Charges on Ions in Gases. 209 



<came to be determined. Also a similar effect in a much less 

 degree was found in the previous experiments on the direct 

 determination of the rates of diffusion at high pressures, 

 which was subsequently found by Zeleny to affect the 

 velocities. There is thus a considerable amount of evidence 

 to show that the experiments on which the determinations of 

 Nxe are based are reliable and consistent. 



Experiments on the same principle as the previous work 

 in which N x e was found in gases at high pressure by deter- 

 mining the rate of diffusion and the velocity under an electric 

 force have recently been made by J. Franck and W. Westphal*, 

 and they confirm to some extent the result obtained at low 

 pressures, inasmuch as they find that the mean value of 

 Nxe for positive ions produced by Rontgen rays is 

 1*4 x 10 10 . According to Professor Millikan a larger 

 number of positive ions ought to have been found than is 

 indicated by this number if my conclusions are correct ; and he 

 suggests that the real reason why the number 1*4 X 10 10 found 

 for N xe exceeds the theoretical number P23 x 10 10 arises 

 from an effect which makes the rate of diffusion too small. I 

 have already explained why this suggestion cannot explain 

 the high value 1*4 x 10 10 obtained by this method. With 

 regard to the number of ions with double charges that are to 

 be expected, I cannot see that these experiments are incon- 

 sistent with those at low pressures, in which larger values of 

 Nxe were obtained with positive ions. It must be remem- 

 bered that when a beam of Rontgen rays traverses a gas 

 there is a considerable secondary effect produced in the gas 

 itself. It would be extremely difficult to separate the ions 

 generated directly by the primary rays from those produced 

 in the primary beam by the secondary rays, which consist of 

 negative ions projected with a large velocity. The effect 

 of these would not be so great in a small volume at low 

 pressure, as they would reach the boundary before all their 

 energy was spent in generating ions in the gas. It is not 

 inconsistent with the determinations at low pressures that at 

 high pressures there is a greater proportion of positive ions 

 having single charges. Professor Millikan concludes from bis 

 experiments that the " greater part of the ionization of a gas bv 

 X-rays is due to the direct action by primary rays," which if it 

 were true would make the above explanation untenable. It is 

 extremely difficult to follow the reasoning which leads to this 

 conclusion, but it is interesting in this connexion to note the 

 results of experiments that have recently been made with gases 



* J. Franck & W. Westphal, Ver. D. Phijs. Ges. Marz 5, 1909. 

 Phil. Mag. S. 6. Vol. 22. No. 127. July 1911. P 



