of Electrons, and the Spectrum of Canal Rays. 683 



(3) Steady translation alters the configuration of the 

 system by an amount proportional to the velocity (to a 

 first approximation) ; the change o£ position of any 

 electron may be represented as due to a disturbing force 

 acting on the original system, but it is of such a type 

 as to produce no periodic waves in the surrounding 

 medium (§ 19). 



(4) Variable motion of the system gives rise to periodic 

 waves only when there is resonance, and when the 

 variations in velocity and orientation of the system 

 are sufficiently rapid (§ 22). Waves are produced by 

 vibrations excited more especially during collisions be- 

 tween different systems. 



§ 25. These conclusions have been obtained on the three 

 assumptions (A), (B), (C). The two first relate to the 

 structure of the system of electrons ; they are made for the 

 sake of simplicity, but, as has been already pointed out, there is 

 much to be said in their favour on account of the permanence 

 of systems which are to serve as models of atoms or ions. 

 It is hardly possible that other assumptions made in their 

 place can yield qualitatively different results. The third 

 assumption is one which is usually made in the kinetic theory 

 of gases and, although possibly only approximate in the 

 present case, must at any rate lead to results of the right 

 order of magnitude. 



The conclusions, such as they are, are entirely at variance 

 with the hypothesis of Stark, that the energy emitted in the 

 displaced lines of the canal rays is transformed from kinetic 

 energy of translation by the direct action of radiation pressure. 

 Obviously the only way in which radiation pressure can 

 reasonably be said to act directly, is by resisting the motion 

 of the canal-ray particles and so consuming work which is 

 emitted as radiant energy. The argument against this 

 hypothesis may briefly be stated thus : — 



Consider a group of similar systems of electrons in motion, 

 which represent a bundle of canal rays, or a quantity of 

 radiating gas. Let E, be the average radiation per second 

 per system, of any type. It is ultimately derived from one 

 or more of three sources : — (1) by direct transformation of 

 kinetic energy of translation by radiation pressure, to the 

 average amount W ; (2) by indirect transformation from the 

 same source, and also from other types of vibrational energy 

 by collisions, to the total average amount K; (3) by absorp- 

 tion, due to resonance, from electromagnetic energy of waves 



