EICHARDS. — RETENTION AND RELEASE OF GASES. 425 



large amount. Lord Rayleigh, in the course of his recent v;ikiable 

 work on the density of gases,* concludes that his carbonic oxide also 

 was free from hydrogen or hydrocarbons ; and indeed there is no reason 

 why this result should not be easily attained if one uses pure materials. 

 Lord Rayleigh explains the existence of hydrogen in ray carbonic oxide 

 by drawing attention to the fact that in my case the gas, dried only 

 by calcic chloride, was passed over red hot platinum sponge. 



This explanation is undoubtedly correct. Many experimenters have 

 shown that the reaction CO + IloO = H3 -|- COo is an easily rever- 

 sible one at high temperatures,! and in the presence of a large excess 

 of carbon monoxide, it is obvious that nearly all of the trace of aqueous 

 vapor would be reduced. P"'or my purpose at the time, the source 

 of the hydrogen was of no importance; it was only necessary to know 

 the exact amount of the impurity, and this was suitably determined. 

 It is satisfactory to know the cause of the discrepancy, however, as 

 well as to know that it is easier to prepare carbon monoxide almost 

 free from hydrogen than seriously contaminated with it. More often 

 than we suspect, perhaps, are our carelessly planned efforts at purifica- 

 tion apt thus to introduce more impurity than they eliminate. 



In a recent number of the Journal of the American Chemical Society, 

 (March, 1898,) F. R. M. Hitchcock has published a preliminary paper 

 upon the occlusion of various gases by the oxides of metals, having 

 apparently overlooked the Harvard investigations. It is satisfactory 

 to see the phenomenon widely recognized ; for an appreciation of this 

 serious cause of possible constant error will result in the publication of 

 fewer erroneous determinations of atomic weights than have appeared 

 in the past. One should also bear in mind the fact that nitric acid 

 is not by any means the only substance which is liable to leave apprecia- 

 ble contamination in material from which it has been expelled by heat. 

 The number of such occlusions is far greater than most analysts 

 imagine ; this fact is not realized because the minute residues are often 

 so hard to detect. 



The evidence of this paper proves that in reality no discrepancy 

 exists between the results of Scott, and Morse and Arbuckle, and 

 those obtained at Harvard. The apparent disagreements were caused 



* Proc. Roy. Soc, LXII. 204 (1898). 



t Bunsen, Liebig's Annalen, LXXXV. 137 (1853). Horstmann, Ibid., CXC. 

 228 (1878). Butsch, Ibid., CCX. 207. And especially Hoitsema, Zeitschr. Phys. 

 Chem., XXV. G88, as well as others. 



