142 PROCEEDINGS OF THE AMERICAN ACADEMY. 



b. Three weeks elapsed before the experiment could be con- 

 tinued, during which time the flasks remained sealed by the cotton 

 wool. In two of them the growth did not seem to have increased 

 much, but in the third the ground was completely filled with mould, 

 which did not seem, however, to be entirely homogeneous. I am not 

 sure that the stopper of this flask may not have been slightly loosened. 

 The odor of garlic was very strong. 



The flasks were connected with silver solutions as before, and air 

 led through each day for 17 days. In four days there appeared on 

 the lower end of the entrance tube of the first silver solution a dark 

 mirror, which increased slightly and was apparently the only deposit 

 formed. The temperature during the first week was about 20°, but 

 afterwards about 25°. At a very rough estimate 170 litres of air 

 were passed through the system. 



The first silver solution was poured out of the bottle, and the latter 

 merely rinsed with water. The solution was heated nearly to boiling, 

 and excess of hydrochloric acid was added. At this point I was 

 struck by the strong odor coming from the warm mixture, it being 

 noticeable at a distance of two feet from the beaker. It strongly 

 resembled the garlic odor of arsenic, and also recalled the odor from a 

 solution of iron in dilute acid. My assistant, in making the previous 

 precipitation with sodic chloride, noticed no odor. The odor grew 

 weaker as the mixture was kept warm, but the argentic chloride was 

 filtered before it had entirely disappeared. The filtrate, after addition 

 of a little more nitric acid, was evaporated with sulphuric acid to 

 fuming, diluted, and added to the Marsh flask which had been run- 

 ning for forty minutes without a trace of arsenic. In ten minutes a 

 mirror began to appear, was very heavy in thirty minutes, and at its 

 maximum in fifty. It was clear and well defined, but too heavy 

 for accurate estimation. I placed it at 0.07 mgr., which is a low 

 estimate. 



Besides the mirror of silver (?) on the end of the tube in the absorp- 

 tion bottle, there was evident on closer inspection a small amount of 

 a nearly colorless (perhaps slightly yellow) gelatinous-looking sub- 

 stance adhering to the bottom and walls of the bottle. This dissolved 

 easily in the few drops of strong nitric acid added to dislodge the 

 heavier black deposit, though the action of the acid upon it was 

 masked by the nitrous fumes from the solution of the mirror. The 

 solution was precipitated with hydrochloric acid, evaporated to fum- 

 ing with sulphuric acid, diluted, and added to a Marsh flask which had 

 been running for 30 minutes without evidence of arsenic. A mirror 



