termed cent, are used by copper assayers to facilitate calculation. The following table gives the actual weights of the conventional weights used in the assaying of copper-ores in Cornwall, &c. : 100 50 25 20 15 12 10 9 8 7 Grains. 400 200 100 80 60 48 40 36 32 28 roy Wei Dwts. & it. Grains. 16 t 16 8 8 4 4 3 8 2 12 2 1 16 1 12 1 8 1 4 Assay Weights Actual Troy Weight. in cents. Grains. Dwts. Grains. 6 = 5 = 4 = 3 = 2 = i = \ = i! - 24 20 16 12 8 =1 o 0-5 = 0-25 = 20 16 12 8 4 2 1 I The assays are reported on the 100 parts, the unit being subdivided into , i, i, and T J ths, and not into decimal parts, so that 281 per cent, produce is equivalent to 28-75 per cent., &c. FLUXES, REAGENTS, &c. They should be kept in common covered earthenware jars, or in a long rectangular wooden box divided into compartments. The follow- ing fluxes are used : 2 An upright block of wood, about 2 to 3 feet high and 14 inches diameter, having a circular plate of cast-iron about 5 to f inch thick and 14 inches diameter, and an inverted border 2 \ inches deep, will be found very useful for many purposes in an assay laboratory. FLUXES, REAGENTS, &c. 459 Borneo. Dried or calcined borax should be used. Crystallized borax or biborate of soda (NaO, 2B0 3 -flO HO) contains 47-1 per cent, of water; for the purposes of assaying it is desirable to expel this water : when crystallized borax is heated, it melts in its water of crystallization and afterwards swells up to a very light, white, porous, extremely bulky mass, which is dried or calcined borax. If the temperature is afterwards raised, it fuses into a clear, transparent, colourless liquid, which, after solidification, constitutes glass of borax, which is free from water. Much of the so-called dried borax of commerce contains a considerable amount of water, which causes it to swell up or intumesce when heated, an effect which never occurs when it is perfectly dried. Borax has the property of forming fusible compounds with the earthy and metallic oxides, such as lime, oxide of iron, &o., and of rendering silicates more fusible by the formation of borosilicates. Glass. Plate or window-glass should be selected, or common green bottle-glass may be used ; but flint-glass, which contains oxide of lead in large quantity, should be avoided. With the exception of the latter, all varieties of glass in commerce consist essentially of silica in combination with a fixed alkaline base chiefly soda and lime. Green bottle-glass owes its colour to the presence of a sensible amount of oxide of iron. By heating it in a muffle or in a crucible, and plunging it while red hot into cold water, it breaks into small pieces and can then be readily reduced to powder. It is useful in assaying ores containing lime and other earthy bases with little or no silica, and serves to economise borax. Quartz in fine powder may be sub- stituted for it, but glass is to be preferred on account of its fusibility. Lime. (CaO). Common powdered unslaked lime should be used; but the best is prepared by strongly heating pieces of Carrara marble in a muffle or large crucible until the carbonic acid is expelled ; the lime should be slaked when cold, and the hydrate thus formed should be again heated to expel the water ; by this means a finely-divided granular variety of dry lime is obtained, which does not absorb car- bonic acid with so much rapidity as some varieties of lime. It acts as a flux for silica, silicates of alumina, &c. Fluor-spar. Fluoride of calcium (Ca Fl). Care should be taken in selecting it to see that it is free from copper-pyrites, blende, and especially from galena; any lumps containing these minerals should be rejected : it is also desirable that it should be free from quartz, as the presence of this substance would interfere with the easy adjust- ment of its proportions. The fluor-spar from Derbyshire and the north of England is the best for assay purposes. It melts at high temperatures, and forms very fusible compounds with sulphate of lime (gypsum), sulphate of baryta (heavy spar), phosphate of lime, and silica (see p. 43 ante). It increases the fluidity of the slag, and aids in imparting to it the property of cracking up readily when cold. Nitre or saltpetre. Nitrate of potash (KO, NO 5 ). It is an anhydrous salt, but the crystals contain small quantities of water mechanically 460 ASSAYING OF COPPER-ORES BY THE CORNISH METHOD. diffused. Common commercial varieties, powdered and dried, should be used. It ma}^ be conveniently obtained in a fine state of division by dissolving it in water, and evaporating to dryness, stirring all the while during the latter part of the process. It fuses at a gentle heat without undergoing decomposition. Fused nitre is known as sal prunella. Nitrate of soda may also be employed. These nitrates act as powerful oxidising agents by virtue of the large amount of oxygen which they contain, whereby the sulphur of metallic sulphides is converted into sulphurous or sulphuric acids, and the metals into oxides. Salt. Chloride of sodium (NaCl). It decrepitates when heated, and fuses at a red heat into a limpid liquid, and afterwards volatilizes in the form of dense white fumes. On account of its comparatively low specific gravity it generally forms the upper layer in a process of fusion. It should be dried and powdered, and should be free from sulphates which might cause it to retain copper. Some assayers use both dried and undried salt. Carbonate of soda. -^-Crystallized carbonate of soda (NaO, C0 2 +10HO). It contains about 63 per cent, of water ; when heated it melts in, and afterwards loses, its water of crystallization, becoming a white porous mass. Thus freed from water, it should be used for assay pur- poses. Bicarbonate of soda, carbonate of potash, or a mixture of car- bonate of potash and soda may be used instead of dried carbonate of soda. The alkaline carbonates form very fusible compounds with silica, &c. Tartar, or Cream of Tartar. Tartrate of potash, generally known as bitartrate of potash (KO, HO, T). When pure it is white ; but the varieties generally used by assayers are more or less coloured, and sold as red or white argol; they are cheaper and have a greater reducing power than cream of tartar ; they contain foreign carbonaceous matter, tartrate of lime, and other impurities present in the wines from which they are derived. Sulphur. It is most convenient to use flowers of sulphur, but pow- dered roll brimstone will answer. "When pure it is completely volatilized at a gentle heat. Charcoal Finely powdered charcoal is a most useful reducing agent; gunpowder charcoal is best, but anthracite or culm powder, or coke dust may also be used, and for some purposes as in calcina- tion are to be preferred, as they burn away less rapidly. Iron-pyrites. (FeS 2 ). Mundic or bisulphide of iron. It should be selected free from copper ; many varieties contain small quantities of this metal, but that occurring in the coal measures is generally free from it, on which account it is to be preferred for assay purposes. Sulphide of iron, prepared by adding sulphur to hot scrap or bar iron, or a mixture of haematite and sulphur, may be used as a substitute for iron-pyrites. Refining or white flux. It is obtained by plunging a red-hot iron rod into the following mixture, when deflagration ensues : REFINING OR WHITE FLUX SAMPLING. 461 By measure. Nitre 3 parts. Cream of tartar 2 parts. Salt 1 part. The mixture for this purpose should be contained in a large crucible or in a vessel of iron. It should continue to be stirred about until all deflagration ceases. A porous mass is left, which should be reduced to powder ; it is generally grey or reddish-grey in colour. The pro- portions used by different assayers vary somewhat ; some omit the salt, others use equal parts of red and white argol in place of cream of tartar. Owing to the varying amount of carbonaceous matter present in the different commercial varieties of argol, the refining- flux may differ considerably in its oxidising power. When a fresh batch of refining-flux is prepared, before use it is better to test its power on a small piece of melted copper : if it is too sharp, i. e. too strongly oxidising, argol may be added ; or if, on the other hand, it is not sufficiently oxidising, nitre may be added, \\hen nitrate of potash is deflagrated with tartrate of potash, the carbon is oxidised by the oxygen of the nitrate, and carbonate of potash is formed. Ee- fining-flux consists essentially of carbonate of potash, and any unde- composed nitrate of potash, tartrate of potash, and impurities present in these substances such as lime, &c. and chloride of sodium. Car- bonate of soda or potash mixed with a small percentage of nitre may be used instead of refining-flux. Some assayers prepare two kinds of refining-flux, one containing more nitre than the other ; the strongest, or that in which there is most nitre, is used for refining very impure coarse-copper buttons. SAMPLING. The method of sampling copper-ores at the mines has already been described. The samples are generally received by the assayer in brown paper parcels of about 1-J Ib. weight, occasionally in mass or in small lumps. The ore should be reduced to powder by grinding and rubbing on a flat iron plate, or by trituratioii in a large mortar, and passed through a sieve of from 40 to CO holes to the linear inch : the sifted sample should be well mixed. When native copper is present in the ore great care must be taken to ensure a fair average sample, as portions of the metal are often retained on the sieve. The best plan then is to take the total weight of the sample and to assay the sifted portion as well as the portion retained on the sieve. The sample may be dried before or after sifting. PRELIMINARY EXAMINATION. Before the sample can be assayed, it must be examined as to the character of the ore, the nature of its associated vein-stuff or gangues, the approximate percentage of copper, &c. ; and in this examination a practical knowledge of mineralogy is of great advantage. It is necessary to determine beforehand, whether, in the subsequent treat- ment of the ore, it will require calcination or not, the addition of nitre 462 ASSAYING OF COPPER-ORES BY THE CORNISH METHOD. or sulphur, &c. The assayer may obtain information for his guidance by resorting to the following expedients : a. Washing or vanning a portion of the ore in the copper ladle, fig. 120, G. b. Testing the ore by means of the mouth blow-pipe, or by chemical tests. c. Or by making a rough preliminary fusion for regulus, &c., on the ore direct. The last two methods may prove of service to the uninitiated until they acquire skill by practice. The first expedient is that occasionally adopted by assayers ; but a good practical assayer with a quick eye will, by simple inspection of the sample, generally, though not invari- ably, decide correctly as to the mode of treating it, and it is only in cases of doubt that even vanning is resorted to. CHIEF CHARACTERISTICS OF THE PKOCESS. The chief peculiarity of the Cornish method of assaying copper-ores is the concentration of the copper in a regulus ; and formerly even rich carbonates, oxides, &c., were assayed on this principle. Each assayer has his own peculiarities in manipulation, follows his own rules as to the nature and quantity of fluxes, &c. ; but whatever difference there may be in minor details, in the practice of different assayers, the Cornish method always comprises the four following operations : 1. Fusion for regulus. 2. Roasting of the regulus. 3. Fusion for coarse copper. 4. Refining. 1. Fusion for regulus, The object of the preliminary roasting or the addition of nitre in fusion will be explained hereafter. The result in either case is the expulsion of a portion of the sulphur as sulphurous and sulphuric acids, and the oxidation of a portion of the iron present in excess above that necessary to form a proper regulus, the oxide of iron thus produced being retained in the slag as silicate or boro-silicate of iron, while the remaining copper, iron, and sulphur, together with small quantities of antimony, zinc, &c., fuse into a regulus. 2. Roasting of the regulus. If the regulus be regarded as a compound consisting of disulphide of copper and sulphide of iron, by the process of roasting the whole of the sulphur is eliminated as sulphurous acid, and simultaneously the copper becomes converted into protoxide of copper, and the iron into sesquioxide and magnetic oxide. The clotting which occasionally occurs is due to fusion of the sulphides, or when the regulus is too fine to the easy fusibility of the disulphide of copper. If the calcination is conducted at too low a temperature, there is a great tendency to form basic sulphate of copper, which is difficult to decompose at a high temperature, such as may be employed in roasting, without causing the substance of the crucible to be acted upon. "When a roasted regulus, containing basic sulphate of copper, is afterwards reduced for coarse copper, disulphide of copper is formed in sensible quantity. Jf galena or sulphide of lead is present in the PROPORTIONS OF FLUXES, &c. 463 regulus, during roasting it is converted into oxide of lead and sulphate of lead, which, on subsequent fusion for coarse copper, are reduced with the separation of metallic lead, which unites with the copper, and with the formation of sulphide of lead. If sulphide of zinc, tin, or antimony be present in the regulus, by roasting they would be converted into oxides; and in the subsequent fusion a portion of these oxides would be reduced to the metallic state and pass into the copper. A portion of any other metallic oxides which may happen to be present may also be reduced and pass into the copper. Any arsenic which might be present would be partly given off as arsenious acid (AsO 3 ), and a portion would remain as basic arseniate, which, on subsequent reduc- tion, would be converted into an arsenide, and pass in part into the copper. 3. Fusion for coarse copper. The temperature employed should be sufficient, by the aid of the fluxes used, to reduce the oxide of copper to metal, and the oxide of iron to the state of protoxide, which should be taken up by the borax and be separated as slag. If the button of coarse copper be examined chemically, it will be found to contain small quantities both of iron and sulphur, and, it may be, of other metals also. The slag retains a small quantity of copper, but is some- times practically clean. 4. liejiniiig. This operation consists of two distinct stages. In the first the metal is fused and kept so until it becomes clear and bright ; sulphur is evolved as sulphurous acid, and the foreign metals present- are oxidized by the oxygen of the air and appear as scum, but traces of these metals still remain in the copper. In the second stage, on the addition of refining flux, a portion of the copper is oxidized, and dis- solves in the metallic copper as dioxide, and copper at tough-pitch is produced, or more generally dry copper of the copper-smelters. Some oxide of copper is retained in the slag, with the oxides of the foreign metals present ; but the greater part of the copper is recovered on subsequently fusing the slag with reducing agents. When copper is melted under salt, it is acted on with the formation of red compound, which has not yet been chemically examined. PROPORTIONS OF FLUXES, &c. Qmntity of Ore taken for Assay : Copper. Grs. Troy. Assay weight marked. Ores tinder ...... 10/ ...... 400 grs. = 100 Ores from 10 to 30% ...... 200 grs. = 50 Ores over ...... 30% ...... 100 grs. = 25 for regains. The fluxes are not usually weighed, but only roughly measured out by means of the flux spoon or ladle, as weighing- would occupy too much time, and would moreover be quite unneces- sary in the case of an expert assayer : however, until experience is gained, it is better to note the weights used. Jn fusion for regulus the proportions of the various fluxes must be so adjusted as to produce a fusible slag with the " ganguo" and oxide of iron resulting from the 404 ASSAYING OF COPPER-ORES BY THE CORNISH METHOD. oxidation of the pyrites, &c., and to ensure that it shall possess the property of easily separating from the regulns ; secondly, to ensure the correct proportions of nitre, sulphur, &c., as may be requisite for the formation of a proper regulus. The following proportions may be taken as an example for the for- mation of the slag : Grains. Grams. Borax f 150 to 200 Glass 150 to 200 Lime 200 Fluorspar 200 These proportions will produce a very good mixture, fusible in itselt and suitable for many ores ; and they may be so varied as to produce a slag of the required nature for any ore. It must be borne in mind that borax is more fusible than glass ; that these two substances tend to produce a vitreous slag, and act as fluxes for lime, oxide of iron, &c. ; that fluor spar increases the fluidity, and aids in imparting to the slag the property of breaking up when cold ; that lime also aids in com- municating this property to the slag, and acts as a flux for silica, quartz, &o. Some assayers omit glass. It is possible to omit any one of the above-mentioned four ingredients, and yet produce a fusible slag. Borax alone will answer as a flux for many ores, but there is no mix- ture which more satisfactorily fulfils the required conditions than the one proposed. The quantity used should be from 600 to 800 grs., which is sufficient to ensure the regulus being well covered with slag as it is poured from the crucible. For the formation of a proper regulus, which should contain about 50 per cent, of copper, we have first to consider if the ore is yellow or grey. In yellow ore, i. e. copper pyrites with or without iron pyrites, &c., more iron and sulphur are present than are necessary to form the desired regulus. In grey ore, i. e. vitreous copper, the addition of iron and sulphur is necessary to produce a proper regulus, and these may be conveniently supplied by adding iron pyrites, or a mixture of oxide of iron and sulphur. TABLK SHOWING THE PROPORTIONS OF IRON AND SULPHUR TO BE OXIDIZED IN ORDER TO OBTAIN FROM YELLOW COPPER-ORE A PROPER REGULUS (WITH ABOUT 50/ o OF COPPER). Formula. Copper. Iron. Sulphur. Total. Copper pyrites or " YeM Cu 2 S+Fe 2 S 3 64 56 64 184 Rpo'ulvis Cu 2 S+FeS 64 28 32 124 To be oxidized FeS 2 28 32 60 Supposing we had pure copper pyrites to deal with, and wished to produce a regulus containing about 50 per cent, of copper, it will be PROPORTION OF FLUX, ETC. 405 seen from the table above that we have to oxidize an amount of iron and sulphur equal to one equivalent of iron-pyrites. This may be done either by partial roasting or " warming," or partly by roasting and partly by the addition of nitre in the subsequent fusion, or simply by fusion with nitre. Either of these three modes may be adopted, and all are resorted to by different assayers, according to the nature of the ore they have to operate on. The first and second methods require considerable experience as to the amount of roasting necessary. The third is the more direct plan: 100 grs. of copper-pyrites will require about 75 grs. of nitre to yield a regulus containing about the correct proportion of copper, and a less amount will be necessary in proportion to the increased amount of earthy matter present. The following results have been actually obtained by experiment : 1. 200 grs. of yellow copper-ore from Cornwall, containing 31-42 per cent, of copper and a small amount of quartz, by fusion with 140 grs. of nitre and 200 grs. of each of the fluxes, borax, glass, fluor-spar, and lime, gave a regulus weighing 116 grs., and containing, by calculation, 54-1 per cent, of copper; it was blue on the exterior, was very tender, and had a red-brown semi-metallic fracture. The slag was dark-green, opaque, semi-vitreous, tender, and free from copper. 2. Another specimen of the same ore, fused with the same propor- tion of nitre and fluxes as in Exp. 1, gave a regulus of similar character weighing 120 grs., and containing, by calculation, 52 36 per cent, of copper. The slag resembled that in the last experiment. The fusion occupied 15 minutes. 100 parts of the ore employed and containing 31 4'2/o f copper require 70/ nitre. Do. pure copper-pyrites do. 34'81/ do. 77 p 8/o do. The regulus produced contained ralher more than 50 per cent, of copper. Iron-pyrites (EeS 2 ), mundic, is often associated or intimately mixed with copper-pyrites, and a portion or the whole of this will very often require to be oxidized. 100 grs. of iron-pyrites, when fused with the above-mentioned fluxes, will require about 180 grs. of nitre to oxidize it, and cause it to pass into the slag. 3. 200 grs. of iron-pyrites from Cornwall, containing 44-68 of iron, and about 2-5 per cent, of silica (quartz), fused with 200 grs. of each of the four fluxes used in Exp. 1 and 400 grains of nitre, gave a black, somewhat glassy, opaque slag free from regulus. The experiment was repeated, and with the same results. 4. 200 grains of the same pyrites, fused with the same proportions of the fluxes used in Exp. 1, and under the same conditions with 335 grs. of nitre, gave a similar slag, and a button of regulus weighing only 2 grs. ; it was fibrous, crystalline, and dark iron-grey in colour. From these experiments -we find that the nitre is rather in excess when it amounts to twice the weight of the iron-pyrites, and that grs. do not quite suffice for 100 grs. of the same kind of pyrites. The preceding results will guide the assayer in adjusting the pro 2 H 466 ASSAYING OF COPPER-ORES BY THE CORNISH METHOD. portion of nitre, but attention must also be paid to the mode of conducting the fusion. The same quantity of nitre, added to an ore, will only yield about the same weight of regulus when the conditions of fusion are the same. Kapidity or slowness of fusion will make a great difference : in the first case the regulus sinks before the nitre has time to complete its action, and in the second case this does not occur. With proper precautions in two experiments with the same ore, the reguluses should only differ a few grains in weight. TABLE SHOWING THE PROPORTIONS OF IRON AND SULPHUR NECESSARY TO BE ADDED IN ORDER TO OBTAIN A PROPER REGULUS FROM VlTREOUS CoPPER-OllE. Formula. Copper. Iron. Sulphur. Total Vitreous or "grey" cop-) Cu 2 S 64 . . 16 80 To be added FeS t . 28 16 44 Regulus . . Cu 2 S + FeS 64 28 32 124 . Vitreous copper-ore, if fused direct with fluxes, would yield a " too fine " regulus, and copper would be lost in the slag : hence, in order to obtain a proper regulus from this kind of ore, iron and sulphur must be added in the form of iron-pyrites (the excess of sulphur in the pyrites will be driven off in the fusion), or a mixture of oxide of iron and sulphur, or sulphide of iron may be substituted. 100 grs. of di sul- phide of copper will require about 60 grs. of sulphide of iron (FeS), about 82 of iron-pyrites, or about 55 of haematite and excess of sulphur, to yield a regulus containing about 50 per cent, of copper. As many ores contain oxide of iron in the matrix, to such sulphur only need be added. Purple copper-ore generally contains sufficient copper to yield a proper regulus, containing from 50 to 60 per cent, of copper, by direct fusion with fluxes, and without the addition of nitre or sulphur. In the adjustment of the nitre to yellow, and of the sulphur to grey, ores, it is customary with some assay ers to employ them in nearly the following order : 1. 3 dwts. of nitre. 0/3 dwts. of nitre. 1 1 do. tartar. Q ( 3 dwts. of nitre. \2to2Jdo.tartar. 3 dwts. of nitre. 2 do. tartar. 1 do. sulphur. (3 dwts. of nitre. 5. J2 do. tartar. ( 2 do. sulphur. 13 dwts. of nitre. 6. m do. tartar. do. sulphur. 200 grs. (50 grs. of the copper-assay weight) of the ore are fused in the usual way with 3 dwts. of nitre. If the regulus is too coarse, 4, 5, and so on up to 9 dwts. of nitre, should be tried. On the other hand, if 3 dwts. of nitre prove too much, one of the mixtures ASSAY CLASSIFICATION OF ORES, &c. 467 2, 3, 4, 5, 6, should be added. It is very seldom these additions have to be resorted to in the order given, as experienced assayers are gene- rally able at once to determine upon the right proportions. It is obvious that in some of these mixtures the ingredients must mutually counteract each other, with respect to the objects for which they are intended. It is alto obvious that a true regulus could not be obtained from a grey ore unless iron is present in some form. As a general rule, sulphur is never added to yellow ores. Fusion, for coarse copper. The reagents generally used for this part of the process are mixtures of tartar and nitre in such proportions that the tartar is always in excess above that required to form carbonate of potash with the nitre. A small quantity of borax is likewise added by some assayers ; others use glass instead, others omit both, and others pursue the common practice of adding salt. In the Metallurgical Laboratory we have been accustomed to use carbonate of soda width tartar or charcoal, or, better still, with a mixture of both. The propor- tions must necessarily vary with the amount of calcined regulus, but the following will serve as a guide : Grains. Grai 1. Carbonate of soda 50 to 150. Tartar 50 to 150. 3 Borax . ,...20 to 30. (Tartar 50 to 200. 5. {Nitre 10 to 50. Borax 20 to 80. In the fusion direct for coarse copper, as in assaying carbonates, oxides, silicates, &c., it is better, if the ore does not contain lime, oxide of iron, &c., to add from 20 to 50 grs. of either of these substances, with a view to keep the slag free from copper, more especially when silica is present. The use of too much borax should be avoided, as it tends to retain the copper in the slag. fusion of slags from refining or from fusion, for coarse copper. They may be fused with from 50 to 100 grs. of tartar, or less tartar with the addi- tion of from 5 to 10 grs. of charcoal ; or the following mixture may be employed : Carbonate of soda about 50 grains. Tartar 50 to 100 do. Charcoal 5 to 10 do. ASSAY CLASSIFICATION OF ORES AND CUPRIFEROUS PRODUCTS. I. Ores, &c., containing over 30 per cent, of copper, which may be fused direct for coarse copper, or, after previous roasting, without fusion for regulus : a. Native copper, bar copper, &c., require refining only. 6. Black and red oxides, blue and green carbonates, silicates, oxy- chlorides, require fusion for coarse copper and refining. c. Various sulphides comparatively pure, or containing arsenic or antimony, rich regulus, &c., require roasting " sweet," fusion for coarse copper, and refining. 3 Or charcoal 15 to 25 grains. 2 H 2 486 PRACTICAL DIRECTIONS FOR CONDUCTING THE PROCESS. II. Ores, &c., containing less than about 30 per cent, of copper, which require a preliminary fusion for regains to obtain the copper in a more concentrated form, and to free it from earthy and other im- purities : a. Ores which require partial calcination, or the addition of nitre, or both, to remove the excess of sulphur, and carry a portion of iron and other foreign metals into the slag : Yellow copper, pyrites, &c., with or without iron-pyrites, blende, galena, &c. b. Ores which require the addition of sulphur, or of sulphur and iron, to yield a proper regains : 1. Vitreous or "grey" copper-ore. 2. Oxides, poor carbonates, &c. ; slags. 3. Poor antimonial and arsenical copper-ores, after having been previously roasted " sweet." c. Ores which yield a proper regulus by direct fusion : Purple copper-ore, mixed sulphides and oxides, &c. To prevent unnecessary repetition, the assay classification will not be followed in the order above stated, but a general description of the method through all stages of the process will be given, leaving it to the judgment of the operator to apply the particular part applicable to the kind of ore submitted to assay. PRACTICAL DIRECTIONS FOR CONDUCTING THE PROCESS. 1. Fusion for regulus. The ore, "raw" or calcined, is intimately mixed with the fluxes, borax, glass, lime; fluor-spar, nitre, iron-pyrites, or a substitute for the latter, in accordance with the directions already given. The mixture is generally made in No. 1 crucible by means of a small steel spatula, and may be covered with a thin layer of borax or of the mixed fluxes. When a preliminary calcination is resorted to, the crucible in which this operation is performed should be reserved for the subsequent fusion; or should a scorifier be employed, it should be reserved for the subsequent calcination of the regulus. In calcination the ore should be exposed to a dull red heat until the blue flame of sulphur ceases, and the surface appears reddish-brown, which usually occurs in about ten minutes. The crucible or scorifier is then taken out and allowed to cool. If much iron-pyrites, blende, &c., is present, the calcination must be prolonged. The crucibles, when placed in the furnace, should be packed well round with coke, up to, but not above, them, so that they may be as uniformly heated as possible. Covers are generally dispensed with; but, if used, they should be placed slightly on one side, so as to leave space for the free escape of the gases evolved, otherwise the fused matter is liable to run over the sides of the crucible. A cover may be useful in preventing particles of fuel from being projected into the crucible. The crucibles being arranged, and the furnace-top closed or left slightly open at the back, the temperature is gradually raised, so that the fusion may be com- pleted in from fifteen to twenty minutes. During the first part of the FUSION FOR REGULUS. 469 fusion, effervescence occurs from the escape of carbonic, sulphurous, and nitrous acids, and the vapour of water ; but it gradually dimi- nishes, until, at the close of the fusion, the surface of the slag is per- fectly tranquil. The crucible is now quickly removed from the fur- nace, and made to receive a rotary motion in order to wash down any particles adhering to the sides, immediately after which the melted contents are poured into an iron mould (fig. 120, H). The interior of the crucible must be examined while hot, to ascertain that no globules of regulus remain attached to the internal surface. As sot>n as the slag has solidified, or is " set," it is seized with the forceps (fig. 120, B), plunged two or three times into cojd water, and then left to cool. This immersion in water cracks up the slag, and causes the regulus to be more easily detached when cold. The use of the hammer should be avoided, as the regulus is very fragile ; and it is moreover unnecessary, if proper precautions have been observed. If any slag adheres to the regulus, it is generally on the upper surface, from which it may be detached by gentle taps with a small hammer, or by means of a spatula-blade. After the regulus is separated, the slag should be examined to see that there are no globules of regulus re- maining in it ; and should any be present, they will generally be found at or near the external surface. The presence of regulus in the slag most frequently arises from the latter not having been sufficiently fluid, or it may be caused by the projection of particles of fuel into the crucible during fusion. When all the conditions are right, the slag will be free from all traces of copper-regulus. To save time and trouble the slag may be poured into one cavity of the mould and the regulus into another ; but considerable practice is required to enable the operator to do this successfully. If the regulus is not easily detached, or if there is any doubt respecting the cleanness of the slag, the latter may be broken in pieces and re-melted with a little sulphur or iron-pyrites, when the small button of regulus which may be thus formed must be added to the regulus previously obtained. The slag is generally opaque, or more or less glassy, and varies in colour from white, grey, pale-green, bottle-green, to black ; the lighter- coloured slags are generally produced from "grey" ores, and the darker-coloured ones from "-yellow" ores, &c., the dark colour being due to oxide of iron. Occasionally the oxide of iron imparts a bluish colour to the slag ; but it differs in tint from that sometimes produced by protoxide of copper. The colour of .the slag should be uniform throughout, and not variegated, mottled, or streaked, as this generally arises from imperfect fusion. It should be free from every trace of red colour, which would indicate the presence of copper as dioxide. A proper slag breaks up easily, or crumbles into small angular frag- ments under slight pressure. This property of the slag is very advan- tageous in causing the easy separation of the fragile regulus, and is due tQ the use of fluor-spar or lime. If the slag is too glassy, it has a tendency to adhere to the regulus, an evil which may be counter- acted on a subsequent fusion by increasing the amount of lime or fluor- spar, or both, as the case may require. If the slag flows thick from 470 PRACTICAL DIRECTIONS FOR CONDUCTING THE PROCESS. the crucible and has a stony aspect when cold, or the crucible is much corroded or cut into at the line of junction with the upper surface of the slag during fusion, there will generally be found to have been excess of lime and sometimes excess of fluor-spar. It is a common practice to use a layer of salt as a cover for the mixture during fusion : salt acts as a lubricator to the sides of the crucible, and allows the gases produced in the reactions to escape more freely from the surface of the slag during fusion. There are, however, some objections to its use. The fumes which it produces and its extreme fusibility prevent the assayer from observing whether the slag underneath has the proper degree of liquidity. When salt is used, on pouring out the fused contents of the crucible, it will be found on the top of the slag, and should be crystalline and white or greyish-white in colour ; but if it has a pink or red colour, it is a sign that copper is present in sensible quantity. A good or proper regulus should be reddish-brown in colour, more or less convex on the upper surface, very tender, full of cracks or fissures, easily broken up into fragments and reduced to powder. The reddish-brown colour is due to the surface being studded with fine particles of metallic copper, or moss-copper in miniature. It is also obtained of a blue colour, when it corresponds to the " blue-metal" of the copper-smelters. It should contain about 50 / of copper ; but the produce may vary from 40 to 60/ . It is not desirable to obtain it richer in copper. A good regulus is more easily calcined than one too eoarse or too fine. A coarse regulus is dull and coarse-grained on the exterior, more or less flat, and sometimes vesicular on the upper surface, is compara- tively hard, and varies in colour from iron-grey to brass-yellow, according to the amount of iron and copper present; but it readily acquires a red, purple, or blue tarnish ; its proper colour can best be seen by inspecting a clean, fresh fracture when quite cold ; the fracture is crystalline, fibrous, or coarsely granular. It is not so easily reduced to powder- as a proper regulus, and the powder is lighter in colour ; it corresponds to the coarse-metal of the copper-smelters. There is little fear of the slags retaining copper when a ccarsB regulus is produced ; but the calcination is not quite so readily effected without the forma- tion of sulphates, and the time of calcination is prolonged in a degree proportionate to the amount of iron and sulphur present above what may be necessary to form the proper regulus. It may be desirable occasion- ally, when operating on a very poor ore or slags to obtain a coarse regu- lus, to increase the bulk, and so facilitate the collection of the copper. If the regulus is too fine, it is more or less round, and the external surface is very smooth and bright, semi-metallic in lustre, and nearly black in colour ; its fracture is glassy or very compact in structure, and dark bluish-grey in colour ; it is comparatively hard when broken : the powder is nearly black, and has a bluish tinge ; it is more difficult to calcine than the proper regulus, owing to its greater tendency to fuse. Moreover, when the regulus is too fine, copper is liable to be retained in the accompanying slag ; and although it is possible to pro- x* l/7 /y>' N \ / CALCINATION OF THE REGULTlL. . . -- - -n^T m T duce a comparatively fine rcgulus with a slap; tree iVmn copper, yet it is not safe or desirable to do so and run the risk of losing copper. Occasionally the regulns will rapidly crumble into fine powder, which may arise from imperfect fusion, from its containing excess of sulphide of iron, or from leaving it too long in contact with water. The presence of other metals, such as lead, zinc, antimony, arsenic, &c., alters the character of the regulus. The following are some analyses of the three kinds of regulus, good, coarse, and fine, obtained from an experienced Cornish copper assayer, and analysed by Mr. R. Daintree, while a student in the Metal- lurgical Laboratory 4 : " Coarse." " Good." " Fine." Copper By analysis. By theory. 27-74 20 '90 By analysis. By theory. 58 -Gl 58-20 By analysis. By theory. 72 -70 73 04 Iron 35 '05 35 70 16*50 17-15 5-42 5 -38 Sulphur 37-10 37-40 24-70 24-40 21-80 21-51 99-89 100-0 99-87 100-00 99-92 100 00 Formula 2Cu 2 S + 3Fe 2 S :j 3Cu ? S + 2FcS GCu 2 S + Ft-S or Cu' 2 S+ IpVS 3 Cu 2 S + FeS Cu 2 S + 'FeS Each of the preceding results is the mean of three determinations of each ingredient. The copper was determined by standard solution of cyanide of potassium ; the iron by standard solution of bichromate of potash ; and the sulphur as sulphate of baryta. On comparing these analyses with those of the products of copper-smelting, they will be found to correspond in composition to coarse, lias, and white metal respectively. 2. Calcination of the regulus. It is first reduced to fine powder in an iron or bronze mortar, fig. 120, D, and, after the removal of the powdered regulus, a little coke-dust or anthracite powder is rubbed down in the mortar and afterwards added to the regulus. Thus finely divided, it is introduced into a No. 2 crucible or scorifier. If a scorifier is used, its internal surface should be previously rubbed over with the unctuous variety of haematite powder, whereby the removal of the calcined regulus is facilitated and loss by adhesion prevented. The scorifier, containing the regulus evenly spread, is placed in a muffle, which may be filled with as many as it will conve- niently hold. The calcination is conducted at a gradually-increasing temperature, the muffle door being left open all the while. The regulus should be stirred occasionally during the process with a roasting spatula on temporary removal from the muffle, or by means of the stirring-rod within the muffle. Calcination takes place more readily in a muffle, and does not require so much attention as when it is effected in a 4 Mr. Daintree was an excellent manipulator. He is now in Australia, where he has long resided. 472 PRACTICAL DIRECTIONS FOR CONDUCTING THE PROCESS. crucible in the furnace, very little stirring being necessary. If it is not pulverulent on completion of the process and shows any sign of clotting, or if there is any other reason for supposing that it has not been perfectly calcined, it must be rubbed down with about 20 grains of anthracite powder, and re-calcined for about 10 minutes. When the calcination is ended, the scorifier is taken out of the muffle ; and, when cool, the contents are transferred to crucibles ready for fusion. The scorifiers may be 1 used over again for the same purpose. The common practice is to eifect the calcination in a crucible, selecting the smaller or middle size according to the quantity of regulus. Several calcinations may be carried on in one furnace at the same time. The furnace is filled with fresh fuel to within a short distance of the top, and the crucibles are then introduced and placed firmly with a slight incli- nation forwards, so that a current of air may pass over the surface of the regulus powder. -A stirring-rod, fig. 120, I, is put into each crucible, with the broad end downwards, the upper end being held between the fingers of the left-hand, or allowed to lean against some support to faci- litate the manipulation. The calcination should commence at -a dull red-heat, which should be gradually, but as quickly as is compatible with a proper calcination, increased to a strong red-heat. Constant stirring is necessary for the first 15 or 20 minutes, but afterwards it need only be occasional. The time required for complete calcination is about 30 minutes, and sometimes three-quarters of an hour. If the tem- perature is too low at first, there is a tendency to form basic sulphate of copper, which gives a greyish tinge to the surface of the powder and prolongs the calcination. Too high a temperature occasions clot- ting, and so prevents oxidation. A\ hen clotting occurs, the regulus must be removed and triturated with a little coke-dust or anthracite, and afterwards put back, into the crucible, and the calcination pro- ceeded with ; but in such cases it is best to throw it away and begin afresh. " At the right temperature the powder will remain in a sandy state and be very easily kept glowing at first. If the powder shows the least sign of softening or of clotting, the temperature should be lowered immediately by raising the crucible for a few seconds and placing under it some pieces of fresh coke. The calcination is com- pleted when it is " sweet," i.e. when the odour of sulphurous acid can no longer be detected. The crucible and rod are now removed from the furnace and allowed to cool. 'When cold the roasted regulus has a greyish-black colour. If any portion adheres to the rod, it is filed or scraped off and returned to the crucible. The same crucibles are used for the subsequent fusion. 3. ^ Fusion for coarse copper. The calcined regulus is mixed in the crucible with the proper fluxes, and then put into the furnace, the fire being "hot." Fusion occurs in about 10 or 15 minutes. The fusion should be rapidly effected, and as soon a,s effervescence ceases the melted contents are poured into a mould ; when the slag is set, the button is dipped into cold water, or the whole is allowed to cool. The slag should be black, glassy, and tender ; if it contains copper it presents streaks or patches of a red colour ; this is best observed by FUSION FOR COARSE COPPER. 473 examining the exterior of the slag or the interior glaze of the crucible used in the fusion. This red colour may be due to deficiency of re- ducing agent, the presence of too much borax, or to too low or too high a temperature during fusion ; or to keeping the copper top long in a state of fusion, when the metal has a tendency to become oxidized and pass into the slag. The interior of the crucible and the slag should be examined to see that no prills or small globules of metal are present : none will be found after a successful fusion. The slags are retained for subsequent fusion, which, though commonly practised, is often unnecessary. 1 The button of coarse copper should be easily detached from the slag and have a clean surface ; if it is black or covered with regulus, the previous calcination has been imperfect, the coating of regulus being due to sulphide left unoxidized or to sulphates formed in the roasting having been again reduced to sulphides, \\hen this evil occurs, the whole should be rejected and the assay begun again. The coarse copper is generally convex on the upper surface and somewhat rough towards the centre ; after nicking it with a chisel it breaks short in the vice when struck, and the fracture is fine-grained, dull, and greyish or orange in tint ; occasionally, when produced from pure ores, it has the properties of fine copper. Its character necessarily varies according to the impurities present. Peculiarities which have been ob- served in connexion with the presence of various metals are as follow : Iron. It is often present in small quantities under 1 per cent., and is not objectionable, as it tends to render the accompanying slag free from copper, and it passes out easily on refining. When it is present in larger quantity the copper is harder and paler in colour, and the convexity of the upper surface of the button is increased : the latter is not easily fractured, and is more difficult to fuse than copper, and occasionally subsides as an irregularly-shaped, imperfectly -fused mass. The conditions which favour the passage of iron into the copper are excess of reducing agent and a high temperature. The results of the following experiment will illustrate the mode of formation of this ferriferous copper. An intimate mixture consisting of Grains. Protoxide of copper (CuO) 200 Haematite 100 Charcoal 100 Carbonate of soda 300 Borax 30 was exposed in a covered Cornish crucible to a high temperature for more than half an hour. A black, glassy slag was produced, pre- senting on its external surface small patches of a red colour ; and a well-melted, clean, bright, metallic button was formed of a pale copper-colour, and weighing 210 grains : it was composed of Copper 71-08 Iron . ,28-23 99-31 Tin. It hardens the copper and renders the exterior browner in 474 PRACTICAL DIRECTIONS FOR CONDUCTING THE PROCESS. colour ; even when present to the extent only of 5 per cent., the frac- ture of the copper is not much altered, though it is somewhat paler in colour ; it is difficult to recognise the presence of tin in small quantity from the physical characters of the copper, except that of increased hardness. Lead. When lead is present even in considerable quantity the button on the exterior will retain its copper colour and be soft and malleable ; but it has a characteristic dull, bluish-grey, granular fracture ; the button is often accompanied by a regulus when lead is present, the sulphate of lead which is formed in the process of calcination becoming reduced during the fusion. Lead is the great curse of the copper assayer. Antimony. It renders the copper hard and brittle, and communicates to the fracture a dull yellowish-grey colour ; when it exceeds 10 per cent, the fracture has an iron-grey colour, and the copper is very brittle. Zinc. It is not easily recognized in small quantity, as in this case its presence does not much alter the character of the copper ; it tends to communicate a yellow colour and fibrous crystalline fracture' to the copper. Arsenic. It renders the copper harder, brittle, and greyer in colour. Nickel. It is occasionally present, when it renders the copper harder and paler in colour. Silver. It often occurs in copper from true grey copper-ores ; but when present in small quantity it cannot be detected by the physical characters of the copper. Gold. It is rarely present in appreciable quantity ; but occasionally it is found accompanying silver. 4. Refining. The crucible which has been used for the fusion of coarse copper is usually employed for refining. The process is con- ducted as follows : The crucible is placed well down in the fire, with the centre directly under the line of junction of the two bricks used to close the furnace ; and, when it has become bright red-hot, the button of coarse copper is dropped quickly into it, the furnace is closed and watched attentively through the crevices between the bricks above mentioned. Fusion soon occurs, and at first slight evo- lution of gas takes place from the dull surface of the metal ; after a short time the coating of oxides rapidly clears off and the surface becomes bright and resplendent round the edges ; and from the centre bright and bluish-green light is reflected, producing the appearance technically termed the "eye" or "star." Some refining flux or refining flux and salt, previously placed ready in the copper scoop, fig. 120, F, is now quickly projected on the top of the fused button, and the furnace is closed ; after the lapse of about 2 minutes the cru- cible is removed and its contents poured out into the mould. The button should come out covered with slag ; and when set it is seized with the forceps and its under surface made to touch the water in the bowl, by which means the slag is readily detached. The whole process of refining should occupy from 5 to 10 minutes. The slag is grey, flesh-coloured, pink, or pale-red, and separates CLEANING OF SLAGS OBTAINED IN REFINING. 475 easily from the button ; occasionally it is blue, bluish-green, or green, from the presence of protoxide of copper ; or it may be deep-red from the presence of dioxide of copper ; and this arises from the copper having been too much oxidized before the refining mixture was pro- jected on it, or from the flux itself being too sharp, i. e. too strongly oxidizing. The button, when " fine," is more or less flat, and its upper surface is coated with a thin orange-red film ; the copper is soft, malleable, tough, and breaks with difficulty, and its fracture is finely silky. More com- monly it is dryer and has an indentation or pit on the upper surface, and when broken across the fracture is granular and has a purplish- red colour. If not sufficiently refined, the copper has more the appear- ance of coarse copper externally and when fractured : in this case the refining process should be repeated. When over-refined, the copper is coated with a thick, red, rose, or crimson-red film, and it is then said to be burnt ; the slag is red, and adheres to the button, which is brittle, and when broken across presents a deep-red somewhat porous fracture. Salt is generally used in refining : it serves to dilute and check the too rapid action of the refining flux, and may also aid in the separation of the antimony, lead, &c. ; but the refining flux may be used alone. When the flux is too " sharp," or not " sharp " enough, it may be modified by adding a small quantity of tartar or nitre respectively, before projecting it into the crucible. When the coarse copper is very impure, from the presence of iron, tin, lead, antimony, zinc, &c., a small quantity of borax may be added with advantage when the button is first put into the crucible ; it dissolves the scum, and the button " cleans " more readily. Occasionally it may be necessary to lift the crucible up, and resort to gentle shaking or tapping to break the crust on the surface of the button. Some assayers subject the coarse copper button, before refining it, to one or more meltings with a refining flux ; the process is called " washing." Lead is difficult to remove in refining without some copper passing into the slag. Antimony is also difficult of separation, and a small quantity of lead is generally added to aid in its removal ; but it may be separated without lead by the use of a refining flux a little more oxidizing. Tm retards the appearance of the " eye " in refining. Silver, which rarely exceeds in amount 1 per cent, of the ore, remains in the copper after refining. The whole, or nearly so, of the silver originally present in the ore becomes thus concentrated in the copper. If present in sufficient quantity materially to affect the per-centage of copper, the button should be assayed for silver by cupellation, and the weight of the silver found deducted from the total weight of the original button of copper. Cleaning of slags obtained in refining. The slag from the refining pro- cess always contains copper. It should be roughly powdered, mixed with some tartar, charcoal, &c., fused, and poured out into the mould. When cold, the product is broken down and carefully examined for small prills or shots of copper. The prills obtained generally weigh 476 SPECIAL MODES OF ASSAYING. from 1 to 5 grs., according to the quantity and purity of the button subjected to refining, and the skill of the operator ; they are added to the button of fine copper and weighed; they are always refined by copper assayers, but their weight is often so small that the loss which they undergo in the process is practically unimportant. The fusion requires from 10 to 15 minutes; the slag should be black. If it is considered necessary to clean the slag from the coarse copper fusion, it may be intermixed with the refining slag, and one fusion will suffice for both. The loss of copper in the Cornish method of assay. If any loss occurs in the preliminary roasting or subsequent fusion for regulus, it is from fine particles being carried off in calcination, or from small entangled shots of regulus being retained in the slag. The chief loss of copper occurs in the fusion for coarse copper and in the subsequent refining ; for if the slags be tested chemically, they will be found to contain small quantities of copper, the proportion varying according to the richness of the ore, and the amount and nature of the foreign metals present, especially lead. The discrepancies between the results of different assayers arise chiefly from the manner in which the pro- cess of refining is carried on. It is the most difficult part of the process, requiring much skill and judgment for its accurate per- formance ; and through inattention or want of experience, a large amount of copper may be lost in this final operation. The actual loss of copper decreases in proportion to the richness of the ore ; but the difference between two assays one by the dry and the other by the wet method increases in proportion to the richness of the ore. This statement must be taken in a very general sense, as the presence of other metals, the nature of the ore, &c., will modify the result of the dry assay considerably. 5 SPECIAL MODES OF ASSAYING PARTICULAR KINDS OF ORES AND- OTHER CUPRIFEROUS PRODUCTS. Native copper. Simple fusion with a small quantity of tartar, or char- coal and carbonate of soda, is often sufficient to yield a button of fine copper ; if not, the button must be subjected to the process of refining. The slags in either case should be cleaned as usual by remelting with a reducing flux. If it is intermixed with oxide, carbonate, &c., the pro- portions of reducing agents should be increased. Bar copper generally requires melting and refining. It is well to glaze the interior of the crucible before use, which is readily done by melting in it a little borax. Oxychloride of copper. This should be fused direct with carbonate of 5 The German meiliod of assaying cop- per-ores I shall not enter upon a descrip- tion of this method in the present volume. The copper, if not present in the state of oxide in the ore, is converted into oxide by roasting sweet. The roasted product is melted with borax and black-flux, whereby the copper is reduced in a state of greater or less purity. The reduced copper is refined by scorification with lead in a muffle; and certain allowances are made to compensate for the loss of copper in the slag. This process of assaying is sensibly inaccurate. FOREIGN METALS OCCURRING IN COPPER-ORES. 477 soda and a reducing agent, such as charcoal, tartar, &c. No object would be gained by roasting. The coarse copper is refined if ne- cessary. Regulus. Kich regulus. 1. It should be reduced to powder, and cal- cined in a scorifier or crucible, with the usual precautions. To com- plete the calcination, it should be afterwards rubbed down with 20 or 30 grs. of coke-dust or anthracite-powder, and re-calcined. If not thus treated, it is liable to retain sulphates which would occasion loss, and yield a regulus, on fusion for coarse copper. 2. Fused for coarse copper. 3. Kefined. As "British regulus" is generally poor in copper, a portion of the iron and sulphur should be first removed by partial calcination, or, as it is technically termed, " warming," or by the addition of nitre ; after which it must be fused for regulus, and the process continued in the usual manner. Slags. They are generally silicates of protoxide of iron, &c., con- taining copper. The best method of cleaning them is to fuse direct, with the addition of iron-pyrites, and so collect the copper in a some- what coarse regulus. FOREIGN METALS OCCURRING IN COPPER-ORES, &c. Tin. It occurs in copper-ores chiefly as oxide of tin (tin-stone), or as sulphide of tin and copper (bell-metal ore). When present as oxide, the greater part will be carried into the slag by fusion with nitre, &c. ; but when it exists as sulphide, or when other sulphides are present in large proportion, the ore should be subjected to a preliminary calcina- tion before fusion for regulus, so as to obtain a richer regulus than would otherwise be produced, as in this case the tin tends to pass into the slag as oxide. Iron. It occurs in the state of iron-pyrites or mundic, arsenical iron-pyrites, mispickel, oxide, &c. By calcination or the use of nitre there is no difficulty in causing any quantity of iron to pass into the slag in the state of oxide. Zinc. It is generally present as sulphide in the form of blende or Hack jack. Calcination or the use of nitre are the only means resorted to to oxidize the zinc and carry it into the slag, care being taken to ensure the presence of sufficient iron and sulphur to form the proper regulus. Lead. It is generally present as sulphide of lead or galena, but occurs in true grey copper-ores. Its presence occasions much trouble to the copper-assayer, because, even by prolonged calcination, it cannot be made to pass into the slag, and will be found in the regulus and coarse copper. The following is the best mode of treating an ore containing lead : If iron-pyrites is 'present, the preliminary roasting or addition of nitre may be omitted, and the ore fused at once for regulus ; and when the latter is in the state of fusion a few grains of iron-wire should be dropped into it, or the end of a thin rod of iron may be kept immersed in it for about ten minutes. The rod should then be withdrawn, and the regulus poured out and left to cool. On detaching the regulus (which should be coarse) when cold, the lead 478 METHODS OF ESTIMATING COPPER BY WET ASSAY. will be found as a button at the bottom, and, if in small quantity, diffused as globules. It can be separated from the regulus by bruising in a mortar, when the flattened particles of the metal may be picked out. The regulus may afterwards be treated as a yellow copper-ore in the usual way. The lead will contain some copper, which can be best estimated by the wet way. If the ore does not contain iron- pyrites, it should be added, as in the case of grey ores, c. There will be some difficulty in applying this method at first, in ensuring the right conditions suitable for each ore ; but it is the most effectual with which we are acquainted, and, if properly conducted, no further trouble will be experienced from the presence of lead in roasting the regulus, fusion for coarse copper, &c. Antimony. It is generally present in some varieties of fahlerz, or true grey copper-ore. In the case of rich ores, the best course is to cal- cine in a scorifier or crucible, taking great care, during the process, to prevent clotting by keeping the temperature low, as the ore is very liable to fuse, by which means the antimony is oxidized, and in a great degree passes off in the state of vapour as antimonious acid (SbO s ). The calcined ore is rubbed down with coke-dust or anthra- cite-powder, and recalcined. The calcined ore is subsequently fused for coarse copper, with the addition of more carbonate of soda, and less borax than usual. If the ore is poor after calcination, it must be fused for regulus with the addition of iron and sulphur, &c., in the usual manner. Arsenic. If present in grey ores, they should be treated in the same way as for antimony. If it exists as mispickel, arsenical iron-pyrites, &c., the ore may be partially roasted, or treated with nitre on fusion for regulus. When present in the regulus, the latter, after calcination, should be mixed with some coke-dust or coal-powder, and re-calcined in order to reduce the basic arseniates always formed during the roasting process, and so to expel the arsenic as completely as prac- ticable. METHODS OF ESTIMATING COPPER BY WET ASSAY. There are various processes described for the determination of copper by wet assay, i. e. by the aid of liquid reagents, but few of them are applicable to the estimation of copper in copper-ores, as they would occupy too much time, or are unsuitable for practical purposes, Wet methods yield more accurate results than can be obtained in the dry way, and it matters not which of the various methods is employed, or in what part of the world the assay is made : provided the sample is correctly taken, the results will closely agree, and this certainly cannot be always affirmed of the dry methods of assay. In this work those wet processes only will be noticed which by long experience in the Metallurgical Laboratory have been found best for practical purposes. Sampling. After the sample has been prepared, as for dry assay, a small quantity of the ore, say aboiit 100 grs., should be taken out, and reduced to a finer powder than for dry assay. BY CYANIDE OF POTASSIUM - APPARATUS. 479 Quantities to be taken for assay : Ores, &c., containing over 30 per cent, of copper 10 grs. Ores containing from 4 to 30 , , , , 20 , , Ores under 4 , , , , 50 , , The balance used for weighing should be capable of carrying 500 grs., and of turning with T -vth of a grain. A chemical balance is the one usually employed. It should have a light counterpoised silver, platinum, or porcelain dish attached to it, in which the ore should be weighed out. BY CYANIDE OF POTASSIUM. This process was first made known by Mr. Henry Parkes in 1851, as applicable to the assaying of copper- ores. It has since been extensively used in South America (the ores consisting chiefly of oxides, carbonates, silicates, and oxychlo- rides), where it has almost superseded the method of precipitation by iron. It has also been applied in this country for the estima- tion of copper in slags, &c. ; but as the copper-ores are bought and sold in England according to the return by dry assay, it has been somewhat limited in its application. Of all the wet processes by means of standard solutions, at present known, it is certainly the best adapted and most direct for assaying copper-ores, as the pre- sence of iron does not interfere with the process, and in the greater number of instances filtration is unnecessary. Like all other me- thods, it requires experience to be skilfully conducted ; but that it is capable of yielding correct results, within 0-1 or 0-2 per cent., has been fully proved by some thousands of assays, made by various persons, which have come under our notice during the course of several years, results very closely approximating having been ob- tained by different manipulators with the same samples of ores. Some difficulties will always be met with in taking up any new process, but no process should be condemned by an operator because he fails to succeed in his first trials. When a solution of cyanide of potassium is slowly added to a blue ammoniacal solution of copper, the latter gradually loses its colour, and finally becomes quite colourless ; upon this reaction, the estimation of copper by cyanide of potassium depends. By ascertaining, by direct experiment, the amount of cyanide of potassium solution required to discharge the colour in an ammoniacal solution containing a given weight of copper, it is easy, by a comparative experiment, to determine the amount of copper in a given weight of ore. This process will be considered under the following heads : Appa- ratus, Standard Solution, and Mode of conducting the Assay. Apparatus. It will be unnecessary, under this head, to enter into any detail in reference to the ordinary apparatus in common use in chemical laboratories, such as flasks, beaker-glasses, funnels, &c. The number of each will be regulated by the number of assays it is required to conduct at one time. The only special piece of apparatus necessary for the assay by standard solution is a burette. That usually employed in this country is known as the English burette or alkalimeter. It 480 METHODS OF ESTIMATING COPPER BY WET ASSAY. should be about 18 inches high, J an inch in internal bore, ths of an inch in external diameter, of 1000 grs. capacity, and graduated to 200 divisions. When a number of assays have to be done at the same time, those known as Mohr's burettes are the best. Fig. 121 represents one of 1 000-grain capacity, and graduated to 200 divisions, fitted up on a stand to facilitate manipulation. Into the upper part is fitted a cork, through which passes a piece of quill glass tubing, which may be used to con- nect the burette, by means of a piece of vulcanized India-rubber tubing, with the bottle or jar containing the standard solution, so that it may be easily filled. To the lower part of the burette a piece of glass tubing drawn out to a point is connected by means of vulcanized India-rubber tubing, and the flow of the standard solution can be regulated with the greatest nicety by means of a pressure-screw (see fig. 121, 6) operating on this connecting piece of India- rubber tubing. The burette can be raised or lowered by means of the arms and screws attached to the upright pillar of the stand. The number of "such burettes must be regulated by the requirements of the assayer. Any number of assays can be con- ducted at one time by one person by the aid of these burettes, of which several may be fitted up on one stand ; or they can be attached to the table by means of bracket-arms. Standard solution. 2000 grs. of cyanide of potas- sium are to be dissolved in 4 pints (35000 grs.) of water. This salt is deliquescent, and should not be left exposed to the air. It is very soluble in water, and the solution is ready for immediate use. The above quantity will produce a standard solution, of which 1000 grs. will be equal to about 10 grs. of copper. That known as gold or photographic cyanide is the best, as the solution prepared from it may be kept for a great length of time without becoming dis- coloured. If common cyanide of potassium is used, about 2880 (6 oz. Troy) will be required for 4 pints of water. The solution prepared from it can only be kept for a few days without becoming discoloured and muddy; but when the solution is rapidly con- Fig. 121. . sumed, it answers very well for assay purposes. An objection raised against the use of cyanide of potassium for standard solution is its liability to decompose. W'hen numerous assays have to be made, this circumstance is not of any mate- rial consequence, as decomposition and consequent alteration of the standard take place very slowly. For instance, a large quantity of solution, prepared and used from time to time in assaying, had the following standards : Oct. 26th, 1858 .... 1000 grs. = 10-06 grains of copper. July 25th, 1859 .... 1000,, = 9-84 March 20th, 1860 .. 1000,, = 9-45 June, 1860 1000,, = 8'92 STANDARD SOLUTION METHOD OF ASSAY. 481 From this it will be seen that it requires a considerable time to make any decided alteration in the strength of the solution. For practical purposes, the standard will not require checking more than once a week. The solution should be kept in bottles of green glass ; lead glass bottles should be avoided ; even the green glass bottles are slowly acted on by it, the glass being decomposed, and a thin scaly deposit formed. The solution of cyanide of potassium is standardized as follows : Three pieces of pure electrotype copper, previously cleaned by means of hydrochloric or dilute nitric acid, washed and dried, are accurately weighed. The weight of the pieces may vary from 5 to 10 grs. each. Each piece is dissolved in a pint flask by dilute nitric acid, and the solution is boiled to expel all nitrous acid fumes ; it is diluted with water to about half-a-pint, and treated with ammonia in excess, when it will become deep blue. The burette is filled to the uppermost division, with the solution of cyanide of potassium. When the solu- tion of copper is quite cold, the flask containing it is placed under the burette, and the solution of cyanide is allowed to run in in small quantity at a time, until the blue colour is almost discharged, and is replaced by a faint lilac tint. The number of divisions is then read off on the burette of solution of cyanide which has been required for decolorization. The second and third pieces of copper are proceeded with in exactly the sajne manner; and from the data obtained the amount of copper, equal to 1000 grs. of the solution of cyanide of potassium, is calculated according to the result of each experiment, and the mean of the three results is taken : this will be the standard. The following tabular form, from results actually obtained, will show how the data should be recorded : ' No. Weight of copper taken. No. of divisions of solution of cyanide of potassium required. Calculated amount of copper 200 divisions (1000 grains) of standard solution is equal to. 1 2 3 5 '055 grains 6-745 ' do. 5-155 do. 1 100-0 132-5 101-0 flean result or standard... 10-110 10-181 10-208 10-166 If the first two results agree, it may be unnecessary to obtain a third. Instead of calculating the standard as above, the calculation may be made so that the number of divisions of solution, equivalent to 10 grs. of copper, is used as the standard for estimating the percentage of copper in the ore. Method of Assay. A known weight of the ore is placed in a flask, or beaker-glass provided with a glass cover, and moistened with strong sulphuric acid; strong nitric acid is then added, and the whole is digested at a gentle heat, nitric acid being added cautiously from time to time, until nitrous acid fumes cease to be given off. The solution is diluted with a small quantity of water, and reheated ; when the ore is completely decomposed the contents are transferred without filtration 2 i 482 METHODS OF ESTIMATING COPPER BY WET ASSAY. into a pint flask, and diluted with water to about or f of a pint ; excess of ammonia is next added, and the solution is allowed to become quite cold. When cold, without regarding the amount of hydrated sesqui- oxide of iron which may be present, the standard solution of cyanide of potassium is gradually and cautiously added in small quantities at a time, with occasionally shaking the copper solution, until the blue colour is nearly discharged, and there remains only a faint lilac-tint. The number of divisions required to produce this effect is read off, and from the standard the per-centage of copper is calculated. Example : 200 divisions of the burette equal 10 grs. of copper; 20 grs. of the copper-ore required for decolorization 105 divisions ; therefore Divisions. Divisions. Copper. Copper. 200 : 105 : : 10 : 5-25 and 5-25 X 5 = 26 '25 per eent. of copper. The decolorization of the solution should take from to f of an hour for completion, according to the amount of copper present. The cyanide being added very slowly, especially towards the end of the process, the last tint should remain permanent, or nearly so, for about 10 minutes. To aid in recognising the tint of the solution a white glazed tile or piece of white paper should be placed under and behind the flask during decolorization. When the ore is a sulphide, with ordinary precaution it will be completely oxidized by sulphuric and nitric acids ; but if any globules of sulphur remain they can be easily picked out after the dilution of the solution, ignited, and the residue, if any, digested in nitric acid, and added to the bulk of the solution. If any difficulty is experienced in dissolving the basic sul- phates formed, some hydrochloric acid may be added with advantage. V\ hen silver is present, it must be got rid of by adding hydrochloric acid and filtering off from the residue. When sesquioxide of iron is present, it imparts a greenish appear- ance to the ammoniacal solution, and the proper tint of the solution is best observed by placing the eye on a level with the top of the liquid ; after a little practice the alteration in the tint of the oxide of iron which occurs will afford a sufficient indication until near completion, the reddish-brown colour becoming more distinct as the assay is pro- ceeded with. If any difficulty is experienced in observing the tint towards completion, the precipitate may be removed by filtration. When the assay is finished the sesquioxide of iron will be free from copper, as the portion at first, retained passes out into the solution during the course of decolorization. The sesquioxide of iron should not be filtered off in the assay, as it would retain a considerable amount of copper, which could not be perfectly extracted from it even by means of strong ammonia. By the cautious addition of cyanide of potassium to an ammoniacal solution of copper mixed with sesqui- oxide of iron, the latter may be wholly deprived of copper. After the copper has been determined, the sesquioxide of iron is filtered off, washed, dissolved in hydrochloric acid, reduced by zinc to the state of protochloride, and the amount of iron estimated by means of a standard solution of bichromate or permanganate of potash. INTERFERENCE OF OTHER METALS. 483 Interference of other metals. The following experiments have been made on this subject. The copper was dissolved in the usual way in nitric acid, diluted with water to about J- a pint, and ammonia added in excess to the solution. In each case the amount of solu- tion of the foreign metal added contained at least 5 grs. of the metal. Weight of Electro- type Copper taken. Estimation made in the presence of Remarks. Number of Grains of Solution of Cyanide of Potassium required. Calculated amount of Copper 1000 Grains of Standard Solution is equal to. Grains. 5-345 5-675 4-230 6-305 8-150 30-735 6-645 6-410 6-90 5-30 5-955 5-335 4-700 5-450 3-955 3-895 {Added as sesquichloride before) the addition of ammonia . f 910 960 720 1070 1350 1800 1160 910 1020 910 790 Grains. 5-873 5-910 5-875 5-890 5-960 5-960 5-948 5-810 5-838 5-863 5-974 ( Sesquioxide of) \ iron . ) I Lead, antimony, ) ( tin, & bismuth 1 Do. do. Nickel i (Added as chlorides before thel \ addition of ammonia / Do. do. do. Added as sulphates before the ad- dition of ammonia, there was no precipitate formed by the addi- tion of the cyanide, and the solu- tion was not decolorized, but re- mained yellow after the addition of 1630 grains of solution. .. Cobalt Added as chloride before the ad- dition of ammonia, a precipitate was produced during the experi- ment, but the solution was not ( decolorized by 2000 grains of standard solution J Antimony .... Added as chloride Bismuth Added as chloride, the precipitate) formed on dilution re-dissolved) on the addition of the ammonia ) (Added as chloride, no precipitate) Lead Tin ! Added as chloride, the precipitate) by ammonia nearly all dissolved, > leaving an opalescent solution ) Added as arsenic acid Arsenic ! Silver 1 Added as precipitated chloride, I the cyanide of potassium acts> upon it during decolorization . . . ) If Added as nitrate, interferes, as inl Silver Zinc Added as chloride, the solution was decolorized by 1300 grains of solution, and the ammoniacal solution became turbid towards . the close of the experiment , From the above experiments it will be seen that the metals which cause any material interference are silver, nickel, cobalt, and zinc. Certain precautions must be observed in the presence of the fol- lowing metals : 2 i 2 484 METHODS OF ESTIMATING COPPER BY WET ASSAY. Iron does not interfere, when present as sesquioxide, except me- chanically ; the presence of a very . large quantity somewhat pro- longs the assay by the time occupied in the precipitate subsiding sufficiently to observe the tint of the solution. AY hen present as arseniate it is soluble in ammonia, and in the presence of copper forms a brownish green solution: this can be easily remedied by the addition of some solution of sulphate of magnesia, when the magnesia combines with the arsenic acid, and the solution, after the lapse of a few minutes, acquires the proper tint. The assay can then be proceeded with, without nitration, in the usual way. Manganese is seldom present in copper-ores in sufficient quantity to interfere. If present, it may be completely precipitated as peroxide by the addition of carbonate of ammonia and a few drops of bromine, and heating the solution, or allowing it to stand some time. When cold the assay can be proceeded with. Nickel and cobalt. Their presence is generally detected during the assay by the solution remaining of a yellowish tinge after the blue colour has disappeared by the addition of the cyanide of potassium. When present the copper must first be separated by precipitation. Zinc. It very often exists in copper-ores in the form of blende. It gives no indication of its presence during the assay. AVhen present it is first converted into cyanide by the cyanide of potassium, before the copper is acted on, and the degree of this interference is pretty constant for a given weight of .zinc : thus a quantity of standard solu- tion of cyanide of potassium equivalent to 1 grain of copper would be equal to about 3 grs. of zinc. Attempts have been made to estimate zinc by colouring the aminoniacal solution by means of a small quan- tity of solution of copper of known strength, and then adding the solu- tion of cyanide until decolonization occurred ; but the standard solution' of cyanide in this case requires to be made so weak that the results were not satisfactory. Arsenic. It does not interfere except when present with iron. Silver. It interferes, but is easily got rid of by the addition of a few drops of hydrochloric acid, and filtration before the addition of the ammonia; the precipitation of chloride of silver is much pro- moted by shaking or the application of heat. Lead, tin, antimony, bismuth, do not interfere. \Vheii zinc, nickel, and cobalt are present, the copper must first be separated : this may be effected by a. Precipitation by means of a piece of iron from the dilute sul- phuric acid solution, nitric acid being absent ; redissolving the pre- cipitated copper in nitric acid, diluting with water, &c., and estimating by the cyanide of potassium. b. Passing sulphuretted hydrogen through the acid solution, filtering and redissolving the precipitated sulphide of copper in nitric acid, &c. c. Precipitating the copper by hyposulphite of soda, and dissolving the di sulphide of copper in nitric acid, &c., and estimating by the cyanide. This method is the quickest and best : the first, however, is a useful practical method. BY PRECIPITATION WITH HYPOSULPHITE OF SODA. 485 BY PRECIPITATION WITH HYPOSULPHITE OF SODA. The ore is decom- posed by boiling in the usual way by means of sulphuric and nitric acids ; the solution is diluted, filtered, and washed ; after filtration, the liquid, in quantity from -j to f of a pint, is heated to boiling ; solution of hyposulphite of soda is then cautiously added in small quantities, when after a few seconds it will rapidly darken ; and if sufficient hypo- sulphite has been added the whole of the copper will be precipitated as disulphide of copper, mixed with free sulphur, and with the evolu- tion of sulphurous acid. The heat is continued, and if the whole of the copper is thrown down from the solution the precipitate will coagulate as it were, and separate easily from the supernatant liquor ; and a white precipitate of sulphur will be produced on further addition of the hyposulphite. The whole is thrown on a filter while hot, and the precipitate is washed w'ith hot water and dried ; when dry it is carefully removed from the filter, without detaching any portion of the paper ; the filter with adhering particles of the precipitate is ignited, and the residue added to the dried disulphide of copper ; the whole is then heated in a covered porcelain crucible, placed over an air-gas- burner, until the excess of sulphur ceases to burn round the upper edge of the crucible. When cold the crucible and contents are weighed, and the weight of the former deducted. The disulphide should have a dull greyish black colour. If any doubt occurs as to the state of the residue, it may be re-ignited in admixture with sulphur, and weighed again. The amount of ash from the filter is subtracted, and from the weight of the disulphide of copper the amount of copper is calculated. 100 parts of disulphide of copper contain 79*85 parts of copper. If any black oxide remains with it, it is of no consequence, as it contains the same amount of copper as the disulphide, and therefore will not interfere with the result. After weighing, the disulphide should be tested to see that other metals are absent. This method may be used for separating copper from nickel, cobalt, zinc, manganese, &c. We have employed it for several years past. Hyposulphite of soda has been used in the quantitative estimation of copper in this country during several years, although the process has only recently been described by Fresenius. In assaying ores consisting chiefly of copper or iron pyrites, instead of treating the ore direct with acids, after weighing, it may be partially roasted in a porcelain capsule in a muffle, or over an air-gasburner for about 10 minutes, by which means a large proportion of the sulphur will be driven off". The roasted ore is subsequently boiled in nitric acid during 20 minutes or an hour, when the copper will be dissolved out, a portion of the oxide of iron remaining insoluble. Where a large number of assays have to be made, this mode is sometimes resorted to for the purpose of economising nitric acid, as the whole of that portion of acid which would be decomposed in the oxidation of the sulphur is thereby saved. It also obviates the trouble occasioned, sometimes, from the globules of sulphur remaining after prolonged boiling of the ore in nitric acid direct. The fact of a portion of the iron being left after calcination and boiling in nitric acid, as insoluble oxide, is 486 METHODS OF ESTIMATING COPPER BY WET ASSAY. advantageous, especially in assaying by the cyanide method. The quantity of copper retained, if any, in the residue amounts to a mere trace, if the ore has been properly treated. In assaying some poor copper-ores, where the copper is diffused through a large quantity of gangue, as in clay-slate, &c., it is some- times difficult, if not impossible, to get out the last portions of the copper by means of acid ; in this case the ore may be calcined, and then fused with bisulphate of potash, &c., and the estimation afterwards proceeded with in the usual way. This plan may also be employed for determining the copper in slags. By way of illustration the following results by different methods, and obtained by different persons with the same ore, are appended : Iron pyrites containing copper pyrites gave By boiling the ore in nitric acid and estimating thel 6-94 copper direct by cyanide of potassium (by E. Smith)/ By calcining the ore previous to boiling in nitric acid ( Q . ~~rj (by E, Smith) ........j By precipitation by hyposulphite of soda, and weighing ) ft . n9 . as disulphide (by Mr. C. Tookey) j By cyanide of potassium, by another person 6 12 By cyanide of potassium, by a third person 5 '98 By dry assays 5*80 BY A STANDARD SOLUTION OF HYPOSULPHITE OF SODA. This process was first described by Mr. E. 0. Brown, 6 and is extensively used for the determination of copper in gun-metal, &c., at the chemical department of the Eoyal Arsenal, Woolwich. It is especially adapted for the estimation of copper in commercial varieties of copper, bronze, &c., where lead or iron is not present in large quantity. It requires some modification to render it applicable to the assaying of copper- ores. The process is founded on the reaction between iodine and hyposulphurous acid, when the products are hydriodic and tetra- thionic acids. The completion of the reaction is manifested by the bleaching effect produced upon a solution of starch added during the time of experiment. Slight differences of temperature, or variations in the mode of manipulation, do not in any way affect the results. For experiments upon this subject, the paper before alluded to should be consulted. The reagents required are 1. A solution of hyposulphite of soda. This is made by dissolving 1300 grs. of the crystallized salt in 4 pints of water, and standardizing by means of weighed pieces of pure electro- type copper by the process hereafter described, the mean results being taken as the standard. 2. Iodide of potassium. The salt may be used in crystals, and should be free from iodate of potash. 3. Solution of starch. This is prepared by boiling starch in a large quantity of water, and allowing it to stand until the insoluble residue has subsided ; the clear supernatant liquor, which may be decanted off, is employed. Quarterly Journal of the Chemical Society, April, 1857. COLORATION-TEST. 487 Process. From 8 to 10 grs. of the copper or alloy are dissolved in dilute nitric acid, and the nitrous acid is expelled by boiling. To the solution, diluted with a small quantity of water, carbonate of soda is added until a portion of the copper remains precipitated. The solution is treated with an excess of pure acetic acid, and poured into a pint- flask, when it is further diluted with water. About 60 grs. of iodide of potassium are dropped into the flask and allowed to dissolve. The standard solution of hyposulphite of soda is now poured in until the greater part of the pure iodine disappears, and the solution acquires a yellow colour. A little of the starch solution is now added, and the addition of the hyposulphite of soda cautiously continued until bleach- ing is completed, or, in other words, until the solution becomes colour- less. The number of divisions on the burette containing the standard solution is read off, and the amount of copper calculated therefrom. As iron is generally present in copper-ores, and the red colour of the acetate of iron renders it somewhat difficult to observe the reaction of the hyposulphite during the process, in order to make the process applicable to the estimation of copper in copper-ores, it is necessary, after having decomposed the ore, &c., and expelled the nitrous acid, to dilute and filter, precipitate the copper from the filtrate by hypo- sulphite of soda, and redissolve the precipitated disulphide of copper in nitric acid. The copper may then be estimated in the manner described. When iodide of potassium is added to the acetic acid solution of protoxide of copper, di-iodide of copper is formed, with the liberation of an equivalent amount of free iodine which dissolves in the excess of iodide of potassium present, the potash combining with the acetic acid. By the addition of the standard solution of hyposulphite of soda, the free iodine is converted into iodide of sodium, with the formation of tetrathionate of soda. The starch solution merely serves to render the termination of the reaction manifest, the blue iodide of starch first formed becoming gradually bleached as the process approaches com- pletion. The reaction may be expressed by the following formula : 2(CuO, A) + 2KI = Cu2I + I + 2(KO, A) I + 2 v NaO, S 2 O 2 ) = Nal + NaO, S 4 O 5 COLORATION -TEST. A ready method of approximately determining the proportion of copper in slags containing only a small quantity of the metal is of much use. Even in the best-conducted copper- works it is desirable that the ore-furnace slag should be occasionally examined che- mically as to its content of copper ; and the smelter should not always be satisfied that it is sufficiently dean when he fails to discover any sensible amount of regulus in shots on its freshly fractured surface. The blue coloration-test by ammonia may be conveniently adopted for this purpose. A series of bottles of colourless glass, and of pre- cisely the same capacity, must be procured, and their form should be quadrangular in preference to cylindrical. These bottles are filled with dilute standard solutions of copper dissolved in excess of am- monia. In the first bottle there should be, say -,V of a grain of copper, 488 METHODS OF ESTIMATING COFFEE BY WET ASSAY. in the second T %, in the third T V, and so on. The intensity of the blue coloration will be proportionate to the amount of copper in the bottles. In testing a slag, it should be reduced to fine powder, and digested with nitro-hydrochloric acid ; and when the decomposition is complete, the solution should be diluted and poured into an empty bottle of precisely the same capacity and shape as those of the series containing the standard ammoniacal solutions of copper. Should it be desired to sepa- rate the silica, the process usually employed in an analysis must be resorted to. The slag, in a finely divided state, is digested with hydro- chloric acid until it is completely decomposed, and then the whole is evaporated to dryness ; hydrochloric acid is poured on the dry residue, and, after the lapse of about half an hour, a little nitric acid or chlorate of potass is added to peroxidize the iron. The silica may now be easily separated by filtration, and the filtrate treated as above described. Ammonia in excess should be added, and the bottle filled up with water, so that there may be the same volume of liquid as in one of the test-bottles. This solution, with respect to intensity of colour, is now to be compared with those in the test-bottles ; and from that with which it most nearly agrees the amount of copper is deduced. It is requisite that the solutions should be very dilute, for otherwise the colour is so intense that no difference can be remarked even in solu- tions containing very different proportions* of copper. Any person who has had slight experience in chemical manipulation will be able, after a few preliminary experiments, to construct a series of these test- solutions as may best suit his convenience. It need hardly be re- marked, that in applying the coloration-test the bottles should be placed in a good light between a side window and the observer. It is not necessary to filter the solution after the addition of the ammonia, provided time be allowed for the complete subsidence of the precipitate of sesquioxide of iron formed. A sensible amount of copper is retained by the sesquioxide of iron, from which it cannot be separated even by a great excess of ammonia. The presence of oxides, such as those of nickel, cobalt, &c., which dissolve in ammonia, producing coloured solutions, will render this method of testing inaccurate. Le Play appears to claim the merit of having first suggested the coloration-test for copper-slags, &c. He thus writes : " After nume- rous trials, I have been led to the following process, which leaves nothing to be desired with respect to rapidity of execution : The principle of this process consists in judging of the quantity of copper contained in a liquor by the intensity of the blue tint which protoxide of copper (CuO), dissolved in ammonia, communicates to a given quantity of water." 7 Now, Heine in 1839 published a descrip- tion of this identical process, which he had employed in determining the amount of copper in slags ; 8 but Le Play's description of it did not appear until nine years afterwards. Mr. Keates informs me that he had used it so long ago as 1830. 7 Precedes Metallurg., p. 454. I fergehalts aus Schlacken. Btrgwerks- 8 Metliode zur Ermittelung des Kup- | fmmd, 1839. 1. p. 33. INACCURACY OF CORNISH METHOD OF DRY ASSAYING. 489 INACCURACY OF THE CORNISH METHOD OF DRY ASSAYING. This is a tender subject with copper-smelters. The fact is, that in the Welsh process of smelting a larger quantity of copper, or surplus, is obtained than is indicated by the produces of the Cornish assayer, that is, exclusive of the copper obtained from the one hundred weight of ore in excess of the ton of 2240 Ibs. (see p. 306), which, in former days, was supposed to be lost in transit from the mines ! and the amount of this surplus is regarded by the smelters as a trade secret. Whatever profit may have been derived from the surplus in what some smelters now designate as the palmy days of copper-smelting, it is certain that it cannot now be properly regarded as any special source of profit ; because it is an essential element in the consideration of the prices of ores, and is, accordingly, paid for by the smelter. In the present day the miner would not generally receive a farthing more for his ore, whatever changes might be effected in the plan of assaying and the mode of sale. Miners believe that they are cruelly victimized by the smelters, and often entertain the fallacious notion that if they could only smelt their own ores their dividends would greatly increase. Smelters consider it desirable to keep miners as much in the dark as possible, and so sus- picion of unfair treatment on the part of the former is engendered in the minds of the latter. My impression is I may be mistaken that there would be a much better understanding between miners and copper-smelters if copper-ores were assayed by a more accurate method than the Cornish, and sold at a given sum per unit (say 1 cwt.) of copper in the ore, according to the market value of copper at the time. I will now proceed to establish the fact that the produce obtained by the Cornish method, however skilfully it may be conducted, is al- ways sensibly below the actual amount of copper in the ore. When at Swansea in 1859, I was anxious to procure portions of the samples of ores which had been supplied to the smelters from the ore-yards prior to a sale at Swansea on a particular day. In my innocence I did not for a moment conceive that any smelter would object to furnish me with the residues of the identical samples which had been operated upon by his private assayer ; and accordingly I applied for some of these residues to certain smelters whom I have long had the pleasure of numbering amongst my personal friends. Great was my surprise when I met with a distinct refusal, on the ground that, being members of the Association, they were pledged to inviolable secrecy in every- thing relating to assays ; but the refusal, I should observe, was accom- panied with every courteous expression of regret. Nay, in my utter ignorance, at the time, of the pledge of secrecy, I had even ventured to ask for the produces rendered by the Association assayers ! (see p. 307, ante). In spite of these unexpected obstacles, I am happy to state that I have succeeded in honourably procuring all the information which I needed, or even desired. Captain Petrie and other gentlemen connected with the ore-yards of Swansea have provided me with samples of all the parcels of ores, &c., sold at Swansea Nov. 15, 1859, the sampling having taken place on the 6th of October preceding. Out of these samples a series has been selected for examination in the Metallurgical 490 COMPARATIVE RESULTS BY CORNISH AND WET METHODS. INACCURACY OF CORNISH METHOD OF DRY ASSAYING. 491 Laboratory. The whole of each sample was dried and reduced to im- palpable powder, and portions were assayed with the greatest possible care. The results will be found in the accompanying table. In the 2nd column the word in italics is the name of the vessel in which the ore was imported. In the 4th column are the produces of the sellers, as published in the * Swansea Ore Circular' for Nov. 15, 1859. In the 5th column are the produces of the Association Assay ers, which are sup- posed to be kept impenetrably secret. I have received these results with permission to publish them. I am not indebted for them to any smelter, or any person in the employ of any smelter, or to any assayer. The fact of my possessing them will satisfy the members of the Associ- ation that, in accordance with the old belief, secrets which are known to many have but little value, and might as well be known to all. In the 6th column are the percentages of copper determined by my col- league, Mr. R. Smith, who conducts our Assay Laboratory, and has during the last ten years had almost daily practice in assaying : all the results were obtained by the process a, described further on. In the 7th column are the percentages of copper determined by a former student of our School of Mines, Mr. W. Weston, who has been con- stantly engaged for two years in analytical work in the Metallurgical Laboratory. The produces marked * were obtained by process 6, and the rest by process c, now to be detailed. a. The ore was decomposed by sulphuric and nitric acids : the solution was diluted with water, and then treated with ammonia in excess. The copper in this ammoniacal solution was estimated by a standard solution of cyanide of potassium, according to Parkes's method. 6. The ore was decomposed by sulphuric and nitric acids ; the solu- tion was diluted with water and filtered ; hyposulphite of soda (NaO, S 2 2 ) was added to the filtrate ; the disulphide of copper, thus precipi- tated, was dissolved in nitric acid; the solution was treated with excess of ammonia, and the copper in this ammoniacal solution was de- termined by Parkes's method. c. The ore was decomposed as above, and the copper precipitated as in process b ; but the disulphide of copper, after solution in nitric acid, was precipitated by potash and weighed as protoxide. It will be perceived from the figures at the bottom of the 10th column that on the average the smelters receive 100 tons of copper, when the Cornish method of assaying would lead to the conclusion that they receive only 91-28 tons; but it must be remembered that there is a considerable and inevitable loss of copper in smelting. The surplus which the smelters have been accustomed to expect from their furnaces, beyond the produce indicated by the Cornish method, is about 9 per cent. ; but part of this excess is derived from the hundred weight of ore which they obtain in addition to the ton. The surplus deduced in the Table is 8'72 per cent. ; but this deduction, it must be borne in mind, is made on the supposition that an equal quantity of each ore was sold at the sale in question, which was very far from being the case. The average excess of copper which the smelters actually receive must be much greater than 8-72 per cent. ; for, in addition to the surplus of 9 per cent, which their furnaces should 492 INACCURACY OF CORNISH METHOD OF DRY ASSAYING. yield, a large amount of copper is lost in the ore-furnace slag. Of late, I know that some disappointment has been experienced on account of this surplus having been sensibly below 9 per cent. ; and this has been attributed to the large importation of rich regulus, &c., in assaying which by the Cornish method there is proportionately less loss of copper than upon poorer ores or products. It is, however, not a little singular that there should be so great a difference as 5-41 between the results of the Cornish method and those of wet processes in a regulus containing more than 60 per cent, of copper, like that of No. 1 7 of the Table. Perhaps the Association Assayers may have heard expressions of disappointment regarding the decrease of the surplus, and have con- sidered it proper to make a certain reduction from the produces actually obtained by them from rich cupriferous substances of this nature, on the ground that the surplus ought to 3 oj 15 15 15 117 8 6 98 12 85 1 I give the addition under each column as it is in the original. A sum of 11. 12s. 6d. has been added to each of the totals ; but for what reason I do not know. I should wish it to be distinctly understood that I simply present the foregoing statements respecting smelting costs as I received them, and that personally I cannot in any degree be responsible for their accuracy. I am in possession of information on this subject which I cannot in honour reveal ; but I may state my impression that in the present day a copper-smelter has not much chance of adding to his wealth, if he is not a shrewd, judicious, and energetic man of business. COST OF COPPER-SMELTING AT ATVIDABERG IN PRODUCING 1 SWEDISH CENTNER OF REFINED (GAAR) COPPER. Materials employed in making 1 centner of refined copper. Centners. Wages in raising and) Ore 15-27 Six-dollars (Mynt). Kix-dollars (Mynt). carrying ore Calcining Residua, slags, &c 6'23 Do. remelting Regulus 5-23 ( Do. crushing, car- ) rying I Black copper . . 1 i 7 /Smelting, refining, and \ { regulus-calcining ...J Coke, total 3-48 ...!. ...... Charcoal, total, for smelt-) ing and refining J b ' b/ cub. feet. Wood, in calcining ores... 6 77 Do. in calcining regulus 7*92 at 66 per centner 10 08 0-21 at 0-10 do. 0-62 at 0-265 do. 1-39 at 1-86 at 0-80 do. do. 0-83 6-36 5-50 at 04 per cub. foot. 27 at 0-04 do. 0-32 Rix-dollars (Mynt) 25-58 18 Rix-dollar (Mynt) = 1 1 do. do. = 13 -33d. 1 Swedish centner = 0-837 English cwt. of 112 Ibs Eng. cwt. of refined copper will cost 1 13s. lljd. very nearly, say 1 14s. ton will cost 34. M. Malmqvist estimates it at about 40 per ton. The total cost of making copper at Atvidaberg, per Swedish centner, including expenses of management, sick-fund, poor-rates, taxes, &c., is 42-16 Rix-dollars (Mynt), or 55 19s. per ton of 2240 Ibs. COMPOSITION OF COMMERCIAL COPPEK. 503 1g 1 ** f * ^-5 g" 2 CO l> TfH CD -rJH t> C5 IT* O IT O 9* rH rH O lit y a) : : o o : o : go :oo : : : : : OS -P rH .-^ .!a 2 rH 0^ CO |i|l i | OS os : : o : : : : : : : o o : : : : : | ^fe pj^ i "f rH 00 rH O rfl |> (N S 2 87. l-li 1 o"c is Hi S co rH : o : : : : ::::::: o :: os S >|ts fls'S S ^ S3 5 co ^ 2*, g. ilif I oo:oo :o :o::oo:::o CS ^3 Jg O o rH ipi tf W *O CO OS CO ^* rH 10 ft !^rrtP3 * r- l-Maj'SsO'.S^C fill 1 'ill 'ffl 504 ANALYSES OF EGYPTIAN AND INDIAN COINS. It is somewhat remarkable that no mention of antimony or arsenic is made in any of the foregoing analyses ; and yet, probably, no metals are more frequently present in commercial varieties of copper. A curious fact may here be stated concerning the occurrence of arsenic in copper of extremely ancient date. A flat piece of copper, sup- posed to have formed the blade of a knife, was discovered in boring a few years ago, at a depth, it is alleged, of 13 feet from the surface, below the site of the statue of Eamesses the Second, who is believed to have reigned in Egypt about 1400 B.C. A portion of the copper was analysed in my laboratory, at the request of Mr. Leonard Horner, by Mr. C. Tookey, and found to be composed as follows : Copper 97-12 Arsenic 2-29 Iron 0-43 Tin with traces of gold 0-24 100-08 The effect of the arsenic would be to communicate hardness ; but this blade, which contained 2 per cent, of arsenic, was, nevertheless, so soft that it could not have been an effective cutting implement. The presence of arsenic in this ancient relic may possibly have been accidental. The late Mr. Henry and myself detected antimony in a specimen of copper which I received from the late Mr. Paul Moore, a large consumer of cfcpper in Birmingham, and which was reputed to have been made exclusively from the fine Burra-Burra ores. In a specimen of copper recently examined in the Metallurgical Laboratory about 30 oz. of antimony to the ton were found : when rolled plates of this copper were cut with shears, the edges had a peculiar roughness. Bismuth does not appear in these analyses, and yet it is occasionally present in very sensible quantity. Iron is generally present in copper ; and, although the proportion is usually very small, yet it should be borne in mind that a considerable amount of this metal may be retained by copper. A singular copper-like ancient coin from India was submitted to me for analysis by my friend Mr. Edward Thomas, so well known for his acquaintance with the coins of India. 4 When broken across, its fracture was finely granular. It was found by Mr. C. Tookey to consist of Copper 94'59 Iron , 5-06 99-65 4 Mr. Thomas has communicated the following information respecting this coin : " An engraving of a coin of similar types to that of which the analysis is given above, is to be found in the ' Journal of the Asiatic Society of Bengal,' vol. vii. plate xxxii. fig. 12, and in Prinsep's 'Essays' (London, 1858), pi. xliv. fig. 12. The obverse, which of a man, with the legend, in old Pali letters, Khatrapasa Pagdmashasa. The reverse exhibits a rude figure of a horse, which, unlike the device of the obverse die, seems to have been merely sunk into the anvil as a catch to fix the planchet. This class of coin is supposed to have been issued by the local sovereigns of the Sub-Himalayan Gangetic provinces has been struck with an eifective and | prior to the introduction of Greek art well-cut die, bears the Buddhist device | See Prinsep's ' Essavs,' vol. i. p. 222." COPPER SHEATHING. 505 COPPER SHEATHING. Sheet copper is largely consumed in sheathing H. M. ships of war ; and extraordinary differences have been observed in the manner, in the degree, and in the rapidity with which different samples of copper are corroded by sea-water. The sheathing may be pretty uniformly corroded over the entire bottom of a ship, or it may only be sensibly attacked here and there on the surface of particular plates; it may be eaten away so as to form irregular holes, sometimes of considerable size ; it may be exposed during many years to the action of sea-water without presenting any marked sign of corrosive action, or in the course only of a few months it may become so honey-combed and full of holes as to be no longer serviceable. I have inspected a large number of specimens of copper sheathing which have been stripped off the bottoms of H. M. ships from time to time, and the results of my observation are such as I have just recorded. When copper is exposed to the action of sea-water, its surface acquires a green coating, which chiefly, I believe, consists of oxy- chloride of copper, formed by the conjoint action of the chlorides in the water and atmospheric air. When this coating has once formed, complex local electrical actions probably occur, and may occasion the formation of compounds either new or as yet very imperfectly under- stood. During many years I have had experiments in progress on such local actions, and the results are not a little complicated and perplexing. I had the opportunity in Paris some years ago of inspecting a piece of old copper sheathing on which were small but distinct crystals of dioxide of copper. A thorough investigation of the composition of the products of corrosive action would be extremely instructive. It is possible that researches of this nature have been made and published ; but, if so, I have not been so fortunate as to meet with the record of them. It has, I believe, been satisfactorily established that in some locali- ties corrosion proceeds more rapidly than in others; and this has been attributed to differences in the quality of the water. Thus, the late Professor Daniell of King's College investigated the question of the rapid destruction of copper sheathing on the African station, and arrived at the conclusion that it was due to the presence of sul- phuretted hydrogen in the water. 8 To the same agent his successor, Dr. Miller, has recently referred the rapid corrosion of yellow-metal (see the article on Brass) which is alleged to occur in the London Docks. The copper-smelters and others who manufacture this metal were so impressed with the belief of the extremely corrosive nature of the water of these docks, that some time ago they jointly agreed to refuse their usual guarantee for the durability of the metal during a 5 It is recorded (p. 57) in the Keport of the Committee appointed to enquire into the state of the Copper Mines and Copper Trade of the Kingdom (ordered to be printed May 7, 1799), that a Liver- pool merchant-vessel was sheathed with copper in April, 1785, had made sixteen voyages from Liverpool to the coast of Africa, thence to the West Indies and back to Liverpool, and, after insignificant repairs, the copper in April, 1799, was still good. 500 COPPER SHEATHING. certain term, in the case of vessels entering and lying in the London Docks. They were supported in this step especially by the evidence of Dr. Miller, who declared that the dock water was strongly impreg- nated with sulphuretted hydrogen, and was therefore likely to act injuriously upon the metal by the formation of metallic sulphides on its surface. But the Company, it should be stated, established con- clusively that the yellow metal sheathing now in use varies remark- ably in the degree of rapidity with which it suffers corrosion. Now whatever truth there may be in the statements as to the greater corrosive action of sea-water in particular localities, it has been esta- blished, on unquestionable evidence, that commercial copper differs remarkably in its power of resisting that action. Yery numerous observations and comparative experiments on this point have been made in H.M. dockyards, and from the results which have been re- corded I insert the following by way of illustration : 6 Oct. 5, 1845, the "Vanguard" was sheathed on the starboard side with 400 sheets of copper (C) smelted and manufactured exclusively from Cornish ores by Messrs. Grenfell and Co., and at the same time the port side was sheathed with 400 sheets of copper (F) made from a mixture of British and foreign ores. May 29, 1849, 19 sheets were taken off from each side and weighed, these sheets having been pre- viously separately weighed and marked for the purpose of identifica- tion. The original weights of the sheets (C) ranged from 7 Ibs. 8 oz. to 9 Ibs. 8 oz., and the original sheets (F) ranged from 8 Ibs. 6 oz. to 9 Ibs. 15 oz. The results are given underneath : Average Loss per Sheet. Average Loss per Sheet per Annum. C 9-32 oz 2-66 oz. F 15-32 4-38 The "Sappho" was wholly sheathed with copper "of Chatham manufacture, 1847 " (C) by which, I presume, is meant copper from various sources remelted, cast, and rolled at Chatham with the ex- ception of 50 sheets, on each side, of copper smelted at Swansea by Messrs. Williams, Foster, and Co. (H). This copper is designated in the Eeports as " hard metal sheathing," and its use was suggested by Mr. Moyle, an experienced practical smelter formerly in the employ of the firm above-mentioned, but now engaged at Chatham Dockyard. This " hard copper" is described as having been " separated from the soft at a certain process of smelting," from which it may be inferred that it consisted of the " bottoms" reduced in the process of making best selected. Of this " hard copper " 30 tons were refined under Mr. Owen's inspection at Swansea in 1846, and in the refining much less lead was added than usual. The refined " hard copper " was rolled into sheets and employed ill several experiments. A portion was submitted to Mr. Prideaux, who states that he found it to contain the following foreign matters : 6 When no particular authority is given j must be understood that they have been for statements in the following pages, it | derived from official documents. COPPER SHEATHING. 507 Per cent. Zinc 0-200 Iron 0-076 Nickel 0-040 Tin 0-019 Lead 0'007 \ 0-491 Silver O'OOl Antimony 0'024 Arsenic 0'124 Manganese traces Silicon 0-035 Aluminium O'OIO Calcium 0-055 0'154 Magnesium 0'007 Potassium and sodium 047 0-645 Mr. Prideaux remarks that the silicon and following ingredients " may have possibly been dissolved from the glass in which the copper was boiled with acid." This possible source of error did not probably occur to Mr. Prideaux until he had completed this most elaborate ana- lysis, for otherwise he would hardly have expended the time necessary to determine these ingredients quantitatively. The " Sappho," thus coppered, had been lying in Portsmouth harbour only 12 months, when her sheathing was found to be so much deteriorated, being in some places eaten away in holes, that it was necessary to remove not fewer than 80 sheets. The average loss per sheet of C was 16f oz., while the average loss of H was only 2f oz. It is recorded that some of the same " hard metal" sheathing was tried on a merchant ship, the " Esk," belonging to Mr. Brocklebank, and reported on most favourably. In the case of H.M. ship " Howe " at Sheerness in 1847, there were two different kinds of copper on her bottom : one the old copper manu- factured at the Portsmouth Metal-Mills, and the other new cake-copper produced from a mixture of British and foreign ores, and supplied by contractors. AYith the former (probably manufactured in about 1832-1833) the average loss per sheet per annum was only 0-79 oz., or 11 oz. in 14 years; whereas with the latter (smelted and manufac- tured in 1843) the average loss per sheet per annum was 4-3 oz., or 15 oz. in 3j years. I find it recorded that copper sheathing was first applied in the Navy in 1761, when one ship was coppered ; a second was coppered in 1765, a third in 1770, four in 1776, nine in 1777, and within three years from the last date all the ships of the British Navy were coppered. It is maintained that the copper formerly employed in the Navy re- sisted the corrosive action of sea-water much better than that produced in recent times, and that manifest deterioration in the quality of copper sheathing commenced about 1832 or 1833. From the peace in 1815 to 1832 no copper had been purchased for the Navy, a sufficient supply of new sheathing having been obtained during this period by re-melting 508 COPPER SHEATHING. and re-manufacturing the old sheathing. Now it is stated that the smelting of foreign in admixture with British copper-ores was intro- duced during the years 1833, 1834, and 1835; and hence the com- mencement of deterioration in the quality of the copper is found to be nearly coincident with the introduction of foreign copper-ores into this country, or at least with the introduction of new copper into the Navy. The following table, extracted from official records, may be interesting to copper-smelters : TABLE SHOWING THE AMOUNT OF CORKOSION IN SHEATHING MADE FROM DIFFERENT KINDS OF COPPER, FROM 1816 TO 1 844 INCLUSIVE. The copper was examined or taken off between July, 1843, and Dec. 1845. Name of the Ship. Station. . Date of Manufacture of the Sheathing. Average Loss per Sheet per Annum. Hamoaze 1816 oz. 0-83 Fountain ( water-tank ) ... Chatham (breakwater 'i vessel) / Plymouth Sound ... do. do. 1817 1817 0-55 0-33 Semiramis . Sea and harbour . . . 1821 0-82 Nereus Hamoaze 1821 0-85 Netly (tender) Sea and harbour ... 1823 0-66 Armada Chatham Hamoaze Plymouth Sound ... 1824 1825 0-75 0-57 Impregnable Sea and harbour 1825 0-66 Netlv (tender^ do do . . . 1826 0-75 Beresford (hov) Plymouth 1828 1-25 Sta- Hamoaze 1829 1-50 Cvckms CSV") Sea 1829 1-75 Li"ht vessel Breakwater 1832 0-50 America Hamoaze 1832 2-25 Forth do 1833 1-00 Grecian Sea . 1835 6-13 Falmouth (lighter) Sea and harbour . . . 1837 1-25 Cyclops (S. V ) . . . . Sea 1838 3-20 Nimrod do . 1838 5 20 Forth Hamoaze 1838-9 11-00 Eiidymion Sea 1840 5-33 Calcutta do 1840 4-50 Vanguard do 1840 0-90 Indus .. , . do ... 1840 4-50 Clarence Hamoaze 1840 2-66 Volage Sea 1841 7-33 Superb Hamoaze 1842 1-33 Royal William. . do 1842 2-56 Pandora . . do 1843 11-33 Melampus Sea ... ... 1843 6-00 Do do 1844 6-00 Persian Hamoaze 1844 4-17 Acorn do 1844 6-50 Daring Sea 1844 1-00 It appears from this table that the transition from good to bad copper was abrupt, and occurred about 1833, as previously stated. But notwithstanding the conclusion to which the results in the fore- going table would lead, it is certain that all the old copper was not good, and that all the new copper is not bad. In support of the first proposition the following evidence may be COPPER SHEATHING. 509 adduced. Mr. Owen had access to the papers of Sir Samuel Bentham, and extracted from them this tabular record : Ship's Name. When taken off. Duration of Sheathing. Remarks. Repulse Dec. 1808 Years. Months. 2 6 ( Copper in state of cor- Dragon Feb. 1807 2 4 \ rosion. Warn thin. Jan 1808 2 8 Do. Jan 1808 2 5 Copper corroded. Dryad June, 1808 2 11 Do. and very thin. Lark Sept 1808 3 Do. worn thin. EXTRACTS FROM DOCKYARD REPORTS. Portsmouth Yard. " H.M. ship 'Intrepid,' stripped in Dec. 1796. The copper was of the most approved sort, but it was only on 4 years. Copper fastened ; ^ of the copper was much corroded, very foul, and in general very bad." Sheerness Yard. Report of the officers to the Navy Board, March 9, 1797: Ship's Xame. Marks on the Sheet. How long on. Loss per Sheet per Annum. Observations. Adamant 28 OZ. ^ &c. ... Years. 8 oz. 1-56 Good. Hind C 28 oz. no mark 10 1-85 Do. Ariadne 28 oz 4 3-22 Middling. Do 32 oz . . 4i 13-65 Bad Proserpine Latona - 28 oz. ? &c. ... 6 7 1-56 1-21 Good. Do. (u Plymouth Yard, Feb. 22, 1797." The officers send to the Navy Board a sheet from the ' Chatham,' and supposed to have been on 18 years; not the least appearance of corrosion or foulness, like the Copper now in use [sic in ital.], except on the nails, which being of mixed metal are very foul with barnacles and long weeds hanging to them. They also send a sheet taken from the bottom of H.M.S. ' Sheerness/ of the im- proved sort of copper, on little more than 3 years, marked ' Owen Williams,' and eaten through in many places, but not so bad as those taken from the same ship and returned to the contractor. From these circumstances, they say, together with many general observations we have had it in our power to make, we are fully convinced that the copper now in use (1797) is adulterated, and that to a very great degree." " The master shipwright recollects the ' Daedalus ' being coppered by him at Liverpool in 1780, and also its being taken off again under him in the river Thames in 1791, a period of 11 years, when it was found to be in a very perfect state without any corrosion, and a square foot of the 32-oz. copper not to have lost in weight more than 510 COPPER SHEATHING. 3 or 4 oz., and that evidently from wear ; whilst he recollects many other instances of copper not having been on more than 2 years, and its being eaten quite through. This, with what here passed under our own notice at this yard, confirms us in our opinion of its being greatly adulterated." The then Xavy Board were so impressed with the deterioration in the quality of the copper used for sheathing, which appeared to have commenced about the year 1786, that they employed Dr. Higgins, of Greek Street, Soho, to make analyses of the copper with a view to ascertain how far the presence of foreign matters might influence the corrosive action. Dr. Higgins reported that he found the copper to be alloyed with about 2J per cent, of lead, and other metals (chiefly tin and antimony) in much smaller quantity ; that copper so alloyed is apt to be injured in repeated annealing, especially when coal is used as fuel, and is more easily penetrated and corroded by the salts contained in sea- water. In support of the other proposition, that all recent copper-sheathing is not bad, the following experimental results may be cited : In October, 1845, the " Superb " was sheathed on the starboard side with 400 sheets of copper produced exclusively from Cornish ores by Messrs. Grenfell and Co. (G), and on the port side with the same number of sheets of old re-manufactured copper rolled at Portsmouth (P). On Feb. 24, 1849, it was reported from Portsmouth Dockyard that both kinds of copper on this ship were found on examination to be remarkably free from corrosion, and less foul than has usually been observed in copper on the bottoms of ships after having been so long afloat. The actual results are as under : Average Loss per Sheet per Annum. G 2-34 oz. P 2-12 Great differences were noted in the degree of corrosion in a few of the plates, especially in the case of G : thus one plate, which originally weighed 8 Ibs. 10 oz., lost 17 oz. ; whilst two other plates, which ori- ginally weighed 9 Ibs. 3 oz. and 9 Ibs., are stated to have lost nothing in weight ; in the case of P the extremes were losses of 3 oz. and 10 oz. in plates which originally weighed 9 Ibs. 3 oz. and 8 Ibs. 3 oz. re- spectively. It is suggested that the great loss in some of the plates of the most corroded (G) might have been caused by bad rolling. There are many persons, both in and out of the House of Commons, who stoutly maintain that in no case is it desirable that the Govern- ment should engage in manufacturing operations, and that it would better consult the interests of the nation by contracting with private manufacturers for any work which may be required to be done ; and grave charges of reckless expenditure and general maladministration in H.M. dockyards have recently been brought against the Admiralty by members of Parliament, eminent engineers, shipbuilders, and others. While it is admitted that in some instances these charges are well founded, yet, in common justice, it deserves to be recorded that the Government has acted wisely in not always trusting too implicitly in COPPER SHEATHING. 511 matters of business to the disinterested patriotism of our manufac- turers. In proof of this the following somewhat startling facts may be adduced : The attention of Sir S. Bentham was directed, in 1803, to the large expense incurred by the Navy in the re -manufacture of old copper- sheathing. At that time the re-manufacture of this old metal was con- ducted by private manufacturers, who were paid at the rate of 4^cL per lb., or 421. per ton. At the recommendation of Sir Samuel, metal-mills were erected at Portsmouth expressly for carrying on this work, when the private manufacturers offered to reduce their charges to 2$d. per lb. ; but it was eifected at the Portsmouth Mills at a cost only of 1 hours, after which it is gradually melted down and exposed to the air during about 18 hours : this process is repeated twice, in precisely the same manner, in a furnace of exactly the same construction. /Hister copper, containing about 95 per cent, of copper, is* thus pro- duced : it is refined in the usual way. ( 518 ) ZINC. HISTOKY. IT is stated that Basil Valentine first employed the word zinc ; but that Paracelsus, the renowned latro- chemist who wrote in the 16th century, was the first to associate the word with a metal pos- sessing the distinctive characters of zinc. 1 It has, however, been admitted by Beckmann and others that this metal was first described in the 13th century by the Dominican monk Albert of Bollstedt, commonly known as Albertus Magnus. But the following passage occurs in Strabo, from which one might at first almost be disposed to conjecture that zinc in its metallic state was not unknown to the ancients : " There is a stone near Andeira, which, being burnt, becomes iron; afterwards when melted in a furnace with a certain earth, it drops false-silver, which, with the addition of copper, pro- duces what is called the mixture, and which some name Orichalcum. 2 False-silver is also found in the neighbourhood of Tenolus." 3 Ori- chalcum, or Aurichalcum, was a metallic substance so closely resem- bling gold in appearance, that Cicero puts the question " whether, if a person should oifer a piece of gold to sale, thinking that he was only disposing of a piece of orichalcum, an honest man ought to inform him that it was really gold, or might fairly buy for a penny what was worth a thousand times as much." 4 Now, the only metallic substance which it is possible to conceive the ancients could have mistaken for gold is the alloy of copper and zinc, termed brass. The alloy of copper and tin, termed bronze, which was well known to the ancients, could certainly never have been mistaken for gold, unless we assume that Cicero referred to a gilded or plated article ; and that the gilding of bronze was practised in very ancient times, is estab- lished by the fact that in the British Museum there are ornamented vessels of bronze from Nineveh, upon the surfaces of which distinct traces of gold still remain. The expression, false-silver, is remarkable, and must have been intended to indicate a metal having a certain degree of resemblance to silver in colour, and probably also in fusibility and hardness. But the only other metals then known, which in colour could be said to resemble silver, were mercury, lead, and tin ; and as 1 GeschichtederMetalle.Zippe. Wien, : pounded of opos, a mountain, and 1857, p. 241. An elaborate account of the copper, history of this metal will he found in the ! 3 Lib. xiii. I am indebted to my friend following work : Geschichte des Zinks : Dr. Smith for the literal translation of in Absicht semes Verhaltens gegen andere j this passage, which is quoted by Watson. Kbrper.etc. G. F. C. Fuchs. Erfurt, 1788. ; Quote d from Watson's Chemical Es- Derived from 'Opeixa\Kos, com- says, 1786, 4. p. 85. ZINC HISTORY. 519 these metals may be easily recognized, if any one of them had been intended by the term false-silver, it would surely have been designated by its proper name. Moreover, two ont of the three, mercury and lead, could not have been intended, because neither would in any degree communicate to copper the properties ascribed to orichalcum. There are, however, difficulties attending this interpretation of the passage of Strabo. Zinc being very volatile at the temperature at which it is reduced from its ores, the expression drops false-silver would seem to be quite inapplicable, unless the furnace to which Strabo refers had, like a modern zinc-furnace, been provided with a suitable distillatory and condensing apparatus ; which is extremely improbable. Yet that, in the absence of any special apparatus of this kind, the vapour of zinc may be accidentally condensed in cracks in the walls of a furnace and trickle down in drops, is proved by the fact that I have received from my friend, Mr. Parry, of the Ebbw Vale Iron-Works, specimens of zinc which have thus condensed and trickled down in drops through cracks near the twyers of one of the blast- furnaces, in which a zinciferous ore had been smelted. In regard to the statement concerning " the stone, which, when burnt, becomes iron," and its fusion with a certain earth, there ' is not a little perplexity. Supposing zinc really to have been indicated by the term false-silver, it is doubtful whether the stone or the earth, or both, yielded the zinc. Various conjectures, more or less plausible, might be offered, which, though interesting in an archaeological point of view, would be out of place in this work, ^Yhatever opinion may be entertained as to the interpretation of the passage in Strabo, there is not, I believe, in the works of any other ancient classical author, a single statement which would lead to the inference that the Greeks or Eomans had the slightest knowledge of the existence of such a metal as zinc. It is possible that Strabo may have incorrectly described the process of making orichalcum ; for in the.present day, when travellers possessing no knowledge of metallurgy, though they may be generally well- informed men, attempt to describe, from their own observation, even the simplest metallurgical process, they almost invariably commit great errors, and fail to render an intelligible account. On the other hand, it is not improbable that zinc may have been discovered at a much earlier period than is supposed, as it appears to have been brought from the East by the Portuguese a century before it was produced as a commercial article in Europe, unless we except Goslar, where, according to Lohneiss, it had been obtained prior to 1617. 4 In the early part of the 17th century the Dutch captured a Portuguese ship with a cargo of zinc, which was sold under the name of speautre, xpiaiiter, speauter, or spialter : hence the word i^peltrum, introduced by l>oyle, and the English word spelter, which is almost the only name for zinc in our workshops at the present day. 6 During the last century 5 Bergman's Physical and Chemical . Trans. London, 1784, 2. p. 313. fi Beckmann's History of Inventions. London, 1814, 3. pp. 91-2. Watson's Chemical Essays, 4. p. 2. Watson states that the metal with which this vessel 520 ZINC HISTORY. zinc was largely imported into Europe, under the name of tutenag, it is said, from the East Indies/ Several localities in the East are specified from which the metal was procured ; and, as China is one of them, and we know that the Chinese possess considerable metallurgical knowledge, it may safely be concluded that zinc was actually produced in China, There is a tradition, moreover, recorded by Bergman, that an Englishman visited China in the last century expressly to learn the art of making zinc ; that he attained his object, and returned home in safety with the secret ; and that some time afterwards works were erected at Bristol for the extraction of zinc by distillation per descensum. 8 By diligent historical research we should probably arrive at positive information on the subject. If we admit that zinc was regularly produced in China as an article of commerce at the period above referred to, it is not unreasonable to suppose that the Chinese practised the art long before that period, possibly in the later times of the Greeks and Ro- mans ; and that the metal may have thence found its way into Europe. In 1721 Henckel published the fact that zinc might be procured from calamine by means of phlogiston, but he concealed the method. In 1742 A. Van Swab extracted it from the ores by distillation, at Westerwick, in Dalecarlia, where it was proposed to erect works to conduct the process on a large scale, but no further step was taken. In 1746 Margraaf, in ignorance of what had been done in Sweden, made known a process of extracting zinc, which he had himself discovered. 9 Dr. Isaac Lawson, in this county, is reported to have first invented a practical method of extracting zinc from calamine, and to have established works for carrying out his invention. 1 Bishop Watson suggests that this Dr. Lawson may have been the man who went to China in order to acquire the art of extracting zinc ; but it is merely a suggestion which is entirely unsupported by evidence. According to the same author, zinc-works were first established at Bristol in about the year 1743, by Mr. Champion, who obtained a patent for his process. In 1739 2 a patent was granted to Mr. John Champion for a certain invention relating to metals, but it had no reference to zinc. Again, in 1758, 3 the same person obtained a patent for making spelter and brass from sulphide of zinc or blende, which was known by the names of " black-jack," "mock-jack," or "brazill:" it was directed that the mineral should be washed, purified, ground fine, and calcined ; and that the product should be mixed with charcoal and made into spelter, or used as a substitute for calcined calamine in the manufacture of calamine-brass. I find no patent of the date mentioned by Watson ; though there can be no doubt from the specification of that last was laden was termed calaem, with which he connects the word calamine. 7 Beckmann, op. cit. 8 Op. cit., p. 814. 9 Bergman, op. cit, p. 313. The pre- ceding statements in this paragraph are given on the same authority. 1 Price, Mineral. Cornub., p. 46. Cliam- bers's Diet., 1753, quoted by Watson, Chem. Ess., 4. p. 36. 2 Sep. 13. No. 569. 3 July 28. No. 726. ANALYTICAL EVIDENCE. 521 mentioned, that Mr. Champion had for some time previously been engaged in producing zinc on the large scale. In about the year ] 7<;<; Watson visited Mr. Champion's copper-works, near Bristol, and saw the process of making zinc, which at the time was kept rigidly secret. Many years afterwards he published an accurate description of this process, which was the same as that hereafter described as the English process, and which to this day, I believe, is still in operation at Bristol. Whatever doubt there may be as to the antiquity of the discovery of zinc, there can be none as to the fact that brass that is, a yellow alloy of copper and zinc was produced early in the Christian era, if not before its commencement. The proof of this fact is established by the analysis of objects of known date, and may be deduced from descriptions of the process of preparation in works of known date. Analytical evidence. Some years ago the Duke of Northumberland presented to the Museum of Practical Geology various coins which had been thrown aside from his collection as duplicates under the inspection of my friend Admiral Smyth. From these I selected for analysis a coin which had a characteristic brass-yellow colour. At my request the late Mr. Burgon, of the British Museum, examined this coin with particular care, and pronounced it undoubtedly genuine. It was a Greek imperial coin of Trajan struck in Caria about A.D. 110, and bearing the name of the magistrate Theodoras. Mr. Burgon was one of the highest authorities on coins of this class. Before the blow- pipe it yielded certain evidence of the presence of zinc. It was ana- lysed in my laboratory by Mr. Tookey, and found to have the following- composition : Copper 77-590 Zinc 20-700 Tin 0-386 Iron 273 98 749 In the collection at the British Museum I examined a coin of Geta struck at Mylasa, in Curia, between A.D. 189-212, which had unmis- takably the colour of true brass. Two other coins containing zinc, from the Northumberland collec- tion, were analysed in the Metallurgical Laboratory some years ago by Mr. T. Philipps. They were examined by Admiral Smyth, who has a high reputation for his knowledge of Eoman coins, and who has furnished me with the following descriptions of them. The first was a second brass coin of Vespasian, struck at Home A.D. 71. Ob- verse, the profile of Vespasian looking to the right. Reverse, inscrip- tion obliterated ; a draped female with a cornucopias and patera ; in the field S. C. (senatus consultu). In the collection of the British Museum I have inspected similar coins which are free from patina, and present the yellow colour of ordinary brass. The second was a large brass imperial Greek coin of Caracalla, struck A.D. 199. Obverse, CEOVHPOC ANTHNEINOC ; the laurelled head of Caracalla. Reverse, obliterated. 522 ZINC HISTORY. Coins of Severus in the British Museum, about 10 years earlier than that here referred to, have a decided brass-yellow colour. ANALYSES. No. 1. Copper 81-97 Zinc 18-68 Tin Lead.. 0-14 Iron .. 12 100-91 No. 2. 74-24 14-42 5-28 G-57 0-40 100-91 Mr. J. A. Phillips has published the following analyses, executed by himself, of ancient coins containing a considerable amount of zinc. 4 Two analyses of each coin were made. 1. Copper 82-26 Zinc 17-31 Tin Lead Iron 0-35 81-07 17-81 1-05 3. 83-04 15-81 0-50 4. 85-67 10-85 1-14 1-73 0-74 5. 79-14 6-27 4-97 9-18 0-23 99-92 99-93 99-38 100-13 99-79 Specific gravity... 8 "52 8-59 8-50 8 '30 No. 1. Large brass of the Cassia family, about B.C. 20 ; metal of a yellow colour. No. 2. Large brass of Nero, A.D. 60 ; reverse, Eome seated; metal of a bright yellow colour. No. 3. Titus, A.D. 79 ; metal yellow and soft. No. 4. Hadrian, A.D. 120; Fortune reduci ; finely patinated ; metal of a fine yellow colour. N.o. 5. Faustina, jun., A.I>. 165; Pietas; without patina; metal of a whitish colour and very brittle. The two following analyses of Roman coins containing zinc are by Genth. 5 Both were struck, and not cast, and both were very tough : 1. Copper 86 92 Zinc 10-97 Tin 0-72 Lead 1-10 Iron 0-18 Silver 0-30 Arsenic and antimony traces 2. 88-58 7-56 1-80 2-28 0-29 0-21 100-19 100-72 Specific gravity g'778 8'754 1. Of the reign of Hadrian ; bronze-yellow, but the colour of a fresh fracture inclining to brass-yellow ; very finely granular. This analysis agrees pretty closely with that of the coin of the same period by Mr. Phillips. 2. Of the reign of Trajan ; colour bronze-yellow, 4 Quart. Journ. of the Chem. Soc. of London, 1852, 4. p. 252 et scq. 5 Jahres-Bericlit for 1 859. 103. Wagner, p. DESCRIPTIVE EVIDENCE. 523 inclining to brass-yellow ; a fresh fracture greyish and very finely granular. Genth remarks that the alloys of which these coins con- sisted were, " without question," obtained direct from the ores, and that only the small quantity of tin was subsequently added ; but this is merely an opinion, which is certainly not beyond question. Gobel has published the following analysis of a Roman coin con- taining zinc in the Dorpat Museum : on the obverse is the head of Tiberius Claudius Caesar, and on the reverse the bust of Antonia Augusta. 6 Copper 72' 20 Zinc... , 27-70 99-90 When the so-called brass coins of numismatists contain a consider- able amount of zinc, like several of those of which analyses have been given, their colour, to an experienced eye, affords almost conclusive evidence of the presence of that metal. Through the kindness of my friend Mr. Vaux, of the British Museum, I have been enabled care- fully to inspect numerous brass coins in the collection under his charge. There are brass coins with the names of Augustus, Drusus, Agrippina, Caligula, and Nero, which have the characteristic colour of true brass. The earliest Roman imperial brass coin of the reign of Augustus has distinctly the colour of true brass. There is a spe- cimen in the British Museum of a piece of rolled metal, under the name of Orichalcum, which was obtained from the melting down of coins of Agrippina and Claudius. In colour and fracture it exactly resembles brass, and before the blowpipe it immediately yields the reactions of zinc. Mr. Vaux showed me a large brass coin of Tra- jan, struck for Cyprus. The date is A.D. 98-116; on the obverse is the emperor's head, and on the reverse a male figure standing, under- neath which is the inscription KOINON KvnpinN. The metal has the characteristic colour of true brass. Gobel asserts that zinc is only found in alloys of Roman origin, and that the bronze objects derived from the Greeks, and their colonies in Italy, Egypt, Asia, &c., invariably consist of copper and tin, or of these metals and lead. 7 Descriptive evidence. In order that this evidence may be appreciated, it is necessary to premise that, until a comparatively recent period, all brass was made by heating metallic copper imbedded in a mixture of calcined calamine and carbonaceous matter. This was effected in large crucibles, which were exposed to a suitable and long-continued heat in furnaces to be hereafter described. The zinc, immediately on its liberation, combined with the copper to form brass. The account given by Pliny of the metallurgy of copper is incom- plete, and, probably, inaccurate ; but it is certainly not more unsatis- factory, and cannot be more erroneous, than descriptions of metallur- f) Ueber den Einfluss der Chcmie auf I von Dr. Fr. Gobel. Erlangen, 1842, p. 29. die Ermittoluug der Volker dcr Vorzt-it | " Op. n't, p. 14. 524 ZINC HISTORY. gical processes which are occasionally published in this country at the present day. Much confusion has arisen from the ambiguous sense in which Pliny employs the word cadmia. It is applied by him to a kind of copper-ore (fit et e lapide asroso quern vocant cadmiam) as well as to certain products volatilized during the process of smelting copper. Of these, two distinct varieties are mentioned : one is described as white, and very light (pompholyx) ; and the other as a solid incrusta- tion upon the walls of the furnace, mixed with sparkling points, and sometimes, also, with charcoal (spodos). These descriptions would be quite applicable to the products formed in furnaces in which zincifer- ous ores are smelted, or in which any matters capable of evolving metallic zinc are sufficiently heated. The vapour of the zinc would be wholly oxidized, partly within the furnace near the mouth or top, and partly after it escaped into the air, and the result would be the forma- tion of a very light, flaky, white oxide, which would, as Pliny describes, adhere to the roof of the smelting-house, and the deposition, in the course of time, of a solid coating, consisting chiefly of oxide of zinc, upon the upper and internal parts of the furnace ; and this coating might be more or less crystalline, and present the appearance of spark- ling points. We have already seen that zinciferous copper-ores are at this day smelted in Sweden, and that a dense deposit, rich in oxide of zinc, is speedily formed on the sides of the upper part of the furnace. From the cadmia of the furnaces remedies for ophthalmic diseases were prepared, and compounds of zinc are still applied in the treat- ment of these diseases. The statements of Pliny concerning the forma- tion and medicinal uses of this artificial cadmia of furnaces are suffi- ciently corroborated by Dioscorides and other ancient authors. Now, whatever may have been the ore of copper which, according to Pliny, was termed cadmia^ to distinguish it from another ore of copper termed chalcitis, it seems extremely probable that the furnace- cadmia was similar in all respects to the zinciferous incrustations, t6rm.Qd.furnace-calamine, of modern furnaces : and if these two substances be not identical, I am quite at a loss to conceive of what the ancient furnace-cadmia could have consisted. My conviction is that they were identical, for in the smelting of any copper-ores free from zinc no such incrustation would be formed as Pliny describes. Granting this identity, the word cadmia, as used by Pliny, means both a particular ore of copper and furnace-calamine. But copper-ore is stated by him to furnish cadmia" metalla aeris multis modis iri- struunt medicinam * * * maxime tamen prosunt cadmia. Fit sine dubio ha3C et in argenti fornacibus, etc." 8 In this passage, from the reference to silver- furnaces, it is probable that cadmia refers to furnace- calamine derived from ores of copper, and, if so, it would follow that zinciferous copper-ores were smelted when Pliny wrote. These ores must have been either non-sulphuretted ores or sulphuretted ores. If s Lib. xxxiv. -l-l. DESCBIPTIVE EVIDENCE. 525 noii -sulphuretted ores, they probably consisted of carbonate of copper in admixture with carbonate of zinc or calamine ; and in the smelting of such ores in a small blast-furnace with charcoal, such as the Romans were accustomed to employ, brass, instead of copper, may have been directly produced. But as it would be impossible to prevent the tem- perature in the lower part of such furnaces from rising sufficiently high to cause the volatilization of much of the reduced zinc, a deposit of cadmia, or furnace-calamine, would be formed on the upper part of the furnace. On the other hand, in the case of sulphuretted ores, the zinc must have existed as blende, and the copper as copper-pyrites, or as some other sulphuretted ore of copper, when the usual process of roasting 9 must have preceded that of smelting ; but brass would, pro- bably, not have been directly produced, and difficulties would have occurred such as we have seen have only recently been surmounted in Sweden. Nevertheless this class of ores would have yielded a deposit of cadmia. In either case the copper-smelter would have had at his command metallic copper and cadmia, that is, a substance con- sisting chiefly of oxide of zinc ; and it would have been easy for him to produce brass from these materials by simply heating them together in contact with the fuel of the furnace. The following passage occurs in Pliny, and much stress has been laid upon it : " Hoc (ses) a Liviano cadmiam maxime sorbet et auri- chalci bonitatem iinitatur." l The word cadmia in this passage must refer to furnace-calamine : it cannot refer to the ore of copper termed cadmia, as the statement that copper absorbs copper-ore and produces a substance which is not copper, but orichalcum, a metallic substance distinct from, copper, would not be intelligible. In the sense of fur- nace-calamine, and supposing orichalcum in this instance to mean brass, the word cadmia is quite appropriate ; and the word "absorbs" exactly expresses what occurs in the preparation of calamine-brass, in wliich the copper may be truly said to absorb the zinc in proportion as it is reduced. In another passage concerning the production of spodos, a variety of fumace-calamine previously mentioned, it is stated that the Cyprian is the best, and that it is produced in the fusion of cadmia and copper- ore together " fit autem liquescentibus cadmia et aerario lapide." * Admitting cadmia to mean furnace-calamine, and the copper-ore to have been a non-sulphuretted one, brass would certainly have been obtained during this fusion, which was, doubtless, effected in furnaces with charcoal as usual. There is a third passage in Pliny which, in my judgment, affords strong corroborative evidence in favour of the conclusion that brass was made in his day. In describing the different kinds of copper and its mixtures he writes: "In Cyprio coronarium et regulare est, utrumque ductile ; coronarium tenuatur in laminas taurorumque felle 9 The process of roasting on wood is clearly referred to by Pliny in the state- ment concerning the treatment of the Capuan copper-ores. 1 Lib. xxxiv. 2. 2 Lib. xxxiv. 34. Lib. xxxiv. 20. 526 ZINC HISTORY. tinctum speciem auri in coronis histrionum preebet." a There were two kinds of Cyprian copper, both ductile ; one called " coronarium," and the other " regulare." The former was beaten into thin leaves, of which crowns, having the appearance of gold, were made for actors. The golden tint was produced by colouring the surface of the metal with ox-bile, which would act like a fine yellow lacquer and greatly enrich the colour of ordinary brass. Dr. Bostock remarks 4 " it is very improbable that the effect could be produced by the cause here assigned." However, that the golden tint mentioned by Pliny would, in a greater or less degree, be produced by the use of bile, I have no doubt ; nor do I conceive there would be any difficulty in causing the bile to adhere to the surface of the metal. It is stated that the German word for native carbonate of zinc, Gal- raei, and the English word calamine, are derived from cadmia, and, if so, traditional evidence is afforded in favour of the conclusion that the cadmia of Pliny and other early writers contained zinc. But as Pliny applied the term to a particular kind of copper-ore, differing from the ore termed chalcitis, it may be that at an early period a zinciferous variety of this ore happened to be smelted when the production of furnace calamine would be observed and that the specific name of the copper-ore became subsequently applied to the product obtained from it by volatilization. The orichalcum of the ancients has been the subject of much discus- sion. That this metallic substance differed from copper, and sometimes closely resembled gold in colour, there can be no reasonable doubt. Some writers have, if I am not mistaken, contended that all orichal- cum was an alloy of copper and zinc. It seems to me, however, most probable that the term orichalcum included various alloys of copper, of which brass was one. At the present day the terms brass and bronze are constantly used as synonymous, both by public writers and in ordinary conversation, yet they refer to alloys of copper essentially different, brass consisting of copper and zinc, and bronze of copper and tin. Thus, guns composed of gun-metal, which is bronze, are invariably designated as brass-guns. So, also, numismatists apply the term brass to coins which do not contain a particle of zinc, but consist wholly of copper and tin. With these facts before us it may easily be imagined that similar confusion in the use of terms prevailed in ancient times, and that orichalcum, like the modern word brass, may have been applied to alloys which in composition differed much from each other. Festus, who is supposed to have written between A. P. 100 and A.D. 422, 5 seems clearly to indicate that orichalcum was brass; for he 4 m ^^ 2Q \ i _>. 3 Ber/elius, Tr. de Ch. 2. p. DOT. 534 REDUCTION OF OXIDE OF ZINC of a vitreous lustre ; but if the tube is less strongly heated, the drops of metal become coated with very pretty small crystals of oxide. 4 When commercial sheet zinc is boiled in water from which air has been pre- viously expelled by boiling, hydrogen is very slowly evolved just below 100C. ; and on continuing the experiment during one or two hours the action decreases. This point was examined by A. Dick in my laboratory. The zinc employed was well cleaned with emery paper, and the hydrogen was collected in an inverted tube. Zinc heated with protoxide of lead. The following experiment was made (S.) : Quantities employed, in grains. Ratio of mixture. Zinc in fine powder. Litharge. Zn, 2PbO = 160 1120 The mixture was heated in a well-covered clay crucible to a strong red -heat during three-quarters of an hour : the products were a button of malleable lead weighing 245 grains, and a dark brown resin-like slag imperfectly melted in the centre, and containing shots of metal. The proportion of lead reduced is nearly the same as Berthier obtained in a similar experiment. 5 Zinc heated with the fixed alkaline carbonates. The carbonic acid is decomposed with the formation of carbonic oxide and oxide of zinc. Zinc heated with the neutral fixed alkaline sulphates. The sulphuric acid is said to be decomposed with the formation of oxide and sulphide of zinc. Zinc heated with carbonic acid, (S.) Dry carbonic acid was passed over zinc in a tube of hard glass heated to redness, when carbonic oxide was copiously evolved, and burned at the end of the tube from which it escaped. Brilliant, minute crystals of zinc were found deposited on the upper surface of the tube, each apparently consisting of an agglo- meration of crystals similar to those described by Nickles, and erro- neously believed by him to be pentagonal dodecahedrons. Reduction of oxide of zinc by carbon and carbonic oxide. The oxide is reduced by either of these agents at a strong red-heat ; and, conversely, metallic zinc at a red-heat reduces carbonic acid to carbonic oxide ; but, so far as is yet known, it has no action on carbonic oxide at any temperature, even the highest. In order that complete reduction should be effected by carbon, it is not necessary that the oxide of zinc and solid carbonaceous matter should be intimately mixed. On the large scale oxide of zinc in a greater or less degree of purity is always the compound from which zinc is directly extracted ; and carbona- ceous matter, such as charcoal, coal, or coke, is always employed as the reducing agent. Frequently both the oxide and reducing agent are in the state of coarse particles, and are very imperfectly mixed. The contact, therefore, between these two matters is very imperfect, and yet cqmplete reduction takes place ; so that it is certain that carbonic oxide must be actively concerned in the process. The operation is effected in large clay vessels. The interstices between 4 Begnault, Ann. des Mines, 3. s. 11. p. 16. 5 Tr. 1. p. 384. BY CARBON AND CARBONIC OXIDE BY HYDROGEN. 535 the particles are filled with atmospheric air, of which the oxygen becomes speedily converted into carbonic oxide, as there is always a large excess of carbonaceous matter. The oxide not in direct contact with solid carbonaceous matter is reduced by carbonic oxide, with the formation of carbonic acid, which is instantly converted into carbonic oxide by contact with the surrounding incandescent carbonaceous matter ; for otherwise the reduction of the oxide of zinc would speedily cease, as metallic zinc at the temperature at which it is liberated re- duces carbonic acid to carbonic oxide with the formation of oxide of zinc. The limited quantity of carbonic oxide existing in the first in- stance in the reducing vessel, acts as a vehicle by means of which the carbon is brought in contact with the oxide of zinc : it is perpetually being converted into carbonic acid and re-converted into carbonic oxide. A current, therefore, of carbonic oxide and of the vapour of zinc will continue to escape during the whole period of the reduction. If any reduced zinc should happen to be re-oxidized by the carbonic acid generated, the oxide of zinc formed, being light and finely divided, would necessarily be brought in direct contact with solid incandescent carbonaceous matter, and be immediately reduced. Redaction of oxide of zinc by hydrogen. Experiment by A. Dick, in my laboratory. A current of hydrogen, after having been exposed suc- cessively to the action of sulphuric acid, chloride of calcium, and hydrate of potass in fragments, was passed over oxide of zinc heated to redness in a tube of hard glass. At first the oxide seemed to volatilize, only a small quantity of metallic zinc being visible ; but on sending a good current of gas through the tube, reduction took place rapidly. However, a little of the reduced zinc was evidently again oxidized in the vapour of water formed, and was carried forwards as a white subli- mate. The metallic zinc condensed in small drops in the cooler part of the tube ; and although these drops must have consisted of compa- ratively pure zinc, yet they were found to dissolve with brisk efferves- cence in dilute sulphuric acid. Deville states that when 10 or 15 grammes of pure oxide of zinc are exposed to the action of a feeble current of pure hydrogen at a very high temperature in a porcelain tube, no reduction takes place : there is merely a change in the position of the oxide of zinc, which is deposited in other parts of the tube in well defined and occasionally large crystals. When, however, the oxide is exposed under precisely the same conditions to a rapid current of hydrogen, reduction occurs, and metallic zinc appears in the tube. In explaining these results, Deville entirely discards the so-called action of mass : " otherwise," he writes, " at the moment when the vapour of water and zinc react upon each other to reproduce oxide of zinc and hydrogen, this gas being predominant in the mixture ought, by reason of its mass, to protect part of the metal from oxidation. This never happens when the operation is conducted with sufficient slow- ness. But all is explained by admitting that variation in temperature causes the affinities to be reversed : thus, in the part of the porcelain tube heated directly, zinc in vapour and water may well coexist ; but in parts less heated, where the zinciferous deposit occurs, affinities ;>:)<; 1IKDUCTION OF OXIDE OF ZINC BY HYDROGEN. change, water is decomposed, and every trace of metallic zinc disap- pears. This takes place in my experiments : but when the hydrogen passes with rapidity, the zone of the tube in which the reverse action occurs is traversed by the mixture of the vapours with such rapidity that the cooling of the matters prevents the ulterior reaction." 6 Con- cisely stated, Deville's explanation is that at a given temperature hydrogen reduces oxides of zinc, and at a -lower temperature zinc decomposes water. Let us call the higher temperature a, and the lower temperature b. Now, if there be no action of mass, as Deville supposes, it would necessarily follow that at a oxide of zinc should be reduced by exposure to a small quantity of hydrogen in admixture with a large quantity of the vapour of water ; and that at 6 metallic zinc should decompose the vapour of water in admixture with a large quantity of hydrogen. But Deville has not adduced any experimental evidence to show that these results occur under the conditions sup : posed. That the action of mass should not be ignored in these phenomena of reduction seems evident from the experiments pre- viously recorded (p. 15), as also from the conclusive reasoning of Gay-Lussac on the reducing action of carbonic oxide. Although it is well known that affinities are affected by temperature, yet it appears to me that the phenomena of reduction for which Deville accounts by a change of affinity consequent on a change of tempera- ture are also perfectly explained by the difference in the relative pro- portions of hydrogen, the vapour of zinc and the vapour of water con- sequent on the difference in the velocity of the current of hydrogen in the two cases under discussion. Sulphur heated with oxide of zinc. It is stated that sulphide of zinc is formed, and sulphurous acid evolved. Iron heated with oxide of zinc. The oxide is reduced by iron at a high temperature, and the whole of the zinc is volatilized. Oxide of zinc heated with silica. The following experiments were made (S.) on the direct formation of silicates of zinc by strongly heat- ing intimate mixtures of oxide of zinc and fine Australian sand of great purity : Quantities employed in each experiment, in grains. Ratio of mixture. Oxide of zinc. Sand. 1. ZnO, SiO 3 = 200 230 2. 2ZnO, SiO 3 240 138 3. 3ZnO, SiO 3 = 240 92 4. 3ZnO, 2SiO* 180 138 Results. 1. The mixture was exposed in a Cornish crucible in a muffle during five hours to a strong heat at or near whiteness. The product was fritted. The frit was pulverized and reheated under precisely the same conditions as in the first firing. The second product, though more strongly fritted than the first, was yet not melted. 2. The mixture was heated under precisely the same condi- tions, and during the same time as No. 1. The product was fritted, 6 Aim. de Cliim. et de Phys. 1855, 3. s. 43, p. 479. SILICATE OF ZINC HEATED WITH CARBON. 537 but not so firmly as No. 1. 3. The mixture was heated under pre- cisely the same conditions, and during the same time as No. 1. The product was fritted, not quite so firmly as No. 1, but more so than No. 2. 4. The mixture was heated under the same conditions as No. 1 during 3^ hours. The product was slightly flitted, but readily crumbled between the fingers : it had a yellowish white colour. A portion from the centre of the frit obtained in the first three of the foregoing experiments was heated in a crucible of platinum foil during two hours in a muffle furnace ; the platinum crucible was placed within a covered clay crucible, and protected from contact with the walls of this crucible by supports of platinum wire. The heat was quite white, and the outer clay crucibles were softened and flattened. Similar experiments were attempted in Deville's furnace, when the platinum was melted, possibly from having been attacked by zinc reduced by the gases of the furnace. 1. No sensible change. 2. Fused, slightly porous, opaque, translucent at the edges, fracture vitreous in lustre, white, with a very slight yellow tinge. 3. Fused, of a greenish yellow tint, more translucent than, but in other respects similar to, No. 2. Native hydrated silicate of zinc, nearly pure, was perfectly melted by exposure in a crucible of platinum foil under the same conditions, when it was converted into a mass which was opaque, stony, and greenish grey in colour. Gelatinous silica was separated by the action of hydrochloric acid on each of these products, so that in every case silicate of zinc had been formed. In the Great Exhibition of 1851 were exhibited, in the French Department, specimens of beautiful colourless glasses, composed of silica, oxide of zinc, and alkaline or earthy bases. Silicate of zinc heated with carbon. The following experiments were made (S.) : (1.) 20 grains of artificial silicate of zinc of the formula 3ZnO, SiO 3 , produced as above described, were mixed in the state of powder with 5 grains of charcoal. The mixture was put into a small covered brasqued crucible, such as are used in the Swedish method of assaying iron ores, and this crucible was enclosed in a covered plumbago crucible, the space between the two being filled with anthracite powder. The mixture was exposed to a white heat during 1 hour. The residue in the brasqued cavity was very light, porous, and friable ; it occupied the same space as the original mixture ; it did not gelatinize with hydrochloric acid ; it was fused with mixed carbonates of potash and soda, and the product was carefully tested for zinc, but no trace was detected ; it contained, however, a little oxide of iron and carbon. Supposing the oxide of zinc to have been completely reduced and the zinc evolved, the residue of silica should have weighed 5 "54 grains, whereas the actual weight was G 3 ; but the difference may be ac- counted for by the oxide of iron and carbon in the residue. (2.) 20 grains of calcined nearly pure native silicate of zinc were mixed with 5 of charcoal, and the mixture was treated exactly like that in No. 1. The residue weighed <>"> grains, and after ignition 5*1, the 538 OXIDE OF ZINC HEATED WITH BORACIC ACID. decrease being due to the removal of a little carbon with which it had been mixed ; it did not gelatinize with hydrochloric acid, nor could a trace of zinc be detected in it by the method of testing adopted in No. 1. The experiment was repeated with exactly the same result. (3.) 50 grains of the same finely-powdered native silicate of zinc were treated under precisely the same conditions in a brasqued crucible without admixture with charcoal. The residue was porous, somewhat slagged on the exterior, and weighed 13*86 grains ; it was quite free of zinc. (4.) 130 grains of the same native silicate in small pieces, about as large as peas, were heated without admixture with charcoal in a brasqued crucible, under the same conditions as those above described'; the residual pieces were agglutinated at the edges, weighed 44 grains, and contained zinc. The conclusion to be drawn from these experi- ments is that oxide of zinc, in finely-divided silicate of zinc, may be completely reduced by exposure to a high temperature in a brasqued crucible, even without having been mixed with charcoal. Silicate of zinc heated with lime and charcoal, (S.) The native hydrated silicate from the United States was employed ; it was tested, and found to be nearly pure, containing only a trace of iron and lime. An inti- mate mixture of 50 grains of calcined native silicate of zinc, 50 grains of lime, and 20 of charcoal, was heated in a brasqued and luted crucible during an hour in an iron-assay furnace at the highest temperature. The residue gelatinized by the addition of hydrochloric acid, and evolved a small quantity of sulphuretted hydrogen, but contained no trace of zinc. Oxide of zinc heated with boracic acid. The following experiments were made (S.) : Quantity employed iu each experiment, in grains. Ratio of mixture. Oxide of zinc. Fused boracic acid. 1. ZnO, BO 3 = 200 175 2. 2ZnO, BO 3 = 240 105 3. 3ZnO, BO 3 = 240 70 4. 3ZnO, 2B0 3 = 240 140 The ingredients were intimately mixed and heated in a covered platinum crucible to strong redness in a muffle during about half an hour. In every case perfect fusion occurred, and file products, which were very liquid, were poured into an open iron ingot mould. Results. 1. The product solidified into a colourless, transparent glass, slightly opalescent on the external surface; the fracture was conchoidal and non-crystalline. 2. The product, when solid, was white, vitreous, crystalline, and translucent; the fracture was largely lamellar, and had a beautiful pearly lustre. 3. The product, when solid, was vitreous, pale yellow, and opaque in mass, but translucent in small pieces ; the fracture was less largely lamellar than that of No. 2, and had a more or less pearly lustre. 4. The product was a beautiful, perfectly transparent and colourless glass, presenting no trace of crystallization. By the addition of a small quantity of boracic acid, tribasic silicate SULPHIDE OF ZINC. 539 of zinc was rendered perfectly fusible, as will appear from the follow- ing experiment (S.) : A mixture of oxide of zinc, silica, and boracic acid was prepared according to the formula 3 (3ZnO, SiO 3 ) + 3ZnO, BO 3 . The proportions taken were oxide of zinc 240 grains, fine sand 69 grains, and fused boracic acid 17J- grains : this corresponds to 5*37 per cent, of boracic acid. The mixture was at first heated in a covered platinum crucible to redness in a muffle, when the product was found to be merely fritted. The frit was then heated under the same conditions to whiteness during half an hour : the product was perfectly melted, and solidified into a white, translucent mass, of which the fracture was largely crystalline, and had a pearly, or rather adamantine, lustre. Oxide of zinc heated with alumina, (S.) An intimate mixture of oxide of zinc and anhydrous alumina, in the ratio of 1 equivalent to 0, was heated in a small covered crucible of platinum foil, enclosed in a covered and luted clay crucible during an hour, in Deville's furnace. The outer clay crucible was softened and flattened. The mixture was agglutinated into a compact, grey, stony substance, which scratched flint-glass. The quantities operated on were 1*6 grain of oxide of zinc and 1 2 grains of alumina. Oxide of zinc heated with protoxide of lead. When oxide of zinc is healed with eight times its weight of litharge, a very liquid product is formed, which is crystalline like litharge, in very small laminae, opaque, and of a very pale yellow colour. But when it is heated with only six or seven times its weight of litharge, the product has a pasty consistency. 7 Oxide of zinc heated with the fixed alkaline carbonates.- Fusible com- pounds are obtained; but in order that they should be thoroughly liquid, a high temperature (50 Wedgwood's pyrometer) must be employed, and the oxide of zinc at most should not exceed of the weight of the mixture. The melted product is homogeneous, crystal- line, translucent, and colourless. 8 Oxide of zinc heated with cyanide of potassium. Oxide of zinc mixed with a large excess of cyanide of potassium was exposed in a tube of hard glass, closed at one end and drawn out at the other after the introduction of the mixture to the highest temperature of a blowpipe air-gas flame. The fused product had while hot the yellow colour of oxide of zinc ; no trace of vapour of zinc was observed to escape from the drawn-out end of the tube, (S.) Xiiljtldde of zinc : Formula ZnS. Precipitated from an aqueous solu- tion of a salt of zinc by sulphuretted hydrogen or a soluble sulphide, sulphide of zinc is an amorphous, white powder, which, heated in a close vessel, loses water, and acquires a pale yellow colour. It 7 Berthier, Tr. 1. p. 515. s Ibid., Tr. 2. p. 507. 540 SULPHIDE OF ZINC. becomes crystalline and agglomerated by exposure to a strong white heat, or at a temperature at which wrought iron may be easily melted ; but it may be said to be practically infusible. When a mixture of finely divided zinc and sulphur is projected into a red-hot crucible, combi- nation takes place with incandescence ; but much of the metal may be preserved from the action of sulphur by becoming coated with infu- sible sulphide of zinc. \Vhen a mixture of zinc turnings and cinnabar is suddenly exposed to a strong heat, detonation occurs like that produced by heating a mixture of a combustible substance and nitre, sulphide of zinc being formed, and the mercury reduced and volati- lized. At a lower temperature the greater part of the cinnabar sublimes. When a mixture of zinc filings or granulated zinc and polysulphide of potassium is heated, as soon as the latter melts, sul- phide of zinc is formed with violent deflagration. Blende appears to be somewhat volatile at very high temperatures without fusion. The following experiment was made (S.) : 200 grains of pulverized Laxey blende, free from matrix, were heated in a small Stourbridge clay crucible in Deville's furnace during li hour: anthra- cite was the fuel employed, and the temperature was extremely high. The crucible containing the blende was closed with a well-fitting cover, over which was placed an inverted crucible ; and the two crucibles thus arranged were put into a plumbago crucible, over which was also placed an inverted clay crucible. The blende became firmly agglutinated, though not fused, into one mass, and lost 9 grains in weight. The upper part of the interior of the crucible containing the blende and the under surface of the cover were coated with dark crystalline matter, of a somewhat metallic lustre ; but none was found in the interior of the inverted crucible immediately above. Precipitated sulphide of zinc was exposed in a small covered clay crucible, enclosed in a well-covered and luted plumbago crucible, to the highest temperature of an iron-assay furnace during an hour. At the upper part of the inner crucible, around the edges of the cover, a ring of matter was deposited perfectly crystalline, translucent, of a fine brown colour, and in lustre exactly resembling common blende : at the bottom of the crucible was a small somewhat porous mass, perfectly crystalline, of a light brown colour, and having the lustre of blende. There was a considerable space between this mass and the adjacent walls of the crucible, which were sensibly acted upon and fused into a dark brown glass. On the internal surface of the crucible above were numerous small clusters of minute, brilliant, brown, trans- lucent crystals, some of which appeared to be sharply defined hex- agonal prisms, with flat ends. This is the form of oxide, and not of sulphide, of zinc. Yet, according to Breithaupt, oxide and sulphide of zinc have the same crystalline form, and the so-called oxy sulphide of zinc, which occurs as a furnace product, is isomorphous with them. 9 In another experiment, in which Laxey blende in powder ZnS =Zn 3 + SO 2 -f- 5ZnS, the residue of blende should weigh 60 grains, i.e. the actual weight found. (3.) A mixture of 40 grains of oxide of zinc and 48 of sulphide, i. e. in the ratio of ZnO : ZnS, after having been heated in the same manner, gave a residue similar to that in No. 2, which weighed 22 grains. Supposing the same reaction to have occurred as in No. 2, the residue should have weighed 24 grains. (4.) A mixture consisting of 40 grains of oxide and 24 of blende, i. e. in the ratio of 2ZnO + ZnS, after having been heated in the same manner, gave a very light residue weighing 4-2 grains, from which only a trace of oxide of zinc was dissolved by heating with ammonia. In this experiment the oxygen of the oxide of zinc was just sufficient to form sulphurous acid with the sulphur of the sulphide. The covers of the inner crucibles in all these experi- ments were coated on the under surface with minute brilliant crystals of various shades of brown. The preceding results seem to establish the fact that, contrary to what has hitherto been asserted, oxide and sulphide of zinc mutually reduce each other at high temperatures, just like oxide and disulphide of copper at much lower temperatures. Nume- rous trials were made in the first instance with clay crucibles not lined with blende, but in no case could a decisive result be obtained, as the substance of the crucible was always much corroded. 2 Jahres-Ber. 1831, 10. p. 119. HEATED WITH VARIOUS METALS. 543 Sulphide of zinc heated with carbon. When sulphide of zinc is subjected to a white heat, either in admixture with charcoal or simply in a brasqued crucible, it wholly disappears, and is probably decomposed, with the formation of bisulphide of carbon. \Vhen the experiment is made with native blende containing iron, there is a residue of sulphide of iron perfectly free from zinc, (S.) Sulphide of zinc heated with various metals, (S.) Iron. Complete reduc- tion occurs at a bright red heat ; the zinc being wholly volatilized, and sulphide of iron formed. Tin. The following experiments were made : A mixture of 96 grains of pulverized Laxey blende and 116 of finely- granulated tin, i.e., in the ratio of 1 equivalent to 1, was exposed in a covered clay crucible to a bright red-heat during half an hour. The product consisted of a button of metal covered with regulus, which w;is easily detached from the former. The button weighed 14-5 grains ; it was harder than tin, yet flattened considerably under the hammer, but not without slightly cracking at the edges ; its fracture was crystal- line-granular, and had nearly the colour of tin ; it contained zinc. The regulus weighed 11 grains ; it was hard, brittle, finely-granular in fracture, and iron-grey in colour. Great volatilization, therefore, both of zinc and tin occurred, the former probably in combination with sulphur. The internal surface of the crucible was not acted upon, and there was no sign of permeation. The experiment was repeated with a mixture of 192 grains of blende and 232 of tin, i.e., in the same ratio as before ; the crucible was exposed to the same degree of heat as in the last experiment, but during a longer time, namely, three-quarters of an hour. The result was similar. The button of metal weighed 16'5 grains. The regulus could not be detached from the crucible and weighed. Both metal and regulus resembled those of the last experiment. Antimony. The following experiments were made : A mixture of 144 grains of Laxey blende and 120 of antimony, i.e., in the ratio of 3 equivalents to 1, was pre- pared by trituration, and exposed in a covered clay crucible to a bright red-heat during half an hour. The product consisted of an aggluti- nated and firmly-coherent mass ; it weighed 245 grains, so that the loss amounted only to 19 grains ; it was easily fractured, when the mass appeared to consist of a uniform mixture of particles of blende and a well-melted metal, resembling antimony; under particular directions of incident light, the fracture taking place in one plane, the lustre was brilliantly metallic. The whole of the internal surface of the crucible above the mass at the bottom, as well as the under-surface of the cover, was lined with minute metallic particles, partly globular and partly crystalline, over which were deposited here and there small radiating groups of minute white acicular crystals. The experiment was repeated with the same results. Hence, antimony does not appear to reduce sulphide of zinc, even at a high temperature. Lead. The following experiment was made : A mixture of 192 grains of pul- verized Laxey blende and 416 of finely-granulated lead, i.e., in the ratio of 1 equivalent to 1, was exposed in a covered clay crucible at a bright red-heat during three-quarters of an hour. The product was an 544 SULPHIDE OF ZINC. imperfectly-fused cavernous mass ; it was hard, brittle, largely crys- talline, dark lead-grey, and metallic in lustre. Not a trace of metallic lead was found. The whole of the internal surface of the crucible was considerably acted upon and melted into a brown porous substance ; the under-surface and edges of the cover were coated with metallic matter resembling galena in fracture. Copper. Two experiments were made, with the following proportions : Quantity employed, in grains. Katio of mixture. Laxey blende. Copper in powder. 1. ZnS + 2Cu = 48 64 2. ZnS + 4Cu = 48 128 The mixtures were exposed in covered clay crucibles to a bright red- heat during half an hour. Results. 1. The product consisted of a button of metal covered with regulus. The metal resembled copper in colour and other respects : it weighed 15 grains. The regulus resembled disulphide of copper, and weighed 67 grains. The internal surface of the crucible and cover was covered here and there with numerous small globules, similar in appearance to the regulus. 2. The product consisted of a button of metal and regulus. The metal resembled yellow brass in colour and fracture : it weighed 85 grains. The regulus resembled disulphide of copper, and weighed 71 grains. Hence, sulphide of zinc is reduced by copper at a high temperature. The proportion of zinc remaining alloyed with the copper will necessarily vary with the degree of heat and duration of the experiment. Sulphide of zinc heated in hydrogen. According to iBerthier it is not acted upon when heated in hydrogen. 3 Sulphide of zinc heated in the vapour of water. Eegnault obtained the following results : When native sulphide of zinc or blende is heated in a glass tube in a current of aqueous vapour, a small quantity of sulphuretted hydrogen is evolved ; but after the experiment had con- tinued during two hours, the blende had scarcely changed in appear- ance. "When, However, the experiment was carried on at a strong heat in a porcelain tube, decomposition was much more easily effected, arid the blende was almost completely desulphurized, oxide of zinc being deposited in small silky masses in that part of the tube from which the steam escaped. 4 Sulphide of zinc heated with carbonic acid, (S.) Dry carbonic acid was passed over blende in small pieces about as large as peas in a porce- lain tube heated to redness. The experiment was continued during 2 hours, and the gas was collected over mercury : it was wholly absorbed by a solution of potass, so that no reduction could have taken place. Possibly reduction may be effected at a much higher tempe- rature. Sulphide of zinc heated with protoxide of copper. The following experi- ments were made (S.) : the mixtures were exposed to a strong red-heat during half an hour in covered clay crucibles : 3 Tr. 2. p. 569. 4 Ann deg Mines, 3. s. 11, p. 46. SULPHIDE OF ZINC. 545 Quantity employed in each experiment, in grains. Ratio of mixture. Laxey blende. Protoxide of copper. 1. ZnS + 2CuO = 120 200 2. ZnS + 3CtiO = 144 360 Results. 1. The product consisted of a button of metal and regulus, covered with a ring of vitreous brown-black slag : the surface of the crucible in contact with this slag was corroded. The metal was like copper, had a dull fracture, and weighed 70 grains. The regulus resembled disulphide of copper in appearance, and weighed 106 grains ; it was tender, easily detached from the button of metal, finely granular in fracture, bluish-grey, and had moss-copper on its surface. 2. The result was similar. The button of metal weighed 201 grains, and the regulus 30 grains. Sulphide of zinc heated' with protoxide of lead. Both elements of the sulphide are oxidized with the formation of sulphurous acid, which escapes, oxide of zinc, and metallic lead. According to Berthier, sulphide of zinc requires to be mixed with not less than 25 times its weight of litharge, in order that the oxide of zinc formed may be per- fectly dissolved in the excess of litharge. When these proportions are employed, the resulting slag is vitreous, resin-brown, with an olive tint in colour, and translucent at the edges ; and 28 per cent, of pure lead is reduced. Berthier strongly heated a mixture of 24-08 gram, of blende and 55-78 gram, of litharge, i.e., in the ratio of 1 equivalent to 2 ; 29-2 gram, of very hard (aigre) lead of a greyish-black colour separated: it contained 1-8 per cent, of sulphur and 0-8 of zinc and iron. The button of lead was covered by a black, somewhat metallic matter, intermediate between a regulus and a slag, and which consisted of sulphides and oxides of zinc and lead. 5 Sulphide of zinc heated with peroxide of manganese. According to Ber- thier, decomposition takes place at a white heat with the formation of sulphurous acid, oxide of zinc, and protoxide of manganese. In attempting to determine this reaction by experiment we failed .to obtain satisfactory results on account of the corrosion of the clay cru- cibles employed. Sulphide of zinc heated with nitre or nitrate of soda. The action is ener- getic, both elements of the .sulphide being oxidized with the formation of oxide of zinc and sulphuric acid, which remains in combination with the alkaline base. Sulphide of zinc heated with carbonate of potass or soda. According to Berthier, at a red-heat chemical action takes place with effervescence, but there is no disengagement of metallic zinc : the product is well melted, homogeneous, opaque, and light brown. \Vhen carbonate of soda is heated with sulphide of zinc in the ratio of 1 equivalent to 1 , the product contains oxide of zinc, sulphide of zinc, and sulphide of sodium, and no sulphuric acid ; but as the sulphide of sodium in the product contains more sulphur than the monosulphide, a portion of the zinc would appear to be oxidized at the expense of the car- 5 Tr. 1. p. 403. 2 N 546 ZINC AND CAEBON ZINC AND PHOSPHORUS. bonic acid of the alkaline carbonate, with the formation of carbonic oxide. When the same experiment is made with the addition of char- coal, oxide of zinc is not produced, but an equivalent proportion of metallic zinc is volatilized. 6 Sulphide of zinc heated with lime. Berthier states that lime decom- poses sulphide of zinc, but only with the aid of carbon. The amount of zinc volatilized increases with the temperature. He exposed a mixture consisting of 6 '03 grammes of sulphide of zinc and 6'32 of carbonate of lime (i. 4, 125. 552 ENGLISH PROCESS OF EXTRACTING ZINC. coincide with that of the tube. This short tube is firmly fixed by means of the cross rod of iron, c, c, in the middle of which is a ring, as indicated by the shading ; at the ends of this rod are the iron rods, d, d, which pass through pieces of iron provided with screws, and let into the side walls. The lower tube is conical, and widens downwards ; its upper end fits on the outside of the lower end of a, and by giving it a twist round, it hangs suspended. At e is a ring of iron riveted on, which serves as a handle. The lower tubes frequently drop off, but may be immediately readjusted. Underneath the lower end of each long tube is placed a vessel of sheet-iron to receive the zinc as it drops. The complete adjustment of the tubes is shown in B, fig. 125. Reduction house. It consists of two stories, the upper one, A, in which the pots are heated, and the lower one, B, called the cave, fig. 125, Feet 12 Fig. 124. in which the zinc is condensed. The furnace is octagonal. Fig. 124, on the left of the line c, D, is a ground plan of half the furnace, and on the right of the same line is a plan of half the furnace on the line E, F, REDUCTION HOUSE. 553 A / Fig. 125. 554 ENGLISH PKOCESS OF EXTRACTING ZINC. in the elevation, fig. 125, which is the floor-level of the upper story. As the furnace on each side of the line c, r>, is in all respects similar, it was not considered necessary to represent more than half of it. The foundation in the cave is contained within a square, of which the side is 1 6 feet, b, b, are vertical piers of brick : c, c, c, are spaces by which the condensing tubes can be conveniently reached ; they are arched over near the top of the cave. There is one of these spaces for each set of tubes. From the floor of the upper story rise eight piers, of which four are shown at d, d, e, e. The fire-place extends across the furnace from g to g ; on each side of it is a flat bed of fire- brick on a level with the floor. In this bed are six holes, each 13 inches square, three on each side, /, /, /, which communicate with the cave below. Over these holes the pots respectively stand. The bed of this part of the furnace is supported by bars of iron, which pass from the upper part of the wall, a, a, to the opposite piers, b, b, fig. 124. On the left of the vertical central line in A, fig. 125, is shown half the furnace in elevation ; and on the right of the same line is shown the other half in section on a line passing through the middle of the fire-place .at right angles to its long axis. Externally, the piers d, e, e, d, incline inwards up to the plane, passing through the line i, K ; and from this plane springs the cone n, n, n, n, which becomes cylindrical near the top. This cone carries off the smoke of the furnace. The piers are united above by the arches m, m, m. Internally, the piers d, e, e, d, are vertical as high as the arches A, h, of which they form the abutments. Between every two adjacent piers is a similar arch, so that there are eight in all, of which those at the ends of the fire-place, g, g, are somewhat narrower than those on each side. From the piers on a level with the abutments of the arches, h, h, springs the flat dome or roof, i, i, which covers in the furnace ; the rise of this dome is only 9 inches. The spaces, g, g, under the arches at the ends of the fire-place, are closed with vertical walls of brick, in each of which is a fire-hole, z, and three smaller holes, a, a, a, which are stopped with bricks easily movable ; through these holes any cracks which may occur in the pots on the sides towards the fire-place may be filled up by plastering on clay with a long tool. The fire-hole has no door, and is kept stopped with coal piled up against it. Through the open spaces under the side arches, h, the pots are introduced and fixed, so that the centre of the hole at the bottom may coincide with the centre of the square hole in the bed ; the spaces are then closed by walls built of large rough pieces of brick, y, y, and in each of these walls are adjusted easily movable bricks, which are indicated by the shading In the elevation, and which are intended for a similar purpose as those at a, a, a. In the roof, or dome, i, and equidistant from the piers, .e, e, is a rectangular opening, o, on each side of which, resting on the dome, is a little vertical wall, p; these side walls, p, are connected at q by a transverse wall sloping forwards and outwards. Across o is placed a large flat quarry, r, A flue is thus formed as indicated by the arrow ; it may be closed more or less by moving the brick, , fig. 131 ; fig. 133, front elevation ; fig. 134, end elevation, in which is the fire-door. The fire-place is arched over; and in the arch are 12 rectangular openings into an arched chamber above, in which the retorts are heated, and which may be termed the retort-chamber. In fig. 131, the draughtsman, in order to make the construction more evident, has taken a liberty, and shown openings actually in this part of the roof of the fire-place, whereas they should have only been indicated by dotted lines. In front, the retort-chamber is closed by an iron door lined on the inside with fire-brick, and suspended and counterpoised as shown in figs. 131, 133. With proper care and under ordinary circum- CLAY NOZZLES OK CONDENSERS. 561 stances a retort will last seven weeks in the hottest part of the furnace, and from 10 to 12 weeks in other parts. Fig. 133. Front elevation. Fig. 131. Vertical section on the line AH, fig. 132. Fig. 132. Horizontal section on the line CD, fig. 131. Fig. 134. End elevation at the fire-door. Clay nozzles or condensers. In the upper part of the mouth of each retort, and resting on a narrow brick supported by the bridge-pieces, is fixed an elbow-shaped clay nozzle, /?, of the form and dimensions shown in fig. 129 (1, 3). There is an opening, 46 { 2-18J 12 ' G1 Refractory matter bricks, clay, &c 1-80 1-83 Bur-iron, cast-iron, &c 0'34 0'32 Repairs 1-36 1-04 General expenses 1'25 1-31 Cost of production 34-56 35*36 Do. per ton 14 12s. 7$d. 14 19s. 4|d It is stated that the slight increase of cost in 1858 was caused by diminished produce of the ore. The difference in the cost of production at the Lydognia Works and those of the Silesian Company is more apparent than real. This Com- pany has its own mines of calamine and coal, and the amount charged for these materials is exactly what was expended in raising them, &c., from the mines ; whereas in the account of the cost of production at the Lydognia Works they are charged at their selling pi~ke at the time. In the latter account, coal is charged at the rate of 7-3 francs per 1000 kilogrammes, while in that of the Silesian Company it is estimated only at 5-61 francs. In like manner the calamine at the Lydognia Works, though con- siderably richer, and yielding 16'92 per cent, of zinc, instead of 14-93, is estimated at 20-71 francs per quintal of the zinc produced, or 14-25 francs higher than at the Silesian Company's Works. By deducting this excess in the price of coal and calamine (3-28 + 14-25 = ), 17-53 from 54'84 the cost of production per quintal of zinc at the Lydognia Works we have the cost of production reduced to 37'31 francs, an amount quite comparable with that of the cost of production at the Silesian Company's Works. The excess of expen- diture, 2-75 francs, at the Lydognia Works, is made up of the following items : Labour, 0-82 franc ; repairs, 0-19 ; and general ex- penses, 1-74. 2 P 578 BELGIAN PROCESS OF EXTRACTING ZINC. THE BELGIAN PROCESS OF EXTRACTING ZINC. This process was first practised in the vicinity of Liege in the early part of the present century, but in what year I am unable to state. The ore formerly treated was exclusively calamine, of which large depo- sits once existed in the vicinity. The construction of the furnace and the distillatory apparatus have remained substantially the same to the present day, though both have been variously modified at different works. I nave examined and compared many of the descriptions of this process ; and I. have had placed at my disposal sketches of one of the most approved furnaces of the Vieille Montagne Company, with all neces- sary measurements. 9 Extensive zinc-works, containing both Belgian and Silesian furnaces, have within a short period been erected near Swansea by Mr. Hussey Vivian. Tn 1848 I saw English, Silesian, and Belgian zinc-furnaces in the works then carried on by the late Mr. Vivian ; but only the first were in operation. The engravings in illustration of the process in this work have been executed from working drawings of one of the largest Belgian furnaces yet erected, and I have reason to believe they are correct. Reduction is effected in vessels or retorts of refrac- tory clay, closed at one end and open at the other, which is termed the mouth. In the mouth is adjusted a clay-nozzle, or condenser, on the free end of which is fitted a vessel of sheet-iron. Retorts and appendages. The retorts are cylindrical, e, e, e, figs. 146, 148. Fig. 146. Frunt Elevation. The furnace is shown as only partially filled with retorts ; the shaded rectangular spaces indicate the empty compartments. 9 By M. Eckart. to whom I have plea- sure in acknowledging ray obligation. Thum has recently published a descrip- tion of the works of the Vieille Montagne Company, of which I have availed myself. Ueber den Zinkhiittenbetrieb der Alten- berger Gesellschaft. Von Ingenieur W. Thum, Berg. u. hiittenm. Zeit. 1859-1860. 11ETORTS AND APPENDAGES. 579 They are composed of the same materials, and fashioned in moulds constructed on the same principle, as those described under the Silesian process. The clay-nozzles, or condensers, have the form and dimen- sions shown in fig. 148, i, i, i, &c. : their transverse section is circular. On the outer ends of "these nozzles are placed sheet-iron vessels, or 5T UNIVEESIT Fig. 147. Vertical Section on the line A B, fig. 150. receivers, of which the form and dimensions are also shown in fig. 148, /, ?, I : in the centre of the closed, or wide, ends of these receivers is a small hole, m. But it has, I believe, been found that this condensing apparatus may be replaced with advantage by the simple clay condenser in use at Stolberg, and which has already been described. In replacing 2 P 2 580 BELGIAN PKOCESS OF EXTKACTING ZINC. old retorts by new ones, the same precautions must be observed as in the Silesian process. Description of ihz furnace. The retorts are arranged in parallel rows one above another in a narrow vertical arched chamber, along the bottom of which is the fire-place. Fig. 146 is the front elevation ; fig. 147 is a vertical section on the line A, B, fig. 150 ; fig. 148 is a vertical sec- tion on the line c, D, fig. 150; fig. 149 is the end elevation at the fire- place door ; fig. 1 50 is a horizontal section im- mediately above the fire-bars on the line E, F, fig. 148. In front is a series of cast-iron plates, or shelves, inclin- ing slightly forwards and downwards, a, a, a, &G., figs. 146, 148. On these plates, at regu- lar distances, fire-brick slabs are placed on edge vertically and. at right angles to the long axis of the fire-place, 6, b, b, figs. 146, 148. On the upper and posterior F edges of these slabs, where they extend be- hind the iron plates, other slabs of fire-brick are placed nearly hori- zontally, c, c, c, &c., fig. 148 ; the anterior edges of these slabs rest on the posterior edges of the iron plates, in the man- ner shown in fig. 148. A series of rectangular compartments is thus Fig. 148. Vertical Section on the line C, D, fig. 150. formed, within each of which the mouths of two retorts are contained. In the back wall of the furnace is a series of ledges, upon which the closed ends of the retorts are supported, (/, d, d, &c., figs. 147, 148. The retorts incline downwards towards the front, e, e, e, fig. 148. By this inclination any corrosive slag which may be produced during the working of the furnace will flow forwards from DESCRIPTION OF THE FURNACE. 581 the hottest towards the coolest part, where it would be less injurious. Clay luting is carefully applied round the mouths of the retorts, so as to prevent as completely as possible the passage of air through any pail of the front of the furnace. In the arch /, /, figs. 14G, 1 47, 148, are three flues, g, g, g, communicating with the stack, h. The clay condensers, z, z, i, fig. 148, are luted into the mouths of the retorts, and supported in front by bricks, 7i, 7e, k. Immediately above the fire, and resting on the walls of the fire- place, is a row of thick empty retorts, o, fig. 148, which are designed to protect the next row of retorts from the too great heat of this part of the furnace. Below the floor, in front, is a pit, n, fig. 148, to receive the residues from the retorts. All the in- ternal parts of the furnace are constructed of good fire-brick. The furnace is securely clamped and braced by iron standards and tie -rods. Furnaces are usually built together in groups of four, with a central stack. 1 The 'furnace represented in the accom- panying engravings differs considerably, in proportions and in magnitude, from that at Moresnet, described and figured by Thum. The former contains 78 retorts, exclusive of the empty ones in the lowest row ; and is about 11 feet wide, 9 feet 6 inches high from the floor to the centre of the arch, and 4 feet from front to back, inside measure ; whereas the latter contains only 61 retorts, exclu- sive of the empty ones in the lowest row, is about 8 feet wide, 10 feet G inches high, and 6 feet 6 inches from front to back, inside measure. The retorts in the former are ;> feet G inches long and 8 inches in dia- meter, outside measure ; and those of the latter about 3 feet 3 inches long and 8 inches in diameter, outside measure. The clay-nozzle and sheet-iron vessel employed at Moresnet diifer somewhat from those shown in the accompanying engravings. In the Moresnet furnace the three uppermost shelves (a, a, &c.) are entirely of cast-iron. In 1857 an attempt was made at Aiigleur to collect the fume which, in spite of the sheet-iron vessels on the nozzles, escapes into the atmos- phere. A hood of sheet-iron was suspended over the front of the furnace, and connected with the stack by a wide sheet-iron pipe. The result Fig. 149. End elevation at the fire- place door. 1 The apparatus used in charging is exactly similar to that previou under the Silesian ** 582 BELGIAN PROCESS OF EXTRACTING ZINC. was not successful : no sensible amount of fume was recovered, and the men suffered much from the heat radiated upon their heads. The following details concerning the working of the Belgian fur- naces of the dimensions stated above, are extracted from Thum's de- scription : The ore employed at Moresnet is calamine, which is calcined in circular kilns 2 m 20 (7 ft. 2 in.) in diameter at the mouth ; 2 m 95 (9 ft. 8 in.) in diameter at the widest part, which is 2 m 14 (7 ft.) below the mouth; and l m 70 (5 ft. 6 in.) in diameter at the bottom, where four openings exist for drawing out the calcined ore. The bottom is conical, and the height of the cone is l m 05 (3 ft. 5 in.) ; and the height of the kiln, from the point of the cone to the mouth, is 5 m 35 (17 ft. 6 in.). Fig. 150. Horizontal Section on the line E, F, fig. 148. The ore is broken in pieces of about the average thickness of O m 15 (6 in.). The kiln is charged with alternate layers of inferior non-caking coal and ore. Calcined ore is drawn from 4 to 6 times in 24 hours, during which period 25,000 kil. (about 25 tons), on the average, are passed through the kiln, with a consumption of from 3 to 4 per cent, of coal. The small ore which is obtained in the process of dressing is calcined in reverberatory furnaces. At Moresnet, large furnaces with double beds, one over the other, are employed for the purpose, as well as smaller furnaces heated by the waste gases of the reduction furnaces, and built on the top of the latter. The construction of the double- bedded furnaces is similar to that of the calciner described at p. 382. The beds are slightly elliptical on the sides, 5 m 00 (16 ft. 5 in.) long, and 2 m 20 (7 ft. 2 in.) broad in the .widest part. The ore is first dried on the top of the furnace, and then allowed to fall upon the upper bed, where it is heated and carefully stirred during 6 hours, after which it is transferred to the lower bed, where it is again calcined during the same length of time. About 8000 kil. (8 tons) of calamine are calcined in one of these furnaces in the course of 24 hours. Both the clay and the calamine are ground under edge-rolls of cast- iron revolving in pairs over beds of cast-iron plates. The rolls are l m 60 (5 ft. 3 in.) in diameter, and about O m 35 (about 14 in.) in breadth ; DESCRIPTION OF THE FURNACE. 583 they weigh 3200 kil. (about 3 tons 2 cwt.) each. Under a pair of such rolls, from 15,000 to 18,000 kil. (about from 15 to 18 tons) of calcined calamine or clay may be ground in the course of 12 hours, and with a consumption of about 3-horse power. At Moresnet the retorts are made by hand : one man will make from 18 to 20 in 12 hours. After having been dried and burnt ready for use, they cost on an average, in 1857, from 2 fr. (Is. Sd.) to 2 fr. 10 c. (Is. 9d) each. At the Vieille Montagne Company's Works at Angleur and St. Leonard in Belgium, the retorts are made as follows : A cylindrical mould, divided longitudinally into two equal parts, which may be clamped firmly together in the usual manner, is rammed full of clay and placed under a boring machine, when the interior is bored out. With two machines of this kind about 2(50 retorts may be made in 12 hours. The price of a retort, dried and burnt ready for use, is 1-eduged to 1 fr. 60 c. (Is. 4d.~). One man will make from 100 to 110 clay nozzles or receivers in 12 hours ; and the average cost of these, in the year above mentioned, was from 14 c. to 16 c. (from 1 T V/. to l T %d.) each. Best fire-bricks, in the same year, cost 55 fr. 19 c. (about 21. 6s.) per 1000 kil. (1 ton). The finely-ground calamine, which should contain about 50 per cent, of zinc, is moistened with water, and intimately mixed with car- bonaceous matter ; and with this mixture the retorts are charged twice daily once in the morning and once at night. The carbonaceous matter consists of the most non-caking coal-slack, which may advan- tageously be mixed with half its weight of cinders or coke-dust : these materials should be as free from sulphur as possible. Before charging a retort, the nozzle is detached and the residue in the interior scraped out, the slaggy matter adherent to the inner surface of the retort being at the same time 'removed as completely as practicable. The matters thus drawn out fall into the pit n, fig. 148, from which they are subse- quently conveyed away. The*workmen, in charging the upper retorts, stand on a moveable platform with steps. After charging, the nozzle, smeared over with clay, is adjusted into the mouth of the retort, and a brick, k, fig. 148, is pushed under it. The sheet-iron receivers are placed on the nozzles when the vapour of zinc begins to appear. After the lapse of 6 hours most of the nozzles will be filled with melted zinc, which is withdrawn by means of a little scraper. This process is repeated before charging afresh. Cracked or perforated retorts must be replaced by new ones, which have been previously gradually heated to redness in an adjoining kiln. This operation generally occurs every morning. The lower rows of retorts, being exposed to a higher temperature than the upper ones, are more heavily charged than the latter. The retorts of the lower half of the furnace generally receive from 12 to 12J kil. (26 to 27J Ibs.), and those of the upper half from 6 to 7 kil. (13 to 15 Ibs.), each. Less rich ores, and such as contain much basic matter like oxide of iron, which would attack the substance of the retorts at a high temperature, are generally introduced into the latter. Zinciferous matters, from which the metal may, on account of their rich- 584 BELGIAN PROCESS OF EXTRACTING ZINC. ness, be separated at a gentle heat, are put into the upper retorts : they consist chiefly of the residues obtained in the treatment of zinc fume in the Montefiori furnace, and the crusts of oxide of zinc which are detached from the nozzles, and which in the course of time may accumulate to such an extent as to obstruct the passage and render them useless. Before using the nozzles it is customary to smear them over with milk of lime, with a view of causing these crusts a more easy separation. At Moresnet, in a furnace such as has been described, 1200 kil. (about 1 ton 2 cwt.) of ore are treated in 24 hours ; and the average yield of zinc, from ore containing 50 per cent., is from 450 to 470 kil. (from 922 to 963 Ibs.), inclusive of that in the fume produced, and of which the greater part is obtained in the state of ingot by simple fusion in the Montefiori furnace. The actual yield of zinc from the ore is from 38 to 39 per cent. In the following table by Thum are presented the average results of the working of one of the Moresnet furnaces during 24 hours, calcu- lated from the actual results obtained during the first four months of 1857 : Charge. Produced. Total yield of zinc % Fuel and materials consumed. Calamine. Carbon aceous reducing matter. Zinc. Zinc fume. Coal. Retorts. Nozzles. Fire-bars. kil. kil. kil. kil. kil. Number. Number. kil. 1035 549 371 37 39-0 1479 3-8 11-5 0-3 Thus the amount of coal consumed (supposing the 549 of reducing agent to have consisted entirely of coal) for 1 ton of zinc is 5 -46, or very nearly 5| tons ; and only 3-98 tons, exclusive of the coal used in reduction. But this does not include the coal consumed in calcination of the ore, and that used in the 'preparatory heating of the retorts ; yet the amount seems very small. In large zinc-works in this country, in 1859, the amount of coal con- sumed as fuel, in the extraction of 1 ton of zinc by the Belgian process, was stated to me to be 6i tons ; but I think it must have exceeded this amount. I am informed (Nov. 1861) that the weekly production of zinc at one large establishment in this country is about 100 tons : both the Silesian and Belgian methods are employed ; the cost of production by the former process is about 171 per ton, and by the latter process about 18?., inclusive of the cost of ore, &c. From the preceding data it appears that with ore containing 50 per cent, of zinc, there is an average loss of zinc of from 11 to 12 per cent, ; but of this at least half is recovered from the residua, the other half being irrecoverably lost, partly by volatilization from the nozzles, partly by the escape of zinc from bad retorts, and partly by imper- fect reduction and the consequent loss of zinc in the residua from the retorts. Supposing that on an average not fewer than 3 retorts have daily to be replaced by new ones, and that each retort receives a CARTNTHIAN METHOD OF EXTRACTING ZINC. 585 |ui\ SITI charge of 10 kil., then the constant loss of zinc by volatilization might be estimated, according to Thum, at 1-2,5 per cent. Each furnace requires 4 men 2 head-men and 2 subordinates of whom 2 are on duty at a time during a shift of 12 hours. The men receive a fixed sum in wages, and, in addition, an amount dependent on the zinc produced. At the end of every month, in addition to their fixed wages, they receive only half the amount which they may have earned in excess: the other half is not paid to them until 15 months afterwards, t. e. a quarter after it has become due. By this arrangement the men are stimulated by a regard for their own interest at all times to produce the greatest yield, to take care of their furnaces, and to retain their situations. CARINTHIAX METHOD OF EXTRACTING ZINC.* This process was formerly practised in Carinthia and Hungary, but whether it has been entirely abandoned I know not. Reduction was effected in clay pipes or retorts set vertically in an arched chamber, heated by the flame of a wood fire, which I will call the reduction chamber. These pipes were 40 inches long, 4 inches in diameter in the clear at the top and 3i at the bottom, and from i to f inch in thickness. The narrow end of .each pipe was fixed in a foot-piece of clay, consisting of a short vertical tube having at the top a square flange, also of clay. The flanges of these foot-pieces formed the bed of the reduction chamber ; they were supported by a trellis of iron bars, and packed close together like the pieces of a mosaic pavement. The , pipes being duly charged with the mixture of zinc- ore and carbona- ceous matter, and closed at the top or wide end, the entrance into the reduction chamber was stopped and the fire lighted. The vapour of the zinc passed downwards, as in the English process, through the foot-piece, and condensed in drops on sheet-iron placed about 12 or 14 inches underneath to receive it. AVhen the charge was worked off, the furnace was allowed to cool, in order that workmen might enter the chamber and prepare for the next charge. Thus the furnace did not work continuously, and there must, consequently, have been much waste of heat and loss of time. Both calamine and blende were employed. The calamine was calcined with wood in open kilns con- sisting of four walls, and was afterwards finely stamped or ground and sifted. The blende, after having been freed as completely as pos- sible by hand from pyrites and vein-stuff, was roasted with wood and small charcoal in open walled kilns. The roasted mass was washed with water in order to dissolve out any white vitriol which it might contain ; after which it was reduced to fine powder and roasted sweet in a reverberatory furnace: this required from 8 to 12 hours, with frequent stirring. The roasted blende was mixed with charcoal-dust MI id lime in fine powder, as obtained by slaking and reheating, and the whole was thoroughly kneaded with water containing in solution 2 Tagebuch einer metallurgisch-tech- I men, etc. Von Christian Fiirchtegott nologischen Keise, (lurch Miihrcn, Boh- Hollimder. Niirnberg, 1824, p. 373. 586 CARINTHTAN METHOD OF EXTRACTING ZINC. common salt and the soluble matter of wood-ashes. The actual quan- tities employed were as follow : For ore sufficient to charge 336 pipes, 14 cubic feet of solution of wood-ashes, 26 Ibs. of common salt, and 76 Ibs. of lime. It was estimated that the solution contained about 4 Ibs. of potash. The potash and the salt served the purpose of glazing the internal surface of the tubes and of thereby rendering them impervious to the vapour of zinc. Small pieces of charcoal of the size of hazel-nuts were mixed with the mass in order to loosen it and pro- mote reduction. Each pipe contained from 5 to 6 Ibs. of ore. The pipes, after having been properly burnt, were filled by means of a small shovel to within about 4 inches of the top or wide end with the mix- ture prepared as above described. Into the empty space thus left small pieces of charcoal were introduced, over which pieces of charcoal dipped in clay were placed crosswise so as completely to close the tube at this end. The charge was prevented from dropping down through the lower or narrow end by pieces of charcoal. A portion of the reduction chamber was set apart for burning unburnt pipes to be used in a subsequent operation. A single furnace at Dolach contained 84 pipes charged and 51 empty. At Dognacska in Hungary there were not fewer than 256 pipes in a single furnace. 3 The furnaces were built either in pairs or four together. After working oft' a charge the furnace was left to cool during 1 or 2 days, after which the whole of the pipes were taken out ; the bad were replaced by freshly burnt ones, and the rest were used again. The furnace was lighted at 9 o'clock in the morning, and at 7 in the evening it began to yield zinc. At noon on the second day zinc was copiously produced ; but the furnace was not considered as in proper working order until about 10 o'clock in the evening of this day. The whole time of distillation generally lasted from 30 to 40 hours. In extracting 1 centner of zinc at Dolach 180 cubic feet of wood were consumed. The selling price of zinc (1 824) was 31. 6s. Sd. the centner (Bavarian), that is, between 601. and 707. per ton. ZINC FUME. In the extraction of zinc from its ores a considerable quantity, as we have seen, of fume in the state of very mobile grey dust, is collected in the sheet-iron vessels attached to the nozzles or condensing-tubes in the Silesian and Belgian processes. It consists essentially of very finely-divided metallic zinc, intimately mixed with oxide ; but it is also stated that certain foreign metals of frequent occurrence in zinc-ores namely, cadmium, arsenic, lead, and (accord- ing to Thorn) antimony become concentrated in it. Thum has published the following analysis, by himself,' of the fume from the Silesian furnaces at Borbeck, in which calamine from Wies- loch is treated in admixture with many varieties of blende : 4 97-82 Iron .......................................... 0-16 Lead ........................................ 0'23 Cadmium and arsenic ..................... 0'08 98-29 The proportion of oxygen in combination with the metals is not given. 3 Karsten, Sys. 4. p. 479. * Berg. u. h. Zeit. 1859, p. 409. THE MONTEFIORI FURNACE. 587 In one place Thum states that the furne is pyprophoric, and, in another, that it takes fire in moist air and burns to oxide. 5 When fresh fume is heated to the melting-point of zinc, without access of air, and then subjected to pressure, the finely-divided particles of zinc unite into a liquid mass, leaving from 8 to 10 per cent, of oxide of zinc. Accordingly, a process (to be described presently) for the extraction of zinc from fume is founded on this principle. The excessively minute particles of metallic zinc in the fume would appear to be kept from uniting during condensation by the interposi- tion of a small quantity of oxide of zinc. It is easy to conceive how the latter may be generated by the action of the vapour of zinc on any carbonic acid which may escape along with it and the carbonic oxide from the retorts. The reduction of carbonic acid to carbonic oxide might thus be effected by the metallic zinc present, with the formation of an equivalent proportion of oxide of zinc, either by an alteration of temperature or a variation in the relative proportions of the carbonic acid, carbonic oxide, and vapour of zinc changes of condition which may be supposed not only possible, but likely to occur. Thum attributes the formation of this oxide to the action of the vapour of water evolved from the coal used as the reducing agent ; but the production of zinc fume takes place equally, I believe, when no water can be present in the retorts. Oxidation may be caused in a certain degree by the presence of atmospheric air in the condensing apparatus ; yet, if I mistake not, and I am by no means sure on this point, fume will be formed even when atmospheric air has been perfectly excluded from this apparatus. It has been recommended to employ the grey fume as paint, espe- cially for the copper bottoms of ships ; but it is generally utilized either by redistilling it in admixture with carbonaceous matter in the reduc- tion furnace, or by adding it to the charge of ore prepared for reduction, or it may be subjected to special treatment in what is termed the Montefiori furnace. If the first plan be adopted, impure zinc is apt to be produced on account of the concentration of foreign metals in the fume ; and the second plan is consequently preferred, as the influence of these foreign metals is rendered insensible by their diffusion through a large mass of zinc. The Montefiori furnace* The vessel or pot in which the fume is heated is a hollow cylinder, closed at one end (the bottom) and open at the other (the mouth). The depth of the cavity is about 27 inches, and its diameter between 7 and 8 inches. On one side, at the bottom, is fixed on at right angles a tube, of which the bore widens upwards towards the junction of the vessel. Both the pot and the projecting tube are made of clay. Two parallel rows of these pots are set vertically on their closed ends in a rectangular chamber, having a partition-wall between them. Underneath the bottom of this chamber, which is flat, is a fire- place with an arched roof, extending in a direction parallel with the two rows of pots above. Flues pass through the roof of the fireplace into the chamber, which by this means is equably heated. The ends of 5 Op. cit., p. 408, 409. 6 Patented in this country by Jacob Levi Elkin, A.D. 1856. No. 270(5. 588 FOREIGN MATTER IN COMMERCIAL ZINC. the projecting tubes of the pots protrude through the outer or side walls of the chamber. The roof of the chamber is flat, and contains openings corresponding to the mouths of the pots, which remain open, and are thus easily accessible from the outside. The pots are provided with cylindrical clay pieces, which fit into them like pistons ; and to these pieces are firmly attached iron rods extending upwards considerably above the mouths of the pots : these rods pass through openings in bars of iron, fixed over the pots at a con- venient height above the roof of the furnace ; and on their upper or free ends suitable handles, with four arms placed crucially, may be fastened, so that the clay pistons may be easily pressed down and made to receive a circular motion. The furnace being heated, and the ends of the projecting tubes having been stopped with clay, zinc fume is put into the pots, and the pistons of clay are pressed down upon it. After gently heating during 2 or 3 hours, and exerting pressure with the communication of an occasional rotatory motion to the pistons, the zinc is obtained in a liquid mass at the bottom of the pots, and may be withdrawn by un- stopping the projecting tubes at the bottom. In each pot about 20 kil. of zinc fume may be thus treated at a. time ; so that in a furnace con- taining 12 of these pots, from 700 to 900 kil. of fume may be treated in the course of 12 hours. The zinc separated directly by this process is stated to be quite as impure as that obtained by the distillation of the fume per se. At Borbeck, where coal containing chloride of sodium is used, chloride of cadmium, in fine transparent crystals, has been observed to be de- posited in very sensible quantities about the orifice of the projecting tubes in the Montefiori furnace. 7 A specimen of fume which I collected on the arches near the con- densing tubes of the English zinc furnace, which I saw in operation in 1859, has been imperfectly analysed in my laboratory by Mr. Weston. Its colour was dirty brownish grey. Blende was the ore employed. The following results were obtained : Oxide of zinc 46*70 Oxide of cadmium 16 23 Sesquioxide of iron 4'45 Alumina 0'75 Lime ; 1-43 Sulphur 0-82 Sulphuric acid 2-85 Organic matter (carbonaceous) 10-01 Hygroscopic water : 1-98 Insoluble residue undetermined . 7 '94 93-16 Unaccounted for 6'84 100-00 FOREIGN MATTER IN COMMERCIAL ZINC, AND ITS INFLUENCE ON THE WORKING QUALITIES OF THE METAL. Commercial zinc is never pure, and varies much in quality. Karsten has ascertained the effect of various metals on zinc, " Thum, op. cit., p. 409. FOREIGN MATTER IN COMMERCIAL ZINC. 589 and has also investigated the composition of different kinds of commer- cial zinc. 8 Sulphur. Karsten never detected a trace of sulphur in any of the numerous specimens of zinc which he examined. According to Dr. Alfred Taylor, commercial zinc generally contains sulphur. A minute quantity of sulphur was detected in every specimen of the various kinds of commercial zinc examined by Eliot and Storer. Arsenic. Karsten only sought for arsenic in varieties of zinc from Silesia, which could not be rolled without cracking; and he was not able to detect a trace. Schauffele states that he found arsenic in several kinds of commercial zinc, and determined the amount by Jacquelain's method, which consists in passing the hydrogen evolved by the action of dilute sulphuric acid upon the zinc through a solution of gold. His results are as follow : 9 Arsenic in 1000 parts of zinc. Silesian zinc contained 0-0085 Vieille Montague ,, 0-00522 Ditto Corfali mine ,, 0-00457 French ,, 0-019 Eliot and Storer appear to have investigated with great care the question of the presence of arsenic in zinc. They impugn the accuracy of the methods adopted by Schauffele, and consider his results unsatis- factory. The conclusion at which they have arrived after a long course of experiment with many different varieties of zinc is, "that much of the zinc of commerce is free from arsenic, or at least contains no arsenic that can be detected by the most delicate tests known for that metal." One specimen of Vieille Montagne zinc was found to be perfectly free from arsenic, but another specimen contained decided traces of it ; it was detected in Silesian zinc ; in zinc extracted from blende by the English process at the Mines-Eoyal Works, near Neath; in zinc extracted by the Silesian process from blende by Messrs. Dill- wyn and Co., Swansea ; in zinc stated to have been extracted from silicate of zinc, near Wrexham ; in zinc produced by Messrs. Vivian, Swansea; one specimen of Pennsylvanian zinc was perfectly free from arsenic, while another contained it ; the New Jersey zinc, extracted from red zinc ore, contained an unusually large proportion of arsenic. Of the four samples of English zinc examined, " that of the Messrs. Vivian contained the most arsenic ; 200 grammes of this spelter yielded an enormous mirror of arsenic in less than 10 minutes, and in a few minutes more a second mirror, large enough to give the characteristic odour." ! A portion of arsenic is retained in the in- soluble matter left by the action of hydrochloric or dilute sulphuric acid on zinc. Tin. Eliot and Storer obtained decided evidence of the presence of tin in the zinc of New Jersey and in that of Messrs. Vivian, while in others they either failed to detect it, or only found " faint traces" of it. Karsten found that when the best Silesian zinc was alloyed with only 1 per cent, of tin, it was brittle at the 8 Ueber die Beimischungen welche die Festigkeit des Zinkes vermindcru Archiv. 1842, 1G. p. 597-032. 9 Liebig u. Kopp, Jahresb. for 1850, p. 320. 1 Op. cit., p. 87. 590 FOREIGN MATTER IN COMMERCIAL ZINC. temperature at which zinc is malleable ; and that it could only be rolled into sheet at a much lower temperature, and then not without splitting considerably at the edges. Bismuth and antimony. Neither was detected by Karsten in Silesian zinc. Copper. Not a trace was found by Karsten in Silesian zinc. Eliot and Storer found not the slightest evidence of the presence of copper in any zinc except that of New Jersey. According to Karsten zinc containing ^ per cent, of copper is harder and more brittle than ordinary zinc ; it can only be rolled with difficulty, when it breaks easily and splits much at the edges, and the sheet zinc produced cannot be corrugated without breaking. Silver. None was detected by Karsten in Silesian zinc. Iron. Zinc and iron alloy easily, and commercial zinc is rarely free from iron. Karsten never found more than 0'24 per cent, of iron in any zinc ; but a specimen of Silesian zinc which I procured at M. Lemanski's zinc rolling-mills in London contained not less than 1*64 per cent. : it was pronounced unfit for rolling. Eliot and Storer found 0-21 percent, of iron in New Jersey zinc, and from O05 to 0*07 per cent, in sheet zinc procured in Berlin. The presence of iron is indicated by minute grey specks on the bright cleavage-planes of the freshly fractured surface of an ingot of zinc. Iron increases the hardness of zinc, and causes it to heat so much in rolling that in order to prevent cracking, it is necessary to allow the metal to cool sufficiently from time to time during the process. In the unannealed state the rolled sheet is extremely rigid and strong. According to Karsten, iron is very injurious in zinc intended for rolling when much lead is present ; and zinc containing much iron and not more lead than exists in the best kinds of zinc, which are almost free from iron, is quite unfit for rolling. The proportion of iron in zinc intended for rolling, Karsten further states, should never exceed 0*2 per cent. ; and it should never be accompanied with a greater amount of lead than is usually found in the best quali- ties of zinc. Zinc may be perfectly freed from iron by re-distillation in clay retorts, care being taken to prevent contact of the distilled metal with iron. Cadmium. Zinc in the state in which it is condensed in the process of reduction, or, as it is termed, rough zinc, always contains cadmium, which, being by far the most volatile and most easily oxidis- able of the two metals, may be separated in a greater or less degree from zinc by simply keeping it melted during a sufficient time in a rever- beratory furnace. In 1829 Mentzel published an account of experi- ments which he had made at the Lydognia Zinc-works in Upper Silesia to ascertain the effect of cadmium on the malleability of zinc. 2 He found that, although cadmium increases the hardness and brittle- ness of zinc, an alloy consisting of 85 per cent, of zinc and 15 of cad- mium could yet be rolled to the thickness of a line (about - 1 1 1 - inch) without cracking ; and he states what seems rather anomalous, namely, that this alloy is not more brittle than zinc containing 10 or only 5 per cent, of cadmium. There is no reason for believing that cadmium, in the small proportion in which it occurs in commercial zinc, exerts 2 Karsten' s Arcliiv. 1829, 1. p. 416. FOREIGN MATTER IN COMMERCIAL ZINC. 591 any sensible effect upon the' malleability of the metal. Lead. This metal is generally, if not always, present in commercial zinc, and, according to Karsten, in proportions varying from 0-3 to 2*0 per cent. The following table contains Eliot and Storer's quantitative determina- tions of lead in various kinds of zinc : Lead Lead per cent. per cent. Silesian 1-46 From Paris 6 0-106 Vieille Montagne 0-292 From Berlin, in sheet 1-297 New Jersey, from red oxide ... 0*079 Wrexham, from silicate of) i.iqo Pennsylvania, from silicate of I ft . ftn zinc f zinc f Mines Royal, from blende ... 0-823 Vieille Montagne, used at the I n , Q , Dillwyn and Co., from blende 1-661 U. S. mint f Messrs. Vivian 1-516 6 Sold as pure by Rousseau freres. The dark grey residue obtained by the action of dilute sulphuric acid on commercial zinc consists chiefly of lead. "When zinc and lead are melted together, they separate more or less completely during solidifi- cation ; and when zinc containing lead is kept long melted without being agitated, the lead tends to subside. But zinc cannot be perfectly deprived of lead in this manner. I am indebted to Messrs. Matthiessen and Bose for the following results. Lead and zinc were melted together in various proportions in a Hessian crucible over a large air-gas burner, stirred during a quarter of an hour with a tobacco pipe stem, and then allowed to remain at rest in a fused state during half an hour, a jet of coal-gas being all the while directed on the surface of the racked metals. The metallic mixture was poured into a porous cell previously heated to redness in a large crucible filled with sand. The weight of each casting was about 300 grammes (about 10 ozs. avoird.), the height 100 mm (4 inches), and the diameter 25 mm (1 inch). After cooling and breaking the cell, the upper part of the metal was broken by the blow of a hammer from the lower one. Solidification of the metals gene- rally occurred in about 2 hours after casting. Experiments were made to ascertain whether a more perfect separation would take place by keeping the metallic mixture liquid during a longer time ; but this was not found to be the case. In each case a portion was taken for analysis from the middle of the upper or zinc end, avoiding the outside ; and another portion from the lower or lead end. TABLE OF RESULTS WITH LEAD AND ZINC. Zinc 50 do Zinc end Lead do. 1-22 Lead 66 6 do Lead end ... . Zinc do. 1-62 Zinc 33-3 do. Zinc end . . Lead do. 1-20 Lead 4 do . Ton . Lead do. 1-20 Zinc 96 do Bottom Do do. 1-17 Lead 96 do To}) . . Zinc do. 1-63 Zinc 4 do. Bottom . Do. do. 1-79 592 FOREIGN MATTER IN COMMERCIAL ZINC. Similar experiments were made with zinc and bismuth, and the results will be found in the following table : TABLE OF RESULTS WITH ZINC AND BISMUTH. Bismuth oO per cent. ... Zinc 50 do. Zinc end Bismuth per cent. 2'42 Do Do. do. 2-48 Bismuth end Zinc do. 13 '85 Bismuth 50 do Zinc end Bismuth do. 2-39 7inr ^0 do Bismuth end ... Zinc do 8-65 Bismuth 80 do Top Zinc do 14-0 Zinc 20 do Do Do. do. 14-1 Bottom Do do ] 2 93 Do Do do ]3-l Bismuth 80 do TOD . . Do. do 16*3 Zinc 20 do Bottom Do do 13-3 Bismuth 80 Zinc 20 do. do ... Top Bottom Do. Do do. do 14-1 8'8 Bismuth 5 do . Top... Bismuth do 2-38 Zinc 95 do. Bottom . Do. do. 2-40 The conclusions which Messrs.- Matthiessen and Bose draw from the preceding results are 1. That at a temperature of a large air-gas burner lead will only dissolve 1*6 per cent, of zinc, and zinc only 1*2 per cent, of lead. 2. That bismuth under the same conditions will dissolve from 9 to 14 per cent, of zinc, and zinc 2'4 per cent, of bismuth. Karsten inferred that lead may be present in zinc in two distinct states one in which it is wholly combined with the zinc, and the other in which an alloy of lead and zinc is mechanically diffused through the mass of the zinc. The first condition is stated to occur when fusion takes place at a high temperature and is followed by extremely rapid solidification : the metal thus obtained is hard, and dis- solves easily in dilute sulphuric acid with the formation of a pulve- rulent residue of lead. The second condition is stated to occur when fusion likewise takes place at a high temperature, but is followed by very gradual solidification : the metal in this case is soft, and dissolves slowly in dilute sulphuric acid with the formation of a flocculent or riband-like residue of lead in combination with a little zinc, which only becomes converted into powder by the prolonged action of the acid. With respect to the influence of lead upon the malleability of zinc and the quality of sheet-zinc there has been great difference of opinion. Zinc containing as much as 3 per cent, of lead has been rolled into sheet without cracking ; and it has been stated that zinc strongly impregnated with lead may be more easily and better rolled into sheet than purer and harder zinc, but that the tenacity of sheet-zinc decreases in proportion to the amount of lead which it contains. Karsten asserts that the presence of about li per cent, of lead renders sheet-zinc tender and unsuitable for most kinds of work ; and suggested that the peculiar states of combination in which he supposed lead to METHODS OF EXTRACTING ZINC COMPARED. 593 be capable of existing in zinc, might exert a decided effect on its malleability ; but we have seen that the malleability of the metal is remarkably affected by the mere conditions of temperature under which it is melted and allowed to solidify, so that possibly Karsten in his experiments may have ascribed to the presence of lead what was due exclusively to varying conditions of temperature. In preparing ingots of zinc for rolling, care is taken to allow the lead to subside as completely as possible from the melted zinc. Pre- paratory to casting into ingots, the zinc is kept melted in large quan- tity in a reverberatory furnace, of which the bed inclines from the fire-bridge, so as to form a tolerably deep cavity or well at the oppo- site end, where there is an opening through which the metal is laded out. A considerable quantity of lead subsides and accumulates at the bottom of this well, and is from time to time removed. The tempera- ture of the melted metal, it is hardly necessary to observe, is consider- ably below that at which ignition would occur. The ingots of zinc are thrown on the upper part of the bed near the fire-bridge, and are allowed to melt down gradually. If in lading any lead should acci- dentally be poured into the ingot mould along with the zinc, it will appear in ribands or strips upon the surface of the rolled sheets. THE VARIOUS METHODS OF EXTRACTING ZINC COMPARED WITH EACH OTHER. I should require more precise and far more numerous data than I possess at present in order to enable me to discuss this subject in a satisfactory manner. Much has been written upon it, different con- clusions have been arrived at, and different reasons have been assigned for the selection of one process in preference to another. Yet there are points on which all persons who have had experience in the extraction of zinc seem to be pretty nearly unanimous. It may be advisable to examine the question under the following heads : Consumption of Fuel, Consumption of Refractory Clay, &c., Cost of Labour, and Nature of the Ore. I shall exclude from consideration the Carinthian process, because the disadvantage resulting from its intermitting character, and the con- sequent alternate cooling and reheating of the furnace, is so manifest as to place it out of comparison with the other three processes described. Consumption of fuel. It is decidedly much greater in the English process than in either the Silesian or the Belgian. We have seen that in the treatment of the poor ores in Silesia, yielding less than 20 per cent, of zinc, not so much coal is consumed as in the treatment of ores in the English furnace, yielding more than 40 per cent, of zinc. AYith .ores of the same character and richness, the consumption of fuel is very sensibly less in the Belgian than in the Silesian furnace. Consumption of fire-clay, fyc. It is greater in the Belgian than in the Silesian process, cceteris paribus. I have not exact information as to the amount of consumption under this head in the English furnace. Cost of labour. The amount of labour required is, I believe, greatest in the Belgian process and least in the English. Moreover, in the former, the work is harder, and the men suffer more from heat, than 2 Q 594 METHODS OF EXTRACTING ZINC COMPARED. in either of the other two processes : indeed it is stated that in hot weather it is sometimes difficult to retain men at Avork at the Belgian furnaces. This circumstance might be expected to increase the cost of labour in tho Belgian process, as compared with the other two. On the other hand, in a given time the largest yield is obtained from the Belgian furnaces. But this advantage is in a greater or less degree counterbalanced by the fact that the Silesian furnaces may be kept un- interruptedly at work during a longer period than the Belgian, which require more frequent reparation than the former. In the management of the Belgian furnaces more skill is said to be required than in the case of the Silesian. Nature of the ore. When an ore contains much foreign matter, such as oxide of iron, which exerts a corrosive action on the retorts, the Belgian process is recommended, because the contact of this matter would be least prolonged in this process, in which the retorts are most frequently cleaned out ; and because, owing to the comparatively steep inclination of the retorts, any corrosive slag which may be formed would flow more readily towards the lower and cooler parts, 'where' it would produce less mischief. The varieties of calcined blende employed are apt to be contaminated with oxide of lead, as well as oxide of iron ; and on this account the Belgian process has been pre- ferred for this class of ores. It is requisite, in the case of blende, that the ore should be calcined as sweet as possible ; and in order to this, it must be reduced to fine powder. The internal surface of the retorts will therefore be exposed to greater contact in the case of blende than in the case of calamine, which is only reduced to the size of hazel-nuts ; but the greater the contact, the greater will be the amount of corrosion of the substance of the retorts an evil which will be less in Belgian than Silesian retorts, for the reasons above assigned. My friend Dr. Astley Price, Ph. D., informs me that the corrosive action of any zinc-ore upon the substance of the retorts may be entirely prevented by heating the ore, mixed with coal-tar, in close vessels, so as to carbonize the tar. He had procured a patent for the process in 1858 ; 3 but previously, in 1856, Messrs. George and John Darlington had claimed the use of tar in admixture with zinc-ores in a provisional specification. 4 In the selection of a locality for the erection of zinc-works, it is essentially requisite not only that the ore should be delivered at a reasonable price, but also that good coal and very superior fire-day should be obtainable at a reasonable price. According to the experience of the Yieille Montagne Company, zinc can only be profitably extracted when the price of coal does not exceed 15 fr. per 1000 kil. (12s. Gd. per ton), and that of fire-clay 35 fr. per 3 kil. (1?. 9s. 2cf.) ; and the ores, at the above-mentioned prices of coal and fire-clay, should on an average not contain less than 45 per cent, of zinc. 5 It has been ascertained at the works of this Company, 3 A - D - 1858. No. 828. -i A.D. 1856. No. 1619 5 Tlmm, op. cit., 1860, p. 101. ALLEGED IMPROVEMENTS IN THE EXTRACTION OF ZINC. 595 where both the Silesian and Belgian processes are conducted, that greater consumption of coal in the former is, with respect to value, equivalent to greater consumption of clay in the latter, 6 and that the cost of production is about the same in both processes. If we know, therefore, the relative consumption and actual prices of clay and coal in these processes, it is easy to select that which, cceteris paribus, will yield the most profit, ALLEGED IMPROVEMENTS IN THE EXTRACTION OF ZINC. Several patents have been obtained for effecting the reduction of zinc from its ores in blast-furnaces, and collecting the volatilized metal in various kinds of condensing apparatus ; but I am not aware whether any of the methods specified have been tried on the large scale. It is, how- ever, certain that not one of them has come into use. Although the ores might be readily reduced in a blast-furnace, yet there would assuredly be great practical difficulties in condensing the metal without consider- able loss. The vapour of the zinc would issue from the furnace in ad- mixture with a large volume of unconden sable gases ; and this condition would render the condensation and economical collection "of the metal extremely difficult, if not impossible, on the large scale. Of this we shall hereafter have ample proof when we examine the subject of the collection of the fume produced in lead-smelting. Costly contrivances have been adopted in order to catch this fume, and yet scarcely one has proved efficient in preventing the escape of an enormous quantity of lead into the atmosphere. We shall also find that precisely the same difficulty is experienced in collecting mercury when it is volatilized in admixture with the gaseous products of the combustion of fuel. While I do not venture to predict that the blast-furnace will never be success- fully applied in the extraction of zinc from its ores, yet I confess I do not see much reason to anticipate improvement in that direction. The large experience which we now possess concerning the metallurgy of zinc would rather encourage us to hope for further improvements in the construction of the reduction -furnaces and in the mode of applying heat. As I have stood by the hot-blast stoves of some of the great South Welsh iron-works, which are heated exclusively by the com- bustion of the waste-gases of the blast-furnaces, it has often struck me that this mode of heating would be peculiarly well adapted to zinc reduction-furnaces. The temperature is remarkably uniform through the interior of these stoves, and there would appear to be no difficulty in raising it to a degree amply sufficient to reduce oxide of zinc ; for I am assured that by suitably regulating the supply of gas and atmos- pheric air, the fire brick lining of the stoves may even be melted. It would be possible, I think, with the use of this gaseous fuel, to con- struct a zinc reduction-furnace of large dimensions, either on the Silesian or Belgian principle, which should combine the most favour- able conditions for the extraction of zinc from its ores, namely, a suffi- 6 Thum, op. cil, p. 101. 2 Q 2 596 METHODS OF ASSAYING ORES OF ZINC ciently high and perfectly equable temperature with the capability of regulating it to a nicety. Some -years ago a gas-zinc-furnace was constructed by Meiitzel on the Silesian principle, at the Lydognia Works/ and was reported by him to be successful ; but as I have found no mention of such a furnace in any recent description of these works, I presume that on further trial the result proved unfavourable. The furnace resembled an ordi- nary Silesian furnace, with the exception that the fireplace was re- placed by a deep chamber or gas-generator, for the production of the gas (see p. 200, ante) ; and there were several air-flues in the walls of this chamber, which proceeded from below upwards and supplied air, necessary for the combustion of the gases, in the immediate vicinity of the retorts. I may be wrong in inferring that the experiment at the Lydognia Works was a failure. However this may be, I still have a strong impression that gas-zinc-furnaces will one day be adopted. It may require a considerable outlay in making the requisite experiments on the construction of such furnaces, and success may only be attained after numerous failures. The silicate of zinc, or electric calamine, of which I believe con- siderable quantities are still thrown away in this country, and which is frequently a constituent of common calamine, is, I am informed by Professor Brush, employed with success as an ore of zinc in the United States ; and some years ago I received from Mr. J. A. Phillips a speci- men of zinc extracted, on the large scale, in North Wales from the same description of ore. This zino, however, was not so free from iron as we were led to anticipate ; but this no doubt depended on the appa- ratus used to effect condensation, or on that used in the process of re- melting, and was in no degree caused by the nature of the ore. We have seen that the oxide of zinc in electric calamine may be completely reduced by charcoal alone at a high, temperature ; but in the retorts of ordinary zinc-furnaces, it is stated that any silicate of zinc which may be present in the ores remains unreduced. To effect reduction by carbon alone, the temperature must not only be very high, but the mixture must be very intimate. Probably the best method of operating on electric calamine would be to mix it thus intimately with the car- bonaceous reducing agent and lime, or lime and clay to furnish base for the saturation of the silica. METHODS OF ASSAYING ORES OF ZINC. The ores with which the assayer has to do are Calamine, raw or calcined. Blende, raw or calcined. Red oxide of zinc. Silicate of zinc. Formerly the assays were made by the " dry way," i. e. by exposing the ore in admixture with charcoal or other reducing agent to a high 7 Karsten, Arclnv. 2. B. 23. METHODS OF ASSAYING ORES OF ZINC. 597 temperature, and determining the amount of zinc by loss ; and when the ores are rich in zinc and free from volatile metallic impurities, the per-centage obtained will approximate to the correct amount. The process is still occasionally resorted to by some assayers, and may be conveniently applied when liquid reagents are not at hand. It may be conducted thus : 50 grs. of ore are calcined " sweet," and the calcined product is weighed and mixed with from 10 to 15 grs. of charcoal- powder (of which the amount of ash which it leaves on burning has been previously ascertained), and the mixture is exposed, in a closed brasqued 8 clay crucible, to a high temperature for about half an hour. The residue from the cavity of the brasque is recalcined and weighed ; and from the weight is deducted the amount of ash derived from the charcoal ; and the difference between the weights found after the first and second calcinations will be equivalent to the amount of oxide of zinc reduced and volatilized. This oxide is composed of 1 part of oxygen by weight, and 4 parts of zinc ; so that the weight lost by volatilization, minus -, will give the amount of zinc contained in the weight of the ore employed in the process. The actual results yielded by an assay on this principle are presented in the following tabulated form : Grains. 50 grains of calamine after calcination weighed 33*50 The calcined residue, after exposure of the roasted ore to a high temperature in admixture with charcoal (minus the weight of asli in the quantity of charcoal used), weighed 10-30 Oxide of zinc reduced and volatilized, i.e. Loss 23*20 Deduct i for oxygen , 4*64 50 grains of ore contain of zinc 18*56, or 37*12 / . A covered clay crucible may be substituted for one lined with carbon ; but in that case a larger amount of charcoal or other carbonaceous re- ducing agent must be mixed with the calcined ore. When much oxide of iron is present, it will be reduced with the formation, if the tempe- rature be high, of minute globular particles of iron, and the result will accordingly be vitiated. 9 When lead is present as galena, it is con- verted, during calcination, into oxide and sulphate of lead, which, on subsequent heating with carbon, become reduced to metallic lead and sulphide of lead respectively, and, in the presence of iron, to metallic lead and sulphide of iron. W hen present as carbonate of lead, the latter is changed by calcination into oxide, which is subsequently reduced by the carbon with the separation of lead. A considerable amount of lead is also volatilized during the fusion ; and the loss from this cause, together with the reductions just mentioned, renders the 8 The brasque should be solid, and may 9 A correction may be made by extract- be conveniently made of charcoal powder ing the globules of iron with ;i magnet, moistened with treacle. (See Crucibles, and deducting ;m equivalent amount of &c<, ante.} \ sesquioxide of iron. 598 METHODS OF ASSAYING OKES OF ZINC- assay worthless. If the ore is a silicate, or contains silicate of zinc, the oxide of zinc, in combination with the silica, will be completely reduced by the charcoal, if the conditions are correct, and a skeleton of silica be left ; and if lime is present, a residue of silicate of lime will be obtained. The proportion of zinc in blende may, in the absence of volatile im- purities, be determined by exposing a known weight of the pulverized blende, in admixture with a small quantity of charcoal, to a high tem- perature in a closed brasqued crucible. If the blende is pure, it will be completely reduced and volatilized ; but when iron is present, a fused button of sulphide of iron will remain. The loss in weight indi- cates the amount of the sulphide of zinc, which contains 67 % of zinc. BY SULPHIDE OF SODIUM. This is the most direct, .quickest, and most practical method for the estimation of zinc in ores of zinc ; and it is accordingly that which is now generally employed for commercial purposes. It is capable of yielding results within a few tenths per cent, of the amount of zinc obtainable by analysis. When a solution of sulphide of sodium is added to an ammoniacal solution of zinc containing a small amount of hydrated oxide of iron, the zinc is precipitated as white sulphide of zinc, and afterwards the suspended oxide of iron beeornes blackened by the slight excess of sulphide of sodium which may have been used. By ascertaining how much of a standard solution of sulphide of sodium is required exactly to precipitate the zinc from the ammoniacal solution of a known weight of the metal, it follows that by a comparative expe- riment on a given weight of ore, the amount of zinc present in an ore may be determined. For this method, solutions of sesquichloride of iron and sulphide of sodium are required. The dilute solution of sesquichloride of iron may be prepared as follows : 50 grs. of haematite in powder (Fe 2 3 ) are to be dissolved in hydro- chloric acid ; or 35 grs. of iron-wire may be dissolved in nitro-hydro- chloric acid, and the solution afterwards diluted with water to 1 pint : 50 grs. of this solution will contain about 0-2 of metallic iron. Standard Solution. 1400 grs. of crystallized monosulphide of sodium (XaS, 9 HO) are dissolved in 4 pints (35,000 grs.) of water. This salt is very soluble in water, and the solution, which should be clear and colourless, is ready for immediate use. The solution should either be filtered, or the clear supernatant liquor decanted off from any small amount of black precipitate which may be occasionally present. It should be kept in bottles of glass free from lead. An objection against the use of this solution is its liability to slow decomposition, so that it requires to be re standardized every third or fourth day. A quantity of solution prepared and employed for assaying had the following standard: 1000 grs. = 10-03 grs. of zinc; after the lapse of 15 days '0 grs. = 9 ;} 1 grains of zinc. In practice, where many assays have to be made, this liability to decomposition is immaterial, as the solii- BY SULPHIDE OF SODIUM. ' 599 tion can readily be re-standardized ; or not more solution need be prepared at a time than is sufficient for a given number of assays. \\hen crystallized sulphide of sodium cannot be obtained, about 2000 grs. of caustic soda are to be dissolved in 1 pint of water, and the solution is to be divided into two equal portions. Through one of these portions sulphuretted hydrogen gas is passed until saturation occurs ; it is then added to the other portion, when the strength of the sulphide of sodium solution thus prepared must be approximately found, after which it must be diluted with water until it is reduced to the proper strength. The solution may be standardized, as follows : 2 or 3 pieces of pure zinc, varying from 5 to 10 grs. each in weight, are dissolved in separate portions of dilute hydrochloric acid, containing a few drops of nitric acid ; each solution is put into a separate flask and diluted with water until it occupies about three-quarters of a pint in volume ; 50 grs. of the dilute solution of sesquichloride of iron, and afterwards ammonia in excess, are added to each solution in succession. The solu- tion of sulphide of sodium from a graduated burette is slowly allowed to run into the cold ammoniacal solutions until the zinc is completely precipitated, and the flocculent sesquioxide of iron blackened. The number of divisions of the solution in each flask used is then read off, and the mean of the numbers is taken as the standard. For example 200 Divisions (1000 grs.) Ziuc. of solution of sulphide Divisions. of sodium afe equal to 5 885 grs. required 117 therefore 10 06 grs. of ziue. 7-9GO 159 10-01 Mean number or standard 10 03 Process. From 10 to 50 grs. of the ore are to be heated with hydro- chloric or sulphuric acid and a small quantity of nitric acid, a large excess of acid being avoided ; when decomposition is complete, the solution is diluted with a small quantity of water ; ammonia and car- bonate of ammonia in excess are poured in, and the whole is digested at a gentle heat during 20 or 30 minutes, and afterwards filtered into a pint flask, and the residue on the filter washed with warm ammoniacal water. To the ammoniacal filtrate, which should measure about three-quarters of a pint, are added 50 grs. by measure of the solution of sesquichloride of iron to which ammonia has been previously added. When the solution is cold, the measured solution of sulphide of sodium from the burette is slowly and cautiously dropped in, while the ammoniacal solution is kept in constant motion, until the white pre- cipitate of sulphide of zinc assumes a greyish black colour from the blackening of the flakes of sesquioxide of iron in suspension. The number of divisions of the solution of sulphide of sodium used is then read off, and from the standard the percentage of zinc is calculated. For example: 200 divisions (1000 grs.) of standard solution = 10-03 grs. of zinc, and 20 grs. of ore required 170 divisions of the same solution. 200 : 170 :: 10-03 : 8-53 8-53 X 5 = 42-65 % of zinc. 600 METHODS OF ASSAYING ORES OF ZINC It is desirable to have the oxide of iron present in a flocculent con- dition : it is therefore better to add ammonia to the measured quantity of sesquichloride of iron, before mixing it with the ammoniacal solu- tion of zinc ; for if the sesquichloride is added direct, the oxide of iron coagulates. Instead of the sesquichloride, tartrate of ammonia and iron or tartrate of potass and iron may be added, when the iron will remain in solution ; good results may be thus obtained, but not better than with the flocculent oxide of iron. Although it is possible to make correct assays, after a little practice, with variable amounts of oxide of iron present in the ammoniacal solution, by stopping the assay at the proper tint, which will vary according to the amount of iron pre- sent, yet the use of the flocculent oxide is preferable. It is necessary to agitate the ammoniacal solution continually during the addition of the sulphide of sodium ; otherwise the oxide of iron is liable to be blackened before the complete precipitation of the zinc as sulphide, by coming in contact with the sulphide of sodium as it enters the solution. If the oxide of iron becomes thus blackened, the assay is spoiled ; but this difficulty will not be experienced -if the assay is properly managed. The change of tint of the oxide of iron from reddish-brown to black is best observed by transmitted light ; and when darkening begins to appear, the sulphide of sodium should be added drop by drop until the proper point, when the blackening will occur rapidly. There is a slight error owing to the retention of a small quantity of zinc in the residue after digestion in ammonia and carbonate of ammonia ; but it is practically unimportant, and is compensated for by the slight excess of sulphide of sodium used in blackening the oxide of iron at the completion of the assay. Precautions to be taken in the presence of the following metals : Iron. This metal is often present in ores of zinc, in calamine as sesquioxide and carbonate of protoxide, and in blende as sulphide. When it occurs in small quantity the assays may be made without any nitration; but it is better always to filter off and add a definite amount of iron to each assay. If the iron present is considerable, on addition of the ammonia, it is apt to retain some oxide of zinc ; and in this case after decanting off the ammoniacal solution it is better to treat it a second time with ammonia and carbonate of ammonia pre- viously to filtration. Manganese. -It is present in red oxide of zinc, calamine, &c. On the addition of carbonate of ammonia, the manganese will only be par- tially separated ; but the addition of a few drops of bromine to the ammoniacal solution will ensure its complete precipitation. Copper. -It is occasionally present as carbonate in calamine, or as copper pyrites in blende. It communicates a violet or blue colour to the ammoniacal solution. When present in small quantity, it may be separated conveniently by adding a few drops of the sulphide of sodium to the hot filtered ammoniacal solution of zinc, until the colour is discharged, and afterwards filtering rapidly. The zinc can now be estimated in the filtrate when cold. If copper is present in con- BY SOLUTION OF AMMONIA AND CARBONATE OF AMMONIA. 601 siderable quantity, the ore must be treated with dilute sulphuric acid, and the copper precipitated from the solution of sulphate of zinc thus obtained by means of a piece of clean bright iron. The precipitated copper may be washed, dried, and weighed, if necessary. The solu- tion freed from copper is heated with a small quantity of nitric acid in order to peroxidize the iron, after which the assay may be proceeded with in the manner described. Lead. It is sometimes present as carbonate of lead in calamine, or as galena in blende. A portion of the metal will be left in the in- soluble residue as sulphate, and the remainder will be precipitated on the addition of the ammonia and carbonate of ammonia, so that lead does not interfere with the assay. Silver. It is occasionally present in blende. In the presence of dilute hydrochloric acid, it will be left in the insoluble residue, which should be separated by filtration before the addition of the ammonia, &c. ; but it is seldom present in sufficient quantity to interfere with the assay. Cadmium. It is occasionally present in small quantity; it is not soluble in the ammoniacal solution containing carbonate of ammonia, and, therefore, does not interfere with the assay. Method of assay by dissolving out the oxide of zinc by means of solution of ammonia and carbonate of ammonia. This method only yields approxi- mate results, but it is nevertheless useful when sulphide of sodium is not at; hand. It answers pretty well for tolerably pure ores; but when other matters are present, especially oxide of iron, the error is increased in proportion to the amount of iron present. The process is conducted as follows : From 20 to 50 grs. of the finely powdered ore are weighed out in a porcelain capsule, and calcined completely in a muffle or over an air-gas burner, and the weight of the calcined ore is taken. This is digested in a beaker or flask at a gentle heat, with a solution of ammonia and carbonate of ammonia, for about 20 minutes. The insoluble residue is collected on a filter, washed with hot ammoniacal water, dried, ignited, and weighed. The weight of the residue thus obtained after digestion in ammonia, &c., deducted from the weight of the calcined ore, indicates the amount of oxide of zinc dis- solved out ; and this, minus deducted for oxygen, indicates the amount of zinc present in the quantity of ore operated on. Example : Grains. Weight of blende taken 25 "Weight of calcined ore before treatment with am- monia, &c 21-23 Weight of calcined residue after treatment with am- monia, &c 3 -90 Loss = Oxide of zinc, soluble in the ammoniacal solution 17-33 Deduct | for oxygen 3'46 Zinc in 25 grs. of blende 13-87, or 55-48 / . G02 METHODS OF ASSAYING ORES OF ZINC. When the ore is finely powdered, and contains only a small pro- portion of foreign matter, there is no difficulty in extracting prac- tically the whole of the oxide of zinc from the calcined ore by the direct action of ammonia and carbonate of ammonia. But when the amount of oxide of iron or earthy matter is large, a portion of the oxide of zinc will be retained in the residue. This especially occurs with oxide of iron, in which case it is better to modify the process thus : After the ore is calcined it must be treated with hydrochloric and nitric acids, and to the solution, diluted with water, ammonia and car- bonate of ammonia are added ; the insoluble residue and the precipi- tate are collected together on a filter, washed, ignited, and weighed : the weight deducted from the weight .of the calcined ore gives the quantity of oxide of zinc soluble in the ammoniacal solution. If the residue is now examined, it will be found to be practically free from zinc. Of course, if the ore contains silicate of zinc, the amount of oxide of zinc existing as such will be extracted at the same time ; so that this modification cannot be resorted to if it is wished to determine the amount of zinc exclusive of that existing as silicate. Sometimes it is desirable to ascertain roughly for smelting purposes in an ore of zinc, how much zinc is present as silicate ; and this may be done by means of the solvent action of ammonia and carbonate of ammonia on the calcined ore, when the silicate of zinc is left comparatively unacted upon; but the digestion should not be continued too long, otherwise the silicate will be attacked and zinc dissolved. Examples of results actually obtained and relating to the preceding description. 1. 20 grs. of finely powdered blende (from Laxey, Isle of Man), nearly pure, after calcination over an air-gas-burner weighed 16*85: after digestion in a mixture of hot ammonia and carbonate of ammonia the ignited residue weighed O26 grs. 16-85 - 0-26 = 16-59 grs. of oxide of zinc dissolved out ; 16 '59 - 3 -32 (oxygen) = 13-27 grs. of zinc ; or 66 35 %. The insoluble residue, after ignition, contained a trace of zinc, and the ammoniacal solution a trace of copper. 2. 25 grs. of calcined calamine (prepared for smelting) weighed after calcination 21-45 grs. : after digesting in ammonia and carbonate of ammonia, the ignited residue weighed 4*86 grs., and the ammoniacal t solution was tinted slightly by copper. 21-45 - 4-86 = 16 -59 grs. of oxide of zinc; 16 "59 - 3 -32 (oxygen) = 13- 27 grs. of zinc; or 53-08%. The 4-86 grs. of residue, after prolonged re-digestion in ammonia and carbonate of ammonia, on re-ignition, weighed 4-75 grs. By boiling in hydrochloric acid, and estimating the zinc dissolved as oxide, the latter weighed 0-56 gr., which is equivalent to 1-78 / of zinc left in the insoluble residue. Three assays 011 the same sample of ore by the sulphide of sodium plan gave as a mean result 53-45 / of zinc. RESULTS OBTAINED. 603 3. 20 grs. of blende, containing 20*48 % of sulphide of iron ( = 13-03 / of iron), after calcination in a muffle, weighed 10*65 grs. The calcined product was digested with twice its weight of carbonate of ammonia, and some ammonia, for about of an hour ; the insoluble residue, collected on a filter, and ignited, weighed 5*97 grs. ; after re-digestion with ammonia, &c., it weighed 5*88 grs. 16-65 5 88 = 10-77 grs. of soluble oxide of zinc ; equivalent to 8 61 grs. of zinc or 43 -05%. The 5*88 grs. of residue, acted on by hydrochloric acid and then digested with ammonia and carbonate of ammonia, gave of insoluble matter, after washing on a filter and ignition, 4*28 grs. The filtrate contained zinc. 5-88 4-28 = 1 '60 grs. of soluble matter, equivalent to 1*28 of zinc, or 6'40 / , retained in the first residue by the sesqui- oxide of iron, and not dissolved out by the direct action of the ammoniacal solution. 43-05 + 6-40 =49-45 / ; analysis gave 49*38 / of zinc. Another experimenter on the same ore obtained the following results : By the direct action of ammonia and carbonate of ammonia on the calcined ore, 42*74 / of zinc. By treating the calcined ore with hydrochloric acid, boiling with ammonia, carbonate of ammonia, &c., and then weighing the insoluble residue, 50*38 / of zinc. 4. 50 grs. of calcined calamine (prepared for smelting), containing silicate of zinc, after calcination weighed 46-20 grs. ; after treatment with ammonia and carbonate of ammonia, and igniting the insoluble residue, it weighed 19*05 grs.; equivalent to oxide of zinc dissolved, 27-15 grs.; equivalent to 21*72 grs. of zinc, or 43*44 / . The residue gelatinized on the addition of hydrochloric acid, and contained a con- siderable amount of zinc (about 11*35%). Two assays by the sul- phide of sodium plan gave 54*65 and 54*94 / of zinc ; the mean of these results is = 54*79 / . 5. 20 grs. of finely-powdered native hydrated silicate of zinc, con- taining 53*33 / f zinc, were digested with a hot solution of ammonia and carbonate of ammonia five times successively, and the decanted solution was each time tested for the presence of zinc ; about three- fourths of the oxide of zinc were extracted on the first digestion ; a small quantity on the second and third ; a very much smaller quantity on the fourth ; and on the fifth a trace of zinc was still extracted. The final residue was decomposed by hydrochloric acid, and the zinc esti- mated as oxide, which weighed 0*22 grs. ; equivalent to 1*10 / of zinc left insoluble after the action of the ammonia. The following experiment was made to determine the amount of oxide of zinc which may be dissolved out of calcined silicate of zinc by a solution of ammonia and carbonate of ammonia : 20 grains of silicate were employed. By the 1st digestion, 26*15 % of oxide was extracted; by the 2nd, 4*15 / ; and on further digestion zinc continued to be dissolved. 604 METHODS OF ASSAYING OKES OF ZINC- TABLE OF KESULTS OBTAINED BY DIFFERENT METHODS USED FOR ESTIMATING ZINC. Percentage of Zinc. NATURE OF THE OKE. I. By Analysis. II. By a standard solution of sulphide of sodium. Hi. By the solvent action of ammo- nia and curb, of ammonia. 1 Blende from Laxey nearly pure . 49-38 54-93 39-44 66-54 49-78 49-3 55-25 /39-33\ \39-56j (53-611 \53-71I (53-38\ \53-61/ /53-21\ \53-35j 66-3 49-5 50-4 53-1 2. Blende, containing sulphides of iron and 3 Same sample as 2 by another person . . 5. Blende, with spathose iron ore, carbonate i of lime &c . I 6. Calcined calamine, containing silicate of, 7. Calcined calamine, containing copper 8 Same sample as 7 by another person 1. The same ore as that used in Experiment 1, p. 602. 2. The result in column 1 was obtained as follows : after the separa- tion of the iron and manganese, the zinc was precipitated as sulphide, which was redissolved in acid, and from the solution thus obtained, it was precipitated as carbonate and after ignition weighed as oxide. 3. In column 3 the result was obtained by a modification of the ammonia plan, i. e. by decomposing the ore with acid before dissolving the oxide of zinc. 4. In column 1 the result was obtained by weighing the zinc as oxide. 5. In column 1 the result was obtained by evaporating the ammo- niacal filtrate to dryness, and afterwards precipitating the zinc by car- bonate of soda, and weighing as oxide of zinc. 7. In column 2 the result was obtained as follows : the copper was precipitated from the hot ammoniacal solution by sulphide of sodium, the solution filtered, and the zinc determined in the filtrate when cold. The same ore was used in Experiment 2, p. 602. l By a standard solution of bichromate of potash, $c. This process depends upon the fact that a solution of sesquichloride of iron is reduced to the state of protochloride, by freshly precipitated sulphide of zinc, with the liberation of sulphur : this reaction is explained by the formula The amount of protochloride of iron formed is afterwards estimated by a standard solution of bichromate of potash or permanganate of potash, as in the wet method of assaying iron ores. The bichromate of potash solution employed for estimating iron is made by dissolving about 305 grs. of the crystals in 4 pints of water, and this may also be used in the assaying of zinc ores. But if it is prepared specially for the purpose, it should be made stronger, in order to obviate the necessity which may sometimes arise of refilling the 3 1. By K. Smith. 2. By R. Smith. 4. I By C. Tookey. I. By C. Tookey. II. By R. Smith. 7. By R. Smith. II. By K. Smith. 5. BY A STANDARD SOLUTION OF BICHROMATE OF POTASH. 605 burette during the process, viz. by dissolving 525 grs. of the bichromate in 4 pints of water; 200 divisions (1000 grs.) of this solution will indicate about!7'2 grs. of iron, or be equivalent to about 10 grs. of zinc. The solution of ferricyanide of potassium, used in conjunction with the bichromate, is prepared by dissolving 2 grs. of the crystals in about 8 fluid ounces of water. The strength of the solution of bichromate of potash may be deter- mined by dissolving in a flask a known weight of pure iron wire (from 5 to 10 grs.) in hot dilute hydrochloric acid, diluting the colourless solution with about i a pint of water, and adding the bichromate, until a drop of the solution taken out ceases to give the faintest trace of blue colour with a drop of the ferricyanide of potassium solution : every 10 grs. of iron present will be equivalent to 5-81 grs. of zinc. Full details will be given concerning this process, under the subject of Iron, in a subsequent volume. Process. A known weight of the ore is to be taken, and the am- inoniacal solution of zinc obtained in the same way as in the sulphide of sodium plan. To the warm ammoniacal solution, sulphide of sodium or sulphide of ammonium in slight excess is added, the precipitated sulphide of zinc is collected on a filter, and washed with warm am- moniacal water. The filter with its contents is transferred to a beaker- glass provided with a glass cover, and after the addition of an excess of a slightly acid, dilute solution of sesquichloride of iron, the whole is digested at a gentle heat for about 10 or 15 minutes ; sulphuric acid is then poured in ; a gentle heat is continued to be applied until the sulphur coagulates ; the solution is filtered and the filter washed with warm water, after which the amount of protochloride of iron present in the filtrate is determined by bichromate of potash as follows : The filtrate, if necessary, is diluted with water to about f of a pint, and a small quantity of hydrochloric acid is added : the standard solution of bichromate of potash is then gradually added, until all the iron is converted into sesquichloride, and a drop of the solution taken out on the end of a glass rod ceases to give the slightest trace of blue colour, with a solution of ferricyanide of potassium (red prussiate of potash) previously dotted out on a white porcelain slab or plate. The number of divisions of the bichromate of potash used is then read off, and from the quantity of iron present as protochloride the amount of zinc is calculated ; 56 parts of iron will be equivalent to 32'53 of zinc. For example : 200 divisions (1000 grs.) of the bichromate indicate 10 of iron, or 5'81 of zinc. Iron. Iron. Zinc. Zinc. 56 : 10 : 32-53 : 5-81 20 grs. of the zinc ore required 180 divisions of the bichromate of potash. Divisions. Divisions. Zinc. Zinc. 200 : 180 : 5'81 : 5'23 5-23 x 5 = 26- 15% of zinc. The above process yields good results ; but that by a standard solu- tion of sulphide of sodium is to be preferred for practical purposes on the ground of rapidity of execution. ( 006 ) B K A S S. 1 IN this work the term brass will be restricted to alloys consisting essentially of copper and zinc ; but it should be more properly con- fined to such as are either decidedly yellow or have the yellowish tint characteristic of brass. The old English name for it is latten. Copper and zinc unite in all proportions and form homogeneous alloys, and the union is attended with the evolution of heat. The number of these alloys in actual use is considerable. There is great confusion with respect to the names by which they are distinguished ;- for in some cases different terms are applied to the same alloy, and in others the same terms are applied to different alloys. Thus the terms tombac, prince's metal, similar, and Mannheim gold are used by some authors to designate alloys consisting of about 85 per cent, of copper and 15 per cent, of zinc ; whereas, according to other authors, princes metal and Mannheim gold are synonymous, and are composed of 75 per cent, of copper and 25 of zinc ; 2 according to another author similar consists of about 71 per cent, of copper and 28^ per cent, of zinc, and Mannheim gold of 80 per cent, of copper and 20 per cent, of zinc ; 3 and, again, according to another author, similor and Mannheim gold are synonymous, and are applied to alloys of copper containing from 10 to 12 per cent, of zinc and from 6 to 8 per cent, of tin. 4 Seeing that such inextricable confusion exists in the employment of the terms above mentioned, it is desirable to discard them altogether, and to indicate the varieties of brass by their per-centage composition. But when the term brass is used without qualification, it is generally intended to indicate an alloy consisting of about 2 parts by weight of copper and 1 of zinc. Birmingham is the principal seat of what may be termed the brass industry of this country. It is stated that the trade was introduced into the town, about A.D. 1740, by the family of Turner, and that the chief supply of the metal was derived from the Macclesfield, Cheadle, and Bristol Companies. 5 1 In a subsequent part of this work I shall examine the subject of the constitu- tion of alloys ; and I adopt this plan be- cause an extensive knowledge of alloys is essential for such an examination, and I think it will be most agreeable and in- structive to the reader to acquire it gra- dually in the consecutive study of the various metals. 2 Supplement to Ure's Diet, of Arts, &c., 1844, p. 35; Gmelin's Handbook, loc. cit. 3 Handbuch der Metall-Legirungen, von Joh. Tenner. Quedlinburg, 1860, p. 15. 4 Regnault, Cours elemen. de Chimie, 2ifeme e - d 3 p 300 5 Button's History of Birmingham. 1795, p. 113. VALUABLE QUALITIES OF BRASS. 607 It appears that the manufacture of brass was for some time in the hands of a few wealthy persons in Birmingham, who acted, as mono- polists usually act, in an impolitic manner, and overreached them- selves. Upon this subject I may introduce the following somewhat quaint observations of Hutton, the historian of Birmingham. These monopolists, " instead of making the humble bow for favours received, acted with despotic sovereignty, established their own laws, chose their own customers, directed the price, and governed the market. In 1780 the article rose, either through caprice or necessity, perhaps ihe former, from 721. a ton to 84/. ; the result was an advance upon the goods manufactured, followed by a number of counter-orders and a stagnation of business. In 1781, a person, from affection to the user, or resentment to the maker, perhaps the latter, harangued the public in the weekly papers, censured the arbitrary measures of the brazen sovereigns, showed their dangerous influence over the trades of the town, and the easy manner in which works of our own might be con- structed. Good often arises out of evil : this fiery match, dipped in brimstone, quickly kindled another furnace in Birmingham. Public meetings were advertised, a committee appointed, and subscriptions opened to fill two hundred shares of 100/. each, deemed a sufficient capital ; each proprietor of a share to purchase one ton of brass annually. Works were immediately erected on the banks of the canal, for the advantage of water-carriage, and the whole was conducted with the true spirit of Birmingham freedom. The old companies, which we may justly consider the directors of a South Sea bubble in miniature, sunk the price from 84. to 56/. Two inferences arise from this measure : that their profits were once very high, or were now very low ; and, like some former monarchs in the abuse of power, they repented one day too late." 6 The new works were those to which reference will hereafter be made as belonging to Mr. Pemberton. In 17D5 the annual consumption of brass, according to Hutton, in Bir- mingham amounted to 1000 tons. The qualities which render brass so valuable may be succinctly stated as follows : It is harder than copper, and therefore better resists wear : it is malleable and ductile in a high degree, so that it may be rolled into thin sheet shaped into vessels of various kinds by means of the hammer raised by the process of stamping into objects such as curtain-bands and drawn out into fine wire : it is well adapted to casting, as it is easily fusible at a lower temperature than copper and is capable of receiving very delicate impressions from the mould : it is stated to resist atmospheric influences better than copper; but when its surface is unprotected by lacquer, it rapidly tarnishes and becomes black ; it has a pleasing colour, and is susceptible of a fine polish ; and, lastly, it has one great advantage over copper, namely, cheapness. The so-called Dutch metal, or leaf, furnishes an excellent illustration of the remarkable malleability of certain kinds of brass : it is only brass containing a large proportion of copper, and is made 6 Op. cit., pp. 114, 115. 608 MALLEABILITY OF BKASS, by beating out the laminated metal in the same manner as gold-leaf; the thickness of the leaves of Dutch metal is stated not to exceed The malleability of brass varies with its composition and with the temperature ; and it is also affected, in a very decided degree, by various foreign metals, though present only in minute quantities. Some kinds of brass are malleable only while cold, others only while hot, and others are not malleable at any temperature. At a temperature sensibly below fusion, all brass, like copper, is brittle, and may be reduced by trituration to powder. When an ingot of brass is broken while hot, its fracture is coarsely fibrous or columnar ; but when broken cold, it should be finely granular, at least in the case of certain descrip- tions. According to Bertliier, zinc is never completely volatilized from copper by heat. 7 A table will be found further on containing informa- tion concerning the working qualities, colours, &c., of numerous alloys of copper and zinc. When the fracture of a cast ingot of certain metals is fibrous, the directions of the fibres will generally be at right angles to the cooli no- surfaces. In the case of a sphere, the fibres will have the direction of radii ; and in the case of a square or rectangular prism, two diagonals will be plainly visible on the transverse fracture, formed by the points of junction of the internal extremities of the fibres. I have excellent illustrations of such arrangements of fibres. According to Storer, " the tendency to shoot into fibres . . . extends over quite a space, from alloys containing 57 or 58 per cent, of copper, or even more, down to those containing 43 or 44 per cent., where it gradually disappears. ... It is remarkable that this inclination to form fibres is strongest in those alloys which contain nearly equal equivalents of zinc and copper, being less clearly marked as one recedes in either direction from this point, until a stringy texture analogous to that of copper is reached on the one hand, and the peculiar pastiness of zinc on the other. In preparing crystals, this pastiness manifests itself decidedly in the alloys immediately below those which are fibrous, becoming more strongly marked as the alloys are richer in zinc; at least so far as my own observations have extended, i.e. to 30 per cent, or less of copper. The fracture of these white alloys is for the most part vitreous." 8 I have ingots of best-selected copper, cast in closed iron moulds, which have not a stringy, but a coarsely fibrous transverse fracture : these ingots are 12 inches long, and exactly 1 inch square, and the fracture was obtained by breaking them while cold. On attempting to roll them, either cold or hot, and with a very gently graduated pressure, they were cracked in innumerable places. There is great variation in colour in the alloys of copper and zinc, from the red of copper at one extreme to the bluish white of zinc at the other. The transition from one to the other is gradual, through intermediate stages of yellow. Tr. cles Essais, 2. p. 603. I by Frank H. Storer ; Memoirs of the On the Alloys of Copper and Zinc, I American Academy, 1860, 8. p. 35. CRYSTALS OF BRASS PROCESS OF STAMPING. 609 P>y fusion, brass is occasionally obtained in tolerably good crystals. I hiring at least twelve years I liave had such crystals in my possession, and they have always appeared to me to resemble the regular octa- hedron ; but unfortunately they are not sufficiently perfect to admit of admeasurement ; and without this no positive statement of any scien- tific value can be made on the subject. Numerous experiments have been made by Mr. Storer, Harvard ( 1 ollege, U. S., On the production of crystals of various kinds of brass, by the usual method of preparing crystals of bismuth ; and the con- clusion at which he has arrived is, that the crystals of brass .are octa- hedrons belonging to the monometric or cubical system, and that zinc is isomorphous with copper, which, as is well known, crystallizes in that system. He obtained the most perfect individual crystals from a quantity of brazier's solder, which consists of equal parts, by weight, of copper and zinc, and occurs in the state of coarse powder, produced by heating the alloy to a sufficient degree, and triturating it while hot. Mr. Storer adduces, in support of his conclusion that zinc belongs to the monometric system, the statement of Nickles, that this observer obtained zinc in the form of pentagonal dodecahedrons ; and he suggests that surely Nickles could not be mistaken on a point of crystalline form, though the eminent crystallographer, Gustav Eose, who doubts the statement of Nickles, might be mistaken. We have obtained crystals of zinc in small spheiical aggregations, which, as far as a judgment can be formed from description, appear to be identical with those described by Nickles. They are now in the possession of Professor Miller of Cambridge, who will probably be able to decide the point in question, and who will, if I mistake not, pronounce a judgment in favour of Eose. During the process of stamping brass that is, of subjecting it to heavy blows in dies, as is done in the manufacture of various articles used in modern decoration it is requisite to anneal the metal from time to time. After the completion of the stamping, the article will remain discoloured with adherent oxide formed in the annealing pro- cess. This is easily detached' by plunging the metal in aquafortis of suitable strength, and then washing with water. A bright metallic surface is thus produced, ready to receive the usual coating of varnish or lacquer. This process of cleansing the surface is known by the name of " dipping." If the brass contains certain impurities, it will not "dip" that is, it cannot be made to acquire a bright Mirface by the application of acid. The colour produced by dipping may be varied by the use of aquafortis (nitric acid) of different strengths. This, I presume, will depend upon the component metals being dissolved in different ratios by acids of varying strength. An alloy consisting of about 95 per cent, of copper, and 5, or somewhat less, of aluminium, acquires, by suitable "dipping" in nitric acid; a colour so closely resembling that of fine gold, that I think it scarcely possible to distinguish one from the other, even when seen side by side. We ascertained this fact several years ago in the Metallurgical Laboratory ; and I have frequently mentioned it in public, and exhi- 2 R 610 BUASS DEAD-DIPPING. bited illustrative specimens of it ; yet I do not find that it has received any attention. Unfortunately the beautiful gold-coloured surface thus produced tarnishes, and can only be preserved by a covering of varnish or lacquer. The process called "dead-dipping" is extensively prac- tised in Birmingham, where I understand it was introduced not many years ago. I saw it practised (Oct. 1854) in the following manner at large brass-works, where it was accidentally discovered : Dead-dipping. This term is applied to the process of producing an agreeable pale yellow dead surface on brass works, such as curtain bands, &c. The brass work, with the adherent black scale produced in annealing after the final stamping, is " pickled " in dilute nitric acid or aqua fortis (say 1 part of acid to 7 or 8 parts of water) : the precise strength is not particularly attended to ; acid which has been em- ployed in the subsequent part of the process being put aside for the purpose. It is left in this acid " pickle " until the scale maybe easily detached by gently rubbing the surface 'with the finger. It is then taken out and immediately washed by moving it about in water; after which it is put into much stronger acid, and left in it until the surface of the metal becomes white with " curd," an appearance which is due to a stratum of small bubbles of gas uniformly diffused over the surface. No precaution is taken to ensure the use of acid of constant strength : it may be about twice as strong as that used in the first step of the process, or may consist of about 1 part by measure of acid and 4 of water. I saw several pieces of stamped work fastened or strung together by a piece of brass wire, and treated as above described ; I was informed, however, that each piece ought to be dipped separately, but that the workmen would be thus careless and operate upon several pieces at a time in spite of every effort to prevent their doing so. As soon as the white stratum or " curd " has properly formed, the brass is washed in water and then roughly dried by moving it about in cold sawdust. It is next dipped, with adherent particles of sawdust, in strong nitric acid for a few seconds, washed in water, and immediately afterwards washed again in water, not heated, containing argol or impure cream of tartar dissolved, which done, it is dried in liot sawdust. The argol solution is said to prevent a brownish discoloration or mottling of the surface which would otherwise occur. This mottling, moreover, is produced if the drying sawdust is not hot enough. The sawdust is kept heated in a large quadrangular iron pan, placed upon a suitable brick stove. Although the workmen are in the habit of subjecting several pieces of work together to the first and second dippings, yet, in every case, in the last each piece must be dipped separately. The dipped surface is now ready for lacquering. By casting copper, heated to a certain temperature, under water, I have obtained an ingot coloured superficially so exactly like brass that any one would mistake it for this alloy. QUALITIES OF VARIOUS ALLOYS OF COPPER AND ZINC- .dddd .fa .d .d .d . .d . . .d . . .d_.d 'd +++++++++++++++++ ++++-H+++ + ++ +4-++ + + + + + R 012 MANUFACTURE OF CALAMINE BRASS. Observations on the preceding Table. The alloys employed by Mr. Mallet were made in a closed, bent, wrought-iron tube, coated and lined with porcelain clay and plumbago. The zinc was gradually brought in contact with the copper : the apparatus excluded air, and was continually agitated, until the alloy was poured into a mould of cast-iron, into which it was cast into a long strip, which solidified instantly. About 7 Ibs. weight of each alloy were formed at once ; and the constitution of each, where any cause of doubt existed, was verified afterwards by an assay. Mr. Mallet states that the addition of only of an equivalent of zinc to the yellow and tough alloy, 2Zn + Cu, renders it white and extremely brittle. It will be found that the specific gravities of the alloys do not increase in regular proportion to the amount of copper which they contain, though their general tendency is to do so : the greatest disturbances occur in the series from 8Cu-fl7 Zn to 8Cu+23Zn. The specific gravities were taken on the alloys just as they were cast, and suddenly cooled in the cast-iron mould ; but it was found that on rolling some of them condensation took place in veiy variable degrees, from which Mr. Mallet infers that sudden cooling produces an effect on the alloysjof copper and zinc analogous to tempering in cold water. PREPARATION OF BRASS. Until a comparatively recent period all brass was made by the old process of cementation, which has been almost entirely superseded by that of alloying zinc in the metallic state directly with copper. This ancient process, which had been practised for centuries, was, never- theless, patented in this country in 1779, by Mr. John Champion, senior! 9 MANUFACTURE OF CALAMINE BRASS. Only a few years ago I saw the old process carried on in Birmingham at Mr. Pemberton's works, the last which survived in that town ; but they have since been pulled down, and I am not aware whether a single calamine brass furnace is now in operation in England. In 1859 I found several of these furnaces still in existence at the copper- works of Messrs. Sims, Nevill, and Company, at Llanelly, Glamorgan- shire, and there are others, I am informed, at Swansea. I am indebted to the kindness of the firm just mentioned for permission to examine them, and take the measurements from which the accompanying wood- cuts have been prepared. I have also pleasure in stating that I received much of the following information concerning the mode of conducting the process from the late Mr. Joseph Stringer, who had been during a long period in their employ, and who had superintended the furnaces. 9 A.D. 1779, Nov. 24. No. 1239. Abridgments-Metals and Alloys, p. 22. MANUFACTURE OF CALAMINE BRASS. 613 Description of the furnace. It exactly resembles, in all essential points, the furnaces represented in the engravings contained in old metal- lurgical treatises. It consists of a circular chamber w, lined with fire- brick, d d ; it is contracted above to a circular opening, the mouth, in which is fixed a cast-iron collar, e e ; it is closed at the bottom by a circular cast-iron plate or bed-plate, a a, in which are 12 holes sym- \<'\g. 151. Vertical Section on the line A 13, fig. 152. <__ __ s _^^__^ ^__ Fig. 152. Horizontal Section on the line C I), fig. 151. metrically arranged round one larger hole in the centre, k, through which the ashes and clinkers may be withdrawn from time to time ; below the bed-plate is the ash pit, n, communicating in front by means of an arched air-way, c, with a long arched passage or vault, ?", through which air is conveyed to the furnace from the outside, and the workmen obtain access to the ash pit. Over the small holes in the bed-plate are 014 MANUFACTURE OF CALAMINE BRASS. placed short nozzles or twyers of cast-iron, //, tapering upwards, the larger central hole being left without a nozzle. The nozzles are 6 inches high, 2-y inches in diameter at the bottom, and 1 inch at the top, inside measure, and f inch in thickness. The space between the nozzles is filled up level with bricks and fire-clay, so as to form a solid bed 6 inches thick. The air which sustains combustion enters through these nozzles, which are used as a substitute for bars. The furnace is enclosed in a solid mass of brick- work, b b ; and for the sake of strength, at the back of the arch, h, over the air- way, c, and in front of the bed-plate, is placed an iron bar resting upon the walls, I' //, which form the sides of this air-way. Several of these furnaces are con- structed in a row, and over the whole is built a room of brick covered in with a large brick cone, open at the top, exactly like an ordinary glass-house. 1 It will be perceived that the furnace has no chimney, except the mouth, which is kept more or less closed, according to cir- cumstances, by a circular cover of cast-iron, or of fire-brick set in an iron frame. In the accompanying woodcuts, the furnace is shown as widening somewhat to the height of 13 inches above the upper surface of the bed-plate. The furnace-builder, whom I interrogated', stated that it did not thus widen upwards, but Mr. Stringer stoutly main- tained that it did ; and I could not succeed in determining this point for myself, which is not very material. Crucibles. The crucibles employed are made of fire-clay. They are circular, and 12 J in. deep, 81 in. wide at the top, and 6} in. in the middle, inside measure ; they are 1 in. thick at the top, and 2 in. thick at the bottom. Mr. Stringer informed me that he was always accus- tomed to employ a slightly larger crucible, which he termed the king- pot, in the centre of the bed-plate ; but another calamine-brass maker of great 'experience tells me that there is no necessity for this variation in the size of the crucibles. According to Mr. Stringer, the king-pot should be 13-J in. deep, 8J in. wide at the top, and 7f- in. wide in the middle ; lj in. thick at the top and 2 in. thick at the bottom. Such a crucible will hold 120 Ibs. of metal, while one of the smaller size will only hold 84 Ibs. Composition of the charge. Calcined calamine or blende calcined sweet and ground fine, 100 Ibs., ground coal 40 Ibs. These are intimately mixed dry, and passed through a sieve of 8 holes to the linear inch. The mixture is then spread level, and 2 gallons of water are poured upon it ; after standing half an hour, it is again well mixed and passed through a sieve of 4 holes to the linear inch, after which it is levelled and thoroughly mixed with 6$ Ibs. of lean-shot copper. The mixture should be sufficiently moist to adhere together by pressure with the hand. It is now ready for charging, and produces brass for the best battery purposes, i.e. brass intended for hammering, &c. Mode of charging. It is supposed that the furnace has become pre- viously heated in the regular course of working. The pots are charged with care moderately lightly, covered, and arranged in the manner 1 See description of the English Zinc Furnace, p. 553. MANUFACTURE OF CALAMINE BRASS. 615 shown in fig. 1 52 ; the central hole in the bed-plate having been pre- viously stopped with clay. Four flat pieces of coal, each about 15 in. long, from 3 to 4 in. thick, and 8 or 9 in. wide, are placed so as to form a cross over the pots, one end of each piece resting on a side-pot and the other on the king-pot. 3 cwt. of coal broken in pieces a little larger than the fist are then carefully put into the furnace, the fall of the coal being broken by placing a pair of tongs in the mouth, which is afterwards partially closed. The coal is thus left to burn until " all the gas is out of it," which will require about 1-J- hour, when the cover is placed over the mouth to within \ in. on one side in order " to harden the coke " formed. After this, the coked coal is cautiously poked down amongst the pots, care being taken to keep the draught holes open and clear. The cover is now placed to within 1 \ in. on one side of the mouth, and the furnace may then be left during 3 hours without a further supply of coal. It is the duty of the foreman to see that the draught holes are properly cleared and the pots kept covered with coke. The heat is gradually raised in the course of the process by removing the cover a little on one side. The process, if properly attended to, will be completed in about 10 hours. The brass which has been formed is now to be collected. The king- pot is taken out and its contents are well stirred with an iron bar having a flat or paddle-end. One of the side-pots is next taken out and treated in the same manner, when the brass, which has subsided to the bottom, is poured into the king-pot. The other side-pots are taken out in succession, and the same process repeated with each. The king-pot is well shaken with the stirring-bar during the pouring. The brass from all the side-pots having been thus collected in the king- pot, the metal in the latter is skimmed and poured into iron ingot moulds. It is scarcely necessary to remark, that as soon as a side-pot has been deprived of its brass, it should be replaced while hot in the furnace. An old calamine brass-maker informs me that good pots lasted on an average 16 days, not being allowed to cool during that period. \\lieii brass of inferior quality is to be made, the weight of the copper in the charge above stated is reduced 1 lb., and spelter is added until the desired quality is attained. The spelter should be broken in small pieces and put into the bottom of the pot before charging with the regular mixture. It is important in the manufacture of calamine brass, that the king or receiving pot should always be the cleanest, and have as little matter adhering to its sides as possible. The paddle is pushed down round the sides, so as to get the metal together, which cannot well be done if there is much concretionary matter round the inside ; and by gently bumping the bottoms of the pots on the floor, the subsidence of the metal is promoted. 2 2 I may mention that, notwithstanding I tion during long periods without inter- the incessant care which a calamiue brass i mission, the occupation does not appear furnace requires, as it is kept in opera- | to be injurious to health. I knew an old 616 MANUFACTURE OF CALAMINE BEASS. The following details relate to Mr. Pemberton's calamine-brass-works formerly in operation at Birmingham. The diameter of a furnace was 3 feet 6 inches at the bottom, and the height 3 feet 6 inches to the collar. Stourhridge clay crucibles were employed, 12 inches deep and 8 inches in diameter. Each furnace contained 9 such crucibles. In later times only two qualities of brass were made, which were distin- guished by the marks B. I. and B. XX. The charges were as follows : B. I. B. XX. Ibs. Ihs. Feathered-shot copper .................... ......................... 64 61 Ground and sifted calcined calamine from Somersetshire 88 97 bushel. bushel. Coal-slack ............................................................ 1 1 Cost of making calamine-brass in the last century. The following details, extracted from Mr. Morris's Journals previously referred to, may now be interesting to brass-makers for the sake of comparison : BRASS HOUSE CALCULATION IN 1781 (AT THE FOREST WORKS). Dr. cwt. qrs. Ibs. 9 . s. d, cwt. qrs. Ibs. . s. 1 20 Coppef shot aU 40 10 1ft By 14 16 Brass at 90 > po 19 86 per ton J 4 per ton f 6 16 Calamine at 6) , , n 10 Ibs. of do. in skimmings \ per ton J at 45 per ton / V% bags of charcoal-dust at I n 30s. per dozen I 33 bags of coal at 35s. per wey 015 4 Jno. Spears' week's wages... 15 His brother do. 12 Helper 060 Wear of furnaces, tools, 63 16 6 grinding and dressing ca- lamine, charcoal, fern, &c. 076 47 17 9 If copper at 96 per ton, add 4 14 3 52 12 If the brass is made richer) by using 6 Ibs. more cop- 1 5 14 4 per every charge j 58 6 4 The copper is estimated to increase 50 per cent, in weight ; and the brass, which is composed of 2 parts of copper and 1 of zinc, to be equal to the weight of the calamine used. calamine brass furnace-man at Mr. Pem- berton's works who had been constantly employed in the business during sixty years, having commenced when a boy ten years old. He had scarcely had one continuous night's rest out of five nights every week during that long period, and yet he appeared vigorous, and by no means unhappy. MANUFACTURE OF CALAMINE BRASS. 617 THE BRASS HOUSE CALCULATION IN 1784. Dr. Cr. cwt. qrs. Ibs. . s. d. cwt. qrs. Ibs. . s. d. 14 2 26 Copper shot at) -,, 1Q _ By 22 11 brass at 05s. l 71 lp 80 per ton f 58 7 per cwt } /l 22 11 Calamine at I ~ 7 -, 5 15s. per ton 1 12 bags of charcoal at 30. I i ^n Q per dozen ) f wey of coal at 35s. per wey 163 Workmen's wages 1 4 Wear of furnaces, tools,) grinding and dressing ca-> 130 laiiiine, fern, &c ) 70 8 11 Profit... I 7 5 The furnaces in one week 71 10 4 In this process oxide of zinc is reduced at a temperature below the melting-point of copper, which, being thus exposed to the action of the vapour of zinc, becomes permeated with this metal and converted into brass. Care must be taken so to regulate the temperature that the copper shall not melt, but remain diffused through the mass of the charge ; for if it were allowed to melt, it would trickle down to the bottom of the pot, in a greater or less degree, and much of the zinc would then escape. By exposure to the vapour of zinc, copper may be converted into brass even below the melting-point of this alloy. In illustration of this fact, the following pretty experiment may readily be made : A little zinc is placed at the bottom of a clay crucible, and covered with a layer of coarsely -pounded fire-brick or burnt fire-clay ; a copper coin is then introduced and surrounded with coarse charcoal-powder ; after which the crucible is closed with a luted cover, and exposed during a considerable time to a gentle red-heat. The surface of the coin will by this means be converted into yellow brass, without oblite- rating the effigy and other characters upon it. In experiments of this kind which we have made, the surface has always had a crystalline or frosted appearance. Calamine-brass was formerly used by button-makers in the manu- facture of gilt buttons, which were gilt by the old process of water-gilding, i. e. by means of mercury (lucus a non, &c.) a designation which would be more appropriate to the modern method of electro-plating. It was preferred for this purpose, because it was said to receive the gold better than brass made from spelter ; and to " stand the soldering better," to which these buttons were subjected in attaching the, shanks. It was also specially used in making the wire-gauze employed in the sieves of papermakers. A thoroughly practised brass-worker in Birmingham most positively maintains to me that he can immediately distinguish calamine-brass from common brass by the peculiar appearance of its polished surface. "Why calamine-brass should differ from common brass, I do not at present understand ; but that such a difference actually existed when the former kind of brass was largely produced, can hardly be denied. Indeed so impressed are some of the Birmingham 618 DIRECT PREPARATION OF BRASS. brass-founders with the fact of this difference, that quite recently I know one large firm has applied to an establishment in Glamorgan- shire for calamine-brass. Perhaps the difference between the two kinds of brass would not now be found, if the manufacture of calamine-brass were resumed ; and that which formerly existed may have depended upon inferiority in the quality of the zinc produced at that time. Calamine-brass, I believe, ceased to be manufactured in Birmingham because its price was sensibly higher than that of common brass. I have met with a statement to the effect that a Mr. Champion ob- tained a patent, about the year 1818, for making brass by exposing plates of copper to the vapour of zinc below the melting-point of brass ; a but in the Abridgments of Specifications relating to Metals and Alloys, I do not find any record of this patent. It has been supposed that the remarkably malleable brass of Nuremberg was produced by a similar method. DIRECT PREPARATION OF BRASS. This is effected either in crucibles, as in ordinary brass-foundries, or in reverberatory furnaces, as in the manufacture of yellow-metal sheathing. The crucibles employed for this purpose have been previously described. The zinc is gradually and cautiously added to the copper when the latter "has just melted. The ingots of copper, before being put into crucibles, should be heated to redness. The furnaces in use in Birmingham are about 10 inches square and 2 feet deep, while those which I have seen in London are round. The flue leading to the stack should be small, and close to the top of the furnace ; but its size must obviously vary with the stack and other conditions. The fuel should be good coke, . and not such as frequently contains a large quantity of corrosive ash. The metal, when well melted, is skimmed and poured into sand-moulds for castings of various kinds ; or, when intended for rolling, into closed iron ingot-moulds, previously warmed, lightly oiled and dusted over with charcoal in the interior. In former times moulds of granite were used for casting ingot-brass. In the making, casting, and remelting of brass, there is always an inevitable loss from the volatilization of zinc, for which a due allowance is made to the founders when they deliver the metal. The Chinese appear to be unacquainted with the art of rolling brass, and, as substitute, cast it into tolerably thin sheets. I have a specimen of one of these, rather exceeding T V inch in thickness, which 1 re- ceived from my friend Harry S. Parkes, so well known in his official capacity in China. It has been analysed in my laboratory by Mr. T. Philipps, and found to have the following composition : Copper 56-59 Zinc ; 38-27 Lead 3.30 Tin !. 8 Iron 1-4.7 100-71 Manuel s-Roret. Alliages Metalliques, Paris, 1839, p. 169. MUNTZ'S METAL. 619 Muntz s metal. This alloy, and its application " for sheathing the bottoms of ships and other such vessels," was the subject of a patent granted to the late George Frederick Muntz, of Birmingham, in 1832. 4 The proportions specially recommended in the specification are 60 per cent, of copper and 40 of zinc ; but these proportions may be varied from 50 up to 63 per cent, of copper, and from 50 down to 37 per cent, of zinc. Best-selected copper and foreign zinc are directed to be used. The metal is cast into ingots and rolled while liot into sheets, which, when finished, are " pickled " in sulphuric acid diluted with water to free them from adherent scale, and afterwards washed in water. In the same year Mr. Muntz obtained a second patent for " an improved manufacture of bolts and other the like ships' fastenings." 5 Precisely the same proportions of copper and zinc are claimed in this patent as in the first. In 1846 a third patent was granted to Mr. Muntz for the use of an alloy consisting of 56 per cent, of copper, 43^ of zinc, and 3f- of lead. 6 In the specification it is directed that only the purest metals should be used, and that the alloy is to be cast into ingots, which are to be rolled at a red heat, and treated in other respects in the manner stated in the specification of the first patent. I am not aware whether the alloy last described has ever been manufactured and applied ; but my impression is that it has not ; and I shall, there- fore, dismiss it from further consideration. I may state that I have succeeded in rolling brass well, which, on subsequent analysis, was found to contain not less than 8 per cent, of lead. The theory assigned by Mr. Muntz for the application of his alloy is, that by exposure to sea-water the zinc is slowly and uniformly corroded over the entire surface, whereby the attachment of barnacles, &c., is prevented. Mr. Faraday informed me that in a specimen of sheathing formed of the alloy in question, which had long been ex- posed to the action of sea-water, he found no zinc remaining. Ex- perience, especially of late, has certainly not confirmed the statement concerning uniformity of corrosive action, as much of the modern sheathing is eaten away in holes, notwithstanding the declaration of copper smelters that in the manufacture of yellow metal they employ only best-selected copper and zinc of the best quality ! Muntz's metal, or yellow-metal sheathing, has entirely superseded copper-sheathing in the merchant service, though the latter is still retained in the Navy. Its special advantages are stated to be, that it keeps the bottoms of ships cleaner and costs considerably less than copper-sheathing. It is now generally made in reverberatory furnaces, the zinc being cautiously added to the melted copper. The melted metal is tapped into a vessel lined with clay, out of which it is laded into suitable closed iron ingot-moulds, the interior of which has been lightly oiled and dusted over with charcoal in the usual manner. Just previously 4 A.D. 1832, Oct. 22. No. 6325. Abridgments of Specifications relating to Metals and Alloys. 5 A.D. 1832, Dec. 17. A.D. 1846, Oct. 15. No. 6347. No. 11410. 620 MUNTZ'S METAL. to tapping, samples of the alloy are taken out, in the same manner as copper-proofs in the process of refining copper, and cast into small ingots, which are passed through rolls while still hot, and are afterwards broken across, when, if the fracture presents the proper appearance, the metal is tapped out forthwith. The fracture should be close and finely granular; but if it does not present the proper appearance, zinc is thrown into the furnace and well mixed with the alloy, after which the fracture is again examined, and if it is right, lading takes place immediately ; but if not, the process of adding zinc and the testing of the fracture must be repeated until the desired quality is attained. The eye of the furnace man requires to be educated for this kind of examination. Although the proper quantities of the two metals may have been put into the furnace in the first instance, yet, from the very nature of a reverberatory furnace, it is impossible to calculate upon the precise amount of zinc which may be volatilized, even in the same furnace at different periods. Hence the necessity of the testing, &c. above described. But as the usual charge of a furnace consists of copper, " new scrap," and old yellow sheathing, of which the average composition is not exactly known, it becomes all the more necessary to follow the course above described in order to produce an alloy of the right quality. I am informed by an experienced yellow metal manufacturer, that the proportion of zinc should not exceed 38 per cent. that if it sensibly exceeds this proportion, the sheathing is apt to become friable and that if it is sensibly below this proportion, it wears away too rapidly. The rolled sheets, after final annealing, are immersed in dilute sulphuric acid, scoured on the surface with flannel and sand, and after- wards washed and dried. I am assured that cast yellow-metal nails of the same composition as the sheathing cannot be used for attaching it to the bottoms of ships. The copper sheathing in the Navy is attached by nails having the following composition : Copper 80-82 Tin 9'30 Zinc 3-88 100-00 Mr. Muntz, like most successful patentees, had to encounter opposi- tion on the part of certain copper-smelters, and to defend his patent- rights in the Courts of Law. He succeeded in obtaining a signal victory over his opponents, which certainly has not always been the fortune of patentees who have benefited either themselves or the world by their inventions. At the expiration of the patent in the' ordinary course of 14 years, Mr. Muntz applied to the Privy Council for an extension of it, when he admitted, if I remember correctly, that it had yielded him a profit of not less than 68,000?. The application was refused. A few years afterwards Mr. Muntz died, and his property was sworn under 600,000?. ! The manufacture of the alloy is still con- ducted on a very large scale near Birmingham, by one or more of his MOLECULAR CHANGES IN BRASS. 021 sons. There are but few if any metallurgical patents which have been so profitable to the patentee as that of Mr. Muntz. Most of the large copper-smelters are now engaged in the manufacture of Muntz's metal. MISCELLANEOUS OBSERVATIONS ON BRASS. Brass intended for rolling. It is stated, and I believe correctly, that the presence of antimony is especially injurious to brass intended for rolling, as it renders the metal brittle and very liable to crack. Brass intended for turning. It is usual to introduce a small quantity of lead into brass intended for this purpose, in order that the turnings may " leave the tool easily." About 3 ozs. of lead are added to 10 Ibs. of brass ; and this addition should be made after the crucible contain- ing the melted brass has been taken out of the furnace for casting, the metal being afterwards well stirred. Chaudet gives the following analysis of brass fit for lathe-work and filing : 7 Copper 65-80 Zinc 31-80 Lead 2 15 Tin 0-25 100-00 The tin is supposed to be accidental. Brass for engraving upon. A little tin is a good addition to brass intended for door-plates, as it gives crispness, and causes the metal to " break up short " under the graver. Brass for the use of braziers. It should not contain less than 66^ per cent, of copper. Composition of a variety of brass from Paris. It was exhibited in the Paris Exhibition of 1855; it was in the form of an ordinaiy table- spoon, and was thinly electro-plated with gold, which peeled oif under the action of nitric acid ; the surface of a portion was completely re- moved by filing, and the remainder analysed in my laboratory by Mr. Burbidge Hambly : it had the following composition : Copper 87-83 Zinc 12-44 Iron 0-35 100-62 Electro-deposition of brass. This is now extensively practised in Bir- mingliam. Molecular changes in brass. Some kinds of brass wire become ex- tremely brittle in the course of time, especially if subjected to vibra- tion. I have seen thick brass wire become almost as brittle as glass in the course of a few weeks, after having been kept extended and sub- jected to vibration. Brass chains used to siipport objects, such as chan- deliers, INDEX. BRASS, defined, 606; history of, 521, 606; manufactured in England, 527, 606. , its valuable qualities, 607. , malleability of, 608. . , crystals of, 609. , process of lacquering, 609, 622 ; of dead dipping, 610. , table of qualities of various alloys of copper and zinc (from Mallet and Karsten), 611 ; observations, 612. , preparation of brass ; manufacture of calamine brass, 612 ; cost of making it in the last century, 616. , direct preparation of, 618. . , Muntz's metal, 619. , miscellaneous observations, for rolling, turning, engraving, &c., 621. , from Paris, its composition, 621. . , electro-deposited, 621. , molecular changes in, 621. Brass-wire, annealing of, 622. casting, defects in, 622. foils, 623. Breeze, 1 14. Breeze-oven, 186. Bricks, fire, 235 ; blue, 240. Bronze, ancient, 519. Buck, Jeremy, his patent for coke-making, 144. Bucker (in copper-smelting), 330. Budd, J. P., his patent for coking, 191. Buhring's patent for coking, 190. Burette (or alkalimeter) for wet assay of cop- per, 480. " Burnt leavings" (copper), 322. Button (in copper assaying), 473, 475. C. u Cadmia," a term used by Pliny, considered, 524. Cadmium found in zinc-blende, &c., 549, 601 ; in commercial zinc, 590. Caking-coal, 91 ; composition of British, 99; of foreign, 100. Calamine (carbonate of zinc), abundant in England, 527, 549 ; of Wiesloch, 548. - , zinc obtained from, 521. Calamine brass, manufacture of, 612 ; highly esteemed, 617 ; works in England, 527. Calcination (of copper-ore), 323, 342. Calciner (in copper-smelting), 314 ; in zinc- smelting, 567. Calcining rods, 457. Calorific power and intensity of fuel, 53-59. Calorimeter, Rumfoi'd's, 54. Calvert's method of desulpluirization of coke 196. Cameron, J., on oxide of tin in copper slag, o 1 4-. Cannel coal, 96 ; its composition, 105. Carbon, reduction by, 14. - and carbonic oxide, calorific power of, 55. - presence of, in copper, 269 ; heated with COBALT. disulphide of copper, 257 ; presence of, in zinc, 546. Cai'bon crucibles, 234. Carbonates, alkaline, heated with oxide of zinc, 539, 545. Carbonate of ammonia employed in assaying zinc-ores, 601. Carbonic acid, sulphide of zinc heated in, 544. Carbonic oxide, its calorific power, 55 ; fur- naces for utilizing it, 198. Carinthian process of extracting zinc, 585. Champion, J., his patent for extracting zinc, 520. Charcoal, 107; absorbs gases, 108 ; analyses of it by Bunsen, Playfair, and Ebelmen, 109; by Violette, 110. Charcoal-burning, various modes, 111 ; in Bavaria, 115; in Sweden, 117, 125; in China, 123. results of, 131 ; summary of directions, 133 ; theory of, in circular and rectangular piles, 134;" cost of, 141. , yield of, 128; by volume, 129; by weight, 130; influence of temperature on yield, 131. Chark, 144. Chevandier's analyses of woods, 63 ; of wood ashes, 68. Chili, copper-smelting in, 331. Chimneys, difficulty of applying formulae for their construction, 207. China, zinc brought from, 520; and rolled brass, 618, Chloride of sodium (salt), used in copper assaying, 460. Choubine's analyses of cupriferous sandstone and pig-iron, 433, 434. Chrysocolla (hydrated silicate of protoxide of copper), 312. Church, Jabez, his patent for coking, 190. Claridge and Eoper's patent for desulphuriza- tion of coke, 195. Clay nozzles or condensers (in zinc-works), 561, 579. Clays described, 208 ; analyses, 209. Clinker, analysis of, 98. Coal, its various definitions, 78 ; analyses, 80, 147 ; errors in, 84 ; compounds gene- rated by its destructive distillation by Hof- mann, 147 ; by Neumann and others, 191 ; evaporative power of coals, 206 ; ashes, 82 ; composition, 106 ; lignites, 85 ; bitu- minous coal, 91 ; caking coal, 91 ; waste of coal, 95 ; free burning and cannel coals, 96 ; fibrous and granular matter in coal, 96; composition of coal used in copper- smelting, 97 ; of British caking coal, 99 ; of foreign, 100 ; foreign non-caking coal, composition, 104 ; composition of ashes, 106; metals in, 106. " Coarse copper," fusion for, 467, 472. metal (copper), 323 ; characters, &c., 342, 345 ; James's patent for melting it with non-sulphuretted ores of copper, 381. Cobalt eliminated from copper, 375. INDEX. 027 COBRE. (Jobre ore and dust (copper), 322. Coke: history, 144; properties, 146; com- position, 146 ; preparation, 147 ; desul- phurizatiou of, 194. Coke-ovens, 157; Cox's, 159; Jones's, 162; Appolt's, 167; Seraing, 182; Davis's breeze-oven, 186. Coking: in circular piles, 149; in ridges, 1 52 ; in large open rectangular kilns, 152 ; theory of, 155. of non-caking coal slack, 189. , products from, 191. , cost of, 196. Coloration-test (in assaying copper), 487. Colours of metals, 1 2. Condensers (in zinc- works), 561, 573. Condensing-tubes (for extracting zinc), 551. Conduction of heat and electricity (table), 9. Cooke, Professor, on crystals of zinc, 528. COPPER: history and physical properties, 241. , native, 309 ; of Lake Superior, 309. , specific gravity, 283 ; electric conduc- tivity, 287. , action of oxygen on, 242. , commercial, its composition, 503. , Egyptian and Indian coins, 504. , dioxide of, 242 ; heated with silica, 243; with silica and alumina, 244 ; with protoxide of lead, 253; with protosulphide of iron and silica, 254. , protoxide of, 243 ; heated with silica, 244 ; with silica and alumina, 245 ; with metallic lead, 253 ; with protoxide of lead, 253 ; with sulphide of lead, 254. , borates of, 245. , disulphide of, 246; heated with other sulphides, 246 ; with access of air, 247 ; in admixture with dioxide, protoxide, or sulphate of copper, 249 ; exposed at high temperatures to hydrogen, 255 ; to vapour of water, 255 ; heated with carbon, 257 ; with iron, 257 ; with zinc, 258 ; with lead, 258 ; with tin, 259 ; with antimony, 260 ; with protoxide of lead, 261 ; with sulphate of lead, 262 ; with nitre, 262 ; with caustic soda, 262 ; with carbonate of soda, 263 ; with baryta or lime, 263 ; with cyanide of potassium, 263. heated with protoxide of lead, 250 ; sulphate of lead, 252 ; sesquioxide of iron, 252 ; peroxide of manganese, 252 ; sulphide of zinc, 544. exposed to the action of vapour of water at high temperatures, 265. and dioxide of copper (diy copper), 264 ; cold- short, red-short, and tough pitch copper, 266. and carbon, 269. , electrotype, 271. , overpoled, 273. and nitrogen, 278. and phosphorus, 279. and arsenic, 281. and silicon, 282. and iron, 473. COPPER. COPPER and zinc, 600 ; table of alloys, 611 . black, analysis of, 422. " hard," for sheathing, 506. COPPER-ORES, mode of dressing and sampling in Cornwall, 300; sale of, 301 ; at Swan- sea, and circular, 302; at Kedruth, and circular, 305; " Settled list," 308. , native copper, 309 ; red oxide, 309 ; black oxide, 309 ; green carbonate (mala- chite), 309 ; blue carbonate, 310 ; vitreous or grey sulphide, 310 ; purple copper, 310 ; copper-pyrites or yellow copper-ore, 310; grey copper-ore, 311; chrysocolla, 312; atacamite, 312; of Devon and Cornwall, 313 ; foreign metals in, 477. ASSAYING: by the Cornish method, 454, 462 ; furnaces and implements, 454 ; fluxes and reagents, 458 ; sampling, 461 ; preliminary examination, 461 ; chaiacteris- tics of the process, 462 ; proportions of fluxes, 463 ; classification of ores and cu- priferous products, 467. , practical directions for conducting the process, 468. , special modes for particular ores, &c., 476. , loss by the Cornish method, 478. , methods of estimating copper by vet assay, 478 ; by cyanide of potassium, 479 ; by precipitation with hyposulphite of soda, 485. , coloration-test, 487. , inaccuiacy of Cornish method of assay- ing, 489. , table of comparative results by the Cornish and wet methods, 490. pyrites, assay of, 464. "rain," 411; amount of copper in, 495. schist, smelting of, 413 ; of Hesse, 426. COPPER-SMELTING in Great Britain, 289 ; profits, 306 ; list of smelters in England and Wales, 307. , loss of, 496. profits in last century, 501. Welsh process of smelting, 314 ; fur- naces employed ; calciner, 314 ; melting furnace, 318 ; reactions during process, 332 ; calcination or roasting, 342, 401 ; composition of coarse metal, 342 ; of ore- furnace slag, 345 ; white metal, 347, 361 ; bine metal, 348, 361 ; moss-copper, 359 ; blister copper, 361 ; pimple copper, 362 ; refining, 366; " best selected " process, 364, 367; elimination of foreign metals, 369. in six operations, 322. at Hafod works, 322. modifications of Welsh process, 326. in Chili, 331. , various alleged improvements: by James, Troughton, Budd, Morgan, Kapier, Rivot and Phillips, and H. Vivian, 382, 386. in blast furnaces, 387. in Sikkim, 388 ; furnaces, 389. 2 s 2 628 INDEX. COPPER. COPPER-SMELTING at Singhana, India, 391. in Japan, 392. in Sweden, 395 ; furnaces, 395 ; calci- nation or roasting, 401; fusion of roasted ore, 402; fusion for black copper, 405; refining, 406 ; toughening, 409 ; fuel, &c., 411,412. in Norway, 411. in Prussian Saxony, 413. in Hesse, 426. in Perm, in Russia, 433. , kernel-roasting, 439 ; in piles, 441 ; in Styrian kilns, 441 ; changes of the ore in kernel-roasting, 444 ; theory, 446. wet methods of extracting copper, pre- cipitation of copper from solution by iron, 447 ; Bankart's process, 447 ; Escalle's, 450 ; Hahner's, 451 ; Henderson's, 451. , loss of copper in smelting, 494. , commercial details : freights, 496 ; weights, 497 ; cost of Welsh method, 499 ; Logan's formula, 499. Copper-sheathing, 505 ; first applied, 507 ; corrosion of, 505, 624; amount of cor- rosion in different kinds of copper on 34 vessels, 508 ; extracts from Dock-yard Reports, 509 ; adulteration of copper, 510 ; trial of copper alloyed with tin, 512. , effects of foreign matters upon copper- sheathing, 514 ; results of Sir H. James's experiments, 515 ; pi'otective influence of phosphorus, 515. in Dutch navy, 516. Copper-smoke, injurious to vegetation, 340 ; proposals to utilize it, 341. " Copper-sponge," 451. Copper-works, cost and capital required, 499. Cornish method of assaying copper-ores, 454 ; its inaccuracy, 489 ; results of compared with those of wet method, 490, 492. copper-works, 296. crucibles, 221. Corrosion of copper-sheathing, 505 ; amount of, in various kinds of copper, 508 ; causes almost unknown, 512 ; difficulty of ascer- taining causes, 513. Corrosive action of zinc-ore upon retorts pre- vented by admixture of tar, 594. Cory's patent for coking, 190. Cox's coke-oven, 159, 197. Crucibles, qualities required, 216; materials of, 208, 216 ; Stourbridge clay, 219, 220 ; Cornish, 221, 455 ; London, 222 ; Hes- sian, 223; French, 224; Belgian, 225; graphite, black lead, or plumbago, 225 ; carbon, 234 ; alumina, 234 ; very small, 228 ; lined with carbon, 229. used in making calamine brass, 614. Crucible-covers, 230. Crucible-stands and tongs, 231, 456. Crystallization of metals, 4. Crystals of copper, 241 ; of zinc, 528 ; of brass, 609. Cupel- pyrometer, 46. Cupriferous sandstone, 433 ; pig-iron, 434. FARADAY. Cyanide of potassium employed in assaying copper, 479 ; standard solution of, 480. heated with oxide of zinc, 539. D. Dalfors, charcoal-burning at, 125. Daniell, Professor F., on melting-points, 46 ; on corrosion of copper-sheathing, 505. Davis's breeze-oven, 186. Dawes's application of gases from decomposi- tion of steam by cai'bon, 203. Dawson's blocks of fuel, 175. " Dead dipping" (brass), 610. Demetrius erects calamine brass-works in Surrey, 527. Despretz's experiments on heat, 3. Deville and Debray melt 55 Ibs. of platinum, 3. Deville' s blast-furnace, 232. Deville on reduction of oxide of zinc, 535. Dillwyn and Co.'s zinc-works, process there, 558. Dinas fire-brick, 236. Dipping brass, 609 ; dead dipping, 610. Dioxide of copper (see under Copper}. Disulphide of copper (see under Copper). Dissolving- vats, for extraction of copper, 448. Distillation, 20; of coal, 141; of zinc, 569. Dockyard reports on copper-sheathing, ex- tracts from, 509. Dry methods of assaying (see Cornish method, and under Copper and Zinc). Ductility of metals, 7. Dundonald, Earl of, his patent for distillation of coal, &c., 193. Diinnstein, 421. Dutch metal or leaf (copper and zinc), 607. Ebelmen on slags, 24 ; on composition of permanent gases from charcoal piles, 136 ; on gases of coke-ovens, 175. Ecton copper-mine, 291. Egyptian copper, 504. Eisensau (in copper-smelting), 403 ; analyses of, 423, 431. Electricity conducted by metals, 9. Electric conductivity of copper, 287. Electric calamine (silicate of .zinc), 549, 596. Electrotype copper, 271, 275. English process of extracting zinc from the ore, 550 (see under Zinc). F. Fallowfield, W. , patent for use of charred peat, 142. " False silver," mentioned by Strabo, 518. Faraday, M., analyses gaseous products of INDEX. 629 FAVRE. copper-ore calciner, 337 ; discovers hydro- fluoric acid in, 371. Favre and Silbermann on calorific powers of fuel, 55, 57. Felspar, artificial, 425. Ferstl's analysis of peat-ashes, 76. Festus on " Cadmia," 526. Fire-bricks, 235 ; Diuas brick, 236. Fire-clays, 208 ; composition, 209-215. Flame, cause of its luminosity, 52. Fluorides and hydrofluoric acid evolved in copper-smelting, 371. Fluor spar as a flux, 43 ; used in assaying copper, 459. Flux and slag defined, 18. Fluxes, 43 ; proportions of, in assaying cop- per, 463 ; refining or white flux, 460. Forbes, D., on copper-smelting in Norway, 439, 444. Forbes, J. D., on conductivity of metals, 10. Fournet on blue colour of slag, 28. Fracture of metals, varieties of, 5 ; of copper, 242; of zinc, 529. Frankland, E., his analysis of gaseous mixture from action of steam on red-hot coke, 194. Freights of copper-ore, 496. Fremy on combustible minerals, 106. French crucibles, 224. Fuel defined, 51 ; calorific power, 53 ; table of, 58; calorific intensity, 59, 203, 205; classification of fuels, 62 ; concluding ob- servations on, 203. (See Wood, Charcoal, Coal, Anthracite, Coke, and Peat.} Fume (in. copper-smelting), 432 ; zinc-fume, 586. Furnaces, materials for, 208 ; Sefstrom's blast-furnace, 231 ; Deville's, 232. employed in copper-smelting, 314. , improvements by Troughton and others, 381 ; blast-furnaces, 386. for assaying copper-ores, 455. used in extracting zinc, 562, 580 ; (for zinc-fume), Montefiori furnace, 587. used in calamine brass-works, 613. Fusibility of metals, 2. G. Gases from charcoal piles, composition of, 136. from coke-ovens, 175 ; economic appli- cation, 178. , combustible, utilization of, 198. Gas zinc-furnace at Lydognia works, 596. Gay-Lussac on reduction by carbonic oxide, 18. Genssane on coal -distillation, 192. Gleiwitz, coking at, 152. Gloucestershire copper works, 292. Gold eliminated from copper, 378 ; 700 pounds' worth, 380. combined with copper, 474. Government establishments, waste of copper- slags, &c., in, 511. Graphite, black lead, or plumbago, its com- position, from different localities, 226. JAMES. Graphite crucibles, 227. Greek coins containing zinc, 521. Grenfell's copper-sheathing, report on, 506, 510. Grill, A., on charcoal-burning at Dalfors, 125. Grove, W. R., on patent laws, 452. H. Hafod copper-smelting works, 322. Hales on products of distillation of coal, 191. Hammergaar copper, 431. " Hard copper " employed in sheathing, ana- lysis of, 506. Haton on copper-smelting at Agordo, 440, 443. Heat, its action on metals, 2, 9, 11, 12; on zinc, 531. Henckel on extraction of zinc, 520. Hesse copper-schist, 426 ; slags, 428. Hessian crucibles, 223. Higgins, Dr., analyses of copper, 510. Hofmann, A. W., products of coal by de- structive distillation, 147. Horn coal, 96. Button, W., his account of brass manufac- ture at Birmingham, 607. Hydrated silicate of zinc (? Smithsonite), 549. Hydrogen, calorific power, 56. evolved by action of water on zinc, 533. , reduction of oxide of zinc by, 535. , sulphide of zinc heated in, 544. and hydro-carbons, utilization of, in iron-smelting furnaces, 203. Hyposulphite of soda employed in wet assay of copper, 485. Indian copper-smelting, 388, 391. copper coin, 504. Iron employed in extracting copper from solu- tion, 447. heated with oxide of zinc, 536 ; with sulphide of zinc, 543. and lime, sesquioxide of, a slag, 43. and sulphur, proportions to be oxidised necessary to obtain a proper regulus from yellow copper-ore, 464; from vitreous copper-ore, 466. combined with copper, 473, 504. with zinc, 591, 600. pyrites in copper-ore, 465. J. James, Sir H., experiments on combination of phosphorus with copper as sheathing, 514. 630 INDEX. JAMES. James's, J., raode of melting " coarse metal " (copper) with non-sulphuretted ores of copper, 381. Jars on separating gold from copper, 379. Japanese copper-smelting, 392. Jones's coke-oven, 162. Jones, E., patent for condensing products of coking, 194. Julien on extraction of zinc by Silesian pro- cess, 574 ; on cost of the process, 575. Kampmann's analyses of sand, 239. Kane, R., analyses of peat ashes, 75. Kaolin (china clay), composition of, 212. Karsten on yield of charcoal, 131 ; on alloys of copper and zinc (table), 611. Kernel-roasting, 439 (see under Copper- smelting}. Kersten's experiment on blue slag, 27. Keswick copper-mine, 289. " King-pot " crucible (in brass manufacture), 614. Kremer's analyses of lignite ashes, 90. L. Lacquering brass, 609, 622 ; paper, 623. " Lana philosophica " (zinc), 532. Lake Superior native copper, 309. Lancashire copper- works, 291. Latten (old name of brass), 606. Lawson, Dr. J., said to have obtained zinc from calamine, 520. Laxey blende, experiments on, 540, 542. Lead, compounds of, with copper (see under Copper). with copper, the curse of the assayer, 474, 475. in commercial zinc, 591. in ores of z^jnc, 601. heated with oxide of zinc, 543. Lenz on conductivity of metals for heat, 10 ; for electricity, 11. Le Play on reduction by carbon, 15 ; analysis of clinker, 98 ; researches on copper- smelting, 332 ; on sulpho-silicates, 344 ; on blue metal, 356 ; analyses of copper slag, 347, 363 ; of bottoms, 365 ; of re- finery slag, 366 ; claims suggestion of colour test, 488 ; on loss of copper in smelting, 497 ; failure of his prediction as to success of copper-smelting in France, 498. Lignites, 85 ; composition of, 86 ; analysis of ashes of, 90 ; Frdmy on, 106. Lime (for assaying copper), 459. heated with sulphide of zinc, 546. Liquation, 20. Logan, Sir W., on cost of copper-smelting, 4:99. London crucibles, 2'J-J. MUNTZ. Liirzeron kernel-roasting (copper), 440, 445. Lustre of metals, 12. Lydognia zinc-works, 577 ; gas zinc-furnace there, 596. M. Malachite (green carbonate of copper), 309. Malleability of metals, 6 ; of copper, 241 ; of zinc. 529 ; of brass, 608. Mallet's table of qualities of various alloys of copper and zinc, 611, 612. Malmqvist, P., revises account of copper- smelting in Sweden, 395. Manganese in copper-ores, 484 (see Copper'} ; in zinc-ores, 600. peroxide of, heated with sulphide of zinc, 545. Mannheim gold (alloy of copper and zinc), 606. Mansfeld schist, analyses of, 418. Margraaf made known a process of extracting zinc, 520. Marsh-gas, calorific power of, 52. Marsilly, De, on desiccation of peat, 77 ; on caking coal, 94 ; on coke, 146. Matthiessen on conductivity of metals for heat, 10 ; on electric conductivity of copper, 287. Metallurgy defined, 1. Metallurgical processes : reduction, 14 ; smelt- ing, 18; roasting or calcination, 19; dis- tillation, sublimation, and liquation, 20. Metals, a conventional term, 2 ; their physical properties, 2 ; action of heat on, 2 ; their specific gravity, 3; crystallization, 4; varieties of fracture, 5 ; malleability, 6 ; ductility, 7 ; tenacity, 7 ; toughness, 8 ; softness, 9 ; conduction of heat and of elec- tricity, 9 ; capacity for heat, 1 1 ; expan- sion by heat, 12 ; opacity, 12 ; lustre, 12 ; colours, 12. in coal, 106. Michaut's coke-oven, 187. Miller, W. A., on corrosion of copper-sheath- ing, 505. Miller, W. H., on constitution of slag, 24; on crystals of peroxide of tin from copper, 375 ; on crystals of zinc, 528. Mineralogy, treatises on, recommended, 309. Mineral charcoal, 189 ; Rogers's, 188. Molecular changes in brass, 621. Montefiori furnace, 587. Moresnet zinc-works, 581, 582 ; results, 584. Morris, J., on cost of copper-smelting, 501 ; of making calamine brass, 616. , R., patent for imitating Japanese copper, 394. Moss-copper, 359. Muntz's metal or yellow metal (for sheathing) ; has superseded copper-sheathing in mer- chant service, 619 ; highly profitable to Mr. Muntz, 620. INDEX. 631 MURDOCH. Murdoch introduces gas lighting, 192. Mushet on coking, 15^. N. Napier, James, experiments in copper-smelt- ing, 336 ; analyses of " coarse metal," 342, 346. Native copper (assay of), 476. of Lake Superior, 309. Neumann on products of distillation of coal, 191. Nickel eliminated from copper, 375. " Nil album " (zinc), 532. Nitrate of soda heated with sulphide of zinc, 545. Nitre (or saltpetre), in copper assaying, 459. heated with copper, 262 ; with sulphide of zinc, 545. Nitrogen combined with copper, 278. Northumberland (Duke of), analyses of coins presented by him to Museum of Practical Geology, 521. Norway, copper-smelting in, 411, 439, 444. Nozzles or condensers in zinc-works, 561. O. Olefiant gas : its calorific power, 53. O'Neill, C., on specific gravity of copper, 287. Opacity of metals, 12. Ores defined, 13 (see Copper-ores and Zinc- ores}. Orichalcum described by Strabo, 518 ; alluded to by Cicero, 518 ; the term probably in- cluded various alloys of copper, of which brass was one, 526 ; described by St. Ambrose and others, 527. Overpoled copper, 273. Oxides (see under Copper and Zinc). Oxy-chloride of copper, 476. Oxygen : its action on copper, 242 ; on zinc, 531. Oxy -sulphide of zinc, 540. P. Paracelsus on zinc, 518. Paraffine oil, 193. Paris, brass brought from : its composition, 621. Parkes, A., patent for adding phosphorus to zinc, 547. Parkes, H., employs cyanide of potassium in assaying copper, 479. Parkes, Messrs., patent for application of phos- phorus to metallic sheathing, 516. Parkes, S., on coke-ovens, 157. Parrot coal, 96. Parry's improvements in coke-ovens, 162. RED. Passau clays: their composition, 210, 226. Patent laws, remarks on, 381, 452. Peat, 73 ; its specific gravity, 73 ; composi- tion, 74 ; ashes, 75 ; desiccation, 77 ; ex- traction and preparation, 77. compressed, 78. Peat charcoal or coke, 142 ; carbonized by super-heated steam, 143. Pemberton's brass-works at Birmingham, 607, 612, 616. Penknife made of copper containing phos- phorus, 280. Percy, J., experiments on constitution of slags, ' 23 ; on their fusibility, 29 32 ; on com- bination of arsenic and zinc, 547. Perm, copper-smelting there, 433. Peroxide of manganese heated with sulphide of zinc, 545. Phillips, R., analyses of air from copper-ore calciner, 337. Phillips, J. A., analyses ancient coins, 522. Philosopher's wool, " I/ana philosophica " (zinc), 532. Phosphorus combined with copper, 279 ; with zinc, 546. , experiments on its protective action on copper-sheathing, 514. Pig-iron, cupriferous, 43. Pimple-copper, 362. Pipe-metal (zinc), 571. Pitch employed in coking, 189. Pitch coal, 85. Plattner on melting-points of slags, 46 ; on roasting disulphide of copper, &c., 247 ; on kernel-roasting (copper), 446. Pliny on certain metallurgical processes, 524, 525. Plot, R., on coke, 144. Plumbago, 226 (see Graphite}. Poling (copper), 325. Pompholyx (oxide of zinc), 532. Pots (used in extracting zinc), 551 ; mode of making, 555. Powell, Gabriel, his letter on Swansea copper-works, 293. I Pouillet's table of melting-points of metals, 3. Precipitating vats (for extracting copper), 448. Prideaux, Mr., analysis of " hard copper," 507. Prills (shots of copper), 475, 493. Prince's metal (copper and zinc), 606. Protoxide of copper (see Copper} ; sulphide of zinc heated with, 544. Protoxide of lead heated with zinc, 534 ; with oxide of zinc, 539 ; with sulphide of zinc, 545. Prussian zinc-works, expenses of, 576. Pyrometer, cupel, 46. R. i Rect-metal (copper), 356. | Red zinc ore (oxide of zinc), 548. 632 INDEX. REDRUTII. Redruth copper-ore circular, 305. Reduction defined, 14; by carbon, 14. Reduction-house (in zinc-works), 552. Refining (in copper assaying), 474. Refining (or white) flux, 460. Refinery slag, 366. Regnault, his analyses of lignites, 86 ; on sul- phide of zinc heated in vapour of water, 544. Regule (copper), 327, 330, 365. Regulus defined, 19. of copper, fusion for, 468 ; calcination of, 471. . Retorts, materials for, 208. in zinc-works, 572, 578. Reverberatory furnace, 19. Risca copper-works, 293. Rivot and Phillips's method of copper-smelt- ing, 385. Roaster slag, 362. Roasting, 19. of copper-ore, 324, 401. Rogers, Mr., on coking, 155. on mineral charcoal, 188. Rohstein, 419. Roman coins, analyses of, 521, 522. Ruel's black-lead crucibles, 227. Rumford on calorific power of fuel, 53 ; his calorimeter, 54. Runner-pig (in copper-smelting), 330. Russian copper-smelting, 433. S. Salt (chloride of sodium), used in assaying copper, 460. Sampling of copper-ores, 461 ; for assay, 478. Sand and sandstones, 238. Sandstone, cupriferous, 433. Scheerer on permanent gases from charcoal piles, 140. Schulze's apparatus for mechanical analysis of soils, 209. Scorifiers or roasting dishes, 456. Sea- water, its corrosive action on copper- sheathing, 505. Secrecy in copper assays injudicious and useless, 489. Sefstrom's experiments on silicates of lime, magnesia, and alumina, 39. blast-furnace, 231. Seguin's table of the tenacity of metals, 8. " Selecting " process (copper), 364. Seraing coke-ovens, 182. Sheathing (see Copper-sheathing). Sheet-zinc, 529. Sheffield, W. H., construction of copper cal- ciner, 318. Silbermann on calorific powers of fuel, 55. Silesian process of extracting zinc from the ore, 558, 573 (see under Zinc). Silica heated with dioxide of copper, 243 ; with protoxide of copper, 254 ; with oxide of zinc, 536. Silicate of zinc heated with carbon, 537 ; with lime and charcoal, 538. SWANSEA. Silicates, their atomic constitution, 21 ; chemical constitution, 23 ; external charac- ters, 25; colour, 27; fusibility, 29; ex- periments of Berthier, 32 ; of Sefstrom, 39. Silicon combined with copper, 282. Silver in copper, 378, 475 ; in zinc-ores, 549, 601. Similor (alloy of copper and zinc), 606. Slags defined, 18 ; mostly silicates, 20 (see Silicates'). Slag, copper : ore-furnace slag, 342, 420 ; white-metal slag, 347 ; blue-metal, 350 ; roaster slag, 362 ; refinery slag, 366. Smelting, 18 (see Copper and Zinc}. Smith, J. T., patent for coking, 190. Smithsonite, 549. Soda heated with disulphide of copper, &c., 262. , hyposulphite of, employed in wet assay of copper, 485. Softness of metals, 9. South Staffordshire coal-pits badly worked, 103. Specific heat of copper, 241. Specific gravity of metals, 3 ; of wood, 68 ; of copper, 283 ; of zinc, 530. "Speise" defined, 19. Spelter, speltrum, &c., names of zinc, 519. Split bricks, 230. Spurstein, 421. Stacks, 207 ; see Chimneys. " Standard," the term used in copper trade explained, 304. Steam, peat charred by, 143. raised by coke-ovens, 182. Stein's analyses of coal, 93. Storer, F. H., on fibrous character of alloys of copper and zinc, 608 ; on crystals of brass, 609. Stourbridge-clay crucibles, 219, 220. Strabo describes orichalcum, 519. Styrian kilns, 441. Sulphide of copper (see under Copper") ; of zinc, 539 (see under Zinc). of sodium, employed in assaying zinc ores ; standard solution, 598. Sulpho-silicates, supposed, 49. Sulphur, combined with copper (see under Copper") ; enormous quantity of it evolved in the air in copper-works at Swansea, 338 ; its proposed utilization, 341. collected in copper-smelting at Agordo, 441. combined with zinc, 589 ; heated with oxide of zinc, 536. Sulphuretted hydrogen in water, its alleged action on copper-sheathing, 505. Sublimation, 20. Swansea, as seen at sea, 336 ; copper-works, 293, 299 ; copper smoke there, 334, 338 ; proposed remedies ; opinions of Mr. Vivian, 338. copper-ore circular, 303, , secrecy of copper assayers useless, 489. INDEX. 633 SWANSEA. Swansea zinc- works, 578. method of extracting zinc from the ore, 550-558. Sweden, copper-smelting in, 395. Swedenborg's " Regnum Subterraneum" re- ferred to, 439. Swedish copper-ores described, 395. Swedish copper-smelter, feat of, 399. Sylvester and Hobson's patent for making zinc wire, &c., 529. T. Tartar (tartrate of potash) used in assaying copper, 460. Temperature, its influence on yield of char- coal, 131 ; on fracture of copper, 366. Tenacity of metals, 7 ; Baudrimont's table of, 8. of copper, 241 ; of zinc, 531. Terreil's experiments on devitrified glass, 25. Thum, on zinc-fume, 584; analysis of, 586, 587. Tin in copper, 473 (see Copper) ; in zinc, 589 (see Zinc). heated with sulphide of zinc, 543. found in copper-sheathing, 510. wasted in tin-works, 512. Tongs for crucibles, 231. Tombac (alloy of copper and zinc), 606. Tools employed in assaying copper, 456 ; in extracting zinc, 567. Torbanehill mineral, trial respecting, &c., 78. Toughness of metals, 8. Trees, growth of, 71. Troughton's improvements in copper-smelting, 381. Turf, 73 (see Peat}. Turner family said to have introduced brass manufacture into Birmingham, 607. Tutenag (a name of zinc), 520. Tutia f oxide of zinc), 532. TyndaU, .1., on calorific conduction, 10. W. Water, proportion of, in wood, 65. in coke, 146. , its action on zinc, hydrogen evolved, 533 (see Vapour'). Water-gilding (buttons), 617. Weights used for selling copper-ore, 497. Welsh copper-works, 293. Welsh process of copper-smelting, 314 (see under Copper-smelting). Wet methods : of extracting copper, 447 ; of assaying copper, 478 (see under Copper) ; of assaying zinc, 598 (see under Zinc). White flux, 460. White-metal, 347, 361. Wiedemann and Franz on conductivity of metals, 10. Williams, Dr. T., thinks copper-smoke at Swansea beneficial, 339. Wilson, T., on treatment of cupriferous resi- dues of pyrites, 517. Witting's analyses of wood ashes, 69, 70. Wood : kinds employed as fuel, 62 ; ele- mentary composition, 63, 204 ; propor- tion of water in, 65, 204; specific gravity of, 68 ; proportion of ashes in, 68 ; growth of, 71 ; directions for cutting and storing, 72. Wood-charcoal, its calorific power, 55. Y. Yellow copper-ore, 322 ; assay of, 464. Yellow metal, 515, 516 (see Muntzs metal). Young, Dr. T., his illustration of tough- ness, 9. Young's, J., paraffine oil, 193. Young, W., invents Dinas fire-brick, 236. Yorkshire copper- works, 290. V. Valentine, Basil, stated to have first used the term zinc, 519. Van Swab, A., extracted zinc from its ores by distillation, 520. Vapour of water, disulphido of copper heated in, 255 ; sulphide of zinc heated in, 544. Vats for extracting copper from solution, 448. Vignoles' patent for charring peat by steam, 143. Vitreous copper-ore, assay of, 466. Vitriol, black, 424. Vivian, Mr., on copper-smoke, 338. , Mr. Hussey, on eliminating from cop- per nickel and cobalt, 375 ; gold and sil- ver, 378. , his method of smelting rich copper slags, 386. ZINC, its history, 518; called spelter, &c., i>19. , analytical evidence of the antiquity of zinc combined with copper (brass), 521 ; descriptive evidence, 523. , physical properties, 528. , specific gravity, 530 ; certain chemical properties: action of oxygen, 531. , oxide of, its preparation, 532. heated with powerfully oxidizing solid re-ageuts, 533. , action of water on, 533. heated with protoxide ot lead, 534 ; with fixed alkaline carbonates, 534 ; with neutral fixed alkaline sulphates, 534 ; with carbonic acid, 534. , oxide of, 532 ; reduction of, by car- bon and carbonic oxide, 534 ; by hydro- gen, 535 ; heated with sulphur, 536 ; with 2 T 634 INDEX. ZINC. iron, 536 ; with silica, 536 ; with lime and charcoal, 538 ; with boracic acid, 538 ; with alumina, 539 ; with protoxide of lead, 539 ; with the fixed alkaline carbonates, 539 ; with cyanide of potassium, 539 ; with sulphide of zinc, 541. ZINC, silicate of, heated with carbon, 537. , oxysulphide of, 540. , sulphide of, 539 ; blende (black-jack), 549 ; heated with other sulphides, 541 ; with access of air, 541 ; with oxide of zinc, 541 ; with carbon, 543 ; with various metals : iron, antimony, lead, 543 ; in hydrogen, 544 ; in vapour of water, 544 ; with carbonic acid, 544 ; with protoxide of copper, 544 ; with protoxide of lead, 545 ; with peroxide of manganese, 545 ; with nitre or nitrate of soda, 545 ; with carbonate of potass or soda, 545 ; with lime, 546. presence of carbon in, 546 ; with phos- phorus, 546 ; with arsenic, 547 ; in copper, 474. and copper, 606 ; table of alloys of, 611. , ores of, 548 ; red zinc-ore (oxide) of zinc), 548; carbonate of zinc (calamine), 548 ; hydrated silicate of zinc (electric calamine), 549. , EXTRACTION OF: English process, 550; roasting the blende, 550 ; pots and con- densing tubes, 551 ; reduction house, 552 ; mode of making the pots, 555 ; of charging the pots and managing the furnace, 557 ; treatment of the rough zinc, 558 ; cost of production, 558. , Silesian process of extracting, 558, 572 ; retorts and appendages, 558 ; an- nealing-oven, 560 ; clay nozzles or con- ZINC. densers, 561; laggins or stoppers, 562; iron appendages, 562 ; the furnace, 562 ; tools employed, 566 ; nature and prepara- tion of the ore, 567 ; calciner, 567 ; dis- tillation of zinc, 569 ; melting distilled zinc, 570; yield of zinc, 571; consump- tion of fuel, 571 ; repairs, 572 ; modifica- tions of details of the process, 572 ; cost of production, 575, 576. Zinc, Belgian process of extracting, 578 ; retorts, &c., 578 ; results, 584. , Carinthian method of extracting, 585. , fume, 586 ; Montefiori furnace, 587. , commercial, foreign matter in, and its influence on the working qualities of the metal, 588. , the methods of extracting zinc com- pared, 593 ; consumption of fuel, fire-clay, &o. ; cost of labour, 593 ; alleged improve- ments, 595. , methods of ASSAYING its ores ; dry way, 596. , wet way of assaying its ores, by sul- phide of sodium, 598 ; process, 599 ; pre- cautions to be taken in the presence of iron, manganese, copper, 600 ; of lead, silver, cadmium, 601 ; by dissolving out the oxide of zinc by means of solution of ammonia and carbonate of ammonia, 601 ; results, 602 ; by a standard solution of bichromate of potash, 605; process, 606. , flowers of (oxide), 532. , manufacture of, into sheets, wire, &c., 529. works in England, 520 ; 1'equisites for their situation, 596. blende (sulphide of zinc), 432 ; found in North America, 549 ; in Europe, 550. LONDON : PKIXTKH BY WILLIAM CLOWES AND SONS, STAMFORD STREET, AND CHARING CROSS. ( 635 ) ERRATA. Page 22, last line, for 6. 2RO, 3SiO 2 read 6. 2RO, 9SiO-. 20, lino 25, far R read RO. ,, 71, line 8, transpose the parenthesis after "metre " in line 7. ,. 167, last line, for wide read thick. 279, line 15 from foot, for red, short read red-short. 282, line 6 from foot, for Cotton read Cottam. . . : ! 1 0, last line, for Cu-S -f Fe 2 S read Cu 2 S + Fe 2 S 3 . 31.2, line 7 of text, for ores read ore. 315, lines 17 and 21, for ii read II. 316, line 3, for i /, &c., by the channels 1 1, as shown read II ly the channels as shown. 331 , lines 39 and 45, for Tougoy read Tongoy. 339, line 18, for dark-coloured read light-coloured. 345, line 8 from foot, for raw read calcined. 347, line 3 from foot, for Slag read Regulus. 442, Fig. 116, transpose 6 to right of third gutter. 471, Table, col. 5, for 26 read 28 15 18 46 54 col. 7, for 04 read 08 38 38 51 54 490, insert reference 2 to second note. 601, lines 18 and 20 from top, omit the word not. 2 U T RETURN CIRCULATION DEPARTMENT TO +> 202 Main Library LOAN PERIOD 1 HOME USE 2 3 4 5 6 ALL BOOKS MAY BE RECALLED AFTER 7 DAYS LOAN Pc.V: ' - PHfQfl TO DUE DATE. RENEWALS: CALL (4 15) 642-3405" DUE AS STAMPED BELOW \ ,IMK U Tf* UNIVERSITY OF CALIFORNIA, BERKELEY FORM NO. DD6, 60m, 1/83 BERKELEY CA 94720 s LD 21-95m-7,'37 "..?KELEY LIB A JA