( 672 ) 
§ 5. As to the relation between electric polarization and electric 
force, this is given by LARMOR in the form 1) 
' K—1 €g oR 
9 Ane? PDA , 
Be Ser &9 dQ 
~ Agee 4 nc? de 
or 
K=4 S, 
M, = : Eg AC 
Ane? Anc? dz (6) 
— 1 &5 dE, ; 
eM, = 5 
4m c* TE Fr 
Now, in my formula (1) the rotational terms are very much 
smaller than the first term o WM, We may therefore, in those terms, 
1 
replace M by = €. Hence 
N if J ; te 1 
eae Le stent (7) 
5 
and, in the case under consideration, 
1 Pile RE 
M, = — Ey a ee 
6 ode, OF ta 
Wat j 3E, 
Me AE Ak + Pr Ey. 
If this is compared with (6), it appears that the formulae of 
LARMOR agree with the particular case k = 0 of my theory, and 
that the coefficients we have introduced are related to each other 
as follows: 
et a A ae 
dre no hi Wa oe es Eden 
(9) 
§ 6. For &=0 my formula (2) gives 
2% 1 Wye  . 
omt (1+ : )i DR 
a value depending on px. On the contrary, LARMOR's result does 
not contain the velocity of translation, but this is only so, because 
his calculation of the angle of rotation is not quite exact. 
As is well known, this angle may be expressed in the velocities 
of propagation of right- and left-handed circularly polarized rays. 
In doing this, we have first of all to assign to the period of 
vibration, taken with reference to a fixed point of the substance, a 
__ 
1) Le, p. 211, As I shall not consider the magnetic rotation, I have put 2, = 0. 
eee 
EE EN 
