372 Dr. G. J. Stoney on the 



supposing electrons (the charges of electricity which are 

 associated with chemical bonds, and which, so long as they 

 are undisguised, are acted on by the disturbance perpetually 

 going on in the surrounding aether) to be carried about by 

 the internal or B events of the molecule. To complete the 

 picture we may suppose most of these electrons to be so con- 

 nected with the internal economy of the molecule that they 

 can only perform evolutions that are resolvable into partials 

 that have definite periods. 



Before this change the internal events were Ba and Be 

 events, but, owing to the introduction of the slight roughness, 

 the Be events have become B6 events ; i. e. they have acquired 

 the power of very slowly, and through a great number 

 of collisions, affecting u, v, and w, which are the A events. 

 But even if the roughness be so slight that they require a 

 million of collisions to impart a sensible proportion of their 

 energy to the A events, they will accomplish a protracted 

 task of this kind in about the seventh part of the thousandth 

 of one second, if, as we have supposed, the rapidity of events 

 in our model is as great as it is in gases at atmospheric 

 pressures and temperatures. Now, observations with the 

 phosphoroscope indicate that the persistence of B6 events, 

 when once excited, is in all the observed cases much greater ; 

 so that the outer surfaces of our model molecules should be 

 very nearly, though not quite, smooth. Other experiments 

 which are in progress seem to show that, in some instances 

 at least, the energy radiated away from phosphorescent events 

 is much less than that which they lose by conduction, i.e. by 

 imparting energy to A or Ba events. Now an excessively 

 small defect in smoothness, such as we have supposed, would 

 not prevent the Boltzmann-Maxwell distribution of energy 

 from being nearly the actual distribution among A and Ba 

 events, wherever the other conditions of the theorem are, or 

 are nearly, fulfilled ; while side by side with them any amount 

 of energy may be maintained in B& events by the electro- 

 magnetic waves which unremittingly traverse the surrounding 

 aether, or may be set up by chemical reaction*. 



* It is easy to contrive other useful models, find to treat them like 

 those in the text. Of this kind are models with spherical, spheroidal, or 

 ellipsoidal outer surfaces, but with centres of inertia, either displaced 

 from the centre of figure, or moving- about it under definable conditions ; 

 or with more than one body inside the outer surface, and either concentric 

 or not concentric with it, and with viscous or other connexions which 

 would provide for various kinds of interaction ; or with the rigid outer 

 surface discarded, and central forces put in its place — a substitution which 

 either need not, or need not more than a little, alter the condition that 

 the average energy shall be a sum of squares ; and many others. Each 





