Mr Adie, On Bismuth. 241 



Schneider's value. Classen attributed the difference to the pre- 

 sence of lead in Schneider's pure metal and his method of purifi- 

 cation was specially directed towards its removal. 



Schneider in 1894 repeated his former work, without purifica- 

 tion by electrolysis and confirmed the value he had previously 

 obtained. He pointed out that the presence of lead would raise 

 and not lower the atomic weight, and that it would require 025 

 per cent, of lead to raise the atomic weight by 0*1 7. 



3. The determination of the equivalent of bismuth thus 

 appeared to present a method of finding out at first whether 

 any separation into different constituents could be effected in 

 bismuth. 



In the analysis of the sulphates the ratio Bi : S0 3 had been 

 determined, and this was used to calculate an equivalent, with 

 the following results : — 



The values for bismuth here given were obtained by reducing 

 the sulphates by means of potassium cyanide in a porcelain 

 crucible, and were compared with those obtained by oxidising the 

 metal, similarly prepared, by means of nitric acid in a platinum 

 crucible. 



Metal from salt Atomic weight 



Bi,0 3 , 4S0 3 , 3H 2 208-2, 207-2 

 Bi 3 3 , 4S0 3 , 7H 2 (2nd crystallisation) 208-2 



„ ,, „ (3rd crystallisation) 208-5 



4. The differences shown by these numbers were not con- 

 sistent enough to be useful in deciding whether any separation 

 had' been effected, but it was noted that recrystallisation always 

 tended to increase the atomic weight, and that a platinum crucible 

 generally gave higher values than porcelain. In two cases the 

 values 205-5, 205-7 in porcelain corresponded with 208*2, 208 - 7 

 in platinum. 



5. It seemed advisable to localise the cause of this variation 

 before proceeding further. About 2 gm. of pure bismuth oxide 

 (B.P.) was fused alone in a porcelain crucible and it gained about 

 06 mg. After fusion, a residue of silica was left when the oxide 



