258 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



ever, than that evolved when the same cupric oxide was dissolved in sul- 

 phuric acid. It was shown, however, that the gas which resulted upon 

 solution in hydrochloric acid and stannous chloride consisted wholly of 

 nitrogen, and that the volume of this nitrogen corresponded to the nitro- 

 gen in the gases set free by solution of the same oxide in sulphuric acid. 

 Although it was evident that the oxygen was absorbed by the stannous 

 chloride, yet, since in the case of all oxides heretofore examined the in- 

 cluded gases consisted chiefly of nitrogen, it is fair to conclude that this 

 was so in the case of ferric oxide. Only a trace even of nitrogen was 

 evolved from the ferric oxide, hence this preparation must have been 

 practically free from included gases. 



In order to avoid any error due to hygroscopicity, the oxide used for 

 each analysis was bottled in dry air in the bottling apparatus, which has 

 already been described elsewhere.* The platinum boat containing the 

 oxide was heated in the hard glass tube while the air was exhausted from 

 the apparatus by means of a Sprengel air pump. Dry air was then 

 admitted, and the boat was pushed into the weighing bottle which was 

 then stoppered and weighed with the usual precautions. Since in two 

 instances repetition of the bottling failed to change the weight of the 

 oxide, subsequently this precaution was omitted. 



The oxide was then reduced in the manner already described, constant 

 weight of the metal being attained much more easily than in the case of 

 ■oxide prepared from ferric hydrate. Five closely agreeing results were 

 obtained from analyses of material prepared as above. 



SERIES II. 



* These Proceedings, 32, 59 (1896). 



