822 
MR. J. LARMOR OR THE ELECTRIC MEDIUM. 
velocity of the waves is (k/K/^)^, so that K = jx^. The equation of propagation when 
the molecules are moving through the stationary mther with velocity v in the direction 
of the wave-motion, is 
P 
(1% 
clP 
+ P 
+ y 
AY 
civ ) 
A. 
civ- 
5 
where ^. 2 , — (A ~ 1) -^i above. Thus, Vj being the velocity of the wave, and V 
the velocity of propagation in free sether, we have just as before 
giving very approximately Vj = V/x“^ -j- u (1 — which is Fresnel’s law. 
The exact expression for merely modifies the first term of Fresnel’s approxima¬ 
tion by a correction involving u®(l — pA’ which does not change sign with v; thus 
in the application to Michelson’s second-order experiment there is no essential 
modification, and his negative result remains outside the scope of this analysis. 
125. An important corollary to the present theory is suggested and confirmed 
by the experiments of PlONTGEN on the convection of excited dielectrics, mentioned 
above (§ 60). When a material dielectric is moved across an electric field, each ion of 
the group which constitutes one of its molecules produces its own convection current, 
composed partly of change of electric displacement in the surrounding free eether, but 
co]npleted and made circuital by the actual convection of the ionic charge itself 
When, as in Rontgen’s experiment, the configuration in space does not change by the 
motion, so that there is no displacement-current in the surrounding aether, it is easy 
to see that the total electromagnetic effect is the same as if the dielectric were 
magnetized to an intensity which is at each point the vector product of its velocity of 
movement and its electric moment per unit volume, the latter being (K — l)/47r 
times the electric force at the place. We have just seen (§ 124) that this is in accord 
with the optical aspect of convection of transparent matter. 
I have much pleasure in expressing my deep obligation to Professor G. F. FitzGerald 
for a very detailed and instructive criticism of this paper with which he has favoured 
me. I have been much guided by his comments in revising the paper, and would 
have made still more use of them but for the length to which it had already run. I 
need hardly state however that he is not to be held responsible for any of the views 
herein expressed. 
My best thanks are also due to Mr. A. E. H. Love for a criticism at an earlier 
stage, from which I derived much advantage. 
