SANGER. — VOLATILE COMPOUNDS OP ARSENIC. 



127 



pound, and from 0.2 to 0.3 oi Q of the total amount in the flask. 

 Hamberg states that by addition of all the amounts separated from 

 the silver solutions, partly as sulphide, partly as metallic arsenic, he 

 finds that the greater part of the arsenical gas was not absorbed, and 

 he thinks that possibly the silver solution was only partially able to 

 decompose the gas. Whatever estimate he may have made from the 

 mirrors is not stated ; hence we cannot know what amount he thinks 

 was recovered. As my own experiments have shown the inadequacy 

 of 2% argentic nitrate for absorption of the volatile compound and 

 the chance for loss in preparing the solution in the above manner 

 for analysis, the cause of loss is quite clear. The quantitative esti- 



* According to the method described by me in these Proceedings, XXVI. 24. 



