136 PROCEEDINGS OP THE AMERICAN ACADEMY. 



A similar experiment, in which 120 mgr. arsenic oxide were used, 

 was completed quantitatively, and by the Berzelius-Marsh method a 

 ring of arsenic was obtained weighing 2.8 mgr., corresponding to 

 4.3 mgr. As 2 5 , or 3.6% of the amount taken. This is unfortunately 

 the only quantitative work in the paper, and I cannot place much 

 reliance on it owing to the inaccuracy of the gravimetric Berzelius- 

 Marsh method with such small amounts. 



Pure cultures were now made of some of the moulds developed 

 in the above experiments, and there were isolated Penicillium glau- 

 cum, Aspergillum glaucum, and Mucor mucedo. Cultivations were also 

 made of Bacillus radiciformis, B. prodigiosus, B. subtilis, and Sarcina 

 lutea. All were then cultivated separately on sterilized arsenical 

 preparations. The odor was noticed only from A. glaucum and Mucor 

 mucedo. The latter, being a widespread and easily cultivated mould, 

 was selected for further experiments. 



Ten Erlenmeyer flasks were fitted with two-holed rubber corks and 

 two right-angled tubes, one passing to the bottom of the flask, the 

 other to just below the cork, each outer end being plugged with 

 cotton wool. Potato pulp containing arsenious oxide and a little tar- 

 taric acid was placed in the flasks, which were sterilized, inoculated 

 with a pure culture of mucedo in agar, and connected in series. In 

 the rear was placed a wash bottle of water, in front a 5% solution 

 of argentic nitrate, and a Bunsen pump drew air through the system. 

 The temperature was that of ordinary summer heat. After. 28 days 

 the presence of arsenic was established in the filtered solution. 



The solution before filtration contained a considerable quantity of 

 a yellow crystalline substance which quickly darkened. This Gosio 

 evidently hoped to connect with the compound 3 AgN0 3 . AsAg 3 of 

 Poleck and Thivmmel,* overlooking the fact that the latter is formed 

 only in very concentrated solution. On testing the substance no arse- 

 nic was found. 



The activity of Mucor mucedo was further shown by varying the 

 methods of culture, the nutritive soil, and the quality and quantity of 

 the arsenic compounds. Scheele's and Schweinfurth green, in pro- 

 portions varying from 0.001 to 0.1% of the pulp, gave reactions, but 

 realgar and orpiment gave no odor. Yet, if the action on the sul- 

 phides was protracted through many months, there was a slight 

 development. After chemical analysis had confirmed the odor, the 

 latter alone was considered a sufficient indication of arsenic in most 



* Archiv d. Pharm., CCXXII. 8. 



