186 Mr. H. Moore on the Ionization in Carbon- 



benzene, the one calculated from observations, the other 

 on the assumption that the emission of corpuscles is an 

 " atomic " phenomenon, is thus well within the limit of 

 experimental error considered possible, but the difference 

 between the values for alcohol requires some further ex- 

 planation. 



In all the other cases, the purest chemicals supplied by 

 Kahlbaum were used, but pure alcohol could not be obtained 

 on account of Excise regulations ; the alcohol used was 

 ordinary " absolute " alcohol, and probably contained 1-2 

 per cent, of water. The presence of a very small quantity 

 of water would reduce the saturation vapour-pressure to 

 below that of pure alcohol, and at 0° 0., a difference of 

 1 mm. in the pressure would cause an error approaching 

 10 per cent. This, together with the error rendered possible 

 by the very small ionization values already discussed, would 

 probably account for the large discrepancy between the two 

 alcohol values. 



The values for the corpuscular radiation from an atom of* 

 oxygen and an atom of carbon, obtained from these equations, 

 may be substituted in the molecular formula for other gases 

 and vapours not experimented with in the present series of 

 observations. The numbers thus obtained, when multiplied 

 by the " corpuscular factor " for the gas or vapour used, 

 should give the ionizations produced in these gases or 

 vapours at 76'0 cm. by a stream of homogeneous (copper) 

 X-radiations in terms of that produced by the same beam of 

 radiation through an equal length of air at 76*0 cm. 



Thus, in the case of oxygen, the amount of corpuscular 

 radiation liberated would be 1*24 times that liberated in air, 

 taking the •'atomic'" value = 0*62 as obtained from these 

 equations. The et corpuscular factor " for oxygen being 1*10 

 (Barkla & Philpot, Phil. Mag., June 1913), the ionization 

 in oxygen as compared with that in an equal length of air 

 should be 1*36 for copper X-radiation. The value actually 

 obtained by these observers is 1*38. 



In carbon di-oxide the corpuscular radiation would be 

 0*2 + 1'24 = 1*44. The corpuscular factor obtained by 

 Kleeman is 1*08, giving an ionization of 1*56 compared with 

 air, the experimental values of Orowther & Owen being 

 1-57 (Crowther, Cambridge Phil. Soc, Feb. 1909 ; Owen, 

 Roy. Soc. Proc, May 1912). 



