246 Mr Adie, On Bismuth. 



13. The above work leads to the conclusions 



(i) Bismuth and bismuth oxide after passing through the' 

 usual methods of purification still contain silicon and silica 

 respectively. 



(ii) Precipitation from acid solution in the form of basic salts 

 or sulphide does not free bismuth from silicon, though the amount 

 is reduced. 



(iii) Distillation of the oxide in chlorine does remove the 

 silica. 



(iv) Since the presence of silicon increases the absorption of 

 oxygen, Classen's number for the atomic weight of bismuth is 

 probably more correct than Schneider's. This is further sup- 

 ported by Dumas' (1859) value of 210, which has hitherto been 

 rejected, based on the ratio of CI : Bi in bismuth chloride. 



I have much pleasure in acknowledging my indebtedness to 

 Mr K. C. Browning who kindly undertook the qualitative work 

 and most of the chlorination, to Mr J. W. Capstick for the 

 specific heat determinations, and to Mr H. Ramage for the spec- 

 trum analysis. 



Note. A preliminary determination of the atomic weight of 

 bismuth using the pure metal gave the value about 208*8 (0 = 16). 



Note on the preparation of chlorine. 



Chlorine is most conveniently prepared for a continuous supply 

 by the action in a Kipp of strong hydrochloric acid on broken 

 pieces of a mixture of 5 parts of roughly powdered potassium 

 permanganate with 3 parts of plaster of Paris. The mixture is 

 moistened, spread into a cake about -|in. thick on a piece of board 

 and scored with cuts about ^ in. apart. The permanganate in 

 becoming reduced to a manganic salt liberates twice as much 

 chlorine as an equal weight of bleaching powder, while the 

 chlorine is free from oxides. 



