122
tray and its contents are placed in the muffleand annealed at a dull red heat, eare being takenthat the temperature does not rise too high, as thecornets are liable to adhere to the platinum eups,■which it must be admitted is a slight objectionto the use of this apparatus; and the only otherfault -which can be found with it is the impossibilityof determining whether any, and if so whichassay pieces contain platinum or palladium, a factwhich was immediately decided in the old method ofparting in separate vessels by the colour impartedto the acid. But when a number of assays ismade on metal of fairly uniform composition, as incoinage operations, the apparatus is invaluable, andcompared with its advantages these faults areunimportant.
The weight of the cornet will indicate very nearlythe amount of gold in the alloy if the assay has beenproperly conducted, but not with absolute exactness.Even when every precaution has been used, the goldcontained in the cornet is never chemically pure;it still retains a very small quantity of lead and ofsilver, and frequently also traces of copper, in con-sequence of which the weight of the cornet may ex-ceed by Tj-jj-g-irth part the true result; and whenit is considered that in assaying alloys the quantitytaken is often not more than 7 or 8 grains, it will, be seen that even this small error might lead toserious disappointment in estimating the proportionof gold in large masses of bullion. To ascertain theamount of the error, and thus arrive at a perfectlycorrect estimate, it is usual to assay, simultane-ously with the alloys, a number (generally three)of what are termed proofs, consisting of weighedportions of perfectly pure gold, approximately equalin amount to that present in the assay pieces. Thecornets obtained from these proof-assays are weighed,and the excess or deficiency in weight over that ofthe gold which was introduced indicates the amountof the correction which it becomes necessary toapply to the other assays. This correction is liableto variation, according to the temperature of thefurnace and other causes; but it usually rangesfrom 0'2 to 05 in 1000. It may be observed thatit is not essential to employ absolutely pure gold,provided the difference from purity be known. Thusthe corrections to be applied to a gold assay willbe readily understood from the following formula,where (1000 representing the weight of alloy origin-ally taken)—
p is the weight of the piece of gold finally obtained.
x, the amount of gold in the alloy, expressed inthousandths.
a, the weight of gold (almost absolutely pure)taken as a check, which approximately equals x
b, loss or gain in weight experienced by a duringthe process of assay.
k, variation of “ check gold ” from absolute purity,expressed in thousandths.
The actual amount of fine gold in the check-piece—
= “ ( X — lS5o)>
and the corrected weight of the cornet will be —
ah .
“ = - p — Tbub— 6
b being added or subtracted according as it is aloss or gain. If a be assumed to be equal to x,this equation becomes—
p b
x — - 1 — —
k
1 + 1000
The actual amount of metal retained, called sur-charge, is greatly influenced by the amount of copperpresent in the alloy under examination, and variesconsiderably with the heat of the furnace and othercircumstances. In an alloy rich in gold, and whichtherefore required only a small quantity of lead, thesurcharge is slight; when, however, a large quantityof copper is present the surcharge becomes negative,in consequence of the absorption of gold by thecupel; and for alloys containing 08 or 70 per cent, ofgold the surcharge is zero. Rossler has shown thatthe gold lost in the cupellation process increaseswith the amount of lead used, and decreases as thesilver present is increased- and these conclusionsare confirmed by the results of experiments recentlymade in the Royal Mint. He also considers that ingeneral the gold lost is greater in amount than thesilver retained by the cornet.
Parting Assays. —It is frequently necessary to de-termine not only the amount of gold or silver con-tained in a given alloy, but also the amount of eachof these precious metals present in the same sample;the assays made with a view to obtain this informationare termed “ parting assays.”
If the quantity of silver in the alloy submitted tothe operation of parting be more than three timesthat of gold, the separation may still be effectedby nitric acid, but the gold will be left in a stateof powder. It may, however, be washed, ignited,and weighed in the manner described in connectionwith the quantitative estimation of gold by the wetmethod, and in the process of parting with sul-phuric acid on the large scale. Should the amountof silver be but small compared with the gold, andthe assayer wish to avoid the trouble of inquartation,that is, the adding of the requisite proportion of silver,the alloy may at once be subjected to the action ofaqua regia, when the gold will be dissolved, and thesilver will be thrown down as an insoluble chloride.The clear solution containing the gold is thendecanted, and when the excess of acid has beenexpelled by evaporation, the gold can be precipi-tated, as already described, by sulphate of iron.
The determination of the amount of silver con-tained in gold by the dry method of assay is a some-what complex operation. One or two gold assays inthe ordinary way are made, and at the same timetwo more assay pieces are passed through the fur-nace, but without the addition of silver. By thismeans all the oxidizable metals are removed in bothcases, and the difference in weight between one ofthe cornets resulting from the ordinary assay, and