Mr Adie, On Bismuth. 245 



After reduction the product was slightly soluble in nitric acid, 

 giving a yellow solution which did not contain iron, uranium, 

 gold, nor any recognised substance. This yellow substance was 

 soluble in fused potassium cyanide, from which it could only be 

 precipitated electrolytically. The mass of yellow substance was 

 very small and was reserved for further investigation. 



The residue was found, on testing, to be free from lead, tel- 

 lurium and antimony, and was submitted to the spectroscope. 

 Its spectrum showed the presence of silicon in large quantity, 

 with traces of silver, copper, manganese, bismuth, and lead. 



The proportion of silica in the residue was now determined 

 by covering it with ammonia in a platinum basin and saturating 

 it with hydrofluoric acid prepared by the action of sulphuric acid 

 on pure ammonium fluoride; the liquid was then evaporated and 

 ignited. 



0'0986 gm. lost 00743 gm. which indicated that about 75 per 

 cent, of the residue consisted of silica. 



11. This tenacious adherence of silica, which dissolves with 

 the oxide of bismuth, is precipitated with the basic salts and 

 sulphide of bismuth, and is reduced and dissolved by bismuth 

 in fused potassium cyanide, does not appear to have been noted, 

 at any rate in connection with the atomic weight determination. 

 The question now arises as to how far the presence of this silicon 

 will affect the atomic weight. If the metallic bismuth contains 

 only 0'07 per cent, the atomic weight will be lowered by 1 unit, 

 and it was found (§7) that when only one-third of the bismuth 

 was precipitated as basic nitrate, it left the smallest amount 

 of residue and the precipitate contained nearly 0'15 per cent, 

 of silica. This percentage was considerably increased by preci- 

 pitating more of the basic salt. 



It will thus be seen that purification by precipitation still 

 leaves enough silica to account for the discrepancy between 

 Classen and Schneider's numbers, and presumably the silica 

 would not be reduced, and deposited, on electrolysis of the nitrate. 



12. Some other evidence can be cited in favour of the view 

 here brought forward, that the silica becomes reduced in presence 

 of the fused bismuth and the resulting silicon or silicide of bis- 

 muth becomes oxidised, and follows the bismuth into solution. 



Bismuth oxide when reduced by potassium cyanide always 

 makes a bead and some black powder. This black powder 

 becomes relatively more plentiful as the proportion of silica in 

 the oxide and the residue left on chlorination increase, e.g. in R,,. 



Bismuth nitrate prepared by dissolving bismuth in nitric 

 acid is grey, and deposits a fine grey powder, which disappears 

 on evaporation and concentration. 



