﻿Ionization by Alpha Kays. 479 



by the first theory is about 3 times that predicted by the 

 second ; for the small value it is about 20 times. It is very 

 difficult to believe that, even if different values of v and V 

 were applied to the two theories, measurements which agree 

 so well with one theory could be made also to agree throughout 

 the whole range of y with another theory, the predictions of 

 which for the same values of v and V are so completely 

 different. I think we may conclude with some certainty 

 that the second theory is not correct, and that the formation 

 of complex ions in dry air is not a common event in the 

 life-history of an electron. It is intended in the future to 

 make similar measurements in other gases and vapours to 

 ascertain whether in them the formation of complex ions 

 occurs with greater frequency. 



17. (4) The values of v and V may be compared with 



those obtained by Townsend. N — - , and in these experi- 

 ments I, though not known very accurately, was within 

 10 per cent, of 0'15. Hence these measurements give 



N =12'l, V' = 20'75, 



Townsend's work gives 



No^ 17-93, V'^19-00, 



The difference between the values of N is certainly greater 

 than can be attributed to experimental uncertainties ; it is 

 probably due to the deficiencies of the theory. For Townsend 

 used values of I so great that the chance of an electron 

 liberated anywhere between the electrodes reaching the 

 boundary of the field without colliding at all was very small, 

 whereas in my work under the most favourable conditions 

 this chance was as great as 0*16. Further experiments are 

 needed with different values of I to ascertain whether a dis- 

 crepancy of this nature is to be anticipated from the inadequacv 

 of the mathematical theory. 



18. (5) We will now consider what light the agreement 

 between theory and experiment which has been found throws 

 upon our chief question, that of the speed of the delta rays. 

 The theory assumes that the delta rays from both metal and 

 gas have no initial velocity. The assumption that they have 

 such a velocity would alter the theory in two ways. 'First, 

 the potential difference through which an electron must fall 

 before it acquires the power of ionization would depend on 

 the direction, relatively to that of the field, in which it was 

 ejected. If the initial velocity were small compared with 



