548 Dr. J. Franck and Dr. W. Westplial on the 



appear only ions with either one positive or one negative 

 charge. This result is contrary to measurements of Towns- 

 end *, who found doubly-charged positive ions in the 

 ionization by Rontgen rays, and contrary also to the con- 

 clusions which we f drew from the study of the velocity 

 and the diffusion of ions, which were also generated by 

 Rontgen rays. We have found that about 9 per cent, of the 

 positive ions had a smaller (not more than half as great) 

 coefficient of diffusion than the rest, and concluded from 

 this that there existed 9 per cent, doubly-charged positive 

 ions in this special case. 



Millikan and Fletcher suppose that our result is caused by 

 slow, so-called intermediate or Langevin ions, whose exist- 

 ence in this case, according to Pollock J, is not wholly 

 impossible. We have also thought of this explanation, but 

 refused it, as apparently only slow positive ions were 

 generated and there was no evidence why, if they were 

 Langevin ions, that negative Langevin ions should not also 

 exist, as in all other cases where these ions appear. 



Now there can be no doubt but that the experimental 

 device of Millikan and Fletcher appears superior to our 

 more indirect method, especially as it gives the possibility of 

 making the charge on individual ions subject to investigation. 

 Nevertheless we are not able to consent to the conclusion of 

 the authors as to the non-existence of the doubly-charged 

 ions in the ionization by Rontgen rays, and we are rather 

 inclined to believe that the conditions for their generation 

 are not yet sufficiently cleared up. We want in the 

 following to state some points which seem to favour our 

 view. 



It seems especially important to point to the fact that 

 in the oil-drop method the observed number of doubly- 

 charged ions does not directly give the amount per cent, 

 of the same. Even if we put aside the violent recombination 

 of this type of ions (that effect, in the case of Millikan and 

 Fletcher, probably being small), there is still this to be 

 considered — L e., that the coefficient of diffusion of the 

 doubly and more highly charged ions is far smaller than 

 that of the univalent ones. And the influence of the 

 diffusion on the probability for the oil-drop to catch an ion 

 follows, even from Millikan's own paper on the elementary 



* J. S. Townsend, Proc. Roy. Soc. A, vol. lxxx. p. 207 (1908). 

 t J. Franck and W. Westplial, Verh. d. D. Phys. Ges. vol. xi. p. 146 

 (1909). 



X J. A. Pollock, Science, vol. xxxix. p. 919 (1909). 



