138 PROCEEDINGS OP THE AMERICAN ACADEMY. 



have a slight action in the course of months or years, being led to 

 this conclusion partly by Hamberg's work, though he has no data of 

 his own to prove the theory. 



The investigation of the chemical nature of the volatile compound 

 was begun on the assumption that arseniuretted hydrogen was formed, 

 as the silver solution was reduced and arsenic found in the solution. 

 On attempting, however, to precipitate argentic arsenite from solution 

 by ammonia, it was found that the excess of alkali liberated a vola- 

 tile substance with an intense garlic odor. The following examina- 

 tion was then made. The filtered silver solution was treated with 

 excess of potassic hydroxide in a flask, and the product of the action 

 led over lime and caustic potash to free it from carbon dioxide. The 

 gas then passed over hot cupric oxide and the combustion product, 

 led into baric hydroxide, caused a turbidity. The residue in the 

 cupric oxide tube was digested for two days in cold dilute potassic 

 hydroxide, filtered and washed. The solution was precipitated by 

 sulphuretted hydrogen and the filtrate acidified with hydrochloric acid. 

 The precipitate from the latter oxidized with nitric acid gave reac- 

 tions for arsenic in the Marsh apparatus. We shall have to take it 

 for granted in the above experiments that the carbon dioxide was 

 all held back by the absorbents, and that the cupric oxide was 

 non-arsenical. 



Another filtered silver solution was made alkaline with potassic 

 hydroxide, and air passed from it into platinic chloride for two days. 

 The air from a series of flasks was also passed into this absorbent for 

 ten days. In both cases no satisfactory results were obtained. 



In the solution after treatment with alkali, arsenic was looked for, 

 but only faint traces were found. This would be against the forma- 

 tion of arseniuretted hydrogen, yet Gosio inclines to the belief that 

 there may be traces of it formed. 



Considerable space is given to discussing the mechanism of the 

 reaction, yet, in view of the fact that the compound has not been 

 isolated, such discussion is purely speculative. Gosio refers to the 

 work of Pollaci,* Selmi,f and Fitz and Mayer t on the formation of 

 hydrogen by action of lower organisms, and quotes the statement of 

 Nencki § that bacteria may decompose water into H and OH, which 

 would cause a hydrogenization and a hydroxylization, a double action 



* Reference not given by Gosio. 



t Presumably the same reference as given by me above. 

 i Ber. d. deutsch. chem. Gesells., XI. 1880; XII. 474. 

 § Ibid., XII. 474. 



