[ 574 ]

When one parcel os platina was cupelled with three orfour fresh quantities of bismuth, the first cupels were al-ways tinged of a blackish rusty hue, the next paler, andthe third for the most part only of the orange yellow colourwhich bismuth itself communicates, and which is consi-derably deeper than the stain imparted by lead.

In many of the cupellations, the surface of the metalwas found covered with a leafy substance like deep co-loured litharge; and sometimes, under the cupelled plate,there was a large quantity of rough spongy greenish matter,adhering in many parts strongly to the platina, running intocavities in its bottom, and in some places lying as it werebetween plates or flakes of the metal. It appeared thatbismuth, in eupellation with platina, does not diffuse orspread itself so thin, or sink so deep into the cupel, as leaddoes; but loads the parts which it touches, in such a man-ner, as to be prevented from extending further, and to becollected there in its femivitrified state; sometimes lying inlarge quantity on the cupel, though a considerable part ofthe cupel at bottom was not tinged with it. This doesnot seem to happen when bismuth is worked- off alone,,and therefore probably proceeds from this metal being aless powerful menstruum than lead for the ferrugineous andother foreign matters blended with platina. Many of thecupellations however went on well, without any appear-ances of this kind, and yielded brittle plates, sometimesdull coloured and sometimes bright, according as less ormore of the bismuth was worked off, of uneven surfaces,with large protuberances irregularly and sometimes ele-gantly disposed. Some of the minutes of these experi-ments having been lost, I cannot recollect whether it waswith large, or with small proportions of bismuth, that theprocess succeeded best.

From.