468 Mr. G. W. Walker on the 



For silver, which is the best conductor we know, 



k is of order 10"^ .'. h!^ is of order 10-«. 



For the Na line, 



2) is of order 10^'^, 



Kq or Kj is of order 10 "-^ 



Hence p^^(? is of order 10 ~^-. 



Hence, if the particles conduct as well as silver, il is clear 

 that the coefficient of 3N . T preserves the value unity from 

 p = to far beyond the violet. 



I think that we may conclude that the term is capal^le of 

 accounting for refraction following Gladstone and Dale's 

 law ; but, on account of the extremely small rate of change 

 with frequency, it is incapable of practically accounting for 

 dispersion. 



(4) Proposed Theory. 



I propose now to modify the supposition that the motion of 

 the molecule in those coordinates which are affected by 

 electrical forces is independent of temperature. It will 

 conduce to clearness if we select the simplest molecule 

 capable of giving the effects we require, and afterwards 

 consider a more complicated molecule. 



I therefore select a molecule which consists of a particle 

 of mass mi with a positive charge e and a particle of mass 11x2 

 with a negative charge —e, and suppose that the force 

 between them is the ordinary electrostatic attraction. 



It is necessary to suppose that the two particles are closely 

 associated, because very large electrical forces are required 

 to produce ionization. 



Since the particles are supposed to be of finite size, there 

 will be a minimum distance within which the particles 

 cannot approach. Let us call this distance between the 

 centres r. We shall therefore suppose that the two particles 

 describe circles about their common centre of inertia, and 

 do not vibrate radially. There will be a pressure between 

 the particles equal to the difference between the electric 

 attraction and the centrifugal force. It is clear that when 

 such a molecule is under the influence of external electrical 

 forces, the effective control against those forces is the 

 rotational energy of the molecule, which plays the part of 

 potential energy. On averaging for all the molecules, 



