SANGER. — VOLATILE COMPOUNDS OP ARSENIC. 141 



As 2 5 . They were then heated for two hours in a steam sterilizer 

 and allowed to stand 24 hours. This was twice repeated. Each 

 flask was then inoculated with a sterilized platinum point from the 

 tube marked " patata bagnata in una soluzioue di As 2 5 ." The 

 three flasks were then connected together and in the rear was joined 

 a side-neck test-tube containing 2% argentic nitrate. In front was 

 placed an additional layer of cotton wool in a chloride of calcium tube, 

 then two silver solutions of the same strength as the first. The joints 

 of the apparatus were wired and were tight. A current of air was 

 drawn very slowly through the system day and night, the amount 

 being measured. In one day the growth at the points of inoculation 

 began, and in three days there was a patch half an inch in diameter 

 around each spot. The silver solution had darkened slightly. The 

 temperature during the trial was about 25°. 



At the end of twelve days, as it became necessary to change the 

 place of 'experiment, the flasks were disconnected and packed for trans- 

 portation. Up to this time no difference could be noted between this 

 experiment and the many preceding, except that the mould was in 

 smaller quantity and apparently more homogeneous. The first silver 

 solution was somewhat reduced, there being a slight black deposit, but 

 the second was not changed. But as soon as the flasks were discon- 

 nected (they were not uncorked) an alliaceous odor could plainly be 

 perceived at the tubes, a point that I had never observed in any pre- 

 vious experiment. The silver solutions were then tested, the method 

 being for certain special reasons slightly modified. The unfiltered 

 solution was precipitated by a very slight excess of sodic chlo- 

 ride, and the filtrate evaporated with sulphuric acid until it fumed 

 strongly. The diluted residue was boiled with a very little sodic 

 sulphite, and the excess of sulphur dioxide expelled. The cooled 

 solution was then introduced into the Marsh flask which had been 

 running one hour without sign of a mirror in the deposition tube. 

 Here may it be said that all reagents used had been subjected to the 

 most rigid test, the stream of hydrogen from the generator not giv- 

 ing any arsenic whatever in a seven-hour run. Twenty-five minutes 

 after introduction of the prepared solution a clearly defined mirror of 

 arsenic was deposited which was the first that I had obtained in my 

 experiments on this subject. The amount was small, about 0.01 mgr., 

 but, taken in connection with the odor, it was very satisfactory con- 

 sidering the small amount of mould and the duration of its action. 

 One hundred and five litres of air had been drawn through. The 

 •second silver solution treated in precisely the same way gave no 

 mirror. 



