137 
As is well known, from the sign of the Faraday and the Zee- 
man effect we can infer that (at least) certain spectral lines, the 
D-lines of Sodium for example, are due to the electromagnetic 
activity of negatively charged corpuscles. In such cases the value 
of e is negative; hence by formula (6) it is evident that the sign 
of the difference 
*1 — X 2 ( 10 ) 
depends solely upon the sign of the quantity F defined by equa¬ 
tion (7). 
§ ö. It now remains for us to collect the results of the fore¬ 
going considerations. Going back to equation (3) of § 4., we see 
that the amplitudes of the right-handed circularly polarized wave 
and of the left-handed circularly polarized wave, denoted by a 1 
and a 2 in §§ 1. and 2., are 
a 0 . exp 
f 2n% 1 z\ 
v r ) ; 
a 0 . exp 
(-T*-'). « 
a 0 being the amplitude of the original linearly polarized wave. 
Let us write 
g = ooy ce N/n 0 2 ; (2) 
thus g is identical with the quantity denoted by the same symbol 
in the paper On the Electromagnetic Theory of Dispersion etc. quoted 
above 1 ). Two cases have now to be examined: 
1. Suppose: 
i 
® 
V 
'X 
then : 
F< 0 ; 
< X 2 Î 
% > a 2 
2. Suppose: 
1 
y-H 
~"o 
A 
in this case: 
F> 0; 
> X 2 7 
a x < a 2 . 
From §§ 1. and 2. we conclude that in both cases we are entitled 
to assume 
_£ _ 1 — exp (— 2uAz/À) ^ 
a If - exp (— 2 n A z/X) 
where A stands for the absolute value of the difference — x 2 ; 
this value can be calculated with the help of equation (6), § 5. 
1 ) Bulletin lnt. for April 1907, p. 324. 
