N0 . II STRUCTURE OF THE ATOM PARSON 5 



electrons in the atom, especially in reference to the hypothesis of 

 atomic numbers, with which they conflict to a certain extent. Then 

 follows a detailed application of the theory to the problems of 

 valence (§§9, 10, 12, 13), with a discussion of the residual magnetic 

 and electric forces due to different groupings of magnetons (§11). 

 §§14, 15 deal with the volumes of atoms, and after this (§16) it is 

 convenient to recapitulate the assumptions of the theory, which is at 

 that stage fully developed. §§17, 18 deal with the moment of the 

 magneton and a few questions connected with it; and §§19-23 con- 

 tain a full treatment of magnetic phenomena. 



§2. Considerations of Magnetism 



The arguments for the substitution of the conception of the 

 magneton for that of the classical electron in orbital motion, in 

 explaining magnetic phenomena especially, are principally concerned 

 with the radiation difficulties involved in the latter conception, 

 although conclusive arguments of another kind (pp. 9, 10) are 

 also available. The radiation difficulties have of course been a 

 matter of common knowledge, but since on account of the apparent 

 impossibility of avoiding them they have largely been ignored, it is 

 worth while to make a critical study of them as they occur in appli- 

 cations of the electron theory to magnetism. 



Of all the theories so far suggested, the present magneton theory 

 is the only one that allows the existence of orbital motion and so of 

 steady magnetic forces in the atom without the accompaniment of 

 radiation processes. Disturbances or irregularities of any kind in 

 the rotation of the magneton's annular charge will give rise to 

 radiations certainly, but these will be non-essential to the chemical 

 and magnetic individuality of the atom, and will be set up always by 

 chance external stimuli, just as all the radiation processes in atoms 

 (not including the emission of a and /S " rays ") are known to be in 

 actual fact. 



The contrary is the case with the classical electron. Every system 

 of such electrons that has as yet been devised to explain magnetic 

 phenomena either permits of continuous radiation or precludes the 

 possibility of the atom giving radiations of at all the same kind as 

 are observed : this will be made clear in what follows. 



To begin with, it has long ago been pointed out by Sir J. J. 

 Thomson that it is out of the question to consider orbits containing 



