486 Prof. C. Bar us on the Ionization 



(Together these values follow also from the present series of 



experiments, but from different hypotheses). 

 Thus 



26K = 434 x 1 6-7 x 2*3 x 10~ 19 x 4 x 10 4 /M x lO" 10 x ■14='260, 



since U = /i' cancels out; 



L = '376 x 1-5 x 50 x -46= 130. 



The following table is thus computed from 



a7V = 26£(l-e- L / v ). 



v= 



1 



2 



3 



4 



5 litres /min. 



#/v= 



•26x1-00 



•26x1-00 



•26X-99 



•26 x -96 



•26X-93 





= •26 



= •26 



= •257 



= •250 



= •242 



Hence tvitJmi the limits of observation (V<4 litres/min.) 

 ^■/V is appreciably linear, compatibly with the evidence of the 

 four series contained in the chart (fig. 3). 



10. The slope, o?/V, of the computed values is, however, in 

 every case definitely above the slopes found by experiment, 

 which, as taken from the chart (fig. 3), are : — 



Series 1, ^/V='13, Series 3, ^/Y=:*10, 



Series 2, a?/V = '12, Series 4, x/Y = '16. 



Quantitatively, therefore, the assumption JJ = k makes the 

 computed data of ^/V from 1*6 to 2*6 times too large. To 

 improve this result one would thus have to write JJ<k. Now 

 since there is evidence * in my second paper that k decreases 

 as the degree of dilution increases ; and since k, when expe- 

 rimentally found, comes out smaller than the value assumed 

 for U, this inference (U<k) is not tenable, k being simply 

 the absorption- velocity for the average concentration in the 



* With me the belief is gradually gaining ground, and will be further 

 qualified in my next paper with spherical condensers, that the absorption- 

 velocity k is a constant irrespective of dilution, and that the evidence to 

 the contrary obtained in my second paper with very wide absorption- 

 tubes (2 inches in diameter) must be otherwise interpreted. These 

 results were at the outset not very satisfactory (colours dull, indicating 

 promiscous condensation) ; and one may reach other conclusions than the 

 immediate ones there stated. In speculating on the origin of the velocity k, 

 considered as a nuclear constant, one is confronted with this question : 

 If the nucleus is larger than the molecule, as I believe it is, it will with 

 increasing size be more liable to the bombardment of a number of 

 molecules at the same time, promiscously in all directions, whereas the 

 molecules themselves collide singly. Therefore, how large must a nucleus 

 be made in order that the molecular velocity of, say, 485 metre/second 

 may sink to the value of only about 1 cm./sec, both as the result of 

 increased inertia, and of the increasingly unfavourable conditions of 

 bombardment. 



