Memoir on various Comlinations of Gold. 413 



parations, I begaij by washing them, until the water by the test 

 of nitrate of silver ceased to indicate any trace of muriatic acid. 

 I then treated them with pure potash, and with the aid of heat 

 I remarked, that the yellow precipitate gradually assumed a 

 deeper tint, and became at last a brown black, while precipitates 

 of this colour underwent no change. By saturating the pota^.h 

 with nitric acid in excess, and pouring nitrate of silver into it, a 

 very sensible precipitate of muriate of silver was formed in the 

 alkalies with which the yellow precipitates had been treated ; 

 but that with which the l)lack precipitates had been treated gave 

 no indication of muriatic acid. 



From these experiments it is evident, that the precipitates by 

 potash owe their difference of colour, to the greater or less quan- 

 tity of acid retained in combination. The yellow precipitates 

 retaining a certain quantity of acid ought, therefore, to be con- 

 sidered as true salts with excess of acid, which excess probably 

 gives them the solubility and styptic taste observed by Vauquelin. 

 The black precipitates, which may be considered as a true oxide 

 of gold, appear to have neither of these j)roperties. 



Barytes added in excess to a solution of gold l)y oxy-muriatie; 

 acid, produces a pale-yellow ])recipitate, variable in intensity, 

 according to the quantity of alkali: it still shows the presence of 

 muriatic acid. When lieatcd the precipitate swells, becomes 

 gray; and lastly, brown black, like that procured by potash : it 

 is now free from acid. The liquor still contains gold in solu- 

 tion. 



Soda, strontian and lime produce analogous results. 



The precipitates, whether from solutions in nitro-muriatic or 

 in oxy-muriatic acid, arc always of the same nature when brought 

 to the black colour by alkalies. Therefore, in these various cir- 

 cumstances the brown black oxide obtained is identical. 



Of the Oxide of Gold, 



Having analysed this oxide by heat, and measured with care 

 the oxygen gas evolved, it resulted from one experiment that 

 lUO parts of gold take 9-82 of oxvgen: a second experiment 

 gave 10-21; and a tiiird lO-OO. 'The mean of the three is 

 10-01; so that 100 parts of the oxide are composed of gold DO'J}, 

 and oxygen 9-10. 



These results agree with Bergman's, who states that 100 parts 

 of gold take 10 of oxygen; but thev differ from others. Proust's 

 results Avere variable: in one operation 100 of gold took S-57 of 

 oxygen ; in a second, the same proportion of gold took 31 of oxy- 

 gen ; and Thomson announces, in his A])pendix, that 100 parts 

 of gold always combine with 8 of oxygen, or with some multiple 



of 



