GIFT OF MICHAEL REESE ASSAYING BY SrOTTISWOODE AXD CO., NEW-STREET SQUARE LONDON" A MANUAL OF PEACTICAL ASSAYING BY JOHN MITCHELL, F.C.S. EDITED BY WILLIAM CBOOKES, F.E.S., PRES. C.S. SIXTH EDITION ILLUSTRATED WITH 201 WOODCUTS LONDON LONGMANS, GBEEN, AND CO. 1888 All rights reserved PBEFACE TO THE SIXTH EDITION. THE last edition of Mitchell's Assaying was published in 1881, and a new edition being required, the opportunity has been taken of x rewriting some of the descriptions which the progress of research had made old or ill adapted to modern requirements. At the same time much new matter has been introduced, and matter which had become obsolete has been omitted. The number of woodcuts has been increased from 188 in the last edition to 201 in the present edition, whilst the number of pages, notwithstand- ing numerous excisions, has grown from 809 to 896. Among the new and important additions to this edition may be enumerated a description of the ' Automatic Sampling Machine,' invented by Mr. D. W. Brunton ; many new gas furnaces and burners for the laboratory, devised by Mr. Fletcher, Messrs. J. J. Griffin, and others ; new blowpipe reagents and operations ; new processes, dosimetric, volumetric, and colorimetric, for the partial and complete assay of iron ores, iron, steel, spiegeleisen, &c. In the copper assay a full description is given, for the first time in this country, of the American system of fire [vi] PREFACE TO THE SIXTH EDITION. assay. The system adopted at Swansea is so interwoven with the customs of the trade that its replacement by a more accurate process is perhaps not to be expected for some time, but for those assay ers who are not bound to this particular process the Lake Superior system can be strongly recommended, as being quick, inexpensive, and comparing favourably in accuracy with the wet methods. In the assay of silver the action of bismuth on the ductility of this metal a subject hitherto overlooked has received considerable attention. Much has also been added on the subject of gold ores, a matter of large and increasing interest to miners and metallurgists ; and improved modes of assaying the precious metal and its detection in poor ores are given. Besides these more important additions and alterations, minor additions are to be found in every chapter. A little more prominence has been given to the English system of grain weights, as English assayers are more familiar with these than with the metric system in use on the Continent ; and, as far as possible, the decimal system has been adopted, the grain being taken as the unit. So many valuable memoirs on assaying and metallurgical subjects are published abroad that it has, however, been found impracticable to adhere universally to grains. W. C. LONDON : April 1888. EXTRACTS FROM THE PEEFACE TO THE THIED EDITION. IN this edition are incorporated all the late important discoveries in Assaying made in this country and abroad, and special care is devoted to the very important Volumetric and Colorimetric Assays, as well as to the Blowpipe Assays. Most of the chapters are entirely rewritten, whilst the chapter on Crystallography being a subject only remotely bearing on Assaying is left out altogether. On the other hand, in some cases it may seem that by treating of purely analytical details the limits -of Assaying have been exceeded. But these departments are so closely related as to make it impossible to fix the line of demar- cation between them. Moreover, chemistry is cultivated by almost all to whom this work is of interest or service, so that it is hoped these amplifications will add to its value. The old equivalents are retained, as they are more generally understood by students of science who do not make chemistry their chief study. The Editor is under many obligations to his friend Dr. Eohrig, M.E., for assistance in revising the manuscript and incorporating into the work the latest continental improve- ments, as set forth in Professor Kerl's Probirkunst. The author of the best work on volumetric analysis which has [viii] EXTKACTS FROM PEEFACE TO THE THIRD EDITION. yet appeared in English, Mr. Sutton, F.C.S., has kindly placed several cuts, &c., at the Editor's disposal, and some descriptions of German processes have been taken from the last English Edition of Fresenius's Quantitative Analysis a work which should be a standard of reference for all students who desire to carry their chemical researches further than is possible to be treated of in a work pro- fessing to deal only with Assaying. LONDON : September 1868 PBEFACE TO THE FIEST EDITION. WHEN the rank our country holds among nations, as regards her Mining interest, is taken into consideration, it must be with all a matter of surprise that no work especially devoted to the elucidation of the processes to be employed in ascertaining the richness in rnetal of any sample of ore (that is, in other terms, its Assay) has of late years appeared before the British public. Indeed, the only work at pre- sent known in England is Berthier's ' Traite des Essais par la Voie Seche,' which, for the mere purpose of inculcating the principles of Assaying, has many disadvantages not the least of which is its being written in a foreign tongue ; and although a knowledge of French is now so very general, yet many are prevented buying scientific works in that language on account of the difficulties of finding equivalents for the technicalities which must necessarily be employed. It is also a very large work, and one contain- ing much matter which the assayer does not need matter relating to the composition of wood and coal ashes, furnace products, &c., which are more especially adapted for the metallurgist. [x] PREFACE TO These considerations, coupled with the paucity of any knowledge of Assaying, excepting that confined to a very limited number of persons, induced the author of the following pages to turn a considerable amount of his at- tention to this subject, more especially as much difficulty was experienced in not having a suitable text book for the use of his pupils. A portion of the following pages was drawn up as a Manual for such a purpose ; but, on con- sideration, it was thought the extension of such a work was so much needed that it was determined to alter the original plan as> far as was consistent with the complete carrying out of the object in view, viz. the production of a Manual embodying information in every branch of Assaying, either by the wet or the dry processes. The following is a sketch of the manner in which this is accomplished, the author having followed the excellent arrangement of Berthier as closely as possible, from whose work also much matter that suited these pages, and which it would have been useless to rewrite, has been inserted. First, the Mechanical and Chemical Operations of Assay- ing are treated in full, inclusive of a description of the apparatus required, their mode of use, &c. Secondly, Furnaces, Fuel, and Crucibles, together with a description of the best Pyrometers, and their applications. Thirdly, the Fluxes, their properties, preparation, use, &c. Fourthly, an Essay on the use of the Blowpipe, and all its appurten- ances ; as Fluxes, Supports, &c. Fifthly, the action of the Fluxes on some Mineral Substances. Sixthly, a method of discriminating many Minerals by means of the Blow- pipe, aided by a few tests by the humid method. Seventhly, the Humid Analysis of many Mineral Substances, their composition, locality, &c. (All the minerals mentioned in the three last heads comprehend such only as generally THE FIRST EDITION. [xi] came under the notice of the Assay er.) Eighthly, the complete Assay of all the common Metals, in addition to which the Assay of Sulphur, Chromium, Arsenic, Heating power of Fuel, &c., is fully discussed ; and ninthly, and lastly, a copious Table drawn up for the purpose of ascer- taining in Assays of Gold and Silver the precise amount, in ounces, pennyweights, and grains, of Noble metal con- tained in a Ton of Ore from the assay of a given quantity This Table is the most complete and copious yet published Not only has it been endeavoured to collect all that is generally known on the subject of Assaying, but many new facts have been added, and such matter entered into, that the success of an assay is rendered much more certain ; and most assays are conducted more rapidly and with greater exactitude than heretofore. It has also been endeavoured to introduce a new system, in which is pointed out the rationale of each process, with the chemical action taking place between the fluxes and the ores in course of assay, so that by paying a careful attention to the matters discussed^ so much of the chemical nature of all ores that can come under the assay er's hand may be known, that the practice by * rule of thumb ' (a rule on which very little dependence is to be placed, ex- cepting after years of the most laborious practice, and a rule which cannot be imparted, excepting the pupil pursue the same unprofitable course) must, it is hoped, be speedily abandoned when, by knowing the chemical properties of the body operated on, the necessary fluxes and processes might be at once indicated, and with a certainty of perfect success. Having premised thus much, the author must beg to express his thanks to his friend Mr. F. Field for the kind [XllJ PREFACE TO THE FIRST EDITIOX. assistance lie afforded him whilst experimenting on the various modes of assay described in the body of the work ; and trusting that any little imperfections which may be detected will not be harshly criticised, but that it may be taken into consideration that the author has attempted to improve a branch of mining knowledge to which unfor- tunately too little attention has been devoted, and to which, if he has added anything useful, he is indebted for the first principles of such knowledge to Berthier's ' Traite des Essais,' for which, to the talented writer of the above work, he is under the most lasting obligation. CONTENTS. CHAPTER I. PAGE Chemical nomenclature Laws of combination, etc. . . 1 Metallic and non-metallic elements .... 2 Oxides ........ 4 Salts ....... .. 4 Binary compounds containing no oxygen . ... 5 Laws of combination ...... 5 Chemical symbols ; their employment and uses . . ,7 CHAPTER II. Preparation of the sample. Weighing . ... 9 Automatic sampling machine .' . . . .11 Anvil and stand, 14. Hammers, 15. Cold chisel, 16. Shears, 17. Pestle and mortar . . . . . . .17 Iron mortar, 17. Wedgwood mortar, 17. Porcelain mortar, 18. Steel mortar, 19. Agate mortar, 20. The sieve ........ 21 Elutriation . . . . . . .22 Washing, dressing, or vanning . . . . .23 The balance . . . . . . .26 Operation of Aveighing, 26. Bullion balance, 27. Rough assay balance, 28. Assay balance, 28. Conditions of accu- racy and delicacy, 29. The weights . . . . . . .34 Assay weights for silver, 35. Assay weights for gold, 35. Method of weighing . . . . .36 Mayer's method, 39. Incinerating precipitates before weighing . . . .41 [xiv] CONTENTS. CHAPTER III. PAGE General preparatory chemical operations . . . .42 Calcination . . . . . . .42 Use of Crucibles, 42. Roasting . . . . . . . .44 Roasting test, 44. Roasting in crucibles, 45. Roasting in platinum capsules, 46. Reduction ...... , . 46 Cementation, 47. Reduction by hydrogen 48. Fusion ........ 48 Solution ........ 49 Glass and platinum forceps, 50. Distillation . . . . . . .51 Liquid distillation, 51. Dry distillation, 53. Sublimation, 54. Scorification Cupellation. . . . . .54 CHAPTER IY. Production and application of heat . . . .55 Calcining furnace . . . . . . .55 Chimney . . . . . . . .56 Evaporating furnaces . . . . . .57 The hood ........ 57 Fusion furnaces. Wind furnaces. . . . .57 Body of furnace, 58. Ashpit, 57. Bars, 58. Chimney, 59. Blast furnaces . . . . . . .60 Royal Institution furnace, 60. Sefstrom's blast furnace, 61. Deville's furnace, 62. Muffle or cupel furnace . . . . . .62 Aufrey and d'Arcet's furnar ^0. Brick furnace for numbers of cupellations, 64. Ur .sal furnace, 65. Furnace operations . . . . .67 Auxiliary a^^aratus . . . . . .67 Poker- or stirring rods, 67. Tongs, 67. Spectacles and look- ing glass, 68. Curved rods, 68. "Wrought iron ladle, 69. Ingot moulds, 69. Fuel for furnaces . . . ... . .69 Coke, 70. Charcoal, 70. Effects produced by wind and blast furnaces . . .72 Oil and gas blast furnaces . . . . . .74 Oil furnaces . . . . . . .74 Description of the apparatus, 74. Oil lamp furnace, 75. Oil reservoir, 75. Crucible furnace, 75. Management of the oil lamp furnace, 76. Power of the oil lamp furnace, 79. Requisite blowing power, 79. CONTENTS. [xv] PAGE Griffin's gas furnace t . . . . .80 Operations in crucibles, 80. Operations in muffles, 89. Skittle-pots, 83. Flue, 83. Distillation of zinc, 84. Extra large gas furnace, 84. Fletcher's universal gas furnace . . . . ,84 Burner, 84. Single jacketed arrangement, 84. Double jacketed chemical furnace, 84. Brown's gas assay furnace . , . . . .87 Fletcher's reverberatory gas furnace . . . .90 Fletcher's muffle and draught crucible furnaces . . .92 Fletcher's new melting arrangement . . . .92 Injector gas furnace, 92. The burner, 93. ' Salamander ' crucibles, 93. Injector gas or spirit furnace, 94. Benzo- line, 95. Gore's gas furnace . . . . . . .95 Griffin's reverberatory gas furnace . , .100 Melting furnace for lead, tin, antimony, etc. . . .104 Bunsen's gas burner, 105. Mounts for crucibles, 108. Gas furnace for boiling or evaporating, 110. Solid flame gas burner, 111. Dirt proof high power burner, 112. Safety Bunsen burner, 113. Lutes and cements . . . . . .113 Fire lute, 113. Fat lute, 113. Roman cement, 114. Plas- ter of Paris, 114. Linseed or almond meal, 114. Lime and egg lute, 114. White lead with oil, 114. Yellow wax, 114. Soft cement, 114. Cement for brass or glass, 114. Cement for mending pestles, 115. Adhesive paste, 115. Waterproof cement, 116. Resinous or hard cement, 117. Caoutchouc, 117. Faraday's directions for luting iron, glass, or earthenware retorts, tubes, etc., 118, Willis's cement, 119. Iron cemep* , -\J 9. Beal's cement, 120. Boiler cement, 120. Bruyei rement, 120. Oxy- chloride of zinc cement, 1 20. Crucibles, cupels, etc. . . . . . V ' 1 20 Hessian, Cornish, Stourbridge, and London clay crucibles, 1 20. ' Salamander ' brand of crucibles and plumbago fittings, 121. Porcelain crucibles, 122. Black-lead crucibles, 123. Charcoal crucibles, 126. Lined crucibles, 127. Lime crucibles, 128. Forbes's experiments with lime and char- coal crucibles, 129. Alumina crucibles, 130. Magnesia crucibles and bricks, 130. Malleable iron crucibles, 135. Platinum crucibles, 135. Preserving platinum crucibles, 135, 137, 138. Dexter's lamp for platinum crucibles, 135. Silver crucibles, 140. Nickel crucibles, 140. Cupels . . 141 Cupel mould, 142. Powder for cupels, 143. Scorifiers, 144. Methods of measuring the heat of furnaces . . .144 [xvi] CONTENTS. PAGE Wedgwood's pyrometer, 144. Daniel's pyrometer, 144. Wilson's pyrometer, 147. Siemens's pyrometer, 148. Comparing the temperatures of two furnaces, 149. CHAPTER Y. Fuel, its assay and analysis . . . . .150 External appearance of fuel, its porosity, compactness, fracture, size, and shape of pieces . . . . .152 Estimation of adhering water . . . . .152 Estimation of specific gravity . . . . .153 Estimation of absolute heating power . . . .154 Pyrometric heating power, 154. Different modes of ascer- taining absolute heating power, 154. Berthier's method, " 155. ' lire's method, 157. lire's calorimeter, 158. Wright's calorimeter, 159. Estimation of specific heating power . . . .160 Estimation of pyrometric heating power . . . ..160 Estimation of the volatile products of carbonisation . .160 Examination of the coke or charcoal left behind on carbonisation 161 Estimation of the amount of ash . . . . .162 Estimation of the amount of sulphur . . . .162 Examination of other peculiarities of fuel . . .164 Calculation of results . . . . . .164 Assay of coal before the blowpipe . . . . .165 Valuation of coal for production of illuminating gas - . .167 CHAPTER VI. Reducing agents . . . . . . .169 Hydrogen gas, 169. Carbon, 170. Black-lead or graphite, 170. Estimation of the value of graphite, 170. Anthra- cite, 171. Coke, 171. Wood charcoal, 171. The fatty oils, 172. Tallow, 172. Resins, 172. Sugar, 173. Starch, 173. Gum, 174. Tartaric acid, 174. Oxalic acid, 174. Ammonium oxalate, 174. Comparative reduc- ing powers of the above agents, 175. Metallic iron, 175. Metallic lead, 176. Oxidising agents . . . . . . .176 Litharge, 176. Ceruse or white lead, 176. Lead silicates and borates, 177. Potassium and sodium nitrates. 177. Assay of saltpetre, 178. Lead nitrate, 180. Manganese peroxide, 181. Copper oxide, 181. Iron peroxide, 181, The caustic alkalies, potash and soda, 181. Potassium and sodium carbonates, 181. Lead, copper, and iron sul- phates, 181. Sodium sulphate, 181. CONTENTS, [xvii] PAGE Desulphurising agents . . . . . .181 Atmospheric oxygen, 182. Charcoal, 182. Iron, 182. Litharge, 182. Behaviour of litharge with sulphides of manganese, 183 ; iron, 183 ; copper, 185 ; antimony, 186 ; zinc, 186 ; lead, 187. Caustic alkalies and their carbonates, 187. Nitre, 188. Lead nitrate, 188. Lead sulphate, 188. Sulphurising agents . . . . . .189 Sulphur, 189. Cinnabar, 189. Galena, 189. Antimony sulphide, 189. Iron pyrites, 190. Alkaline persulphides, 190. Fluxes . 191 Non-metallic fluxes . . . . . .191 Silica, 192. Lime, magnesia, alumina, and their silicates, 192. Alumina, 193. Borax, 193. Glass, 193. Analysis of ' different kinds of glass, 194. Fluor-spar, 195. Potassium and sodium carbonates, 195. Potassium nitrate, 196. Common salt, 196. White flux, black flux, and raw flux, 197. Argol or cream of tartar, 199. Salt of sorrel, or potassium binoxalate, 199. White and mottled soap, 199. Reducing power of the various fluxes, 200. Metallic fluxes . . . . . . .201 Litharge and ceruse, 201. Glass of lead (lead silicate), 201. Lead borate, 201. Lead sulphate, 201. Copper oxide, 201. The iron oxides, 201. CHAPTER VII. The blowpipe and its uses ...... 202 Blowpipe, 203. Trumpet mouthpiece, 204. Nipples, 205. Lamps and oil, 206. Faraday's directions for using the blowpipe, 206. Oxidation, 209. Reduction, 209. Auxiliary blowpipe apparatus . . . . .210 Supports, 210. Charcoal, 210. Platinum, 211. Wire, 211. Forbes's instruments for preparing charcoal, 211. Plati- num spoon, 212. Aluminium support, 212. Asbestos cardboard, 213. Reagents and fluxes . . . . . .213 Blue litmus paper, 213. Reddened litmus paper, 213. Brazil-wood paper, 213. Turmeric paper, 213. Nitric acid, 213. Zinc, 213. Copper, 213. Iron wire, 213. Potassium cyanide, 214. Sodium carbonate, 214. Re- duction of metallic oxides, 216. Borax, 218. Sodium ammonio-phosphate, 219. Potassium nitrate, 220. Potas- sium bisulphate, 220. Sodium bisulphate, 221. Vitrified boracic acid, 221. Cobalt nitrate, 221. Nickel oxalate, a [xviii] CONTENTS. PAGE 221. Copper oxide, 221. Silica, 222. Turner's flux, 222. Ammonium fluoride, 222. Calcium fluoride and sulphate, 223. Bone ashes, 224. Proof lead, 224. Tinfoil, 224. Dry silver chloride, 224. Tincture of iodine, 229. Dry silver iodide, 231. Soda-paper, 233. Forbes's soda-paper apparatus, 233. General routine of blowpipe operations . . . .234 Size of the assay, 234. Order of blowpipe operations, 234. Discrimination of minerals . . . . .237 Crystalline form, 237. Mode of fracture, 238. Lustre, 239. Colour and streak, 239. Hardness, 240. Specific gravity, 241. Sonstadt's solution, 242. Fusibility, 244. Chemical, characters, 244. Colour of borax bead, 246. Appearance of reduced bead, 247. Requirements for testing minerals, 248. Description of minerals . . . . . .248 Quartz and the silicates, 248. Rock crystal, 248. Amethyst, 248. Rose quartz, 250. Cairngorm, 250. False topaz, 250. Chalcedony, 250. Carnelianand sard, 250. Agate, 250. Onyx or sardonyx, 250. Flint or hornstone, 250. Jasper, 250. Bloodstone, 250. Opal, 250. Talc, 250. Mica, 251. Chlorite, 251. Serpentine, 251. Meerschaum and ne- phrite, 251. Augite and horneblende, 252. Chrysolite, or olivine, 252. Tourmaline, 252. Garnet, 253. Topaz, 253. Beryl or emerald, 254. Zircon or hyacinth, 254. Felspar, 255. Mica, 255. Zeolites, 255. Corundum or sapphire, 256. Spinel, 256. Chrysoberyl, 257. Diamond, 257. Graphite or black-lead, 258. Coal, 258. Apatite, 258. Fluor-spar, 259. Calcspar, 259. Magnesite, 260. Dolomite, 260. Aragonite, 260. Rock salt, 261. Solu- ble sulphates, 261. Nitre, 261. Gypsum, selenite, ala- baster, 261. Heavy spar or barytes, 262. Sulphur, 262. Tin ore, 262. Molybdenite, 263. Bismuth, 263. Anti- mony sulphide, 263. Arsenic, 264. Arsenic sulphide, 264. Native iron, 265. Iron pyrites, 265. Arsenical pyrites, 265. Magnetic iron, 266. Specular iron, hema- tite, or micaceous iron, 266. Red ferric oxide or red ochre, 266. Brown ferric oxide, or brown ochre, 267. Titanic iron, 267. Chrome iron, 267. Green-earth, 267. Iron carbonate, 268. Manganese ores, 268. Arsenical nickel, 268. Smaltine or tin-white cobalt, 269. Cobalt bloom, 269. Blende, 269. Zinc carbonate, 270. Zinc silicate, 270. Galena, 270. Lead carbonate or cerusite, 270. Pyromorphite, 271. Lead sulphate, 271. Cinnabar, 272. Native mercury or quicksilver, 272. Native copper, 273. Vitreous copper, 273. Copper pyrites, 273. Grey copper, 274. Black cupric oxide, 274. Red CONTENTS. [xix] PAGE cuprous oxide, 274. Copper carbonates, blue and green, 275. Platinum, 275. Gold, 275. Silver, 276. Silver sulphide, 276. Antimonial and arsenical silver ores, 276. Horn silver, 277. Lead and antimony sulphides, 277. Mercury, lead, silver, or copper selenides, 277. Millerite or nickel sulphide, 277. White nickel, 278. Rutile, 278. Sphene, 278. Wolfram, 278. Pitchblende, 278. Mag- netic iron pyrites, 279. Determination of minerals . . . . .279 Scheme for the determination of minerals .... 280 Group I. Minerals which have a metallic lustre, and which give off sulphur . . . . . . .281 Group II. Minerals which have a metallic lustre, and which give off either an odour of garlic without sulphur, or white fumes which have not a garlic or sulphurous odour . 283 Group III. Minerals which have a metallic lustre, and which give off no fumes ...... 284 Group IV. Minerals which possess a non-metallic lustre, a coloured streak, and which give off fumes or odour when heated before the blowpipe .... 286 Group V. Minerals which possess a non-metallic lustre, a coloured streak, but which give off no fumes or odour before the blowpipe . . . . .287 Group VI. Minerals which have a non-metallic lustre, and which are scratched by quartz, showing a white streak . . 288 Group VII. Minerals which have a non-metallic lustre, and which are not scratched by quartz . . . .291 Coloured flames . . . . . . .294 Blue flames, 294. Green flames, 294. Yellow flames, 295. Red flames, 295. Chlorine, 295. Lead, 295. Arsenic, 295. Selenium and antimony, 295. Bromine, 295. Boracic acid, 295. Tellurium, 296. Copper, 296. Iodine and copper, 296. Phosphoric acid, 296. Baryta, 296. Zinc, 296. Soda, 296. Water, 296. Strontia, 296. Lithia, 296. Lime, 297. Potash, 297. CHAPTER VIII. Volumetric analysis . . . . . . .298 Reactions of volumetric analysis, 298. Principle of volume- tric analysis, 298. Standard solutions, 302. Instruments and apparatus, 302. The burette, 303. Mohr's burette, 304. Gay-Lussac's burette, 304. Britton's burette, 305. Modes of operating, 305. The pipette, 306. Measuring flasks, 307. a 2 [xx] CONTENTS. CHAPTER IX. PAGE The assay of iron ....... 308 Ores of iron ....... 308 Magnetic iron ore, 308. Red hematite, 308. Brown hema- tite, 308. Spathic iron ore, 308. Titaniferous iron ore, 308. Franklinite, 308. Clay band and black band iron stone, 308. Assay of iron in the dry way ..... 309 Classification of iron ores, 309. Fluxes, 309. Air furnaces, 311. Crucibles, 311. The charge, 313. Ores of unknown com- position, 313. Ores previously analysed, 314. The opera- tion, 315. Effects produced by the following substances : Manganese, 317. Titanium, 317. Phosphorus, 317. Sul- phur, 317. Chromium, 317. Calculation of results . . . . . .319 Assay of iron and its ores in the wet way . . .321 Dr. Penny's process . . . . . .321 Reduction by zinc, 327. Reduction by stannous chloride, 328. Reduction by ammonium bisulphite, 328. Titration of iron with sodium hyposulphite, 330. Oudemans's method, 330. Haswell's method, 330. The complete assay of iron ores ..... 332 The mechanical treatment, 332. The chemical treatment, 334. Estimation of phosphoric acid, 334. Sulphur and iron, 338. Silica, ferric oxide, alumina, manganese, lime, and magnesia, 341. Nickel, cobalt, and zinc, 344. Esti- mation of ferrous oxide, 346. Ferrous oxide in insoluble silicious matter, 347. Alumina, 349. Calculation of the analysis, 349. Carbonic acid, 349. Water and carbon in carbonaceous matter, 351. Alkalies, 355. Copper, lead, arsenic, and antimony, 356 Titanic acid, 358. Estima- tion of specific gravity, 360. Gooch's perforated crucible . . . . .361 Estimation of carbon, silicon, phosphorus, &c., in metallic iron and steel, 363. Boring and sampling, 363. Estimation of total carbon, 367. Weyl's method, 368. Estimation of gra- phite, 369. Dr. Eggertz's method, 369. Mr. Tosh's method, 370. Estimation of combined carbon, 371. O. Arnold's method, 373. J. Blodgett Britton's method, 374. Mr. Stead's method of estimating minute quantities of carbon, 376. A new form of chromometer, 380. Preparation of inorganic standards for the colorimetric carbon test . . . . . . . .381 Day standard colours, 382. Night standard colours, 386. Estimation of sulphur in iron and steel .... 388 CONTENTS. [xxi] PAGE Dr. Eggertz's method, 388. J. Wiborgh's method, 393. Preparation of the cloth, 394. The colour scale, 398. Details of the process, 401. Estimation of sulphur in iron ores, 405. Estimation of silicon in iron and steel . . . .407 Mr. Turner's method . . . . . ,413 Estimation of basic cinder and oxides in manufactured iron . 414 Mr. Bettel's process . . . . . .415 Estimation of phosphorus in iron and steel , . .416 M. Tantin's method, 416. J. B. Mackintosh's method, 417. J. Lawrence Smith's method, 417. Quantity of iron em- ployed, 419. Solution, 420. Concentration of the phos- phorus, 420. Separation of the phosphorus, 421. Estimation of manganese in iron . . . . .422 E. Riley's method, 424. The direct method, 424. The in- direct method, 425. W. Kalman and A. Smolka's method, 426. Dr. Peter's method, 427. Mr. Galbraith's method, 429. Estimation of titanium in iron . ... . . 430 Mr. Riley's method, 430. Mr. Bettell's method, 432. Turner's table of the hardness of iron and steel 433 CHAPTER X. The assay of copper ....... 434 Classification of minerals and substances containing copper . 434 Class I. Sulphuretted ores or products with or without sele- nium, antimony, or arsenic, 434. Copper glance, 434. Chalcopyrite, 434. Erubescite, 434. Bournonite, 434. Fahlerz, 434. Covelline, 434. Wolfsbergite, 434. Do- meykite, 434. Copper regulus, copper speiss, etc., 434. Class II. Oxidised ores and products, 434. Red copper, 434. Malachite, 434. Azurite, 434. Cyanosite, 434. Phos- phate of copper, 434. Arseniates, 434. Chromate, vana- date, and silicate of copper, slags, etc., 434. Class III. Copper and its alloys, 434. Classification of different methods of assaying copper . . 435 A. Assay in the dry way, 435. a. For rich ores and products of Class I., 435. English copper assay ...... 435 Moissenet's description, 435. Ticketing in Cornwall, 435. De la Beche's sketch of the method, 436. Division adopted, 436. Section I. Reactions, 437. Two kinds of assays roasted and raw sample, 437. [xxii] CONTENTS. PAGE. 1. Regulus, 438. Pyrites, 438. Very poor pyrites, 438. Variegated copper ore, 439. Sulphide of copper, 439. Carbonated minerals, 439. Native mixture, 439. 2. Calcining, 440. 3. Coarse copper, 440. 4. Washings, 441. 5. Testing, refinery, 442. 6. Slags for prill, 443. Section II. Manipulations, 443. Crucibles used in Cornwall, 444. Wind furnace, 444. Fusion for regulus, 445. Cal- cining the matt, 448. Coarse copper, 449. Washings^ 450. Testing and refining, 450. Prill, 450. Section III. Some minerals and substances of a special nature ; influence of foreign metals, 451. Stanniferous minerals, 451. Antimonial minerals, 451. Zinciferous minerals, 452. Plumbiferous minerals, 452. Regulus of Chili, 452. Slags of copper, 452. Old copper, 452. Section IY. Summary considerations comparison of the results with the analysis by the wet way, 453. Principal causes of loss in the Cornish method, 454. Assay of copper in the dry way, 454. b. For ores and products of Class II., 455. Lake Superior fire assay ...... 455 Dr. Peter's description, 455. Sampling, 456. Fluxes, 457. Furnace, 457. B. Assays in the wet way, 461. a. Colorimetric copper assay, 461. Heine's method, 461. Le Play's method, 46 7. T.O. Cloud's method, 467. Endemann's method, 467. Carnelly's method, 470. Standard copper solution, 470. Solution of, 472. Jacquelain's and Von Hubert's colorimetric assay, 472. b. Volumetric copper assays, 476. Fleck's modification of Mohr's method, 476. E. 0. Brown's method by sodium hyposulphite, 479. c. Electrolytic copper assay, 480. Estimation of copper in the Mansfield ores by Dr. Steinbeck's process ....... 480^ Extraction of the copper from its ores, 480. Separation of the copper, 481. Quantitative estimation of the precipi- tated copper, 482. Special observations on this method, 483. Estimation of copper in the Mansfield ores by M. C. Luckow's process . . . . . . : , 487 Roasting the ore, 489. Solution of the roasted product, 489. Precipitation of the copper, 490. Weighing the copper, 493. Assay of copper pyrites . . . . . .494 CONTENTS. [xxiii] PAGE Detection of traces of copper in iron pyrites and other bodies . 497 Estimation of arsenic in copper . . . . .499 CHAPTER XI. The assay of lead . . . . . ,502 Classification of minerals and substances containing lead . . 502 Action of reagents on lead sulphide, 502. Oxygen, 502. Me- tallic iron, 502. Alkalies and alkaline carbonates, 503. Potassium nitrate, 503. Argol, 503. Assay of substances of the first class (sulphides, antimonides, etc.) 504 Processes generally adopted : 1. Fusion with potassium carbonate, 504. 2. Fusion with black flux, 510. 3. Fusion with metallic iron, 511. 4. Fusion with sodium carbonate, or black flux and metallic iron, 514. 5. Roasting and reducing assay, 515. Assay with black flux and iron, 517. Roasting and reduction assay with iron, 518. Roasting and fusing with black flux, 518. Level's fusion assay with potassium ferrocyanide and cyanide, 519. 6. Assay with sulphuric acid, 519. 7. Assay of galena in the wet way, 521. Assay of substances of the second class . . . .523 Humid assay of ores of the second class . . . .526 Assay of substances of the third class . . . .527 Humid assay of substances of the third class . . . 528 Maxwell Lyte's process, 529. Mascazzini's process, 529. Jannesay's process, 529. Alloys of lead (Class IV.) 530 Assay with sulphuric acid, 530. Level's fusion assay with potassium ferrocyanide and cyanide . 531 Estimation of lead by means of standard solutions . . 531 1. Flores Dumonte's method, 531. 2. Schwartz's method, 532. 3. Buisson's volumetric process, 533. 4. Diehl's volumetric process, 534. CHAPTER XII. Assay of tin ..... 537 Tin ores ...... 537 Tin oxide, 537. Crystallised tin oxide, 537. Disseminated tin oxide, 537. Sandy tin oxide, 537. Concretionary tin oxide, wood tin, 537. [xxiv] CONTENTS. PAGE Analysis of a sample of tin oxide from Cornwall, 538. Remarks on tin ore and the minerals which may be mistaken for it, by Dr. A. Leibius, 538. Assay of pure tin oxide ...... 539 Method used in Cornwall, 540. Method by means of potas- sium cyanide, 540. Assay of tin oxide mixed with silica .... 542 Assay of tin ores containing silica and tin slags . . .542 Assay of tin ores containing arsenic, sulphur, and tungsten . 543 Approximative assay ...... 544 J. H. Talbot's method for assaying tin in the presence of tungsten, 546. Estimation of tin by the humid method .... 546 Klaproth's process, 546. J. B. Hallet's process, 548. M. Moissenet's process, 548. Assay of tin in gun and bell metal .... 549 E. Burse's method, 550. Estimation of tin by means of a standard solution . . 550 M. Gaultier de Claubry's process, 550. M. Lenssen's process, 551. M. Stromeyer's process, 551. Method for the analysis of tin ore, 552. Another method, 553. Decomposition of tin slags, 553. MM. Mohr and Terreil's method, 554. Rammelsberg's method, 554. CHAPTER XIII. Assay of antimony . . . . . .556 Classification of antimonial substances .... 556 Class I. Native antimony and all antimonial substances con- taining oxygen or chlorine, but little or no sulphur : Native antimony, 556. Antimony oxide, 556. Antimo- nious acid, 556. Antimonic acid, 556. Class II. Antimony sulphide and all antimonial ores contain- ing sulphur : Antimony sulphide, 556. Antimony oxy- sulphide, 556. Haidingerite, 556. Assay of ores of the first class ..... 556 Assay of ores of the second class . . . . .557 1. Estimation of the pure antimony sulphide (antimonium crudum) . . . . . . . 557 2. Estimation of regulus of antimony . . . 558 Two methods, 558. a. By roasting and fusing the oxidised matter with black flux, 558. b. By fusing the crude ore with iron, or iron scales, with or without the addition of black flux, 558. F. Becker's method, 563. Mr. Button's method, 563. Detection of antimony in sublimates .... 563 CONTENTS. [XXV] PAGE To distinguish arseniuretted hydrogen from antimoniuretted hydrogen ....... 564 Separation of tin from antimony and arsenic . . . 564 0. Winckler's method, 564. Assay of alloys of lead and antimony (type metal) . . 565 CHAPTER XIV. Assay of zinc ....... 567 Classification of bodies containing zinc . . . .567 Class I. Zinc ores in which the metal exists "as oxide not combined with silica, 567. Class II. Zinc ores in which the metal exists as oxide, but partly or wholly combined with silica, 567. Class III. Zinc ores in which the metal is partly or wholly combined with sulphur, 567. Class IV. Alloys, 567. Assay of ores of the first class . . . . .567 Assay of zinc by the humid process in ores of the first class . 571 Assay of ores of the second class . . . . .572 Wet assay of zinc in ores of the second class . . ,572 Assay of ores of the third class . . . . .573 Humid wet assay of zinc in ores of the third class . . 573 Assay of cupriferous blende ..... 574 Alloys (Class IY.) 574 Volumetric assay of zinc . . . . . .575 Galetti's process, 575. Schaffner's method, modified by C. Kunzel, 577. a. Solution of the ore, and preparation of the ammoniacal solution, 577. 6. Preparation and standardising of the sodium sulphide solution, 578. c. Assay of the zinc in the solution of the ore, 579. d. Further modification of the process, 580. H. Schwarz's method, 581. Carl Mohr's method, 582. J. Drewson's method, 583. Separation of copper from zinc, 584. CHAPTER XV. Assay of mercury ....... 587 Assay of mercurial ores ...... 587 Berthier's method, 590. Eschka's method, 591. Assay for the amount of cinnabar in an ore . . . 592 Electrolytic assay of mercury ..... 594 Volumetric estimation of mercury . . . .595 M. J. Personnel method, 595. M. Rivot's method, 597. G. Attwood's method, 598. XXvil CONTENTS. CHAPTER XVI. PAGE Assay of silver ....... 600 Classification of argentiferous substances . . . .600 Class I. Minerals containing silver, 600. Class II. Metallic silver and alloys, 600. General observations on the assay of ores and substances of Class I. . . . . . .600 Fusion with oxidising reagents ..... 603 Litharge, 603. Special directions for the crucible assay of ores and substances of Class I. ....... 605 Preliminary assay for dividing all substances of this class into three sections, 606. Assay of reducing power of argol, 607. Assay of oxidising power of potassium nitrate, 607. Assay of litharge for silver, 608. Assay of ores of the first section ..... 608 Assay of ores of the second section . . . .609 Assay of ores of the third section ..... 609 Scorification . . . . . . .610 The roasting, 612. The fusion, 612. The scorification, 613. Special instructions for the scorification assay of ores of Class I. . 615 Assay in scorifier, 615. Assay of substances of Class I. admixed with native or metallic silver ........ 616 Cupellation . . . . . . .617 Pliers and microscope, 623. Amalgamation . . . . . . .626 Substances of the second class . . . . .627 Separation of silver from galena . . . . .627 General remarks on the assay of the alloys of silver and copper . 627 Cupellation, 627. D'Arcet's results, 628. Quantity of lead required, 629. Loss of silver in the assay of coined alloys, 630. Special instructions for the assay of the alloys of silver and copper 631 Assay for approximative quantity of alloy, 631. Assay proper of silver bullion, 631. Assay of alloys of copper and silver, 632. Alloys of platinum and silver . . . . .632 Alloy of platinum, silver, and copper .... 633 Assay of native silver, rough silver left on sieve during pulverisa- tion of silver ores of first class, and native alloys of silver . 633 Dr. W. Dyce's process for separating gold and silver from the baser metals ....... 633 Assay of silver bullion by the wet method . . .634 CONTENTS. [xxvii] PAGE Gay-Lussac's method ...... 634 Measurement of the solution of common salt, 636. Measure of the normal solution of salt by weight, 636. Preparation of the decime solution of common salt, 638. Preparation of the decime solution of silver, 640. Weighing the normal solution of common salt, 641. Preparation of the normal solution of common salt when measured by weight, 641. Preservation of the normal solution of common salt, 646. Application of the process described in the determination of the standard of a silver alloy, 647. Correction of the standard of the normal solution of salt when the tempera- ture varies, 649. Table of corrections for variations in temperature of the normal salt solution, 651. Table for the assay, by the wet method, of an alloy containing any proportions whatever of silver, by the employment of a con- stant measure of the normal solution of common salt . 651 Tables for determining the standard of any silver alloy by employ- ing an amount of alloy always approximatively containing the same amount of silver . . . . 654, 673 Assay of pure, or nearly pure silver, the temperature of the normal solution of salt being that at which it was standard- ised ........ 674 First example, 674. Second example, 675. Third example, 676. Graduation of the normal solution of salt, the temperature being different to that at which it was wished to be graduated . 677 First example, 677. Second example, 678. Approximative determination of the standard of an unknown alloy ........ 678 Modes of abridging manipulation . . . . .679 Bottles, 679. Stand, 680. Water-bath, 680. Flue, 681. Agitator, 681. Table, 683. Cleaning the bottles, 684. Reduction of silver chloride obtained in the assay of alloys by the wet method . . . . . . . 684 Preparation of pure silver . . . . . .685 Modifications required in the assay of silver alloys containing mercury ....... 68fr Method of taking the assay from the ingot, 687. Method of assaying silver bars adopted in the assay offices of H.M. Indian mints, 687. The chloride process, 688. Apparatus and appliances required, 697. Effect of bismuth on the ductility of silver . . . 699' Titration o^ silver in presence of other metals . . .711 Copper, 713. Mercury, 714. Palladium, 714. Determination of silver in galena by Volhard's process . .714 David Forbes on blowpipe assay of silver . . . .715- Apparatus, 716. Concentration of the silver lead, 717. Cu- [xxviii] CONTEXTS. PAGE .pellation, 720. Estimation of the weight of the silver globule obtained in cupellation, 723. The scale for this purpose, 724. Cupellation loss, 727. Table modified by Plattner, 728. Classification of argentiferous substances . . . .729 A. Metallic alloys capable of direct cupellation, 730. a. Consisting chiefly of lead or bismuth, 730. 6. Consisting chiefly of silver, and alloys of silver with gold and copper, 731. c. Consisting chiefly of copper, 732. B. Metallic alloys incapable of direct cupellation, 733. a. Containing much copper or nickel, with more or less sulphur, arsenic, etc., 733. b. Containing tin, 734. c. Containing antimony, tellurium, or zinc, 735. d. Containing mercury ; amalgams, 737. e. Containing much iron, 737. Alloys of silver and copper, 739. CHAPTER XVII. Assay of gold ....... 740 Classification of substances containing gold . . .740 Class I. Ores containing gold, 740. Graphic tellurium, 740. Folliated tellurium, 740. Class II. Alloys of gold, 740. Native gold and aurides of silver, 740. Artificial alloys of gold, 740. Assay of gold ores . ...... 741 Preparation of the sample, 742. Collection of the gold and silver, 742. Crucible assay, 742. The charge, 743. Size of lead button, 743. Preliminary assay of ore, 744. Roasting the ore, 745. Fusion, 747. Scorification assay, 747. The lead button, 749. Cupellation, 749. Estimating the weight of minute spheres of gold . . . 753 Oeneral observations on the assay of gold ores . .. .756 Gold and copper, proportion of lead, 756. Examination on the touchstone, 757. Table for the proportion of lead to be employed in the cupellation of gold and copper, 759. Gold, silver, platinum, and copper, 759. Gold alloyed with silver, 761. Inquartation, 761. Surcharge, 763. Mr. Seine's method, 763. Makin's method, 765. Aqua regia, 766. Yon Jiiptner's method, 767. Standard of the alloys of gold . . . . .768 Assay of gold coin and bullion . . . . .769 Preliminary assay, 769. Assay proper, 769. Parting assays, 771 CONTENTS. PAGE Assay of pyrites for gold . . . . . .773 Treatment of gold and silver bearing copper ores . . .775 Detection of minute traces of gold in minerals . . . 777 CHAPTER XVIII. Assay of platinum . . . . . . .781 Platinum in its native state . . . . .781 Analysis of platinum ores . . . . . .782 Bunsen's method, 782. Platinum and palladium, 782. Ru- thenium, rhodium, and iridium, 785. C. Lea's process, 790. Deville and Debray's process, 795. Sand, 795. Osm- iridium, 796. Platinum and iridium, 797. Palladium, iron, and copper, 798. Gold and platinum, 799. Rhodium, 799. Analysis of platinum ores from various sources, 799. Guyard's process for extracting metals from platiniferous residues, 799. Analysis of osm-iridium, 803. Wolcott Gibbs's process, 804. Nelson Perry's process, 807. CHAPTER XIX. Assay of bismuth ....... 809 Varieties of bismuth ores ...... 809 Native bismuth . . . . . .809 Assaying bismuth ores . . . . , .810 Assaying bismuth in ores containing a large amount of copper . 810 Refining crude bismuth . . . . . .813 Purification of the reduced bismuth . . . .814 Purification of bismuth from arsenic, 814. Purification of bismuth from antimony, 815. Purification of bismuth from copper, 815. Purification of bismuth from sulphur, 816 Volumetric assay of bismuth . . . . .817 R. W. Pearson's process, 817. M. P. Muir's process, 818. CHAPTER XX. Assay of chromium ...... 820 Principal ore of chromium . . . . . 820 Assay of chrome ore ...... 820 Genth's process, 820. O'Neill's process, 822. W.Gibbs's pro- cess, 823. Clark's process, 824. H. N. Morse and W. C. Day's process, 825 Estimation of chromium by means of standard solutions . . 827 Estimation of chromium in iron and steel, 827. J. O. Arnold's process, 827. W. J. Sell's process, 830. [XXX] CONTENTS. CHAPTER XXI. PAGE Assay of arsenic . . . . . . .831 Minerals from which arsenic is produced . . . .831 Assay for arsenic . . . . . . 831 Approximative method, 832. Mr. Parnell's method, 832. CHAPTER XXII. Assay of manganese . . . . . .833 Commercially valuable minerals containing manganese . .833 Valuation of manganese ores . . . . .833 Sherer and Rumpf 's method, 834. Mohr's method, 837. Otto's method, 837. Bunsen's method, 837. J. Pattinson's process, 837. CHAPTER XXIII. Assay of nickel and cobalt ores . . . . .839 Ores of nickel . . . . . . .839 Ores of cobalt . . . . . . .839 Hadow's process for separating nickel and cobalt . . .839 Decomposition of cobalt speiss . . . . .843 Assay of nickel ores ...... 845 Assay of commercial metallic nickel .... 845 Assay of cobalt ores ...... 846 Liebig's method of separating nickel and cobalt . . . 846 Wolcott Gibbs's improvement . . . . .847 Terreil's method . . . . . . .847 Fleitmann's quantitative assay of small proportions of cobalt in nickel. ....... 849 Plattner's method for detecting nickel before the blowpipe . 849 Assay of complex nickel and cobalt ores .... 850 A. Olassen's method, 850. Ores containing sulphur, arsenic, nickel, cobalt, and iron, 851. Alloys of zinc, copper, and nickel, 853. Ohl's method for the assay of nickel speiss . . .854 Lead and copper speiss, 855. Speiss containing little or no lead or antimony, 855. Speiss containing much lead or antimony, 856. CHAPTER XXIV. Assay of sulphur ....... 859 Commercially valuable sulphur minerals . . . .859 Assay of sulphur in iron and copper pyrites . . .859 CONTENTS. [xxxi] PAGE Assay of sulphur in the dry way . . . . .859 Assay of sulphur in the wet way . . . . .861 C. R. A. Wright's process, 861. Pearson's method, 862. Hougeau's process, 864. Deutecom's method, 864. Holland's method, 865. Breckmann's process, 866. CHAPTER XXV. Discrimination of gems and precious stones . . . 867 Explanation and introduction, 867. Principal sources of recognition, 867. Colourless stones . . . . . . .867 Diamond, 867. The matrix of the diamond, 868. Quartz, 874. White zircons, 875. White sapphire, 876. White topaz, 876. Comparative table of the weights of colour- less stones, 878. Use of the tatye, 878. Yellow stones . . . . . . .879 Yellow zircon, 879. Yellow sapphire, 879. Cymophane (chrysoberyl), 879. Yellow topaz, 880. Yellow tourmaline, 880. Yellow emerald, 881. Comparative table of weights of yellow stones, 883. Yellow quartz, 884. Brown and flame-coloured stones ..... 884 Zircon (hyacinth), 884. Yermeil garnet, noble garnet, al- mandine, 884. Comparative table of weights of brown or flame-coloured stones, 885. Essonite, cinnamon stone, 886. Tourmaline, 886. Red and rose-coloured stones ..... 886 Red sapphire, oriental ruby, 886. Deep red garnet, noble garnet, 886. Spinel ruby, 886. Reddish topaz, 887. Red tourmaline, 887. Comparative table of weights of red and rose-coloured stones, 887. Blue stones ....... 888 Blue sapphire, 888. Disthene, cyanite, 888. Blue topaz, 888. Blue tourmaline, 889. Blue beryl, 889. Dichroite, water sapphire, 889. Turquoise, 889. Comparative table of the weights of blue stones, 890. Violet stones .890 Violet sapphire, 890. Violet tourmaline, 890. Violet quartz, amethyst, 890. Comparative table of weights of violet stones, 891. Green stones . . . . .891 Green sapphire, 891. Peridot, crysolite, 891. Green tourma- line, 892. Emerald, 892. Aqua-marine, 892. Chryso- prase, 892. Comparative table of weights of green stones, 896. [xxxii] CONTENTS. Stones possessing a play of colours (chatoyant) . Sapphire, 894. Garnet, 894. Cymophane, 894. Antique emerald, 894. Quartz, 894. Felspar, nacreous felspar, fish-eye, &c., 895. Comparative table of weights of stones possessing a play of colours (chatoyant), 895. Glass and artificial gems ...... Inferior brilliancy, 896. Inferior hardness, 896. Fusibility, 89 6. PAGE 893 896 APPENDIX. TABLE I. Showing the quantity of fine gold in 1 oz. of any alloy to ^ of a carat grain of the mint value of 1 oz. of each alloy . . ii TABLES A, B, and C. To convert mint value into bank value when the standard is expressed in carats, grains, and eighths ... xx TABLE II. Table of relative proportions of fine gold and alloy, with the respective mint values of 1 oz. of each alloy when the standard is expressed in thousandths . . . xxi-xxxii TABLE. To convert mint value into bank value when the standard is expressed in thousandths . . xxxiii TABLE III. Assay table, showing the amount of gold and silver, in ounces, pennyweights, and grains, contained in a ton of ore, &c., from the weight of metal obtained in an assay of 200 grains of mineral .... xxxiv-xlvii INDEX xlix . A MANUAL OF PEACTICAL ASSAYING, CHAPTEE I. CHEMICAL NOMENCLATURE LAWS OF COMBINATION, ETC. IN a treatise intended to be used principally by the prac- tical assayer, it is neither necessary nor possible to give more than a brief outline of the elements of chemical nomenclature and of chemical combination. A knowledge of practical chemistry is undoubtedly of great value to the assayer ; indeed, no one can attain to any degree of eminence in this branch of industry unless he has had some amount of practice in the laboratory ; but it will be beyond the scope of this volume to teach the elements of chemistry. Such instruction in chemistry must be sought for in books which are specially devoted to the science. The student should, above all, endeavour to acquire a practical knowledge of experimental chemistry by going through a course of instruction in a laboratory. CHEMICAL NOMENCLATURE. Every material substance with which we are acquainted consists of one or more bodies, termed elements, from the fact that with our present means of research we are unable to reduce them to a more simple form. Thus, if a piece of common iron pyrites be exposed B L> CHEMICAL NOMENCLATURE to certain chemical operations, it will be found to consist of two substances, each physically and chemically distinct from the other and from the original substance. One body is sulphur, an opaque yellow substance, fusing at a very low temperature, igniting readily, and burning with a peculiar suffocating odour. The other constituent is iron, a well-known metallic substance, requiring an intense heat for fusion, and not burning at a red heat. If we perform any experiment which, in the present state of knowledge, ingenuity could suggest, we are totally unable to cause either the sulphur or the iron to assume a more simple or elementary state of existence. We can with ease cause either of them to enter into new combinations with other bodies, and these compounds we can decompose as in the case of the pyrites and obtain both sulphur and iron again in their separate forms with all their characteristic proper- ties ; but nothing more than this can be effected : hence we are led to the belief that both sulphur and iron are elements, or bodies containing only one kind of matter. The following table gives the elements discovered up to the present time. Those substances whose names are printed in italics have hitherto been found of no practical use ; and those marked with an asterisk (*) are often found native, or unassociated with mineralising elements. Names of the Elements Aluminium Antimony Arsenic Barium Beryllium * Bismuth Boron Bromine Cadmium Ctesium Calcium * Carbon Cerium Chlorine Elements. - - i Symbols Atomic Weights Names of the Elements Al 27 Chromium Sb 119-6 Cobalt As Ba 74-9 136-9 *Copper Didymium Be 9 Erbium . Bi 207-5 Fluorine . B 11 Gallium . Br 80 Germanium Cd 112 *Gold . Cs 132-7 Hydrogen . Ca 40 Indium C 12 Iodine Ce 141-2 *Iridium Cl 35-5 Iron . Symbols Cr Co Cu Di E F Ga Ge Au H In I Ir Fe A.tomic Weights 52-5 58-6 63-5 146 166 19 69-9 72-3 196-2 1 113-4 126-5 192-5 56 CHEMICAL NOMENCLATURE. Elements cont. "ssas" ! s "" b <"> Atomic Weights Names of the Elements Symbols Atomic Weights Lanthanum La 138-5 Selenium . Se 78-9 Lead. . Pb 207 Silicon Si 28 Lithium . Li 7 *Silver Ag 108 Magnesium j Mg 24 Sodium Na 23 Manganese : Mn 54-8 Strontium Sr 87-3 Mercury . Hg 199-8 * Sulphur S 32 Molybdenum Mo 95-9 Tantalum Ta 182 Nickel . ! Ni 58-6 Tellurium Te 127-7 Niobium (Co Terbium . Tb ? lumbium, Cb) Nb 93-7 Thallium . Tl 203-7 Nitrogen . N 14 Thorium . Th 232 Osmium . Os 195 Thulium . Tm ? Oxygen . O 16 Tin . Sn 118 Palladium . j Pd 106-2 Titanium . Ti 50-2 Phosphorus \ P 31 Uranium . U 239-8 Platinum . | Pt 194-3 Vanadium V 51-1 Potassium K 39 Wolfram(Tung Rhodium . i Rh 104-1 sten) W 183-6 Rubidium Rb 85-2 Ytterbium Yb 172-6 Ruthenium Ru 103-5 Yttrium . Y 89-6 Samarium j Sa 150 Zinc . Zn 65 Scandium '< Sc 44 Zirconium Zr 90-4 The first column contains the name of the element ; the second, the symbol, in which all chemical changes and decompositions are most readily understood ; and the third, the atomic weight. These atomic weights are not given beyond the first place of decimals, to avoid tedious calcu- lation ; for all practical purposes they may be considered accurate. Of the compounds of these elements, only those will be discussed which are likely to fall under the notice of the assayer. The principal compound bodies with which the assayer will have to deal are acids, oxides, salts, and binary, sub- stances containing no oxygen. When a body combines in more than one proportion with oxygen, that compound containing the least oxygen takes the termination ous^ that containing the most ic\ thus,: sulphurous acid, sulphuric acid ; arsenious acid, arsenic acid ; ferrous oxide, ferric oxide ; mercurous u -2 CHEMICAL NOMENCLATURE. oxide, mercuric oxide ; in only one proportion (or when they form only one basic oxide) they are distin- guished by the termination ic, as potassic oxide, aluminic oxide. OXIDES are binary oxygen compounds; they may be divided into three series. The first series comprises those oxides which do not possess the property of combining with acids to form salts they are termed indifferent oxides ; the second contains those capable of uniting with acids to form salts, and called salifiable oxides or bases ; the third comprises those oxides which have acid characters, and form salts by uniting with bases. When an elementary body combining with oxygen forms but one oxide, it is simply called the oxide of that element. Thus we say zinc oxide, potassium oxide, alumi- nium oxide. If the body is capable of combining with oxygen in many proportions, the words proto-, sesqui-, bin-, or per-, &c., precede the term oxide, to express the progressive amounts of oxygen. Most metals form one salifiable oxide, and many of them have two ; these are now gener- ally distinguished by the terminations ous and ic, in the same manner as are the acids. Thus we have protoxide of lead, iron, copper, tin, &c. ; sesquioxide of aluminium, iron, or chromium, &c. ; binoxide or peroxide of manganese, copper, mercury, &c. ; and when we speak of them as salifiable bases, ferrous and ferric oxides ; mercurous and mercuric oxide. Some metals unite with oxygen in still higher proportions ; these compounds are almost always acids, such as chromic acid, stannic acid, antimonic acid, &c. SALTS are formed when an acid unites with a base, and usually the properties of the acid and the base are recipro- cally neutralised ; thus an acid which before combination possesses the power of reddening blue litmus, loses it on combining with the base, and, in like manner, a base which would at first change reddened litmus paper to blue loses this property as the acid saturates it. In this case the acid and base have combined to form a salt. LAWS OP COMBINATION. 5 The names of salts are governed first by the nature of the acid ; secondly, by the salifiable nature of the base ; and, thirdly, by the proportions in which the acid and base are combined. Acids terminating in ic form salts ending in ate. Acids terminating in ous form salts terminating in ite ; and the new names having these terminations are added to the name of the oxide. Thus sulphuric acid and iron protoxide form sulphate of iron protoxide, ferrous sulphate, or, more commonly, iron protosulphate ; arseni- ous acid and iron protoxide form arsenite of iron prot- oxide, ferrous arsenite, or iron protarsenite ; nitric acid and iron sesquioxide form nitrate of iron sesquioxide or ferric nitrate. When the salt exists in the neutral state its name is .formed as above, but if the proportion of acid is greater than in neutral salts, it is termed an acid salt: thus potassium bisulphate is sometimes called acid sulphate of potassium. If, on the other -hand, the base is in excess, the name is preceded by the words sub or basic ; thus, lead subacetate or basic- acetate of lead. Binary compounds containing no oxygen exist very largely in nature, and it is from them that the greater part of our copper, lead, silver, &c., is obtained. When a non-metal combines with a metal to. form a compound which is neither acid nor basic, its name is derived from the non-metal by the addition of the termination uret or ide. The latter term is, however, gradually displacing the former. Thus the compounds of sulphur with iron and chlorine with silver are res- pectively called iron sulphuret or sulphide, and silver chloride. If a non-metal combines with a metal in more than one proportion, the same rule is followed as with the oxygen compounds : thus we have ironjpn>fo-sulphide, iron sesqui-sulphide, andiron Si-sulphide (ordinary iron pyrites or mundic). Laws of Combination. On examining the compounds which the same substances afford by their union in dif- ferent proportions, it has been noticed that the propor- (5 LAWS OF COMBINATION. tions of the elements existing in each compound are definite ; a certain weight of one substance will only com- bine with a certain weight of another substance, and the lowest combining weight of any of the elementary bodies is termed its atomic weight, and is represented by the numbers in the third column of the table of elementary substances. As before stated, all substances combine in fixed or definite proportions ; thus, if 223 parts of oxide of lead are analysed, they will be found to consist of 207 parts of lead and 16 of oxygen. Again, the analysis of 18 parts of water or oxide of hydrogen would give 2 parts of hydro- gen and 1 6 of oxygen ; now, taking hydrogen as unity, we have 207 as the equivalent of lead, and 16 as that of oxygen. If we follow oxygen further in its combinations, it will be seen that' 16 parts of oxygen combine with 1 part of hydrogen. 207 lead. 40 calcium. , 118 tin. 63-5 copper. The above numbers, therefore, represent the equivalents of the respective elements. Again, the equivalent of sulphur is 32, and this represents the weight of sulphur which will combine with the above weights of hydrogen, lead, calcium, tin, or copper to form sulphides of the respective bases. 35*5 parts of chlorine, or 78*9 parts of selenium, also combine with the same weights, viz. hydrogen 1, lead 207, &c., to form chlorides and selenides. Compounds like these are of the simplest class, and consist of single equivalents only ; there are, however, many compounds containing more than two equivalents, in which case the following laws are followed. In one class of compounds the quantity of one of the constituent elements remains constant, while each new compound is formed by the successive addition of another CHEMICAL SYMBOLS. 7 atom of the other constituent element ; and it must also be borne in mind that no element will combine with another in less than its atomic weight. Another series will commence with two atoms of an element united with an uneven number of atoms of another element ; thus we can have binary compounds in the proportion of 2 to 3, 2 to 5, or 2 to 7. The atomic weight of a compound body is the sum of the atomic weights of the elements forming it : thus sulphuric anhydride is composed of one atom or 32 parts of sulphur, and 3 atoms, or 48 parts, of oxygen ; its atomic weight is therefore 80. The atomic weight of any com- pound body may be ascertained by adding together the atomic weights of its constituent elements. Owing to the invariable law of the constancy of chemi- cal compounds, we are enabled to calculate the reaction which occurs between two or more bodies when decompo- sition takes place: thus 174 parts of potassium sulphate contain 80 parts of sulphuric anhydride and 94 parts of potassium oxide ; and if it were desired to obtain lead sulphate by the decomposition of lead nitrate by adding to it the above quantity of potassium sulphate, the exact amount of lead nitrate required would be readily found by adding together the equivalent of the elements forming nitric acid and lead oxide. CHEMICAL SYMBOLS : THEIR EMPLOYMENT AND USES The symbol of an element standing alone signifies one atom of that element. Thus : S implies not only the element sulphur, but 32 parts of sulphur ; a small figure on the right-hand side of the symbol indicates the number of atoms to be represented ; thus, S 2 is equal to two atoms, or 64 parts of sulphur. Two symbols placed thus, FeS, indicate a compound of equal equivalents of iron and sulphur. Separation of elements by the sign + or a comma is employed to show the union of two compound bodies ; thus the com- pound of silver sulphide and lead sulphide may be thus written : AgS + PbS, or AgS,PbS. A large figure on the same line as the symbol, and on its left side, multiplies the CHEMICAL SYMBOLS. whole of the symbols to the first comma or + sign : thus, 2AgS,PbS, or 2AgS + PbS, represents a compound of two atoms of silver sulphide with one of lead sulphide. If, however, it be thus written, 2(AgS, PbS), it means two atoms of the whole of the elements which are inclosed in the brackets. PREPARATION OP THE SAMPLE. CHAPTER II. PREPARATION OF THE SAMPLE WEIGHING. THE selection and preparation of the sample is the first and most important 'operation in assaying. It is of little use for the operator to ascertain with accuracy the per- centage of every individual constituent in the mineral operated on, if the sample does not truly represent the average of the ore. It should be borne in mind that samples of mineral are generally selected for their richness, and represent the most favourable portions of the ore ; and no pains should be spared to secure a sample for analysis which will truly show the bulk of mineral whose value is required to be known. The assayer must always bear in mind the object which his experiments have in view. If they are to ascertain the actual percentage of one or more constituents existing in a certain stone, his labours are comparatively easy, all that is required being to reduce the whole of the specimen to the finest possible state of division, and, having well mixed the powder, to analyse a portion of it. But if it is desired to find out the composition of a special mineral or crystal, the greatest possible care must be taken to remove the whole of the gangue or other im- purities, and to obtain for analysis those portions only which represent with greatest accuracy the pure mineral. To effect this the surrounding rock is first removed as carefully as possible, and then the specimen is crushed into coarse pieces on a sheet of clean paper. By means of a pocket magnifier and a pair of pincers, clean, typical pieces of the mineral are then to be selected for analysis. 10 PKEPARATIOX OF THE SAMPLE. If, however, as will most frequently be the case, the object of the assay be to ascertain the average value of a mineral lode or heap of ore, then the assayer must proceed differently. The portion experimented upon must truly represent, in the respective amounts of its valuable ma- terial, gangue, quartz, and earthy matters, the whole bulk of that of which it professes to be a sample ; and this having been secured, the whole must be carefully pow- dered and passed through fine sieves, taking care that every portion of the mineral goes through. If this be not attended to, it will frequently happen that the few grains left out are sufficient to vitiate the whole assay ; this is especially apt to be the case when examining ores the valuable ingredients of which are of a ductile or malle- able nature, such as auriferous quartz. In this case it fre- quently happens that the great, bulk of gold exists in the form of one or two small pieces, and these being flattened and beaten out in the operation of powdering will almost certainly be left upon the sieve. In cases like this it is better to collect and assay such pieces separately, and esti- mate their proportion to the whole weight of the sample, than to attempt to powder and distribute them uniformly! SAMPLING. The important operation of sampling neces- sarily precedes any process of assay. Dr. Peters, in his ' Modern American Methods of Copper Smelting,' gives the following useful description of the operation of sampling as practised in the great mining centres of the New World. By sampling we seek to obtain within the compass of a few ounces a correct representative of the entire quantity of ore, which may vary in amount from a few pounds to several thousand tons. As a rule it will lessen the chance of serious error, in very large transactions, to divide the lot into parcels of not over fifty tons each, and sample each of these lots by itself. The utmost care and vigilance in sampling and assaying should be required at every smelting works, both in the interest of the works and in that of the ore-seller. Until quite recently, it has been customary to sample lots of ore by quartering them down, rejecting a certain PREPARATION OF THE SAMPLE. 11 proportional part at each successive operation, and re- ducing the size of the ore fragments as the quantity to operate on diminishes. This is a laborious and expensive method, and in the case of finely pulverised ores may well be replaced by the use of the ' split shovel,' or one of the many automatic sampling machines that have been invented. But since the establishment of public sampling works at most of our great mining centres, where the correctness of the sample is guaranteed by the works, which dis- tribute packages of each lot of ore to the agents of the various rival smelting companies for them to assay and bid upon, the vast quantities of ores handled, and the im- portance in many instances of retaining the lump form of the ore as essential to the subsequent metallurgical opera- tions, have imperatively demanded some method of auto- matic sampling that shall be rapid, accurate, and equally applicable to ores in both the pulverised and lump form. The means hitherto employed all depend upon the same general principle of cutting or dividing a falling stream of ore by means of flanges, fingers, ore travelling buckets, in such a manner as to obtain a certain desired proportion of it for a sample. While many of these devices work admirably upon pulverised ore, free from dampness or foreign obstructing substances, they are apt to give entirely unsatisfactory results upon a mixture of fine and coarse ores, while the presence of strings, chips, rags, &c., usually clogs them and deranges their working. Mr. D. W. Brunton, of Denver, Colorado, has invented an automatic sampling machine that is apparently free from all the defects enumerated, and which has been shown by practical trial to be equally applicable to coarse, fine, or mixed ores, while it cannot be clogged by foreign bodies of any reasonable size.* Brunton overcomes these difficulties by deflecting the entire ore stream to the right or left, while falling through * See Transactions of the American Institute of Mining Engineers, vol. xiii. p. 639, for drawings and full description of this sampler. 12 PREPARATION OF THE SAMPLE. a vertical or inclined spout. By a simple arrangement of movable pegs, in connection with the driving gear, the proportion of the ore stream thus deflected into the sample bin may vary from 10 to 50 per cent. ; the latter amount only being required in coarse ores of enormous and very variable richness, while for ordinary lump ores from 10 to 20 per cent, is the maximum required. Instead of passing the sample-stream of ore into a bin, this system may be still further perfected by leading it directly to a pair of moderately fine rolls, the product of which is elevated to a second similar sampling-machine from which the final sample drops into a locked bin. Six months' constant experience with this sampler has shown that 10 per cent, of 20 per cent., or 2 per cent, of the original ore parcel is usually quite sufficient ; though in exceptional cases, 15 per cent, of 30 per cent., or \\ per cent, of the ore may be required. The two machines are driven at different speeds, to prevent any possible error that might arise from isochronal motion, and by careful tests of this machine in resampling lots of ore the limit of error has been found less than one-fourth of 1 per cent. ; while even the best hand- sampling may vary 2 per cent. The fact that the division is one of time and not of ore is one of the most important features of this valuable in- vention, as it consequently is forced to deflect the exact proportion of the ore-stream, for which it is set, whether coarse or fine, wet or dry, Hght or heavy. MOISTURE. The determination of the moisture present in any given parcel of ore is also a matter of much im- portance ; and probably more inaccuracies attend this apparently simple process than any other of the pre- liminary operations. This determination must, of course, take place as nearly as possible at the same time that the entire ore parcel is weighed, as otherwise the sample may lose or gain moisture. In lump ores, it is difficult to obtain a correct sample even for moisture, without some preliminary crushing, PREPARATION OF THE SAMPLE. 13 and to save labour, it is best to use a portion of the regu- lar assay sample for this purpose ; the accurate weighing of the entire ore parcel being postponed until just before or after the sampling, and the portion reserved for the moisture determination being placed in an open tin vessel, contained in a covered metal case, having an inch or two of water on its bottom, in which sample tins stand. From one-fourth to one-half pound of the sample is usually weighed out for this determination, and dried under frequent stirring, and at a temperature not exceeding 212. While it is always important to keep within the limit of temperature just mentioned, it is especially the case with certain substances which oxidise easily. Among these are finely divided sulphides, and, above all, the pulverulent copper cements obtained from precipitating copper with metallic iron from a sulphate solution. Such a sample, containing actually 5J per cent, of moisture, showed an increase of weight of some 2 per cent., on being exposed for thirty minutes to a tempera- ture of about 235 Fahr. Certain samples of ore especially from the roasting furnace are quite hygroscopic, and attract water rapidly after drying. In such cases the precautions used in analytical work must be employed, and the covered sample weighed rapidly in an atmosphere kept dry by the use of strong- sulphuric acid. The sampling of the malleable products of smelting, such as blister copper, metallic bottoms, ingots, &c., can only be satisfactorily effected by boring a hole deeply into a certain proportional number of the pieces to be sampled. Where such work is only exceptional, an ordinary ratchet hand-drill will answer, but in most cases a half- inch drill run by machinery is employed. The chips and drillings are still further subdivided by scissors, and, as even then it is difficult to obtain an abso- lutely perfect mixture, it is best to weigh out and dissolve a much larger amount than is usually taken for assay, 14 PREPARATION OF THE SAMPLE. taking a certain proportion of the thoroughly mixed solu- tion for the final determination. The ore must always be reduced to a powder, more or less fine, according to the nature of the chemical operation or assay to which it is to be subjected. This division is effected by means of the anvil, hammer, pestle and mortar, sieve, method of elutriation, or other means generally in use for the preparation of any fine powder. The actual process to be adopted must vary according to the nature of the different bodies under examination. In some cases simple crushing is sufficient ; in others the ore will have to be pounded in a mortar ; whilst occasionally it is neces- sary to reduce it to the very highest degree of fineness by elutriation. There are other operations us strictly mechanical as are the above, viz. washing, dressing, and vanning a sample of ore, the end and aim of which is to separate, in a suitable vessel, by means of water and difference of specific gravity, the earthy or useless and, in some cases, objectionable portion from the heavier metallic and valu- able portion. This operation is almost always employed on the larger scale in dressing ores for the smelter. The tools and materials employed in preparing the sample in the assay laboratory are the anvil (and stand), vice, hammer, files, cold chisel, shears, pestle and mortar, *t eel-crushing mortar, sieve, &c. THE ANVIL (fig. 1). The anvil is most useful in size when it weighs about 28 Ibs. ; but one of 14 Ibs. will suffice. The anvil recommended is of the shape usually employed by the blacksmith. The anvil-stand is constructed of stout wood, about two inches in thickness, and forms a cube of about two feet square. It contains three or four drawers, which serve to hold the hammers, cold chisel, shears, files, &c., which are required in an assay office. In the centre the anvil is fixed, and in one corner: a vice may be also secured. In general the anvil and hammer are employed for the purpose of breaking a small fragment, from a mass of ore PREPARATION OF THE SAMPLE. 15 for examination, or ascertaining whether the button or prill of metal produced in an assay be malleable or other- FlG. 1. wise. The anvil is also exceedingly useful as a support for a crucible while breaking it to extract the metallic or other valuable contents. THE HAMMERS (figs. 2 and 3), of which two are requisite, ought to have one end flat and FIG. 2. FIG. 3. square and the other pick- or wedge-shaped. The horizontal wedge end of fig. 2 is useful for breaking open crucibles and in detaching small fragments from a specimen of ore. The flat end serves for ascertaining the mal- leability of buttons of metal. This hammer should weigh about 1 Ib. The larger hammer, fig. 3, should weigh about 4 Ibs., and is employed for breaking coke suffi- ciently fine for the use of the fur- nace, and detaching fragments from refractory minerals, in both of which cases either 16 PREPARATION OF THE SAMPLE. end may be employed, as may seem most serviceable to the operator. The flat end of this hammer is also used for driving a cold chisel in separating masses of gold, silver, copper, lead, &c., for assay. This hammer has a vertical pick or wedge end. Very hard and stony materials which have to be broken on the anvil (and all such ought to be so treated) scatter many fragments, to the certain loss of a proportion of the substance, and the probable injury of the operator ; this can be prevented by wrapping the mineral in a piece of stout brown paper, or if necessary in several folds. The fracture can then be safely attempted. This latter precaution must be specially taken in frac- turing gold quartz, or hard rock containing metallic silver, FIG. 4. FIG. 5. as the loss of a very minute quantity of metal would involve a considerable error in the result afforded by the assay. All minerals, unless very friable, must be reduced to a moderate size say that of a walnut by means of the anvil and hammer, before pulverisation ; otherwise, if the reduction be attempted in a mortar, it is nearly certain to be injured ; moreover, the operator will find his labours much abridged by using the anvil for this purpose. The anvil can also be made very serviceable in repoint- ing worn or burnt-out tongs &c. It need scarcely be added that it must be placed as far as possible away from bottles or other frangible articles, otherwise accidents may occur by the forcible projection of fragments of crucibles, stones, &c. THE COLD CHISEL (fig. 4) is employed for cutting off me- tallic masses for assay. It should be five or six inches long and about half an inch wide, which is the best size for general use. However, for some purposes, as cutting copper PREPARATION OF THE SAMPLE. 17 and other very tough metals, it is convenient to have a chisel only a quarter of an inch wide, as these metals are so much more difficult to cut, and the small chisel meets with the least resistance. Small shears (fig. 5) are also exceedingly useful in cutting off pieces of sheet metal, such as lead, for cupel- lation, scorification, &c. THE PESTLE AND MOKTAR. Mortars are made of various materials, as cast-iron, bronze, porcelain, agate, &c. ; the assay er requires one of cast-iron, one of porcelain, and one of agate. The iron mortar (fig. 6) ought to be of the capacity of from three to four pints ; the porcelain (Wedgwood ware) (fig. 7) may contain about two pints. The ease with which FIG. 6. FIG. 7. a mortar may be used depends much upon its form, and opinion is greatly divided on the subject. Faraday* says that the pestle should be strong, and the size of its upper part sufficient to allow of its being grasped firmly in the hand, and below to permit a considerable grinding surface to come in contact with the mortar. Its diameter in the lower part may be about one third or one fourth of the upper diameter of the mortar. The curve at the bottom should be of shorter radius than the curve of the mortar, that it may not touch the mortar in more than one part, whilst at the same time the interval around may gradually increase, though not too rapidly, towards the upper part of the pestle. The bottoms of all mortars ought to be of considerable * Chemical Manipulation, p. 149. 18 PREPARATION OP THE SAMPLE. thickness, in order to withstand the smart blows they will occasionally have to receive. Berzelius recommended a mass of pumice-stone for cleansing porcelain mortars. It is used with water as a pestle, and in course of time will be worn to the shape of the mortar ; its action will then be more speedy. Iron mortars can be best cleaned by friction with a little fine sharp sand, if washing be not sufficient to re- move the adhering substance. Great care must be taken to dry mortars perfectly, especially those of iron, other- wise they will become rusted, and the rust will contaminate the substances pulverised in them. The iron mortar is principally of use in the reduction of the masses of mineral (broken on the anvil, as before described) to a state of coarse powder, in order to render the substance more readily capable of pulverisation, strictly so called. In the use of the iron mortar, all friction with the pestle ought to be avoided, and the body within it must be struck repeatedly and lightly, in a vertical direc- tion, taking care to strike the large pieces, so that all may be equally reduced. This can be carried on until the whole is about the size of fine sand. It is transferred to the porcelain mortar, where direct blows must be carefully avoided. The process is now carried on somewhat differently ; the pestle is to be pressed with a moderate force, and a circular motion given to it, taking care every now and then to lessen and then to enlarge the circles so as to pass over the whole grinding surface of the mortar, and insure the pulverisation of the mass of mineral submitted to operation. In general, the finer the state of division to which a mineral is reduced, the more accurate and ex- peditious will be its assay ; and in preparing a mineral for assay by the wet method, no labour ought to be spared on this point. Pulverisation is rendered much easier by operating on a small quantity at once, and removing it very often from the sides and bottom of the mortar by means of a spatula. The quantity operated on at one time must be regulated by the hardness and friability of PREPARATION OF THE SAMPLE. 19 the substance whose pulverisation is to be effected. The harder it is, the less must be taken, and vice versa. In the use of the iron mortar fragments are occasionally projected. This may be prevented by covering the upper part of the mortar with a cloth. This applies also to the porcelain mortar, for the dust of some minerals has a disagreeable taste and smell. Indeed, in some cases the ambient powder is highly deleterious, as in the pulverisa- tion of arsenical nickel, cobalt, and other ores. Here the simple cloth is not a sufficient protection ; it should be slightly damped with water, and tightly tied round the mortar, and firmly held round the pestle, when nothing can escape. Some minerals can be pulverised with greater ease if they are ignited and suddenly quenched in cold water. Amongst them may be named flint, and many other sili- ceous matters, as gold quartz. In the pulverisation of charcoal for assays, it will be found useful to heat it, as hot charcoal is more readily pulverised than cold. In some instances the powder obtained in the iron or porcelain mortar is not fine enough ; recourse should then be had to the agate mortar, in which the mineral, in as fine a state of division as the larger mortars will give it, is ground in small portions at a time, until it is reduced to an impalpable powder. When small specimens or rare minerals are being ope- rated upon, if it is especially desirable to avoid loss, it is advisable to use a steel mor- tar (fig. 8) for the prepara- tory reduction of the mineral to coarse powder. A B, C D, and E F represent the three component parts of the mor- tar ; these may be readily taken asunder. The substance to be crushed (having, if practicable, first been broken into c 2 L>0 PREPARATION OF THE SAMPLE. small pieces) is placed in the cylindrical chamber E F; the steel cylinder, which fits somewhat loosely into the chamber, serves as a pestle. The mortar is placed upon a solid support, and perpendicular blows are repeatedly struck upon the pestle with a hammer, until the object in view is attained. (Fresenius.) In the selection of agate mortars, they must be examined to see that they have no palpable flaws in them ; very slight cracks, however, that cannot be felt, do not render the mortar useless, although they increase the danger of its destruction by a chance blow. THE SIEVE. The operation of sifting is employed when a very fine powder is required, or when a powder of uniform size is needed. Sieves of various materials and different degrees of fineness are necessary. The larger sieve, for preparing coke for the blast furnace, is made of stout iron wire, and must have its meshes from 1 inch to 1J inch square. The fine coke, which is sifted from that which is the proper size for the blast furnace, may be mixed with that of ordinary size, and employed economi- cally in the muffle furnace. For the preparation of minerals a set of three sieves should be provided, each one finer than the other. The coarsest may contain 40 holes to the linear inch, the finer or medium sieve 60, and the finest from 80 to 100. The coarsest sieve is used for preparing galena for assay ; the medium for copper, tin, iron, and other like ores ; and the finest for gold and silver ores, or for preparing any substance for the wet assay, as, in the latter case, the finer the state of division the substance attains, the more rapid will be its solution or decomposition by the liquid agents employed. The sieve fig. 9 is made of wood, over which is strained in the ordinary manner brass wire-gauze of the necessary degree of fineness. When in use, j5, fig. 10, is fitted into the lower part of A (same figure). This contrivance pre- vents all loss of the fine powder. If the matter to be sifted be offensive or deleterious to the operator, a sieve termed the drum or box-sieve may be employed (see fig. 10), where C represents a cover fitting over the sieve. If SIFTING. 21 small, this may be used in the ordinary way ; but if large, its method of use is rather peculiar, and requires some practice to fully develop its powers. One side of the under edge must be FIG. 9. FIG. 10. held by one or both hands according to its size, whilst the other rests on a table or a bench. A semicircu- lar oscillating motion must now be com- municated to it by moving the hands up and down at the same time that they are being alternately brought into approximation with the sides of the operator. In cases of necessity, a sieve may be readily extem- porised. Place the powder to be sifted in a piece of fine lawn or muslin, according to the fineness required, tie it up loosely, and shake or tap the powder, with, its muslin or other envelope, on a sheet of paper, and the sifting will be rapidly and easily accomplished. The sieve is also extremely serviceable in the separation of some ores from their gangues or vein-stones, especially if the latter be stony and hard. This point must be par- ticularly noted, as it is the cause of much variance between the results of different assayers ; for instance, part of the same sample of ore might be sent to two assayers, and the produce made by one would be 8^ per cent., and that by the other 9 or 9^, or, in some cases, even more. This dis- crepancy generally arises from the cause above mentioned. In the one case the workman has rejected part of the hard gangue, and so rendered the residue richer ; whilst in the other he has pulverised the whole, making the produce less, but giving more accurately the amount of .metal in the substance submitted to assay. A knowledge of this fact is also very useful from another point of view. Suppose it were wished to separate '2*2 SIFTING. in a speedy manner, as perfectly as possible, any friable mineral, such as galena or copper pyrites, from its matrix by mechanical means, it might be accomplished by the use of the sieve, as follows : Place a small quantity of the mineral in an iron mortar, and strike, repeatedly, slight vertical blows. When it is tolerably reduced, sift it, and it will be found that what passes through is nearly pure mineral, with only a small quantity of matrix ; repeat the pounding and sifting operations, until, after a few repetitions, that which remains in the sieve is nearly pure gangue. Native metals, as gold, silver, and copper, are also* partially separated after the manner above described. The fine particles of metal, during the process of pounding and trituration, become flattened, and cannot pass through the sieve, whilst the more brittle portions pass through and are separated. ELUTRIATION. This process can only be employed for those bodies which are not acted on by water ; and it must be remembered that many substances which are usually considered to be insoluble in water are, when in a very finely divided state, acted upon to a greater or less extent. The operation is thus effected : The substance is reduced to the finest possible state of division by any of the fore- going processes ; it is then shaken up with a quantity of water in a glass or other vessel, After a few moments' repose, the supernatant liquid, retaining in suspension the finer particles of the pulverised substance, is poured off, and the grosser parts, which have fallen to the bottom of the vessel, are repulverised, and again treated with water. By repeating these processes a powder of any required degree of fineness may be obtained. It is seldom, however, that a substance is required for assay by the dry way, in such a minute state of division. In the humid or wet method it is occasionally very useful. If the supernatant water is roughly decanted off, where the powder to be elutriated is light, the least disturbance of the vessel containing it occasions the distribution of the portion which has settled, throughout the liquid, and the ELUTRIATION. 23 consequent mixture of fine and coarse particles. This can be avoided by the employment of the syphon. The operation is then thus conducted : The syphon is filled with water, and the shorter end placed in the liquid whose transversion is to be effected : the forefinger of the right hand, which, during this time, has been applied to the longer end of the instrument, is now removed, when the water will flow out until it is level with the immersed end of the syphon. Fresh water can then be added, the powder stirred up again, and the operation of decantation by the syphon carried on as long as requisite. WASHING, DRESSING, OR VANNING. This operation is ex- ceedingly useful for discovering the approximate quantity of pure ore, such as galena, copper pyrites, oxide of tin, native gold or silver, in any sample of earthy matter or ore in which it may be disseminated. The theory of the operation about to be described is easily understood. Bodies left to the action of gravity in a liquid, in a state of rest, experience a resistance to their descent which is proportionate to their surface, whatever may be their volume and density. Hence it follows, firstly, that of equal volumes the heaviest fall most rapidly; secondly, that of equal densities those having the largest size move with the greatest speed ; for in particles of un- equal size and like form the weight is proportional to the cube of the dimensions, whilst the surface is only propor- tional to the square of these dimensions ; hence in small particles the surface is greater in relation to the weight than in the large particles. Thirdly, of equal densities and volumes, particles offering the largest surface (those which are scaly and laminated, for example) undergo more re- sistance in their motion than those which, approaching the spherical form, have less surface. The adhesion of the liquid to the particles of bodies held in suspension is also an obstacle to their subsidence. This force, like the dynamic resistance, is proportional to the surface and in- dependent of the mass or volume ; whence it follows that, in a fluid in motion, of bodies having equal volumes, the least dense acquire the greatest rapidity of movement, and 24 DRESSING OR VANNING. are deposited at the greatest distance from the point of departure ; whilst with equal densities the smallest grains are carried farthest ; and lastly, with equal densities and volumes, the particles exposing most surface traverse the greatest space. It is, therefore, evident that the most advantageous condition for separating, by washing, two substances of unequal specific gravity or density is that the heavier shall be in larger grains than the lighter ; this unfortu- nately, however, is a condition that can very seldom be fulfilled, as the heaviest substances are those metallic minerals whose frangibility is nearly always greater than the earthy matters accompanying them as gangues. This being the case, it is very important so to arrange that the fragments of the various mixed substances shall be nearly of the same size. This may be effected by very frequently sifting the mineral during the process of pulverisation, reducing it also more by blows than by grinding, so as to get as little fine powder as possible, as that is nearly certain to be washed away during the process. The operation of washing or vanning may be performed by one of two methods. In the first, a small stream of running water is employed ; in the second, water is added to the substance to be washed, and poured off as necessary. In the first process, a vessel somewhat resembling a banker's gold scoop (but longer in proportion) is employed ; the mineral to be washed is placed in the upper part, and a small quantity of water added, with which the mineral is thoroughly and carefully moistened, and mixed with the fingers. The scoop must then be so inclined that a fine stream of water from any convenient source (say a tap) may fall just above the upper part of the mixture of mineral and water ; then, firmly holding the larger and upper end of the scoop with the left hand, and sustaining the lower part with the right, it is shaken frequently in the direction of its longitudinal axis. At each shake all the particles in the scoop are so agitated that they become suspended in the water, and the current of liquid running from the tap into the scoop moves them all in its own direction ; DRESSING OR VANNING. 25 but they are deposited at different distances from the point at which the water enters, the heaviest being carried through but a very small space. It is now soon seen that the mineral assumes a heterogeneous surface ; at the upper part, the heavy portions are seen nearly pure ; the light substances, on the other hand, are nearly without mixture at the lower end, and in the intermediate part the heaviest portion of the mixture is nearest the upper end. If the washed matter were now to be divided into horizontal layers, the heaviest matter would be found at the bottom, and the lightest on the surface. Things being in this state, the scoop must be made to oscillate on its axis, so that the latter remain immovable, and in a slightly in- clined position. In this manner, the layer of water running over the surface of the mineral agitates that part only, and carries off all light substances there deposited in the previous operation. When necessary, these matters may be removed by the finger, and made to run into a vessel placed below the scoop, in which all the water and matters carried off are received. This operation, however, must not be hurriedly performed, so as to mix the parts already separated : each layer must be removed separately, commencing with the upper one. This being done, the scoop must be alternately kept in motion by shakings, as at first, and then on its axis, and the washing off of the finer particles renewed, and so on until the separation is effected as far as may be judged necessary. At the commencement of the operation, the water carries out of the scoop the lightest particles, as organic matter, clay, &c, ; at a little later period the water carries with it a small but definite quantity of the heavier portion, the proportion of which increases as the operation proceeds, until at last the greatest possible care is required. It is always better to rewash the latter portion which passes off from the scoop ; hence the necessity of allowing all the wash-water passing from it to collect in a vessel placed for that purpose. In the second method of washing, a tin, zinc, or wooden pan is employed. It should be circular, one or two feet in 20 DRESSING OR VANNING. diameter and three or four inches deep ; the sides should descend in a conical manner, so that the bottom is not more than four inches in diameter, and the angle between it and the sides as sharp as possible. The substance to be examined is placed in the washing- dish, the latter filled with water, and the mineral well mixed with it until perfectly moistened as before. After a moment or so the muddy water is poured off, and the operation repeated until the water passes off clear. When this happens, only so much water must be placed in the pan as will leave a slight layer on the mineral. Now, by holding the pan in one hand, and shaking it with the other, the greater part of the heavy mineral, gold or otherwise, will fall below the sand. If now the pan be inclined towards the hand which is shaking it, the lighter portions, even if tolerably large, will flow off with the water, leaving the heavier matters in the angle, from which, with ordinary care and a little practice, it is difficult to disturb them* If there be a large quantity of earthy matter, this may be (after sufficient shaking) removed by the finger, as in the first-described process. By careful repetitions of these processes, the whole, or nearly the whole, of the sandy and earthy matters may be removed, and the gold or other mineral left nearly pure. This is the plan employed in prospecting for gold, diamonds, and other gems, and in some cases for their commercial extraction. In Cornwall and other mining counties this operation is very cleverly and carefully performed on the miner's common shovel, and the richness of any particular sample of either tin, lead, or copper is thereby determined with a very near approach to accuracy. THE BALANCE. OPERATION OF WEIGHING. At least three balances will be required in a laboratory where general assays are per- formed. The first must be capable of carrying three or four pounds in each pan, and must turn with a quarter of a grain. This may be of the form of the bankers' or bullion balance (fig. 11), and may be employed in weigh- WEIGHING. ing samples of gold quartz or silver ore containing metallic grains capable of being separated by the sieve (see p. 21) ; the second (fig. 12), or rough assay balance, is similar to FIG. 11. 28 WEIGHING. the apothecary's scales ; it should take 1,000 grains in each pan, and turn with one tenth of a grain. This serves for weighing samples of ore and fluxes for assay, and for determining the weight of buttons or prills of lead, tin, iron, copper, &c., obtained in an assay. The third and most delicate^ or true, assay balance (fig. 13) should carry about 1,000 grains; must turn distinctly and accurately with the -roW tn of a grain. This is employed in the assay of gold and silver, bullion, and in the assay of minerals containing gold and silver ; also for general analytical purposes. The first two balances may be used, with ordinary care, by any one ; but the third balance, in its use and adjustment so as to maintain its extreme accuracy, requires some particular instructions, which necessarily involve the principle of the balance. These have been so admirably given by Faraday, in his * Chemical Manipulations,' that we can do no better than transcribe them : 4 The theory of this balance is so simple that the tests of its accuracy will be easily understood and as easily practised. It may be considered as a uniform inflexible THE BALANCE. 29 lever, supported horizontally at the centre of gravity, and supporting weights at equal distances from the centre by points in the same horizontal line with the centre of gravity. If the weights be equal the one will counter- poise the other ; if not, the heavier will preponderate. In the balance, as usually constructed, there are certain departures from the theory as above expressed some from the impossibility of execution, and others in conse- quence of their practical utility ; and a good balance may be said to consist essentially of a beam made as light as is consistent with that inflexibility which it ought to possess, divided into two arms of equal weight and length by a line of support or axis, and also terminated at the end of each arm by a line of support or axis, intended to sustain the pans. These three lines of support should be exactly parallel to each other in the same horizontal plane, and correctly perpendicular to the length of the beam ; and the plane in which they lie should be raised more or less above the centre of gravity of the beam, so that the latter should be exactly under the middle line of suspension. It will be unnecessary in this place to speak of the coarse faults which occur in the ordinary scales these will be easily understood ; and from what has to be stated of the examination of the most delicate instrument, the impos- sibility of avoiding them without incurring an expense inconsistent with their ordinary use will be as readily comprehended.' Two principal things have to be attended to in the selection of a balance its accuracy and its delicacy. The accuracy depends upon the following conditions : 1. The arms should be equal to each other in length. The length of each is accurately the distance from the middle to the distant knife-edge, all the edges being considered parallel to each other, and in the same plane. The two arms should accord perfectly in this respect. This equality may be ascertained in two or three ways. Suppose the balance with its pans to vibrate freely, and rest in a horizontal position, and that after changing the pans from one end to the other the balance again takes its horizontal 30 WEIGHING. state of rest in such a case an almost certain proof is obtained of equality in length of the arms. They may, however, be equal, and yet this change of the pans from end to end may occasion a disturbance of equilibrium, because of the unequal distribution of weight on the beam and pans ; but to insure an accurate test, restore the pans, and consequently the equilibrium, to the first state : put equal, or at least counterpoising, weights into the pans, loading the balance moderately, and then change the weights from one pan to the other, and again observe whether the equilibrium is maintained ; if so, the length of the arms is equal. Equality of weight is not so necessary a condition, although this should be obtained as accurately as possible. One arm with its pan may be considerably heavier than the other, but from the disposition of the weight in the lighter arm towards the extremity, or in the heavier towards the middle of the beam, the equilibrium may be perfect, and therefore no inaccuracy be caused thereby in the use of the balance. Instruments are usually sent home In equilibrium, and require no further examination as to this particular point than to ascertain that they really are in adjustment, and that after vibrating freely they take a horizontal position. 2. The beam must be of such a form and strength that it will not bend when loaded with the greatest weight the balance is intended to carry. All well-made modern balances are sufficiently rigid in this respect, and may be safely trusted to carry their full weight without flexure of the beam. The beam should also be as light as practicable. 3. The knife-edges supporting the pans, and the centre one on which the beam vibrates, must be accurately in the same line. The delicacy of a balance likewise depends upon several conditions. The centre of gravity must be very little below the fulcrum. If it be considerably depressed, then, upon trying the oscillations of the balance by giving it a little THE BALANCE. 31 motion, they will be found to be quick, and the beam will soon take its ultimate state of rest ; and if weights be added to one side, so as to make it vibrate, or to bring it to a certain permanent state of inclination, the quantity required will be found to be comparatively considerable. As the centre of gravity is raised the oscillations are slower, but producible by a much smaller impulse ; the beam is a longer time before it attains a state of rest, and it turns with a smaller quantity. If, however, the centre of gravity coincides with the fulcrum or centre of oscillation, then the balance is said to se^ that is, the smallest possible weight will turn the beam ; the oscillations no longer exist, but one side or the other preponderates with the slightest force, and the valuable indication which is furnished by the extent and velocity of the vibrations is lost. The case in which the centre of gravity is above the fulcrum rarely if ever occurs. Such a balance, when equally weighted, would set on the one side or the other ; that side which was in the slightest degree lower tending to descend still further, until obstructed by interposing obstacles. In balances intended to carry large quantities (as in the balance for weighing gold quartz, &c.) it is necessary to place the centre of gravity lower than in those for minute quantities, that they may vibrate regularly and readily. This is one cause why they are inferior in deli- cacy, for, as a consequence of the arrangement, they will not turn except with a larger weight. Balances are also liable to set when overloaded. Thus, if a balance be equally weighted in each pan, but over- loaded, it will, if placed exactly horizontal, remain so, .but the slightest impulse or depression on one side destroys the equilibrium ; the lower side continues to descend with an accelerated force, and ultimately remains down, being to all appearance heavier than the other. Generally speaking, the more delicate a balance the sooner this effect takes place ; this is one limit to the weight it can properly carry. 32 WEIGHING. The vibrations of a balance vary with the quantity of matter with which it is loaded : the more the weight in the pans, the slower the vibrations. These should be observed, and the appearances retained in the mind, in consequence of the useful indications they afford in weighing. A certain amplitude and velocity of vibration would indicate to a person used to the instrument nearly the weight required to produce equilibrium ; but the same extent and velocity, with a weight much larger or smaller, would not be occasioned by an equal deficiency or redund- ancy of weight, as in the former case. The weight also required to effect a certain inclination of the beam, or to turn it, should be known, both when it is slightly and when it is heavily loaded. Thus, if the instrument turns with T ^Vo tn of a grain, with 1,000 grains in each pan, or with T o~o-g-^- o^h of the weight it carries, it may be considered perfect. The friction of the knife-edges must be as slight as possible. Most of the faults in the working of a balance, if ordinarily well made, depend upon imperfections in the middle knife-edge and the planes upon which it rests. The edge is made either of agate or ste.el, preferably the former, and should be formed out of one piece, and finished at once, every part of the edge being ground on the same flat surface at the same time. In this way the existence of the two extreme or bearing parts of the edge in one line is insured ; but when the two parts which bear upon the planes are formed separately on the different ends of a piece of agate or steel, or, what is worse, when they are formed on separate pieces, and then fixed one on each side the beam, it is scarcely possible they should be in the same line : and if not, the beam cannot be correct. These knife-edges usually rest on planes, or else in curves. The planes should be perfectly flat and horizontal, and exactly at the same height ; the curves should be of equal height, and their axes in the same line. If they are so, ancl the knife-edge is perfect, then the suspension will be accurately on the line of the edge, and reversing the beam THE BALANCE. 33 will produce no change. The balance must always be kept perfectly level by means of the three screws on which it stands, and adjusted by the spirit-level or plum- line with which it is furnished. The balance should be kept in a well-lighted dry room, quite away from acid or other vapours. The case should be kept closed as much as possible, and a glass vessel full of lumps of good quick-lime should be kept in it. When the lime falls to powder it should be renewed. In order to test the accuracy and delicacy of a balance, remove the pans and their end supports, and notice how the beam oscillates. When it has been found to oscillate with regularity, and gradually to attain a horizontal posi- tion of rest, it should be reversed -that is, taken up and turned half-way round, so as to make that which before pointed to the right now point to the left. The beam should then again be made to oscillate, and if it perform regularly as before, finally resting in a horizontal position, it has stood a severe test, and promises well. Then re- place the pans and repeat the tests, noticing the time required for each oscillation. When the pans are hung upon the beam, the balance should of course remain horizontal. They should be tried by changing, then by reversing the beam, and afterwards by changing the pans again. The pans are best suspended by very thin platinum wire, so as to avoid hygrometrical influence upon them. Afterwards load the balance with the full weight it is intended to carry say 1,000 grains in each pan, and notice if the indications are as rapid upon adding or subtracting the smallest weight as they were when the pans were empty. Tests of this kind are quite sufficient for the purpose of the assayer, who, having ascertained that his balance, whether slightly or fully laden, vibrates freely, turns deli- cately, and has not its indications altered by reversing the beam or changing counterpoising weights, may be perfectly satisfied with it. The irregularities which may be discovered by these tests are best corrected by a workman ; but as in all the D 34 WEIGHING. best balances now made adjusting screws for these pur- poses are provided, it has been thought advisable to introduce here such matter as, after careful perusal, will enable every one to adjust and examine his balance properly ; so that, in the absence of a skilled workman, it may without much danger be put into working order by the assay er himself, if accidentally damaged by rough treatment. THE WEIGHTS. Various kinds of weights are necessary for the different balances required by the assay er. For the larger balance, Troy-weights from 4 Ibs. to ^ grain will be requisite ; for the second size, weights from 1,000 grains to y^th part of a grain ; and for the assay balance, weights from 1,000 grains to ToVo tn f a g ram - The best material adapted for weights is unquestion- ably platinum. This is, however, too expensive for its general adoption, and therefore brass weights are almost invariably employed down to the ten- or twenty-grain weight, the smaller ones only being of platinum. On the Continent weights are generally made of silver, and if of brass are electro-gilt. For the smallest weights of all (those below 0*10 grain) aluminium is often used, its lightness, and consequently greater bulk, enabling these small weights to be made considerably larger than if they were of platinum. The riders are generally of silver-gilt wire. The slight tarnish which gradually forms on brass weights may be disregarded until it becomes thick. Weights ought never to be touched with the fingers, and should, when not in use, be kept tightly fastened in their box, away from all acid fumes. The most convenient series in which to have the weights is 600, 300, 200, 100, 60, 30, 20, 10, 6, 3, 2, 1, -6, -3, -2, -1, &c. This is preferable to the series formerly employed, as it admits of the use of a less number of weights to arrive at any required amount. According to Deville and Mascart receptacles lined with velvet are not adapted for the preservation of weights, as dust is deposited in the velvet and acts upon the weights when taken out or put in. Small boxes of ivory or smooth wood are preferable. Peculiar weights are necessary for the assay of gold and silver bullion in England (with the exception of assays GOLD AND SILVER ASSAY POUNDS. 35 for the Bank of England ; see Gold assay), gold being reported in carats, grains, and eighths, and silver in ozs. and dwts. The most convenient quantity of either of the precious metals for assay is 12 grains. The quantity taken, however, is of no very great consequence ; but whatever its real weight, it is denominated in England the assay ' pound' This assay ' pound ' is then subdivided into aliquot parts, but differing according to the metal. The silver assay ' pound ' is subdivided, as the real Troy pound, into 12 ounces, each ounce into 20 pennyweights, and these again into halves (the lowest report for silver), so that there are 480 different reports for silver, and therefore each nominal half-pennyweight weighs ^th part of a Troy grain, when the ' pound ' is twelve grains. Assay Weights for Silver. Assay grains. 12 11 6 . 3 2 1 0-500 0-250 0-150 0-100 0-050 0-025 The gold assay 'pound is subdivided into 24 carats, each carat in to 4 assay grains, and each grain into eighths, so that there are 768 reports for gold; and the assay * pound ' weighing 12 Troy grains, the lowest report, or -|th assay grain, equals -g^th Troy grain ; thus Assay Weights for Gold. Silver ozs. dwts. grs. 12 11 600 300 200 100 10 050 030 020 010 12 ( carats. 24 22 12 Joid grs. o eighths. . . o . . o Assay grains. 12 11 . . . .6 3 3 o o Iffths 2 1 o 2 1 . . . . o '. '. '. '. ffths 6 . - 6 -ths o o 3 e?ths 2 . 1 D 2 36 THE WEIGHTS. In cases where the very smallest weights have to be employed, great care must be taken in seizing them with the forceps, as they are apt to spring away and be lost. In the assay balance (fig. 13) the use of weights less than -j^thofa grain is avoided by a very ingenious contrivance. Each side of the beam is equally divided into ten parts, and over the beam on each side is placed a sliding rod, as represented in the figure. The object of these rods is to carry, in the direction of the beam, the small bent piece of wire (letter c, fig. 13) called a rider, which serves in lieu of the smallest weights the T ^th and the yo^^h- These riders are thus employed : one weighing y^th of a grain is placed on the cross-piece of the extremity of the sliding rod just mentioned, and the rod thus furnished is brought gradually along the beam from the centre to the end, until the rider can be deposited on the division on the beam marked 10 ; the balance is then loaded on that side with a weight equal to T V^h of a grain. If now the rod be advanced to the centre of the balance, and the rider dropped on the mark 5, the half of T Vth of a grain will be pressing on that side of the balance, or, in other words,, 05 of a grain ; and when the rider is at the marks 1, 2, 3, 4, respectively, '01, -02, -03, -04 of a grain will be indicated. With a rider weighing y^^th of a grain thou- sandths of grains may be indicated : thus the last rider placed on the marks 1, 2, 3, 4 would equal -01, -0002 y 003, -004 grain, &c. THE METHOD OF WEIGHING. The operation of weighing is very simple ; but as in the hands of the assayer it becomes one of great frequency, the facilities for its performance require to be mentioned. It should in the first place be as- certained before every operation that the balance is in order, so far as relates to its freedom of vibration, and also that no currents of air are passing through the case, so as to affect its state of motion or rest, a situation being chosen where such influence may be avoided. In most cases there is a small projecting arm on the upper part of the beam, which, being turned either to the right or left hand side of the beam as required, serves to establish perfect equili- WEIGHING 37 brium. Perfect equilibrium is, however, a matter of no consequence if the assayer observes one or two simple rules. He should never on any account weigh by the direct method, that is, he should never obtain the weight of a substance by putting it at once into one pan and then counterpoising it by adding weights to the other pan. This method is only to be relied on when the balance is of rare perfection, and is used by no one but the assayer him- self. The plan of weighing by difference should invariably be adopted. By this means the weight of any body can be readily ascertained, no matter whether the arms of the balance are of unequal length or the pans out of equilibrium. In the first place, it should be a rule that one pan, preferably the left, be reserved for the substance to be weighed, and the other pan be set apart for the weights. Supposing the weight of a portion of mineral is re- quired. First place a clean watch-glass, or platinum 16> method of effecting solu- tions will be given in those cases where it is necessary. Where it is necessary to manipulate in acid or other solutions, the glass and platinum forceps de- scribed by David Forbes, F.E.S., in the 'Chemical News' for October 2, 1868, will be found useful. The accompanying woodcut, fig. 15, shows them in front and side view, and will require but little explanation. They are made as follows : an ordinary pair of strong surgical forceps are taken and the points cut off; a small piece of sheet brass, bent into a cylinder, is then soldered to each arm as shown at a ; these cylin- ders being formed by merely bending the brass round, so as to leave an open slit about one twentieth of an inch wide in front. Two glass rods, such as are used for stirrers, as long as the glass arms of the forceps are in- tended to be, and just as thick as will enter these brass cylinders when pressed with some force, are rounded by the blow-pipe at the one end, whilst the other, when softened, is somewhat flattened between the glassblower's pliers, as seen in the woodcut. In order to complete the DISTILLATION. 51 forceps it is now only necessary to push each of these rods into its corresponding brass cylinder or socket, the longitu- dinal slits of which, by imparting a certain amount of elasticity to the sockets, cause them to grasp the glass rods firmly, and retain them without any cement or other fixing. The relative lengths of the arms are easily adjusted by slipping one rod more or less forward, whilst the points can be made to hold and meet accurately, by rubbing them down on a piece of sandstone. Such forceps may, of course, be made to any convenient size ; the one figured in the woodcut is drawn to exactly half-size, and is found to be of very useful dimensions for general analytical work, especially when manipulating in nitro-hydrochloric acid, nitrate of silver, and other solutions which would have acted upon metals, horn, ivory, &c. Fig. 16 represents another convenient form of forceps, also drawn to one half the real size, with long platinum points soldered to the steel body at a\ these have been also found of great service in general laboratory operations, especially when hydrofluoric acid is in question. DISTILLATION. There are two distinct classes of this operation : in the one, liquids are submitted to experiment with the object generally of separating them from sub- stances which are non-volatile, and will consequently be left behind when the liquid comes over. Belonging to this class may be mentioned the distillation of nitric acid, the preparation of distilled water, and the separation of mercury from gold and silver amalgam. In the other kind of distillation, which goes by the name of dry distil- lation^ solid bodies, as wood, coal, &c., are subjected to heat in order generally to ascertain the amount of gas or other volatile matter given off in the course of an experi- ment, from a certain quantity of the coal or other sub- stance operated upon. In liquid distillation (as in the purification of nitric acid, &c.), retorts are used. The best form for general use is that which is furnished with a stopper at the upper part of the body, a (fig. 17), through which the liquid is introduced ; the neck of the retort is then placed in that E 2 52 DISTILLATION FIG. 17. of a receiver, b, over which a piece of wet cotton or woollen cloth is placed, and which must be kept cold by means of a stream of water from a funnel, c, the shaft of which is partially plugged up with cotton wool. Heat is then applied to the retort, and as much of the liquid as is desired is distilled over into the re- ceiver. It is advisable not to fill the retort more than two thirds full, and to apply the heat at first very gently, other- wise there is a risk of breaking the vessel. A more convenient form of apparatus for distillation and condensation is shown at fig. 18, in which a Liebig's FIG. 18. condenser is attached to the retort. Fig. 19 will show the construction of the condensing apparatus. The cold water passes into the funnel above, is conveyed at once to the lowest end of the condenser, whilst the heated water passes off by the upper tube. Distilled water is a most important agent in the labora- tory ; and, as much is needed, it is better to have a still DISTILLATION. 53 specially adapted for its production. Such a one is de- picted at fig. 20, where A is the body of the still ; B the furnace in which it is , /,i ,.-,-. ! , FIG. 19. set (the still may also be placed in the portable furnace, fig. 25, p. 65) ; C the still head ; D E the neck ; F the worm ; I J K L the worm-tub containing cold water to condense the steam generated in the still ; M N the pipe to lead fresh cold water to the FIG. 20. bottom of the worm-tub, while the warm water runs off .at the top, as in Liebig's condenser ; and P the vessel in which the distilled water is received. In the dry distillation of bodies, earthenware, glass, or iron retorts are employed ; but for small operations a tube of wrought-iron, about one inch internal diameter, and plugged at one end, is found to be a convenient form of apparatus. It is placed with the substance contained in it in a furnace, and a small tube, either of glass or pewter, is fixed by means of a perforated cork to the open end of the large tube. The gas given off during the operation may be collected by the aid of a pneumatic trough. 64 SUBLIMATION. SUBLIMATION. This operation is a kind of distillation- in which the product is obtained in the solid form. The apparatus which may be employed for this purpose are tubes, flasks, capsules, or crucibles. Florence flasks are exceedingly useful ; they may be sunk in a sand bath, and the sublimed substance received directly into another flask, or by passing through an intermediate tube. Sometimes, however, it is difficult to entirely remove the sublimed sub- stance ; and in order to avoid this inconvenience, Dr. Ure has proposed the following very excellent subliming appa- ratus : It consists of two metallic or other vessels, one of which is flatter and larger than the other. The substance to be sublimed is placed in the smaller vessel, and its opening is covered by the larger filled with cold water, which may be replaced from time to time as it becomes hot. The sublimed substance is formed on the lower part of the upper vessel. A large platinum crucible, filled with cold water, and placed on the top of a smaller one, answers the purpose of the before-mentioned apparatus very well. SCORIFICATION CuFELLATiON. These operations will be described under the head of Silver Assay. FURNACES. 55 CHAPTEE IV. PRODUCTION AND APPLICATION OF HEAT. FURNACES for assay purposes may be heated either by solid fuel, oil, or gas, and they may be divided into wind and blast furnaces. In the former the fire is urged by the or- dinary draught of a chimney, and in the latter by means of bellows or artificial blast. We shall commence with the former, as they are in most common use. They are of various kinds, according to the purposes for which they are required. The three principal kinds are those for fusion, calcination, and cupellation. Coal, coke, and charcoal are the fuels employed, and the merits of each will be par- ticularly discussed. Blast furnaces are only employed for the purpose of fusion, although their forms are various ; charcoal and coke are the fuels most in use, but oil and gas blast furnaces are used in small laboratory operations, and for many purposes they are preferable to other furnaces, on account of their freedom from dust and dirt, and the perfect control the operator possesses over the heat. Furnaces consist of certain essential parts viz. first, the ash-pit, or part destined to contain the refuse of the combustible employed ; secondly, the bars on which the fuel rests ; these are sometimes made movable, or are fixed to a frame ; the former arrangement is more convenient, as it allows clinker and other refuse matters to be readily removed ; thirdly, the body of the furnace in which the heat is produced ; and lastly, in wind furnaces, the chimney by which the heated air and gaseous products of combus- tion are carried off. CALCINING FUKNACE. Calcining furnaces are small and shallow, because a high temperature is not required. They 56 CALCINING FURNACES. may be made square or circular ; the former are most readily constructed, and, where many crucibles are to be heated at once, they are preferable to the circular ; but the latter give the greatest degree of heat with the least pos- sible consumption of fuel, and are to be preferred on that account where one crucible only is to be ignited. The body of the furnace is best made with good bricks, lined with Welsh lump, fire-bricks, or a mixture of Stour- bridge clay and sand. It is also desirable that a plate of iron with a ledge be placed over the upper part of the furnace to protect the brickwork from blows with crucible tongs, &c., and to keep it in its place when disturbed by sudden alterations of temperature. The bars of the fur- nace may be either in one single piece, or made up of several bars of iron fastened to a frame. They ought to be as far as practicable from each other, and must not be too large, although large enough not to bend under the weight of the fuel and crucibles when they become hot, and they must not be so far removed from each other as to allow the coke or charcoal to fall through easily. Lastly, the more readily the air can find access to the centre of the fuel, the higher will be the temperature pro- duced in the furnace ; very simple assays occasionally fail, only because the bars are either too large or too close together. CHIMNEY. Calcining furnaces generally have no fixed chimney, but are covered with a movable one when a greater degree of heat is required. This chimney may be about five feet high, the diameter of the furnace at the bottom, and tapering off to about two thirds of that dia- meter at the top. It is made of strong plate iron, fur- nished with a wooden handle. The lower part is provided with a door, by means of which the interior of the furnace may be examined without disturbing the whole arrange- ment of the chimney, and consequent cooling of the con- tents of the furnace. If, during the course of any experiment, noxious or offensive vapours are expected to be given off, the furnace must be so arranged that they may be introduced into a WIND FUKflACE. 57 flue, by fastening a piece of iron plate pipe, furnished with an elbow joint, on to the movable chimney before spoken of. EVAPORATING FURNACES. The furnaces just described answer exceedingly well in the absence of gas, for heating small flasks, evaporating basins, &c., when surmounted by a tripod stand or sand bath. This is necessary, as many assays by the dry way are preceded and followed by cer- tain operations in the wet way. THE HOOD. In order to prevent certain gases or va- pours from fires, evaporating basins, &c., from entering into the laboratory, a large metal covering, termed a hood, is employed, terminating in a chimney having a good draught. It is best made of sheet or galvanised iron. FUSION FURNACE WIND FURNACE. The wind furnace, properly so called, is a furnace provided with a chimney, and capable of producing a very high temperature. Wind furnaces are generally square, but, if more than four crucibles are to be heated at one time, they may be made rectangular, the chimney being placed at one of the long sides. When the furnace is required to hold but one pot, it may, however, be made circular. The body of the furnace ought to be made of good bricks, solidly cemented with clay, and bound by strong iron bands. The bricks must be very refractory, and capable of sustaining changes of temperature without cracking. They are ordinarily made with the clay used in the manufacture of crucibles. In some cases bricks are not used for the lining of this kind of furnace ; for in- stance, a mould of wood is placed in the centre, and the open space between the surface of that and the outer brick- work is filled with a paste of very refractory clay, each layer being well beaten down. When the space is filled the case is withdrawn, and the crust of clay dried with much precaution, every crack that may be caused by unequal desiccation being filled up as fast as formed. This method of manufacture is very applicable to circular furnaces. In every case, however, it is necessary to border the edge with a band of iron to prevent injuries from tongs, or pots. 58 WIND FURNACE. By using a mixture of 1 part of refractory clay and 3 to 4 parts of sifted quartz sand, no cracks are formed during desiccation. This mixture is used on the Continent for the interior fittings of Sefstrom's blast furnace, as well as for larger blast furnaces for manufacturing purposes. It is said to stand a high temperature exceedingly well. Makins * recommends for small furnaces the second kind of bricks, known as Windsor, or in the trade P.P. bricks. ' These are of a red colour, very siliceous, but soft, easily cut and shaped, and yet standing heat very well The best method of cutting them is by a piece of zinc roughly notched out as a saw, and then the more accurate figure required may readily be given them by grinding upon a rough flat stone. In this way the small circular furnace formerly made by Newman, and sold by him as his " universal furnace," is lined by cutting the bricks with care to the radii of the circle they are to form, when they key in, like an arch, and so need no lining whatever.' THE ASH-PIT is an open space under the bars, which serves as a receptacle for ashes, clinkers, &c., produced during the time the furnace is in use. It should have the same area as the furnace, and be completely open in front, so that the air may have free access ; it is well, however, for the sake of economy, to furnish this opening with a hinged door, having a register plate fixed in it, so that the draught may be reduced, or entirely shut off, in order that the fire may be extinguished when desirable, and fuel saved which would otherwise be burnt in waste. On the one hand, it is well to have the power of cut- ting off access of air into the body of the furnace by the lower part, either to put out the fire entirely, or to dead en it whilst putting in a pot ; and, on the other, to attain the maximum of temperature, we must have the means of allowing the air to pass with the greatest possible facility into the furnace. In order to do this it is necessary to furnish the ash-pit with doors, or valves, whereby the quantity admitted may be regulated as desired. It is ad- vantageous to lead the air to the ash-pit from a deep and * Makiris's Metallurgy, p. 88. WIND FURNACE. 50 cold place, by means of a wide pipe. A chimney of less height will then be required. THE BARS are made in one piece, or are made up of movable pieces of metal; the latter arrangement is the most convenient. Wherever a wind furnace is in use, the upper opening is closed by a cover made of a fire-tile r encircled with iron. THE CHIMNEY is a very essential part of a wind furnace : it is on its height and size that the draught depends, and, in consequence, the degree of heat produced within the furnace. In general, the higher and larger the chimney r FIG. 21. the stronger is the draught ; so that, by giving it a great elevation, exceedingly high temperatures may be obtained. But there is a limit which it is useless to pass in a furnace destined for operations by the dry way ; and, besides this, the building a very high chimney presents many difficulties and much expense, so that in laboratory operations, where a very strong current of air is required, recourse is had to a pair of double bellows. A temperature can be produced in a wind furnace sufficiently strong to soften the most 60 WIND FURNACE. refractory crucibles, by means of a chimney from six to forty feet high. Chimneys are generally made square or rectangular, and have interiorly the same dimensions as the body of the furnace. About two feet above the upper part of the furnace they are furnished with a register or damper, by means of which the current, of air may be regulated or entirely stopped at will. The damper is a plate of iron sliding into a small opening across the chimney. A wind furnace of the kind above described is repre- sented by fig. 21. The left-hand figure in 21 is the plan, the middle the elevation, and the right is a sectional view. A the body of the furnace in which the crucibles to be heated are placed, G the bars, and P the ash-pit ; the cover is formed of a thick fire-tile of the requisite size, firmly encircled by a, stout iron band, and furnished with a handle for con- venience in moving it ; B the flue, C the chimney, E the damper ; H a hood over the furnace, supported by iron bands h h h ; M the handle of a ventilator T 7 , which serves to carry off hot air and fumes from furnace when open ; and finally, /S, a small sand bath, in which to set the red- hot crucibles when taken from the fire ; one foot square inside the fireplace of the furnace is a good and convenient size ; the remainder will then be in proportion. BLAST FURNACES. In this species of furnace the air necessary to keep up the combustion is forced through the fuel by means of a blowing apparatus, instead of being- introduced by the draught of a chimney as in the wind furnace. The most convenient apparatus for forcing air into a furnace is a double bellows ; a fan may be used, but it is not so powerful. The quantity of air passing into a furnace varies with the length of the assay, and ought to increase gradually as the temperature becomes higher. The following is the description of a most excellent blast furnace which has been in use for some years in the laboratory of the Royal Institution : The temperature BLAST FURNACE. 61 produced by it is extraordinary, considering the small amount of time and fuel employed. It is sufficiently powerful to melt pure iron in a crucible in ten or fifteen minutes, the fire having been previously lighted. It will effect the fusion of rhodium, and even pieces of pure platinum have sunk together into one button in a crucible subjected to its heat.* All kinds of crucibles, including the Cornish and Hessian, soften, fuse, and become frothy in it ; and it is the want of vessels which has hitherto put a limit to its application. The exterior (fig. 22) consists of a black-lead pot, eighteen inches in height, and thir- teen inches in external diameter at the top ; a small blue- pot of seven and a half inches external diameter at the top has the lower part cut off so as to leave an aperture of five inches. This, when put into the larger part, rests upon its lower external edge, the tops of the two being level. The interval between them, which gradually in- creases from the lower to the upper part, is filled with pulverised glass-blowers' pots, to which p IG< 2 2. enough water has been added to moisten the powder, which is pressed down by sticks so as to make the whole a com- pact mass. A round grate is then dropped into the furnace, of such a size that it rests about an inch above the lower edge of the inner pot ; the space beneath it, therefore, constitutes the air- chamber, and the part above, the body of the furnace. The former is 7-J inches from the grate to the bottom, and the latter 7-^ inches from the grate to the top. Finally, a longitudinal hole, conical in form, and 1^ inch in diameter in the exterior, is cut through the outer pot, forming an opening in the air-chamber at the lower part, its use being to receive the nozzle of the bellows by which the draught is thrown in. Sefstrom's blast furnace, obtainable at most chemical- instrument makers, is also very powerful and convenient ; it consists of a double furnace. It is made of stout sheet- * Faraday. 02 CUPEL FURNACE. iron, lined with fire-clay, and is used with coke, or char- coal and coke, broken into pieces of about a cubic inch in size. The blast of air is supplied by a powerful blowing- machine. It will readily produce a white heat. Indeed the limit to its power seems to be the difficulty of finding crucibles or interior furnace fittings which will stand the temperatures produced in it without softening. Kersten states that he increases the heat in Sefstrom's blast furnace by using a hot blast. H. Ste.-Claire Deville has employed for melting platinum a furnace 12 inches high, and 11 inches wide, which rests upon a cast-iron plate full of holes. This is connected with a forge bellows. After blowing for a few minutes, the temperature of the furnace will have reached the highest possible degree, but this zone of maximum heat only extends to a small height above the bottom of the furnace. Above this point a considerable quantity of car- bonic oxide gas is formed, which burns with a very long flame. The heat produced in this furnace is so high that the best crucibles melt, and only crucibles made of good and well-burned lime can be used. THE MUFFLE OR CUPEL FURNACE is a furnace in the centre of which is placed a small semi-cylindrical oven, which is termed the muffle. These furnaces were in use as early as the thirteenth century. Their construction and dimen- sions depend 1. On the temperature which the furnace is intended to produce ; 2. On the number of cupellations required to be per- formed at one time ; and 3. On the kind of fuel to be used. The muffles are mostly made of refractory clay, gener- ally of one piece, but it is best to make them of two pieces ; the bottom being one, and the cover or vault the other. Muffles are sometimes made of cast-iron, cast in one piece. They are useful in small furnaces intended for cupellations only. Muffle furnaces must always be provided with a good chimney to carry off the hot gases. CUPEL FURNACE. The muffle, being completely surrounded by ignited fuel, acquires a very high temperature, and in its interior all operations requiring the presence of air, and which cannot be carried on in contact with carbonaceous matters, may be performed such as roastings, scorifications, and cupellations. When from ten to twenty cupellations ha.ve to be effected at one time, large brick furnaces are employed ; and, in consequence, much fuel is consumed to waste in these when only a few cupellations are required. This has occasioned many persons to endeavour to form small furnaces, where one or two cupellations may be carried on with the smallest possible quantity of fuel. MM. Aufrye and d'Arcet have contrived a furnace which is capable of fulfilling all these conditions. The furnace is elliptical, about 7 inches wide and 18 high; its ash-pit has but one circular opening, and its height is such that, when the furnace is placed upon it, and .the whole upon a table, the assayer can, when seated, readily observe the course of the assay within the muffle. The hearth has five openings, in one of which the muffle is placed ; in another a brick to support it ; a third is for the purpose of introducing a poker to stir the ashes, and make them fall through the grate-holes : this can be closed with a small earthen plug ; and lastly, there are two round holes, placed in its largest diameter, to facilitate the in- troduction of air, either by draught or a pair of bellows, as the case may require. The support for the fuel is generally a plate of earthenware, pierced with holes, and bound round with iron wire to keep it together in case it cracks by changes of temperature ; but it is better to use an iron grating. The dome of the furnace has a circular opening, which can be closed by a plug of earthenware ; this opening serves for the introduction of the fuel. A chimney is necessary to increase the draught; it is made of sheet-iron, and may be from 1^ to 2 feet in height, and ought to fit the upper part of the dome very exactly. At its base there is a small gallery, also of sheet-iron, in which it is 64 CUPEL FURNACE. intended to place the new cupels, so that they may be strongly heated before introduction to the muffle. This saves many of them from fracture. MM. Aufrye and d'Arcet have estimated the quantity of charcoal necessary to heat this furnace. The following are comparative experiments : Silver Lead Time Standard Charcoal employed, employed, of assay, used, No. grains grains minutes thousandths grains 1 1 4 12 947 173 2 1 4 11 950 86 1 4 13 949 93 1 4 10 949 60 Coke or charcoal may be used in this furnace, but the fire must be lighted by means of charcoal alone, as coke is very difficult to inflame in a cold furnace. When it is red- hot it may be fed with coke, or, better still, a mixture of coke and charcoal. Where great numbers of cupellations have to be made at once, the following form of brick furnace is requisite : Fig. 23 shows an elevation of the furnace ; fig. 24 shows a section. The interior of the furnace is of fire- brick ; the exterior, of common brick. The upper part is protected by a plate of iron, and the upper opening,, through which the fuel is introduced, is covered, when necessary, by a large fire-tile strongly encircled with an iron band, to which are attached two handles, by which the whole can be moved. The muffle opening, as seen partially open in the dia- gram, can be entirely closed by means of two sliding doors y made of sheet iron, running in a stout wrought-iron frame, built into the brickwork. Immediately below the muffle entrance are two movable bricks ; these close the openings through which the fire-bars are introduced ; and still lower down is the ash-pit door, furnished with a register for the better regulation of the current of air required by the furnace. In fig. 24 is shown a brick built into the back of the furnace, on which the close end of the muffle is supported. This brick may, however, be replaced by a crucible or fire-brick standing on the bars of the furnace: A very useful furnace for small operations is one which UNIVERSAL FURNACE. has been aptly termed the ; universal furnace,' as it is capable of performing all that is required of any furnace FIG. 23. FIG. 24. in an assay (see figs. 25 and 26,^elevation and section). Tt is much to be recommended for its durability, cheapness, FIG. 25. FIG. 26. and its small size compared with the heat it can produce. It is constructed externally of sheet iron, very stout, and F 66 FURNACE OPERATIONS. is lined with fire-brick, not cemented together, but ground and keyed as an arch, so that it can never fall out till it is completely useless. Its height is about 2J feet and diameter 1 foot ; internal diameter 8 inches and depth of fireplace 1J feet. It is furnished with five doors, one in the ash-pit and four in the body of the furnace, two in the front, one above the other, and two opposite each other, at the sides. The cover serves as a sand-bath, and when that is taken off there is a series of cast-iron rings, fitting the top of the furnace, where basins can be placed either for the purpose of evaporation, calcination, or roasting. The two opposite holes serve for the introduction of a tube in operations where it is requisite to pass a gas over any body at a red heat. In the lower hole in front can be placed a muffle for roastings and cupellations, introducing fuel and crucibles by the upper one ; it also serves as an opening through which the state of the furnace can be seen, or the progress of any assay observed. Iron, manganese, nickel, and cobalt can be fused in this furnace when it has a flue of about thirty feet in height attached to it, and by closing the ash-pit door the dullest red heat, for gentle roastings, can be obtained. FURNACE OPERATIONS. Crucibles must be carefully supported in the fire, and must always be covered. They must stand solidly, and be at equal distances from the sides and bottom of the furnace, so as to receive a like share of heat, and they must be completely surrounded with fuel. If a crucible is supported on the grate-bars of a furnace, the draught of cold air will prevent the bottom from getting hot. If it is supported on the fuel, the bottom gets heated quickly, but the fuel on burning away allows the crucible to fall down, and may cause the loss of the contents. For these reasons a crucible should always be supported on a piece of fire-brick about three or four inches high. In many cases an old crucible inverted will serve as a convenient support. The fire must be got up gradually, so as to FURNACE APPARATUS. 67 prevent the sides of the furnace and the crucibles within from cracking from the sudden increase of heat. No time is saved by urging the fire strongly at first, for crucibles are bad conductors of heat, and a high temperature at the commencement scarcely causes the heat to penetrate to the interior faster than a moderate redness. After the furnace has arrived at a full red heat, more air may be given, and in from about twenty minutes to one hour the assay will be finished. During the time that the fur- nace is in full action, the cover must be occasionally re- moved to add more fuel, if any open spaces occur round the crucibles, also to press the fuel close to the pots. When the pots are taken out they may be placed on the anvil or in a sand-bath, and allowed to cool gradually before they are broken to examine their contents. In commencing a second assay immediately in the same furnace, certain precautions must be taken to insure success. In the first place, all ash and clinker must be removed from the grate by means of a crooked poker ; secondly, the fuel must be pressed down firmly ; and lastly, a layer of fresh combustible must be placed on the fire, and before that is ignited the crucibles must be ' O arranged upon the support and the spaces about them be filled with coke or charcoal, as the case may be, and the assay be proceeded with as before. In executing many assays, one after the other, a great saving of fuel is effected, for the furnace is not allowed time to cool. AUXILIARY APPARATUS. Ordinary assay furnaces require very few instruments ; they are, firstly, pokers or stirring rods, made of stout bar-iron ; these may be straight, as for stirring the fuel from the top of the furnace, so as to fill up cavities formed by uneven combustion; or curved, for clearing the bars from below from clinker and ashes. Straight and curved tongs are also required ; for small crucibles the blacksmith's common forge tongs are the most suitable ; tongs with semicircular ends (see fig. 27) are very serviceable for larger crucibles. The tongs a are particularly adapted for removing large cupels or calcining F 2 68 FURNACE TOXGS. dishes from the muffle ; the tongs b and c are used for lifting heavy crucibles from the wind or blast furnace. In case the eyes of the operator are weak, it is advisable to make use of a pair of deep neutral-tint spectacles. Most of the radiant heat from the interior of a furnace may be FIG. 27. cut off by holding before the face a large sheet of window glass'; or the operator may look at the reflected image in a looking-glass, instead of looking direct into the furnace itself. Some assayers recommend the use of masks for the face and gloves for the hands, but these are not needed. In cupel furnaces, both curved and straight pokers or stirring rods are required ; also a curved rod made of lighter iron, to be used in closing the sliding doors, slightly moving cupels, &c. The tongs used vary in form (see fig. 28). a represents very light elastic tongs or pincers FIG. 28. employed in the introduction of lead and other matters to the cupel ; b shows the tongs for holding the scorifier ; the curved part fits the lower part of the scorifier, and the upper or single part passes over the upper part of the scorifier, so that its contents may be emptied into the proper mould without fear of its slipping from the opera- INGOT MOULDS. tor's grasp ; c represents the tongs used in moving cupels ; they are slightly curved, so that the cupels from the back part of the muffle may be removed without disturbing those FIG. 29. in front. Pig. 29 shows the plan and section of the ingot mould, into which the contents of the scorifiers are poured ; it is made of thin sheet iron, and the depressions for the reception of the fused lead slag^ and ore are hammered out. FIG. 30. Fig. 30 is a wrought-iron ladle, in which lead clippings, &c., are melted, in order to obtain a fair average of a large Fio 31 , quantity ; and fig. 31 represents the ingot mould into which the fused lead or other metal is poured. Other special apparatus will be described under the assay in which they are required. Furnaces are heated with anthracite, coke, and charcoal, and sometimes with a 70 FUEL FOR FURNACE. mixture of the two latter ; coal is very seldom employed,, and therefore will not be much spoken of; coke is the principal combustible used in assaying. Calcining fur- naces ought to be heated with charcoal alone, because coke employed in small quantities ignites and burns with too much difficulty. All fuel, contains certain fixed matters which remain after combustion, and which constitute the ash. This ash fuses or agglutinates together, and when a certain quantity is formed, if it be not removed, the fire will decrease in intensity, and finally die out. As all com- bustibles do not contain the same amount of ash, they should be carefully selected ; those containing the least are to be preferred, in the first place because, weight for weight, they contain more available fuel, and, secondly,, because they can be used in a furnace a longer time with- out the formation of so much clinker. The composition of the ash likewise merits much attention. Charcoal contains, in general, from 3 to 4 per cent, of ash, the chief components of which are lime and potash carbonates. Certain other matters are also present, as phosphoric acid, iron oxide, manganese, &c., but these are in very minute proportions. The ash is not fusible per se, and, if it does not meet with any substance capable of combining with it, it passes through the bars as a white powder ; but when the potash predominates, it exercises a corrosive action on the bricks with which the furnace is lined, as also on crucibles, lutes, &c., by the formation of a fusible potassium silicate, which in course of time runs down the sides of the furnace, and chokes the bars. Whenever the ash is in very small proportion to the charcoal, its fusion is rather useful than otherwise, be- cause it forms a species of varnish, which, penetrating the surface of the bricks and lutes, gives them solidity by binding them together with a cement, forming part of their substance. The proportion of ash which coke contains is very variable; some commercial samples contain from 8 to 10 per cent., while others, made from very pure coal, give but 2 to 3 per cent. ; so that this fuel also ought to be CHARCOAL COKE. 71 carefully chosen. The nature of this ash is different from that of charcoal ; it consists principally of iron oxide and clay. The former is produced from the pyrites which coal generally contains. The clay is similar to the car- bonaceous schists, not very fusible by itself, but neverthe- less capable of softening, When pure, it forms a slag, which attacks neither the bricks nor crucibles. This happens very rarely ; it is more often that iron oxide predominates, and this by contact with the carbonaceous matter becomes reduced to the state of protoxide, and is then not only very fusible, but exercises on all argillaceous matters a very corrosive action, so that crucibles are very seriously injured, and the sides of the furnace require frequent repairs. Weight for weight, coke and charcoal give out nearly the same quantity of heat ; but in equal bulks, the former develops much more heat, because its density is greater. From this difference in the calorific power of coke and charcoal, it results that in the same furnace the former produces a greater degree of heat than the latter ; and at high temperatures the difference has been proved to be nearly 10 per cent. In order to account for this, we must consider, firstly, that in a given space the quantity of heat produced in a certain time (and, in consequence, the tem- perature) depends on the amount of fuel burnt, and in- creases with its weight ; secondly, that combustion takes place but at the surface of the masses, whatever may be the nature of the fuel ; from which may be deduced that the weight of fuel burnt in an unit of time ought to be exactly proportionate to its density ; and in consequence, the densest fuels, furnishing the most food for combustion, ought to give out the greatest heat. But, as for the same reason they consume a larger proportion of oxygen, they require, in order to produce the maximum effect, a more rapid and stronger current of air. It is clear, from what has been stated concerning the relative properties of coke and charcoal, that when the former can be procured of good quality, and especially when the ash contains but little oxide of iron, it ought to 72 EFFECTS OF FURNACES. be preferred to charcoal, for assays requiring a high tem- perature. This being an important subject, it has been thought advisable to devote a special chapter to the assay of fuel. A very essential condition in obtaining the maximum effect of a furnace, the importance of which can alone be appreciated by experience, is to choose pieces of fuel of a suitable size. If, on the one hand, a shovelful of coke or charcoal be taken at random, it generally contains the dust and dirt found in most fuel, and which, by filling the interstices, prevent the air from passing as required, and consequently render the combustion slow. On the other hand, if a furnace be filled with large pieces, considerable spaces are left between them, so that but a comparatively small surface is exposed to the action of the atmospheric oxygen, and a correspondingly small quantity of fuel is consumed in a given time ; so that the maximum heat can never be obtained. In order to produce the desired result, it is necessary that the pieces shall have a certain mean size, and experience has proved that pieces about 1 inch to H inch diameter produce the best effect. These may be selected by sifting the coke through two strong wire sieves, one of which has meshes about 1 J inch square, and the other about 1 inch square. The coke which passes through the larger one, but will not go through the smaller sieve, will be the right size for the furnace. THE EFFECTS PRODUCED BY WIND AND BLAST FURNACES. Assays by the dry way can be made either in wind or blast furnaces. In either of them, the degree of heat depends upon the volume of air which passes through the fuel in the same time ; but, cceteris paribus, large furnaces produce more heat than small ones, because comparatively less heat is lost by radiation in the former than the latter. In a wind furnace, the maximum of heat is limited by the size of the chimney, and in a blast furnace by the dimensions of the bellows ; but by weighting the latter, FOCUS OF MAXIMUM TEMPERATURE. 73 more or less, the force of the blast and, in consequence, the temperature can be increased, to a considerable extent. In this respect blast have the advantage over wind furnaces. In the latter, the draught increases in proportion as the heat becomes more intense in the furnace, so that the temperature producible increases progressively. In a blast furnace, the bellows can be weighted and worked as heavily as possible at once, and, by opening all the apertures for receiving air, the maximum temperature can be produced more rapidly than in a wind furnace ; but this is of little use, because, as heat passes very slowly through the substance of a crucible, when the object is to fuse its contents it must be heated gradually, so as to avoid running the risk of softening the crucible before its contents are acted upon, or even scarcely made warm. Wind furnaces are, however, much more serviceable and economical than blast, because they work themselves, and do not require the service of a man to attend to the bellows. A blast furnace is used in a laboratory, in cer- tain cases ; for instance, when a single crucible has to be submitted to an intense heat, and when the furnace is small, and the bellows large, in which case the operation resembles a blow-pipe assay. In whatever manner the air is introduced into any kind of furnace, whether wind or blast, it is evident that the quantity of heat developed in equal-sized furnaces depends upon the quantity of air introduced in the same time ; but the degree of temperature is not the same in different parts of the furnace, and the distribution of heat varies according to the manner in which the air is intro- duced into the midst of the fuel. The side over which the air passes is kept cold by the current, on which account fire-bars last a long time without becoming oxi- dised, but the heat rapidly augments up to a certain distance from the bars, at which place it arrives at its maximum ; above that it diminishes rapidly, because the air is nearly deprived of its oxygen. Experiment has proved that this maximum is about 2J to 3 inches above the bars or tuyeres. 74 OIL AND GAS FURNACES. In common wind furnaces the air enters through the spaces between the horizontal bars which form the bottom of the furnace, and the crucibles are placed on a stand which rests on these bars. By this means the lower and centre part of the crucibles, in which parts the matter to be fused is placed, are exactly situated in the maximum of heat ; but the stand being constantly kept cold, by the contact of a current of air, establishes a continual draining or carrying away of heat from the interior of the crucibles outwards, so that the substance submitted to assay can only arrive at the maximum temperature after a length of time, and the maximum then is always inferior to that in the mass of fuel. It is on this account that assays in a blast or wind furnace generally occupy from one hour to two hours. The author has found that the time may be reduced to half that just stated, if a good solid foundation of fuel be made, and the crucible placed on that, and well surrounded by coke, constantly kept close to the pot and the sides of the furnace ; in this manner the cooling effect of the stand is removed, and the consequent maximum effect of the furnace is produced, but then there is danger of the supporting fuel being burnt away from the crucible and the latter getting upset. OIL AND GAS BLAST FURNACES. It sometimes happens that metallurgists and assay era have occasion to melt metals at a white heat, but do not wish to heat a large furnace for the purpose. In these cases either the gas or oil furnaces now to be described will prove very useful. OIL FURNACES. Mr. Charles Griffin, the well-known chemical instrument maker, has described an oil lamp, which is not only as powerful in action as the best gas furnaces, but almost rivals them in handiness and economy. DESCRIPTION OF THE APPARATUS. The oil-lamp furnace is represented in perspective by fig. 32, and in section by fis. 33. It consists of a wick-holder, an oil-reservoir, and GRIFFINS OIL BLAST FURNACE. 75 a fire-clay furnace ; to these must be added a blowing- machine for the supply of atmospheric air. The oil-reservoir is represented at a, fig. 32 ; it is made of japanned tin-plate, mounted on iron legs, and fitted with a brass stop-cock and delivery-tube. Its capacity is a little more than a quart. The wick-holder is represented at b, and the upper surface of it by the separate figure old, into a furnace full of lighted coal : take it out when reddish white, and expose it to a current of cold air pro- duced by a bellows or otherwise : if it stand these trials, it may be heated afresh and plunged red-hot into water, and, if it be not broken, placed immediately in the fire. The best pots support all these operations without break- ing ; but it often happens that they are filled with innu- merable small fissures, through which fused matters can pass. This can be ascertained by fusing rapidly in the assay pot a quantity of litharge : if these fissures be pre- sent, the fused oxide will readily filter through them. CHARCOAL CRUCIBLES. As all oxidised matters act readily on clay pots, and a great number of the metals and their compounds adhere to them, they have long since been replaced, under certain circumstances, by charcoal cru- cibles, which do not possess these disadvantages. The older assayers used merely a piece of charcoal, with a hole made in it, and then bound round with iron or other wire. The use of these has, however, been abandoned for some time, and earthenware crucibles lined with charcoal have been substituted (see fig. 61, a, 6, and c). These may be considered as charcoal pots enveloped with refractory clay ; they are solid, always free from cracks, and easy of preparation, and they have the same properties as the solid charcoal crucibles without their inconveniences. In order to prepare these crucibles, the charcoal must be chosen carefully, so as to contain no foreign substances ; it must be pulverised and passed through a sieve ; the powder moistened with water or treacle, mixed with a spatula, and then kneaded with the fingers until it just adheres, and forms adhesive lumps without being sufficiently wet to adhere to the hand. Some advise the addition of gum to the water with which the charcoal is moistened. The LINED CRUCIBLES. 127 crucible is moistened slightly by being plunged into water, and withdrawn as speedily as possible, and about half an inch in depth of the charcoal paste, prepared as above, is placed in it ; the paste is then pressed firmly down, by means of a wooden pestle : the blows are to be slight at first, and then increase in force until it is as firm as possible: another layer is then applied and pressed as before, and the process repeated until the crucible is quite full, taking great care to render all as firm as possible, especially at the sides. In order to make each layer adhere firmly to the other, the surface must be scratched rather deeply with the point of a knife before a new layer is applied. When the crucible is completely filled, a hole is to be scooped in the charcoal of about the form of the pot. The sides are then rendered smooth by friction w r ith a glass rod. This is absolutely necessary, so that the metallic globules produced in an assay may not be retained by the asperities of the lining, but may be readily enabled to unite into one button. When a lined charcoal pot is well made, its sides are very smooth and shining. For ordinary use, the lining may be f ths of an inch thick at the bottom and ^th or so at the sides : but in some cases, for instance, when the substance to be fused is capable of filtering through the lining and attacking the pot as a flux, it must be at least twice the above thickness in every part. Lined crucibles have many advantages over plain 128 LIME CRUCIBLES. crucibles. The lining gives them greater solidity, and prevents a loss of shape when softened ; for plain crucibles are always three-fourths empty when their contents are fused, on account of contraction in volume : the pots then have nothing to sustain their sides when they soften towards the end of the assay, at which period the highest temperature is employed. Besides, vitreous matters do not penetrate the lining, and, exercising no action on it, can be obtained in a state of purity, and the exact weight determined : if fused in a plain pot, the weight could not be ascertained, because a portion would adhere to the sides, and the resulting mass would not be pure, having taken up a portion of the crucible in which the fusion was effected. The lining, too, effects the reduction of certain metallic oxides by cementation, and does away with the necessity of adding powdered charcoal to the body to be reduced. This property is very valuable, because, when an oxide is reduced by mixing it with charcoal, an excess must always be employed, and this excess remains with the metal, and prevents its exact weight from being ascertained. No oxidising substances or bodies which readily part with oxygen (oxide of copper, for examp]e) must be calcined in a plumbago or charcoal-lined crucible, unless indeed the chemical union of the charcoal with the oxygen is desired. LIME CRUCIBLES. Some years ago Deville proposed the use of crucibles cut out of solid blocks of pure lime, in order to prevent the introduction of carbon and silicon into metals and alloys during the process of fusion. The results of experiments made with such crucibles were found to be extremely satisfactory, and metals fused therein, as iron, manganese, nickel, cobalt, &c., were obtained far purer and more malleable and ductile than when fused in the usual clay or brasqued crucibles. Un- less, however, the crucibles required were of very small size, it was found difficult to obtain blocks of lime for shaping them sufficiently large and free from flaws ; and experiment showed a considerable loss, both by breakage when shaping them, and by their cracking when in the \v TY }} IJME CRUCIBLES. 129 furnace. In order to obviate this, trials were made with clay crucibles lined with lime, but ineffectually, as these crucibles invariably melted down before the requisite heat was arrived at a result due to the action of the lime itself upon the outer clay crucible. Mr. David Forbes, F.R.S.,has published in the c Chemi- cal News ' the result of some very valuable experiments on this subject. The arrangement he proposes fully answers the purpose, the crucibles being capable of stand- ing the heat of melted wrought iron or cobalt without fusion or cracking, as well as of being made of any reason- able size. A clay crucible of somewhat larger capacity than the desired lime one is filled with common lamp-black, com- pressing the same by stamping it well down. The centre is then cut out with a knife until a shell or lining of lamp- black is left firmly adherent to the sides of the crucible, and about half an inch or less in thickness, according to the size of the crucibles ; this lining is now well rubbed down with a thick glass rod until its surface takes a fine glaze or polish, and the whole cavity is then filled up with finely powdered caustic lime, thoroughly pressed down, and a central cavity cut out as before ; or the lime powder may be at once rammed down round a central core of the dimensions of the intended lime crucible. This lime lining is naturally rather soft before being placed in the furnace, but upon heating, it agglutinates, and forms a strong and compact crucible, which is prevented from acting upon the outer one by the interposed thin layer of lamp-black, and at the end of the experiment it generally turns out as solid and compact as those made in the lathe. Experiments made with such crucibles, even up to dimensions containing several pounds of metal, have proved them extremely well suited for these operations, and doubt- less similar crucibles could be made, lined with magnesia or alumina as required. In some cases ordinary black- lead crucibles, lined with powdered lime, magnesia, or alumina, might possibly be found to answer. Having frequently used lime crucibles in metallurgical K 130 ALUMINA CRUCIBLES. operations, and having met with the inconvenience pointed out by Mr. Forbes, the editor can appreciate the great value of his improvement. It is one which cannot fail to be extensively adopted in metallurgical laboratories. In certain particular experiments, crucibles are lined with other bodies besides charcoal and lime, such as silica, alumina, magnesia, or chalk, by merely moistening their respective powders with water, and applying the paste as above described for the charcoal. A slight layer of chalk lessens the liability of attack from fused litharge. ALUMINA CRUCIBLES are strongly to be recommended in many metallurgical operations. They are made in the following manner. Ammonia alum is ignited at a full white heat, when it leaves behind pure alumina in a dense compact form : this is to be finely powdered. To a solution of another portion of ammonia alum in water, ammonia is added, when alumina is precipitated in a gela- tinous state : this is to be washed until free from sulphate of ammonia. The dense alumina is then mixed with water and worked up into a paste, the precipitated gelatinous alumina being kneaded in from time to time ;' this gives coherency : and when sufficient has been added (which must be ascertained experimentally), the mass may be moulded into shape. These crucibles require slow and careful drying ; but they well repay all the care which is bestowed on them, for they do not readily crack, are attacked by very few fluxes, give out no impurities to metals which are melted in them, and are infusible at the highest heat of the furnace. MAGNESIA CRUCIBLES AND BRICKS. M. H. Caron has pointed out the advantages which would accrue to metal- lurgy from the employment of magnesia as a refractory material. Formerly the high price of this earth appeared likely to confine it to the laboratory. Now, circumstances have happily changed ; recent modifications introduced into the manufacture of cast steel, and especially the employment of Siemens's furnace and Martin's process, absolutely demand more refractory bricks than those at present in use, irrespective of price. On the other hand, MAGNESIA CRUCIBLES AND BRICKS. 131 native carbonate of magnesia, which formerly cost 10/. the ton, may now be obtained at the price of 21. 15s. delivered at Marseilles or Dunkirk. Calcination at the place where the carbonate is obtained may still further reduce its price. 1 The following is M. Caron's process for its agglomeration, which may be employed by the chemist for the ready preparation of refractory vessels of all forms ; by the physicist to obtain pencils for oxy- hydrogen lighting purposes ; and also by the manufac. turer, to replace, in some cases, fire-bricks which have become unsuitable for carrying out different processes of heating. The magnesia which he employs comes from the island of Elba, where it is found in considerable quantities as a native carbonate, white, very compact, and of great hard- ness. This carbonate contains traces of lime, silica, and iron ; it is, besides, interspersed sometimes with serpentine and large plates of silica, which would diminish the in- fusibility of the substance, and render it especially unfit for oxyhydrogen illumination if their removal is neglected. These plates are, however, easily recognised, and may be readily separated, even in working on the large scale. In the case of refractory bricks, the presence of a small quantity of these foreign bodies would, at the most, give rise, under the influence of the highest temperatures, to a slight vitrification, offering no serious inconvenience. Before powdering the carbonate, it is advisable to bake it at the temperature necessary for the expulsion of the carbonic acid ; the material then becomes very friable, and may be pulverised more easily. It is then possible to separate the serpentine and silica, which do not become friable under the influence of heat. This preliminary treatment does not permit of the agglomeration of the magnesia, and, even were this difficulty to be overcome, a temperature higher than that of the original calcination would cause an enormous contraction, producing fissures and alterations in shape which would interfere with the 1 This preliminary calcination requires less heat than burning lime, and diminishes the weight of the carbonate one-half. K 2 132 MAGNESIA CRUCIBLES AND BRICKS. use of this substance. It is therefore indispensable, before moulding the magnesia, to submit it to a very intense heat, at least equal to that which it is intended to support sub- sequently. Thus calcined it is not plastic, its appearance is sandy, and compression does not cause it to acquire any cohesion ; a mixture of magnesia, less calcined, imparts to it this quality. The quantity of the latter to be added necessarily varies with the degree of calcination of the two magnesias ; it is scarcely one-sixth of the weight of that which has been exposed to the temperature of melting steel. It only now remains to moisten it with 10 or 15 per cent, of its weight of water, and strongly compress it in iron moulds, as adopted in making artificial fuel. The brick produced in this operation hardens on drying in the air, and becomes still more resisting when it is subsequently calcined at a red heat. The same process would appear practicable, varying the form of the moulds, for obtaining crucibles of great capacity ; but compression is difficult in large masses, as well as when the moulds have a large surface, as the magnesia adheres strongly to the sides. Although M. Caron has been able to obtain small crucibles for the laboratory, he does not consider this process adapted to make the large crucibles employed in steel melting. In this and other cases it is preferable to ag- glomerate the magnesia in the wet way. To endow magnesia with a sort of plasticity, advantage is taken of a property of this earth pointed out in ' Ber- zelius's Chemistry.' When strongly calcined and then moistened, it hardens in drying. This fact is doubtless due to a hydration which takes place without sensible in- crease of temperature. When solidified in this manner the magnesia only loses the assimilated water at a high temperature. Then calcination not only does not disin- tegrate it, but, on the contrary, confers upon it a hardness and resistance comparable to those of ordinary crucibles after their baking. This being understood, the appli- cation of this fact is obvious. Thus, the magnesia to be employed in the manufacture of crucibles should only be PLATINUM CRUCIBLES. 133 moistened, moulded into shape, dried, and then ignited. For the construction of steel-melting furnaces, a paste of moistened magnesia should be plastered over the walls ; it will become ignited in due course without any particular precautions being taken. It sometimes happens, however, either owing to the magnesia being too much or too little hydrated, or owing to its containing siliceous matter, that the vessels before or after firing do not possess quite the desirable solidity ; they should then, to acquire this, be simply moistened in a cold saturated solution of boracic acid, dried, and then fired as before. This operation does not render the magnesia more fusible ; it only causes the grains of the substance to cohere more strongly together. Very pure, strongly calcined, and finely pulverised magnesia may be employed in the form of paste (barbotine), and yields the most delicate and translucent crucibles, as well as the sharpest and most complicated impressions. MALLEABLE IRON CRUCIBLES are often very serviceable in assays of fusibility, and of certain selenides and sulphides, also in assays of galena or ordinary lead ore. They are either made of hammered sheet iron, or by plugging up small iron tubes, such as gun-barrels, &c. The latter are pre- ferable, because thick solid crucibles can be used a number of times, whilst the others are necessarily very thin and can be used only once. Whenever iron crucibles are em- ployed at a very high temperature, they must be placed in those of earthenware, which protect them from the oxidising action of the air ; but when they are not heated above the temperature of a copper assay, they may be used naked, if tolerably thick. For assays at the above temperature, cast-iron crucibles may be employed with advantage, instead of wrought-iron, because they are very nearly as good, and much less expensive. PLATINUM CRUCIBLES. Platinum crucibles are invalu- able in a laboratory. Few pieces of apparatus are used so frequently by the chemist. Their chief use is in the ignition of precipitates and the decomposition of siliceous minerals by fusion with alkaline carbonates. They are 134 PLATINUM CRUCIBLES. preferable to porcelain, as not being fragile and being more readily heated to redness over the gas or spirit flame. Their most convenient size is 1^ inch high and 1J inch wide at the top. In employing a crucible for the incineration of filters in quantitative assays by the wet way, it sometimes happens (as, for instance, with chloride of silver or sulphate of lead) that the employment of platinum is inadmissible. In these cases thin porcelain crucibles must be used. The analyst will, however, frequently experience difficulty, owing to the extreme slowness w^ith which, in many cases, the last portions of the carbon of a filter are consumed when ignited in a porcelain crucible. It does not appear, how- ever, that the following simple method of obviating the difficulty, as practised in the laboratory of Professor Sheerer, in Freiburg, has ever received the publicity which it deserves. Whenever a filter upon which a substance capable of injuring platinum has been collected, has to be incinerated, the porcelain crucible or capsule in which the process is to be conducted should be placed within a vessel of platinum of similar form, and the whole ignited in the usual way. Whether the greatly accelerated rapidity of combustion of the carbon which ensues depends upon a more equal distribution of heat brought about by the greater conducting power of the metal an explanation which is current for the somewhat analogous case of copper- coated glass flasks or whether, as seems probable, the power of the porcelain vessel to absorb heat be really increased by the interposition of the platinum ; whether both these causes be of influence, or the result depends upon another less apparent reason ; or, finally, whether vessels of some other metal would not be preferable to those of platinum, are questions which are open to dis- cussion. Fresenius gives the following excellent directions as to the preservation of platinum crucibles. The analyst should ; acquire the habit of cleaning and polishing the platinum crucible always after using it. This should be done, as recommended by Berzelius, by friction with moist sea-sand, PLATINUM CRUCIBLES; 135 whose grains are all round, and do not scratch. The writer has found this method to answer extremely well. The sand is rubbed on with the finger, and the desired effect is produced in a few minutes. The adoption of this habit is attended with the pleasure of always working with a bright crucible, and the profit of prolonging its existence. This mode of cleaning is all the more necessary when one ignites over gas-lamps, since at this high temperature crucibles soon acquire a grey coating, which arises from a superficial loosening of the platinum. A little scouring with sea-sand readily removes the appearance in question, without causing any notable diminution in the weight of the crucible. The ordinary Bunsen burner is known to act upon the surface of platinum vessels brought into contact with the inner line of the flame ; the metal loses its polish, becom- ing superficially porous and spongy, and requires the use of the burnisher to bring it back to its original state. This alteration of the surface Mr. Dexter has found to be at- tended with a change of weight, so that for some years he has used a lamp of different construction for the heating of platinum crucibles in analytical operations. Such a lamp may be made by removing the air-tube of a common Bunsen lamp, and putting in its place a somewhat longer one of glass or iron of about 12 millimetres internal dia- meter. The gas jet should have a single circular aperture, and be in proper proportion to the diameter of the tube, which may be held in any of the ordinary clamp supports. The tube being raised sufficiently above the jet to allow free entrance of air, and a full stream of gas let on, a ' roar- ing ' flame is produced, of which the interior blue cone is pointed, sharply defined, and extends only about half an inch from the top of the tube. A polished platinum sur- face is not acted upon by this flame, provided it be not brought into contact with the interior cone. In the Bunsen burner, as usually made, the supply of air depends upon the diameter of the tube, the holes at its base being more than sufficient to supply the draught. With the wider tube it is necessary to limit the admission of air by 130 -PLATINUM CRUCIBLES. depressing the tube upon the lamp when the force of the gas is diminished. Otherwise the proportion becomes such that an explosive mixture is formed ; for this reason it is more convenient to use an arrangement in which the access of air can be regulated by an exterior tube sliding obliquely downward over the air-apertures. The gas jet should be on a level with the top of these apertures, which must be much larger than those of the ordinary Bunsen's burner. On account of the liability to explode and burn at the jet inside, the lamp is not well adapted for ordinary use ; but for ignition of crucibles, working of glass, &c., it has proved efficient and practical. In connection with some sensible remarks upon the before-mentioned use of sand in cleaning platinum cru- cibles, Erdmann explains in the following way the cause of this grey coating which forms upon platinum crucibles whenever they are ignited in the flame of Bunsen's gas- burner. This coating has given rise to much annoyance and solicitude among chemists. Indeed, it has often been asserted that the use of Bunsen's burner is unadvisable in quantitative analysis, since by means of it the weight of platinum crucibles is altered and the crucibles themselves injured. The coating is produced most rapidly when the crucible is placed in the inner cone of the flame, and the more readily in proportion as the pressure under which the gas is burned is higher. Having found it advantageous to maintain, by means of a special small gas-holder, a pressure of four or five inches upon the gas used in his own laboratory, Erdmann observed that the strong gas-flame thus afforded immediately occasioned the formation of a dull ring upon the polished metal placed in the inner flame, this ring being especially conspicuous when the crucible becomes red-hot ; it increased continually, so that after long-continued ignition the whole of the bottom of the crucible was found to be grey and with its lustre dimmed. The ring is caused neither by sulphur, as some have believed, nor by a coating of inorganic matter, but is simply a superficial loosening of the texture of the platinum, PRESERVATION OP PLATINUM CRUCIBLES. 137 in consequence of the strong heat, whence it first of all appears in the hottest part of the flame. In consequence of the serious damage which the gas furnace causes, many chemists now discard gas and ignite platinum crucibles over specially constructed spirit lamps. In conjunction with Pettenkofer, Erdmann instituted several experiments, which have left but little doubt that the phenomenon depends upon a molecular alteration of the surface of the metal. If a weighed polished crucible be ignited for a long time over Bunsen's lamp, the position of the crucible being changed from time to time, in order- that the greatest possible portion of its surface shall be covered with the grey coating, and its weight be then de- termined anew, it will be found that this has not increased. The coating cannot be removed either by melting with bisulphate of potash or with carbonate of soda. It dis- appears, however, when the metal is polished with sand ; the loss of weight which the crucible undergoes being exceedingly insignificant, a crucible weighing 25 grammes having lost hardly half a milligramme. When the grey coating of the crucible is examined under the microscope, it may be clearly seen that the metal has acquired a rough, almost warty, surface, which disappears when it is polished with sand. Platinum wires, which are frequently ignited in the gas-flame for example, the triangles which are used to support crucibles become, as it is known, grey and brittle. Under the microscope they exhibit a mul- titude of fine longitudinal cracks, which, as the original superficial alteration penetrates deeper, become more open, or, as it were, spongy, until finally the wire breaks. If such wire is strongly and perseveringly rubbed with sand, the cracks disappear, ana the wire becomes smooth and polished ; for the grains of sand, acting like burnishers, restore the original tenacity of the metal, very little of its substance being rubbed off meanwhile. The loosening effect of a strong heat upon metals is beautifully exhibited when silver is ignited in the gas-flame, a thick polished sheet of silver immediately becoming dull white when thus heated. Under the microscope the metal appears swollen 138 CLEANING PLATINUM CRUCIBLES. and warty. Where it has been exposed to the action of the inner flame along its circumference, this warty con- dition is visible to the naked eye. A stroke with the burnishing stone, however, presses down the loosened par- ticles, and reproduces the original polish. This peculiar condition which the surface of silver assumes when it is ignited, is well known to silversmiths ; it cannot be re- placed by any etching with acids, and it must be remem- bered that what is dull white in silver appears grey in platinum. If each commencement of this loosening is again de- stroyed, the crucibles will be preserved unaltered, other- wise they must gradually become brittle. Crucibles of the alloy of platinum and iridium are altered, like those of platinum, when they are ignited. It is, however, some- what more difficult to reproduce the original polish of the metal by means of sand, as might be expected, from the greater hardness of the alloy. The sand used should be well worn. When examined under the microscope no grain of it should exhibit sharp edges or corners ; all the angles should be obtuse. If there are spots on the platinum crucibles which cannot be removed by the sand without wearing away too much of the metal, a little potassium bisulphate is fused in the crucible, the fluid mass shaken about inside, allowed to cool, and the crucible finally boiled with water. There are two ways of cleaning crucibles soiled outside ; either the crucible is placed in a larger one, and the interspace filled with potassium bisulphate, which is then heated to fusion, or the crucible is placed on a platinum-wire triangle heated to redness, and then sprinkled over with powdered potassium bisulphate. Instead of the bisulphate, borax may be used. Never forget at last to polish the crucible with sea-sand again. A remarkably rapid and perfect method of cleaning platinum apparatus consists in gently rubbing upon the dirty metal a small lump of sodium-amalgam. Sodium has the curious property of lending to mercury the power of ' wetting ' platinum in so complete a manner that the CLEANING PLATINUM CRUCIBLES. 139 positive capillarity between platinum and an amalgam con- taining even only one per cent, of sodium appears to be as great as that between mercury and zinc, with this important difference, however in the former case, the 4 wetted ' metal does not suffer the least trace of amalga- mation. Even when foreign metals, such as lead, tin, zinc r silver, are purposely added to the sodium -amalgam, the platinum surface suffers no disintegration. When the amalgam has been rubbed on with a cloth until the whole surface is brilliantly metallic, water is ap- plied which oxidises the sodium and allows the cohesion of the mercury to assert itself. On wiping the mercury off, the platinum surface is left in admirable condition for the burnisher. When the crucible is clean it is placed upon a clear platinum-wire triangle, ignited, allowed to cool in the desiccator, and weighed. This operation, though not in--, dispensable, is still always advisable, that the weighing of; the empty and the tilled crucible may be performed under as nearly as possible the same circumstances. In using platinum crucibles, it must be remembered that certain substances must not be ignited in them. Berzelius says that ' it is improper to ignite in platinum vessels the caustic alkalies or the nitrates of any alkaline base, such as lime, baryta, or strontia, because the affinity of the alkali for platinum oxide causes a very considerable oxidation of the metal ; and after the saline matter is re- moved, the surface of the metal is found to be honey- combed.' The alkaline sulphides or the alkaline sulphates mixed with charcoal are inadmissible, because the sulphides so- formed attack platinum even more energetically than the caustic alkalies ; so are metals whose fusing-point is lower than that of platinum, because an alloy would be formed. Gold, silver, and copper may be heated to dull redness in platinurn vessels without danger ; but fused lead cannot come in contact with platinum without destroying it. A drop of fused lead, tin, zinc, or bismuth, placed on red-hot platinum, always produces a hole. Neither can a phosphide 140 SILVER CRUCIBLES NICKEL CRUCIBLES. or phosphoric acid mixed with charcoal be ignited in vessels of platinum, because a platinum phosphide is pro- duced, which is an exceedingly brittle compound. In analyses by the wet method, nitro-hydrochloric acid (aqua regia), even when very dilute, must not be allowed to come in contact with platinum, and, as a general rule, liquids containing either free chlorine, bro- mine, or iodine must not be boiled in platinum capsules. SILVER CRUCIBLES can only be used at temperatures below full redness. They are not affected by caustic alkalies, but must not come in contact with sulphur or be heated over coke, coal, gas, or other fuel containing sulphur. NICKEL CRUCIBLES. Recently nickel has been intro- duced as a material for crucibles. It is well known that pure nickel is one of the toughest of all the metals, .and that it fuses only at very high temperatures. It lias a fine grain, takes a high polish, and is very compact and unalterable. These qualities have led to its being employed for crucibles and evaporating dishes. Mr. Wanklyn has published some notes on the behaviour of these vessels in his laboratory. He finds that for many purposes crucibles of pure nickel are quite as serviceable as platinum crucibles, and they are much cheaper, costing only about one-tenth as much as platinum. They stand the action of alkalies remarkably well there was no alteration in the weight of the crucible after caustic pot- ash had been fused in it. Hydrochloric acid in the cold, whether dilute or concentrated, may be used to clean out these crucibles and no alteration in weight is the result. Cold oil of vitriol is likewise without action ; but con- centrated nitric acid attacks them, causing rapid loss of weight. Nickel dishes are especially useful for taking water-residues and milk- solids, and indeed for these pur- poses nickel is not inferior to platinum. Mr. Bertram Blount considers that the uses of nickel for chemical purposes are confined to (i.) Dry ignitions (i.e. not fusions) in an oxidising flame, provided the heat be not too intense. CUPELS. 141 (ii.) Fusions with caustic alkalies. From which may probably be deduced (iii.) Fusions with barium hydrate. (iv.) Fusions with alkaline carbonates. The inability of nickel to stand high temperatures, in contact with a reducing flame, limits its employment con- siderably, and demands caution in dealing with it. CUPELS. These are vessels in which the operation termed cupellation is carried on. They must be made of such substances as are not acted upon by certain fused oxides, as those of lead or bismuth, and their texture has to be sufficiently loose to allow of the oxides penetrating their substance readily, and yet be sufficiently strong to bear handling without breaking. There are several substances of which cupels might be made, which will fulfil all these conditions, but only one is in general use, viz. the ash of burnt bones. This consists principally of calcium phosphate, with a little calcium car- bonate and fluoride. Berzelius found that bones of oxen contained 57 parts of calcium phosphate for every 3'8 parts of calcium carbonate, whilst, according to Barros, sheep bones contain 80 parts of calcium phosphate to 19 parts of calcium carbonate. When bones are burnt whole they likewise contain mineral matter derived from the cartilage, such as alkaline sulphates and carbonates. The greater part of the calcium carbonate is likewise converted into caustic lime. The bones of sheep and horses are best for cupels. In getting rid of the organic matter, it is advisable to boil them repeatedly in water before burning them. This dis- solves a great part of the organic matter. If the bones are not rendered quite white by the first ignition, but con- tain a little carbon, they should be ground up, moulded into shape, and burned again. Care should be taken not to heat the bone earth too strongly. In this case the bones will have a smooth, glassy fracture, and will not be sufficiently spongy or absorbent to make good cupels. When the bones are burnt white throughout, they must 142 MANUFACTURE OP CUPELS. be finely ground, sifted, and washed several times with boiling distilled water till all soluble salts are removed. The finest particles of the powdered bone earth will remain longest suspended in the washing waters. This must be allowed to settle separately, and should be reserved for giving a final coating to the surface of the cupels ; this coating acts, to a certain extent, like a fine filter, and may be applied to all cupels, although the body of the cupel is made of different materials. For the body of the cupels, the bone-ash should be -about as fine as wheat flour. If too coarse, litharge con- taining silver will be absorbed into its pores, and will occa- sion a loss of silver. Cupels must neither crack nor alter in texture at a white heat. It is very important that they should not contain carbon, and therefore, in making them, the bone earth must not, as sometimes recommended, be mixed with beer, or water containing adhesive substances. Nothing but pure water should be used, and the mixture should be just sufficiently moist to adhere strongly when well pressed, but not so moist as to adhere to the finger or the mould employed to fashion the cupels. The mould (fig. 62) con- FIG. 62. sists of three pieces : one a ring, b, having a conical opening ; another a pestle, a, having a hemispherical end fitting the larger opening of the ring ; and the third, c, a piece of turned metal, into which b fits ; c serves to form an even bottom to the cupel. In order to mould the cupels, proceed as follows : Place the ring on the lower piece e, and fill it with the com- position : then place the pestle upon it, and force it down as much as possible : by this means the moistened bone ash will become hardened, and take the form of the pestle ; the latter must then be driven as much as possible by repeated blows from a hammer, until quite home. The surface of the cupel may then have sifted over it a little of the very fine levigated bone- ash, and the pestle hammered again on it. It is then to be turned lightly round, so as to smooth the inner sur- MANUFACTURE OF CUPELS. 143 face of the cupel, and withdrawn : the cupel is removed from the mould by a gentle pressure on the narrowest end. When in this state, the cupel must be dried gently by a stove ; and lastly, ignited in a muffle, to expel all moisture. It is then ready for use. There are two or three points to attend to in manufac- turing the best cupels. First, the powdered bone-ash must be of a certain degree of fineness ; secondly, the paste must be neither too soft nor too dry ; and thirdly, the pressure must be made with a certain degree of force. A coarse powder, only slightly moistened and compressed, furnishes cupels which are very porous, break on the least pressure, and, as before mentioned, allow small globules of metal to enter into their pores. When, on the contrary, the powder is very fine, the paste moist, and compressed strongly, the cupels have much solidity, and are less porous ; the fine metal cannot penetrate them, but the operation proceeds very slowly : besides, the assay is likely to become dulled, and incapable of proceeding without a much higher degree of tempera- ture being employed. Cupels for assaying silver bullion are sometimes made of equal parts of bone-ash and beechwood-ash ; and for assaying gold, 2 parts of beechwood-ash and 1 part of bone-ash are used. The hemispherical cavity of both these kinds are coated with the fine levigated powder of bone-ash . Beechwood-ash is preferred for the manufacture of cupels on account of the larger proportion of phosphoric acid it contains. According to Hertwig, beechwood-ash contains in 100 parts : Potassium carbonate 11-72 Sodium carbonate 12-37 Potassium sulphate . 3-49 Calcium carbonate 49-54 Magnesium carbonat :* 7-79 Calcium phosphate 3-32 Magnesium 2-92 Iron 0-76 Aluminium ,, 1-51 Manganese Silica 1-59 2-46 144 PYROMETRY. SCORIFIERS. A scorifier (fig. 63) is a vessel made much in the shape of a cupel, but of crucible earth. The proper use both of cupels and scorifiers will be explained under the head of silver- assaying. METHODS OF MEASURING THE HEAT OF FURNACES. As much of the accuracy of an assay depends on the tempera- ture at which it is made, and as the temperature required FIG. 63. varies with each metal, it is very desirable to possess some means of ascertaining the heat of the furnace more accu- rately than by the eye. Many persons have devised instruments, called pyrometers, for this purpose ; the earliest being those of Mr. Wedgwood and the late Pro- fessor Daniel, of King's College. We shall not give a description of Wedgwood's pyro- meter, as its indications are inaccurate, from the fact that the clay cylinders, whose contraction serves to measure the temperature, will contract as much as by the long con- tinuance of a low heat as by the short continuance of a high one. Hence the degrees of heat measured by Wedg- wood's pyrometer have been enormously exaggerated. It was long since noticed that it did not produce comparable effects; and this was supposed to proceed wholly from the impossibility of obtaining clay perfectly alike for each experiment. Daniel's pyrometer is composed of a rod of platinum simply laid in a groove made of refractory clay, and baked in the highest degree of heat. This rod rests at one end DANIELS PYROMETER. 145 on the edge which terminates the groove, and at the other on a lever with two arms, the larger of which forms a needle on a graduated arc of a circle ; so that the removal of this needle from its position marks the additional length which the metal acquires by the heat. The following is Daniel's description of his pyrometer : ' It consists of two parts (see fig. 64), which may be distinguished as the The register is a solid bar of black- register and the scale. FIG. 64. J lead or earthenware highly baked. In this a hole is drilled, into which a bar of any metal, a, six inches long, may be dropped, and which will then rest upon its solid end. A cylindrical piece of porcelain, , called the index, is then placed upon the top of the bar, and confined in its place by a ring or strap of platinum passing round the top of the register, which is partly cut away at the top, and tightened by a wedge of porcelain. When such an arrange- ment is exposed to a high temperature, it is obvious that the expansion of the metallic bar will force the index for- ward to the amount of the excess of its expansion over that of the black-lead, and that when again cool it will be left at the point of greatest elongation. What is now required is the measurement of the distance which the index has been thrust forward from its first position, and this, L 146 DANIEL S PYROMETER. though in any case but small, may be effected with great precision by means of the scale c.' * This is independent of the register, and consists of two rules of brass accurately joined together at a right angle by their edges, and fitting square upon the two sides of the black-lead bar. At one end of this double rule a small plate of brass projects at a right angle, which may be brought down upon the shoulder of the register formed by the notch cut away for the reception of the index. A movable arm is attached to this frame, turning at its fixed extremity on a centre, and at its other carrying the arc of a circle, whose radius is exactly five inches, accurately divided into degrees, and thirds of a degree. Upon this arm, at the centre of the circle, another lighter arm is made to turn, one end of which carries a vernier with it, which moves upon the face of the arc, and subdivides the former graduation into minutes of a degree ; the other end crosses the centre, and terminates in an obtuse steel point, turned inwards at a right angle. When an observation is to be made, a bar of platinum or malleable iron is placed in the cavity of the register ; the index is to be pressed down upon it, and firmly fixed in its place by the platinum strap and porcelain wedge. The scale is then to be applied by care- fully adjusting the brass rule to the sides of the register, and fixing it by pressing the cross piece upon the shoulder, and placing the movable arm so that the steel part of the radius may drop into a small cavity made for its reception, and coinciding with the axis of the metallic bar. The minute of the degree must then be noted which the vernier indi- cates upon the arc. A similar observation must be made after the register has been exposed to the increased tem- perature which it is designed to measure, and again cooled, and it will be found that the vernier has been moved forward a certain number of degrees or minutes. The scale of this pyrometer is readily connected with that of the thermometer by immersing the register in boiling mercury, whose temperature is as constant as that of boil- ing water, and has been accurately determined by the * Daniel's ' Chemical Philosophy,' p. 111. DANIEL S PYROMETER, 147 thermometer. The amount of expansion for a known number of degrees is thus determined, and the value of all other expansions may be considered as proportionate. By Daniel's pyrometer the melting-point of cast iron has been ascertained to be 2,786, and the highest tem- perature of a good wind furnace 3,300 Fahrenheit points which were estimated by Mr. Wedgwood at 17,977 and 21,877 respectively. The following is a list of the melting-points of some of the metals as ascertained by Professor Daniel ; and it is obvious that in an assay of each particular metal the tem- perature employed must exceed by a considerable number of degrees its melting-point. The table is, therefore, very useful. Fahr. Tin melts at - - - - - - - 4-22 Cadmium 442 Bismuth 497 Lead . 612 Zinc . 773 Silver 1860 Copper 1996 Gold . 2016 Cast iron 2786 Cobalt and nickel are rather less fusible than iron. Mr. S. Wilson* has described an ingenious process of measuring high temperatures. He exposes a given weight of platinum or Stourbridge clay to the action of the heat which is to be measured, and then quenches it in a definite weight of water at a certain temperature. Thus, if the piece of platinum weigh 1,000 grains and the water 2,000 grains at 60 F., and should the heated platinum when dropped into the water raise its temperature to 90, then 90- 60 =30 ; which, multiplied by 2 (because the weight of the water is twice that of the platinum), gives 60 the temperature to which a weight of water equal to the platinum would have been raised. To convert this into Fahrenheit degrees we must multiply by 31J, which is the specific heat of water as compared with platinum, that of the latter being 1. Therefore 60 x 31 J 1875, which will be the temperature of the furnace. Philosophical Magazine,' ser. iv. vol. iv. p. 157. L2 148 COMPARISON OF HIGH TEMPERATURES. FIG. 65. One or two other methods of measuring high tempera- tures applicable to special cases may here be mentioned. Mr. C. W. Siemens, C.E., F.E.S., has invented an inge- nious pyrometer, the principle of which is, that as the electrical conductivity of platinum, iron, and other metals decreases as they rise in tem- perature, their increase of resistance to the passage of the current is a measure of the heat to which the metals are subjected. The principle of construction may be ex- plained by the aid of fig. 65, in which F A B is a tube of pipe-clay, and the length between the projections A and B has a screw-shaped spiral groove cut on its outer surface ; the length of this part of the tube is about 3 inches. A spiral of fine platinum wire lies in the groove, each turn of the platinum spiral being thus pro- tected from lying in contact with its neighbour by the projecting edges of the groove, by which plan of insulation the current is forced to pass through the whole length of the fine wire. D is a little platinum clam, connected with one pole of the battery, and the position of this clam on the spiral regulates the length of pla- tinum wire through which the current shall pass. By this plan of adjustment all the pyro- meters constructed by Mr. Siemens are made to agree with each other. At F the ends of the thin platinum wire are connected with very thick platinum wire, and higher up, near E r where the heat of the furnace is less felt, the thick platinum wires are connected with thick copper wires, shown at P ; from E to F these connecting wires are pro- tected by clay pipes, as shown in the cut. When this arrangement has to be used, the whole of it is dropped into a thick metal pipe made of iron, copper, or platinum, according to the heat of the furnace to be tested. The lower end of this outer pipe is shown at KM, and when it is used the spiral A B lies inside it at N M. At COMPARISON OF HIGH TEMPERATURES. 149 R there is a very thick collar of metal in which the heat accumulates, and this prevents the cooling action of the length K R (most of which does not enter the furnace) from interfering with the accuracy of the indications. The ends of the wires p are connected with a suitable and very delicate electrical apparatus, by which the increasing elec- trical resistance of the hot spiral is measured. A good plan for comparing the temperatures of two furnaces is to prepare alloys of platinum and gold, con- taining definite quantities, say 5, 10, 15, 20 per cent. &c., of gold. These fuse at intermediate temperatures between gold and platinum. By placing small angular chips of these alloys separately in muffles, and noticing which are melted, which softened only, and which resist the action of the heat, an idea of the power of the furnace is obtained. In this way the amount of heat required to perform any operation may be registered for future reference, by sim- ply recording that it was sufficient just to melt, say, a 20 gold 80 platinum alloy. 150 CHAPTEE V. FUEL : ITS ASSAY AND ANALYSIS. BEFORE treating of the assay of metals and metalliferous- ores, it is advisable to devote some space to the important subject of fuel. The substances employed as fuel, although all of vegetable origin, are derived either from the veget- able kingdom (wood), or from the mineral kingdom (peat, brown coal, coal, anthracite). These natural fuels can be converted into artificial fuels by heating them more or less out of contact with the air (charcoal, turf-charcoal, coke). The essential elements of combustible matters are car- bon, oxygen, and hydrogen ; nitrogen being present some- times, but only in small proportions. These constitute the organic part ; various salts and silica constitute the inorganic part, or ash. The valuable constituents of fuel, on which its calorific and reducing powers depend, are the carbon and hydrogen, and it is upon the combustion or union of these elements with oxygen to form carbonic acid and water that the heating effect of the fuel depends. The more oxygen a fuel contains, the less carbon and combustible gases it will yield, and the more hydrogen, the more combustible gases. The proportion of hydrogen to oxygen in wood . is 1 turf . 1 fossil wood 1 coal . 1 anthracite 1 6 4 2-3 1 The more oxygen, the less carbon the fuel contains, thus : Anthracite contains about 90 per cent, carbon Coal 80 Brown coal 70 Fossil wood and turf 60 Wood 60 ASSAY OF FUEL. 151 The more carbon a fuel contains, the greater heat it pro- duces, and the more difficult it is to ignite. The greater the amount of hydrogen in a fuel, the more inflammable it will be, and the larger flame it gives, the hydrogen being evolved below a red heat. But the more carbon present the less flame. These differences are shown in a blazing fire and a glowing fire. In a flame the hottest part is at the periphery, whilst in a glowing fire the greatest heat is in the immediate contact of the burning surface. An elementary analysis of coal teaches little with regard to the nature or practical value of the combustible. A proximate analysis, on the contrary, enables us to learn something in regard to the real nature of the coal. The moisture and ash are not only diluents of the fuel, but are in themselves obstacles to its effectiveness ; the vapo- risation of the moisture causes a serious loss of heat, whilst the ashes, by hindering complete combustion and by the heat they contain when dropped through the grate, con- stitute another loss. By furthermore determining the total amount of volatile matter we learn both the per- centage of coke in the fuel and the amount of carbon (fixed combustible) and bitumen (volatile combustible matter). Although neither of these two products can be considered as simple chemical . compounds, it is never- theless of the utmost practical importance to know these two quantities, because of the great value of coke and gas in manufactures. The assay of fuel comprises the following examina- tions : 1. The examination of the external appearance of the fuel, its porosity or compactness, its fracture, the size and shape of the pieces composing it. 2. Estimation of the adhering water. 3. Estimation of the specific gravity. 4. Estimation of the absolute heating power. 5. Estimation of the specific heating power. 6. Estimation of the pyrometric heating power. 7. Estimation of the volatile products of carbonisa- tion. :152 ASSAY OF FUEL. 8. Examination of the coke or charcoal left behind on carbonisation, both with regard to quality and quantity. 9. Estimation of the amount of ash, and its composi- tion. 10. Estimation of the amount of sulphur. 11. Examination of any other peculiarity which may be noticed during the burning or carbonisation of the fuel. 1. EXTERNAL APPEARANCE OF THE FUEL, ITS POROSITY, COMPACTNESS, FRACTURE, SIZE, AND SHAPE OF PIECES. From the outward appearance of a fuel, its cleavage, and an examination of the embedded earthy matter, iron pyrites, gypsum, &c., its applicability to any special purpose may be judged. Its degree of inflammability, together with the pressure of blast which it will bear in the furnace, partly depend on the more or less compactness of the fuel. iThe amount of loss which it will suffer in transport de- pends upon its friability. Playfair and De la Beche * estimated the amount of this loss in coal by rotating in a barrel different qualities of coal for the same time. The powder produced was separated and weighed, and in this way the friability or cohesion of a fuel could be ex- pressed in percentages. Schrotter f made the same experi- ments with brown coal. The size and form of the pieces composing the fuel is important, as on this depends the space occupied in its stowage an important point for steam-vessels. This space cannot be calculated from its specific gravity, but must be ascertained by direct measurement. The space occupied will be smallest when the form of the lumps is cubical. 2. ESTIMATION OF THE ADHERING WATER. The water contained in a fuel exerts great influence on its heating power. It not only increases its bulk, but it acts in- juriously by abstracting a certain quantity of heat required for its evaporation, and it also causes imperfect combus- * Dingl. ex. 212, 262; cxiv. 346. Liebig's * Jahresber.,' 1847-1848, p. 1117; 1849, p. 708. t Wien. Akad. Ber. 1849, Nov. and Dec. p. 240. Liebig's ' Jahresber., 1849, p. 709. ASSAY OF FUEL. 15U tion. For this reason, wood, turf, and brown coal never give so high a temperature as coal, anthracite, and coke. The estimation of the adhering water is effected by drying a certain weight of the pounded fuel in a water- bath at 212 F., or in an air-bath at 220. It may also be ascertained by placing a certain weight of the powdered fuel in a glass tube, heating to 212, and passing over it air dried by means of chloride of calcium, till the fuel no longer loses weight. The amount of water which the dried fuel will absorb from the atmosphere in twenty-four hours should also be estimated, in order to ascertain its hygroscopic qualities. 3. ESTIMATION OF THE SPECIFIC GRAVITY. The specific gravity of a fuel depends on its density and the amount of ash, and it appears also to be in proportion to its greater or less inflammability. Of two equal volumes of carbonised fuel, the one will produce the greatest heating effect which has the greatest specific gravity, provided the density is not produced by mineral constituents. The estimation of the specific gravity is difficult, and sometimes uncertain, owing to the cleavage of the fuel, and the entanglement of air in its pores. The best way of obviating this difficulty is as follows : Coarse fragments, freed by means of a sieve from all small particles, and averaging l-10th c.c. in volume, are introduced into a fifty-gramme flask provided with a thermometer stopper. The constants for this flask for temperatures varying from 50 to 80 F. are previously carefully estimated. The true sp. gr. corresponds to the coal perfectly soaked, so that all its pores are filled with water. That requires, on the average, 12 hours, permitting two es- timations per day, one in the morning, another in the evening. That this precaution is important may be seen from the following example : a sample of coal gave the sp. gr. 1-309 at 64 F., immediately after filling the flask with water : after about 12 hours' soaking, the sp. gr. had increased to 1-328, for the same temperature. According 154 ASSAY OF FUEL. to this latter estimation, a cubic foot of this coal would weigh 82-76 Ibs. ; according to the former, only 81*58, or 1-18 Ibs. less. This shows a considerable degree of po- rosity of the coal, and indicates the absurdity of giving the weight in pounds of a cubic foot of coal with four decimals, although no statement in regard to temperature or time of weighing is made. 4. ESTIMATION OF THE ABSOLUTE HEATING POWER. The value of a fuel for any purpose depends chiefly on its price and the quantity required for that purpose. The quantity required depends on the heating power possessed by a certain weight of fuel (its absolute heating power) or that possessed by a certain volume (its specific heating power). The less oxygen, ash, and water the fuel contains, the greater its heating power will be, and this will also increase in proportion to the carbon and hydrogen present. Whether the combustion is effected quickly or slowly, the amount of heat produced will be the same, but the degree of temperature attained will be very difficult. This latter constitutes the pyrometric heating power. The estimation of the absolute heating power of a fuel may be effected a. By heating a definite quantity of water from 32 F. to 212 ; . b. By ascertaining how much fuel is required to melt a known weight of ice ; c. By ascertaining how much water may be evaporated by 1 Ib. of different kinds of fuel ; d. By ascertaining how much the temperature of a room increases by burning a certain weight of a fuel in a stove. e. By ascertaining the elementary composition of the fuel, and calculating how much oxygen will be required to convert the carbon and hydrogen into carbonic acid and. water ; the quantity of heat produced will be in pro- portion to the amount of oxygen consumed. /. By Berthier's method. g. By lire's method. BERTHIER'S METHOD FOE ASSAY OF FUEL. 155 According to Berthier, the most convenient method for ascertaining the comparative calorific power of any combustible matter is by means of litharge. He says : It has been proved by the experiments of many philosophers that the quantities of heat emitted by combustible sub- stances are exactly proportioned to the amounts of oxygen required for their complete combustion. Whence, after the elementary constitution of any combustible is known, its calorific power is easily estimated by calculation. For instance, it is only necessary to ascertain the quantity of oxygen absorbed in the conversion of all its carbon into carbonic acid, and all its hydrogen into water, and compare that quantity with that which is consumed in burning a fuel whose calorific power is well ascertained. Such a fuel is pure charcoal. By adopting the principle just pointed out, it may be conceived that, without knowing the composition of a fuel, its heating power may be ascertained by estimating the amount of oxygen it absorbs in burning. This can be done in a very simple and expeditious manner, if not exactly, at least with sufficient exactitude to afford very useful results in practice. It is as follows : many metallic oxides are reduced with such facility that when heated with a combustible body, the latter burns completely, without any of its elements escaping the action of the oxygen of the oxide, if the operation be suitably per- formed. The composition of the oxide being well known> if the weight of the part reduced to the metallic state be taken, the quantity of oxygen employed in the combustion can be ascertained. In order to collect the metal and separate it from the non-reduced mass, it, as well as its oxide, must be fusible. Litharge fulfils these conditions, and experiment has proved that it completely burns the greater part of all ordinary fuels ; the only exceptions are some very bituminous matters containing a large proportion of volatile elements, a portion of which escapes before the temperature is sufficiently high to allow the reduction to take place. The experiment is made as follows : 10 grains of the finely powdered or otherwise divided fuel is mixed 156 BERTHIER'S METHOD FOR ASSAY OF FUEL. with about 400 grains of litharge. The mixture is care- fully placed in an earthen crucible, and covered with 200 grains more litharge. The crucible is then placed in the fire and gradually heated. When the fusion is perfect, the heat is urged for about ten minutes, in order that all the lead may collect into a single button. The crucible is then taken from the fire, cooled, broken, and the button of lead weighed. Sometimes the button is livid, leafy, and only slightly ductile ; in/ which case it has absorbed a little litharge. This can be partially prevented by fusing slowly, and adding a little borax. Two assays, at least, ought to be made, and those re- sults which differ more than a grain or two ought not to be relied on. The purer the litharge, the better the re- sult ; it ought to contain as little minium as possible. It is an excellent plan to mix up the litharge of commerce with one or two thousandths of its weight of charcoal, and fuse the whole in a pot ; when cold, pulverise the litharge, which will now be deprived of minium. Pure carbon produces, with pure litharge, 34 times its weight of lead, whilst hydrogen gives 103 '7 times its weight of lead ; that is to say, a little more than three times as much carbon. We can, therefore, from these data, find the equivalent of any fuel, either in carbon or hydrogen. When a fuel contains volatile matters, the quantity can be ascertained, as before pointed out, by ignition in a close tube or crucible. If, further, we ascertain the pro- portion of lead it gives with litharge, it is easy to calculate the equivalent in carbon of the volatile matters, and, in consequence, to ascertain its calorific value. Supposing that a substance gives by distillation C parts of coke, or carbon, having deducted the weight of the ash and of volatile substances, and that it produces P parts of lead with litharge. The quantity C of carbon would give 34 x C of lead ; the quantity of volatile matter would give but P 34 x C ; it would be equivalent to * |^ of carbon : whence it follows that the quantity of heat developed by the charcoal, the volatile matter, and the unaltered com- DR. URK ON BERTH IERS METHOD. 157 bnstihle, will be to each other as the numbers 34x0,. P-34xC, and P. Dr. Ure* says, speaking of the above method of assay,. ' On subjecting this theory to the touchstone of experi- ment, I have found it to be entirely fallacious. Having mixed very intimately 10 grains of recently calcined char- coal with 1,000 parts of litharge, both in line powder, I placed the mixture in a crucible, which was so carefully covered as to be protected from all fuliginous fumes, and exposed it to distinct ignition. ' No less than 603 grains of lead were obtained, whereas, by Berthier's rule, only 340 or 346*6 were possible. On igniting a mixture of 10 grains of pulverised anthracite with 500 grains of pure litharge previously fused and pulverised, I obtained 380 grains of metallic lead. In a second experiment, with the same anthracite and the same litharge, I obtained 450 grains of lead ; and in a third,, only 350 grains. It is therefore obvious that this method of Berthier's is altogether nugatory for ascertaining the quantity of carbon in coals, and is worse than useless in judging of the calorific qualities of different kinds of fuel.' This discrepancy in the results obtained by Dr. Ure is very perplexing, and does not at all accord with Berthier's experience, as shown by his experiments, or by the author's on the subject. The latter never had a difference of more than 50 grains, and in general only two or three, which latter result is satisfactory. The only precaution he found necessary was to heat very gradually until the mixture was fully fused, and then to increase the fire to bright redness for a few minutes. Further experiments have been made by the author on this subject, and he has succeeded most perfectly in estimating the value of a fuel. With the litharge of com- merce, which contains much minium, the process is never exact : results have been obtained differing as much as 40 or 50 grains when the litharge employed had not been purified, and to purify it completely is a troublesome * ' Supplement to the Dictionary of Arts, Mines, and Manufactures.' 158 UEE'S CALORIMETER, process. This difficulty may be completely obviated, however, by substituting for litharge, white-lead, using for each 10 grains of fuel 700 grains of white-lead, which FIG. 66. DIMENSIONS. A B=3J in. ; diameter f in. C D=2 in. C B, socket for AB. Diameter across D=4 in. E F=6 in. F G=5 in. ; diameter H G = l| in. A B weighs 39 grammes. The remainder of the apparatus, including the stopcock, weighs 391 grammes. SCALE OF 12 INCHES. 12 are well mixed with it, and 300 grains of pure white-lead to cover the mixture. When the whole is heated, the carbonate of lead de- composes, forming pure lead oxide, which is then reduced, as in the former case. By this process the results corre- spond to 1 grain in the quantity of lead produced from a WRIGHT'S CALORIMETER. 159 given sample of fuel. Of course great care must be taken that the white-lead is genuine. Commercial samples are frequently adulterated with lead sulphate and barium sulphate, lead oxychloride, zinc oxide, &c. This is a serious drawback to this otherwise excellent modification. WRIGHT'S CALORIMETER. The instrument known as Wright's calorimeter gives very accurate results, and is the one most generally used now in experiments on the heating power of fuel, in all but the most refined investi- gations. It is shown in the accompanying figure (fig. 66). The copper cylinder A B is filled with a mixture of 20 grains of the combustible, and 240 of the deflagrating com- pound, which is composed of three parts of potassium chlorate and one of potassium nitrate. A little piece of cotton soaked in potassium chlorate is placed partly in the mixture, the other end projecting above the top of the cylinder ; this is ignited, quickly covered with the bell- shaped part of the apparatus, and immersed in a measured quantity of water. As constructed, the whole metallic apparatus weighs 6,642'7 grains, and with this weight 290-1 grains of water are used. The temperature is re- corded before and after making the experiment. During the deflagration the stopcock is closed ; it is, however, opened before taking the temperature the second time. A tenth of the temperature that the water is raised by the combustion is added for errors that are incidental to the use of the instrument. If the instrument is made of the weight above given, the result is obtained by a very simple calculation. Each Fahrenheit degree by which the temperature of the water has been augmented corresponds to a pound of water converted into steam. EXAMPLE. Fahr. Temp, of water before making experiment = 56 after the combustion . = 65 ~9 + Ath produced by the combustion = 9'1 One pound of the coal will convert into steam (maximum effect) 9'1 Ibs. of water at 212 F. 100 PYROMETKIC EXAMINATION OF FUEL. 5. ESTIMATION OF THE SPECIFIC HEATING POWER. This represents the heat produced from a certain volume of fuel. It may be ascertained by multiplying the absolute heating power by the specific gravity. 6. ESTIMATION OF THE PYROMETRIC HEATING POWER. By pyrometric heating power is meant the degree of tem- perature which may be obtained by completely burning the fuel. This heating power not only depends upon the composition of the fuel, but chiefly on the time required for its combustion, and this again depends on the looseness and inflammability of the fuel. The absolute heating power of hydrogen is greater than that of carbon, but with regard to the pyrometric heating power it will be found that the reverse is the case. Carbon burned in contact with the air to carbonic acid will produce a heat of 2,558 C. ; if burned to car- bonic oxide it only produces 1,310 ; hydrogen burning to water will produce a heat of 2,080. From this we learn that fuel rich in carbon, such as anthracite, coal, and coke, will produce a greater pyrometric effect than fuel rich in hydrogen, as wood, &c. Density is an essential quality of fuel required to pro- duce great pyrometric effect. This is proved in the follow- ing way. When atmospheric air first acts on the carbon con- tained in fuel, carbonic acid is formed, and the tempera- ture rises to a certain degree, but on passing over glowing coal, carbonic acid becomes converted into carbonic oxide, and this causes a portion of the heat at first produced to become latent. This conversion of carbonic acid into car- bonic oxide is more easy and complete as the fuel used is more inflammable ; and as v a greater quantity of heat is thereby rendered latent, it follows that the heating power of such a fuel is inferior. This accords with general ex- perience ; for it is well known that coke is able to produce a greater heat than charcoal. Good methods for estimating pyrometric heating power were given in the last chapter. 7. ESTIMATION OF THE VOLATILE PRODUCTS OF CARBONI- ASSAY OF FUEL. 161 SATION. The amount of volatile matter yielded on car- bonising a fuel depends partly on the composition of the fuel, and partly on the temperature employed. If a fuel rich in oxygen and hydrogen is quickly heated, it will yield the greatest amount of volatile products. These are partly liquid (tar, naphtha, and acetic acid or ammoniacal water), and partly gaseous (carbonic oxide, carbonic acid, and light and heavy carburet ted hydrogen). The more oxy- gen a fuel contains, the more carbonic acid and carbonic oxide it will produce ; the more hydrogen it contains, the more illuminating gas it yields. The applicability of a sample of coal to the production of illuminating gas depends on these conditions. Coal distilled at a low temperature yields much tar and comparatively little gas, and when a very high temperature has been used, less tar and more gas is produced, but the great heat will have reacted on the gas and injured its illuminating qualities. If the coal contains pyrites, the gas will contain sulphur compounds. The amount of water produced is generally larger than that of the tar. In order to estimate the amount of volatile matter given off from any particular sample of coal, proceed in the fol- lowing manner : Place a given weight, say 200 grains, of the coal in an iron tube closed at one end, to the other end of which adapt, by means of a cork, a glass or other tube, which must be led into an inverted jar full of water standing in the pneumatic trough. Eaise the tem- perature very gradually to redness, and continue the heat until no more gas is given off, then ascertain its quantity in cubic inches, with due correction for temperature and pressure. 8. EXAMINATION OF THE COKE OR CHARCOAL LEFT BEHIND ON CARBONISATION. The amount of coke or charcoal yielded by a sample of fuel is found by the last operation. This residue is the amount of coke which that particular sample of coal produces ; and its weight, divided by two, gives the percentage of coke. The process of coking, charring, or carbonising fuel, whilst it drives off some of the valuable hydrocarbon con- M 162 EXAMINATION OF COKE OF FUEL. stituents, also gets rid of all the aqueous elements. And therefore the coke or charcoal which is left behind has its value greatly increased when high temperatures are required, although, from the absence of flame-yielding constituents, it is much more difficult to ignite. The degree of inflammability of coke or charcoal is relatively the same as that of the raw fuel from which they were produced. The more inflammable a fuel has been, the more inflammable will be the coke or charcoal pro- duced from it. The temperature employed in the carbonisation, as has been already explained, exerts great influence on the yield of coke. If the fuel contains iron pyrites, part of the sulphur goes off in the volatile portion, but from one-fourth to one-half is retained in the form of iron sulphide. 9. ESTIMATION OF THE AMOUNT OF ASH. In order to ascertain the amount of ash : Fully ignite about 50 grains of coal in a platinum capsule, allowing the air to have free access all the time until nothing but ash is left. Its amount may then be ascertained by weighing : good fuel should contain little ash. It may vary from 1 to 10 per cent., but if it exceeds 5 per cent, it becomes deleterious. The chemical composition of the ash also influences the quality of the fuel to some extent. 10. ESTIMATION OF THE AMOUNT OF SULPHUR. This is an important operation in the assay, as a coal containing sulphur cannot be employed for particular operations, and, indeed, those which contain much sulphur ought only to be used for the commonest purposes. This assay is most important to ironmasters as well as to steamboat and other companies, who consume fuel under steam boilers ; and the coal they purchase should always be subjected to this particular test, as sulphur has a corroding and destroying action on iron and copper. Where sulphurous coals are continually burnt under boilers, the metal of the latter becomes deteriorated, and the boiler is rapidly rendered useless. Sulphur exists in coal in the form of iron pyrites ; this can generally be detected by its brassy colour. Some DETERMINATION OF SULPHUR IN FUEL. 1G:>> coals and lignites also contain calcium sulphate, and in rare cases barium sulphate. The process for the estimation of the amount of sulphur in coal is not difficult. 1 part of the coal is to be finely pulverised, and then mixed with 7 or 8 parts of potassium nitrate, 16 parts of common salt, and four parts of potassium carbonate, all of which must be perfectly pure ; the mixture is then placed in a platinum crucible and gently heated. At a certain temperature the whole ignites and burns quietly. The heat is then increased until the mass is fused : the operation is finished when the mass is white. It must, when cold, be dissolved in water, the solution slightly acidulated by means of hydrochloric acid, and barium chloride added to it as long as a white preci- pitate forms. This precipitate is barium sulphate, which must be collected on a filter, washed, dried, ignited, the filter burnt away, and the remaining barium sulphate weighed : every 116 parts of it indicate 16 of sulphur. Dr. Price has drawn attention to a source of error which lias hitherto escaped notice in the estimation of sul- phur, where fusion of the substance with nitre is the pro- cess employed. This author has found that unless great care be taken to prevent the fused mass passing over to the outside of the vessel, and so coming in contact with the flame or products of combustion, an appreciable and, in some cases, serious error will arise, owing to the sul- . phuric acid produced from the sulphurous acid in the flame a product of the oxidation of the sulphide of car- bon contained in the gas combining with the potassium of the fused salt. Several experiments have been made to ascertain the amount of error that may be occasioned from the above cause. In one instance the flame issuing from a Bunsen's burner was made to strike against a little fused nitre on the under side of a small platinum dish, when, in three quarters of an hour, as much sulphuric acid was obtained as is equivalent to 12 milligrammes of sulphur. As a check on these experiments, nitre was fused by the flame of the spirit lamp, when, as was to be anticipated, not a trace of sulphuric acid could be detected upon the IT 2 164 DETERMINATION OF SULPHUR IN FUEL. addition of a barium salt to the aqueous solution of this fused mass, rendered acid by hydrochloric acid. In es- timations of sulphur in coke or coal, great care should, therefore, be taken to prevent any of the fused saline con- tents of the crucible from getting on to the outside. In fusing pig iron with nitre, a process recommended by some for the estimation of the sulphur it contains, the mass, especially if the iron be rich in manganese, invariably creeps over to the outer wall of the crucible ; and it is, therefore, impossible to obtain correct results when the operation is conducted over the gas flame. The assayer should for these reasons always employ a spirit flame in preference to gas in sulphur estimations. Mr. Teikichi Makamura recommends the following pro- cedure : 3 or 4 parts of the mixed alkaline carbonates, or of sodium carbonate, are intimately mixed with one part of coal in very fine powder in a large platinum dish. The mixture is heated at first very gently, a spirit lamp being used instead of a Bunsen, to prevent possible absorption of sulphur ; the heat is then raised slowly without attain- ing that of visible redness, until the surface becomes only faintly grey. No smoke or odorous gases should escape during the whole of the oxidation. The tem- perature is now raised to a faintly red heat for sixty minutes, at the end of which time the mass is perfectly white, or reddish if iron be present. The mass is not to be stirred during the ignition. The residue is heated with water, filtered, and the sulphates estimated in the ordinary way. 11. EXAMINATION OF OTHER PECULIARITIES OF FUEL. Besides the above-named examinations, the assayer should notice the degree of inflammability of the fuel, and whether any particular smell is evolved during combustion ; whether the coal is good for coking purposes ; whether it burns with a large smoky flame or a luminous flame ; whether it burns quietly or with decrepitation ; and whether the ash is dusty or fusible, and likely to accumulate and clog up the grate-bars. 32. CALCULATION OF RESULTS. It may be sufficient here ASSAY OF FUEL. 165 to state that, beside the percentage composition of the -coal, it is proper to reduce the composition to the com- bustible =100, in order to obtain a comparative estimate of the character of the fuel itself (in regard to the propor- tion of bitumen and carbon), and of the amount and quality of the impurities (ash and moisture). It has also been shown that, for considerable areas of the coal-field, the sum of the constituents on the scale of combustible = 100 is the proper calorific equivalent, and that the per- centage of the combustible in the fuel gives a proper estimate of its value. ASSAY OP COAL BEFORE THE BLOWPIPE. The blowpipe method is well adapted to the assaying of .coal. Not only does the portableness of the apparatus make it very con- venient for use away from home, wherever the balance can be set up ; but its use at home is quite as satisfactory on the score of exactness as the assay with the muffle or retort, or large platinum crucible, and large balance. Mr. B. S. Lyman gives the following directions for carrying out this assay : Besides the ordinary pieces of the blowpipe apparatus, as made at Freiburg, all that needs to be made expressly for the coal assay is a small covered platinum crucible of the same size and shape as the clay crucibles of that appa- ratus ; and there must be a little ring of German silver, for the crucible to stand on, about three-eighths of an inch across and half that in height. Such a crucible cover and ring weigh about 40 grains more than the ordinary metallic cup that rests on the pan of the balance ; the crucible and ring without the cover weigh less than 30 grains more than the cup. If it be desired to estimate the amount of hygroscopic moisture in the coal, a small dry- ing bath must be made too ; but the hygroscopic water in ordinarily well-dried coals (not brown coals) is of little importance. The size of the crucible allows the coking of 3 to 10 grains of coal, according to the dryness of the coal and the extent of its swelling up when heated ; and as the blowpipe balance weighs within l-l,000th of a grain, it 16G ASSAY OF FUEL. is easy to weigh within much less than l-10th of one per cent, of the amount of coal assayed much nearer, in fact, than the exactness of the coke assay in other respects. On this point, indeed, the blowpipe assay is quite as good as the assay with the larger balance, especially the muffle assay, where the coal must be brushed into a clay recep- tacle after weighing, and the coke or ashes brushed off from it before weighing ; while here the crucible is weighed each time without removal of its contents, and without danger, therefore, of losing anything or adding any dust. It may be objected that the smallness of the amount of coal that can be assayed with the blowpipe makes it a less trustworthy indicator of the general com- position of the coal than a larger assay ; but the size of the lumps or powder assayed may be made finer accord- ingly, so that, when mixed up, an equally just sample of the whole mass would be got for the small assay as for the large. Any one who has had a little experience, both in the use of the blowpipe and in the ordinary muffle assay of coal, will scarcely need any further teaching for the coal assay with the blowpipe. For others it is worth while to say that the coal may be assayed either in a fine powder or in little lumps, and either with a slowly increasing or with a quickly increasing heat. A quick heat will give less coke by several per cents., but will often make a dry coal cake together that would not cake with a slow heat. The cover of the crucible should be left open a little, for the easy escape of the gas, but covered enough to prevent any flying off of solid material. The heat should increase to redness, and as soon as the escaping gas stops burning the heat should be stopped. As some coals part with their gas more quickly than others, of course no definite time can be fixed for heating all coals ; but the burning of the gas is a good enough sign. Care should be taken not to let the coke take up moisture from the air before weighing, as it will quickly do if it has a chance. Of course, owing to the different effects of quick or slow heating, a certain uniformity of result, even with perfectly uniform samples ASSAY OF FUEL. 167 of coal, can only be got without error by practice and by mechanical skill, by reproducing with nicety the same conditions in successive assays. After the coke has been weighed, it can be heated again with very free access of air, say with the crucible tilted to one side, and the cover off, until everything is thoroughly burnt to ashes ; and these should be re-heated until no change for the less is made in the weight. With free-burning, soft (semi-bituminous) coals this burning to ashes is very slow, so that it is very fatiguing or even im- possible to carry it out with a blowpipe ; but in that case the crucible may be heated over a Bunsen gas-burner or an alcohol lamp, and left to glow for hour after hour. The coking is far more conveniently done in the same way than by blowing with the mouth. As an illustration of the degree of accuracy which this method may be expected to give, the author adduces a pair of blowpipe assays, made five years ago, of some West Virginia asphaltum, that seemed itself to be much more uniform in composition than coal from different benches in one bed is apt to be : Volatile Matter Coke Ashes No. 1 . . 47-29 per cent. 52-71 per cent. 1-65 per cent. No. 2 . . 46-93 53-07 1-81 Mean . . 47-11 52'89 1'73 VALUATION OF COAL FOR THE PRODUCTION OF ILLUMINATING GAS. Take 100 grains of the coal in small lumps, so that they may be readily introduced into a rather wide com- bustion tube. This is drawn out at its open end (after the coal has been put in) so as to form a narrow tube, which is to be bent at right angles ; this narrower open end is to be placed in a wider glass tube, fitted tight into a cork fastened into the neck of a somewhat wide-mouthed bottle serving as tar vessel (hydraulic main of the gasworks). The cork alluded to is perforated with another opening, wherein is fixed a glass tube, bent at right angles for con- veying the gas first through a chloride of .calcium tube, next through. Liebig's potash bulbs containing a solution of caustic potash, having lead oxide dissolved in it. Next 168 ASSAY OP FUEL. follows another tube, partly filled with dry caustic potash, and partly with calcium chloride ; from this last tube a gas-delivery tube leads to a graduated glass jar standing over a pneumatic trough, and acting as gas-holder. Before the ignition of the tube containing the coal is proceeded with, all the portions of the apparatus are carefully weighed and next joined by means of india-rubber tubing. After the combustion is finished, which should be carefully con- ducted, so as to prevent the bursting or blowing out of the tube, the different pieces of the apparatus are discon- nected and weighed again. The combustion-tube has to be weighed with the coal after it has been drawn out at its open end, and with the coke after the end of the com- bustion when it is again cold ; and for that reason care is required in managing it. We thus get the quantity of tar, ammoniacal water, carbonic acid, and sulphuretted hydrogen (as lead sulphide) ; and the gas is measured by immersing the jar in water, causing it to be at the same level inside and out. Empty the Liebig's bulbs into a beaker, and separate the lead sulphide by filtration, wash carefully, dry at 212 F., and weigh. From the lead sul- phide the sulphuretted hydrogen present is calculated. This process, devised by the late Dr. T. Eichardson, of Newcastle-on-Tyne, was found by him to yield very trust- worthy results, so as to be suitable for stating what quantity of gas a ton of coal thus analysed would yield. 109 CHAPTEE VI. SEDUCING AND OXIDISING AGENTS FLUXES, ETC. IN some assay operations in the dry way, bodies are heated in suitable vessels per se ; but it is more often necessary to add to the bodies submitted to assay other substances, which are varied according to the nature of the change to be effected. These substances may be divided into five classes : I. reducing agents ; IE. oxidising agents ; III. desulphurising agents ; IV. sulphurising agents ; and lastly, V. fluxes properly so called. I. EEDUCING AGENTS. The substances belonging to this class have the power of removing oxygen from those bodies with which it may be combined. In assaying, the substance under examina- tion is generally fully oxidised either naturally or artifi- cially before reduction is required to be effected. The most common reducing agents are as follows : 1. Hydrogen gas. 2. Carbon. 3. The fatty oils, tallow, pitch, and resins. 4. Sugar, starch, and gum. 5. Tartaric acid. 6. Oxalic acid. 7. Metallic iron and lead. HYDROGEN GAS The most common method of prepar- ing this gas consists in dissolving zinc in dilute sulphuric acid. But as this plan gives the gas in the moist state, it must be dried by being allowed to bubble through oil of 170 EEDUCING AGENTS. vitriol or by being passed through a bottle containing fragments of dried calcium chloride before it is used for assaying purposes. This gas will only be required in very accurate assays, which are generally performed where there are ample conveniences for generating pure hydro- gen gas. The gas is colourless, invisible, and inodorous when absolutely pure. It is a most powerful reducing agent, and reduces a great number of metallic oxides at a red or white heat, viz. the oxides of lead, bismuth, copper, antimony, iron, cobalt, nickel, tungsten, molybdenum, and uranium. When any metal is required in a state of abso- lute purity, this is the only reducing agent admissible, as others give the metal combined with a certain proportion of carbon. CARBON Found in large quantities in the mineral kingdom, but generally combined with other bodies. In a state of purity it constitutes the diamond. The diamond, like all other species of carbon, is unacted on by the highest possible temperature when in close vessels. It burns in atmospheric air and oxygen gas, but requires for combus- tion a higher temperature than ordinary charcoal. After the diamond, the varieties of carbon found in nature or artificially prepared are 1. BLACK-LEAD, or GRAPHITE. This is a mineral found in beds in the primitive formations, principally in granite and mica-schist. It is generally mixed with earthy sub- stances, and rarely yields less than 10 per cent, of ash. Before employing it for reduction purposes it should be purified. Lowe * has given an excellent plan for effecting this object. ESTIMATION OP THE VALUE OP GRAPHITE. G. C. Witt- stein f heats 15 grains of the sample to dull redness, and takes the loss as water. The residue is ground up with 45 grains of a mixture of equal equivalents of potassium and sodium carbonates, introduced into a platinum cru- cible, covered with 15 grains of potash or soda, and slowly heated to redness. The crust which is formed must be * Polyt. Centr. 1855, p. 1404. f Dingler's Pqlyt. Journal, 216, 45. REDUCING AGENTS. 171 from time to time pushed down with a platinum wire. After fusion for half an hour it is let cool, softened with water, heated almost to a boil for a quarter of an hour, filtered, washed well, and the liquid set aside. The residue on the filter is dried, placed in a small flask, the ashes of the filter added, and about 50 minims of hydrochloric acid (sp. gr. 1-12) poured in. On the mixture becoming slightly gelatinous a few more drops of the acid are added, the digestion is continued for an hour, the mixture diluted with water, filtered, and washed. Pure graphitic carbon remains on the filter, and is dried, slightly ignited, and weighed. The acid filtrate mixed with the former one of an alkaline character is evaporated to dryness, and in it silica, alumina, ferric oxide, &c., are estimated in the usual manner. 2. ANTHRACITE. This is another species of fossil carbon much resembling ordinary coal, but differing from it by burning with neither smell, smoke, nor flame. 3. COKE. This is the residue of coal after all the volatile matter is expelled. It is generally iron-black, and has nearly a metallic lustre ; it is difficult to inflame, and burns well only in small pieces, but gives a very in- tense heat. Oven or furnace coke is preferable, as it is harder, lasts longer, and is more economical in use. 4. WOOD CHARCOAL. This is obtained by burning the woody part of plants with a limited supply of air, so as to drive off all their volatile matters, and leave merely their carbon. It is this kind that is generally employed in assays. It ought to be chosen with care, well pulverised, passed through a sieve, and preserved in well-stopped vessels. Wood charcoal is never perfectly pure, generally containing, besides ash, a proportion of hydrogen and watery vapour ; these bodies are not generally prejudicial, but in some experiments ash must not be present ; in that case pure charcoal may be procured by heating sugar to redness in a closed crucible. The advantage of carbon as a reducing agent consists in its great affinity for oxygen, which at a red heat sur- passes that of most other substances. Charcoal by itself 172 REDUCING AGENTS. possesses two inconveniences : first, it lias the property of combining with many metals ; and in the second place, it is infusible, and cannot combine with vitreous substances. The property it possesses of combining with iron, nickel, cobalt, &c., is of no consequence to the assay er, for the increase of weight it gives is not material, excepting under the circumstances to be hereafter pointed out ; but its infusibility and inability to combine with fluxes is a very serious inconvenience ; for after the reduction, that portion which has not been consumed remains disseminated with the grains of metal in the fused slag, and prevents the separation of all the metal, and the consequent formation of a good button ; a large quantity of charcoal can thus irreparably injure an assay. This inconvenience does not happen, however, when an oxide is reduced by cementa- tion in a lined crucible, but there are some cases in which this mode of reduction is inadmissible. Coke should never be used as a reducing agent in assays when it is possible to avoid it. It often contains a very large proportion of earthy and other extraneous matters (more particularly sulphur, which is very in- jurious). Coke is never so good as wood charcoal as a reducing agent, because it burns more slowly. When it is used, the temperature employed for an assay must be much increased. Coal is nearly always inconvenient, because it swells by heat ; nevertheless, as it is not required in very large quantities, it is sometimes employed, being finely powdered and sifted previous to use. THE FATTY OILS. The name oil is generally given to those bodies that are fat and unctuous to the touch, more Or less fluid, insoluble in water, and combustible. They all become solid at various degrees of temperature. There are some which, at the temperature of our climate, have constantly a solid form, as butter, palm oil, cocoa- nut oil, &c. ; TALLOW is an animal product analogous to the fatty oils in properties. EESINS. The greater part of the resins are solid, but REDUCING AGENTS. 173 some are soft. They are brittle, with a vitreous and shining fracture, and are often transparent. They are very fusible, but cannot be raised to their boiling-point without partial decomposition. Although all the bodies just mentioned consume in their combustion a large quantity of oxygen, they cannot generally effect the total reduction of an oxide on account of their volatility ; so that, before the temperature at which the reduction takes place can be attained, the greater part of the reducing agent has been expelled. They generally act only by virtue of the small carbonaceous residue produced by the action of heat ; so that their use is very limited and uncertain. Whenever they are em- ployed as reducing agents, without covering the substance, a loss is experienced, on account of the bubbling and boil- ing caused by their decomposition; this will always take place unless the contents of the crucible be covered with charcoal powder. Oils are very serviceable in the reduc- tion of a large mass of oxide by cementation ; in this case, after the oxide has been placed in the crucible, as much oil is added as the oxide and the lining of the crucible will soak up. Fat or resin is also used to prevent the oxida- tion of the surface of a metallic bath (as in the fusion of bar-lead samples), by coating the metal, and preventing the action of the atmospheric oxygen. SUGAR in its decomposition by heat leaves a much larger proportion of carbon than the oils, fats, or resins ; so that it would appear serviceable as a reducing agent. There are some cases in which it may be used with advantage, but it undergoes a great increase in volume when heated ; so that losses in an assay may occur by the use of this agent. To purify sugar from mineral ingredients it should be re-crystallised from alcohol. It then may be used as such, or after carbonisation. It yields about 14 per cent, of charcoal : this is pure carbon, and leaves no resi- due when burnt ; it is, therefore, preferable to wood char- coal in cases where no foreign matter should be introduced into the assay. STARCH, well dried, and, better still, terrified, is em- 174 KEDUCIXG AGENTS. ployed with advantage as a reducing agent, and is better than sugar, as it neither fuses, swells up, nor spirts, and in many cases is even preferable to charcoal, because it is in such a fine state of division that it can be more readily and intimately mixed with the substance to be reduced. Wheat and rye flour have nearly the same qualities as starch. They are sometimes used. GUM decrepitates slightly by heat, softens, agglomerates, and boils, without spirting. The gums can be employed as reducing agents under the same circumstances as sugar and starch, but the two latter are preferable, because they con- tain no earthy substances. TARTARIC ACID is the reducing agent in cream of tartar, or argol, of which so frequent use is made ; but the acid is never employed by itself. When heated in close vessels it fuses and decomposes, giving off combustible gases, leav- ing a little charcoal. It burns when heated in contact with air, giving rise to a peculiar and not unpleasant odour. OXALIC ACID fuses at a temperature of 208 without decomposing, but when heated to 230 it is decomposed, giving rise to carbonic acid, carbonic oxide, and a little formic acid vapour ; and when heated strongly some por- tions volatilise without decomposition : it does not leave a carbonaceous residue. The property which oxalic acid possesses of not leaving a residue would render it remarkably valuable for the re- duction of the metallic oxides in cases where the slightest trace of carbon is to be avoided, if its reducing power were greater ; but it decomposes at a low temperature, and in burning absorbs but a small quantity of oxygen, especially when it has not been dried ; so that even for the most easily reducible oxides a large proportion must be em- ployed. When it is combined with a base, as potash in potassium binoxalate, its reducing power is much aug- mented. AMMONIUM OXALATE, when heated in close vessels, is decomposed. Its reducing power is nearly double that of oxalic acid. COMPARATIVE REDUCING POWER. 175 COMPARATIVE REDUCING POWER OP THE ABOVE AGENTS. In order to give an idea of the comparative reducing power of the agents just described, the result of some assays made on them by Berthier, by means of litharge, are given below. By heating the same weight of each reducing agent with an excess of litharge, buttons of leads were obtained whose weights were proportional to the quantity of oxygen absorbed, and by comparing them with each other the re- ducing power of each flux is given ; by taking for unity the weight of the reagent, calculation has proved that 1 part of pure carbon reduces from litharge 34-31 of lead. The following are the results of Berthier's experi- ments : Hydrogen. Pure carbon Calcined wood charcoal Amber resin Ordinary wood charcoal Animal oil Tallow Resin Sugar Torrified starch Common starch Gum Arabic Tartaric acid Ammonium oxalate . Oxalic acid 104-00 34-31 31-81 30-00 28-00 17-40 15-20 14-50 14-50 13-00 11-50 11-00 6-00 1-70 90 It must be borne in mind that these numbers do not represent the quantities of oxygen each reagent would ab- sorb in complete combustion ; but that it only indicates the quantity of metal produced by equal weights of the reagents. In assaying, however, it is rarely that these agents are used by themselves ; they are generally mixed with a flux properly so called, which will be more particularly de- scribed under the head of Fluxes. METALLIC IRON removes oxygen from the oxides of lead, bismuth, copper, &c., but is rarely added for that especial purpose ; and when it does produce this effect it is gener- ally secondary, because it previously existed in the matter subjected to assay, or was added for some other purpose. 176 OXIDISING AGENTS. METALLIC LEAD reduces but a very small number of oxides, but it reduces many to the minimum of oxidation ; it also decomposes some sulphates and arseniates. II. OXIDISING AGENTS. The oxidising agents in general use are as follows : - 1. Oxygen, atmospheric or combined. 2. Litharge and ceruse. 3. Lead silicates and borates. 4. Potassium nitrate. 5. Lead nitrate. 6. Manganese peroxide. 7. Copper oxide. 8. Iron peroxide. 9. The caustic alkalies. 10. The alkaline carbonates. 11. Lead, copper, and iron sulphates.- 12. Sodium sulphate. LITHARGE is a fused protoxide of lead, and is generally obtained from the silver-lead works. When melted, it oxidises nearly all the metals, except mercury, silver, gold, palladium, platinum, &c., and generally forms very fusible compounds with the oxides. These two properties cause it to be a very valuable agent in separating silver and gold from all the substances with which they may be mixed. Litharge is occasionally mixed with a little of the red oxide of lead ; the presence of this in large quantities be- comes injurious, as it has the property of oxidising silver. Ordinary litharge can be easily freed from this oxide by fusing it and pouring it- into a cold ingot mould, then pulverising, and carefully keeping it from contact with air, as it readily absorbs oxygen, and if it be allowed to cool in the atmosphere it will nearly all be converted into the red oxide. CEKUSE, or WHITE-LEAD, is a carbonate of lead protoxide. OXIDISING FLUXES. 177 As it does not contain the slightest traces of red oxide, it may be used where the presence of that substance may be inconvenient ; but it is troublesome to use, as it is much less dense than litharge ; large vessels must be employed in consequence ; besides, it generally contains a small quantity of lead acetate or subacetate, and sometimes metallic lead separates from it on ignition, which is, in some cases, disastrous to the result of an experiment. When ceruse is employed, a certain quantity must be fused, to ascertain if any metallic lead be produced ; and, on the other hand, it must be examined to ascertain if it be adul- terated with barium sulphate. When it is pure it dissolves completely in acetic or nitric acid. LEAD SILICATES AND BORATES behave as litharge, but they oxidise less rapidly. They may be prepared by fusing together 1 part of silica or boracic acid with 1 part of litharge. The borates are more fusible than the silicates, but their use is attended with inconvenience, as they swell very much in fusing. POTASSIUM AND SODIUM NITRATES fuse at a temperature below redness without alteration, but when heated more strongly they give up oxygen. The action of these salts, when fused, is very energetic, because they have a great tendency to decompose, and because they contain a large quantity of oxygen. They are used as oxidising agents in the purification of the noble metals, and for ' preparing some fluxes. They ought always to be employed in a state of purity. Saltpetre often contains impurities. On this account an estimation of the real amount of potassium nitrate often becomes necessary, not only in cases where saltpetre is to be used for docimetric purposes, but also when used in certain technical operations, viz. the manufacture of gunpowder, enamel, &c. If saltpetre is very impure, it may easily be purified by recrystallisation to such a degree that it will only con- tain 2 to 3 per cent, foreign substances (chiefly sodium chloride). An exact assay of saltpetre is most difficult, and the 178 IMPURITIES IN SALTPETRE. different modes in use are not quite accurate, on ac- count of the nitrate of soda generally present as an impurity in saltpetre, and not easy to estimate by means of reagents. Manufacturers often intentionally mix the raw salt- petre with soda-saltpetre, and it is also often manufac- tured from a mixture of soda-saltpetre and potassium carbonate. ASSAY OF SALTPETEES. The following are the different modes of assaying saltpetre. a. To estimate the nitric acid directly, 150 grains of the well-ground sample are mixed with six times its weight of well-powdered silica, placed in a platinum capsule and dried in the water-bath till there is no further loss of weight. The temperature is then raised to dull redness for about half an hour, by which time the whole of the nitric acid is expelled. The difference in weight gives the nitric acid. Sodium nitrate may be treated in the same manner. The most usual, though not the most satisfactory, pro- cess for the assay of potassium and sodium nitrates is to estimate the total impurities, viz. water, insoluble matter, alkaline chlorides and sulphates, and the residue being taken as pure nitrate. b. Gay-Lussac's mode of assaying saltpetre consists in converting the potassium nitrate into potassium carbonate, and in estimating its amount volu metrically by means of standard sulphuric acid. 2-639 grains of saltpetre are mixed with 1 grain of ignited pine-root, and 12 grains of ignited and finely pulverised sodium chloride (the latter is added in order to moderate the combustion), and this mixture is heated in a platinum crucible. After cooling, the mass is extracted by water, and either a standard solu- tion of sulphuric acid or oxalic acid is added to the solu- tion. The sulphuric acid is prepared by mixing 70 grains of sulphuric acid, sp. r gr. 1*84, with 600 grains of water, and to this mixture so much water is added again that 100 measures of it will saturate 6-487 grains of potassium carbonate. The number of measures used for saturation ASSAY OF SALTPETRE. 179 will then indicate directly the percentage of potassium carbonate. The following foreign substances in raw saltpetre should be estimated. Water. 150 to 300 grains of air-dried, '.'finely pul- verised saltpetre are heated in a porcelain crucible to 120 C. 5 and the resulting loss is calculated as water. Mechanically mixed Impurities. The substance ob- tained in the former assay is dissolved in hot water, and filtered through a dried and weighed filter. The residue is well washed with hot water, dried on the filter at 250 F., and weighed. On deducting the weight of the filter, there will be left the weight of the mechanically mixed impurities (alumina, silica, calcium carbonate, iron peroxide, &c.), which usually amount to 2 to 5 per cent. Lime and Magnesia. These substances are precipitated as carbonates by sodium carbonate, in the former filtered solution raised to the boiling-point ; the carbonates are then dissolved in hydrochloric acid, and neutralised with ammonia. The lime can be precipitated by oxalic acid, and filtered off; the magnesia which remains in solution may then be precipitated by sodium phosphate. The amount of lime in East Indian raw saltpetre which has been once crystallised varies between 0*21 and 0-26 per cent., the amount of magnesia between 0-26 and 0'28 per cent. Chlorine. Thirty or 40 grains of raw saltpetre are dis- solved in about an ounce of pure warm water in a flask furnished with a tight-fitting stopper, and the amount of chlorine is estimated by a standard solution of nitrate of silver. The solution, after being warmed and acidulated with nitric acid, is mixed gradually with the solution of silver ; after each addition of the latter it is to be shaken and then allowed to rest. (For particulars see the chapter on the assay of silver in the wet way.) The amount of chlorine found by this assay is cal- culated as being derived from f potassium chloride and ^ sodium chloride, so that 1 part of chlorine corresponds to 1-927 part of metal (1-285 potassium, 0-642 sodium). N 2 180 ASSAY OF SALTPETRE. Experience has proved that East Indian saltpetre contains potassium and sodium chlorides in these proportions. Sulphuric Acid. 100 to 120 grains of raw saltpetre are dissolved in six ounces of water, and from this solu- tion, heated to the boiling-point, the sulphuric acid is precipitated by means of a solution of barium nitrate. The precipitated barium sulphate is filtered off, washed , ignited, and weighed. The amount of sulphuric acid in East Indian raw saltpetre varies between 0-05 and (Ml per cent. Sodium Nitrate. This estimation is most difficult, arid the following modes are recommended. a. Longchamps's mode is based upon the decomposition of soda-saltpetre by potassium chloride, producing sodium chloride and potassium nitrate. The saltpetre is mixed with potassium chloride, and the solution evaporated down. By this operation, first sodium chloride and after- wards saltpetre become separated. The latter is washed, dried at 150 C., and weighed. Werther has recommended a similar mode. 3. If the saltpetre does not contain certain oxides, such as alumina, lime, &c. (or if, previously present, they have been precipitated), a solution of potassium antimo- niate will precipitate the soda contained in the saltpetre solution. The precipitate consists of sodium antimoniate, 100 parts of which contain 84*39 antimonious acid, and 15-61 soda. 7. The presence of soda is also to be ascertained by washing saltpetre with a saturated solution of pure potash- saltpetre. This saturated solution will then contain a pro- portionally large amount of sodium nitrate. If a small quantity of the solution is made to crystallise upon a watch-glass, soda-saltpetre, showing a rhombohedricform, may be detected by means of a microscope, while potash- saltpetre crystallises in prisms, and sodium and potassium chlorides in cubes arranged in the form of steps. LEAD NITRATE acts in a similar way to the two last- mentioned salts. It is prepared by dissolving litharge in nitric acid, and crystallising the solution. DESULPHURISING REAGENTS. 181 MANGANESE PEROXIDE is easily reduced to the state of protoxide by many metals, and is a very powerful oxi- dising agent ; but is rarely employed, because its com- pounds are very infusible. It is employed occasionally in the purification of gold and silver. COPPER OXIDE is not much employed as a Hux, but is often contained in substances submitted to assay ; it then acts as an oxidising agent. A great number of metals, ven silver, reduce it to the minimum of oxidation ; and other metals, as iron, for instance, totally reduce it. IRON PEROXIDE. This, like copper oxide, sometimes acts incidentally as an oxidising agent. THE CAUSTIC ALKALIES, POTASH AND SODA, fuse below a red heat, and volatilise sensibly at a higher temperature. Charcoal, at a high temperature, decomposes the water combined with the hydrates of potash and soda, convert- ing them into carbonates, but an excess at a white heat decomposes the carbonate, and potassium or sodium is the product. POTASSIUM AND SODIUM CARBONATES are very much em- ployed as agents in the assay by the dry way. They have the power of oxidising many metals, as iron, zinc, and tin, by the action of the carbonic acid they contain ; part of it being decomposed, with the formation of carbonic oxide. LEAD, COPPER, AND IRON SULPHATES. These three salts at a high temperature oxidise the greater number of the metals, even silver, the sulphuric acid giving off oxygen and sulphurous acid. They are used in the assay of gold. SODIUM SULPHATE is not used by itself as a reagent, but is a product in many operations ; it is either formed in the course of an assay, or is contained as an impurity in some of the bodies used. III. DESULPHURISING EEAGENTS. 1. The oxygen of the atmosphere. 2. Charcoal. 3. Metallic iron. 4. Litharge. 5. The caustic alkalies. 182 LITHARGE. 6. The alkaline carbonates. 7. Nitre. 8. Lead nitrate. 9. Lead sulphate. 1. THE OXYGEN OF THE ATMOSPHERE acts as a desulphuris- ing agent in roasting, combining with the sulphur present,, forming sulphurous acid or sulphuric acid, sometimes both. 2. CHARCOAL decomposes many sulphides by taking their sulphur to form sulphide of carbon. It acts in this manner with the mercury, antimony, and zinc sulphides. It is only employed as an auxiliary to the desulphurising power of the alkalies and their carbonates. 3. IRON separates sulphur from lead, silver, mercury, bismuth, zinc, antimony, and tin, but only partially decom- poses copper sulphide. It is generally used in the state of filings, or nails ; the latter are preferable, and ought to be kept free from rust. Oxide of iron may be used if it be mixed with the requisite quantity of charcoal to reduce it. Cast iron must not be employed, as it has very little affinity for sulphur. 4. LITHARGE exercises a very energetic action on sul- phides, even at a low temperature. If it be employed in sufficient proportion, the sulphide acted on is wholly decomposed. The sulphur is often disengaged as sulphur- ous acid, and the metal remains alloyed with the lead proceeding from the reduction of a portion of the litharge, or combines as oxide with that portion of the litharge which is not reduced. The quantity of litharge requisite for the decomposition of a sulphide is considerable, and varies according to its nature ; some sulphides require 34 times their weight. When less than the requisite quantity is used, only a portion of the sulphide is decom- posed, and a corresponding quantity only of lead reduced, whilst the remainder of the sulphide forms, with the litharge and the metallic oxide which can be produced, a compound belonging to the class of oxysulphides, which is generally very fusible. ACTION OF LITHARGE ON SULPHIDES. 183 When the sulphides have a very strong base, as an alkali or alkaline earth, no sulphurous acid is given off by the action of litharge, but all the sulphur is converted into sulphuric acid. Litharge is a very valuable reagent, and its use is nearly exclusively confined to the assay of sulphides containing the noble metals, as these metals are thus obtained as alloys of lead, which are afterwards assayed by cupellation. The following is an account of the behaviour of this reagent with the ordinary sulphides. Manganese Sulphide requires at least six times its weight of litharge to produce a fusible compound, and thirty times its weight to desulphurise it completely. The sulphur and metal oxidise simultaneously, and a manganese protoxide is formed, which partly peroxidises, taking a brownish tint in contact with the atmosphere. Berthier assayed the four following mixtures : Manganese sulphide . . 5 5 5 5 Litharge . ... 20 30 100 150 The first produced an infusible, greyish-black, scori- forin mass, in which small plates, having the look of galena, could be discovered. It was composed of the sulphides and oxides of manganese and lead. Much sulphurous acid was given off during the operation. The second fused to a soft paste, and gave 17' 5 of lead, and a compact, vitreous, opaque slag, of a very deep brown colour. The slag contained about half its weight of manganese sulphide. The third fused readily, and produced 31-5 of ductile lead, and a transparent, vitreous slag, of a deep hyacinth- red. The fourth produced 33'T of lead, exceedingly ductile, and the desulphurisation was complete. Iron Sulphide. -Thirty; parts of litharge are sufficient to scorify iron protosulphide ; the metal is converted into the protoxide. 184 ACTION OF LITHARGE ON SULPHIDES. The four following mixtures Iron protosulphide ... 10 10 10 10 Litharge 60 125 250 300 gave, the first a pasty, scoriform mass, colour metallic grey, and very magnetic. It was composed of the sul- phides and protoxides of iron and lead. The second gave a very fluid metallic black slag, very magnetic, opaque, and possessing great lustre, and 36 of lead. The third gave a compact vitreous transparent slag of a fine resin-red, and 67 of lead. The last yielded a similar slag to the former, but con- taining no sulphur, and 70 of lead. Native iron pyrites was treated with the following proportions of litharge : Iron pyrites . . 10 10 10 10 10 10 Litharge. . . 60 125 200 300 400 500 The mixtures fused very readily with an abundant disengagement of sulphurous acid. The first produced only a metallic button, divisible into two parts : the lower was the larger, and was a lead subsulphide ; the other looked like compact galena, but was magnetic ; it was composed essentially of the sulphides of iron and lead, but probably contained a small quantity of their oxides. The second and third gave black vitreous opaque slags which stained the crucibles brown, together with lead, having a granular fracture and a deep grey colour ; the first button weighed 35, and the second 40. Both samples of lead were contaminated with a small quantity of slag, and contained from y^Vo ^ Tiro f sulphur, and a small quantity of iron. The slags from the last three mixtures were vitreous, transparent, and of a fine resin-red colour : the buttons of lead equalled 45-4, 54*8, and 86 parts. A much larger proportion of litharge does not produce more than 86 of lead, proving that 50 parts of litharge completely effect the desulphurisation of iron pyrites. ACTION OP LITHARGE OJS T SULPHIDES. 185 Copper Sulphide. The following mixtures of sulphide of copper and litharge Copper sulphide . .10 10 10 10 10 Litharge .... 20 30 50 100 250 fuse very readily, giving off an abundance of sulphurous acid. The slags formed were compact, vitreous, opaque, or translucid, and more or less bright red. The copper which they contained was at the minimum of oxidation. The first three mixtures gave metallic buttons com- posed of uncombined lead and sulphide of copper. The fourth gave 28 of lead, with a little adhering sulphide of copper. The fifth gave 38*5 of pure ductile lead, the exact quantity that ought to be reduced from litharge by the transformation of the above quantity of copper sulphide into suboxide and sulphurous acid. Copper sulphide does not combine with litharge ; this is an exception to the general rule. It requires about twenty-five times its weight of litharge to decompose it completely. When litharge is combined with a certain quantity of copper protoxide, it -has no action on the sulphide of that metal. The desulphurisation of copper pyrites requires about 30 parts of litharge. Copper pyrites .... 10 10 10 10 Litharge ..... 50 100 200 300 were fused together. In the first assay the fusion was accompanied with much ebullition, and the mass remained pasty : 6 parts of ductile lead were produced, and a matte similar to galena, but deep grey, with small facets, and a brownish-black vitreous slag. In the second, much ebullition and swelling up took place : 35 of lead, 45 of matte, and a deep brown vitreous slag were produced. In the third assay, 49 of lead was the result. It was covered by a thin layer of matte, and a very shining, dee]) brown, vitreous, translucid slag. 18G ACTION OF LITHARGE ON SULPHIDES. The last mixture fused readily, almost without ebul- lition, and gave 72 of lead, and a compact shining slag, of a bright grey, and without the least trace of matte ; the desulphurisation was complete. Antimony Sulphide has a great tendency to combine with litharge, and it must be heated with at least 25 parts to effect its desulphurisation. By mixing these two sub- stances in the following proportions Antimony sulphide .10 10 10 10 10 Litharge . . .38 60 100 140 250 the first three mixtures afforded very fluid slags, compact, deep black, and slightly metallic, and. buttons of ductile lead, weighing 2, 9, and 26 parts. These slags resemble the black litharge produced at the commencement of a cupellation. The fourth mixture gave a transparent compact slag, vitreous and shining, having a splendid hyacinth-red colour, and 50 parts of lead. The last produced 57 of lead, proving the desulphuri- sation to be complete. The antimony, in this case, exists as protoxide in the slag. M. Fournet has observed that antimony sulphide has the property of carrying copper sulphide, and even silver sulphide, into the compounds formed with litharge. In one of the experiments which he made, a double sulphide, composed of equal parts of silver sulphide and antimony sulphide, was fused with three times its weight of litharge, and gave, first, a button of lead mixed with silver ; secondly, a matte like galena ; and thirdly, a black slag. This slag was analysed, and found to contain from 8 to 9 per cent, of silver. It is probable that all the sulphides, having a strong tendency to combine with lead oxide, have, like antimony sulphide, the property of determining the scorification of a certain quantity of silver sulphide ; like all the sul- phides, which, in a state of purity, are completely decom- posed by lead oxide. Zinc Sulphide must be fused with twenty-five times its ACTION OF LITHARGE OX SULPHIDES. 187 weight of litharge to be decomposed. The following mix- tures were heated together : Blende .... 24-08 12-08 10 10 Litharge ... . 55-78 83-68 100 250 However strongly the first mixture was heated, it always remained pasty ; 29*2 of a greyish-black lead were produced, which contained -018 of sulphur and -008 of zinc. The button was covered by a metallic-looking black substance, intermediate between a matte and a slag ; it was composed of zinc and lead sulphides and oxides. The second mixture gave 35 -5 of lead and a fluid slag, which was compact, opaque, and black. The third gave 43 of lead, and a deep grey slag. The last produced 65 of pure lead, and a vitreous slag, of an olive colour, and translucid on the edges. Lead Sulphide. Galena and litharge, at a heat just sufficient to fuse them, combine and form an oxysulphide; but if the temperature be increased, the two bodies react on each other, and are mutually decomposed. If 2,789 parts of litharge be employed to 1,496 of lead, or 1,865 of litharge to 1,000 of galena, nothing but pure lead is obtained. If more litharge be employed, a portion is not decomposed, and covers the lead. If less be employed, the galena is not completely decomposed, and the lead is covered by a matte of subsulphide. But when litharge is combined with a certain prp- portion of sulphides or metallic oxides, it completely loses its oxidising power on galena, even at a white heat ; so that it can be combined with this substance as with the other sulphides, without effecting its total decomposition. 5, 6. CAUSTIC ALKALIES AND THEIR CARBONATES. All the sulphides are decomposed by caustic alkalies and their caroonates; but in the latter case carbonaceous matter must be present. In the absence of charcoal there are some sulphides, as of copper, on which they have no action. In these decompositions alkaline sulphides are formed, and combine with and retain a certain quantity of the sulphide submitted to experiment. The proportion 188 CAUSTIC ALKALIES AND THEIR CARBONATES. of the sulphide which remains in combination with the alkaline sulphides depends on many circumstances. It is always less when a large proportion of alkali or carbonate has been employed, as it is also when a high degree of temperature has been employed ; and the presence of charcoal always much diminishes the proportion. When the metal of a sulphide is very volatile, as mercury or zinc, the decomposition may be perfect- Potash, as it is sold in commerce, generally contains foreign substances, viz. silica, peroxide of iron, potassium sulphate, chloride, phosphate, and silicate, soda-salts, &c., and also water. A partial purification of the potash may be effected by dissolving it in water, which will not dissolve some of the above-named foreign substances. Soda also is never free from foreign substances. Ammonium carbonate is used for decomposing metallic sulphates which are formed during the roasting process of several sulphur minerals. Ammonium sulphate is then formed, which is volatile when slightly heated. 7. NITRE, SALTPETRE, OR POTASSIUM NITRATE has a very powerful action on the sulphides ; in fact, if not modified by the addition of some inert substance, as an alkaline carbonate or sulphate, explosion may take place, and a portion of the contents of the crucible be thrown out. Where an excess of nitre is used, all the sulphur is con- verted into sulphuric acid, and every metal but gold and silver is oxidised. When only the exact quantity of nitre is employed that is to say, just as much as is sufficient to burn all the sulphur in the sulphides of those metals which are not very oxidisable, as those of copper, silver, and lead, the metal is obtained in a state of purity, and the whole of the sulphur converted into sulphuric acid ; but with the sulphides of the very oxidisable metals the oxygen of the nitre is divided between the sulphur and the metal. 8. LEAD NITRATE possesses the combined properties of nitre and litharge. It is not much used. 9. LEAD SULPHATE is not used as a reagent, but is often SULPHURISING REAGENTS. 189 formed in the assay of lead ores. It decomposes lead sulphide by burning the sulphur. It acts on many other sulphides in a similar manner. IV. SULPHURISING REAGENTS. 1. Sulphur. 2. Cinnabar, or mercury sulphide. 3. Galena. 4. Antimony sulphide. 5. Iron pyrites. 6. The alkaline persulphides. 1. SULPHUR fuses at 226, and at 284 is very liquid. It has very powerful affinities, and combines with the greater number of the metals. That kind generally known as flowers of sulphur ought to be employed ; and before use, the presence or absence of earthy matters should be ascertained by exposing it to a dull red heat in a crucible. The sulphur will go off, and the earthy impurities will be left behind. Sulphur is principally used in the preparation of the alkaline sulphides and in the assay of some of the noble metals. 2. CINNABAR is decomposed by many of the metals, and is a better sulphurising agent than sulphur itself, as it is less volatile. 3. GALENA. Many metals, as iron, copper, &c., separate sulphur from lead, while some others, as silver, gold, &c., do not ; so that if galena be heated with an alloy of various metals, some of which decompose it, and some do not, the former are transformed into sulphides, and the latter com- bine with the metallic lead which is produced. It is often employed for this purpose. It is a common ore, and readily procured. The samples employed must contain no antimony sul- phide, and alHhe matrix must be carefully separated by sifting and washing. 4. ANTIMONY SULPHIDE yields its sulphur to many of the metals, but it is only used in the separation of gold 190 SULPHURISING REAGENTS. from some alloys. In this operation the sulphur combines with the alloyed metals, and the antimony with the gold, for which it has much affinity. ; 5. IRON PYRITES is a persulphide which loses half its sulphur at a white heat. It is much employed in metal- lurgical operations, but not in assaying. 6. ALKALINE PERSULPHIDES can support a tolerably elevated temperature without losing sulphur, but they have a great tendency to do so, and to this is due their sulphurising power. By their means almost every metal can be made to combine with sulphur. When an alkaline persulphide is heated with a metal, or an oxide of a metal mixed with charcoal, a fused compound, a mixture of the sulphide of the metal and an alkaline sulphide, is obtained. When they are in combination they are held together by very feeble affinities, and their decomposition is gener- ally eifected by the mere action ; of water, which dissolves the alkaline sulphide and leaves the other perfectly pure. But with gold, molybdenum, tungsten, antimony, &c., the compound is stable and soluble in water ; and it is from this fact that the alkaline sulphides are sometimes em- ployed in the assay of auriferous substances. In order to effect a sulphurisation by means of the alkaline sulphides, it is much better to use equivalent mixtures of sulphur and alkaline carbonates than to pre- pare them beforehand. To obtain potassium persulghide, 46 parts of potassium carbonate, and 54 of flowers of sulphur, must be fused together ; and for sodium persul- phide, 40 parts of dry sodium carbonate must be heated with 60 parts of sulphur. > When the mixture is fused in a plain crucible, potas- sium sulphate or sodium sulphate is formed, because part of the alkali gives up its oxygen to a portion of the sulphur, which becomes sulphuric acid ; but when char- coal-lined crucibles are used, the carbon combines with the oxygen of the alkali, and no sulphate is produced. FLUXES. 191 V. FLUXES. Fluxes are used in the following cases : 1st. To cause the fusion of a body either difficultly fusible or infusible by itself. 2ndly. To fuse foreign substances mixed with a metal, in order to allow the latter to separate by its difference of specific gravity. ordly. To destroy a compound into which an oxide enters, and which prevents the oxide being reduced by charcoal. Zinc silicate, for instance, yields no metallic zinc with charcoal, unless it be mixed with a flux capable of combining with the silica. 4thly. To prevent the formation of alloys of some metals with others, as, for instance, in the case of a mix- ture of the manganese and iron oxides ; when a suitable flux is employed, the iron is obtained in a state of purity, whereas if no flux had been added an alloy would have been obtained. Gold and silver can be separated from many other metals by means of a flux. 5thly. To scorify some of the metals contained in the substance to be assayed, and obtain the others alloyed with a metal contained in the flux, as gold or silver with lead. 6thly. A flux may be employed to obtain a single button of metal, which otherwise would be obtained in globules. Fluxes are divided into non-metallic and metallic ; the non-metallic fluxes are 1. Silica. 2. Lime. 3. Magnesia. 4. Alumina. 5. Calcium and aluminium silicates. 6. Glass. 7. Borax (sodium biborate). 8. Fluor-spar (calcium fluoride). 9. Potassium carbonate. 192 SILICA, LIME. 10. Sodium carbonate. 11. Nitre (potassium nitrate). 12. Common salt (sodium chloride). 13. Black flux and its equivalents. 14. Argol (potassium bitartrate). 15. Salt of sorrel (potassium binoxalate). 16. Soap. The metallic fluxes are 17. Litharge (lead oxide) and ceruse (lead carbonate). 18. Glass of lead (lead silicate). 19. Lead borate. 20. Lead sulphate. 21. Copper oxide. 22. Iron oxides. 1. SILICA is employed frequently to cause the fusion of some gangues in assays made at an elevated temperature. Silica combines with all the bases, and forms with them silicates, which are more or less fusible. Quartz is the beet form of silica to use. For that pur- pose it must be strongly heated, and then quenched in cold water. It can then be easily pulverised. In case the quartz takes a yellow or reddish colour on ignition, it must be digested with hydrochloric acid. 2, 3, 4, 5. LIME, MAGNESIA, ALUMINA, AND THEIR SILICATES. No simple silicate is readily fusible, so that lime, magnesia, or alumina are employed, according to circumstances, to reduce a simple silicate to such a condition that it will readily fuse in an assay furnace. Sometimes it may be requisite to use all the above-mentioned earths. Pure lime, when exposed to atmospheric air, attracts carbonic acid and water so quickly that, in practice, pure calcium carbonate is used in the form of chalk, calcareous spar, or marble, if they are pure. Calcium carbonate fre- quently contains foreign substances, viz. iron, manganese, alumina, silica, and also magnesium carbonate. A certain quantity of magnesium carbonate is, in many cases, advan- tageous, and alumina and silica are not disadvantageous. GLASS, BORAX. 193 Alumina is never used. in the pure state. Washed china-clay which, on burning, becomes white, is used instead. Clay generally contains from 20 to about 40 per cent, alumina, and, if it is used for the formation of sili- cates, a quantitative analysis of its components should first be performed. 6. GLASS is a very useful flux in certain assays, and, being a saturated silicate, it will serve by itself either as a slag or merely as a covering. The kind employed must contain no easily reducible metallic oxides, and it must especially be free from arsenious acid and lead oxide. The subjoined analyses of glass from Bodemann KerVs ProUerkunst will be found useful. (See p. 194.) 7. BORAX. That kind with 10 atoms .or 47*1 per cent, water effloresces when exposed to atmospheric air ; and the other kind, with 5 atoms or 30 per cent, water, does not effloresce, and crystallises in octahedrons. This difference is immaterial for assaying purposes, but it is of importance in purchasing borax. When borax is heated, it loses its water of crystalli- sation and undergoes an enormous increase of volume ; at a higher temperature it fuses and forms a transparent glass, which becomes dull on the surface by exposure to air. Only the fused vitrified borax ought to be used in assays. It must be reduced to powder, and kept in well- closed vessels. Borax may be regarded as containing free boracic acid ; it is an excellent and nearly universal flux : it has the property of forming, like boracic acid, fusible com- pounds with silica and nearly all the bases, and is prefer- able to boracic acid because it is much less volatile. It may be used at a high or a low temperature. It is employed in the assay of gold and silver, because it fuses and com- bines with most metallic oxides, or in obtaining a regulus that is to say, to separate the metals, their arsenides and sulphides, from any stony matter with which they may be mixed ; because this salt is neither oxidising nor desul- phurising. It is also employed in the assay of iron and tin ores, as in the presence of charcoal it retains but traces o 194 ANALYSES OP GLASS. CQ ft H FH ^5 P^3 . W e p fe o QQ w S *? 9 9 T 1 XO "tf O rH rH C- O rH . CO :6 . : | co 6 jt $ O o 6 . . C5 C5 . .0 : : TH o : ' o cT fl rH O 00 CM CO OCMrH OTH 03 i 6 CM 10 Tin 00 CO . t- CO CM 6 CM rH CM C CM CO rH CO rH T^ rH O rH <3 rH o .CM" .cp . rH - O . CM :6 % . . CM . : CM : O O OOCM , . . . . O CM : CM 6 rH I 1 CM CO CO ^ (jq (jq rH rH O C5 t- O cp Oi t>- . rH rH CM rH ^ CO rH CO t- rH CO TH CO rH 00 O o t* rH t" 00 rH rH rH 1 : cb : rH rH 00 CM CO O t- rH : CM CO t^ . rH O O . : cb ob cb " CM 9 CO s 00 CM O O CO <* rH t- (M O t^ CO rH CM X 8 00 C rt< . O O t- O O t- h> '. O OD rH CM >b CM rH rH T 1 cO o t- o rH O Tt< CM rH rH C5 2o? rH rH $ S OOrH CM I-H . . . O CO t- Ci : : : CO rH rH CO o" S So o cp W3 O CO.C- 99 t- ft CO rHO CM gcOO^t-OCO t>- 1> t- t- CO t> L- L C~ CO . Si^^^^g i L y wine (Warrington) -j ^-y-^ I 'Q by acids (Struve) (Peligot) inalysis (Rowneyand f 0) * If ii is r ^. 11 11 s^ ories in Paris (Salvetat) . . ) ble wine-glass ss (Peli decomp white hollow plate (mirror) gl glass goblet, not bes for organic on mbus Otto Gla goblet from Petersburg, of which 8 n extracted by muriatic acid at the French hollow wine-glass ... Frenchglasa for the apparatus of the labor gla las las bee l French window French mirror-g g irr ir English m CMrj5 rH rH CO COt-OO rH rH rH . CM ' FLUXES FLUOR-SPAR, ALKALINE CARBONATES. 19o of their oxides, and, indeed, much less than generally remains with the silicates. 8. FLUOR-SPAR or CALCIUM FLUORIDE is rarely employed in assays, but in certain cases is an excellent flux, as will be hereafter shown. 9, 10. POTASSIUM CARBONATE and SODIUM CARBONATE. It has been already shown that they possess oxidising and desulphurising power ; they will now be considered as fluxes. They are decomposed in the dry way by silica and the silicates, with the separation of carbonic acid. The presence of charcoal much facilitates this decomposition. They form fusible compounds with many metallic oxides. In these combinations the oxide replaces a cer- tain quantity of carbonic acid ; but they are not stable they are decomposed by carbon, which reduces the oxides, or by water, which dissolves the alkali. On account of their great fusibility, the alkaline car- bonates can retain in suspension, without losing their fluidity, a large proportion of pulverised infusible sub- stances, as an earth, charcoal, &c. The alkaline carbonates ought to be deprived of their water of crystallisation for assaying purposes ; in fact, it would be better to fuse them before use. They must in all cases be kept in well-stopped vessels. They may be used indifferently, but sodium carbonate is to be preferred, as it does not deliquesce, and is gener- ally much cheaper. The alkaline carbonates of commerce always contain sulphates and chlorides. In some cases this causes no in- convenience, but there are many circumstances in which the presence of sulphuric acid would be injurious. Potassium carbonate can readily be procured free from sulphate and chloride by means of pure nitre and charcoal, as follows : Pulverise, roughly, 6 parts of pure nitre, and mix with 1 part of charcoal ; then project the mixture, spoonful by spoonful, into a red-hot iron crucible. The projection of each spoonful is accompanied by a vivid deflagration, and potassium carbonate is found in a fused o 2 190 FLUXES XITRE, COMMOX SALT. state at the bottom of the crucible. It must be pulverised, separated from excess of charcoal, and kept in a dry state for use. Sodium carbonate may be obtained in much the same way, substituting sodium nitrate for potassium nitrate. Either carbonate may also be obtained in a sufficient state of purity by repeatedly crystallising the commercial car- bonates. 11. POTASSIUM NITRATE. Its properties have already been pointed out. The presence of silica or of silicates much assists its decomposition. 12. COMMON SALT (SODIUM CHLORIDE) is recommended either mixed with flux, or placed above it, for the purpose of preserving the substance beneath from the action of the atmosphere, or to moderate the action of such bodies as cause much ebullition. It is very useful in lead assays, and is much used in the assay of silver by the wet way. It must be previously pounded, and heated to dull redness in a crucible, to prevent its decrepitation. Common salt, though containing calcium and magne- sium sulphates and chlorides, is in most cases sufficiently pure for assaying purposes. If intended for copper assays it must be previously purified from sulphates. Plattner * has examined the influence of common salt upon different oxides and sulphates. It does not act upon uncombined lead and zinc oxides. Lead sulphate, when melted with it at a dull red heat, becomes liquid, and evolves vapour of lead chloride. By raising the tempera- ture, and by giving more draught of air, the evolution of such vapour is increased. Common salt acts upon zinc sulphate in the same way. Antimony oxide and antimo- nious acid heated with it at a dull red heat evolve vapour of antimony chloride, though not in a great quantity. Copper sulphate melted with salt at a red heat becomes converted into copper chloride and sodium sulphate. Copper chloride becomes vaporised if air is admitted, and it becomes converted into copper subchloride by raising the temperature a little, chlorine being then evolved. * >. 3 ' 80 Sodium carbonate Charcoal . Sodium carbonate u L Charcoal . *" * Sodium carbonate c/v i Sugar . 10 / Sodium carbonate 80 1 Sugar . . ^ 2 ' 8( Sodium carbonate Starch Sodium carbonate 80 ) aA Starch > ' ' ' ' ' ' 2 71< as holding the largest charcoal borer, a section and plan of which are shown in b. This large borer is employed for forming the deep holes in the charcoal used in the blowpipe furnace, and which serve to contain the clay crucibles or capsules in which the assays are fused. The blast-holes in the charcoal inside the blowpipe furnace are bored out by the gouge-shaped borer d, which also serves for making small holes or grooves in charcoal for general purposes. The smaller borer c is most useful, particularly in boring out the holes for receiving the soda paper cornets containing the assay for reduction. The saw-knife e also fits into the same handle, and is used for trimming and sawing across the charcoal pieces, having coarse saw-teeth in front, whilst the back presents a sharp knife-edge. The figures are all drawn to one-half of the real size.' Platinum. This metal is much employed as a support in cases where charcoal would be injurious by its reducing power. It is used in three forms, viz. wire, foil, and as a spoon or small capsule. Wire. A moderately strong wire of platinum, about 2 inches long, and curved at one end, is used with great advantage in many quantitative examinations. The curve serves as a support in all experiments on tests of oxidation and reduction, where alteration of colour only is to be observed. This support can be relied on, for it is totally p 2 212 CHARCOAL APPARATUS. free from the false varieties of colour which are too often perceptible when the assay rests on charcoal. In the treatment of metals, or in reduction tests, where an easily melted body is to be operated upon, charcoal must, how- ever, be used. It is necessary to have at hand several platinum wires, so as to proceed to another experiment without being obliged to forcibly remove the adhering borax glass, or to wait for its solution in hydrochloric acid, which is the better mode. If the platinum loop melts with the reagent, it must be cut away, and a new one formed. A wire can be used for a very long time, and when it becomes too short to be held between the fingers the straight end may be fastened into a cork, or a piece of glass tubing. The platinum spoon (see fig. 72) and foil are used in much the same way ; but as charcoal and the platinum wire FIG. 72. answer every purpose, it will be unnecessary to de- scribe their use further : small iron spoons of the above form are also made, and are very useful in cases where the presence of iron is not objectionable. Other instruments, as forceps, hammer, anvil, agate mortar, scissors, &c., are sufficiently familiar to every- body not to require description. Special apparatus required for any operation will be described in the course of the processes. Colonel Eoss recommends as a support in certain cases a thin rectangular slip of aluminium plate, 4 inches by 2, half an inch of the lower end of which is turned up at an angle of 80 as a rest for the assay. A fragment of the substance of the size of a pea is laid upon the edge close to the angle, and heated very slightly about half an inch from the top of a pure blue flame. The sublimate obtained should be examined from time to time, increasing the heat after each examination, till nothing more is obtained. For a full description of this method, by means of which BLOWPIPE EEAGENTS AND FLUXES. 213 substances existing in mixture or combination may be usefully separated by taking advantage of their different degrees of volatility, the reader is referred to 'Alphabetical Manual of Blowpipe Analysis ' (Triibner & Co.) Asbestos cardboard, according to Mr. W. M. Hartley, may also serve for supports. This substance resembles greyish cardboard, but has a soapy feel, like steatite ; it can be used for making crucible supports, sand-baths, muffles, retort supports, &c. ; it can be cut with cork borers or scissors ; by moistening with water it can be moulded to any shape. After moistening it should be gradually dried and ignited, to get rid of organic matter. It stands the ordinary wear and tear of the laboratory well. It is formed principally of asbestos fibres. It can be obtained from the manufactory, 31 St. Vincent Place, Glasgow, at 4s. a pound. REAGENTS AND FLUXES. BLUE LITMUS PAPER is used for detecting free acid in solution, its colour being changed to red. REDDENED LITMUS PAPER is used for the detection of free alkali, its colour being restored to blue. BRAZIL-WOOD PAPER is used for detecting hydrofluoric acid, being tinged straw-yellow when immersed in a very dilute solution of this acid. TURMERIC PAPER is used for detecting free alkalies : the change produced is very characteristic, its bright yellow colour becoming dark brown. NITRIC ACID is employed in the solution of various metals, alloys, and ores, and for the discrimination of certain precipitates ; also as an oxidising agent. ZINC is principally employed for the reduction of antimony and tin. COPPER is used for the reduction of mercurial salts and for the detection of arsenious acid. It is also used to detect the presence of nitric acid. IRON WIRE is employed to precipitate many metals, and in the separation of sulphur and the fixed acids from any 214 BLOWPIPE REAGENTS AND FLUXES. substance with which they may be combined. The metals which can thus be precipitated, or deprived of sulphur, are copper, lead, nickel, and antimony. For instance, if a small piece of iron (pianoforte) wire be placed in a sub- stance in fusion, and acted upon by the blowpipe, it becomes covered with the reduced metal. The latter sometimes appears as small globules. Iron has the property of reducing phosphorus from phosphoric acid or the phosphates, giving rise to a phos- phide of iron, which forms on fusion a white, brittle, metallic globule. POTASSIUM CYANIDE. This is a most useful flux. MM. Haidlen and Fresenius say, ' We have examined its action on many oxides, sulphides, salts, &c., in reference to its use as a reagent combined with the blowpipe. We prefer, in general, a mixture of equal parts of anhydrous soda and potassium cyanide. This mixture was employed on account of the great facility with which the pure cyanide fuses. In acts, in general, so very similarly to pure soda, that it would be superfluous to describe singly the changes which each individual body appeared to undergo when exposed to its action. We cannot, however, pass over the following especial advantages which it possesses as com- pared with soda. First, reductions are obtained with such great facility that the least practised operator may execute reductions which would otherwise be very difficult : for instance, the reduction of tin from either its oxide or sulphide ; and secondly, that the fused mixture of potas- sium cyanide with soda is so easily absorbed by the char- coal, that the grains of reduced metal can always be most distinctly perceived, and may be most easily separated therefrom for further examination. 7 SODIUM CARBONATE, called, for the sake of brevity, Soda. The plain carbonate or the bicarbonate may be in- differently employed ; but in either case it is absolutely necessary that it be free from sulphates. There are two objects in view in the employment of soda as an auxiliary to the blowpipe : first, to ascertain if the substances combining with this body be fusible or REAGENTS IN THE DRY WAY. 215 infusible ; and secondly, to facilitate the reduction of certain metallic oxides. In the fusion of substances with soda there are many things to observe. The necessary quantity must be taken on the moistened point of a knife, and kneaded in the palm of the hand, so that it may form a coherent mass. If the body under examination be pulverulent it must be incor- porated with it, but if in lump it must be placed upon it, forcing it slightly into the moistened soda ; then carefully heated on the charcoal with a gentle flame, until thoroughly dry ; and lastly it may be fused. It generally happens that the soda, at the instant of fusion, is absorbed by the charcoal : but this does not hinder its action on the assay ; for if it be fusible with soda, the latter comes to the surface and attacks it, finally forming a liquid globule. If the substance be infusible in soda, but decomposable by it, it alters its appearance without entering into fusion. But, however, before pronouncing any substance to be infusible by soda, the flux ought to be mixed with the pulverised substance. If in these trials too little soda be taken, a portion of the substance remains solid, and the rest forms a covering of transparent glass ; if too much, the bead of glass becomes opaque on cooling. - It sometimes happens that the assay contains a substance which, being insoluble in the glass of soda, prevents it becoming transparent. Then, in order that we may fall into no error respecting the nature of the glass, it becomes necessary in the two last-mentioned cases to add a new quantity of the body under examination, and then ascertain if a limpid globule cannot be obtained. In general it is the best method to add the soda by successive small doses, and note the changes produced by each addition. It sometimes happens, in this kind of assay, that the glass becomes coloured at the moment of cooling, and finally takes a yellow or deep hyacinth-red ; it even becomes occasionally opaque and yellowish-brown. These phenomena indicate the presence of sulphur, either in the assay or the soda employed. If the same colour be constantly produced by the same soda, it is a proof that it contains sulphate of soda ; it must then 2K5 SEDUCTION BEFORE THE BLOWPIPE. be discarded ; but if it gives generally a colourless glass, it is the substance under assay that contains sulphur or sul- phuric acid. Reduction of Metallic Oxides. This species of assay, by which quantities of reducible metals, so small as to escape ordinary wet analyses, can be detected, is the most important discovery Gahn made in the application of the blowpipe. If a small quantity of native or artificial oxide of tin be placed on charcoal, it requires a long blast and a skil- ful operator to produce a grain of metallic tin ; but if a small quantity of soda be added, the reduction takes place- readily, and so completely with pure oxide, that the whole is transformed into a button of tin. From this it is certain that the presence of soda favours the decom- position. The action, however, can be explained thus, as Berzelius himself hints : the red-hot charcoal reacts upon the sodium carbonate, producing by its reduction a certain amount of sodium, which by its strong attraction for oxygen seizes on that contained by the metallic oxide which is required to be reduced. If the metallic oxide contain an unreducible substance, the reduction of the former becomes difficult ; but if a little borax be added the reduction takes place as usual. This assay is very easy of execution, and the metal is readily recognised, as by previous assays the nature of it is somewhat ascertained, and the reduction but confirms the previous idea. If the metallic oxide be mixed with such a quantity of non-reducible substances that its nature cannot be ascer- tained by previous experiment, how can it be proved that a reducible metal is present ? Gahn has solved this question in a very simple manner. ' After having pulverised the substance to be assayed, it is kneaded in the palm of the hand with moistened soda, and the mixture placed on charcoal and exposed to a good reducing flame ; a little more soda is then added, and the blast recommenced. As long as any portion of the sub- REDUCTION BEFORE THE BLOWPIPE. 217 stance remains on the charcoal, soda is added in small por- tions, and the blast continued until the charcoal has ab- sorbed the whole of the mass. The first quantities of soda serve to collect the metallic particles scattered in the sub- stance to be assayed, and the final absorption of the latter completes the reduction of any that may remain in the state of oxide. ' This done, the burning charcoal is extinguished with a few drops of water ; then having cut out the part which absorbed the soda and assay, grind it to a very fine powder in an agate mortar. This powder is then washed with 'water to carry away the finest portion of the charcoal. The grinding and washing are repeated until all the char- coal is washed away. If the substance contained no metallic body, nothing will remain in the mortar after this last washing. But if it contained the smallest quantity of reducible matter, it is found at the bottom of the mortar, as small brilliant plates if it be malleable, or as a fine pow- der if it be brittle or not fusible. In either case the bottom of the mortar is covered by metallic traces, resulting from the friction of the particles of metal against its sides (pro- vided that the quantity of metal contained in the sample be not too small). The flattening of almost imperceptible globules of any malleable metal converts them into shining discs of a perceptible diameter. In this manner may be discovered by the blowpipe, in an assay of ordinary size, less than a half per cent, of tin, and even less than that of copper.' The following points in this class of assay ought to be particularly attended to. First, to produce the strongest possible flame, taking care that it covers every part of the assay < Secondly, to leave none of the metal in the charcoal, or lose the smallest quantity in the collection. Thirdly, to well grind the carbonaceous mass. Fourthly, to decant very slowly, so that only the lighter parts may be carried away by the water. Fifthly, not to judge of the result until the whole of the charcoal has been removed, for a small quantity remaining suffices to hide the metallic particles ; and, moreover, the particles of charcoal, viewed 218 BORAX, in a certain light, have themselves a metallic lustre, which will deceive an inexperienced eye. Sixthly and lastly, not to trust to the naked eye, however plain the sample may be, but always examine by the aid of a good micro- scope. The metals reducible by this process are (besides the noble metals) molybdenum, tungsten, antimony, tellurium, bismuth, tin, lead, thallium, copper, nickel, cobalt, and iron. Amongst these, antimony, bismuth, thallium, and tellurium volatilise easily when they are exposed to a strong heat. Selenium, arsenic, cadmium, zinc, and mer- cury volatilise so completely that they cannot be collected except by means of a small subliming apparatus. The reduction can always be effected the first time when the assay contains from 8 to 10 per cent, of metal ; but in proportion as the standard decreases more attention and care must be paid to the washing and recognition of the reduced metal in the mortar. A good system of prac- tice in this experiment is to employ any cupreous sub- stance, and make on it a great number of experiments, taking care to mix it each time with a substance contain- ing no copper ; thus the metallic value will diminish at each new assay, until at last no copper can be found. If the substance to be assayed contains several metals, the reduction of their oxides must be made in globo, and a metallic alloy obtained. Some, small in number, are reduced separately. For instance, copper and iron give a regulus of each metal ; copper and zinc, the first gives a regulus of copper, whilst the latter volatilises. But when the result of the operation is an alloy, recourse must be had to the reactions produced by other fluxes to ascertain its constituents. BORAX (SODIUM BIBORATE). The borax of commerce must be dissolved in hot water and recrystallised before it can be used in blowpipe analysis. Borax may be employed either in crystals, the requisite size for an assay, or in a pulverulent form ; in this case it may be taken up on the moistened point of a knife. Some operators prefer fusing the borax before use, in order to FUSION WITH BORAX. 210 V y-> , . "~,-\K\\Pv . drive off its water of crystallisation, and thus tumefaction of the crystal on charcoal. Borax is employed in the solution or fusion of a variety of substances. It is best to commence by acting upon a scale of the substance to be examined ; because if a pow- der be employed the resulting action cannot be so well ascertained. The following phenomena are to be carefully watched, for in treating any substance with borax it must be particularly noted whether the fusion takes place rapidly or otherwise ; without motion or with effervescence ; if the glass resulting from the fusion is coloured, and if that colour changes in the oxidising or reducing flame ; and, lastly, if the colour diminishes or increases on cooling, and if, under the same circumstances, it loses or retains its transparency. Some substances possess the property of forming a limpid glass with borax, which preserves its transparency on cooling, but which, if slightly heated in the exterior (oxidising) flame, becomes opaque and milk-white, or coloured when the flame strikes it in an unequal or inter- mittent manner. The alkaline earths, as yttria, glucina, zirconia ; the oxides of cerium, tantalum, titanium, &c., belong to this class. In order to be certain of this result we must assure ourselves that the glass is saturated to a certain point with either of the above class of bodies. The same thing, however, does not happen with silica, alumina, iron, manganese, &c., oxides, and the presence of silica prevents the production of this phenomenon with the earths ; so that alone they present this peculiar appearance with borax ; but when combined with silica (as natural silicates, for instance) no such effect is produced. This operation has received the name of flaming r , and any substance thus acted upon is said to become opaque by flaming. SODIUM AMMONIA-PHOSPHATE (MICROCOSMIC SALT) is obtained by dissolving 16 parts of sal-ammoniac in a very small quantity of boiling water, and mixing with it 100 parts of crystallised sodium phosphate, dissolving the whole with heat, filtering the boiling liquid ; during cooling the double 220 FUSION WITH MICROCOSMIC SALTS. salt crystallises. When microcosmic salt is not pure it forms a glass which becomes opaque by cooling. It is then necessary to dissolve it in a small quantity of w r ater and recrystallise it. It may be collected in large crystals, or in a pulverulent state. The crystals are in general of a suitable size for ordinary assays. Placed on charcoal, and submitted to the blowpipe flame, it bubbles and swells up, giving off am- monia ; that which remains after this treatment is an acid sodium phosphate, which fuses readily, and forms on cool- ing a transparent and colourless glass. As a reagent, it acts principally by its free phosphoric acid ; and if the salt be employed in preference to the acid, it is because it is less deliquescent, costs less, and passes readily into the charcoal. By means of microcosmic salt we ascertain the action of free acids on any substance we may wish to assay. The excess of acid it contains combines with all bases, and forms a class of double salts, more or less fusible, which are examined as to their transparency and colour. In consequence this flux is used more particularly in the detection of the metallic oxides, most of which impart to it very characteristic colours. This flux exercises on acids a repulsive action. Those which are volatile, sublime ; and those which are fixed remain in the mass, dividing the base with the phosphoric acid, or yielding it up entirely ; in w r hich case they are suspended in the glass without being dissolved. In this respect microcosmic salt is a good test for silicates ; for by its aid silica is liberated, and appears in the glass as a gelatinous mass. POTASSIUM NITRATE (NITRE), in long and thin crystals, is employed in hastening the oxidation of those substances which do not readily combine with oxygen in the exterior flame. It is used as follows : The point of a crystal is thrust into the fused bead ; but in order to prevent the cooling of the latter the crystal is held by a pair of pliers, so that when the bead begins to cool it may be withdrawn, the bead reheated, and the crystal employed as before, until the desired effect is produced. POTASSIUM BISULPHATE is employed in the detection of FUSION WITH BISULPHATES. 221 lithia, boracic acid, nitric acid, hydrofluoric acid, bromine, and iodine. It separates baryta and strontia from the earths and metallic oxides. SODIUM BISULPHATE. Professor J. Lawrence Smith* 1 has suggested the use of sodium bisulphate as a substitute for the potassium bisulphate in the decomposition of minerals, especially the aluminous minerals. He finds that ' while the soda salt gives a decomposition at least as complete as the potash salt, the melted mass is very soluble in water, and in the future operations of the analyses there is no embarrassment from a deposit of alum. 6 The ordinary commercial article is not sufficiently pure for use, and he prepares it from pure sodium carbo- nate, or sodium sulphate that has been purified by recrys- tallisation. In either instance pure sulphuric acid is added in excess to the salt in a large platinum capsule, and heated over a flame, until the melted mass, when taken up on the end of a glass rod, solidifies quite firmly. The mass is then allowed to cool ; moving it over the sides of the capsule will facilitate this operation. When cool it is readily detached from the capsule, is then broken up, and put into a glass stoppered bottle. In almost every instance where we have been in the habit of using potassium bisul- phate the sodium bisulphate can be substituted.' VITRIFIED BORACIC ACID is used to ascertain the pre- sence of phosphoric acid and small portions of copper in lead alloys. For quantitative analysis it is generally used to ascertain the quantity of copper contained in a lead ore, and also the amount of copper united with various metals. COBALT NITRATE in solution ought to be free from arsenic and nickel, and the solution must be moderately strong. It is used as a test for alumina, magnesia, tin, and zinc, by the blowpipe. NICKEL OXALATE is used in qualitative examinations for the detection of potash in a salt which also contains soda and lithia. COPPER OXIDE is employed to detect the presence of hydrochloric acid and chlorine. * ' American Journal of Science and Arts.' 222 AMMONIUM FLUOEIDE. SILICA is, with soda, an excellent test for the presence of sulphuric acid ; and when in combination with borax or soda, separates tin from copper. TURNER'S FLUX. This is a mixture of potassium bisul- phate and fluor-spar. It is used for producing coloured flames before the blowpipe. AMMONIUM FLUORIDE has been proposed as a blowpipe reagent by Professor N. W. Lord, of the Ohio State Uni- versity. The use of bisulphate of potassium and fluor- . spar as a reagent for developing the flame coloration of boron is well known ; but the alkali present prevents the application of the method for liberating some other bodies in the same way. Fluoride of ammonium, on the contrary, having all the value of fluor-spar as a source of fluorine, admits of much easier application, and is a most useful reagent for detecting the alkalies, boron, and other similar bodies in their mineral combinations. The method of using the reagent is simple. For testing felspar, or similar silicates,. a little of the powdered mineral is mixed with this reagent, then placed on a piece of platinum and moistened with sulphuric acid ; the mixture allowed to stand a few moments, or else gently warmed, taken upon a loop of platinum wire, and tested either in the blowpipe flame or in a Bunsen burner, being dried a little on the wire first. The alkali flame is nearly as well shown as with the pure salts. As the fluoride of ammonium is permanent, is easily obtained free from alkalies and boron, and can be kept indefinitely in a small wooden box, it is always easy to use. As a test for boron the reaction is of surprising deli- cacy. The fact that the fluoride of boron is volatile at a temperature far below that required for alkalies, permits thus its detection in borax or any alkaline compound.. To a drop of sulphuric acid placed on a platinum crucible cover, a few grains of the fluoride should be added, and then the mineral (powdered) to form a paste. This is taken as before described on a platinum loop. It should be heated gently until it stops ' sputtering ' from escape of free acid and water (but on no account heated to CALCIUM FLUORIDE. 223 redness), then brought not in, but near the lower part of the flame of a Bunsen burner or a good blowpipe flame. A bright green coloration is at once given, untinged by soda-yellow. The coloration is, of course, evanescent, and disappears before the assay is red-hot. A little prac- tice is needed to find the right part of the flame, to get the right heat, and at the same time to draw the volatile boron compound into the heated zone. This reaction shows boron very strongly in all speci- mens of tourmaline. With a hand spectroscope the appli- cation of this method gives instant proof of the existence- of boron, potassium, sodium, and lithium, even in very small traces in rocks. Borax treated in this way gives a bright green flame,, almost like copper. CALCIUM FLUORIDE (FLUOR-SPAR) AND CALCIUM SULPHATE (GYPSUM). These two bodies (deprived of water) are used to indicate the presence of each other. If a small piece of gypsum be ignited in contact with a similar piece of fluor-spar, they soon liquefy at their points of contact ; they then combine, and form, by fusing, a colourless and transparent bead of glass, which becomes enamel white on cooling. Calcium fluoride is thus employed as a test for gypsum, and vice versd. Mr. S. D. Poole prefers a mixture of calcium sulphate (selenite) and fluor-spar (in the proportions of two- parts of the former to one of the latter) to Turner's flux described above. It forms an easily fusible bead, which by itself gives only a very faint dull red tint (calcium) to- the flame, but which renders the presence of many elements which give colour most beautifully evident. Thus small portions of lepidolite, petalite, &c., mixed with this flux impart the fine carmine tint of lithium ; copper and stron- tium show their well-known colours, especially after con- tinued blowing. Potassium and sodium minerals (felspar and albite) are at once distinguished. This flux is of more limited applicability than that of Turner, because there is no provision in it for the libera- tion of hydrofluoric acid. 224 BONE ASH. It serves, when mixed with bisulphate of potash, to detect lithia and boracic acid in their various combina- tions. BONE ASHES are employed in the cupel lation of gold and silver. Harkort reduced them to many states of minute division by the processes of sifting and washing. The bones are burned until they become perfectly white, and then freed from any carbonaceous matter that may have ad- hered to them. This being done, they are pulverised in a mortar, and the finer portions separated by a sieve. The remaining powder is then thrown upon a filter, and treated with boiling water, which extracts the soluble matter. The washing, which is then resorted to, is for procuring the bone ashes of a more uniform degree of fineness. The mass from the filter is mixed with water in a cylindrical glass, allowed to settle for a few minutes, and then de- canted ; the coarser powder is deposited at the bottom of the vessel, while the finer passes over suspended in the water. By repeated decantations in this way deposits are obtained of different degrees of fineness ; the last, or that which remains longest floating through the liquid, being the finest. The resulting powders must be kept in sepa- rate bottles. The coarser ashes are used for the cupella- tion of rich silver ores, and the finer for assaying ores in which only a minute quantity of gold or silver is present. PROOF LEAD is made use of in cupelling argentiferous or auriferous substances ; it must be free from silver. Dumas states that the best method of obtaining lead in this desir- able state is to decompose the best white-lead by means of charcoal, a it is then impossible for it to contain any other metal. TINFOIL is employed to reduce certain peroxides to the state of protoxide. When it is used, a small roll, about a quarter of an inch long, is plunged into the fused button, and heated strongly in the reducing flame : the desired effect is then produced. DRY SILVER CHLORIDE. Herr H. Gericke proposes the employment of this compound in qualitative blowpipe SILVER CHLORIDE IX BLOWPIPE ASSAYS. 225 assays. In an elaborate paper on this subject, communi- cated to the 'Chemical Gazette' (vol. xiii. p. 189), he says : ' Amongst the phenomena which characterise different bodies before the blowpipe, and serve for their distinction, the colour of the flame is of no small importance. This power of colouring the blowpipe flame is not, however, exhibited by all bodies with sufficient intensity to enable them to be distinguished by it with certainty ; and certain substances are consequently usually employed, such as muriatic acid with baryta, strontia and lime, or sulphuric acid, partly to form and partly to set free volatile com- pounds. By this means, however, although the intensity of the coloration is heightened, its duration is not in- creased, as these acids, and particularly muriatic acid, evaporate for the most part before they have acted suffi- ciently, so that the coloration lasts only for a few mo- ments. This defect may be got over by the employment, instead of the volatile muriatic acid, of a chloride, which will retain the chlorine at a high temperature, so that it may only be set free by degrees in small quantities, while the body forming its base may be without action upon the colouring power of the body under investigation. For this purpose chloride of silver appears to be the best, espe- cially as it may readily be prepared in a state of purity. The best plan is to stir it with water into a thick paste, and keep it in a bottle. ' In regard to the action of chloride of silver upon the coloration of the blowpipe flame, I have investigated several compounds of potash, soda, lithia, lime, baryta, strontia, copper, molybdenum, arsenic, antimony, and lead, and mixtures of these substances. Chloride of silver, of course, has no action upon borates and phosphates, both of these acids being amongst those which offer the most resistance to the action of heat. ' For a support, I employed first of all platinum wire, but this is soon alloyed by the metallic silver which sepa- rates, and thus rendered useless in testing metals. Silver wire is too readily fusible, and also difficult to obtain free Q 226 USE OF CHLORIDE OF SILVER from copper, which may give rise to errors when in con- tact with chloride of silver. For these reasons, iron wire is best fitted for experiments with chloride of silver, as from its cheapness a new piece may be employed for each experiment, while the silver may readily be obtained in the form of chloride from the broken pieces, If the size of the fragment under examination be sufficient, the plati- num forceps may be employed. ' The results at which I arrived, by the employment of chloride of silver, in comparison with those obtained without this reagent, are as follows : ' With potash compounds, such as saltpetre, potashes, &c., the flame is decidedly of a darker colour with chloride of silver ; and even in ferrocyanide of potassium, which, when treated by itself with the blowpipe, colours the flame blue, the addition of chloride of silver produces a distinct potash coloration. ' The action of chloride of silver upon soda salts is not so favourable ; for although with some, as nitrate of soda, common soda, and labradorite, the flame acquires a more intense yellow colour by the addition of chloride of silver, this reagent produces no observable difference with other soda compounds, such as sulphate of soda and analcime. This also applies to the compounds of lithia, some of which give a finer purple-red colour on the addition of chloride of silver, whilst upon others it has no such effect. ' With lime compounds chloride of silver acts favour- ably upon the colouring power. Thus the addition of chloride of silver to calcareous spar or gypsum (in the reduction flame) gives the flame a more distinct yellowish- red colour, but stilbite gives no coloration either with or without chloride of silver. With fluor-spar the coloration cannot well be observed, as this decrepitates too violently under the blowpipe. ' The action of chloride of silver upon compounds of baryta and strontia is decidedly advantageous, as both the intensity of the coloration and its duration leave nothing to be desired. Siliceous celestine, which, when heated by itself in the forceps, scarcely coloured the flame, imme- IN BLOWPIPE ASSAYS. 227 diately produced a permanent red coloration, when heated with chloride of silver. 4 Although it appears from the preceding statements, that the employment of chloride of silver presents no advantage with some substances, it may be used with good results in the treatment of mixtures of alkalies and earths. 4 Thus, with petalite alone the lithia coloration is first produced, and a slight soda coloration is afterwards obtained ; whilst with chloride of silver the soda coloration appears very distinctly after that of the lithia. With lithion-mica alone a very distinct lithia coloration is presented ; but in the presence of chloride of silver a colour is first produced which may lead to the conclusion that potash is present, but the lithia coloration is weak- ened. Ryacolite, heated by itself in the blowpipe flame, only gives a distinct soda coloration ; but with chloride of silver a slight potash coloration is first produced, and the colour of soda then appears very distinctly ; the lime contained in it cannot, however, be detected by the colora- tion of the flame. 4 Chloride of silver may be employed with still greater advantage with the following metals, but in these cases it is particularly necessary that the operator should become familiar with the colour produced by each individual substance. 4 With copper compounds, such as red copper ore, malachite, copper pyrites, sulphate of copper, &c., when contained in other minerals so as to be unrecognisable by the eye, the employment of chloride of silver may be of the greatest service, as the smallest quantities of copper, when treated with chloride of silver under the blowpipe, give a continuous and beautiful blue colour to the flame. With chloride of silver the presence of copper may be distinctly ascertained by the blowpipe, even in a solution which is no longer coloured blue by the addition of ammonia. 4 The employment of chloride of silver will be equally advantageous with molybdenum, as in this case also the Q 2 228 USE OF CHLORIDE OF SILVER flame gains greatly in intensity. Arsenic, lead, and anti- mony are already sufficiently characterised, the former by its odour, the two latter by their fumes ; but even with these metals chloride of silver may be employed with ad- vantage to render their reactions still more distinct. It is only necessary to observe, that the greenish -blue flame of antimony appears greener and more like that of molyb- denum under the influence of chloride of silver. 'Chloride of silver may also be employed with com- pounds containing several of the above-mentioned metals. ' If bournonite be heated in the oxidation flame of the blowpipe, a fine blue flame is first produced, which in- dicates lead with certainty ; if chloride of silver be now applied, copper is also readily shown. The antimony contained in bournonite cannot be ascertained by the coloration of the flame ; but this may easily be detected upon charcoal, or in a glass tube open at both ends. ' Native molybdate of lead, without chloride of silver, only gives a blue colour to the blowpipe flame ; with chloride of silver this blue coloration of the lead comes out more distinctly, but at the same time the tip of the flame, particularly when the reduction flame is employed, appears of a beautiful yellowish-green colour from molyb- denum. ' With mixtures of arsenic and copper, or antimony and copper, the flame first acquires a greyish-blue or greenish-blue colour from the oxidation of the arsenic or antimony ; the copper may then be very easily detected by chloride of silver. This applies also to mixtures of arsenic and molybdenum, or antimony and molybdenum ; with .chloride of silver the yellowish-green flame of molybdenum appears distinctly. It will be more difficult to analyse mixtures of arsenic and lead, or antimony and lead, in this manner ; and if a compound contain both arsenic and antimony, these two bodies are not to be distinguished with chloride of silver under the blowpipe. ' From these experiments it appears that in blowpipe testing it is advantageous to employ chloride of silver instead of muriatic acid. IN BLOWPIPE ASSAYS. 229 ' Chloride of silver is particularly to be recommended in testing metallic alloys for copper. Thus, to test silver for copper, chloride of silver may be applied to the ends of silver wires, and on the application of heat the smallest quantity of copper will furnish the most distinct reaction. This is as sensitive as any of the known copper reactions, and may be performed quickly and easily. In testing metallic alloys for traces of copper, it may be advisable to submit those which contain antimony, zinc, lead, and other volatile metals to roasting^ so as to drive off these metals before the addition of chloride of silver.' TINCTURE OF IODINE. Messrs. Wheeler and Ludeking have given, in the ' Transactions of the Royal Society of Canada,' a series of experiments proving that tincture of iodine can be made a very valuable blowpipe reagent. The tincture of iodine is prepared by making a saturated alcoholic solution of the element, which dissolves very readily to a dark red liquid in this menstruum. The usual blowpipe charcoal support is replaced by long thin tablets of plaster-of-Paris, in order to develop the true colours of these varicoloured iodides on a white back- ground. These tablets are prepared by mixing plaster-of- Paris with water to a thin fluid paste, which is poured over a smooth flat surface (as a plate of glass) that has been previously oiled to prevent its adhesion. In a few minutes it will set into a hard cake ; but before this takes place, when it has become stiff, it is divided by a knife or spatula into pieces about 4 by 1-^ inches for use. These tablets are the supports on which these iodide reactions are made by putting the substance on one end, then moistening with the tincture, and blowing with the blue flame, when the volatilised iodides are deposited on its cold surface when suitably inclined. The oxidising flame must be employed in order to prevent the deposition of soot, which tends to interfere by its black film. Description of the Iodide Coats. Arsenic. A reddish-orange coating. Lead. A chrome-yellow coating. 230 TINCTURE OF IODINE. Tin. A brownish- orange coating. Silver. A faint greyish : yellow white cold, bright yellow when hot ; close to assay. Antimony. An orange-red coating. Mercury. A yellow and scarlet, the yellow changing completely to scarlet on standing. Selenium. A reddish-brown coating. Tellurium. A purplish-brown coating. Bismuth. A chocolate brown, fringed with red near the assay. Cobalt. A greenish brown, edged with green ; brown coat evanescent, changing into faint green, especially when breathed upon. Molybdenum. A deep ultramaririe-blue coating ; close to assay, that is, a permanent oxide. Tungsten A faint greenish blue near assay ; that is, a permanent oxide (brought out stronger by dropping more tincture on tablet after the operation). Copper. A white coating. Cadmium,. A white coating ; becomes bright golden- yellow on blowing ammonium sulphide vapours over it. Zinc. A white coating that soon disappears. As the copper, cadmium, and zinc iodides are white, the tablets should first be coated with a film of soot (obtained by holding the tablet in a smoky flame) or else charcoal should be used in order to give a black back- ground to the white coats. If to the peculiar velvety chocolate-brown coating of bismuth a drop of dilute ammonia be added, or ammonia vapour be blown over it, the brown disappears, leaving a brilliant red coating. Many of these coatings are more or less evanescent, disappearing on prolonged exposure at ordinary tem- peratures. In studying these coats it will be observed that we can now detect tin in the presence of zinc, which has hitherto been impossible with the blowpipe, and that we have a very striking and characteristic reaction for molybdenum. The other coatings are more or less characteristic indivi- dually, and will find favour as confirmatory tests. For DRY SILVER IODIDE. 231 mixtures and complicated cases the iodide reactions will not supersede the standard methods. DRY SILVER IODIDE. Mr. P. Casamajor has introduced silver iodide into blowpipe work with marked advantages. It gives characteristic sublimates of iodide very beauti- fully and quickly, and it has the advantage over tincture of iodine that it is a dry powder, easily kept in bottles which need not close very perfectly. The iodide coatings of mercury, bismuth, and lead are familiar to all chemists who use a blowpipe. They are obtained by V. Kobel's mixture of equal parts of sulphur and of potassium iodide. Silver iodide has over this mixture the advantage that there is no sulphur to give deposits when operations are carried on in glass tubes, and no fumes of sulphur dioxide. It requires less time to obtain the coatings, and they have a more distinct appear- ance. In experimenting with silver iodide, mixed with various metallic compounds, the iodide coatings are best deposited in open glass tubes of about 4 inches in length and J inch in diameter. The substance to be tested is mixed into a paste with the silver iodide. A small portion of this mixture is placed at one end of the open tube, and the blowpipe flame is blown on it for a short time. The iodide coatings immediately appear and are seen through the glass. The glass tube may be held by a tongs or simply by a piece of paper as the blowing is not suffi- ciently prolonged to heat the glass tube beyond what paper will stand. A small quantity of powdered charcoal or lampblack, mixed with silver iodide and the substance to be tested, gives the characteristic coatings more quickly than dis- tinctly. The following metals have given iodide coatings in glass tubes : Mercury. In tubes, as on other supports, the yellow and red iodides are produced simultaneously, and streaks of bright red are seen on a yellow ground. Bismuth. Yellowish red near the end of the tube and thick brown coating beyond. DRY SILVER IODIDE. Lead and Tin. Both these metals give bright yellow deposits which retain their colour when cold. These de- posits cannot be distinguished one from the other, both being equally bright. In the case of tin a very strong smell of iodine is given off, which is possibly due to stannic iodide. Arsenic. Near the end of the tube to which the flame is applied there is a yellow deposit ; beyond this a white coating of arsenious acid. The yellow portion turns white on cooling, but becomes yellow again when the tube is heated over a flame. Antimony. The orange-red coating given by this metal becomes quite faint on cooling, but the colour becomes bright again when the tube is heated. Zinc. The deposit is white both when cold and when hot ; the fumes are not very abundant, much less so than those due to lead or tin. Iron makes a deposit which may be considered as characteristic from the fact that the yellow coating in the tube is streaked with distinct dashes of brown. The yellow portion becomes white on cooling, but the brown streaks do not change. Thallium. A yellow coating is deposited as with most metals. After this has taken place, if a reducing flame touches the deposit, this fades, leaving a grey tinge with an edge of purple. This seems to be characteristic of thallium. Ccesium, Rubidium, and Lithium have not given de- posits which can be called characteristic. The deposit from caesium differs from those of the other two metals in being less volatile. The caesium deposit does not extend far beyond the heated end of the tube. By increasing the heat it melts, but does not move forward. Chromium gives a white coating which remains at the hot end of the tube. The portion nearest to this end by further heating becomes of a pale reddish brown. Manganese. Yellow hot, but white when cold, like deposits from many other metals. Molybdenum. Beyond the yellow coating, which turns SODA-PAPER. 233 white on cooling, are distinct blue streaks, which are very characteristic. These were first observed by Messrs. Wheeler and Ludeking, by treatment with tincture of iodine on tablets of plaster-of-Paris. With a glass tube, the blue streaks extend through the whole length of the tube. Manganese and Uranium give deposits which are yellow when hot, and white when cold. These are too common to be characteristic. Deposits have been obtained on charcoal and on thin sheets of iron, either on the metallic surface or on a coat of soot, by the use of silver iodide and metallic compounds. Some of these deposits are very good, but they are not so uniform for the same metal as deposits obtained in glass tubes. SODA-PAPER. Mr. Forbes writes as follows in the ' Chemical News ' : ' As it would be impossible to submit any powdered substance to the direct action of the blow- pipe flame without its suffering mechanical loss, some means must be employed for holding the particles together until they are so agglutinated by the heat that no such loss need be apprehended ; this is secured by the use of the soda-paper envelope or cornet, as devised by Harkort. For this purpose slips of thin slightly sized writing paper, about 1 J inch long by 1 inch broad, are steeped in a solution of one part crystal- lised pure carbonate of soda (free from sulphate) in two parts of water. When dried these are used for forming small cylindrical cornets, by rolling them round the ivory cylinder, fig. 73 d, previously described. A bottom is formed to them by folding down a portion of their length on to the end of the cylinder, which is then pressed firmly into the corresponding mould in the blowpipe anvil, and which, upon the withdrawal of the F IG . 73. 234 SODA-PAPER. cylinder, serves as a support until they are filled with the assay from the scoop in which the assay and flux have been mixed. After pressing the assay down, the super- fluous paper is cut off, leaving only sufficient when folded down upon the contents of the cornet, to form a paper cover to the top similar to the hollow of the cornet. The assay is then ready for placing in a bore in the charcoal, formed by the charcoal borer c, fig. 71, and is then submitted to a reducing fusion.' GENERAL ROUTINE OF BLOWPIPE OPERATIONS. Size of the Assay. The morsel operated on is suffi- ciently large when the effect of the heat and the fluxes added can be distinctly discerned. The size of the assay-piece generally recommended is much too large ; its size ought to be about that of a mustard seed ; that of the flux added, about the size of a hemp-seed. It should in general be previously reduced to fine powder. When a large piece is employed, the experiment con- sumes much more time and requires much more labour than a smaller piece. It is only in reductions that a larger piece may be successfully employed, because in that case the more metal produced, the more readily can its nature be ascertained. Having thus endeavoured to fix the size of the assay, we will now lead our readers to the operations necessary in blowpipe analysis, and in the order in which they are to be performed. First. The substance is heated in the closed tube, or mattrass, over a spirit lamp. It may, by this treatment, decrepitate or give off water, or some other volatile sub- stance. Secondly. It is heated gently on charcoal, by aid of the blowpipe ; and, as soon as warm, withdrawn from the heat, and the odour given off ascertained : volatile acids, .arsenic, selenium, or sulphur, may be present. The odour thus produced by the oxidising flame must be compared with that produced by the reducing flame ; if there is any difference, it must be carefully noted. Sulphur, selenium, OPERATIONS IN BLOWPIPE ANALYSIS. 235 <&c., are best detected in the oxidising flame, and arsenic in the reducing flame. Thirdly. The substance is examined as regards its fusibility. If it be in grains, it is better acted upon on charcoal, notwithstanding its liability to escape on the first insufflation, when not very fusible. But if we can choose the form, it is better to knock off a small splinter, by means of the hammer, and hold it in the flame by the platinum- pointed pincers. A fragment with the most pointed and the thinnest edges ought to be selected. By thus acting, we can always ascertain at a glance if the substance be fusible or not. Infusible substances retain their sharp points and angles, which can be ascertained immediately by means of a microscope. The points are merely rounded in bodies of difficult fusibility, and in substances of easy fusion are rendered globular. Certain substances, and particularly some minerals, change both aspect and form when exposed to the blowpipe flame, without entering into fusion ; some swell up like borax ; some of them fuse after tumefaction ; others keep in that state without fusion. Some minerals give off a sort of foam on fusing, giving rise to a kind of blebby glass, which, although transparent itself, does not appear so, on account of the multitude of air-bubbles it contains. This bubbling and tumefaction takes place in the greater part of the minerals only at that temperature at which all the water is disengaged ; and these ramifications appear to proceed from a new molecular arrangement, produced by the action of heat on the constituent parts of the substance. It cannot be said that the expansion of a particular part of the substance or its formation into gas, gives rise to this, because it most often happens in those substances which contain no such substance. The minerals which generally give these indications are the double silicates of lime or alkali, and alumina. It sometimes disappears after a few instants, and occasionally lasts as long as the substance is kept in fusion. In the latter case, it seems that the assay takes carbonic acid from the 236 PRELIMINARY BLOWPIPE OPERATIONS. flame, which carbonic acid is transformed by the charcoal into carbonic oxide, and it is that gas which causes the bubbles. In the employment of fluxes, it is necessary to continue the blast for a sufficiently long time, because some sub- stances appear infusible at the commencement of the oper- ation, and gradually yield to the influence of the flux, and in about two minutes enter into full fusion. The substance is best added in small quantities, and no new dose must be introduced until the former one is acted upon, so that at last the glass arrives at that degree of saturation that it can dissolve no more : it is at this particular point that the reactions are most vivid and plain. Beads of glass, not so saturated, do not give such decided indications. Occasionally, in operating with a flux in the reducing flame, it happens that the assay-bead reoxidises during the cooling of the charcoal, and thus the labour of a preceding operation is lost. In order to obviate this inconvenience, the charcoal is turned over, so that the bead may fall in a yet liquid state on some cold body, as a porcelain plate. When the colour of the bead is so intense that it ap- pears opaque, its transparency can be proved by holding it opposite to the flame of a lamp ; the reversed image of the flame can then be seen in the bead, tinged with the colour imparted to the flux by the body under experiment. The globule may also be flattened by a pair of pliers before it cools, or it may be drawn into a thin thread. In either of the last-mentioned cases its colour can readily be ascer- tained. Minerals exposed to the exterior and interior flame, either with or without fluxes, present a variety of pheno- mena, which ought to be carefully noted, and which, collectively, form the result of the assay. The smallest circumstance must not be overlooked, because it may lead us to ascertain the presence of a substance not suspected. It is necessary, in all cases, to make two assays, and com- pare the separate results ; because it sometimes happens that an apparently trivial fact had been overlooked in the first series of operations, which materially conduces to the good result of the experiment. DISCRIMINATION OF MINERALS. 237 DISCRIMINATION OF MINERALS.* Selecting those minerals which are remarkable for their wide distribution, and at the same time those which are of value, whether rare or frequent, the number which it is advisable for the explorer to know is reduced to about seventy-five, and it is only to this limited number that reference will be made in this chapter. The distinctions in minerals lie in their chemical and physical characters ; but as the appearance of a mineral may afford little or no indication of what these are, they have to be discovered by testing. These characters will be described in detail, and the modes of testing will be indicated, where necessary. The characters include crystalline form, mode of fracture, colour, lustre, transparency, taste, odour, feel, hardness, specific gravity, fusibility, and chemical com- position. Crystalline Form. The crystalline form is an important guide in many cases, but crystals are subject to so many modifications, that this character ceases to be a practical one except to those persons who are familiar with its geometrical laws. Stil] it is very useful to know a few common primary forms such as the following : CUBE. 6 sides or faces, each square (fig. 74). OCTAHEDRON. 8 sides, each an equal-sided triangle (fig. 75). DODECAHEDRON. 12 sides, each rhombic, (shaped like the diamond of a pack of cards) (fig. 76). TETRAHEDRON. 4 sides, each triangular (fig. 77). * This method of distinguishing the various minerals, ores, etc. which are likely to be met with during an exploration is partly condensed from a very valuable pamphlet by Alexander M. Thompson, D.Sc., Professor of Geology at the University of Sydney, entitled ' Guide to Mineral Explorers in Distin- guishing Minerals, Ores, and Gems.' The instructions given will not be found to involve greater difficulties than can be overcome by a little practice and perseverance. The appliances which are needed may be bought for a few shillings, and will be found too trifling in weight and bulk to incommode the traveller. 238 DISCRIMINATION OF MINERALS. EHOMBOHEDRON. 6 sides, each rhombic (fig. 78). PRISM. Any column with three or more sides ; when FIG. 74. FIG. 75. FIG. 76. FIG. 77. FIG. 80. FIG. 79. Fm. 81. placed on its base, it may stand straight or oblique ; it may terminate abruptly with a flat face, or come off to a point, blunt or sharp, like a pyramid (figs. 79, 80, 81). Mode of Fracture. The mode in which a mineral breaks when smartly struck with a hammer, or pressed with the point of a knife, is a character of importance. Many can only be broken in certain directions ; for instance, a crystal of calc-spar can only be split parallel to the faces of a rhombohedron ; many crystals break more easily in one direction than in DISCRIMINATION OF MINERALS. 239 others. Whenever a mineral breaks with a smooth, flat, even surface, it is said to exhibit cleavage. Cleavage always depends upon the crystalline form. But minerals often break in irregular directions, having no connection whatever with the crystalline form, and this kind of breaking is called fracture ; the broken surfaces are usually irregular or conchoidal (i.e. with concave and convex out- lines, like shells). Lustre. Some minerals have a brilliant lustre like that of metals ; in others the lustre resembles that of glass, silk, resin, or wax ; while others are dull, or destitute of lustre. The lustre of the diamond is called adamantine. Colour and Streak. Minerals may be colourless, white, black, or of any colour, either dull or brilliant. The same mineral may present a variety of similar tints, or even distinct colours. It often happens that a mineral, which when viewed in a solid mass possesses a distinct colour, affords a powder which has a colour different from that of the solid mass, or is even destitute of colour, that is to say, white or nearly so. The dust formed on scratching a mineral with a knife, or by a splinter of quartz, or by a diamond, is termed the ' streak ' : it has usually the same colour as the powder. The colour of a mineral and its streak may cor- respond ; or the mineral and its streak may possess diffe- rent colours ; or the mineral may be coloured, while its streak is colourless. For instance, cinnabar has both a red colour and a red streak. Specular iron has a black colour, but a red streak. Sapphire has a blue colour, but a white or colourless streak. The streak of most minerals is dull and pulverulent, but a few minerals exhibit a shining streak, like that formed on scratching a piece of lead or copper. This kind of streak is distinguished by the name of Metallic. 240 DISCRIMINATION OF MINERALS. In judging the colour of a mineral a surface quite free from tarnish should be chosen. The colour of a transparent gem can be seen to best advantage by immersing it in water, about half an inch below the surface. Hardness. This character is of great importance in distinguishing minerals ; it implies the degree of facility with which the particles may be separated by cutting or scratching. The diamond is the hardest substance known, as it will scratch all others. Talc is one of the softest minerals. Other minerals possess intermediate degrees of hardness. To express how hard any mineral is, it becomes necessary to compare it with some known standard. Ten standards of different degrees have been chosen, and are given in order in the following scale : 1. Talc. 6. Felspar. 2. Gypsum. 7. Quartz. 3. Calc Spar. 8. Topaz, 4. Fluor Spar. 9. Corundum, or Sapphire. 5. Apatite. 10. Diamond. The hardness of a mineral may often be found by drawing the point of a steel knife across it. For instance, the slightest pressure will suffice to scratch talc ; fluor-spar is not so easily scratched as calcite ; the greatest pressure is needed to scratch felspar ; and quartz does not yield to the knife at all. The hardness may also be found by scratching one mineral with another ; thus, the diamond will scratch all other minerals ; corundum scratches topaz, topaz scratches quartz, quartz scratches felspar ; and so on. If on drawing a knife across a mineral it is impressed as easily as calcite, its hardness is said to be 3. If a mineral scratches quartz, but is itself scratched by topaz, its hardness is between 7 and 8. In trying the hardness of a mineral a little judgment is necessary ; for instance, a sound portion of the mineral DISCRIMINATION OF MINERALS. 241 must be chosen ; a sharp angle used in trying to scratch : a streak of dust on scratching one mineral with another may come from the waste of either, and it cannot be determined which is the softer, until after wiping off the dust, and viewing with a lens. The use of the above scale, however, implies an ac- quaintance with the ten standard minerals. It is very desirable to have specimens of these minerals for reference. But a collection of the following common minerals, which can easily be procured, will be found quite enough for ordinary purposes viz. : calc-spar, felspar, quartz, topaz (white or yellow), and corundum or sapphire. By the test of hardness clear distinctions may be drawn between minerals which resemble each other ; for instance, iron pyrites and copper pyrites are similar in appearance, but copper pyrites can easily be scratched with a knife, whereas iron pyrites is nearly as hard as quartz and cannot be impressed with a knife at all. In determining the hardness of small stones, it is most convenient to fasten them by heating upon a stick of seal- ing wax. Specific Gravity. By specific gravity is meant the comparative weight of equal bulks. Water is taken as the standard of compari- son ; the specific gravity of a mineral is a number showing how many times it is, bulk for bulk, heavier than water. The specific gravity of water is called 1, of gold 19, implying that, if equal bulks of gold and water were taken, the gold would weigh 19 times as heavy as the water. The specific gravity of a mineral can be found by weighing it first in air in the usual manner, and then observing how much of its weight it loses, when suspended from the arm or pan of a balance, and allowed to hang freely in water. If a piece of quartz weighing 26 grains is attached by a hair or thin cotton to the scales, and weighed whilst hang- ing in water, it will be found to weigh only 16 grains ; it thus loses 10 grains, or if of its entire weight. Similarly gold would lose -fa of its weight. R 242 DISCRIMINATION OF MINERALS. Minerals differ very widely in the proportion of weight which they lose in water, but the same mineral invariably loses the same proportion ; for instance : Quartz loses -J-g- of its weight ; topaz ^g- ; sapphire ^ ; zircon i-g. ; tin ore f [}. These proportions depend upon the specific gravity of these minerals. The specific gravity of water is called 1, .of quartz 2*6, of topaz 3*5, of sapphire 4*0, of gold 19. In estimating how much weight a mineral loses in water, a very delicate balance is required when the weight in air is under 10 grains ; but for portions weighing heavier than this, a common balance turning readily to a grain may be used for practical purposes. The mineral must be sound throughout, and free from any pores or cracks, and its surface should be rubbed over with water, before im- mersing it, to prevent bubbles of air adhering, which would falsify the result. In careful trials distilled or rain water should be used. A trial .of specific gravity can have no value, unless it is made on a pure portion of a mineral, quite free from any adhering foreign matter. A rough estimate of specific gravity can be formed from the feeling of pressure in shaking any mass loosely in the palm of the hand : in this way it can be judged whether the specific gravity is high or low. EULE. The rule for finding the specific gravity is to divide the weight of the mineral in air by its loss of weight in water. Example: A piece of quartz weighed 1,398 grains in air, and 862 grains in water ; here the loss of weight is 536, and the weight in air divided by this number is 2 -6, which is the specific gravity of quartz. For separating from each other minerals of different specific gravities R. Breon proposes the use of melted mix- tures of lead and zinc chlorides in different proportions, so as to yield liquids ranging from 2' 4 to 5*0 in specific gravity. Sonstadt* has devised a very ingenious and useful method of taking specific gravities of minerals, when a rough and ready method of discriminating between bodies * ' Chemical News,' March 20, 1874, vol. xxix. p. 127. DISCRIMINATION OP MINERALS. 243 of similar appearance but of different specific gravities is all that is needed. He takes a solution of potassium iodide saturated at the common temperature, and dissolves in it as much iodide of mercury as possible. It will then dissolve more potassium iodide, then more mercuric iodide, and so forth. The iodides dissolve very slowly at the last, and as it is best not to accelerate the solution by the application of heat, considerable time must be allowed when a liquid of maximum strength is required. The solution, after filtering, is fit for use, and it may be obtained without danger of its crystallising above zero, of sp. gr. 3*085 at 12 C. This liquid is transparent, very mobile, filters easily', is of an amber colour, gives no preci- pitate on addition of water, and does not readily lose or gain water on long exposure to the atmosphere. The solution undergoes no perceptible chemical change by free exposure to the air for many months. It may be diluted to any extent, and then concentrated by heat without injury. When this liquid is diluted so that quartz will just float in it, any mineral heavier than quartz will of course sink. As many minerals closely resemble quartz in external characters, and can be distinguished more readily by their specific gravity than in any other way, this method of testing the specific gravity furnishes a very quick and certain means of discrimination. A small bottle of the liquid, containing say half an ounce, and carried in the waistcoat pocket, will enable the mineralogist to make a great many examinations in his walks as to whether the specimens examined are above or below a certain specific gravity. The smallest distinctly visible particle of a mineral is enough for an estimation. After the bottle becomes clogged with specimens, the liquid may be poured out for further use, and if the washings are added and the whole concentrated there need be no sensible waste. Mr. Hardman suggests that with this liquid it would be possible to separate completely the three constituent minerals of granite mica, felspar, and quartz weigh them, and estimate absolutely their percentage. Professor Church has used Sonstadt's solution exten- R 2 244 DISCRIMINATION OF MINERALS. sively for years in the separation of rare minerals from their siliceous gangue. He says he has often identified the minerals of a crushed fragment of rock by introducing a drop of the solution into the ' live box ' under the microscope. By using solutions of increasing strength, or adding the solid salt, the felspars, the quartz, and then some of the heavier minerals may be successively brought to the surface and thus into focus. Fusibility. Some minerals can be easily fused ; others only with difficulty, while others resist the highest heat which can be applied to them. There are such wide differences between the various degrees of fusibility of minerals, that this character helps greatly in distinguishing them. The fusibility is most readily tested by holding a small splinter of the mineral with a forceps in a candle flame, urged by the blowpipe ; or the mineral may be laid upon a piece of charcoal, and the flame directed upon it by the blow- pipe. Some minerals fly to pieces when heated, others swell up, or give off peculiar and characteristic odours. Chemical Characters. The chemical composition of a mineral is one of the most important facts to know respecting it; the various processes for estimating this form a science by them- selves ; but only the tests depending upon the use of the blowpipe will be described here. A great deal can be learnt respecting a mineral by a few simple trials with the blowpipe. The only requirements are a common blow- pipe, a candle, a forceps or pliers, a piece of No. 22 plati- num wire, two inches long, dried sodium carbonate, dried borax, and potassium cyanide. The charcoal selected for these experiments should be free from cracks and openings. By dry sodium carbonate is meant not merely dry to the touch, but quite free from water ; this may be prepared from recrystallised common washing soda, by expelling the water which it contains ; the washing soda is put in a shallow clean iron dish, and placed over a clear fire until a white dry powder is formed ; too strong a heat might DISCRIMINATION OF MINERALS. 245 fuse the dry powder. A quarter of an ounce may be kept in a well-corked bottle or tube for use. Sodium bicarbon- ate may be used instead without previous heating ; or if the bicarbonate be moderately heated, it loses weight, and becomes sodium carbonate, quite free from water like the above. The borax is to be dried in the same way ; a quarter of an ounce will be enough ; it is convenient to keep the platinum wire in the same tube. Unless these tubes are well corked these chemicals reabsorb moisture. For test- ing tin ore it is useful to have a little potassium cyanide kept in a bottle with the cork and rim well covered with melted bees'-wax ; it would otherwise liquefy, by -absorption of moisture, and become useless. It is a most dangerous poison, and the greatest caution must be observed in its use. Three kinds of effect can be produced by the blowpipe : first, it can be used as a source of heat for testing fusibility ; secondly, it can add oxygen to a mineral ; thirdly, it can take oxygen away. The three principal means of chemically testing minerals before the blowpipe are (1) with borax; (2) on charcoal, usually with the addition of sodium carbonate ; (3) by holding in the point of oxidation. '. EXPERIMENT No. 1. Many metals impart a colour to fused borax, by which their presence can be recognised. To try this experiment a bead of fused borax must first be obtained on the platinum wire. The end of the wire is bent into a loop or ring about the twelfth part of an inch in diameter. The wire is then heated in the blow- pipe flame, and dipped whilst hot into the borax : the por- tion of borax that adheres is then fused on to the wire in the blowpipe flame, and the hot wire is again dipped ; this is repeated until the loop contains a glass-like bead of fused borax. If the bead has become cloudy, the soot causing this may be burned off in the oxidising point of the flame. Having thus obtained a clear, colourless, transparent bead, -the next step is to add to it a minute portion of the mineral which is to be tested. By touching a little of the finely 240 DISCRIMINATION OF MINERALS. pulverised mineral with the borax bead, while softened by heat, enough will adhere to the bead for a first trial. The bead is then kept at a white heat in the oxidising point of the flame for a few seconds, and on removal its colour is noted, both whilst hot and when cold. If no colour is imparted a fresh trial may be made with a larger quantity of the powder ; but if the bead is opaque owing to the depth of colour, as is often the case, a fresh experiment must be made, using a still smaller quantity of the powder. The colour can only be fairly judged in a perfectly trans- parent bead. If no colour can be obtained in the oxidising flame, further experiment with the borax bead is needless ; but if a colour is obtained, it is then advisable to try the effect of the reducing flame upon the same bead. The following observations and inferences may result from this test : Colour of Bead in Showing the OXIDISING FLAME. EEDUCING FLAME. presence of Green (hot) ; Blue (cold). Red. COPPER. Blue (hot and cold). Blue. COBALT. Amethyst. Colourless MANGANESE. Green. Green. CHROMIUM. Bed or Yellow (hot). 1 Yellow or colourless (cold). f Bottle Green. IRON. Violet (hot) ; Ked-Brown (cold). Grey and Turbid, NICKEL. difficult to ob- tain. This mode of testing may often be used to prove the presence of the above-mentioned metals. It requires some practice before trustworthy results can be obtained in reducing ; the reduced bead, if brought out of the flame at a white heat into the air, may at once oxidise ; but this may be prevented by placing it inside the dark inner cone of an ordinary candle flame, and allowing it to cool partially there. EXPERIMENT No. 2. The mode of testing with sodium carbonate on charcoal is performed as follows : A sound piece of charcoal half an inch square is chosen, and a neat cavity is scooped out on its surface, into which is placed a mixture containing the pulverised mineral to be tested, with three or four parts of sodium carbonate, the whole not exceeding the bulk of a pea ; after lightly pressing DISCRIMINATION OF MINERALS. 247 the mixture into the cavity, the blowpipe flame may be cautiously applied to it ; and afterwards, when the mix- ture no longer shows a tendency to fly off, the charcoal may be advanced nearer to the blowpipe, and finally be kept at as high a temperature as possible in the reducing part of the flame. In testing for tin ore, a piece of potassium cyanide about the size of a pea may be placed upon the mixture after the first application of heat, and the further applica- tion of heat may then be continued. This treatment is designed to extract metals from minerals ; it favours in the highest degree the removal of oxygen. But like the borax test, jt is limited in its appli- cation ; it can only be used to detect certain metals. The failure of the test, in any case, must not be looked upon as a conclusive proof of the absence of the particular metal sought : for instance, copper can be easily abstracted from copper carbonate by this test, but not from copper pyrites. Still the test is a most valuable and indispensable one to the mineralogist. The test is complete when the metal is obtained as a globule, in the cavity of the charcoal ; in many cases the globule will be found surrounded with the oxide of the metal, forming an incrustation on the charcoal ; and the colour of such incrustation should be carefully noted, both at the moment of removal, from the flame, and after cooling. By pressing the globule between smooth and hard surfaces, it can be seen whether the metal is flattened out (or malleable), or crushed to pieces (brittle). The following observations and inferences may result from this test: GLOBULE. INCRUSTATION. Presence of Yellow (malleable). None. GOLD. White do. Do. SILVER. Ked do. Do. COPPER. White do. White. TIN. Do. do. Bed (hot) ; Yellow (cold). LEAD. Do. (brittle). Do. do. BISMUTH. None. Yellow (hot) ; White (cold). ZINC. White, brittle, giving off White. ANTIMONY. fumes when removed from the flame. 248 DISCRIMINATION OF MINERALS. EXPERIMENT No. 3.- In addition to these substances there are others which occur abundantly in minerals, and which may be recognised by the blowpipe with the greatest ease ; for instance, sulphur and arsenic. These may be discovered by heating a fragment of the mineral, supported on a piece of charcoal, or held in a forceps, in the oxidising point of the flame, and noticing the odour which is given off; a smell of burning sulphur indicates that the mineral contains that substance, and white fumes having a garlic odour indicate the presence of arsenic. Mercury, antimony, and other substances may escape as fumes when heated in this manner. Requirements for Testing Minerals. The requirements already mentioned are blowpipe, candle, forceps, platinum wire, dried borax, dried sodium carbonate, potassium cyanide, charcoal ; also the minerals for comparisons of hardness namely, calc-spar, felspar, quartz, topaz, sapphire. In travelling it is well to dis- pense with the grain scales and weights for taking specific gravities. It would be dangerous to attempt to travel with nitric acid. In addition to these it will be found useful to have a steel pocket-knife, one blade of which should be kept magnetised, which may be easily done by touching it occasionally with a strong magnet ; a small iron spoon for heating minerals, such as cinnabar, over a candle flame ; also a small pocket lens. DESCRIPTION OF MINERALS. QUARTZ AND THE SILICATES. Perhaps no other mineral presents such a great variety of forms and colours as quartz, and no mineral occurs in greater abundance. When pure it consists of silica only, but it is usually contaminated with other ingredients, principally alumina, iron, and clay. The impure varieties of quartz compose most of the pebbles and sand of the soil. DISCRIMINATION OF MINERALS. 249 Out, of the six hundred minerals at present known, at least two hundred and fifty contain silica in greater or less proportion, and are hence termed silicates. Silicates are indispensable in the manufacture of glass, porcelain, earthenware, and for other purposes ; but they exist in such profusion that their economic value is ex- ceedingly trifling. The great majority of the silicates are purely objects of scientific interest ; a few forms of silica &re esteemed as gems, such as the precious opal, and some varieties of quartz ; also the silicates, topaz, emerald, zircon or hyacinth, garnet and carbuncle, the tourmaline, .and some others. Handsome varieties of serpentine are often used in ornamental stonework. 1. QUARTZ. Chemical Composition. Silica. Crystallised in six-sized prisms, terminated by pyramids. Sides of the crystal often marked across with fine parallel lines. Transparent or opaque. Colourless, or of various colours. Glassy lustre. Fracture irregular, conchoidal. Specific gravity, 2'6. Hardness, 7. Cannot be scratched with a knife ; scratched by topaz, zircon, sapphire, and diamond, and thus easily distinguished from these gems. Quartz scratches glass with facility ; felspar and many other minerals can be scratched by quartz. The irregular fracture, the fine parallel markings, and the hardness gene- rally suffice to distinguish quartz. Infusible in the blowpipe flame. The following are the chief varieties of quartz ; the differences are due either to their mode of formation or the presence of impurities. They have the same general characters as pure quartz, 1. TRANSPARENT VARIETIES : BOCK CRYSTAL. Pure, transparent, colourless quartz. Used for spectacle glasses and ornaments. AMETHYST. Transparent. Of a rich violet or purple colour. Used as a gem. 250 DISCRIMINATION OF MINERALS. EOSE QUARTZ. Seldom perfectly transparent. Of a rosy tint. CAIRNGORM, or SMOKY QUARTZ, is transparent, with a smoky tinge. FALSE TOPAZ has a yellow pellucid colour, distinguished from topaz by its inferior hardness. 2. SEMI-TRANSPARENT VARIETIES : CHALCEDONY. Pale colour and waxy lustre. Kesem- bling icicles in some instances, the frothy surface of a liquid in others. CORNELIAN and SARD have red tints. AGATE exhibits cloudy or moss-like patches, or a number of lines arranged in circular and angular patterns. ONYX or SARDONYX is made up of regular layers one above another of different colours, often black, white, and red. It is much used for cameos. FLINT or HORNSTONE. Common dull varieties. 3. OPAQUE VARIETIES : JASPER is quartz rendered opaque by clay, iron, and other impurities ; it is of a red, yellow, or green colour ; sometimes the colours are arranged in ribands, or in other fantastic forms. It is used for ornamental work. BLOODSTONE is green jasper with splashes of red re- sembling blood-spots. 2. OPAL. Chemical Composition. Silica and water. Never crystallised. Fracture conchoidal. Specific gravity, 2'2. Hardness, 6. Can be scratched by quartz, and thus distinguished from it. Infusible. It is generally milk-white. Precious or Noble Opal exhibits a beautiful play of colours, and is a valuable and rare gem. The common varieties are of no value. 3. TALC. Chemical Composition. Silica, magnesia, water (pro- toxide of iron). DISCRIMINATION OF MINERALS. 251 Usually in irregular layers. Nearly opaque. White or green ; pearly lustre ; greasy feel. Specific gravity, 2-7. Hardness, 1. Easily impressed by the nail. But impure varieties are much harder. Infusible. Yields no water when heated in a glass tube. Is not attacked by boiling sulphuric acid. Its greasy feel and pearly lustre readily distinguish it. Mica, which is often confounded with it, is not so soft, has not a greasy feel, and can be split into very thin trans- parent layers. Steatite is a variety often applied to useful purposes. 4. CHLORITE. Chemical Composition. Silica, magnesia, alumina, pro- toxide of iron, water. Often forming rocks. Opaque. Green of various shades. Lustre pearly or dull. Hardness, 1 to 2. Very easily cut with a knife. Infusible. In glass tube yields water. Boiling sul- phuric acid extracts from it magnesia, alumina, and fer- rous oxide. ' Abundant. Of no value. 5. SERPENTINE. Chemical Composition. Silica, magnesia, water (ferrous oxide). Often 'forming rocks. Never crystallised. Opaque. Green. Lustre resinous or dull. Streak white. Hard- ness, 3. Can be cut with a knife. Specific gravity, 2'5. Infusible except in thin edges ; turns white in blowpipe flame. Powder decomposed by sulphuric acid like chlorite. Gives off water in glass tube. Some varieties form handsome stone for slabs and ornamental work. Meerschaum, which is soft and earthy, and nephrite (the New Zealand Maori greenstone), which is as hard as quartz, both contain silica and magnesia. 252 DISCRIMINATION OF MINERALS, 0. AUGITE AND HORNBLENDE. Chemical Composition. Silica, lime, magnesia, ferrous oxide (occasionally alumina and fluorine). In four, six, or eight-sided prisms, exhibiting cleavage in some directions. Usually opaque. Green, black, or white. Glassy, pearly, or resinous lustre. Specific gravity, 3 to 3-5. Hardness, 5 to 6. Can be scratched with a knife, using pressure. Scratched by quartz. Dark varieties fusible. Augite usually occurs in stout six or eight- sided prisms with roof- like terminations ; hornblende in long slender prisms. Asbestos is a variety composed of separable fibres like flax. Abundant in igneous and other rocks. Of no value. Asbestos is used as a fireproof material, and the long silky variety is woven into fireproof fabrics. 7. CHRYSOLITE, OR OLIVINE. Chemical Composition. Silica, magnesia, protoxide of iron. In prisms ; but usually in grains or lumps resembling bottle glass. Transparent or opaque. Green, yellow, or black. Glassy lustre. Fracture conchoidal. Hardness, 6 to 7. Cannot be scratched with a knife. Specific gravity, 3-4. Infusible. In appearance, hardness, and infusibility may resemble tin ore, but is very much lighter in weight. Differs from obsidian (volcanic glass) in its infusibility and superior hardness. Occurs in basalt and lava. Of no value. 8. TOURMALINE. . Chemical Composition. Silica, alumina, magnesia, boracic acid, fluorine, oxides of iron (lime and alkalies). In prisms, with three, six, nine, or more sides, fur- rowed lengthwise, terminating in low pyramids. Com- DISCRIMINATION OF MINERALS. 25$ monly black and opaque ; rarely transparent, and of a rich red, yellow, or green colour. Glassy lustre. Fracture uneven. Specific gravity, 3vl. Hardness, 7 to 8. Can- not be scratched with a knife. Not scratched by quartz. Infusible. When the smooth side of a prism is rubbed on cloth it becomes electric and can attract a small piece of paper ; if the prism is as wide as a pipe-stem, when one side is heated for a moment in the blowpipe flame the opposite side be- comes electric and can attract paper until the heat spreads uniformly through the crystal. (On this account tour- maline is said to be pyro- electric.) Occurs in granite and slate. Of no value ; except the fine-coloured transparent varieties, which are used as gems and for optical purposes. 9. GARNET. Chemical Composition. Silica, alumina, lime, iron, mag- nesia, manganese. Crystallised in dodecahedrons, never in prisms. Trans- parent or opaque. Generally red ; also brown, green, yellow, black, white. Glassy or resinous lustre. Fracture conchoidal or uneven. Specific gravity, 3'5 to 4-3. Hard- ness, 6-5 to 7'5. Cannot be scratched with a knife. Fusible with more or less difficulty. Eed varieties impart a green colour to borax bead owing to presence of chromium. Common in gneiss and schists. Fine-coloured trans- parent varieties (carbuncle, cinnamon stone, almandine) are used in jewellery. 10. TOPAZ. Chemical Composition. Silica, alumina, fluorine. In prisms, ''sometimes furrowed lengthwise, variously terminated, breaking easily across with smooth brilliant cleavage. Transparent or semi-transparent. White, yellow, greenish, bluish, pink. Glassy lustre. Specific gravity, 254 DISCRIMINATION OF MINERALS. 3'5. Hardness, 8. Scratches quartz. Is scratched "by sapphire. Infusible, but often blistered and altered in colour by heat. When smooth surfaces are rubbed on cloth they become strongly electric and can attract small pieces of paper, but rough surfaces do not show this. The brilliant cleavage of topaz distinguishes it from tourmaline and other minerals. Occurs in granite. Used in jewellery. The topaz be- comes electric by friction much easier than other gems, such as the balas ruby, which it may resemble. The white topaz resembles the diamond ; but, unlike diamond, it can be scratched by sapphire. 11. BERYL, OR EMERALD. Chemical Composition. Silica, alumina, glucina. In six-sided prisms. Usually green. Transparent or opaque. Glassy lustre. Fracture uneven. Specific gravity, 2'7. Hardness, 7 to 8. Scratches quartz. Infusible, or nearly so, but becomes clouded by heating. Occurs in granite and schist. Valuable for jewellery when transparent and rich glass green (emerald), or sea green (aquamarine). Opaque crystals of large size, exceed- ing a ton in weight, have been found in North America. 12, ZIRCON, OR HYACINTH. Chemical Composition. Silica, zirconia. In square prisms, terminated by pyramids, and in octa- hedrons. Often found in pebbles and grains. Transparent or opaque. Wine or brownish red, grey, yellow, white. Glassy lustre. Fracture usually irregular, but in one direction it can be split so as to exhibit a smooth, even cleavage face having an adamantine lustre like the diamond. Specific gravity, 4'0 to 5'0. Hardness, 7'5. Scratches quartz ; is scratched by topaz. Infusible ; the red varieties, when heated before the blowpipe, emit a bluish phosphorescent light, and become permanently colourless. DISCRIMINATION OF MINERALS. 255 Occurs in syenite, granite, basalt. Clear crystals used in jewellery, in jewelling watches, and in imitation of diamond. It may be distinguished from diamond by its inferior hardness, and in not becoming so readily electric by friction. 13. FELSPAR. Chemical Composition. Silica, alumina, potash or soda (lime). Crystallised or in irregular masses. Opaque. Usually flesh-red, or white, or of various dull tints. Glassy or pearly lustre. Fracture irregular ; but in some directions it splits with an even, glimmering, cleavage face. Specific gravity, 2-3 to 2-8. Hardness, 6. Easily scratched by- quartz. Cannot be scratched with a knife without greatest pressure. In thin edges fusible with difficulty. Abundant in granitic and porphyritic rocks. No value. By its decomposition it forms porcelain clay or kaolin. 14. MICA. Chemical Composition. Silica, alumina, magnesia, potash, iron. Always crystallised in thin plates, which may be split into extremely thin flexible layers. Transparent in thin layers. Brown or black. Lustre glassy, pearly, or metallic. Streak white. Specific gravity, 2 -7 to 3*1. Hardness, 2 to 2- 5. Very easily scratched with a knife. Infusible. Differs from talc in not having a greasy feel, in being harder, flexible, and affording thinner layers perfectly transparent. Abundant in granite and schist ; fine particles common in sandstone. Applied to various uses when in large plates, otherwise of no value. Was formerly used instead of glass for windows. 15. ZEOLITES. This name is used for a large class of silicates, com- prising from fifty to a hundred different minerals, which all contain water as an essential ingredient, and which 256 DISCRIMINATION OF MINERALS. melt readily before the blowpipe, and boil up owing to the disengagement of steam. They occur filling pores and cavities in basalt, lava, and other rocks. They are usually white and well crystallised. Can be scratched with a knife. Of no value. Silica, alumina, lime, soda, potash, and water are the principal ingredients. 16. CORUNDUM, OR SAPPHIRE. Chemical Composition. Alumina. In six-sided prisms, often irregularly shaped. Some- times in granular masses. Transparent or opaque. Blue, black ; also red, green, yellow, white. Glassy lustre, some- times pearly. Fracture uneven or conchoidal. Specific gravity, 3-9 to 4'2. Hardness, 9. Easily scratches topaz and quartz. In hardness it is only inferior to the diamond. Infusible. Occurs in river sands ; in granite, felspar, magnetic iron, basalt. As a gem it stands next in value to the diamond r but its tint must be brilliant and clear. The blue variety is called Sapphire, the most esteemed shade being deep velvet blue ; the blood-red variety is the Oriental ruby, which can be easily distinguished from other red gems by its superior hardness ; the bright yellow variety is the Oriental topaz, distinguished by its hardness from the topaz, yellow tourmaline, and false topaz ; the bright green is the Oriental emerald ; the bright violet, Oriental amethyst ; these varieties readily scratch the emerald (No. 11) and amethyst (see quartz, No. 1) ; one variety exhibits a six- rayed star inside the prism, and it is called the Asterias. Dull crystals are called corundum, and grey or black granular varieties emery ; these two kinds are used for polishing powder. The ruby is the most highly prized form of this mineral. 17. SPINEL. Chemical Composition. Alumina, magnesia. In octahedrons, occurring only crystallised. Usually red and transparent ; also white, blue, green, yellow, brown, black; the dark shades usually opaque. Lustre glassy. DISCRIMINATION OF MINERALS. 257 Fracture conchoidal. Specific gravity, 3*5 to 4*0. Hard- ness, 8. Scratches quartz. Infusible, and thus distinguished from garnet, which it may resemble. Colour altered transiently by heat. Dis- tinguished from zircon by its superior hardness and inferior specific gravity. Occurs in river sand ; in igneous rocks, gneiss, lime- stone. The bright transparent varieties are used in jewel- lery. When red it forms the common spinel, or balas- ruby, which is distinguished from the Oriental ruby by its inferior hardness ; bright green, chlorospinel ; orange, rubicelte ; violet, almandine-ruby ; black, pleonast. 18. CHRYSOBERYL. Chemical Composition. Alumina, glucina. In prisms or tables. Transparent or semi-transparent. Green. Lustre glassy. Fracture conchoidal ; imperfect cleavage. Specific gravity, 3*5 to 3-8. Hardness, 8'5. Infusible and unaltered before the blowpipe. Distinguished from beryl by its specific gravity, its tabular crystallisation, and its entire infusibility. Occurs with beryl in river sand, gneiss, and granite. Pellucid and fine opalescent varieties are used as gems. 19. DIAMOND. Chemical Co mposition. Carbon . In octahedrons, tetrahedrons, dodecahedrons, and forms related to these ; the faces of the crystal sometimes curved. Transparent. Colourless, yellow, red, green, blue, white, brown, or black. Lustre adamantine. Breaks with smooth cleavage planes parallel to the octahedral faces. Specific gravity, 3*5 ; loses I0-35ths of its weight in water. Hard- ness, 10. It is the hardest substance known, and scratches all other minerals and gems. Infusible. It burns and is consumed at a high tempe- rature. Becomes strongly electric when rubbed, and can then attract light objects ; other gems do not exhibit this pro- s 258 DISCRIMINATION OF MINERALS. perty unless polished. Some varieties, after exposure to the sun, are said to give out light when placed in the dark. The distinguishing characters of diamond are, its hard- ness, crystalline form, clean fracture, its brilliant reflection and adamantine lustre, the facility with which it may be electrified by friction, and a peculiar sound on rubbing together. Diamonds have been found in quartz- conglomerate and micaceous sandstone, but are mostly obtained in river beds. 20. GRAPHITE, OR BLACKLEAD. Chemical Composition. Carbon . In six-sided prisms ; but usually in uncrystallised wavy layers. Opaque. Black. Lustre metallic. Specific gravity, 2. Hardness, 1 to 2. Very easily cut with a knife. Has a greasy feel ; marks paper like a lead pencil. Infusible. Burns slowly away. Molybdenite and foliated tellurium resemble graphite ; the former has a paler colour than graphite, and the latter is very easily fusible. Occurs in gneiss and slate. Valuable for lead pencils and crucibles. 21. COAL. Coal and carbonaceous shale differ from all the minerals which they may resemble by burning before the blowpipe and leaving a white or brown ash. The quantity of ash affords an estimate of the purity of the coal. 22. APATITE. Chemical Composition. Phosphoric acid, lime, fluorine. In six-sided prisms. Also in masses. Transparent or opaque. Usually green. Sometimes white, yellow, blue, red, brown. Lustre resinous. Fracture conchoidal or uneven. Specific gravity, 3*2. Hardness, 5. Can be scratched with a knife, using pressure. Infusible, except on very thin edges. Some kinds DISCRIMINATION OF MINERALS. 259 phosphoresce when heated. The pure mineral in powder dissolves slowly in nitric acid without effervescence. The crystals may resemble beryl, which, however, is too hard to be scratched with a knife. Occurs in gneiss, slate, limestone. Of value from its use in the manufacture of artificial manures. 23. FLUOR-SPAR. Chemical Composition. Fluorine, calcium. In cubes or octahedrons. Also in masses. Transparent or opaque. White or light violet, blue, green, or yellow; sometimes layers of different tints in the same piece. Lustre glassy. Breaks with smooth cleavage planes parallel to the octahedral faces. Specific gravity, 3'0 to 3*2. Hard- ness, 4. Can be scratched with a knife, but not so easily as calcite. Fusible with difficulty ; generally flies to pieces when heated. Some varieties phosphoresce. Occurs in veins with lead and silver ores. Used in etching glass, and as a flux in smelting ; sometimes for ornaments, but is too brittle. Abundant in many countries, and of little value. 24. CALC-SPAR, OR CALCITE. Chemical Composition. Carbonic acid, lime. In rhombohedrons and other crystalline forms. Also massive, earthy, or fibrous. Transparent or opaque. White when pure ; often tinted. Lustre glassy, or dull. Breaks with smooth cleavage planes parallel to the rhombohedral faces. Specific gravity, 2-5 to 2*8. Hardness, 3. Easily scratched with a knife, streak white. Infusible before the blowpipe, but emits a strong light. When burned, as in a kiln, it forms quicklime. Effervesces when vinegar is poured upon it. It completely dissolves in nitric acid with rapid effervescence. Calc-spar is one of the most abundant minerals ; it occurs in cavities and veins of all kinds of rock. The term 8 2 260 DISCRIMINATION OF MINERALS. calc-spar or calcite is restricted to the crystallised variety. In an imperfectly crystallised and compact form it exists in large rocky masses and beds ; all marbles and limestones consist of it, mixed more or less with impurities. Chalk and stalactites are nearly pure calcium carbonate. All varieties of calcium carbonate may be easily distinguished by being scratched with a knife, giving a white streak whatever the colour of the mass may be, by effervescing with an acid, and by being infusible. 25. MAGNESITE. Chemical Composition. Carbonic acid, magnesia. In rhombohedrons. Also globular, compact, earthy. Transparent or opaque. White, yellow, brown. Lustre glassy. Cleavage parallel to rhombohedral faces. Specific gravity, 2-8 to 3*0. Hardness, 4 to 5. Infusible. Dissolves in nitric acid, with very slow effervescence. Does not yield quicklime when burnt. Occurs with serpentine and limestone. Used in pre- paring Epsom salts. 26. DOLOMITE. Chemical Composition. Carbonic acid, lime, magnesia. In rhombohedrons, faces often curved. Often granular or massive. White or dull- tinted. Glassy or pearly. Specific gravity, 2-8 to 2-9. Hardness, 3-5 to 4. Infusible. Effervesces in nitric acid, and dissolves more slowly than calc-spar. Yields quicklime when burnt Occurs in extensive beds of various ages like limestone. Used as a building stone, and in the manufacture of Epsom salts. Difficult to distinguish from calcite without chemical analysis. 27. ARAGONITE. Chemical Composition. Same as calc-spar. It differs from calc-spar in its crystalline form, which is usually difficult to discern. It often occurs in fibrous DISCRIMINATION OF MINERALS. 261 clusters or in tangled branches. Specific gravity, 2'9. Hardness, 3-5 to 4-0. It has the general characters of calc-spar, but may be distinguished from it by falling to powder in the blowpipe flame, as well as by its superior hardness. 28. ROCK SALT. Chemical Composition. Sodium chloride. Has the characters of ordinary table salt, but is more or less impure. Occurs in beds interstratified with sand- stones and clays, which are usually of a red colour and associated with gypsum. In the county of Cheshire, where salt mines are worked, the surface indications are brine springs supporting a vegetation like that near the sea-coast ; also occasional sinking of the soil caused by the removal of the subterranean bed of salt, by spring water in some cases, and by mining operations in others. Small and unimportant quantities of salt are often found encrust- ing various rocks in dry weather. 29. SOLUBLE SULPHATES. Aluminium sulphate alone or combined with potassium (alum), and magnesium sulphate (Epsom salts), are often found encrusting rocks. They are soluble in water, and easily distinguished by their taste. 30. NITRE. Chemical Composition. Potassium nitrate. Nitre or saltpetre is another soluble mineral. It has a cooling taste. It can be easily distinguished by the vivid manner in which it burns on red-hot charcoal. 31. GYPSUM, SELENITE, OR ALABASTER. Chemical Composition. Sulphuric acid, lime, water. In prisms with oblique terminations ; sometimes resem- bling an arrowhead. Transparent or opaque. White or 262 DISCRIMINATION OF MINERALS. dull-tinted. Glassy, pearly, or satin lustre. Cleavage occurs easily in one direction. Specific gravity, 2-3. Hard- ness, 2. Very easily cut with a knife. Fusible with difficulty. In the blowpipe flame it be- comes white and opaque without fusing, and can then be easily crumbled between the fingers. Nitric acid does not cause effervescence. Occurs in fissures and in stratified rocks, often forming extensive beds. When burnt it forms plaster-of-Paris ; it is also used for ornaments, and as a manure. 32. HEAVY SPAK, OR BARYTES. Chemical Composition. Sulphuric acid, baryta. In tabular glassy crystals. Also in dull masses. Trans- parent or opaque. White or tinted. Specific gravity, 4-3 to 4*8 ; its great comparative weight readily distinguishes it. Hardness, 2-5 to 3 -5. Splinters fly off the crystals when heated in the blow- pipe flame. Fusible with difficulty. Not acted upon by acids. Occurs with various ores. Used as a white paint. 33. SULPHUR. Crystallised or massive. Yellow. Eesinous lustre. Specific gravity, 2-1. Hardness, T 5 to 2-5. Fusible; burns with a blue flame and well-known odour. Occurs in volcanic regions, and in beds of gypsum. 34. TIN ORE. Chemical Composition. Tin, 78'4 ; oxygen, 21-6. In four-faced prisms and pyramids, having an ada- mantine lustre. Also in masses and grains (stream tin) usually dull; sometimes resembling wood (wood tin). Semi-transparent or opaque. Brown or black ; streak and powder pale brown. Fracture uneven. Specific gravity, 6*8 to 7*0 ; the great comparative weight is an important character to observe in distinguishing tin ore from other . DISCRIMINATION OF MINERALS. 263 minerals. Hardness, 6'0 to 7*0. Cannot be scratched with a knife, and may thus be distinguished from blende, which it resembles in lustre and infusibility. Infusible. When mixed in powder with sodium carbon- ate, placed on charcoal and covered with a small piece of potassium cyanide, and then heated in the blowpipe flame, a malleable globule of metallic tin is obtained. Occurs in veins, and disseminated in granite, schist, slate, and porphyry ; and in alluvial deposits. It is a valuable ore, and the sole commercial source of the metal. 35. MOLYBDENITE. Chemical Composition. Molybdenum, 58*9 ; sulphur, 41-1. In thin plates, like graphite (No. 20). Lustre metallic. Colour, lead grey. Specific gravity, 4*5 to 4-6. Hardness, 1*0 to 1*5. Easily scratched by the nail. Infusible. Tinges blowpipe flame faint green. Heated on charcoal for a long time, it gives off a faint sulphurous odour, and becomes encrusted white. Occurs in granite, syenite, and chlorite schist. Not applied to any particular uses. 36. BISMUTH. Chemical Composition. Metallic bismuth. Sometimes crystallised in rhombohedrons closely re- sembling cubes, but generally massive. Lustre metallic. White, with a tinge of red, liable to tarnish. Brittle. Specific gravity, 9-6 to 9-8. Hardness, 2*0 to 2-5. Easily scratched with a knife. Easily fusible. Sometimes gives off an odour of garlic, owing to admixture of arsenic. Occurs with cobalt, silver, and tin ores in granite and 'slate rocks. Bismuth is a very valuable metal. 37. ANTIMONY SULPHIDE. Chemical Composition. Antimony, 72*9 ; sulphur, 27*1. Usually in long columnar or fibrous crystals. Also 264 DISCRIMINATION OF MINERALS. massive and granular. Lustre metallic. Lead colour. Often tarnished. Streak metallic. Specific gravity, 4'6 to 4- 7. Hardness, 2. Very easily scratched with a knife. Easily fusible. Before blowpipe gives off white vapours and an odour of sulphur, and is entirely volatilised. When the corner of a large piece of ore is fused, the border of the fused part is often tinted red. When heated on char- coal with cyanide of potassium, it gives a globule of metallic antimony, which is brittle, has a crystalline surface, burns when strongly heated, emitting white fumes, and can be entirely volatilised. Occurs in veins in granite and slate, alone, or with ores of silver, lead, and other metals. This ore is the principal commercial source of the metal. 38. ARSENIC. Chemical Composition. Metallic arsenic. Seldom distinctly crystallised. Usually in fine granular or spherical masses. Colour white, usually with a black tarnish. Streak white, metallic. Brittle. Specific gravity, 5-7 to 5-8. Hardness, 3-5. Before the blowpipe it quickly volatilises without fusing, giving off white fumes having an odour of garlic. Occurs in veins with lead and silver ores. 39. ARSENIC SULPHIDE. Chemical Composition. Sulphur, arsenic. In crystals, or massive. Yellow or red. Semi-transparent or opaque. Eesinous or glassy lustre. Specific gravity, 3*5. Hardness, 1*5. Volatilised before the blowpipe with a blue flame. Occurs in veins with arsenical ores ; in beds of clay, limestone, and gypsum, and has been observed in lava. Used as a pigment and in making fireworks, but objection- able for each of these uses, owing to its highly poisonous character. DISCRIMINATION OF MINERALS. 2(35 40. NATIVE IRON. Chemical Composition. Iron, with a small percentage of nickel. Occurs in meteorites. Resembles ordinary iron. Mal- leable. Is attracted by a magnet. Specific gravity, 7*0 to 7-8. 41. IRON PYRITES. Chemical Composition. Iron, 46 '7 ; sulphur, 53'3. In cubes and allied forms ; sides often marked by fine parallel lines. Also massive. Brass yellow. Lustre metallic. Fracture irregular. Specific gravity, 4*8 to 5*1. Hardness, 6 to 6- 5. Cannot be scratched with a knife ; scratched by quartz ; scratches glass with great facility. Strikes fire with steel (the origin of the term pyrites). Before the blowpipe it burns with a blue flame, giving off an odour of sulphur, and ultimately fuses into a, black magnetic globule. Abundant. Used as a source of sulphur and sulphuric acid ; occasionally auriferous. This ore and arsenical pyrites form the ' mundic ' of miners. It is easily dis- tinguished from copper pyrites by its hardness ; copper pyrites being easily cut with a knife. Distinguished from gold by its hardness and in not being malleable, and in giving off* sulphurous odours in the blowpipe flame. 42. ARSENICAL PYRITES (MISPICKEL). Chemical Composition. Iron, 34'4 ; arsenic, 19-6 ; sulphur, 46 -0. In flattened prisms. Also massive. White. Lustre metallic. Streak grey. Fracture uneven. Specific gravity, 6-0 to 6*3. Hardness, 5- 5. Cannot be scratched with a knife ; scratched by quartz. Heated before the blowpipe it gives off white arsenical fumes of a garlic odour, and ultimately fuses into a black globule. Abundant in mining districts ; sometimes auriferous. This ore and iron pyrites form the ' mundic ' of miners. DISCRIMINATION OF MINERALS. 43. MAGNETIC IRON. Chemical Composition. Iron, 72-4 ; oxygen, 27*6. In octahedrons and dodecahedrons. Also in masses (lodestone) and in grains. Black. Lustre metallic. Streak or powder black. Fracture irregular. Specific gravity, 5-0 to 5-2. Hardness, 5-5 to 6 -5. Not scratched with a knife. Magnetic ; it can attract iron filings. Is itself attracted by a magnet. Infusible. With borax bead gives the indications of iron. Occurs in many rocks, sometimes in beds, or forming mountainous masses ; common in river sands. Used as an ore of iron. 44. SPECULAR IRON, HEMATITE, OR MICACEOUS IRON. Chemical Composition. Iron, 70 ; oxygen, 30. In tabular crystals or scales. Also fibrous, massive, granular, earthy. Colour black. Streak or powder, red. Lustre, metallic or dull. Specific gravity, 4*5 to 5-3. Hardness of crystals, 5-5 to 6*5. Not scratched with a knife. Earthy varieties softer, and can be scratched with a knife. Infusible. With borax bead gives the indications of iron. An abundant ore of iron. Oft en gradually changes into red or brown ochre. 45. RED FERRIC OXIDE, OR RED OCHRE. Chemical Composition. Iron oxide, and more or less water. An uncrystalline earthy variety of the preceding, often mixed with clay. Colour, bright or dull red. Can generally be scratched with a knife. Blackens when heated, but regains its red colour on cooling. With borax bead gives the indications of iron. Abundant ore of iron. DISCRIMINATION OF MINERALS. 267 4G. BROWN FERRIC OXIDE (YELLOW OR BROWN OCHRE). Chemical Composition. Iron oxide, water. Like the last, but of a brown, yellow, or black colour. Eartliy, fibrous, stalactitic. Scratched with a knife. Blackens before the blowpipe. With borax bead gives the indications of iron. An abundant ore of iron. 47, TITANIC IRON. Chemical Composition. Iron oxides and titanic acid in variable proportions. In octahedrons or in tabular plates. Also in grains. Black. Lustre metallic. Streak or powder black. Specific gravity, 4- 5 to 5*3. Hardness, 5 to 6-5. Not scratched with a knife. Infusible. With borax gives the indications of iron. With microcosmic salt, which is often used instead of borax in an exactly similar way, it gives a red bead in the reducing part of the flame, but rather a large quantity of the mineral must be used to obtain this result. It is sometimes magnetic. Its black streak or powder distinguishes it from specular iron, which it often resembles. Common in some river sand. 48. CHROMIC IRON. Chemical Composition. Chromium sesquipxide, ferrous oxide (alumina, magnesia). In octahedrons. Usually massive. Black. Lustre faintly metallic. Streak or powder dark brown. Fracture irregular. Specific gravity, 4-4 to 4'6. Hardness, 5'5. Not scratched with a knife. Infusible. With borax bead gives the characteristic indications of chromium. Occurs in serpentine. Used in the preparation of chromium colours. * 49. GREENEARTH. Chemical Composition. Iron silicate, and other ingre- dients. Has a green earthy appearance, often resembling 268 DISCRIMINATION OF MINERALS. an ore of copper, but is readily distinguished from copper by blowpipe tests, and by not forming a blue solution in nitric acid. 50. IRON CARBONATE. Chemical Composition. Carbonic acid, 37 '9; ferrous oxide, 62-1. In rhombohedrons ; faces often curved. Usually massive, globular, fibrous, or encrusting. Light or dark brown. Glassy or pearly lustre. Streak white. Specific gravity, 3 -7 to 3 -9. Hardness, 3*5 to 4-5. Scratched with a knife. Infusible. Blackens when heated. With borax bead gives the indications of iron. Dissolves in nitric acid with effervescence when heated. Occurs in beds and nodules in stratified rocks ; in veins and cavities. It is often mixed with clay (clay ironstone). Abundant ore of iron. 61. MANGANESE ORES. Chemical Composition. Various manganese oxides. Crystallised or massive. Black. Lustre unmetallic ; dull or shining. Powder or streak brown or black. Spe- cific gravity, 4 to 5. Hardness generally below 3. Very easily scratched with a knife. Infusible. With borax bead gives the characteristic indications of manganese. Widely distributed. Used in chemical manufactures. 52. ARSENICAL NICKEL. Chemical Composition. Nickel, 44 ; arsenic, 56. Usually in masses of a pale copper colour and metallic lustre. Specific gravity, 7*2 to 7*8. Hardness, 5 to 5-5. Scratched with a knife, using pressure. Before the blowpipe on charcoal it melts, giving out white arsenical fumes having a garlic odour. It is readily distinguished by its pale copper red colour and its arsenical fumes when heated. DISCRIMINATION OF MINERALS. 26$ Occurs in veins in granite and slate, with ores of cobalt, silver, copper, bismuth, lead. A valuable source of metallic nickel. 53. SMALTINE (TIN-WHITE COBALT). Chemical Composition. Cobalt up to 24 per cent., arsenic. In octahedrons, cubes, dodecahedrons, and allied forms. Also massive. Tin-white or steel-grey. Lustre metallic. Streak greyish -black. Fracture uneven. Specific gravity, 6-3 to 6-6. Hardness, 5-5 Fusible. In the blowpipe flame gives off arsenical fumes (odour of garlic). With borax bead gives the characteristic indications of cobalt. In nitric acid forms a pink solution. Eesembles mispickel and iron pyrites, but is at once distinguished by the test with borax bead. Its arsenical fumes distinguish it from iron pyrites, and its crystalline form from mispickel. Occurs in veins in slate and gneiss. A valuable ore of cobalt. 54. COBALT BLOOM. Chemical Composition. Cobalt oxide, 37*6 ; arsenic acid, 384 ; water, 24'0, In oblique crystals, with a highly perfect cleavage like mica. Also in incrustations. Eed or pink, grey, green. Lustre brilliant pearly. Transparent or opaque. Specific gravity, 2-9 to 3-1, Hardness, 1-5 to 2. Very easily cut with a knife. Fusible in blowpipe flame, evolving arsenical fumes. When heated on charcoal it gives off an odour of arsenic. With borax bead gives indication of cobalt. Occurs in beds and veins with other ores of cobalt. A valuable ore of cobalt. 55. BLENDE. Chemical Composition. Zinc, 66*7 ; sulphur, 33 -3. In dodecahedrons, octahedrons, and allied forms. Also massive. Yellow, red, brown, black. Lustre adamantine, 270 DISCRIMINATION OP MINERALS. resinous, or waxy. Transparent or opaque. Breaks with brilliant cleavage faces in some directions. Specific gravity, 4*0 to 4*1. Hardness, 3 -5 to 4*0. Easily scratched with a knife. Infusible. Emits a strong light when heated, but no odour of sulphur is perceptible. It is easily distinguished by its waxy lustre, softness, infusibility, and perfect cleav- age. It dissolves at once in nitric acid. Occurs with lead and copper ores. It is the 'black jack' of miners. An ore of zinc, but more difficult to smelt than the carbonate and silicate. 56. ZINC CARBONATE (CALAMINE). Contains 52 per cent, of zinc. Usually in crusts or masses. White, green, or brown. Opaque. Pearly or glassy. Specific gravity, 4*1 to 4'5. Hardness, 5. Can be scratched with a knife, using a little pressure. Infusible. On charcoal becomes yellow whilst hot, white on cooling. Dissolves rapidly with effervescence when- heated with nitric acid. Occurs with galena and blende. A valuable zinc ore. 57. ZINC SILICATE (SMITHSONITE). Contains 53 per cent, of zinc. In prisms, or massive. White, greenish, bluish, or brownish. Glassy lustre. Transparent or opaque. Specific gravity, 3- 3 to 3*5. Hardness, 5. Infusible. Shines with a green light in the blowpipe flame. Does not effervesce with nitric acid, but dissolves, leaving a jelly of silica. Occurs with zinc carbonate. A valuable zinc ore. 58. GALENA. Chemical Composition. Lead, 86*6 ; sulphur, 13*4. In cubes. Also granular, massive. Lead colour. Metallic lustre. Streak metallic. Breaks into cubical fragments with bright cleavage faces. Specific gravity, DISCRIMINATION OF MINERALS. 271 7*4 to 7'7. Hardness, 2-5. Very easily scratched with a knife. Easily fusible. Before the blowpipe on charcoal is reduced to a metallic globule of lead, giving off an odour of burning sulphur. Occurs in granite and stratified rocks. Often associated with copper and other ores. The principal ore of lead. It usually contains a small quantity of silver. 59. LEAD CAKBONATE (CERUSITE). Contains 77 per cent, of lead. In prisms, sometimes united in four or six-rayed crosses. White or grey'." Transparent or opaque. Lustre glassy. Specific gravity, 6-4 to 6 -6. Hardness, 3- 5. Flies violently to pieces in the blowpipe flame. If placed in a cavity on charcoal, and covered with sodium carbonate, then carefully fused by the flame, it yields a globule of metallic lead. In nitric acid it dissolves with effervescence. Usually occurs with galena. It is a valuable lead ore. 60. LEAD SULPHATE (AJSTGLESITE). Contains 68 per cent, of lead. In slender brilliant crystals upon galena. Also massive. White or grey. Transparent or opaque. Specific gravity, 6-3. Hardness, 3. Before the blowpipe fusible, but apt to decrepitate on charcoal ; with sodium carbonate yields a globule of me- tallic lead. Differs from lead carbonate in not dissolving with effervescence in nitric acid. Usually occurs with galena, and results from its decom- position. 61. PYROMORPHITE (LEAD PHOSPHATE). Chemical Composition. Lead oxide, 74'0 ; phosphoric acid, 15-8 ; lead chloride 10-2. In stout prisms, grouped together. Also massive. Bright green or brown. Opaque or semi transparent. 272 DISCRIMINATION OF MINERALS. Lustre resinous. Streak white. Fracture irregular. Specific gravity, 6*9 to 7*1. Hardness, 3-5 to 4-0. Easily scratched with a knife. Easily fusible. With sodium carbonate on charcoal the lead is reduced. Soluble in nitric acid. Occurs sparingly in veins with galena. 62. CINNABAR. Chemical Composition. Mercury, 86'2 ; sulphur, 13 -8. In granular, compact, and earthy masses. Sometimes in crystals, exhibiting adamantine cleavage faces. Opaque or semi-transparent. Vermilion or brownish red. Specific gravity, 8-0 to 8-2. Hardness, 2-5. Very easily scratched with a knife. Before the blowpipe it volatilises, giving off a strong odour of burning sulphur. Mixed with dried sodium car- bonate, and heated over a candle-flame, in an iron spoon, it gives off vapours of mercury, which may be condensed on a gold coin held half an inch above the mixture. The surface of the coin appears whitish at first, but when rubbed between the fingers becomes brilliantly amalga- mated. With care, this test easily detects one per cent, of cinnabar in an ore. The mercury is removed from the gold coin by gentle heating. The blowpipe tests distin- guish it at once from red oxide of iron and all other red minerals. Occurs in talcose and argillaceous rocks. It is the principal source of the mercury of commerce. 63. NATIVE MERCURY, OR QUICKSILVER. The metal in a pure state is rarely found. It occurs disseminated in liquid globules through sandstone and other rocks, in cavities of which it may accumulate. It is easily recognised. A rock suspected to contain mer- cury may be tested by simply heating it as described under cinnabar ; but without the addition of carbonate of soda. DISCRIMINATION OF MINERALS. 273 64. NATIVE COPPER. Usually in strings, plates, or irregular masses : some- times crystalline. Like ordinary copper, but often tar- nished. Specific gravity, 8-9. Easily scratched with a knife. Malleable, can be flattened out under a hammer. Occurs with copper ores. 65. VITREOUS COPPER. Chemical Composition. Copper, 79*8 ; sulphur, 20*2. Sometimes in prisms, but usually massive. Blackish lead grey, tarnished. Streak metallic. Specific gravity, 5'5 to 5 -8. Hardness, 2-5 to 3. Very easily scratched with a knife. Fusible. Before the blowpipe gives off an odour of sulphur. When heated on charcoal, a malleable globule of metallic copper remains, tarnished black, but rendered evident on flattening under a hammer. With borax bead gives the indications of copper. Dissolves in nitric acid, forming a blue solution. (These tests distinguish it from sulphide of silver.) Occurs with other copper ores. A valuable ore of copper. 66. COPPER PYRITES. Chemical Composition. Copper, 34*6 ; iron, 30'9 ; sul- phur, 34-9. In tetrahedrons or octahedrons. Usually massive. Brass yellow, often tarnished. Lustre metallic. Streak unmetallic, blackish green. Fracture uneven. Specific gravity, 4*1 to 4*3. Hardness, 3'5 to 4'0. Easily scratched with a knife. Fusible. Gives off an odour of sulphur before blow- pipe. Does not give the indications of copper with borax bead, or when heated upon charcoal with sodium carbon- ate. Dissolves in nitric acid, forming a blue solution. Distinguished from iron pyrites by being easily cut with a knife ; and from gold by not flattening under a hammer, and by its greenish powdery streak. Occurs in granite and slate in lodes or veins. Valuable ore of copper. T 274 DISCRIMINATION OF MINERALS. 07. GREY COPPER. This term includes a variety of ores having a common crystalline form, generally the tetrahedron also ; a definite chemical formula, though the ingredients are numerous and may be variously combined within certain limits. Sulphur is an invariable ingredient ; and arsenic or anti- mony, one or both, must be present ; the other ingredients are copper, iron, zinc, lead, silver, or mercury, in variable proportions. The copper ranges up to 40 per cent. ; and in some kinds as much as 30 per cent, of silver has been found. It also occurs massive. Steel grey to iron black. Lustre metallic. Streak black, or dark red when zinc is present. Fracture uneven. Specific gravity, 4'5 to 5-2. Hardness, 3 to 4. Can be easily scratched with a knife. Fusible. Before the blowpipe gives off an odour of sulphur, also white inodorous fumes of antimony, and occasionally arsenic. Copper cannot be detected by the blowpipe tests. It dissolves in nitric acid, forming a greenish brown solution. Occurs with copper pyrites, galena, and blende. This ore is wrought for copper, and occasionally for silver. 68. BLACK CUPRIC OXIDE. Heavy black powder or mass. Soft. Easily distin- guished from manganese by affording the indications of copper by the blowpipe tests. It results from the waste of various copper ores. Valuable as a source of the metal. 69. RED CUPROUS OXIDE. Chemical Composition. Copper, 88-8 ; oxygen, 11*2. In octahedrons and dodecahedrons. Also in granular and earthy masses. Eed. Lustre adamantine, metallic, or earthy. Streak red. Semi-transparent or opaque. Exhibits cleavage parallel with octahedral faces. Specific gravity, 6. Hardness, 3'5 to 4-0. Can be scratched with a knife. Before the blowpipe on charcoal it yields a globule of metallic copper. With borax bead gives the indications DISCRIMINATION OF MINERALS. 275 of copper. Forms a blue solution in nitric acid. These tests distinguish it from red ferric oxide. Occurs in granite and slate, with copper ores and galena. Valuable source of the metal. 70. COPPER CARBONATES, BLUE AND GREEN. Chemical Composition. Oxide of copper, carbonic acid, water ; the percentage of metallic copper about 56. In crystals, but usually in fibrous, silky, globular, incrusting masses. Blue or green. Opaque. Glassy, silky, or dull. Specific gravity, 3-7 to 4'0. Hardness, 3'5 to 4*0 ; can be scratched with a knife. Blacken when heated. On charcoal are reduced to a globule of pure copper. Give the indications of copper with borax bead. Soluble in nitric acid with effervescence, forming a blue solution. Copper Silicate resembles the carbonate, and is distin- guished by dissolving in nitric acid without effervescecce. Occur with copper ores, and result from their decom- position. Valuable sources of the metal. 71. PLATINUM. In flattened or angular grains or nuggets, which are malleable. Steel-grey. Lustre metallic. Specific gravity, IT to 19. As heavy as gold, and, therefore, easily distin- guished and separated from lighter materials. Infusible. Insoluble in nitric acid. Occurs in quartz veins, but principally in alluvial deposits with gold. Used chiefly for chemical apparatus. Of great value. 72. GOLD. In dust, grains, or nuggets in river sand ; or in wiry, branching, and irregular forms in quartz. Pale or deep yellow. Malleable. Specific gravity, 15 to 19. Hard- ness, 2-5 to 3-0. Fusible without blackening, and without giving off any odour. Imparts no colour to boiling nitric acid. T 2 270 DISCRIMINATION OF MINERALS. The minerals commonly accompanying gold are iron pyrites, arsenical iron, iron and manganese oxides, galena., zinc blende, and copper pyrites in quartz veins ; and magnetic iron, titanic iron, chromic iron, tin ore, quartz, zircon, topaz, corundum, diamond, in alluvial deposits. 73. SILVER. In strings, plates, and branching forms penetrating quartz, porphyry, slate, granite. Silver white, but usually tarnished black. Malleable. Specific gravity, about 1O5, Hardness, 2-5 to 3v Fusible without giving off any odour. Soluble in nitric acid, and on adding salt to the solution a white curd is thrown down which blackens on exposure to sun- light. 74. SULPHIDE OP SILVER. Chemical Composition. Silver, 87 ; sulphur, 13. In dodecahedrons or allied forms. Also massive. Black. Opaque. Lustre metallic. Streak shining. Specific gravity,* 7*2. Hardness, 2-0 to 2'5. Very easily cut with a knife. Very fusible, giving off an odour of sulphur when heated. Before the blowpipe on charcoal, with or with- out sodium carbonate, it yields a white globule of metallic silver which can be flattened under a hammer. The ore .is soluble in nitric acid, and on adding salt to the solution a white curdy precipitate is thrown down y which blackens on exposure to sunlight. Occurs in veins in granite, porphyry, and slate, with arsenic, silver, and lead ores. 75. ANTIMONIAL AND ARSENICAL SILVER ORES. Several ores of silver contain arsenic and antimony, as well as sulphur ; the percentage of silver in these ores varies from 12 to 68. Bed, grey, or black. Lustre ada- mantine or metallic. Red streak. Specific gravity, 5 to 0. Hardness, 2 to 3. Easily scratched with a knife. DISCRIMINATION OF MINERALS. i>77 Fusible. Before the blowpipe give off an odour of sulphur, or arsenical fumes of a garlic odour, or fumes of antimony. Heated on charcoal with sodium carbonate, afford a globule of metallic silver. Nitric acid extracts the silver from these ores, forming a solution in which salt throws down a white curd, black- ening on exposure to sunlight. 76. HORN SILVER. Chemical Composition. Silver, 75-3 ; chlorine, 24*7. In veins with silver ores. Greenish. Waxy lustre. Hard- ness, I'O to 1'5. Cuts like wax or horn. Very easily fusible. Heated with sodium carbonate, 'On charcoal, it yields a globule of metallic silver. 77. LEAD AND ANTIMONY SULPHIDES. Lead- or steel-grey. Lustre metallic. Hardness not exceeding 3. Specific gravity about 6. Fusible. Give off an odour of sulphur and white fumes of antimony before the blowpipe ; and with sodium carbonate on charcoal afford a bead of metallic lead. 78. MERCURY, LEAD, SILVER, OR COPPER SELEXIDE. Gives off a strong odour of horse-radish, due to sele- nium, when heated before the blowpipe. The metal may be found in some cases by sodium carbonate on charcoal, in others by the use of nitric acid. 79. MILLERITE (NICKEL SULPHIDE). Chemical Composition. Nickel, 64- 9 ; sulphur, 35*1. In delicate needles. Brass-yellow. Metallic lustre. Fusible. Hardness, 3 to 5. Imparts to borax the -colour of nickel. 278 DISCRIMINATION OF MINERALS. 80. WHITE NICKEL. Chemical Composition. Nickel, 28-3 ; arsenic, 71*7. Same colour, hardness, specific gravity, and crystalline form as smaltine, No. 53. Often contains cobalt, and graduates into smaltine, with which it is usually found. 81. RUTILE. Chemical Composition. Titanium and oxygen. In crystals and masses. Eed brown. Streak paler. Lustre sub-metallic. Hardness, 6'0 to 6-5. Specific gravity, 4-2 to 4-3. Infusible. With borax bead, yellowish-green or colour- less (oxidising), dirty violet (reducing). 8-2. SPHENE. Chemical Composition. Titanic acid, silica, lime. In thin wedge-shaped crystals. Yellow, green, brown. Transparent or opaque. Lustre adamantine. Hardness. 5 to 5*5. Specific gravity, 3-2 to 3*6. Transparent yellow, bead with borax when hot. Fusible with difficulty on edges. 83. WOLFRAM. Chemical Composition. Tungstic acid, iron, manga- nese. Crystals or masses. Brownish black. Lustre shining or dull. Opaque. Hardness, 55. Specific gravity, 7 to 8. Fusible with difficulty. With borax gives the colour of iron. Characterised by its great weight. Found often with tin ores. 84. PITCHBLENDE. 'Chemical Composition. Uranium, 84'S ; oxygen, 15*2.. Massive. Black, opaque. Resinous lustre. Streak black. Hardness, 5-5. Specific gravity, 6*4 to 6*7. Infusible. With borax gives yellow bead. Dissolves in hot nitric acid, forming a yellow solution. DISCRIMINATION OF MINERALS. 279 85. MAGNETIC IRON PYRITES. Chemical Composition. Iron, 60'5 ; sulphur, 39-5. Usually massive. Bronze yellow, tarnished. Metallic lustre. Feebly magnetic. Hardness, 3*5 to 4-5. Fusible. DETERMINATION OF MINERALS. The following scheme* will show how a mineral may be determined by means of the tests and characters already indicated. The scheme first separates minerals into two divisions namely, those which possess a metallic lustre, and those which do not ; and these are further subdivided into groups by means of other characters. The various members of each group are enumerated in the tables, which also give the mode of discrimination. In most cases it is easy to decide whether a mineral possesses a lustre like a polished metal, but where the lustre is less decided there is room for a difference of opinion ; besides, a mineral may possess a metallic lustre in some specimens, and not in others. To provide against this, any mineral about which difficulty may arise is placed in both divisions, and even in several groups if necessary. The mineral blende affords an instance. From inspection of the scheme it will be observed that valuable metalliferous minerals possess, as a rule, a metallic lustre, are heavy in proportion ta their bulk, and are easily scratched with a knife. The chief exception is tin ore, which has not a metallic lustre, and which cannot be impressed with a knife. In testing it is necessary to bear in mind that a piece of ore may consist of a single mineral only, either pure, or mixed with a readily distinguishable gangue ; or it may be composed of finely disseminated particles of two or more metalliferous minerals. But by search in localities where such mixtures occur, pure and isolated specimens of each * Drawn up by Dr. A. M. Thompson. 280 DETERMINATION OF MINERALS kind may generally be found. Moreover, after a little practice the complication of characters which mixtures present will not interfere with the recognition of their components. The scheme is not expected to prove an infallible means of determining minerals in all possible instances, though previous experience has shown that it is of use to those who have learned to apply the tests recommended in the former part of this chapter. To avoid error, it is necessary in all cases, before coming to a decision, for the observer to be convinced that the mineral under examination ac- tually possesses all the characters of the mineral which it is supposed to be. Serious mistakes may arise by drawing inferences from the scheme only, and neglecting to confirm all the other characters. In performing the experiments recommended, the greatest care and accuracy are needful ; to the careless and superficial observer, or whoever is not proof against self-deception, the scheme can be of little use. The characters of each mineral have been described in the preceding pages. SCHEME FOR THE DETERMINATION OF MINERALS. DIVISION I. Minerals possessing a METALLIC lustre. Experiment. Heat a fragment of the mineral supported on charcoal, or held with a forceps, in the oxidising point of the blowpipe flame. 1. An odour of burning sulphur is given off. (See Group I.) 2. An odour of garlic is given off, or white fumes not having an odour of sulphur. (See Group II.) 3. No fumes or odour given off. (See Group III.) Note. An odour of horse-radish indicates Selenium (78). DIVISION II. Minerals which do NOT poss'ess a metallic lustre. Experiment. Scratch the mineral with a knife or sharp BEFORE THE BLOWPIPE. 281 fragment of quartz or sapphire, according to its hardness, and observe the colour of the streak. Note. If the streak is metallic the unmetallic lustre is due to tarnish. See Division I. A. The streak is black or coloured. Experiment. Hold a fragment of the mineral on char- coal, or with a forceps, in the oxidising point of the blowpipe flame. 1. Fumes or odour given off. (See Group IV.) 2. No fumes or odour given off. (See Group V.) B. The streak is uncoloured that is to say, white, or nearly so. Experiment. Try to scratch a smooth surface of the mineral with a splinter of quartz. 3. The mineral can be scratched by quartz. (See Group VI.) 4. The mineral cannot be scratched bv quartz. (See Group VII.) Note 1. The determination must always be made upon pure minerals free from any adhering foreign matter. Note 2. Coal and carbonaceous shale are distinguished from minerals which they may resemble by burning in the blowpipe flame, leaving a more or less copious ash. GROUP I. MINERALS WHICH HAVE A METALLIC LUSTRE, AND WHICH GIVE OFF SULPHUR. Experiment. Try to scratch the mineral with a knife, and observe the streak. I. Not scratched with a knife ; scratches glass witli great facility. Pale brass-yellow or white. Iron pyrites , 41. II. Scratched with a knife ; does not scratch glass easily. * Streak unmetallic. 282 DETERMINATION OF MINERALS a. Brass-yellow. Solution in nitric acid blue. Copper pyrites, 66. b. Black or red. Solution in nitric acid colourless, but on adding salt water a white curd of precipitate is thrown down which blackens on exposure to sunlight. Mineral gives off white fumes before the blowpipe. Antimonial or arsenical silver ores, 75. c. Lead- or steel-grey. Abundant white fumes before the blowpipe. On charcoal with sodium carbonate yields a globule of metallic lead. Lead and antimony sulphides, 77. d. Eed. Wholly volatile before the blowpipe. Mixed with carbonate of soda and heated in an iron spoon over a candle flame, white vapours are given off which may be condensed on a gold coin held a little above the mixture, covering its surface with a brilliant amalgam when rubbed. Cinnabar i 62. Note. The streak of Grey Copper (see below) may be unmetallic, often red. ** Streak metallic. Experiment. Try fusibility in bio wpipe flame. A. Infusible : e. Lead- grey. Very soft. In thin leaves. Tinges the blowpipe flame faint green. Molybdenite, 35. Note. Magnetic Iron Pyrites and Vitreous Copper are fusible with diffi- culty. B. Fusible : (Easily, except k and m.) Experiment. Heat a small piece (size of linseed), free from gangue, on charcoal without adding sodium carbonate,, hold first in oxidising point and finish in reducing point. f. Wholly volatile, emitting abundant white fumes. Mineral has a lead-colour. Crystallised in long prisms, Antimony sulphide, 37. g. Globule of metallic lead remains. Mineral has a lead-colour. Cubical cleavage. Galena, 58. h. Abundant white fumes given off. Mineral has a lead or steel colour ; and when heated on charcoal with sodium carbonate affords a globule of metallic lead. Lead and antimony sulphides, 77. BEFORE THE BLOWPIPE. 283 i. Globule of metallic silver remains (facilitated by adding a little sodium carbonate). Mineral has a black colour. Is soluble in nitric acid ; on adding salt water to the solution a white curd is thrown down which blackens on exposure to sunlight. Silver sulphide, 74. k. Globule of metallic copper remains (facilitated by adding a little sodium carbonate) ; discovered by crushing the residue and flattening it under a hammer. Mineral ha? a black colour. Does not give off copious white fumes before the blowpipe. It forms a blue solution with nitric acid. Vitreous copper, 65. I. Slaggy globule remains, usually attracted by a magnet, not yielding copper when crushed. Mineral has a black or steel colour. It gives off copious white fumes before the blowpipe. It forms a greenish-brown solution with nitric acid. Grey copper + 67. m. Greyish-black magnetic bead remains. Mineral has a bronze or copper colour. Can be scratched with a knife ; is attracted by a magnet ; does not form a blue or green solution with nitric acid. Magnetic iron pyrites, 85. n. Black magnetic globule. Mineral has the form of brass yellow needles. Nickel sulphide, 79. N.B. Before determining the name of a mineral, it is necessary to com- pare it with the description to which the number refers. GBOUP II. MINERALS WHICH HAVE A METALLIC LUSTRE AND WHICH GIVE OFF EITHER AN ODOUR OF GARLIC WITHOUT SULPHUR, OR WHITE FUMES WHICH HAVE NOT A GARLIC OR SULPHUROUS ODOUR. 1. The 1 fumes evolved have an odour of garlic (arsenic). * Not scratched with a knife. Scratches glass. a. White metallic. Eesidue after roasting gives indi- cations of iron, with borax bead. Crystallised in flattened prisms. Mispickel, 42. b. White metallic. Eesidue after roasting gives indi- 284 DETERMINATION OF MINERALS cations of cobalt, with borax bead. Crystallised in octa- hedrons. Smaltine, 53. Note. White Nickel resembles Smaltine, and is found with it. ** Scratched with a knife. Does not scratch glass. c. Pale copper red. Arsenical nickel, 52. d. Carmine red. Arsenical silver ore, 75. e. White (as streak shows), tarnished black ; wholly volatilised by heat. Arsenic, 38. Note. Metallic Bismuth is frequently associated with Arsenic. Possibly Grey Copper and Arsenical Silver may be looked for here ; for their charac- ters see Group I. 2. The furnes evolved have not a garlic odour. Possibly grey copper, antimonial silver ore, lead and antimony sulphides, or cinnabar : but these ores generally give off a faint odour of sulphur in addition to white fumes. For their characters, see Group I. N.B. Before determining the name of a mineral, it is necessary to com- pare it with the description to which the number refers. GROUP III. MINERALS WHICH HAVE A METALLIC LUSTRE, AND WHICH GIVE OFF NO FUMES. I. The mineral is malleable (can be flattened under a hammer). a. Yellow. Gold, 72. b. Eed. Copper, 64. c. White, rusty surface ; strongly attracted by a magnet. Iron, 40. d. White, feebly or not attracted by a magnet. In- fusible. Insoluble in nitric acid. Platinum, 71. e. White, often tarnished ; not attracted by a magnet. Fusible. Soluble in nitric acid ; on adding salt water to the solution, a white curd is thrown down, which blackens on exposure to sunlight. Silver, 73. II. The mineral is brittle (breaks to pieces under a hammer). BEFORE THE BLOWPIPE. 285 Experiment. Observe the colour imparted to a borax bead in oxidising and in reducing. 1. Violet in oxidising, colourless in reducing. a. Manganese ores, 51. 2. Red (hot), yellow (cold), in oxidising, bottle-green in reducing. Scratch the mineral with quartz and observe the colour of the streak. b. Brown streak ; mineral not attracted by a magnet. Brown iron oxide, 46. c. Eed streak ; not attracted by a magnet. Specular iron, 44. d. Black streak ; not attracted by a magnet. Titanic iron, 47. e. Black or brown streak ; strongly attracted by a magnet. Magnetic iron, 43. Note. Possibly Wolfram (83) ; non -magnetic ; streak reddish-brown or black ; remarkable for its great weight. 3. Green (hot], blue (cold), in oxidising, m^in reducing, Red, cuprous oxide, 69. 4. Green in oxidising, green in reducing. /. Chromic iron, 48. 5. Colourless. * Infusible. Minerals HARDER than calc-spar. g. Eeadily soluble in nitric acid. Scratched with a knife. Blende, 55. h. Insoluble in nitric acid. Not scratched with a knife. With potassium cyanide upon charcoal, yields a globule of metallic tin. Tin ore, 34. Minerals SOFTER than calc-spar. i. Black, like black-lead. Specific gravity about 2. Graphite, 20. 286 DETERMINATION OF MINERALS k. Lead-colour. Specific gravity about 4:5. linger blowpipe flame pale green. Molybdenite, ** Fusible. 35. Heated on charcoal forms an incrustation which is red whilst hot, yellow when cold. Bismuth, 36. 6. Note. Green in oxidising, dirty violet in reducing, indicates Entile (81) ; yellow in oxidising, green in reducing (difficult to obtain in a candle flame), afforded by a black mineral, soluble in boiling nitric acid, forming a yellow solution, indicates Pitchblende (84). Note. Mica may possibly be sought for in this group. It is distinguished by its lightness of weight, and its capability of being split into very thin, trans- parent, flexible layers. N.B. Before determining the name of a mineral, it is necessary to com- pare it with the description to which the number refers. GROUP IV. MINERALS WHICH POSSESS AN UNMETALLIC LUSTRE, A COLOURED STREAK, AND WHICH GIVE OFF FUMES OR OB-OUR WHEN HEATED BEFORE THE BLOWPIPE. a. Colour and streak yellow, entirely volatile, burning with a blue flame and sulphurous odour. Sulphur, 33. b. Colour red, orange, or yellow, entirely volatile, burns on charcoal with a blue flame and garlic odour. Arsenic sulphide, 39. c. Streak red. Strong odour of garlic or white fumes (of antimony) when heated. Dissolved in nitric acid ; on adding salt to the solution a white curd is thrown down which blackens on exposure to sunlight. Arsenical or antimonial silver ore, 75. d. Streak red. Wholly volatile before blowpipe. Mixed with sodium carbonate and heated in an iron spoon over a candle flame, vapours are given off which may be condensed on a gold coin held a little above the mixture, covering its surface with a brilliant amalgam when rubbed. Cinnabar, 62. Sfjj. Before determining the name of a mineral, it is necessary to com- pare it with the description to which the number refers. BEFORE THE BLOWPIPE. 287 GEOUP V. MINERALS WHICH POSSESS AN UNMETALLIC LUSTKE, A COLOURED STREAK, BUT WHICH GIVE OFF NO FUMES OR ODOUR BEFORE THE BLOWPIPE. Experiment Observe the colour imparted by the mineral to a borax bead, both in oxidising and reducing. 1. Green (hot), blue (cold), in oxidising ; red in reduc- ing (difficult to obtain). a. Colour of mineral blue or green. Copper carbonate , 70. b. Colour of mineral red. Red cuprous oxide , 69. c. Colour of mineral black. Black cuprous oxide, 68. 2. Green in oxidising, green in reducing. d. Colour of mineral black. Chromic iron, 48. 3. Red (hot), yellow (cold), in oxidising : bottle-green in reducing. e. Colour of mineral green. Greenearth, 40. /. Colour of mineral brown ; blackened during heating. Brown ferric oxide, 46. g. Colour of mineral red ; blackened during heating. Red ferric oxide, 45. Note. Iron Arseniate is green or yellow, Iron Phosphate is blue. 4. Amethyst in oxidising, colourless in reducing. h. Colour of mineral brown or black. Manganese -ores, 51. Note. Manganese Carbonate is rose red or brownish, streak white. 5. Colourless in both oxidising and reducing. i. Easily scratched with a knife. Cleavage. Soluble in nitric acid. Blende, 55. k. Not scratched with a knife. No cleavage. Heated with potassium cyanide upon charcoal yields a malleable globule of metallic tin. Tin ores, 34. Note. Cobalt Bloom (54) has a red colour and streak ; it gives the indi- cations of Cobalt with borax bead. Entile, Wolfram, and Pitchblende might be considered to have an unmetallic lustre. See under Group III. Njg. Before determining the name of a mineral, it is necessary to com- pare it with the description to which the number refers. 283 DETERMINATION OF MINERALS GEOUP VI. MINERALS WHICH HAVE AN UNMETALLIC LUSTRE, AND WHICH ARE SCRATCHED BY QUARTZ, SHOWING A WHITE STREAK. I. Minerals soluble in water (having a taste). a. Taste of common table salt. Salt, 28. b. Sweetish astringent taste. Alum, 29. c. Bitter taste of Epsom salts. Epsomite, 29. d. Cooling taste. Causes vivid combustion when thrown on a piece of red-hot charcoal. Nitre, 30. Note. Zinc (white), Iron (green), and Copper (blue) sulphates have a nauseous metallic taste. II. Minerals insoluble in water ; but soluble in nitric acid (best determined by placing a very little of the powder in a test tube, pouring on it a few drops of the acid, and heating if requisite). N.B. All of these can be scratched witli a knife. * Easily fusible. a. Green, yellow, or brown. Hardness, 3 to 4. Glo- bule of lead, on charcoal with sodium carbonate. No effervescence in nitric acid. Pyromorphite, 61. b. White or grey. Hardness, 3 to 4. Flies violently to pieces in the blowpipe flame. Globule of lead on char- coal with sodium carbonate. Effervesces in nitric acid. Cerusite, 59. Note. White or grey. Effervesces in nitric acid. Specific gravity 4'3. Does not fly to pieces when heated ; yields no metal on charcoal, WITHERITE BARIUM CARBONATE. ** Infusible. c. Hardness, 3. Dissolves with very brisk efferves- cence in cold nitric acid. Does not crumble to powder in the blowpipe flame. Calc-spar, 24. d. Hardness, 3- 5 to 4. Dissolves like calc-spar ; but crumbles to powder in the blowpipe flame. Aragonite, 6. e. Hardness, 3 -5 to 4. Dissolves much slower than calc-spar ; but difficult to distinguish without chemical analysis. Dolomite, 26. BEFOEE THE BLOWPIPE. 289 f. Hardness, 4 to 5. No effervescence in cold nitric acid, very slight in hot. Magnesite, 25. g. Hardness, 3'5 to 4*5. Effervescence in hot nitric acid. The mineral blackens when heated in the blowpipe flame ; it gives the indication of iron with borax. Iron carbonate, 50. h. Hardness, 5. Effervescence in hot nitric acid. Heated on charcoal w r ith sodium carbonate it forms a crust which is yellow whilst hot. Ckdamine, 56. i. Hardness, 5. Dissolves in nitric acid without effer- vescence, leaving a jelly of silica. In blowpipe flame it shines with a green light. Zinc silicate. 57. j. Hardness, 3-5. Effervesces in hot nitric acid. Flies violently to pieces in the blowpipe flame. On charcoal with sodium carbonate yields a globule of lead. Cerusite, 59. Jc. Hardness, 3- 5 to 4. Brilliant cleavage. Eapidly soluble in nitric acid with effervescence. Heavy (specific gravity, 4). Blende, 55. 1. Hardness, 4- 5 to 5. Slowly soluble in nitric acid without effervescence. Apatite, 22. Note. Hardness, 3'5. Effervesces in hot nitric acid. Tinges the blow- pipe flame crimson. Strontianite (Strontium carbonate). HI. Minerals insoluble in both water and nitric acid. * Easily fusible. HARDNESS. TO to 1*5. Like wax. On charcoal with sodium carbon- ate yields a globule of silver. Horn silver, 76. 2-5 to 3-0. White or grey. On charcoal with sodium carbonate yields a globule of lead. Angle- site, 60. 3-5 to 4-0. Green or brown. On charcoal with sodium carbonate yields a globule of lead. Pyro- morphite, 61. 3*5 to 6'5. Usually white. Swell up before the blowpipe. Occur in cavities of rocks. Zeolites, 15. 5-0 to 6-0. Black, green, or white. Prisms, fibrous or flax-like. Hornblende and augite, 6. u 290 DETERMINATION OF MINERALS HARDNESS. 6 -5 to 7 *5. Usually red. Crystallised in dodecahedrons. Garnet, 9. Note. Glass (artificial) may be included here ; easily distin- guished by its appearance, easy fusibility, and inferior hard ness, 4*5 to 5*5. ** Infusible, or not easily fusible. rO to 1-5. Light green. In layers. Unctuous. Talc, 3. 1-5. Dark green. In layers or granular. Some- times fusible to a black glassy bead. Chlo- rite , 4. 2'0 to 3*5. Green. Turns white in blowpipe flame. Feels often greasy. Serpentine, 5. 2-0. Usually white. After heating becomes opaque, and can be rubbed to powder between the fingers. Fusible with difficulty. Gypsum, 31. 2-0 to 2-5. Grey or black. Can be split into extremely thin semi-transparent flexible layers. Mica, 14. 3-0 to 3-5. White or grey. Tabular crystals or massive. Heavy (specific gravity, 4-3 to 4-8). Splinters fly off crystals when heated. Fusible with difficulty. Heavy spar, 32. 4-0. Cubes or massive. Octahedral cleavage. Flies to pieces when heated. Specific gravity, 3 to 3-2. Fusible with difficulty. Fluor-spar, 23. 5-0. Six-sided prisms, massive. Green or white. Specific gravity, 3 to 3 -3. Generally flies to pieces when heated. In fine powder, slowly soluble in nitric acid. Apatite, 22. 5-0 to 6-0. Prisms, fibrous or flax-like. Cleavage. Green, black, or white. Generally fusible. Specific gravity, 2-9 to 3-5. Hornblende and augite, 6. Note. Possibly Epidote. Yellowish green. Hardness, 6'0 to 7-0. 6-0. Prisms or massive. Flesh-colour, white, or tinted. Cleavage. Specific gravity, 2-3 to 2-8. Felspar, 13. BEFORE THE BLOWPIPE. 291 HARDNESS. () to 7*0. Masses or grains like glass. Black or green. Specific gravity, 3'3 to 3'6. Chrysolite, 7. 6-5 to 7' 5. Dodecahedrons. Usually red. Generally fus- ible. Eed infusible varieties impart a green colour to borax bead. Specific gravity, 3*5 to 4-3. Garnet, 9. 5-5 to 6'5. Never crystallised. White, tinted, or chatoyant. Specific gravity, 1-9 to 2-3. Opal, 2. 7'0. Six-sided prisms. Specific gravity, 2 -6. Quartz, 1. Note. Spliene (82) occurs in acute thin crystals. Yellow, green, or brown. Specific gravity, 3'2 to 3'6. Chiastolite and other Aluminium Silicates occur in prisms. Brown or white. Opaque. Hardness, 6 to 7'5. Infusible. Specific gravity, 3 to 3*7. N.B. Before determining the name of a mineral, it is neces- sary to compare it with the description to which the num- ber refers. GROUP VIL Including the Gems. MINERALS WHICH HAVE AN UNMETALLIC LUSTRE, AND WHICH ARE NOT SCRATCHED BY QUARTZ. Note. Tin ore (34) may occur here. Hardness, 6 to 7. Bemarkable for its great weight. Specific gravity, 6'8 to 7'0. Brown or black. With cya- nide of potassium, on charcoal, yields a globule of Tin. GEMS. If crystallised, the determination is greatly facilitated, but if the crystallisation is not evident, the hardness will be found a sufficiently near indication in the following scheme. The requisites in determining the hardness of gems are good specimens of sapphire, topaz, and quartz possessing smooth bright surfaces, as well as sharp points or corners. The hardness is determined by drawing the sharp points of these three test-minerals over smooth bright surfaces of the mineral under trial. (See hardness.) Or it may be ascertained with great certainty by means of a small angular fragment broken off the mineral to be tried. For convenience of holding, it should be mounted on a stick of sealing-wax, which is best done by previously heating the fragment, held with a forceps over a candle flame, and applying it, whilst hot, to the wax ; the hard U 2 292 DETERMINATION OF MINERALS ness may then be ascertained by drawing the fragment thus mounted across smooth surfaces of the sapphire, topaz, and quartz, and observing whether a scratch is thus produced. In this way the hardness of a mineral not larger than a grain of sand may be determined. The facility of cleavage is an important character in the topaz and the diamond. The electrical properties are characteristic in the dia- mond, topaz, and tourmaline. The specific gravity is a most important aid in distin- guishing gems, but it cannot be attempted with an ordinary pair of grain scales in the case of minerals weighing under ten grains at the very least. Many varieties of zircon and spinel, and in some cases the topaz and emerald, can readily be distinguished by the effect of heat, which is applied most conveniently by the blowpipe. The diamond may be easily distinguished by the use of a small writing or scratching diamond, which fails to mark the faces of a real diamond when drawn lightly across, but scratches all other gems with facility. After practice, the sound caused by tapping two diamonds together becomes characteristic. 1. CRYSTALLISED IN PRISMS. HARDNESS. 7*0. Six-sided prisms or pyramids ; sides of prism finely marked across. No cleavage. Frac- ture fconchoidal. Infusible. When two pieces are Drubbed together in the dark, they emit a phosphorescent light and a peculiar odour. Specific gravity, 2 -5 to 2*8. Quartz, 1. 7*0 to 8*0. Prisms, three, six, nine, or more sided ; fur- rowed lengthwise. Black or coloured. Opaque or transparent. Smooth sides of prisms become electric by friction. No cleavage. Infusible or nearly so. Specific gravity, 3'0 to 3- 3. Tourmaline, 8. BEFORE THE BLOWPIPE. 293 FARDNESS. 7*5. Prisms, four-sided, not furrowed. Wine-red, brown, or white. Decolourised permanently by heat. Adamantine cleavage, but rather difficult to obtain. Infusible. Specific gravity, 4 to 5. Zircon, 12, 7 to 8. Prisms, six or more sided. Usually green. Imperfect cleavage. Fusible with difficulty in thin edges ; becoming clouded by heat. Specific gravity, 2'7 ; its low specific gravity combined with its hardness is characteristic. Beryl or Emerald, 11. 7'5 to 8. Sides of prisms often finely marked lengthwise. Cleavage across the prisms brilliant and easily obtained. Smooth surfaces become strongly electric by friction. Infusible, but sometimes blistered and altered in colour by heat. Specific gravity, 3-5. Topaz, 10. 8'5. Prisms or tables. Green. Infusible and un- altered by heat. Specific gravity, 3-5 to 3 '8. Chrysoberyl, 18. 9'0. Prisms, six-sided or irregular. Cleavage across the prisms, but difficult to obtain. Fracture irregular. Blue, green, black, red, yellow, brown, or white. Infusible. Easily distin- guished by its great hardness, scratching all other gems except diamond. Specific gravity, 3*9 to 4- 2. Sapphire, ruby, or co- rundum, 16. 2. CRYSTALLISED IN OCTAHEDRONS OR DODECAHEDRONS. HARDNESS. 7'5. See characters above. Zircon, 12. '6-5 to 7-5. Dodecahedrons. Usually red. Most varieties are easily fusible ; red infusible varieties impart a green colour (due to chromium) to borax bead. Specific gravity, 3-1 to 4-3. Garnet, 9. 294 COLOURED FLAMES. HARDNESS. 8-0. Octahedrons. Usually red or black, also blue,, green, yellow, and brown. Some red varie- ties become opaque- and black when heated ; rose-red varieties become green when heated,, but regain their original colour on cooling. Specific gravity, 3 -5 to 4'0. Spinel, 17. 10-0. Octahedrons, dodecahedrons, or modifications of these forms. Crystalline facets often curved. Cleavage perfect. Lustre, brilliant adamantine. Usually colourless or straw- coloured. Not water-worn. Strongly elec- tric by the slightest friction. Specific gravity, 3-5. Diamond, 19. N.B. Before determining the name of a mineral, it is neces- sary to compare it with the description to which the number refers. For further particulars respecting gems and precious stones see the concluding chapter of this volume. COLOURED FLAMES. There are a great number of sub- stances best detected by the colours they impart to the flame of the blowpipe. Indeed, so important is this point that it has been thought advisable to collect all the facts known on this subject into one place, rather than scatter them over the work. These experiments are best made in a dark room, and with a very small flame.* BLUE FLAMES. Large intense blue . Copper chloride. Pale clear blue . . Lead. Light blue . . Arsenic. Blue ... . Selenium. Greenish blue . . Antimony. Blue mixed with green Copper bromide. GREEN FLAMES. Intense emerald green . Thallium. Very dark green, feeble . Ammonia. Dark green . . . Boracic acid. Full green . . . Tellurium. Full green Copper. Emerald green mixed with blue . \ PP er j did ?- ( Copper bromide. Pale green . . . . Phosphoric acid. Very pale apple -green . . . Barium. Intense whitish green . . . Zinc. Bluish green Tin binoxide. * Griffin's ' Blowpipe Analysis,' p. 148. COLOURED FLAMES. 295 YELLOW FLAME. Intense greenish yellow . . . Sodium. RED FLAMES. Intense crimson .... Lithium. Eed Strontium. Keddish purple Calcium. Violet Potassium. Chlorine, combined with copper, gives an intense blue flame. This phenomenon may be produced as follows : Take a piece of thin brass wire, and bend one end of it several times upon itself; place upon this some microcosmic salt, and fuse it until it has acquired a green colour. Then add to it the substance suspected to contain chlorine, and place it in the oxidising flame, just at the point of the blue flame ; if any chloride be present a splendid blue colour will be produced. Lead. The blue colour produced by this metal is readily obtained. Fragments of a mineral must be held in the tongs, and powder may be assayed on charcoal. Arsenic, in the metallic state, gives rise to a light blue flame. Selenium and Antimony, when treated in the same manner, afford characteristic flames. Bromine. If any substance containing bromine be placed in a bead of fused microcosmic salt on the brass wire, and then in the oxidising flame, a bright blue flame with emerald-green edges will be produced. Boracic Acid. The following is Dr. Turner's process for the detection of boracic acid. The substance is to be mixed with a flux composed of 1 part of fluor-spar and 4J parts of potassium bisulphate. This mixture is to be made to adhere to the moistened end of a platinum wire, and held at the point of the blue flame ; at the instant of fusion a dark green flame will be produced. It may also be produced by merely dipping the mineral in sulphuric acid, and exposing it to the blowpipe blast. In case a very small quantity of boracic acid is contained in a mineral, the following process may be employed : The substance must be fused with potassium carbonate on 296 COLOURED FLAMES. charcoal, moistened with a drop or two of sulphuric acid, and then a few drops of alcohol ; the latter will burn with a green flame when exposed to the flame of the blowpipe. Tellurium. The peculiar flame given by this metal is produced by heating a portion of its oxide on charcoal in the reducing flame. Copper. All the compounds of copper, except those in which bromine and chlorine enter, give a beautiful green flame. The soluble salts give it per se, but the insoluble require moistening with sulphuric acid. Iodine and Copper. To the bead of microcosmic salt on the brass wire add any compound containing iodine, and a bright green flame will be produced when the mass is heated in the oxidising flame. Phosphoric Acid. The phosphates, when moistened with sulphuric acid, give a light green tint to the outer flame. Barium. The soluble salts of barium give a light green colour to the outer flame when moistened with water. Zinc, when exposed to the blowpipe flame, burns with an intense whitish-green light. Sodium. Any salt of soda, or substance containing soda, being exposed to the outer flame, gives a brush of intensely coloured flame, of a fine amber or greenish-yellow. Water. Certain minerals containing water give a feeble yellowish tint to the flame due to the presence of a little sodium. Strontium. All the salts of this substance which are soluble in water give a crimson tint to the flame, which does not endure after the substance is fused. Strontium carbonate must be moistened with hydrochloric acid, and strontium sulphate must be reduced to a state of sulphide by ignition with charcoal ; it must then be moistened with hydrochloric acid, after which treatment it will exhibit the characteristic flame. Lithium. All that has been said of strontium applies to lithium, but the coloured flame given by lithium is perma- nent, whilst that afforded by strontium is evanescent. COLOURED FLAMES, 297 Calcium acts as strontium. Potassium, treated as sodium, gives a purplish light. This is, however, very liable to be masked by the intense yellow communicated by small quantities of sodium, which are almost always present with potassium. The potassium flame can be seen even in presence of sodium by looking through cobalt-coloured glass. CHAPTER VIII. VOLUMETRIC ANALYSIS. THE main feature of volumetry is not so much analysis in the proper sense of the term, as the quantitative estima- tion of one principal constituent of a substance. This estimation is done by means of solutions, con- taining a certain quantity of reagents in a certain volume, which is called a standard solution, the quantity used of such solution being measured by graduated tubes (burettes, pipettes, &c.) The reaction of a volumetric analysis can be of three different kinds, according to the reagent used and to the substance to be estimated. 1. The substance to be analysed being an acid or a base, it can be saturated by a suitable standard solution (saturation-analysis, used for acids, potash, soda, &c.) 2. The substance to be assayed maybe precipitated by the standard solution, and the completion of the process is observed when no further precipitate occurs (precipitation- analysis, e.g. Pelouze's copper assay, Gay-Lussac's silver assay). 3. The substance to be estimated becomes, by the standard solution, either oxidised or reduced, and by the performance of this process certain colours will appear or disappear, from which the completion of the process is to be observed (oxidation or reduction-analysis, e.g. Schwarz's copper assay). These processes of volumetric analysis are frequently used in assaying. The principle of volumetric analysis may be fully ex~ plained by the following examples given by Fresenius.* * Fresenius's ' Quantitative Analysis,' fourth edition, p. 76. PEINCIPLES O'F VOLUMETRIC ANALYSIS. 299 ' Suppose we have prepared a solution of chloride of sodium of such a strength that 100 c.c. will exactly pre- cipitate 1 grm. silver from its solution in nitric acid, we can use it to estimate unknown quantities of silver. Let us imagine, for instance, we have an alloy of silver and copper in unknown proportion, we dissolve 1 grm. in nitric acid, and add to the solution our solution of chloride of sodium, drop by drop, until the whole of the silver is thrown down, and an additional drop fails to produce a further precipitate. The amount of silver present may now be calculated from the amount of solution of chloride of sodium used. Thus, supposing we have used 80 c.c., the amount of silver present in the alloy is 80 per cent. ; since, as 100 c.c. of the solution of chloride of sodium will throw down 1 grm. of pure silver (i.e. 100 per cent.), it follows that every c.c. of the chloride of sodium solution corresponds to 1 per cent, of silver. ' Another example. It is well known that iodine and sulphuretted hydrogen cannot exist together ; whenever these two substances are brought in contact decomposi- tion immediately ensues, the hydrogen separating from the sulphur and combining with the iodine (I + HS = HI + S). Hydriodic acid exercises no action on starch paste, whereas the least trace of free iodine colours it blue. Now, if we prepare a solution of iodine (in iodide of potassium) con- taining in 100 c.c. 0-7470 grm. iodine, we may with this decompose exactly O'l grm. sulphuretted hydrogen ; for 17 : 127:: 0-1 : 0-7470. ' Let us suppose, then, we have before us a fluid con- taining an unknown amount of sulphuretted hydrogen, which it is our intention to estimate. We add to it a little starch paste, and then, drop by drop, our solution of iodine, until a persistent blue coloration of the fluid indicates the formation of iodide of starch, and hence the complete decomposition of the sulphuretted hydrogen. The amount of the latter originally present in the fluid may now be readily calculated from the amount of solution of iodine used. Say, for instance, we have used 50 c.c. of iodine solution, the fluid contained originally 0-05 sul- 300 PRINCIPLES OF VOLUMETRIC ANALYSIS. phuretted hydrogen ; since, as we have seen, 100 c.c. of our iodine solution will decompose exactly O'l grm. of that body. ' Solutions of accurately known composition or strength, used for the purposes of volumetric analysis, are called standard solutions. They may be prepared in two ways, viz. (a) by dissolving a weighed quantity of a substance in a definite volume of fluid ; or (b) by first preparing a suitably concentrated solution of the reagent required, and then estimating its exact strength by a series of experiments made with it upon weighed quantities of the body for the estimation of which it is intended to be used. ' In the preparation of standard solutions by method a, a certain definite strength is adopted once for all, which is usually based upon the principle of an exact correspond- ence between the number of grammes of the reagent contained in a litre of the fluid, and the equivalent number of the reagent (H=l). In the case of standard solutions prepared by method b, this may also be easily done, by diluting to the required degree the still somewhat too concentrated solution, after having accurately estimated its strength ; however, as a rule, this latter process is only resorted to in technical analyses, where it is desirable to avoid all calculation. Fluids which contain the equivalent number of grammes of a substance in 1 litre are called normal solutions ; those which contain y 1 ^ of this quantity, decinormal solutions. ' The estimation of a standard solution intended to be used for volumetric analysis is obviously a most im- portant operation ; since any error in this will, of course, necessarily falsify every analysis made with it. In scien- tific and accurate researches it is, therefore, always ad- visable, whenever practicable, to examine the standard solution no matter whether prepared by method a or by method />, with subsequent dilution to the required degree by experim en ting with it upon accurately weighed quantities of the body for the estimation of which it is to be used. PRINCIPLES OF VOLUMETRIC ANALYSIS. 301 ' In the previous remarks no difference has been made between fluids of known composition and those of known power ; and this has hitherto been usual. But by accept- ing the two expressions as synonymous, we take for granted that a fluid exercises a chemical action exactly correspond- ing to the amount of dissolved substance it contains ; that, for instance, a solution of chloride of sodium containing 1 eq. NaCl will precipitate exactly 1 eq. silver. This pre- sumption, however, is very often not absolutely correct. In such cases, of course, it is not merely advisable, but even absolutely necessary, to estimate the strength of the fluid by experiment, although the amount of the reagent it contains may be exactly known, for the power of the fluid can be inferred from its composition only approximately,, and not with perfect exactness. If a standard solution keeps unaltered, this is a great advantage, as it dispenses with the necessity of estimating its strength before every fresh analysis. ' That particular change in the fluid operated upon by means of a standard solution, which marks the completion of the intended decomposition, is termed the FINAL REACTION. This consists either in a change of colour ', as is the case when a solution of permanganate of potash acts upon an acidified solution of protoxide of iron, or a solution of iodine upon a solution of sulphuretted hydrogen mixed with starch paste ; or in the cessation of the formation of a precipitate upon further addition of the standard solution, as is the case when a standard solution of chloride of sodium is used to precipitate silver from its solution in nitric acid ; or in incipient precipitation , as is the case when a standard solution of silver is added to a solution of hydrocyanic acid mixed with an alkali ; or in a change in the action of the examined fluid upon a particular reagent, as is the case when a solution of arsenite of soda is added, drop by drop, to a solution of chloride of lime, until the mixture no longer imparts a blue tint to paper moistened with iodide of potassium and starch paste, &c.' The only condition on which the volumetric system of analysis can be carried on successfully is, that the greatest 302 APPARATUS FOR VOLUMETRIC ANALYSIS. care is exercised with respect to the graduation of the measuring instruments, and the strength and purity of the standard solutions. A very slight error in the analytical process becomes considerably magnified when calculated for pounds, hundredweights, or tons of the substance tested. The end of the operation in this method of analysis is in all cases made apparent to the eye. (Button.) STANDARD SOLUTIONS. A very useful form of bottle for their preservation is the ordinary wash-bottle, or any common bottle fitted with the same arrangement of tubes. The mouth end of the blowing tube should be furnished with a tightly fitting- india-rubber cap to prevent alteration of the standard by evaporation. A similar cap over the point will be useful, although not absolutely necessary, or it may be closed by a small cork fitting over it. Burettes can then be filled with the solution without its frothing, and if the tube which enters the liquid does not reach the bottom of the bottle, the sediment, if any, is not disturbed ; another ad- vantage is that the solution does not come into contact with the cork, nor can any dust enter. The Instruments and Apparatus. The Burette, or graduated tube for delivering the stan- dard solution, may be obtained in a great many forms under the names of their respective inventors, such as Mohr, Gay-Lussac, Binks, &c. ; but as some of these possess a de- cided superiority over others, it is not quite a matter of in- difference which is used, and therefore a slight description of them may not be out of place here. The burette, with india-rubber tube and clip, contrived by Dr. Frederic Mohr of Coblentz, shown in figs. 82 and 83, has the preference over all others for general purposes. The advantages possessed by this instrument are, that its constant upright position enables the operator at once APPARATUS FOR VOLUMETRIC ANALYSIS. 303 to read off the number of degrees of test solution used for any analysis. The quantity of fluid to be delivered can be regulated to the greatest nicety by the pressure of the thumb and finger on the spring clip* and the instrument FIG. 82. FIG. 83. not being held in the hand, there is no chance of increas- ing the bulk of the fluid by the heat of the body, and thus leading to incorrect measurement, as is the case with Binks's or Gay-Lussac's form of instrument. The principal disad- vantage of these two latter forms of burette is, that a correct reading can only be obtained by placing them in an up- right position, and allowing the fluid to find its perfect level. 304 THE BURETTE. FIG. 84. FIG. 85. The preference should, therefore, unhesitatingly be given to Dr. Mohr's burette wherever it can be used ; the greatest drawback to it is that it cannot be used for potassium permanganate, in consequence of its india-rubber tube, which decomposes the solution. We are again indebted to Dr. Mohr for another form of instrument to overcome this difficulty, viz. the foot burette, with india-rubber ball, shown in fig. 84. The flow of liquid from the exit tube can be regu- lated to a great nicety by pressure upon the elastic ball, which is of the ordi- nary kind sold for children, and has two openings, one cemented to the tube with shellac, and the other at the side, over which the thumb is placed when pressed, and on the removal of which it refills itself with air. Gay-Lussac's burette, supported in a wooden foot, may be used instead of the above form, by in- serting a good fitting cork into the open end, through which a small tube bent at right angles is passed. If the burette is held in the right hand, slightly inclined towards the beaker or flask into which the fluid is to be measured, and the mouth applied to the tube, any portion of the solu- tion may be emptied out by the pressure of the breath, and the disadvantage of holding the instrument in a horizontal position, to the great danger of spilling the contents, is avoided ; at the same time the beaker or flask can be held in the left hand and shaken so as to mix the fluids, and by this means the end of the operation is more accurately determined . THE BUEETTE. 305 FIG. 86. FIG. 87. Fig. 85 will show the arrangement here described. Mr. J. Blodget Britton has described, in the ' Chemical News ' for August 5, 1870, a burette for use in determining iron in metals and ores. Figs. 86 and .87 represent two of the kind, but of different patterns, mounted on walnut-wood stands; the former is for metals, and the latter, of smaller capa- city, for ores. Securely fastened to the upright B is a graduated tube A, having its lower part drawn out in the usual manner, but bent outwards at an angle of about 25. E is a piece of cork riveted into a sheet steel spring D, which presses tightly, by means of the latter, against the vent of the tube, c is a thumb- screw passing through the frame and bearing against the spring. The tube of fig. 87 has a capacity of 100 c.c., and is graduated into tenths, or 1000 ; but that of fig.. 86 has a capacity of 150 c.c., and is graduated into twen- tieths, though only at its lower and narrow part. Modes of operating. Place a small narrow-necked funnel in the tube, as shown by the figures ; pour in the solution to be used until it quite reaches the funnel, and then remove the latter to carry away any floating bubbles ; turn the thumbscrew and bring the top line of the solution exactly to the zero line of the scale ; stop the flow, and after- wards touch the point of the cork with a glass rod to take from it any adhering drop. The instrument is then ready x 306 THE PIPETTE. FIG. 88. 50CC 10 CC for use. By means of the thumbscrew the dropping may be controlled with extreme nicety or instantly stopped. Cork, after a little use, becomes quite inert towards potassium permanganate. The occasional application of some pure tallow to the end of the tube and cork will be quite effectual in preventing any of the fluid from running upwards by capillary at- traction. For everyday use in the labo- ratory, as well as for very accurate determinations, this burette will find favour. The Pipette. The pipettes used in volumetric analysis are of two kinds, viz, those which deliver one certain quantity ,only, and those which are graduated so as to deliver various quantities at the discretion of the analyst. In the former kind, or whole pipette, the graduation may be of three kinds ; namely, 1st, in which the fluid is suffered to run out by its own momentum only. 2nd, in which it is blown out by the breath. 3rd, in which it is allowed to run out to a definite mark. Of these methods the last is preferable in point of accuracy, and should therefore be adopted if possible. The next best form is that in which the liquid flows out by its own momentum, but in this case the last few drops empty themselves very slowly ; but if the lower end of the pipette be touched against the beaker or other vessel into which the fluid is poured, the flow is hastened considerably, and in graduating the pipette it is preferable to do it on this plan. In both the whole and graduated pipettes the upper COLOEIMETRIC ANALYSIS. 307 end is narrowed to about J inch, so that the pressure of the moistened finger is sufficient to arrest the flow at any point. Fig. 88 shows two whole pipettes, one of small and the other of large capacity, and also a graduated pipette of medium size. The Measuring Flasks. These indispensable instru- ments are made of various capacities ; they serve to mix up standard solutions to a given volume, and also for the subdivision of the substance to be tested by means of the pipettes, and are in many ways most convenient. They should be tolerably wide at the mouth, and have a well - ground glass stopper, and the graduation line should fall just below the middle of the neck, so as to allow room for vshaking up the fluid. Colorimetric Analysis is also used in assaying. It is based upon the fact that a coloured solution appears the more intense the more of the colouring substance it con- tains. If, therefore, a solution containing a certain amount of a substance, and being in consequence of a certain intensity of colour, is prepared, it will be possible to obtain the solu- tion under assay of an equal intensity of colour by appro- priate dilution. By measuring the volume of the assay solution and taking into consideration the amount of the standard solu- tion, the quantity of the substance contained in the assay solution may be readily calculated. 308 CHAPTEE IX. THE ASSAY OF IRON. THE ores of iron, properly so called, always contain the metal in the oxidised state, and in various degrees of purity. The oxides and carbonates are the only minerals of iron which can be used as ores in the blast-furnace. These are associated with different impurities or foreign materials in greater or less proportion. The following is a list of the principal ores of iron, with the maximum percentage of metallic iron that could occur in each if it were absolutely pure, as in its formula : MAGNETIC IRON ORE, Fe 3 4 =Fe 2 3 + FeO, 72-41. RED HEMATITE and specular ore (anhydrous iron, ses- quioxide or ferric oxide), Fe 2 3 , 70-00. BROWN HEMATITE, limonite (brown iron ore), 2Fe 2 3 ,3H 2 0, 59-92. SPATHIC IRON ORE (iron carbonate), FeO,C0 2 , 48-22. TITANIFEROUS ORE (ilmenite), FeO,Ti0 2 , + nFe 2 3 . FRANKLINITE, 3(FeO,ZnO,MnO) + (Fe 2 3 ,Mn 2 3 ), 45-16. Besides these may be mentioned the varieties of impure iron carbonates, known as clay or clay-band ironstone and black-band ironstone. Clay-band ironstone sometimes re- sembles compact limestone, sometimes greyish hardened clay. Its great specific gravity, its effervescing on the addi- tion of an acid, and acquiring a brown-red colour on roast- ing, are sufficient means of identifying it. The following is the result of an analysis of this class of ore by the author ; the specimen was from Ireland, county Leitrim : THE ASSAY OF IRON. 309 Ferrous oxide . 51-653 Ferric oxide 3'742 Manganese oxide -976 Alumina ....... 1-849 Magnesia -284 Lime -410 Potash -274 Soda -372 Sulphur . -214 Phosphoric acid -284 Carbonic acid . 31-142 Silica 6-640 Carbonaceous matter and loss . . . 2-160 100-000 Black-band is a combustible schistose variety of this ore. The following analysis is also by the author : Ferrous oxide ....... 20-924 Ferric oxide ....... -741 Manganese oxide . . . . . . 1-742 Alumina ........ 14*974 Magnesia . . . . . . -987 Lime . . . . . . . -881 Phosphoric acid ...... '114 Silica ..... "'"-. . . 26-179 Sulphur ........ -098 Carbonic acid . ..... . 14-000 Carbonaceous matter ..... 16*940 Water and loss 2-420 100-000 Besides these iron ores the following substances, con- taining iron and used as fluxes, require assaying : Granite, Chlorite, Basalt, Pyroxene; Amphibole, and also some kind of slags (finery cinder, tap cinder, &c.) A. THE ASSAY OF IRON IN THE DRY WAY. Iron ores very seldom occur in a pure state, and the ores may be arranged for their assay in the dry way (and also for smelting) into five classes. 1. Iron ores containing silica, lime, and another base, which ores are fusible per se. 2. Iron ores containing predominantly silica. 3. lime. 3. alumina. 310 THE ASSAY OF IRON IN THE DRY WAY. 5. Iron ores containing a large amount of magnesia ;. these ores are most difficultly rendered fluid. The flux used for assaying (and also melting) varies according to the nature of the predominant compound, and the quantity used according to the amount of that com- pound. If the composition of the ore is known, it is easy to ascertain the amount of flux necessary to form a slag with the bases or silica present ; in most cases an extra quantity of the flux should be added, in order to produce a sufficient volume of slag to cover the button. According to Dr. Percy,* blast-furnace cinder of the following formula may be taken as a type of the kind of slag desirable : Al 2 3 ,Si0 3 2(3CaO,Si0 3 ). Its approximate composition per cent, is as under : Silica ..... 38"| T2 parts Alumina . . . 15 > or about < 1 part Lime .... 47; L3 parts The following mixtures of various fluxes, when fused,, produce a slag which may be regarded as approximating to the above composition : Quartz 1 1 ^| T36-5 per cent. China Clay ". 2 f Silica . \Alumina 0-92 / 0-82 0-82 f J = < 15*5 Lime . . 2 . . 2-5J 148-0 ' Glass . 2* I Silica . ( Materials = Alumina 1-75^ tO-75 \- r35 = 4 16 Lime . , 2^ 2-5 J L50 Shale or fireclay Lime . 3 f Silica . \Alumina . . wn 0-9 ^> 2-5 J f35 - J 17 Us > According to Bodemann a compound of 56 % silica, 30 % lime, and 14 % alumina forms a slag most easily ren- dered fluid, but as it is found that this slag itself is not sufficiently fusible in a small assaying furnace (air-furnace), an addition of fluor-spar is made to the mixture, and in some cases (the iron ore being very difficultly rendered * Percy's ' Metallurgy,' p. 240. t 30 % say of alkalies, liine, &c., on account of their fusibility, are taken as? equivalent to so much alumina. THE ASSAY OF IKO^ IN THE DEY WAY. 311 fluid) some borax is added, or a mixture of borax and fluor- spar. An exact knowledge of the mineralogical properties of the iron ores, and a due experience, will enable the assay er to properly adjust the fluxes without resorting to an analysis to find what amount of silica and bases are present. In some iron-works of Germany the following propor- tions of fluxes are used : For magnetic ore, red hematite ~) r on n/ t, n r /a (very rich) ' ' c " a ^ anc * 25 %fluor-spar argillaceous brown iron ore 20 to 40 30 to 40 bog iron ore ... 50 50 spathose iron ore . . 10 to 15 20 to 25 finery cinder . . . 20 to 25 20 to 25 Air-furnaces are best adapted for assaying iron ores where many assays are required. The furnace should have a cross section of 18 in. x 18 in., and depth to grate-bars 21 inches. Flue 7 in. x 7 in. Anthracite is the best fuel to use ; then coke. In the assay of an iron ore it is required to reduce the oxide of iron to the metallic state, or rather to that of cast iron, to collect it in a button, and to form with the foreign materials of the ore by means of fluxes a fusible slag; that will not retain any of the iron in combination or in the form of pellets. ; Naked crucibles, either* of clay or black-lead, or cru- cibles lined with charcoal, are employed ; the latter are preferable. The button of metal does ,not adhere to naked pots, but the slag adheres very strongly ; so much so that it cannot be detached with any degree. of accuracy for weighing (which in some of M. Berthier's processes is of importance). Black-lead pots allow neither the slag nor button to adhere, but the former dissolves much argilla- ceous matter from the pot, so that its weight is greatly" increased, and the assay cannot be verified. In naked crucibles charcoal must always be added to the assay, to reduce the iron oxide ; in which case, if an excess be.added, it prevents the button from completely forming, so that 312 THE ASSAY OF IRON IN THE DRY WAY. globules remain in the slag (with care this may, however, be avoided). Neither do naked crucibles resist the fire as well as those lined with charcoal, because the lining sup- ports the sides when they soften. The charcoal lining also allows the assay to be finished without adding any reagent to the ore ; the button can be readily taken out, because it does not adhere to the charcoal ; and lastly, the earthy matters in the ore, which have formed a slag, may be col- lected and weighed. If we have added any flux to the ore, the total weight can also be ascertained. The method of lining crucibles with charcoal brasque and conducting the assay is given as follows by Mr. Blossom : * The brasque has a composition of four parts of finely pulverised charcoal to one part of molasses. This must be thoroughly kneaded until a ball of it made in the hands resists, to a sensible degree, an attempt to pull it apart. The crucibles are packed full by driving the brasque in with a mallet ; a conical cavity, of sufficient size for the charge, is then cut out, and the brasque dried in an oven. Care must be taken not to burn the molasses, for the brasque would in that case crumble, and be useless. These are the best crucibles for iron assays, because they combine the following advantages : Being lined with charcoal, none need be mixed with the charge to reduce the oxide, whereas in naked crucibles charcoal must be added, and is liable to prevent the com- plete collection of the iron in a button by holding little pellets in suspension. The slag neither adheres to a charcoal lining nor takes up any material from it, while it does adhere to ordinary naked crucibles, and dissolves argillaceous matter from black-lead crucibles. In the former case, then, the slag may be weighed as a verification of the assay, while in the latter this is, of course, impossible. The lining serves as a support for the crucible, which, under the high heat employed, is very apt to be softened and crushed beneath the weight of the fuel. * ' Chemical News,' April 5, 1872. THE ASSAY OF IRON IN THE DRY WAY. 313 The Charge. In making up the charge it is only necessary to con- sider the materials required as fluxes for the foreign matters of the ores. It may be well to sprinkle a little charcoal into the charge as a precaution, but none is abso- lutely required. Two cases may arise, in which we have (1) ores of unknown composition, and (2) ores previously analysed. The assay in both cases gives us a clue to the nature of the iron that may be obtained from the ore, and to the character and proportion of the fluxes to be added in the blast-furnace, in order that we may produce a fusible slag free from iron. In the former case we obtain the additional information of the approximate percentage of iron, though the iron assay is seldom, if ever, made for this purpose. Eecourse is had to the more accurate chemical analysis, which gives us the exact proportions of the substances which affect the iron injuriously or other- wise. In all the assays a constant weight of ore, 300 grains, is taken. 1. Ores of Unknown Composition. In the assay of an ore the composition of which is unknown, we employ one or more preliminary assays, and, if satisfactory results be not obtained from either, we make another assay with a charge modified according to the indications of the preliminary assay. The follow- ing charges may be used to advantage in the preliminary assay : Preliminary Assay Charges. i. Ii. m. iv. Silica .... 75 30 120 75 grains. Lime .... 75 120 45 75 Ore . . . . 300 300 300 300 1. The first charge is employed for the purer ores, those containing very little gangue such as some varieties of magnetic ore, red and brown hematites, specular and micaceous ores. 314 THE ASSAY OF IRON IN THE DRY WAY. 2. Ores containing silica ; some varieties of brown hematite, magnetic ore, &c. 3. Ores containing lime, magnesia, or protoxide of man- ganese carbonates, &c. ; calcareous hematites, spathic iron. 4. Ores containing silica and alumina ; clay ironstones, black-band, &c. The principle involved in all the charges is that of furnishing for a base, lime ; for an acid, silica ; and vice versa. The choice of a charge, therefore, depends on the acid or basic nature of the gangue of the ore. The mate- rials of the gangue might possibly be associated in such proportions as to flux themselves, but this would happen rarely. Ores containing titanium require the addition of fluor- spar to the charge, in quantity varying from 15 to 300 grains, according to the amount of titanium present. 2. Ores previously Analysed. When we know the percentage composition of an ore r it is a very simple matter to calculate a charge for the dry assay. Good results are obtained from a charge so propor- tioned as to yield a slag corresponding to the following formula of a blast-furnace cinder, as given by Percy : H 2 3 represents alumina, and BD, lime, magnesia, and other bases. Its approximate percentage composition is as fol- lows : Silica ...... 38 "| r 2^ parts E 2 O 3 (alumina) . . . . 15 ^> or about < 1 part KO (lime, magnesia, &c.) . . . 47 J [9 parts We have, then, only to establish the latter relation between the component materials of the gangue, to obtain,. on fusion, the above slag. Let us take the following example : 300 grains Ore Difference The Ore contains Per cent. contain Required to be added Silica .:....' 165 - 4'95 25' 20-05 grains. Alumina . . 1-94 5-82 10- 4-18 Lime, MgO, &c. . 4'51 13-53 30- 6-47 THE ASSAY OF IKON IN THE DRY WAY. 315 Silica is supplied by ground quartz. For the bases EO represented in the furnace slag and in the ore by lime, magnesia, &c., we add pure unslaked lime. The alumina is added in the form of kaolin, which may be assumed to contain equal parts of alumina and silica. Allowance must be made in adding silica for that introduced with the kaolin. It happens sometimes that the ore contains more than is required of one of the ingredients of the slags, or the silica introduced with the kaolin may, when added to that already present, increase the quantity beyond the re- quirement. In either case make up a new slag with the excess. The charge having been weighed out, must be tho- roughly mixed on glazed paper ; after placing it in the crucible, the conical cavity is closed with a piece of char- coal, and the whole top of the crucible is covered with a luting of fireclay. The latter is mixed with ^ ^ part fine sand, and is made plastic with borax water. Hair is sometimes employed to prevent the luting from cracking off when dry ; but no trouble is experienced from this source if the luting be properly made and applied. It should not be put on too thick, should be lapped over the edges of the crucible, and thoroughly dried before placing- the crucible in the furnace. .. Four crucibles are introduced at one time, and rest upon two fire-bricks placed one upon the other, to keep the crucibles in the very midst of the glowing coals. If the crucibles do not rest steadily on the bricks, it is well to support them with a little luting, to prevent their being knocked over in the fire. A low fire may be kindled before, the introduction of the crucibles, or it may be kindled around them. The fuel is added gradually until it fills the furnace above the tops of the crucibles ; the fire is then maintained at its maximum temperature for 2-J 3-| hours, according to the refractory nature of the ore. Ores containing much titanium may even require 4 hours, while carbonates containing manganese may fuse well in 2-| hours, or in less time. Three hours will generally be suifi- 316 THE ASSAY OF IRON IN THE DRY WAY. cient for ores that do not contain much titanium. When the fire has burned out, the bricks and crucibles are removed in one mass, cemented together by the slag of the fuel. The crucibles are detached, and the exteriors broken with a hammer ; on inverting and tapping the brasque lining, the slag and the button of cast iron will fall into the hand, when, if they adhere together, a slight tap will suffice to separate them. Before separation, how- ever, they ' should be carefully cleansed and weighed ; if necessary, the slag may then be broken, and any particles of iron it retains mechanically may be extracted with a magnet. The weight of the iron being deducted from the weight of the slag and button, we obtain the weight of the slag. This ought to approximate closely to the weight of the fluxes introduced and the corresponding material of the ore. If a large amount of iron has combined with the slag it will be indicated by the excess in weight. Titanium and manganese enter the slag almost completely ; hence if they are present, allowance must be made for them. Duplicate assays are made, and the two results should not differ more than 0'3 0-4 of one per cent. The slag ought to be well fused, colourless, transparent, and vitreous, or white, light-grey, bluish-grey , opaque, and semi-vitreous, resembling porcelain or enamel. A good button will be well formed, and will separate completely from the slag. If the metal be of good quality, the button, when. wrapped in a piece of thin tin-plate, and struck on the anvil, will flatten slightly before breaking. It ought to be grey or greyish-white, and the grain fine, or tolerably fine. A button of bad iron breaks readily without changing form, sometimes even pulverising : the metal is generally white and crystalline on the broken surface. The following are some of the characters that may be observed in slags, and their indications with reference to the charge : A perfectly transparent slag of greenish tint indicates an excess of silica. A stony rough slag, or one that is crystalline in fracture and dull in lustre, indicates an excess of bases. ASSAY OF IRON IN THE DRY WAY. 317 If the product, instead of being melted, is only fritted, and contains the reduced iron interspersed as a fine grey powder, both silica and alumina are deficient in the flux, lime and magnesia being in excess. The latter is one of the most refractory substances found in iron ores, and, when present in quantity, requires an addition of both silica and lime. A vesicular slag, with the iron interspersed in malleable scales, indicates the presence in the ore, of iron and manganese silicates, or an excess of silica, which react on the carburetted iron as it forms, producing malleable iron and carbonic acid. This trouble is corrected by the addition of lime. Manganese in small quantity gives an amethystine tint to the slag ; in larger proportion it produces a yellow, green, or brown colour. Titanium often produces a resinous, black, and scoria- ceous slag, sometimes curiously wrinkled on the outside. It is covered on the outside with a metallic pellicle of the cyano-nitride of titanium with its characteristic copper colour ; sometimes the slag is vitreous and of a bluish tint. Chromium gives a dark resinous slag, sometimes surrounded by a thin metallic coating. The following are some characters of the button depend- ent on the substances named : Phosphorus. A hard, brittle, white metal, known as cold-short. Sulphur. A strong, reticulated, mottled structure, and red-short iron. Manganese. A button smooth exteriorly, hard and non-graphitic : it breaks under the hammer, and presents a white crystalline fracture. Titanium. The button is smooth on the outside, and breaks under the hammer with a deep grey fracture, dull or crystalline. It adheres strongly to the slag. The button is covered sometimes with titanium cyano-nitride with its characteristic copper colour. Titanium is said to increase the strength of the metal. It may be present to the extent of one per cent. Chromium. Sometimes the button is smooth, well 318 THE ASSAY OF IRON IN THE DRY WAY. fused, with a brilliant crystalline fracture and tin- white colour ; at other times it is white, only half fused, or it may even form a spongy mass of a clear grey colour, according to the quantity of chromium contained in the iron. Many alloys of iron and chromium will scratch glass. Berthier recommends the following method for esti- mating the other, chiefly slag-forming, components of iron ores. The operations of this method are comprised in roasting or calcining, to drive off any volatile or combust- ible matters, and in treating the ore with certain acids, the object of which is to ascertain the amount of insoluble matter, by difference of weight, before and after the action has taken place. The hydrated ores are calcined to estimate water ; and those containing manganese, to reduce it to a fixed and known state of oxidation (sesquioxide). The carbonates are roasted to expel carbonic acid, and the ores from the coal formations to burn the combustible matter with which they are mixed. Slags and dross are also roasted to free them from charcoal. A simple calcination sometimes is sufficient, as in the case of carbonates ; but where mixtures of ferric and ferrous oxides are to be assayed, they must be subjected to a long roasting, in order to convert all the contained ferrous oxide into ferric oxide. Diluted and cold nitric or acetic acid is employed for minerals whose matrix is purely calcareous or magne- sian, as these acids dissolve the earthy carbonates, without attacking either stones, clay, or the iron oxides. The residue is to be well washed, dried, and weighed, and the amount of carbonates calculated by the difference. It is now to be treated with boiling hydrochloric acid, or, what is preferable, aqua regia. The ores which contain sub- stances insoluble in these acids are generally of a clayey or flinty nature. These are to be weighed, and according to their weight that of the flux to be added in the assay is estimated, as will be shown hereafter. It must be borne in mind, however, that the clays THE ASSAY OF IRON IN THE DRY WAY. 319 are not absolutely insoluble in hydrochloric acid, for a certain quantity of alumina is always dissolved, which is generally greater in proportion to the amount present in the iron ore. The ores containing titanium are boiled with concen- trated sulphuric acid, after they have been reduced to the finest possible state of division. All the iron, titanium, and manganese oxides are dissolved, and the stony gangues which resist the action of this acid can be estimated. The utility of this estimation will be pointed out as we proceed. When all the operations necessary for each particular case have been completed, we know the proportion of vola- tile substances, of substances soluble in acetic acid, and those insoluble in hydrochloric and sulphuric acids, con- tained in the substance under assay. Let A be the weight of the rough or non-calcined ore ; B the weight of the same calcined ; C the weight of the fluxes in a rough state ; D the weight of the same calcined ; P the weight of matter insoluble in hydrochloric or sul- phuric acid ; E the weight of the fixed substances soluble in acetic or nitric acid a weight which can be readily calculated when we know the loss which the ore, not treated by acids, suffers by calcination, and the residue of the treatment of this substance by acetic or nitric acid ; M the weight of the button of metal and scattered globules ; S the weight of the slag ; and the loss of weight in the assay, which represents the quantity of oxygen disengaged during the reduction. The following is the disposition of the data from which, at one view, all the useful results of the assay can be de- termined. In the assay has been employed A, rough ore = calcined ore . . . . B B, of rough fluxes added = fixed flux . . D Total of fixed matter B + D The result has been 320 THE ASSAY OF IRON IX THE DRY WAY. Metal M\ T , i M Slag-S / T tal M + Loss Fluxes D Verifiable matters S D Substances insoluble in hydrochloric acid, &c. . T Substances soluble in hydrochloric acid, &c. . S D T Substances soluble in acetic acid ... R, Substances insoluble in acetic acid, and solu- ble in hydrochloric acid .... S D T R When the iron in the substance assayed is in a known degree of oxidation, and when but little manganese is present, the quantity of oxygen ought to correspond very nearly with the quantity of metal M produced : if it does, the assay must be correct. A rigorous correspondence between the two numbers, however, cannot always be obtained, because the iron is not pure, but always contains carbon, so that in ordinary assays the ferric oxide loses but from twenty-eight to twenty-nine per cent, of oxygen. On the other hand, the quantity of iron remaining in the slag makes up in part for the carbon combined with the metal reduced ; but when the assay has been made with a suitable flux, the quantity of oxide remaining is very small, and never exceeds one per cent, of the weight of the slag. When the iron is in an unknown degree of oxidation, the loss produced in the assay gives the degree, if it has been made without accident ; but if there is any doubt, and the result is of importance, the assay must be recommenced for verification. If the ferruginous matter contain manganese, and if that metal be in the state of protoxide, the verification just described can be made without modification, because the manganese dis- solved in the slag is always at the minimum of oxidation ; and when a sufficient quantity of flux is employed, the amount reduced is of no consequence. But when the manganese is in the state of red oxide, it parts with a certain quantity of oxygen on being reduced to the minimum of oxidation f quantity estimated in the loss 0), so that a perfectly accurate verification cannot be made. Nevertheless, the difference between the loss THE ASSAY OF IKON IN THE WET WAY. 821 0, and the quantity of oxygen calculated from the metal M, cannot be very great, because the red oxide of man- ganese loses but *068 of oxygen in its transformation to protoxide. If the assay has been made with care, the loss of oxygen indicates the amount of iron in a very approxi- mate manner, and nearly always with an exactitude which is surprising to those not accustomed to this kind of operation. Titanic acid behaves in iron assays exactly as the oxides of manganese ; it disengages at most but -06 of oxygen when dissolved in the earthy glasses in contact with charcoal. B. THE ASSAY OF IRON AND ITS OEES IN THE WET WAY. This will be subdivided into the following sections : a. Assay of iron ores, pig iron, and steel, for the metallic iron they contain. $. Complete assay of iron ores. 7. Estimation of carbon, sulphur, phosphorus, silicon, &c., or metallic iron and steel. a. Dr. Penny's Process. The following method of estimating the amount of iron in a sample by means of a normal solution has been contrived by Dr. F. Penny, who was led to substitute potassium bichromate for potassium permanganate, as recommended by Marguerite. The reason of employing the bichromate is that it is an unchangeable salt, whilst the permanganate sometimes undergoes decomposition, so that its strength is variable, and each series of experiments made with it requires a separate verification by means of a weighed quantity of pure iron. This inconvenience is avoided in Dr. Penny's method, which is described in his own words as under : . ' I shall proceed to describe the method of employ- ing the potassium bichromate for the determination of the amount of iron in clay-band and black-band ironstone. I shall be purposely minute, as I particularly desire that the process may be serviceable to those who, from their Y 322 THE ASSAY OF IRON IN THE WET WAY. pursuits in life, are interested in the value and quality of ironstone, and who may be imperfectly acquainted with analytical operations. 4 A convenient quantity of the specimen is reduced to coarse powder, and one-half at least of this still further pulverised, until it is no longer gritty between the fingers. The test solution of potassium bichromate is next prepared. 44*4 grains of the salt in fine powder are weighed out, and put into an alkalimeter (graduated into 100 equal divisions), and tepid distilled water afterwards poured in until the instrument is filled to 0. The palm of the hand is then securely placed on the top, and the contents agitated by repeatedly inverting the instrument, until the salt is dis- solved and the solution rendered of uniform density throughout. It is obvious that each division of the solution thus prepared contains 0-444 grain of bichromate, which corresponds to ^ a grain of metallic iron. The potassium bichromate used for this process must of course be purchased pure, or made so by repeated crystallisation, and it should be thoroughly dried by being heated to incipient fusion. ' 100 grains of the pulverised ironstone are now intro- duced into a Florence flask, with 1^ oz. by measure of strong hydrochloric acid, and \ an ounce of distilled water. Heat is cautiously applied, and the mixture occa- sionally agitated, until the effervescence caused by the escape of the carbonic acid ceases ; the heat is then increased, and the mixture made to boil, and kept at moderate ebullition for ten minutes or a quarter of an hour. During these operations it will be advisable to incline the flask, in order to avoid the projection, and consequent loss, of any portion of the liquid by spirting. About 6 oz. of water are next added, and mixed with the contents of the flask, and the whole rapidly transferred to an evaporating basin. The flask is rinsed several times with water, to remove all adhering solution. '.Several small portions of a weak solution of pure red potassium prussiate (containing one part of the salt to 40 of water) are now dropped upon a white porcelain slab, DE. PENNY'S PROCESS. 323 which is conveniently placed for testing the solution in the basin during the next operation. ' The prepared solution of potassium bichromate in the alkalimeter is then added very cautiously to the solution of iron, which must be repeatedly stirred, and as soon as it assumes a dark greenish shade it should be occasionally tested with the red potassium prussiate. This may be -easily done by taking out a small quantity on the top of a glass rod, and mixing it with a drop of the solution on a porcelain slab. When it is noticed that the last drop communicates a distinct red tinge the operation is ter- minated. The alkalimeter is allowed to drain for a few minutes, and the number of divisions in the test-liquor consumed read off. This number multiplied by two gives the amount of iron per cent, in the specimen of ironstone, assuming that, as directed, 100 grains have been used for the experiment. The necessary calculation for ascertain- ing the corresponding quantity of protoxide is obvious. ' When black-band ironstone is the subject of ana- lysis, or when the ore affords indication, by its appear- ance or during the treatment with hydrochloric acid, that it contains an appreciable quantity of carbonaceous matter, then the hydrochloric acid solution must be filtered before being transferred to the basin, and the filter, with the insoluble ingredients, must be washed in the usual way with warm distilled water, slightly acidu- lated with hydrochloric acid until the filtrate ceases to give a blue colour with the red potassium prussiate. In those cases, also, where the presence of iron pyrites in the ironstone is suspected, it will be necessary to remove the insoluble matter by filtering before applying the bichromate solution ; but with ironstones in which the insoluble ingredients are merely clay and silica, filtration is not essential. 6 Now it is evident" that the foregoing process, so far as I have described it, serves for the determination of that portion of iron only which exists in the ore in the state of protoxide. But many specimens of the common iron- stone of this country contain appreciable quantities of T 2 324 THE ASSAY OF IRON IN THE WET WAY. ferric oxide, which, being unacted upon by the potas- sium bichromate, would escape estimation by the present method. By an additional and easy operation, however, the amount of metallic iron in the ingredient may be likewise determined. It is only necessary to reduce it to the minimum state of oxidation and then to add the bichromate as previously directed. ' The best and most convenient agent for effecting the reduction of the ferric oxide is sodium sulphite. The only precaution to be observed is to use it in sufficient quantity, and at the same time to take care that the iron solution contains excess of acid. When the reduction is complete, a few minutes' ebullition suffices to decompose the excess of sodium sulphite, and effectually to expel every trace of sulphurous acid. ' In order to test the exactness of this mode of estimating the iron of the peroxide, I made several ex- periments with peroxide prepared from known quantities of pure iron wire. The peroxide was thoroughly washed, dissolved in hydrochloric acid, reduced with sulphite of soda, and after complete expulsion of the excess of sulphurous acid, the solution was diluted with water and treated with potassium bichromate. I select three of the experiments : ' Exp. I. 10 grs. of iron consumed 8-87 of bichromate. II. 18 15-94 III. 25 22-15 ; The mean of all my experiments on this point gives the ratio of 100 of iron to 88'6 of bichromate. ' Whenever, therefore, the ore of iron contains ferric oxide it will be necessary to add sodium sulphite to the hydrochloric acid solution before the addition of the test- liquor from the alkalimeter. The sulphite should be dis- solved in distilled water, and added to the solution of iron in small successive portions, until a drop of the liquor gives merely a rose-pink colour with potassium sulpho- cyanide, which indicates that the reduction of the ferric salt is sufficiently perfect. The liquor is now heated till the odour of sulphurous acid is no longer perceptible. DR. PENNY S PROCESS. 325 These operations should be performed while the solution is in the flask, and before it is filtered or transmitted to the basin. ' I will here mention, for the guidance of those who may not be fully aware of the reactions of the oxides of iron, that the existence of an appreciable quantity of peroxide in the ironstone may be readily discovered by dissolving (as directed in the process) 39 or 40 grs. of the ore in hydrochloric acid, diluting with about 8 oz. of water, filtering, and testing a portion of the solution with potassium sulphocyanide. If a decided dark blood-red colour is produced, the quantity of ferric oxide in the stone must be determined ; but if the colour is only light red or rose-pink, the proportion is exceedingly small, and for practical purposes not worth estimating. Of course, when the specimen of ironstone has an ochrey or a reddish appearance on the surface or in the fracture, the presence of a large proportion of ferric oxide is indicated, and its exact quantity must be determined.' The details of Dr. Penny's process have been carefully examined by Mr. R. W. Atkinson, with the result of elimi- nating several small sources of error which interfered with accuracy. In all cases where volumetric methods are employed, the first and most important point to be considered is how to obtain the standard solution invari- ably of an accurately known strength, and it is from a variation in the methods employed for this end that a large amount of the variation in the results is due. Presuming that potassium bichromate is the salt almost universally employed in England, it seems simple enough to weigh out the exact amount of salt and dissolve it in a known volume of pure distilled water. Dr. Penny recommends the fusion of the bichromate for the purpose of driving off com- pletely the water entangled in the crystals ; but if this be done, however carefully the heat be regulated, on dis- solving in water, and allowing to settle, a green deposit of chromic oxide will be found at the bottom of the flask or bottle, showing that a small amount of decomposition takes place during fusion, and that the value of the stan- 326 THE ASSAY OF IKON IN THE WET WAY. dard must be lower than the calculated value. Instead of fusing the salt, it is much better to grind up the very purest crystals and to dry them in a steam oven for several hours before weighing out. But, however the solution be made up, it is always safer to standardise it before use. The great difficulty in the use of salts of iron for stan- dardising is to obtain one which will not alter in composi- tion by keeping for two or three months. Ferrous sulphate is out of the question ; and most specimens of the iron and ammonium sulphate are liable to oxidation, though in a less degree than the ferrous sulphate. We have been for- tunate enough to get from Messrs. Hopkin and Williams a sample of granulated iron and ammonium sulphate, which after several months' use still gives the theoretical percentage of iron, ammonia, and sulphuric acid. Pre- suming that the strength of the bichromate solution in terms of iron is accurately known, a fair sample of the ore having been ground sufficiently fine to pass com- pletely through a sieve of 120 meshes to the linear inch, part of it is dried at 212, and when cold portions of about fifteen or sixteen grains are weighed out for analysis. Unless the balance is very rapid in its action, it will be found necessary to re-dry the portions first weighed out if accurate and concordant results are to be expected.. The ore is in such a very fine state of division that it greedily absorbs moisture from the air during the opera- tion of weighing out : this applies most forcibly to the soft red ores, like Campanil and Vena Dulce, but it is also true of the harder Eubio and other brown ores. The weight of the portion being accurately known, it is trans- ferred to a conical flask, and digested at a gentle heat on a hot plate, with from 150 to 230 grains strong hydro- chloric acid, the flask being closed with a watch-glass. It has been asserted by K. F. Fohr that ferric chloride is volatile at about 212, but this is contrary to experience. It is true that yellow drops of ferric chloride may some- times be seen depending from the under surface of the watch-glass, but this is only the case when the solution is REDUCTION BY ZINC. 327 accompanied by a boiling of the liquid, and is doubtless due to spirting. If the digestion is carried on so as to avoid bubbling, the drops on the cover are always colour- less, and no loss of iron from volatilisation need be feared. When the ore has been completely dissolved, the next step is the reduction from the ferric to the ferrous state, to effect which there are three reducing agents mainly em- ployed, viz. zinc, stannous chloride, and sulphurous acid in the form of one of its salts. These will be dealt with in the order given. Eeduction by means of zinc is capable of giving trustworthy results, provided that pure zinc free from iron is employed ; or if the zinc contains iron, provided that the amount thus introduced be allowed for. But the use of zinc is open to other objections : in the first place, it dissolves only slowly, and thus unduly retards the ope- ration, which, when a number of analyses are to be carried out, is a matter of no small moment. In the second place, the titration is further retarded by the slowness with which the blue colour with potassium ferricyanide is deve- loped, in consequence of the presence of zinc chloride in the solution. And lastly, the colour towards the end of the titration becomes so faint, even when fully developed, that it is impossible to distinguish the presence of an amount of iron less than one or two tenths per cent, of the iron contained in the ore. Consequently, although reduction by means of zinc permits the analyst to obtain uniformly concordant results, without the risk of error, its action is too slow, and its indications are not sufficiently delicate for the most accurate work. But, however much rapidity of work may be an object to be kept in view, there can be no doubt but that reduc- tion by zinc is greatly to be preferred to the second method, viz. reduction by means of stannous chloride. Indeed, the latter method, as usually performed, is open to the grossest abuses, and ought to be prohibited by any authority which may seek to reform the present methods of analysis. One reason why stannous chloride has become a favour- 3-28 THE ASSAY OF IRON IN THE WET WAY. ite reducing agent is the rapidity with which the analysis can be made. Starting with a sample of ore the percentage of moisture and iron in the ore can be found by this method (with the above qualifications as to accuracy) in from two to three hours ; but it is contrary to the best interests of a chemist to seek rapidity of work at the expense of accuracy, and the abandonment of this method of reduction is strongly recommended. Of all methods, reduction of the ferric salt by the use of a concentrated solution of ammonium bisulphate is the most accurate and trustworthy. Sodium bisulphite is some- times used, but is not nearly so satisfactory as the am- monium salt, as it is more difficult to separate the last traces of sulphurous acid from the former than from the latter. The mode of manipulation is as follows : The ore having been dissolved in hydrochloric acid in the conical vessel, the solution is diluted with acidified water and filtered into pear-shaped flasks, the filters being thoroughly washed with hot acid water. The filtered ferric chloride is next carefully neutralised with ammonia, strong at first and afterwards dilute, until a faint reddish precipitate remains permanent. Two or three drops of strong hydrochloric acid are washed round the inner neck of the flask, and as the acid flows down it spreads out, dissolving any par- ticles of ferric hydrate which may have remained on the sides of the flask. When the solution is quite clear, and of a faint reddish colour, 75 to 90 grains of a strong solu- tion of ammonium bisulphite (sp. gr. 1-06) are added, the flask shaken, and boiling water added. On shaking the flask the colour entirely disappears, and the flask is then put over a burner. A small piece of thick platinum wire is introduced to assist the boiling, and about 25 or 30 grains of dilute sulphuric acid (1 acid to 6 water) are added to acidify the solution and to assist in the expulsion of the excess of sulphurous acid. After the liquid is once in a state of ebullition it is kept boiling briskly for thirty minutes (less time is sufficient, but it is always well to err on the safe side), during which time nearly the required amount of bichromate is run out into the dish. At the REDUCTION BY SULPHUROUS ACID. .320 end of the half-hour the boiled solution is added to the potassium bichromate, and the titration is carried out as usual. By proceeding as above there is only one loophole for the introduction of error, viz. in the length of time allowed for boiling off the sulphurous acid ; but if the conditions as given above are fulfilled, constant and accurate results may be relied upon. The above method has this advantage, viz. that the solution is practically one of ammonio-ferrous sulphate, a salt which is one of the most stable of all the ferrous salts. It is, therefore, less liable to become oxi- dised by exposure to the air in transferring to the basin than the acid solution of ferrous chloride obtained by the two previous modes of reduction. A further advantage lies in the fact that the end reaction with potassium ferri- cyanide is beautifully clear and delicate ; so that -there is no difficulty in distinguishing the addition of ^ c.c. of bichro- mate (strength 1 c.c. = O077 grain iron), equivalent to 0-0038 iron. Numerous experiments have shown that perfectly constant results can be obtained by this method, the same percentages of iron in a given ore having been found with different standard bichromate solutions after the lapse of several months. It is rarely the case that three experi- ments carried out simultaneously give percentages of iron differing by more than 0' 05 to 0-07 per cent, of the ore ; but the main advantage which the method of reduction by ammonium bisulphite possesses is that results are found which are perfectly independent of any unconscious bias on the part of the operator, and the author feels convinced that were this method constantly and generally used, we should hear less of differences of two per cent, and three per cent, beween two chemists' analyses of the same sample of ore. The existence of ' sellers ' and ' buyers ' chemists is a disgrace to the profession, and anything which is likely to put an end to the scandal, even in one trade, ought to be welcomed. One other point in the volumetric estimation of iron remains to be noticed. The solution of potassium ferri- cyanide slowly decomposes when the bottle containing it 330 TITRATION OF IRON is exposed to diffused daylight, and a yellowish sediment is deposited. This change is very greatly retarded, if not entirely prevented, by protecting the solution from light by covering the bottle with an inverted tin canister. Con- nected with this decomposition of the ferricyanide solution is the fact, already well known, that when a mixture of that solution with one of ferric chloride is exposed to day- light reduction takes place, and the solution turns blue. But it is not so generally known how rapidly this takes place, and that if the drops of ferricyanide to which the completely oxidised iron solution has been added be allowed to remain exposed to the light (protected from dust by means of a glass plate) for ten or fifteen minutes, a distinct blue tinge will be developed. It is important to remember this, for in titrating the blue colour requires two or three minutes to become fully developed when the amount of ferrous salt remaining in the solution is very small, and in order to prevent the drop turning blue by reduction under the influence of daylight, it is advisable to keep the slab covered with a black cloth or with a flat tin cover. By this means it is possible so to protect the mixture from change that no blueness is perceptible after the lapse of an hour or more, provided that all the iron in the solution to be titrated has been fully oxidised. TITRATION OF IRON WITH SODIUM HYPOSULPHITE. A. C. Oudemans, jun., has proposed to estimate iron in the acid solution of the chloride to which a little solu- tion of copper sulphate and potassium sulpho-cyanide has been added by dropping in a solution of sodium hyposul- phite of known strength until the red colour of the iron sulpho-cyanide has disappeared, and estimating the excess of hyposulphite by titrating back with iodine solution. A. E. Haswell modifies this method so as to avoid the possibly disturbing separation of copper sulpho-cyanide, and to dispense with the back titration with iodine. Ac- cording to his experiments, the iodine-starch reaction often takes place too early, before all the hyposulphite has been converted into sodium tetrathionate. He explains this WITH SODIUM HYPOSULPHITE. 331 occurrence by the tendency of cupric iodide to split up into cuprous iodide and free iodine, and thus produce a premature blue coloration which after a time disappears again as the cuprous iodide recornbines wdth the free iodine to form the cupric iodide. Haswell mixes the moderately acid solution of ferric chloride in presence of a cupric salt with a few drops of a dilute solution of sodium salicylate, and then reduces with sodium hyposulphite. The deep violet colour of the solution fades gradually and becomes colourless in presence of a very slight excess of the reducing agent. The excess of sodium hyposulphite is then oxidised with a dilute solu- tion of sodium bichromate. The limit of the reduction is sharply marked by the faint violet colour which indicates the oxidation of a trace of the iron. It must be remem- bered that strong hydrochloric acid destroys the colour produced by salicylic acid in ferric chloride, which, how- ever, is restored on moderate dilution with water. For the execution of the method there are required a solution of sodium hyposulphite, standardised by means of a solution of ferric chloride of known strength ; a solution of potassium bichromate about half the strength of the sodium hyposulphite ; a solution of copper, prepared by dissolving two grammes cupric ammonium chloride in 100 c.c. water ; and a solution of sodium salicylate con- taining about 5 grammes of the salt per litre. Five or ten c.c. of the iron solution are measured into a small flask, slightly acidulated with hydrochloric acid, and mixed with 1 to 2 c.c. of the copper solution, and a few drops of the sodium salicylate. If the colour resulting is not a pure violet, but an olive brown, the liquid is diluted with water and the hyposulphite is added until the liquid appears perfectly colourless on standing with the back to the window and looking through the flask at a sheet of white paper. It often happens that on adding more sodium salicylate a faint coloration reappears, but it is removed by a drop of .hyposulphite. It is then titrated back with the bichromate until a faint viftlet coloration appears. 332 COMPLETE ASSAY OF IRON ORES. Complete Assay of Iron Ores. The methods employed in the analysis of iron ores have been thoroughly investi- gated by Mr. A. A. Blair. The following account of the method he recommends is taken from ' Mining Industries of the United States.' The treatment of samples of iron ore naturally divides itself into two parts, the mechanical and the chemical, and it will be described under these heads. The care with which the identity of every sample is preserved throughout will be shown, and the methods by which the estimation of the different elements is rendered as accurate as the state of chemical knowledge would allow will be given in detail. THE MECHANICAL TREATMENT. The sample is first taken from the bag and placed upon a large piece of clean strong paper, and the label removed from the box and put in a little card-rack fixed over the bench. Half a dozen small chips, representing, as nearly as possible, the different varieties of the ore, are put aside and carefully labelled, to be used for specific gravity estimations and for thin sections for the micro- scope. The large steel mortar, fig. 89, is half filled with the ore, the leather cover adjusted, and the machinery started- This mortar was cast at the Mid vale Steel Works of Phila- delphia, and was made of an exceptionally fine quality of propeller steel, containing over one per cent, of carbon and about fifteen-thousandths of one per cent, of phosphorus. The pestle was forged and hardened, and the wear, after crushing between 15,000 and 20,000 pounds of ore, was scarcely perceptible in either mortar or pestle. The mortar weighs about 70 pounds, and the pestle with the stem and weight about 25 pounds. The tappets A are faced with raw hide, which stand the wear of the cams, H, remarkably well, much better than either hard or soft iron or steel, the dust from the ore causing the latter to cut very fast. The shaft makes about 90 revolutions per minute, and the ore, when the mortar is about half filled, feeds itself, so THE MECHANICAL TREATMENT. 333 that without any attention 25 pounds of hard ore in lumps are reduced almost to powder in .about one and a half hours. The stem B is hooked up, so that the pestle clears the FIG. 89. top of the mortar A. The pulley D raises the mortar clear of the block, and by means of the traveller E the mortar is emptied on the chilled iron plate F, as figured in the sketch. The ore is ground to a fine powder on this chilled plate by the hardened steel muller c, any of it falling off the plate being caught in the sheet-iron troughs G. It is thoroughly mixed and quartered, and the result- ing sample, reduced finally to about 6 to 8 ounces, is placed in a clean dry bottle. This bottle has the number of the sample etched upon it, the same number in figures half an inch high pasted on the neck, and the label which came in the sample bag pasted on its side. The bottle is not taken into the grinding-room until the sample is being ground, and is always previously numbered, so that when it is brought upstairs with the label the num- bers can be compared and the label pasted on. A por- 334 COMPLETE ASSAY OF IRON ORES. tion of the sample is ground in the agate mortar A (fig. 90), with an agate pestle B, fitted in a flexible shaft c, and FIG. 90. CEILINC LINE revolving at the rate of 700 times a minute, transferred to a ground-glass stoppered bottle, and dried at 100 C. It is then ready for analysis. THE CHEMICAL TREATMENT. Estimation of Phosphoric Acid. For ores low in phosphoric acid 10 grms. are dissolved in hydrochloric acid ; the solution is evaporated to dry- ness on the sand bath, re-dissolved in dilute hydrochloric acid (one part of acid to two of water), filtered, the filtrate treated with 10 c.c. ammonium sulphite * to reduce the iron to the ferrous condition, ammonia added until the solution is nearly neutral, and then heated until it has decolourised. Five c.c. of strong hydrochloric acid are * Made by saturating ammonia with sulphurous acid, generated by heating powdered charcoal and strong sulphuric acid in a flask. The mixture is made of the consistency of cream, and the gas passed first through a wash- b ottle containing water. ESTIMATION OF PHOSPHORIC ACID. 335 added to decompose any excess of ammonium sulphite and the sulphurous acid driven off by passing carbonic acid through the boiling solution. When the last trace of sulphurous acid is driven off, sulphuretted hydrogen is passed through the boiling solution to precipitate any arsenic, the sulphide of arsenic filtered off, the beaker placed in cold water, and when thoroughly cooled, dilute ammonia added until a slight permanent green precipitate of ferrous hydrate remained after stirring. Acetic acid is added until the precipitate dissolved (a few drops should be sufficient) the solution, diluted with hot water to about 800 or 900 c.c. And if the precipitate is white, a dilute solution of ferric chloride or bromine water is added until it becomes red. If it is necessary to add much ferric chloride a few drops of ammonic acetate are also added to insure the decomposition of all the former salt. The solution is then heated to boiling, boiled a few minutes, filtered rapidly, and washed once with hot water. The filtrate should pass through the filter perfectly clear, and if the precipitate is red, any subsequent cloudi- ness of the filtrate is of no consequence. If, however, the filtrate passes the filter cloudy, it should be returned to the main portion, a few drops more of ferric chloride added, and the solution again boiled and filtered. The precipitate of ferric phosphate, hydrate, and basic acetate is dissolved in hot dilute hydrochloric acid (1 1) on the filter, and the large beaker cleared of any adhering precipitate with hydrochloric acid, received in a small beaker, and evaporated nearly to dryness. Five to ten grms. of citric acid are dissolved in the least possible quantity of hot water, and filtered into a small beaker into which are also filtered about 5 c.c. of magnesium mix- ture,* and the whole is added to the solution of the ferric phosphate. This solution is then neutralised by ammonia and cooled. When perfectly cold, from one- third to one-half its bulk of strong ammonia is added, * Made by dissolving equal weights of sulphurous acid and ammonium chloride in the least possible quantity of water, filtering, adding bulk of strong ammonia, stirring occasionally, and allowing it to stand for some days before using. 336 COMPLETE ASSAY OF IRON ORES. and the solution is stirred until the precipitate of ammo- nium-magnesium phosphate begins to form. After standing for some time the solution is stirred again, and the stirring is repeated at intervals for an hour. It is allowed to settle overnight, filtered on the asbestos felt in Dr. Gooch's pierced crucible, washed with dilute ammonia (1 3), dried on the pump, moistened with a drop or two of ammonium nitrate in ammonia, dried and ignited until the glow passes over the precipitate, cooled in a desiccator, and weighed as magnesium phosphate. This precipitate is then dissolved in dilute hot hydrochloric acid, the felt is washed on the pump, and the crucible is heated to redness and re-weighed. This weight, unless the precipitate of magnesium phosphate contained silica, agrees perfectly with the original weight. In the latter case, however, the last weight of the crucible is subtracted from the weight of the crucible with the precipitate to obtain the weight of magnesium phosphate. The following table is used for cal- culating the percentage of phosphoric acid or phosphorus, instead of the factors 0-6396 or 0-2793. When the amount of magnesium-ammonium phosphate is large it is dissolved, boiled, and re -precipitated, as a small amount of magnesia is liable to be carried down mechanically with it. It is found necessary to avoid heating the magnesium phosphate after the glow has passed, as this salt is liable to attack the asbestos slightly at high temperatures, causing a small loss in weight of the felt on the subsequent solution of the precipitate in dilute hydrochloric acid. For ores containing titanic acid it is necessary to modify this process in several particu- lars. It is found that under certain circumstances phos- phoric acid combines with titanic acid, forming a salt (possibly a phospho-titanate) very insoluble in hydrochloric acid, so that in almost every instance phosphoric acid is found in the insoluble silicious residue when this contains titanic acid. Also, upon neutralising the main solution after adding ammonium sulphide in ores containing titanic acid, a fine white precipitate resembling barium sulphate is formed, which usually remains after the subsequent ESTIMATION OF PHOSPHORIC ACID. 337 TABLE FOR PHOSPHORUS AND PHOSPHORIC ACID. Mg p P 2 5 Mg P PA Mg p P 3 5 1 0-003 0-006 35 0-098 0-224 68 0-190 0-434 2 0-005 0-013 36 0-100 0-230 69 0-193 0-441 3 0-008 0-019 37 0-103 0-237 70 0-195 0-448 4 0-011 0-026 38 0-106 0-243 71 0-198 0-454 5 0-014 0-032 39 0-109 0-249 72 0-201 0-460 6 0-017 0-038 40 0-112 0-256 73 0-204 0-467 7 0-019 0-045 41 0-114 0-262 74 0-207 0-473 8 0-022 0-051 42 0-117 0-269 75 0-209 0-479 9 0-025 0-057 43 0-120 0-275 76 0-212 0-486 10 0-028 0-064 44 0-123 0-281 77 0-215 0-492 11 0-031 0-070 45 0-126 0-287 78 0-218 0-499 12 0-033 0-077 46 0-128 0-294 79 0-221 0-505 13 0-036 0-083 47 0-131 0-300 80 0-223 0-512 14 0-039 0-089 48 0-134 0-307 81 0-226 0-518 15 0-042 0-096 49 0-137 0-313 82 0-229 0-524 16 0-045 0-102 50 0-139 0-319 83 0-232 0-531 17 0-047 0-108 51 0-142 0-326 84 0-235 0-537 18 0-050 0-115 52 0-145 0-332 85 0-237 0-544 19 0-053 0-121 53 0-148 0-339 86 0-240 0-550 20 0-056 0-128 54 0-151 0-345 87 0-243 0-556 21 0-059 0-134 55 0-154 0-352 88 0-246 0-563 22 0-061 0-141 56 0-156 0-358 89 0-248 0-569 23 0-064 0-147 57 0-159 0-364 90 0-251 0-576 24 0-067 0-153 58 0-162 0-371 91 0-254 0-582 25 0-070 0-159 59 0-165 0-377 92 0-257 0-588 26 0-073 0-166 60 0-167 0-384 93 0-259 0-595 27 0-075 0-173 61 0-170 0-390 94 0-262 0-601 28 0-078 0-179 62 0-173 0-396 95 0-265 0-607 29 0-081 0-185 63 0-176 0-403 96 0-268 0-614 30 0-084 0-192 64 0-179 0-409 97 0-271 0-620 31 0-086 0-198 65 0-181 0-416 98 0-274 0-627 32 0-089 0-204 66 0-184 0-422 99 0-276 0-633 33 0-092 0-211 67 0-187 0-428 100 0-279 0-639 34 0-095 0-217 addition of hydrochloric acid, and consists of titanic acid with some phosphoric acid. If, after dissolving the precipitate of ferric phosphate, &c., thrown down by ammonic acetate, the solution is allowed to run to dry- ness, there is found after re-solution in hydrochloric acid a granular white or yellowish residue, absolutely insoluble in hydrochloric acid, which consists essentially of titanic and phosphoric acid. This reaction affords a very delicate test for titanic acid, for if the residue obtained in this latter case be collected on a small filter, dried, ignited, and fused with sodium carbonate, and the fusion treated with hot water, a residue of sodium titanate will remain, which is insoluble in water. This residue z 338 COMPLETE ASSAY OF IRON ORES. collected on a small filter, dissolved in a little hot dilute hydrochloric acid, and treated with zinc in a test-tube, gives the very characteristic violet colour due to titanic oxide. When titanic oxide is found it is necessary to fuse the insoluble silicious residue, the residue left in the filter from the solution of the precipitate by ammonic acetate, and the residue left from the re-solution of this latter, after it has run to dryness, with sodium carbonate. The fused mass is treated with hot water and filtered ; the filtrate containing, besides silica and alumina and sodium carbonate, all the phosphoric acid, while the titanic acid remains on the filter as sodic titanate, insoluble in water. Potassium salts should not be used, as potassic titanate is decomposed by water. The filtrate is acidulated with hydrochloric acid, .evaporated to dryness, dissolved in water with a little hydrochloric acid, filtered, a few drops of solution of ferric chloride added, and the ferric phos- phate precipitated by ammonic acetate. This precipitate is filtered, washed, dissolved in hydrochloric acid, and the phosphoric acid estimated by magnesium mixture as usual. Sulphur and Iron. One grm. of ore is fused with 10 to 12 grms. of sodium carbonate and a little nitre, the fused mass is run well up on the sides of the crucible, and the cru- cible is chilled. The mass is then detached from the crucible and transferred to a tall beaker, or the crucible with its contents is placed in the beaker and treated with boiling water. When the fused mass is entirely disintegrated (the crucible, if placed in the beaker, having been washed off and removed) the ferric oxide is allowed to settle. If the solution is coloured by alkaline man- ganate a few drops of alcohol are added, and the solution is allowed to stand until the colour disappears. If the solution after the disappearance of the colour due to the manganate is yellow, it is an indication of the presence of chromium in the ore. It is then decanted through a filter and washed twice with hot water by decantation. SULPHUR AND IRON. 339 The filter is washed once or twice with hot water, and the filtrate is acidulated with hydrochloric acid, evaporated to dryness, re-dissolved in hot water with a few drops of hydrochloric acid, filtered from the silica, and the sulphuric acid precipitated by barium chloride in the boiling filtrate, After standing over- night in a warm place the barium sulphate is fil- tered on the felt, washed thoroughly with hot water (and ammonium acetate if it is large), and weighed as barium sulphate, which multiplied by 0-1373 gives the sulphur. The crucible in which the fusion was made is treated with hydrochloric acid, and when the adhering ferric oxide is dissolved a little hot water is added, and the whole is poured on the filter to dissolve any ferric oxide which might have been in suspension in the solution of the fused mass. This is allowed to run into the beaker containing the ferric oxide, the crucible is rinsed, and the filter is washed with hot water. The whole is evaporated to dryness to render silicic acid insoluble, re-dissolved in hydrochloric acid, and transferred to a small flask of about 50 c.c. capacity. A small funnel is put in the neck of the flask, and 3 grms. of granulated zinc in small lumps are added carefully to the solution. By heating the solution and adding a few drops of hydro- chloric acid from time to time the ferric chloride is soon reduced to ferrous, which is shown by the solution be- coming colourless. After washing down the neck of the flask and the funnel with a fine jet of water, if the addition of a few drops of hydrochloric acid fails to impart any colour to the solution, the reduction is considered com- plete. Fifteen c.c. of dilute sulphuric acid (1 1) are added little by little to the solution, and when the zinc has all dissolved the neck of the flask is filled nearly to the top with hot water. The flask is then placed in cold water, and when the solution is thoroughly cooled it is washed out into an oblong white dish of about 1,500 c.c. capacity and diluted to about one litre. A solution of potassium permanganate is run z 2 340 COMPLETE ASSAY OF IRON ORES. FIG. 91. in from a burette, the representation of which is given in fig. 91. This form of burette, the invention of Mr. Thomas H. Garrett, is the most satis- factory one I have ever used. The solution of permanga- nate is carefully standard- ised by means of ferric chlo- ride of known strength, made by dissolving wrought iron of known composition in ni- tric acid, driving off the nitric acid by repeated evapora- tions with hydrochloric acid, diluting, filtering into a clean glass-stoppered bottle, and estimating the ferric oxide gravimetrically in a weighed amount of the solution. The strength of the ferric chlo- ride solution being thus ac- curately known, a portion is weighed out into one of the little flasks previously mentioned, reduced with zinc, sulphuric acid added, and the strength of the perman- ganate solution necessary to colour 3 grms. of zinc treated with hydrochloric acid and sulphuric acid, and diluted to the same volume as the solution of the ore, is subtracted before calculating the value of the permanganate solution , and also from the amount of permanganate solution re- quired for the ore before calculating the percentage of iron. By using sulphuric acid in addition to the hydro- chloric acid as described above it is found that the end reaction with potassium permanganate is as sharp as if no hydrochloric acid is used, and the results obtained are SILICA, FERRIC OXIDE, ALUMINA, ETC. 341 in all cases as accurate as could be desired. For many ores it is possible to dissolve the finely ground sample directly in the flask in hydrochloric acid without previous fusion, but for those containing ferrous silicates or pyrites it is generally necessary to resort to fusion for the accu- rate estimation of the total iron. Silica, Ferric Oxide, Alumina, Manganese, Lime, and Magnesia. One gramme is dissolved in hydrochloric acid, evapo- rated to dryness, re-dissolved in dilute hydrochloric acid, and evaporated a second time to render the silica in- soluble. It is then re-dissolved in hydrochloric acid and water (1 3), filtered on a small ashless filter, washed, dried, ignited, and weighed as insoluble silicious matter. This is fused with five times its weight of dry sodium carbonate, treated with hot water, and washed into a platinum dish, the crucible treated with acid and carefully rinsed into the dish, the whole acidulated with hydro- chloric acid, evaporated to dryness, treated with water and a little hydrochloric acid, and evaporated to dryness. The mass is then treated with hydrochloric acid and water (] 5), heated to boiling, and filtered on a small ashless filter. The filtrate is dried, ignited, and weighed. The silica is treated in the crucible with hydrofluoric acid * and sulphuric acid, evaporated to dryness, ignited, and weighed. The difference between this and the weight of the precipitate is silica. Any residue obtained by the treatment with hydrofluoric acid and sulphuric acid is examined. It might consist of barium sulphate or titanic acid if either of these exists in the ore, or of a little alumina or ferric oxide or of a small amount of sodium sulphate if the silica is not carefully washed. The filtrate from the silica is treated in a platinum dish with ammonia, boiled until it smells but faintly of * The hydrofluoric acid is re-distilled with the addition of a little sul- phuric acid and potassium permanganate from a platinum still, and collected in a platinum bottle, as the crude hydrofluoric acid always contains ferric oxide, besides various sulphur compounds. 342 COMPLETE ASSAY OF IRON ORES. ammonia, filtered, washed, dried, ignited, and weighed as alumina with or without a tinge of ferric oxide. To the filtrate, ammonium oxalate is added, the solution is boiled and allowed to stand overnight, filtered, washed, and ignited at a high temperature, and weighed as lime. This filtrate is evaporated down, sodium and am- monium phosphate are added, and the solution is well stirred to precipitate the magnesium-ammonium phos- phate. After standing overnight it is filtered, washed, ignited, and weighed as magnesium phosphate. Whence 111 : 40 = weight of magnesium phosphate to weight of magnesia. The sum of the weights of the silica, alumina, lime, and magnesia should about equal that of the in- soluble silicious matter. When there is a deficiency, alkalies are looked for in another portion. If there is an excess, and the precipitated alumina is red, iron in insoluble matter is estimated. The filtrate from the insoluble silicious matter is nearly neutralised with sodium carbonate solution, 2 grammes of sodium acetate are added, and the whole after being diluted to about 700 c.c. with hot water is boiled, and the precipitate ferric oxide, &c., is filtered on a washed filter. The precipitate is washed two or three times on the filter, and then transferred back to the beaker with a platinum spatula, the filter is washed with hydro- chloric acid and finally with water, the whole being re- ceived in the beaker with the mass of the precipitate. Sufficient acid is added to dissolve the precipitate, and the operation is repeated, the filtrates being added to- gether. The solution of the precipitated ferric oxide, &c.,, is evaporated to dryness to render insoluble any silica from the reagents ; re-dissolved in dilute acid, and filtered into a large platinum dish ; the solution is boiled, am- monia is added, and the precipitate is collected on an ashless filter, washed thoroughly on the pump with hot water, dried, ignited (finally over the blast), and weighed as iron and aluminium phosphate. The filtrates from the acetate precipitations are evaporated down to about 300 c.c., 2 or 3 grammes of sodic acetate are added, the- SILICA, FEEKIC OXIDE, ALUMIXA, ETC. 343 solution is heated to boiling, and sulphuretted hydrogen is passed through to precipitate any copper, nickel, cobalt, and zinc. The precipitate by sulphuretted hydrogen is filtered off', and after all smell of sulphuretted hydrogen has been boiled off, bromine water is added to the solution. When the precipitated manganic oxide has collected at a gentle heat, and while the solution is still coloured by bromine, it is boiled until colourless, filtered, washed several times, and the manganic oxide is dissolved on the filter in hydrochloric acid with the addition of solu- tion of sulphurous acid, which causes its very rapid solu- tion. This solution is evaporated to dryness, re-dissolved in hydrochloric acid and water ; a slight excess of am- monia is added, nearly all smell of ammonia is boiled off, the solution is filtered, any slight precipitate of ferric oxide, &c., is re-dissolved and re-precipitated and filtered as before. The filtrates are added together, and excess of sodium- ammonia phosphate is added with enough hydrochloric acid to render the solution decidedly acid, and after boiling for some time an excess of am- monia is added to precipitate the manganese-ammonium phosphate. This is boiled until the precipitate becomes crystalline and the solution smells but slightly of am- monia, when it is filtered, washed with cold water, ignited, and weighed as manganese phosphate, which multiplied by 0-5 gives oxide of manganese. To the filtrate from the precipitate of manganic oxide by bromine is added the ammoniacal filtrate from the final precipitation of the iron, alumina, and phosphoric acid ; the whole is evaporated down to about 400 c.c., and the lime and magnesia are precipitated as described in the analysis of the insoluble silicious matter. There are several sources of error which it is found very necessary to guard against in the above analysis namely, the contamination of the distilled water by silica when it is boiled in glass, the strong action of ammoniacal solu- tions on the beakers, and the presence of silica, alumina, ferric oxide, lime, and magnesia in many reagents, and 344 COMPLETE ASSAY OF IRON OKES. especially in sodium carbonate. Distilled water which has been heated overnight in a Bohemian flask on the sand- bath is found to contain 52 milligrammes of solid residue to the litre, 26 milligrammes of which is silica ; and ferric oxide precipitated from a solution to which water boiled in a flask from two to six hours has been added, contains after ignition as much as 8 per cent, of silica. To avoid this tin-lined copper flasks are used for heat- ing distilled water, and to avoid the error due to the action of ammonia salts and ammoniacal solutions on the beakers, all the precipitations and evaporations are made in platinum. We were fortunate enough to obtain some remarkably pure sodium carbonate, containing only 2 milligrammes of silica, 1^ milligramme of alumina, 2 milligrammes of lime, and 2 milligrammes of magnesia to the 100 grammes from Messrs. Powers and Weightman, Philadelphia. Nickel, Cobalt, and Zinc. Three grammes of ore are treated as if for the esti- mation of manganese, and the sulphides of copper, nickel, cobalt, and zinc are precipitated from the boiling solution of the acetates by sulphuretted hydrogen. This precipitate is collected on a small filter, washed with sulphuretted hydrogen water containing a little acetic acid, dried, ignited in a porcelain crucible, and transferred to small beaker. It is then treated with hydrochloric acid and a little nitric acid, evaporated to dryness, re- dissolved in hydrochloric acid, diluted, boiled, and the copper precipitated by sulphuretted hydrogen. The copper sulphide is filtered off and washed with hot water, the filtrate containing the cobalt, nickel, and zinc is evaporated to dryness, re-dissolved in a few drops of water containing not more than two or three drops of hydrochloric acid, and an excess of potassium nitrate acidulated with acetic acid added. After standing for a day or two the potassic-cobaltic nitrite is filtered off, washed once or twice with a strong solution of potassic acetate, and then, after removing the filtrate, with alcohol NICKEL, COBALT, AND ZINC. 345 to displace the alkaline salts. The precipitate is then ignited very carefully in a porcelain crucible, treated with sulphuric acid to decompose all the nitrate, made alkaline with ammonia, filtered, and the cobalt is precipitated by the battery ; or the ignited precipitate is dissolved in a few drops of hydrochloric acid, transferred to a small beaker, dilated, and any alumina and iron present is precipi- tated by boiling for several hours with an excess of sodic acetate. The precipitate of alumina and iron is filtered off, the filtrate is made alkaline with ammonia, and the cobalt is precipitated as sulphide by ammonium sul- phide. The precipitate is allowed to settle, filtered off on a small ashless filter, washed, dried, ignited, and weighed as cobaltous sulphide, or treated with sulphuric acid and zinc, and weighed as cobaltic sulphide, which multiplied by 0*4872 gives cobalt oxide. The filtrate from the potassium- cobaltic nitrite is acidulated strongly with hydrochloric acid and heated to decompose all the nitrite, and the nickel oxide is preci- pitated by an excess of soda or potash, filtered, and the filtrate tested for zinc oxide with ammonium sulphide. If any zinc sulphide is found, the nickel oxide is re- dissolved and precipitated as before, filtered, and the filtrate and washings are added to the first filtrate. The zinc sulphide is allowed to settle, filtered, washed with water containing sulphide of ammonium, re-dissolved in hydrochloric acid, and evaporated to dryness. This is treated with dilute hydrochloric acid, filtered, and the zinc is precipitated by solution of sodium carbonate, filtered, washed, dried, ignited, and weighed as zinc oxide. The precipitate of nickel oxide is dissolved on the filter in hydrochloric acid, the filtrate run to dryness with sulphuric acid, diluted, excess of ammonia is added, filtered, and the nickel precipitated by the battery ; or the hydrochloric acid solution is evaporated to dryness, dissolved in a drop or two of hydrochloric acid, diluted, and boiled with an excess of sodium acetate. Any pre- cipitate of alumina, &c., is filtered off, excess of sulphide of ammonium is added, then an excess of acetic acid 346 COMPLETE ASSAY OF IRON ORES. and sulphuretted hydrogen is passed through the boil- ing solution. The precipitated nickelic sulphide and sulphur are collected on a small ashless filter, dried, ignited, heated with a little ammonic carbonate, and weighed as nickelous sulphide. Estimation of Ferrous Oxide. One gramme of finely ground ore is weighed into a flask A (fig. 92) of about 100 c.c. capacity, fitted with a FIG. 92. rubber stopper, through which pass two glass tubes as shown in the cut. Dry carbonic acid is passed in through the tube B to expel the air through the tubes c and D, the latter dipping beneath the surface of the water in the beaker E, and when this is accomplished the stop- per is removed, 15 c.c. of strong hydrochloric acid is quickly poured in, the stopper is replaced, and the ore is dissolved with the aid of heat, the current of carbonic acid being continued uninterruptedly. When the solution ESTIMATION OF FEEEOUS OXIDE. 347 of the ore is accomplished the source of heat is re- moved, and, the current of carbonic acid being temporarily stopped, the water in the beaker B is allowed to run back into the flask A until the latter is nearly filled, when the current of carbonic acid is turned on again, and allowed to continue until the solution in the flask is thoroughly cooled. This is accomplished by removing the tripod c, placing a dish nearly filled with cold water under the flask, and replacing the tripod. The solution is then washed out into the dish used for titrating, into which 3 grms. of zinc dissolved in 15 c.c. of the dilute sulphuric acid (1 1) are previously poured, and the whole is diluted to 1 litre. The amount of ferrous oxide is then estimated by the potassic permanganate solution. The amount of solution required to colour about 1 grm. of ferric chloride diluted to the same bulk, and containing the same amount of sulphuric acid, is subtracted from the amount required for the titration before calculating the amount of ferrous oxide. Ferrous Oxide in Insoluble Silicious Matter. When the insoluble silicious matter contains iron in the ferrous condition for instance, in the form of epidote 1 grm. is treated with hydrochloric acid diluted, and the insoluble matter is collected on the felt in a pierced crucible, carefully dried, transferred with the felt to an ordinary platinum crucible, and treated in the apparatus of fig. 93, with hydrofluoric acid and hydrochloric acid in a current of carbonic acid. When entirely decomposed it is allowed to cool, the current of carbonic acid being kept up, and then transferred quickly to a dish containing about 1 litre of water and 15 c.c. of the dilute sulphuric acid and zinc ; the crucible is washed out, and the amount of ferrous oxide is estimated by the standard solution of potassium permanganate. 348 COMPLETE ASSAY OF IRON ORES. Sulphuric Acid. Sulphuric acid may exist in the ore in the form of barium sulphate, calcium sulphate, or as sulphates of iron, FIG. 93. &c., formed in the roasting or weathering of an ore con- taining pyrites. When it exists as barium sulphate it is found in the analysis of the insoluble silicious matter in the residue remaining after the treatment of the silica with hydrofluoric acid. In this case the residue is fused with sodium carbonate, treated with hot water, and filtered. The filtrate is acidulated with hydrochloric acid, and the sulphuric acid is precipitated by barium chloride and estimated as barium sulphate, which multiplied by 0*3433 gives sulphuric anhydride. The barium carbonate on the filter is dissolved in dilute hydrochloric acid, sulphuric acid is added, and the barium sulphate is finally weighed, which multiplied SULPHUEIC ACID. 34$ by 0-6567 gives baryta. In the other cases 10 grammes of the ore finely ground are treated with water contain- ing a little hydrochloric acid, filtered, the sulphuric acid is precipitated in the filtrate as barium sulphate, and the weight is determined with the usual precautions. Alumina. The total amount of the soluble ferric oxide, alumina, and phosphoric acid is added to the ferric oxide and alumina found in the insoluble silicious matter, and from this is subtracted the phosphoric acid which gives the total iron and alumina. From this is subtracted the ferric oxide found by titration ; the difference is alumina. The accuracy of this result depends of course upon that of the volumetric estimation of the ferric oxide. The comparison of many analyses shows that the error due to this need never exceed a few hundredths of 1 per cent., and no direct method is found to yield results equally accurate or concordant. Calculation of the Analysis. The sulphur in the sulphuric acid found as such is subtracted from the total sulphur ; so much as is neces- sary to form the sulphides of copper, nickel, cobalt, zinc, and antimony with the amounts of these elements in the ore, and supposed to exist in this condition, is sub- tracted from this, and the remaining sulphur is calcu- lated to iron disulphide. The amount of iron required for this is calculated to ferric oxide, and subtracted from the ferric oxide found by titration. From this is taken the amount of ferric oxide calculated from the ferrous oxide found in the estimation of ferrous oxide, and the result is the amount of ferric oxide existing in this state in the ore. Carbonic Acid. Three grammes of ore are weighed into the flask A of the apparatus (fig. 94), and the connections are made as there shown ; 10 c.c. of strong sulphuric acid are 350 COMPLETE ASSAY OF IRON ORES. added to 65 c.c. of water and poured into the bulb-tube B, the stopcock c having first been closed. After the potash bulb and the drying-tube are weighed they are attached to the apparatus, the tube s is filled with fused calcium chloride, being added to prevent the drying-tube R from, absorbing moisture from the atmosphere. The WATEE AND CARBON. 351 TJ-tube o is empty, p is filled with pumice saturated with anhydrous copper sulphate, and Q with dried calcium chloride. When the connections are all proved to be tight, N is fitted into the neck of the bulb-tube with a piece of rubber tubing, and the acid in the bulb is allowed to run into the flask very slowly, and when it is all in, a slow current of air is forced through the appa- ratus by means of the bottles L L. The air is freed from carbonic acid by potassic hydrate in the tube M. As soon as the current of air is started, the flask A is heated gradually, and finally the solution is boiled until the bend of the tube o is filled with condensed water. It is then allowed to cool while the current of air is con- tinued. The potash bulb and the drying-tube are de- tached and weighed with the usual precautions ; the increase of weight is of course the carbonic acid due to the carbonates contained in the ore. Water and Carbon in Carbonaceous Matter. Many ores besides black-bands contain carbon ; for in- stance, several ores from New Jersey contain graphite, and nearly all limonites and magnetites contain carbon in organic matter, probably from the organic acids which originally held the iron in solution. As pyrites is also of common occurrence in such ores it is necessary to devise some method by which the water of composition and the carbon could be estimated in the presence of pyrites. In attempting to estimate the water by heating the ore in a current of air, some sulphuric acid is always formed and driven over into the drying-tube, and some organic matter is certain to be dissolved if the ore is treated with acid for the estimation of the carbon. It is found by careful experiments that when a carbonate of any kind is fused with potassium anhydrous chromate in excess the carbonic acid is all expelled together with any water present, while pyrites treated in the same way is oxidised with formation of potassium sulphate, which is not decomposed even at a bright red heat. 352 COMPLETE ASSAY OF IRON OEES. A very practical and easy way to apply this in the treatment of ores is afforded by the use of Dr. Gooch's tubulated crucible* (fig. 95), and the process is as follows : For ores containing much water or carbonic acid I FIG. 95. HES gramme, and for others 3 grammes are weighed into the crucible A of fig. 96, and carefully mixed, by means of a rod or wire, with 7 grammes of potassium bichromate, which has been previously fused and powdered, the cap B is adjusted, and the whole is placed in the air-bath and heated to 100 for a short time. The crucible is then placed on the platinum triangle c, and connected by means of a cork with the weighed tube D, containing dried calcium chloride, and the weighed potash bulb E and drying-tube F are attached, the latter guarded by the calcium chloride tube G, as seen in the sketch. The cap for the crucible consists of a conical cover H drawn out vertically into a tube i, into which is burned a horizontal tube J, of the same diameter. Through the top of the tube I passes the tube K to a distance of half an inch below the bottom of the cap, the end being slightly bent ; K is burned into I at its point of entry a, sealing i * ' Chemical News,' vol. xiii. p. 326. WATER AND CARBON. 353 at this point. A straight glass tube M, drawn tapering, is fused to the platinum tube j at b and another piece, N, is bent at a right angle to the platinum K at c. A piece of rubber tubing, closed with a piece of glass rod at one end, is drawn over the end of the tube N, the space around the cap in the flange d is filled with small A A 354 COMPLETE ASSAY OF IRON ORES. pieces of neutral anhydrous -sodic tungstate, which are fused by means of a blowpipe flame, making an air-tight joint. The mixture in the crucible is kept cool during this operation by dipping the end in a beaker of cold water. The expansion of the air in the apparatus during the heating causes it to bubble through the potash bulb, and the reflux of the solution in the bulb as the apparatus cools is a good index of the tightness of the joints. When the joints are shown to be all tight the cap is removed from N, and the bottle L is connected with N, as shown in the sketch. A slow current of air, freed from carbonic acid and water by passing through the tube Q, containing potash and calcium chloride, is then started through the apparatus, and the crucible is heated very gradually and cautiously by the burner o. As the steam is gradually driven into the drying-tube it is not allowed to condense at y, but is driven forward into the calcium chloride by the heat of a small alcohol lamp, applied to the drying-tube at this point. When the greater part of the water has been thus driven over, the crucible is heated by a horizontal flame from the blast-lamp p, above the level of the mixture to prevent the latter from frothing up and stopping the end of the tube K. The bottom of the crucible is gradually heated to a dull red, and allowed to remain at this temperature for about fifteen minutes, when the lamps are turned out and the whole is allowed to cool in the current of dried air from the bottle L. The tubes are then re-weighed ; the increase in the weight of D is the weight of the water of com- position in the ore, and that of the potash bulb and drying- tube the weight of the total carbonic acid. This latter, of course, is the sum of the carbonic acid, due to the car- bon of the carbonaceous matter or graphite, and of that due to the carbonic acid in the carbonates. The amount of the latter (previously estimated) is subtracted from the total carbonic acid, and the difference multiplied by 0-2626 gives the carbon in carbonaceous matter or the graphite. ALKALIES. 355 Alkalies. The alkalies are generally confined to the insoluble silicious matter in the ores, but are occasionally found in the portion soluble in hydrochloric acid. In the former case 5 or 10 grammes of the ore are digested in hydro- chloric acid until only the silicates remained undissolved, diluted, and filtered. The insoluble residue is washed, dried, ignited, and treated in the crucible with hydrofluoric and sulphuric acids until it is entirely decomposed ; it is then evaporated down until all the hydrofluoric acid and nearly all the sulphuric acid has volatilised. It is then treated with hot water and a little hydrochloric acid if necessary ; the clear solution is transferred to a small platinum dish, the alumina and ferric oxide are pre- cipitated by ammonia, and the solution is boiled until nearly all ammonia is driven off. It is then filtered into another platinum dish, evaporated to dryness, and heated until all the ammonium salts are volatilised. The residue is treated with a little water, a few drops of ammonium oxalate, and excess of ammonia, and the solution is boiled and filtered. The filtrate is evaporated to dry- ness, and the residue is heated to dull redness. It is then treated with water, filtered, and an excess of barium hydrate * is added to precipitate the sulphuric acid and magnesia. The solution is boiled and filtered from the barium sulphate, and the filtrate is evaporated to dry- ness after the addition of a little ammonium carbonate and hydrate. The residue is treated with a little water, filtered to get rid of the barium carbonate, and the filtrate is evaporated to dryness and heated carefully to vola- tilise any ammonium salts. The residue is treated with water, filtered into a small weighed platinum dish ; the filtrate is acidulated with hydrochloric acid, evaporated to dryness, heated carefully to a very low red, and weighed as quickly as possible as alkaline chlorides. The alkaline chlorides are then dissolved in water, any residue is * Made by treating ordinary barium carbonate with tolerably strong nitric acid, and washing the barium nitrate with nitric acid. The barium nitrate is dried and fused until all the nitrous acid is driven off. A A 2 356 COMPLETE ASSAY OF IRON ORES. filtered off, weighed, and its weight subtracted from the weight of the chloride, an excess of platinum chloride is added, and the solution is evaporated on the water-bath until the syrupy liquid solidifies on cooling. The residue is treated with alcohol, 95 per cent, filtered on the felt, washed with 95 per cent, alcohol, dried at 120 C., and weighed as platinum-potassium chloride, which multiplied by 0*1931 gives potassium oxide. The weight of the platinum-potassium chloride, multiplied by 0-3056 to re- duce to potassium chloride, is subtracted from the weight of the alkaline chlorides, and the difference (sodium chloride), multiplied by 0-5299, gives sodium oxide. The filtrate from the platinum-potassium chloride is evaporated to dryness in a platinum dish with the addition of a little oxalic acid to decompose all the platinum chloride ; the residue is treated with water, filtered, and any magnesia is precipitated by microcosmic salt and weighed as magnesium phosphate, which multiplied by 0- 4*2 8 gives the amount of magnesium chloride to be sub- tracted from the weight of chlorides to get the amount of alkaline chlorides. For the estimation of alkalies in the portion of the ore soluble in hydrochloric acid 1 gramme is treated with hydrochloric acid, diluted, filtered into a platinum dish, and heated to boiling ; a slight excess of ammonia is added, and the whole is evaporated to dryness to render the ferric oxide very granular and easy to wash. It is then treated with water, a drop or two of ammonia is added, and the ferric oxide, &c., filtered off and washed. The filtrate, after the addition of a drop or two of sulphuric acid, is evaporated to dryness, and the ammoniacal salts are driven off by ignition. The residue is treated with water, a little ammonium oxalate is added to precipitate the lime, and the alkalies are estimated, as in the former case, after the precipitation of the calcium oxalate. Copper, Lead, Arsenic, and Antimony. Ten grammes of the very finely ground ore are treated with hydrochloric acid with the addition of potassium COPPER, LEAD, ARSENIC, AND ANTIMONY. 357 chlorate until only the silicious residue remains unacted upon. It is then diluted, filtered, reduced with ammo- nium sulphite, all the excess of sulphurous acid is boiled off, and sulphuretted hydrogen is passed through to satu- ration. The precipitate is allowed to settle in a warm place, and when the smell of sulphuretted hydrogen has nearly passed off, it is filtered on the pump, washed, and treated on the filter with potassium sulphide to dis- solve out the sulphides of arsenic and antimony. The filter containing the sulphides of copper and lead is dried and ignited in a porcelain crucible. The ignited sulphides are transferred to a small beaker and treated with hydrochloric acid and nitric acid, excess of sulphuric acid added, and the whole is evaporated down until sul- phuric acid begins to volatilise. Water is then added, and, if any lead sulphate separates out, an equal bulk of alcohol, and the whole is allowed to stand some hours, when the lead sulphate is filtered off, dissolved in basic ammonium tartrate or acetate, and the lead sulphide pre- cipitated by sulphuretted hydrogen is filtered, washed, and ignited. It is treated in the crucible (porcelain) with nitric and sulphuric acids, and finally ignited and weighed as lead sulphate, which multiplied by 0-6832 gives lead. The solution containing the copper is evaporated down to drive off the alcohol, washed out into a platinum crucible, and the copper is precipitated on the cru- cible by the battery in the apparatus shown in fig. 97, washed with water and alcohol, and weighed as copper. The potassium sulphide solution containing the sulphides of arsenic and antimony is acidulated with hydrochloric acid, and allowed to stand in a warm place until the sul- phides and free sulphur have collected and the solution smells but faintly of sulphuretted hydrogen. It is then filtered on the felt, and the precipitate, if it contains FIG. 97. 358 COMPLETE ASSAY OF IKON OKES. much free sulphur, is treated with carbon disulphide r transferred with the felt to a small beaker, and treated with hydrochloric acid and potassium chlorate, or aqua regia, which dissolves the arsenic and antimony very readily. A little tartaric acid is added to keep the antimony in solution, the whole is diluted and filtered, excess of ammonia is added, and then magnesium mixture, and the arsenic precipitated as ammonio-magnesium arseniate with the same precautions that were used in the precipi- tation of phosphoric acid. The precipitate is filtered on the felt, washed with dilute ammonia, dried at 100 C., and weighed as ammonio-magnesium arseniate. It is then heated very gradually at first, and finally to a full red, and weighed as magnesium arseniate. The first weight obtained multiplied by 0*3947, and the second by 0-4839, gives arsenic. If the heat is applied carefully and slowly enough at first the latter result is most apt to be -the correct one. The filtrate which contains the antimony is acidulated with hydrochloric acid, and the antimony is precipitated by sulphuretted hydrogen with the usual precautions, filtered on a small disc of paper on the bottom of the perforated crucible, dried, and the pre- cipitate and paper are treated in a porcelain crucible with fuming nitric acid, evaporated to dryness, ignited, and weighed as antimony oxide, which multiplied by 0'7922, gives antimony. Titanic Acid. As titanic acid is found to interfere with the estima- tion of phosphoric acid, so the latter is found in many cases to absolutely prevent the precipitation of the former. When the ore contains only a small amount of titanic acid 5 grammes are treated with hydrochloric acid, evaporated to dryness, re-dissolved in dilute hydro- chloric acid, filtered, the insoluble residue dried, ignited, and treated in the crucible with hydrofluoric and sulphuric acids. The solution in the crucible is evaporated down until all the hydrofluoric and sulphuric acids are driven TITANIC ACID. 359 off, ignited, and the residue is fused with sodium car- bonate. The fused mass is treated with hot water, which dissolves the sodium phosphate, aluminate, and excess of sodium carbonate,* leaving insoluble the sodium titanate and any ferric oxide, calcium carbonate, barium carbonate, &c. The insoluble portion is filtered off, washed, dried, and put aside to be added to the titanic acid in the soluble portion. The filtrate obtained after the first treatment of the ore with hydrochloric acid, which contains the great mass of the iron with the titanic acid which dissolves with it, is deoxidised by ammonium sulphite, the excess of sul- phurous acid driven off, and the ferric phosphate, titanate, and small excess of ferric oxide are precipitated by am- monium acetate exactly as in the estimation of the phos phoric acid. This precipitate is filtered off, washed; dried, and fused with sodium carbonate, the fused mass is treated with water, filtered, and the insoluble portion containing ferric oxide and sodic titanate is washed and dried. The two filters containing the sodic titanate are separated as carefully as possible from the adhering pre- cipitates, ignited in a platinum crucible, the precipitates are added, and the whole is fused at a low heat with 15-20 times the weight of acid potassium sulphate. The heat is gradually increased to a dull red, and kept at this point until the fusion was perfectly clear and of a reddish-brown colour. It is then cooled and treated with a large excess of cold water, containing a few drops of sulphuric acid, with constant stirring until nothing re- mains undissolved, except, perhaps, a little silica. It is then filtered, washed with cold water, the filtrate is diluted to about 500 c.c., and boiled in a large platinum dish or beaker with the constant addition of strong aque- ous solution of sulphurous acid, until all the titanic acid is precipitated. The precipitate is allowed to settle, the clear solution is decanted through a filter, the pre- cipitate is boiled up with water, thrown on the filter, and washed as titanic acid. The filtrate is boiled again with the addition of sulphurous acid, and any further pre- * Any tungsten in the ore would be in this solution as sodium fcungstate. 360 COMPLETE ASSAY OF IRON ORES. cipitate filtered off and added to the first. It is found .that when the solution is kept at the boiling-point during the filtration, and the precipitate washed with boiling water, there is no tendency to pass the filter ; whereas, if the solution is allowed to get cold, the filtrate is nearly always cloudy. To prevent the precipitation of ferric oxide with the titanic acid, it is necessary to keep an excess of sulphurous acid in the solution ; but if in spite of this precaution the ignited titanic acid is coloured with ferric oxide, it is fused again with potas- sium sulphate and re-precipitated. In the analysis of ores containing much titanic acid it is preferable to fuse the ore at once with sodium carbonate, treat with water, filter, wash and dry the residue, separate and burn the filtrate, fuse the residue and filter ash with potassium sulphate, and estimate the titanic acid as before. > 98< Estimation of Specific Gravity. A great many experiments were made both on lumps and on powdered ore, and it was decided that in commer- cial samples good results could be obtained only by using the powdered material. The work was done and the flask herein described was made by Mr. James Hogarth. About 20 grammes of the powdered sample are weighed out and transferred to the specific gravity flask, the con- struction of which is shown in fig. 98. This flask is made with the view of overcoming two difficulties which occur in using the common flask, viz. the ex- pansion and overflow consequent on transferring the flask at 60 P. to the higher temperature of SPECIFIC. BRAVITV FLASK the balance-case, and the necessity for waiting until the SPECIFIC GRAVITY. 361 finely divided mineral has settled before the stopper could be inserted without loss of material. These ends are successfully met by melting on a capillary tubulus to the lower part of the neck, and grinding in a stopper having a small bulb above the capillary to allow for expansion. Having transferred the weighed quantity of ore to the flask, enough water is added to cover it, and the tem- perature is raised almost to the boiling-point by placing the flask in a water-bath. To insure complete expulsion of air the flask is now placed under a bell-jar, connected with the aspirator, and allowed to boil for a few moments at the reduced pressure. It is then filled up with water almost to the tubulus, cooled, tlie stopper is inserted, and by suction it is filled up slightly above the mark on the capillary part of the stopper. When the thermo- meter is stationary at 60 F. the volume is adjusted by touching the capillary end of the tubulus with blotting- paper, or by bringing a drop of water to it as the case might require. The flask is then dried, transferred to the balance-case, and, after a sufficient time has elapsed to allow it to take the temperature of the room, weighed. If W be the weight of ore taken, W the weight of ore 4- water at 60, and K the weight of the flask + water at 60 then Q Sp.gr- = = To obtain K the flask is nearly filled with boiled water and treated exactly as described above. The Perforated Crucible. In processes in which the use of filter-paper is delete- rious or inconvenient, nitrations and ignitions have been made with the aid of the asbestos filter devised by Mr. Gooch, which consists essentially of a felt of asbestos deposited upon a perforated surface of platinum.* * For a full description of the process see ' Chemical News,' xxxvii. D. 181. 362 COMPLETE ASSAY OF IRON ORES. FIG. 99. PERFORATED CRUCIBLE. The mode of preparing the filter is as follows: Asbestos of fine silky flexible fibre is scraped longitudi- nally (not cut) to a fine short down, which is purified by boiling in strong hy- drochloric acid and washing by decanta- tion. A platinum crucible with bottom perforated with nu- merous fine holes is fitted to the upturn- ing end of a piece of soft rubber tubing, the other end of which is stretched over the top of a ' ' f l funnel as shown in fig. 99, and the neck of the latter passes through the stopper of a vacuum flask in the usual way. The vacuum pump having been put in action, a little of the prepared asbestos suspended in water is poured into the crucible, and the pump is attached, when the asbestos at once assumes the condition of a firm compact layer, which is washed with ease under the pressure of the pump. After washing the felt, the crucible with the felt ad- hering is removed from the funnel, ignited, cooled, and weighed as usual ; then set again in place in the rubber- holder, taking care that the pump is working, and the liquid to be filtered is poured into the crucible. The crucible, felt, and adhering precipitates are, after washing, ignited, or merely dried, as the case demands, with no further care than the nature of the precipitate itself necessitates. During the ignition of precipitates it is frequently necessary to prevent the contact of reducing gases with the perforated bottom, and this is accomplished by placing the crucible either upon a platinum crucible cover or within a second crucible, or better by using a form of CARBON ETC. IN IRON AND STEEL. 363 crucible which is conical in shape and provided with a cap removable at pleasure, to cover the bottom during ignitions. When a larger filtering surface than the bottom of a crucible affords is needed a cone of platinum, per- forated and fitted to a rubber seat, is substituted for the crucible. 7. Estimation of Carbon, Sulphur, Silicon, Phosphorus, fyc., in Metallic Iron and Steel. Boring and Sampling. Samples of metallic iron for ana- lysis are always obtained by boring, and the proper sam- pling of these borings is as difficult a task as the corresponding one of sampling a heap of ore. As is well known, cast-iron borings are a mixture of small particles of iron with more or less of finely divided graphite separated from the sur- faces of these small particles during the process of boring. The amount of graphite thus mechanically mixed is in all cases quite large enough to cause serious difficulty when the problem is to obtain an average sample of a given lot of borings for carbon estimation. This diffi- culty arises from the necessity of obtaining a uniform mixture of the heavy and comparatively coarse borings with the finely divided and light graphite, and of remov- ing a sample for analysis without disturbing the uniformity of this mixture. The difficulties here presented have been investigated by Mr. Porter W. Shinier. It was first sought to over- come the difficulty by having the borings made very fine, so that they might have more nearly the size of the par- ticles of graphite. It was not practicable, however, to secure this uniformity of size, since much of the graphite is in the form of the finest dust. At all events, the dupli- cate estimations of carbon in these fine borings, made with every precaution by combustion in oxygen, frequently showed differences too large to be accounted for in any other way than by imperfect sampling. Very coarse borings were also tried, with the idea that the graphite separated from these large and heavy pieces 364 ASSAY OP IRON ORES IN THE WET WAY. would perhaps be inappreciable ; but it was found that quite enough was separated to vitiate the results. Various methods of mixing were tried, both in the bottle and on glazed paper ; but in all these mixtures it was found that duplicate samples seldom contained the same proportion of the mechanically mixed graphite ; and when they did there could be no certainty that it was the true proportion. All this was shown many times over by the failure of duplicates to agree closely, the difference being sometimes as much as O2 per cent. It was finally clear that nothing was to be hoped from these methods. Even if we could secure a perfect mixture containing in every part its due proportion of graphite, it would run a great risk of being destroyed the moment a spatula was inserted to remove a sample ; for the least agitation causes some of the graphite to fall through between the coarser bor- ings. The sampling difficulty was finally overcome in this way : The borings were carefully poured out into a large porcelain crucible or dish, and enough alcohol is added to merely moisten them (2 c.c. alcohol to 450 grains). Mix thoroughly for about five minutes. Bernove with a spatula as many samples of the moist borings as are needed, weigh- ing them out roughly so as to get approximately the weights desired. Dry off the alcohol and weigh accurately* The samples thus obtained perfectly represent the average of the original sample, for the alcohol simply serves tempo- rarily to hold the graphite where it belongs, upon the surface of the borings. As the alcohol evaporates from the original sample the graphite again falls away from the borings, and the moistening must be repeated when new samples are to be removed. The following total carbon results show the im- provement : i. ii. Mixed dry in the bottle . > . > . 3'68 3'84 p.c. Moistened with alcohol and mixed . . 3'97 3-99 The first results are an extreme case of unsatisfactory duplicates ; and they also show the tendency of the gra- CARBON, SULPHUR, SILICON, PHOSPHORUS, ETC. 365 phite to go to the bottom of the bottle when mixed dry. The mechanically mixed graphite, approximately estimated by removing the iron by means of a magnet from samples taken moist, was 0-76 and 0-79 per cent, in dupli- cate estimations. After shaking up the dry borings a few times, and re- moving nearly all from the bottle, the remaining sample contained 1*29 per cent, mechanically mixed graphite. The other estimations, such as phosphorus and silicon (iron excepted), are perhaps not appreciably affected when samples are taken dry. But whenever samples are thus taken, the borings which remain are for ever vitiated for carbon estimations which shall truthfully represent the original sample. In very careful work, therefore, and especially when there is any prospect of subsequent carbon estimations to be made, it is neces- sary to moisten the borings when weighing out samples for phosphorus and silicon, or to have two sets of borings made, one of which is to be reserved for carbon esti- mations. For iron estimations the samples should always be taken moist. It was suggested by Mr. Frank Firmstone, Superin- tendent of the Glendon Iron Works, to determine whether or not there is any appreciable loss of graphitic dust while making cast-iron turnings according to the method devised by him, in which the turnings are taken equally from the whole section of a piece of pig iron, the outside being first removed. Through the kindness of Mr. Firmstone, Mr. Shinier was enabled to procure from the same piece of pig iron a sample of turnings taken dry, a second sample kept moist with alcohol to prevent loss of dust, and several sections -^ inch in thickness. The following are the total carbon results from these samples : I. Turnings taken dry, but sampled moist. II. Turnings taken moist. III. Thin section. i. ii. in. 4-132 4-122 4-031 4-123 4-123 4-038 4-111 4-103 4-108 3G6 ASSAY OF IRON ORES IN THE WET WAY. The results under I. and II. show that, with care in catching the turnings, and avoidance of draughts of air, there is no appreciable loss of graphitic dust. The first two results on the thin section are low, because of in- complete decomposition of the iron, when, after forty hours, hydrochloric acid was added to complete the solu- tion. The third result is about right ; but complete decom- position was obtained in this case by frequent and long- continued stirring. While, therefore, it is possible to obtain the correct total carbon by use of a thin section, it offers no advantage in point of time, since borings may be dissolved in neutral and cold copper and ammonium chloride solution in fifteen to twenty minutes by constant stirring. Some chemists, with a keen appreciation of the sam- pling difficulty, have proposed to overcome it by using clippings of the iron from -J to y 1 ^ inch in thickness. While it is no doubt possible, in most cases, to obtain the true result by this means, it is open to the objection of being very slow when these clippings are to be dissolved in neutral copper and ammonium chloride solution. Fur- thermore, the weight of iron taken for carbon estima- tion in cast iron being limited by convenience to about 50 grains, a very few clippings make up this weight, and these represent only those parts of the sample from which they happen to come. While there may not be an appreciable difference in the composition or in the condi- tion of the carbon of the different parts of a section of pig iron, it is yet possible that there may be, and in this case it is easy to be on the safe side. When turnings are taken equally from the whole sec- tion of a piece of pig iron, and are then moistened with alcohol and thoroughly mixed, every part of the section has its due representation in a sample of 50 grains. It may not be out of place to add that, in sampling ores, and, in fact, any similar mixture of particles of different size, composition, and specific gravity, it is advisable to moisten with water before mixing. This secures a uniform mixture of coarse and fine parts from which a true sample ESTIMATION OF THE TOTAL CARBON. 367 can easily be removed, prevents loss of dust, and is in every way more satisfactory than mixing dry. In pulverised ores it is very common for the finest part to consist principally of lighter gangue material ; and the problem of securing a true sample of such a mixture is similar to that of obtaining a true sample of a mixture of cast-iron borings and graphite. Carbon. This is a problem of considerable difficulty, and to secure accurate results many special precautions are necessary, owing to the large preponderance of the iron over the carbon present. The carbon may be present in two forms as combined carbon, and as free or graphitic carbon. The estimation may be of the total carbon present, or of either the combined or graphitic separately, and the method of analysis adopted will have to be selected accordingly. The following is a description of the most satisfactory processes which have been devised for these estimations. A. Estimation of the Total Carbon. Mr. Addison B. Clemence employs the following method for estimating carbon in steels. The cut (fig. 100) shows the form of apparatus adopted. It is made of platinum. The following is the process : Dissolve from 50 to 80 grains of the borings in double chloride of copper and ammonium, using 560 grains of the salt to 40 ounces of water for 46 grains of steel. After the separated copper is completely dissolved, filter on to a plug of asbestos placed at 5, and wash thoroughly with hot water. Any carbon that adheres to the sides of the tube may be swept down by moistened asbestos. The tube is then placed in an air-bath, and dried at a temperature of from 350 to 380 F. for about one hour. A hard rubber cork through which is passed a glass tube is in- serted at e, the oxygen gas passing from c to a. Around the tube at c is a single thickness of filter-paper, about two inches wide, kept wet by a stream of water supplied by a reservoir on a shelf above. Heat is applied at b for one half-hour, at the end of which time the potash bulb is ready for the balance. One Bunsen burner is sufficient 368 ASSAY OF IKON ORES IN THE WET WAY. for the combustion. The same precautions are taken to dry the gas before entering the platinum tube, as well as before entering the potash bulb, as in the case with the porcelain tube method. ; Six burners of a combustion furnace will consume 14 ft. of gas in one half- hour, while one Bunsen burner will con- sume 2 ft. in the same time. The following table shows some re- sults obtained by the " platinum tube " process : A is a Swedish Bessemer iron contain- ing 0*10 per cent, of carbon ; B, an American Bessemer, with 0-18 per cent. ; and C, an American Bessemer, with 0'50 per cent. ; all obtained by the "porcelain tube" process. <----* A. 0-09 0-10 0-10 0-10 B. 0-18 0-18 0-16 0-17 c. 0-49 0-51 0-49 0-44 Weyl's very ingenious method for the estimation of the total carbon is founded upon the fact that a piece of iron at- tached to the positive pole of a galvanic battery, and suspended in hydrochloric acid, is dissolved, while the hydrogen is given off at the negative pole. The for- mation of hydrocarbons, and a consequent loss, is in this manner prevented. One great advantage in this method is that the iron does not require to be in pow- der. A piece of iron 2 to 4 grammes in weight, attached to the positive pole of ** a Bunsen's cell, is suspended in dilute hydrochloric acid, just below the surface of the liquid. From the negative pole hydrogen passes off, while the ESTIMATION OF GRAPHITE. 369 iron dissolves quite quietly, and the strong solution of ferrous chloride formed may be seen falling in a regular stream through the lighter liquid. The iron is dissolved in about twenty-four hours, and the carbon is left behind in the same shape as the piece of metal from which it was derived. In Weyl's earlier experiments it was found that some of the liberated carbon at the positive pole was carried over to the negative pole by the mechanical working -of the stream. To prevent this, a diaphragm of bladder or parchment paper is interposed between the two, which entirely obviates the possibility of loss in this way. B. Estimation of the Graphite. Dr. Eggertz reduces 1 gramme of iron to small pieces, mixes with 5 grammes of pure iodine and 5 c.c. of water in a small flask, covered with a watch-glass, and placed in ice-cold water before adding the iron. It is to be kept for twenty-four hours at C., and frequently stirred meanwhile. By keeping the liquid cold, no carburetted hydrogen is produced. The greater the amount of silicon in the iron the greater is the tendency to the production of carburetted hydrogen. The residue of carbon and silica left after the iron is dis- solved is collected on a filter of known weight, when it is dried at from 95 to 100 C., and washed thoroughly with hot water. After twelve hours, it is to be washed with a mixture of hydrochloric acid and twice its volume of water, heated to 70 or 80 C., until the filtrate ceases to give a blue colour with ferrocyanide solution. The object of leaving the residue for twelve hours is to allow any small particles of iron remaining undissolved by the iodine being oxidised by atmospheric air, and prevent disengage- ment of hydrogen when the hydrochloric acid is added. After the hydrochloric acid is washed out of the filter, it is dried, with its contents, at 95 to 100 C., until constant in weight. This weighing gives the amount of the carbo- naceous residue and silica (but not the whole of the silica, because some part of it would have been dissolved), and by burning the carbon away and weighing the silica the weight of the carbonaceous residue may be ascertained. B B 370 ASSAY OF IRON ORES IN THE WET WAY. The carbonaceous residue may consist either of graphite or of the compound of carbon, iodine, and water already mentioned, if the carbon was combined with the iron. To ascertain which is the case, 1 gramme of iron is dissolved in 15 c.c. of hydrochloric acid (1*12 density) in a flask covered with a watch-glass, and, when the iron is dissolved, the solu- tion is boiled for half an hour. All the carbon combined with the iron is disengaged in the form of carburetted hydrogen gas, while the graphite and silica remain. If the carbona- ceous residue left after dissolving the iron comes in con- tact with atmospheric air before the liquid is boiled, it is so altered that it is not dissolved and disengaged as gas. The graphite that remains after boiling the liquid is collected on a filter of known weight, washed, dried, and weighed. It is then burnt, and the residual silica weighed to ascertain the quantity of graphite. Very satisfactory results have been obtained by this method. The differences do not amount to more than Ol per cent. When the quantity of carbon to be estimated is very small, more than 1 gramme of iron must be used in the analysis. Mr. Tosh gives the following process for the estimation of graphite. Two to three grammes of iron are treated with dilute hydrochloric acid, and when the solution approaches completion a considerable quantity of strong acid is added to separate the last portions of iron and manganese. The insoluble matter, consisting mostly of graphite, is collected on a carefully weighed filter, washed with hot water, dilute hydrochloric acid, solution of caustic soda, and hot water again, successively, and lastly with alcohol and ether to remove oily hydrocarbons. (By wash- ing with dilute acid and with alkali, the iron and silica or oxide of silicon are separated.) After drying at 120 C., the filter and graphite are weighed, and burned away. The small residue (a mere trace of silica or titanic acid) is weighed, and this weight subtracted from the first gives the amount of graphite. The results obtained agree very closely. In washing the graphite with solution of soda, there is always a brisk effervescence, due to the oxidation of oxide ESTIMATION OF COMBINED CARBON. 371 of silicon to silicic acid, by decomposition of water, with consequent liberation of hydrogen. C. Estimation of Combined Carbon. When steel, or pig-iron containing carbon in chemical combination, is dis- solved in nitric acid, a soluble brown colouring matter is formed, whose colouring power is very intense, and the solution assumes a tint which is dark in proportion to the quantity of the chemically combined carbon. Iron and graphite (or free carbon) do not influence this colouration ; for the solution of nitrate of iron is only slightly greenish, unless extremely concentrated, and graphite is insoluble in nitric acid. Thus, in dissolving two pieces of different steels of the same weight in nitric acid, taking care to dilute the darker solution until the two liquids present exactly the same colour, it is very evident that the more highly carburetted steel will furnish the larger quantity of liquid, and that the proportion of the volumes will indicate the relative pro- portion of colour in the two steels. If now the composi- tion and content of carbon of one of the steels is known, the absolute percentage of carbon in the other steel may be immediately deduced. Dr. Eggertz has applied these reactions to a method of estimating the combined carbon. To obtain trustworthy results certain precautions must be taken. In a cylindrical test-tube, dissolve gradually in the cold 10 centigrammes of wrought-iron, steel, or cast iron, reduced to a fine powder, in 1-J to 5 c.c. of nitric acid of 1*2 sp. gr. The use of nitric acid containing hydrochloric acid must be avoided, because the solution of iron would have a yellow tint. In proportion as the metal contains more carbon, more nitric acid must be used. After some time, when the chief part of the metal appears to be attacked, place the tube in a water bath, and warm it to 80 C., in such a position that only the lower part of the tube is in contact with the warm water ; a movement takes place in the acid, which favours its reaction upon the metal; a slight disengagement of carbonic acid from all the particles of carbon may be observed. The operation should always B 2 372 EGGERTZ'S COLOUR TEST FOR THE be conducted under the same circumstances as to heat and length of time. The evolution of gas having ceased (in operating upon steel, the reaction must continue two or three hours), place the tube in a large vessel filled with water, to bring the solution always to the same temperature. This precaution is indispensable, because the same liquid is darker when warm than when cold. Afterwards, pour off, as exactly as possible, the clear liquid into a graduated burette. Upon the black residue remaining in the tube pour some drops of nitric acid, and heat carefully over a lamp. If there is no further liberation of gas, the residue consists of nothing but graphite or silica. Cool the new solution, and add it to that which is already in the burette. The liquid is then diluted with water until its colour corresponds exactly with that of the normal liquid, which latter should be of such a degree of concentration that each c.c. represents O'OOOl grm. of carbon. If, for instance, this normal liquid is prepared from cast steel containing exactly 0*85 per cent, of carbon, 1 decigramme of that steel must be dissolved in 8*5 c.c. of nitric acid ; lOOgrms. of steel containing 85 centigrammes of carbon would thus be dissolved in 8500 c.c. of the normal solution, 100 c.c. of that solution would represent 1 centigramme of carbon, and, consequently, 1 c.c. of the normal solution would represent O'OOOl grm. of carbon. To compare the normal solution with the solution of iron under examination, it should be contained in a tube of the same kind, and when the two tubes are held together by daylight before a thin sheet of paper, the colour should be exactly the same in both of them. For the preparation of normal colour solu- tions see pp. 382-388. The Eggertz colour test for estimating carbon in steel is perhaps used more extensively commercially than any other analytical process. This is particularly the case in Bessemer steel works, where every blow is tested to ascer- tain its percentage of carbon. The number of estimations thus amounts even in a small works to 25 daily. In carrying out the analytical process, which is too ESTIMATION OF CARBON IN STEEL. 373 FIG. 101. well known to require description, it is of course necessary to preserve the identity of each blow-number ; and this is effected by dissolving the steels in a square shallow copper water-bath provided with a lid perforated with holes in which the tubes are inserted in a standing position, the first hole indicating the starting-point of the day's blows. This form of bath is objectionable from several points of view. Mr. J. Oliver Arnold has consequently designed a bath which two years' constant use has proved to have none of the disadvantages connected with the old form of bath. The bath consists of a cylinder of glass 5 inches high by 5 inches in diameter, closed at one end to contain the water ; this is heated on an iron pi ate or sand-bath nearly to boiling-point (90-95). On the top is a perforated cover of glazed white earth- enware ; this cover contains 17 holes, numbered 1 to 15 in burnt-in black figures, and the other two holes are marked S for the standard steels. The tubes containing the steels and nitric acid are suspended in the water to the proper depth by slipping on the tube a rather tightly fitting cylinder of india-rubber f inch long ; this india-rubber serves the double purpose of keeping the ends of the tubes off the bottom of the bath, bumping thus being avoided, and of preventing the bobbing up of the tubes by flota- tion (see figure). The apparatus, being of glass and glazed earthenware, can be washed perfectly clean every day. The tubes are suspended vertically in the water to any desired depth. 374 ESTIMATION OF CAEBON IN STEEL. The clear numbering enables the selection of any given blow about which 4 a hastener ' has been received with- out chance of error, and the progress of the solution of the steels in the tubes can be watched without removing them. Except from mechanical breakage the apparatus is practically indestructible. In cases where the daily car- bons exceed 15, a second and third bath may be employed, the numbering following on, or one large bath could be made ; this, however, might prove inconveniently large. In cases where only a few odd carbons are required a smaller bath perforated for 6 steels and a standard would suffice. Several modifications of Eggertz's method have been proposed. The most successful of them, affording exceed- ingly accurate results, was communicated to the ' Journal of the Franklin Institute ' for May 1870 by J. Blodget Britton, and has been tested for a considerable time. In- stead of a single tube, containing a standard solution for comparison, as suggested by Eggertz, a number of tubes- FIG. 102. A 8aJJjJLIlLttl having their solutions differently standardised, one from the other, are employed. These are arranged securely in a wooden frame, with spaces between for placing the tube containing the solution to be tested, and forming together a convenient portable instrument called a colorimeter a representation of such an instrument being shown in the annexed cut. The position of the tube containing the solution to be tested is shown at A. The tubes are fths of an inch in diameter, and 3-| inches in length, filled with water and alcohol coloured with roasted coffee, and herme- tically sealed. The solution in the tube to be left has its colour to correspond exactly with one produced by 1 BRITTOIs T S METHOD. 375 gramme of iron, containing 0'02 per cent, of combined carbon, dissolved in 15 c.c. of nitric acid. The solution in the tube next to it has its colour to correspond with one produced by the same quantity of iron, but containing 0-04 per cent, of combined carbon ; and so with each of the other tubes, increasing 0*02 per cent, of carbon in regular succession to the right, the last reaching 0*3 per cent., as indicated by the figures on the upper rail of the instrument. On the back of the instrument, and for the purpose of partially screening the light and allowing the different shades of colour to be distinctly discerned, there is tightly stretched between the rails some fine white parchment paper. This screen is not shown by the cut, but it serves a very important purpose. The process is conducted as follows : 1 gramme of the finely divided metal is put into a tube of about 1-J inch in diameter and 10 inches long, and digested for fifteen or twenty minutes in 10 c.c. of nitric acid of a little more than 1-20 specific gravity, free from chlorine. The solution is then cautiously poured into a beaker, and a small portion of metal which remains undissolved and adheres to the bottom of the tube is treated with 5 c.c. of fresh acid, exposed to a gentle heat till completely dissolved, and added to the other. The contents of the beaker, when sufficiently cool, are filtered through two thicknesses of German paper (not previously moistened, and of a diameter not exceeding 4^ inches) into a tube about 5 inches long and of precisely the same diameter as those in the instrument. After the filtered solution has remained for some minutes at the tem- perature of the atmosphere, and its colour become fixed, the tube is placed in the instrument and the carbon esti- mated by a comparison of shades : the estimation may be made readily as close as O'Ol per cent. Heat should not be applied in the first instance to facilitate the solution of the metal, because a high temperature is apt to cause a slight loss of colour. Two thicknesses of paper are taken, because one alone is liable to break ; and the paper should be used dry, for, if previously wetted, the water will weaken the colour of the solution ; and it ought to be cut 376 ESTIMATION OF CARBON IN STEEL. to a size not exceeding 4-^ inches, to prevent undue absorp- tion. If the metal to be examined contains more than 0-3 per cent, of carbon, O5 gramme, or less of it, may be taken, or the solution may be diluted with an equal volume or more of water and the proper allowance made ; or an instrument of higher range may be used. On the other hand, if the metal contains a very small percentage of carbon, 2 grammes of it may be taken. For preparing the standard solutions (the normal ones begin to lose colour after some hours), caramel dissolved in equal parts of water and alcohol, as suggested by Eggertz, answers well ; but with roasted coffee as the colouring matter the true shades may be obtained. (See also pp. 382-88.) An important paper by Mr. J. E. Stead was read in 1883 before the Iron and Steel Institute on a new method of estimating minute quantities of carbon in iron and steel, and on a new form of chromometer. The following a,re extracts from Mr. Stead's paper : 1. ESTIMATION OF MINUTE QUANTITIES OF CARBON. As is well known, it is impossible to estimate with accuracy minute quantities of carbon by the ordinary colour method, owing to the colour of the nitrate of iron present, which interferes so as to make it impossible to judge of the colour due to carbon. Having been engaged in some careful investigations on the nature of the colouring matter which is produced by the action of dilute nitric acid upon white iron and steel, it was found it had the property of being soluble in potash and soda solutions, and that the alkaline solution had about two and a half times the depth of colour pos- sessed by the acid solution. This being so, it was clear that the colouring matter might readily be separated from the iron, and be obtained in an alkaline solution, by simply adding an excess of sodium hydrate to the nitric acid solution of iron, and that the colour solution thus obtained might be used as a means of estimating the amount of carbon present. Upon trial this was found to be the case, and that as small a quantity as 0-03 per cent, carbon could be readily estimated. STEzVDS METHOD. 077 The method as now in use is conducted as follows : Standard solution of nitric acid, T20 sp. gr. Standard solution of sodium hydrate, 1*27 sp. gr. One gramme of the steel or iron to be tested is weighed off and placed in a 200 c.c. beaker, and, after covering with a watch-glass, 12 c.c. of standard nitric acid are added. The beaker and contents are then placed on a warm plate, heated to about 90 or 100 C., and then allowed to remain until dissolved, which does not usually take more than ten minutes. At the same time a stan- dard iron containing a known quantity of carbon is treated in exactly the same way, and when both are dissolved 30 c.c. of hot water are added to each, and 13 c.c. soda solution. The contents are now to be well shaken, and then poured into a glass measuring-jar and diluted till they occupy a bulk of 60 c.c. After again well mixing and allowing to stand for ten minutes in a warm place, they are filtered through dry filters, and the filtrates, only a portion of which is used, are compared. This may be done by pouring the two liquids into two separate measur- ing-tubes in such quantity or proportion that upon looking down the tubes the colours appear to be equal. Thus if 50 m.m. of the standard solution is poured into one tube, and if the steel to be tested contains, say, half as much as the standard, there will be 100 m.m. of its colour solution required to give the same tint. The carbon is therefore inversely proportional to the bulk compared with the standard, and in the above-assumed case, if the standard steel contained 0-05 percent, carbon, the following simple equation would give the carbon in the sample tested : Q-" = 0-025 per cent. 100 Experiments were made upon a steel which contained 0-11 per cent, carbon, to ascertain what the influence would be of heating the nitric acid solution for an increas- ing length of time after dissolving on the bath, and it 378 ESTIMATION OF CARBON IN STEEL. was found that the carbon colour was not materially affected by heating the acid solution twice as long as was necessary for completely dissolving the carbon compound, and that although the iron is dissolved in five minutes it is evident that some of the carbon compound at first formed escapes solution in that period. The next point was to ascertain what effect the use of an excess of nitric acid in dissolving the steel w^ould have on the colouring matter. It was seen that 6 c.c. acid in excess does not materially affect the estimation, but when this is exceeded the colour is reduced in quantity. It now became important to know if a greater or less quantity of soda solution would have a different solvent power on the colouring matter. To ascertain this, four separate portions of the soft steel were treated alike in dissolving, but to the solutions different quantities of soda solution were added. It was here seen that, as before stated, 13 c.c. sodium hydrate solu- tion is capable of effecting solution of the colouring matter. By using a less amount, however, by experiment it was found that the colour is precipitated with the iron oxide. It is very well known that, in the old colour method, very slight traces of hydrochloric acid, if present, alter the character of the colour to such an extent as to make the colour estimation unreliable. It therefore was of interest to ascertain if the same would occur in the alkaline method. Four portions of steel were treated as usual, excepting that to one portion a single drop of hydrochloric acid was added when being dissolved, to a second five drops, and to a third ten drops, but to the last portion no hydrochloric acid was added. Th.e following are the results obtained, viz. : Hydrochloric Acid. 1 drop 5 drops 10 drops None Carbon . 0-105 p. c. 0-090 p. c. 0-078 p. c. 0-110 p. c. Second Test. 1 drop 5 drops 10 drops None Carbon . 0-356 p. c. 0*338 p. c. 0-324 p. c. 0-410 p. c. The colour in each case, and even in that in which the larger quantity of hydrochloric acid was added, was the STEAD S METHOD. 379 same in quality although differing in quantity, showing (1) that the presence of chlorides is harmless, and (2) that nitro-hydrochloric acid, even in small quantities, pre- vents the formation of the full amount of colouring matter capable of being produced by nitric acid alone. A large number of samples of low carbon iron have been examined by the alkaline method, including samples of iron taken from the Bessemer converter at the end of the blow before any addition of spiegeleisen. The results are given here Blown Iron taken from the Bessemer Converter. No. 1 0-040 p. c. carbon. No. 2 No. 3 No. 4 No. 5 No. 6 Average Average by combustion Standard soft steel . Pure iron wire Cleveland iron ship plates Ditto ditto The colour solutions from these low carbon irons are different in tint from those obtained from the higher car- bon steels ; and it is important that a low carbon iron be used as a standard for comparison. When high carbon steel is heated to redness and chilled, it is well known that the colour from the chilled steel is very much less in quantity than that from the same steel before hardening. The difference, however, is not nearly so marked when there is little carbon present in the steel, as was proved by the following results, viz. : Several samples of iron and steel after being drilled were heated to redness and chilled in water, the results before and after being as follows : 0-036 0-045 0-039 0-061 0-048 0-045 0-048 New Colour Method Combustion 0-120 p. c. carbon. 0*122 p. c. 0-038 , 0-055 > 0-035 Soft steel = soft 0-168 p. chilled in cold water 0-158 chiUed in hot water . 0-168 Staffordshire square iron bar . O'llO , chilled 0-100 , flat . 0-069 . chilled 0-069 , Soft steel 0-077 : chilled. . . 0-071 , Difference >0-010 p. c. 0-010 None. 0-006 380 ESTIMATION OF CARBON IN STEEL. It is not often that soft iron or steel is chilled before being placed in the hands of the analyst, but it is satisfac- tory to know that even if they were the results by colour would not be rendered useless. When using the new method Mr. Stead found that some steels give a much yellower colour than others, and in course of investigation he discovered that there are present in all nitric acid steel solutions two distinct colour- ing matters, which have been separated arid obtained in a nearly pure state, one of which is bright yellow, resembling potassium chromate, the other being of a dark brown-red colour. In some steel solutions the yellow colour pre- ponderates, and in others the brown. 2. A NEW FORM OP CHROMOMETER. In comparing colour solutions there are two methods of procedure. The first is that generally adopted in making estima- tions of the carbon by the acid colour process, in which the darker solution is diluted with water until the colours of the two solutions are equal in density that is to say, until the colour is equal per cubic centimetre. The diluted volume is then noted, and the amount of carbon read off in 0*1 per cent. c.c. In the alkaline method it is better to use the second plan, already described, of comparing directly the relative density of the colour solution without dilution, and ascertaining the lengths of the two columns of liquids, which, when examined from the surface, give the same depth of colour. The carbon in this process is, as compared with the stan- dard, inversely proportional to the length of the liquid column. If a fixed length of liquid column be used of the solu- tions of carbon and a variable standard column, then, by using a suitable standard solution, the carbon may be de- duced from the length of the latter required to make a colour column equal in depth to the former and the per- centage read directly from a graduated scale. The instru- ment arranged by Mr. Stead is made on this principle; it is extremely simple and easily constructed. It consists of two parallel tubes, which may be of any suitable diameter, STEAD'S CHROMOMETER. 381 one of which is contracted at a point 9 inches from the top, and is open at both ends. The lower end passes through an india-rubber cork to the bottom of a 4-oz. bottle, which contains the standard colour solution. A second tube of smaller diameter also passes through the cork into the bottle, the outside end of which is in com- munication with a large syringe. Just above the contracted part of the first-mentioned tube a small glazed cylinder of china clay rests. By pressing the syringe the liquid can be forced from the bottle below up this tube. The second tube is about 9 inches long, and is closed at the lower end. At this end a small glazed clay cylinder is also placed. When this tube is placed parallel to the first, the length from the open upper ends to the flat surfaces of the clay prisms is equal in each. A small looking-glass at an angle of 45 is fixed above the open ends of the tubes, and the standard tube is graduated into O'Ol part to O15 part. The method of working with the apparatus is very simple. The colour solution to be compared is placed in the second tube, with which it is filled up to a certain fixed mark. It is only now necessary to squeeze the syringe and force the liquid up the first tube until the colours in the two columns are equal, as can be seen by looking into the mirror above. The height of the standard solution is read off on the graduated scale, which will be the percentage of carbon in the steel or iron under examination. The preparation of Inorganic Standards for the Colorimetric Carbon Test. Wherever the amount of work renders it practicable the plan of using permanent standard solutions in con- nection with the colorimetric carbon test affords such manifold advantages that it is to be strongly recommended. That it has not attained a wider and more general applica- tion is mainly due to the fact that, as the method is generally employed, the difficulty attending the produc- 382 ESTIMATION OF CARBON IN STEEL. tion of the colours, and their doubtful stability when produced, have been only too evident to those who have undertaken their composition. The old method of using standard solutions made from sugar, coffee, &c., has always been hampered by the tendency of the colours to change upon exposure to the light, and the consequent liability to serious error unless they are closely watched. Appreciating these difficulties while recognising the de- sirability of permanent colours, Mr. Magnus Troilius, at the suggestion of Professor F. L. Erckmann, investigated the properties of cobalt, copper, and ferric chlorides with the view of determining their efficacy in the production of standard solutions. His results seem to have been satisfactory, but his treatment failed to eliminate many of the objections that have rendered the use of organic colours undesirable. The following mode of procedure in the making-up of inorganic standards has been found by Mr. W. Eobinson, of Joliet, Illinois, to be most effi- cient in point of accuracy, ease, and celerity of production and stability. DAY STANDARD COLOURS. Mr. Eobinson uses as a range in illustration the car- bon-contents included between the extremes 0*5 per cent, arid 0-3 per cent, as representatives of Bessemer rail steel that branch of the industry in which the uni- formity and continuity of the product renders the use of permanent standard solutions specially advisable. The principles formulated below will evidently apply to either a higher or lower carbon range. The salts employed are the same as those used by Mr. Troilius, viz. the neutral chlorides of cobalt, copper, and iron. Solutions of these are made as follows : Cobalt Chloride. Dissolve in water containing 1 per cent, of free hydrochloric acid (1*12 specific gravity) in the ratio of 1 gramme salt to 1 c.c. water. Copper Chloride. Dissolve in water containing 1 per cent, hydrochloric acid (1-12 specific gravity) in the ratio of 1 gramme of the salt to 1^ c.c. water. PREPARATION OF STANDARD COLOURS. 383 Ferric Chloride. Dissolve in water containing 2 per cent, hydrochloric acid (1-12 specific gravity) in the ratio of 1 gramme salt to 1 c.c. water. The solution in all cases is best effected by the appli- cation of a gentle heat ; and filtering through a ribbed filter assures freedom from turbidity. Free acid is used to prevent the otherwise enhanced liability to decomposi- tion of the chlorides. Should any difficulty be encoun- tered in obtaining a clear solution of the ferric chloride, subsidence of the liquid for a few hours, and then decan- tation, will obviate the trouble. Mr. Eobinson proceeds : ' Our purpose is now to make up a set of colours that shall correspond to those made by the nitric acid solution of steels containing in car- bon every alternate point between the extremes of 0-5 and 0-3 per cent. It is clearly impracticable to have steel standards with this gradation of carbon ; while, on the other hand, to make a single colour comparable with some one steel, and then by dilution make up others, corresponding to all the desired percentages, is liable to introduce an error. I have two steel bars that are kept as standards, the carbon composition of which, as shown by repeated combustions, is 0*501 per cent, and 0-304 per cent, respectively. As the limit of error in the colorimetric test can be considered as about 0-01 per cent, for this grade of steel, I am safe in calling the first a 0-5 and the second a 0-3 per cent, metal. From the 0-5 and 0-3 standard steels are now made up nitric acid solutions under the same conditions, as regards amount of steel taken, acid used, time of heating and temperature, as those to which the regular tests are to be subjected. I am in the habit of dissolving 0-5 grammes of drillings in 12 c.c. nitric acid (1-20 specific gravity), and subjecting it in a steam-bath to a temperature of 98 to 100 C. for fifteen minutes. 'The next step is to make up from the chlorides a colour which shall exactly match the 0'5 per cent, steel solution, and another which, diluted to J- ths of its strength, shall precisely correspond to the 0'3 per cent, steel. The 384 ESTIMATION OF CAEBON IN STEEL. former we will call a brown 0-5 colour, and the latter a green 0-5 colour. I use the term " green 0'5 " for the latter, because it will be found to be a colour which, though of a greener cast, corresponds in intensity, before diluting, to the solution of the 0-5 steel. These are pre- pared as follows : 4 For the brown 0*5 colour, mix the chloride solutions and water containing 0-8 per cent. (1-12 specific gravity) hydrochloric acid in the ratio of 87 '5 c.c. water, 5*7 c.c. cobalt chloride, 25 c.c. copper chloride, and 43 c.c. ferric chloride. For the green 0'5 colour take 82-6 c.c. water containing 0'8 per cent, hydrochloric (1-12 specific gravity), 5-4 c.c. cobalt, 5 c.c. copper, and 7 c.c. ferric chloride solutions. The first of the colours so produced will ap- proximately match the 0-5 steel solution, while the second, diluted to J-ths of its strength with water containing 0*8 per cent, hydrochloric acid (1*12 specific gravity), will approximately correspond to the 0-3 steel solution. The absolute identity of the colours may be obtained by the addition of the chloride or chlorides, the colour of which corresponds to the shade that is wanting. The compari- son of the green 0*5 colour with the 0-3 steel is readily accomplished by the aid of a reading-tube suitably cali- brated 6 c.c. and 10 c.c., for instance In the final match- ing care should be exercised that the same light is used as is to be regularly worked with the same window if possible as this assures similarity of conditions as regards transmitted and reflected light. With correct brown and green 0*5 colours so obtained the alternate intermediate points are easily acquired by mixing the two and diluting with water containing 0'8 per cent, hydrochloric acid (1-12 specific gravity) by the aid of a burette, as per following table : DAY STANDARD COLOURS. 385 Per cent. Carbon No. c.c. Brown 0'5 colour No. c.c. Green 0'5 colour No. c.c. Water 0-5 10 0-48 8-64 0-96 0-4 0-46 7-36 1-84 0-8 0-44 6-16 2-64 1-2 0-42 5-04 3-36 1-6 0-4 4 4 2 0-38 3-04 4-56 2-4 0-36 2-16 5-04 2-8 0-34 1-36 5-44 3-2 0-32 0-64 5-76 3-6 0-3 6 4 This table is based on the following considerations : 1. That theoretically, from a given solution represent- ing carbon ol known value, other solutions representing definite lower carbon percentages may be made by dilut- ing with water. 2. That practically this will give colours of shades at variance with those aimed at. The explanation is clear. We have first to decide upon some definite bulk as a suitable amount of the solutions to make up. With the 5 inch by -| inch reading-tubes adopted here, 10 c.c. has been found convenient. To make the O48 standard, for instance, we must have such a 0*5 colour that the shade produced upon diluting with the proper amount of water will match that made from a 0-48 steel. This neither the green nor the brown 0-5 alone will do. The former is too green, the latter too brown. The operation of determining in the mixture the correct proportion of the two requisite to give the desired colour (for the 0'48 colour as an example) is purely a mathematical one. The computation is as follows : Amount of green : amount of brown : : 0'02 : 0'18 ; 0-02 and 0-18 being the relative distances of the 0*48 colour from the extremes, 0-5 and 0'3. The proper ratio of the brown and green 0*5 colours being found, the amount of the mixed chlorides and water to make a total of 10 c.c. is easily calculated thus : 0-5 : 0-48 : : 10 : x 10-9-6 = 0-4 CC 386 ESTIMATION OF CARBON IN STEEL. Hence 9'6 will be the number of cubic centimetres of the combined chloride solutions and 0-4 the number of water. Any of the other gradations in this or any range is found in a like manner. If it is desirable to make either a higher or lower, all that is necessary is to accur- ately know the carbon-contents of two suitable steels and proceed as above. It is always well to keep the colours from, exposure to the light as much as possible. When it is necessary to use the colours constantly, as in testing every heat at a Bessemer plant, a reading-rack capable of being closed is to be recommended. This avoids the neces- sity of removing to a dark place. Such a rack, provided with a front and back door, swung on hinges fastened to the bottom, and arranged with a thin plate of fine ground glass for a reading background, has proved very effective. For this purpose I have found the glass to possess advan- tages over the mediums ordinarily employed. The solu- tions must of course be corked tight. With due regard to the precautions and directions here laid down, a set of reliable inorganic standard solutions can be made from the above chlorides that will be found to be constant for four or five months of continued use, and often for even a longer period. When decomposition does set in, the colours gradually grow dark. NIGHT STANDARD COLOURS. As the method is commonly conducted, standards are weighed from the steel direct. When tests have to be made after every blow, this is obviously inconvenient, and renders it unadvisable to leave the work in unskilled hands. With the purpose of placing the night-tests as nearly as possible on an equal footing with the day-tests, the follow- ing investigation was undertaken by the writer. The day chloride standards, when compared by both kerosene and gaslight, were found to appear much lighter than the corresponding steel solutions. The arc light gave the same result. Monochromatic light as obtained by the soda bead and various coloured glasses was tried without NIGHT STAND AED COLOURS. 387 success. The magnesium light gave the same reactions as daylight, but was laid aside as impracticable, and the idea of using the day-standards for night work was then aban- doned. Night-standards that proved entirely satisfactory were finally made up from the same solutions and on ex- actly the same principle as the day-standards, with the simple modification that the brown 0*5 and green 0'5 were matched with the (>5 and O'o steel solutions in a dark room, by the aid of kerosene light. To make the brown 0-5 night-standard, mix the solutions in the following ratio: 7*4 c.c. cobalt, 1-5 c.c. copper, and 4*3 c.c, ferric chloride solution with 86*8 c.c. water containing 0*8 per cent, hydrochloric acid (1-12 specific gravity). For the green 0*5 night-standard, mix in the ratio of 7*7 c.c. cobalt, 2*2 c.c. copper, and 4'5 c.c. ferric chloride solution with 85-6 c.c. water containing O8 per cent, hydrochloric acid (1-12 specific gravity). The extremes being made up, the intermediate colours are produced by the same table as was used for the day -standards. The colour and intensity of the brown and green 0*5 thus prepared will be approxi- mately correct. The final shading must be accomplished in the same manner as for the day-solutions. Colours so produced will be found to be considerably darker than the corresponding day- colours, when compared by day- light. For this work and for the regular night-tests I use a rectangular camera, made of sheet iron, arranged on the back with a door, open in front, and so cut, about midway between -the ends, as to admit a light wooden rack holding the colours. Three medium -sized kerosene-lamps, placed on a stand in front, furnish the light. The rack is pro- vided with a plate of ground glass as a reading back- ground, and is so adjusted that a slide between the glass and the lamps protects the standards from the light when not in use, while the door on the back of the camera serves to darken the other end. With standards made as above, I have found no difficulty in turning out as accur- ate work by night as by day, -and with the requisition of no greater skill and ability than is required of any day c c 2 388 ESTIMATION OF SULPHUR IN IRON AND STEEL. ca,rbon-boy. In these night-standards it will be noticed that there does not exist that sharp contrast in shade between the 0'3 colours produced by diluting the brown and the green 0*5 as is found in the day-standards. It is almost entirely a case of comparative intensity, when, un- like the day-standards, the shade is of but little moment. The kerosene light seems to eliminate to a great degree the power of distinguishing variations in the green tint ; and for this reason a very good set of night-colours could be made by simply diluting the brown extreme with water alone. It will not be as accurate a match, however, as if the green is used also. It takes a little time to familiarise the eye to night-reading ; but when once accustomed it has been my experience that as close, if not closer, discri- minations can be made between the colours by night as by day. In comparing by dilution, the failure to take into consideration the distinction of shade as well as intensity has often been the source of considerable error. With standard solutions made as directed, this liability is reduced to a minimum ; and this, with the manifest advantages of ease and celerity, renders inorganic standards a desidera- tum for any laboratory dealing with continuous work of this kind. Estimation of Sulphur in Iron and Steel. The plan usually adopted is to dissolve a weighed quantity (about 3 grms.) of the metal in strong nitric acid, adding a little hydrochloric acid occasionally, and evaporating the solu- tion to dryness. Dissolve the residue in very dilute warm hydrochloric acid, and precipitate the sulphuric acid in the solution by means of chloride of barium. If the pre- cipitated barium sulphate, after washing once or twice by decantation, has a yellow or brown colour, owing to the presence of iron mechanically carried down, heat it, before filtering, with dilute hydrochloric acid. Dr. Eggertz, to whom analytical chemistry is indebted for the colorimetric process of estimating combined carbon in iron and steel, has devised a very expeditious plan for estimating the sulphur. He takes one decigramme of cast iron, wrought iron, or steel, cut up or pulverised, and EGGERTZS METHOD. 889 passed through a sieve with holes not larger than 0*6 m.m., and introduces it, by means of a glass or glazed-paper funnel, into a flask about 0-15 metre high and 5 centimetres diameter, previously containing 1 grm. of water and 0*5 grm. of concentrated sulphuric acid ; or, in preference, 1-5 grin, of sulphuric acid, sp. gr. 1-25, and whose volume (1-5 c.c.) has been marked on the flask. A piece of polished silver plate (18 m.m. long, 7*5 m.m. wide, and 1 m.m. thick, with a hole at one end), composed of 75 per cent, of silver, 25 per cent, of copper, and attached to a thin platinum or silver wire, is quickly introduced into the flask, so that it may be a little below the neck ; a cork is put in so as to hold the wire without completely closing it. It is allowed to stand fifteen minutes at the ordinary tem- perature, and the silver plate is then removed. If the iron contains sulphur, the plate is coloured by the sulphuretted hydrogen gas disengaged during the solution of the iron in the dilute sulphuric acid ; and, according to the amount of sulphur present, the coloration of the plate passes to a coppery yellow, a bronze brown, a bluish brown, or a blue. These colorations are estimated with the greatest accuracy especially that of the silver plate alone, No. 1 ; that of coppery yellow, No. 2 ; that of bronze brown, No. 3 ; that of blue, No. 4. The intermediate degrees may be represented by decimals, thus : 2 ! 5 if the coloration is between 2 and 3; 3 p l if the plate is one-tenth towards the blue ; 3*5 if it is as much blue as brown ; 3*9 if the brown coloration is feeble. As the normal coloration, No. 2, Dr. Eggertz has adopted that of the bronze called yellow metal, newly rubbed with fine sand on leather. (This metal consists of 60 parts of copper and 40 of tin.) For the coloration No. 3, a convenient alloy has not yet been found. A bronze, con- sisting of 85 parts of copper and 15 parts of zinc, does not quite represent the colour which should be obtained, for when freshly cleaned it is too bright, and at last takes a bluish coloration'. For the colour in question it is better to use a plate of silver which remains in the flask during the solution of the iron, until it has become as brown as 390 ESTIMATION OF. SULPHUR IN IRON AND STEEL. possible,- and a slight bluish colour begins to be perceived ;, the plate is then removed and preserved in a well-closed tube. The colour No. 4 resembles that of a watch-spring. If the amount of sulphur is very considerable, this colora- tion passes to a clear bluish grey. By passing the plate of silver over a bottle of sulphide of ammonium the desired number can be easily obtained. To obtain in these assays for sulphur the proper tint on a silver plate, it is necessary to take certain precautions. The plate is to be held in pincers and cleaned as well as possible by rubbing it with a soft leather on which is placed a little fine rotten-stone. Contact with the fingers is avoided by means of a piece of paper, and the plate is to be carefully dried with a piece of filtering-paper. If the plate, by cleaning or by the action of the burnisher, has been purified on its surface, this should be carefully removed by again rubbing with the leather, for pure silver is less sensitive to the action of the gas than that of the given standard. Thus it has sometimes been found that the silver employed for coinage furnishes less homoge- neous plates, of which those parts richest in copper assume more quickly the blue coloration. On this account, the plates should be compared between themselves, by intro- ducing them at the end of a wire into a flask in which iron is dissolving containing from 0-05 to 0-08 per cent, of sul- phur. On introducing the plate, care must be taken not to turn the side but the edge against the strongest current of gas, whicn would otherwise colour one face of the plate stronger than the other. The plate should be rapidly in- troduced into the flask after the introduction of the iron, as then a very strong disengagement of sulphuretted hydrogen, immediately takes place. After a first experiment the flask is to be filled with water several times, so as to get rid of the odour of sulphuretted hydrogen. If a steel mortar is employed to pulverise the iron, the whole of the piece selected should be reduced to a very fine powder. The mortar should be well cleaned each time, taking care to remove the disc from the bottom. Changes in temperature between 15 and 25C. seem to have no sensible influence on the coloration of the metal ; if the temperature exceeds. EGGERTZS METHOD. 391 40 the plate becomes moist and gives false indications. Some practice is required to judge of the coloration of the plate, but it may be easily acquired. Generally the best plan is to place the standard plates of tints 1, 2, 3, &c., on a sheet of white paper by the side of the plate under experiment, exposing them to a good light near a window (but not sunlight), and to examine them with a lens. The colorations between 2 and 4 are the most difficult to recognise ; but with a little experience none will vary more than Ol ; so that, for instance, the colora- tion may be estimated between 3'5 and 3 '6. The following is somewhat an approximation between the different colorations upon the silver plate and the amount of sulphur in a great number of different samples of iron : Number of Percentage for Coloration of Sulphur 1-0 0-00 1-2 0-01 2-0 0-02 2-5 0-03 3-0 0-04 3-1 0-05 3-2 0-06 3-3 0-07 3-5 0-08 3-6 0-09 3-7 0-10 3-8 0-12 3-9 0-15 4-0 0-20 It is evident that in this way the exact quantity of the sulphur is not estimated ; but several years' experience has shown that if these experiments are made with care, and the quantity of sulphur does not exceed O'l percent., the results are near enough for all practical purposes. Iron which does not colour the silver plate will sometimes produce a coloration if we double the quantities of iron and acid. With half the quantities of acid and sulphurised iron, silver generally gives a little more than half the real quantity of the sulphur which is present. Amongst experiments on the estimation of sulphur in iron, the following deserve mention : 1. The quantity of sulphur in wrought iron is often so small that it produces S92 ESTIMATION OF SULPHUR IN IRON AND STEEL. no coloration on the silver plate ; this iron, therefore, not being red-short, may be employed for all kinds of uses. It must not, however, be forgotten that the quantity of sul- phur is not equally distributed throughout a piece of iron, but that it may vary considerably in different places. On experimenting with the turnings obtained from a portion of an iron bar which was visibly red-short, a stronger tint is often obtained on the silver plate than when using other parts of the bar. The fragments obtained from red-short iron in boring a horse-shoe do not often give on the silver plate a deeper coloration than 2, and it appears to follow that ordinary wrought iron which contains 0'02 per cent, of sulphur in certain parts cannot conveniently be em- ployed for this purpose. If the red-short iron gives to the plate a slighter and more feeble coloration than 2, it may be supposed that the breaking is due less to sulphur than to an insufficient working of the cast iron, the crude pieces in wrought iron entirely free from sulphur often acting as if they were red-short. In general it appears certain that the quantity of sulphur in iron is more injurious when the iron has been badly worked. In a hard iron melted in a steel crucible we may, in spite of its containing O04 per cent, of sulphur, make holes like those in a horse-shoe without any trace of cracks, which may undoubtedly be attributed to the homogeneity and good working of this iron ; the quantity of phosphorus being only 0'3 per cent. The lower portion of an English rolled rail, without fault, contained 0*11 per cent, of sulphur and O03 per cent, of phosphorus ; a portion was cut off which was so red- short that it could not be made use of. 2. The amount, in sulphur, of steel of the best quality is such that the colorations on the silver plate vary only between 1 and 1'5. As in the case of wrought iron, the quantity of sulphur often varies in different parts of the same piece of steel, and that also appears to be the casein a little less decided manner in cast steel. 3. The quantity of sulphur in cast iron is rarely so little as not to colour the plate. In the greater number of Swedish cast irons, this quantity is such that the silver plate varies in coloration between 2 and 3. In iron for NEW COLORIMETE1CAL PROCESS. 30:} gun-castings it is between 3'3 and .3*7, and sometimes more. In cast iron the quantity of sulphur is often dis- tributed unequally ; there is generally more on the surface than is met with below. If the coloration of the silver- plate does not- exceed 3, we can assume that the cast iron refined in the ordinary manner will not give red-short iron, especially if the refining is done carefully. But as in different methods of refining, different quantities of sulphur may be removed from the iron, and in general more if the iron is the result of a light charge of the blast furnace, it cannot be said beforehand that all cast iron which commu- nicates a bluish coloration to the plate will necessarily give red-short iron. This will be the case, however, with cast iron which colours the plate as deep a blue as that of a watch-spring. In cast iron, which gives a red-short wrought iron, without rendering the silver plate more than brown, it is probable (the iron having been well re- fined) that the cause is owing to the presence of other substances than sulphur ; but this occurrence is very rare. Many circumstances appear to show that the quantity of sulphur in iron diminishes with time, at least on the surface, and under favourable conditions. New Colorimetrical Estimation of Sulphur in Iron. The above-described colorimetrical method of esti- mating the amount of sulphur in iron, worked out by Prof, v. Eggertz, has been, and still is, indisputably of great utility to the Swedish iron manufacture. The method since then has come into general use in our ironworks. Mr. J. Wiborgh sought to perfect such a colorimetrical process as would be applicable also for greater amounts of sulphur than that for which the Eggertz method is in- tended. This new colorimetrical sulphur test fulfils the demands that in general may be put upon the practical requirements intended in the method namely, it is quick and easy to execute ; also, it has sufficient accuracy to give the amount of sulphur, whether in cast iron, steel, or wrought iron. The following is Mr. Wiborgh's description of the method. 394 ESTIMATION OF SULPHUR IN IKON AND STEEL. Basis of the Method. Iron dissolves completely in diluted sulphuric acid or hydrochloric acid, and the gases evolved hydrogen, carburetted hydrogen, and sulphuretted hydrogen pass through a cloth impregnated with a metallic salt, and through the action of sulphuretted hydrogen a metallic sulphide is formed, which colours the cloth. From the intensity of the colour afterwards the amount of sulphur in the iron is decided. I proceed here upon the assumption that a given sur- face is always coloured equally strongly by a fixed quantity of sulphur. But to obtain an equal quantity of sulphur from two specimens of iron which have unequal amounts of sulphur, evidently the amounts of iron weighed out must be inversely proportionate to the amounts of sulphur. Therefore a w r eight W of an iron with sulphur amount S gives the same colour as a weight W of another iron with sulphur amount S' ; so must WS = WS'; and as S 7 is the amount of sulphur sought, WS S'-'w 7 '- If you have thus an iron (normal iron) with the amount of sulphur accurately known, you can, by varying the amounts of it weighed out, produce a colour-series in which for every colour you know the product W S, This colour-series constitutes a scale with the aid of which the unknown amount of sulphur S' in another iron can be estimated by dividing the same colour's known product, W S, by the weight of the iron used for the assay. The apparatus shown in the accompanying figure con- sists of a small boiling flask, A, provided with a close-fitting caoutchouc stopper, m, in which are placed a glass cylin- der, R, with one end drawn out to a tube, p, and the other end having a flat polished flange, G, also a funnel tube, T, to introduce the acid. This latter consists properly of two, T and T', united with a caoutchouc tube, over which NEW COLOR IMETRICAL PROCESS. 395 FIG. 103. is placed a nipper tap, K, with a screw, so that the quan- tity of acid to be passed into the flask may be accurately regulated. Upon the cylinder flange is placed a caoutchouc ring, N, and upon this the prepared cloth o. That the ring may close tightly against the flange and cloth, lay upon the cloth another caoutchouc ring, N', of the same size as the former, and outermost a wooden ring, s, which is pressed against the flange by the spring B. These caoutchouc rings ought to have a precise inside diameter, for upon the size depends how much of the sur- face of the cloth will be coloured ; as showing, by the drawing, it is made less than the opening of the cylinder, because it is easier to obtain such rings of a precise size than the glass cylinder. In the apparatus represented the diameter of the ring o is 55, and the opening of the cylinder 58 millimetres. In order that no steam may condense upon the ring s and run down to the cloth, it is best to make this ring of wood rather than of glass or metal, of the size and form shown in the drawing. By the arrangement now described neither gas nor steam can come from the apparatus without first passing through the cloth. To heat the flask to boiling it is placed upon a sand- bath, E, which rests upon the stand D, and is heated by means of a gas or spirit lamp, p. Preparation of the Cloth. In the beginning I used, instead of cloth, unglazed paper, such as filter-paper ; but I soon found that cloth 396 ESTIMATION OF SULPHUR IN 'IRON AND STEEL. was much more serviceable. Such paper certainly per- mits steam and gas to pass ; but it is tender, and easily breaks asunder with the least incaution while boiling. The cloth used is common, fine, white cotton calico. Linen is less suitable, for it is thin, and therefore does not absorb so easily all the sulphuretted hydrogen as evolved. The preparation of the cloth is simply to moisten it with a solution of a metal salt. For this may be employed either lead, silver, copper, cadmium, or antimony salts ; but of all these I have found cadmium salts the most serviceable. Silver and lead salts are certainly very sensitive to sulphuretted hydrogen ; but the combinations of these metals with sulphur are black, and colour the cloth too strongly, so that it will be necessary to employ small weighings of the sample or extravagantly large apparatus. Copper salts certainly give brownish and considerably softer colours ; but, in consequence of the property of the sulphide to enter into variable combinations with the oxide, they are easily oxidised, and not permanent. Antimony has few soluble salts ; and, besides, in one case these are seen to be less sensitive to sulphuretted hydro- gen : they are little fit for this purpose. A sufficiently dark colour is obtained by mixing the salts as, for ex- ample, cadmium and lead salts but has not led to any good results, as the colour comes out uneven, darker upon some places than upon others, so that the cloth has a more or less flannel-like aspect, varying between yellow and black. Eor these reasons I have selected cadmium salt alone for the preparation of the cloth, and this the rather as the sulphur combination of cadmium is a particularly beau- tiful and constant yellow colour ; also, that the affinity of this metal for sulphur is so great that it surpasses that of lead. When the cloth is impregnated with a mixture of cad- mium and lead acetates, equal parts of each, and then exposed to a small quantity of sulphuretted hydrogen, the cloth at first will be almost exclusively yellow, and NEW COLORIMETRICAL PROCESS. 397 by the passage of more sulphuretted hydrogen it becomes dark, which shows well that the gas is sooner decom- posed by cadmium than by lead. Again, concerning which of the cadmium salts ought to be selected for impregna- tion of the cloth it appears to be a matter of indifference which is employed, although the colour tint will be some- what different as one or other salt is used. Cadmium nitrate gives a singularly high and strikingly beautiful orange colour ; the sulphate is somewhat weaker, with brownish yellow tint, while the chloride and acetate give lighter colours. The colours occasioned by the varying amounts of sul- phur may perhaps be most easily judged separately when the cloth is prepared with cadmium nitrate ; but not- withstanding I have likewise used the acetate, because I think it likely that this salt, which has the weakest and most volatile acid, ought to give colours that would be least changeable, which property is of considerable weight. Tn order that all the sulphuretted hydrogen may be absorbed by the cloth, it must contain a certain amount of cadmium salt in proportion to the largest quantity of sulphur which may possibly be present ; otherwise a por- tion of the sulphuretted hydrogen will pass through the cloth, colouring yellow not only the under side but also the upper side. A solution of 5 grms. crystallised cadmium nitrate in 100 c.c. distilled water is of suitable strength. Cloth of the proper fineness, prepared with such solution, allows not a trace of sulphuretted hydrogen to pass through, for the cloth was coloured only on the under side, and when double folds were used the upper had not the faintest colour. The general influence of the degree of concentration of the solution is that the colours from strong solutions lie more upon the surface of the cloth, and that of weaker solutions penetrates deeper into the cloth ; and this causes a certain unevenness in the aspect of the colour, although the amount of sulphide of cadmium in both 398 ESTIMATION OF SULPHUR IN IRON AND STEEL. cases is alike. The influence of a smaller variation in the degree of concentration is not perceptible. The cloth is prepared by cutting, on a round model, several folds to about 80 m.m. diameter. These are laid in the solution of cadmium acetate, care being taken that each circle is thoroughly saturated by the solution. After some minutes take out the pieces, and spread them out on a clean cloth until completely dried ; then place them for safety in a suitable box. The Colour Scale. According to the quantity of sulphuretted hydrogen which reacts upon the prepared cloth, so is it covered with different amounts of cadmium sulphide, and receives a more or less yellow colour. The sensibility in this respect is so great that indeed T oVo tn P art f a ni.grm. of sulphur is able to communicate to the surface of a square centi- metre of the white cloth a certainly weak but yet very manifest yellow colour. Increase successively the amount of sulphur by T^OO^ P art f a m -g rm - per square centi- metre ; you obtain by this small addition of sulphur from the first a clear distinction in the intensity of the. colour ; but in proportion as the strength of the colour increases with the augmented amounts of sulphur, the difference will be all the more difficult to observe. When the amount of sulphur increases to about 0'02 m.grm. per square c.m., the cloth is now strongly coloured, and to produce a manifestly distinct difference of colour intensity it requires two or three times greater excess of sulphur than with the lower colour estimations. Consequently you cannot with advantage make use of very strong colours, because the difficulty of their comparison is considerably increased. In order to estimate the amount of sulphur in an iron, you must first have a colour series, or scale of colours, whereon every colour number represents a certain amount of sulphur per cent., presupposing that a certain quantity of iron has been weighed out for assay. This colour series can be easily procured by the help of NEW COLORIMETRICAL PROCESS. 399 a normal iron in which the amount of sulphur has been estimated beforehand with the greatest accuracy. How great the amount of sulphur this normal iron may con- tain is quite immaterial, but use the most trustworthy among the known methods of estimating sulphur in iron by oxidation of the sulphur, and thereafter precipitating by chloride of barium. With small amounts of sulphur this is to a certain degree unsafe, so it is better to use for the scale an iron with a pretty high amount of sulphur, such as about O'l per cent., because the influence of a small error in this estimation of the sulphur will be less. If you have a known amount of sulphur S, also an iron of which 0'4 grm. is weighed out, and its amount of sul- phur S', to obtain the same colour from both of the irons must be w x s = 0-4 x S' Or w _0'4xS / . In this formula place instead of S' the successive amounts of sulphur per cent., as 0*005, Q-01, 0-02, &c. As you know how much of the normal iron in every case ought to be weighed, so you have a series of colours that are entirely the same as those which will be obtained if you had different irons with their respective amounts of sulphur and 0'4 grm. of each weighed ; you get, in other words, a colour scale, where colours give directly the amount of sulphur the iron contains. Make up such a scale for the apparatus shown in the sketch, with the inner ring 55 m.m. diameter, under the supposition that 0'4 grm. has been weighed for testing. It is not desirable to go farther than this, so that the highest colour number corresponds to 0*1 per cent, sul- phur, because otherwise the colours will be too strong. Nevertheless, you must not take between the high colour numbers a greater difference in the amount of sulphur than between the lower, if the different colour intensities are to be clearly separated from each other. 400 ESTIMATION OF SULPHUR IN IRON AND STEEL. For example, you can allow the scale to be composed of seven colour numbers, taken thus No. 1 corresponding to 0*005 per cent sulphur. 2 0-01 3 0-02 4 0-03 5 0-05 6 0-07 7 0-10 If thus 0-4 gramme iron is weighed for assay, one can, with the help of the above scale, estimate the amount of sulphur up to O'l per cent. The accuracy wherewith the estimations are done is, for the lower of amounts of sul- phur at least, equal to 0*005 per cent., and for higher O'Ol per cent. ; for the differences between the colours of the scale are here so great that one with the greatest ease can estimate a colour lying between two colour numbers. The same scale may, however, be used to estimate quickly whatever amount of sulphur is desired, if only the weight of iron used for assay be varied. It is obvious that absolutely the same colour which an iron gives from 0'4 gramme must be got from another iron with half as much sulphur if 0-8 gramme has been weighed ; and that, in general, if 0'4 n is weighed, the colour scale comes to represent sulphur amounts which are J of those that answer to the weight 0'4 gramme. One ought therefore to use in general greater amounts for assay, when low amounts of sulphur are to be esti- mated with great accuracy, also lesser quantities for high amounts of sulphur. In case the colour comes out too strongly lying upon the margin or beyond the greatest amount of sulphur in the scale, the assay ought to be made with a lesser weight. To avoid calculation one can, under every colour num- ber, place not only the amount of sulphur which corre- sponds to that weighed out in making the scale, but also that which corresponds to some other weighings which possibly may occur. Note the following weighings, W, also the corresponding amounts of sulphur, S, per cent. : NEW COLORIMETKICAL PROCESS. 401 0-8 gramme. 0-005 per cent. 0-4 , 0-01 0-2 0-1 0-08 0-04 0-02 0-02 0-04 0-05 0-1 0-2 With these different weighings are thus given the colour numbers of the amounts of sulphur from 0-005 per cent, to 0-2 per cent. Having by the above method prepared the different colour numbers which form the scale, they may be arranged in order upon small white drawing-paper, bound, and pre- served in a suitable portfolio. When cadmium acetate is used for the preparation of the cloth, the colours will be very constant. They have undergone no sensible change for several months, though preserved only in the above manner. One such scale ought to be made for every apparatus, and observe that the cloth used for the preparation of the scale must be the same sort as that afterwards used for testing. Details of the Process. Rinse every part of the apparatus with water, so that no acid remains from the former assay. The boiling-flask is half filled with distilled water ; then fit the apparatus together and place on the sand-bath, which is heated by means of a gas- or spirit-lamp, so that the water in the flask comes to gentle boiling. While heating, weigh out the iron to be tested. This may certainly be in the form of small pieces, but, that the solution may not proceed too slowly, it is best for the iron to be in the form of filings, turnings, or powder. Difficultly soluble iron, such as white iron high in silicon, phosphorus chrome iron, &c., ought always to be finely pulverised. The weighed sample is brought, by means of a small funnel and hair-pencil, into the test-tube r. This test- tube has the mouth widened. It is set in a loop made on the end of a platinum wire w, which is bent over the edge of the test-tube as shown in the figure, so that the platinum wire may not slip over the tube and rest on the D D 402 ESTIMATION OF SULPHUR IN IKON AND STEEL. bottom of the flask ; this ought to be avoided, because the test-tube may take such a position that the access of the acid to the iron maybe made difficult. The platinum with which the test-tube is thus fastened ought to be rather stiff about 0'3 m.m. diameter. After the water has boiled two minutes, and the air has been expelled, take out the stopper with the cylinder attached. The test-tube with the sample is lowered into the flask, where it rests on the bottom, being held upright by the platinum wire. Put the apparatus together again, and upon the cylinder flange lay the above-mentioned caoutchouc ring, 55 m.m. in diameter, and over it the prepared cloth, another caoutchouc ring, and lastly the wooden ring, all of which are pressed together by the spring clamp. As soon as the cloth is laid on screw up the clamp K on the funnel-tube, and steam must now pass through the cloth. Partly to get the air as much as possible driven out of the apparatus, and partly to moisten the cloth, maintain the water in gentle boiling for about eight to ten minutes before any acid is introduced. Then fill the funnel-tube with diluted sulphuric acid (for example, J volume sul- phuric acid, 1*83 sp. gr., and |- volume water). Open care- fully the screw-clamp K, and allow the acid to drop into the flask. For 0-4 gramme iron use about 10 c.c. dilute acid. As soon as acid comes down, the solution of the iron begins ; steam and gases pass through the cloth, and in proportion as the solution proceeds, also according to the amount of sulphur in the iron, the under side of the cloth is more strongly coloured yellow. After all the iron is dissolved, boil the solution further from five to ten minutes, to drive out the sulphuretted hydrogen which yet remains in the apparatus. Then open the spring clamp, remove the bottom ring, and lay the cloth upon a piece of filter-paper to dry completely. It only remains now to estimate the amount of sulphur by comparing the colour of the cloth with the colour scale. That, the liquid in the flask during the whole period be maintained at an even but gentle boiling is very essential NEW COLORIMETRICAL PROCESS. 403 for this process. The boiling ought to be so strong that the steam is always seen to pass through the cloth, but by no means so violent that the cloth becomes stretched by the boiling. For when it happens that the cloth, in consequence partly of this stretching and partly from the condensed steam, becomes more and more close, and takes a strongly convex form, then the pressure in the apparatus increases, so that on the introduction of the acid the gases formed rush through the funnel-tube. With cautious boiling one need never fear such an occurrence. In order to avoid oxidation of the sulphuretted hydro- gen formed during the solution of the iron, it is of importance that before the acid is introduced into the flask the air be as completely as possible expelled from the water and the apparatus, also that the boiling be so strong that plenty of steam always accompanies the gases which pass through the cloth. The cloth in the assay ought to be coloured evenly, for if the colour is uneven it very considerably increases the difficulty of estimating the colour strength. Whether the colour comes out even or not depends chiefly upon the construction of the glass cylinder. It ought to be so accurately made that the centre line of the tube coincides with that of the cylinder. Further, the tube ought to be short and of a conical form, 7 to 8 m.m. diameter at the lower end. If the tube be too wide the cloth is always coloured un- evenly ; and again, if it be too small, drops of water con- densed in the tube are cast up on the cloth, which give it a spotted aspect. The drawing represents the cylinder so set in the caoutchouc stopper that the mouth of the tube is even with the under side of the stopper ; but it is still better to allow the end of the tube to be 5 to 10 m.m. under the stopper. It is certainly now more difficult for the water to leave the tube, and causes especially at first, before the cylinder becomes warm a weak bubbling ; but this has no hurtful influence, but contributes to the more even coloration of the cloth. Observe, further, that the glass cylinder be placed by D D 2 404 ESTIMATION OF SULPHUR IN IRON AND STEEL. the eye as vertical as possible, also that the apparatus be riot placed in a draught. The time required for a sample is from half to three- quarters of an hour, according as the iron is more or less soluble. This new method, in many special experiments, has given most satisfactory results, in that it is independent of the amounts of carbon and silicon in the iron, indicating high as well as low amounts of sulphur in iron, as seen in the accompanying table, with sulphur estimations in different sorts of iron. These have been performed by this method, and controlled by accurate estimations of the sulphur by the wet method. No. Sorts of Iron tested for Sulphur by Wiborgh's Colorimetric Method and by Chloride of Barium By Wiborgh's Method. Sulphur, p.c. By Wet Method with Chloride of Barium .A. Sulphur, p.c. Estimated by 1 White charcoal cast iron 0-005 0-005 A. Tamm. 2 Spiegeleisenfrom Siegen,with 0*05 per cent, copper . 0-005 0-006 J. W T iborgh. 3 Bar iron, with 0*076 per cent. arsenic . . . . 0-007 0-008 Y. Lagervall. 4 Casting 0-0075 0-005 A. Tamm. 5 Grey charcoal cast iron 0-0075 0-005 A. Tamm. 6 White 0-012 0-01 J. Wiborgh. 7 Casting ..... 0-012 0-014 A. Tamm. 8 Half- white charcoal cast iron 0-018 0-018 J. Wiborgh. 9 Malleable casting . 0-015 0-013 N. Lagerfelt. 10 White charcoal cast iron,with 0*071 per cent, arsenic 0-02 0-025 Y. Lagervall. 11 White charcoal cast iron 0-02 0-02 J. Wiborgh. 12 Malleable casting . 0-023 0-024 E. V. Zweigbergk. 13 White charcoal cast iron 0-023 0-024 Y. Lagervall. 14 ?> > 0-025 0-022 A. Tamm. 15 > J> JJ 0-028 0-029 J. Wiborgh. 16 Cast iron melted with copper, containing 1'55 p.c. Cu 0-039 0-038 5 17 Siemens-Martin iron 0-04 0-037 >5 18 5> >5 0-05 0-047 }> 19 White charcoal cast iron,with 0-015 per cent. Cu 0-06 0-061 JJ 20 Steel 0-07 0-068 21 Siemens-Martin iron . 0-1 0-093 1J 22 Grey cast iron 0-135 0-134 Y. Lagervall. 23 Cannon cast iron . 0-15 0-145 J. Wiborgh. 24 White cast iron from Horde, with 1-88 per cent. P . 0-21 0-19 E. Aquilon. 25 Malleable casting . 0-35 0-34 J. Jungner. 26 WTiite cast iron from coke . 0-45 0-46 J. Wiborgh. 27 " 0-7 0-66 P. G. Lidner. KEW COLORIMETRICAL PROCESS. 405 I have also tested if some impurities in iron such as copper and arsenic had any influence upon this method. For this purpose several sorts of iron containing copper and arsenic were specially selected from the collection of the Mining School ; and in the table it is found I. In the samples 2, 16, and 19 copper has no action ; also II. In samples 3 and 10 the same is probably the case with arsenic, for the certainly somewhat large difference in the sulphur estimation which occurs in sample 10 must be sought in other circumstances, as sample 3 even with its small amount of sulphur and high amount of arsenic shows no noticeable difference. The apparatus may be made of whatever size is wished, but in relation to use in Sweden it is seldom necessary to examine iron with a greater amount of sul- phur than O'l per cent. I regard an apparatus with an inner ring of 55 in.m. as diameter of a suitable size. On the other hand, if one has in general to test iron with high amounts of sulphur, as cast iron made with coke, &c., I would recommend that the apparatus be made somewhat larger. Estimation of Sulphur in Iron Ores. Five grammes of the mineral, ground as finely as possible in an agate mortar, are treated with potassium chlorate and hydro- chloric acid. After desiccation and extraction with hydro- chloric acid and water, the insoluble substances may be lead, calcium, barium, and strontium sulphates, silica, and undecomposed mineral. By stirring well and filtering the liquid whilst warm, the two former salts may generally, however, be dissolved. The filtration should be performed through a double filter, to prevent the pulverised mineral from passing through. When the clear portion of the solution has been poured upon the filter, add to the in- soluble matter 5 c.c. of hydrochloric acid and 5 c.c. of water ; then leave the mixture for two hours on the water- bath at a boiling temperature, when, if care be taken to stir well, the calcium sulphate will be completely dissolved. Wash the insoluble portion in warm water, and pour it on 40G ESTIMATION OF SULPHUR IN IRON ORES. a filter, taking care to place below a flask specially to^ receive that portion of mineral which may have passed through the filter. The filtered liquid, whose volume is about 50 c.c., should be rapidly boiled and mixed with 2 c.c. of a saturated solution of barium chloride. (This amount is sufficient to precipitate the sulphuric acid formed from 0*1 grm. of sulphur.) After cooling add to- the mixture 5 c.c. of ammonia, sp. gr. 0-95, then stir well with a glass rod, and leave the whole to rest at the ordi- nary temperature for twenty-four hours. The clear solu- tion should be decanted as completely as possible on to a strong filter, and the precipitate stirred up with about 20- c.c. of cold water, and then left to itself until it has quite settled. If warm water is used without having added a few drops of hydrochloric acid, a little oxide of iron will be precipitated. The clear liquid is likewise thrown on to the filter, and the operation is repeated several times with cold water, and then two or three times with boiling water,, without which precaution the barium sulphate will pass through the filter. Finally collect the precipitate, and wash it with warm water. The last drops of this water, on being evaporated on a watch-glass, ought not to leave more than a scarcely visible white ring. The precipitate is then dried, heated to redness, and weighed. If it is coloured with iron oxide it must be washed with a little hydrochloric acid, dried in the water-bath, and taken up with a few drops of acid and w r ater, and then the pre- ceding operation repeated (w r ashing, drying, heating, and. weighing). If the precipitate has only a feeble red colour, which is often the case, this latter operation will be un- necessary. To dissolve the lead sulphate which may occur in the insoluble portion, remove this from the filter with the end of a feather, introduce it into a flask, and pour over it 10 c.c. of concentrated ammonium acetate. The solution, after having been strongly agitated and heated in the' water-bath, is carefully poured on to a filter. Then wash the residue with a little ammonium acetate, and repeat the treatment until a few c.c. of solution, acidulated with ESTIMATION OF SILICON IN IRON AND STEEL. 407 a little acetic or hydrochloric acid, is not rendered turbid when warmed with barium chloride solution. The ni- trate is then diluted, slightly acidified, and the sulphuric acid precipitated by means of barium chloride. After the lead sulphate has been removed, there may still occur barium and strontium sulphates in the insoluble portion. To decompose these salts the residues must be dried, heated to redness, and weighed, and then fused with five times their weight of pure dry sodium carbonate. The mass is digested with water over a water-bath, the liquid filtered, and the residue washed with warm water. The silicic acid is separated from the solution, which contains the sodium silicate, carbonate, and sulphate, by adding hydro- chloric acid and drying on the water-bath. After filter- ing, precipitate the solution with barium chloride. To ascertain if the iron mineral contains gypsum or other soluble sulphates, take 5 grammes, place them in 20 c.c, of hydrochloric acid and 60 c.c. of distilled water, and digest them for three hours on the water-bath, with frequent agitation. The filtered solution is mixed with barium chloride and 15 c.c. of ammonia ; then proceed as already described. If there be present in the mineral grains of iron- or copper-pyrites, or galena, they will only give traces of sulphuric acid in this operation. Estimation of Silicon in Iron and Steel. All who are occupied in the analysis of iron and steel are aware how very uncertain the estimation of silicon is when the method hitherto used for its separation is followed, because cast iron, and even bar iron and steel, are never found absolutely free from intermingled slag. This slag is de- composed by the ordinary method of dissolving the iron in acids, and its silica then augments the amount of silica formed from the silicon contained in the iron or steel* This cannot be said of every sort of cast iron, but these sometimes contain blast-furnace slag, although pig iron containing slag may be considered as rare. It ought also to be mentioned that crystallised silicon has been found in crystallised cast iron from Krain, in the form of small silvery plates, which are neither acted upon by boiling 408 ESTIMATION OF SILICON IN IRON AND STEEL. aqua regia nor by ignition in oxygen gas ; but they are converted into silica by fusing with potassium and sodium carbonates. Crystallised silicon is insoluble in hot solutions of sodium carbonate, but is soluble, with development of hydrogen, in hot solutions of caustic potash, and also in hot hydrofluoric acid. The accurate estimation of the silicon in iron and steel has been effected by Dr. Eggertz, who, after fruitless efforts to dissolve iron in highly diluted organic or inorganic acids, which should have no effect on the refinery slag, finally discovered such a solvent in bromine, which, when mixed with water, dissolves the iron without the slightest action on the accompanying slag. But as experimenting with bromine in large quantities is very disagreeable, trials were made to use iodine instead ; and this, like bromine, has been proved to have no effect on the slag, nor on iron oxide or proto-sesquioxide, or manganese proto-sesquioxide. At the same time, bromine dissolves iron quicker than iodine does, and is, perhaps, more easily obtainable in the requisite state of purity. Moreover, as continued experiments have shown that a solution of sodium carbonate can separate finery slag from the silica which has been formed by the use of iodine or bromine on the silicon contained in the iron, the follow- ing method for the estimation of silicon and slag in bar iron or steel has been used and considered successful ; the same method may be employed for cast iron, because blast-furnace slag, when such is found, is not perceptibly changed by iodine or bromine, nor by solutions of sodium carbonate. Three grammes of bar iron or steel which have passed through a sieve of 0*2 of a line are taken for analysis. Fifteen grammes of iodine are added in small portions at a time to 15 c.c. of water in a beaker of 100 c.c. capacity. The water must be previously boiled to expel the air, which would otherwise oxidise the iron. The iodine is stirred in the water with a glass rod, in order to get rid of the air which has accompanied it, and the floating iodine particles are allowed to sink. The beaker with the iodine and water, which is kept covered with a ESTIMATION OP SILICON IN IRON AND STEEL. 409 watch- glass, is cooled in ice before the iron is put in, and during the solution it is kept at the temperature of C. For the first few hours it must be well stirred every hour, or oftener, with a glass rod, but afterwards not so fre- quently. In consequence of the low temperature and the careful admixture of the iron (by which heat is prevented), the solution may be performed without the least de- velopment of gas, and the iron has less tendency to become oxidised by the air when at this low tempera- ture. By pressure, and by agitating with the glass rod, the solution of the iron particles which collect at the bottom of the beaker is much facilitated ; if no lumps of the particles are visible, the beaker may be kept at the ordinary temperature or, preferably, in ice water. If some of the solution has dried on the sides of the beaker or on the glass rod, it must be well moistened with the same solution before water is added. About 30 c.c. of water, which should be very cold in order to pre- vent the formation of basic salts, are added to the solution ; it is then well stirred, left to settle, and the fluid with the lighter particles of graphite is poured on to a filter of 2 in. diameter ; the filtration is kept up without interruption until there remains only a somewhat heavy dark powder of slag, &c., at the bottom of the beaker ; about 5 c.c. of water, with a few drops of hydrochloric acid, are now poured in and stirred with the glass rod ; if hydrogen is given off, it is an indication that there is still some metallic iron undissolved. The acidified water is quickly poured on the filter, in order not to act on the slag. If a de- velopment of gas is perceived, a little iodine, with sodium carbonate and water, is added for the complete solution of the iron, and the residue is thrown on the filter and washed with cold water, until a drop of the filtrate gives no reaction with a solution of 0-2 per cent, of potassium ferrocyanide contained in a small porcelain crucible. Solutions containing 0- 00001 gramme of ferric oxide per c.c. show in this way very distinct reactions, particularly if a drop of dilute nitric or hydrochloric acid be added. The filtrate is evaporated to dryness, in which operation 410 ESTIMATION OF SILICON IN IRON AND STEEL. some of the iodine is sublimed away. Thirty c.c. of hydrochloric acid, 1*12 sp. gr., are then added, and it is again evaporated in order to obtain the silica which may be dissolved in it. The filter, previously dried and weighed, is again dried and weighed when containing the precipitate. It is then ignited, and the residue weighed. After ignition, the residue is boiled in a solution of soda, in order to extract the silica, and weighed. It should be observed that some part of the silica which has been formed from the silicon in the iron may possibly unite with the slag during the drying and ignition. In conse- quence of this, it is difficult to extract it by means of a soda solution, whence this method is not to be recom- mended in exact estimations of silicon. When using bromine as a solvent 6 c.c. must be taken to 8 grammes of finely powdered iron or steel, with 60 c.c. of water, which has been previously boiled and cooled to C. ; and this temperature must be preserved by placing the beaker in ice water until the solution is complete, which usually takes place in two or three hours ; it is cautiously stirred once or twice with a glass rod ; if stirred hastily, the solution proceeds too violently. The further operations are conducted in the same manner as when using iodine. The bromine is most conveniently preserved under water, and is taken up by a pipette, which is in- troduced into the bottle, the upper end being kept closed by the finger. When it is preferred to dissolve iron or steel in lumps, instead of in powder, this may be done ; but it is not then necessary to place the beaker in ice water, as the metal is less violently acted upon in this form. Several days are required for the solution ; the iron, and particularly the pieces of steel, must be occasionally cleaned from the graphite which adheres to their surface. In order to estimate the silica (formed from the silicon in the iron) and slag, the filter, which contains graphite (in combination with iodine or bromine and water), silica, and slag, is unfolded, whilst it is still wet, on a watch-glass. The contents are washed away from the filter (these should ESTIMATION OF SILICON IN IEON AND STEEL. 411 only occupy the lower half of the filter whilst in the funnel) with a very fine jet from a wash-bottle (so as not to obtain too much water) into a platinum or silver crucible of the capacity of 30 c.c. The loosening of the mass may be facilitated by a fine camel-hair brush. The water in the crucible is evaporated to about 6 c.c., then mixed with 3 c.c. of a saturated solution of sodium carbonate free from silica, and the crucible put in a water-bath, It is kept in the boiling water 1 hour, during which time the liquid is stirred two or three times, and the insoluble mass crushed with a platinum spatula. The supernatant liquid is carefully poured from the insoluble mass on to a small filter, and to the mass in the crucible are added 1 c.c. of saturated solution of sodium carbonate, and 2 c.c. of water. When this has been boiled 1 hour, the whole contents of the crucible are thrown on the filter and washed. The solution of silica in soda is acidified by hydrochloric acid, and mixed with the iron solution, and the whole evaporated to dryness on a water-bath. When the solution attains" the thickness of ordinary syrup, it is stirred very often with the glass rod until the mass becomes a dry powder, and heated until the smell of hydrochloric acid has nearly gone off; the beaker is then placed in boiling water for 6 hours, 15 c.c. of hydrochloric acid of 1-12 sp. gr. are then added, and the beaker left on the water-bath 1 hour* As soon as the red powder is entirely dissolved 50 c.c. of water are added ; and when no crystals of iron chloride are visible, the solution is thrown on a filter and washed with cold water, warm water forming basic iron salts which make the silica appear red. The filter containing the silica is dried and ignited in a porcelain crucible, gradually in- creasing the temperature to a full red heat, and weighed ; if the silica is coloured red by ferric oxide, a little hydro- chloric acid, 1*19 sp. gr., must be poured into the crucible, One decigramme of ignited and pure silica obtained from analysis will dissolve in the above manner in 6 c.c. of a saturated soda solution and 12 c.c. of water. If any residue is observed after the second boiling, this arises from some impurity which has united in small quantities 412 ESTIMATION OF SILICON IN IKON AND STEEL. with the silica, rendering it insoluble. When strong hydro- chloric acid is boiled with this insoluble silica, it may afterwards be dissolved. When the solution of silica is diluted with water to the volume of 50 c.c. at the ordinary temperature, it has no tendency to come into the form of jelly. Quartz-powder is dissolved by the preceding method, but very slightly, but ignited titanic acid and finery slag are not acted upon, and the tersilicate slag from blast-furnaces but very little. When the silica is quickly exposed to a high tempera- ture a considerable loss may arise from the spirting of the water combined with the silica. Silica dried at 100 C. has been proved to retain 1 equivalent of water to 3 equivalents of silica that is, about 6 per cent, of water, which is lost by a strong ignition. The ferric oxide is easily dissolved in the heat of a water-bath. The silica is again thrown on a filter, w r ashed, dried, ignited, and weighed. To ascertain the purity of the silica, it may be mixed in a platinum crucible with ten times its weight of pure ammonium fluoride, diluted with water to the thickness of syrup. The water must be evaporated on a water-bath, and the crucible heated, with a cover on it, by a gradually increasing heat over a spirit-lamp to a full red. If nothing is left in the crucible, the silica was pure, and has passed off as silicon fluoride ; but, if anything remains, the opera- tion with ammonium fluoride must be repeated until a constant weight is obtained. When iron contains tungsten, some tungstic acid is formed, and this accompanies the silica for the most part, being dissolved by the soda solu- tion, but it is not volatilised by the use of ammonium fluoride. Yanadic acid also accompanies the silica, behav- ing as tungstic acid. Instead of using ammonium fluoride, it is preferable to use hydrofluoric acid, with which the silica is moistened, and the evaporation is conducted on a water-bath. The mass left on the filter from the soda solution may be composed of besides graphite slag, iron oxide, titanium oxide, &c. (but not copper, at least when the iron does not contain more than 1 per cent.); this is ME. TURNERS PROCESS. 413 dried, ignited, and weighed. The method of separating iron oxide and slag, when the iron or steel contains both these, is not yet known. If the composition of slag were always alike (which it is not) it would be easy to calculate its amount from either the silica or iron oxide obtained in the analysis. In a piece of very red-short Bessemer iron which contained no sulphur, by several experiments 0'3 per cent, of iron oxide has been obtained, and only traces of silicon. After ignition, the iron oxide may possibly be found as sesquioxide. The amount of oxygen, in case the redshortness is due to this, as it probably is, amounts to less than Ol per cent. When the iron or steel for analysis contains titanium, a part of this substance follows the slag in the form of titanic acid. If such is the case, this must be melted with ten times its weight of acid potassium sulphate, by which it is dissolved ; the mass is dissolved in cold water, and the solution precipitated by boiling ; the weight is determined, and subtracted from that of the slag. Mr. Thomas Turner estimates silicon in iron and steel in the following way : For a sample of cast-iron, 2 grammes of metal in the form of borings are placed in a clean beaker of about 8 ounces capacity, and covered with about 50 c.c. of water; 5 c.c. of sulphuric acid are then added, and the beaker kept covered while solution proceeds. The liquid is then evaporated till quite solid, after which the residue is ex- tracted with 100 c.c. of boiling water. The insoluble portion is washed with boiling hydrochloric or nitric acid, and afterwards with water so long as any iron is extracted. It is necessary in all cases to test the filtrate by potassium ferrocyanide or thiocyanate. The insoluble portion is dried, ignited, and weighed in the usual way. In the analysis of steels containing only a small quan- tity of silicon, results by this method agree fairly well with those obtained by other trustworthy processes. In pig- iron of specially good quality good results are also obtained. The conclusions arrived at from Mr. Turner's experi- ments are as follows : 414 BASIC CINDER IN IRON. 1. That with cast irons of specially good quality the silicon can be correctly estimated by evaporation with dilute sulphuric acid. 2. With phosphoric irons the residue obtained, though white, is often impure, and should be further treated in order to obtain accurate results. 3. "With phosphoric irons containing titanium, the silica is contaminated not only with iron, but also with titanic oxide and phosphoric acid. The residue may be very nearly white and still contain 20 per cent, of sub- stances other than silica. 4. On treatment with aqua regia the colour of the residue is usually an indication of its purity. The amount of silicon in grey charcoal pig-iron is about 2*7 per cent, and in spiegeleisen 0-8. per cent. The amount of silicon in pig from coke blast-furnaces is rarely more than 4 per cent. The least quantity of silicon in grey cast iron is about 0*2 per cent., and in white (spiegel- eisen) O'Ol per cent. The amount is usually from 1 to 2 per cent, in cast iron suitable for the Bessemer process, and in pig-iron for puddling about 0'5 per cent. The amount of silicon in iron of different degrees of hardness from the same charge of the blast-furnace ought to be pretty well judged by the fracture, after some estimations have been made by analysis. Estimation of Basic Cinder and Oxides in Manufac- tured Iron. The principal value attached to estima- tions of slag and oxides in defective iron is the almost absolute certainty of discovering whether to the chemical constituents of the metal or to its careless manufacture may be attributed the faults observed. When we con- trast the ' life ' of an iron rail of fair chemical purity with ' mild ' steel rails of nearly the same composition, it is forcibly suggested to even the most superficial observer, that the duration of a rail is proportionate to the cohesion of its metallic particles, due principally to freedom from mechanical impurities. Engineers have ever been alive to the necessity of suitable machinery for expressing the fluid cinder from the iron as it comes from the puddling BASIC CINDER IN IRON. 415 and re-heating furnaces ; but how rarely is the estima- tion of cinder asked for at the hands of the analyst! We occasionally find extremely pure irons and steels pure at least by the result of routine analysis which are con- demned when subjected to mechanical tests. In some cases a crystalline structure of the metal may weaken the iron ; but it is usually the cinder which, by breaking the lines of continuity in molecular structure of the metal, determines its fracture. We regularly estimate the ordinary chemical consti- tuents of a manufactured iron : why not its contained oxides of manganese and iron, and basic cinder, which, when present in excess, weaken the iron and so detract from its value ? The cause of the omission seems to be that no process with the exception of Eggertz's is suitable for technical work. Fresenius's process is, without doubt, accurate ; but what process can be admitted into an industrial labo- ratory (where a complete analysis is required in two days or less), which for each experiment requires a rod of the metal and several days for the completion of the analysis ? An estimation of cinder, &c., by the following me- thod, devised by Mr. Bettel, only occupies an hour or so, and has the advantage of yielding constant results : Five grammes of the borings are heated with a solution of 10 c.c. bromine and 35 grammes potassium bromide in 150 c.c. water. Heat is continued until the iron is dissolved, the solution filtered through a 4-inch Swedish paper (pre- viously washed with hydrochloric acid and boiling water), drained, and washed with a solution of sulphurous acid containing 5 per cent, hydrochloric acid. When the filtrate is practically free from iron, wash with boiling water containing \ per cent, hydrochloric acid ; then with pure water rinse into small platinum dish ; evaporate to low bulk, and dissolve out silica by means of hot solution of carbonate of soda. Boil, dilute, filter, wash with hot water ; then with a ^-per-cent. solution of hydrochloric acid ; finally with water. Dry, ignite, and weigh. Esti- 416 BASIC CINDER IN IRON. mate the silica in the residue in the ordinary manner. The bromine in the first filtrate may be recovered as bromide of potassium in an obvious way. Some analysts object to the bromine process on account of the vapour contaminating the air of the laboratory. To those Mr. Bettel recommends the following process, which has given good insults : Five grammes of the iron rather finely divided are dissolved in 60 c.c. clear solution of cupric chloride (1 in 2) mixed with 100 c.c. saturated solution of potassium chloride. When no particles of iron can be felt by the aid of a glass rod, add 50 c.c. of dilute hydrochloric acid (1 in 20) ; boil and filter through a 4-inch Swedish paper previously moistened with hot hydrochloric acid (1 in 3), then with saturated solution of potassium chloride. Wash the residue on the filter with potassium chloride solution till all the copper is removed, then with hot dilute hydrochloric acid (1 in 50), finally with hot water. Separate the silica as before, ignite and weigh. If the copper obstinately adheres to the paper, as sometimes happens, slip over the tube of the funnel a piece of india-rubber tubing with clip (or plugged with glass rod) ; fill up funnel with strong liquid ammonia ; cover, and allow to remain for half an hour ; then proceed with dilute acid, &c., as before. The results of both processes agree with Fresenius's galvanic method. Estimation of Phosphorus in Iron and Steel. The im- portance of ascertaining the quantity of phosphorus in iron is very great ; for although the presence of a very small quantity of phosphorus in cast iron does not produce any sensible modification in its proprieties, it nevertheless loses its most essential qualities when the proportion of phosphorus amounts to some thousandths. Almost all the methods hitherto proposed consist in treating cast iron with oxidising agents in such a manner as to cause the phosphorus to pass into the condition of phosphoric acid, which is estimated as a magnesian compound. Several sources of error are inherent to this method, to avoid which M. Tantin liberates the phosphorus as an hydrogen ESTIMATION OF PHOSFHOKUS IX IRON AND STEEL. 417 compound. Experiment shows that there is not the least trace of phosphorus in the residue after the complete at- tack of the cast iron by hydrochloric acid, which fact is not surprising if it be considered what strong affinities phos- phorus has for hydrogen. The hydrogen phosphide pro- duced by the action of hydrochloric acid upon cast iron is almost always accompanied by sulphuretted, arseniu- retted, and carburetted hydrogen. In order to effect the separation of these gases, first pass them into a WoulfFs flask containing a solution of potash, which absorbs the sulphuretted hydrogen ; the other gases are then made to traverse a solution of silver nitrate, which transforms the arseniuretted hydrogen into silver arsenite, soluble in the slightly acid liquid, whilst the phosphuretted hydrogen precipitates the silver solution and forms an insoluble phos- phide. The phosphorus being thus completely separated from the sulphur and arsenic, its estimation is effected in a simple manner. The phosphide of silver is treated with aqua regia, and thus transformed into silver chloride and phosphoric acid, which is precipitated in the state of am- monia-magnesian phosphate. Mr. J. B. Mackintosh has conducted several careful experiments on M. Tantin's method, and finds that to secure accuracy the steps to be followed are : 1. Solution in hydrochloric acid in a stream of oxygen or air, absorbing the escaping gases in permanganate acidi- fied with sulphuric acid. 2. Heating the solution to boiling, stopping the passage of the oxygen current, and carefully adding an excess of sulphurous acid solution, and continuing to boil till the precipitated manganese binoxide in the absorption-flask is re-dissolved. This boiling should last several minutes to insure the completion of the reaction. 3. Disconnecting the junctions between the absorption- flasks and solution-flask, and between the solution-flask and the acid-flask, and allowing to cool. 4. Mixing solutions, filtering out residue which is placed (with filter-paper) in a porcelain dish, oxidising E E 418 ESTIMATION OF PHOSPHORUS IN IRON AND STEEL. with nitric acid and potassium chlorate, and evaporating to dryness.* 5. Boiling the solution till the excess of sulphurous acid is expelled, adding a few c.c. of permanganate to per-, oxidise a little of the iron, and precipitating basic acetates. Boiling the nitrate for other precipitates, to insure getting all the phosphoric acid present. 6. Dissolving these precipitates in hydrochloric acid and adding to the solution of the residue, in which by this time the paper will have been thoroughly destroyed. 7. Evaporating to dryness for silica, and proceeding as usual with the molybdate precipitation. The form of apparatus used in these estimations is shown in fig. 104. It consists of a flask, A, to hold the acid ; FIG. 104. a flask, B, for the iron ; and flasks c and D, of which c is generally left empty to catch condensed steam, and the two flasks D contain the absorbent liquid. The absorbent liquid in D is a solution of potassium permanganate acidified with nitric acid. The iron having been placed in B, and the necessary amount of dilute hydrochloric acid to dissolve it in A, the whole apparatus is filled with oxygen, a current of which is kept passing through the whole time. When the air is all expelled * It is better to filter out the residue instead of filtering it out with the first basic acetate precipitate, because it is difficult to work accurately on a muddy liquid, and because there is a probability of loss of phosphorus by the continued action of the hot acid solution on the residue after the sulphurous acid has been expelled. By proceeding as directed this chance of loss is removed. ESTIMATION OF PHOSPHORUS IN IRON AND STEEL. 419 the flask A is inverted above the level of B, so that its contents flow into B and the solution of the iron com- mences. The solution is heated to boiling for some minutes ; and when all has dissolved with the exception of the insoluble portion, the apparatus is disconnected, the insoluble residue filtered from the solution, the solution -oxidised with nitric acid and treated as described above. The saving of time by this process is the greater as the percentage to be estimated is less. In a steel of low percentage, where it is necessary to use 10 grms. for the -estimation, necessitating the use of 120 c.c. of nitric acid for solution, by the old method, the time taken for the evaporation of this amount of liquid, for the subse- quent thorough drying of the residue, and for the reso- lution and reduction of the iron to the ferrous state, is evidently much greater, and the operations followed are much more tedious, than by this method, where the phos- phorus is concentrated at the start in but a few milli- grammes of iron in very little bulk of solution, enabling the subsequent evaporation to dryness and thorough drying of the residue to be performed in a very short period. In M. Tantin's experiments he used cast iron only. Mr. Mackintosh's have been conducted on pig and wrought iron. It is possible that the difference in the results may be partially accounted for by different modes of occurrence of phosphorus similar to the different forms of carbon ; but this is only offered as a suggestion. In calculating the percentage of phosphorus from the weight of the yellow precipitate, Mr. Mackintosh uses in all cases the figure of 1'63 per cent, phosphorus in the precipitate. Dr. J. Lawrence Smith was engaged off and on for two or three years, examining the question of the estimation of phosphorus in iron and steel, making several hundreds of variously modified experiments, and repeating the details of processes adopted by various chemists. The following method was ultimately adopted as affording the most speedy and accurate results. Quantity of Iron employed. It is customary to employ E E 2 420 ESTIMATION OF PHOSPHORUS IN IRON AND STEEL. 1 grm. for pig iron, and 2 to 3 grms. for malleable iron and steel ; but Dr. J. L. Smith employs but 1 grm. for all varieties of iron ; for even where the iron or steel contains one- thousandth and less of phosphorus, as satisfactory results are obtained as where 2 and 3 grms are employed. Solution. The iron, say 1 grm., is placed in a por- celain capsule of about 100 or 150 c.c., and 3 or 4 c.c. of water added ; the capsule is placed on a water- bath, and 10 to 15 c.c. of aqua regia are added little by little ; the aqua regia is prepared in advance in the usual way with 2 parts hydrochloric acid and 1 part nitric acid. The contents of the capsule are now evaporated to dry- ness over the water-bath, or more speedily on an iron plate ; the capsule with its contents is then placed in an air-bath and heated from 140 to 150 C. for from 30 minutes to 1 hour, thus rendering all the silica insoluble ; 3 or 4 c.c. of hydrochloric acid with an equal quantity of water are added to the dry residue, and then warmed gently over a water-bath or lamp ; the iron is re-dissolved, a little more water added, the solution filtered with the filter-pump, the filtrate placed on a narrow graduated measure of 100 c.c. capacity, and sufficient water added to make the liquid contents 100 c.c. ; the whole is well shaken to make the solution uniform. The next step is to concentrate all the phosphorus into a limited amount of the iron. Concentration of the Phosphorus. From 90 to 92 c.c. of the last solution is placed in a capsule of 300 or 400 c.c. capacity, either of porcelain or platinum the latter by preference and 100 c.c. of water added ; the iron oxide is now reduced to iron protoxide by ammonium sulphite.* Two or three centimetres of the ammonium sulphite are added to the iron solution and the contents of the capsule are boiled until all the sulphurous acid is driven off, this stage of the process being recognised by the sense of smell. By putting a small drop of the solution on the * Equal parts of ammonia and water are placed in a bottle and an excess of sulphuric acid passed through ; the operation lasts for several hours, using a mixture of charcoal and sulphuric acid. Once prepared, it keeps very well, when kept from the light. J. LAWRENCE SMITH'S PROCESS. 421 end of a glass stirrer into a weak ammonia solution we readily recognise the complete conversion of the oxide, for the precipitate is nearly white. Of course during the whole of the above process the solution is acid, with the excess of hydrochloric acid. Ammonia is now added slowly to the warm solution until a little of the greenish precipitate remains undissolved ; about 20 c.c. of acetic acid is now added to the solution (which immediately re- dissolves the precipitate), and then 1 or 2 c.c. of ammonia acetate solution; finally, the 8 or 10 c.c. of original solu- tion remaining in the graduated glass is added with 200 or 300 c.c. of water. The whole contents of the large capsule is boiled gently from one-half to one hour, and if necessary the water renewed as it is evaporated. The result is the for- mation of a basic per-salt of iron containing practically all the phosphorus that was originally in the gramme of iron used. Separation of the Phosphorus from the above Precipi- tate. With a filter-pump on a 3^-inch filter, the last pre- cipitate is collected in 15 or 20 minutes. The precipitate is not washed, but a mixture of 5 or 6 c.c. of hydrochloric acid, with an equal quantity of water, is warmed in the capsule in which the boiling has taken place, so as to dissolve the adhering oxide of iron ; the hot acid solution is thrown on the filter in the funnel, detached from the pump ; the filtrate is readily dissolved, and passes into some convenient vessel, and the filter washed once or twice ; this solution is placed in a porcelain capsule and evapo- rated to dryness over a water bath or on a hot plate. The former is preferable, although it takes a longer time. To the dry, but not over-heated residue is added 1 to 2 c.c. of nitric acid, with an equal quantity of water. This will furnish a clear solution if there be no titanium in the iron ; if the latter be present, there will be formed a floccu- lent precipitate that can be readily separated by a filter prior to the last treatment. The last Treatment. The solution now need not be more than 10 or 20 c.c., to which ammonia is to be added 422 ESTIMATION OF PHOSPHORUS IN IRON AND STEEL. until the precipitate first formed is no longer re-dissolved ; then add a few drops of nitric acid to clear up the solution completely, in which the phosphorus is supposed to have been concentrated. 30 c.c. of molybdic acid solution is now added to the last solution in a small beaker, which is then warmed for 15 or 20 minutes to a temperature of 80 0., and agitated with a glass rod. The phosphorus is precipitated as the double ammonia-salt, and settles as a chrome-yellow powder in less than 30 minutes, and is ready for collection on a double filter ; * although it is better to allow two hours or more time to elapse before filtering and washing with the filter-pump. As the filter is very small, it is readily washed with a little distilled water. After washing, the double filter is placed in an air-bath heated to about 120 C., and in about 30 minutes weighed by separating the filters ; the complete dryness is verified by a second heating in the air-bath. Of the phospho-molybdate every ] 00 m.g. will contain 1-63 m.g. of phosphorus, or 3*74 m.g. of phosphoric acid. The result of this method of analysis will indicate a very minute quantity of phosphorus less than what is contained in the iron, but so small as not to affect the practical result, and will be more accurate, certain, and speedy than if estimated as magnesian phosphate. Estimation of Manganese in Iron. After the silicon is estimated in the iron or steel by Eggertz's method (p. 408), the manganese may be estimated in the same- amount of material. The filtrate from the silicate is diluted with water until it measures 400 c.c., and accurately divided into two portions of 200 c.c. each, one of which is set on one side, and in the other the manganese is estimated in the following manner : (In the case of wrought iron * When filtering a precipitate to be weighed on the filter, a double filter is used, each of the same size ; they are weighed one against the other and exactly balanced by the weights ; on the lighter one a -f mark is put with pencil, and the number of m.g. that it is lighter than the other. As only a 2 or 3 inch filter is used, the difference in weight between the filters does not usually exceed 10 or 20 m.g. A number of these double filters (with the differ- ence marked on them) may conveniently be kept ready for this or any other purpose. ESTIMATION OF MANGANESE IN IRON. 423 and steel, where 3 grammes are taken, the solution is diluted to 200 c.c., and the manganese estimated without dividing the solution.) A saturated solution of sodium carbonate is added to the solution until it becomes nearly neutralised, appearing of a deep brown colour ; water containing 5 per cent, of sodium carbonate is then added, drop by drop, until a slight turbidity occurs in the solu- tion ; and if, on standing in the cold, this turbidity rather increases than diminishes, sufficient carbonate has been added (if too much sodium carbonate has been added, and a precipitate is thrown down, the excess must be neutralised by hydrochloric acid) ; to the slightly turbid solution add 1-^ c.c. of hydrochloric acid, and heat on the water-bath until the solution becomes clear ; dilute with about half as much water as the volume of the solution, and add 30 c.c. of a saturated solution of sodium acetate ; boil for a quarter of an hour ; allow the precipitated iron to settle, and decant the clear liquid through a filter, washing the iron by decantation with boiling water containing -J per cent, of sodium acetate ; finally, throw the iron o$ the filter, and continue the washing until bromine water gives no reaction, showing that all the manganese has passed through the filter ; evaporate the filtrate down to 400 or 500 c.c. ; and at the temperature of 50 C. add a few drops of bromine to precipitate the manganese, and keep it near to that temperature for twelve hours, stirring occasionally with a glass rod. The solution after the addition of the bromine becomes of a yellow or brownish colour, but should be perfectly colourless before filtering. The man- ganese is now thrown on a filter which has been dried at 100 C., and accurately weighed, washed with cold water containing 1 per cent, of hydrochloric acid, dried at 100 C., and weighed. The precipitate is a hydrated manga- nese oxide, containing 59'21 per cent, of manganese. The .precipitate may also be ignited in a porcelain crucible at a white heat, and is then an anhydrous manganese oxide, containing 72-05 per cent, of manganese. If copper is present it must be removed previous to the precipitation of the manganese ; or the amount of copper found in the 424 ESTIMATION OF MANGANESE IN IRON. ignited oxide, and then an equivalent amount of copper oxide, subtracted from the total weight of the precipitate. In using the first method, 20 grains of the finely divided spiegeleisen are completely dissolved in hydro- chloric acid, diluted, and a current of sulphuretted hydrogen passed through the liquid. After standing for twelve hours, the solution is filtered and washed with water containing sulphuretted hydrogen ; the filtrate is boiled, 10 grs. of potassium chlorate added, the iron sepa- rated, and the manganese estimated in the usual manner. If the method used be that of estimating the copper in the precipitate, the estimation must be made with the greatest care, on account of the small quantity of copper present ; the solution must be decanted immediately the zinc is completely dissolved, and excess of acid must be carefully avoided ; otherwise the film of copper will par- tially re-dissolve. It is evident that if the precipitation be effected by ammonium sulphide or sodium carbonate separation or estimation of the copper is likewise neces- sary. Mr. E. Eiley ('Chemical News,' April 27, 1877) gives the following instructions on the estimation of manganese in spiegeleisen, and in many auriferous iron ores : There are two methods now in use (a) The Direct Method. The pulverised spiegeleisen (about 1 grm.) is dissolved in dilute nitric acid, sp. gr. 1*2, a little potassium chlorate and hydrochloric acid added to destroy the soluble organic matter from the combined carbon ; the solution, diluted to about a litre, is neutralised with sodium or ammo- nium carbonate, sodium or ammonium acetate added, the solution boiled, the basic iron peracetate allowed to settle, and filtered off. This precipitate is re-dissolved in hydro- chloric acid, and the process repeated to insure complete separation of the manganese. The filtrates are evaporated to 1-^ litre, allowed to cool, 2 to 4 c.c. bromine added, the solution well shaken, 0*88 ammonia added in excess, the solution heated gradually for an hour, boiled for a few minutes, the precipitate allowed to settle, filtered (the filtrate should be evaporated and tested for manganese), MR. EILEY'S PROCESS. 425 dried, and ignited in a muffle, or over a gas blowpipe for half an hour. (b) The Indirect Method. The finely powdered spiegel- eisen (about 1 grin.) is dissolved in dilute sulphuric or in hydrochloric acid, the liquid diluted with recently boiled and cooled distilled water, and the iron estimated volu- metrically ; to the percentage of iron thus obtained 5 per cent, is added for carbon and impurities ; the difference is assumed to be manganese. The results obtained by this method are usually too low, from the formation of soluble organic matter during the process of solution. This error can be obviated by using nitric acid for a solvent, evaporating to dryness and heat- ing ; the iron and manganese oxides are then dissolved in hydrochloric acid, the solution largely diluted, and reduced with sodium sulphite. The results thus obtained agree very closely with the direct method. Thus, for all practical purposes, the indirect method is sufficiently accurate, and can be accomplished in one hour, the direct method re- quiring five or six hours. The author strongly recom- mends the use of ammonium acetate and carbonate, instead of the corresponding soda salts in the direct method ; and proves by check experiments with pure manganese sul- phate, &c., the statements of Fresenius and others, that the presence of ammoniacal salts prevents the complete pre- cipitation of manganese by bromine and ammonia, to be erroneous. On the other hand, if soda salts be used, the ignited precipitate will contain soda. The author con- siders that sulphur cannot exist in spiegeleisen. He es- timates the carbon by dissolving the iron in neutral copper chloride, and after complete solution of the iron and precipitated copper, the carbon is filtered on asbestos, and burnt with copper oxide in a current of oxygen. The carbon estimations by the colour test are unsatis- factory for high percentages of carbon. According to the author, the percentage of carbon varies with the percentage of manganese. 426 ASSAY OF MANGANESE IN SPIEGELEISEN. Assay of Manganese in Spiegeleisen, Ferro-manganese, and the most important ores. W. Kalman and Alais Smolka have ascertained that manganous oxide when opened up with a flux of borax and potassium-sodium carbonate with access of air yields an oxidation product containing 5 atoms available oxy- gen and 6 atoms manganese. The flux is obtained by melting in a platinum capsule 2 parts borax glass and 3 parts of the double carbonate, and pulverising the very hygroscopic mass while still warm. There is be- sides required a solution of ferrous sulphate and a per- manganate solution. [The former is obtained by dissolving about 100 grammes ferrous sulphate in 1,000 c.c. water, acidulating with sulphuric acid, filtering, and mixing the solution with 100 c.c. of pure undiluted sulphuric acid. The permanganate solution is standardised for iron, 1 c.c. of the solution preferably representing 0-0025 grm. iron. Its standard for manganese can be calculated from the following proportion : iron standard : x 10-56 : 6-55. For applying the method, from 0-15 to 0*30 grm. of the sample, very finely ground, is ignited for fifteen minutes in an open platinum crucible with a Bunsen burner, and then more strongly with a blast. By the ignition the manganese is chiefly converted into mangano- manganic oxide. The crucible is let cool, covered, and about 20 parts of the flux weighed in. Heat is slowly applied till the mixture is melted, care being taken that not much spirts up upon the lid, as such portions become oxidised to manganate and make the result too high. The contents of the crucible are kept in a state of fusion for fifteen to twenty minutes, the lid is then removed, the crucible placed slanting, and the fusion is continued for five minutes longer, stirring with a platinum spoon. Equal quantities (10 to 15 c.c.) of the iron solution are poured into two beakers and diluted. The contents of the crucible are dissolved in the solution in one of the ASSAY OF MANOANESE IN SPIEGELEISEN. 427 beakers, adding a little strong sulphuric acid if requisite, when the manganese compound formed oxidises a part of the iron. The solution of ferrous sulphate in both glasses is then titrated with permanganate. The difference multiplied by the manganese standard gives the propor- tion of manganese in the sample. In cases where the manganese exists as a silicate the results are only approxi- mate. It is suitable for all cases where the manganese in the sample may be converted into mangano-manganic oxide by simple ignition, where the proportion of man- ganese is at least 1 to 2 per cent., and where no other substance is present which can become capable of giving up oxygen to a ferrous solution, such as chlorine. The accurate estimation of manganese in spiegel- eisen is of commercial importance ; as, being the most important constituent, the value of the material is fre- quently judged by the percentage of that element alone, while the error introduced by the presence of copper is aggravated by the fact that not only is copper worthless but absolutely injurious. The following method for the assay of manganese in iron and steel by Mr. Samuel Peters, Bay State Iron Works, South Boston, is not new in principle, but has given very satisfactory results : Dissolve O'l grm. pig iron or steel in 3 or 4 c.c. nitric acid, about 1-2 sp. gr., and boil gently in a long test-tube (about 8 inches long and f-inch diameter) for five or ten minutes, or until solution is complete ; then add an excess of plumbic oxide, say 0-2 or O3 grm., and boil again two or three minutes.* Cool the tube and its contents in water. Filter through asbestos, washing out the test-tube and the residue on the filter with dis- tilled water until all the colour has been washed through. Transfer to a graduated tube (f-inch in diameter), hold- ing 50 or 60 c.c., graduated in 0*2 c.c., and compare with a standard solution of permanganate held in a tube for that purpose. The comparison is made in the same * It is unnecessary to filter off graphite in pig iron before boiling with plumbic oxide. 428 ASSAY OF MANGANESE IN SPIEGELEISEN. manner as that in the Eggertz method when estimating combined carbon in steel, &c. The solution under com- parison is then diluted and well mixed with distilled water (by pouring the contents of the graduated tube into a small dish, and then transferring to the tube again), until its colour is exactly of the same intensity as the standard solution. Having attained to this point, the number of c.c. is noted, and the result is obtained by multiplying each c.c. by O'OOOl. Each c.c. is equivalent to 01 per cent, manganese when Ol grm. of iron is taken for analysis. For irons containing O'lO to 0*35 per cent, manganese, Ol grm. is the proper quantity ; but if there be, say, 0-8 to TOO per cent., it is best to take 0-1 grm. and divide the solution (before adding the lead peroxide) in four equal parts, and use O25 for the estimation, taking another 0'25 for a second estimation. In case of a high percentage, as 1 per cent., if O'l grm. is taken the re- sults are too low on account of some of the manganese escaping oxidation. This agrees with the observations of others. With an unknown iron, one or two trials with O'l grm., or half that quantity, will point out the probable amount, and so be a guide for the next trial. If the amount of iron taken does not yield more colour than corresponds to 25 to 35 c.c. of standard hue, it may be safely said that all the manganese is oxidised. It is as well to take this volume as the guide to the quantity of iron to be taken. The quantity of manganese in the liquor to be tested should not exceed 0'4 of a millgramme, and certainly not over half a milligramme. By taking 0*1 grm. of a spiegeleisen containing nearly 12 per cent, manganese, and diluting to 50 c.c., and taking 2 c.c. or 0'04 for the estimation of the manganese, very nearly the proper amount of manganese is obtained. This seems to show that if the division of the solution can be accurately made, and the bulk of the coloured liquid can be kept down well, the amount of manganese in spiegeleisen can be estimated very fairly. Combined carbon in large quantity does not inter- ASSAY OF MANGANESE IN SPIEGELEISEN. 429 fere with the accuracy of the method, for a steel con- taining 2 per cent, combined carbon and only 0-8 per cent, manganese was found to give good results by this method. The standard is made by diluting a permanganate of potash solution of known strength until each c.c. = 0*00001 grm. manganese. For example, a T n F solution will contain 3-16 grm. permanganate in 1000 c.c. or 0*0011 grm. manganese per c.c. ; if this be diluted 110 times it will give the required strength. The standard is contained in a tube of the same bore as the one used for the analysis ; or else the standard is put in the latter one, and a solution of permanganate put into a tube of nearly the same bore, and diluted until it exactly corresponds with the standard solution, when it will serve as a standard. Permanganic acid of the proper hue keeps better than permanganate of potash of the same hue, and is of course easily made by adding nitric acid to the latter. The time occupied in obtaining a result by this method is very short (about half an hour), and it is a method that will prove of advantage in analysing steel made by the Bes- semer and Siemens-Martin processes. Mr. Galbraith (< Chemical News,' February 4, 1876) has given the following simple and accurate process for the assay of manganese in spiegeleisen : 1 gramme of the spiegeleisen is dissolved in nitric acid (sp. gr. 1-20) in a small round-bottomed flask, and evaporated to dryness. When dry, the flame, which may be either a spirit-lamp or a Bunsen burner, is turned so that the bottom of the flask is cherry-red for ten minutes. It is then allowed to cool very gradually. At this point a weighed quantity of ammonio-ferrous sulphate, or ferrous sulphate of a known strength, is put into the flask and then heated with rather dilute hydro- chloric acid. The contents of the flask very soon dissolve ; but it is well to keep shaking the solution while it is being heated, to prevent loss of chlorine. It only remains now to estimate the iron left unoxidised, in order to arrive 430 ESTIMATION OF TITANIUM IN IRON. at the quantity of manganese, which can be done with potassium bichromate solution. It is feared that the ferrous solution may get oxidised by exposure to the air ; a small piece of marble put into the flask, which can also be fitted with a cork and tube, will readily prevent that. In four successive experiments the following results were obtained : No. Fe Equal to Oxidised Manganese p. c. 1 0-2018 19-82 2 0-2103 20-65 3 0-2396 23-53 4 0-2435 23-88 No. 2 gave by ammonium acetate method 20-55 per cent., which was done with great care. No. 4 is a repeti- tion of No. 3. It is evident, of course, that there is nothing original or new in the above method ; but it contrasts very favourably with the usual methods of separating the iron with sodium or ammonium acetate, and precipitating the manganese from the filtrate with bromine. It is not at all trouble- some, does not take long, and has the advantage that the only chemicals and apparatus required are those which are necessary for the assay of iron ores. Estimation of Titanium in Iron. The detection of titanium in iron is easy, although its estimation is difficult. The best results have been obtained by following Kiley's plan.* This is essentially as follows : A weighed portion of the iron borings are treated with fuming nitric acid in a flask, a few drops of hydrochloric acid added from time to time, the whole being well boiled. The contents of the flask are then transferred to a porcelain dish, evaporated to dryness, and heated strongly. On cooling, it will be found that the iron oxide readily detaches itself from the dish, and can be easily transferred to a beaker, the portions left on the dish being dissolved in hydrochloric acid, and poured on the contents of the beaker ; the dish may be rinsed out, if necessary, with strong hydrochloric acid. The contents of the beaker are boiled for from two to three * Chemical News, viii. 226, 233. ESTIMATION OP TITANIUM IN IEON. 431 hours, until complete solution of the iron is effected ; and as some quantity of hydrochloric acid is required for this, the best plan is to allow a large portion of the excess of acid to evaporate in the beaker, retaining only as much as is requisite to keep the iron in solution. The silica is filtered off in the usual way, after diluting with water and adding a few drops of hydrochloric acid on the filter to dissolve the basic salt formed by the water. By this means the silica can be obtained nearly white after burning off the graphite, and very little iron will be found with it unless much phosphorus be present, as the silica invariably contains more or less iron phosphate from the insoluble iron phosphide, which cannot be completely dissolved out by hydrochloric acid. Before estimating the titanium the residue from the silica should be fused with potassium bisulphate, dissolved in water, and added to the solution of iron in which the titanium is to be estimated. The solu- tion is reduced with sodium sulphite, and the excess of sulphurous acid is driven off by boiling. The solution is then near]y neutralised with ammonia, and ammonium or sodium acetate added ; if there is only a small quantity of phosphoric acid, there will always be sufficient peroxide of iron to precipitate it, but if not, a few drops of nitric acid must be added so that the precipitate produced is distinctly red, and the solution boiled and filtered as quickly as possible. The residue is fused with potassium bisulphate, or, where nitric acid is used, this is driven off with sulphuric acid. The result of the fusion with potas- sium bisulphate is dissolved in cold water (when a little iron phosphate, which remains insoluble, is separated), boiled for some hours, and allowed to stand a night in a warm place, when the titanic acid is filtered off and washed with dilute sulphuric acid, dried, ignited, and weighed. The above process is not very satisfactory for the quantitative estimation of titanic acid. The iron phos- phate (insoluble in the potassium bisulphate) cannot be washed without its passing through the filter ; and very frequently, also, the small amount of iron keeps up the 432 ESTIMATION OF TITANIUM IN IRON. titanic acid, as iron even in small quantities has a very great effect in preventing the precipitation of titanic acid ; so that it is always advisable to add a little sodium sulphite, which reduces the iron oxide and facilitates the precipitation of the titanic acid. Titanium may, however, be found more satisfactorily and more readily during the process usually adopted to estimate the amount of graphite in pig iron, provided a large quantity of the pig be operated on. About 200 grains of the pig are to be dissolved in dilute hydrochloric acid ; when the pig is nearly all dissolved, and the action of the acid has ceased, more hydrochloric acid is added, and the solution well boiled, so as to thoroughly extract all the iron. The solution is then thrown on dried counterpoised filters encircling each other, and the filter well washed to remove all the iron. It is then treated with dilute potash, and washed once ; then re-treated with it so as to entirely remove the silica. The potash is thoroughly washed out, and the filter treated with hydrochloric acid, thoroughly washed, and dried at 250 F. until the weight is constant. This gives the graphite, on burning which a residue of a dirty light brown colour is left, which, fused with potas- sium bisulphate, and subsequent treatment as above ex- plained, is seen to be nearly pure titanic acid. Mr. W. Bettell (' Chemical News,' Aug. 22, 1873) pro- poses the following modification of Mr. D. Forbes' process (' Select Methods in Chemical Analysis,' second edition, pp. 194, 195) for the estimation of titanic acid, which may not be unacceptable to those engaged in the analysis of titanic ores : Fuse about 0-5 grm. of the finely powdered ore with 6 grms. of pure potassium bisulphate (which has been recently fused and powdered) in a platinum crucible at a gentle heat, carefully increased to redness, and continued till the mass is in a state of tranquil fusion. Eemove from the source of heat, allow to cool, digest for some hours in 5 or 6 oz. of cold distilled water (not more than 10 oz. is to be used, as it generally causes a precipitation of some titanic acid) ; filter off from a little ESTIMATION OF HAKDNESS. 433 pure white silica, dilute to 45 or 50 oz., add sulphurous acid till all the iron is reduced, then boil for six hours, replac- ing the water as it evaporates. The titanic acid is precipitated as a white powder, which is now to be filtered off, washed by decantation, a little sulphuric acid being added to the wash-water to prevent it carrying titanic acid away in suspension. Dry, ignite, allow to cool, moisten with solution of ammonium carbonate, re-ignite, and weigh. The titanic acid is in- variably obtained as a white powder, with a faint yellow tinge, if the process has been properly carried out. This method of fusing with potassium bisulphate (' Select Methods in Chemical Analysis,' 2nd edition, p. 212) is preferable to all others for decomposing difficultly soluble iron ores. The Hardness of Iron and Steel. Mr. T. Turner has given the following useful table of the hardness of different varieties of iron and steel in comparison with that of other bodies. The figures repre- sent the weight in grammes necessary to produce a scratch with a diamond on drawing its point over the smooth sur- face of the metal. Substances Steatite Lead (commercial) Tin . Eock salt Zinc (pure annealled) Copper (pure anne Calcite . Softest iron Fluor-spar Mild steel Tyre steel Good cast iron Bar iron Apatite . Hard cast-iron scrap Window -glass Good razor-steel Very hard white iron Relative hardness ' 1 ) 1 2-5 4 ed) 6 aUed t 8 12 15 * L - 19 21 20^24 - - 2124 24 34 rap 36 60 \ 60 ron 72 F F 434 CHAPTER X. THE ASSAY OF COFFEE. IN the assay of copper by the dry way, all minerals and substances containing that metal may be divided into three classes. CLASS I. Comprises Sulphuretted Ores or Products, with or without Selenium, Antimony, or Arsenic. Copper glance, Cu 2 S, containing 79'7 p. c. of copper Chalcopyrite, Cu 2 S, Fe 2 S 3 , 34-4 Erubesoite, 3Cu 2 S, Fe 2 S 3 , 55'7 Bournonite, 3Cu 2 S,SbS 3 + 2(3PbS,SbS 3 ) 12-7 Fahlerz, 4(Cu 2 S,FeS 5 ZnS,AgS,HgS).(SbS3,AsS 3 ,Bi 2 S 3 ) 3048 Covelline CuS, 66'7 Wolfsbergite, Cu 2 S,SbS 3 , 24'9 Domeykite, Cu 6 As, 71'6 Copper regulus, Copper speiss, &c. CLASS II. Oxidised Ores and Products. Red copper, CojO, containing 88'7 per cent, of copper Malachite, 2CuO, + H 2 0, 57'3 Azurite, 2CuO,C0 2 + CuO,H 2 0, 55-1 Cyanosite, CuO,S0 3 +5H 2 Q, 25'3 Phosphate of copper, 30 56 Arseniate of copper, , , 25 50 Chromate, Vanadate, and Silicate of Copper ; Slags, &c. CLASS III. Copper and its Alloys. The different methods of assaying copper are more numerous than those for any other metal. They are in some cases similar to each other, and in others based upon very different principles. CLASSIFICATION OF THE COPPER ASSAYS. 435 These methods may be divided into : A. ASSAY IN THE DRY WAY. a. For Rich Ores and Products of Class I. I. English Copper Assay. b. For Ores and Products of Class II. 1. Lake Superior Fire Assay, B. ASSAYS IN THE WET WAY. a. Colorimetric Copper Assay. 6. Volumetric Copper Assay. c. Electrolytic Copper Assay. A. ASSAYS IN THE DRY WAY. a. For Rich Ores and Products of Class I. I. ENGLISH COPPER ASSAY. M. L. Moissenet has given, in the 'Annales des Mines,'* a very complete description of the Cornish method of assaying copper by the dry way. The following is from a translation by Mr. W. W. Procter. Each of the large Swansea copper-works keeps an assayer at Cornwall, whose duty it is to estimate the richness in copper of all the lots of minerals in the county sold every Thursday at the Ticketing, and of all the samples of foreign minerals and copper products which may be useful to the smelter. The copper being obtained in the state of a prill or metallic button, the impurities (generally tin, antimony, &c.) are thus made evident, and the hammer soon proves the quality of the metal which we ought to expect to obtain by metallurgic treatment. As for the accuracy of the method, as far as regards the whole of the metal obtained, we shall revert to this later on. We would, however, observe that, within certain limits, the method would not be less practical on account of being inexact ; * Vol. xiii. p. 183. p p 2 436 ENGLISH COPPER ASSAY. for we must not forget that it has chiefly for its object to teach the smelter the value of the mineral, even more than its true richness. For example, if we get too low an assay from a sample of 2 or 3 per cent., we should only from this assent to the opinion of the metallurgist, whose interest it is not to work upon very poor minerals. The same remark will apply to the case of minerals very antimonial, &c. Besides, in the description of the method we shall discover the principal phases of the Welsh process ; so that it is more just to con sider the Cornish assay as a metallurgy on a small scale than as a scientific laboratory method. From thence result also the necessity of long practice and the almost uselessness of theoretical knowledge-for those who purpose employing this method alone. Sir Henry de la Beche (' Eeport on the Geology of Cornwall,' &c., p. 595), in giving a sketch of the method, declares it to be rather rough and uncertain, and fails not to add at the conclusion a translation of a passage relative to the assay of copper pyrites from M. Berthier's treatise on assays by the dry way. These drawbacks upon the scientific value of the Cornish method cannot injure the power of facts ; they constitute but another reason which we may have for giving an ac- count of the manner in which the first basis of the valua- tion of the greater part of the copper minerals has been fixed since so long a period. Division Adopted. The rather complex operations through which we have to pass will be better apprehended by explaining in succession 1. The order of the operations, the nature and influence of the fluxes employed, the kind of products obtained (reactions). 2. The manipulations to which each operation gives rise, the furnaces and apparatus used, the characters of the principal products during the chief phases and at the end of each (manipulations). We shall add to these 3. Some information upon the influence of the princi- ENGLISH COFFEE ASSAY. 437 pal foreign metals (tin, antimony, zinc, lead), and upon the treatment of some special coppery matters. 4. Summary considerations on the result of the English method compared with those of the analysis by the wet way. SECTION I. EEACTIONS. At the very outset we distinguish two kinds of assays. 1. The roasted sample. 2. The raw sample. The first only applies to cupreous pyrites or to samples essentially formed of it that is to say, which contain sulphur in excess ; the process begins by a roasting. In the raw assay we dispense with the roasting ; we have recourse to the addition of reagents, either oxidising or sulphurising, according to the minerals ; we endeavour to place them by these mixtures in the condition of a pro- perly roasted pyritic mineral. From this point, at least in general, the operations become identical. They consist in 1. Fusion for regulus (regulus). 2. Calcining the regulus (calcining). 3. Fusing for coarse copper (coarse copper). 4. One or two fusions with fluxes (washings). 5. Trial by striking with a hammer, last refining (test- ing, refining). 6. Treatment of slags for prill. All the slags except those of the fusion for regulus have been preserved. The fusion No. 6 gives a small sup- plementary button of copper, which again undergoes, if necessary, one or two washings. As we have said, the roasting is used only for pyrites. We shall return later on to the duration and the circum- stances of this operation. Its evident aim is to drive off the excess of sulphur, so as to cause the whole of the copper, with a part only of the iron which abounds in the pyrites, to pass into the state of sulphide at the time of the fusion for regulus. 438 ENGLISH COFFEE ASSAY. I. Eegulus. 1. Pyrites. The fusion for regulus of a properly roasted pyrites is made by mixing with it equal volumes of the three fluxes borax, fluor-spar in powder, lime slaked in powder of each one ladle, and covering the mixture with a layer of moist common salt. The matters composing the gangue of the roasted mineral consist prin- cipally of quartz, silica, and in general of more alumina and magnesia than lime ; oxide of iron, resulting from the roasting of the pyrites, is also present. The borax only serves to give fusibility, the fluor-spar contributes to the same end by forming a fluosilicate. Otherwise it does not play an important part in the de- composition that is to say, there is probably no produc- tion of fluoride of silicon and calcium ; for this last base is added here in considerable proportion, so as to form im- mediately a silicate which may combine with the fluoride of calcium. The ferric oxide being so reduced as to pass into the slag, and the different metallic oxides to pass into the re- gulus, yield oxygen, which reacts on the remaining sulphur. The disengagement of sulphurous acid which results from this, joined to the water contained in the fluxes, justifies to a certain extent the use of a layer of common salt, designed to prevent the boiling over. Besides this, the common salt, being without action on the metallic sul- phides, does not here produce those important effects which it exerts in the later fusions. If the pyrites appear insufficiently roasted, we must add a little nitre, the oxidising action of which again gives off sulphur ; the opposite case, that of a roasting too much prolonged, is rare ; we remedy it by the addition of sulphur and tartar. 2. Very poor Pyrites. In a very poor pyrites that of Bear Haven, in Ireland, for example the proportion of sulphur does not require us to have recourse to the roast- ing ; we employ the three fluxes and one ladle of nitre. ENGLISH COPPER ASSAY. 439 3. Variegated Copper Ore. Peacock ore contains less sulphur in proportion to the copper than pyrites ; we also fuse with a little nitre. 4. Sulphide of Copper. The sulphur is here insuffi- cient. We add together sulphur -J to 1 ladle, according to the valuation ; tartar J to ^ ladle that is to say, half the volume of the sulphur. The tartar is a powerful re- ducing agent, and is supposed in small quantities to favour the action of the sulphur by preventing its disengage- ment as sulphurous acid by the oxidising matters in the mineral ; but if used in excess, it acts as a desulphuriser, as well by its carbon as by its alkali. 5. Carbonated Minerals. The addition of sulphur and carbon is evidently still more necessary here. 6. Native Mixture : f sulphide of copper, ^ pyrites. We add, in this case, nitre for the pyrites, and sulphur and tartar for the sulphide of copper ; although these reagents appear sure to neutralise each other, it is possible that their simultaneous employment may be logical. The nitre pro- bably decomposes the pyrites, which would without it fuse and give a very ferrous regulus, whilst the free sulphur would be of little use, on account of the sulphide of copper. Be this as it may, this is the plan adopted. During the progress of the fusion for regulus we have still to introduce other matters, some incidentally, and others in all cases. If a blue flame persists in escaping from the crucible, an index of the formation of sulphurous acid, we project into it sulphur 1 ladle, tartar ^ a ladle. When the fusion appears almost finished, in order to render the bath more liquid, and to facilitate the collection of the button, we throw in a little dried salt, and a flux composed before- hand of lime, a little fluor-spar, and a very little borax that is to say, of the elements in different proportions of the mixture introduced originally. The regulus obtained is composed principally of copper, iron, and sulphur. We shall return to the aspect and the richness which it ought to have according to the minerals treated. 440 ENGLISH COPPER ASSAY. II. Calcining. The calcination of the regulus is one of the most im- portant operations ; it ought to be quite complete. III. Coarse Copper. To the calcined regulus is added nitre J ladle, borax 1 ladle, charcoal -J ladle, dry salt 1 ladle (these quantities remain the same, whatever mineral may be assayed) ; tartar 2 ladles for a regulus of medium richness. Covering of moist salt, 2 ladles. The nitre is designed to burn the sulphur which may have escaped the calcining, and to insure the passage of the easily oxidisable metals, especially of iron, into the slag in the state of oxides. It is, besides, in too small pro- portion to act upon the copper, especially in presence of reducers whose effect is certainly later than the deflagra- tion of the nitre. The borax plays simply the part of a flux. The dry salt has for its object to give fluidity to the slag. Unfortunately, if the addition of the salt attains this object, it also determines from this operation a sensible loss of copper by carrying it away with the saline vapours. We shall insist upon this point in describing the washing. The charcoal and the tartar are especially the important reagents in the fusion. The tartar, at the same time that it is one of the most energetic reducers, is also a flux and a desulphuriser. Its use is, then, perfectly justified here, only the proportion of tartar added ought to be regulated according to the quantity of copper, which the weight and aspect of the regulus permit the experienced assayer to estimate sufficiently closely ; an excess of tartar would re- duce the foreign metals, and produce in consequence a very impure coarse copper. When the fusion appears complete, we throw in some white flux,* which gives fluidity to the slag, and deter- * This white flux is prepared in the laboratory by mixing in a mortar, tartar 3 volumes, nitre 2 volumes, salt a little, then determining the combus- tion by the introduction of a red-hot iron rod, which is turned round until the matter ceases to deflagrate. ENGLISH COPPER ASSAY. 441 mines by its partial decomposition, from which a disen- gagement of carbonic oxide results, a stirring up of the materials. These two effects facilitate the collection of the metallic button. The potassium carbonate begins also, without doubt, from this operation to refine the metal a little by attacking the iron, zinc, and tin already reduced. M. Berthier (' Essai par la Yoie Seche,' vol. i. p. 393) points out this reaction : ' A part of the carbonic acid which it contains being decomposed and changed into car- bonic oxide, a compound is formed consisting of alkali, carbonic acid, and metallic oxide, &c. Lead, copper, and antimony are not attacked. IV. Washings. In the operation of washing we put into the crucible, at the same time as the coarse copper, the following fluxes : White flux, 1 ladle ; dry salt, 2 ladles. It is evident that the white flux is here employed as an oxidiser of the foreign metals, and with a view to the application of the above-mentioned reaction. As for the salt, it is both useful and injurious. If it were only used with the view of augmenting the fluid mass so as to preserve the metal from contact of air, &c., it would be advantageously replaced by an excess of white flux ; but it can form volatile chlorides with the arsenic and antimony which the copper has retained in the form of arsenide and antimonide. Common salt is, then, to be regarded as one of the principal agents of purification put in operation by the Cornish method. On the other hand, the loss of copper which arises from the carrying off of this metal by the vapours of common salt cannot be doubted. M. Ber- thier has found that by heating equal weights of copper and salt until the complete volatilisation of the latter, 3 per cent, of the metal is carried off. In the event of the coarse copper appearing too im- pure, we take care to add a little nitre. According to the appearance of the button we recommend the washing or not. 442 ENGLISH COPPER ASSAY. Y. Testing, Refining. The button of metal is flattened on an anvil. We thus recognise tin by the hardness, and antimony by the brittleness, of the alloy. The button is then put alone in the crucible. When it presents a proper appearance that is, when the edges assume a bright colour, the centre, which the assay er calls the eye, being dark ^we hasten to put into the crucible the fluxes, which are the same as for washing, only taken in rather smaller quantity. In general, when we have operated well, the button obtained is of a fine colour, and is regarded as pure ; if we have passed the eye, it is covered with a layer of red oxide ; if, on the contrary, we have put in the fluxes too soon, the button is dull. It is easy to give an account of the reactions which take place during the refining, and which differ a little from those of the washing. In heating the button alone in the air in the crucible, it is intended to submit it to an oxidation which ought to act sufficiently on all the foreign metals more oxidisable than copper without acting too much on the latter. The proper point is indicated by the appearance of the eye ; the projection of the fluxes puts an end to the atmos- pheric oxidation, and causes the scorification of the oxides which expel part of the carbonic acid of the potassium carbonate, for which they substitute themselves, and give rise to triple compounds of metallic oxides, alkali, and carbonic acid. The lead, tin, iron, and zinc oxides behave thus. When we have passed the eye, there has been a con- siderable formation of oxide, which leaves the button reddened, as I have indicated. At the same time the slag is strongly coloured red or green. If, on the contrary,, the fluxes have been thrown in too soon, the oxidation has been insufficient, and then the refiner just falls back upon the preceding operation of washing an operation less efficacious, and even without result, in the case of lead and antimony. ENGLISH COPPER ASSAY. 443 As for the physical phenomenon of the eye, perhaps it corresponds to the very short instant when the oxides, less dense than the copper, are concentrated at the top of the button, .and there make a dark spot before attaining a temperature sufficiently elevated to acquire the brightness of the metal itself. We may add that the minerals of Cornwall, generally more impure than foreign minerals, require a notably longer time for the appearance of the eye. Extra Accidental Washing. More often the refining gives a definite product, put aside to be weighed with the prill extracted from the slag ; let the button be clear, burnt, or dull. Even if the metal appeared too impure we would not recommence the refining, but would have recourse to an extra washing by putting at once into the usual crucible, besides the button and the usual fluxes, the slag from the refining. VI. Slags for Prill. All the slag from the fusion for coarse copper in- clusively having been preserved, we fuse them altogether with : Tartar . . .1 ladle "1 Simple reducing Charcoal . . . traces / mixture. We obtain a small globule, variable with the circum- stances of the different operations which have allowed more or less copper to pass into the slag. If the prill is not very small, and its appearance indicates a metal not sufficiently pure, we submit it to one or two washings as above. SECTION II. MANIPULATIONS. The sample, which has been taken with the utmost care, arrives at the laboratory rather coarsely powdered, still wet, and wrapped in strong packing-paper ; the paper is opened and placed near a furnace on the cast-iron plate which covers it ; the drying is rapidly done there. The first requisite is to discover the kind or kinds of minerals, so as to employ the warm or raw sample. 444 ENGLISH COPPEE ASSAY. For this purpose we throw one or two large pinches of the mineral into a flat-bottomed copper dish, and we wash it very easily by putting in water several times and giving a rotatory motion to the matters at the same time that we incline the dish so as to cause the muddy parts to run from the gangue. The small metallic fragments remain distinctly visible, and we can often discern by simple in- spection the presence of foreign metals. We weigh 400 grs. of the dried mineral a quantity upon which the assay is made. The crucibles used in Cornwall are of three sizes : 1. Large. 2. Large second. 3. Small second. The small seconds have externally the internal dimen- sions of the large, into which they fit as into a nest ; the first and third are sold the one in the other, and called nested. They are the most used. The large serve for the roasting and the fusion for regulus: the small seconds for calcining the regulus and all the fusions which follow. The large seconds are only employed in place of the former when we have to treat a very large regulus. The crucibles are of a kind rather wrinkled, and as if fused superficially ; they present the appearance of coarse stoneware pottery. Their form, moderately wide, permits us to make use of them successively for the roasting and the fusion for regulus, and gives them sufficiently great stability in the fire of a wind-furnace. They are, besides, very resisting. They are made at Truro and Eedruth. The wind-furnace has for its principal dimensions Inches Length from front to flue . . . 10 Breadth Depth to the bars . Opening of the flue 8 14 8 2 A sufficiently large space is reserved underneath the fire, where the ashes accumulate without inconvenience, but opening only by a contracted framework so as not to allow too free an access of cold air. ENGLISH COFFEE ASSAY. 445 The furnace serves either for roas tings or for fusions ; in the latter case we cover it with two mounted bricks, very easy to manage, and allowing us to only half-open it when we wish to inspect the contents of the crucibles. We can conduct ten roastings at once ; the crucibles are marked by a brush with colcothar mixed with water. The furnace having been recharged with coke, we put the crucibles on the top, and after a few minutes, the substances beginning to get warm, we stir them by means of iron rods. Each crucible receives a rod, which we leave standing there (leaning against the chimney) during the whole period of the roasting, so as to avoid the loss which would take place if we withdrew the rod. From time to time we renew the surfaces by lightly taking hold of the rod with the left hand by the upper end, whilst the right forefinger and thumb make it turn at once upon itself and round the crucible. The duration of the roasting varies essentially with the nature and the richness of the mineral ; it is never less than six or seven minutes, and may reach half an hour. When from the sandy appearance of the matters we con- sider the operation finished, we withdraw the crucible, raise the iron rod with care, and expose the crucible to the air, allowing its contents to cool slowly. The roasting has succeeded when the surface has the brown-red colour of iron oxide and the bottom only is black. In this case we proceed to the fusion for regulus by simply adding the three fluxes (borax, fluor-spar, and Jime) ; if the bottom of the crucible appear too black, we ought to complete the oxidising action by the addition of a little nitre. Fusion for Regulus. The different substances above indicated are taken from the box with a slightly concave ladle of If diameter, then mixed in the crucible with a stirring-knife. We ought to allow the heat of the wind-furnace to fall and then to recharge, so as to have a gentle fire at the com- mencement of the fusion for regulus. The crucibles 446 ENGLISH COPPER ASSAY. placed upon the coke, and supported against the walls of the furnace, which we then close with the two bricks. After about a quarter of an hour, we open the front brick, so as to observe the progress of the operation ; it is at this stage that we throw the sulphur and tartar into those crucibles from which a blue flame is disengaged. Some minutes later that is to say, nearly seventeen minutes from the commencement we add the salt and the flux destined to collect the regulus ; then (twenty minutes from the beginning) we run into a metal mould, not greased. We make, in general, several fusions at once four, for example ; we have in consequence two moulds into which we pour the contents of the crucibles in an adopted order, so as to avoid all confusion. The matters, very rapidly solidified, are detached simply by a blow, and fall in order on a metal plate fixed in front of the laboratory window. We immediately seize them with the copper tongs, put them into a basin of the same metal, and im- merse them for a moment in cold water, where it is important not to leave them too long. This immersion allows us then to separate very easily the slag from the button of regulus, itself very brittle. For this purpose the mass is put on the metal plate, and by means of a hammer we strike with care all round the slag, which breaks off pretty cleanly. We hasten to detach from the surface of the regulus the slag which may remain adherent, using a small hand-chisel, without the hammer. The slags are broken, and if we find any prills of regulus they are added to the principal button. Sometimes in these breakings, and especially in those analogous for the last fusions, we surround the substances by an iron ring, placed on the metal plate, so as to avoid loss of splinters. In a general way, the slags of the fusion for regulus are rejected. We shall see further on how it may become necessary to flux them again when the mineral contains blende. The aspect of the regulus is characteristic, and it is easy to arrive at a pretty close estimation of its richness, and consequently of the degree of success of the operation, by simple inspection of the regulus. ENGLISH COPPER ASSAY. 447 No. 1 . A very poor regulus (coarse) ; that is to say, too much charged with iron, is bronzed and dull; the opera- tion following would not be able to carry off the excess of iron, at least without a corresponding loss of copper. A like regulus evidently results from an imperfect warming, or from an excess of sulphur, or from an insufficiency of nitre, as the case may be. It contains less than 40 per cent, of copper. There is nothing for it but to reject it. No. 2. A regulus of good appearance is in general bronzed but rather shining ; it appears finer. Its richness varies from 40 to 60 per cent. No. 3. From oxides, carbonates, and from some mine- rals charged with impurities (Sn, Sb) we desire to obtain a fine bluish button of a greater richness 65 to 75 per cent. We perceive, indeed, that from oxides and carbon- ates, to which we have only to add sulphur, and which also by their nature do not, like pyrites, contain combined iron, it is easy to obtain a richer regulus without fearing any loss of copper. As for the stanniferous and antimonial minerals, we shall return to them further on. No. 4. In every case a regulus, the richness of which rises to 80 per cent., and of a very shining grey-blue appearance, ought to be rejected, its richness indicat- ing the loss of a certain quantity of copper left in the slag. Here is, in the preceding order, the result of the ana- lyses of four buttons whose description agrees with that which we have just given, excepting, perhaps, No. 2, whose fracture is rather reddish : No. Copper Iron Balance ; sulphur and traces of foreign metals 1 2 3 4 Coarse, to be rejected . Good in general (rather too fine) . Good for a carbonate, &c. . Too fine, to be rejected 36-00 60-00 65-60 80-16 32-90 14-70 10-50 2-10 31-10 25-30 13-90 17-74 If we compare these products with those obtained in 448 ENGLISH COPPER ASSAY. tlie metallurgy of copper by the Welsh method, we find (Le Play, ' Annales des Mines ') : Matts of the operations H. V. IV. VIII. h Si a a d 2 (-H Different metals 1 a "9 m co 3 ( Coarse matt (fusion of poor mine- j II. \ rals, raw or calcined) 34-6 34-1 1-5 29-8 31-3 UCu 2 S + Fe 2 S 3 + 4(Fedif.met.)S j I/ Blue matt (fusion of the cal- \ cined coarse matt with I minerals of mean richness) [ 57-2 18-5 1-0 23-3 24-3 1 0-8Cu + 3Cu 2 S + 2(Fe. d.m.)S J 1 Beddish variety, matte mince } \ l-3Cu + 3Cu 2 S + 2(Fe. d.m.) S ) 61-6 15-8 0-6 22-0 22-6 /White matt (fusion of the calcined I course matt with rich minerals, carbonates, and oxides) IV J Metal very pure type very blue variety . 77.4 64-8 0-7 9-0 0-9 3-6 21-0 22-6 21-9 26-2 mean . 73-2 6-3 20-5 8Cu 2 S + FeS (Matt (roasting of extra white] VIII A matt VII.) 81-1 0-2 18-5 ( O2Cu + Cu 2 S J These numbers show the evident analogy, the identity almost, of the products of the laboratory and those of the works ; we may sum up by saying that the regulus ought to be richer than coarse metal, and in the case of ordinary minerals to approach if not to attain (as in the case of sample No. 2) to the composition of blue metal. For carbonated and oxidised minerals we arrive directly at the very bluish variety of white metal. Finally, in no case must we have a button as rich as regulus matt. Calcining the Matt. The matt is pounded fine in a bronze mortar ; we avoid loss of fragments by means of a perforated cover and a cloth which surrounds the pestle. To facilitate the pulver- isation, and avoid the sulphide greasing, we add in the mortar a small piece of coke. The pounded matt is care- fully turned upon a sheet of paper, the mortar wiped out with a hare's foot, and the substance put into a small second or large-second crucible. The calcining is con- ENGLISH COPPER ASSAY. 449 ducted as the roasting of a mineral ; it generally lasts longer, for the expulsion of the sulphur is to be as com- plete as possible. It is necessary to regulate the fire with the greatest care, so as to avoid all agglomeration, and to stir almost continually. When the matter adheres to the rod, we withdraw the crucible for a moment ; this incon- venience is chiefly produced if we have not detached the slags sufficiently from the matte ; the calcining is then much longer, the flames remain blue a long time, and the fumes which are disengaged have an odour which is -not purely that of sulphurous acid. When the fumes and the odour cease, and the matter has taken a sandy appearance, we raise the heat ; then withdraw the crucible, and allow it to cool slowly in the air as when roasting. The mean duration of calcining is half an hour. Coarse Copper. The fluxes above indicated are taken from a box No. 1, except the dry salt called the refining flux, which forms part of a second box. The ladle for this box No. 2 is a little larger than for the first : it has a diameter of 1^ inch. At the beginning of the operation the furnace is well filled and lighted ; the same fire ought to suffice for all the following fusions, which it is very important to conduct with great rapidity. After a moment, and if there is any frothing, we throw in some dry salt, which calms the ebullition. At the end of ten minutes, the fusion ap- pearing complete, we throw in a pinch of white flux. A little after we withdraw successively each of the crucibles, pouring them carefully and by a single turn into each of the principal cavities of the metal mould. These moulds ought, this time only, to be greased with a cloth impreg- nated with suet. The crucibles are immediately put back again into the fire. We detach the slag as previously, seize' each one successively with the copper tongs, and plunge it for an instant into a basin full of water. The rest is effected as for the regulus, only the slags are preserved on the metal G G 450 ENGLISH COPPER ASSAY. plate, and in the order in which we have detached them. The button of copper obtained appears more or less black ; T have already indicated the influence of the tartar in excess. Washings. We place the button and the fluxes in a large copper shovel, lengthened and narrowed at the end, called a scoop, and we pour them into the crucible, which is already at a red heat. As the fusion is made in five or six minutes, it would be inconvenient to prolong it on account of the loss occasioned by the carrying off of copper with the vapours of common salt. The pouring is made with care first into one of the large cavities, then as soon as the metal has fallen there we finish by casting the slag into one of the small lateral cavities. This last slag, probably rich in copper, is less fluid, and would adhere to the button, which would be difficult to cleanse. The two buttons being detached from the mould, we immerse the small one first, then finish as in the preceding operation. Testing and Refining. The crucible has again been put back into the furnace, after the pouring ; the button tried by the hammer is put into the crucible by means of the tongs. At the end of about three or four minutes it attains the colour of the vessel, the eye manifests itself, and we rapidly throw in the fluxes put into the scoop beforehand. The pouring is done as for the washing, with the small button of slag kept apart. In general we get a button regarded as pure, clean copper ; if not, as I have said, we proceed to an extra washing by adding exceptionally in the scoop the last slag obtained. Prill. The crucible this time has been left out of the furnace ; put into it all the slags, collected for this purpose from the metal plate into the scoop, and upon which we have ENGLISH COPPER ASSAY. 451 put reducing reagents. The fusion lasts a quarter of an hour ; pour all at once into the large cavity : before the cooling, by means of a transverse blow, get rid of the upper beds which are still liquid, and composed principally of common salt. Then operate as above. Collect the prill, which again undergoes, if necessary, a washing. SECT. III. SOME MINERALS AND SUBSTANCES OF A SPECIAL NATURE INFLUENCE OF FOREIGN METALS. STANNIFEROUS MINERALS. Most often we only perceive the presence of tin in a copper mineral when testing with the hammer, which reveals the nature of the bronze ; when we proceed to the refining of such a stanniferous button, it is impossible to obtain the characteristic eye ; that is to say, the surface of the metal becomes quite clear, and we scarcely open the furnace when it again becomes obscure. We free it from tin by two or three extra wash- ings. If we suspect tin from the known produce of the mineral, or the inspection of the sample in the basin, we endeavour to obtain a fine regulus, which is accomplished in the case of a warm sample by prolonging the calcining, and for the raw sample by putting in more nitre or less sulphur. It is clear that tin can only enter the regulus by virtue of the excess of sulphur necessary to the formation of the coppery matte, and that by restraining this excess of sulphur we diminish the chance of tin entering the button. The fine regulus ought to contain 70 to 75 per cent, of copper, as for the carbonated copper minerals. ANTIMONIAL MINERALS. Antimony is also detected in the testing, the metal being rendered very brittle, We then add one or two grammes (15 to 30 grains) of lead in the refining operation. There forms an alloy of lead and an- timony heavier than copper, which is poured into the small cavity of the mould. When we suspect antimony we act as for tin that is to say, we produce a fine regulus, a most careful roasting expelling the antimony ; then we have to make two washings, and in the second to add the metallic lead. G G 2 452 ENGLISH COPPER ASSAY. We cause, then, three influences to act with a view of expelling the antimony : 1. Slow oxidation at a low temperature, disengaging antimony. 2. Eepeated chloridations, whence a formation of vola- tile chlorides. 3. Affinity of the lead and mechanical separation of the alloy. ZINCIFEROUS MINERALS. One of the metals which is most troublesome is zinc. We recognise it by the appear- ance of the regulus and by its colour, which is that of blende. Perhaps once out of ten the regulus collects suffi- ciently to be able to detach it ; in this case we pound it, add to it the slags, and borax 1 ladle, nitre ^ ladle. We fuse anew, and obtain a good regulus, for the nitre has caused the zinc to pass into the slag in the state of oxide. Most often the zinciferous regulus does not collect, and there is nothing for it but to begin anew by making a very prolonged roasting of at least half an hour. PLUMBIFEROUS MINERALS. Lead is not injurious, for it does not alloy with copper. The warming is also pro- longed. Lead passes into the regulus, which facilitates the collection of the matter. In the last operation the lead easily passes into the slag ; it also, in case of need, carries off antimony. Thus the copper obtained from lead minerals is most malleable. Special Cupriferous Products. EEGULUS OF CHILI. These are treated as those which we obtain by the fusion for regulus. Their richness, which rises to nearly 60 per cent., requires us to add much tartar in the fusion for coarse copper. SLAGS OF COPPER. To obtain regulus w^e add to the slag sulphur, tartar, and nitre, this last maintaining metals other than copper in the state of oxide in the slag. OLD COPPER. For turnings, waste of workshops, &c., yielding 97 to 98 per cent, by the assay, and containing, in fact, not much foreign matter except a little mixed dust ENGLISH COPPER ASSAY. 453 or dirt, we take care first to glaze the crucible by fusing in it a little borax and nitre ; then we treat the matters by a simple washing, the slags of which we work for prill. This last is often very considerable. SECT. IV. SUMMARY CONSIDERATIONS COMPARISON OF THE EESULTS WITH THE ANALYSIS BY THE WET WAY. After this detailed account of the numerous operations which the metal undergoes before attaining the state of button and prill, it would be superfluous to insist upon the practical difficulty of the Cornish method. Nevertheless, in experienced hands, and in the case of daily practice, it is still a rapid method, allowing us to treat almost uniformly the different varieties of copper mineral, and at the least to remedy during the operation itself the impurities which show themselves. As to the metallurgic accuracy, below is a small table showing comparatively the produce by the dry way (es- timated by a Cornish assayer) and that which we have obtained by the most precise methods of the wet way. It comprehends six samples, whose richness varies within sufficiently great limits. Nature of the sample Dry way Wet way Difference and produce D. W W-D Regulus of Chili ..... 564 = 56-250 58-40 2-150 Green copper carbonate of Castile . 9f = 9-750 11-52 1-770 Variegated copper, Huel Damsel . Pyrites, West Wheal Seton . . 10i = 10-500 8|= 8-375 11-30 8-40 0-800 0-025 United Mines .... 8 = 8-000 10-38 2-380 Devon Great Consols 4|= 4-625 5-60 0-975 S(W-D) . . 8-100 6 1-350^ By adding the result given by the last five minerals we find = 9-44 b By taking the ratio 454 ENGLISH COPPER ASSAY. we see that we must add to the richness indicated by the Cornish assay about f of that result, and by taking the ratio that the loss is -J of the copper, if we consider a mineral of 9 or 10 per cent. Without wishing to draw a conclusion altogether general from so small a number of analyses, we neverthe- less think they suffice to show that the Cornish method occasions losses always sensible and sometimes considerable. We think we may assert that upon the whole of the Cornish minerals whose mean richness varies from 6 to 7 per cent., the loss by the assay is not less than 20 per cent, of the contained copper, and that for certain pyrites of 3 to 4 per cent, it attains 30 and 40 per cent, of the metal. The principal causes of these losses are (1st) the quantity more or less great of copper left in the slag of the regulus ; (2nd) and especially the carrying away of copper by the vapours of common salt in the fusion for coarse copper, the washing or washings, the refining and the treatment of the slag for prill. In consequence we think they ought to bear princi- pally on the oxidised minerals for which we make a rich regulus, and still more on the impure minerals, which besides a rich regulus have undergone several washings. Thus the minerals of Algeria, grey copper, assayed some years ago at the School of Mines, have given a produce much higher than that indicated by the Cornish assayers. It may be said, therefore, that the results of the Cornish assay do not fall short of the truth by a fixed quantity,, but become more and more inaccurate the poorer the ore. The smelter gets out of the ore more metal than the assay indicates. The most frequent alloys of copper, i.e. brass, German silver, gun-metal, &c., cannot be assayed in a reliable manner in the dry way ; German silver because the ENGLISH COPPER ASSAY. 455 nickel could not be removed at all, or only with great difficulty, and the rest because zinc and tin give such diffi- cultly fusible oxides that they could not be properly removed in the refining. In the alloys of copper with silver -, gold, and platinum, the copper may be estimated from, the loss arising from cupellation with lead. In all assays of copper in the dry way the silver or auriferous silver contained in the assay sample cannot be removed, and it is generally pretty completely collected in the copper obtained. These copper assays give no- where any indications whether gold or silver is present or not ; and the amount of these metals which may be present must therefore be both sought for and estimated by a special assay for them. If they are found, and in sufficiently large quantity, they are deducted from the weight of the copper. The dry assay is mostly found in practice in smelting works, where, even in the hands of less scientifically edu- cated than skilful assayers, with the character of the assay substance once known, and suitable practice in following out the separate manipulations, it gives results which suffice for the business of working copper in the large way. b. For Ores and Products of Class II. THE LAKE SUPERIOR FIRE ASSAY. We are indebted to Dr. E. D. Peters's work on ' Modern American Methods of Copper Smelting ' for the following excellent description of the dry method of assay in use in America.* The ordinary English fire-assay is little suited to American conditions. The Lake Superior fire-assay, on the contrary, is not only quick and inexpen- sive, but compares favourably in accuracy with the best * In his description of this mode of assay Dr. Peters acknowledges his indebtedness to Mr. Maurice B. Patch of Houghton, Michigan, for valuable assistance in the preparation of this section on the Lake Superior assay. The position held by Mr. Patch as chemist to the Detroit and Lake Superior Copper Company is a sufficient guarantee of the accuracy of the following description. 456 AMERICAN COPPER ASSAY. wet methods. It is so peculiarly adapted to the conditions that have given it birth, that no American work on the metallurgy of copper would be complete without a detailed account of it, especially as our docimastic literature up to this time has made little mention of it. In the Swansea assay, the substance under treatment consists usually of a mixture of sulphides and gangue rock, which necessitates a series of calcinations and fusions, culminating in a button of impure copper, which has still to be refined at a con- siderable loss. The Lake Superior assay er has the simpler problem of dealing only with native or oxidised compounds of copper that can be reduced to the metallic state at so low a temperature as to preclude the adulteration of the copper button with any other metallic substances, and thus obviate the necessity of any refining process. In spite of the apparent simplicity of this method, it demands a good deal of skill and experience to obtain correct re- sults ; but these once acquired, no assay can excel it in accuracy and celerity. A glance at the composition of the substances operated on will render clear the objects to be accomplished. The material assayed consists of the concentrates from the jigs, tables, buddies, and other concentrating machines. This material is technically termed ' mineral,' and varies greatly in richness, composition, and size of particles, ranging in copper from 10 to 97 percent., and in some instances con- taining a gangue of nearly pure ferric oxide, while in others it is highly silicious. Nearly all grades of mineral contain a considerable proportion (from 3 to 10 per cent.) of metallic iron from the stamp heads, while a sample containing 50 per cent, of titanic iron-sand is no unusual occurrence. It can readily be seen that no small skill is required so to flux these various mixtures as to obtain a clean and fusible slag, and a button of copper free from iron or other metals that may be reduced with compara- tive ease, and thus yield a far too high result. Sampling. The mineral is received from the mines, packed in strong barrels, weighing in the damp condition in which it arrives from 500 to 2,000 pounds, its weight AMERICAN COPPER ASSAY. 457 depending on its degree of concentration, the character of its accompanying gangue, &c. As this material is to be refined at once, the barrels are emptied on the iron plates that form the floor in the neighbourhood of the charging- door of the refining-furnace. After the contents of each barrel have been thoroughly and separately mixed, a small sample is taken from every package and put into a tightly covered copper can. Only the samples from casks of the same grade of mineral are placed in any one can, as each quality is assayed by itself, although six or more different grades of mineral may go to make up the sixteen barrels that usually form a furnace charge. If two or more furnaces are simultaneously in operation, the samples of the same grade are mixed together, to avoid the unnecessary multi- plication of assays ; the cans containing a little water in the bottom, into which the tight copper cans are set, to prevent any loss of moisture in the sample, which might occur despite the close cover. Fluxes. Sodium bicarbonate, borax, potassium bitar- trate (cream of tartar), ferric oxides, sand, and slag from the same operation are used to flux the gangue and other worthless constituents, and effect the proper reduction of the copper. The chief impurity to be dreaded is sulphur, for which reason the best quality of sodium bicarbonate must be purchased, and potassium bitartrate must be used instead of argols. The borax and soda are prepared by being melted in iron ladles, to drive off their water of crystallisation, and then pulverised through a twenty- mesh screen ; clean, well-fused slag from former operations is reduced to the same degree of fineness, while. the oxide of iron flux is prepared by pulverising selected fragments of specular iron through a fifty-mesh sieve. Any clear quartz sand answers for the silica needed. Furnace. A common natural draught melting-furnace is used, an inside measurement of 9-J by 18 inches being large enough to accommodate six Hessian crucibles. These are set in rows of three on two thin fire-bricks, the latter resting on the longitudinal grate-bars, and serving to raise the crucibles to the zone of greatest heat. Soft coal, broken 458 AMERICAN COPPER ASSAY. to egg size, forms the customary fuel, and is carefully filled in around the charged crucibles, which are not placed in the furnace until the latter is in full heat. The crucibles employed are four inches high and three inches in diameter, and are provided with well-fitting covers made at the works from a mixture of fireclay and sand ; these are the more necessary because the assay often fills the crucible to within half an inch of the top. The skill of the assayer is nowhere more evident than in the fluxing of the different grades of mineral, the com- position of which was briefly noticed in the opening para- graph of this chapter. It is, of course, familiar to all chemists that sodium bicarbonate and ferric oxide act as powerful bases, while the electro-negative elements are represented by borax and sand : the potassium bitartrate exercises a strong reducing action, as well as furnishing an active base. The slag equalises the entire mixture, being capable of neutralising a considerable amount of either base or acid, and it covers the molten metal and protects it from oxidation. It is not to his skill in fluxing alone that the assayer trusts ; of almost equal importance are the degree of temperature maintained and the length of time that the assays are left in the furnace. Good results can only be obtained by shortening the period of fusion to the utmost. This demands a very hot furnace at the outset, good fuel, and a lively draught. Under these conditions an easily fusible assay will pro- bably be entirely finished in 20 minutes, while from 25 to 30 minutes are required for different samples. It is quite safe to assert that if the time necessary for a perfect fusion is increased to 40 minutes, the resulting button will contain sufficient impurities reduced from the slag to give a result from 2-|- to 6 per cent, too high. This assay is applicable to silicates as well as oxides and native copper, and the results obtained from the assay of both refining and blast furnace slags cannot be excelled in accuracy by any other method. A table of the different weights of fluxes used in assay- ing the various grades of mineral from the Peninsular AMERICAN COFFEE ASSAY. 459 Copper Company's works is annexed, as well as the mix- ture adopted for reverberatory slags for very silicious ores. Minerals Weight, grains Borax, grains Soda, grains Slag, grains Potassium bitartrate, grains Sand, grains Iron ore, grains Xo. Per cent. copper 1 1 92 1,000 60 55 200 300 2 86 1,000 60 60 180 300 3 60 500 100 80 300 4 83 500 150 160 300 150 5 20 500 190 200 300 175 # 35 500 140 140 300 100 t 5 to 20 500 200 200 300 The percentage of slag-forming materials being so small in Nos. 1 and 2, it requires but a slight amount of borax and soda to flux them, while an addition of neutral slag is necessary to protect the molten copper. A smaller quantity of the ore is weighed out in the succeeding assays, as they are so poor in copper that a large amount of flux is required by the great quantity of gangue, so that the capacity of the ordinary crucibles would be greatly ex- ceeded if 1,000 grains were used. No. 3 mineral contains just sufficient ferric oxide to form a good slag with the mixture given ; while in Nos. 4 and 5 this substance, as well as metallic iron, increases to such an extent as to require the addition of a considerable proportion of sand to flux this base and to prevent the adulteration of the button with metallic iron. The sample of Calumet and Hecla tail-house mineral given is typical of the treatment of very silicious material. There is nothing remarkable in the considerable proportion of borax (an acid flux) used with even highly quartzose ores ; for, in addition to the fluxing powers of the soda that it contains, a boro-silicate is very much more fusible than a simple silicate. No peculiarities exist in the execution of this assay ; the ore and fluxes are thoroughly mixed on glazed paper and covered with a thin layer of potassium bitartrate after being poured into the crucible. In the * Calumet and Hecla tail-house mineral. t Rich slag from refining. 460 AMERICAN COPPER ASSAY. No. 1 mineral, which is nearly as coarse as split peas, fragments of iron frequently exist, which come from the stamp heads, and must be picked out of the sample after weighing out for assay : not that cast iron will alloy with copper, but that the fragments will be found imbedded in the copper button after cooling. The results obtained by this method are surprisingly accurate. Duplicate estimations of the lower grade samples seldom vary more than 0*1 or 0'2. A difference of 0*4 per cent, is a rare occurrence even in the higher classes of mineral, where the size of the metallic fragments renders the sampling, and even the weighing out, of a correct assay a matter of some uncertainty. A few results from Mr. Patch's notes will confirm these statements. An average series of tests on cupola slags by the colorimetric method for the period of a month, duplicated by the fire assay, gave a result 0-05 per cent, lower for the latter test, the slag containing about 0*5 of one per cent. As an illustration of the results of this system when applied to very rich ore, a comparative test was made for eight days on No. 1 Calumet and Hecla mineral, with the following results : Battery assay .... 89-100 per cent. Fire assay . . . 88-812 A similar test on No. 2 Calumet and Hecla mineral : - Battery assay .... 77*590 per cent. Fire assay . 77'657 A similar test with various samples : No. Battery assay Fir r89-501 >aa.ftn mean = 89-544 ^ . >mean = 89*92 oy-/u L 89-70 I f 77-40] > mean - 77-740 J ^'.^ >mean = 77'50 L77-40J It is a somewhat curious fact that the slight loss of about 0-25 per cent, of copper which results from the passage of a minute portion of the metal into the slag ASSAYS IX THE WET WAY. 461 is just about counterbalanced by the impurities in the copper button from the reduction of ferric oxide, the amount of which is indicated by the following analyses of copper buttons the only weighable impurity being iron : Copper, per cent. Copper, per cent. Copper, per cent. 09-83 99-76 99-51 99-84 99-80 99-87 99-52 99-46 99-79 This account of a little-known process will doubtless remove the impression sometimes held by chemists, that the Lake Superior copper assay is a clumsy and imperfect operation, and unworthy any advanced system of metal- lurgy. B. ASSAYS ix THE WET WAY. a. Colorimetric Copper Assays. These are based upon the fact that ammonia, added in excess to the solutions of salts of copper, produces a beautiful azure-blue colour, whose intensity depends upon the quantity of copper dissolved. By comparing the shades of blue colour in equally thick layers of the dis- solved ammoniacal assay substance (assay fluid) with a normal or standard ammoniacal fluid whose copper contents are known, the quantity of copper in the former can be calculated when its volume is measured. To Heine, the superintendent of the smelting works in Mansfield, belongs the merit of having first successfully employed this reaction for the estimation of small per- centages of copper ; and later it has been also extended by Jacquelain, Von Hubert, and Muller, to the estimation of larger quantities of copper. 1. HEIXE'S COLOEIMETRIC METHOD. For the estimation of the quantity of copper in bodies poor in this metal, e.g. in slags, lead matte, litharge, crude lead, and 'other plumbiferous metallurgical pro- ducts, tin, cupelled silver, &c. in short, in all substances 462 THE ASSAY OF COPPER. which contain from a trace to about 1 per cent, or a little more of copper, this method is the most advantageous to be used. After the assay sample has been reduced to as fine a state of mechanical subdivision as possible, which with slags is best attained by sifting or washing them, one centner (3-4 grammes) of it is weighed out and dissolved, or so completely decomposed by a suitable acid that in the residue, which is to be filtered and well washed, no more copper remains behind. For this purpose nitric acid or aqua regia is employed, according to the character and particular behaviour of the substance, and the nitric acid is concentrated or somewhat diluted, as may be required. The solution is either immediately, or after the copper has been first precipitated by sulphuretted hydrogen gas and again dissolved, strongly supersaturated with caustic ammonia, and the precipitate, if any, thereby produced, digested in caustic ammonia for a considerable time, with frequent stirring at a very gentle heat (30-40 G.), then filtered off and thoroughly washed. According to the quantity of copper present, and according to the degree of dilution, the solution obtained will appear more or less strongly coloured blue. The volume of the solution is measured in graduated vessels, and the intensity of the colour compared with and estimated from fluids, which have been previously prepared as standard fluids, and which for a definite volume contain a definite, accurately weighed quantity of copper, that has been dissolved in nitric acid, precipitated by caustic ammonia, and redis- solved in excess of the same. From the measured volume, and the intensity found by comparison, the quantity of copper is then found by calculation. Heine proposes standard fluids with one, two, three, and four assay loth of copper in one ounce (two loth, commercial weight) of the ammoniacal fluid. These four standard fluids are all-sufficient. If the French weights and measures are used, standard fluids are taken with -001, -002, -003, -004 gramme of copper to every twenty -five cubic centimetres of the fluid. COLORIMETRIC ASSAYS. 463 The graduated vessels (cylinders) required for the preparation of the standard fluids, as well as for the measuring of the assay fluid, can be easily prepared by the assay er himself. One quarter of an ounce of water is weighed out a number of times in succession and poured into the cylinder, and each time the height of the fluid is marked in a durable manner on the glass with a diamond, or by etching it with hydrofluoric acid vapour, &c. Also earthen or porcelain measures, that are prepared and marked for the volumes that hold one, two, three, four, &c., ounces of water, may be used. It is not practicable to replace the volumetric measure- ment by weighing, for the quality and quantity of those substances which are soluble in acids and not precipitated by ammonia, or are again dissolved by it, may vary greatly in the assay. In the formation of the normal fluids, two assay pounds of chemically pure (galvanic) copper are weighed out on a good balance, dissolved in nitric acid, the solution super- saturated with caustic ammonia, and placed in a graduated cylinder, which is divided to whole, half, and quarter ounce volumes of water, and then water enough is added to bring the fluid to the sixteen-ounce mark. The fluid then contains -f-g-=4 loth of copper per ounce. Six ounces of this four-loth solution are then taken, two ounces of water added to it, and eight ounces of fluid obtained, with ^ = 3 loth of copper to one ounce of water. The two-loth solution is formed in a similar way by diluting four ounces of the four-loth solution to eight ounces ; the one-loth, by diluting four ounces of the four-loth normal fluid to six- teen ounces. In the measuring of the assay fluid it is estimated within one-eighth of an ounce, which is suffi- ciently close. If in the dilutions a mistake is actually made of one-sixteenth of an ounce, the maximum of possi- bility, the error amounts to about two cubic centimetres, which in a whole mass of fluid of 200-500 cubic centi- metres has no influence upon the solution that can be detected with the eye. The preservation of the standard fluids, as well as the 464 THE ASSAY OF COPPER. comparison of the blue assay fluids with them, must take place in glass vessels closed with ground-glass stoppers. These vessels must have the same form and size, consist of the same colourless glass, and have an equal thickness of glass in the smooth side walls. The last condition is obtained the surest by grinding. This grinding, however, which notably increases the cost of the glasses, is not in- dispensably necessary if the vessels are carefully formed and blown in a good glass-house. An oblong form is most advantageous for the vessels. They hold about an ounce and a half of fluid, and are about two inches long, two and a half inches high, and one inch wide, with walls about one-eighth of an inch thick. The glasses are very advantageously formed from an unblemished sheet of plate glass of equal thickness throughout, by cementing the sides together and the insertion of a glass neck. The assayer has in the form of vessel indicated a triple control in the comparison of the assay fluid with the normal solution according as he looks through the fluid in three different directions. The digestion of the assay sample with acid may take place in any suitable vessel whatever a glass flask, a beaker covered with a watch-glass, &c., only no thumping and spirting of the fluid should be possible in the process. The nitric acid, &c., must be added little by little. The time required for this may vary greatly. The solution of cupelled silver* skimmings , &c., with nitric acid is finished in a short time ; on the other hand, in the examination of difficultly decomposable slags, with which concentrated nitric acid or aqua regia will always be used, the digestion often requires to be continued at a warm temperature for two to three times in twenty-four hours. The mass must be frequently stirred with a glass rod, because many slags decompose rapidly with evolution of heat, form a thick jelly, and deposit a crust on the bottom of the glass sub-, mono-, and bi-silicate slags mostly decompose readily, * With cupelled silver, after dissolving in nitric acid, the silver may be precipitated with sodium chloride, the silver chloride filtered, washed, and the solution then mixed with caustic ammonia. COLORIMETRIC ASSAYS. 465 higher silicates resist complete decomposition by aqua regia and then a preliminary solvent ignition or fusion with potassium carbonate or calcined sodium carbonate, or, better, a mixture of both, is, necessary, precisely in the manner given in the wet assay of copper. Here also it does no harm if some of the substance of the crucible remains adhering to it. The decomposition of the slags by acid is complete when in the stirring with a glass rod no more grating can be perceived. After hot water has been added to the decomposed assay, the residue is collected on a filter, well washed out, without diluting the filtrate too largely, and the copper precipitated from the solution, if necessary, with sulphu- retted hydrogen gas, especially when a notable quantity of alumina and iron is present, whose slimy precipitates from the immediate precipitation with ammonia always retain copper. This precipitation of the copper has also the advantage that, as cobalt and nickel do not precipitate with it, the colouring effects which they would produce, if present, are removed. Since the copper sulphide re- quires for its solution but a few drops of nitric acid, in the succeeding treatment of the solution with ammonia, but a small quantity of ammoniacal salt is formed, and the specific gravity of the coloured fluid varies but very little from that of water and the normal solution. With the increase of the specific gravity of the assay solution, its volume is considerably increased, and therefore it gives too large a measure in the direct precipitation with am- monia. If the precipitation with sulphuretted hydrogen gas is completed in four to six hours, the copper sulphide is filtered out, thoroughly washed with cold water contain- ing sulphuretted hydrogen, the filter dried, ignited in a porce 7 Crucible, the copper oxide formed warmed with a few dro of nitric acid or aqua regia, supersaturated with ainm< .a, filtered, and well washed, till the washings are no longer tinged bluish. A precipitation of the copper with iron wire, from a solution evaporated with sulphuric acid, and a re-solution H H 466 THE ASSAY OF COPPER. of the copper in nitric acid, consumes less time. If the copper is not precisely precipitated, errors of some 30 per cent, and more of the whole amount of copper may occur. By repeated solution of the iron precipitate and precipitation with ammonia, all the copper cannot, how- ever, be extracted. In the examination of litharge, the solution in nitric acid may be dispensed with. The copper oxide can be at once extracted from it with caustic ammonia ; however, the litharge and ammonia must then be allowed to work at least twenty-four hours on each other, with very diligent stirring, and, moreover, the litharge must be ground very fine. The ammoniacal solution obtained from the assay is now well stirred, so that it may mix with perfect unifor- mity with the last washings ; then, either the whole, or a part of it, is placed in a clean assay glass, and compared with the standard fluids in similarly formed glass vessels standing on a sheet of white paper. Should it correspond with none of them in the intensity of its colour, the whole of the fluid is diluted somewhat with water, until this is the case. Its volume is thereupon measured in the glass vessel graduated in ounces, &c., and noted. For a check, the dilution may be carried still farther till the colour of the assay corresponds to the next more faintly coloured standard fluid, and then the increased volume be measured anew. This might perhaps be still again repeated, but it becomes more and more uncertain. The calculation of the percentage of copper, from the intensity and the volume found, then presents no further difficulty. Suppose that the assay fluid agrees with the normal so- lution of four loth of copper to the ounce of water, and its quantity amounts to five ounces, then the quantity of copper in the centner of the assay substance is 5 x 4=20 loth. This fluid further diluted till it equals the normal solution with three loth of copper, must measure six and two- third ounces if the obtained value of twenty loth is to be confirmed. According to Heine's experiments, the possible error COLOEIMETEIC ASSAYS. 467 of observations in the comparisons and measurement de- scribed amounts as a maximum with the stronger normal solutions (with sixteen loth and over) to from three-quarters to one loth, with the weaker ones to scarcely half a loth of copper. In a centner of the assay substance, one loth of copper '03 per cent, can still be estimated with cer- tainty. Le Play's Method. Le Play estimated in finely pul- verised and carefully washed copper slags, the copper in one gramme of the poorest slags to within half a milli- gramme, and of the richest slags to within one milli- gramme, by using twenty-six standard fluids with various percentages of copper in cylindrical vessels. The com- parison of colours in round vessels is more uncertain than in oblong ones, since in the former the light is dissipated and shadows are produced. T. 0. Cloud's Method. According to Mr. T, 0. Cloud the copper cannot be completely extracted from the finest ground slags, even after three days' digestion with aqua regia, and subsequent treatment with sulphuric acid. He recommends fusing the slag with four parts of mixed potassium and sodium carbonates and J part potassium nitrate. The fused mass is extracted with dilute sul- phuric acid, the liquid evaporated down, and the copper estimated galvanically or colorimetrically. Endemanris Method. As a delicate test for small quan- tities of copper, Dr. Endemann (' Annalen der Chernie ') adds to the dilute solution concentrated hydrobromic acid, when a dark brownish-red or violet colour is at once produced. This reaction is so delicate that T -^Q- milligramme of copper can be detected with certainty. One drop of a solution containing this small quantity of copper is brought on a watch-glass, then one drop of hydrobromic acid is added, and the solution is then allowed to evaporate slowly by standing the glass in a warm place. When the whole has been concentrated to about one drop, this will distinctly show a rose-red colour. The colour thus produced is about three or four times as distinct as the one which is H TT 2 408 THE ASSAY OF COPPEE. obtained by the addition of potassium ferrocyanide. Of other metals which are examined in this direction, we find only iron to be apt to interfere with this reaction, and then only when it is present in considerable quantity. This reaction may also be utilised as a colorimetric test for the quantitative estimation of small quantities of copper. If a substance contains so little copper that the fluid does not equal the most faintly coloured standard fluid in intensity of colour, the assayer must endeavour to remedy the matter by evaporating till this is the case. An evapora- tion is, however, to be avoided if possible, first because of the loss of time, and also because other precipitates, carbo- nate of lime, &c., are apt to be caused by it, and because, when it has to be continued too long, so much ammonia is apt to be volatilised, that a new addition of it becomes necessary. This method of assaying soon finds the limits of its accuracy in an increasing percentage of copper in the assay sample, since with fluids rich in copper, and there- fore strongly coloured blue, the errors of observation soon amount to several loth. And to seek then to better one- self by diluting largely, yields no more accurate results^ since a small error of observation in estimating the intensity of the colour, is so much the more multiplied in the calculation of the value by the greater number of the ounces. If nickel is contained in the assay substance, the assay cannot be conducted in the way prescribed, since the nickel is extracted by the acids, and dissolves also in caustic ammonia with a blue colour. The assay may also become uncertain from the presence of much m,anganese, cobalt, or chromium, since they render the hue of the blue colour dingy. Chromium may be completely removed by a slight boiling of the ammoniacal fluid ; not so cobalt. The presence of vanadium or molybdenum does no harm. If nickel, or much cobalt and manganese, is contained in the assay substance, the solution obtained by acids and filtered, though not further diluted, must first be decom- COLOKIMETE1C ASSAYS. posed by metallic iron. What is thrown down by the iron is collected on a small filter, washed thoroughly, and then, together with the filter, treated with dilute nitric acid. When the copper is all dissolved, this solution is supersatu- rated with caustic ammonia and then managed as above. With higher percentages of copper the process of the Swedish copper assay is used for estimating the value. The precipitation of the copper may also be performed with sulphuretted hydrogen gas. Le Play removes the injurious influence of manganese, nickel, and cobalt, by allowing the green or violet-coloured ammoniacal solution to stand open to the air for some time in a moderately warmed drying furnace, whereby a few variously coloured gelatinous flocks are gradually deposited, and the fluid, after the addition of a few drops of ammonia, then becomes pure blue. According to Jacquelain and Von Hubert, nickel and cobalt are in a simple way rendered perfectly harmless by gradually adding white pulverised marble to the solution of the assay substance, until the effervescence ceases, and then warming the whole on the sand-bath, whereby all the copper is perfectly precipitated as carbonate, while nickel and cobalt remain dissolved. It is now filtered, washed, the residue dissolved in nitric acid, and the solution treated as already explained, with ammonia. By the addition of potassium carbonate to the ammoniacal fluid, and heating, all the manganese precipitates, while the copper remains dissolved in the excess of ammonia, and can be separated from the manganese precipitate by filtration. The manga- nese must have been present as oxide in the original solu- tion in order that the precipitation by potassium carbonate may be perfect. The assayer may convince himself whether nickel or cobalt is present, by slightly supersaturating a blue ammoniacal solution^ obtained by the ordinary process of assaying, with hydrochloric or sulphuric acid, then pre- cipitating the copper completely with iron, filtering the residual solution, concentrating somewhat, if necessary, and then supersaturating with ammonia. If the fluid 470 THE ASSAY OF COPPEK. remains colourless, neither of the two metals is present : a blue colour indicates nickel, a red one cobalt. Sometimes the normal solutions, which when freshly prepared appear azure blue, assume a greenish hue, which renders the comparison difficult, if not impossible. Copper nitrate produces with ammonia a pure azure blue, copper sulphate a lilac colour, and copper chloride greenish hues. Sulphuric and hydrochloric acid are therefore to be avoided as much as possible in the solution. But, nevertheless, an assay fluid may sometimes e.g. by standing some time in the air, or by slow filtration become green, in which case the colour is destroyed by a few drops of nitric acid, and ammonia added anew. But sometimes also the greenish colour disappears, if the solution stands in a covered vessel in the air, or by the addition of a few drops of red cobalt ammonio-oxide. According to Mliller, also, the colour stands in the closest connection with the quantity of ammonia employed, and it therefore leads to greater accuracy in the assay if- a titrated solution of ammonia is used, and the volume of the ammoniacal fluid noted, which, after neutralisation of the residual free acid, is used for the solution of copper. The solution appears more intense when viewed with a grey background than with a white one. A greenish blue colouring becomes the more noticeable the greater is the excess of ammonia, or the more ammoniacal salts are in the solution. Camelly's Method. Dr. T. Camelry (Manchester Philo- sophical Society) gives the following colorimetric method for estimating small quantities of copper : The method of analysis consists in the comparison of the purple-brown colours produced by adding to a solu- tion of potassium ferrocyanide first, a solution of copper of known strength, and secondly, the solution in which the copper is to be estimated. The solution arid materials required are as follows : (1) Standard Copper Solution. Prepared by dissolv- ing 0-393 grammes of pure CuS0 4 .5H 2 in one litre of water, 1 c.c. is then equivalent to 0-1 mgrm. Cu. COLORIMETRIC ASSAYS. 471 (2) Solution of Ammonium Nitrate. Made by dissolv- ing 100 grms. of the salt in one litre of water. (3) Potassium Ferrocyanide Solution. Containing one part of the salt in 25 parts of water. (4) Two glass cylinders, holding rather more than 150 c.c. each, the point equivalent to that volume being marked on the glass. They must, of course, both be of the same tint and as nearly colourless as possible. (5) A burette, marked T L c.c., for the copper solution, a 5 c.c. pipette for the ammonium nitrate, and a small tube to deliver the potassium ferrocyanide in drops. The following is the method of analysis : Five drops of the potassium ferrocyanide are placed in each cylinder, and then a measured quantity of the neutral solution in which the copper is to be estimated into one of them (A), and both filled up to the mark with distilled water, 5 c.c. of the ammonium nitrate solution added to each, and then the standard copper solution runs gradually into (B), till the colours in both cylinders are of the same depth, the liquid being well stirred after each addition. The number of cubic centimetres used is then read off. Each cubic centimetre corresponds to 0*1 mgrm. of copper, from which the amount of copper in the solution in question can be calculated. The solution in which the copper is to be estimated must be neutral, for if it contains free acid the latter lessens the depth of colour, and changes it from a purple-brown to an earthy brown. If it should be acid it is rendered slightly alkaline with ammonia, and the excess got rid of by boiling. The solution must not be alkaline, as the brown coloration is soluble in ammonia and decomposed by potash ; if it is alkaline from ammonia this is remedied as before by boiling it off, while free potash, should it be present, is neutralised by an acid and the latter by am- monia. Lead when present in not too large quantity has little or no effect on the accuracy of the method. The precipi- tate obtained on adding potassium ferrocyanide to a lead salt is white, and this, except when present in compara- 472 THE ASSAY OF COPPER. tively large quantity with respect to the copper, does riot interfere with the comparison of the colours. When copper is to be estimated in a solution contain- ing iron the following is the method of procedure to be adopted. To the solution a few drops of nitric acid are added in order to oxidise the iron, the liquid evaporated to a small bulk, and the iron precipitated by ammonia. Even when very small quantities of iron are present this can be done easily and completely if there is only a very small quantity of fluid. The precipitate of ferric oxide is then filtered off, washed once, dissolved in nitric acid, and repre- cipitated by ammonia, filtered, and washed. The iron pre- cipitate is now free from copper, and in it the iron can be estimated by dissolving in nitric acid, making the solution nearly neutral with ammonia, and estimating the iron by the method given in the paper before referred to. The filtrate from the iron precipitate is boiled till all the am- monia is completely driven off, and the copper estimated in the solution so obtained as already described. 2. JACQUELAIN'S AND VON HUBERT'S COLORIMETRIC ASSAYS. Heine's method, for the reasons stated, is suitable only for the estimation of small quantities of copper. Jacque- lain has extended it to the examination of all cupriferous substances, and this process has been further perfected by Yon Hubert. According to the latter, a solution of any cupriferous accurately weighed substance is prepared, mixed with ammonia in excess, the ammoniacal solution (assay solution) measured at a definite volume, and a small, likewise measured portion of the measured solution diluted with water, until its blue colour shows an equal intensity with the blue colour of another solution (normal solution), also cupriferous and ammoniacal, whose copper contents are known once for all. Then, from the quantity of water added, in order to make the two fluids equal to each other in the intensity of their blue colours, the amount of copper in the substance under examination can be estimated by calculation. The normal solution is prepared by dissolving -5 of a COLORIMETRIC ASSAYS. 473 gramme of chemically pure copper in dilute nitric acid, adding ammonia in excess, and diluting with distilled water until the whole at 12 C. amounts to one litre = 1000 cubic centimetres. The solution is filtered, and pre- served in a flask provided with a glass stopper ground in to fit it. For the preparation of the assay fluid, with substances whose percentage of copper ranges from 1-5 to the highest per cent., two grammes, and with the poorer substances five grammes, are brought into ammoniacal solution with the precautions specified in Heine's assay. This solution, with over 5 per cent, of copper, is measured at two hun- dred cubic centimetres, with 2 to 5 per cent, of copper at one hundred and fifty cubic centimetres, and with 2 per cent, and under, at one hundred ; and also, as may be required, at 90, 80, 60, 50 c.c., according to the intensity of the fluid. Only with an extremely small quantity of copper is the assay fluid evaporated to a smaller volume, .in order to be able to conduct the colorimetric test with accuracy. The comparison of the intensity of colour of the assay fluid with the normal fluid is accomplished in two different ways, according as the former, when measured at a defi- nite volume, is darker or lighter than the latter. This can be seen if a small arbitrary portion of each is poured into a glass tube of nine millimetres interior diameter, twelve centimetres in length, and uniform thickness, and the two tubes are held in parallel positions over a piece of white paper so that they rest firmly on it, and are inclined to it at an angle of about 45, and direct light falls upon them. Shadow should not fall upon the tubes. (a) The Assay Fluid is Darker than the Normal Solution. By means of a pipette, five cubic centimetres of the normal solution are placed in a glass tube closed at the bottom and not graduated, and seven millimetres in interior diameter and twelve centimetres long. Since 1000 c.c. of the normal solution contain -5 of a gramme of copper, five cubic centimetres contain exactly '0025, and the ratio 474 THE ASSAY OF COPPER. 5 : -0025 expresses once for all the known proportion of copper in the normal solution. Five cubic centimetres of the definitely measured assay fluid are now also placed in a beaker and gradually diluted with water till they show the same intensity of colour as the normal solution. In the comparison the assay fluid must be in a similar tube to that containing the normal solution. With richer proportions a greater accuracy is attained in this comparison if the assay fluid is so far diluted that its intensity still appears as little as possible darker than that of the normal solution, and then water added carefully, drop by drop, till its intensity is judged as little as possible lighter than that of the nor- mal solution, whereupon the mean of the two volumes noted is taken as the correct value. The measuring of the diluted assay fluid is performed in glass tubes of nine millimetres interior diameter and fifty centimetres in length, which from their lower closed end to the circular mark designated by 0, hold exactly five cubic centimetres, and from upwards are divided into cubic centimetres and tenths. If, for example, two grammes of the assay substance have been weighed out, the assay fluid measured at 200 cubic centimetres, and five cubic centimetres of it diluted to 8-2 cubic centimetres, in order to obtain an equal inten- sity of colour between the normal and assay fluid, then the percentage of copper, #, follows from this according to the following chain of ratios : 100 per cent. 200 c.c. assay fluid. 8-2 c.c. diluted assay = normal solution. 0025 gramme of copper in normal solution. x = 8'2 per cent, of copper. (b) The Assay Fluid is Lighter than the Normal Solution. In this case five cubic centimetres of the normal solution are diluted till their intensity is equal to that of the assay solution that has been measured at a definite volume, and for the comparison larger tubes of nine millimetres' interior diameter are used. COLORIMETRIC ASSAYS. 475- If, for example, two grammes of the assay substance have been weighed out, 150 cubic centimetres of assay fluid obtained from it, and to get the same intensity of colour, five cubic centimetres, of the normal solution diluted to 8 -4 cubic centimetres, the quantity of copper x amounts, according to the following chain of ratios, to 2-205 per cent. : 100 per cent. x 2 8-4 5 150 c.c. assay solution. 5 c.c. normal solution. 0025 gramme of copper. x = 16-8 | 37-5 = 2-205 per cent. This assay is adapted for all cupriferous substances,, since nickel, cobalt, and manganese, which would influence the result unfavourably, can be removed without particu- lar difficulty. It is also easy to be learned by those less practised in analytical operations, can be completed in a few hours, and is far less expensive than the dry assay. From two to one-tenth per cent, of copper can also be determined by it with accuracy. Heine, however, prefers his method when a small per- centage of copper is to be estimated, since by it even one loth of copper in the centner = -03 per cent., can be esti- mated, and there is less liability to error. While in slag assays with nine to eighteen loth of copper in the centner,, by Heine's method, errors of half a loth are not to be avoided, variations of more than one loth occur by Von Hubert's process. The latter works with a too deeply coloured normal fluid, corresponding to a solution of over fourteen loth of copper to one ounce of water, while Heine does not exceed four loth. The process is surer if the fluids are diluted and thicker layers of them compared, and the hue thus made artificially deeper, than if small quantities of stronger fluids are compared and the hue made artificially lighter by comparing them in thinner layers, or especially in tubes, where the light is dispersed and shade produced. The comparison in oblong glasses is therefore to be pre- ferred to that in tubes. 476 THE ASSAY OF COPPER. By a comparison of Yon Hubert's assay with tnat of the Oberhartz, it appears that, as Yon Hubert's experi- ments themselves have shown, both give equally accurate results for substances not too poor in copper (i.e. contain- ing not less than 5 per cent.) The Oberhartz assay allows a direct estimation of the copper, requires less ap- paratus, is also very simple, and can be completed in a shorter time. Since different individuals are differently susceptible to colours, and the blue colour of the copper ammoriio-oxide, in consequence of causes yet unknown, sometimes inclines more or less to greenish, and thereby renders observation difficult, therefore, for the sake of greater certainty, though not of greater accuracy, those assays by which an estimation of the copper is possible by weight should in general be preferred to the colori- metric methods, and this is the case with the Oberhartz assay down to two per cent. With smaller percentages the colorimetric assay must be called to our aid. It is not yet settled that with higher percentages of copper the principle of colorimetry is a correct one ; that is, that the intensity of the colour is directly proportional to the quan- tity of the colouring agent. Since ammoniacal solutions poor in copper often show a dash of green colour, Yon Hubert prepares a normal solution for such by dissolving one decigramme of copper and diluting to one litre of fluid. b. Volumetric Copper Assays. FLECK'S MODIFICATION OF MOHR'S METHOD.* The proposal to take the action of solution of potas- .sium cyanide on ammoniacal solution of copper, as the foundation of a method for estimating copper, is due to Carl Mohr.f * This process is given by Fresenius, condensed from the ' Polytechn. Oentralbl.' 1856, 1313. t Annal. d. Chem. u. Pharm. 94, 198 ; Fr. Mohr's Lehrbuch der Titrier- rnethode, 2, 91. VOLUMETRIC ASSAYS. 477 In carrying out this estimation according to the direc- tions of Mr. Parkes, a solution of potassium cyanide is slowly added to a blue ammoniacal solution of copper, when the latter gradually loses its colour, and finally becomes quite colourless ; upon this chemical reaction the estimation of copper by cyanide of potassium depends. By ascertaining by direct experiment the amount of potassium cyanide solution required to discharge the colour in an ammoniacal solution containing a given weight of copper, it is easy by a comparative experiment to estimate the amount of copper in a given weight of ore. For the preparation of the standard solution 2,000 grains of fused potassium cyanide are to be dissolved in two quarts of water, to produce a solution of which 1,000" grains measure will be equal to about ten grains of metallic copper. The solution should be preserved in green glass stoppered bottles, and kept as much as possible away from the light : it is liable to a slow decomposition, which will necessitate the standard being checked at intervals of one or two weeks. In order to standardise the solution, a burette, holding 1,000 grains measure, is filled to the zero mark, and a piece of pure electrotype copper, previously cleaned by means of dilute nitric acid, washed and dried, is accurately weighed. About eight grains may be con- veniently taken ; this is dissolved in a pint flask by dilute nitric acid, and, after the energy of the first action has sub- sided, the solution is warmed and ultimately boiled to expel all the nitrous acid fumes. It is diluted with cold water to the bulk of nearly half a pint, treated with ammonia in excess, and to the deep blue solution the cyanide is added from the burette until the colour is so nearly discharged that a faint lilac tint only remains. This will generally become quite bleached on standing at rest for a short time, so that the cyanide must not be added too hastily towards the end of the operation. It will be advisable to control the standard by a second experiment upon another weighed portion of copper, and to stop short of bleaching entirely the faint lilac tint of the solution. A piece of white paper folded and placed under and behind the flask during the 478 THE ASSAY OF COPPER. decolorisation, will aid in recognising the proper tint of the solution. In applying this process to the examination of copper ores, a known weight of the finely powdered sample is introduced into a beaker provided with a glass cover, and moistened with strong sulphuric acid ; strong nitric acid is then added, and the whole digested on a sand-bath until nitrous fumes are no longer given off. Should a small quantity of sulphur be separated in the treatment of py- ritic ores, the small globules may be taken out, burnt, and the residual copper dissolved in a few drops of nitric acid and mixed with the remainder. Water is now to be added and left in contact for a short time to extract all the metallic salt from the insoluble residue, which need not be filtered off ; and so, likewise, when ammonia is added in the next place, any ferric oxide which may thus be precipitated is left in the solution, for it is apt to contain a small proportion of copper when first thrown down ; but this is entirely removed by the potassium cyanide later in the experiment. When the ore contains much iron it is considered desirable to remove the hydrated peroxide by filtration, in order to be enabled to observe with greater precision the last effects of the potassium cyanide ; and in the event of requiring to know the amount of iron present in the ore, the precipitated ferric oxide on the filter is redissolved in dilute sulphuric acid, reduced to the state of ferrous oxide by metallic zinc, and then tested in the usual way with a standard solution of potassium bichromate. The metals which interfere with this mode of valuing copper ores, are silver, nickel, cobalt, and zinc. The first may readily be separated by adding a few drops of hydro- chloric acid to the original solution : the other metals may be excluded by following one of the methods pointed out by the author for that purpose. Fleck proposes the following modification in this pro- cess : Instead of caustic ammonia, use a solution of ammo- nium sesquicarbonate (1 in 10), warm the mixture to VOLUMETEIC ASSAYS. 479 about 60, and, in order to render the end reaction plainer, add 2 drops of solution of potassium ferrocyanide (1 in 20) : the blue colour of the solution is not altered by this addition, nor is its clearness affected. The value of the potassium cyanide solution is first estimated by means of copper solution of known strength, and it is then employed on the copper solution to be examined. On dropping the potassium cyanide into the blue solution warmed to 60, the odour of cyanogen is plainly perceptible, and the colour gradually disappears. As soon as the ammoniacal double salt of copper is destroyed, the solution becomes red from the formation of copper ferrocyanide, without any precipitate appearing, and with the addition of a final drop of potassium cyanide this red colour in its turn vanishes, so that the fluid now appears quite colourless. The method thus modified yields, it is true, better, but still only approximate, results.* Where such are good enough, the method is certainly convenient. E. 0. BROWN'S METHOD BY SODIUM HYPOSULPHITE. The process described by Mr. E. 0. Brown is particu- larly applicable to the estimation of copper in gun- metal, brass, and other alloys which contain no large amounts of iron and lead. It is founded on the reactions between salts of copper and the neutral iodides, and on the conversion of the liberated iodine into hydriodic and tetrathionic acids by a standard solution of sodium hypo- sulphite. The completion of the second reaction is manifested by the bleaching effect produced upon the blue iodide of starch by the addition of the hyposulphite. A convenient strength of solution for this purpose may be made by dis- solving 1,300 grains of the crystallised salt in two quarts of water. The potassium iodide must be free from iodate ; and a clear solution of starch employed. * In six experiments, in which he had purposely added different quantities of carbonate of ammonia, Fleck used for 100 c.c. copper solution, in the minimum 15-2, in the maximum 15-75, in the mean 15'46 c.c potassium cyanide solution. 480 THE ASSAY OF COPPER. From eight to ten grains of the copper or alloy are dissolved in dilute nitric acid, and the red nitrous fumes expelled by boiling. The copper nitrate is converted into acetate by adding sodium carbonate until a portion of copper remains precipitated, and then re-dissolving in acetic acid. The solution is diluted with water, and about 60 grains of potassium iodide in the form of crystals dropped into the flask, and allowed to dissolve. The standard solution of sodium hyposulphite is now poured in from a burette, until the greater part of the dark- coloured free iodine disappears. A little of the starch solution is now added to make its presence more apparent, and the addition of the hyposulphite continued until the bleaching is completed, when the pale yellow colour of the copper subiodide will alone be visible. The amount of copper in the ore or alloy is calculated from the number of divisions indicated upon the burette. Copper ores containing much iron (which interferes by reason of the dark red colour of the acetate) may be dis- solved in nitric acid, and treated with sulphuretted hydro- gen to precipitate the copper, the sulphide being collected on a filter, washed, and re-dissolved in nitric acid to pro- duce a solution suitable for testing by this process. Or the hyposulphite may itself be employed to furnish a pre- cipitate of copper disulphide. c. Electrolytic Copper Assay. ESTIMATION OF COPPER IN THE MANSFELD ORES BY DR. STEINBECK'S PROCESS. This method embraces three distinct operations, viz. : 1. The extraction of the copper from the ore; 2. The separation ; 3. The quantitative estimation of that metal. 1. The Extraction of the Copper from the Ore. A proof centner, equal to 5 grammes of pulverised ore, is put into a flask, and there is poured over it a quantity of from 40 to 50 c.c. of crude hydrochloric acid, of a specific gravity of 1-16, whereby all carbonates are converted into chlorides STEINBECK S PROCESS. 481 while carbonic acid is expelled. After a while there is added to the fluid in the flask 6 c.c. of a normal nitric acid, pre- pared by mixing equal bulks of water and pure nitric acid of 1*2 sp. gr. As regards certain ores, however, specially met with in the district of Mansfeld, some, having a very high percentage of sulphur and bitumen, have to be roasted previously to being subjected to this process ; and others, again, require only 1 c.c. of nitric acid instead of 6. The flask containing the assay is digested on a well-arranged sand-bath for half an hour, and the contents only boiled for about fifteen minutes, after which the whole of the copper and other metals occurring in the ore are in solution as chlorides. The blackish residue, consisting of sand and schist, has been proved by numerous experi- ments to be either entirely free from copper, or at the most only 0-01 to O03 per cent, has been left undissolved. The extraction of the copper from the ore, according to this method, is complete even in the case of the best quality of ore, which contains about 14 per cent, of metal ; while, at the same time, the very essential condition for the proper and complete separation of the metal, viz. the entire absence of nitric acid, or any of the lower oxides of nitrogen, is fully complied with. 2. Separation of the Copper. The solution of metallic and earthy chlorides, and some free hydrochloric acid, obtained as just described, is separated by filtration from the insoluble residue, and the fluid run into a covered beaker of about 400 c.c. capacity ; in this beaker has been previously placed a rod of metallic zinc, weighing about 50 grammes, and fastened to a piece of stout pla- tinum foil. The zinc to be used for this purpose should be as much as possible free from lead, and at any rate not contain more than from 0-1 to 0*3 per cent, of the latter metal. The precipitation of the copper in the me- tallic state sets in already during the filtration of the warm and concentrated fluid, and is owing chiefly to the com- plete absence of nitric acid completely finished in from half to three-quarters of an hour after the beginning of the filtration. If the fluid be tested with sulphuretted I I 482 THE ASSAY OF COPPEE. hydrogen, no trace even of copper will be detected ; the spongy metal partly covers the platinum foil, partly floats about in the liquid, and, in case either the ore itself or the zinc applied in the experiment contained lead, small quantities of that metal will accompany the precipitated copper. After the excess of zinc (for an excess must be always employed) has been removed, the spongy metal is repeatedly and carefully washed by decantation with fresh water, which need not be distilled, and care is taken to collect together every particle of the spongy mass. 3. Quantitative Estimation of the Precipitated Copper. To the spongy metallic mass in the beaker, wherein the platinum foil is left, since some of the metal adheres to it, 8 c.c. of the normal nitric acid are added, and the copper is dissolved, by the aid of moderate heat, as copper nitrate : in the event of any small quantity of lead being present, it will of course be contaminated with lead nitrate. When copper ores are dealt with which contain above 6 per cent, of copper, which may be sufficiently judged from the larger bulk of the spongy mass of precipitated metal, 16 c.c. of nitric acid, instead of 8, are employed for dissolving the metal. The solution thus obtained is left to cool, and next, immediately before titration with potassium cyanide, mixed with 10 c.c. of normal solution of ammonia, prepared by diluting 1 volume of liquid ammonia, sp. gr. 0*93, with 2 volumes of distilled water. In the case of ores which yield over 6 per cent, of copper, and when a double quantity of normal nitric acid has consequently been used, the solution of copper in nitric acid is diluted with water, and made to occupy a bulk of 100 c.c. ; this bulk is then divided exactly into two portions of 50 c.c. each, and each of these separately mixed with 10 c.c. of the liquid ammonia solution just alluded to, and the copper therein volumetrically esti- mated. The deep blue-coloured solution of oxide of copper in ammonia only contains, besides ammonium nitrate, any lead which might have been dissolved, having STEINBECK'S PROCESS. 483 been precipitated as hydrated lead oxide, which does not interfere with the titration with potassium cyanide. The solution of the last-named salt is so arranged that 1 c.c. thereof exactly indicates 0-005 grm. of copper. Since, for every assay, 5 grins, of ore have been taken, 1 c.c. of the titration fluid is, according to the following, proportion 5 : O'OOo :: 100 : O'l equal to O'l per cent, of copper ; it hence follows that, by multiplying the number of the c.c. of potassium cyanide solution used to make the blue colour of the copper solution disappear, by O'l, the per- centage of copper contained in the ore is immediately indicated. As may be imagined, at the laboratory of the mine- owners at Eisleben, such a large number of assays are daily executed that, in this case, there can be no reason to fear a deterioration of the cyanide solution, of which large quantities are used and often fresh made ; but for security's sake the solutions are purposely tested for control at least once every week. According to the described plan, six assays can be made within 4 hours ; and during a work- ing day of from 7^ to 8 hours, twenty assays have been often quite satisfactorily made by the umpires, as well as by the workmen at Eisleben. Special Observations on this Method. Dr. Steinbeck considered it necessary to test this method specially, in order to see what influence is exercised thereupon by (1) ammonium nitrate, (2) caustic ammonia, (3) the presence of lead oxide. The copper used to perform the experiments for this purpose was pure metal, obtained by galvano- plastic action, and was ignited to destroy any organic matter which might accidentally adhere to it, and, next, cleaned by placing it in dilute nitric acid. Five grammes of this metal were placed in a litre flask, and dissolved in 266 '6 c.c. of normal nitric acid, the flask and contents gently heated, and, after cooling, the contents diluted with water, and thus brought to a bulk of 1,000 c.c. exactly. Thirty c.c. of this solution were always applied to test and titrate one and the same solution of potassium cyanide under all circumstances. When 5 grammes of ore, con- i i 2 484 THE ASSAY OF COPPER. taining on an average 3 per cent, of copper, are taken for assay, that quantity of copper is exactly equal to 0-15 gramme of the chemically pure copper. The quantity of normal nitric acid taken to dissolve 5 grammes of pure copper (266 -6 c.c.) was purposely taken, so as to corre- spond with the quantity of 8 c.c. of normal nitric acid which is applied in the assay of the copper obtained from the ore, and this quantity of acid is exactly met with in 30 c.c. of the solution of pure copper. As regards No. 1 and No. 2 (see above), the influence of double quantities of ammonium nitrate and free caustic ammonia (the quantity of copper remaining the same), and the action of dilute solution of potassium cyanide there- upon, will become elucidated by the following facts : a. Thirty c.c. of the normal solution of copper, con- taining exactly 0-15 gramme of copper, were rendered alkaline with 10 c.c. of normal ammonia, and are found to require, for entire decoloration, 29-8 c.c. of potassium cyanide solution. A second experiment, again with 30 c.c, of normal copper solution, and otherwise under identically the same conditions, required 29-9 c.c. of cyanide solution. The average of the two experiments is 29*85 c.c. b. When to 30 c.c. of the normal copper solution 8 c.c. of normal nitric acid are first added, and then 20 c.c. of normal ammonia solution, instead of only 8, whereby the quantity of free ammonia and of ammonium nitrate is made double what it was in the case of the experiments spoken of under a, there is required of the same cyanide solution 30-3 c.c. to produce decoloration. A repetition of the experiment, under exactly the same conditions, gave 30-4 c.c. of the cyanide solution employed ; the average of both experiments is, therefore, 30-35 c.c. The difference between 30-35 and 29-85 is equal to 0-5 c.c., and that figure is therefore the coefficient of the influence of double quantities ; and supposing this to happen with the ores in question, it would only be equiva- lent to 0-05 per cent, of metallic copper. It is hence clear that slight variations of from 0-1 to 0'5 c.c. in the measuring out of 8 c.c. of normal nitric acid, used STEINBECK S PEOCESS. 485 to dissolve the spongy copper, and of 10 c.c. of normal ammonia, in order to render the nitric acid copper solu- tion alkaline, are of no consequence whatever for the technical results to be deduced from the assay. It should, moreover, be borne in mind that the quantities of free .ammonia and of ammonium nitrate in the actual assay of ores, for which always a quantity of 5 grammes of ore is taken, vary according to the richness or poverty of the ores in copper ; and the quotation of the following results of experiments proves that the influence of these sub- stances is only very slightly felt in the accuracy of the results : Eight c.c. of the normal nitric acid have been found to contain, by a series of experiments, 1*353 gramme of anhydrous nitric acid ; and this quantity of acid is -exactly neutralised by 7*7 c.c. of normal ammonia solu- tion, which contains 0-6515 gramme of ammonium oxide ; and 10 c.c. of the said normal solution contain 0-846 gramme of ammonium oxide. One gramme of metallic copper requires, for complete oxidation, 0-2523 gramme of oxygen, and this quantity of -oxygen is given off by 0*5676 gramme of anhydrous nitric acid ; while, at the same time, binoxide of nitrogen is disengaged. From these data can be calculated (1) the quantity of nitric acid which becomes decomposed when variable quantities of metallic copper are dissolved therein; (2) what quantity of nitric acid is left to form neutral nitrate of ammonium ; and (3) what quantity of free ammonia will be left after a portion of that alkali has been combined with, and therefore neutralised by, copper oxide ; and any remaining free nitric acid. It is found that the quantitative variations between ores containing 1 per cent, or 6 per cent, of metal vary very little from the normal quantities exhibited by ores containing 3 per cent, of metal. The relation is as 1 : 2 ; -and, for technical purposes, this has been proved not to be a disturbing quantity. When, however, larger quantities of ammoniacal salts are present in the fluid to be assayed for copper by means 486 THE ASSAY OF COPPER. of a titrated solution of potassium cyanide, and especially when ammonium carbonate, sulphate, and, worse still, chloride are simultaneously present, these salts exert a very disturbing influence. The presence of lead oxide in the copper solution to be assayed has the effect of producing, on the addition of 10 c.c. of normal ammonia, a milkiness along with the blue tint ; but the presence of this oxide does not at all interfere with the estimation of the copper by means of the cyanide, provided the lead be not in great excess ; and a slight milkiness of the solution even promotes the visi- bility of the approaching end of the operation. Dr. Steinbeck, however, purposely made some experi- ments to test this point, and his results show that neither 50 nor 100 per cent, of addition of lead exerts any per- ceptible influence upon the estimation of copper, from its ores or otherwise, by means of potassium cyanide. A small quantity of accidentally occurring lead will not, therefore,- affect the results, and this the less so as, gener- ally, no ores of both metals occur together wherein both are met in sufficient quantity to make it worth while working the ore for both metals at the same time. Since it is well known that the presence of zinc very perceptibly influences the action of a solution of potassium cyanide, when applied to the volumetrical estimation of copper, Dr. Steinbeck considered it necessary to institute some experiments in order precisely to ascertain with what quantity of zinc present along with copper this in- fluence commences to become perceptible. The solution of zinc applied was made by dissolving the metal in the smallest possible quantity of nitric acid ; and 1 c.c. of that solution contained 0*001 gramme of zinc. The results of the experiments show that the presence of zinc does not interfere with the visibility of the end of the reaction, viz. the decoloration of the copper solution. They also prove that a small quantity of zinc, less than 5 per cent, of the quantity of copper present, or 0*0075 gramme by weight of zinc, does not at all affect the action of the solution of potassium cyanide ; but when the quantity of zinc - LUCKOW'S PROCESS. 487 increases, a very perceptible effect is seen upon the solution of cyanide : it is therefore necessary to bestow due care while washing the spongy copper after it has been precipitated by means of zinc from its solution. Since it has been ascertained that the action of the solution of potassium cyanide in researches of thi's kind is also affected by an increased temperature of the solu- tion of copper which is to be titrated, it is strictly neces- sary never to operate with warm ammoniacal solutions of copper, but to suffer the same to cool down to the ordinary temperature of the air of the laboratory. While 30 c.c. of copper solution, containing 0*15 gramme of copper, and 10 c.c. of normal ammonia solu- tion, required at the ordinary temperature 30 c.c. of cyanide solution, the same quantities required, at between 40 and 45 C., 28*8 c.c. of solution of cyanide ; and at. 45 C., 28'9 c.c. of the same solution, thus proving the injurious effect of warm solutions. ESTIMATION OF COFFEE IN THE MANSFELD ORES BY M. C. LUCKOW'S PROCESS. This gentleman has applied to the quantitative estima- tion of copper a new method, based upon the. precipitation of the metal in the metallic state from solutions containing either free sulphuric or nitric acid, by means of a galvanic current. It is a great advantage of this method, that, while the copper is precipitated, it is simultaneously separated from metals witji which it is often found alloyed ; some of these, such as tin and antimony, are separated by treatment with nitric acid in an insoluble form ; while others, like silver, can easily be removed in the form of chloride. It is,, at the same time, another advantage that the state in which the copper is obtained admits of its being accurately weighed and estimated, while a great number of opera- tions, which require much time and various apparatus, are at the same time got rid of. Although M. Luckow had previously discovered a 488 THE ASSAY OF COPPER. method of electro-metallic analysis from fluids containing free sulphuric acid, his researches on the same subject, in the case of free nitric acid, belong to a recent period. These researches brought very unexpectedly to light the curious fact that even a weak galvanic current had the power of completely precipitating copper in a pure metal- lic state from nitric solutions, provided they did not con- tain more than O'l gramme of anhydrous nitric acid to the c.c. (nitric acid of 1'2 sp. gr. contains 0-32 gramme of anhydrous nitric acid to the c.c.) ; while it was found that the action was, at the same time, more regular, and less dependent upon the power of the current than when free sulphuric acid was present. The following more com- monly occurring metals are not precipitated by galvanic action from acid solutions : Zinc, iron, nickel, cobalt, chromium, the metals of the earths and alkalies. The fol- lowing are precipitated (a) in the shape of peroxides, at the positive electrode : completely, lead and manganese ; in- completely, silver. When the solution contains traces of manganese, it becomes, in consequence of the formation of a salt of manganese peroxide, or of permanganic acid, deeply violet-coloured. This very sensitive reaction for man- ganese also takes place when small quantities of chlorine are present. The presence in the fluid of oxalic and tartaric acids, and other readily oxidisable organic sub- stances, and such protoxides as are readily peroxidised for instance, iron protoxide retards the formation of peroxides, as well as the occurrence of the reaction of manganese. (b) Mercury, silver, copper, and bismuth are precipi- tated at the negative electrode in a metallic state. When mercury is present in the solution simultaneously with copper, the former metal is separated before the latter, in the fluid metallic state. As soon, however, as the pre- cipitation of copper commences, an amalgam of the two metals is formed, when mercury is also present. Silver is precipitated almost simultaneously with copper ; bismuth only begins to be precipitated after the greater portion of the copper has been separated. A complete separation of LUCKOWS PROCESS. 481) silver only ensues when some such substance as tartaric or other similar acid is simultaneously present in the solution. The separation of the three last-named metals by means of galvanic action is, therefore, unsuccessful ; but, fortu- nately, there are many other means to accomplish this end completely. (c) Metallic arsenic is only precipitated slowly, and long after the complete separation of copper, if arsenic acid happen to be present. The same remark applies to antimony, since it is well known that small quantities of antimonic acid are soluble in nitric acid. The operations, according to Luckow's plan, are 1. Eoasting the ore; 2. Solution of the roasted product; 3. Precipitation of the copper ; 4. Weighing the copper. 1. Eoasting the Ore. Care should be taken to obtain a finely ground average sample of the ore. Then weigh off in small porcelain capsules, previously counterpoised, quantities of from 1 to 3 grammes ; these quantities are then placed on the inverted lid of an iron crucible, on the inner surface of which the powdered ore is heated over the flame of a Bunsen gas burner. The powder may be carefully stirred up with a platinum wire, to promote the access of air during the roasting ; the ignition of bituminous matter and sulphur will be ended in about seven minutes. Ores which do not contain bitumen at all need not be roasted. It has been already stated that in the case of poor copper ores (and those of the Mansfeld district are gener- ally so), the quantities to be weighed off for assay should not vary according to a presumed percentage of copper. Two grammes are therefore taken, and, instead of roasting the ore on the lid of an iron crucible, small porcelain crucibles are used for that purpose. 2. Solution of the Roasted Product. The iron lid is suffered to cool, the roasted powder placed on a piece of glazed paper, and any powder adhering to the lid is re- moved by means of a camel's- hair brush on to the paper. The powder is next transferred to small beaker glasses, and about 2 or 3 c.c. of nitric acid, of 1-2 specific gravity, are added, along with about 10 to 15 drops of concentrated 490 THE ASSAY OF COPPER. sulphuric acid. The beakers are then placed on a sand- bath and moderately heated, at first ; but when the con- tents have become nearly dry, the heat is increased, so as to evaporate and expel all sulphuric acid. The beakers should be covered with perforated watch-glasses. This operation requires from about three-quarters of an hour to one hour. The addition of sulphuric acid is made in order to increase the oxidising action of the nitric acid, and also to convert any lime which may happen to be present in the ore into a difficultly soluble salt. It is very useful, also, to add from 10 to 20 drops of hydrochtoric acid to the mixture of the two acids just alluded to, since the rapidity of the evaporation is thereby increased, and the occasional spirting about of the fiuid is lessened. The process just described may be modified, first, by the use of porcelain capsules, the contents of which are easily transferred to beakers with flat bottoms, and not higher than about 2 inches altogether. It is better, also, to use sulphuric acid, prepared with equal bulks of con- FIG. 105. centrated acid and water, and to measure off 4 c.c. for each assay ; while for each assay, moreover, 6 c.c. of nitric acid and about 25 drops of hydrochloric acid are taken. Instead of covering the beaker with a perforated watch-glass, the upper part of a funnel is used, as represented in fig. 105 ; with this arrangement the sulphuric acid evaporates far more readily, and loss by spirting is pre- vented. The beaker is heated on a well-arranged sand- bath. 3. Precipitation of the Copper. As soon as the beaker after removal from the sand-bath has become quite cool, the funnel which has been used as a cover is washed on both sides, inner and outer, with nitric acid of 1*2 sp. gr., diluted with six times its bulk of pure water ; the sides of the beaker are next likewise washed, and it is then filled to about half its height with the same acid. A few drops of a concentrated solution of tartaric acid are added (this acid is best kept in solution in open vessels, only slightly LUCKOWS PROCESS. 491 covered with a piece of paper) ; this having been done, the wire spiral, represented in fig. 106, is carefully placed in the beaker. This spiral consists of a piece of platinum wire, about l-12th of an inch thick, and 7-J- inches long, two-thirds of its length being so wound that the straight end of the wire projects as if it were the axis of the centre of the spiral. The convolutions of the spiral are so large that they touch the sides of the beaker, while the straight portion just touches the centre of the bottom of the vessel. When the heating has been carefully attended to, the acid fluid added to the contents of the beaker, after evaporation to dryness, will generally be quite clear ; if it FIG. 106. FIG. 107. FIG. 108. happens to be turbid, 1 or 2 c.c. of a concentrated solution of barium nitrate may be added, and the thorough mixing of this saline solution with the acid contents of the beaker promoted by gently moving up and down the platinum spiral just alluded to, and allowing the fluid to rest for -a few minutes after. The copper present in the mass left at the bottom of the beaker gradually dissolves, and it is not actually requisite to wait before applying the galvanic current until it is all dissolved. The next point is to place in the beaker the platinum 492 THE ASSAY OF COPPER. foil, represented at fig. 107, of which the dimensions are length, 2^ inches ; width, 1J inch. The lower end of this platinum foil should be kept about 1-1 6th of an inch apart from the convolutions of the spiral. When the beaker is only half filled with liquid, the platinum foil is immersed in the same for more than three-fourths of its height. The wire fastened to this foil is fixed, by means of a screw a, to the arm a b of the stand, represented in fig. 108 ; the other screw, b, serves to fasten a copper wire proceeding from the zinc end of the galvanic battery. When the small screw clamp c (fig. 108) has been fastened to the platinum wire placed in the beaker, another wire is fastened in the top opening of the clamp, and this wire connected with the copper end of the battery, and the galvanic circuit thus closed. In a few moments after this has been done, the platinum foil, bent in the shape of the cylinder and placed inside the beaker, as before described, will be observed to become covered with a coating of metallic copper, while from the platinum wire spiral bubbles of gas escape, which facilitate, to some extent, the solution of the copper oxide in the dilute acid. In order to ascertain whether the whole quantity of the copper has been precipitated, some more dilute nitric acid is added to the fluid in the beaker glass. If, in ten minutes after this, no more metallic copper is separated on the clean portions of the platinum foil, the operation is finished. It must be here observed that continued practice has proved that the addition of a concentrated solution of barium nitrate acts injuriously on the process, as the metallic copper, which becomes separated, gets mixed with some insoluble barium sulphate, which increases the weight of the substance to be weighed. The time occupied by the complete precipitation of the metal varies according to the force of the galvanic current. It takes from three to even eight hours. In order to make this point certain, all test assays are left, for eight hours consecutively, to the action of the galvanic current, expe- rience having proved that, after that lapse of time, even LUCKOWS PROCESS. 493 with a weak current, the precipitation was so complete that all chemical reagents for detecting the presence of copper failed to discover the most minute trace of that metal. 4. Weighing the Copper. The platinum cylinder to which the copper adheres, and the platinum wire spiral, are disconnected from the galvanic apparatus, the plati- num cylinder carefully removed from the beaker and immediately plunged into a beaker filled with fresh cold water, and rinsed therein ; next washed with alcohol, by means of a washing bottle, and then dried in a drying ap- paratus, and weighed after cooling. Since the platinum cylinder has been- very accurately weighed before the experiment, its increase in weight will, of course, be that of the copper obtained. The process here described has been somewhat modi- fied and greatly improved upon at Eisleben, where it is in constant use, by the employment of a series of galvanic elements. It is, in the first place, found better not to disconnect the galvanic current while the copper is yet in contact with acid, so that, instead thereof, the acid fluid in the beaker is replaced by turning in a stream of water,, and suffering the same to run over the sides of the beaker, and to be received into a proper vessel to hold it. In this manner all the acid is displaced, without risk of any very small quantity of copper becoming acted upon by the acid during the brief period elapsing between the disconnecting of the galvanic current and the removal from the beaker of the platinum cylinder and spiral wire. These parts, on being removed, are carefully washed, first with boiling water, next with alcohol, and then dried at a temperature of about the boiling-point of water. The cylinder is then weighed, the copper coating is removed by means of nitric acid ; the platinum is next washed in water, dried, and again weighed. There are in use at Eisleben nine galvanic batteries (lead and zinc elements) ; these yield eighteen assays ready for weighing in twenty-four hours ; and it would not be difficult for the person there employed to work with twelve- 494 THE ASSAY OP COPPER. batteries each of three elements. In place of the Mei- dinger elements, which do not remain constant for months, a thermo-electric apparatus has been introduced, with the best results. The results obtained are highly satisfactory. The following observations may be made in reference to this method : (a) The quantity of ore taken for trial is 2 grammes ; this is found sufficient, while it consumes less acid. (b) The evaporation of the acid is carried on to com- plete dryness on the sand-bath. Spirting of the fluid is easily prevented. When the copper has been precipitated properly it will show its peculiar colour on the surface, and the good success of the operation may also be judged from the fact that no saline matter adheres to the platinum ; the com- plete absence of this saline matter has been found to be evidence of perfect removal of the copper from the fluid. The process just described is especially applicable for rather poor ores, such as do not contain above 7 or 8 per cent, of copper. Each assay, from beginning to end, takes ten hours for complete analysis ; but it is evident that the greater portion of this period does not give active employ- ment to the assayer. The expense of working this process, after the apparatus has been once purchased, is very small. The process may also be applied to analyse richer ores, and also alloys of copper, with some slight modifications which will readily suggest themselves. Assay of Copper Pyrites. The following method of treating copper pyrites has been found more advantageous than the ordinary process of oxidising the mineral with aqua regia, and subsequently evaporating the solution re- peatedly with hydrochloric acid, or with sulphuric acid, to expel the last traces of nitric acid. It is thus described by Mr. F. P. Pearson in the ' Chemical News ' : Place a weighed quantity of the powdered mineral, to- gether with some potassium chlorate, in a porcelain dish. (Five grammes of a variety of a pyrites containing about 18 per cent, of copper was found to be enough for one ASSAY OF COPPER PYRITES. 495 analysis ; and a quantity of potassium chlorate equal to a small teaspoonful was added to the ore.) Invert a small glass funnel with bent stem in the dish above the pyrites, and pour upon the latter rather more ordinary strong nitric acid than would be sufficient to completely cover the powder. Place the dish upon a water-bath, and, from time to time, throw into it small quantities of potassium chlorate. The doses of the chlorate must be repeated at frequent intervals, until free sulphur can no longer be seen in the dish. If need be, add nitric acid, also, from time to time, to replace that lost by evaporation. As a general rule, it is safer and more convenient to heat the mixture on a water-bath than upon sand, though the oxidation of sulphur can be effected more easily and quickly when the mixture of nitric acid and chlorate is heated to actual boiling than at the temperature obtain- able by means of a water-bath. When the last particles of sulphur have been destroyed, remove the inverted funnel from the dish, rinse it with water, and collect the rinsings in a beaker by themselves. Allow the liquid in the evapo- rating-dish to become cold, pour upon it a quantity of ordinary strong hydrochloric acid rather larger than the quantity of nitric acid taken at first, evaporate the mixed solution to dryness, and heat the dry residue to render silica insoluble, in case any silica be present. Pour water upon the cold residue, and, without filter- ing the liquor, wash the contents of the dish into the beaker which contains the rinsings of the funnel. Heat the liquid in the beaker nearly to boiling, add to it about 25 c.c. of a strong aqueous solution of ferrous sulphate slightly acidulated with sulphuric acid, and keep the mix- ture at a temperature near boiling during four or five minutes, in order to destroy the small quantity of nitric .acid which may have escaped decomposition in spite of the dry evaporation with hydrochloric acid. The ferrous salt seldom acts instantaneously, but the reducing action proceeds rapidly and satisfactorily when once begun. If need be, add more of the ferrous solution, little by little, until the entire contents of the beaker 496 THE ASSAY OF COPPER. become dark-coloured or black, and no more gas is dis- engaged. In order to be sure that all the nitric acid has been reduced, it is as well, after the mixture of liquid and solu- tion of ferrous sulphate has been duly heated, to place a drop of the mixture upon porcelain, and test it with potassium ferrocyanide. In general, however, the colora- tion of the liquid in the beaker, due to the formation of nitrous or hyponitric acid, will be a sufficient indication that the sulphate of iron has done its work. The nitrous fumes quickly disappear from the liquid at a subsequent stage of operations when metallic iron is immersed in the solution. When enough of the ferrous sulphate has been added, filter the mixed solution into a wide beaker, precipitate the copper in the metallic state upon a sheet of iron in the usual way, and ignite the copper in a porcelain crucible, in a current of hydrogen, before weighing it. By means of the ferrous salt, the last traces of nitric acid may be got rid of far more quickly, conveniently, and certainly than by the old system of evaporating the pyrites solution with several successive portions of hydrochloric acid. By treating the pyrites with potassium chlorate and nitric acid it is easy to oxidise and dissolve every particle of the sulphur in the mineral, so that no portion of the latter can escape decomposition by becoming enveloped in free sulphur. When aqua regia is used, on the other hand, or a mixture of potassium chlorate and hydrochloric acid, a certain proportion of sulphur almost invariably remains undissolved, and might easily enclose portions of the mineral, so as to prevent them from the solvent action of the acids. For the Estimation of Copper and Sulphur in Copper Pyrites, E. Fresenius (' Zeitschrift fur Anal, Chemie,' 1877, p. 355) proceeds as follows : ' After drying at 100 C., and carefully preparing the sample, he takes for the estimation of the copper 5 grammes pyrites, heats with 6 to 7 c.c. hydrochloric acid (sp. gr. 1'17), adding gradually nitric acid of sp. gr. 1-37, till SULPHUR IX COPPER PYRITES. 497 no more action ensues, and then digests for some hours at a gentle heat. The contents of the flask are poured into a porcelain capsule, the flask is twice rinsed, out with 10 c.c. hydrochloric acid (sp. gr. 1*12) into the capsule, and is then set aside. The contents, of the capsule are then evaporated almost to dry ness on the water-bath, 20 c.c. hydrochloric acid of sp. gr. 1/12 are added, heated, mixed with water, and filtered into a boiling-flask holding 500 c.c. The solution-flask is also rinsed upon the filter with water. The filtrate is dried, incinerated in a porcelain crucible, and the residue (in part lead sulphate) is treated with 1 c.c. aqua regia, evaporated to dryness, the residue treated with 5 c.c. hydrochloric acid (1*12 sp. gr.) slightly diluted, and the solution, which may contain a little copper, is filtered into the main solution. The solution is then precipitated with sulphuretted hydrogen at 70 C., the precipitate is filtered, washed, dried, mixed with the thoroughly burnt ash of the filter, heated with 5 c.c. nitric acid (of sp. gr. 1-2), filtered, and washed. The filter is incinerated, the small residue is again heated with 2 c.c. of the same nitric acid, diluted, filtered to the main solu- tion, and washed. The solution is mixed with 12 c.c. dilute sulphuric acid (1 part acid to 5 water), evaporated to expel nitric acid, filtered to remove lead sulphate, washing with water acidulated with sulphuric acid. The copper is then precipitated with sulphuretted hydrogen at 70 C., the precipitate washed, dried, ignited along with the ash of the filter in a current of hydrogen, and weighed as copper sulphide. 6 For the estimation of sulphur the author fuses ^ grm. of the sample with 10 parts of a mixture of 2 parts sodium carbonate and 1 part potassium nitrate, and estimates the sulphuric acid formed. For burnt ores he uses 4 parts sodium carbonate to 1 part of potassium nitrate.' Professor Chapman, of Toronto, gives the following directions for the detection of minute traces of copper in iron pyrites and other bodies : Although an exceedingly small percentage of copper K K 498 THE ASSAY OF COPPER, may be detected in blowpipe experiments by the reducing process, as well as by the azure-blue coloration of the flame when the test matter is moistened with hydrochloric acid, these methods fail in certain extreme cases to give satisfactory results. It often happens that veins of iron pyrites lead at greater depths to copper pyrites. In this case, according to the experience of the writer, the iron pyrites wi]l almost invariably hold minute traces of copper. Hence the desirability, on exploring expeditions more especially, of some ready test by which, without the necessity of employing acids or other bulky and difficultly portable reagents, these traces of copper may be detected.* The following simple method will be found to answer the pur- pose : The test substance, in powder, must first be roasted on charcoal, or, better, on a fragment of porcelain,f in order to drive off the sulphur. A small portion of the roasted ore is then to be fused on platinum wire with phosphor-salt ; and some potassium bisulphate is to be added to the glass (without this being removed from the wire) in two or three successive portions, or until the glass becomes more or less saturated. This effected, the bead is to be shaken off the platinum loop into a small capsule, and treated with boiling water, by which either the whole or the greater part will be dissolved ; and the solution is finally to be tested with a small fragment of potassium ferrocyanide. If copper be present in more than traces, this reagent, it is well known, will produce a deep red precipitate. If the copper be present in smaller quantity that is, in exceedingly minute traces the precipitate * In blowpipe practice as far, at least, as this is possible the operator should make it an essential aim to render himself independent of the use of mineral acids and other liquids and inconvenient reagents of a similar character. If these reagents cannot be dispensed with altogether, their use, by improved processes, may be greatly limited. t In the roasting of metallic sulphides, &c., the writer has employed, for some years, small fragments of Berlin or Meissen porcelain, such as result from the breakage of crucibles and other vessels of that material. The test substance is crushed to powder, moistened slightly, and spread over the sur- face of the porcelain ; and when the operation is finished, the powder is easily scraped off by the point of a knife-blade or small steel spatula. In roasting operations, rarely more than a dull red heat is required ; but these porcelain fragments may be rendered white-hot, if such be necessary, without risk of fracture. ESTIMATION OF ARSENIC IN COPPER. 499 will be brown or brownish- black ; and if copper be en- tirely absent, the precipitate will be blue or green assum- ing, of course, that iron pyrites or some other ferruginous substance is operated upon. In this experiment the pre- liminary fusion with phosphor-salt greatly facilitates the after solution of the substance in potassium bisulphate. In some instances, indeed, no solution takes place if this preliminary treatment with phosphor-salt be omitted. Estimation of Arsenic in Copper. The estimation of the small quantity of arsenic always present in commercial copper, or the separation of a very small quantity of arsenic from a large amount of copper, is a matter of considerable difficulty, as the ordinary methods of separation fail to give accurate results. Having had a very extensive experience in the analysis of copper, and knowing the extraordinary discrepancies which occur in analyses of the same sample of copper by different chemists, which can only arise from the use of imperfect methods, Mr. A. Humboldt Sexton proposes method which he has found to give quite accurate results. The copper is dissolved in nitric acid, a small quantity of solution of ferric nitrate added, the solution nearly neutralised with sodium hydrate (not ammonia), and excess of sodium acetate added. The solution is then heated to boiling, and filtered as rapidly as possible : the precipitate after being well washed is dissolved in hydrochloric acid, the solution made alkaline with ammonia and saturated with sulphuretted hydrogen, and filtered from the precipi- tated iron sulphide. The filtrate is acidified with hydro- chloric acid and allowed to stand in a warm place for some time. The arsenic and antimony sulphides are filtered off, dried at 100 C., the precipitates removed completely from the paper into a small beaker, treated with red fuming nitric acid, a few drops of hydrochloric acid being added as soon as the action has ceased. It is then diluted, filtered, the arsenic precipitated as ammonia- 500 THE ASSAY OF COPPER. magnesium arseniate, and weighed as usual. If the preci- pitated sulphides cannot be perfectly removed from the filter-paper, the paper must be treated with nitro-hydro- chloric acid, filtered, and the filtrate added to the nitric acid solution. This method is very accurate, and each stage has been carefully experimented upon. It requires, however, some special precautions. When the sodium acetate is added, the colour of the solution should change from pale blue to dark green ; this shows that the solution has been sufficiently neutralised. The beaker must be removed from the heat immediately the solution begins to boil ; if the solution be left boiling (sometimes when it is not), a greenish white precipitate of basic copper acetate- falls. This can generally be removed by the addition of a few drops of hydrochloric acid, but in cases where it has separated on the surface of the beaker, or where it will not readily dissolve, it is best to throw out the solution and commence again. This is very troublesome to those using this method for the first time, but after a little experience has been gained it very rarely happens. The precipitate should have the dark red colour of ferric acetate ; if it is paler it is due either to there not being sufficient iron, or to the co-precipitation of some basic copper acetate. The filtrate should be blue or pale green : sometimes it is dark green and turbid, from the presence of iron acetate carried through the filter ; in that case the first portions must be passed through the filter again. The precipitate must be washed till it is free from copper, and when it is dissolved in hydrochloric acid the solution must have the yellow colour of ferric chloride. If it is at all green, the solution must be neutralised, a little more sodium acetate added, and the iron and arsenic re-precipitated. Mr. Sexton has made a larger number of experiments in order to ascertain the amount of iron necessary for complete precipitation. With equal quantities of iron ESTIMATION OF ARSENIC IN COPPER. 501 and arsenic, a small quantity of arsenic remained in solu- tion, and the iron-arsenic precipitate was of a pale colour. With 1-5 parts of iron to 1 of arsenic the precipitation was complete. In order to make sure, it is well to add about twice as much iron as it is expected there is arsenic present. Then, even if a little iron remains unprecipitated, all the arsenic will be thrown down. Since copper sulphide retains so much arsenic, it might be expected that iron sulphide would act in a similar manner, but it does not ; if there be no copper present the precipitate is quite free from arsenic, but if copper is present a considerable quantity of arsenic may be retained. Hence the importance of thoroughly washing the acetate precipitate, and re -precipitating it if necessary. 502 CHAPTER XI. THE ASSAY OF LEAD. ALL minerals and substances containing lead may, for the purposes of assay by the dry way, be divided into four classes : Class I. comprises sulphides, antimonial or otherwise (galena, &c.) Class II. includes all plumbiferous substances contain- ing neither sulphur nor arsenic, or mere traces only of these elements (litharge, minium, lead carbonate, native and artificial, lead fume, cupel bottoms, furnace hearths, lead slag, &c.) Class III. comprises all substances into whose com- position either sulphuric, arsenic, chromic, or phosphoric acid, or a mixture of either, enters (pyromorphite, wolfram- ite, &c.) Class IV. Alloys of lead. CLASS I. Before describing the different modes of assaying sub- stances of this class, it will be as well to pass in review the action of various reagents on sulphides of lead, in order that the rationale of the assay of those ores may be better appreciated. Action of Oxygen. If galena be roasted at a very gentle temperature, care being taken to avoid fusion, it- will be converted into a mixture of lead oxide and lead sulphate, with evolution of sulphurous acid, thus: 2(PbS) + 70 =PbO + PbO,S0 3 + S0 2 . Action of Metallic Iron. This metal completely and ALKALIES AND ALKALINE CARBONATES. 503 readily decomposes lead sulphide, giving metallic lead in a pure state, thus : On the one side we have lead sulphide and metallic iron, on the other metallic lead and iron sulphide. The Alkalies and Alkaline Carbonates decompose lead sulphide, but only partially ; pure lead is separated, and at the same time a very fusible grey slag is formed, which contains an alkaline sulphate and a compound of lead sul- phide and an alkaline sulphide. A certain proportion of the alkali is reduced by the sulphur, which is converted into sulphuric acid, so that no lead oxide is produced. This reaction may be thus expressed : 7(PbS) + 4(K 2 0)=4Pb + K 2 0,S0 3 + 3(PbS,K 2 S). Potassium Nitrate completely decomposes lead sul- phide, with the reduction of metallic lead and formation of potassium sulphate and sulphurous acid, thus : If the nitre be in excess, the lead will be oxidised in pro- portion to the excess present ; and if there be a sufficiency added, no metallic lead at all will be produced. Argol. The presence of carbonaceous matter much favours the decomposition of galena, by determining the reduction of a larger quantity of potassium, and thereby the formation of a larger quantity of alkaline sulphide. With 4 parts of argol to 1 part of sulphide, 80 parts of lead are reduced. If the reaction were complete, the decomposition would be as follows : For the reactions of lead oxide (litharge) and lead sul- phate on sulphide of lead, see pages 187 and 188. From the reactions above given, it will be seen that there are many substances capable of completely reducing the lead from its sulphide, and yet few can be used safely 504 THE ASSAY OF LEAD. with any advantage, as so to use them would imply a knowledge of how much sulphur and lead were in the ore to be assayed, in order to tell the precise quantity of either of the reagents required ; for it is evident that if either more or less of some were added, a faulty result would be the consequence : so that some systematic mode of assay, which may be suitable for all classes of galena, whether mixed with other sulphides or with gangue, must be contrived. To facilitate this we now proceed to give an outline of the methods generally adopted in the assay of lead ores by various processes. 1. FUSION WITH POTASSIUM CARBONATE. This plan is used at the Oberhartz, and described by Kerl as follows : One centner of the very finely pulverised assay sub- stance is weighed out, mixed with three to four times its weight of pure, dry, and finely pulverised potassium car- bonate, and covered over, in a small clay crucible (fig. 61), with a layer of decrepitated sodium chloride about one- fourth of an inch thick. The assays thus prepared are placed in the thoroughly heated muffle of a large assay furnace (figs. 23, 24) having a strong draught. They re- main in the highest temperature of the furnace, with the mouth of the muffle closed with glowing coals, till they have come into perfect fusion (about twenty to thirty minutes). The draught opening is then closed, and at the same time the muffle opened, until the temperature has fallen so far that the crucibles appear brownish red, and the vapours above them have greatly diminished, or have disappeared. At this heat the crucibles whose contents must, however, always remain in perfect fusion are main- tained, according to the fusibility and composition of the assay sample, and the draught of the furnace for a longer or snorter time (ten to twenty-five, generally ten minutes). This period, during which the heat is allowed to remain low, is called the cooling of the assay. The furnace is now again brought back to its first FUSIOX WITH POTASSIUM CAEBONATE. 505 temperature by completely opening the draught and closing the muffle. Ten to fifteen minutes of this last heating are in most cases sufficient. Only poor ores, &c., which contain also a pretty large quantity of arsenic, or of iron, zinc, and copper sulphides, are allowed to continue hot five to ten minutes longer. If many assays are to be made, it will be found advan- tageous to mix those which contain larger quantities of foreign sulphides, or, by reason of their earthy contents, are difficultly fusible, with more or less borax ; or, instead of this, to place them in the back and hotter part of the muffle, while those that are very rich in lead and easily fusible are placed in front, since the latter will be hot enough here, and more easily reached by the air than those deeper in the muffle. The crucibles, when cold, are broken, the lead buttons obtained are freed from all adhering slag or substance of the crucible, and if the assay were otherwise successful their weight found. The assays should not be too rapidly cooled, because the slag is thus easily cracked, and the still half-fluid button lying below is apt to be broken into several pieces. In a successful assay, the lead melted together to a button deports itself under the hammer and knife like pure lead, and possesses also its colour. If the slag shows, upon its surface of separation from the metallic button, lead- grey spots with metallic lustre, it will generally also be found that a thin layer of not completely decomposed glistening lead sulphide or subsulphide has at the same time deposited itself upon the button. This layer, if the above appearance presents itself in a high degree, can be rubbed off or removed in fine scales. The lead button itself then shows upon its surface a high metallic lustre, which does not have the colour of pure lead, but a darker and blackish hue. Assays of this kind are to be rejected ; they have not been allowed to remain cool long enough, or they have in the process become too cold: they give the amount of lead too low, and often very considerably so. In assays which have stood too long in the furnace 506 THE ASSAY OF LEAD. in the last fusing heat, a very bright button of lead is also found ; but here the layer of undecomposed lead sulphide is wanting, as also the glistening spots on the surface of the slag surrounding the button. If the influence of the heat and air continues too long, then, besides a loss through volatilisation of the lead, a slagging of the lead oxide may take place. A button that is brittle, laminated, and bril- liantly white in the fracture, indicates an insufficiency of flux, or the presence of antimony and arsenic. In successful assays the lead button generally has a bluish appearance, which, although not dull, is at the same time not strongly brilliant. The slag must be completely homogeneous, and must have settled down uniformly towards the bottom of the crucible, so that it does not stick in a thick layer to the upper part of the sides of the crucible. It shows by this that it has been in proper fusion. It must have covered over the button in a thick layer (about one-fourth of an inch thick). The sodium chloride covering, or a more or less colourless slag that is formed containing sodium chloride and potassium carbonate, overlies in a still thicker layer the true dark-coloured slag containing the foreign metallic oxides. A porous slag containing metallic globules indicates a too small quantity of flux or too low a temperature ; a brilliant vitreous slag, too high a temperature and a slagging of lead. An assay and its duplicate must, moreover, give equal results. Lead matte and lead fume are smelted, with the addi- tion of borax and coal-dust, with potassium carbonate, and with the first the heat is allowed to last somewhat longer (perhaps to three-quarters of an hour) than with ores. The potassium carbonate assay gives for lead matte, with its not inconsiderable lead contents (30 per cent, and over), pretty satisfactory results. The theory of this lead assay appears from the fol- lowing. If perfectly pure galena is intimately mixed with three or four times its weight of good dry potassium carbonate, placed in a clay retort, and this so arranged in the muffle of the assay furnace that its neck projects from the mouth FUSION WITH POTASSIUM CARBONATE. 507 of the muffle, while in the opening of the neck a glass tube is closely fitted, which goes into a receiver, from which it is further prolonged in a second tube, it will be observed that at first only a little water collects in the receiver, pro- ceeding from the small quantity of moisture always present in the potassium carbonate. Later, with an incipient red heat in the retort, a gas is disengaged, which upon closer investigation proves to be pure carbonic acid gas, i.e. free from sulphurous acid. The disengagement of gas becomes more active with a stronger red heat, without yielding gases of different composition, but ceases again after a while. In order to obtain assurance of a complete decom- position, the retort may be kept for an hour at a very strong red heat. After the cooling and breaking of the retort, some pure lead oxide and carbonate is found deposited in the neck of it, then a pure lead button upon the bottom, and over this a brown slag, free from little globules of lead. It consists in by far the greatest part of potassium sulphide and still undecomposed potassium carbonate, but also in small part of potassium silicate de- rived from the silica of the retort. If this slag is treated with water till nothing further will dissolve, the substances named can be easily shown to exist in the solution. The solution is colourless, and when supersaturated with acids disengages sulphuretted hydrogen, but throws down no sulphur. In the treatment of the slag with water, lead sulphide remains behind in black flocks, even the superfi- cial character of which shows that it is not undecomposed galena, but lead sulphide separated from a chemical com- bination. If the brown slag from the retort is placed in a small uncovered crucible and brought back into the hot muffle of the assay furnace and melted, then after some time, whether the slag was covered with sodium chloride or not, a button of lead again separates at the bottom of the cru- cible, and the brown slag now shows itself decolourised. If the crucible is removed from the furnace too soon, only the upper layer of slag is decolourised, and that lying below is still completely unchanged. The decolourised slag 508 THE ASSAY OF LEAD. consists of potassium carbonate and sulphate, and no longer contains any trace of potassium sulphide. In the above-described lead assay, the process in the strong preliminary heat proceeds as in the retort, i.e. the potash of the potassium carbonate is reduced to potassium, while it yields its oxygen to the sulphur of the galena and with it forms sulphuric acid ; the liberated potassium takes up sulphur from another portion of galena, forming potas- sium sulphide. The galena would now in this double way soon lose all its sulphur, if a combination a sulphur salt of potassium sulphide with lead sulphide did not form, which resists all further action of the potassium carbonate [4 (K 2 0,C0 2 ) + 7PbS = 4Pb + 3 (K 2 S,PbS) +K 2 0,S0 3 + 4COJ. The carbonic acid of the thus decomposed potassium car- bonate escapes together with that set free by the sulphuric acid formed, and causes a puffing up of the mass, by which globules of lead already separated are raised up with it, and may perhaps remain with some of the slag sticking to the upper crucible walls. They would here oxidise and produce yellow spots. The covering of sodium chloride is designed to guard against loss of lead in this and similar ways. It serves in a certain manner to rinse down the sides of the crucible. The atmospheric oxygen, in the open crucible, is not entirely excluded by the covering of sodium chloride. In the cooling of the assay, it oxidises the sulphur salt con- tained in the upper part of the slag, forming potassium sulphate and a portion of lead sulphate. The latter, during the last high heat, decomposes the lead sulphide still remaining in the slag, in such a way as to produce metallic lead (PbS + PbO,S0 3 = 2Pb + 2S0 2 ). The reduced particles of lead separate well from the slag thus rendered thinly fluid. Mattes must be allowed to cool longer than ores. The potassium carbonate assay presupposes in general great practice and close attention on the part of the assayer ; and moreover, if one wishes to find the correct value at once, without fruitless preliminary examinations, and without the necessity of repeating the assay, a general FUSION WITH POTASSIUM CARBONATE. 509 knowledge of the constituents of the assay sample, so far, for example, as this can be obtained by the aid of minera- logy, is necessary. The assay after this method, which requires but little preparation, can only be conducted in the muffle furnace, but then in pretty large number (as many as fifty at once). For its success it is indispensably necessary that the cooling of the assay be allowed and stopped again at the right time and in the proper degree. If it is allowed to cool too long, too much sulphate of lead is formed in proportion to the lead sulphide still present in the slag, and in the last heating up, by the action of the two upon each other, easily scorifiable oxide of lead is produced (PbS + 3PbO,S0 3 =4PbO + 4S0 2 ). If the cooling is too soon interrupted, only a small part of the lead sulphide in the sulphur salt is oxidised, and, by the action of the oxidised portion upon the lead sulphide, lead sub-sulphide is produced, which either remains in the slag or settles upon the lead button (2PbS + PbO,S0 3 = Pb 2 S + 2Pb + 2S0 2 ). Experience gives the only means at hand to guide us here, but leaves us easily in the lurch, so that the result of the assay becomes more doubtful than in some of the methods- hereafter described. With substances containing antimony this assay deserves the preference over the others, since most of the antimony remains in the slag in the state of sulphide and oxide. An addition of saltpetre works advantageously. Arsenic and arsenic sulphide mostly go off in fumes during the smelt- ing, but nevertheless always cause the formation of a brittle metallic button. Copper sulphide remains in great part in the slag, but a part of the copper is desulphurised and goes into the lead. If the quantity of copper present is very considerable, the button of metal may be considered as black copper, and refined, and the loss thereby occur- ring reckoned as lead. Iron protosulphide, which occurs, for example, in lead matte, is decomposed by potassium carbonate, forming metallic iron, which desulphurises the galena. Iron pyrites , on the other hand, occasions the forming of a large quan- 510 THE ASSAY OF LEAD. tity of potassium sulphide and, in consequence of this, of a sulphur salt. It follows, therefore, from the above, that ores which contain much foreign sulphides are not suited to this method of assaying, since they cause the production of a large amount of potassium sulphide, which always retains lead sulphide. By an addition of saltpetre to the potas- sium carbonate these sulphides may, indeed, be partially decomposed : only an oxidation of the lead is apt to be produced, as well as a mechanical loss by the violent action of the saltpetre. From pure galena, by the potassium carbonate assay, 80 per cent, of lead at most can be obtained. Calcined sodium carbonate is inferior to potassium carbonate as a desulphurising agent, and always yields a few per cent, less lead than the latter. According to .Phillips, 75 to 77 per cent, of lead is obtained from galena with sodium car- bonate. With potassium cyanide, under certain circum- stances, the same result can be obtained as with potassium carbonate, and it does not require so high nor so long- continued a temperature ; still it offers no real advantage over potassium carbonate. An addition of 30 to 35 per cent, of saltpetre to an assay, with which ten parts of sodium carbonate are used, promotes, indeed, the desul- phurising of the lead, but also increases its loss. At the Oberhartz smelting-house the lead button is weighed out to pounds, and a difference of five pounds is allowed between different assayers. It is also a custom, though not a correct one, to allow as many pounds differ- ence as there are tens of pounds in the weight of the lead button obtained. Thus, with a lead contents of thirty and seventy pounds, the difference in the separate assays might amount to three and seven pounds respectively. 2. FUSION WITH BLACK FLUX. A modification of the preceding method of assaying, which is sometimes employed, consists in using, instead of the potassium carbonate, an equal quantity of black flux, FUSION WITH METALLIC IRON. 511 or indeed of argol, or in mixing a few per cent, of pow- dered charcoal or flour with the potassium carbonate, or in replacing it in part by argol. Too great an addition of carbon diminishes the fusibility of the mass, and hinders the flowing together of the separated particles of lead. By using argol the operation lasts longer, because the mass remains pasty until most of the tartaric acid has been decomposed ; but a greater product of lead is obtained. The chemical reaction during the operation is thereby modified so that the carbon of the black flux exerts an influence upon the potash, and partially reduces it to potassium ; the potassium, thus set free, works now, as before, upon the galena. The latter is thus, without the influence of the air, more completely decomposed than by potassium carbonate alone, and the smelting is, therefore, conducted in covered crucibles (fig. 61) in the wind fur- nace. But since there is also potassium sulphide formed, and this dissolves lead sulphide, it is more advisable, for the completest possible separation of the lead, to perform the smelting in open crucibles in the mulfle, in order to allow the atmospheric oxygen to work at the same time on the assay. The product of lead from pure galena does not generally exceed 76 to 79 per cent. At the Victor-Frederick smelting works in the Hartz, one centner (= one hundred and fourteen assay pounds) of galena is mixed with three or four times as much black flux, and with pyritic ores ten pounds of borax-glass are added. The mixture is covered with sodium chloride, heated for about twenty-five minutes in the muffle furnace with a charcoal fire, and then, after the mouth of the muffle has been opened for about five minutes, taken out of the furnace. 3. FUSION WITH METALLIC IRON. Schlutter and many of the older assayers were aware that iron would desulphurise galena, and ever after advised the addition of a certain quantity of that metal to the different fluxes which they used in lead assays ; but it was 512 THE ASSAY OF LEAD. at the practical School of Mines, at Montiers, that iron was first employed alone. The process is extremely convenient and easy of exe- cution ; it always succeeds, and requires no troublesome attention. The fusion takes place quietly, without frothing or bubbling ; and as the whole substance employed requires but little space, very small pots may be employed, or a very large quantity assayed. But this process can only be employed for pure galenas, or those which contain at most a few per cent, of gangue. When galena is heated with iron, the metal is trans- formed into protosulphide, whence it follows, that to de- sulphurise galena 22 '6 per cent, is required ; but expe- rience has shown that it is better to employ a little more, and 30 per cent, can be used without inconvenience. The iron employed ought to be in the state of filings, or wire cut very small. The mixture is placed in a crucible, which is three-fourths filled ; the whole is covered with a layer of salt, sodium carbonate, or black flux, and exposed to a full red heat. After the flux is perfectly fused, the pot may be cooled and broken, and a button is obtained, which at first sight has a homogeneous aspect, but on being struck with the hammer separates into two distinct parts. The lower part is ductile lead : the upper, a very brittle matte, of a deep bronze colour, and slightly mag- netic. Pure galena yields, by this process, 72 to 79 per cent, of lead, so that there is a considerable loss, which loss is entirely due to volatilisation. Berthier says that it does not appear possible to avoid this loss, which amounts from 6 to 13 per cent., giving as a reason that it is probable galena begins to sublime before it arrives at the proper heat for decomposition. Antimonial galenas, or galenas mixed with iron pyrites, may be assayed in the same manner ; but then a sufficiency of iron must be added to reduce the antimony to the metallic state, as well as to reduce the iron pyrites to the minimum of sulphurisation. If the galena be mixed with blende, the greater portion remains in the slag, because it is only decomposed by iron at a very high temperature. FUSION WITH METALLIC IRON. 513 Blende being infusible by itself, much diminishes the fusibility of the mattes produced ; and if it exists in very large quantity, it can even hinder their complete fusion ; in which case some iron protosulphide and metallic iron must be added to the assay, to make the slag more fusible. All minerals are at a minimum of sulphurisation when existing in mattes from metallurgical works ; therefore much less iron may be used in their assay than if they were pure ores. In very rich lead mattes, in which the lead exists as subsulphide, from 10 to 12 per cent, is sufficient. A small excess of iron may be employed with- out inconvenience ; but if a larger proportion be added than is necessary to execute the desulphurisation, the matte contains some iron in the metallic state, and loses its liquidity, and in consequence retains some globules of lead. The usual mode of assaying lead ores (galena) in the lead mills is by a modification of this process : in lieu of placing the ore in an earthen crucible, and adding nails or filings, a given weight of the ore is projected into a red- hot wrought- iron crucible, which is kept in the fire for about a quarter of an hour, or until all the galena seems decomposed. The lead thus reduced is poured into a mould ; and if the scoriaceous matter be not well fused, the iron crucible is returned to the fire and heated still more strongly, and any lead that may be separated is poured into the mould and weighed with the rest. This is a very rude and imperfect process, and gives only tolerable results with pure galenas, but is quite unsatisfac- tory with those containing much earthy matter, as not above half the lead is obtained, owing to volatilisation and exposure to the air, and the loss of globules in the slag. This process succeeds much better when a flux is added ; this may be argol, or sodium carbonate, or a mixture of both (see next process). L L 514 THE ASSAY OF LEAD. 4. FUSION WITH SODIUM CARBONATE OR BLACK FLUX, AND METALLIC IRON. When galena is heated with an alkaline flux, out of contact of air, the slag contains a double sulphide of lead and the alkaline metal employed : if iron be thrown into this fused mixture metallic lead separates, and the iron combines with the sulphur formerly combined with the lead, and the slag will contain a double alkaline sulphide, containing iron sulphide instead of lead sulphide, thus : PbS + K 2 S + Fe - Pb 2 + FeSK 2 S. Any earthy substances the ore may contain will be dissolved by the alkaline flux, without very much impair- ing its fluidity. All these facts being considered, it may be readily seen that the assay of all earthy bodies contain- ing lead sulphide may be made in this manner, with as much accuracy as this method of assay can be capable of. Either sodium carbonate or black flux may be employed as the alkaline reagent, and more of either of those substances must be employed, in proportion to the in- creased quantity of earthy matters the ore contains. Two parts are nearly always more than sufficient for poor ores, and are best for all cases, because an excess of flux does not diminish the yield of lead ; nevertheless it is sometimes convenient to employ, for the latter class, but half a part. As to the iron, it is employed only to separate that part of the lead which has been dissolved in the state of sulphide by the alkali, but not decomposed ; so that much less may be employed than is necessary for the decomposition of the w r hole amount. Experiment has shown that the maximum amount of lead from pure galena may be obtained by the use of the following mixtures : 2 parts of black flux, or sodium carbonate, and 10 to 12 of iron. 1 part of black flux, or sodium carbonate, and 20 of iron ROASTING AND REDUCING ASSAY. 515 \ a part of black flux, or sodium carbonate, and from 25 to 30 of iron. When black flux is employed, and the iron is in the state of filings, it would be inconvenient to employ too much of the latter, especially if the assay were heated very strongly, because the button of lead might be con- taminated with iron ; but when sodium carbonate is used with small iron nails instead of filings, the excess of iron is not inconvenient, but rather useful, because the desul- phurisation is certain to be complete.* The following changes take place in both cases. That portion of iron filings mixed with the sodium carbonate which has not been sulphurised is brought to the state of oxide by the carbonic acid of the alkaline carbonate, and remains combined or neglected in the slag ; so that the proportion of iron is never too great, and never becomes mixed with the lead. When black flux is employed, the same oxidation does not take place, on account of the presence of carbonaceous matter, so that the portion of filings not combined with sulphur, and which is merely held in suspension in the flux, passes through it with the globules of lead to the bottom of the crucible ; but if, instead of filings, small nails are employed, they only suffer corrosion at their surface, without change of form or soften- ing, and after the assay are found fixed in the surface of the button of lead, so that they can be detached very readily, and, according to Berthier, without loss of lead. This, however, we have found no easy task, and have always sustained a notable loss. 5. ROASTING AND REDUCING ASSAY. This mode is preferable for ores and substances which contain a considerable quantity of foreign sulphides, or arsenides and antimonides, and a greater or less amount of * The French assayers use a piece of plate iron in the shape of a horse- shoe, which is moved about in the melted mass until no more globules of lead attach themselves to it. In Germany a mass of iron wire is used. What iron is not consumed by the assay is found still hanging together in a single mass. L L 2 516 THE ASSAY OF LEAD. earthy matter. It is used in many parts of Germany (at the Unterhartz), and is described by Kerl thus : Two assay centners of ore, matte, &c., are heated at first at a low red heat in the muffle, on a roasting dish that has been previously rubbed with chalk. After ten or fifteen minutes they are taken out of the furnace, then again roasted at a moderate temperature for ten or fifteen minutes with frequent turning of the dish. The assay is then once more taken from the furnace, allowed to cool, rubbed up in an agate mortar, and again roasted for half an hour, whereupon it is taken out of the furnace ; tallow is added while it yet glows, and it is again brought to a strong red heat. The rubbing up and calcining with tallow are repeated several times more, and when afterwards the assays have been exposed for two hours continuously to a strong red heat, with the mouth of the muffle almost en- tirely closed, if no more sulphurous acid vapours escape, the roasting is considered as finished. This lasts from six to twelve hours. The roasted sample is then portioned out with the balance, each portion mixed with three or four parts of black flux and an equal quantity of borax and glass, placed in a small crucible covered with sodium chloride, furnished with a little piece of coal as a cover, and smelted in the wind furnace for about a quarter of an hour after the fire is well ignited. Assays that have worked well give nearly equal malleable buttons that do not contain matte, and a black uniformly fused slag. The purpose of the roasting is to convert the metallic sulphides, arsenides, and antirnonides into oxides. But since, in the process, sulphates, antimoniates, and arseniates are produced, we seek to destroy these by repeated cal- cining with tallow (see above), instead of an intermixture of coal-dust or flour. By melting the roasted assay with its charge at not too high a temperature, the lead oxide is reduced, and the foreign oxides and earths contained in the sample are, by the aid of the potash in the black flux, as well as of the borax and glass, slagged off. If sulphates or sulphides have remained behind in the roasted ore, they will in the smelting be partially desulphurised by the ASSAY WITH BLACK FLUX. 517 action of the oxides, especially of the iron oxide. An addition of metallic iron would in this respect be advan- tageous. The roasting is a lengthy process, and one which causes a not unimportant loss of lead. If it is not done tho- roughly, then in the reduction smelting, sulphur salts are formed, which always retain lead, as also a plumbi- ferous matte which surrounds the lead button. By the use of too high a temperature in the smelting a great part of the foreign oxides is reduced, and the lead becomes contaminated. The reduction, however, cannot be entirely avoided, even with a rightly conducted temperature. Galena melts less easily than metallic lead if the air is excluded ; but is much more volatile than the latter, and is decomposed by fusion into a higher sulphide which is volatile, and a lower one (Pb 2 S) which remains as a residue. Galena by roasting gives a mixture of lead oxide and sulphate, from which last the sulphuric acid cannot be separated, even at a fusing temperature. Lead sulphate becomes soft by heat, fuses at a bright white heat, and is converted by carbon, with a considerable loss of lead through volatilisation, into lead oxide, metallic lead, or lead subsulphide, acording to the quantity of carbon used and the temperature employed. With lead oxide the sulphate easily fuses together. ADDITIONAL REMARKS ON THE LEAD ASSAY. Comparison of the Different Methods for the Docimastic Esti- mation of Lead in their Application to Various Products. Markus has made the following comparative experi- ments with the methods of assaying lead ores most in use at the Austrian smelting works at Joachimsthal. a. Assay with Slack Flux and Iron. One assay centner (5*7 grammes) of the finely rubbed, sifted, and dried assay substance was mixed with two assay centners of black flux, made of sixteen saltpetre and forty argol, and sixty of borax-glass in a mixing capsule, and put into a clay crucible, on the bottom of which a piece of thick iron wire, 518 THE ASSAY OP LEAD. one inch long and forty centners in weight, had been placed in a vertical position. The crucible charge, covered over with two centners of decrepitated sodium chloride, was smelted in a mineral coal muffle furnace, with the mouth of the muffle closed, and the draught half open, at a mode- rate temperature, the temperature then lowered for six to seven minutes by opening the mouth of the muffle, then the muffle closed again for an equal period, and the final heat then given. The cessation of the low crackling of the assay was now carefully attended to, and this, ceasing after seven to eight minutes, indicated the completion of the assay, The duration of the assay was twenty minutes. b. Roasting and Reduction Assay with Iron. One assay centner of galena was roasted, at first at a low tempera- ture, for about thirty minutes on a roasting dish, and the dish then pushed into the back part of the muffle for six to eight minutes to destroy the sulphates formed. The roasted ore was rubbed fine, intimately mixed with three hundred centners of black flux, and fifty centners of borax- glass, placed in a crucible with a piece of iron at the bottom, weighing twenty centners covered with salt, and smelted as above. c. Roasting and Fusing with Black Flux. One centner of the roasted ore was smelted as before with three hun- dred centners of black flux and fifty centners of borax, but without iron. The results obtained proved 1 . That with all those products which contain tolerably pure lead sulphide, especially with high percentages, the iron assay, a, gives in a remarkably predominant degree the most lead (as high as 96 per cent, of all the lead present). 2. With impure lead ores, which contain more foreign sulphides, the assay a gives likewise the highest per- centage, though the assays b and c give only a few per cent. less. 3. If foreign sulphides are present in predominant ASSAY WITH SULPHURIC ACID. 519 quantity, the methods of b and c give a slightly higher percentage than that of a. LevoVs Fusion Assay with Potassium Ferrocyanide and Cyanide. According to Levol, the method of assaying galena for its lead by smelting it with black flux and iron is defective in two respects. First, it is difficult to choose precisely the quantity of iron required for the reduction of the lead, and a lack or excess of it either gives too little lead or a button containing iron ; and second, in order that the reaction may be complete and the lead unite to a button, we are compelled to use a very high temperature, at which lead volatilises. The first defect can indeed be removed by the use of iron crucibles, but these are easily rendered unserviceable, and require a pouring out of the fused mass, and then globules of lead are apt to remain in the slag. By the use of a mixture of fifty parts of potassium cyanide and one hundred of anhydrous potassium ferro- cyanide to one hundred of galena, the loss of lead diminishes to from 2 to 2^ per cent., probably in con- sequence of the easy fusibility of the mixture, and the extremely fine division of the iron in the potassium ferro- cyanide. With antimonial galena this process is not applicable, as the antimony is reduced and goes into the lead. Potassium cyanide alone, gives, by reason of the greater quantity of metallic sulphide which it retains, a smaller product of lead. 6. ASSAY WITH SULPHURIC ACID. The assay sample is rubbed as fine as possible. A suitable quantity of it is then weighed out for an assay, and boiled with four to eight times its weight of oil of vitriol until all is decomposed. All excess of sulphuric acid is then evaporated in a porcelain capsule, under a flue with a good draught, and the mass carried to dryness. Boiling 520 THE ASSAY OF LEAD. sulphuric acid decomposes the sulphides, changing iron, copper, nickel, zinc, &c., into salts which dissolve readily in water, and also at the same time changing the lead sulphide into sulphate, which in water, especially when cold, and containing free sulphuric acid, is practically in- soluble. The composition of the ore is in general ascer- tained by first heating it with nitric acid or aqua regia, and then with the addition of sulphuric acid, evaporating to dryness. The dry mass, when cold, is moistened with a small quantity of sulphuric acid, then cold water ; it is afterwards, by the aid of a small brush, brought without loss upon a small filter, and washed with cold water until the filtrate is colourless. Unnecessary prolonging of the washing is to be avoided, for lead sulphate is not absolutely insoluble. The filter, with its contents, is dried in the funnel, until it can be easily taken out of it without tear- ing. It is now put immediately into the clay crucible in which the lead sulphate is afterwards to be reduced, and this is placed in a very gentle stove warmth. (Some potassium carbonate may be first poured into the bottom of the crucible.) When completely dry, the crucible with the cover laid over it is very gently heated, so that the filter carbonises, which very soon happens, as the free sulphuric acid is not completely soaked out. The filter is now stirred up with a little rod, black flux or potassium carbonate with coal-dust and iron are introduced into the crucible, and intimately mixed with the lead sulphate and the rest of the insoluble residue. About four or five times the volume of the whole residue is taken of black flux, and the assay is further treated as prescribed in the portion which follows upon the assaying of lead sulphate. In this way the lead is concentrated, and the foreign sulphides, which were specified above as the cause of the failure of the assay in such cases, completely removed. The result obtained in this way is satisfactory, and de- serves the same confidence as one obtained in favourable circumstances by the ordinary lead assay from an ore with a medium or high percentage of lead. ASSAY OF GALENA. 52J 7. ASSAY OF GALENA IN THE WET WAY. When in contact with metallic zinc, galena is readily decomposed by acids. Even oxalic, acetic, and dilute sulphuric acids are capable, when hot, of decomposing galena metallic lead being deposited and sulphuretted hydrogen gas set free, while with hydrochloric acid the decomposition is peculiarly rapid and complete. Galena is easily decomposed, also, even in the cold by dilute nitric acid in presence of zinc ; but the reaction differs in this case from that just described not metallic lead but free sulphur is deposited, while lead nitrate goes into solution. The reaction with zinc and hydrochloric acid has been employed with advantage by Mr. F. H. Storer, Professor of Chemistry in the Massachusetts Institute of Technology, for assaying galena, particularly the common American variety, which contains no other heavy metal besides lead. The details of the process are as follows : Weigh out 2 or 3 grammes or more of the finely powdered galena. Place the powder in a tall beaker, together with a smooth lump of pure metallic zinc. Pour upon the mixed mineral and metal 100 or 150 c.c. of dilute hydrochloric acid which has been previously warmed to 40 or 50 C. ; cover the beaker with a watch-glass or broad funnel, and put it in a moderately warm place. Hydrochloric acid fit for the purpose may be prepared by diluting 1 volume of the ordinary commercial acid with 4 volumes of water. For the quantity of galena above indicated, the lumps of zinc should be about an inch in diameter by a quarter of an inch thick ; they may be readily obtained by dropping melted zinc upon a smooth surface of wood or metal. The zinc and acids should be allowed to act upon the mineral during fifteen or twenty minutes in order to insure complete decomposition. Any particles of galena which may be thrown up against the cover or sides of the beaker should of course be washed back into the liquid. It is 522 THE ASSAY OF LEAD. well, moreover, to stir the mixture from time to time with a glass rod. When all the galena has been decomposed, as may be determined by the facts that the liquid has become clear, and that no more sulphuretted hydrogen is evolved, decant the liquid from the beaker into a tolerably large filter of smooth paper, in which a small piece of metallic zinc has been placed. Wash the lead and zinc in the beaker as quickly as possible with hot water, by decantation, until the liquid from the filter ceases to give an acid reaction with litmus paper ; then transfer the lead from the beaker to a weighed porcelain crucible. In order to remove any portions of lead which adhere to the lump of zinc, the latter may be rubbed gently with a glass rod, and after- wards with the finger or a piece of caoutchouc, if need be. Wash out the filter into an evaporating dish, remove the fragment of zinc, and add the particles of lead thus collected to the contents of the crucible. Finally, dry the lead at a moderate heat in a current of ordinary illuminating gas, and weigh. The lead may be conveniently dried by placing the crucible which contains it in a small cylindrical air-bath of Eammelsberg's pattern, provided with inlet and outlet tubes of glass, reaching almost to the bottom of the bath. When the process is conducted as above described, the lead undergoes no oxidation ; hence there is no occa- sion for igniting the precipitate in a reducing gas. The precipitate needs only to be dried out of contact with the air. If desirable, the sulphur in the galena can be esti- mated at the same time as the lead, by arresting the sulphuretted hydrogen in the ordinary way. If the mineral to be analysed is contaminated with a silicious or other insoluble gangue, the metallic lead may be dissolved in dilute nitric acid after weighing, and the insoluble impurity collected and weighed by itself. In the case of galenas which contain silver, antimony, copper, or other metals, precipitable by zinc, the proportion of each metal must be estimated by assay or analysis in the usual ASSAY OF SUBSTANCES OF THE SECOND CLASS. 523 way, after the total weight of the precipitated metals has been taken. Besides galena, almost any of the ordinary lead com- pounds may evidently be assayed by the method above described. For example, metallic lead may be precipitated quickly and completely from the sulphate, chromate, nitrate, oxide, and carbonate and with peculiar ease from the chloride by means of zinc and hydrochloric acid. The method would also furnish an easy qualitative test for the detection of baryta in white-lead. When applied to the analysis of lead nitrate, it would probably be best to decompose the nitrate by means of a solution of sodium chloride before adding the zinc and hydrochloric acid. In all these cases the decomposition of the lead salt by the zinc is so complete that no trace of coloration is produced when sulphuretted hydrogen is added to the liquid decanted from the metallic lead. Attempts to estimate sulphur and lead in the same portion of galena, by means of the reaction of zinc and dilute nitric acid above described, have not given satisfac- tory results. The free sulphur obtained by treating galena with zinc and ordinary nitric acid, diluted with three, four, and five volumes of water, always retains a small quantity of lead, while a certain amount of sulphuric acid is found in the clear liquid. It is, in short, well-nigh or quite impossible to avoid the secondary reactions between zinc and lead nitrate, and between sulphuric and nitric acid, which set in as soon as, or just before, the last traces of the galena have been decomposed. CLASS II. Assay of Substances of the Second Class. The assay of these substances is very simple indeed. Litharge, minium, lead carbonate, &c., may be assayed by simple fusion with carbonaceous matter ; but when the operation is thus conducted, loss of lead is sustained : it is therefore better 524 THE ASSAY OF LEAD. to add some flux which will readily fuse, and allow the globules of reduced lead to collect into one button. No flux fulfils this condition better than a mixture of sodium carbonate and argol, which is to be intimately mixed with the assay. The following is the best mode of procedure : To 200 grains of the finely pulverised substance add 100 grains of argol, and 300 of sodium carbonate, and inti- mately mix : place the mixture in a crucible which it about half fills, and cover with a layer of common salt about J inch thick ; submit the crucible to a very gradu- ally increasing temperature, keeping the heat at low redness for about a quarter of an hour ; then urging it to bright red until the contents of the crucible flow freely ; take it from the fire and shake, tap it as directed in the copper assay, and either pour the contents into the mould or allow to cool in the crucible. If the operator be pressed for time, the mould may be used, but it is recom- mended to allow the assay to cool in the crucible, for unless the operator be very careful, and have had some considerable practice, he is very liable to lose a small .quantity of metal in the pouring. After the contents of the mould or crucible, as the case may be, are cold, the lead may be separated from the slag by repeated gentle blows from the hammer : if any of the slag or crucible adhere to the button, the latter may be readily freed from it by placing the button between the finger and thumb with its edge on the anvil, and then gently hammering it. The lead will be so altered in shape under the hammer that the slag or crucible readily falls off; and by continu- ing the process the whole may be removed. The cleaned button may then be hammered into a cubical form, and is ready for weighing. In the assay of lead great care must be taken in the management of the temperature, as lead is sensibly volatile above a bright red heat, even when covered with flux, and still more so if any portion be uncovered from want of sufficient quantity of flux ; neither must the assay remain in after the flux flows freely, for a loss may thereby occur from oxidation, by decomposition of sodium carbo- ASSAY OF SUBSTANCES OF THE SECOND CLASS. 525 nate, as explained in the reduction of copper ores and the copper-refining process. For the rationale of this mode of assay, refer to page 199, which explains the decomposition of lead oxide, with the production of metallic ]ead, carbonic acid, and water, by the agency of a substance, like argol, containing both carbon and hydrogen. Cupel bottoms, some lead fumes, and silicious slags require a modified treatment in the assay, as the sub- stances mixed with the oxide of lead (more particularly bone-ash in the cupel bottoms) are very infusible ; and if the flux already mentioned as applicable to the other matters belonging to this class were employed, a very high temperature would be necessary ; and as lead, as already stated, is sensibly volatile above a bright red heat, an evident loss of that metal would be the result. Cupel bottoms may be thus assayed : 400 grains of the finely pulverised bottoms to be mixed with 200 grains of argol, 400 grains of sodium carbonate, and 400 grains of pulverised fused borax ; the mixture is placed in a crucible as already directed, covered with salt, and the fusion con- ducted as just described. Lead fumes and silicious slags require only half their weight of fused borax, with 200 argol, 400 sodium car- bonate, and 400 substance (fume or slag), covered with salt. The addition of the borax, which is a most powerful flux, causes the fusion of the assay to take place almost as readily with the last-named refractory substances as with the former easily fusible and reducible matters. The assay, however, is rather more subject to ebullition or boiling over the sides of the crucible ; hence it must be carefully watched ; and the instant it appears likely to do so the crucible must be removed from the fire, gently tapped on the furnace top, and when the effervescence has subsided returned to the furnace, and this operation repeated until the fusion proceeds tranquilly. The lead obtained in these assays, if the ore or sub- stance contained any foreign metal, is never pure ; if silver, 526 THE ASSAY OF LEAD. copper, tin, or antimony be present, the whole of either of these metals will be found alloyed with the lead pro- duced ; but if the ore contains zinc, and it be heated sufficiently, only a trace remains ; nevertheless the zinc carries off with it a considerable quantity of lead. The following experiments will show what an influence the presence of zinc has upon the return of lead : 100 parts of litharge, 100 parts of zinc oxide, 300 parts of black flux, were fused together, and 84 parts of lead were the result. 100 parts of litharge, 100 parts of zinc oxide, 600 parts of black flux, were fused together, and but 70 parts of lead were pro- duced instead of 90, which the pure litharge ought to have given. Hence it will be seen that the more zinc is reduced, the more lead is volatilised. If iron oxide be present in the assay, it is reduced, but it remains in suspension in the slag, and the lead does not contain a trace when it has not been too strongly heated. If the assay be made at a very high temperature, the iron may be fused, and then the lead will be ferruginous ; this may be ascertained by means of the magnet. A similar result was obtained by many assayers, who thought for a long time that lead and iron could thus combine together ; but by careful examination it is easily ascertained that the ferruginous buttons are but mechanical mixtures of lead and iron in grains. Indeed, by careful hammering, nearly all the iron may be removed from the lead, so that it loses its magnetic properties. The manganese oxides, when mixed with the ore, are changed into protoxide, which remains in the flux, and is not reduced. Wet Assay of Ores of the Second Class. Pulverise the substance very finely, and to 100 grains placed in a flask add one ounce of nitric acid diluted with two ounces of water (if minium be the substance to be analysed, it ASSAY OF SUBSTANCES OF THE THIRD CLASS. 527 must be first heated to redness, so as to reduce the whole of the lead it contains to the state of protoxide), and gently heat, gradually raising the temperature to the boiling- point : when all action seems to have ceased, pour the contents of the flask into an evaporating basin, and evapo- rate to dryness with the precautions directed in the analysis of iron ore. Allow the dry mass to cool, add a little dilute nitric acid, gently warm for an hour, then add water, boil and filter. The whole of the lead now exists in the solu- tion as nitrate : thus, say lead carbonate has been the substance under analysis, then PbO,C0 2 + N0 5 =PbO,N0 5 + C0 2 . To the filtered solution containing the nitrate as above, add solution of sodium sulphate, or dilute sulphuric acid, until no further precipitation takes place ; insoluble sul- phate of lead will now be thrown down : this must be allowed to completely subside by standing in a warm place ; and when the supernatant liquid is quite bright, the sul- phate may be collected on a filter, washed, dried in the water-bath, and weighed. It contains 68-28 per cent, of metallic lead. The decomposition of the lead nitrate by sodium sul- phate may be thus expressed : PbO,N0 5 + JSTa 2 0,S0 3 =PbO,S0 3 + Na 2 0,N0 5 . Estimation of lead by standard solution will be described at the end of this chapter. CLASS III. Assay of Substances of the Third Class. In the assay of bodies belonging to this class a reducing agent must be employed ; but if that alone be used, the sulphates and arseniates produce sulphides and arsenides, and not pure lead. The action of another reagent is therefore necessary, in order to deprive the lead of the sulphur and arsenic with which it is combined. 528 THE ASSAY OF LEAD. There are two reagents employed for the sulphates the alkaline carbonates and metallic iron ; but for the arseniates and arsenites iron must be employed, because the alkaline carbonates have no action on the arsenides. In all cases black flux is employed ; this furnishes a reducing agent for the oxides, and a flux for the earthy matters. Iron is added when the arsenites or arseniates are assayed ; but that metal may either be employed or not when the sulphates are operated upon. It is, how- ever, always better to use it. When a mixture of black flux and iron is employed, the assay is made in exactly the same manner as that of the sulphides (large nails are preferable whenever the use of iron is indicated in a lead assay). With the sulphate, the iron sulphide formed combines in the slag with the alkaline sulphide ; but it is not so with the arseniates and arsenites. The arsenide produced mixes neither with the lead nor the slag, but gives rise to the formation of a brittle matte which adheres slightly to the button of lead. When only black flux is employed, either of the two following processes may be adopted : First, the ore can be fused with four parts of common black flux ; then, as in the case of sulphides, the excess of carbon determines the formation of a large quantity of an alkaline sulphide, and consequently produces a desulphurisation of the lead. Secondly, it may be fused with such a proportion of black flux, containing only the requisite proportion of carbon to reduce the lead oxide, or with an equivalent mixture of sodium carbonate and charcoal. Pure lead sulphate fused with one part of sodium carbonate and 4 per cent, of char- coal gives 66 of lead ; but in order to employ this method the richness of the ore must be known, and the dry way is then useless, excepting for the estimation of the silver these substances always contain. Wet Assay of Substances of the Third Class. These are treated in precisely the same manner as those of the preceding class. In treating lead ores with nitric acid a loss generally results from the formation of insoluble lead sulphate. The ASSAY OF SUBSTANCES OF THE THIED CLASS. 529 solubility of these salts in sodium hyposulphite renders it possible to avoid this inconvenience. After treatment with nitric acid, J. Grceme (Bulletin de la Soc. Chimique de Paris, November 5, 1873) proposes to exhaust the residue with boiling water, until the soluble salts and the acid are completely eliminated. It is then digested in the cold with a concentrated solution of sodium hyposulphite. After this treatment has been twice or thrice repeated the residue is exhausted again with water ; the lead is then precipitated from the filtrate by sulphuretted hydrogen or ammonium sulphide ; to facilitate the agglomeration of the precipitate and its washing, it is heated in the water-bath. The sul- phide is then converted into sulphate, and its weight added to that of the sulphate obtained directly. F. Maxwell Lyte thinks the following process for the management of the assay of lead in ores will be found convenient, particularly where, as is often the case, the lead to be estimated is mixed as sulphate with the matrix insoluble in acid. He dissolves the sulphate or chloride, as the case may be, in ammonium acetate, makes the solution as neutral as possible, and estimates the lead by a standard solution of bichromate (a half-decinormal solution answers well) with a silver nitrate indicator. A. Mascazzini, previously to reducing the galena or other lead ore to the metallic state, converts the lead present in the ore into sulphate by igniting it in a porcelain crucible with ammonium sulphate, after which the ore is treated in the usual manner, The flux preferred by the author is that recommended by Plattner, consisting of 13 parts of potassium carbonate, 10 of dry sodium car- bonate, 5 of previously fused borax, and 5 of well-dried starch. To detect galena in mixtures, M. E. Jannesay throws upon coarsely powdered galena a fragment of potassium bisulphate, which gives a distinct evolution of sulphuretted hydrogen. If the two bodies are ground , together the odour becomes almost insupportable. Potassium bisul- phate kept in fusion for half an hour produces the same M M 530 THE ASSAY OF LEAD. effect, perhaps with less intensity. Sulphuric acid, mixed or even heated with galena, does not give rise to a sen- sible disengagement of sulphuretted hydrogen. Blende gives a sulphydric odour, but less intense. Antimony, iron, mercury, and silver sulphides give off no sensible odour. Boulangerite, zinbrenite, and in general the sul- phides in which lead and sulphur do not form an isolated combination, do not evolve sulphuretted hydrogen with the potassium bisulphate. CLASS IV. ALLOYS OF LEAD. ASSAY WITH SULPHURIC ACID. No docimastic assay is known for exhibiting the lead isolated from its alloys. In individval cases a serviceable result may be attained, if the metal with which the lead is combined be estimated, and its quantity then deducted. This method is, however, in general the more unreliable the smaller is the quantity of lead, or when the lead is alloyed with several metals ; so that then the quantity of lead can often only be estimated by the partial or com- plete aid of the wet way. For many products (e.g. crude lead, hard lead con- taining antimony or arsenic plumbiferous copper, &c.) the assay with sulphuric acid described on page 519 is suitable. One assay centner of the substance is decom- posed by nitric acid or aqua regia, then, with the addition of sulphuric acid, evaporated to dryness, and the dry mass treated as above directed. If the residue consists only of lead sulphate, it can be brought upon a weighed filter, and from the weight of the residue after drying, the amount of lead may be calculated. 100 parts lead sulphate contain 68-33 parts lead. DOCIMASTIC ESTIMATION OF LEAD. 531 ESTIMATION OF LEAD BY MEANS OF STANDARD SOLUTIONS. 1. FLORES DUMONTE'S METHOD. Estimation of Lead by means of Standard Solutions. This process is due to M. Flores Dumonte, and may be thus described : This mode of analysis is analogous to that proposed by Pelouze for the estimation of copper ; advantage is taken of the fact that oxide of lead is soluble in caustic potash in the same manner that copper oxide is soluble in ammonia ; and from either solution the respective metal is precipitated by means of a standard solution of sodium sulphide. The solution of sodium sulphide may be conveniently made by dissolving one ounce of sodium sulphide in one quart of water, and estimating how much of it is necessary to precipitate twenty grains of lead. To this end weigh off twenty grains of lead, dissolve them in nitric acid, di- lute with water, and add excess of caustic potash until the oxide of lead first thrown down is completely dissolved. The solution must now be heated to ebullition, and the sodium sulphide gradually added from the burette ; at each addition a black precipitate of lead sulphide falls. The liquid is then boiled for a short time, by which means . it brightens ; more sodium sulphide is then added, and the whole again boiled, and these operations alternately con- tinued until no further coloration or blackening is pro- duced by the last drop of sulphide. The number of divisions used is then read off, and the calculation made as in Pelouze's copper assay, substituting lead for copper. Having thus standardised the solution of sodium sul- phide, the assay of a sample of ore may be thus made : If the ore belong to the first class, dissolve it in dilute nitric acid and evaporate to dryness ; to the dry mass add excess of caustic potash solution, and boil ; after about a quarter of an hour's ebullition, filter and throw down the lead as directed, with the standard solution ; from the amount used calculate the quantity of lead present ; if the ore be of the second or third class, treat with strong nitric M M 2 53-2 THE ASSAY OF LEAD. acid and sodium carbonate, as already directed. The lead carbonate so produced may be dissolved in either nitric or acetic acid, and to the solution thus obtained add caustic potash, &c. 2. SCHWARTZ'S METHOD. Dissolve 14*730 grammes of pure potassium bichromate in sufficient water to form one litre. One cubic centimetre of this solution precipitates 0-0207 gramme of lead. In the estimation of pure lead a certain quantity of it should be dissolved in a minimum of nitric acid, the solution diluted with water, carefully neutralised with ammonia or sodium carbonate, excess of sodium acetate added, and the solution precipitated by the potassium bichromate solution. When the precipitation approaches its end, or when the precipitate commences readily to subside, some drops of a neutral solution of silver nitrate are deposited on a porcelain plate, and the potassium chromate solution only added by two or three drops at a time to the liquid under examination ; after each addition the whole is well stirred, allowed to subside, and a drop of the clear supernatant liquor added to one of the drops of the silver solution. As soon as the potassium bichromate is in excess the two drops form a red colour, while the precipitated lead chromate has no effect on the silver test, but simply floats on the top as a yellow precipitate. Should the solution assume a yellow colour before the silver reaction has commenced, it would indicate that not sufficient sodium acetate had been added in the first in- stance, and it w r ould be necessary to add this now, and also a cubic centimetre of a normal lead solution, contain- ing 0-0207 of lead as nitrate. The slight turbidity which first takes place soon goes off, and the operation may be proceeded with as before. One cubic centimetre must naturally, in such instance, be deducted from the amount of chrome solution, on account of the extra addition of lead. Bismuth alone seems to interfere with the reaction, BUISSON S VOLUMETRIC PROCESS. 533 .and behaves very like lead with chromic acid, and if present it requires a different mode of proceeding. The higher oxide of mercury is not precipitated by potassium bichromate, not even in an acetic solution, while the lower oxide is ; and, as it is difficult to peroxidise all the mercury when united with lead, even by long-continued boiling in nitric acid, it becomes necessary to evaporate and calcine the residue till all the mercury is volatilised. To obviate the formation of red-lead, the calcined residue has to be moistened with a few drops of oxalic acid, and again dried and carefully calcined and dissolved in acetic acid ; after this the lead may be estimated as usual. To avoid the above calcinations, the mercury may be precipitated from the nitric acid solution by means of hydrochloric acid, and the liquid boiled till the calomel is converted into the higher chloride. Copper, cadmium, zinc, iron, and cobalt do not in the least interfere with the reaction, provided the iron is per- oxidised. Of the different acids, hydrochloric acid some- what disturbs the last silver reaction, but by using larger drops, and allowing the reaction of silver chloride to go off, we obtain the usual silver chromate reaction. Lead sulphate has first to be converted into the state of carbonate, by boiling with sodium carbonate, when it may be dissolved in acetic acid. Lead phosphate and arsenite, -or other lead salts insoluble in acetic acid, may be dis- solved in nitric acid, and estimated according to the older .method. .< o. BUISSON'S VOLUMETRIC PROCESS FOR ESTIMATING LEAD. This process is based on the precipitation of lead by potassium bichromate and the decomposition of the excess of the reagent used by potassium iodide in a liquid acidulated with sulphuric acid. 0'5 to 1 gramme of the pulverised mineral to be tested is dissolved in dilute nitric .acid. The solution is saturated with potash, and the preci- pitate redissolved in acetic acid. The iron is removed by boiling. To the solution separated from the iron 25 c.c. 534 THE ASSAY OF LEAD. of potassium bichromate is added, and water to bring the volume up to 250 c.c. After standing for some time the solution is filtered through a dry filter-paper. To 100 c.c. of .the clear liquid an excess of dilute sulphuric acid and potassium iodide are added in such a way as to redissolve the iodine set free, and then a few cubic centimetres of solutions starch. By means of a graduated burette, sodium hyposulphite is added until decolourisation of the starch iodide takes place. The difference of standard obtained by treating the bichromate alone, and after precipitation with a known weight of lead, gives a basis for calculating the amount of metal contained in the substance tested. Silver, bismuth, copper, and baryta should be separated from the lead before applying this process. 4. W. DIEHL'S PKOCESS. W. Diehl proposes the following process for the volumetrical estimation of lead ( 4 Zeitschrift fur Analy- tische Chemie '). He employs a ^ normal solution of potassium bichromate, containing 7 '38 grnis. per litre, each c.c. representing 0*01035 grin, of lead, and a solution of sodium hyposulphite, containing 4 to 5 grms. per litre. To determine the relation between these two solutions, 20 to 30 c.c. water are mixed with 20 to 25 dilute sul- phuric acid (1 vol. monohydrated acid, and 2 vols. of water) ; a certain excess of sulphuric acid is indispensable, hydrochloric acid being less convenient. The liquid is brought to a boil, and the solution of hyposulphite is added drop by drop. The solution becomes gradually paler in colour. Towards the end, after the addition of a few drops, it is let boil up again. The end of the reaction may generally be distinguished by the liquid turning per- fectly colourless, a result occasioned by a single drop. In order to judge of the colour, the flask towards the end of the operation may be set in a porcelain capsule. When very large quantities of bichromate are used the liquid does not become perfectly colourless, but slightly greenish. In assaying ores in this manner they are dissolved DIEHL'S PKOCESS. 535 in aqua regia and dilute sulphuric acid, the solution con- centrated till the sulphuric acid begins to evaporate, diluted with water, boiled to dissolve ferric sulphate, let cool, and filtered through a smooth filter, washing with water containing sulphuric acid. To the residue in the flask as little as possible of which is thrown upon the filter is added about 15 c.c. of a solution of neutral am- monium acetate, and about 50 c.c. of water. The whole is then boiled, and filtered through the same filter, into which a drop of ammonia has been put, into a flask. The same operation is then repeated with 5 c.c. ammonium acetate, and the residue is finally well washed with boiling water, to which a little of the same salt has been added. Thorough washing is necessary, since filters have been found to retain ammonium and lead acetate and tartrate with con- siderable obstinacy. It is then advisable further to wash the filter from its margin downwards with a little boiling dilute hydrochloric acid (1 part hydrochloric acid at sp. gr. 1-12 with 10 parts of water), and then to wash again with hot water. In this manner every trace of lead is re- moved from the filter. A thin filter-paper should be used, and should be washed previously. The solution of lead sulphate in ammonium tartrate thus obtained is then titrated in the cold with potassium bichromate ; with the aid of heat, ammonium acetate dissolves a little lead chromate. The precipitate settles readily if the flask is shaken, and the end of the reaction can be observed to within 0-2 to 0*4 c.c. An excess of at least 2 c.c. of the chromate solution should be added, in order to obviate the solubility of the lead salt. It is advisable in every experiment to take as closely as possible an equal quantity. After thoroughly shaking, it is allowed to stand for half an hour and filtered. If the liquid passes through turbid, a few drops of a solution of sodium acetate, acidulated with acetic acid, are added. If, after all, a little lead chromate passes through the filter, the filtration is repeated. The precipitate is washed four times with cold water, and the solution is acidulated with sulphuric acid and titrated as above. 636 THE ASSSAY OF LEAD. Ammonium acetate is preferable to all other ammonium salts as a solvent for lead sulphate. It is applied in a neutral or faintly acid state. Free ammonia renders the solution turbid. 1 grm. lead sulphate requires 15 c.c. of the liquid acetate for solution. Ammonium tartrate cannot be used. 537 GHAPTEE XII. THE ASSAY OF TIN. THIS metal is always found by the assayer in the state of oxide. Tin Oxide (Sn0 2 ). The appearance of this mineral gives no indication, excepting to an experienced eye, that metallic matter enters largely into its composition ; yet its great density would lead one to suppose such to be the case. Its colour varies from limpid yellowish white to brownish black and opaque, passing from one to the other by all intermediate shades. It usually possesses a peculiar kind of lustre which cannot be readily described, but once seen can scarcely be mistaken. It occurs crystallised in square prisms, terminated by more or less complicated pyramids. These crystals, derived from the octahedron, are often macled or hemi tropic, so that they often possess re-entering angles, which is to a certain extent characteristic. The principal varieties are the following : 1. Crystallised Tin Oxide is found in more or less voluminous crystals of the colour and form as above. 2. Disseminated Tin Oxide. This variety occurs in grains of various sizes, sometimes so small as not to be visible to the naked eye. It is found in the primitive rocks. 3. Sandy Tin Oxide forms pulverulent masses often of great extent ; in appearance it is merely a brown sand. 4. Concretionary Tin Oxide, Wood Tin. This variety occurs in small mamellated masses, the fibrous texture of which resembles that of wood ; hence the name. 538 THE ASSAY OF TIN. The following is an analysis of a sample of tin oxide from Cornwall : Tin 77-50 Oxygen 21-40 Iron -25 Silica . -75 The following remarks on tin ore and the minerals which may be mistaken for it are from the pen of Dr. A. Leibius, Senior Assayer of the Sydney branch of the Eoyal Mint. The colour of native tin ore varies from white to pink, ruby-red, grey, greyish-black to black ; it therefore is cer- tainly no very reliable criterion for distinguishing tin ore. A safer characteristic is the weight, or specific gravity, that of tin ore being 6-8 to 7*0. Unfortunately, however, the specific gravity of iron tungstate is nearly the same as that of tin oxide in fact, a little higher, being 7*19 to 7-55. Titaniferou.s iron has a specific gravity of from 4 f 5 to 5-0, and magnetic iron 4-9 to 5*2. Basaltic hornblende and iron silicate have also been mistaken for tin ore ; but the specific gravity of the former being only 3-1 to 3-4, and that of the latter 3-8 to 4*2, ought to have prevented such mistakes. The colour of the powdered ore forms a much better criterion than that of the unpowdered ore. The powder of good tin ore varies only from whitish-grey to dark drab, while iron tungstate powders reddish-brown, and titaniferous iron black. Most of the minerals appear to be mistaken for tin ore on account of their dark granular pieces having been taken for such ; but Mr. Leibius has seen a sample which con- sists of blackish pieces with about 50 per cent, of small indistinct crystals of a pink and dark ruby colour, with a few small white crystals. The whole mixture being pretty heavy has certainly at first sight all the appearance of good tin ore. Even on closer inspection, when the darker portion just referred to might have been recognised as an iron compound, the ruby-coloured portion might readily pass muster for tin ore unless chemically examined. On further examination the whole sample was found to be ASSAY OF PURE OXIDE OF TIN. 539> free from tin. It was found to consist 1. Black portion, about 50 per cent, of the sample, having a specific gravity of 4-47, was found to be titaniferous iron. 2. Buby-coloured and dark red pieces, about 50 per cent, of sample, with a specific gravity of 4*57, were found to be zircons or hya- cinths, showing the characteristic property, mentioned in Professor Thompson's excellent ' Guide to Mineral Ex- plorers ' (see ante, pp. 254,293), of becoming completely and lastingly colourless when exposed to heat before the blowpipe. 3. Besides these zircons were found a few small topazes and garnets, and also a small sapphire. The specific gravity of the mixed sample as received was 4-55. As already mentioned, there was actually no tin in the sample, and it forcibly illustrates the necessity of the precaution, in dealing with tin ore, to have it carefully assayed. Assay of Pure Tin Oxide. Pure tin oxide may be very readily assayed in the following manner : Weigh off 400 grains, place them in either a black-lead or charcoal- lined crucible, cement on a cover by means of Stourbridge clay, and place in the fire. The heat should for the first quarter of an hour be a dull red, after which it may be raised to a full bright red for ten minutes, and the crucible removed with care so as not to agitate or disturb the con- tents ; tapping in this case must not be resorted to. When the crucible is cold, remove the cover, and a button of pure tin will result ; this weighed and divided by four gives the percentage. If the operation has not. been care- fully conducted it sometimes happens the tin is not in one button, but disseminated in globules either on the charcoal lining or on the sides of the black-lead pot ; in this case the charcoal on the one hand, or the black-lead crucible on the other, must be pulverised in the mortar and passed through a sieve ; the flattened particles of tin will be re- tained by the sieve, and can be collected and weighed. If any small particles escape the sieve, they may be separated from the lining or crucible by vanning. If a charcoal or black-lead crucible be not at hand, an -540 THE ASSAY OF TIN. ordinary clay pot may be used, but not so successfully, excepting under certain circumstances to be hereafter described. Indeed, in Cornwall the ordinary mode of con- ducting this assay is in a naked crucible, thus : About 2 ounces of the ore are mixed with a small quantity ol culm, and projected into a red-hot crucible. If the ore seems to fuse or work sluggishly, a little fluor-spar is added, and after about a quarter of an hour's fusing at a good high temperature the reduced and fused tin is poured into a small ingot mould, and the slag examined for metal by pounding and vanning. This method never gives the whole of the metal. To effect this, without fear of mis- chance in the assay sometimes occurring, as already de- scribed with both black-lead and charcoal-lined crucibles, it may be thus conducted ; always supposing the oxide to be pure, or nearly so, or at least containing little 'or no silicious matter. To 400 grains of ore add 100 grains of argol, 300 grains of sodium carbonate, and 50 grains of lime ; mix well together, place in a crucible which the mixture half fills, cover with a small quantity of sodium carbonate and 200 grains of borax. Place the whole in the furnace with the necessary precautions, raise the heat very gently, and keep it at or below a dull red heat for at least twenty minutes ; then gradually increase until the whole flows freely. Ee- move the crucible, tap it as for copper assay, and allow to cool. When cold, break it, and a button of pure metallic tin will be found at the bottom, and a flux perfectly free from globules and containing no tin. There is. yet another process, which is more easy of execution ; but the reagent employed is more expensive, not so readily obtainable, and more difficult to keep without decomposing than any of the substances above employed. This reagent has been mentioned, in another part of this volume, as a blowpipe flux, and, in the assay of copper ores by standard solutions, as potassium cyanide. This is the most effective reducing flux for tin ores yet known. It acts by absorbing oxygen to form a compound known as potassium cyanate. ASSAY OF PURE OXIDE OF TIN. 541 The assay, by means of this substance, may be made in ten minutes. This method of estimating the value of tinstone has been frequently practised by the writer, and has uniformly furnished correct results with but little expenditure of time and labour. The method of operating is as fol- lows : The sample, having been carefully selected, is first crushed by the hammer in a steel mortar, and then further reduced to powder in an agate mortar. 100 grains is a convenient quantity to be taken for analysis, and it is always advisable to make two independent experiments upon the same sample of ore, with the view of having a control, and the highest result obtained is that upon which to place reliance, since the error must always be on the side of loss rather than excess. A couple of small Hessian crucibles, of about 3 oz. capacity, are prepared in the first instance by ramming into the bottom of them a small charge of powdered potassium cyanide sufficient to form a layer of about half an inch in depth ; the weighed quan- tities of tin ore are then intimately mixed with from four to five times their weight of the powdered cyanide, and the mortar rinsed with a smull quantity of the pure flux, which is laid upon the top of the mixture. The crucibles are then heated in a moderate fire, or over a gas blowpipe, and kept for the space of ten minutes at a steady fusion ; they are then removed, gently tapped to facilitate the formation of a single button, and allowed to cool. Upon breaking the crucibles the reduced metal should present an almost silvery lustre, with a clean upper layer of melted flux. It is advisable to dissolve the latter in water, in order to be certain as to the absence of any trace of reduced metal or heavy particles of the original ore. There is always contained in the commercial cyanide a sufficient quantity of alkaline carbonate to secure the perfect fusion of the silicious gangue, and other like impurities in the tin ore, but the operator should assure himself of the absence of copper and lead in the ore, either by preliminary treat- ment with hydrochloric acid, in which tinstone is ab- solutely insoluble, or by testing the button of reduced tin 542 THE ASSAY OF TIN. after hammering or rolling for such metallic admixture. We have usually found a minute trace of iron, and some- times gold in the melted buttons, but not so much as to add appreciably to their weight. When worked with ordinary care, this process may be relied upon as giving numbers true to within -J per cent., and we do not know any other method which exceeds this in accuracy and rapidity of execution. The following are a few analytical results taken at random from a number of ores assayed in this manner : Tin per cent. L ~~n. Sample No. 1 . . . . 45-6 45-8 No. 2 .... 57-2 57-6 No. 3 . . . . 68-4 68-7 Assay of Tin Oxide mixed with Silica. Although tin oxide is completely reducible by charcoal or other carbo- naceous matter, yet it has such an affinity for silica, that whenever that substance is present the metal cannot be wholly reduced, excepting at the highest temperature of a wind furnace. The following experiments will show the influence of silica on the return of tin in an assay of oxide of that metal with black flux. Ore . . 100 100 100 100 100 Quartz . 25 66 100 150 300 The first gave 52 per cent, of tin ; the second, 43 per cent. ; the third, 28 per cent. ; the fourth, 10 per cent. ; and the last, nothing. The slags also produced in the treatment of tin ores in the large way give no return with black flux. This mode of assay, however, has been recommended by some, but from the foregoing experiments is proved to be perfectly fallacious : that is, unless the quantity of silica present be very small in comparison to the amount of tin oxide ; and even when the latter is present in four times the quantity of the silica, as in experiment No. 1, a loss of 20 per cent, of tin is sustained. Assay of Tin Ores containing Silica and Tin Slags. It having just been shown how injuriously the presence of silica influences the produce of tin, both in ores and slags, ASSAY OF TIN ORES CONTAINING SILICA. 543 other methods of assay than those just described must be adopted for such substances. These will now be detailed. Tin ores containing silica may be treated by two methods : in the first the silica must be carefully separated by vanning ; if the ore be well pulverised this is the best and most expeditious method. In conducting this assay take 400 or more grains of the pulverised ore, according to its richness (if poor, as much as 2,000 grains may be taken), van it carefully, dry the enriched product, which will, if the operation has been properly conducted, be nearly pure oxide of tin, and assay it as already described for ores containing no silica. The other process of assay may be thus conducted, and is dependent upon the fact that iron displaces tin from its combination with silica : thus, if a compound of tin oxide and silica be heated to white- ness with metallic iron, a portion of the iron oxidises and replaces the tin oxide, which was previously in combina- tion with the silica as a tin silicate, and metallic tin and iron silicate result, the tin so reduced combining with any metallic iron that may be in excess, and the button thus obtained is an alloy of tin and iron, whilst the slag is entirely deprived of tin. In this kind of assay mix 400 grains of the silicious tin oxide with 200 grains of iron oxide (either pulverised hematite or forge-scales will answer this purpose), 100 grains of pounded fluor-spar, and 100 grains of charcoal powder : place the mixture in a crucible and cover with a lid, gradually heat to dull redness, and keep at that temperature for half an hour, then heat to whiteness for another half-hour, and remove the crucible from the fur- nace, allow to cool, and break. The button so obtained is to be treated in the wet way, as hereafter described. The assay of tin slags is conducted in the same manner, or simply by mixing the pulverised slag with 20 per cent, of iron filings, and fusing. Assay of Tin Ores containing Arsenic, Sulphur -, and Tungsten (Wolfram). In the assay of such ores it is neces- sary to remove arsenic, sulphur, and tungsten before attempting to obtain the tin in a pure state by the dry 544 THE ASSAY OF TIN. assay. Ores of tin which contain either one or all of these substances are most common ; hence this mode of treatment will be generally required. Most assayers usually submit the ore to the same mode of treatment which it undergoes on the large scale by cal- cination, or rather roasting, by which the greater part of the arsenical and pyritic matter is removed ; this process fails, however, to remove the whole of these substances, and does not at all affect the tungsten. The following process is therefore preferable, and is founded on the fact that arsenical and other pyrites, as well as iron tungstate (wolfram usually accompanying tin ores), are completely decomposed by nitro-hydrochloric acid (aqua regia) at the boiling temperature, the oxide of tin alone not being affected : Take 400 grains or more of the impure tin sample, place them in a flask, and add 1^ ounce of hydro- chloric acid, and -J an ounce of nitric acid, heat gently for about half an hour, and then boil until the greater part of the mixed acids has evaporated ; the sulphur and arsenic will by this time be converted into sulphuric and arsenic acid, and the wolfram completely decomposed, its iron and manganese having become soluble, and its tungstic acid remaining in the insoluble state with the oxide of tin and any silica that may be present. Allow the flasks and con- tents to cool, add water, allow to settle, and decant, and so on until the water passes off tasteless. The insoluble matter in the flask is now tin oxide, silica, and tungstic acid ; to remove the latter, digest for an hour at a very gentle heat with one ounce of solution of caustic ammonia, with occa- sional agitation ; add water, and van the remainder to separate silica ; nothing remains now but tin oxide, with perhaps a little silica : this is now to be dried and assayed as directed for ores containing little or no silica. If only an approximative assay be needed, it may be accomplished after this treatment by taking the specific gravity of the remaining oxide ; so that all ores of tin may be thus roughly assayed, it being premised that the above operation has been so carefully performed that nothing but tin oxide and silica remain. The specific gravity of ASSAY OF TIN OEES CONTAINING AESENIC, ETC. 545 the thus purified ore is to be taken. All now that is neces- sary to be known is the specific gravity of tin oxide, its percentage of pure tin, and the specific gravity of silica, and a simple calculation gives the result. The following is the formula : Let a represent the specific gravity of tin oxide. b silica. c the mixture left after treatment with acid, &c. w weight of rough tin oxide or mixture left after treatment with acid, &c. x tin oxide. y silica. And y = c (a b) Or in arithmetical form thus : 1. From the specific gravity of the rough tin oxide (mixture of tin oxide and silica) deduct the specific gravity of the silica. 2. Multiply the remainder by the specific gravity of the tin oxide. 3. Multiply the weight of the rough tin oxide by the last product, which will make a second product, which may be called P. 4. From the specific gravity of tin oxide deduct the specific gravity of silica. 5. Multiply the difference by the specific gravity of the rough tin oxide. 6. Take this product for a divisor to divide the above product P : the quo- tient will be the weight of pure tin oxide in the rough oxide, and the quantity of metal can now be readily calculated. The following is an assay worked out in this manner : 400 grains of the ore are treated with nitro-hydrochloric acid and ammonia as above described, washed, and dried. Suppose the dried matter weighs 250 grains. The 250 grains thus obtained are placed in the specific gravity bottle, and the specific gravity is found to be 5*4. Specific gravity of tin oxide (approximate) . * . 6*9 silica' . . 2*6 Sp. Gr. Bough Oxide 5-4 2-8 Weight of Bough Oxide 250 Sp. Gr. Pure Oxide 6-9 4-3 Sp. Gr. Silica 2-6 Sp. Gr. Pure Oxide 6-9 19-32 Sp. Gr. Silica 2-6 Sp. Gr. Bough Oxide 5-4 4830 =208-4 2-8 19-32 4830 4-3 23-22 23-22 N N 546 THE ASSAY OF TIN. 208*4 grains is therefore the weight of pure oxide in the 400 grains of ore. Now tin oxide contains 78-61 parts of pure tin, and 208-4 x 78-61 100 So that 400 grains of rough tin ore contain 163*72 grains of pure tin, and =40-93. The rough sample first operated on contains, therefore, 40*93 per cent, of metallic tin. Mr. J. H. Talbott's Process for assaying Tin in the Pre- sence of Tungsten. Another satisfactory method of assaying tin in the presence of tungsten has been described by Mr. J. H. Talbott. The method is based on the fact already mentioned that tin oxide is reduced by potassium cyanide with great facility ; while tungstic acid undergoes no reduction, even when heated with the cyanide at a high temperature. The tin and tungsten oxides are to be heated in a porcelain crucible with three or four times their weight of commer- cial potassium cyanide previously fused, pulverised, and thoroughly mixed with the two oxides. The mass is kept fused for a short time, when the tin separates in the form of metallic globules, while the tungstic acid unites with the alkali of the potassium cyanate and carbonate present. After cooling, the mass is to be treated with hot water, which dissolves the alkaline tungstate and other salts, and leaves the tin as metal ; this, is to be separated by filtra- tion, washed, dried, and weighed as tin oxide, after oxida- tion in the crucible with nitric acid. The tungstic acid may be estimated by difference, or be precipitated by mercury protonitrate, after boiling the solution with nitric acid to decompose the excess of potassium cyanide present, and then re-dissolving the precipitated tungstic acid by means of an alkali. Estimation of Tin by the Wet Method. There are several methods of effecting this analysis, the chief diffi- culty- being found in the intractable nature of the tin oxide, it resisting the action of all acids. This, however, may be overcome as first shown by Klaproth, who found that very finely levigated tin oxide was soluble in hydrochloric acid ESTIMATION OF TIN IN THE WET WAY. 547 after a prolonged fusion with caustic potash. The following is his process : Fifty grains of the tin ore, reduced to the most minute state of division by levigation or otherwise, are mixed with four times its weight of caustic potash. The best mode of mixing is to place the caustic potash in a silver crucible, add its own weight of water, and apply a gentle heat until the potash is dissolved ; then stir in tin ore, and gradually evaporate to dryness, stirring all the time to prevent loss by spirting, as in the analysis of ironstone : when thoroughly dry, enclose the silver crucible in one of clay, and submit the whole to a dull red heat for at least half an hour ; rather more than less renders the perfect solution of the tin oxide more certain. When cold, act on the contents of the crucible with dilute hydrochloric acid, transfer the liquid and any undissolved matter to a flask, add some strong hydrochloric acid, and boil for half an hour. If at the end of this time any of the tin ore remains unacted on, it must be separated by decantation or otherwise from the solution, dried, again fused with potash, and then treated with hydrochloric acid, in which it will now be found totally soluble. This second operation will not be needed if care has been taken to reduce the ore to the finest possible state of division at first. The solution, however obtained, is to be evaporated to dryness, and when cold treated with a small quantity of hydrochloric acid, allowed to stand for half an hour, then water added, boiled, and filtered : the whole of the tin will pass through in solution as tin chloride, and any silica or tungstic acid that may be present will remain in the filter. If the or$ contains copper, lead, and iron, these metals will also be in solution at all events, the lead partially so ; but if the ore has, previously to its fusion with caustic potash, been treated with aqua regia, as already described, then it will contain tin alone. It is always better thus to separate foreign matters before attempting the solution of the tin, as the after process is thereby simplified. Supposing, however, that the rough ore has been submitted to fusion with potash and then dissolved, the solution must be treated Nur O -> 548 THE ASSAY OF TIN. thus : A bar of zinc must be placed in the solution, which will in course of time precipitate tin, copper, and lead ; when all the metals are thus thrown down the zinc is washed and removed, the precipitated metals well washed and dried. To the dried metals strong nitric acid is now to be added, the mass gently heated, and then evaporated to dryness : when cold it is moistened with dilute nitric acid, water added, and the whole filtered. Lead and copper will pass through the filter as soluble nitrates, and the tin will be found in the filter as insoluble peroxide ; this is to be well washed, dried, ignited, and weighed. It contains 78-61 parts of metallic tin. The amount of tin thus obtained, when multiplied by two, will represent the percentage of the ore. If, before the action of caustic potash, the ore has been submitted to the action of aqua regia, sulphuretted hydrogen may be passed through the solution of tin chlo- ride, when tin sulphide will be precipitated ; this is to be washed, dried, gently calcined in a platinum crucible until all smell of sulphurous acid has ceased, allowed to cool, reheated with a fragment of ammonium carbonate, as in the case of roasting copper sulphide, and when cold weighed as pure oxide of tin. The calculation for metal is made as above. Mr. J. W. B. Hallet has found that tinstone is very easily dissolved by fusion with three or four times its weight of potassium fluoride. The mineral must be finely pulverised. The fused mass is treated directly in the cru- cible with sulphuric acid to expel fluorine, after which, by adding water, filtering, and boiling the filtrate, the whole of the tin is thrown down as stannic acid, which is to be separated from traces of iron in the usual manner. This method of dissolving the ore of tin is much more convenient than fusion with caustic alkalies, or with sulphur and sodium carbonate. M. Moissenet precipitates the metal from a solution of the chloride by means of zinc, and then melts the precipi- tated metal in stearic acid. His process comprises five operations : ASSAY OF TIN IN GUN- AND BELL-METAL. 549 1. Purification of the ore by treatment with aqua regia. 2. Eeduction of the residue in the presence of charcoal. 3. Solution of the tin and iron in hydrochloric acid. 4. Precipitation of the tin by means of zinc. 5. Fusion of the precipitate into a button in stearic acid. The precipitation of tin by zinc is very rapid, and takes place in strongly acid solutions ; but the amount of acid and the dilution of the chloride influence the condition of the precipitate. In some solutions it appears in brilliant needles, but in very dilute solutions, and always towards the end of an operation, it is only a muddy deposit. The author recommends that a button of zinc be suspended in the liquid by means of a copper wire. When the precipi- tation is finished, the metal is collected and pressed into a porcelain capsule. On applying heat the lump so formed melts in a few minutes if a piece of stearine is added to it. Assay of Tin in Gun- and Bell-metal. The follow- ing process was employed for some years in H. Sainte- Claire Deville's laboratory: Dissolve about 5 grms. of the alloy in strong nitric acid contained in a flask provided with a funnel in the neck to prevent loss by spirting. When quite dissolved boil the strong solution for about twenty minutes ; dilute with two or three times its bulk of water, and boil again for the same time. Separate the insoluble tin oxide by decantation or filtration, and weigh after calcining it. (The tin oxide is sometimes rose- coloured, owing to the presence of minute traces of gold ; this may be disregarded.) The nitric acid solution freed from the tin is evaporated on a small platinum or porce- lain dish, and the residue is calcined at a dull red heat. In this manner a mixture of oxides is obtained in a suf- ficient quantity to serve for at least two analyses. About 2 grms. of the finely pulverised oxides are placed in a small platinum or porcelain boat, and thence introduced into a small glass tube closed with a good cork suitable for weighing. The boat, the tube, and the cork having been previously weighed, the weight of the oxides is obtained after they have been heated to dull redness in 550 THE ASSAY OF TIN. the apparatus, through which a current of dry air circu- lates. After having weighed the whole the current of air is replaced by dry hydrogen, and the tube is heated over a lamp until the contents cease to lose weight. It then contains unreduced zinc oxide together with copper, lead, and iron in the metallic state ; the colour of the product shows the operator when the experiment is concluded. On weighing again the loss of weight indicates with great accuracy the amount of oxygen contained in the oxides of these three metals. If the iron and lead are present in inappreciable quan- tities, multiplying this loss by 5 will give very nearly the weight of copper present, and, in consequence, the composition of the alloy itself. For bronze, bell-metal, gun-metal, &c., E. Burse ('Zeitschrift fur Analytische Chemie,' 1878, p. 58) pro- ceeds as follows : 1 grm. of the alloy, cut in pieces, is placed in a beaker, covered with 6 c.c. nitric acid of sp. gr. 1-5 : 3 c.c. water are then slowly poured in, and the beaker quickly covered. When the whole is dissolved it is heated to a boil and diluted with 500 c.c. boiling water. The tin oxide, after it is completely settled, is washed with boiling water, and weighed. The filtrate, for the expul- sion of nitric acid, is evaporated with two grms. sulphuric acid, and the copper, after the addition of sulphurous acid, is precipitated with a solution of 2 grm. ammonium sulpho- cyanide. The copper sulphocyanide, after settling, wash- ing, and drying, is weighed as such, or is converted into sulphide. The filtrate is concentrated with the addition of nitric acid, and mixed with ammonia in excess. If iron oxide is deposited, it is filtered off and weighed. The ammoniacal solution of zinc is mixed with ammonium sulphide, avoiding excess ; the sulphide when deposited is filtered off, dried, and, after ignition with sulphur in a current of hydrogen, weighed. Estimation of Tin by means of a Standard Solution. The first process to be described is due to M. Gaultier de Clauby, and may be thus performed : The standard solu- tion is made by dissolving 100 grains of iodine in 1 quart ESTIMATION OF TIN BY A STANDARD SOLUTION. 551 of proof spirit (specific gravity "920), and is thus standard- ised. Ten grains of pure tin are dissolved in excess of hydrochloric acid, the solution boiled, and allowed to cool : the burette is now filled with the solution of iodine which is gradually added to that of the tin until the former ceases to be decolourised ; as soon, therefore, as the tin solution assumes a faint yellow tinge, sufficient iodine has been added : the quantity thus found sufficient is then noted, and the amount of tin each division of iodine solution is equivalent to is calculated as for iron, copper, and the other standard solutions. In the actual assay of tin ore by means of this solution it is necessary that the whole of the tin present be reduced to the state of protochloride : this may be readily effected by boiling the solution of tin for a quarter of an hour with excess of metallic iron, and filtering. To the solution so obtained the iodine is added as above. The tin ore is dis- solved by any of the methods already described. M. Lenssen* estimates tin by means of the iodine .solution, but he operates in a liquid containing double potassium and sodium tartrate, and sodium bicarbonate in excess. The results M. Lenssen obtained by this method are satisfactory, by using the atomic weight of tin generally adopted (59). We shall see farther on why M. Lenssen's results agree. M. Stromeyer,f having recently occupied himself with the same subject, has succeeded in solving the difficulty. The solution of stannous chloride is carefully introduced into an excess of ferric chloride. The salt of iron becomes reduced to a minimum, according to the following equa- tion : 2Sn + 2(Fe 2 Cl 6 ) = 2SnCl 2 + 4FeCl 2 . It is then estimated by permanganate, as if it were a salt of iron protoxide. The results M. Stromeyer obtains in this way are very exact. The author adds that such a method of estimating is applicable only in the absence of copper or iron, as these two metals decompose potassium * Annalen der Chemie und Pharmacie,' vol. cxiv. p. 114. t Ibid. vol. cxvii. p. 261. 552 THE ASSAY OF TIN. permanganate as well as the tin ; but it may be of great use in the estimation of commercial salts of tin. A method for the analysis of tin ore consists in reduc- ing the finely powdered ore, in a hard porcelain vessel^ in a current of hydrogen gas. A bright red heat must be applied and maintained for a considerable time. The sample is allowed to cool in a current of hydrogen and weighed, the operation being preferably repeated to make sure that the reduction is complete. The loss of weight gives the quantity of stannic acid, tin peroxide (1 part by weight of oxygen representing 4- 673 parts tin per- oxide), if no other oxide is present which may be reduced at the same time. But as tinstone contains almost always iron oxide, part of the loss is due to this compound, so that the crucible, with its contents, should be digested in a beaker with hydrochloric acid. When dissolved, it is diluted, filtered into a large flask, the quantity of the residue (silica, &c.) estimated, the liquid supersatu- rated with ammonia, a sufficient quantity of ammonium sulphide added with flowers of sulphur, so as to convert the stannous sulphide into stannic sulphide, which then dissolves in the ammonium sulphide. The whole is di- gested in the flask, loosely stoppered, till the black iron sulphide is separated from the yellow liquid, filtered under cover, washed with the addition of a little ammonium sulphide, dissolved in hydrochloric acid, oxidised with potassium chlorate, and the ferric oxide (iron peroxide) precipitated with ammonia. The tin may be directly estimated by precipitating the solution of sulphide with dilute sulphuric acid, using a slight excess. The vessel is loosely covered with paper, and very gently heated, till it no longer smells of sulphuretted hydrogen. The yellow tin sulphide is washed upon the filter, dried enough to enable it to be taken out of the funnel along with the paper, and heated very gently, for a considerable time, in a porcelain crucible, at first covered and then open. When the odour of sulphurous acid is no longer percep- tible the contents of the funnel are moistened with a few drops of nitric acid, and are then gradually heated, with ESTIMATION OF TIN BY A STANDARD SOLUTION. 553 excess of air, ultimately to full redness. The tin sulphide is thus converted into stannic acid, which is heated for a short time with a little ammonium carbonate to remove every trace of sulphuric acid. In another method for the assay of tinstone, the finely ground sample is mixed in a porcelain crucible with 3 parts sodium carbonate, and 3 parts of sulphur, and the mixture is melted, covered, over the, lamp. When cold, the crucible is laid in water, and heat is applied till the mass dissolves, iron or other electro- positive metals remaining as a black sulphide, which is filtered off, washed, and when dry ignited in the air, in order to convert it into oxide. From the dilute alkaline solution the tin sulphide is precipitated by dilute sul- phuric acid, as directed above. In case of impure tin ores, the insoluble sulphide may consist of iron, zinc, copper, and bismuth ; whilst the soluble, along with the tin, may contain arsenic, tungsten, and molybdenum. A process for the separation of tin from tungsten has been given at p. 546. Tin slags are decomposed by means of aqua regia, and the tin, &c., precipitated by a current of sulphuretted hydrogen. Tin furnace products may be in part alloys of tin and iron, lead, tungsten, cobalt, and arsenic. They are finely powdered, digested with aqua regia, the tin thrown down with sulphuretted hydrogen, and the tin sulphide treated as already directed. If the proportion of tungstic acid is considerable, after digestion in aqua regia and diluting, the tungstic acid separates along with some stannic acid. This deposit is filtered off, washed, dried, and repeatedly ignited with sal-ammoniac, in a covered porcelain crucible, till the weight of the residual tungstic acid is constant. The stannic acid is found as difference. Sulphuretted hydro- gen is then passed into the acid solution (from which the tungstic acid has been deposited) ; the remaining tin is thus precipitated as sulphide, and is filtered off. Iron and manganese are estimated in the filtrate. (F. Rarn- melsberg's Quant. Chem. Analysis.) 554 THE ASSAY OF TIN. We have seen that M. Stromeyer, by a happy modifi- cation, has reduced the estimation of tin to that of iron. Applying the same principle, a salt of copper may be substituted for a salt of iron. An equivalent quantity of copper can thus be estimated in place of tin ; and M. Mohr's as well as M. Terreil's * experiments show that copper can be very exactly estimated by permanganate of potash. A double decomposition takes place on protochloride of tin being added to nitrate or chloride of copper in excess ; a salt of suboxide of copper forms, and the tin passes to the maximum state of oxidation, according to the following equation : 4CuO + 2SnCl 2 = 2Cu 2 + SnCl 4 + Sri0 2 - To estimate tin it is, then, sufficient to transform it into protochloride, to add to it a solution of nitrate of copper slightly in excess, before diluting it with water, and to titrate the liquid obtained by permanganate of potash. There are then three different processes for estimating tin by potassium permanganate : 1. To operate with water freed from air by boiling, protecting it from access of air while cooling. 2. To oxidise protoxide of tin in an alkaline medium. 3. To decompose stannous chloride either by a salt of iron, as proposed by M. Stromeyer, or by a salt of copper. Alloys of tin and lead, such as solder, inferior tinfoil, and the coating of terne plates, may be treated thus, according to Eammelsberg : Oxidise with nitric acid ; evaporate to dryness in the water-bath ; heat the residue rather more strongly ; moisten with nitric acid when cold ; dilute and separate the stannic acid (tin peroxide) by filtration, washing till the filtrate has no longer an acid reaction. When dry the precipitate is detached from the filter, and placed in a porcelain crucible; the paper is burnt on the lid, the * Comptes-Rendus, vol. xlvi. p. 230. ESTIMATION OF TIN BY A STANDARD SOLUTION. 555 ash added to the contents of the crucible, moistened with a few drops of nitric acid, heated and finally ignited. The tin is calculated from the stannic acid. In the filtrate the lead is estimated as sulphate by the addition of sulphuric acid. (See ' Separation of Lead from Copper and Zinc.') 556 CHAPTEE XIII. THE ASSAY OF ANTIMONY. ANTIMONIAL substances susceptible of being assayed by the dry way are divisible into two classes. CLASS I. In this class are comprised native antimony and all antimonial substances containing oxygen or chlorine, and but little or no sulphur. These substances are the following : Native antimony, Sb, Antimony oxide, Sb 2 O 3 , Antimonio.us acid, Sb 2 O 4 , Antimonic acid, Sb 2 5 . CLASS II. includes antimony sulphide and all antimo- nial ores containing much sulphur. Antimony sulphide, Sb 2 S 3 , Antimony oxysulphide, Sb 3 + 2Sb 2 S 3 , Haidingerite, 2Sb 2 S 3 + 3FeS. ASSAY OF ORES OF THE FIRST CLASS. Antimony oxides are readily reduced by charcoal, so that their assay presents no difficulty. The assay is con- ducted in precisely the same manner as that of lead oxide ; only, as antimony is much more volatile than lead, the heat must be managed with care, and the assay taken from the fire as soon as finished. When all suitable pre- cautions are taken, the loss of antimony is not very con- siderable ; but Berthier says it is never less than from 5 to 6 per cent. This, we think, is too high. The pure protoxide gives 77 per cent, of metal, and antimonious ASSAY OF ANTIMONY SULPHIDE. 557 acid 75. The reduction is readily made, without addition, in a charcoal crucible ; but when the substance to be assayed is mixed with impurities, some flux must be added. It succeeds equally well with 3 parts of black flux, with 1 part of tartar, with 1 part of sodium carbo- nate, and 15 per cent, of charcoal, or any other equivalent reducing flux. When the substance under assay contains iron oxide, the latter oxide is more or less reduced, and the metallic iron alloys with the antimony. Oxidised minerals which contain but a small quantity of sulphur can also be assayed in this manner ; because the sulphide gives up to black flux the small quantity of antimony which it contains, so that but little remains in the slag. The common glass of antimony produces by this method of assay 70 per cent, of antimony, and occa- sionally even more than that. The ores of this class occur very seldom, and are only in rare cases subject to assaying. ASSAY OF ORES OF THE SECOND CLASS. As pure antimony sulphide (antimonium crudum) as well as metallic antimony (regulus of antimony) are mer- cantile substances, the assays of the ores of this class have for their object the estimation of both these bodies. 1. ASSAY OF PURE ANTIMONY SULPHIDE (ANTIMONIUM CEUDUM). Antimony sulphide is almost the only mineral from which antimonium crudum is produced. This mineral generally occurs intermixed with very refractory gangue (gneiss, quartz, limestone, &c.) Antimony sulphide fuses readily at a low red heat, and is not changed during fusion if atmospheric air is precluded. At a white heat it volatilises without change of composition. The assay of antimony sulphide is now effected by a liquation process, i.e. by heating the mineral sufficiently in order to melt the antimony sulphide, and, by this means, 558 THE ASSAY OF ANTIMONY. to separate it from the refractory gangue. The produc- tion of antimony sulphide on a large scale is done in the same way. For the purpose of assaying, two pots or crucibles are used, one standing in the other, and leaving sufficient space between the two to receive the fused antimony sulphide. The bottom of the inside crucible is furnished with holes. The mineral to be assayed is put into the inside crucible, the latter is then closed with a cover, and hermetically luted ; the joints of the two crucibles are also luted. The under crucible is then put on the hearth of a furnace, enclosed with ashes or sand, in order to keep it cool, and the upper crucible, as far as it is outside of the under crucible, is covered with coal, and heated to a moderate red heat. The antimony sulphide will then melt and collect in the under crucible, from which it may be taken out, after cooling, and weighed. 2. ASSAY OF EEGULUS OF ANTIMONY. This assay may be made in two ways : first, by roast- ing and fusing the oxidised r L matter with black flux ; secondly, by fusing the crude ore with iron, or iron scales, with or without the addition of black flux. The roasting of antimony sulphide requires much care, for it is very fusible and volatile, as is also the oxide its decomposition gives rise to. The heat ought to be very low during the operation, and the substance con- tinually stirred. When no more sulphurous acid is given off, we may feel [assured that it is perfectly roasted, because no sulphate is ever formed in this operation. The roasted sulphide is then fused with three parts of black flux, or its equivalent. Metallic iron very readily separates all the sulphur from antimony sulphide ; but as iron sulphide has a specific gravity near that of antimony, the separation is very difficult to manage : a strong fire must be employed when the desulphurisation is complete, to keep the whole body in full fusion, for a considerable time. With these precautions, two buttons are obtained, which separate very ASSAY OF REGULUS OF ANTIMONY. 55 well : the one white, and in large plates, which is anti- mony ; and the other a bronze-yellow, a little brighter than the ordinary iron sulphide, because it is mixed with a little metallic antimony. During the operation a con- siderable portion of antimony is always volatilised, which, by this process, is an inconvenience impossible to avoid. It is, nevertheless, practised in the large way in some factories ; but a good result is not generally obtained. It,, however, appears that when all the necessary precautions are taken, it can be employed with advantage. The first precaution which is indispensable is, mixing with the sulphide only the precise proportion of iron necessary to effect its decomposition, which quantity amounts to about 42 per cent, of its weight. If more be used, the antimony, having a great tendency to play the part of an electro -negative element, will combine with the surplus, and an iron antimonide results, part of which will remain in the antimony and part in the slag. Further, the iron ought to be in the finest possible state of division. If the masses be large, a portion of antimony sulphide is volatilised before they can be fully attacked. In general, 63 per cent, of antimony can be extracted from its sulphide by the aid of iron in the small way, but on the large scale it seems that 55 per cent, is the maximum. Cast iron cannot be employed instead of wrought,, because sulphur has very little action on it. The desul- phurisation is imperfect, and the slag adheres to the reduced metal. One of the greatest inconveniences in separating sulphur from antimony by means of iron is the strong heat necessary to separate the slag from the metal. This may be remedied by making the slag more fusible and less heavy, by the addition of some flux, as an alkaline carbonate or sulphate. If antimony sulphide be fused with an alkaline car- bonate and charcoal, a regulus is obtained, and a slag com- posed of an alkaline sulphide and antimony sulphide. If metallic iron be thrown into this slag whilst in fusion, all 560 THE ASSAY OF ANTIMONY. the antimony separates immediately, and a new slag is formed as fluid as the former, containing iron sulphide and a sulphide of the alkaline base employed. If, instead of the above process, the iron be mixed intimately with the antimony sulphide and carbonated alkali, the result is the same 100 parts of sulphide, 42 of metallic iron, 50 of sodium carbonate mixed with one-tenth of its weight of charcoal, or 50 of black flux, give 65 to 66 of regulus ; with the same proportion of iron, and only 10 of flux, only 62 per cent, can be obtained. In these two cases the fusion takes place very rapidly and without bubbling, and the slag, which is very liquid, separates readily from the metal. By employing 1 part of alkaline flux, the proportion of iron can be reduced from 2 5 to 30 per cent., and the product of metal is always from 65 to 66 per cent. Hence, in making an assay of antimony sulphide, it is always better to employ a smaller quantity of iron than is necessary to complete the desulphurisation, and make up for it by increasing the quantity of flux : then it may be insured that no excess of iron will be present. The alkaline sulphates are decomposed into alkaline sulphides by the agency of charcoal at a slightly elevated temperature. The sulphides of the alkaline metals, by combining with the other metallic sulphides, augment their fusibility very considerably. Thus when sodium sulphate, mixed with about one-fifth of its weight of charcoal, is added to a mixture of antimony sulphide and metallic iron, the metallic antimony separates very rapidly, and the slag almost instantly becomes perfectly" fluid. But it must be noted that the presence of an alkaline sulphide diminishes the product of regulus, unless the proportion of iron be augmented at the same time. For instance, with 100 parts of antimony sulphide, 42 parts of iron, 100 parts of sodium sulphate, 20 parts of charcoal, ASSAY OF EEGULUS OF ANTIMONY. 561 but 22 parts of regulus were furnished ; but with 100 parts of antimony sulphide, 42 parts of iron, 10 parts of sodium sulphate, 2 parts of charcoal, 62 parts of antimony were easily obtained. Instead of metallic iron, pure iron oxide may be used, or any ferruginous matter whatever, provided it is rich ; but it is necessary to add, at the same time, an alkaline flux and charcoal to reduce the iron oxide. Not less than 40 parts of iron scales can be employed for 100 of antimony sulphide ; and then, on the addition of 50 to 100 parts of sodium carbonate and 8 to 10 of charcoal, about 56 of regulus are obtained ; but if with 100 parts of sodium carbonate from 14 to 15 parts of charcoal be employed, 65 parts of antimony are the result. By augmenting the proportion of scales, that of soda may be diminished. Thus, if from 56 to 60 parts of scales, 10 of soda, and 10 of charcoal be employed, 50 parts of regulus are the result ; and if the proportion of soda be 50, and that of carbon 10, from 65 to 66, and even 67, parts of regulus are obtainable. The fusion always takes place quickly, and the slags are very fluid. When antimony sulphide is fused with forge slag (ferrous silicate), sodium carbonate, and charcoal, a very white crystalline regulus, in large plates, is obtained ; to- gether with a bronze-yellow matte, and a black, opaque, vitreous slag, shining like jet, in which the greatest portion of the alkali employed appeared to be concen- trated. These three substances separate very readily from each other. 100 parts of antimony sulphide, 80 parts of forge slag, 50 parts of sodium carbonate, 10 parts of charcoal, produced very readily 60 parts of regulus. o o 569! THE ASSAY OF ANTIMONY. Antimony sulphide may also be analysed by boiling with aqua regia. The residue consists of sulphur and gangue. It is to be washed and dried, then weighed and ignited. The loss will be sulphur, and the remainder is pure gangue. Water is then added to the filtered solution, which will cause the precipitation of some of its contained antimony as oxy chloride : this must be separated by filtration. The solution is then to be saturated with potassium carbonate, and a new precipitate will be formed. The solution is to be filtered, and made slightly acid ; then barium nitrate must be added to it to separate its sulphur as barium sulphate, which is to be washed, dried, and weighed ; its weight indicates the amount of sulphur : 116 parts are equal to 16 parts of sulphur. The precipitate by water of oxychloride which remains on the filter is redissolved by hydrochloric acid, and its antimony separated in the metallic state by means of zinc. The precipitate formed by potassium carbonate may con- tain lead, copper, iron, and antimony. It must be treated with nitric acid ; this dissolves everything but the antimony, which may then be estimated as antimonic acid. It is always best, before conducting the analysis of antimony sulphide, to moisten it with very dilute hydro- chloric acid, in order to dissolve a portion of the calcium carbonate which may form part of the gangue. As the composition of the antimony sulphide is constant, the following process is sufficient in the assay of an antimonial ore : Boil the ore, after treatment with dilute hydrochloric acid, with concentrated hydrochloric acid, which dissolves only antimony sulphide, and precipitate the metal as oxy- chloride by means of water. Or, after all gangue soluble in dilute hydrochloric acid has been removed, the residue may be weighed, and then acted on by boiling hydrochloric acid, until all action ceases. The residue must be well washed with weak hydro- chloric acid, dried, ignited, and weighed ; the loss of weight corresponds to the percentage of pure antimony sulphide, which contains 72-7 -per cent, of metal. DETECTION OF ANTIMONY IN SUBLIMATES. 5(33 Franz Becker (' Zeitschrift flir Analyt. Chemie,' 1878, p, 185) mixes 1 part of the ore with 3 parts sodium car- bonate and 3 parts sulphur, melts in a porcelain crucible, extracts with hot water, decomposes the filtrate with hydrochloric acid, and converts the antimony sulphide into oxide in the usual manner. . The following method of estimating antimony is given by Mr. Sutton : The oxide of the metal, or any of its compounds, is brought into solution as tartrate by tartaric acid and water ; the excess of acid neutralised by neutral sodium carbonate, then a cold saturated solution of sodium bicarbonate added in the proportion of 20 c.c. to about CM grm. of Sb 2 3 ; to the clear solution starch liquor and -^ iodine are added until the blue colour appears ; the colour disappears after a little time, therefore the first appearance of a per- manent blue is accepted as the true measure of iodine required. 1 c.c. yV iodine=0-0061 grm. Sb. Detection of Antimony in Sublimates. In the examina- tion of mineral bodies for antimony, the test substance is often roasted in an open tube for the production of a white sublimate. Dr. E. Chapman, Professor of Miner- alogy at Toronto, recommends for the detection of anti- mony in this substance the following process a method more especially available when the operator has only a portable blowpipe-case at his command : The portion of the tube to which the chief part of the sublimate is attached is to be cut off by a triangular file, and dropped into a test-tube containing some tartaric acid dissolved in water. This being warmed, or gently boiled, a part at least of the sublimate will be dissolved. Some potassium bisulphate either alone, or mixed with some sodium carbonate and a little borax, the latter to prevent absorption is then to be fused on charcoal in a reducing flame ; and the alkaline sulphide thus produced is to be removed by the point of the knife-blade, and placed in a small porcelain capsule. The hepatic mass is most easily o o 2 564 THE ASSAY OF ANTIMONY. separated from the charcoal by removing it before it has time to solidify. Some of the tartaric acid solution is then to.be dropped upon it, when the well-known orange- coloured precipitate of antimony sulphide will at once result. In performing this test, it is as well to employ a somewhat large fragment .of the test substance, so as to obtain a thick deposit in the tube. It is advisable also to hold the tube in not too inclined a position, in order to let but a moderate current of air pass through it ; and care must be taken not to expose the sublimate to the action of the flame, otherwise it might be converted almost wholly into a compound of antimonious and anti- monic acids, the greater part of which would remain undissolved in the tartaric acid. A sublimate of arsenious acid, treated in this manner, would of course yield a yellow precipitate ; easily distinguishable by its colour y however, from the deep orange antimonial sulphide. The crystalline character, &c., of this sublimate would also effectually prevent any chance of misconception. To distinguish Arseniuretted Hydrogen from Antimo- niuretted Hydrogen. On passing a mixture of these two gases through a -tube containing solid pieces of caustic potash, these become covered with a brilliant metallic coating of antimony, whilst the arseniuretted compound escapes undecomposed. A lye of potash, density 1-250, only acts very slightly in a similar case. The fragments of potash which have become metallised by the deposit of antimony are altered in the air ; they soon become white in water, the metallic coating falling to the bottom ; but when it is attempted to collect them on a filter they dissolve before the liquid has even passed through. In the clear filtrate, antimony is found in solution. Separation of Tin from Antimony and Arsenic. Dr. Clemens Winckler (' Zeitschrift fur Analyt. Chemie,' 1875, p. 163) proceeds as follows : If the substance is an alloy it is dissolved in a mixture of 4 parts hydrochloric acid, and 5 parts water, with the addition of a sufficient quan- tity of tartaric acid. If the mixed metals exist as sul- DETECTION OF ANTIMONY IN SUBLIMATES. 565 phicles they are collected on a filter, washed, and dis- solved upon it in dilute potash lye. The nitrate is mixed with tartaric acid, treated with a current of chlorine, or with bromine in slight excess, and neutralised with hydro- chloric acid. In either case the solution is introduced into a beaker, diluted to 300 or 400 c.c., and so much solution of calcium chloride of a known strength is added that the subsequently precipitated calcium carbonate may exceed the tin present by about 15 times in weight. The liquid is then neutralised with potassium carbonate, potassium cyanide is added, and afterwards a slight excess of potas- sium carbonate, so that the lime may be totally precipi- tated. The liquid is then heated till it begins to boil ; it is allowed to settle, the liquid is poured upon a filter without disturbing the sediment, which is then treated with fresh water, boiled up, allowed to settle, and the clear liquid poured upon the filter. In this manner the bulk of the antimony is removed. The precipitate in the beaker is dissolved in a little concentrated hydrochloric acid, tartaric acid is added, the liquid again neutralised with potassium carbonate, and precipitated with potassium cyanide. After boiling, the liquid is poured as above through the same filter, three successive portions of water are added, heating each time to a boil, and the precipitate is finally brought upon the filter and completely washed. All the arsenic and antimony are now found in the filtrate, and all the tin, with an excess of calcium carbonate, in the precipitate. ASSAY OF ALLOYS OF LEAD AND ANTIMONY (Type Metal). The comminuted alloy is digested in a flask with concen- trated nitric acid till the antimony is oxidised. It is then slightly diluted with water, supersaturated with ammonia, and mixed with an excess of ammonium bi-hydrosulphide, which must be concentrated and yellow. If recently prepared a small quantity of flowers of sulphur is added. The whole is digested in the flask for some time, heating at last almost to a boil. The flask is then corked up and let stand till the deposit, which must be of a pure black, 566 THE ASSAY OF ANTIMONY. has subsided, while the supernatant liquid is yellow. If this is not the case more ammonium bi-hydrosulphide is needed. When cold the undissolved lead sulphide is separated by nitration from the solution of the antimony sulphide, washed with cold water, dried, placed in a tared porcelain crucible, the filter burnt separately to ashes, which are added to the precipitate along with some sul- phur, and the whole is ignited in a current of hydrogen. After the crucible with its contents has been weighed the addition of sulphur and the ignition are repeated, in order to find if the weight is constant. The lead is calculated from the weight of the lead sulphide. The solution of antimony sulphide is mixed, drop by drop, with hydrochloric acid, stirring all the time till the reaction becomes slightly acid. The beaker is covered with a glass plate and allowed to stand till the antimony sulphide is deposited. It is digested at a very gentle heat till the odour of sulphuretted hydrogen has ceased ; when cold it is filtered upon a filter dried at 120 C., and washed with cold water to which a few drops of hydrochloric acid have been added. When air-dry it is kept at the heat of 120-130C. in the hot-air oven till its weight is constant. A weighed portion of this substance is placed in a bulb- tube and heated in a current of dry carbonic acid, at first very gently, and afterwards at 200-300. The black-grey tin sulphide remains, w r hich is allowed to cool in a current of carbonic acid and weighed. 567 CHAPTER XIV. THE ASSAY OF ZING. ALL bodies containing zinc, usually found in the assay office, may be divided into four classes : Class I. Zinc ores, in which the metal exists as oxide not combined with silica : Earthy zinc oxide, ZnO. Manganiferous zinc oxide, ZnO + MnO. Zinc aluminate, Gahnite, ZnO,6AL 2 O 3 . Franklinite, 3(FeO,ZnO) + (Fe 2 3 ,Mn 2 3 ). Anhydrous zinc carbonate, ZnO,CO 2 . Hydrated zinc carbonate, ZnO,3H 2 + 3ZnO,C0 2 . Class II. Zinc ores, in which the metal exists, as in the former class, as oxide, but partly or wholly combined with silica : Anhydrous zinc silicate. Hydrated zinc silicate, electric calamine. Class III. Zinc ores, in which the metal is partly or wholly combined with sulphur. Zinc sulphide (blende, Black Jack), ZnS. Zinc oxysulphide. (This is rare.) Zinc sulphate, ZnO, S0 3 ,7H 2 0. Zinc selenide, ZnSe. (Very rare.) Class IV. Alloys. ASSAY OF ORES OF THE FIEST CLASS. In order to reduce the zinc oxide contained in sub stances of this class, it is sufficient to mix them with char- coal and expose them to a white heat. 568 THE ASSAY OF ZINC. At the moment of reduction the zinc is in a vaporised state. Its vapours, however, are readily condensable, so that the operation may be conducted in an ordinary retort, and all the metal is deposited in the neck without loss. It seems from this that nothing is so easy, at first sight, as the assay of zinc oxide ; but it is not so. It is very easy to reduce all the oxide, but it is not so easy to col- lect all the zinc ; nor is it easy to condense it all in the metallic state, and in consequence to estimate the pre- cise proportion in the ore submitted to assay. This difficulty consists, first, in the deposit being extended over a large surface and often adhering very strongly to the sides of the retort, so that it is nearly im- possible to detach it ; and, secondly, as the neck of the retort is open, the air, having access to it, brings to the state of oxide all the vapour nearest the end of the neck. The proportion of zinc oxidised is larger in proportion to the smallness of the quantity submitted to assay, and is always very considerable where no more than 200 to 400 grains are operated upon. It is not, therefore, in the extraction of the zinc from its oxide that the assay is rendered partially uncertain, but in its collection. The distillation of zinc requires a very high tempera- ture, and cannot be performed in retorts of glass ; those of earthenware must be employed. It is not necessary to lute these retorts when they are of good quality ; and they are better thin, because they heat more rapidly, and are not so likely to crack. After the mixture of oxide and charcoal has been introduced into the retort it is placed in the fire. The neck ought to have a long tube of glass, with a narrow bore adapted to it, so as to collect all the zinc which may escape from the wide part of the neck of the retort. This disposition is also convenient, as it does not allow such a free access of air. It is heated gradually until it is white inside ; the zinc is reduced and volatilised, and condensed in the neck : the greater the heat, the nearer the orifice. The metal can be ASSAY OF OEES OP THE FIRST CLASS. 569 detached readily from the neck, if it be well blackleaded inside. It is necessary, from time to time, to observe the state of the neck, because when very narrow it is often obstructed, and, if not cleaned out with an iron rod, an explosion might be caused. When the operation is finished the apparatus is allowed to cool, the retort taken out, placed carefully on its side and broken, in order that any particles of zinc which have condensed in its dome may be removed. If the approximate proportion of metallic zinc alone be required, all is collected and fused at a very gentle heat in a crucible with some black flux ; but if the true quantity of zinc is to be estimated it must be done in a more exact manner. The deposit must be collected with all possible care ; the neck must then be broken to pieces, and every piece having adhering to it either zinc or oxide must be placed on one side, and digested in hot nitric acid, which takes up those substances. If any be in the glass tube, it must be carefully cleansed by means of acid, and the solu- tion added to that produced by the digestion of the broken neck, and the deposit mechanically collected, in nitric acid. The solution is then evaporated gradually to dryness and heated to redness. The nitrate, by these means, is decom- posed and transformed into oxide, four-fifths of the weight of which is equal to the quantity of metallic zinc produced in the assay. The foregoing is the method of estimation by dis- tillation ; the following is the method of estimation by difference. Two plans of assay in this manner may be adopted : first, at an ordinary assay temperature ; secondly, at a very high temperature, as that of an iron assay. In all cases it is necessary to commence with the expulsion of all volatile bodies the ore may contain. If water or carbonic acid alone be present, simple calcination will do ; but if carbonaceous matter, roasting must be had recourse to. When the assay is made at an ordinary assay tempera- ture, the sample is finely pulverised and mixed with from 15 to 20 per cent, of equally finely pulverised charcoal, 570 THE ASSAY OF ZINC. and pressed into a crucible, on which a cover is placed, but not luted, and rapidly heated to whiteness. When no more zinc vapour is disengaged it is cooled, and the mixture in the pot collected. The residue ought to be pulverulent ; but as it is mixed with some charcoal, it is roasted, and then weighed. It is evident that the loss represents the zinc oxide. The charcoal added, it is true, leaves a small quantity of ash, but it is too small to be worth accounting for. In making the assay in the manner described, it is to be feared that a small quantity of the oxide might remain un- decomposed, and that a part of the residue might adhere to the crucible, and could not be detached ; and, lastly, there is always a degree of uncertainty as to the state of oxida- tion in which the iron will exist after roasting. No incon- venience of this nature presents itself when the assay is made at a very high temperature. This mode is the most exact of all, and leaves nothing to be desired. The assays of zinc at a high temperature are made exactly as those of iron. They are made in a charcoal crucible, with the addition of fixed fluxes, suitable to effect the fusion of the gangues mixed with the zinc oxide, if they be not fusible by themselves. The button is weighed ; it is a compound of slag and grains of iron, which are col- lected and their weight ascertained, and, by the difference, that of the slag. The weight of oxygen which the iron has lost during its reduction is then added to it, and by deducting from the substance the weight of the button and the oxygen so obtained we have the proportion of zinc oxide reduced in the assay. On the other hand, by deducting from the weight of the slag the weight of flux added, the weight of earthy substances and irreducible oxides which were mixed with the zinc oxide is ascer- tained. These results can be shown in a tabular form, in the following manner : Let m be the weight of the crude ore, n the weight of the calcined ore, r the weight of the flux added, f the Aveight of the cast iron, s the weight of the slag, o the ASSAY OF OKES OF THE FIRST CLASS. 571 weight of oxygen combined with the iron, calculated from the weight of metal produced, ' z the weight of the zinc oxide, then: in crude ore = calcined ore . n r fixed fluxes ...... r n + r Gives metal . . /\ Tfttfll f , >> Gives slag . . i/S^ }/+*** Zinc oxide n + r f s o Flux added . . . r Earthy matter . s - r The following is an actual experiment by Berthier : 100 crude ore = calcined ore ..... 83'3 10 kaolin (china clay) acted on by acids . . . lO'O 7 marble = lime ....... 4-0 97-3 Gave metal . 45-3 \ Tfl fi1 .o^ Gave slag. . 16-3 j ^ W*j*and 26094-62 x 507 = 132296-7234 and 815018-7831 + 132299-7234 2500 = 378-9 oz. or 378 ozs. 18 dwts. (nearly) per ton of the original sample,, before pulverising and sifting. In every case of assay yet described, it may be men- tioned that if the sample contain gold, the whole of that metal will be found with the silver, as obtained by cupel- lation, and may be separated as stated in the chapter on the Assay of Gold. Cupellation. Cupellation is an operation that has been known from time immemorial ; it has many characters in common with scorification, and is effected in nearly the same manner. Like that, it has for its end the separation of silver and gold from different foreign substances, by means of lead ; but it differs in this, that the scoriae pro- duced are absorbed by the substance of the vessel named 618 THE ASSAY OF SILVER. a cupel, in which the operation is made, instead of remain- ing on the melted metal, the latter remaining uncovered and in contact with the air, so that the extraneous metals are not only oxidised, but also all the lead ; and there re- mains nothing but the pure metals, silver and gold, or an alloy of them, in the cupel. Cupellation requires, as an indispensable condition, that the slag should have the property of penetrating and soaking into the body of the substance forming the cupel ; it is, therefore, only applicable to a certain number of substances, and not to all, like scorification. Lead and bismuth oxides, in a state of purity, are the only oxides which possess the property of soaking into the cupel ; but by the aid of one or the other, various oxides which by themselves form infusible scorise on the cupel, acquire the property of passing through it : therefore, on making a cupellation, it is necessary to fuse the substance with a sufficient proportion of lead or bismuth, so that the oxides they produce may combine with the oxides of all the foreign metals produced in the operation, and carry them into the body of the cupel. This proportion varies with the nature of the sub- stances cupelled, and other circumstances. The quantity required in ordinary cases will be mentioned hereafter. The cupels or porous vessels in which the operation is made ought to have a sufficiently loose texture to allow the fused oxides to penetrate them easily, and at the same time to possess sufficient solidity to enable them to bear handling without fracture ; and, moreover, they ought to be of such a nature as not to enter into fusion with either lead or bismuth oxide. For a description of their mode of manufacture, see p. 141. The following is the method in which an ordinary -cupellation is conducted : The furnace being heated, the bottom of the muffle is covered with cupels, placing the largest towards the end ; and if they are required to be heated as quickly as possible, they may be placed upside down, and turned, at the instant of use, by means of the tongs. When the interior of the muffle is reddish-white, ASSAY OF SUBSTANCES OF THE FIRST CLASS. 610 the matters to be cupelled may be introduced. When the cupels have been placed in their proper position, great care must be taken from the commencement to blow out of them all cinders, ashes, and other extraneous substances which may have fallen into them. The substance to be cupelled is sometimes an alloy, which can pass without addition of lead, and sometimes a compound, to which lead must be added. In the first case, the alloy is taken hold of by a small pair of forceps, and deposited gently in the cupel. In the second case, the substance to be cupelled is enveloped in a sheet of lead of suitable weight, and placed, as before, in the cupel ; or the necessary quantity of lead may be first placed in the cupel, and when the lead is fused, the substance to be cupelled added, taking care not to agitate the melted mass and cause loss by splashing. If the substance to be cupelled is in very small pieces, as grains or powder, it must be enveloped in a small piece of blotting-paper, or, still better, in a piece of very thin sheet lead, giving it a slightly spherical form, and dropping it gently into the mass ot molten metal in the cupel. Sometimes the substance is gradually added, by means of a small iron spoon ; but it is preferable to use paper or thin lead, as just recom- mended. When the cupels are filled, the furnace is closed, either by the door or by pieces of lighted fuel, so that the fused metals may become of the same temperature as the muffle. When this point has been gained, air is allowed to pass into the furnace ; the metallic bath is then in the state termed uncovered ; that is, it presents a convex surface, very smooth and without slag. When the air comes in contact with it, it becomes very lustrous, and is covered with luminous and iridescent patches, which move on the surface, and are thrown towards the sides. These spots are occasioned by the fused oxide of lead which is continu- ally forming, and which, covering the bath with a very thin coating of variable thickness, presents the phenome- non of coloured rings. The fused litharge, possessing the power of moistening 020 . THE ASSAY OF SILVER (so to speak) the cupel, is rapidly absorbed by it when sufficiently porous, so that the metallic alloy is covered and uncovered every instant, which establishes on its sur- face a continual motion from the centre to the circumfer- ence. At the same time a vapour rises from the cupels which fills the muffle, and is produced by the vapour of lead burning in the atmosphere. An annular spot is soon observed on the cupel around the metal, and this spot increases incessantly until it has reached its edges. In proportion as the operation proceeds, the metallic bath of silver-lead diminishes, becoming more and more rounded ; the shining points with which it is covered become larger and move more rapidly ; lastly, as the whole of the lead separates, the button seems agitated by a rapid movement, by which it is made to turn on its axis ; it becomes very lustrous, and presents over its whole surface all the tints of the rainbow : suddenly the agitation ceases, the button becomes dull and immovable, and after a few instants it takes the look of pure silver. This last part in the operation of cupellation is termed the brightening, fulguration, or coruscation. If the button be taken from the muffle directly after the brightening, it may throw off portions of its substance; this must be avoided, especially when the button is large. The button, when covered by mammillated and crystalline asperities, is said to have ' vegetated.' The cause of this effect seems to be, that when the fused buttons are suddenly exposed to the cold air, the silver solidifies on the surface, whilst that in the interior remains liquid. The solid crust, contracted by cooling, strongly compresses the liquid interior, which opens passages for itself, through which it passes out, and around which it solidifies when in contact with the cool air. But it sometimes happens that, when the contraction is very strong, a small portion of the silver is thrown off in the shape of grains, which are lost. After brightening, the cupels must be left for a few minutes in the furnace, and drawn gradually to the mouth, before they are taken out, so that the cooling may be slow and gradual. These precautions are nearly superfluous CUPELLATION. 621 when the buttons are not larger than the head of an ordi- nary pin. As silver is sensibly volatile, it is essential, in order that the smallest possible quantity be lost, to make the cupellation at as low a temperature as may be. On the other hand, the heat ought to be sufficiently great, so that the litharge may be well fused and absorbed by the cupel ; and, moreover, if the temperature be too low, the opera- tion lasts a very long time, and the loss by volatilisation will be more considerable than if 'the assay had been made rapidly at a much higher temperature. Experience has proved that the heat is too great when the cupels are whitish and the metallic matter they contain can scarcely be seen, and when the fume is scarcely visible and rises rapidly to the arch of the muffle. On the con- trary, the heat is not strong enough when the smoke is thick and heavy, falling in the muffle, and when the litharge can be seen not liquid enough to be absorbed, forming lumps and scales about the assay. When the degree of heat is suitable the cupel is red, and the fused metal very luminous and clear. In general, it is good to give a strong heat at the com- mencement, so as to well uncover the bath, then to cool down, and increase the heat at the end of the operation for a few minutes, in order to aid the brightening. There can be no inconvenience in urging the temperature at first, be- cause the silver-lead is then poor, and much precious metal cannot be lost by volatilisation. The increase of fire given towards the end is for the purpose of separating the last traces of lead, from which it is very difficult to free the silver ; but this strong fire must not be continued long, otherwise there might be a notable loss by volatilisation. When the assay of very poor argentiferous matters is made, the heat can be kept up nearly all through the cupellation. It generally succeeds better when the tem- perature is too high than too low. The force of the current of air which passes through the muffle is another very important thing in the success of the operation. Too strong a current cools the cupel, 022 THE ASSAY OF SILVER. oxidises too rapidly, and the assay would be spoilt. With too feeble a current the operation proceeds slowly, the assay remains a long time in the fire, and much silver is lost by volatilisation. When the litharge is produced more rapidly than it can be absorbed by the cupel, or when it is not liquid enough, which may happen from the furnace being too cold, or when other oxides, produced at the same time, diminish its fusibility, it accumulates gradually on the fluid metal, forming at first a ring which envelopes its circum- ference, and which, gradually extending, covers the whole surface ; at this period the assay becomes dull, and all movement ceases. When the operation is carefully at- tended to, it is nearly always possible to avoid this acci- dent. If at the first moment any signs are manifested of this evil, the temperature of the muffle must be raised, either by shutting the door, or placing in it burning fuel ; the assay will, in a little time, resume its ordinary course. But when the cause of the mishap is supposed to be the abundance of foreign oxides in the assay, a fresh propor- tion of lead must be added. It can be ascertained whether an assay has passed well by the aspect of the button. It ought to be well rounded, white, and clear, to be crystalline below, and readily de- tached from the cupel. When it retains lead it is brilliant below and livid above, and does not adhere at all to the cupel. Ill order to detach the button, seize it with a strong pair of pliers (see fig. 109), and examine with a microscope, (see fig. HO), brushing it to detach small particles of litharge which may adhere to it, and place it in the pan of a balance (fig. 13, page 28), which will indicate the T _i__tn of a grain. The weight of the silver furnished by the lead or litharge employed in the operation ought to be subtracted from the amount of silver obtained ; so that it is necessary to ascertain the richness of these matters beforehand, as they are never completely free from silver. The poorest of them contain from TWoo tn to iw^o- tn - Sometimes an equal quantity of lead is placed in another CUPELLATION. 623 cupel, and the silver thus obtained placed in the balance- pan containing the weights. Cupellation does not give the exact proportion of silver contained in an alloy. There is always a loss, and this loss- is always greater than that which takes place in the large way, as in the latter process a greater quantity is always FIG. 109, FIG. 110. obtained than that estimated by the assay. The loss of silver is traceable to three causes .: 1st, to volatilisation ; 2ndly, to oxidation ; Srdly, and lastly, to the absorption of minute globules of silver into the body of the cupel. It is certain volatilisation takes place, because a notable quan- tity of silver is always found deposited on the sides of the furnace and chimney in the shape of dust ; and silver, which is volatile by itself, becomes much more so when alloyed with lead, and is carried away by the vapours of the latter, and found in the pulverulent deposits, termed lead smoke or fume, which proceed from the combustion of the latter 624 THE ASSAY OF SILVER. metal in the air. Nevertheless, this cause of loss is not very important, for it is rare that the fume contains more than -roiro-oth f silver, and accurate experiments have proved that in cupellation in the small way not more than two to three per cent, of lead is volatilised. It is certain that a portion of the silver found in cupels which have been used for assays exist in the state of oxide, for no part of their mass is free it is found even in the bottom ; besides it is known that the lead carbonate precipitated from lead acetate made from litharge contains silver, and a notable quantity of that metal is found even in the lead sulphate prepared by means of alum from the acetate (unless the sulphate is repeatedly washed with water). It has been remarked that the centres of cupels which have been used for assays are richer in silver than the parts near the circumference, and that under the button there is a spot of bright yellow, which appears to be a compound of silver. But the most important cause of loss in an assay is the property which the alloys of silver and lead possess of introducing themselves into the pores of the cupel. The quantity thus lost is in proportion to the coarseness of the cupel. For the same quantity of silver, the loss which takes place in an assay varies according to the nature of the alloy, and the circumstances under which the assay is made ; so that it is not possible to form accurate tables of correction. This loss is much augmented with the quantity of lead employed, but without its being proportionate ; so that when scorification is had recourse to, it is advantageous to continue the operation for some length of time, in order that the metallic button may be reduced to the smallest suitable volume. In the assay of rich alloys, the proportion to the total amount of silver is very small, but notable ; and it has been calculated for the alloys of copper employed in the arts at ^ J^th ; but in the assay of poor ores, such as galena and other minerals treated in the large way, the loss is very great, for it is usually as high as -5^-0 th. By extracting the lead from cupels used in this class of assay, the metal furnished contains from about g- - oVo~o tn CUPELLATION. 625 to -s-innnnrti 1 f silver. The following experiment will give an idea of the influence of the proportion of lead on the loss of silver : 100 grains of commercial litharge were fused with 10 grains of black flux, and gave 27 grains of lead, and a slag ; this was pulverised and reduced in the same crucible with 15 grains of black flux, and a second button was produced weighing 45 grains. These two buttons, being cupelled separately, gave, the first -0035 and the second -001 only of silver. Three new quantities of 100 grains of the same litharge were fused ; the first with -J a part of starch, the second with 2^, and the third with 10 of the same reducing agent. The resulting buttons of lead weighed respectively 5, 28, and 79 grains. These buttons were cupelled, and furnished "0035, '0035, and -003 respectively. From these experiments it will be seen that when the litharge is not reduced completely, there remains a notable proportion of silver in the scoriae ; but, nevertheless, in order to extract the largest possible quantity, the whole must not be reduced. Indeed, but a twentieth part only need be reduced, because more precious metal is lost in the cupellation of a large quantity of lead than remains in the portion not reduced. The loss of silver in large cupellations is less than that which takes place in an assay, because in the large way the litharge, or the greater part of it, is run off; whilst in an assay the cupel totally absorbs it, so that the latter presents, rela- tively to the same mass of lead, a very much smaller surface in the large than in the small way ; now it can be readily seen that the quantity of silver lost by absorption into the pores of the cupel must be proportioned to its surface, all things being equal. It has been ascertained by experiment that a cupel absorbs about its own weight of litharge ; so that from this fact a cupel of the proper size may be chosen, when the weight of lead to be cupelled is ascertained. It is always better to have the cupel about ^ or \ as heavy again as the lead to be cupelled. The various metals found in an alloy, which can be submitted to cupellation, scorify in proportion to their s s 626 THE ASSAY OF SILVER. oxidisability. Those most oxidisable scorify with the greatest rapidity, and vice versa ; so that those which have the greatest affinity for oxygen accumulate in the first portions of litharge formed, which, by that means becoming less fusible, sometimes lose the property of penetrating the cupel ; hence the reason why cupellations always present more difficulties at the commencement of the operation than towards the end, when the litharge formed is nearly pure lead oxide, and can contain only copper oxide. The appearance of the cupel used in an assay will give indications of the metals the alloy contained. Pure lead colours the cupel straw-yellow, verging on lemon- yellow. Bismuth, straw- yellow passing into orange-yellow. Copper gives a grey, dirty red, or brown, according to its proportion. Iron gives black scorise, which form at the commencement of the operation, and are generally found at the circumference of the -cupel. Tin gives a grey slag. Zinc leaves a yellowish ring on the cupel, producing a very luminous flame, and occasioning losses by carrying silver in its vapour, and by projecting it from the cupel in its ebullition. Antimony and lead sulphate in excess give litharge-yellow scoriae, which crack the cupel ; but, when not produced in too great a proportion, are gradually absorbed by the litharge. If the lead alloy submitted to cupellation is found to produce this effect, a fresh portion must be mixed with its own weight of lead and scorified : the button so obtained can now be cupelled. Amalgamation. There are a certain number of argen- tiferous matters which can be assayed by amalgamation, as they are treated in the large way by that method. Amongst these are native silver, chlorides, sulphides, and arsenio-sulphides, which contain neither lead nor copper. But this process is seldom had recourse to, because it is long, troublesome, and less exact than those just described. ASSAY OF THE ALLOYS OF SILVER AND COPPER. 627 Substances of the Second Class. Native silver. Alloys of copper and silver. Alloys of other metals and silver (artificial). Silver antimonide. Silver telluride. Auriferous silver telluride (see gold). Silver hydrargyride (amalgam), Silver auride (see gold). Silver arsenide. The following method of separating silver from galena is given in the ' Chemical News,' vol ii. p. 239. ' Galena consists, as is well known, of lead sulphide, mixed with a variable proportion of silver sulphide, and both these substances fuse together, or melt at a bright red heat. Now, it so happens that, when silver sulphide is fused with lead chloride, what is called a double de- composition takes place ; that is to say, silver chloride and lead sulphide are formed. Consequently, if we fuse together a quantity of argentiferous galena and lead chloride, we shall remove the whole of the silver from the galena, and replace it by lead sulphide. This, then, is the process : mix together the galena and lead chloride in the proportion of 100 parts of galena, 1 part of lead chloride, and 10 parts of sodium chloride or common salt ; or, if the galena be very argentiferous, add a larger amount of lead chloride. The whole is then fused together, when the silver chloride and common salt rise to the surface, and may be skimmed off, and the desilverised galena falls and may be run out from the bottom. The mixture of silver chloride and salt may then be decomposed by lime and charcoal, or in any other manner, so as to reduce the silver and a portion of the surplus lead chloride, by which a metal- lic mass will result, suitable for the operation of the cupel.' General Remarks on the Assay of the Alloys of Silver and Copper. The assay of these alloys is nearly always accomplished (at least in England) by cupellation. This assay is most important, as it is by the results obtained in the manner hereafter described that the price or value of all kinds of silver bullion is decided. This class of cupellation is effected without difficulty, because the copper oxide forms so slowly, that the litharge s s 2 628 THE ASSAY OF SILVER. is always enabled to pass it into the body of the cupel. After having weighed the lead and placed it in the cupel, as soon as it is perfectly fused place in it the alloy to be assayed, wrapped either in blotting-paper or thin leaf-lead. It is essential, in this class of assay, to employ a sufficient quantity of lead to carry away all the copper. We may always be sure of succeeding, whatever the alloy may be, by employing the maximum proportion of lead, that is to say, the quantity necessary to pass pure copper ; but as the loss which the silver undergoes increases with the length of the operation and with the mass of the oxidised matters, it is indispensable to reduce this loss as much as possible by reducing the proportion of lead to that which is strictly necessary. Long experience has proved that silver opposes the oxidation of copper by its affinity, so that it is necessary to add a larger amount of lead in pro- portion to the quantity of silver present. M. D'Arcet has obtained the following results by accurate experiments : Standard of silver Quantity of copper alloyed Quantity of lead necessary Relation of lead to copper . 1000 lo*hs 950 50 3 eotoi 900 100 7 701 800 200 10 50 1 700 300 12 40 1 600 400 14 35 1 500 500 16 to 17 321 400 600 16 17 27 1 300 700 16 17 23 1 200 800 16 17 20 1 100 900 16 17 18 1 pure copper 1000 16 17 16 1 It is remarkable that below the standard of 500, the same proportion of lead must be employed, whatever that of copper. This fact is repeatedly verified by experiment. Whenever fine silver is fused in a cupel, it is always neces- sary to add lead, in order to cause the button to unite and form well. If less than fV ns f ^ eac ^ be employed, the button will be badly formed ; the litharge cannot separate but by the action of a very strong heat, and a ASSAY OF THE ALLOYS OF SILVER AND COPPER. 629 considerable loss of silver ensues. If, on the contrary, Y 3 ths of lead is exceeded, the cupellation goes on well, but the loss is greater on account of the duration of the process. These proportions also ought to vary with the temperature. M. Chaudet has found that, to cupel an alloy containing T 9 ^^ths of silver, 5 parts of lead are required in the middle of the muffle, 10 in the front, and only 3 at the back. The proportion of copper carried off by litharge varies not only with the temperature, but even for the same temperature in relation to the amount of copper and lead the alloy contains. By cupelling 100 parts of copper with different proportions of lead in the same furnace, M. Karsten obtained the following results : Lead added Copper remaining after cupellation Quantity of lead consumed in carrying off 1 of copper 100 78-75 3- 200 70-12 7-1 300 60-12 7-7 400 49-40 7-9 500 38-75 8-1 600 26-25 8-15 700 19-75 8-00 800 8-75 8-70 900 5-62 9-50 1000 1-25 10-10 1050 o-oo 10-50 From which we see that the lead carried away from -j^th to yVth of its weight of copper. Much less lead can be employed in a cupellation by making the alloy maintain its richness of copper throughout the operation. This can be accomplished by adding to the alloy in the cupel small doses of lead, in proportion as that first added dis- appears by oxidation. If, for example, an alloy composed of 4 parts of copper and one of silver be fused with 10 of lead, by adding successive small doses of the latter, as already pointed out, but 7 parts will be consumed, although in the regular way from 16 to 17 would be employed. The proportion of copper oxide contained in the litharge increases each instant, and goes on incessantly increasing when an alloy of copper and lead is cupelled which con- tains an excess of copper. According to M. Karsten, this (330 THE ASSAY OF SILVER. proportion is always about 13 per cent, at the commence- ment, and 36, or more than a third, at the end of the operation. In the assay of the coined alloys of copper and silver, the loss of silver may even amount to five thousandths ; but the loss is variable, and is proportionately greater as the standard of the alloy is lower. The following Table contains the results of many experiments made on this subject : Loss, or the quantity of fine Exact standard Standard found by cupellation metal to be added to the stan- dard as obtained by cupellation 1000 998-97 1-03 975 973-24 1-76 950 947-50 2-50 925 921-75 3-25 900 896-00 4-00 875 870-93 4-07 850 845-85 4-13 825 820-78 4-22 800 795-70 4-30 775 770-59 4-41 750 745-38 4-52 725 720-36 4-64 700 695-25 4-75 675 670-27 4-73 650 645-29 4-71 625 620-30 4-70 600 595-32 4-68 575 570-32 4-68 550 545-32 4-68 . 525 520-32 4-68 500 495-32 4-68 475 470-50 4-50 450 445-69 4-31 425 420-87 4-13 400 396-05 3-95 875 371-39 3-61 350 346-73 3-27 325 322-06 2-94 300 297-40 2-60 275 272-42 2-58 250 247-44 2-56 225 222-45 2-55 200 197-47 2-55 175 173-88 2-12 150 148-30 1-70 125 123-71 1-29 100' 99-12 0-88 75 74-34 0-66 50 ! 49-56 0-44 25 24-78 0-22 ASSAY PKOPER OF SILVEE BULLION. 631 These numbers, however, are not constant, and vary with the circumstances under which the assays are made : two assays made from the same ingot, by the same assayer, may differ as much as four or five thousandths. Tillet has remarked that the cupels can retain double as much silver as is lost ; which proves, as has already been men- tioned, that the silver obtained by cupellation is not per- fectly pure, but may retain as much as 1 per cent, of lead. Special Instructions for the Assay of the Alloys of Silver and Copper. As before stated, peculiar weights are employed in the assay of silver bullion ; and the silver assay pound, with its divisions, will be found described at pages 34-35. In the ' General Eemarks on the Assay of the Alloys of Silver and Copper,' it will be seen that the alloy must be cupelled with a quantity of lead, varying with the amount of copper present in the alloy. Standard silver cupels very well with five times its weight of lead ; but when the approximate composition of the alloy is not known, it must be estimated by a preliminary assay. Assay for Approximative Composition of Alloy. Weigh off 50 grains of pure or test lead ; place them in a cupel previously made red-hot ; when the ]ead is fused, and its surface covered with oxide, place in it by means of the light tongs (a, fig. 28, page 68) 2 grains of the alloy under assay, wrapped in a small piece of thin paper. Allow the cupellation to go on according to the instructions, and with all the precautions already given, and when complete, weigh the resulting button, and, according to its weight, add lead in the actual assay in the quantity that is suffi- cient, as exhibited in the Table at page 628. Assay Proper of Silver Bullion. In this assay the ope- rator requires silver known to be standard, and pure lead. With the possession of the above substances the assay is thus proceeded with : Place the 12 grains weight(=l Ib.) in the scale pan, and exactly counterbalance it with stan- dard silver. This is to serve as a check. Remove the weight, and in its place add so much of the alloy to be 632 THE ASSAY OF SILVER. assayed that the balance is again equal. In one cupel, that destined to receive the check sample, place 60 grains of lead ; and in another cupel place such a number of grains of lead as may be found necessary by the preliminary assay. When the lead in both cupels is fused, add the silver alloy, and cupel with the necessary precautions. When the buttons in the cupels are cold, seize them with the pliers, and if neces- sary cleanse them with a hard brush, and place one in each balance-pan. If they exactly balance each other, the alloy operated on is standard silver ; if, however, it weighs less than the button produced from the check sample by the weight equivalent to 2 pennyweights, then it is 2 penny- weights worse than standard : on the other hand, if it be heavier by the same weight, it is 2 pennyweights better than standard. Silver is also reported as so much fine : thus standard silver may be reported as 11 ounces 2 penny- weights fine, and so on. In case extreme accuracy be re- quired, correction must be made according to the standard as shown by the Table at page 630. The standard silver in England is T V^- fine. Assay of Alloys of Copper and Silver. In the treat- ment on the large scale of copper ores containing silver, the contained silver is found alloyed with the copper, and it often falls under the assayer's province to estimate the quantity of precious metal. An assay of this kind is most conveniently accomplished by scorification before cupel- lation, thus : Prepare four scorifiers ; weigh into each of them 50 grains of the alloy, 50 grains of fused borax, and 600 grains of lead, and proceed as already described under the head ' Assay of Ores of the First Class by Scorifica- tion.' When the four buttons of lead are obtained, place them together in another scorifier, and submit to the furnace until the contents of the scorifier are completely covered with oxide ; pour as usual, and cupel the re- sulting mass of lead. Alloys of Platinum and Silver. -If any substance con- taining platinum as well as silver were assayed as already described, the button resulting from the cupellation would, in addition to the silver, contain the whole of the platinum. SEPARATING SILVER FROM THE BASER METALS. 633 In such a case the button so obtained must be thus treated : If the alloy contain much platinum, it must be fused with twice its weight of silver ; then treated with hot nitric acid ; evaporate the solution nearly to dryness ; add water and hydrochloric acid, until no further precipitation of silver as a white curdy precipitate (silver chloride) takes place. The silver chloride may be collected either on a filter or by decantation. The solution containing the platinum is treated with excess of sal-ammoniac solution until no further precipitation takes place ; the solution evaporated to dryness. When cold, dilute alcohol is added ; and the insoluble yellow matter (platinum ammonio- chloride) collected on a filter, washed with alcohol, dried, and ignited. The ignited residue is metallic platinum, which is weighed. The loss of weight which the alloy from cupel has sustained represents the amount of silver previously alloyed with it. Alloy of Platinum, Silver, and Copper. Treat such an alloy as above ; and the liquid, filtered from the platinum ammonio-chloride, will contain the copper. Acidulate it with hydrochloric acid, add metallic zinc, and proceed as directed under the head ' Wet Copper Assay.' Native Silver, Rough Silver left on Sieve during Pulveri- sation of Silver Ores of First Class, and Native Alloys of Silver as Antimonides, c. are treated by scorification and cupellation in precisely the same manner as just de- scribed for alloys of copper and silver. Dr. W. Dyce proposed, in ' Tilloch's Philosophical Magazine ' for 1805, the following process for separating gold and silver from the baser metals : ' Hitherto the process has always been, as far as I have understood it, attended with considerable difficulty in the execution ; but, by that which I am about to describe, it is done with exact certainty. It was discovered and com- municated to me by a gentleman in the neighbourhood. The process consists in mixing not less than two parts of powdered manganese with the impure or compound metal, which should be previously flattened or spread out so as 634 THE ASSAY OF SILVEE. to expose as large a surface as possible, and broken or cut into small pieces for the convenience of putting the whole into a crucible, which is then to be kept in a sufficient heat for a short time. On removing the whole from the fire, and allowing it to cool, the mixture is found to be converted into a brownish powder, which powder or oxide is then to be mixed with an equal proportion of powdered glass, and then submitted in a crucible to a sufficient heat, so as to fuse the whole, when the perfect metals are found at the bottom in a state of extreme purity a cir- cumstance of no small importance to the artist and the chemist, the latter of whom will find no difficulty in sepa- rating the one from the other, with so little trouble com- pared with the usual processes that I have no doubt it will always be practised in preference to the cupel.' Assay of Silver Bullion by the Wet Method. From what has been stated under the head of ' Cupellatiori,' it will be observed that there are many sources of error ; such as volatilisation of the precious metal, its oxidation in the presence of excess of lead oxide and atmospheric oxygen, and, lastly, its absorption into the body of the cupel either as oxide or metal, or in both states. These losses, as before stated, vary with the temperature, the amount of lead employed, and the texture of the cupel ; and, as may be seen from the table of corrections as drawn up by D'Arcet, give a very erroneous assay, unless the addition necessary for each standard be made. Considerable attention was called to this matter in France some years since, and a Special Commission was appointed to examine the subject thoroughly, and, if possible, to devise some means of assay which might be both easy and accurate. The result of this examination was the invention of a process of assay at once elegant and trustworthy : and as a full account of this method has not, to the author's knowledge, been translated and pub- lished in this country,* he has prepared the present from M. Gay-Lussac's Eeport, which formed a part of a com- * Some portion of this report has been published in Dr. Ure's ' Dictionary of Arts, Mines, and Manufactures.' ASSAY OF SILVER BULLION IN THE WET WAY. 635 munication from M. Thiers to Earl Granville, and which appeared in the original language in the year 1837, in a Eeport on the Eoyal Mint. The process of assay about to be described consists in estimating the fineness of silver bullion by the quantity of a standard solution of common salt necessary to fully and exactly precipitate the silver contained in a known weight of alloy. This process is based on the following principles : The alloy, previously dissolved in nitric acid, is mixed with a standard solution of common salt, which precipi- tates the silver as chloride, a compound perfectly insoluble in w r ater, and even in acids. The quantity of silver chloride precipitated is esti- mated not by its weight, which would be less exact and occupy too much time, but by the weight or volume of the standard solution of common salt necessary to exactly precipitate the silver previously dissolved in nitric acid. The term of complete precipitation of the silver can be readily recognised by the cessation of all cloudiness when the salt solution is gradually poured into that of the nitrate of silver. One milligramme of that metal is readily detected in 150 grammes of liquid ; and even a half or a quarter of a milligramme may be detected, if the liquid be perfectly bright before the addition of the salt solution. By violent agitation during a minute or two, the liquid, rendered milky by the precipitation of silver chloride, becomes sufficiently bright after a few moments' repose to allow of the effect of the addition of half a milligramme of silver to be perceptible. Filtration of the liquid is more efficacious than agitation ; but the latter, which is much more rapid, generally suffices. The presence of copper ,. lead, or any other metal, with the exception of mercury (the presence of the latter metal requires a slight modifi- cation of the process, which will be hereafter pointed out) r in the silver solution, has no sensible influence on the quantity of salt required for precipitation : in other words, the same quantity of silver, pure or alloyed, requires for 636 THE ASSAY OF SILVER. its precipitation a constant quantity of the standard salt solution. Supposing that 1 gramme of pure silver be the quan- tity operated on, the solution of salt required to exactly precipitate the whole of the silver ought to be of such strength that, if it be measured by weight, it shall weigh exactly 100 grammes, or if by volume 100 cubic centi- metres. This quantity of salt solution is divided into 1000 parts, called thousandths. The standard of an alloy of silver is generally the number of thousandths of solution of salt necessary to precipitate the silver contained in a gramme of the alloy. Measurement of the Solution of Common Salt. The solution of common salt will hereafter be termed the normal solution of common salt. It can be measured by weight or volume. The measure by weight gives greater FIG ill precision, and it has the special advan- tage of being independent of temperature ; but it requires too much time in nume- rous assays. The measure by volume gives a sufficient exactitude, and requires much less time than the measure by weight ; it is, indeed, liable to the in- fluence of temperature, but tables for correction will be appended. Measure of the Normal Solution of Salt by Weight This solution should be so made that 100 grammes will exactly precipitate 1 gramme of pure silver dissolved in nitric acid. In order to point out the method of taking the weight it must be supposed to have been previously prepared. After the process of taking the weight is described, the mode of preparing the solution will be given. The solution is weighed in a burette (fig. Ill) whose capacity is from 115 to 120 grammes of the solution, and divided into grammes. These divisions are MEASUKEMENT OF THE SOLUTION OF COMMON SALT. 637 for the purpose of approximate! vely estimating the weight of solution, so as to shorten the operation of weighing. The burette is represented as closed by a cork, B, in order to prevent evaporation of the solution when the instrument is not in use. It is also easy to remedy the inconvenience of evaporation, by rinsing the burette with a small quantity of the fresh solution. On pouring the solution from the orifice, (9, of the burette, each division will furnish from 8 to 10 drops ; and consequently the weight of a drop is about a decigramme. The burette is filled with solution to the division o ; it is then tared in a balance capable of turning with a centigramme. The burette is then re- moved, and its place supplied with a weight equivalent to the amount of solution required 100 grammes, for in- stance. The solution is then gradually poured from the burette into a bottle appointed for its reception, until the equilibrium is nearly established. It is not easy to attain the point exactly, as no smaller quantity than a drop can be poured from the burette. This, however, is a matter of indifference ; it suffices to know the exact weight of the solution poured out : suppose it to be 99 gr. 85 c. The mode of more nearly approximating the required weight of 100 grammes will now be pointed out. It must be remarked that it is not the amount of water contained in the 100 grammes that is of consequence, but only the quantity of salt found in solution ; this should exactly represent 1000 thousandths of pure silver. If now 100 grammes of the normal solution be mixed with 900 grammes of water, it is evi- dent that 1 gramme of this new solution is equivalent to a decigramme of the first, and consequently it will be easy to obtain 100 grammes of the normal solution, or rather the 1000 thousandths of salt it ought to contain ; it will now be sufficient to add to the 99 grammes already poured from the burette, 1 gramme of the new solution. It can be weighed, like the normal FIG. 112. 038 . THE ASSAY OF SILVER. solution, to a drop nearly, in the burette (fig. 112), which is of such a diameter that each small division represents a decigramme of liquid, and consequently a centigramme of the normal solution ; but it is more readily measured by volume, preparing it in the manner to be hereafter pointed out. To avoid all confusion, a solution to be termed a decime solution of common salt is one containing the same quantity of salt as the normal solution, in a weight or volume ten times greater. A decime solution of silver is a solution of silver equi- valent to the latter, both mutually suffering complete decomposition. Preparation of the Decime Solution of Common Salt. One hundred grammes of the normal solution of common FIG. 113. FIG. 114. salt are weighed in a flask (fig. 113) containing a kilo- gramme of pure water, when filled up to the mark a b, or 1000 cubic centimetres ; this quantity is made up with pure water, taking care to agitate the whole well, to render the mixtuise, homogeneous. A cubic centimetre of this solution represents 1 thousandth of silver. This quantity is readily obtained by means of a pipette (fig. 114), gauged so that when filled up with water to the mark c d, it shall allow 1 gramme, or 1 cubic centimetre, to run freely, the small quantity of liquid remaining in the pipette not forming part of the gramme. In pouring the liquid by drops, a little more or a little less than PREPARATION OP THE DECIME SOLUTION OF SALT. 639 twenty may be counted, according to the size of the orifice, o. This number will not vary more than one drop. Half a cubic centimetre will consequently be re- presented by 10 drops, and a quarter by 5. The precision arrived at by this method of measurement suffices, since the possible error on the cubic centimetre will be but one- twentieth of that quantity, or one-twentieth of a thou- sandth ; if, however, many measures be required, then compensation must be made. The decime solution of common salt requisite for assays must be kept in a bottle (fig, 114) closed by a cork, tra- versed by the pipette, firmly fixed in a FlG< 115< hole bored for that purpose. To mea- sure a thousandth with the pipette, the bottle is held with one hand, and the pipette with the other (fig. 115). The pipette is taken from the solution after its upper orifice has been closed by the forefinger ; the lower orifice is then inclined against the edge of the flask to remove the liquid, which without this precaution would remain there : the mark c d is then raised to the level of the eye, and by a suitable pressure of the forefinger on the upper orifice, which may be obtained by giving the pipette a slight alternating circular movement between the fingers, the solution is allowed to run out gradually. The instant the concave surface of the liquid is at the level c d, the pipette is firmly closed by pressure of the forefinger on its orifice, which is held above the bottle into which the solu- tion is to be poured, and the forefinger removed so that it can be emptied. It is here necessary to remark that in order to regulate the slow and regular runnings of the liquid from the pipette, by the pressure of the forefinger, the latter ought to be neither too moist nor too dry : if too dry, it will not perfectly close the orifice, even by strong pressure ; if too moist, it prevents the entrance of air, and the liquid will not run, or if it do, it will be irregularly. 640 THE ASSAY OF SILVER. This observation should not be lost sight of in the use of the large burettes mentioned hereafter. Preparation of the Decime Solution of Silver.*- The decline solution of silver is prepared by dissolving 1 gramme of pure silver in nitric acid, in a flask holding 1 litre (see fig. 113), and then diluting the solution with distilled water, so that, cooled at the ordinary temperature FIG. 116. of the air, it shall occupy exactly the volume of one litre. It is measured in precisely the same manner as the decime salt solution. Weighing the Normal Solution of Common Salt. To execute this operation with rapidity, a balance similar to that represented at fig. 116 is employed. The arms are divided as in the assay balance described at p. 28 ; each of the arms, c B, is furnished with a rider, c, of such a weight (about 5 decigrammes) that, moved from the right or the left of the centre, c, of each arm, it indicates PKEPAKATIOX OF THE NORMAL SALT SOLUTION. 641 two decigrammes. The space traversed by the rider is divided into twenty equal parts, representing an equal number of centigrammes. We will take for example the weighing of 100 grammes of normal solution of common salt, which is that most fre- quently made in the estimation of the standard of all varieties of argentiferous matter. There are two weights, one, P, equal to the tare of the burette when full of solution to the mark o, the other, P', equals 100 grammes. The burette is filled with solution, and placed on the right-hand pan of the balance, on which it is kept in position by the collar d 0, and through which it is passed before placing it on the pan. The tare, P, of the burette is- supposed to be on the opposite side. If the equilibrium be not perfect, it is effected by the rider on the left ; the burette is then removed, and 100 grammes of the solution (either more or less to one or two decigrammes) poured out. The burette is then again placed in the balance, with the 100-gramme weight P', the upper part of which is slightly concave to receive the bottom of the burette, in order to prevent it sliding off. The equilibrium is again established by the aid of the rider on the right, If, for instance, it is found necessary to remove the rider, which represents 15 centigrammes, 15 divisions towards B, the weight of the solution poured out of the burette will be equal to 100 gr. - 0-15 gr. = 99'85 gr. If, on the other hand, it is necessary to move the rider six divisions towards c, the weight of the solution will be 100 gr. + 0-06 gr.=100-06 gr. The above method of weighing the salt solution ap- pears to be the most convenient that can be employed, although it is not very expeditious. Other methods of weighing and measuring will be given in an appendix to this article. Preparation of the Normal Solution of Common Salt when measured by weight. After having pointed out the method of weighing the normal solution of salt, and of taking very small quantities, its preparation will be de- scribed. T T 042 THE ASSAY OP SILVER. Supposing the salt as well as the water to be employed are pure, the two substances have only to be taken in the following proportions : 0-5427 kilogramme of salt and 99*4573 kilogrammes of water, to form 100 kilogrammes of solution, of which 100 grammes will exactly precipitate 1 gramme of silver. But instead of pure salt, which is difficult to procure, and which besides rapidly alters by the absorption of atmospheric moisture, it is preferable to employ a concentrated solution of commercial salt, which can be prepared in large quantities, and kept for use as needed. The quantity of salt it contains can be ascer- tained " s by evaporating a portion to dryness, and by a few experiments it is easy to estimate in what proportion it shall be mixed with water to produce a solution, 100 FIG 117 grammes of which shall exactly precipitate 1 gramme of silver. Suppose, for example, that the salt solution contains 250 grammes of salt per kilogramme, and that it is necessary to prepare 100 kilogrammes of the normal solution. Now, since for the preparation of this quantity 0-5427 kilogramme of pure salt is required, we have the following pro- portion : 0-25 : 1 :: 0-5427 : ^=2-1708 kilogs. To this last weight enough water is added to make up 100 kilogrammes, that is to say, 97-8292 kilogrammes, which quantity can be readily measured by means of a flask containing 5 or 6 kilogrammes previously gauged. The mixture must be well agitated by means of the agitator (fig. 117), which is made of an osier twig, split into four branches, to the extremities of which is attached a small square piece of silk. This'substance is employed to avoid the separation of filaments which would ensue from the use of any other material. This agitator can be intro- PREPARATION OP THE NORMAL SALT SOLUTION. G43 duced into very small openings, and is exceedingly ser- viceable in agitating large masses of liquid. When well mixed, the solution must be assayed. To effect this, dissolve 1 gramme of silver in nitric acid, sp. gr. 1-290, in a stoppered bottle (fig. 118) Fm. us. holding about 200 grammes of water, tare the burette (fig. 1] 1) filled with the solution, and pour rather more than less into the bottle ; in proportion as the salt employed is impure, more than 100 grammes will be re- quired to precipitate 1 gramme of silver. The mixture is at first milky, but, by vigorously shaking the bottle, having its stopper firmly fixed, for about a minute, and then allowing it to remain at rest for a short time, the liquid will become perfectly bright : two drops of the solution must then be poured into it from the burette : if a cloudiness is pro- duced, it is agitated again to brighten it, and two drops more added. This must be continued until the last two drops added give no precipitate. The operation is then terminated, and nothing remains to state but the result. Supposing the total weight of solution poured from the burette is 101-88 grammes, the last two drops must not be reckoned, because they produce no effect ; the two pre- ceding drops were necessary, but in part only ; that is to say, the number of drops to be deducted is less than four, and more than two, or rather it is the mean term, three. Or the weight of a drop can be known exactly by taking that of a dozen: suppose it is equal to 0*082 gramme, three times that number must be deducted, or 0*255 grammes from 101-880 grammes: there will remain 101-625 grammes, representing the quantity of normal solution necessary to precipitate 1 gramme of silver. The solution is thus found to be too weak ; to bring it to its proper standard it is necessary to remove 1-625 gramme of water from the 101-625 grammes of solution, or, what is the same thing, to add to the normal solution a certain quantity of the concentrated solution of common T T 2 644 THE ASSAY OF SILVER. salt, which quantity may be found by the following pro- portion : 100 : -1-625 :: 2-1708 kilogrs. of silver solution : ^ = 0*0353. After the addition of this quantity of salt to the normal solution, a fresh assay is made, proceeding in precisely the same manner as before ; taking care, however, to pour from the burette a weight of solution slightly under 100 grammes, or 1000 decigrammes ; for instance, 998*4 decigrammes, because it is not possible, in pouring the solution by drops, to arrive at the exact weight, 1000 decigrammes. To ascertain the true standard in the most exact manner possible, a decime solution must be prepared by weighing 100 grammes of the normal solution, and diluting it with pure water, so that it shall occupy one litre : a cubic centimetre of this solution will represent a decigramme of the normal solution. This decime solution will not be rigorously exact, since the normal solution has not been truly standardised ; but it is easily perceived that the error thus committed is very small, and that it may be neglected. Nevertheless, as soon as the normal solution is perfectly standardised, it is better to prepare another decime solution. A decime solution may be immediately obtained by dissolving 0*5427 gramme of pure sea-salt in such a quan- tity of water that the whole will occupy one litre ; yet the first process is preferable. With the decime solution the assay may be thus con- tinued, remembering that the pipette described at fig. 114 is a cubic centimetre containing 20 drops ; that the half, therefore, is represented by 10 drops, and the fourth by 5. To the 998-4 decigrammes of normal solution already added, pour one pipette and 12 drops of the decime solution, which will exactly complete the weight of 1000 decigrammes of normal solution. The mixture is agitated to brighten it, and one-thousandth of common salt or one pipette of the decime solution added. If this causes a cloudiness, it is agitated and a second thousandth added. This last should produce no opalescence. The weight PREPARATION OF THE NORMAL SALT SOLUTION. 645 of normal solution necessary to exactly precipitate one gramme of silver will be between 1000 and 1001 deci- grammes ; that is to say, the mean will be equal to 1000-5. The standard of the normal solution is then too weak by half a thousandth ; to correct this a quantity of concen- trated salt solution must be added equal to half a thou- sandth of that already added (2-1708 + 0-0353 = 2-2061 kilogrammes) ; that is to say, 1-1 gramme. A new assay is then made for verification. When the standard of a solution is very nearly arrived at, it is well to employ filtration to detect the slightest opalescence, at least when sufficient time is not .allowed for the liquid to become perfectly bright. The surest method, when the standard is nearly attained, is to place some of the liquid in two test-glasses, and pour into one a few drops of the decime solution of common salt, and into the other a corresponding number of drops of the decime solution of silver nitrate. It may then be determined on which side the opalescence is manifested, and the assay of the normal solution may be continued after the mixture of the liquid in the two glasses, since the two quantities of the decime solutions of common salt and silver nitrate mutually decompose each other, and do not interfere with the assay. Once the standard of the normal solution is definitely fixed, the sum of the quantities of the concen- trated solution of common salt which have been employed, as well as those of the water, must be noted, and in the preparation of a new normal solution the proportions found as above would only have to be mixed to obtain at once a solution having very nearly its true standard. In determining the standard of the normal solution, supposing that it were always too weak, it would be necessary to add to the solution a certain quantity of common salt ; but if the true amount had been exceeded, and it had been found too strong, the solution would have to be precipitated with the decime solution of silver ; and knowing the number of cubic centimetres or thou- sandths of silver which had been necessary to ; precipitate the excess of common salt, it could be estimated what 646 THE ASSAY OF SILVER. amount of water must be added to reduce the normal solution to standard. For instance, if 2 thousandths of the decime solution of silver had been consumed, 2 thou- sandths of its weight of water would have to be added to the total amount of solution ; that is to say, 0-2 kilo- gramme or 200 grammes. Preservation of the Normal Solution of Common Salt. The most suitable vessel for containing the normal solution of common salt is one of glass, because that cannot affect the standard. Large glass bottles, termed carboys, are found in commerce. These bottles contain from 50 to 60 litres, and are very applicable for this purpose. Fig. 119 represents one of these bottles fixed in a stand formed of a sieve hoop. It is graduated into litres or kilogrammes of water, and a paper scale fixed on its side shows at any time the quantity of contained liquid. It is closed by an hydraulic valve, made of sheet iron, but the bell or cover is of glass. The detail of this valve is shown at fig. 120. FIG. 119. The a ^ r can on ty en ter the bottle by the narrow tube T 7 , and cannot pass out by it ; consequently evaporation is not to be feared. The neck of the valve should be about a deci- metre deep, into which mercury should be poured, but only to about one-third of its height. The solution is drawn from the bottle by the syphon S. This is furnished with a stopcock ; but this syphon being brittle, at least when not of metal, is not convenient in use, since it is incorporated with the- bell of the valve : it is, therefore, preferable to pierce the bottom of the bottle (fig. 121), and fix a metal tube (T) by means of a plate moulded on the bottom and cemented to it. This tube is raised a little above the bottom of the bottle, and covered by a small cup, the object of which is to protect it from any of the mercury which might fall into it. It is APPLICATION OF GAY-LUSSAC's PKOCESS. 647 terminated at its other extremity by a very narrow tube, so that the flow of the solution may not be too rapid- FIG. 120. FIG. 121. Fifl. 122. Hereafter a metal reservoir will be described which has all the advantages of a glass vessel without its inconve- niences. Application of the Process described in the Estimation of the Standard of a Silver Alloy. The alloy is supposed to be that made into coin, the mean standard of which is fixed at 900 thousandths, but which may vary from 897 to 903 thousandths without ceasing to be legal (French standard for coin). One gramme is dissolved in the bottle (fig. 118) by about 10 grammes of nitric acid, sp. gr. 1-290. This quantity of nitric acid can be readily taken by means of the pipette P (fig. 122), which contains 7'7 grammes of water to the mark a b. The solution may be accelerated by placing the bottle in a small pan of hot water, the bottom of which must be covered with a piece of cloth, so as to prevent contact of the glass and metal. The solution finished, and the flask slightly cooled, the 648 THE ASSAY OF SILVER. nitrous vapour must be removed 4 by a blower (see fig. 123), the nozzle of which is formed of a piece of bent glass tube, connected by a cork with a copper socket D, having a screw inside. This operation ought to be effected, as well as the solution of the alloy in nitric acid, under a chimney with a strong current of air, to carry off the nitrous vapour. The burette (fig. 99), being filled with the normal solution of common salt, and tared, about 90 grammes are poured into the solution of the alloy ; say 89 -85 grammes. After agitating the liquor, a cubic centimetre of the decime solution of common salt is added representing one-thousandth of silver. If a cloudiness be observed, agitate again, and add a second thousandth of common salt, and so on, until the last thousandth gives no precipitate. Suppose it to be the fourth : that must not be counted, because it has produced no effect ; and only a half of the third must be taken, because only a portion of that was necessary. The standard of the alloy would be consequently equal to If it be desirable to approach still nearer to the true standard of the alloy, half-thousandths must be added until the last half-thousandth gives no precipitate ; and in order to avoid all confusion, it is better to write with chalk on a blackboard the thousandths of common salt, preceding them by the plus sign +, and on the other side the thou- sandths of the silver nitrate, preceding them by the sign minus. In the above example, after the addition of the 4 thousandths of common salt, the last of which has pro- duced no cloudiness, 1^ thousandth of nitrate of silver is added, which destroys 1^ thousandth of common salt, and brightens the liquid. If another half-thousandth of nitrate of silver praduce no precipitate, it is not taken into account, and is struck off from the table. From whence it is con- cluded that the quantity of nitrate of silver necessary to APPLICATION OF GAY-LUSSAC'S PROCESS. 649 destroy the excess of common salt is more than 1 and less than li ; that is to say, nearly the J of a thousandth, and is equal to 1J. Thus the number of thousandths of salt really used is 41-25 = 275. The standard of the alloy, therefore, is 898-50 + 275 = 901-25. Another example, everything else remaining as above : Suppose the first thousandth of salt did not precipitate. This is a proof that too much normal solution of common salt has been employed, and that there is an excess of salt in the liquid. Add one- thousandth of silver, and agitate : things are now as at first, but it is nevertheless known that it is with nitrate of silver the process must be con- tinued. One-thousandth has been added, which produced a precipitate ; the second does not. The standard of the .alloy is consequently 898-5 0-5 = 898. To approach still nearer to the real standard, destroy the last 2 thousandths of silver by 2 thousandths of common salt, and add half a thousandth of silver a cloudiness is produced, as already known ; but another half-thousandth does not precipitate. The standard of the alloy is therefore 898-50-0-25 = 898-25. This process, on which it would be useless to enlarge further at present, because many other parts of the pro- cess to be presently described apply to it, is general, and gives exactly the standard of an alloy when it is known approximately, which can always be ascertained by a previous rough assay. Correction of the Standard of the Normal Solution of Salt when the Temperature Varies. It has been assumed that, in the estimation of the standard of the normal solution of salt, the temperature has remained constant. Assays made under these circumstances need no correc- tion ; but if the temperature changes, the same measure of solution will not contain the same amount of salt. Sup- posing the solution of salt has been standardised at 15. If, at the time an experiment is made, the temperature is 18, for instance, the solution will be found too weak, since it has expanded, and the pipette holds less than its proper weight. If, on the other hand, the temperature 650 THE ASSAY OF SILVEE. falls to 12, .the solution becomes concentrated, and is found too strong. It is therefore necessary to estimate the correction to be made for any variation of temperature that may occur. To this end the temperature of a solution of common salt has been gradually raised from 0... .5. ...10... .15. ...20.... 25.. ..30 degrees, and three pipettefuls of the solution exactly weighed at each of the above temperatures. One- third of the total weight gives the mean weight of the contents of a pipette. The corresponding weights of a pipetteful of solution are then entered, and form the second column of the following table, called ' Table of Correction for the Variations of Temperature in the Normal Solution of Salt/ By this table correction may be made for any temperature between and 30, when the solution of salt has been standardised within the same limits. Sup- pose, for example, the solution had been standardised at 15, and that at the time it was used its temperature was 18. On referring to the second column of the table, it will be seen that the weight o a measure of solution at 15 is 100-099 gr. ; and at 18, 100-065 gr. ; the difference 0-034 gr. is the quantity of solution taken too little, and consequently it must be added to the normal measure, so that it may be equal to one thousand thousandths. If the temperature of the solution had fallen to 10, the differ- ence of weight between a measure at 10 and a measure at 15 will be 0-019 gr., which must, on the contrary, be de- ducted from the measure, as it has been taken in excess. These differences of weight of a measure of solution at 15, and that of a measure for any other temperature, form the column 15 in the table, where they are expressed in thou- sandths. They are written on the same horizontal line as the temperatures to which each corresponds, with the sign + when they are to be added, and the sign when to be subtracted. The columns 5, 10, 20, 25, 30 have been calculated in the same manner, to meet cases in which the normal solution had been graduated at each of the above- named temperatures. Thus, to calculate the column 10, take the number 100-118 from the column of weights as CORRECTION OF THE STANDARD OF SOLUTION OF SALT. 651 a point of departure, and find the difference for all the other numbers in the same column. An application of this Table will be given hereafter. TABLE OF CORRECTIONS FOR VARIATIONS IN TEMPERATURE OF THE NORMAL SALT SOLUTION. Temperature Weight 5 10 15 20 25 30 Degrees Grammes Mill. Mill. Mill. Mill. Mill. Mill. 4 100-109 o-o -o-i + 0-1 + 07 + 1-7 + 2-7 5 100-113 o-o -0-1 + 0-1 + 0-7 + 1-7 + 2-8 6 100-115 o-o o-o + 0-2 40-8 hl-7 + 2-8 7 100-118 + 0-1 o-o + 0-2 + 0-8 + 1-7 +2-8 8 100-120 + 0-1 o-o + 0-2 + 0-8 + 1-8 + 2-8 9 100-120 + 0-1 o-o + 0-2 + 0-8 + 1-8 + 2-8 10 100-118 + 0-1 o-o + 0-2 + 0-8 4-1-7 + 2-8 11 100-116 o-o o-o + 0-2 + 0-8 + 1-7 + 2-8 12 100-114 o-o o-o + 0-2 + 0-8 + 1-7 + 2-8 13 100-110 o-o -o-i + 0-1 + 0-7 + 1-7 + 2-7 14 100-106 -0-1 -0-1 +0-1 + 0-7 +1-6 + 2-7 15 100-099 -0-1 -0-2 0-0 + 0-6 +1-6 + 2-6 16 100-090 -0-2 -0-3 -0-1 + 0-5 + 1-5 + 2-5 17 100-078 -0-4 -0-4 -0-2 + 0-4 + 1-3 + 2-4 18 100-065 -0-5 -0-5 -0-3 + 0-3 + 1-2 + 2-3 19 100-053 -o-o -0-7 -0-5 + 0-1 + 1-1 + 2-2 20 100-039 -0-7 -0-8 -0-6 0-0 + 1-0 4-2-0 21 100-021 -0-9 -1-0 -0-8 -0-2 + 0-8 + 1-9 22 100-001 -M -1-2 -1-0 -0-4 + 0-6 4-1-7 23 99-983 -1-3 -1-4 -1-2 -0-6 + 0-4 + 1-5 24 99-964 -1-5 -1-5 -1-4 -0-8 + 0-2 + 1-3 25 99-944 -1-7 -1-7 -1-6 1 -1-0 o-o + 1-1 26 99-924 -1-9 -1-9 -1-8 1 -1-2 -0-2 + 0-9 27 99-902 -2-1 -2-2 -2-0 -1-4 -0-4 + 0-7 28 99-879 -2-3 -2-4 -2-2 -1-6 -0-7 + 0-4 29 99-858 -2-6 -2-6 -2-4 -1-8 -0-9 -t-0-2 30 99-836 -2-8 -2-8 -2-6 -2-0 -1-1 0-0 TABLE FOR THE ASSAY, BY THE WET METHOD, OF AN ALLOY CONTAINING ANY PROPORTIONS WHATEVER OF SILVER, BY THE EMPLOYMENT OF A CONSTANT MEASURE OF THE NORMAL SOLUTION OF COMMON SALT. The process by which the normal solution of salt is measured by weight is applicable to the assay of every kind of alloy, since it suffices to take a weight of the solu- tion corresponding to the presumed standard of the silver, and complete the assay by means of the decime solu- tion ; the process by volume, however, has not the same 652 THE ASSAY OF SILVER. advantage, because the volume of normal solution cannot be varied in the same manner as the weight. This incon- venience, however, is of no very great consequence, for, by keeping the volume of normal solution constant, it suffices to vary the weight of the alloy, taking in each particular case a weight which contains approximative] y one gramme of pure silver. Suppose the alloy has a standard of about 900 thousandths, we have the following proportion : 900 thousandths : 1000 of alloy :: 1000 thousandths: x = 1111-1. If that weight be now taken to ascertain the standard of the alloy, it may be found, for instance, that to the mea- sure of 1000 thousandths of salt it is yet necessary to add 4 thousandths of salt to precipitate the whole of the silver ; that is to say, that 11.11-1 of alloy really contain 1004 of silver. From this result the real standard of the alloy may be found to be 903-6, by the following equation : 1111-1 : 1004 : : 1000 : x = 903-6. But such calculations, however simple, should be .avoided where numerous daily assays are made, not only on account of the time consumed, but still more from the errors to which such operations are necessarily exposed. Fortunately, all these inconveniences may be avoided by the use of tables, which entirely free the assay er from calculation. Wishing in weighing the alloy to avoid fractions of thousandths, and only making use of tenths and half-tenths of thousandths, the weight of alloy increases, starting from a gramme, from 5 to 5 thousandths, and the correspond- ing standard for each of these weights has been sought, all containing one gramme of pure silver. Thus the weight 1020 of alloy, in which there are 1000 of silver and 20 of copper, corresponds to the standard 980-39, obtained by the proportion 1020 : 1 000 : : 1000 : x = 980-39. EXPLANATION OF THE FOLLOWING TABLES. 653 On this principle are formed the first and second columns of the table marked Salt. The first contains the weight of each alloy, and the second its corresponding standard. The following columns, 1, 2, 3, to 10, give the standard of the alloy, when, instead of the 1000 milli- grammes of silver it was supposed to contain, it really contained 1, 2, 3, &c. more, and consequently 1, 2, 3, (fee- milligrammes of copper less. Another table, constructed in the same manner as the preceding, and marked Nitrate of Silver, gives the standard of the alloy when, under the weight given in the first column, it contains 1, 2, 3, 4-^ 559-9 560-4 561-0 561-6 562-1 562-7 558-3 558-9 559-4 560-0 560-6 561-1 556-8 557-3 557-9 558-4 559-0 559-6 555-2 555-8 556-3 556-9 557-5 558-0 553-7 554-3 554-8 555-4 555-9 556-5 552-2 552-7 553-3 553-8 554-4 554-9 550-7 551-2 551-8 552-3 552-9 553-4 549-2 549-7 550-3 550-8 551-4 551-9 672 THE ASSAY OF SILVER. COMMON Weight of Assay in Milligrs. 0. i. O " 3. 4. 1835 545-0 545-5 546-0 546-6 547-1 1840 543-5 544-0 544-6 545-1 545-6 1845 542-0 542-5 543-1 543-6 544-2 1850 540-5 541-1 541-6 542-2 542-7 1855 539-1 539-6 540-2 540-7 541-2 1860 537-6 538-2 538-7 539-2 539-8 1865 536-2 536-7 537-3 537-8 538-3 1870 534-8 535-3 535-8 536-4 536-9 1875 533-3 533-9 534-4 534-9 535-5 1880 531-9 532-4 533-0 533-5 534-0 1885 530-5 531-0 531-6 532-1 532-6 1890 529-1 529-6 530-2 530-7 531-2 1895 527-7 528-2 528-8 529-3 529-8 1900 526-3 526-8 527-4 527-9 528-4 1905 524-9 525-4 526-0 526-5 527-0 1910 523-6 524-1 524-6 525-1 525-6 1915 522-2 522-7 523-2 523-8 524-3 1920 520-8 521-3 521-9 522-4 522-9 1925 519-5 520-0 520-5 521-0 521-6 1930 518-1 518-6 519-2 519-7 520-2 1935 516-8 517-3 517-8 518-3 518-9 1940 515-5 516-0 516-5 517-0 517-5 1945 514-1 514-6 515-2 515-7 516-2 1950 512-8 513-3 513-8 514-4 514-9 1955 511-5 512-0 512-5 513-0 513-5 1960 510-2 510-7 511-2 511-7 512-2 1965 508-9 509-4 509-9 510-4 510-9 1970 507-6 508-1 508-6 509-1 509-6 1975 506-3 506-8 507-3 507-8 508-3 1980 505-0 505-6 506-1 506-6 507-1 1985 503-8 504-3 504-8 505-3 505-8 1990 502-5 503-0 503-5 504-0 504-5 1995 501-3 501-8 502-3 502-8 503-3 2000 500-0 500-5 501-0 501-5 502-0 TABLE FOR THE WET ASSAY OF SILVER. 673 SALT continued. 5. G. I . 8. 9. 10. 547-7 548-2 548-8 549-3 549-9 550-4 546-2 546-7 547-3 547-8 548-4 548-9 544-7 545-3 545-8 546-3 546-9 547-4 543-2 543-8 544-3 544-9 545-4 545-9 541-8 542-3 542-9 543-4 543-9 544-5 540-3 540-9 541-4 541-9 542-5 543-0 538-9 539-4 539-9 540-5 541-0 541-5 537-4 538-0 538-5 539-0 539-6 540-1 536-0 536-5 537-1 537-6 538-1 538-7 534-6 535-1 535-6 536-2 536-7 537-2 533-2 533-7 534-2 534-7 535-3 535-8 531-7 532-3 532-8 533-3 533-9 534-4 530-3 530-9 531-4 531-9 532-4 533-0 528-9 529-5 530-0 530-5 531-0 531-6 527-6 528-1 528-6 529-1 529-7 530-2 526-2 526-7 527-2 527-7 528-3 528-8 524-8 525-3 525-8 526-4 526-9 527-4 523-4 524-0 524-5 525-0 525-5 526-0 522-1 522-6 523-1 523-6 524-2 524-7 520-7 521-2 521-8 522-3 522-8 523-3 519-4 519-9 520-4 520-9 521-4 522-0 518-0 518-6 519-1 519-6 520-1 520-6 516-7 517-2 517-7 518-2 518-8 519-3 515-4 515-9 516-4 516-9 517-4 517-9 514-1 514-6 515-1 515-6 516-1 516-6 512-8 513-3 513-8 514-3 514-8 515-3 511-4 512-0 512-5 513-0 513-5 514-0 510-1 510-7 511-2 511-7 512-2 512-7 508-9 509-4 509-9 510-4 510-9 511-4 507-6 508-1 508-6 509-1 509-6 510-1 506-3 506-8 507-3 507-8 508-3 508-8 505-0 505-5 506-0 506-5 507-0 507-5 503-8 504-3 504-8 505-3 505-8 506-3 502-5 503-0 503-5 504-0 504-5 505-0 XX 674 THE ASSAY OF SILVER. APPLICATION. Assay of Pure, or nearly Pure, Silver, the Temperature of the Normal Solution of Salt being that at which it was standardised. First example. Let the ingot of silver have an ap- proximative standard of from 995 to 1000 thousandths. Take one gramme ; dissolve it in ten grammes of nitric acid, in the bottle, fig. 118. Then pour into the bottle an exact measure of the normal solution of salt, and brighten by agitation. The silver not being supposed to be quite pure, the standard is not further sought for by the decline solution of salt, but that of silver nitrate is employed. One thousandth of this latter solution is poured into the bottle ; it becomes cloudy, and is well agitated. A second and a third thousandth also give a precipitate, but not so a fourth. From these data the following is the method of ascertaining the standard of the alloy : The last thousandth of the decime solution of silver, having produced no cloudiness, is not to be counted. The third was necessary, but only partially so ; consequently the number of thousandths of silver necessary to decom- pose the excess of salt is more than 2 and less than 3 ; in other words, it is equal to the mean, 21 ; but since 2-| thousandths of silver have been required to complete the precipitation of salt representing 1000 thousandths of silver, it is evident that the silver submitted to assay con- tained 21 thousandths of alloy, and that its standard, to within nearly half a thousandth, is but 997-J. If it be considered necessary to arrive nearer the true standard, the following proofs must be employed : Pour into the solution 11 thousandths of salt, which will decom- pose a like number of thousandths of silver.* After due It has already been stated how a thousandth of the decime solution may be subdivided by the number of drops furnished by the pipette. If, for in- stance, it contains 20 drops, 10 will give the half, 5 the quarter, &c. Half a thousandth may also be obtained by diluting the solution with its volume of water, and using a whole pipetteful. This latter plan has been found the best in practice. ASSAY OF PURE, OR NEARLY PURE, SILVER. 675 .agitation, add half a thousandth of silver nitrate. Sup- posing a cloudiness is produced, no further addition must be made ; for it is already known that above the third thousandth no precipitate is formed in the liquid by silver nitrate, and consequently only half of the last half-thou- sandth must be calculated, as only a portion of it was necessary. From which, the entire number of thousandths of silver nitrate being 4J, and those of salt 1-J, there remains 2f for the number of thousandths of nitrate of silver addded to the normalsolution ; and consequently the standard of the aUoy is 1000 -2f =997J. If, on the other hand, the last half-thousandth of the silver nitrate had produced no cloudiness, it would not have to be reckoned, and only half of the preceding half-thousandth would have been taken. Thus from the 4 thousandths of silver nitrate employed a quarter of a thousandth is deducted ; and from the difference, of, is yet deducted 1-J of salt, the final remainder being 2^ thousandths of silver nitrate which have been added to the normal solution : the standard of the alloy would be 1000 -2J= 997 j. Although the above-described operation is very simple, yet it is desirable, in order to avoid al] confusion, to note in writing such thousandths of salt or silver nitrate as are added. The thousandths of salt indicating an increase of standard should be preceded by the sign + ; and the thou- sandths of silver nitrate announcing a diminution of stan- dard, by the sign . Second Example. Suppose the ingot has a presumed standard of 895 thousandths, and the temperature of the normal solution supposed invariable. Find in the table of standards (Salt Table), first column, that which approaches the nearest to 895 ; it will be found to be 896- 9, corresponding to the weight of 1115 milligrammes. This weight of the alloy is taken and dissolved in nitric acid, a measure of normal solution of salt added, and the whole well agitated. The operator is, however, doubtful whether the assay must be proceeded with by the decime salt solution, or the silver nitrate decime solution. If the former produces a precipitate^ x x2 676 THE ASSAY OF SILVER. it is gone on with ; but if it does not precipitate, that already added is decomposed by a similar addition of the second, and the solution rendered bright by agitation. A starting-point has now been arrived at for the continuance of the assay, for it is known that the silver nitrate solution must be employed. Suppose, then, that the alloy, after the addition of the measure of normal solution, yet gives a precipitate with the decime solution of salt. The first 5 thousandths produce a precipitate, but not the sixth, which conse- quently is not counted. The fifth has only been partially required, so that it is more than 4 thousandths, and less than 5, or the mean, 4^, is the quantity required to entirely precipitate the excess of silver in the alloy sub- mitted to assay. But by neglecting at first the fraction 0*5, seek in the Salt Table of Standards the number found on the longitudinal line of the weight 1115, under column 4 ; it is 900-4, and on adding 0-5 to this number we have 900-9, or 901, for the required standard. Supposing, however, that the same alloy, after the addition of the normal measure of salt, gives a precipitate with silver nitrate, and that the 3 first thousandths pro- duce a cloudiness, but not the fourth. The number of thousandths of silver nitrate really necessary for complete precipitation will be very nearly 2-J. To ascertain the real standard of the alloy of which 1115 thousandths were supposed to contain about 1000 thousandths of silver, take the number found in the horizontal line 1115, and in the column 2 of the Silver Nitrate Table. This number, which is 895-1, diminished by the fraction 0-5, gives 894*6 for the standard of the alloy to within half a thousandth. Third Example. The actual temperature of the normal solution of salt being 18, when it was standardised at 15. The ingot of silver submitted to assay has an approxi- mative standard of 795 thousandths. Find in the Salt Table of Standards, first column, that which is nearest to it ; it is 793*7, corresponding to the weight 1260. This weight of the alloy is taken, and the operation pro- ceeded with as already described. Supposing it had GRADUATION OF THE NORMAL SOLUTION OF SALT. 077 required 6'5 thousandths of salt to precipitate the whole of the silver contained in the alloy to within half a thousandth, the required standard, without correction for temperature, will be 798-4 + 0'4 = 798'8. But, making this correction, recourse must be had to the table, page 651, column 15 : the number 0'3, which will be found in the horizontal line 18 and the column 15, possesses the sign ; consequently it must be deducted from 798*8, and the remainder, 798'5, will be the standard weight. If the temperature of the solution, instead of being 3 higher than at the time it was standardised, was 3 lower, that is 12, the correction must be added, and would be equal to +0-2. The standard of the alloy would conse- FrG 12 4. quently be 798-8 + 0-2 = 799. Graduation of the Normal Solution of Salt, the Temperature being different to that at which it was wished to be graduated. Two equally ready processes can be employed. The one consists in reducing the temperature of the solution to the desired degree before standardising ; the other, in estimating its standard without regard to the temperature of the solution, and then correcting its influence by the aid of the tables of correction already given. First Process. Place the liquid to be graduated in a bottle, F, fig. 124. Intro- duce a thermometer, and heat to a deter- minate degree, say 20 for instance. This done, place the jet of the pipette in the bottle ; raise the liquid by aspiration by means of the conical tube, T, fig. 124, which is adapted to the opening of the air-cock, R. As soon as the liquid is raised a little, above the mark a 6, which determines the capacity of the pipette, close the stopcock, and complete 678 THE ASSAY OF SILVEK. the measurement as usual. This same means of filling the pipette by aspiration may be employed to fill it either with caustic alkali or nitric acid, as the case may be, to cleanse it instead of taking it to pieces. Second Method. The solution of salt being supposed at a temperature of 16, and it is desired to graduate it at that of 20. Proceed with the standardising without regard to temperature ; but when it is obtained in each trial assay, it is necessary to make the correction required by the temperature. If, for example, in an approximative assay the standard of the solution was expressed by 1001*5 this standard would not only be too weak by 1*5 thousandth, but, according to the table of temperatures, by yet another 0-5, for the solution is weakened by this quantity, by passing from 16 to 20. The standard, if taken at this last temperature, would be too low by 2 thousandths, and must consequently be corrected. If, on the other hand, the standard of the solution were too high instead of too low, and expressed by 998*5 at the temperature of 16 ; at that of 20, the solution being weakened by 0'5, the standard would only be but one thousandth too high, and it must be corrected by that quantity. Approximative Estimation of the Standard of an Unknown Alloy. It has always been supposed, in the experiments already detailed, that the approximative standard of the alloy sub- mitted to assay was known : and this, indeed, is nearly always the case. If, however, this be unknown, two means are available for obtaining the necessary know- ledge. A decigramme of the alloy is cupelled with one gramme of lead ; or if it be desirable not to use the cupel, it may be ascertained by the wet method, in the follow- ing manner : The assayer supposes the standard of the alloy known to be about a twentieth, and it can always be found nearer than that by touch, density, &c. A weight relative to its MODES OF ABRIDGING MANIPULATION. 679 supposed standard is taken, and its standard sought by adding the decime liquid by 10 thousandths at a time, by means of pipettes of this capacity (see fig. Fm 125 125). The term of complete precipitation is soon passed, and the standard of the alloy to about 5 thousandths is thus ascertained. The approximate standard to 2-J- thousandths may be obtained by adding only 5 thousandths oi solution at a time. Suppose the alloy 840 thousandths. . Take the weight 1190, corresponding to this standard, and proceed "as in an ordinary assay, adding each time, for example, a pipette of 10 thou- sandths of salt solution. It is found the fifth pipette gives no precipitate, and consequently the number of thousandths of salt for the pre- cipitate of the silver to within 5 thousandths is 35. The 1199 of alloy will therefore contain 1000 + 35 = 1035 of silver ; and the approxi- mative standard will be obtained by the pro- portion 1120: 1035:: 1000:0=869-7. Modes of Abridging Manipulation. In the statement already given of the mode of con- ducting the assay by the wet method, only such instructions have been given as were necessary for its full comprehen- sion, and everything that might call away or FIG. 126. fatigue the attention has been omitted. Never- theless, here it will be convenient to describe some methods of abridging the necessary mani- pulations, supposing that ten, or at least five, assays are made at once. Bottles. It is necessary to have these all, as nearly as possible, of the same height and diameter. They are marked progressively on the shoulder, as are also their stoppers (fig. 126), thus 1, 2, 3, 4, &c. They are taken successively by tens, in the natural order. 680 THE ASSAY OF SILVER. The stoppers are placed on a support, numbered in the same manner (fig. 127). The support is pierced with ten holes, 127. distinguished in precedence by a mark between the fifth and sixth. Stand. Each ten flasks are in turn placed in a case or stand of japanned tin-plate (fig. 128), having ten compart- ments numbered from 1 to 10. Each of these compart- ments is cut anteriorly to about half its length, so as to FIG. 128. FIG. 129. FIG. 130. allow the numbers, of the bottles to be seen. The same stand serves for all the series, by making the same units correspond : thus, No. 23 of the third series is placed in stand No. 3, &c. When each flask is charged with the alloy, about 10 grammes of nitric acid, 40 C., are mea- sured by a pipette (fig. 125) introduced into the bottles by means of a funnel with a large neck (fig. 129). The whole are then exposed to the heat of a water-bath, to facilitate the solution of the alloy. Water-bath. This is an oblong tin-plate vessel, cal- culated to receive 10 bottles (fig. 130). It has a move- able double bottom, pierced with small holes,, the prin- MODES OF ABE1DGING MANIPULATION. 681 cipal object of which is to prevent the fracture of the bottles by isolating them from the bottom of the vessel, which is immediately exposed to the heat. On the moveable bottom are soldered the cylinders c c, three or four centimetres in height, and above which, at the distance of eight centimetres, is a sheet of tin plate, p p, pierced with ten holes, corresponding to the cylinders, and connected with the movable bottom by the supports, s s. These cylinders, and the sheet of tin plate, are destined to isolate the bottles, F F, one from the other in the bath, and to keep them some time suspended over it, when the water is boiling, before complete immersion. The water-bath may be replaced by a steam-bath ; the bottles will then be sup- ported by a grating above the surface of the water. The solution of the alloy in the nitric acid takes place rapidly, and as it gives rise to an abundant evolution of nitrous vapour, it must be made under a flue having a good draught. Flue. This is represented at fig. 131. C C is a flue resting on a table or support, T T, about 90 centimetres high. The anterior side in the FlG 1B1 figure is removed to show^ the water-bath 5, and the furnace F. The opening, 0, of the flue is closed by the wooden door, p, movable on two eccentric pivots, which keep it up during the solution, and allow it to fall so that the flask may be placed upon it. The nitrous vapour is removed from the bottles with the blower (fig. 123). The hood, ff, prevents the diffusion of the nitrous vapour in the laboratory.. Agitator. Figure 132 gives a sufficiently exact idea of this apparatus, and dispenses with a long description. It has ten cylindrical compartments, numbering from 1 to 10. The bottles, after solution of the alloy, are placed in it in the order of their numbers. The agitator is then placed by the side of the pipette, by which is measured the normal 682 THE ASSAY OF SILVER. solution ol salt, and into each flask is poured a pipetteful of the solution. The bottles are fitted with their stoppers, previously moistened with distilled water (fig. 133); they are then fixed in order with wooden wedges (fig. 134). The agitator is suspended to a spring, 7?, and a rapid alternating movement given to it with both hands, by which the solution is agitated, and in less than a minute rendered FIG. 132. FIG. 133. FIG. 134. as clear as water. The movement is assisted by a spiral spring, B, fixed to the agitator and its stand. The agitation finished, the wedges are removed, and placed in the vacant spaces between the compartments. The agitator is taken from the spring, and the bottles^placed in order on a table prepared to receive them. Table. This table (fig. 135) has a double bottom ; the upper is pierced with ten holes, a little larger than the MODES OF ABKIDGING MANIPULATION. 083- diameter of the bottles, and of such a distance from the lower portion, or false bottom, that the flasks do not rise above its edge, or at least but little. This disposition is to FIG. 135. FIG. 136. protect the silver chloride from the light, for it decomposes in contact with water, and a little hydrochloric acid is produced, which requires for its precipitation a certain quantity of silver nitrate, and so lowers the standard of the alloy. This cause of error is, however, not very great, at least when the light does not fall directly on the chloride ; but it is easy to avoid, and should not be neg- lected. The disposition already pointed out does not at all complicate the process, and is moreover useful, as it pre- vents the fracture or upsetting of the bottles. When but one bottle is operated on, it is placed for agitation in a japanned tin-plate cylinder, which is held as shown at fig. 136. On placing the bottles in their respective places on the table, a brisk circular movement is given to them, so as to remove any silver chloride adhering to the sides ; their stoppers are removed and sus- pended by spring pincers, a a. These are formed of sheet-iron wire (see fig. 137). A thou- sandth of the decime solution is then poured into each bottle, and before this has been completed there will have FIG. 137. <>84 THE ASSAY OF SILVER. formed in the first bottles where there is any precipitate, a well-marked nebular layer about a centimetre in thickness. At the back of the table is a black board, PP, divided into compartments numbered from 1 to 10, on each of which is marked with chalk the number of thousandths of decime liquid added to the contents of the corresponding bottle. The thousandths of salt announcing augmentation of standard are preceded by the sign + , those of silver nitrate by the sign . Lastly, the black board carries a small shelf pierced with holes, 1 1, and these receive the funnels or drain the bottles ; on this shelf also are fastened the pincers for sus- taining the stoppers. Cleaning the Bottles. The assays terminated, the liquid from each flask is poured into a large vessel in which there is always a slight excess of common salt ; and when it is full, the clear supernatant fluid is removed by means of a syphon. (Immediately will be given the means of reducing the silver chloride so collected to the metallic state.) The bottles, to the number of ten, are first rinsed with the same water passed from one to the other, then a second, and then a third time with fresh water. They are then placed to FIG 138 drain on the board just mentioned, and the stoppers are placed in a stand by series of tens (see figs. 138 and 127). It is important to remark, that when a glass has been rinsed with distilled water, care must be taken not to rub it with the fingers, for water poured in such a vessel would always be clouded on the addition of silver nitrate. This effect is due to the chlorides contained in the perspiration, and is of course more to be guarded against in summer. Reduction of Silver Chloride , obtained in the Assay of Alloys by the Wet Method. Silver chloride can be reduced without sensible loss, after haying been well washed, by plunging it into scraps PREPARATION OF PUKE SILVER. 685 of iron or zinc, and adding dilute sulphuric acid in suffi- cient quantity to set up a slight disengagement .of hydro- gen gas. The whole can be left to itself, and in the course of a few days the silver is completely reduced. This point can be easily estimated by the colour and nature of the product, but better still by treating a small quantity with ammonia, which, if the chloride is perfectly reduced, will give no precipitate or cloudiness on treatment with an acid. The chlorine remains in solution in the water combined with zinc or iron. The residue must now be washed ; the first washings are made with acidulated water, to dissolve iron oxide which might have formed, and the following with ordinary water : after having completed the washing as much water as may be left is decanted, the mass dried, and a little powdered borax added. Nothing now remains but to fuse it. The powdered silver being voluminous, it is placed by separate portions in the crucible, in propor- tion as it sinks. The heat should at first be moderate, but towards the end of the operation should be sufficiently high to reduce the silver and slag to a state of complete liquidity. If it be found that not quite all the chloride was reduced by the iron or zinc, a little potassium or sodium carbonate may be added to the powdered silver. The standard of silver thus obtained is from 999 to 1000 thousandths. Preparation of Pure Silver. Take the silver prepared as above, dissolve it in nitric acid, and leave the solution some time in perfect rest in water, to deposit any gold it might contain. Decant the solution, and precipitate with common salt, well wash the precipitate, and reduce it, when the resulting silver will be pure. M. Gay-Lussac here gives a description of a process for the precipitation of chlorine from nitric acid for use in the mode of assay already described ; but as that acid in a state of ordinary purity forms an article of commerce, and can be obtained at most operative chemists, the process will riot be here reproduced. C86 THE ASSAY OF SILVER. Modifications required in the Assay of Silver Alloys containing Mercury. Whenever mercury is present in solution with silver, it is thrown down as insoluble chloride, and the assay ren- dered inaccurate. The presence of mercury in silver can be readily detected by the remarkable change which occurs in silver chloride on exposure to light (viz. blackening) when free from mercury ; but if the smallest quantity of the latter metal be present, no blackening will ensue. This source of error was removed by M. Levol in the following manner : The sample being dissolved, as usual, in nitric acid, it was supersaturated with 25 cubic centimetres of caustic ammonia ; the pipetteful of normal solution was then added, and the excess of ammonia supersaturated with 20 cubic centimetres of acetic acid, and the operation continued in the usual way. Some little time after the publication of this, M. Gray- Lussac examined the above process himself, and very con- siderably simplified it. He says : ' After having confirmed by several experiments the accuracy of M. Level's pro- cess, I thought it might be simplified by adding to the nitric solution of silver the ammonia and acetic acid at one and the same time, but in sufficient quantity to saturate the whole of the nitric acid, both that in combination with the silver and that in the free state. Ten grammes of ammo- nium acetate, were added, with a little water, to the silver dissolved in nitric acid, and the assay finished in the ordinary manner. The quantity indicated by synthesis was found very accurately, although 100 thousandths of mercury had been added/ Finally, M. Gay-Lussac found that 10 grammes of sodium acetate, in crystals, also fully answered the purpose ; and as that is a very cheap com- mercial salt, it is the best adapted for overcoming the diffi- culty in this class of assay, as regards the presence of mercury. METHOD OF ASSAYING SILVER BARS. 087 FIG. 139 Method of Taking the Assay from the Ingot. The ingots are so rarely perfectly homogeneous, even taking as a starting-point the standard 950 thousandths, that the differences remarked between the assays of samples made in different places should rather be attributed to the above cause than to the assay itself. It is important, there- fore, to take a sample in a uniform manner, and from the same depth, on the upper surface of the ingot as on the lower. This condition is per- fectly fulfilled by boring the ingots with a kind of drill, similar to that employed by the smith, and which is re- presented at fig. 139. The ingot, Z, is placed in a copper tray, C\ and in order to retain the borings, which might otherwise be thrown out, the drill,/, is surrounded by a casing, m, which does not impede its motion, and stands freely on the ingot. After a few turns of the drill, the first borings, which are not pure, are removed by means of a feather, and only those following are collected and reserved for assay. If it be desirable to try the lateral faces, it is necessary to employ a pressure screw, to keep the ingot in the position that may be deemed necessary. Method of Assaying Silver Bars adopted in the Assay Offices of II.M. Indian Mints. According to the report of H. E. Busteed, M.D., H.M. Madras Army, refining assay master, Calcutta Mint, the 038 THE ASSAY OF SILVER. silver bars are assayed by estimating the silver as chlo- ride. The cupellation assay is not correct enough to satisfy the sellers and purchasers : the Gay-Lussac assay is not used, because most of the bullion contains mercury, lead, and other base metals ; because a previous cupel- lation assay is required ; because the high temperature of the climate Causes an evaporation of the salt solution ; and because a large number of persons would be necessary on account of the large daily number of assays. The credit is due to Mr. J. Dodd, a former assay master of the Calcutta Mint (and a surgeon in the Madras army), of having overcome those difficulties of manipulation, inasmuch as he modified and simplified them, and in short so systematised the whole practical working of the process as to render its application to the assaying of silver, to any amount, easy, accurate, and economical. The process is technically known as the ' chloride process.' It is thus described by Dr. Busteed in the ' Chemical News ' for November 1871 : The samples (or ' musters ') for assay are, to save time, first approximately weighed by an assistant ; they are then placed (each sample in duplicate) in small shallow saucers of polished copper, and so brought in batches of 40 on a board, containing in numerical order receptacles for the little saucer, to the assay master, who, in the delicate assay balance, exactly brings each sample to the one required weight.* As each sample is weighed, it is transferred from the platinum skiff of the balance to a bottle on the left hand of the assayer, by means of a small copper funnel. The bottles f for this purpose are held in readiness for the musters by an assistant, and, on receiving them, are removed into the laboratory in batches of six. On being taken to the laboratory, they are ranged on a circular platform or turn-table, and there one of the * The amount of this weight will be more particularly referred to farther on. t The chief appliances will be described more fully at the conclusion of the description of the process. THE CHLOKIDE PROCESS. 689 (European) assistants adds by means of a pipette 1^ drm. of nitric acid to each bottle, which are then (without their stoppers) transferred to a sand-bath and exposed to a considerable degree of heat, till solution of the contents is effected. The specific gravity of the nitric acid used is generally 1-200, i.e. in the case of known alloys of only copper and silver, such as the standard meltings, coins, &c. ; but when the nature of the alloy is uncertain, such as bazaar silver, or some sycee (where the presence of mercury may be suspected), a stronger acid of sp. gr. 1*320 is used. It has been found, too, by experience, that the chlorides from fine bar silver agglomerate better when the solution has been effected in the stronger acid. When the samples have been completely dissolved,* the bottles are brought back to the platform, and there each receives through a glass funnel f about six ounces of cold distilled water. There is then added to each bottle, through a glass pipette as before, 1^ drm. of hydrochloric acid, sp. gr. TOGO, which immediately converts the silver present into the characteristic white precipitate of silver chloride, which forms in slow-falling curdy volumes. The stoppers (previously dipped in distilled water) are then carefully replaced, and the bottles are allowed to stand for five minutes. The bottles are next well shaken two and two by the laboratory workman for three or four minutes, till the chloride aggregates and rapidly falls down ; any particles which may remain attached to the neck or upper part of the bottles are washed down by a quick circular motion, and, more distilled water being added to within about two inches of the neck (great caution being observed in removing and returning the stoppers), the bottles are then allowed to rest each in its assigned place on the platform for four hours. * A slight residuum of gold, as a black powder, is very generally seen. t The portion of this which enters the neck of the bottle is protected or sheathed, with an inch of india-rubber tubing, to prevent chipping, if struck against the neck of the bottle. Y Y 690 THE ASSAY OF SILVER. At the expiration of that period, the clear supernatant liquid (blue-coloured when copper is present) is removed by a glass syphon, which is lowered to within an inch of the deposited chloride, the greatest care being taken that none of it is drawn up into the syphon. As each platform is made to revolve on its centre, according as each bottle is syphoned, the operator sitting in one place brings the platform round till the next bottle in order gets under the syphon, which is thus in turn lowered into each. The fluid escapes from the long leg of the syphon through a funnel fitted in the table to ajar placed underneath. After the first syphoning, the bottles are immediately filled again with distilled water, and each gets a quiet cir- cular motion for a few moments, and the precipitate is again allowed to settle as evenly as possible ; this time it will be sufficient to allow them to rest for two hours, when they are again syphoned as before and the stoppers returned. Under ordinary circumstances these two washings are sufficient ; but if the silver is evidently ' coarse,' a third or fourth washing is similarly given. When it is considered that the chlorides have been sufficiently washed, the bottles are placed for half an hour in a reclining position on their platforms ; this causes the chloride to fall and settle to one spot, and renders its removal from the bottles more easy. Meantime a pneumatic trough has been got ready, capable of containing a batch of twenty inverted bottles ; the trough is filled with distilled water ; for each bottle there is placed on the floor of the trough a small porcelain saucer holding a little Wedgwood crucible or cup, each numbered to correspond to the bottles. A laboratory workman then removes the stoppers from the bottles, and hands them one by one to an assistant at the trough, who, placing his forefinger over the mouth of each bottle, inverts it over its corresponding cup, and does not remove his finger till the neck of the bottle has passed down through the water and well into the cup ; then the finger being taken away, the bulk of the chloride falls by its own weight to the bottom of the cup. THE CHLORIDE PROCESS. 691 The bottle is held in the position by two rings, one (the larger) above the other, which are fixed to the sides of the trough ; this arrangement retains each bottle in situ, at the proper slant, and admits of the operator gently revolving or slightly raising the bottle with his left hand, while with the right he patiently taps the bottom and sides till the whole of the chloride has been dexterously got out ; the finger is then again placed over the mouth, and the bottle raised up through the rings and handed (mouth upwards) to the assayer, or to the supervising assistant standing by, who carefully examines it to see that every particle of chloride has been dropped into the cup. When this part of the manipulation has been neatly done, none of the chloride falls over into the saucer which is placed as a precautionary measure under each cup. When the chloride falls into the cup, it is in an uneven lumpy state, and not in a favourable condition for being uniformly dried ; it has, therefore, next to be broken up. For this purpose the cups (containing the chlorides, and water to the brim) on removal from the trough are taken in batches on a tray to an assistant seated at a steady table, who first carefully decants off about half the water, and then with a finely polished glass rod (four inches long and one-third of an inch thick) gently stirs and beats the lumpy precipitate, while revolving the cup on the table ; this causes it to lie evenly and loosely at the bottom of the cup as a purplish grey powder, not too fine. He next washes the rod over the cup with distilled water from a drop bottle, lest any of the chloride may be adhering to it, and sprinkles a drop or two from it on to the surface of the water in each cup, so as to cause to sink any minute particles that may happen to remain floating. He then, after an interval of ten minutes, drains off about three-fourths of the supernatant water, which he lets run down the rod into a vessel near him, arid with a tap or two of the rod on the outside of the cup to still further loosen the deposit, this part of the manipulation is concluded. The crucibles are next taken to the drying chamber, TT 2 092 THE ASSAY OF SILVEE. where a steam bath is ready to receive them ; on the per- forated upper plate of this they are ranged, and allowed to remain for about an hour. This gradually, and without spurting, frees the chlorides from moisture, which may be known by their caking, i.e. leaving the sides of the cups round the edges and forming at the bottom of each a loose cake, resembling somewhat a gun-wad. The crucibles are then arranged on a hot-air plate and there exposed to a temperature of between 300 and 350 (F.) for about two hours, till thoroughly dried, when they are ready for weighing.* When the above manipulations have been carefully and satisfactorily gone through, each little cup contains an unbroken, tolerably firm cake of chloride of silver, lying unattached, which admits of being easily grasped with a pair of forceps, and cleanly lifted out of the cup and conveyed to the skiff of the assay balance in which it is weighed. The cups are generally brought from the laboratory to the assayer at the balance in batches of eight or ten. A 6 standard,' synthetically pre- pared of pure silver and copper, and an assay pound of pure silver, are introduced with each day's set of assays, and their chlorides dried with the others, and the analysis of them verified before weighing the rest. Occasionally these ' checks ' are also fused and weighed in a porcelain capsule, but the weight found never differs from that of the chloride merely dried as above. Once or twice a month, the silver is recovered from the accumulated chlorides, which are well pounded in a mortar and brought to a powder and then mixed with a proper proportion of chalk and charcoal, and put into a wrought-iron crucible and reduced with heat. The metallic silver so recovered is transferred to the mint. Under the circumstances of the solution and of the precipitation as detailed above, should any gold happen to be present in the sample operated on, it is not dissolved, and therefore becomes entangled with the precipitated * The chlorides are weighed warm, to obviate the risk of their absorbing moisture ; a precaution especially necessary in the heavy monsoon weather in this country THE CHLORIDE PROCESS. 693 silver chloride and dried and weighed with it, and ac- cordingly comes to be regarded and valued as silver. In this the chloride process resembles that by cupellation, which likewise takes no distinguishing cognisance of gold; and both these processes contrast in this respect with the volumetric one, which is a rigid analysis for silver alone ; so that, strictly speaking, an assay conducted by either of the first-named methods ascertains the proportion present of ' the precious metals,' i.e. silver and gold.* Should mercury be present it does not interfere with the result, when the solution has been effected in excess of nitric acid with strong heat. Thus the mercury be- comes per oxidised, and hydrochloric acid forms no pre- cipitate in solutions of mercuric salts ; any mercuric chloride resulting from the combination would remain in solution, and be washed away in the course of the process. Should lead happen to be present, hydrochloric acid gives no precipitate in a dilute solution, the lead chloride being soluble in a certain proportion of distilled water ; but even were the proportion of lead to silver tolerably large, and the lead chloride happened to be thrown down, the repeated washings would dissolve and get rid of it. With regard to the weight of the small portion taken to represent the mass, the system prevails in the Indian mints of taking samples for assay by granulating a small portion of the contents of each melting-pot : when the metal is in a thorough state of fusion and has just been well stirred, a small ladleful of the molten metal is quickly poured from a tolerable height into a vessel of water, and .the granules so formed received on a strainer, lifted out, and perfectly dried. f The weight of this specimen repre- * Much of the silver which finds its way to the Indian mints is rich in gold; for instance, sycee contains on an average somewhat about 12 grs. in the troy pound. This in minting operations is considered as silver, and as such it enters into the coinage. There being as yet no refineries established here, through which such silver could pass to the mechanical departments of the mints, the silver coins made during a period when a heavy importation of sycee had been worked up, contain as much as 4 or 6 grains of gold in every 32 tolas or 1 pound troy. t The introduction into the Calcutta Mint of this system of taking musters is attributable to Dr. Boycott, late Assay Master, and to Dr. Shekelton, 694 THE ASSAY OF SILVER. senting each pot was first fixed at 24 grs., technically called the ' assay pound : ' this in the case of pure silver yielded 31-87 grs. of silver chloride, while the same quan- tity of Indian standard silver (which is -j^-ths silver plus T Vth copper= 916-66 in 1000) yielded ^th less, or 29-21 grs. : on the weight of chloride being ascertained in each case, a table which was calculated and prepared for the purpose was referred to, and the equivalent fineness assigned to the -Jdwt., plus the odd grs., when any. But when it became desirable to prepare for the decimal form of notation, a number more convenient than 31-97 was looked for to represent purity or 1000, and 25 was fixed on as a desirable starting-point, particularly as the quan- tity of pure silver yielding that amount of chloride, viz. 18*825 grs., was quite large enough to represent each pot.* The weight, therefore, of the ' assay pound ' in use at present is 18*825 grs. This produces (with chlorine) in the case of pure silver 25 grs. of silver chloride.f But to obviate the necessity of constant reference to a calculated table to find the equivalent in pure silver of the amount of silver chloride found in each case, it was inge- niously arranged to stamp each of the assay weights not with its actual weight, but with the figures representing the proportion per mille of pure metal which such a weight of chloride so found corresponds to : thus, sup- who, by a number of interesting experiments, satisfied themselves that samples so taken represent the mass of mixed metal to be valued much more fairly than samples of the same mass cut or gouged from it after it has been poured and allowed to cool in the ingot moulds, where a partial separation of the copper from the silver seems to take place ; the result being, according to the above experiments, that in the case of ingots cast in upright moulds all the outside is much below the average fineness of the mass on assay, and the centre much above it. This refers to alloys of silver and copper mixed in or about the proportion of ' standard.' According to M. Levol, however, it would appear that when an alloy of silver and copper in which the proportion of the latter is very high (viz. over 28 per cent.) has been melted, poured, and allowed to cool, an opposite result to the above is found, viz. the outside of the ingots is above the average fineness. An assay, therefore, from a granulated sample must give a much nearer approximation to truth than one from a cut sample. * The average weight of the contents of each melting-pot is 12.500 tolas, or about 390 pounds troy, so that the specimen taken to represent this is but about the 119,000th part ; each sample is assayed in duplicate. t The basis for these numbers was founded on the proportion in which, according to Turner, silver combines with chlorine, viz. 100 parts with 32'80. THE CHLORIDE PEOCESS. 695 posing a melting of five-franc pieces was being assayed, and the chloride resulting from the assay pound operated on weighed 22-5 grs. (showing the actual pure contents in the sample to be 16-94 grs.), instead of referring to a table to see the equivalent per mille-age of silver, that weight which is actually 22*5 grs. has 900 marked on it, and the assayer simply reads the c touch ' from it. Accordingly, the assay weights are as follows : Actual weight in grs. Figures marked on the weights 25-000 ....... 1000-00 22-910 (Std.) 916-66 22-500 900-00 20-000 . . ... . . 800-00 17-500 . . ... . . . 700-00 15-000 600-00 12-500 . 500-00 10-000 400-00 7-500 . 300-00 5-000 200-00 2-500 100-00 1-250 . ' . .... 50-00 1-000 40-00 0-750 30-00 0-500 20-00 0-250 10-00 0-125 5-00 0-100 4-00 0-075 . 3-00 0-050 2-00 0-025 1-00 Assay lb., weight=18*825 grs. An ordinary day's work consists of eighty assays,* estimating imported bullion to the value of four lacs of rupees, and standard meltings and coins to the value of five lacs. But on emergencies, in time of heavy pressure, by working extra hours, as many as 164 assays have been daily conducted, estimating to the value of eight lacs of rupees, and standard coins and meltings to the value of fourteen lacs. Such is an outline of the method of assay worked on a large scale ; of course successful results from it cannot be expected unless each step in the manipulation be con- ducted with great care and accuracy, and only then after much practice and experience. * Exclusive of any gold assays which may be going on. 698 THE ASSAY OF SILVER. The natives of this country possess great aptitude in acquiring the skill and consequent lightness of touch so essential for delicate manipulation ; this, added to their characteristic patience, makes them admirable subordi- nates in an assay laboratory, under judicious supervision ; moreover, their labour is cheap ; so that, on the whole, the process seems to be especially suitable for an Indian mint. When bar silver is imported from the Continent, the assays of it, made" here, almost invariably correspond most closely with those previously made of it in Paris by the volumetric method. But were further proof needed of the practical accuracy of this system, it is to be found in the very close proximity to the legal standard at which the large Indian coinage has been maintained for many years, as annually reported by the assayers to the Eoyal Mint of Great Britain, who test the fineness of the Indian pyx coins by the French process. Without this me tBod (improved and made more per- fect, as it has been, in the hands of successive assay officers), it would have been very difficult for the assay establishments of the Indian mints to have dealt with, in the same time and with the same accuracy, the immense importation of silver to India during "the last fifteen years. In the single year 1865-66 there was poured into the Indian mints, and manufactured into coin, silver alone reaching in value to the prodigious amount of over four- teen millions sterling. The system which enabled the assay officers to value such a rapid and heavy influx with accuracy, and with satisfaction to the importer on the one hand, and to the mint (the buyer) on the other, and to faithfully maintain the immense resulting coinage close to legal standard, has been put to a severe test. If success be the criterion of merit, the twenty years' large experience of this method gained in the Indian mints goes to show that it is worthy of a yet wider field of utility. THE CHLORIDE PROCESS. 697 Apparatus and Appliances required. (1.) The bottles used in this process are of thin (but strong) white glass, and contain about 12 fluid oz. ; about 6 inches in height and 2^ inches in diameter at the bottom, which should present a perfectly even, level floor ; they are without any abrupt shoulder, but become gradually pyramidal from about half-way up to the neck ; this shape favours the easy dropping out of the chloride. The neck is about one inch in length, polished on its inner surface ; the stoppers are of ground glass, polished, with globular heads, and are made to fit with the utmost accuracy and smoothness. The bottles and stoppers are numbered, to correspond with the number on the muster board and also on the cups. (2.) The ' cups' are Wedgwood crucibles, smooth and thin, about 1^ inches in height, 1^ inches in diameter above, and a little less than 1 inch in outside diameter at the bottom. The floor should be perfectly level, and neither it nor the sides should present any roughness likely to retain the chloride. The cups are all numbered. (3.) The porcelain saucers are shallow, f of an inch in depth ; the upper diameter is about 4 inches, the lower 2-J inches. (4.) The turn-table is a circular board of about 3 feet in diameter, fenced by a brass railing (or by a simple ledge) ; its centre is occupied by a raised platform about 2 feet in diameter, between which and the rail the bottles (20 on each) stand, the round outer edge of the platform having semi-lunar niches cut in it, into which the bottles fit ; opposite to each niche on the platform is a little concavity in which the stoppers rest when not in the bottles. Each turn-table is made to revolve on its centre in either direc- tion, and is raised about 6 inches above the long general table on which all are supported ; close to each a funnel is fitted into the lower (supporting) table for conducting away the fluid syphoned from each set of bottles. (5.) The trough is a basin of cast iron (painted) ; it 698 THE ASSAY OF SILVER. may be oblong or round, raised to about the height of 3 feet from the ground ; when round and large enough for twenty bottles, space and distilled water may be economised by having a platform in 'the centre. This is convenient for resting the bottles oh after the chlorides have been got out. A trough of this kind may be about 2^ feet in diameter, having a space 7 inches broad and 4 deep all round between the circumference of the basin itself and the outer edge of the island platform. Into this space is poured distilled water to the depth of 3 inches. From the rim of the trough hang as many brass supports as there are bottles to be inverted ; there are two circular clasps connected at the back to a bar common to both ; one, the larger, is 1-| inch above the smaller and lower one, which is under water ; they are open in front (or towards the centre of the basin) to about J of an inch in width. The openings of both are in the same line, owing to the lower (smaller) segment being projected towards the centre by an abrupt curve in the connecting bar, by which they hang from the brim. This arrangement receives and fixes the inverted bottles in the required position. The distilled water is removed from the trough by the withdrawal of a plug. These troughs are sometimes made to revolve on the centre. (6.) The drop bottle used for washing down the glass rod when breaking up the chlorides, and for sprinkling the surface of the cups, is small-sized, round, so as to be easily grasped ; it holds about 6 oz. The stopper is hollow, with two small tubes leading from its head, one opposite to the other. Glass is so liable to break or chip, that a hollow silver stopper is now generally substituted. (7.) The steam-bath is simply a square vessel made of sheet copper, between 3 and 4 inches deep, the top or upper plate of which has a number of circular openings about two-thirds of the diameter of a Wedgwood crucible. There is also a steam escape pipe leading from the centre below to about a foot in height. They are of various sizes, to contain from 10 to 150 pots : they are raised or moved by two lateral handles. EFFECT OF BISMUTH ON SILVER. G99 (8.) Hot-air plate of thin sheet iron bored with holes for the reception of the crucibles, raised by iron feet about 1-| inch above the furnace plate. It is furnished with a square tin cover, which fits over it. This is provided with lateral apertures for the escape of heated air, and with a tube from its roof for the reception of a thermometer. The drying furnace on which the above rest is sur- mounted by a hood, the glazed door of which slides up and down by weights and pulleys ; the plate is heated by means of gas jets ; it has a good draught to carry off the nitrous fumes, as on it the musters are dissolved in the first instance on a sand bath. (9.) The forceps for removing the cake of chloride from each cup to the skiff of the balance should not be too sharp in its grasp ; it is much improved by having the blades tipped for about an inch from the points with platinum about ^ inch in width. (10.) It is a convenience to have the assay weights arranged in a set of ivory compartments in the weight box ; on the floor of each compartment are engraved the figures corresponding to those engraved on the weight which occupies it ; by this means the assayer has merely to glance at his weight box to see what weights are in the pan of the balance, and to read off the ' touch ' when each chloride is counterpoised. Effect of Bismuth on the Ductility of Silver. The effects produced by small quantities of bismuth on the ductility of silver have been carefully investigated by Surgeon-Major J. Scully, Assay Master, Calcutta. The following are extracts from an elaborate paper which he published in the Chemical News for 1887 : It is well known that alloys of silver and bismuth, in certain proportions, are brittle. In Dr. Percy's valuable work on Metallurgy (Silver and Gold Part I.) it is stated that alloys of silver w^ith bismuth, in the proportion of 50 per cent, and 33 per cent, of the latter metal, are brittle ; while an ore of silver and bismuth, called 700 THE ASSAY OF SILVER. Chilenite, in which bismuth occurs only to the extent of 14*4 to 15*3 per cent., is said to be malleable. The least amount of bismuth, however, which will injuriously affect the ductility of silver, for example, in such an operation as the lamination of silver bars for coinage, does not, so far as I am aware, appear to have been experimentally investigated. It may here be mentioned at the outset that an alloy of silver and bismuth may, by careful hammering, be extended considerably, so as to pass muster as malleable ; although, if subjected to lamination by means of steel rolls, the same alloy will crack at the edges and thus show a deficient ductility, as compared with pure silver or some silver-copper alloys. It is to the deficiency in ductility, as tested by rolling, of silver con- taining only very small proportions of bismuth that I here wish to call attention. My attention was first prominently directed, about two years ago, to the injurious effects caused by small quan- tities of bismuth in silver by the circumstance that some silver bullion, in the shape of English refined bars of as high a fineness as 990 per mille, proved so brittle as to be unfit for mintage. Attention was first attracted to this matter by the peculiar behaviour under assay of the granulated samples taken from this silver after melting. The appearances noticed under assay will be referred to presently, but they led to the bullion being at once tested for brittleness. A bar, about 21 inches long, 2-25 broad, and 1 inch thick, was hammered out at one end without cracking, but on being passed through the rolls it cracked badly at the edges and was pronounced to be ' brittle.' in the Mint sense of the term. The bullion was then re- melted in five plumbago pots, and a partial refinement of it attempted in the ordinary way with nitre, about 8 to 10 Ibs. of this salt being used for each pot. The resulting silver bars were not appreciably improved by this treat- ment ; hammering again proved an inconclusive test, but a bar of the size I have mentioned broke in two by merely dropping on the floor of the melting-room. In the meantime the assay had shown that the brittle EFFECT OF BISMUTH OS SILVER. 701 bullion contained bismuth, and that this was the only substance present likely to be the cause of brittleness. The Indian process of assaying silver has already been described (p. 688) by Dr. Busteed. The main features of the process may here be briefly recapitulated for the purpose I have in view. A fixed weight of the silver bullion to be assayed is dis- solved in an assay bottle, by means of nitric acid aided by heat ; the solution is diluted with water and an excess of hydrochloric acid is added, to precipitate all the silver pre- sent as chloride. The silver chloride having been caused to aggregate and settle by vigorous shaking, the bottle is filled up with water, and the supernatant fluid is subse- quently syphoned off, to remove all the now dissolved matter which may have been contained in the bullion. Under these conditions of solution, precipitation, and dilution with water, chemists will readily understand that even a small trace of bismuth, if it be in the silver, will reveal its presence by the partial formation of insoluble oxy chloride of bismuth. Now, in the assay of the brittle bullion under consideration, solution in nitric acid had been readily and completely effected by the aid of heat ; antimony and tin were consequently absent. After the addition of water and hydrochloric acid, however, the solution in the assay bottles could not be cleared by shaking ; the bulk of the silver chloride collected at the bottom of the bottles, but the supernatant fluid remained turbid. Tin and antimony being excluded, only two metals could produce this result in the wet assay of silver namely, mercury and bismuth. To determine which of these is the interfering metal it is only necessary to note the effect of solar light on the silver chloride formed ; when mercury is present the silver chloride maintains its pure white colour unaltered, while in the presence of bismuth the chloride immediately acquires the well-known purple colour under the influence of daylight. Our assays, then, being turbid after precipitation and yet the silver chloride blackening readily under the influence of 702 THE ASSAY OF SILVER. daylight, it was evident that bismuth was present. The turbidity produced was due to the partial formation of bismuth oxychloride ; and this compound diffusing itself in its characteristic manner through the solution had broken up part of the silver salt into very fine powder, so that some hours had to elapse before the supernatant fluid cleared by the gradual subsidence of both bismuth oxy- chloride and the finely divided silver chloride. The assay was of course thus rendered unreliable, since the silver chloride to be weighed, and on which the calculation of the fineness rested, was contaminated with bismuth oxy- chloride. A cupellation assay of this bullion was at once had recourse to for ascertaining its fineness. So far, then, this tender of silver bullion seemed to establish the following points: 1. Silver bullion of as high fineness as 990 per mille is rendered unfit for coinage purposes by an amount of bismuth which, in this particular case, could not possibly have exceeded 1 per cent., and was probably less than that proportion. 2. Hammering a bar of silver bullion is not a good test for detecting brittleness, as far as Mint purposes are concerned. 3. The toughening of silver bullion 990 fine, and con- taining only a small amount of bismuth, by the aid of nitre in plumbago melting-pots, is not readily affected. 4. The presence of a trace of bismuth in silver of high fineness is immediately detected in the ordinary course of assay by the Indian method, but this bismuth interferes with the perfect accuracy of the results obtained by that process. A comprehensive research seemed, therefore, called for to elucidate the whole subject, and the necessity for this investigation has since been emphasised by the fact that silver bullion contaminated with bismuth has frequently found its way to the Mint since its first discovery here. The points to be investigated seemed naturally to group themselves under the following heads : 1. Is our ordinary wet assay of silver susceptible of EFFECT OF BISMUTH ON STLVEE. 703 such easy modification as will enable us to obtain perfectly accurate results by it, in presence of bismuth, without having recourse to the confessedly less accurate assay by cupellation ? And, how may small quantities of bismuth in silver be readily estimated with the despatch indispens- able for mint operations ? H. What is the smallest amount of bismuth in silver that will render it unfit for coinage, when present in bars of the Indian standard fineness of 916*6 ? And, what is the amount of bismuth that may be tolerated in such bars without materially injuring the ductility ? III. How is silver bullion containing bismuth, which may be tendered to the Mint, to be dealt with, supposing that establishment accepts any metal that is brittle ; and how is the presence of bismuth in refined bars to be accounted for ? 1. As the purity of the bismuth to be used in the experiments now to be detailed was a matter of first im- portance, I may briefly mention the steps taken to insure the purity of the metal. Eefined bismuth was dissolved in nitric acid, precipitated as basic nitrate by diluting largely with distilled water, the nitrate digested in solu- tion of caustic potash, and then well washed, dried, and reduced by heating with charcoal in a clay crucible. A series of synthetical assays, made by dissolving together pure silver and pure bismuth, the latter in the proportion of from 1 to 5 thousandths, showed that our ordinary process of assay, under such conditions, gave unreliable results, there being a surcharge, or higher report than should have been obtained, which varied from 0'7 to 2- 7 mils., when the proportion of bismuth was from 3 to 5 thousandths. A modification in our process of assay was evidently required if it were to be used for estimating the fineness of silver bullion containing bismuth ; and the neces- sary steps to this end were, after repeated experiment, found to consist in adding the smallest possible amount of hydrochloric acid for the precipitation of the silver, and increasing the amount of nitric acid in which it was first dissolved. We use ordinarily for the precipitation of an 704 THE ASSAY OF SILVER. assay pound of silver 5-4 c.c. of hydrochloric acid, of sp. gr. 1075 ; but 2*5 c.c. of acid of this strength suffices for the complete precipitation of an assay pound of even fine silver ; so that we have here at once a means of diminishing the tendency of any bismuth in the silver to form insoluble oxy chloride. If, in addition to diminishing the amount of hydrochloric acid, we added a considerable excess of nitric acid to the solution (which acid would not in any way interfere with the silver chloride formed), all risk of the partial formation of insoluble bismuth salts seemed removed. This in fact has proved to be the case, and the successful modified process for the assay of silver containing bismuth is as follows : The assay pound of silver bullion containing 'bismuth is dissolved in 5-5 c.c. of nitric acid, sp. gr. 1200, with the aid of heat, about 5 ounces of water are added, and then 10 c.c. of nitric acid, sp. gr. 1320. The silver is now precipitated by the addition of 2-5 c.c. of hydro- chloric acid, and after vigorous shaking the supernatant fluid will be found perfectly clear, and it will remain so when the bottle is filled up with water, all the bismuth present being in solution. Whenever samples of silver now show the presence of bismuth during the assay, a fresh set is taken up and worked by the modified process, the delay thus caused not amounting to more than a few minutes. It may be mentioned here that all our assays are reported to one-tenth of a millieme (0*1 per mille). Having thus ascertained the presence of bismuth in silver bullion, and put in practice a modification of the assay process which renders us indifferent to its presence, it is still of importance to ascertain the exact proportion of bismuth which is present in the bullion, and, to be of practical use for mint work, this estimation must be effected rapidly and as simply as possible. The ordinary directions given for the separation of bismuth in the pre- sence of silver, by first removing the latter as chloride and then precipitating the bismuth as carbonate, do not, I find, give accurate results when silver is present in such EFFECT OF BISMUTH ON SILVEK. 705 overwhelming proportions as obtain in the cases under consideration. I have therefore adopted the following plan, which a number of synthetically prepared solutions have proved to give quick and good results, though sometimes the amount of bismuth present is very slightly underestimated. The ordinary silver assay having given a rough visual estimate of the amount of bismuth likely to be present, enough of the bullion is taken to yield a fairly weighable amount of bismuth oxide in the final result. The bullion is dissolved in a small amount of nitric acid, the solution carefully diluted, and an excess of ammonium carbonate at once added, the precipitation being aided by heating. The carbonates of silver and copper at first formed are re-dissolved, and the carbonate of bismuth after a time settles completely at the bottom of the beaker. The con- tents of the beaker are then passed through a filter, of which the weight of ash yielded by incineration is known, and the carbonate of bismuth on the filter washed quite free of all traces of silver. The filter is then dried, its contents transferred to a porcelain crucible for ignition, the filter-paper being ignited separately, treated with a drop or two of nitric acid to re-oxidise any bismuth oxide reduced by contact with the carbon of the filter, and the ash added to the crucible. From the weight of bismuth oxide thus found, after deducting the weight of the filter ash, the amount of metallic bismuth present in the sample of bullion taken for analysis can be at once found. There are only two metals likely to interfere with the accuracy of the process here described namely, cadmium and lead ; the carbonates of both these metals being as insoluble in excess of the precipitant employed as bismuth carbonate. Cadmium is very unlikely to be found in silver bullion and its consideration may be neglected, but if the presence of lead is suspected the carbonate filtered from the silver solution is dissolved in nitric acid, evapo- rated down with the addition of sulphuric acid, and the lead sulphate formed (if any) collected and weighed in the usual way. The bismuth is again precipitated as z z 706 THE ASSAY OF SILVER. carbonate and treated as before directed. Many experi- ments have been made with synthetically prepared mix- tures of silver, copper, lead, and bismuth, the latter two metals being in very small proportion to the silver, so as to imitate the composition of some refined bars. Ullgreen's plan for the separation of the carbonates of lead and bismuth, by dissolving them in acetic acid and then pre- cipitating the bismuth by means of a lead rod, does not- work satisfactorily and requires too long a time for the precipitation. II. As it seemed likely that a large number of ex- periments would be required to estimate accurately the smallest amount of bismuth that would injure the ductility of our coinage alloy, and the still smaller proportion that would not sensibly affect this ductility, it was determined to begin the inquiry by a number of laboratory experi- ments on small bars of silver, before trying the effects of bismuth on ordinary coinage bars and with the procedure for lamination carried out in the Mint. These laboratory experiments were made in the following way : Pure silver prepared for assay check purposes, or an alloy of silver and copper of which the exact composition had been estimated by assay, was melted in a clean plum- bago crucible under charcoal. When the metal was in fusion the necessary amount of bismuth was rolled in a piece of paper, carried down at once to the bottom of the silver bath, and then thoroughly mixed with the silver by stirring. The calculated composition was confirmed by assay of the silver. When this mixture had been accom- plished, the contents of the crucible were poured into an open iron ingot mould, and after cooling either quickly by plunging the casting into water or slowly in contact with the mould the bar so cast was tested for brittleness by hammering and rolling. The bars cast were of two sizes, one set being 3'75 inches long, 1-125 broad, 0-375 thick, and weighing about 6*2 troy ounces ; and another set 2-69 inches long, 1-125 broad, 0-25 thick, and weigh- ing about 4-1 troy ounces. When reduced to the fullest extent by rolling, these bars were converted into straps EFFECT OF BISMUTH ON SILVER. 707 about 0/015 inch in thickness. In laminating them they were twice annealed, first after having undergone four pinches in the rollers, and again after the tenth pinch from the beginning. Similarly shaped bars of silver, without bismuth, were occasionally laminated in the same way to obtain a sure means of comparison. Before any result was accepted as to brittleness or its absence, the bar under experiment was always re-melted and tried at least a second time. The number of experiments in this series amounted to fifty -three, and the following is a summary of the results obtained : Fine silver when alloyed with only 1 per mille (one- thousandth part of its weight) of bismuth, and the casting rapidly cooled by plunging it into water as soon as it has set, has its ductility, as tested by lamination, sensibly but slightly impaired, the straps resulting from rolling having slightly jagged edges. When the proportion of bismuth is increased to 2, 3, 4, and 5 per mille, the plan of cooling remaining the same, the raggedness of the edges of the straps was somewhat increased, but not very markedly. If, however, the casting was allowed to cool down com- pletely, but very slowly, in contact either with the mould or a stone floor, the results were very different. Under this condition of cooling, a bar composed of fine silver with 4 per mille of bismuth was completely brittle ; it was readily broken, and its fracture was strongly crystalline. On laminating it small cracks appeared all over the sur- face on the second pinch, the bar emitting a crackling sound under the rolls, much like the ' cry ' of tin, and on the fourth pinch the bar cracked deeply at the edges. This remarkable effect on the molecular structure of this alloy of silver and bismuth, as due solely to the mode of cooling the casting, was repeatedly verified on the same metal by re -melting and cooling rapidly and slowly alter- nately. The case seems analogous to that of bronze, where slow cooling of the alloy after casting is said to make it hard and brittle. Fine silver with 6 per mille of bismuth, rapidly cooled, was distinctly cold-short and crystalline on fracture ; the z 2 708 THE ASSAY OF SILVEK. bar cracked on the surface at the fourth pinch. With 7 per mille of bismuth these evidences of diminished ductility were slightly more pronounced. With 8 per mille of bismuth the silver was still more brittle, the bar broke readily when hammered, and cracked all over the surface on the fourth pinch from the rolls. With 9, 10 r and 11 per mille of bismuth, the bar of silver could be readily broken in two by merely striking it against the edge of an anvil, the fracture was coarsely crystalline, and the bar, in one case, proved to be very red-short, a mere tap from the tongs sufficing to break it in two when heated for the purpose of annealing. Although these bars were so very brittle, it was still possible to roll them into thin straps after careful annealing ; but the edges of the straps so produced were deeply jagged and indented by cracks. These bars also all emitted the peculiar crack- ling noise under the rolls which has before been men- tioned. An alloy containing 990 parts of silver and 10 of copper then had added to it successively 1, 2, 3, 4, and 5 per mille of bismuth, the castings being rapidly cooled. The remarks already made with reference to fine silver alloyed with the same proportions of bismuth would apply here almost exactly that is to say, the bars were rolled out to a thickness of 0-015 with somewhat ragged edges, so that although ductility, as thus tested, was impaired, it was only slightly so. With 6 per mille of bismuth (fineness of metal on assay 983-9) the edges cracked a little, and, after annealing and rolling out, the strap had decidedly jagged edges and was split for some distance at one end. The bars containing 4, 5, and 6 per mille of bismuth were now re-melted and allowed to cool slowly and completely in the mould. They were all found to be highly brittle, broke easily under the hammer the fracture being granular and not crystalline and on being rolled they cracked badly, all over the surface and at the edges, on the first or second pinch ; in one case the bar broke in two on the second pinch. That these very different results were again solely due to the manner of EFFECT OF BISMUTH ON SILVER. 709 cooling was proved by re-melting and rapidly cooling the castings, when the same metal proved comparatively ductile, as first stated. Silver of the Indian standard of 916*6 per mille (the rest being copper) to which 2 per mille of bismuth was added, gave on lamination straps with slightly jagged edges and proved to be red-short. With 4 per mille of bismuth the bars showed a few surface cracks on being rolled, and the resulting straps had decidedly jagged edges. Slow cooling of these castings did not affect their ductility, thus showing a marked contrast to what had been observed in the case of fine silver and the alloy con- taining only 10 per mille of copper. When the amount of bismuth was increased to 5 per mille, the copper present remaining at 834 per mille, the bars were de- cidedly brittle and cracked readily on hammering the fracture being again granular, and not crystalline as in the case of fine silver. On lamination both surface and edge cracks developed after four pinches from the rolls, and in annealing one of these bars the whole surface blistered considerably, no doubt owing to the temperature having been carried a little too high. Standard silver with 10 per mille of bismuth, reducing the fineness as ascertained by assay to 906*6, was very brittle, the bars breaking easily under the hammer, and on the fourth pinch from the rolls splitting and cracking all over the surface. In the course of these latter experiments it was ascertained that, with from 83-5 to 70 per mille of copper present, slow or rapid cooling of silver alloys containing bismuth made no appreciable difference in their ductility. The foregoing experiments having furnished some in- formation as to the amount of bismuth that might be expected to injure our coinage alloy, it was now decided to test that point practically, by operating on coinage bars subjected to the regular procedure for the manufacture of rupees in the Calcutta Mint. The experiments made in this connection were fourteen in number. The bars used here for coinage weigh about 253 troy ounces, and are about 20 inches long, 2-25 broad, and 1 inch thick ; they 710 THE ASSAY OF SILVER. are cast in vertical iron moulds. In lamination they are first reduced by eleven pinches to a thickness of 0'23 in. ; they are then annealed, and finally reduced by twelve additional pinches to a thickness of 0*06 inch. A number of bars, poured from a pot of which the contents had proved on assay of a granulated sample to be 916-6 fine, were selected for the experiments, and as a preliminary step one of the bars was laminated to test its ductility, It rolled out with smooth ' wire ' edges, and indeed its ductility was beyond suspicion, as it resulted from a melting of good coins. Another bar of the same batch was now melted and 1 per mille of bismuth added to it, the result of the addition being checked, in this and all following cases, by the assay of a granulated sample of the metal, taken after thorough stirring. At the eighth pinch both edges of the lower half of this bar began to crack, and at the eleventh pinch these cracks extended towards the middle line of the strap for about a quarter of an inch, and occurred at about every half inch of the edge. After annealing, and in the subsequent lamination to a thickness of 0-065 inch, these cracks increased con- siderably in number, but did not become sensibly deeper. The strap as finished was pronounced unfit for coinage purposes ; for although two blanks could have been cut from its width, the edges were too jagged to admit of the blanks being obtained exactly along the line from which it was desired to cut them this position being attained by means of a fixed lateral guide against which the edge of the strap had to be maintained in cutting. With 2 per mille of bismuth the results obtained on rolling were not much worse than with 1 per mille. But the side cracks opened out more, and here again it was noticed that the lower portion of the bar (upper and lower here having reference to the casting in upright moulds) was somewhat less ductile than the upper part. With 3 per mille of bismuth (fineness on assay 913'8) the bar began to crack on both edges at the ninth pinch ; at the eleventh pinch there were many cracks quite a quarter of an inch deep, and after annealing the bar these TITRATION OF SILVER IN PRESENCE OF COPPER, ETC. 711 cracks increased at every pinch, so that at the twenty-first pinch the strap was cracked all along both edges very badly. It would only have been possible to obtain one blank from the width of this strap. As it was perfectly clear that no further experiments were required with larger proportions of bismuth, the subsequent trials were made on coinage bars containing 0-5 per mille, 0-25 per mille, and, by dilution of the latter bars with standard silver, to even half and a quarter of the lesser proportion just stated. Here the results were rather discordant ; they appear to have been somewhat influenced by the state of different rolls, and by quick or slow annealing. The general outcome of the tests, how- ever, was that although some of the straps containing the proportions given of bismuth were jagged at the edges, and so would have yielded a diminished percentage in outturn of good blanks, others were not materially worse than the average of straps without any bismuth at all. As a result of this part of the inquiry, it may, I think, be fairly concluded that if our coinage bars contain less than 0-5 per mille of bismuth their ductility will not be mate- rially affected. It must be borne in mind that these results only apply to bars of the size and shape of those experi- mented on, and with the particular treatment in lamina- tion above detailed. With thinner bars and a different method of rolling, different results may be expected. The system of cutting out blanks has also to be considered, for in some mints straps with saw edges are not so pre- judicial as in others. Titration of Silver in Presence of other Metals by means of Ammonium Sulpho cyanide. Mr. J. Volhard gives the following process for the volumetric assay of silver. In a nitric solution of silver to which a soluble ferric, salt has been added, a permanent redness does not appear on the gradual addition of a dilute solution of ammonium or potassium sulphocyanide, until all the silver has been, 712 THE ASSAY OF SILVER. thrown down as a sulphocyanide. If we know how much of the sulphocyanide solution is required to precipitate a known weight of silver we can estimate volumetrically the quantity of silver present in any acid argentic solution, the ferric salt serving as an indicator. For the prepara- tion of the standard solution the author uses ammonium sulphocyanide, though Lindermann prefers the potassium salt ; both in dilute solution are permanent. But the am- monium salt is less frequently contaminated with chlorides, which interfere greatly if present in more than mere traces. The solution may be conveniently made of such a strength that 1 c.c. indicates 1 centigramme of silver. The weighed quantity of the salt is dissolved in water, and diluted in a graduated flask so far that 7*5 grins, (or 8 grins, if the salt appears very damp) may be contained in each litre. For the precipitation of 1 grm. of silver 0-704 grm. of pure ammonium sulphocyanide is requisite ; 0'5 grm. of chemically pure silver is then weighed out, dissolved in 8 or 10 c.c. of pure nitric acid of sp. gr. 1*20, and after the complete solution of the metal the nitrous acid is expelled by boiling or prolonged heating on the sand-bath, and the solution is allowed to cool. It is then diluted with 200 c.c. of water, and 5 c.c. of a cold saturated solution of ammonia-iron alum are added. If the colour of the ferric salt is perceptible, a little pure colourless nitric acid is added till it disappears. The sulphocyanide solution is then added from a burette. At first a white precipitate is produced, which remains suspended in the liquid like silver chloride, rendering it milky. On the further addition of sulphocyanide each drop produces a blood-red cloud, which quickly disappears on agitation. As the point of satura- tion is reached, the silver sulphocyanide collects in flocks, and the liquid grows clearer, without becoming perfectly limpid, as long as a trace of silver remains in solution. As soon as all the silver is precipitated the flocculent precipi- tate quickly deposits, and the supernatant liquid becomes quite clear. The sulphocyanide solution is added by drops till this point is attained, and till a very faint light-brown colour appears in the liquid, which does not vanish on TITRATION OF SILVER IN PBBSBNCB OF COPPER, ETC 713 repeated agitation. The colour is most easily perceived if the liquid is held, not up to the light, but against a white wall turned away from the window. For a repetition of the experiment it is convenient to use a silver solution of a known strength. For this pur- pose 10 grms. of pure silver are dissolved in nitric acid in a litre flask, the nitrous acid is expelled, the liquid is allowed to cool, and diluted up to the volume. For use 50 c.c. are taken with a pipette. If the above-mentioned proportions are preserved for 0'5 grin, silver, or 50 c.c. of the silver solution, somewhat less than 50 c.c. of the sul- phocyanide solution will be required, whence the needful dilution can be calculated. If 48*5 c.c. sulphocyanide have been used, then to every 48-5 c.c. of the solution 15 c.c. of water must be added. The solution is perfectly permanent, even on being kept for two years. The silver solution must have a decidedly acid reaction from free nitric acid, which it is unnecessary to neutralise, although it is important that the proportion of acid in different experiments should be approximately equal. It is, however, essential for the con- stancy of the results that the ferric salt should be in large excess, and should be used approximately in one and the same proportion to the total volume of the fluid. It must also be remembered that the colour of ferric sulphocyanide is destroyed by nitrous acid at common temperatures, and by nitric acid on the application of heat. Hence follows the necessity of completely expelling all nitrous acid, and of allowing the liquid to become cold before the operation is begun. The nitric acid used in this process should be kept in the dark. Copper. The proportion of copper in an alloy may reach 70 per cent, without in the least affecting the accu- racy of the process. Beyond this limit the recognition of the final reaction is somewhat doubtful, since after the precipitation of the silver the liquid is rendered opaque and discoloured by the formation of black copper sulpho- cyanide. The only remedy in such cases is to add a known weight of pure silver to the alloy, so as to reduce 711 THE ASSAY OF SILVER. the proportion of copper below 70 per cent., as is done in Gay-Lussac's process. Mercury. In presence of this metal silver cannot be titrated with sulphocyanide solution. This defect is of little consequence, as mercury is readily expelled. Palladium renders the titration of silver with sulpho- cyanide inaccurate, this metal appearing in the result as silver. This is an important circumstance from a technical point of view, since palladium, though frequently present in silver, occurs in very small quantities. As regards other metals soluble in nitric acid and found along with silver in alloys and ores, lead, cadmium, thallium, bismuth, zinc, iron, and manganese are without influence. The recognition of the final reaction in solu- tions strongly coloured by salts of cobalt or nickel, requires some practice. At first a few drops of sulpho- cyanide will always be added in excess. It is then recom- mended to titrate back with a solution of silver, when the pure colour of the cobalt or nickel will appear so suddenly and distinctly, that it will conversely be easy to seize the exact point, when the colour of the solution is modified to a yellowish brown by the addition of the colour of fer- ric sulphocyanide. When the change of shade has been observed four or five times by titrating backwards and forwards with exactly corresponding solutions of silver and of sulphocyanide, such a certainty in the recognition of the final reaction will be attained that there can be no doubt as to the addition of a half drop more or less. The presence of tin, antimony, and arsenic does not interfere with the accuracy of the process. Mr. C. A. M. Balling gives the following instructions for the direct estimation of silver in galena by Vol- hard's process : From 2 to 5 grins, of the galena, accord- ing to its supposed richness in silver, are very finely ground and intimately mixed in a porcelain mortar with from three to four times its weight of a flux composed of equal parts of soda and saltpetre, placed in a porcelain crucible, covered, and heated over a burner to thorough fusion, BLOWPIPE ASSAY OF SILVER. 715 when the mixture is well stirred with a glass rod. It is then let cool and placed in an evaporating dish partly filled with water, in which the melted matter is softened, dissolved out of the crucible into the dish, which is then heated, and the watery solution is filtered into a flask. The residue on the filter, after being well washed, is rinsed back into the dish, very dilute nitric acid is added, and the whole evaporated to dryness. The dry residue is taken up in water acidulated with nitric acid, heated, and filtered into the same flask in which is the aqueous solu- tion. The residue is washed with hot water, the filtrate is allowed to cool in the flask, ferric sulphate or iron alum is added, and the liquid is titrated. BLOWPIPE ASSAY OF SILVER. The following very complete method for the blowpipe assay of silver and its ores is given by David Forbes, F.K.S., in the 'Chemical News,' JSTos. 380, 384, 392, 396, 398, and 412. The blowpipe assay of silver ores was first described in 1827 by Harkort,* and subsequently considerably im- proved by Plattner. This assay process is in all cases based upon the reduction to a metallic state of all the silver contained in the compound in question along with more or less metallic lead, which latter metal, when not already present in sufficient quantity in the substance itself under examination, is added in the state of granu- lated lead to the assay previous to its reduction. The globule of silver-lead thus obtained, if soft and free from such elements as would interfere with its treatment upon the cupel, may then be at once cupelled before the blow- pipe until the pure silver alone remains upon the bone-ash surface of the cupel ; but if not, it is previously submitted to a scorifying or oxidising treatment upon charcoal until all such substances are either slagged off or volatilised, and the resulting silver-lead globule cupelled as before. * 'Die Probirkunst mit dem Lothrohre.' Freiberg, 1827, I. Heft (all published). 716 THE ASSAY OF SILVER. As, therefore, the final operation in all silver assays is invariably that of cupelling the silver-lead alloy obtained from the previous reduction of the substance, effected by methods differing according to the nature of the argen- tiferous ore or compound under examination, it is here con- sidered advisable to introduce the description of the silver assay by an explanation of this process. In the ordinary process of cupellation in the muffle, bone-ash or other cupels are employed of a size large enough to absorb the whole of the litharge produced from the oxidation of the lead in the assay. This, however, should not be the case when using the blowpipe ; for as .the heating powers of that instrument are limited, it is found in practice much better to accom- plish this result by two distinct operations the first being a concentration of the silver-lead in which the greater part of the lead is converted by oxidation into litharge re- maining upon, but not, or only very slightly, absorbed by, the bone -ash cupel : and the second in cupelling the small concentrated metallic bead so obtained upon a fresh cupel until the remaining lead is totally absorbed by the cupel and the silver left behind in a pure state. By this means a much larger weight of the silver-lead alloy can be submitted to assay, and, for reasons hereafter to be explained, much more exact results are obtained than would be the case when the cupella- tion is conducted at one operation in the ordinary manner. The apparatus used by the author for these operations" are shown to a scale of one-half their real size in the woodcut fig. 140 (a to d). In fig. 140, a represents in section a small cylindrical mould of iron, seven- tenths of an inch in diameter, and about four-tenths high, in which is turned a cup-shaped nearly hemisphe- rical depression two-tenths of an inch deep in centre, the inner surface of which is left rough, or FIG. 140. FOKBES'S BLOWPIPE ASSAY. 717 marked with minute ridges and furrows for the purpose of enabling it to retain more firmly the bone-ash lining which is stamped into it by means of the polished bolt, also shown in the figure. This mould rests upon the stand d, having for this purpose a small central socket in its base, into which the central pivot of the stand enters. This socket is seen in the ground plan, 5, of the base of the mould, which shows likewise three small grooves or slots made to enable a steady hold to be taken of it, when hot, by the forceps. The stand itself is composed of a small turned ivory or wood base, fixed into a short piece of strong glass tubing, which, from its non-conducting powers, serves as an excellent handle. In the centre of the base a slight iron rod rising above the level of the glass outer tube serves as a support for the cupel mould, into the socket in the base of which it enters. Bone-ash is best prepared by burning bones which have previously been boiled several .times, so as to extract all animal matter. The best bone-ash is made from the core- bone of the horns of cattle well boiled out and burned. The ash from this is more uniform than from the .other bones, which have in general a very compact enamel-like exterior surface, whilst the interior is of a much softer nature. Concentration of the Silver-lead. A cupel is prepared by filling the above-described cupel mould with bone-ash powder not finer than will pass through a sieve containing from forty to fifty holes in the linear inch, and should be well dried and kept in an air-tight bottle, and the whole pressed down with, the bolt, using a few taps of the hammer, It is then heated strongly in the oxidising blowpipe flame,. in order to drive off any hygroscopic moisture. The bone- ash surface of the cupel, after heating, should be smooth, and present no cracks ; if the reverse, these may be re- moved by using the bolt again and reheating.* The silver- lead, beaten on the anvil into the form of a cube, is placed * These precautions are very important, as the slightest trace of moisture in the substance of the bone-ash would inevitably cause a spurting of the metal during the operation. 718 THE ASSAY OF SILVER. gently upon the surface of the bone-ash, and, directing a pretty strong oxidising flame on to its surface, it is fused, and quickly attains a bright metallic appearance, and commences to oxidise with a rapid rotary movement. (Occasionally, when the assay is large, and much copper or nickel present, the globule may, under this operation, cover itself with a crust of lead oxide or solidify ; in such cases direct the blue point of a strong flame steadily on to one spot of the surface of the lead globule, until it commences oxidising and rotating. In some cases where much nickel is present, an infusible scale, impeding or even preventing this action, may form, but will disappear on adding more lead say from three to six grains, according to the thick- ness of this scale or crust.) When this occurs the cupel is slightly inclined from the lamp, and a fine blue point obtained by placing the blowpipe nozzle deeper into the flame, and the lamp is directed at about an angle of 30 on to the globule not, however, so near as to touch it with the blue point, but only with the outer flame, so moderating it as to keep the assay at a gentle red heat, and not allowing the rotation to become too violent. This oxidising fusion should be carried on at the lowest temperature sufficient to keep up the rotatory movement, and to prevent a crust of litharge accumulating upon the surface of the globule, but still sufficiently high to hinder the metallic globule from solidifying. Should this, how- ever, happen, a stronger flame must be employed for a moment until the metal is again in rotation ; such inter- ruptions should, however, be avoided. The proper tem- perature can only be learned by practice. A too high temperature is still more injurious, causing the lead to volatilise, and, if rich in silver, carry some of that metal mechanically along with it. The litharge, also, instead of remaining on the cupel, would be absorbed by the bone- ash, and as the surface of the metallic globule is covered by a too thin coating of fused litharge, some silver may be absorbed along with the litharge. In this operation, in order to avoid loss of silver, the fused globule should be always kept in contact with the melted litharge. FORBES'S BLOWPIPE ASSAY. 719 By the above treatment, the air has free access to the assay, and the oxidation of the lead and associated foreign metals goes on rapidly. The surface of the melted globule, when poor in silver, shows a brilliant play of iridescent colours, which does not take place when very rich in silver. The litharge is driven to the edge of the globule, heating itself up and solidifying behind and around it. When the globule becomes so hemmed in by the litharge as to present too small a surface for oxidation, the cupel is moved so as to be more horizontal (having been previously kept in an inclined position), thus causing the lead globule to slide by its own weight on one side, and expose a fresh surface to the oxidising action. When the lead is pure, the litharge formed has a reddish-yellow colour, but if copper is present it is nearly black. In concentrating silver-lead, it must be remembered that an alloy of lead and silver, if in the proportion of about 86 per cent, silver along with 14 per cent, lead, when cooled slowly in the litharge behaves in a manner analogous to the spitting of pure silver, throwing out a whitish-grey pulverulent excrescence rich in silver. For this reason, therefore, the concentration above described should be stopped when the globule is supposed to contain about six parts silver along with one part in weight of lead. In case, however, this limit should have been ex- ceeded, it is advisable at once to push the concentration still further until the silver globule contains but very little lead. In practice with poor ores it is usual to concentrate the lead until the globule is reduced to the size of a small mustard -seed, or in rich ores to some two or three times that size. Upon arriving at this point, the cupel is with- drawn very gradually from the flame, so that the cooling shall take place as slowly as possible until the globule has solidified in its envelope of litharge. If cooled too quickly, the litharge, contracting suddenly, would throw out the globule, or even cause it to spirt; in such case it should be touched by the point of the blue flame, so as to fuse it to a round globule, which is cooled slowly, as before de- scribed. The globule is now reserved for the next opera- 720 THE ASSAY OF SILVER. tion, for which purpose it is, when quite cold, extracted from the litharge surrounding it. Cupellation.- The bone- ash required for this process should be of the best quality and in the most impalpable powder, prepared by elutriating finely ground bone ash, and drying the product before use. The cupel, still hot from the last operation, is placed upon the anvil, and the crust of litharge, with its enclosed metallic bead, gently removed, leaving the hot coarse bone- ash beneath it in the mould ; upon this a small quantity of the elutriated bone-ash is placed, so as to fill up the cavity, and the whole, whilst hot, stamped down by the bolt, previously slightly warmed, with a few taps of the ham- mer. The cupel thus formed is heated strongly in the oxidising flame, which should leave the surface perfectly smooth, and free from any fissures or scales ; if such appear, the bolt must again be used, and the cupel re- heated. In this process it is very important that the cupel should possess as smooth a surface as possible, whilst at the same time the substance of the cupel beneath should not be too compact, so as thereby to permit the litharge to filter through and be readily absorbed, leaving the silver bead upon the smooth upper surface. The bead of silver-lead obtained from the last operation is taken out of the litharge in which it is embedded, and, after removing any trace of adherent bone-ash or litharge, is slightly flattened to prevent its rolling about upon the surface of the cupel. It is now put into the cupel prepared as before de- scribed, placing it on the side farthest from the lamp and a little above the centre of the cupel, which is now in- clined slightly towards the lamp, and is heated by the oxidising flame directed downwards upon it, this causing the globule, when fused and oxidising, to move of itself into the centre of the cupel. The cupel is now brought into a horizontal position, and the flame, directed on to it at an angle of about forty-five degrees, is made to play upon the bone-ash surface immediately surrounding the globule, without, however, touching it, so as to keep this part of FORBESS BLOWPIPE ASSAY. 721 the cupel at a red heat sufficiently strong to insure the globule being in constant oxidising fusion, and at the same time to cause the perfect absorption of the litharge, so as to prevent any scales of litharge forming upon the surface of the cupel under the globule, which would impede the oxidation, as well as prevent the silver bead being easily detached at the conclusion of the operation. Should the heat at any time be too low and the globule solidify, it must be touched for an instant with the point of the flame and proceeded with as before. Should (in consequence of the bone-ash not having been sufficiently heated to absorb the litharge perfectly) a little litharge adhere pertinaciously to the globule, or a particle of the bone-ash cupel attach itself, the cupel should be slightly inclined, so as to allow the globule to move by its own weight on to another and clean part of the cupel, leaving the litharge or bone-ash behind it ; but, if not sufficiently heavy to do so, a small piece of pure lead may be fused to it in order to increase its weight, and so allow of the same proceeding being adopted. By slightly inclining the cupel-stand, and moving it so as to present in turn all parts of the surface surrounding the globule to the action of the flame, the cupellation pro- ceeds rapidly. If, however, the assay contains very little silver, it will be found necessary to move the globule from one spot to another on the cupel, in order to present a fresh surface for absorbing the litharge formed ; this is done by simply inclining the cupel-stand, remembering that the bone- ash surrounding the globule must always be kept at a red heat, without ever touching the globule itself by the flame. In assays rich in silver a play of iridescent colours appears some seconds before the ' brightening,' which dis- appears the moment the silver becomes pure ; as soon as this is observed the cupel should be moved in a circular manner, so that the globule is nearly touched all round by the point of the blue flame, and this is continued until the surface of the melted silver is seen to be quite free from any litharge, upon which it is very gradually withdrawn from the flame so as to cool the assay by degrees very slowly, in order to prevent ' spitting.' 3 A 722 THE ASSAY OF SILVER. When the silver-lead is very poor, this play of colours is not apparent, and as soon as the rotatory movement of the globule ceases, the heat must be increased for an instant, in order to remove the last thin but pertinacious film or scale of litharge, and subsequently the assay is cooled gradually ; when cold it should, whilst still upon the cupel, be examined by a lens, to see whether the bead possesses a pure silver colour, as, if not, it must be reheated. Frequently, when the ' brightening ' takes place, the silver globule is found to spread out, and, after cooling, although of a white colour, is found to appear somewhat less spherical or more flattened in shape than a correspond- ing globule of pure silver would be. This arises from the presence of copper still remaining in the silver, and in such cases a small piece of pure lead (about from one- half to one and a half grain in weight, according to size of assay) should be fused on the cupel along with the silver, and the cupellation of the whole conducted as before on another part of the cupel, when the silver globule will be obtained pure, and nearly spherical in shape. Sometimes the silver globule in ' brightening ' may still remain covered with a thin film of litharge, although otherwise pure ; this arises from too little heat having been employed in the last stage of the operation, and consequently the bead should be re-heated in a strong oxidising flame until this litharge is absorbed, and the globule, after slow cooling, appears pure. If the instructions here given be strictly attended to, it will be found after some practice that very accurate results may be obtained in the blowpipe assay for silver, and that no difficulty will be found in detecting the presence and estimating the amount of silver present, even when in as small a quantity as half an ounce to the ton. When sub- stances containing very little silver or less than that amount are examined, several assays should be made, and the silver-lead obtained concentrated separately, after which the various globules should be united and cupelled together in one operation. It is hardly necessary to remark that the lead employed in assaying should be free from silver, or, if not, its actual FORBES'S BLOWPIPE ASSAY. 723 contents in silver should be estimated, and subtracted from the amount found in the assay. Assay lead containing less than one quarter of an ounce to the ton of lead can readily be obtained, or can be made by precipitating a solution of acetate of lead by metallic zinc, rejecting the first portion of lead thrown down. In all cases the lead should be fused and granulated finely the granulated lead for use in these assays being previously passed through a sieve containing forty holes to the linear inch. It is also useful to have some lead in the form of wire, as being very convenient for adding in small portions to assays when on the cupel. Estimation of the Weight of the Silver Globule obtained on Cupellation. As the amount of lead which can, by the method before described, be conveniently cupelled before the blowpipe is necessarily limited, the silver glo- bule which remains upon the bone-ash surface of the cupel at the end of the operation is, when substances poor in silver have been examined, frequently so very minute that its weight could not be estimated with correctness by the most delicate balances in general use. The blowpipe balance employed by the author turns readily with one-thousandth of a grain, but could not be used for estimating weights below that amount. Globules of silver of far less weight than one-thousandth are distinctly visible to the naked eye a circumstance which induced Harkort to invent a volumeti|ical scale based upon the measurement of the diameters of the glo- bules, which scale in practice has been found of very great utility in the blowpipe assay of silver. The scale for this purpose which is employed by the author is shown in full size in the woodcut on p;. 724. This figure represents a small strip of highly polished ivory about 6J inches long, f inch broad, and inch in thickness, on which are drawn, by an extremely fine point, two very fine and distinct lines emanating from the lower or zero point, and diverging upwards until, at the distance of exactly six English standard inches, they are precisely four hundredth parts of an inch apart. This distance 3 A 2 724 THE ASSAY OF SILVER. (six inches) is, as shown in woodcut, divided into 100 equal parts by cross lines numbered in accordance from zero FIG. i4i. upwards. It is now evident if a small globule of silver be placed in the space between these two lines, using a magnifying glass to assist the eye in moving it up or down until the diameter of the globule is exactly contained within the lines themselves, that we have at once a means of estimating the diameter of the globule itself, and therefrom are enabled to calculate its weight. As the silver globules which cool upon the surface of the bone-ash cupel are not true spheres, but are considerably flattened on the lower surface, where they touch and rest upon the cupel, it follows that the weight of glo- bules corresponding in diameter to the extent of divergence at the different degrees of the scale cannot be calculated directly from their diameters as spheres, but require to have their actual weight experimentally determined in the same manner as employed byPlattner. .. -The table here appended has been cal- culated by the author, and in one column shows the diameter in English inches corre- sponding to each number or degree of the scale itself, and in the two next columns the .respective weights of the flattened spheres which correspond to each degree or diameter ; for convenience these weights are given in the different columns in decimals, both of English grains and of French grammes. These weights are calculated from the following data, found as the average result of several very careful and closely approximating assays, which showed that globules of silver exactly corresponding to No. 95 on this scale, or 0*038 inch in diameter, possessed a weight of 0-0475573 grains or 0-003079 grammes. Erom this the 100 98 96 94 92 _: 90 : 88 " 1- : 86 84 82 80 ~ 78 76 ~ 74 72 70 68 - 66 -f 64 62 - 60 68 - 56 7 :~- 50 : 48 - 46 44 v 42 - 40 - 38 36 - '34 ; 32 - 5 28-^ W 24 22 SO 18 -^ 16 -3 14 12 | i 10 6 ~ o * FORBES 8 BLOWPIPE ASSAY. 725 No. on scale Greatest diameter in inches Weight of globule in grains Weight of globule in grajnmes 1 0-0004 0-00000005 0-000000003 2 0-0008 0-00000044 0-000000028 3 0-0012 0-00000149 0-000000096 4 0-0016 0-00000355 0-000000229 5 0-0020 0-0000069 0-00000044 6 0-0024 0-0000119 0-00000077 7 0-0028 0-0000190 0-00000120 8 0-0032 0-0000284 0-00000184 9 0-0036 0-0000403 0-00000262 10 0-0040 0-0000554 0-00000359 11 0-0044 0-0000736 0.00000478 12 0-0048 0-0000958 0-00000620 13 0-0052 0-0001218 0-00000789 14 0-0056 0-0001522 0-00000985 15 0-0060 0-0001872 0-00001203 16 0-0064 0-0002272 0-00001471 17 0-0068 0-0002725 0-00001764 18 0-0072 0.0003234 0-00002094 19 0-0076 0-0003804 0-00002463 20 0-0080 0-0004437 0-00002872 21 0-0084 0-0005137 0-00003327 22 0-0088 0-0005906 0-00003823 23 0-0092 0-0006748 0-00004367 24 0-0096 0-0007668 0-00004964 25 0-0100 0-0008667 0-00005611 26 0-0104 0-0009749 0-00006311 27 0-0108 0-0010918 0-00007068 28 0-0112 0-0012176 0-00007883 29 0-0116 0-0013528 0-00008758 30 0-0120 0-0014976 0-00009696 31 0-0124 0-0016524 0-00010698 32 0-0128 0-0018176 0-00011677 33 0-0132 0-0019934 0-00012817 34 0-0136 0-0021801 0-00014114 35 0-0140 0-0023786 0-00015397 36 0-0144 0-0025879 0-00016755 37 0-0148 0-0028097 0-00018190 38 0-0152 0-0030437 0-00019705 39 0-0156 0-0032903 0-00021302 40 0-0160 0-0035550 0-00022983 41 0-0164 0-0038230 0-00024751 42 0-0168 0-0041096 0-00026606 43 0-0172 0-0044111 0-00028553 44 0-0176 0-0047250 0-00030589 45 G'0180 0-0050546 0.00032725 46 0-0184 0-0053991 0-00034955 47 0-0188 0-0057590 0-00037285 48 0-0192 0-0061344 0-00039716 49 0-0196 0-0065258 0-00042250 50 0-0200 0-0069335 0-00044890 51 0-0204 0-0073581 0-00047638 52 0-0208 0-0077799 0-00050495 53 0-0212 0-0082580 0-00053464 54 0-0216 0-00873438 0-00056549 55 0-0220 0-00922854 0-00059748 56 0-0224 0-0097412 0-00063067 726 THE ASSAY OF SILVEK. Jfo. on scale Greatest diameter in inches Weight of globule in grains Weight of globule in grammes 57 0-0228 0-0102725 0-00066506 58 0-0232 0-0108228 0-00070021 59 0-0236 0-0113922 0-00073753 60 0-0240 0-0119815 0-00077570 61 0-0244 0-0125901 0-00081513 62 0-0248 0-0132119 0-00085588 63 0-0252 0-0138901 0-00089797 64 0-0256 0-0145440 0-00094141 65 0-0260 0-0152311 0-00098623 66 0-0264 0-0159472 0-00103245 67 0-0268 0-0166828 0-00108010 68 0-0272 0-0174414 0-00112918 69 0-0276 0-0182220 0-00117974 70 0-0280 0-0190256 0-00123177 71 0-0284 0-0198529 0-00128535 72 0-0288 0-0207035 0-00134041 73 0-0292 0-0215782 0-00139704 74 0-0296 0-0224469 0-00145525 75 0-0300 0-0234010 0-00151504 76 0-0304 0-0243496 0-00157645 77 0-0308 0-0253224 0-00163950 78 0-0312 0-0263228 0-00170422 79 0-0316 0-0273484 0-00177060 80 0-0320 0-0284000 0-00183869 81 0-0324 0-0294789 0-00190852 82 0-0328 0-0305838 0-00198008 83 0-0332 0-0317162 0-00205340 84 0-0336 0-0328768 0-00212851 85 0-0340 0-0340649 0-00220549 86 0-0344 0-0349739 0-00228400 87 0-0348 0-0364422 0-00235938 88 0-0352 0-0378008 0-00244730 89 0-0356 0-0390138 0-00253168 90 0-0360 0-0404368 0-00261797 91 0-0364 0-0417943 0-00270790 92 0-0368 0-0431930 0-00279642 93 0-0372 0-0446162 0-00288860 94 0-0376 0-0460718 0-00298279 95 0-0380 0-0475573 0-00307900 96 0-0384 0-0465239 0-00317728 97 0-0388 0-0506249 0-00327759 98 0-0392 0-0522069 0-00338020 99 0-0396 0-0538215 0-00348452 100 0-0400 0-0554688 0-00359138 respective weights of all the other numbers or degrees on this scale were calculated, on the principle that solids were to one another in the ratio of the cubes of their diameters. This mode of calculation is not, however, absolutely correct in principle, for the amount of flattening of the under sur- face of the globule diminishes in reality with the decreasing FORBES'S BLOWPIPE ASSAY. 727 volume of the globule. In actual practice, however, this difference may be assumed to be so small that it may be neglected without injury to the correctness of the results. The smaller the diameter of the globule, the less will be the difference or variation in weight in descending the degrees of this scale, since the globules themselves vary in weight with the cubes of their diameters : for this reason, also, all such globules as come within the scope of the balance employed should be weighed in preference to being measured, and this scale should be regarded as more spe- cially applicable to the smaller globules beyond the reach of the balance. Cupellation Loss. This term is applied to indicate a minute loss of silver, unavoidably sustained in the process of cupellation, which arises from a small portion of that metal being mechanically carried along with the litharge into the body of the cupel. The amount of this loss in- creases with the quantity of lead present in the assay (whether contained originally in the assay or added subse- quently for the purpose of slagging off the copper, &c.) ; it is relatively greater as the silver globule is larger, but represents a larger percentage of the silver actually con- tained in the assay, in proportion as the silver globule obtained diminishes in size. It has, however, been experi- mentally proved that, in assays of like richness in silver, this loss remains constant when the same temperature has been employed, and similar weights of lead have been oxidised in the operation. In the blowpipe assay this loss is not confined to the ultimate operation of cupellation, but occurs, though in a less degree, in the concentration of the silver-lead, and in the previous scorification of the assay, had such operation preceded the concentration. The total loss in the blow- pipe assay is found, however, to be less than in the ordinary muffle assay, since in the latter case the whole of the oxi- dised lead is directly absorbed by the cupel. In mercantile assays of ore it is not customary to pay attention to the cupellation loss, and the results are usually stated in the weight of silver actually obtained. Where, 728 THE ASSAY OP SILVER. however, great accuracy is required, especially when the substances are very rich in silver, the cupellation loss is added to the weight of the silver globule obtained, in order to arrive at the true percentage. The amount to be added for this purpose is shown in the annexed table, which is slightly modified from Plattner's : Actual per- centage of silver found by assay Cupellation loss, or percentage of silver to be added to the actual percentage found by assay in order to show the true percentage of silver contained in same. The entire amount of lead in or added to the assay being the folio wing multiples of the original weight of assay : 1 2 3 4 5 6 I 8 1 11 13 16 1 99-5 / 0-25 0-32 0-39 0-45 0-50 90 0-22 0-29 0-36 0-42 0-47 0-69 0-83 80 0-20 0-26 0-33 0-39 0-44 0-64 0-75 70 0-18 0-23 0-29 0-35 0-40 0-58 0-68 0-82 60 0-16 0-20 0-26 0-30 0-36 0-52 0-61 0-74 50 0-14 0-17 0-23 0-26 0-32 0-46 0-54 0-65 40 0-12 0-15 0-20 0-22 0-27 0-39 0-46 0-55 0-62 35 0-11 0-13 0-18 0-18 0-25 0-36 0-42 0-50 0-57 30 0-10 0-12 0-16 0-16 0-22 0-32 0-38 0-45 0-51 25 0-09 0-10 0-14 0-14 0-20 0-29 0-34 0-40 0-45 20 0-08 0-09 0-12 0-12 0-17 0-25 0-29 0-35 0-39 0-45 15 0-07 0-08 0-10 0-11 0-15 0-20 0-23 0-28 0-32 0-37 12 0-06 0-07 0-09 0-10 0-13 0-17 0-19 0-23 0-26 0-22 10 0-05 0-06 0-08 0-09 0-11 0-15 i 0-17 0-20 0-23 0-27 9 0-04 0-05 0-07 0-08 0-10 0-14 0-16 0-18 0-21 0-25 8 0-03 0-04 0-06 0-07 0-09 0-13 0-15 0-16 0-18 0-22 7 0-02 0-03 0-05 0-06 0-08 0-12 i 0-13 0-14 0-16 0-20 6 0-01 0-02 0-04 0-05 0-07 0-10 0-11 0-12 0-14 0-17 5 0-91 0-03 0-04 0-06 0-09 0-10 0-11 0-12 0-14 4 0-02 0-03 0-05 i 0-07 0-08 0-09 0-10 0-11 3 0-01 0-02 0-04 0-05 0-06 0-07 0-08 0-09 2 0-01 0-03 0-04 0-04 0-05 0-06 0-07 1 , 0-01 0-03 0-03 0-04 0-04 0-05 The use of this table is best explained by an example, as the following : An assay to which there had been added, in all, five times its weight of assay lead, gave a globule of silver equivalent to six per cent. Upon referring to the table it will be seen that the cupellation loss for this would be 0*07 ; consequently the true percentage of silver contained in the assay would be 6'07. This table is only extended to whole numbers, but fractional parts can easily be calculated from the same. FOEBES'S BLOWPIPE ASSAY. 729 When the globules of silver are so minute that they cannot be weighed, but must be measured upon the scale, the cupellation loss should not be added, since, as a rule, it would be less than the difference which might arise from errors of observation likely to occur when measuring their diameters upon the scale. In the case of beginners, it will be found that the cupel- lation is usually carried on at too high a temperature, and that thereby a greater loss is occasioned than would be accounted for by the above table. After some trials the necessary experience will be acquired in keeping up the proper temperature at which this operation should be effected. It now becomes necessary to consider in detail the processes requisite for extracting the silver contents (in combination with lead) from the various metallic alloys of silver which are met with in nature or produced in the arts. In considering these, the following classification of the substances will be found convenient : METALLIC ALLOYS. A. Capable of direct Cupellation. a. Consisting chiefly of lead or bismuth ; silver-lead and argentiferous bismuth, native bismuthic silver. 6. Consisting chiefly of silver : native silver, bar silver, test silver, precipitated silver, retorted silver amalgam, stan- dard silver, alloys of silver with gold and copper. c. Consisting chiefly of copper : native copper, copper ingot, sheet or wire, cement copper, copper corns, copper- nickel alloys. B. Incapable of direct Cupellation. a. Containing much copper or nickel, with more or less sulphur, arsenic, zinc, &c. ; unrefined or black copper, brass, German silver. 6. Containing tin ; argentiferous tin, bronze, bell-metal, gun- metal, bronze coinage. c. Containing antimony, tellurium, or zinc. d. Containing mercury : amalgams. e. Containing much iron : argentiferous steel, bears from smelting furnaces. 730 THE ASSAY OF SILVER. A. METALLIC ALLOYS CAPABLE OF DIBECT CUPELLATION. a. Consisting chiefly of Lead or Bismuth. In estimat- ing the silver contained in these alloys, it is only requisite to place a clean piece of the same, weighing about from one to ten grains according to its probable richness in silver, upon a cupel of coarse bone-ash, and proceed by concentration and cupellation exactly as has been already described under these heads. Should the substance be not altogether metallic, or not free from adherent slag, earthy or other extraneous matter, it should previously be fused on charcoal with a little borax in the reducing blowpipe flame, and the clean metallic globule then removed from the charcoal, and treated as before. In order to remove the globule from the adherent borax-glass, it may be allowed to cool, and then detached ; or, after a little practice, it will be found easy, by a quick movement of the charcoal, to cause the globule, still melted, to detach itself completely, and drop on the anvil in the form of a single somewhat flattened globule, without suffering any loss of lead adhering to the charcoal. In the case of argentiferous bismuth alloys the process is carried on in all respects the same as if silver-lead were being treated. As, however, the bismuth globule is very brittle, care must be taken when separating the concen- trated globule from the litharge, as, if not carefully done, a loss may easily be sustained from a portion of the glo- bule remaining behind adherent to the litharge. It is better, therefore, to remove the litharge by degrees from the globule with the aid of the forceps. Argentiferous bismuth, free from lead, when cupelled alone, invariably leaves a globule of silver, having a dull frosted surface. If, however, at the end of the operation a small quantity of lead 4 (i to ^ a grain) be added, and fused along with it, the silver globule then obtained will be perfectly bright and free from all bismuth. In the case of native bismuthic silver it is advisable to FOKBES'S BLOWPIPE ASSAY. 731 fuse the previously weighed mineral with a little lead and borax-glass on charcoal in the reducing flame, so as to free it from any adherent earthy matter, and then proceed by concentration and cupellation, as before described. b. Consisting chiefly of Silver: native silver, bar, test, and precipitated silver, retorted silver amalgam, standard silver, silver coin and other alloys of silver with gold and copper. These alloys may be at once fused with lead on the cupel itself, and the operation finished as before described. In general, however, it is better to fuse the weighed assay previously with the requisite amount of pure lead and a little borax-glass, say from a quarter to half the weight of assay, in the reducing flame at a low heat on charcoal until the globule commences to rotate. This insures having a perfectly clean button of silver-lead, which is then cupelled in the ordinary manner. In most cases the quantity of lead to be added need not exceed that of the weight of the alloy, but when several percentages of copper are present in the assay, as in the case of many coins, &c., the lead should be increased to some three, or even five, times the weight of the assay in proportion to the amount of copper actually contained in the substance under examination, and which will be treated of more at length under the head of copper-silver alloys. When no more lead has been added to the assay than its own weight, the cupellation may be concluded in one operation by inclining the stand, and so moving the glo- bule to a clean part of the cupel ; but when more copper is present, it is preferable to concentrate first and cupel subsequently, in order thereby to reduce the cupellation loss to its minimum. In the concentration as much copper as possible should be slagged off with the lead, which is effected by inclining the cupel somewhat more than usual, so that its surface may be less covered up with the litharge and exposed as much as possible to oxidation, by which means the litharge, as it forms, is enabled to carry off more of the copper contained in the silver-lead. 732 THE ASSAY OF SILVER. Should the silver globule after cupellation show indi- cations of still containing copper, as before noticed, when treating of cupellation, a small quantity of lead must be fused along with it, and the cupellation finished as usual. As at the present time no means are known by which silver can be separated from gold by the use of the blow- pipe in all cases of alloys containing gold, this metal remains to the last along with the silver, and the result in such cases always indicates the combined weight of both these metals contained in the alloy under examination. The wet assay must be resorted to for effecting their separation. c. Containing chiefly Copper : native copper, ingot, wire or sheet copper, cement copper, copper coins, copper-nickel alloys. Under the most favourable conditions in cupella- tion, the amount of lead requisite, when converted into litharge, to slag off one part of copper along with it as oxide, amounts to between seventeen and eighteen times its weight. In the blowpipe assay it is usual to add to any cupriferous alloy an amount of pure lead equal to twenty times the amount of copper contained in the alloy, in order to insure the whole of the copper being separated in the litharge. In the case of nickel the amount of lead required is somewhat less than with copper, but in prac- tice the same amount of lead may be employed. When the copper is quite clean the requisite amount of lead may be added to it in a single piece on the cupel, fused and cupelled as usual, after previous concentration of the silver-lead to a small-sized globule. It is generally found, however, that traces of iron, slag, gangue, or other foreign matter, are present ; and, conse- quently, it is usually advisable to fuse the assay along with the requisite amount of lead, and about one-half its own weight of borax-glass in the reducing flame, until the whole of the substance is seen to have perfectly combined or alloyed with the lead, and the globule has entered into brisk rotation, whilst at the same time no detached metallic globules are seen in the borax-glass. The concentration of the silver-lead and cupellation FORBES'S BLOWPIPE ASSAY. 733 are then conducted as usual, taking care when concentrat- ing to incline the cupel-stand so as to expose as much as possible of the metallic surface of the melted globule to the oxidising action of the air, with a view of enabling the litharge whilst forming to carry off as much copper along with it as possible. Should the silver globule obtained after cupellation spread out, or appear to the eye more flattened than usual with globules of pure silver, it indicates that some copper still remains, and a small piece of assay lead (-J to 1 grain weight) should be placed alongside it whilst still on the cupel, fused together, and the cupellation finished on a clean part of the same cupel as usual. Precipitated or cement copper, especially that which is in the crude state, and has not been melted and run into ingots, is often very impure, containing so much iron, lead, arsenic, earthy matter, &c., as not to admit of direct cupellation, and in such case should be treated as pertain- ing to class B. a. B. METALLIC ALLOYS INCAPABLE OF DIRECT CUPELLATION. a. Containing much Copper or Nickel, with frequently some little sulphur, arsenic, zinc, iron, cobalt, fyc. ; as unre- fined or black copper, brass, German silver, fyc. As the presence of these extraneous matters would interfere with the cupellation, either by causing a loss of silver-lead projected from the cupel upon the evolution of the volatile substances present, or by forming oxides which could not be absorbed by the cupel, it is necessary to eliminate such substances by a scorification with borax on charcoal, previous to concentration or cupellation. In the case of unrefined and black copper, the portion used in the examination is placed in the scoop with twenty times its weight of assay lead, and its own weight of pow- dered borax-glass, mixed with the spatula, and transferred to a soda-paper cornet. It is then fused on charcoal in the reducing flame, which should be constant and uninter- rupted, until all particles have completely united, and a 734 THE ASSAY OF SILVER. brisk rotation sets in, which is kept up for a short time, when the silver-lead globule, which should appear bright on the surface after cooling, is concentrated and cupelled precisely as is directed under A. c. By this preliminary scorification the sulphur, arsenic, and zinc are volatilised, and any lead, cobalt, or iron slagged off into the borax- glass. In the assay of brass and German silver, the quantity employed is fluxed with its own weight of borax-glass, but only requires ten times its weight of assay-lead. The operation is commenced as before, but the globule is kept somewhat longer in rotation (always keeping the flame directed only on to the borax glass), so as to allow the zinc present to be completely volatilised, which is evident when the surface of the silver-lead becomes bright, on which the heat is increased for a few moments to expel the last traces of that metal, and the silver-lead thus obtained is concen- trated and cupelled as before. The silver globule obtained from the cupellation of substances rich in copper generally requires the addition of a small quantity of lead and re-cupellation (as before described), in order to insure its freedom from copper. b. Containing Tin: argentiferous tin, bronze, bell and gun metal, bronze coinage, fyc. Alloys of silver with other metals containing tin do not admit of being cupelled, since the oxide of tin formed by the oxidation of that metal is not absorbed by the bone-ash of the cupel along with the litharge ; it consequently remains upon the surface of the cupel, and if present in any quantity interferes with the operation. As tin is not volatile when heated on charcoal either in the oxidising or reducing blowpipe flame, it cannot be so dissipated, and, in consequence, the entire amount of tin contained in any alloy under examination must be removed by oxidation or scorification from the silver-lead, previous to its being submitted to cupellation. For this purpose, 1 part of the stanniferous alloy is fluxed with from 5 to 15 parts of granulated assay lead (ac- cording to the amount of copper suspected to be present in the alloy), O5 part anhydrous sodium carbonate, and 0'5 FORBES'S BLOWPIPE ASSAY. 735 part pulverised borax-glass, made up as usual in a soda- paper cornet, and the whole at first gently heated in the reducing flame until the soda-paper is charred and the alloy has afterwards united with the lead to form a single globule, whilst the borax and soda have combined as a glass or slag in which the soda prevents the easily oxidis- able tin becoming oxidised to any extent before a perfect alloy has been formed with the lead, which then contains the whole of the silver. As soon as this is effected, the blowpipe flame is altered to an oxidising one, and the metallic globule is kept at the point of the blue flame, which should touch it so as to cause the tin to become oxidised and be at once taken up by the glass surrounding it. Should, however, it be seen that minute globules of metallic tin made their appearance on the outer edge of the slag or glass,* the operation must be at once discon- tinued, and the assay allowed to cool ; after cooling the metallic globule is detached from the slag surrounding it, and, being placed in a cavity on charcoal, is fused in the reducing flame along with a small piece of borax-glass and afterwards treated with the oxidising flame exactly as before (and if necessary, which is seldom the case unless when treating argentiferous block-tin, this operation may again require to be repeated), until it is seen that the surface of the metallic silver-lead globule does not any longer become covered with a crust or scales of tin oxide, but presents a pure and bright metallic surface. The silver-lead globule is now quite free from tin, and can be cupelled and the amount of silver estimated as usual. c. Metallic alloys containing much antimony, tellurium, or zinc : antimonial silver and argentiferous antimony, tel- luric silver, and argentiferous zinc. Alloys of antimony with silver when treated on charcoal in the oxidising flame give off all their antimony, leaving the silver behind as a metallic globule having a frosted external appearance ; * This occurs when the flux has become so saturated with tin oxide that it cannot take up any more. 736 THE ASSAY OF SILVEft. telluric silver, on the contrary, however, when treated in a similar manner, only evolves a part of its tellurium, and even after cupellation with lead a small amount of tellurium generally remains behind alloyed with the silver. All these compounds may be assayed as follows : One part of the alloy is placed in a soda-paper cornet along with 5 parts of granulated assay lead, and 0-5 part of pulverised borax-glass, and fused in the reducing flame until the globule and slag are well developed ; the oxidising flame is now directed on to the globule, causing the whole of the zinc, along with most of the antimony and part of the tellurium, to volatilise before the lead com- mences oxidising. The last traces of antimony are re- moved with some difficulty, during which operation some portion of the lead becomes oxidised. On cooling, the globule is separated from the slag and concentrated upon a coarse bone-ash cupel as usual, and if 110 tellurium were present in the concentrated silver-lead, this may now be cupelled as usual. If tellurium be present as is seen by the concentrated globule of silver-lead possessing a dark-coloured exterior it must be re-melted with 5 parts of assay lead and again concentrated ; and these operations, if necessary, must be repeated until the surface of the concentrated globule is found to be clean and bright, as is usual with pure silver- lead, when it may be cupelled fine and the silver globule weighed. It sometimes happens, even after all these precautions have been taken, that the silver globule after cupellation shows a crystalline, greyish-white, frosted appearance, from its still containing tellurium ; in such cases its own weight of assay lead (in one piece) should be placed beside it on the cupel, melted together, and the globule again cupelled fine on another part of the surface of the same cupel. In assaying substances very rich in tellurium the results ob- tained are, however, not very satisfactory, and may be as much as one or two per cent, too low, even after employ- ing all precautions. FOKBES'S BLOWPIPE ASSAY. 737 d. Compounds of Silver with Mercury : arquerite, native and artificial amalgams and argentiferous mercury. The assay of these compounds is very simple. A weighed quantity of the liquid or solid amalgam is placed in a small bulb-tube, and heated over the lamp very gradu- ally in order to avoid spirting and to allow the mercury to volatilise quietly ; * the heat is increased by degrees as long as any mercury is driven off, and the residue is heated for some time at a red heat in order to drive off as much mercury as possible without fusing the glass or causing the residual silver to adhere to it. The mercury expelled condenses above the bulb on to the upper part of the tube, and by gently tapping will collect in globules, which, by carefully turning the tube, unite and can be poured out of the tube ; after which the silver, left behind as a porous mass, may be removed from the tube, and after being fluxed with an equal weight of granulated assay lead and half its weight of borax-glass, must be fused on charcoal in the reducing flame, and the button, on cooling, cupelled as usual. Should, however, much copper have been present in the amalgam, a pro- portionately larger amount of assay lead is required to be added. When the argentiferous residue is extremely small, as is often the case when assaying argentiferous mercury, this may adhere firmly to the glass of the tube. On such occasions this part of the tube must be cut off with the adherent residue, and the whole fused in a strong reducing flame along with its own weight of granulated assay lead, and with half its weight of anhydrous sodium carbonate. Upon cooling, the globule of silver-lead thus obtained is cupelled as usual. e. Compounds chiefly consisting of Iron: argentiferous steel ; cast iron ; bears from smelting furnaces. Compounds consisting principally of iron with a small percentage of silver, although occasionally produced in the arts inten- In the case of solid amalgams, which often spirt very violently, this may be obviated by wrapping the assay in a small piece of tissue paper, and heating it in a blowpipe crucible, when all the mercury is given off quietly, leaving the silver behind. SB 738 THE ASSAY OF SILVER. tionally, as, for example, the so-called silver-steel, are commonly found on the blowing-out of furnaces used in the smelting of silver and copper ores, and are frequently rich in silver, as is the case with the bears from the silver furnaces at Kongsberg in Norway. An alloy of iron with silver is occasionally also found appearing in small quanti- ties on the surface of melted silver in the process of cast- ing, and in some cases at least this may be due to the action of the melted silver on the iron rods used for stirring up the molten metal. As iron cannot be made to alloy itself with lead before the blowpipe, it becomes necessary to extract the silver by a more indirect process than is used in the case of other alloys containing that metal. In order to remove the iron the alloy must first be converted into iron and silver sul- phide, and to effect this the iron or steel must be reduced to powder, or fragments none greater than about a quarter of a grain in weight ; for which purpose steel when har- dened may require to be softened previously. One part of the finely divided iron or steel is now mixed with O75 part sulphur, eight parts granulated assay lead, and one part pulverised borax-glass : the mixture after being placed in a soda-paper cornet is carefully fused in a cavity on charcoal in the reducing flame, until the whole appears as a fluid globule containing both the lead and iron in combination with the sulphur. Without removing either this globule or the glass surrounding it from the charcoal, an amount of borax-glass in one or more frag- ments (in all about equal in weight to the original amount of iron employed), is now added (in order to combine with and slag off the whole of the iron), and fused along with the former globule, after which the whole is submitted to a strong oxidising flame until the impure lead globule shows itself protruding from the slag. The charcoal is then inclined, so that the lead is alone subjected to the action of the outer flame, in order to vola- tilise the sulphur, and at the same time oxidise the iron which goes into the slag ; this operation is continued until the globule of lead appears with a bright metallic surface ; ASSAY OF ALLOYS OF SILVER AND COPPER. 739 should it on cooling, however, be found to possess a black colour, and to be brittle, it must be still further oxidised as before described. The silver-lead thus obtained will now be found to contain all the silver, and at the same time to be free from both iron and sulphur, and can be cupelled as usual. No notice is here taken of alloys of silver and gold, since these metals cannot be separated before the blowpipe by any process yet known ; and in all cases where gold may be present in an alloy treated as here directed for obtaining its contents in silver, the gold also will be found to follow along with the silver, and must be parted from that metal by the wet method, in order to enable the true amount of silver present in the substance to be ascer- tained. Alloys of Silver and Copper. A quantity of from 1 to 2 grms. is placed in a small flask or beaker, treated with a sufficient quantity of pure nitric acid of a moderate strength, covered with a watch-glass and digested at a gentle heat till totally dissolved. The solution is then diluted with water, and the silver precipitated with hydro- chloric acid, which is added drop by drop as long as any- thing is thrown down. It is then allowed to digest at a very gentle heat till the supernatant liquid is clear. The liquid is then poured upon a dried and counterpoised filter ; the silver chloride is stirred up Avith a little water, and brought upon the filter, the glass being perfectly rinsed out by aid of the washing-bottle. A few drops of nitric acid may be usefully added to the rinsing-water. The washing is afterwards continued with pure water till the droppings have no longer an acid reaction. The filter with its con- tents is dried, and from its weight that of the silver is calculated. 740 CHAPTEE XVII. THE ASSAY OF GOLD. FOR the purposes of assay, all substances containing gold may be divided into two classes, as in the case of silver. The First Class comprises ores containing gold in a mineralised form. The Second Class, comprises all alloys of gold, native or artificial. CLASS I. Minerals. Graphic Tellurium, (AuAg) Te 3 , containing 30 per cent. of gold, and 10 per cent, of silver. Foliated Tellurium, (Pb,Au,Ag) 2 (Te,Sb,S) 3 containing about 9 per cent, of gold. CLASS n. Alloys of Gold. Native Gold, Au Ag, containing 65-99 per cent, of gold. Palladium Gold, AuPd, containing about 86 per cent, of gold. Rhodium Gold, AuEh, containing 59-66 per cent, of gold, Gold Amalgam, (Au,Ag) 2 HgJ, containing 38 per cent, of gold. Artificial Alloys, Gold coin, jewellery, &c. Of the foregoing list, native gold alone occurs in nature in sufficient abundance to acquire any great commercial value. It is commonly found in a quartzose gangue, and nearly always associated with one, or more, of the following . NATIVE GOLD. 741 minerals : iron and copper pyrites, mispickel or arsenical pyrites, blende, galena, many antimonial minerals, and nearly all the primitive rocks. All auriferous slags, amal- gamation residues and tailings, belong to this class. If silver or platinum is associated with the gold in the mineral subjected to assay, it will be found combined with the gold obtained by cupellation ; and all gold so obtained must be submitted to the ' parting process,' which see further on. When gold is associated in quantity with quartz, its percentage can be approximately ascertained in the same manner as that of pure tin-stone when mixed with quartz (see page 545). If possible, a fragment of the gold must be detached from the quartz, and its specific gravity taken : if this be not possible, and the gold is nearly fine, the number 19 may be adopted. It is better, how- ever, to estimate experimentally the specific gravity of both quartz and gold. Native Gold and Aurides of Silver (Native) are found in variously contorted and branched filaments, in scales, in plates, in small irregular masses, in the crevices or on the surface of common ferruginous and other quartz. The author has received two specimens of gold one from Wales, and the other from the Britannia Mine, Devon and found both to be absolutely fine gold. Artificial Alloys of Gold. The only one of these alloys which will be specially noticed here is the standard gold of this realm. It is composed of 22 parts of fine gold and 2 parts of alloy, constituting 22-carat or standard gold. ASSAY OF GOLD ORES.* The assay of gold ores embraces the following steps : 1. Preparation of the sample. 2. Collection of the gold in a button of metallic lead. 3. Cupellation of the lead button, by which the lead is oxidised and absorbed by * For some portions of the following method of assaying gold ores, the Editor is indebted to Mr. T. M. Blossom, who has had great experience in the laboratory of the School of Mines, Columbia College, U.S.A. (See ' Chemical News,' Nos. 607, 609, 628, 635, 636.) 742 THE ASSAY OF GOLD. the cupel, leaving behind a bead of the gold and silver. 4. Weighing the bead. 5. Inquartation, parting, and cupellation of the gold residue. 6. Weighing the gold bead. 1. Preparation of the Sample. It is essential, in the first place, to obtain a fair average sample of the ore, otherwise the results of the assay may be commercially worthless. Selection must be left to the judgment of the assayer. The sample must be dried, if necessary, care being taken not to roast it. It must then be pounded in an iron mortar and passed through a sieve of eighty meshes to, the linear inch. If any native metal, in the form of scales or filaments, remain upon the sieve, take the weights, separately, of what has passed through and of what is left upon the sieve. The latter must be assayed according to * Assay of Alloys,' and the result referred to the whole amount of ore. It is essential that the whole of the sample, except the malleable portion, be passed through the sieve. Mix thoroughly the sifted ore. 2. The Collection of the Gold and Silver in a button of" metallic lead is effected in a crucible, or in a scorifier, whence two methods of assay : ~\a) Crucible assay. (b) Scorification assay. The former is applicable to all ores ; the latter is lima ted /practically, by the small size of scorifiers, to the richer ores. -(a) .Crucible Assay. An ore of gold and silver is com- posed of precious metal, gangue, and oxides, sulphides, &c., of foreign metals. To collect the precious metals in a button of lead, the ore is mixed with litharge, suitable fluxes, an oxidising or a reducing agent, and fused in the crucible. Litharge is reduced to metallic lead ; the latter seizes upon the precious metals and collects in a button at the bottom of the crucible, while the foreign materials form, with the fluxes, a fusible slag above the lead button. The crucible is broken when cold, and the malle- able button detached from the slag by hammering on an anvil. GOLD OKES. 743 The Charge. The weight of ore taken for an assay depends upon its supposed richness or poverty, since it is required to obtain finally a bead of precious metal of con- venient size for weighing, and, at the same time, neither too large for cupellation (vide cupellation) nor so small as seriously to affect the calculated results, through losses sustained 'in the assay. As a rule, it is usual to take 500, 1000, or 2,000 grains for gold ores. The ores require the following reagents : Litharge, sodium carbonate, and one of the reducing agents, argol and charcoal, or an oxidising agent, as nitre, with invariably . a cover of salt one-quarter to half an inch in depth. Borax, silica, and other reagents are very useful at times, but no general rule can be given for their employment. This matter must be left to the judgment of the assay er, guided by the known properties of the reagents and by the composition of the ore. It is well, in this connection, to bear in mind the principle that for basic impurities an acid flux is needed, and for an acid gangue a basic flux. As a rule, employ a weight of litharge twice that of the ore, and of sodium carbonate, the same as of ore. These proportions also may be modified to advantage, according to the composition of the ore. The propor- tion of nitre, or of reducing agent, depends upon the re- ducing power of the ore, hence it is variable in every case. These reagents are added to control the size of the lead button. Size of the Lead Button. There are two limits to the size of the button : (1) it must be large enough, or, in other words, enough litharge must be reduced throughout the mass to collect all the precious metal, and, at the same time, (2) there should not be a useless excess of lead, which would occasion loss of silver in the subsequent cupellation. These two conditions cannot always be fulfilled, but in this case the latter must be sacrificed. It has been found that a button of 225 grains is the best size for a weight of ore from 500 to 2,000 grains. This is also a convenient size for a cupellation. A button 744 THE ASSAY OF GOLD. that is too large for cupellation can, however, always be reduced in size by scorifying. These requirements necessitate, in many instances, a preliminary assay of the ore to estimate its reducing power. The reducing power of an ore is due to the presence of sulphur, arsenic, antimony, zinc, &c., but generally to sulphur contained in pyrites, &c. Preliminary Assay of Ore. Charge. Ore . . . . .20 grains. Litharge . . ; .' ' .250 Sodium Carbonate . .100 Salt . . . . . Cover. A duplicate assay need not be made. Warm the crucible before placing it in the fire, which should be perfectly bright, and should be urged to effect complete fusion in the shortest possible time. When the contents of the crucible are in quiet fusion, it must be withdrawn, to prevent further reducing action of the furnace gases. Tap the crucible gently, and when cold break it open. Three cases may arise here. Twenty grains of ore may yield 1. No lead, or less than thirty grains. 2. Thirty grains lead. 3. More than thirty grains. The reducing power is stated thus : 20 grains ore =# grains lead. Let us suppose that we shall take for the regular assay 150 grains of ore, and that the reducing power is found to be 20 grains ore=15 grains lead: 150 grains (about) ore will in this case reduce 112-5 grains lead, and as the required button is 225 grains, we must add enough argol, or charcoal, to reduce 112-5 grains in addition ; taking argol as 5*6, we shall require 112'5-^5'6 grains, or 20 grains, or charcoal 112-5-*-28 = 4 grains. If the reducing power correspond to the third case, a similar calculation will indicate how much nitre is needed GOLD ORES. 745 to oxidise part of the sulphur, arsenic, or other reducing agent, and thus prevent the reduction of a button larger than 225 grains. One part of lead requires 0*25 part of nitre to oxidise it. In the second case, 150 grains of ore would reduce a button of 225 grains, and neither argol nor nitre would be required. The character of slag obtained in the preliminary assay may also suggest some modification of the regular charge. If it be earthy, for instance, we would add borax-glass or silica. Experience will often enable the assayer to judge of the reducing power with sufficient exactness, without an extra assay. Cases will arise, however, in which he must make a preliminary assay, or roast the ore, and it is always best to roast when there is a large amount of sulphur in the ore. He will then have an ore of no reducing power. (Case I.) Ores to be Roasted. Ores containing a large amount of sulphur, or arsenic, antimony, or zinc, should always be roasted. In the former case, if the ore be not roasted, there will be danger of the formation of oxysulphides, which are very fusible, but are not decomposed even at a white heat, and enter the slag carrying silver with them. A large quantity of nitre also subjects the contents of the crucible to the liability of boiling over : even should this mishap not occur, the great evolution of nitrous and sul- phurous vapours puffs up the mass throughout the cru- cible, and the globules of lead afterwards reduced may be left adhering to the sides, not being washed down by the retreating charge. Arsenic and antimony produce arseni- ates and antimoniates, which carry silver into the slag. Zinc increases the loss of silver by volatilisation and also in the slag. Roasting the Ore. The ore may be roasted conveniently in a cast-iron pan over the crucible furnace. There ought to be a hood over the furnace, to carry off the fumes of sulphurous, arsenious, and other acids. The pan should be covered with a coating of red ochre, or, still better, of chalk. The former is put on wet with a brush. An excellent even coating of the latter may v be obtained by 746 THE ASSAY OF GOLD. making a chalk paste, of proper consistence, in the pan, and turning with the hand so as to spread the paste while the pan is being held over the fire to dry. The coating prevents a loss of ore and injury to the pan through the sulphides attacking the iron. The weighed sample must be spread over the pan and stirred with a bent wire until all danger of fusion is past. The pan must be heated gradually at first, not above a dull red heat for some time, and may be brought, finally, to a full red, or higher heat. Too high a heat at the outset might cause the fusion of sulphides and the forma- tion of matter troublesome to roast. A rapid disengage- ment of arsenic, antimony, or zinc would cause, also, a mechanical loss of silver, by carrying it off in their vapours. Should fusion take place, it is better to weigh out a fresh portion of ore and roast again with more care. If this be not done the fused portions must be pulverised in a mortar and re-roasted. Generally, in roasting sulphides, the operation may be considered finished when, after keeping the pan at a full red heat for some time, no more fumes of sulphurous acid can be perceived. As there is danger, however, of the formation of sulphates, especially if copper pyrites be present, it is best always to mix some ammo- nium carbonate with the ore, after the fumes have ceased, and to cover the pan. The ore is thus confined in an atmosphere of ammonium carbonate, which decomposes the sulphates with formation of volatile ammonium sul- phate. Arsenic and antimony require the addition of fine charcoal, to reduce any arseniates and antimoniates that may have been formed during the roasting. Care must be taken to burn out all the charcoal. If the ore contain a very fusible sulphide, as antimony- glance or galena, it may be mixed with some fine sand previous to roasting The roasting of ores may also be done in the muffle in an earthen saucer. GOLD ORES. 747 Fusion. The charge prepared according to the foregoing direc- tions is thoroughly mixed and placed in a crucible, which it should not more than two-thirds fill. A hot fire should be employed, and the crucible removed when complete fusion has taken place. This ought to require from twenty to twenty-five minutes. The crucible is tapped, as usual, and broken when cold. (b) Scarification Assay. The reagents necessary for a scorification assay are test-lead and borax-glass. The ore is mixed with these in suitable proportions, the mixture put in a scorifier and fused in a muffle. The operation affords an alloy of lead with the precious metals, and a slag composed of litharge with the impurities and gangue of the ore. The assay might be made with lead only, but it is advantageous to add some borax. There will be required of lead only enough to render the slag liquid and to furnish lead for the button. The proportions of both lead and borax will vary, and should be greater in pro- portion as the gangue and metallic oxides are more diffi- cult of fusion. The following table exhibits the propor- tions found by experience to be best adapted to the different gangues. The proportions are referred to one part of ore. Parts Test-Lead Parts Borax 8 8 5-6 16 16 12-16 10-15 10-15 0-25-1-00 0-15 0-10-0-50 0-10-1-00 0-10-0-15 0-10-0-20 0-10-0-20 Character of Gangue Quartz ose Basic (Fe 2 O 3 ,AL0 3 CaO), &c Galena . Arsenical Antimonial Fahlerz Iron pyrites Blende . No preliminary roa.sting is required. As in the cru- cible assay, the weight of ore taken depends very much upon its richness, but is generally a third, sixth, or tenth of an assay ton.* If one scorifier will not contain the charge, it is best to weigh equal fractional parts for the * An assay ton is about 450 grains. 748 THE ASSAY OF GOLD. number required, rather than to weigh the whole charge and roughly divide it between the scorifiers. In case of an accident to one of the scorifiers, the known loss can easily be restored, if the former course be followed. The exact parts that lead and borax perform in scorification can be understood best from a description of the operation in detail. Three distinct periods maybe noted in the working : 1. Eoasting ; 2. Fusion ; 3. Scorification. A strong- heat is maintained, at first, in order to melt the lead. 'This is effected by closing the muffle and regulating the draught. As soon as the lead is fused the muffle is opened, and the ore is seen floating upon the surface of the lead. 1. The roasting now commences, and is continued at a moderate heat until no more fumes are seen, and the ore has disappeared. 2. The heat should now be raised in order to fuse thoroughly all the material in the scorifier. When the fusion is complete, clear white fumes of lead may be seen arising from the scorifier, there is an intermittent play of colours across the bright surface of the lead, and the slag produced encircles the metallic bath like a ring. The borax plays an important part just here, by giving liquidity to the slag, thus permitting it to be thrown to the side as fast as formed, exposing a clear surface of lead -for oxida- tion. If borax be not added and the ore contain a diffi- cultly fusible gangue, the scoriae will float in detached masses over the lead, impeding the oxidation, until suffi- cient litharge has been formed to give it liquidity. 3. When the fusion is complete, the heat may be lowered and the third period of scorification continued until the ring of slag, ' which is continually growing smaller, closes over the residue of metallic lead. The heat .should again be raised, to liquefy the slag and allow the metallic lead to settle, after which the scorifier is removed from the furnace. If the button is wanted immediately for cupellation, the contents of the scorifier may be poured into an iron or copper mould coated with reel ochre. It CUPELLATIOiV. 749 is thus soon cooled. Otherwise allow the contents to solidify in the scorifier, and break this for the button when cold. Hammer the button as usual. The whole assay occupies about thirty-five minutes. In making the charge, it is customary to mix the ore with a part of the test-lead and with borax, and to cover this in the scorifier with the rest of the lead. This will prevent loss of ore prior to fusion. Too much borax should not be added at first. If a large amount is needed, it is better to mix only a portion with the ore ten grains, for instance and to introduce the rest wrapped up in paper, as needed during the operation. Some of it may be reserved for the final heating after the lead is slagged over. If too much were added at first, the lead would be slagged over before the necessary reactions had taken place. The Lead Button. The lead button submitted to cupel- lation must be malleable and of the proper size for the cupel. A good cupel will absorb its own weight of litharge, but it is better to use a cupel one-quarter to a third as heavy again as the lead-bottom. The cupels in ordinary use weigh about 250 grains, hence a button of 180 to 230 grains is the proper size for them. If the button be too large it may be reduced in size by scorifica- tion. In case of doubt it is better to scorify, since there is less loss in this operation than in cupellation. A brittle button may be due to the presence of sulphur, arsenic, antimony, zinc, or litharge. In either case it must be scorified before cupellation, and with test-lead if necessary. If the button contain copper it must be scorified until no more copper can be seen on hammering out the button, If nickel be present the button cannot be cupelled. This, however, will occur but rarely. 3. Cupellation. The lead button obtained by the fore- going operations is next cupelled. This operation is similar in many respects to scorification, but differs from it in that the scoriae formed are absorbed entirely by the cupel, leaving a pure bead of the precious metals. This property of the absorption of scorise is an indispensable condition in cupellation, so that, unlike scorification, it is 750 THE ASSAY OF GOLD. limited to a few substances capable of being absorbed by the cupel. The oxides of lead and bismuth alone can be absorbed in a state of purity, but they can carry along with them certain proportions of other substances which, by themselves, would form infusible scoria. It is thus that we are enabled to get rid of small amounts of copper, iron, arsenic, &c., remaining in the lead button, Bismuth is seldom, if ever, used for this purpose. The proportion of lead required to carry off the different impurities varies according to circumstances. In the case of lead buttons from an assay, there is always an excess of lead, if they have been properly purified. Other cases will be treated under the head of Assay of Alloys, where the necessary tables will be given. The operation of cupelling a lead button is conducted t as follows : The cupels must be carefully dried before use, and must be free from cracks, which would cause a loss of precious metal. The bottom of the muffle should be covered with sand to prevent injury to it in case of the up- setting of a cupel. A cupel having been selected, it should be wiped carefully with the finger, and all extraneous matter blown out. It may then be placed in the muffle. Before the introduction of the lead button, the muffle should have attained a reddish-white heat, and the cupel should be of the same temperature. This being attained, the button is placed in the cupel with a pair of forceps, gently, so as not to injure it. The muffle should now be closed, either by a door or by a piece of lighted charcoal, to bring the fused button to the same temperature. This done, the muffle is opened and air allowed to enter ; the button, which at first appears bright and uncovered, is soon covered with a film of oxide moving in luminous patches over the surface, and being continually thrown toward the edge, where it is absorbed by the cupel. White fumes of lead arise, and the button is surrounded by a ring of the absorbed litharge, which continually widens until it reaches the edge of the cupel. The button thus gradually diminishes in size by oxidation and absorption, CUPELLATIOtf. 751 and becomes more convex ; the luminous patches become larger and move more quickly ; and when, finally, the last of the lead is absorbed, the button appears to revolve rapidly on its axis, becomes very brilliant, and is suffused with all the tints of the rainbow ; the movement is sud- denly checked, the button becomes dull for a few instants, and then presents the appearance of the pure precious metals. The latter part of the operation is called the 4 brightening ' of the button. Should the bead be large, and composed, in great part at least, of silver, it must be removed slowly and gradually from the furnace, to pre- vent loss by ' spitting,' which might happen if the cupel were removed at once from the muffle. When a cupel is withdrawn directly after brightening, the button is liable to be covered by mammillated and crystalline protuber- ances, and is then said to have ' vegetated ' or ' spit.' Portions of the metal are sometimes thrown off and lost. Whatever the cause may be, this does not occur if the button be withdrawn gradually, so as to permit a slow and gradual cooling. In case the bead be very large, say 2 to 5 grains, it is well to cover it with a hot cupel, which will retard the cooling and prevent the loss of small particles that may be thrown off despite all care. If the bead is not larger than the head of an ordinary pin, the danger of vegetation is slight, and no great precautions need be taken in its removal. It is well to raise the heat of the furnace just before the brightening, or to push the cupel into a hotter part of the muffle, in order to aid the brightening and get rid of the last traces of lead, which are somewhat difficult to remove. The button should also, for the latter purpose, be heated strongly a few seconds after brightening. Silver is sensibly volatile at a high heat ; the loss of gold is slight, and may be disregarded. In the assay of ores this loss need not be considered, but in bullion assay a correction is necessary, for which a table will be given under Assays of Alloys. The loss of silver increases with the temperature, but we must avoid, in cupellation, the two extremes of a high heat and quick work and a low 752 CUPELLATION. heat with prolonged work. Of the two extremes the latter is worse. The following are indices of favourable work- ing : The muffle is reddish-white, the cupel is red, the fused metal very luminous and clear, the lead fumes rise slowly to the top of the muffle, and the litharge is com- pletely absorbed by the cupel. The heat is too great when the cupels are whitish, when the fused metal is seen with difficulty, and the scarcely visible fumes rise rapidly in the muffle. The heat is too low when the fumes are thick and fall in the muffle, and when the un absorbed litharge is seen forming lumps and scales about the button. The degree of heat should bear some relation to the richness of the alloy, and may be greater according as the lead is poorer in silver. By bearing this in mind the assayer may often hasten the operation without detriment to the assay. The draught of air through the muffle must also be regulated. Too strong a current cools the cupel and oxidises the lead faster than it can be absorbed, thus endangering the assay. Too slow a current prolongs the operation and increases the loss by volatilisation, In ordinary work this matter, however, occasions very little trouble. It happens sometimes that the material in the cupel becomes solidified in the midst of an operation, stopping all further action. This disaster is called the ' freezing ' of the button, and is occasioned by the following conditions : a production of litharge more rapidly than it can be ab- sorbed by the cupel, and infusible scoriae, due to a cold furnace, or to an excess of foreign oxides. In either case the scoriae gradually extend over the surface of the button until it is entirely covered, when further movement ceases. This disaster may sometimes be prevented or remedied by raising the heat of the muffle ; if this fails, or if the acci- dent be due to foreign oxides, an addition of pure lead must be made to the assay ; in either case, the results are unreliable. An assay that has passed well, furnishes a bead well- rounded and clear, crystalline below, and readily to be WEIGHT OP MINUTE SPHERES OF GOLD. 753 detached from the cupel. If the bead contain lead, it is brilliant below, and does not adhere at all to the cupel. If the bead exhibit rootlets passing down into the sub- stance of the cupel, the results are inaccurate. ESTIMATING THE WEIGHT OF MINUTE SPHERES OF GOLD. M. G. A. Gozdorf has given, in the ' Chemical News ' for 1886, an ingenious plan for the production and measure- ment of gold and other minute metallic spheres to esti- mate their weight. In making assays for gold where the amount of gold is very small a little silver is required in which the gold may be collected. As nearly all commercial litharge con- tains silver it is rarely necessary to add any separately for this purpose. Having obtained a prill in which the amount of gold is a third or less than the silver, the prill is boiled in dilute nitric acid in a porcelain capsule to dissolve the silver, and where the amount of gold is more than 1 dwt. to the ton a second boiling in strong nitric acid should be given. If care is taken in using dilute acid at first and boiling gently the gold will be left in one piece of a nearly black colour. The acid is now decanted off and the gold washed two or three times in distilled water. The gold may be now placed on an aluminium or other polished metal plate by invert- ing the capsule and leading the last drop of water and the gold with a glass rod on to the plate, the water is drawn off by a piece of filter-paper, and the plate gently heated till dry. Having thus obtained the gold in a pure state a bead is made of boracic acid on a platinum wire loop and pressed on the gold while still red-hot ; the gold adheres without difficulty, and by heating the bead before the blowpipe the gold is obtained as an almost perfect sphere. Should the resulting sphere of gold be very minute it is better to measure it under the microscope while in the bead, but if large enough to be seen with the naked eye it can be measured more accurately after dissolving the 3c 754 THE ASSAY OF GOLD. boracic acid bead in a watch-glass with hot water and placing the sphere of gold on a glass slide. The plan of measuring minute prills of silver and gold to estimate their weight was first introduced by Harkort, who used an ivory scale engraved with two fine lines meet- ing at an acute angle and divided into fifty equal parts (see fig. 141, p. 724). According to the fifth edition of Plattner's 'Probir- kunst,' p. 520, Goldschmidt estimates the weight of silver and gold prills by measurement with the micro- scope. Harkort and Plattner, in making scales for the esti- mation of the weight of gold and silver prills, weighed prills corresponding to the larger divisions of the scale, and from their weight calculated the weight for the smaller divisions. These prills were taken direct from the cupel, and at the point of contact are flattened ; but as the amount of flattening is not always the same, and hardly varies in extent with the size of the prill, and as the con- verging lines on the scale cannot be very sharply defined, this method is not capable of the same accuracy as where the almost perfect spheres are measured with a microscope. No other flux seems to possess advantages equal to those of boracic acid for obtaining a sphere of gold. Borax and other fluxes are so fluid when hot that the gold is very liable to alloy with the platinum wire : this rarely occurs with boracic acid, on account of its great viscosity, even when white-hot. Boracic acid is also easily soluble in water, so that the gold spheres can be separated without loss of time. The following rules and figures may be useful to any one wishing to adopt the system here described. 1. The weight of a sphere increases as the cube of the diameter. 2. The weight of a sphere of any substance of which the specific gravity is known is obtained by mul- tiplying the weight of a unit sphere of water by the specific gravity of the substance and the cube of the diameter. WEIGHT OP MINUTE SPHERES OF GOLD. 755 Constants for Use with Gramme Weights. 1. Weight of a sphere of water 0-01 m.m. in diameter -0-000,000,000,523,6 of a grm. 2. Weight of a sphere of gold 0-01 m.m. in diameter= 0-000,000,010,210,2 of a grm. 3. Weight of a sphere of gold 0'0# m.m. in diameter, a* x 0-000,000,010,210,2 of a grm. 4. If 20 grms. of ore are taken for assay the number of grains of gold per ton is found by x* x 0*008004, in which ^?=the diameter of the sphere of gold in hundredth s of a millimetre. Constants for Use with Grain Weights. 1. Weight of a sphere of water 0*001 inch in diameter, 0-000,000,132,4 of a grain. 2. Weight of a sphere of gold 0*001 inch in diameter, 0-000,002,582 of a grain. 3. Weight of a sphere of gold 0*00# inch in diameter, x* x 0-000,002,582 of a grain. 4. 200 grains of ore being taken for assay, the num- ber of grains of gold per ton is found by x 3 x 0-2045288, in which #==the diameter of the sphere of gold in thousandths of an inch. By taking 978 grains for assay # 3 = grains of gold per ton. To test the accuracy of the above figures a compara- tively speaking large sphere of gold from an assay was measured and found to be 0*593 m.m. or -^ of a m.m. in diameter; 59*3 3 x 0*000,000,010,210,2=0-002,129 of a grm. When weighed on a very delicate balance it was found to weigh 0*0021 grm., and as this balance does not indicate beyond the fourth decimal the results may be considered identical. This sphere indicated gold in the sample tried at the rate of 3 oz. 9 dwts. 13 grs. per ton. The smallest sphere of gold yet measured was 0*024 m.m. 3 o 2 756 THE ASSAY OF GOLD. in diameter, and by applying the above rule the weight would be 2-4 3 x 0-000,000,010,210,2 x 15-43235 (to convert grammes to grains) = 0- 000, 002,178, or a trifle over two millionths of a grain. Spheres of silver may be obtained and measured in a similar manner. The boracic acid acts slightly on the silver, but the quantity dissolved is inappreciable, as the action is not prolonged. The specific gravity of silver being 10-53, the weight of l-100th m.m. would be 0-000,000,000,523,6x10-53 = 0-000,000,005,513,508 of a grm. In a test assay made with silver the sphere measured 0-57=f$j- of a m.m., from which the weight deduced would be 0-001,020,96 of a grm., the balance showing the weight as 0*0010 of a grm. Copper, lead, and other metals cannot be melted in boracic acid on platinum wire without dissolving to a perceptible amount, but may with care be melted in sodic carbonate, and, by dissolving the latter in hot water, the sphere of copper, &c., obtained and measured. GENERAL OBSERVATIONS ON THE ASSAY OF GOLD ORES. Gold and Copper, Proportion of Lead. The alloys of gold and copper are cupelled like the alloys of gold and silver ; but as copper has a very great affinity for gold, it is necessary to use a larger proportion of lead to insure its oxidation when combined with gold than when united with silver. This proportion varies according to the standard and the temperature. It is admitted that for the same standard there must, under similar circumstances, be twice as much lead used in the cupellation of gold as for that of silver. Thus, 14 parts, at least, ought to be em- ployed in common furnaces for an assay of gold coin which contains 0*1 of copper. There is no inconvenience in employing a little more, as it does not increase the loss of gold. However great the proportion of lead may be that is added to the cupreous gold for the purpose of cupellation, the button retains always a very small quantity of copper, which a fresh cupellation does not free it from, EXAMINATION ON THE TOUCHSTONE. 767 .and which occasions what is termed the surcharge. This .surcharge, being very slight, can be neglected in assays of minerals ; but it is necessary to take notice of it in the assay of alloys. But it is known that the presence of silver much facilitates the separation of copper from gold, and it is rare that an alloy of cupreous gold does not con- tain a little silver, which must be separated : and when that is not the case, a small quantity of that metal can be introduced into the alloy, so as to be in about the propor- tion of 3 parts to 1 of gold. When an assay is to be made of an alloy of gold and copper, a sufficient quantity of silver is to be added to fulfil this condition according to the presumed standard, which is estimated approxi- matively by a preliminary assay, and then cupelled with lead. Examination on the Touchstone. This method is based upon the fact that the richer an alloy is in gold the more clearly does a streak drawn with it on a black ground present a pure gold-yellow colour, and the less is it attacked by pure nitric acid or by a test acid. This test -acid consists of ninety-eight parts pure nitric acid of 1*34 sp. gr. (37 Beaume), two parts pure hydrochloric acid of 1-173 sp. gr. (21 B.), and twenty-five parts distilled water. To judge of the richness of the alloy to be examined, its streak is compared with marks drawn with alloys (the touch-needles) whose richness is accurately known. In order to get correctly the streak of the alloy to be tested, the surface of the metal must first be some- what filed away, since this may be impure, or, as with coins and jewellery, it may have been made somewhat richer by boiling with acid, and the so-called colouring of the goldsmith, and a clean fracture is rarely to be obtained. .Five series of prepared touch-needles are required. The first series consists of copper and gold, and is called the .red series, and the proportion of gold increases by half- carats in the successive needles. The second series, the white series, contains needles of gold and silver, in which the proportion of gold likewise increases by half-carats. The third series, a mixed one, contains needles in which the 758 THE ASSAY OF GOLD. quantities of silver and copper are equal, and the propor- tion of gold also increases by half-carats. The fourth con- sists also of needles for a mixed series, in which the silver is to the copper as 2 : 1, and the gold increases by half- carats ; and the fifth is also formed of needles for a mixed series, in which the quantity of silver is to that of the copper as 1 : 2. Moreover, in mints and stamping bureaux, alloys are used which correspond precisely to the legal standards. The testing upon the touchstone begins by determining to which series the alloy to be examined be- longs. Then those touch-needles are rubbed against the stone whose marks most nearly approximate in colour to that of the alloy. The marks must form a thin continuous layer. A drop of pure nitric acid is now placed upon them with a glass rod, and its comparative effect observed. The acid is allowed to work a short time, and then wiped off, in order to see whether the streak appears unchanged, or whether it has more or less disappeared. The test acid above is also used. This is so composed that it does not work at all upon an alloy containing eighteen carats and more of gold, and with such an alloy the streak, after using the acid, will not be wiped off with a fine linen rag, provided that stone and acid had a temperature of 10 to 12 C. Pure nitric acid produces almost no effect upon an alloy of fifteen or sixteen carats fine, and over. The test- ing on the touchstone can indeed make no pretension to accuracy, especially where the amount of gold is small, but it yields sufficiently useful results for a preliminary test. It requires, however, a sharp and very practised eye. Moreover, the preparation of the touch-needles is weari- some, as the required proportion is not always quickly reached, nor are good malleable alloys always obtained. The touchstone, therefore, is in general only used where frequent gold assays are to be made of alloys varying in richness, or where (as frequently with gold plate) aiL examination on the touchstone will suffice. CUPELLATION OF GOLD AND COPPER. 769 TABLE FOR PROPORTION OF LEAD TO BE EMPLOYED IN THE CUPELLATION OF GOLD AND COPPER. Gold in alloy 1000 thousandths Lead required Eatio of lead in the assay to the copper, pages 243, 244.) * ' Chemical News,' March 20, 1874. 874 GEMS AND PRECIOUS STONES. QUARTZ. Specific gravity, 2 -55 to 2*7 ; hardness, 7. Quartz occurs in many forms, and has often by inexperienced persons been mistaken for the diamond, owing to the lustre of its crystals and its considerable hardness. It, however, can always be distinguished from the diamond by its crys- FIG. 150. FIG. 148. FIG. 149. FIG. 151. FIG. 152. talline faces, hardness, and specific gravity (see example in Table I.) It usually occurs in six-sided prisms, more or less modified, terminated with six-sided pyramids. Traces of cleavage are seldom or ever apparent. Some of its salient forms are shown in figs. 148, 149, 150, 151, 152, 153, 154. COLOURLESS STONES. 875 FIG. 153. FIG. 154. Some crystals are as pellucid as glass ; others, however, assume all the shades of colour mentioned in the case of the diamond. Composition (Si0 2 ) : Pure silica or silicic acid. WHITE ZIRCON. Specific gravity, 4'44 to 4-8 ; hardness, 7'5. This stone is often found crystallised in nature in four-sided prisms, terminated by four-sided or rhomboidal or triangular FIG. 156. FIG. 157. FIG. 155. pyramids, and other forms. See figs. 155, 156, 157, 158, 159. These stones are often employed in jewellery under the name of ' rough diamonds.' They often occur brownish- red and brown, red, yellow, and grey ; these varieties will 876 GEMS AND PEECIOUS STONES. be treated under their appropriate heads. It can be readily distinguished from the diamond and quartz by hardness FIG. 158. FIG. 160. arid specific gravity ; also by the action of strong hydro- chloric acid, which, if dropped on the diamond or quartz, and allowed to remain for a little time, produces no change, but if a zircon be so treated, the spot on which the acid was placed remains dull. Composition (Zr 2 3 ,Si0 2 ) : Zirconia . 67'2 Silicic acid . . . . . . 33'5 100-7 WHITE SAPPHIRE. Specific gravity, 3-97 to 4-27 ; hardness, 9. This stone, in hardness, is next to the diamond. The sapphire occurs variously coloured ; other colours will be discussed under their appropriate heads. It crystallises in the rhombohe- dric system, usually in six-sided prisms, but often so very rough as not to be readily distinguishable. May be dis- tinguished by gravity and hardness from all the preceding. Composition (A1 2 3 ) : Pure alumina. WHITE TOPAZ. Specific gravity, 3'54 ; hardness, 9. This variety of topaz, known for its limpidity by the term ' gouttes d'eau," when polished has nearly the same lustre as the diamond ; COLOURLESS STONES. 877 the topaz, however, occurs of many colours see hereafter. It crystallises in the right rectangular prismatic system. Some of its natural forms are shown in figs. 161, 162, 163, 164, 165, 166. It is readily rendered electric, and retains its electricity for a very considerable time ; it is also pyro-electric, or becomes electric when heated, a property by which it is FIG. 161. FIG. 162. FIG. 163. FIG. 164. FIG. 165. FIG. 166. distinguished from the diamond, its specific gravity being so similar that it cannot be made available as a means of discriminating between the two stones. From the other stones in this group, with the exception of the sapphire, it is readily distinguished by its hardness and gravity, and from the latter by its gravity and pyro-electricity. Composition : Silica Alumina Fluorine 34-2 57-5 7-8 99-5 878 GEMS AND PRECIOUS STONES. TABLE I. COMPARATIVE TABLE OF THE WEIGHTS OF COLOURLESS STONES WEIGHED IN AlR AND WATER. Weight in Weight in Water Air White White White White White Grains Zircon Sapphire Topaz Diamond Quartz 1 0-775 0-766 0-716 0-715 0-611 4 3-10 3-06 2-86 2-86 2-42 8 6-20 6-12 5-72 5-72 4-86 12 9-30 9-18 8-58 8-58 7-31 16 12-40 12-25 11-55 11-45 9-75 20 15-50 15-31 14-42 14-31 12-19 24 18-60 18-37 17-28 17-17 14-64 28 21-70 21-44 | 20-16 20-13 17-08 32 24-80 24-51 23-01 22-90 19-53 36 27-90 27-57 25-88 25-76 11-98 40 31-00 30-64 28-75 28-63 24-43 44 34-10 33-71 31-61 31-49 26-88 48 37-20 36-76 34-47 34-35 29-32 52 40-30 39-82 37-34 37-21 31-77 56 43-40 42-89 40-20 40-17 34-21 60 46-50 45-95 43-06 42-94 36-66 64 49-60 49-01 45-93 45-80 39-11 68 52-70 52-07 48-90 48-66 41-56 72 55-80 55-14 51-77 51-52 44-00 76 58-90 58-21 54-63 54-38 46-44 80 62-00 61-28 57-49 57-24 48-88 84 65-10 64-34 60-35 60-12 51-32 88 68-20 67-41 63-22 62-97 53-76 92 71-30 70-47 66-08 65-30 56-21 96 74-40 73-54 68-94 68-69 58-65 100 77-50 76-60 71-80 71-55 61-09 Specific Gravity \ 4-44 4-27 3-54 3-52 2-55 Example of the use of Table /.* A colourless stone, weighing 40 grains in air, is reduced to 24-43 in water. Look in the first column to 40, and then trace along its horizontal line until a number very nearly approaching 24-43 is found ; refer then to the heading of the table, above the number found, and the name there expressed will be that of the stone examined. Supposing, however, the weight of the stone be 41 grains, still the number 24-43 will be the nearest in the table, and -611 must be added to it, as that sum would be the weight of 41 grains * The Tables of Comparative Weights were calculated by Brard. YELLOW STONES. 879 of quartz or water. From the numbers obtained by cal- culation also can the specific gravity be estimated. If this course be pursued, refer to the bottom line of the table for a corresponding number, and to the heading of the table for the name of the stone. When the weight is any even number of grains (that is, without fractions), the readiest way is to refer to the table (first column) for the number of grains, and then to the horizontal line to corre- sponding number obtained, which is the weight in water. Diamond and topaz, however, have very nearly equal densities, and a second characteristic must be had recourse to in order to determine the nature of two stones which have an equal weight in water. This auxiliary character is the development of electricity by heat, a phenomenon exhibited by the topaz but not by the diamond. The test of hardness may be also resorted to. YELLOW STONES. YELLOW ZIRCON ( JARGON). The crystalline form, characteristics, and composition of this stone have been described under the head ' White Zircon.' YELLOW SAPPHIRE. Characteristics, &c., described under ' White Sapphire.' CYMOPHANE (CHRYSO BERYL). Specific gravity, 3'65 to 3-89 ; hardness, 8-5. The cymophane is nearly as hard as the sapphire, harder than the topaz and the emerald ; it readily scratches quartz. Its colour is greenish-yellow, and has been placed in the list of yellow stones rather than green, because usually the yellowish tint is the most decided. This tint, which is very agreeable in itself, is often relieved by a small spot of light of a bluish- white tinge, which moves from point to point of the stone as the position of the latter is varied. It is S80 GEMS AND PRECIOUS STONES. rarely found in regular crystals, but more generally occurs in rolled and rounded masses. For some of its iorms, however, see figs. 167, 168, 169, 170. FIG. 167. FIG. 168. FIG. 169. FIG. 170. Composition : No. 1 is a sample from the Brazils ; No. 2, from Siberia. i. Alumina 78-10 Glucina 17-94 Iron oxide . . . . . 4*46 Chromium oxide .... Copper and lead oxides ... 100-50 2. 78-92 18-02 3-12 0-36 0-29 100-71 YELLOW TOPAZ. The general characteristics of this stone are described under ' White Topaz.' YELLOW TOURMALINE. Specific gravity, 3-00 to 3-22 ; hardness, 7 to 7'5. The tourmaline becomes electrical by heat ; one portion of a crystal attracts light bodies, the other repels them. Its YELLOW STONES. 881 colour is very varied. The tourmaline has a vitreous fracture. It occurs in semicrystalline prisms of irregular FIG. 171. FIG. 172. FIG. 173. FIG. 174. FIG. 175. FIG. 176. form, generally deeply striated, and in prisms of six or more sides, variously terminated, one end usually differing from the other. Figs. 171, 172, 173, 174, 175, and 176 represent some of the forms of this mineral. YELLOW EMEKALD. Specific gravity, 273 to 2-76 ; hardness, 7-5 to 8. The emerald occurs of many colours ; its tint par excellence is- 3 L 882 GEMS AND PKECIOUS STONES. green ; but there are many varieties tinged more or less yellow or blue, and they even occur white. Its fracture FIG. 178. FIG. 177. FIG. 179. FIG. 180. FIG. 181. FIG. 183. FIG. 182. is vitreous, brilliant, and undulating. Its common form is the hexahedral prism, sometimes deeply striated longi- YELLOW STONES. 883 tudinally. It readily cleaves parallel to all the planes of its primary form the hexahedral prism. The above are some of the forms it assumes : 178, 179, 180, 181, 182, and 183. Composition : figs. 17 Glucina Silica Alumina Iron oxide 15-50 66-45 16-75 60 99-30 The mium ie green varieties contain a small quantity of chro- oxide. TABLE II. COMPARATIVE TABLE OF THE WEIGHTS OF YELLOW STONES WEIGHED IN AlE AND WATER. Weight in air Weight in Water Grrftins Yellow Yellow Yellow Yellow Yellow Yellow Yellow Zircon Sapphire Cymophane Topaz Tourmaline Emerald Quartz i 0-775 0-766 9-738 0-716 0-690 0-633 0-611 4 3-10 3-06 2-95 2-86 2-76 2-53 2-42 8 6-20 6-12 5-90 5-72 5-52 5-06 4-86 12 9-30 9-18 8-85 8-58 8-28 7-59 7-31 16 12-40 12-25 11-80 11-55 11-04 10-12 9-75 20 15-50 15-31 14-75 14-42 13-80 12-65 12-19 24 18-60 18-07 17-70 17-28 16-56 15-19 14-04 28 21-70 21-44 20-65 20-15 19-32 17-72 17-08 32 24-80 24-51 23-60 23-01 20-08 20-25 19-53 36 27-90 27-57 26-55 25-88 24-84 22-77 21-98 40 31-00 30-64 29-50 29-75 27-60 25-30 24-43 44 34-10 33-71 32-45 31-61 30-36 27-83 26-88 48 37-20 36-76 35-40 34-47 33-12 30-36 29-32 52 40-30 39-82 38-35 37-34 35-88 32-89 31-77 56 43-40 42-89 41-30 40-20 38-64 35-43 34-21 60 46-50 45-95 44-25 43-06 41-40 37-94 36-66 64 49-60 49-01 47-20 45-93 44-16 40-47 39-11 68 52-70 52-08 50-15 48-90 46-92 43-00 41-56 72 55-80 55-14 53-10 51-77 49-68 45-53 44-00 76 58-90 58-21 56-05 54-63 52-44 48-07 46-44 80 62-00 61-28 59-00 57-49 55-20 50-60 48-88 84 65-10 64-34 61-95 60-35 57-96 53-13 51-32 88 68-20 67-41 64-90 63-22 60-72 55-66 53-76 92 71-30 70-47 67-85 66-08 63-48 58-19 56-21 96 74-40 73-54 70-80 68-94 66-24 60-72 58-65 100 77-50 76-60 73-75 71-80 69-00 63-25 61-09 Specific Gravity j- 4-44 4-27 3-89 3-53 3-22 2-72 2-55 3 L 2 884 GEMS AND PRECIOUS STONES. YELLOW QUARTZ. For the characteristics, hardness, &c., of this mineral, see ' White Quartz/ BKOWN AND FLAME-COLOURED STONES. ZIRCON (HYACINTH). For characteristics, &c., see ' White Zircon/ VERMEIL GARNET, NOBLE GARNET, ALMANDINE. Specific gravity, 4 to 4*2 ; hardness, 6-5 to 7*5. There are very many varieties of garnet, variously coloured ; but FIG. 184. FIG. 185. FIG. 186 their crystalline form a rhombic dodecahedron, more or less modified is a distinguishing characteristic. The colouring matter of the garnet is iron. Figs. 184, 185, 186, 187, and 188 represent some of its crystalline forms. BROWN AND FLAME-COLOURED STONES. FIG. 187. FIG. 188. 885 Composition : Silica Alumina . Iron oxide Manganese oxide 33-75 27-25 36-00 25 97-25 TABLE III. COMPARATIVE TABLE OF THE WEIGHTS OF BROWNISH AND FLAME- COLOURED STONES WEIGHED IN AIR AND WATER. Weight in Air Weight in water Grains Hy acinthine Zircon Vermeil Garnet Essonite Tourmaline 1 0-775 0-750 0-710 0-690 4 3-10 3-00 2-87 2-76 8 6-20 6-00 5-74 5-52 12 9-30 9-00 8-61 8-28 16 12-40 12-00 11-48 11-04 20 15-50 15-00 14-35 13-80 24 18-60 18-00 17-22 16-56 28 21-70 21-00 20-09 19-32 32 24-80 24-00 22-96 22-08 36 27-90 27-00 25-83 24-84 40 31-30 30-00 28-70 27-60 44 34-10 33-00 31-57 30-36 48 37-20 36-00 34-44 33-12 52 40-30 39-00 37-31 35-88 56 43-40 42-00 40-18 38-64 60 46-50 45-00 43-05 41-40 64 49-60 48-00 45-92 44-16 68 52-70 51-00 48-79 46-92 72 55-80 54-00 51-66 49-68 76 58-90 57-00 54-53 52-44 80 61-00 60-00 57-40 55-20 84 65-10 63-00 60-27 57-96 88 68-20 66-00 63-14 60-72 92 71-30 69-00 66-01 63-48 96 74-40 72-00 68-88 66-24 100 77-50 75-00 71-75 69-00 Specific Gravity } 4-44 4-00 3-54 3-22 886 GEMS AND PKECIOUS STONES. ESSONITE, CINNAMON STONE. Specific gravity, 3*5 to 3 -6. This stone has an agree- able orange-yellow tinge, which becomes a warm and brilliant tint when 'the mass is large. This stone is not usually found crystalline, but in irregular forms and masses, which are characterised by fissures in all direc- tions. Composition : Silica 38-80 Alumina 21-20 Lime 31-25 Iron oxide with small quantities of Potash "1 Q-Q and Magnesia / 97-75 TOURMALINE. For the characteristics of this mineral see 'Yellow Tourmaline.' KED AND KOSE-COLOURED STONES. EED SAPPHIEE (ORIENTAL RUBY). For characteristics, crystalline form, &c., see ' White Sapphire.' DEEP RED GARNET, NOBLE GARNET. For characteristics, &c., see ' Vermeil Garnet.' SPINEL RUBY. Specific gravity, 3*5 to 3-6 ; hardness, 8. The spinel readily scratches quartz, but is scratched by the sapphire. Its special colour is red, approaching a rose tint : this tinge, however, undergoes various modifications, such as scarlet, red, rose, yellowish-red, and reddish-purple : it is also found blue and black. Its fracture is flattish- conchoidal, with a splendent vitreous lustre. It occurs crystallised in regular octahedrons, sometimes having their edges replaced as in macles : sometimes it assumes the globular form. The spinel may be distinguished from the true ruby and the garnet by hardness and specific EED AND KOSE-COLOUKED STONES. 887 gravity ; and from reddish topaz, which possesses nearly the same specific gravity, by its electric properties. Composition of spinel ruby : Silica -02 Alumina . 69-01 Magnesia 26'21 Iron protoxide 0'71 Chromium oxide . . . . . 1-11 99-06 EEDDISH TOPAZ. For characteristics, &c., see ' White Topaz.' KED TOURMALINE. For characteristics, &c., see 4 Yellow Tourmaline.' TABLE IV. COMPARATIVE TABLE OF THE WEIGHTS OF EED OR EOSE-COLOURED STONES WEIGHED IN AIR AND WATER. Weight Weight in Water in Air Grains Red Sapphire (True Ruby) Deep Garnets Spinel Smoke or Red Topaz Red Tourmaline 1 0-766 0-750 0-722 0-716 0-690 4 3-060 3-700 2-880 2-860 2-760 8 6-120 6-000 5-770 5-720 5-520 12 9-180 9-000 8-660 8-585 8-280 16 12-250 12-000 11-550 11-550 11-040 20 15-310 15-000 14-440 14-420 13-800 24 18-370 18-000 17-330 17-280 16-560 28 21-440 21-000 20-220 20-150 19-320 32 24-510 24-000 23-110 23-610 22-080 36 27-570 27-000 26-000 25-880 24-840 40 30-640 30-000 28-880 28-750 27-600 44 33-710 33-000 31-770 31-610 30-360 48 36-760 36-000 34-660 34-470 33-120 52 39-820 39-000 37-550 37-340 35-880 56 42-890 42-000 40-440 40-200 38-640 60 44-950 45-000 43-300 43-060 41-400 64 49-010 48-000 46-220 45-930 44-160 68 52-080 51-000 49-110 48-900 46-920 72 55-140 54-000 51-990 51-770 49-680 76 58-210 57-000 54-880 54-630 52-440 80 61-280 60-000 57-770 57-490 52-200 84 64-340 63-000 60-660 60-350 57-960 88 67-410 66-000 63-550 63-220 60-720 92 70-470 69-000 66-440 66-080 63-480 96 73-540 72-000 69-330 68-940 66-240 100 76-600 75-000 72-220 71-800 69-000 Specific Gravity } 4-270 4-000 3-600 3-530 3-220 888 GEMS AND PRECIOUS STONES. BLUE STONES. BLUE SAPPHIEE. For characteristics, &c., see ' White Sapphire.' DISTHENE, CYANITE. Specific gravity, 3'5 to 3*7 ; hardness, 5 to 7. Fine specimens of disthene possess a bright blue colour, which passes insensibly into a deep sky blue. Its transparency is nearly perfect, and it presents small pearly reflections, which add to the beauty of its colour. The primary form of its crystals is a doubly oblique prism, and they cleave very readily in the direction of their length. It can be FIG. 189. FIG. 190, FIG. 191. readily distinguished from the sapphire by its being less hard, as also by its specific gravity. Figs. 189, 190, and 191 represent some of its crystalline forms. Composition of a specimen from St. Gothard : Silica 43-0 Alumina 55-0 Iron oxide -5 98-5 BLUE TOPAZ. For characteristics, &c., see fc White Topaz.' Blue topaz and disthene, having the same specific gravity, may by that test alone be confounded with each other ; but the appearance of each is so different that they can rarely be confounded. If, however, the electrical test be applied, no fear of mistaking one for the other need be entertained, as only the topaz becomes electrical. BLUE STONES. BLUE TOURMALINE. For characteristics, &c., see ' Yellow Tourmaline.' BLUE BERYL. For characteristics, &c., see 'Emerald.' The tint and appearance of this stone and that of the blue topaz are so similar that they cannot be distinguished by that test ; their specific gravities, however, are so different that they may, by this simple means, be readily discriminated. DICHROITE, WATER SAPPHIRE. Specific gravity, 2*56 to 2-65 ; hardness, 7 to 7'5. The chief characteristic of this stone is that it possesses a double colour ; that is, it is a fine blue or a normal yellow, as it is viewed in the direction of its base, or the planes of a hexahedral prism, which is its crystalline form. It can be thus readily distinguished, as also by its having nearly the same specific gravity as quartz, and thus being the lightest of the blue stones. Composition : Silica. . . Alumina Magnesia . Iron protoxide . Manganese protoxide Loss in fire (water ? ) 48-35 31-71 10-16 8-12 33 60 99-27 TURQUOISE. Specific gravity, 2-8 to 3 ; hardness, 5 to 6. This stone has not been placed in the list of specific gravities, as it can be so readily detected by its appearance. It is bright or greenish-blue in colour ; its aspect is earthy or compact. It scratches apatite, and even glass ; but is scratched by quartz. It occurs filling fissures, or forming concretions in siliceous and argillo-ferruginous rocks. Composition : Phosphoric acid 17-86 Alumina . Silica Iron peroxide . Lime . Water and fluoric acid 10-01 8-90 36-82 0-15 25-95 99-69 890 GEMS AND PRECIOUS STONES. TABLE V. COMPARATIVE TABLE OF THE WEIGHTS OF BLUE STONES WEIGHED IN AIR AND WATER. Weight iii Water Weight in Air | Grains Blue Sapphire Disthene Cyanite Blue Topaz Tourmaline Blue Beryl Dichroite, Water Sapphire 1 0-766 0-717 0-716 0-690 0-633 0-622 4 3-06 2-87 2-86 2-16 2-53 2-49 8 6-12 5-74 5-72 5-52 5-06 4-98 12 9-18 8-61 8-58 8-28 7-59 7-47 16 12-25 11-48 11-45 11-04 10-12 9-96 20 15-31 ' 14-35 14-42 13-80 12-65 12-45 24 18-37 17-22 17-18 16-56 15-19 14-94 28 21-44 20-09 20-05 19-32 17-72 17-43 82 24-51 22-96 22-91 20-08 20-25 19-92 36 27-57 25-83 25-78 24-84 22-77 22-41 40 30-64 28-70 28-65 27-60 25-30 24-90 44 33-71 31-57 31-51 30-66 27-83 27-39 48 36-76 34-44 34-37 33-12 30-36 29-88 52 39-82 37-31 37-24 35-88 32-89 32-37 56 42-89 40-18 40-10 38-64 35-43 34-88 60 45-95 43-05 42-96 41-40 37-94 37-35 64 49-01 45-92 45-83 44-16 40-47 39-84 68 52-08 48-79 48-80 46-92 43-00 42-33 72 55-14 51-66 51-67 49-68 45-53 44-82 76 58-21 54-53 54-53 52-44 48-07 47-31 80 61-28 57-40 57-49 55-20 50-60 49-80 84 64-34 60-27 60-25 57-96 53-13 52-29 88 67-41 63-14 63-12 60-72 55-66 54-78 92 70-47 66-01 65-98 63-48 58-19 57-27 96 73-54 68-88 63-84 66-24 60-72 59-76 100 76-60 71-75 71-70 69-00 63-25 62-25 Specific Gravity | 4-27 3-54 3-53 3-22 2-72 2-65 VIOLET STONES. VIOLET SAPPHIRE. For characteristics, &c., see 4 White Sapphire.' VIOLET TOURMALINE. For characteristics, &c., see ' Yellow Tourmaline.' VIOLET QUARTZ, AMETHYST. For characteristics, &c., see ' White Quartz.' GREEN STONES. 891 TABLE VI. COMPARATIVE TABLE OF THE WEIGHTS OF VIOLET STONES WEIGHED IN AIR AND WATER. Weight in Air Weight in Air Grains Violet Sapphire Violet Tourmaline Amethystine Quartz (Amethyst; 1 0-766 0-690 0-611 4 3-06 2-76 2-42 8 6-12 5-52 4-86 12 9-18 8-28 7-31 16 12-25 11-04 9-75 20 15-31 13-80 12-19 24 18-37 16-56 14-64 28 21-44 19-32 17-08 32 24-51 20-08 19-53 36 27-57 24-84 21-98 40 30-64 27-60 24-43 44 33-71 30-36 26-88 48 36-76 33-12 29-32 52 39-82 35-88 31-77 56 42-89 38-64 34-21 60 45-95 41-40 36-66 64 49-01 44-16 39-11 68 52-02 46-92 41-56 72 55-14 49-68 44-00 76 58-21 52-44 46-44 80 61-28 55-20 48-88 84 64-34 57-96 51-32 88 67-41 60-72 53-76 92 70-47 63-48 56-21 96 73-54 66-24 58-65 100 76-60 69-00 61-09 Specific Gravity 4-27 3-22 2-55 GBEEN STONES. GREEN SAPPHIRE. For characteristics, &c., see 'Yellow Emerald.' PERIDOT, CRYSOLITE. Specific gravity, 3*3 to 3-5 ; hardness, 6*5 to 7.. This stone has a more or less deep olive or yellowish -green colour. It is more generally found in rolled grains than in regular prismatic crystals. It is possessed in a very high degree of double refraction. Figs. 192, 193, 194, and 195 represent some of its crystalline forms. 392 GEMS AND PEECIOUS STONES. GEEEN TOUEMALINE. For characteristics, see ' Yellow Tourmaline. EMEEALD. For characteristics, see ' Yellow Emerald.' FIG. 192. FIG. 193. FIG. 194. FIG. 195. AQUA-MAEINE. This stone possesses a very pale green tinge. For other characteristics, see ' Yellow Emerald.' CHEYSOPEASE. This mineral is a green-coloured quartz, and can be readily recognised by referring to the characteristics of quartz. GREEN STONES. TABLE VII. COMPARATIVE TABLE OF THE WEIGHTS OF GREEN STONES WEIGHED IN AIR AND WATER. Weight 1T1 A ! Weight in Water in Air Grains Green Sapphire Peridot Green Tourmaline Emerald Aqua-marine Chrysoprase i 0-766 0-708 0-690 0-633 0-633 0-611 4 3-06 2-83 2-76 2-53 2-53 2-42 8 6-12 5-66 5-52 5-06 5-06 4-86 12 9-18 8-49 8-28 7-59 7-59 7-31 16 12-25 11-32 11-04 10-12 10-12 9-75 20 15-31 14-16 13-18 12-65 12-65 12-19 24 18-37 16-99 16-56 15-19 15-19 14-64 28 21-44 19-82 19-32 17-72 17-72 17-08 32 24-51 22-65 22-08 20-25 20-25 19-53 36 27-57 25-48 24-84 22-77 27-77 21-98 40 30-64 28-32 27-60 25-30 25-30 24-43 44 33-71 31-15 30-36 27-83 27-83 36-88 48 36-76 33-98 33-12 30-36 30-36 29-32 52 39-82 36-81 35-88 32-89 32-89 31-77 56 42-89 39-64 38-64 35-43 35-43 34-21 60 45-95 42-48 41-40 37-94 37-94 36-66 64 49-01 45-31 44-16 40-47 40-47 39-11 68 52-08 48-14 46-92 43-00 43-00 41-56 72 55-14 50-97 49-68 45-53 45-53 44-00 76 58-21 53-80 52-44 48-07 48-07 46-44 80 61-28 56-64 55-20 50-60 50-60 48-88 84 64-34 59-47 57-96 53-13 53-13 51-32 88 67-41 62-30 60-72 55-66 55-66 53-76 92 70-47 65-13 63-48 58-19 58-19 56-21 96 73-54 67-96 66-24 60-72 60-72 58-65 100 76-60 70-80 69-00 63-25 63-25 61-09 Specific Gravity | 4-27 3-42 3-22 2-72 2-72 2-56 STONES POSSESSING A PLAY OF COLOUES (CHATOYANT). In the following list of stones no regard has been paid to absolute colours, but only to the play of colours the stones exhibit. This play or reflection is of two kinds : in some, as the sapphires, it appears as a white star with six rays, on a blue, red, or yellow ground ; or on a purple ground in the garnet. In others it is but a point or mass of pearly light, which sometimes appears to occupy the whole of the stone, and varies according to the inclination given to the stone. The cymophane, cry soli te quartz, Egyptian emerald, felspar, and cat's eye belong to this class. GEMS AND PRECIOUS STONES. The specific gravities of such stones as the opal, &c., have not been given, as their appearance sufficiently charac- terises them. SAPPHIRE. For characteristics, &c., see 'White Sapphire.' GARNET. For characteristics, &c., see ' Vermeil Garnet.' See ' Cymophane.' CYMOPHANE. ANTIQUE EMERALD. For characteristics, &c., see ' Yellow Emerald.' QUARTZ. See ' White Quartz.' FIG. 196. FIG. 197. FIG. 198. FIG. 199. FIG. 201. FIG. 200. STONES POSSESSING A PLAY OF COLOURS. 895 FELSPAR, NACREOUS FELSPAR, FISH-EYE, ETC. Specific gravity, 2'3 to 2-5 ; hardness 4'5 to 5. This species of felspar has a lamellar texture. It will be seen by the lowness of its specific gravity that it cannot be readily confounded with other stones. In appearance its transparency is nebulous, and it presents pearly white reflections, which float about and vacillate in proportion as its position changes. The foregoing are some of the forms of felspar ; see figs. 183, 184, 185, 186, 187, and 188. Composition : Potash 5-26 Silica 52-90 Lime 25-20 Water ....... 16-00 Fluoric acid 0*82 100-18 TABLE VIII. COMPARATIVE TABLE OF THE WEIGHTS OF STONES POSSESSING A PLAY OF COLOURS (CHATOYANT). Weight in Water Weight in Air Grains Sapphires Garnets Gymophane Antique Emerald Quartz Felspar 1 0-766 0-750 0-738 0-633 0-611 0-592 4 3-06 3-00 2-95 2-53 2-42 2-37 8 6-12 6-00 5-90 5-06 4-86 4-74 12 9-18 9-00 8-85 7-59 7-31 7-11 16 12-25 12-00 11-80 10-12 9-75 9-47 20 15-31 15-00 14-75 12-65 12-19 11-84 24 18-37 18-00 17-70 15-19 14-64 14-20 28 21-44 21-00 20-65 17-72 17-08 16-57 32 24-51 24-00 23-60 20-25 19-53 18-94 36 27-57 27-00 26-55 22-77 21-98 21-31 40 30-64 30-00 29-50 25-30 24-43 23-68 44 33-71 33-00 32-46 27-83 26-88 26-05 48 36-76 36-00 35-40 30-36 29-32 28-42 52 39-84 39-00 38-35 32-89 31-77 30-79 56 42-89 42-00 41-30 35-43 34-21 33-15 60 45-95 45-00 44-25 37-94 36-66 35-52 64 49-01 48-00 47-20 40-47 39-11 37-88 68 52-07 51-00 50-15 43-00 41-56 40-25 72 55-14 54-00 53-10 45-53 44-00 42-62 76 58-21 57-00 56-05 48-07 46-44 44-99 80 61-28 60-00 59-00 50-60 48-88 47-36 84 64-34 63-00 61-95 53-13 51-32 49-73 88 67-47 66-00 64-90 55-66 53-76 52-10 92 70-47 69-00 67-85 58-19 56-21 54-47 96 73-54 72-00 70-80 60-72 58-65 56-84 100 76-60 75-00 73-75 63-25 61-09 59-21 Specific Gravity | 4-27 4-00 3-89 2-72 2-55 2-45 896 GEMS AND PRECIOUS STONES. GLASS AND AETIFICIAL GEMS. Glass is often used to imitate gems, but can be easily distinguished by the following characters : 1. Inferior brilliancy. Though in many cases artificial gems have a fine lustre, they are invariably soft. The materials which communicate brilliancy to glass impair its hardness ; and the result is that glass gems, when examined by a lens, are generally found to have blunt or jagged corners and edges, and surfaces covered with minute, irregular scratches. This is invariably the case after the glass gem has been a little in use, and thus the brilliancy soon becomes impaired. 2. Inferior hardness. Artificial gems can be scratched with a knife, using a slight pressure. Faint scratches are made visible by breathing upon them, whereas true gems retain the original brilliancy which they possessed on leaving the lapidary's wheel. The polished faces of true gems cannot be scratched with a knife ; moreover, after long wear, they show no signs of scratches. 3. Polished gems become readily electric by friction, particularly the topaz and diamond ; but glass imitations require much longer friction to produce the same effect, and also retain the electric power for a shorter time. 4. Glass is fusible in the blowpipe flame. APPENDIX. TABLE L Showing the Quantity of FINE GOLD in 1 oz. of any ALLOY to of a Carat Grain and the MINT VALUE of 1 oz. of each Alloy. FINE GOLD, Per Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Drvts. Grs. Carats GTS. E 'sjhtlis s. d. 1 000 24 4 4 11-4545 19 23 375 23 3 7 4 4 10-1271 19 22 750 23 3 6 4 4 8-7997 19 22 125 23 3 5 4 4 7-4723 19 21 500 23 3 4 4 4 6-1448 19 20 875 23 3 3 4 4 4-8174 19 20 250 23 3 2 4 4 3-4900 19 19 625 23 3 1 4 4 2-1626 19 19 000 23 3 4 4 0-8352 19 18 375 23 2 7 4 3 11-5078 19 17 750 23 2 6 4 3 10-1804 19 17 125 23 2 5 4 3 8-8529 19 16 500 23 2 4 4 3 7-5255 19 15 875 23 2 3 4 3 6-1981 19 15 250 23 2 2 4 3 4-8707 19 14 625 23 2 1 4 3 3-5433 19 14 000 23 2 4 3 2-2159 .0 19 13 375 23 1 7 4 3 0-8885 19 12 750 23 1 6 4 2 11-5610 19 12 125 23 1 5 4 2 10-2336 19 11 500 23 1 4 4 2 8-9062 19 10 875 23 1 3 4 2 7-5788 19 10 250 23 1 2 4 2 6-2514 19 9 625 23 1 1 4 2 4-9240 19 9 000 23 1 4 2 3-5965 19 8 375 23 7 4 2 2-2691 19 7 750 23 6 4 2 0-9417 19 7 125 23 5 4 1 11-6143 19 6 500 23 4 4 1 10-2869 19 5 875 23 3 4 1 8-9595 19 5 250 23 2 4 1 7-6321 19 4 625 23 1 4 1 6-3047 19 4 000 23 4 1 4-9772 19 3 375 22 3 7 4 1 3-6498 19 2 750 22 3 6 4 1 2-3224 GOLD-VALUING TABLE. Ill FINE GOLD, Per Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce O*. DivU Grs. Carats Grs. Eighths s. d. 19 2-125 22 3 5 4 1 0-9950 19 1-500 22 3 4 4 11-6676 19 0-875 22 3 3 4 10-3402 19 0-250 22 3 2 4 8-0127 18 23-625 22 3 1 4 7-6854 18 23-000 22 3 4 6-3579 18 22-375 22 2 7 4 4-0305 18 21-750 22 2 6 4 3-7031 18 21-125 22 2 5 4 2-3757 18 20-500 22 2 4 4 0-0482 18 19-875 22 2 3 3 19 11-7208 18 19-250 22 2 2 3 19 10-3934 18 18-625 22 2 1 3 19 8-0660 18 18-000 22 2 3 19 7-7386 18 17-375 22 1 7 3 19 6-4112 18 16-750 22 1 6 3 19 4-0838 18 16-125 22 1 5 3 19 3-7563 18 15-500 22 1 4 3 19 2-4289 18 14-875 22 1 3 3 19 0-1015 18 14-250 22 1 2 3 18 11-7741 18 13-625 22 1 1 3 18 10-4467 18 13-000 22 1 3 18 8-1193 18 12-375 22 7 3 18 7-7919 18 11-750 22 6 3 18 6-4644 18 11-125 22 5 3 18 4-1370 18 10-500 22 4 3 18 3-8096 18 9-875 22 3 3 18 2-4822 18 9-250 22 2 3 18 0-1548 18 8-625 22 1 3 17 11-8274 18 8-000 22 3 17 10-5000 18 7-375 21 3 7 3 17 .8-1725 18 6-750 21 3 6 3 17 7-8451 18 6-125 21 3 5 3 17 6-5177 18 5-500 21 3 4 3 17 4-1903 18 4-875 21 3 3 3 17 3-8629 18 4-250 21 3 2 3 17 2-5355 18 3-625 21 3 1 3 17 0-2081 18 3-000 21 3 3 16 11-8806 18 2-375 21 2 7 3 16 10-5532 18 1-750 21 2 6 3 16 8-2258 18 1-125 21 2 5 3 16 7-8984 18 0-500 21 2 4 3 16 6-5710 17 23-875 21 2 3 3 16 4-2436 17 23-250 21 2 2 3 16 3-9162 ?> M 2 IV GOLD-VALUING TABLE. FINE Per GrOLD, Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Dn-ts. Grs. Carats Grs. Eighths s. d. 17 22-625 21 2 1 3 16 2-5887 17 22-000 21 2 3 16 1-2613 17 21-375 21 1 7 3 15 11-9339 17 20-750 21 1 6 3 15 10-6065 17 20-125 21 1 5 3 15 9-2791 17 '19-500 21 1 4 3 15 7-9517 17 18-875 21 1 3 3 15 6-6243 17 18-250 21 1 2 3 15 5-2968 17 17-625 21 1 1 3 15 3-9694 17 17-000 21 1 3 15 2-6420 17 16-375 21 7 3 15 1-3146 17 15-750 21 6 3 14 11-9872 17 15-125 21 5 3 14 10-6598 17 14-500 21 4 3 14 9-3324 17 13-875 21 3 3 14 8-0049 17 13-250 21 2 3 14 6-6775 17 12-625 21 1 3 14 5-3501 17 12-000 21 3 14 4-0227 17 11-375 20 3 7 3 14 2-6953 17 10-750 20 3 6 3 14 1-3678 17 10-125 20 3 5 3 14 0-0404 17 9-500 20 3 4 3 13 10-7130 17 8-875 20 3 3 3 13 9-3856 17 8-250 20 3 2 3 13 8-0582 17 7-6^5 20 3 1 3 13 6-7308 17 7-000 20 3 3 13 5-4034 17 6-375 20 2 7 3 13 4-0759 17 5-750 20 2 6 3 13 2-7485 17 5-125 20 2 5 3 13 1-4211 17 4-500 20 2 4 3 13 0-0937 17 3-875 20 2 3 3 12 10-7663 17 3-250 20 2 2 3 12 9-4389 17 2-625 20 2 1 3 12 8-1115 17 2-000 20 2 3 12 6-7840 17 1-375 20 1 hr t 3 12 5-4566 17 0-750 20 1 6 3 12 4-1292 17 0-125 20 1 5 3 12 2-8018 16 23-500 20 1 4 3 12 1-4744 16 22-875 20 1 3 3 12 0-1470 16 22-250 20 1 2 3 11 10-8196 16 21-625 20 1 1 3 11 9-4921 16 21-000 20 1 3 11 8-1647 16 20-375 20 7 3 11 6-8373 16 19-750 20 6 3 11 5-5099 GOLD-VALUING TABLE. FINE Per GrOLD, Ounce CARAT G-OLD, Per Ounce STERLING VALUE, Per Ounce Oz. Dmts Grs. Carats Grs. Eighths s. d. 16 19-125 20 5 3 11 4-1825 16 18-500 20 4 3 11 2-8551 16 17-875 20 3 3 11 1-5277 16 17-250 20 2 3 11 0-2002 16 16-625 20 1 3 10 10-8728 16 16-000 20 3 10 9-5454 16 15-375 19 3 7 3 10 8-2180 16 14-750 19 3 6 3 10 6-8906 16 14-125 19 3 5 3 10 5-5632 16 13-500 19 3 4 3 10 4-2357 16 12-875 19 3 3 3 10 2-9083 16 12-250 19 3 2 3 10 1-5809 16 11-625 19 3 1 3 10 0-2534 16 11-000 19 3 3 9 10-9260 16 10-375 19 2 7 3 9 9-5986 16 9-750 19 2 6 3 9 8-2712 16 9-125 19 2 5 3 9 6-9437 16 8-500 19 2 4 3 9 5-6163 o 16 7-875 19 2 3 3 9 4-2889 16 7-250 19 2 2 3 9 2-9615 16 6-625 19 2 1 3 9 1-6341 16 6-000 19 2 3 9 0-3067 16 5-375 19 1 7 3 8 10-9793 16 4-750 19 1 6 3 8 9-6518 16 4-125 19 1 5 3 8 8-3244 16 3-500 19 1 4 3 8 6-9970 16 2-875 19 1 3 3 8 5-6696 16 2-250 19 1 2 3 8 4-3422 16 1-625 19 1 1 3 8 3-0148 16 1-000 19 1 3 8 1-6874 16 0-375 19 7 3 8 0-3599 15 23-750 19 6 3 7 11-0325 15 23-125 19 5 3 7 9-7051 15 22-500 19 4 3 7 8-3777 15 21-875 19 3 3 7 7-0503 15 21-250 19 2 3 7 5-7229 15 20-625 19 1 3 7 4-3955 15 20-000 19 3 7 3-0681 15 19-375 18 3 7 3 7 1-7407 15 18-750 18 3 6 3 7 0-4133 15 18-125 18 3 5 3 6 11-0859 15 17-500 18 3 4 3 6 9-7585 15 16-875 18 3 3 3 6 8-4311 15 16-250 18 3 2 3 6 7-1036 VI GOLD-VALUING TABLE. FINE GOLD, CARAT GOLD, STERLING VALUE, Per Ounce Per Ounce Per Ounce 6b. JJivts. Grs. Car a 's Grs. Eighths s. d. 15 15-625 18 3 1 3 6 5-7762 15 15-000 18 3 3 6 4-4488 15 14-375 18 2 7 3 6 3-1214 15 13-750 18 2 6 3 6 1-7940 15 13-125 18 2 5 3 6 0-4666 15 12-500 18 2 4 3 5 11-1392 15 11-875 18 2 3 3 5 9-8117 15 11-250 18 2 2 3 5 8-4843 15 10-625 18 2 1 3 5 7-1569 15 10-000 18 2 3 5 5-8295 15 9-375 18 1 7 3 5 4-5021 15 8-750 18 1 6 3 5 3-1747 15 8-125 18 1 5 3 5 1-8473 15 7-500 18 1 4 3 5 0-5198 15 6-875 18 1 3 3 4 11-1924 15 6-250 18 1 2 3 4 9-8650 15 5-625 18 1 1 3 4 8-5376 15 5-000 18 1 3 4 7-2102 15 4-375 18 7 3 4 5-8828 15 3-750 18 6 3 4 4-5554 15 3-125 18 5 3 4 3-2279 15 2-500 18 4 3 4 1-9005 15 1-875 18 3 3 4 0-5731 15 1-250 18 2 3 3 11-2457 15 0-625 18 1 3 3 9-9183 15 o-ooo 18 3 3 8-5909 14 23-375 17 3 7 3 3 7-2634 14 22-750 17 3 6 3 3 5-9360 14 22-125 17 3 5 3 3 4-6086 14 21-500 17 3 4 3 3 3-2812 14 20-875 17 3 3 3 3 1-9538 14 20-250 17 3 2 3 3 0-6264 14 19-625 17 3 1 3 2 11-2990 14 19-000 17 3 3 2 9-9715 14 18-375 17 2 7 3 2 8-6441 14 17-750 17 2 6 3 2 7-3167 14 17-125 17 2 5 3 2 5-9893 14 16-500 17 2 4 3 2 4-6619 14 15-875 17 2 3 3 2 3-3345 14 15-250 17 2 2 3 2 2-0071 14 14-625 17 2 1 3 2 0-6796 14 14-000 17 2 3 1 11-3522 14 13-375 17 1 7 3 1 10-0248 14 12-750 17 1 6 3 1 8-6974 GOLD-VALUING TABLE. Vll FINE G-OLD, Per Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Divts. Grs. Carats Grs, Eighths s. d. 14 12-125 17 1 5 3 I 7-3700 14 11-500 17 1 4 3 1 6-0426 14 10-875 17 1 3 3 1 4-7152 14 10-250 17 1 2 3 1 3-3877 14 9-625 17 1 1 3 1 2-0603 14 9-000 17 1 3 1 0-7329 14 8-375 17 7 3 11-4055 14 7-750 17 6 3 10-0781 14 7-125 17 5 3 8-7507 14 6-500 17 4 3 7-4233 14 5-875 17 3 3 6-0958 14 5-250 17 2 3 4-7684 14 4-625 17 1 3 3-4410 14 4-000 17 3 2-1136 14 3-375 16 3 7 3 0-7862 14 2-750 16 3 6 2 19 11-4588 14 2-125 16 3 5 2 19 10-1313 14 1-500 16 3 4 2 19 8-8039 14 0-875 16 3 3 2 19 7-4765 14 0-250 16 3 2 2 19 6-1491 13 23-625 16 3 1 2 19 4-8217 13 23-000 16 3 2 19 3-4943 13 22-375 16 2 7 2 19 2-1669 13 21-750 16 2 6 2 19 0-8394 13 21-125 16 2 5 2 18 11-5120 13 20-500 16 2 4 2 18 10-1846 13 19-875 16 2 3 2 18 8-8572 13 19-250 16 2 2 2 18 7-5298 13 18-625 16 2 1 2 18 6-2024 13 18-000 16 2 2 18 4-8750 13 17-375 16 1 7 2 18 3-5475 13 16-750 16 1 6 2 18 2-2201 13 16-125 16 1 5 2 18 08927 13 15-500 16 1 4 2 17 11-5653 13 14-875 16 1 3 2 17 10-2377 13 14-250 16 1 2 2 17 8-9103 13 13-625 16 1 1 2 17 7-5829 13 13-000 16 1 2 17 6-2554 13 12-375 16 7 2 17 4-9280 13 11-750 16 6 2 17 3-6006 13 11-125 16 5 2 17 2-2732 13 10-500 16 4 2 17 0-9458 13 9-875 16 3 2 16 11-6184 13 9-250 16 2 2 16 10-2909 Vlll GOLD-VALUING TABLE. FINE Per GOLD, Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Divts. 6frs. Carats Grs. M \ghtJis s. d. 13 8-625 16 1 2 16 8-9635 13 8-000 16 2 16 7-6363 13 7-375 15 3 7 2 16 6-3089 13 6-750 15 3 6 2 16 4-9815 13 6-125 15 3 5 2 16 3-6541 13 5-500 15 3 4 2 16 2-3267 13 4-875 15 3 3 2 16 0-9992 13 4-250 15 3 2 2 15 11-6718 13 3-625 15 3 1 2 15 10-3444 13 3-000 15 3 2 15 9-0170 13 2-373 15 2 7 2 15 7-6896 13 1-750 15 2 6 2 15 6-3622 13 1-125 15 2 5 2 15 5-0348 13 0-500 15 2 4 2 15 3-7073 12 23-875 15 2 3 2 15 2-3799 12 23-250 15 2 2 2 15 1-0525 12 22-625 15 2 1 2 14 11-7251 12 22-000 15 2 2 14 10-3976 12 21-375 15 1 7 2 14 9-0702 12 20-750 15 1 6 2 14 7-7428 12 20-125 15 1 5 2 14 6-4154 12 19-500 15 1 4 2 14 5-0880 12 18-875 15 1 3 2 14 3-7606 12 18-250 15 1 2 2 14 2-4332 12 17-625 15 1 1 2 14 1-1057 12 17-000 15 1 2 13 11-7783 12 16-375 15 if i 2 13 10-4509 12 15-750 15 6 2 13 9-1235 12 15-125 15 5 2 13 7-7961 12 14-500 15 4 2 13 6-4687 12 13-875 15 3 2 13 5-1413 12 13-250 15 2 2 13 3-8138 12 12-625 15 1 2 13 2-4864 12 12-000 15 2 13 1-1591 12 11-375 14 3 7 2 12 11-8316 12 10-750 14 3 6 2 12 10-5042 12 10-125 14 3 5 2 12 9-1768 12 9-500 14 3 4 2 12 7-8494 12 8-875 14 3 3 2 12 6-5220 12 8-250 14 3 2 2 12 5-1946 12 7-625 14 3 1 2 12 3-8671 12 7-000 14 3 2 12 2-5397 12 6-375 14 2 7 2 12 1-2123 12 5-750 14 2 6 2 11 11-8849 GOLD-VALUING TABLE. FINE GOLD, Per Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Dwts. Grs. Carats Grs. Eighths s. d. 12 5-125 14 2 5 2 11 10-5575 12 4-500 14 2 4 2 11 9-2301 12 3-875 14 2 3 2 11 7-9027 12 3-250 14 2 2 2 11 6-5752 12 2-625 14 2 1 2 11 5-2478 12 2-000 14 2 2 11 3-9204 12 1-375 14 1 7 2 11 2-5930 12 0-750 14 1 6 2 11 1-2656 12 0-125 14 1 5 2 10 11-9382 11 23-500 14 1 4 2 10 10-6107 11 22-875 14 1 3 2 10 9-2833 11 22-250 14 1 2 '2 10 7-9559 11 21-625 14 1 1 2 10 6-6285 11 21-000 14 1 2 10 5-3011 11 20-375 14 7 2 10 3-9737 11 19-750 14 6 2 10 2-6463 11 19-125 14 5 2 10 1-3188 11 18-500 14 4 2 9 11-9914 11 17-875 14 3 2 9 10-6640 11 17-250 14 2 2 9 9-3366 11 16-625 14 1 2 9 8-0092 11 16-000 14 2 9 6-6818 11 15-375 13 3 7 2 9 5-3544 11 14-750 13 3 6 2 9 4-0269 11 14-150 13 3 5 2 9 2-6995 11 13-500 13 3 4 2 9 1-3721 11 12-875 13 3 3 2 9 0-0447 11 12-250 13 3 2 2 8 10-7173 11 11-625 13 3 1 2 8 9-3899 11 11-000 13 3 2 8 8-0625 11 10-375 13 2 7 2 8 6-7350 11 9-750 13 2 6 2 8 5-4076 11 9-125 13 2 5 2 8 4-0802 11 8-500 13 2 4 2 8 2-7528 11 7-875 13 2 3 2 8 1-4254 11 7-250 13 2 2 2 8 0-0980 11 6-625 13 2 1 2 7 10-7705 11 6-000 13 2 2 7 9-4431 11 5-375 13 1 7 2 7 8-1157 11 4-750 13 1 6 2 7 6-7883 11 4-125 13 1 5 2 7 5-4609 11 3-500 13 1 4 2 7 4-1335 11 2-875 13 1 3 2 7 2-8061 11 2-250 13 1 2 2 7 1-4786 GOLD-VALUING TABLE. FINE GOLD, Per Ounce CAKAT GOLD, Per Ounce STERLING VAI/~E, Per Our-* Oz. Drvts. Grs. Carats Grs. Eighths s. d. 11 1-625 13 1 1 2 7 0-1512 11 1-000 13 1 2 6 10-8238 11 0-375 13 7 2 6 9-4964 10 23-750 13 6 2 6 8-1698 10 23-125 13 5 2 6 6-8416 10 22-500 13 4 2 6 5-5142 10 21-875 13 3 2 6 4-1867 10 21-250 13 2 2 6 2-8593 10 20-625 13 1 2 6 1-5319 10 20-000 13 2 6 0-2045 10 19-375 12 3 7 2 5 10-8771 10 18-750 12 3 6 2 5 9-5497 10 18-125 12 3 5 2 5 8-2223 10 17-500 12 3 4 2 5 6-8948 10 16-875 12 3 3 2 5 5-5674 10 16-250 12 3 2 2 5 4-2400 10 15-625 12 3 1 2 5 2-9126 10 15-000 12 3 2 5 1-5852 10 14-375 12 2 7 2 5 0-2578 10 13-750 12 2 6 2 4 10-9303 10 13-125 12 2 5 2 4 9-6029 10 12-500 12 2 4 2 4 8-2755 10 11-875 12 2 3 2 4 6-9481 10 11-250 12 2 2 2 4 5-6207 10 10-625 12 2 1 2 4 4-2933 10 10-000 12 2 2 4 2-9659 10 9-375 12 1 7 2 4 1-6384 10 8-750 12 1 6 2 4 0-3110 10 8-125 12 1 5 2 3 10-8366 10 7-500 12 1 4 2 3 9-6562 10 6-875 12 1 3 2 3 8-3288 10 6-250 12 1 2 2 3 7-0014 10 5-625 12 1 1 2 3 5-6740 10 5-000 12 1 2 3 4-3465 10 4-375 12 7 2 3 3-0191 10 3-750 12 6 2 3 1-6917 10 3-125 12 5 2 3 0-3643 10 2-500 12 4 2 2 11-0369 10 1-875 12 3 2 2 9-7095 10 1-250 12 2 2 2 8-3821 10 0-625 12 1 2 2 7-0546 10 0-000 12 2 2 5-7272 9 23-375 11 3 7 2 2 4-3998 9 22-750 11 3 6 2 2 3-0724 GOLD-VALUING TABLE. FINE GOLD, Per Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Drvts Grs. Carats Grs. Eighths s. d. 9 22-125 11 3 5 2 2 1-7450 9 21-500 11 3 4 2 2 0-4176 9 21-875 11 3 3 2 1 11-0901 9 20-250 11 3 2 2 1 9-7627 9 19-625 11 3 1 2 1 8-4353 9 19-000 11 3 2 1 7-1079 9 18-375 11 2 7 2 1 5-7805 9 17-750 11 2 6 2 1 4-4531 9 17-125 11 2 5 2 1 3-1257 9 16-500 11 2 4 2 1 1-7982 9 15-875 11 2 3 2 1 0-4708 9 15-250 11 2 2 2 11-1434 9 14-625 11 2 1 2 9-8160 9 14-000 11 2 2 8-4886 9 13-375 11 1 7 2 7-1612 9 12-750 11 1 6 2 5-8338 9 12-125 11 1 5 2 4-5063 9 11-500 11 1 4 2 3-1789 9 10-875 11 1 3 2 1-8515 9 10-250 11 1 2 2 0-5241 9 9-625 11 1 1 1 19 11-1967 9 9-000 11 1 1 19 9-8693 9 8-375 11 . 7 1 19 8-5419 9- 7-750 11 6 1 19 7-2144 9 7-125 11 5 1 19 5-8870 9 6-500 11 4 1 19 4-5596 9 5-875 11 3 1 19 3-2322 9 5-250 11 2 1 19 1-9048 9 4-625 11 1 1 19 0-5774 9 4-000 11 1 18 11-2500 9 3-375 10 3 7 1 18 9-9225 9 2-750 10 3 6 1 18 8-5951 9 2-125 10 3 5 1 18 7-2677 9 1-500 10 3 4 1 18 5-9403 9 0-875 10 3 3 1 18 4-6129 9 0-250 10 3 2 1 18 3-2855 8 23-625 10 3 1 1 18 1-9580 8 23-000 10 3 1 18 0-6306 8 22-375 10 2 7 1 17 11-3032 8 21-750 10 2 6 1 17 9-9758 8 21-125 10 2 5 1 17 8-6484 8 20-500 10 2 4 1 17 7-3210 8 19-875 10 2 3 1 17 5-9936 8 19-250 10 2 2 1 17 4-6661 Xll GOLD-VALUING TABLE. FINE GOLD, Per Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Diets . Grs. Carats Grs. Eighths s. d. 8 18-625 10 2 1 1 17 3-3387 8 18-000 10 2 1 17 2-0113 8 17-375 10 1 7 1 17 0-6839 8 16-750 10 1 6 1 16 11-3565 8 16-125 10 1 5 1 16 10-0291 8 15-500 10 1 4 1 16 8-7017 8 14-875 10 1 3 1 16 7-3742 8 14-250 10 1 2 1 16 6-0468 8 13-625 10 1 1 1 16 4-7194 8 13-000 10 1 1 16 3-3920 8 12-375 10 7 1 16 2-0646 8 11-750 10 6 1 16 0-7372 8 11-125 10 5 1 15 11-4098 8 10-500 10 4 1 15 10-0823 8 9-875 10 3 1 15 8-7549 8 9-250 10 2 1 15 7-4275 8 8-625 10 1 1 15 6-1001 8 8-000 10 1 15 4-7728 8 7-375 9 3 7 1 15 3-4454 8 6-750 9 3 6 1 15 2-1179 8 6-125 9 3 5 1 15 0-7905 8 5-500 9 3 4 1 14 11-4631 8 4-875 9 3 3 1 14 10-1357 8 4-250 9 3 2 1 14 8-8083 8 3-625 9 3 1 1 14 7-4809 8 3-000 9 3 1 14 6-1535 8 2-375 9 2 7 1 14 4-8260 . 8 1-750 9 2 6 1 14 3-4986 8 1-125 9 2 5 1 14 2-1712 8 0-500 9 2 4 1 14 0-8438 7 23-875 9 2 3 1 13 11-5164 7 23-250 9 2 2 1 13 10-1890 7 22-625 9 2 1 1 13 8-8616 7 22-000 9 2 1 13 7-5341 7 21-375 9 1 7 1 13 6-2067 7 20-750 9 1 6 1 13 4-8793 7 20-125 9 1 5 1 13 3-5519 7 19-500 9 1 4 1 13 2-2245 7 19-875 9 1 3 1 13 0-8971 7 18-250 9 1 2 1 12 11-5697 7 17-625 9 1 1 1 12 10-2422 7 17-000 9 1 1 12 8-9168 7 16-375 9 7 1 12 7-5874 7 15-750 9 6 1 12 6-2600 GOLD-VALUING TABLE. Xlll FINE Per GrOLD, Ounce CARAT G-OLD, Per Ounce STERLING VALUE, Per Ounce Oz. Ihuts. Grs. Carats Grs. Eighths s. d. 7 15-125 9 5 1 12 4-9326 7 14-500 9 4 1 12 3-6052 7 13-875 9 3 1 12 2-2778 7 13-250 9 2 1 12 0-9503 7 12-625 9 1 1 11 11-6229 7 12-000 9 1 11 10-2954 7 11-375 8 3 7 1 11 8-9680 7 10-750 8 3 6 1 11 7-6406 7 10-125 8 3 5 1 11 6-3132 7 9-500 8 3 4 1 11 4-9857 7 8-875 8 3 3 1 11 3-6583 7 8-250 8 3 2 1 11 2-3309 7 7-625 8 3 1 1 11 1-0035 7 7-000 8 3 1 10 11-6761 7 6-375 8 2 7 1 10 10-3487 7 5-750 8 2 6 1 10 9-0213 7 5-125 8 2 5 1 10 7-6938 7 4-500 8 2 4 1 10 6-3664 7 ^3-875 8 2 3 1 10 5-0390 7 3-250 8 2 2 1 10 3-7116 7 2-625 8 2 1 1 10 2-3843 7 2-000 8 2 1 10 1-0568 7 1-375 8 1 7 1 9 11-7294 7 0-750 8 1 6 1 9 10-4019 7 0-125 8 1 5 1 9 9-0745 6 23-500 8 1 4 1 9 7-7471 6 22-875 8 1 3 1 9 6-4197 6 22-250 8 1 2 1 9 5-0923 6 21-625 8 1 1 1 9 3-7649 6 21-000 8 1 1 9 2-4375 6 20-375 8 7 1 9 1-1100 6 19-750 8 6 1 8 11-7826 6 19-125 8 5 1 8 10-4552 6 18-500 8 4 1 8 9-1278 6 17-875 8 3 1 8 7-8004 6 17-250 8 2 1 8 6-4730 6 16-625 8 1 1 8 5-1455 6 16-000 8 1 8 3-8181 6 15-375 7 3 7 1 8 2-4907 6 14-750 7 3 6 1 8 1-1633 6 14-125 7 3 5 1 7 11-8359 o 6 13-500 7 3 4 1 7 10-5085 6 12-875 7 3 3 1 7 9-1811 6 12-250 7 3 2 1 7 7-8536 XIV GOLD-VALUING TABLE. FINE Per GOLD, Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Dwts. GTS. Carats Grs. Eighths s. A 6 11-625 7 3 1 I 7 6-5262 6 11-000 7 3 1 7 5-1988 6 10-375 7 2 7 1 7 3-8714 6 9-750 7 2 6 1 7 2-5440 6 9-125 7 2 5 1 7 1-2166 6 8-500 7 2 4 1 6 11-8892 6 7-875 7 2 3 1 6 10-5617 6 7-250 7 2 2 1 6 9-2343 6 6-625 7 2 1 1 6 7-8069 6 6-000 7 2 1 6 6-5795 6 5-375 7 1 7 1 6 5-2521 6 4-750 7 1 6 1 6 3-9247 6 4-125 7 1 5 1 6 2-5973 6 3-500 7 1 4 1 6 1-2698 6 2-875 7 1 3 1 5 11-9424 6 2-250 7 1 2 1 5 10-6150 6 1-625 7 1 1 1 5 9-2876 6 1-000 7 1 1 5 7-9602 6 0-375 7 7 1 5 6-6328 5 23-750 7 6 1 5 5-3054 5 23-125 7 5 1 5 3-9779 5 22-500 7 4 1 5 2-6505 5 21-875 7 3 1 5 1-3231 5 21-250 7 2 1 4 11-9957 5 20-625 7 1 1 4 10-6683 5 20-000 7 1 4 9-3409 5 19-375 6 3 7 1 4 8-0134 5 18-750 6 3 6 1 4 6-6860 5 18-125 6 3 5 1 4 5-3586 5 17-500 6 3 4 1 4 4-0312 5 16-875 6 3 3 1 4 2-7038 5 16-250 6 3 2 1 4 1-3764 5 15-625 6 3 1 1 4 0-0490 5 15-000 6 3 1 3 10-7216 5 14-375 6 2 7 1 3 9-3941 5 13-750 6 2 6 1 3 8-0667 5 13-125 6 2 5 1 3 6-7393 5 12-500 6 2 4 1 3 5-4119 5 11-875 6 2 3 i 3 4-0845 5 11-250 6 2 2 1 3 2-7571 5 10-625 6 2 1 > l 1 3 1-4297 5 10-000 6 2 1 3 0-1022 5 9-375 6 1 7 1 2 10-7748 i o 5 8-750 6 1 6 1 2 9-4474 GOLD-VALUING TABLE. XV FINE Per GrOLD, Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce to. Dmts. Grs. Carats Grs. Eighths s. d. 5 8-125 6 1 5 I 2 8-1200 5 7-500 6 1 4 1 2 6-7926 5 6-875 6 1 3 1 2 5-4652 5 6-250 6 1 2 1 2 4-1377 5 5-625 6 1 1 1 2 2-8103 5 5-000 6 1 1 2 1-4829 5 4-375 6 7 1 2 0-1555 5 3-750 6 6 1 1 10-8281 5 3-125 6 5 1 1 9-5007 5 2-500 6 4 1 1 8-1733 5 1-875 6 3 1 1 6-8458 5 1-250 6 2 1 1 5-5184 5 0-625 6 1 1 1 4-1910 5 o-ooo 6 1 1 2-8636 4 23-375 5 3 7 1 1 1-5362 4 22-750 5 3 6 1 1 0-2088 4 22-125 5 3 5 1 10-8813 4 21-500 5 3 4 1 9-5539 4 20-875 5 3 3 1 8-2265 4 20*250 5 3 2 1 6-8991 4 19-625 5 3 1 1 5-5717 4 19-000 5 3 1 4-2443 4 18-375 5 2 7 1 2-9169 4 17-750 5 2 6 1 1-5894 4 17-125 5 2 5 1 0-2620 4 16-500 5 2 4 19 10-9346 4 15-875 5 2 3 19 9-6072 4 15-250 5 2 2 19 8-2798 4 14-625 5 2 1 19 6-9524 4 14-000 5 2 19 5-6250 4 13-375 5 1 7 19 4-2975 4 12-750 5 1 6 19 2-9701 4 12-125 5 1 5 19 1-6427 4 11-500 5 1 4 19 0-3153 4 10-875 5 1 3 18 10-9879 4 10-250 5 1 2 18 9-6605 4 9-625 5 1 1 18 8-3331 4 9-000 5 1 18 7-0056 4 8-375 5 7 18 . 5-6782 4 7-750 5 6 18 4-3508 - 4 7-125 5 5 18 3-0234 4 6-500 5 4 18 1-6960 4 5-875 5 3 18 0-3686 4 5-250 5 2 17 11-0411 XVI GOLD-VALUING TABLE. FINE GrOLD, Per Ounce CARAT G-OLD, Per Ounce STERLING VALUE, Per Ounce Oz. Drvts Grs. Carats Grs. Eighths s. d. 4 4-625 5 1 17 9-7137 4 4-000 5 17 8-3863 4 3-375 4 3 7 17 7-0589 4 2-750 4 3 6 17 5-7315 4 2-125 4 3 5 17 4-4041 4 1-500 4 3 4 17 3-0767 4 0-875 4 3 3 17 1-7492 4 0-250 4 3 2 17 0-4218 3 23-625 4 3 1 16 11-0944 3 23-000 4 3 16 9-7670 3 22-375 4 2 7 16 8-4396 3 21-750 4 2 6 16 7-1122 3 21-125 4 2 5 16 5-7848 3 20-500 4 2 4 16 4-4573 3 19-875 4 2 3 16 3-1299 3 19-250 4 2 2 16 1-8025 3 18-625 4 2 1 16 0-4751 3 18-000 4 2 15 11-1477 3 17-375 4 1 7 15 9-8203 3 16-750 4 1 6 15 8-4929 3 16-125 4 1 5 15 7-1655' 3 15-500 4 1 4 15 5-8380 3 14-875 4 1 3 15 4-5106 3 14-250 4 1 2 15 3-1832 3 13-625 4 1 1 15 1-8558 3 13-000 4 1 15 0-5284 3 12-375 4 7 14 11-2009 3 11-750 4 6 14 9-8735 3 11-125 4 5 14 8-5461 3 10-500 4 4 14 7-2187 3 9-875 4 3 14 5-8913 3 9-250 4 2 14 4-5639 3 8-625 4 1 14 3-2365 3 8-000 4 14 1-9090 3 7-375 3 3 7 14 0-5816 3 6-750 3 3 6 13 11-2542 3 6-125 3 3 5 13 9-9268 3 5-500 3 3 4 13 8-5994 3 4-875 3 3 3 13 7-2720 3 ' 4-250 3 3 2 13 5-9446 3 3-625 3 3 1 13 4-6171 3 3-000 3 3 13 3-2897 3 2-375 3 2 7 13 1-9623 3 1-750 3 2 6 13 0-6349 GOLD -VALUING TABLE. XV11 FINE GOLD, Per Ounce CARAT GOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Dwts. r. ? . Carats r*. Eighths I s. d. 3 1-125 3 2 5 12 11-3075 3 0-500 3 2 4 12 9-9801 2 23-875 3 2 3 12 8-6527 2 23-250 3 2 2 12 7-3250 2 22-625 3 2 1 12 5-9978 2 22-000 3 2 12 4-6704 2 21-375 3 1 7 12 3-3430 2 20-750 3 1 6 12 2-0156 2 20-125 3 1 5 12 0-6882 2 19-500 3 1 4 11 11-3607 2 18-875 3 1 3 11 10-0333 2 18-250 3 1 2 11 8-7059 2 17-625 3 1 1 11 7-3785 2 17-000 3 1 11 6-0511 2 16-375 3 7 11 4-7237 2 15-750 3 6 11 3-3963 2 15-125 3 5 11 2-0688 2 14-500 3 4 11 0-7414 2 13-875 3 3 10 11-4140 2 13-250 3 2 10 10-0866 2 12-625 3 1 10 8-7592 2 12-000 3 10 7-4318 2 11-375 2 3 7 10 6-1044 2 10-750 2 3 6 10 4-7769 2 10-125 2 3 5 10 3-4495 2 9-500 2 3 4 10 2-1221 2 8-875 2 3 3 10 0-7947 2 8-250 2 3 2 9 11-4673 2 7-625 2 3 1 9 10-1399 2 7-000 2 3 9 8-8125 2 6-375 2 2 7 9 7-4850 2 5-750 2 2 6 9 6-1576 2 5-125 2 2 5 9 4-8302 2 4-500 2 2 4 9 3-5028 2 3-875 2 2 3 9 2-1754 2 3-250 2 2 2 9 0-8480 2 2-625 2 2 1 8 11-5205 2' 2-000 2 2 8 10-1931 2 1-375 2 1 7 8 8-8657 2 0-750 2 1 6 8 7-5383 2 0-125 2 1 5 8 6-2109 1 23-500 2 1 4 8 4-8835 1 22-875 2 1 3 8 3-5561 ] 22-250 2 1 2 8 2-2286 3 N XV111 GOLD -VALUING TABLE. FINE Per GrOLD, Ounce CAKAT (TOLD, Per Ounce STERLING VALUE, Per Ounce Oz. Dwts. Grs. Carats Grs. Eighths 5. d. 1 21-625 2 1 1 8 0-9012 1 21-000 2 1 7 11-5738 1 20-375 2 7 7 10-2464 1 19-750 2 6 7 8-9190 1 19-125 2 5 7 7-5916 1 18-500 2 4 7 6-2642 1 17-875 2 3 7 4-9367 1 17-250 2 2 7 3-6093 1 16-625 2 1 7 2-2819 1 16-000 2 7 0-9545 1 15-375 1 3 7 6 11-6271 1 14-750 1 3 6 6 10-2997 1 14-125 1 3 5 6 8-9723 1 13-500 1 3 4 6 7-6448 1 12-875 1 3 3 6 6-3174 1 12-250 1 3 2 6 4-9900 1 11-625 1 3 1 6 3-6626 1 11-000 1 3 6 2-3352 1 10-375 1 2 7 6 1-0078 1 9-75C 1 2 6 5 11-6803 1 9-125 1 2 5 5 10-3529 1 8-500 1 2 4 5 9-0255 1 7-875 1 2 3 5 7-6981 1 7-250 1 2 2 5 6-3707 1 6-625 1 2 1 5 5-0433 1 6-000 1 2 5 3-7159 ] 5-375 1 1 7 5 2-3884 1 4-750 1 1 6 5 1-0610 i 4-125 1 1 5 4 11-7336 1 3-500 1 1 4 4 10-4062 1 2-875 1 1 3 4 9-0788 1 2-250 1 1 2 4 7-7514 1 1-625 1 1 1 4 6-4240 1 1-000 1 1 4 5-0965 1 0-375 1 7 4 3-7691 23-750 1 6 4 2-4417 23-125 1 5 4 1-1143 22-500 1 4 3 11-7869 21-875 1 3 3 10-4595 21-250 1 2 3 9-1321 20-625 1 1 3 7-8046 20-000 1 3 6-4772 o 19-375 3 7 3 5-1498 18-750 3 6 3 3-8224 (iOLD-VALUING TABLE. XIX FINE Per GrOLD, Ounce CARAT GOLD, Per Ounce STERLING VALUE Per Ounce Os. Dwts. Grs. Carats Grs. Eighths s. d. C 18-125 3 5 3 2-4950 17-500 3 4 3 1-1676 16-875 3 3 2 11-8401 16-250 3 2 2 10-5127 15-625 3 1 2 9-1853 15-000 3 2 7-8579 14-375 2 7 2 6-5305 13-750 2 6 2 5-2031 13-125 2 5 2 3-8757 12-500 2 4 2 2-5482 11-875 2 3 2 1-2208 11-250 2 2 1 11-8934 10-625 2 1 1 10-5660 10-000 2 1 9-2386 9-375 1 7 1 7-9112 8-750 1 6 1 6-5838 8-125 1 5 1 5-2563 7-500 1 4 1 3-9289 6-875 1 3 1 2-6015 6-250 1 2 1 1-2741 5-625 1 L 11-9467 5-000 1 10-6193 4-375 7 9-2919 3-750 6 7-9644 3-125 5 6-6370 2-500 4 5-3096 1-875 3 3-9822 1-250 2 2-6548 0-625 1 1-3274 XX GOLD-VALUING TABLE. To convert MINT VALUE into BANK VALUE when the Standard is expressed in Carats, Grains, and Eighths. This can be readily accomplished for every report by the following Tables : TABLE A. CARATS VALUE IN PENCE CARATS VALUE IK PENCE 1 0681 13 8863 2 1363 14 9545 3 2045 15 1-0227 4 2727 16 1-0909 5 3409 17 1-1590 6 4090 18 1-2272 7 4772 19 1-2954 8 5454 20 1-3636 9 6136 21 1-4318 10 6818 22 1-5000 11 7500 23 1-5681 12 8181 24 1-6363 TABLE B. CAEAT GRAINS VALUE IN PENCE CARAT GRAINS VALUE IN PESCE 1 2 0170 0340 3 4 0511 0681 TABLE C. EIGHTH CARAT GRAINS VALUE IN PENCE EIGHTH CARAT GRAINS VALUE IN PENCE 1 0021 5 0106 2 0042 6 0127 3 0063 7 0149 4 0085 8 0170 GOLD-VALUING TABLE. XXI Table A gives the difference in price between Mint and Bank value for each carat up to fine gold ; Table B the same for carat grains ; and Table C the same for eighths of carat grains. Now as the Bank value of gold is 3 17s. 9d. per oz. standard against Mint value of 3 17s. lO^d., it follows by calculation that fine gold would fetch, Bank price, only 4 4s. 9-8182d, instead of 4 4s. ll-4545