﻿6(32 Prof. H. A. Wilson on the 



in the total volume. This has led Prof. Richards * to suggest 

 that atoms are compressible and that when two combine the 

 attraction between them produces a diminution of volume. 

 Thus we should expect p to increase with m. Also as m 

 gets larger we should expect the changes in p due to the 

 addition of one more atom to become smaller, so that for 

 very large values of m, p should obtain a limiting value. 



As to n it is clear that an increase in the total elm rge on 

 the atom must tend to increase its volume for the electric 

 intensity at the surface tends to pull it outwards. We 

 should therefore expect v to diminish with a and to increase 

 up to a limiting value with m. 



The various series in the spectra of a particular element 

 can be represented by formula? giving v as a function of two 

 integers. Thus for hydrogen we have 



\/r m ) 



where X is a constant and n and m aie integers. 

 For other elements we have according to Ritz 



N N 



(/* + a + b[n 2 y (m + a! + ft'/m*)*" 



If we give n a particular value we get a series of lines by 

 changing m. 



Recent investigations seem to show that these formula? do 

 not give the observed values exactly, so that it seems that 

 they are after all only empirical and without very much 

 physical significance. They do show, however, that v is a 

 function of two integers in many cases. Also we observe 

 that increasing m in the above formula? makes v increase up 

 to a limit, while increasing n diminishes v. 



AVe know too little of the nature of the positive electricity 

 to be able to formulate a satisfactory theory of the variations 

 in its density. The following is merely intended as an 

 illustration to show that the theory proposed can lead to 

 formula? like Balmer's. 



Let us suppose that the positive sphere behaves as though 

 it had a surface-tension T, and that the relation between 

 p and the pressure p due to T is p=(otp)i where a. is a 

 constant. {Suppose also that T is a function of p given by 



V'-x/'p) 



Vp. 

 where a and b are constants. 



* Faraday Lecture, 1911. 



