RICHARDS. — RETENTION AND RELEASE OP GASES. 419 



ought to be combined with the metal is always in a free state, and that 

 this condition, by a recurring process of dissociation and association, per- 

 mits relieving the high pressure of the oxygen in the cell. It is possi- 

 ble that water occluded in the oxide may play the part of a go-between, 

 but there is no experimental evidence of this. It is no new idea to 

 explain the diffusion of hydrogen through palladium, or of carbon 

 through iron in the cementation process, and other similar phenomena, 

 by a similar mechanical conception. 



Quite in accord with this point of view are a number of facts. In the 

 first place cupric oxide, the most easily reducible of the oxides investi- 

 gated, parts with its occluded oxygen with the greatest ease, at the same 

 time obstinately retaining the nitrogen. The progressive increase in the 

 tendency to retain oxygen in the cases of zinc and magnesium corre- 

 sponds to the increase in the difficulty of reducing these oxides. In 

 other words, the least stable oxide is the one which has the feeblest 

 hold upon the occluded oxygen. Cupric oxide begins to lose this 

 impurity with great rapidity at 520°, zincic oxide loses it less rapidly 

 even at 750°, and magnesic oxide is still more obstinate. It is probable, 

 moreover, that in the cases of zinc and magnesium there is a slight 

 amount of transpiration, for a small part of the nitrogen was found to 

 leave at the higher temperatures. This transpiration ought to assist the 

 oxygen also to escape, although in somewhat less measure than the nitro- 

 gen ; but in spite of this possibility of more rapid escape, the oxygen 

 is still held more firmly than in the case of cupric oxide. In this latter 

 substance all the nitrogen is retained indefinitely at temperatures below 

 850°, forming the constant gas residue noted in the first experiments; 

 only when tlie definite structure of the oxide is broken up, and the disso- 

 ciation tension of the cupric oxide becomes so great that cuprous oxide 

 begins to be formed, does the nitrogen take its flight. On the other 

 hand a reducing atmosphere is needed to force zincic oxide to relinquish 

 all its gas, unless an excessively high temperature is employed. Forty- 

 three years ago Sainte-Claire Deville and Rivot showed that at a red 

 heat such an equilibrium exists between zincic oxide, zinc, hydrogen, 

 and water as to enable zinc oxide to be sublimed in a current of hydro- 

 gen, the dissociated oxide reforming when cooled.* This reaction un- 

 doubtedly enables the occluded gas to ^cape by destroying the structure 

 of the solid when gases from the flame are present. 



Again, the fact that many solids, the oxides in question among them, 



* Ann. Chem. Phys., (3.), XLIII. 7, 477. 



