POTASSIUM NITEOSOSULPHATE. 39 



however, was not only a little different from that adopted by Hantzsch, 

 but we had to estimate the nature of the change in another way. He 

 exploded the salt and noted its loss in weight, and from this calculated 

 how much of it had given off nitric oxide, and how much nitrous 

 oxide. We could not do this, since much of the residue was always 

 blown out of our tube. Instead, therefore, we determined by the 

 iodine method the amount of sulphite in a weighed quantity of the 

 residue. There was always a trace of nitrite to be found. 



We have now noticed all the points of supposed difference be- 

 tween Pelouze's nitrososulphate and what Hantzsch has designated 

 the Raschig salt, and have said enough, we believe, in the light of our 

 own experiments, to make it reasonably certain that the two salts are 

 one. We will now turn to Raschig's other salt. 



This salt, which Raschig and Hantzsch have erroneously taken 

 to be identical with Pelouze's potassium nitrososulphate, has two 

 specific characters : — it gives a barium salt insoluble in water, and it 

 yields dipotassium oximidosulphonate when its solution in weak 

 potash is made to crystallise, either by cooling or by evaporation. It 

 may be mentioned here, parenthetically, that we have found it im- 

 possible to get any oximidosulphonate from pure potassium nitrososul- 

 phate. To get this salt of his, Raschig had to evaporate over sul- 

 phuric acid until crystalline crusts formed, the solution produced by 

 passing nitric oxide through potassium sulphite. Other workers have 

 always got their product crystallising out during the absorption of the 

 gas. 



Hantzsch has tried in vain by various means, to prepare this 

 sab. As for ourselves, in whatever way w T e have worked, we have 

 always got Pelouze's nitrososulphate and no other along with it, 

 though we have used a freezing mixture of ice and salt, the cold 

 of melting ice, or the prevailing temperature, have had much potassium 



