1 
THE ELECTiHC AND LUMINIFEROUS MEDIUM. 
809 
velocity of the electric system is taken greater and greater the permeability, in the 
direction of its motion, of the uniaxial medium of the analogy becomes less and less, 
and the field therefore becomes more and more concentrated in the equatoreal plane. 
When the velocity is nearly equal to that of radiation, the electric displacement 
forms a mere sheet on this plane, and the charge of the nucleus is concentrated on 
the inner edge of this sheet. The electro-kinetic energy of a current-system of this 
limiting type is infinite (§ 52), and so is the electrostatic energy ; thus electric inertia 
increases indefinitely as this state is approached, so that the velocity of radiation is 
a superior limit which cannot be attained by the motion through the sether of any 
material system. 
Again, the steady electric field carried along wdth it by a system rotating about a 
fixed axis with angular velocity oj is to be obtained by changing c//clt in the elastic 
equations into d/dt — codjdO', where 6' denotes relative azimuth around the axis ; 
they therefore assume the form 
o 
6)" 
o 
(/ 9 , ^0 = 
of which the solution would be difficult. And the equations of the relative steady 
field for the most general case of uniform combined translation and rotation of an 
electric system, supposed still of invariable shape, are expressed in like manner, by 
taking the central axis of the movement as the axis of x, in the form 
c2 
cd 
cP \ 
(d dev 
= 0 . 
The circuital character of will allow us to reduce these three variables in 
cases of symmetry to a single stream-function, of which the slope along the normal at 
the surface of the nucleus must be null. 
Any deviation from this steady motion of a molecule, produced by disturbance, 
will result in radiation which will continue until the motion has again become steady. 
If we roughly illustrate by the phenomena of the Solar system, the mean circular 
orbits of the planets will represent the steady motion, while disturbances introduce 
planetary inequalities which would give rise to radiation of corresponding periods. 
An apparent obstacle to the application of this hypothesis to the theory of the 
spectrum is that such a steady motion is not unique, its periods depend on the energy 
of the system; but, from whatever cause, the chemical energy of a molecule (wdrich 
is electric, therefore mthereal) has a definite value quite independent of the amount of 
matevkd kinetic energy that may be involved in its temperature and capacity for 
heat. The periods of the vibrations would thus be fixed by the electric eneigy ; wdiile 
the prevailing character of the disturbances, which determines the relative intensities 
of the radiations, would depend on temperature. If there are lines in any spectrum 
which have this kind of origin, wm should expect to find simple linear relations 
between the veci'proccds of their periods or wnive-lengths, as in the Planetary Theory. 
MDCroXCTV.—A. 5 L 
