360 RICHARDSON— DYNAMICAL EFFECTS OF [April 22, 



There is another point in this connection that is not without in- 

 terest. On any theory of absorption the natural periods of a sub- 

 stance are characterised by conferring on it either intense absorp- 

 tion or intense opacity. It is therefore evident that they can be 

 detected very readily by experiment. From an analysis of the nat- 

 ural periods of a large number of substances which has been carried 

 out by Drude it appears that there are two types of vibrations which 

 occur. In the one the electron forms the vibrating system and in the 

 other one of the constituent atoms or a group of atoms vibrate as a 

 whole. Owing chiefly to the enormous difiference between the mass 

 of an electron and that of an atom there is an enormous difference 

 between the frequency of the two types. The electronic type 

 always gives rise to frequencies in the ultra-violet part of the spec- 

 trum and the atomic type to natural frequencies in the infra-red. 

 It is therefore not an accidental circumstance that almost all chem- 

 ically pure substances which are not conductors of electricity are 

 transparent in the visible spectrum. 



The action of Roentgen rays on matter is a subject of great inter- 

 est. According to the ether pulse theory of these rays elaborated 

 by Sir J. J. Thomson, the relation between the Roentgen rays and 

 sodium light is similar to that between a series of sharp cracks and a 

 musical tone. And on the modern view of the nature of white light 

 the difference between white light and the Roentgen rays is one of 

 degree rather than kind. The cracks corresponding to Roentgen 

 rays are much sharper than those which correspond to white light. 

 According to the principles of harmonic analysis which we owe to 

 Fourier it should be possible to resolve either of these kinds of 

 radiation into simple harmonic elements. I have estimated that the 

 average frequency of these elements for the Roentgen rays would 

 be 10,000 times greater than that for those which form white light. 

 This estimate is based on the view that the kinetic energy of the 

 electrons emitted by bodies under the action of ultra-violet light and 

 Roentgen rays is a function of the frequency of the equivalent vibra- 

 tions. The experimental results indicate that the functionality is a 

 linear one and there is considerable theoretical support for this view 

 Some investigators have maintained that the square root of the 

 energy is proportional to the frequency; but even if this extreme 



