818 
MR. J. LARMOR OR A DYNAMICAL THEORY OF 
the vortex theory, is therefore a very indefinite question. In the absence of any such 
clue, a guiding principle in this discussion has been to clearly separate off the material 
energy involving motions of matter and heat, from the electric energy involving 
radiation and chemical combination, which alone is in direct relation to the aether. 
The precise relation of tangible matter, with its inertia and its gravitation, to the 
aether is unknown, being a question of the structure of molecules; but that does not 
prevent us from precisely explaining or correlating the effects which the overflow of 
aethereal energy will produce on matter in bulk, where alone they are amenable to 
observation. 
Optical Dispersion; and Moving Media. 
122. The view of optical dispersion developed in the first part of this paper, on the 
basis of MacCullagh’s analysis, has its foundation in the discreteness of the medium, 
the dispersion being assigned to residual terms superposed on the average refraction. 
The cause of the refraction itself is found in the influence of the contained molecules, 
which are constituted in part at least of mobile electrons and so diminish the 
effective elasticity of the medium. Now if these molecules formed a web permeating 
the medium, with connexions of its own, this web would act as an additional support, 
and the optical elasticity would, if affected at all, be increased. But it is different if 
the molecules are so to say parasitic, that is if they are configurations of strain in the 
rether itself, and their energy is thus derived directly from the aether and not from an 
independent source. To more clearly define the effective elasticity in that case, let us 
suppose a uniform strain of the type in question to be imparted to the medium by the 
aid of constraints ; it follows from the linearity of the elastic relations that the stress 
involved in this superposed strain will be that corresponding to the elastic coefficient 
of the free aether, for there is by hypothesis no web involved with extraneous elasticity. 
Now suppose the constraints required to maintain this pure strain to be loosened ; the 
molecules will readjust themselves into a new equilibrium position which involves less 
energy, and this diminution of the total energy of the strain implies a diminution of 
the corresponding effective elastic coefficient. This analysis has to do with the 
statical elasticity; in electrical terms it corresponds to the explanation of Faeaday 
and Mossotti as to how it is that the ratio of electric force to electric induction is 
diminished by the presence of polarized molecules. If, however, in a problem of 
vibration, the disj^lacement of the medium involved in the molecules thus settling 
down into a new conformation of equilibrium, after the constraints are removed, is 
comparable with that involved in the original strain, the kinetic energy of the 
medium will be affected by the molecules as well as the strain energy, and the cir¬ 
cumstances of propagation will depend on the period of the waves. As the present 
theory involves altered effective elasticity but unaltered effective inertia, this 
dependence can be but slight; in other words the orientation of the molecules does 
not Involve any considerable additional kinetic energy of displacement of the medium 
