N0 . u STRUCTURE OF THE ATOM PARSON 21 



complex assumptions than heretofore : it merely substitutes an elastic 

 coherence for a rigid coherence, and it has the advantage of enriching 

 the atom with additional degrees of freedom. 



In such a sphere, each magneton will, by electrostatic attractions, 

 condense positive electricity in and around itself, and thus its electro- 

 static action on other magnetons will be weakened : the first effect 

 of endowing the positive sphere with elasticity will therefore be a 

 general diminution in volume under the action of the electric and 

 magnetic forces. In order that magnetons may not entirely neutral- 

 ize themselves in this way, it must be further supposed that the 

 elastic tension that obtains in the isolated positive sphere becomes 

 zero when the charge density has increased to a certain value, and 

 then changes sign, becoming a compression and combining with the 

 electrostatic repulsion to oppose a further increase in charge density : 

 such change of sign is of course connoted in the ordinary use of the 

 term "elastic." 



It is possible to make a somewhat elaborate study of the conditions 

 in such an atom, but they are very complex and hard to discuss with 

 any definiteness. Apart from the diminution of volume under the 

 action of the electric and magnetic forces, the elastic sphere will 

 apparently still behave, under static conditions, in much the same way 

 as the rigid sphere ; i. e., there will probably be no spontaneous separa- 

 tion into groups. This statement is no more than a well-considered 

 guess, because the complicated nature of the dependence of the repul- 

 sive forces in the elastic sphere upon the nature of the elasticity makes 

 it very difficult to decide whether or not there can be conditions which 

 would give us an unstable equilibrium in the case of even distribu- 

 tion. A spontaneous separation requires, of course, that at the point 

 of even distribution the rate of change of the magnetic forces shall 

 be greater than the rate of change of the combined electric and 

 elastic forces as the magnetons move towards group formation. 



There is, however, one important respect in which the two cases 

 differ. Molecular collisions will cause much more irregular dis- 

 turbances in an elastic than in a rigid sphere. Such disturbances 

 will lead to the momentary formation of separate groups. Under 

 these circumstances, the groups that form most often and have the 

 longest average existence will be the smallest groups that can possess 

 a minimum of magnetic energy and also great symmetry and mobil- 

 ity — the last being especially important under dynamic conditions. 

 There is thus a strong probability of an average state of grouping 

 into eights in the atom (see §6). 



