(66) 
1 y ] 
Dmv 8" 8 En 2 
DNG ee Fe E99? IE FET =| - (23) 
Here I have put 
Pe. hae a = 8; 
and 
PERI — 92 = Sp. 
It is easily seen that, with the assumptions we have made concer- 
ning the magnitude of the different terms, the equations (22) and 
(23) imply the existence of two absorption-bands, corresponding to 
WD] = 0 and Sg = 0. 
These bands are precisely the outer components of the triplet, 
one is led to by the elementary theory of the ZErMAN-effect. 
The breadth of each of these lines will be equal to that of the 
original absorption-band; in virtue of our assumptions it will be 
much smaller than the distance of the two lines. 
Now it is clear that such a thing would be impossible, if the 
modification of the propagation of light were so small as PoINcARÉ 
finds, namely of the order of #?, if R is the strength of the field. 
If, by the action of the magnetic field, the maximum of absorption 
is shifted to a place, where the absorption was originally insensible, 
we have to do with a finite change at this point of the spectrum. 
§ 10. In order to examine this more closely, we must return 
to the equations of motion themselves, from which the formulae 
(22) and (23) have been deduced. Let the magnetic force be 
parallel to the axis of z (A= B—0, C= R) and let the propaga- 
tion of light take place along the axis of z. Then the complex 
expressions, which satisfy the equations of motion and whose real 
parts are the values of U, V, W, &, 1, ¢, ete, will contain the 
common factor . 
AL 
bij 1 ; = 
LL ee . 
aD) a o 
There will arise no confusion, if we use the letters U, V, ete. 
themselves to represent these complex expressions. 
Let the vector P be perpendicular to the axis of z. Then 
W=0, and Z=0. 
- 
pn ai dn tn 
