6i PROCEEDINGS OF THE AMERICAN ACADEMY. 



a higli temperature in air, but that reoxidatiou of the mouoxide took 

 place while it was cooling. He showed this by igniting in air a crucible 

 containing the black oxide and cooling the crucible suddenly by plunging 

 it into cold water. Under these conditions a film of black oxide was 

 formed on the surface only, while the lower portions remained in the 

 form of brown monoxide. Evidently, to prevent reoxidation it was only 

 necessary to keep the oxide away from contact with oxygen. The method 

 adopted by Russell was to beat the oxide in a current of an inert gas until 

 constant weight was established, and then to reduce the compound iu 

 hydrogen, thus obtaining an indirect ratio between oxygen and cobalt. 



This method of procedure, which was adopted by all subsequent experi- 

 menters on the oxide, seemed open to two serious disadvantages. The 

 first, the difficulty of preparing nitrogen or carbon dioxide free from 

 traces of oxygen, is one which can be surmounted, although not easily. 

 The possibility that small quantities of liberated oxygen may become 

 mechanically entangled in the oxide, in such a way as not to be speedily 

 removed by the current of gas, is a danger even less easily avoided. 

 This oxygen would of course reunite with the monoxide upon cooling, 

 thus making its weight too great. 



As an alternative to heating in a current of inert gas, ignition in a 

 vacuum suggested itself as offering distinct advantages, since neither of 

 the dangers just mentioned would be incurred. For this purpose a 

 glazed porcelain tube, of twenty-eight millimeters internal diameter, 

 similar to that used for fusing silver in a vacuum, was employed.* The 

 tube was closed as usual with Hempel hollow brass stoppers and heated 

 in a perforated Fletcher furnace by means of a large blast lamp, while 

 an ordinary Sprengel air-pump served to exhaust the gases from the tube. 



Very pure cobalt, which had been prepared by various methods during 

 the previous work, was converted into the oxide as follows. From the 

 solution of the metal iu pure nitric acid the pink granular cobaltous 

 hydrate was precipitated by adding an excess of pure freshly redistilled 

 ammonia, and subsequently digesting the mixture upon the steam bath 

 in a platinum bowl. This precipitate was collected upon a Gooch cru- 

 cible in which a circular piece of hardened filter paper was used instead 

 of an asbestos mat. After drying and converting the mass into the black 

 oxide by ignition with an alcohol lamp this material was used directly 

 for analysis. The metal which served as the source of this oxide was 

 known to be free from alkalis and silica,t and as neither the nitric acid 



* Richards and Parker, These Proceedings, XXXII. G3 (1896).. 

 t These Proceedings, XXXIII. 120 (1897). 



