and must be repeated; while, if too high, the volatilisation loss is much increased. The best rule is probably that universally accepted among American assayers-viz., to leave a slight but perceptible ring of feathery litharge crystals round the edge of the cupel, marking the area covered by the original lead button. Beads of silver cupelled at this heat always retain slight traces of lead, but the plus error thus introduced is much less than the minus loss by volatilisation (not to speak of the much greater loss by cupel absorption), so that a rough approximation to accuracy is obtained.

Just before the end of the operation the button exhibits a bright play of colours, caused by the formation of very thin films of litharge, and at this point the cupel should be pushed back into the hottest part of the muffle and left there for a minute or two to expel the last traces of lead. It is then withdrawn gradually towards the front of the muffle to prevent loss by "spitting" or "sprouting" at the moment of solidification, and, if at all large, is best covered with another hot cupel upside down, so as to cool it as slowly as possible.

The bead is removed from the cupel when cold by sharp-nosed pliers, and squeezed and brushed to remove adhering bone ash, if large enough; if too small, it is carefully picked up by fine forceps, placed on a blowpipe anvil and flattened by a blow from the light hammer, after which it is weighed on the assay balance to within mg. Very minute beads are weighed on the gold assay balance, which is sensitive to within 5 mg., corresponding 10 of an oz. gold on A.T. taken.

to

1

Parting. The bead should be flattened before parting, and, if it does not already contain at least two and a-half times as much silver as gold, a small fragment of pure silver foil is added and fused together with the bead on a cupel by means of the blowpipe flame.

Parting may be done either in a small 3-inch test tube, or in a porcelain crucible. The former method is preferable when the bead is large, owing to the danger of loss by projection when boiling a quantity of acid in a crucible; but small beads are best parted in a crucible, because of the danger that particles of gold may stick to the sides of the test tube and escape observation, and because only one acid (of 1-20 sp. gr.) is required for small beads. Large beads are parted by boiling, first in acid of 1.16 specific gravity till apparently no more action takes place; the acid is then carefully decanted off and replaced by acid of 1.27 specific gravity, with which boiling is continued for five minutes longer. The acid is then poured off and the gold washed twice with hot water, being allowed to settle very thoroughly* each time before decantation, after which the tube filled with water is inStress must be laid upon this; v. Ledoux, Trans. A.I.M.E., vol. xxiv.,

p. 873.

verted over a porcelain crucible, which is then also filled with water, and, after tapping the tube a few times, it is allowed to stand for a minute or two. When no more particles can be discerned settling through the water, the tube is removed and the water poured off, the last drops being drawn up by a capillary glass tube, and the crucible dried on a hot plate. When quite dry the gold is ignited over a spirit lamp, or in the muffle, and weighed on the "gold balance" to 01 mg., its weight being subtracted from that of the combined gold and silver bead obtained by cupellation to get at the silver.

Assay of Matte and other Coppery Substances for Silver.-At Western (American) works the sample is scorified down with large excess of lead in 10 lots of A.T. each, collecting the buttons in pairs at the end, and assaying separately the cupels and slag of each pair. This method gives the best results for gold, but higher results for silver are usually given by the following process, which has been in use for many years at copper works where silver assays have to be made on rich copper ores and furnace products:

Take from 1 to 2 A.T. of pulverised matte or copper borings, dissolve in strong nitric acid free from chlorine, evaporate in a porcelain dish over the sand bath to expel all excess of HNO3, then dilute to about 1 litre, add 20 c.c. of a standard solution of lead acetate, then 1 gramme NaCl dissolved in water; stir or shake well, then add 2 c.c. of strong H,SO, or, better, 5 grammes of Na,SO, dissolved in water; mix well, and allow to stand twelve hours. Decant off the clear liquid, filter off the precipitated PbSO, which will have carried down with it all the AgCl and free gold in suspension, and wash it well with hot water, picking out the globules of sulphur with a forceps. Burn these in a porcelain crucible, add the residue to the precipitate, which dry and mix in a small glass or porcelain mortar with thrice its own weight of lead flux, adding also the ashes of the filter carefully burnt. Charge the mixture into a small crucible with a borax cover and an iron nail to keep out copper from the button, and fuse it in the muffle like a lead assay. If the button is over 10 grammes in weight, or if hard from the presence of copper, it is scorified down (lead being added in the latter case), and the resulting button is cupelled and parted in the usual way. The hard part of the cupel, together with the slag from the fusion of the precipitate and the scorifier slag, if any, are pulverised, mixed with 2 parts lead flux, 2 parts borax glass, and part litharge, and fused down in the same crucible. The button of lead is cupelled, and the silver obtained added to the original assay.

Whitehead recommends separating the gold before precipitating the silver, and using KBr instead of NaCl for this *Chem. News, 1892, vol. lxvi., p. 19.

purpose. His method requires two additions of lead acetate and sulphuric acid, two separate settlings of twelve hours each, and two fusions and cupellations, one for the gold, and one for the silver. The author has found it more convenient to get both Au and Ag together in a single button after only twelve hours settling, but the use of KBr (or K I) is a decided improvement, as it obviates the danger of losing gold through Cl set free, and gives higher results for Ag owing to the lesser solubility of AgBr (Ag I) than of AgCl in solutions of copper nitrate. It is best to perform the assay in duplicate from the beginning so as to get a check on both silver and gold.

Losses of Silver in the Commercial Assay.-Whenever accurate results are desired, it is necessary to take the whole of the slags produced in the assay (whether from crucible or scorifier, or both) together with the hard portion of the cupel, and, after powdering, make a separate assay of the mixture as of a basic lead ore, adding some fluorspar to assist in melting the refractory bone ash. The lead button is cupelled at a low temperature, and the silver bead obtained is added to the original assay. Where extreme accuracy is desired, it is even necessary to repeat the correction. The assays made at metallurgical works for ascertaining the value of purchased ores (and generally for all other purposes) are almost invariably direct uncorrected assays; hence, the works have a considerable margin to work upon, and the working results obtained appear to be more satisfactory than they really are.

Some interesting recent researches, however, show that the commercial assay for silver is even more inaccurate than has been supposed. For details the student should consult the original papers, but it may be briefly stated that of the three sources of error, viz. :-(1) Loss from absorption by the cupel, (2) loss in the slags, and (3) loss by volatilisation-it is clearly shown by Dewey that the first is generally by far the most important, the others following in the above order. Some of his results are given in the following table compared with similar results by Stetefeldt. †

The loss of gold in slag and cupel is much less than that of silver, as shown by the experiments of Miller and Fulton, but in practice the differences in commercial gold assays are much larger in proportion, owing, no doubt, to errors of manipulation and to the smaller quantity of metal operated upon. Furman § found losses in cupelling amounting to from 1 to 7 per cent. of the silver, and 0 to 12 per cent. of the gold; he places the aver

* Dewey, Journ. Amer. Chem. Soc., 1894, vol. xvi., pp. 505 to 516; Stetefeldt, Trans. A. I. M. E., vol. xxiv., pp. 530 to 538; Furman, ibid., pp. 735 to 742; Oemichen, Zeitschr. f. angewandte Chemie, 1893, p. 723. + Loc. cit., p. 537. S. M. Q., vol. xvii., Jan., 1896, p. 169.

§ Trans. A.I.M.E., vol. xxiv., p. 735; also E. and M. J., Nov. 3, 1894.

age loss of silver at 2:58 per cent., which agrees fairly well with the results of Dewey.

*

On lead bullion Oemichen gives cupellation losses varying from 5 to 10 per cent. of the silver, and from 2 to 5 per cent. of the gold; Hofman † gives the loss on a pure 44 ozs. bullion as 3.5 per cent. of the silver. In the cupellation of argentiferous copper and lead, Rössler ‡ found losses as under :

Ledoux found losses of 175 per cent. in assaying copper

*Zeitschr. f. angewandte Chemie, 1893, p. 723.

+ Trans. A.I.M.E., vol. xxiv., p. 868.

Quoted by Hofman, Metallurgy of Lead, p. 249.

§ Trans. A.I.M.E., vol. xxiv., p. 875.

bars with 92 ozs. silver, and of 1·12 per cent. in the case of matte with 182 ozs. silver. In direct scorification of pig copper containing from 70 to 80 ozs. per ton with test lead the author has found the uncorrected results to range from 5 to 8 per cent. lower than the assay by the combined wet and dry method above described, the chief source of loss here being, no doubt, the enormous amount of scorifier slag produced.

CHAPTER II.

THE PROPERTIES OF LEAD AND ITS PRINCIPAL COMPOUNDS.

LEAD as a material for the manufacture of sheet and pipe has been known from very early times, and was worked very largely in Spain and Greece many centuries before the Christian era.

Physical Properties. *-Colour.-Lead is bluish-gray in colour, and a freshly cut surface shows a bright lustre, which rapidly tarnishes on exposure to the air.

Hardness.-Lead figures at only 570 on Bottone's scale, lying between tin and thallium and far below all the other common metals; it is so soft as to be readily scratched by the finger nail, while it readily marks paper with a grey streak. Only the pure metal, however, exhibits this quality to perfection, the commercial article contaminated with Sb, As, Cu, and Zn being distinctly harder, the degree of hardness increasing with the amount of impurity present.

Malleability and Ductility.-Lead is very malleable, being readily rolled into thin foil; according to Spring,† lead filings can be formed into a solid block under a pressure of 13 tons to the square inch, while two and a-half times that pressure liquefies the metal. It is, however, low in the scale of ductility, owing to its want of tenacity, which prevents it from being drawn into fine wire.

Specific Gravity.-Lead is the heaviest of all the common metals, its specific gravity varying between 11.254 and 11-395, according as it is cast or rolled. Knab gives 11.352 and Reich 11:37 as the specific gravity of pure cast lead at 0° C,, but the specific gravity of the commercial article is always lower on account of the impurities.

* v. Percy, Metallurgy of Lead, 1870, where the properties of this metal are given in great detail.

+ Quoted by Roberts-Austen, Introduction to the Study of Metallurgy, 4th edition, 1898, p. 79.