146 PROCEEDINGS OF THE AMERICAN ACADEMY. 



The conditions necessary for the action of the moulds, especially 

 the amount of arsenic and the quality of the culture ground, explain 

 why the action has not been earlier discovered. Though it is not 

 certain but that many moulds decompose arsenical matter, yet the 

 intense action is thus far confined to four, Penicillium brevicaule, 

 Macor mucedo, Aspergillum glaucum, and Aspergillum virens. Fur- 

 ther research may succeed in finding others. It is worthy of note 

 that the most active mould is that which was discovered joii decaying 

 paper, and it must not be forgotten that a small amount of arsenic in 

 a wall paper may be quite as good a source of the volatile compound 

 as a very large amount. 



I have not made any investigation into the nature of the compound, 

 as such work would be trespassing on Gosio's field. I think the 

 chances are that no arseniuretted hydrogen is formed, but that we 

 have to deal solely with an organic compound of arsenic. This may 

 perhaps form a "molecular" compound with the argentic nitrate, 

 which, when the latter is acted on by hydrochloric acid or an alkali is 

 set free. The odor of the solution was noticed by Hamberg, Gosio, 

 and myself. Hamberg does not say whether it appeared after adding 

 hydrochloric acid or in neutral solution. Gosio obtained it on adding 

 alkali, and did not apparently detect it in the acid solution, while I 

 discovered it in the desilverizing. Hamberg's yellow deposit may be 

 identical with that noticed by Gosio and me, although Gosio does not 

 asfree with us in the finding of arsenic. 



Hamberg and Gosio infer that the dilute argentic nitrate solutions 

 absorbed only a part of the compound, and my experiments confirm 

 this inference. Potassic permanganate probably oxidizes it com- 

 pletely. The non-absorption of the silver solution is an argument 

 against the compound being arseniuretted hydrogen, for I have satis- 

 fied myself by experiment that a 2% argentic nitrate solution absorbs 

 that gas completely when in small amounts, even when the gas is very 

 greatly diluted with air.* 



The use of argentic nitrate may partially account for some of the 

 negative results of former investigators. With such an absorbent 

 and the use of the simple Marsh test a comparatively large amount 

 of the compound might have escaped notice. 



* Although the argentic nitrate absorbs the arseniuretted hydrogen com- 

 pletely, yet when small quantities of arsenic are used, the amount recovered 

 from the solution is rarely over 50% of the amount taken. Some rather odd 

 results were obtained in investigating this question, the consideration of which 

 is reserved for another paper. 



