RICHARDS. RETENTION AND RELEASE OP GASES. 423 



oxide even at high temperatures must be an infinitesimal far less iu value 

 than the tension of the gas in the vacuum needed for the production 

 of X-rays. 



According to the phase rule and the law of mass action these re- 

 actions must take place in the gaseous piiase, — the only one present 

 of variable composition. As an example, the case of copper is given 

 below : — 



4 CuO U 2 CugO + O2 



Vapor. Vapor. Gas. 



4 CuO 2 Cu,,0 



Solid. Solid. 



The occluded oxygen is supposed to escape by the oscillation, back- 

 ward and forward, of this heterogeneous reaction. The ability of the 

 system to be first dissociated and then associated in another place where 

 more oxygen is present presupposes, however, great internal activity 

 in the substance of the hot solids ; and this internal activity is the 

 most interesting conclusion to be drawn from the phenomena under 

 discussion. 



In short, no matter what point of view one adopts, or what mechanical 

 picture one forms of the reaction, the necessity of ascribing the rapid 

 escape of the oxygen to internal rearrangement seems to be inevitable. 



Professor Morse, in another letter received since the foresoinw was 

 written, kindly tells me that, in the course of some admirable work upon 

 the atomic weight of cadmium which he has just completed, he has had 

 occasion to measure the gases occluded by cadmic oxide made from the 

 nitrate. As one would expect, this oxide holds its occluded oxygen 

 more firmly than copper, but less firmly than zinc. In fact, the be- 

 havior of the substance, according to Professor Morse's description, 

 probably accords well with the hypothesis under consideration. His 

 interesting results will soon be in print, and will speak for themselves. 

 The fact that Rogers and I found no gas in our cadmic oxide in 1892 

 is easily accounted for by the circumstance that we used a bright red 

 heat in the ignition, the oxide parting with both gases by the disinte- 

 gration of its original structure, as the oxide of copper did with Scott. 

 Since cadmium is volatile at high temperatures, however, it was natural 

 that its dissociated oxide should have a wider opportunity of rearrange- 

 ment than in the case of copper ; and hence the crystalline structure 

 of the ignited oxide observed in 1892. Nickel, another metal whose 



