1887.] Bismuth on the ductility of Silver. 123 



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 mer- 

 cury is present the silver chloride maintains its pure white colour un- 

 altered, while in the presence of bismuth the chloride immediately ac- 

 quires 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 daylight, it was evi- 

 dent 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 oxychloride 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 oxychloride. 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 par- 

 ticular case, could not possibly have exceeded 1 ^er cent., and was proba- 

 bly less than that proportion. 



2. Hammering a bar of silver bullion is not a good test for de- 

 tecting brittleness, as far as mint purposes are concerned. 



3. The toughening of silver bullion 990 fine, and containing only 

 a small amount of bismuth, by the aid of nitre in plumbago melting 

 pots, is not readily effected. 



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 emphasized by the fact that silver bullion contaminated with bis- 

 muth 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 : — 



I. Is our ordinary wet assay of silver susceptible of 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 bis- 



