242 Mr Adie, On Bismuth. 



was dissolved in nitric acid. There was also a gain of 0'8 mg. 

 when the same weight of the oxide was heated with silica and 

 sodium silicate, but there was a loss on heating the oxide with 

 sodium metaphosphate and with borax. 



This peculiarity of behaviour in presence of silica only, sug- 

 gested three explanations : (i) that the equivalent of bismuth was 

 lowered by the presence of a substance which is only slowly 

 oxidised when in the form of silicate (formed from the crucible), 

 (ii) that the bismuth in the form of silicate absorbed more oxygen 

 than corresponds with the oxide Bi 2 3 , (iii) that the gases from 

 the flame diffusing through the platinum reduce some of the 

 oxide. The metal so formed would alloy with the platinum and 

 so raise the equivalent. 



It was noted that the platinum became brittle and that a 

 smell suggesting phosphoretted or siliciuretted hydrogen was 

 developed on heating in platinum. 



To test these hypotheses other methods of purifying bismuth 

 were now examined. 



6. The pure (B.P.) oxide was dissolved in pure nitric acid, 

 and precipitated in four portions by water, followed by the 

 addition of caustic potash. In one case 110 gm. of the oxide 

 were dissolved in 300 c.c. of nitric acid (1*36) filtered through 

 glass wool and precipitated by the successive addition of (i) 5 litres 

 of distilled water, (ii) 20 gm. of caustic potash, (iii) 40 gm. of 

 caustic potash, (iv) 40 gm. of caustic potash. Four fractions were 

 obtained which were carefully vvashed and air-dried and were 

 found to weigh 38 gm., 18 gin., 34 gm. and 19 gm. respectively. 

 The first three of these fractions were white and crystalline, the 

 fourth was yellowish and gelatinous. 



At this stage a method of further purification was tried, based 

 on the volatility of the chloride of bismuth, viz. the distillation 

 of the oxide in a current of chlorine, by which lead chloride at 

 any rate would be separated and left behind. A fluffy, light, non- 

 volatile residue was left, but the greater part distilled over. 



7. In order to combine the two processes as advantageously as 

 possible, the four fractions precipitated as above from the nitrate 

 solution were separately distilled in chlorine with the following 

 results: 



Weight of 

 precipitate taken BiCl 3 distilled Eesidue 



(i) 38 gm. 37-6 gm. 0-05 gm. 



(ii) 18 gm. 17-1 gm. 0-03 gm. 



(iii) 34 gm. 29-3 gm. 1-26 gm. 



(iv) 19 gm. 9*0 gm. 040 gm. 



