932 
from which, again assuming a mean value A, follows: 
(Z) > B, = 3,738.10, Asm = — 0,838.10? 
(ID) — 3,938.10, — — 0,784.10? 
To 
py? = (1) 47,485 10°, == (Il) 46,161.10", 
correspond the wavelengths 
Ay = (I) 0,0000274 em, = (ID 0,0000272 cm. 
From the third term can only be derived : 
= B. = (1) 0,0706 . 10°, - = (1) 0,0647. 10°°. 
First of all we can draw the conclusion from these results that 
the found coefficients cannot lay claim to great accuracy. Small 
deviations in the dispersion curve from these series give already 
pretty great deviations in the coefficients and the values derived 
from them. Only 2, agrees for the two series. 
It further appears that really the dispersion of the index of 
refraction as well as that of the constant of rotation can be repre- 
sented by the assumption of free frequencies: 
1. one far in the ultraviolet, so that the corresponding terms 
can be considered constant, with a normal magnetic resolution, 
2. one in the ultraviolet at a moderate distance from the visible 
spectrum, 
3. one in the ultrared. 
For the second wavelength, which gives terms which have a 
preponderating influence on the dispersion, we must assume a nega- 
tive magnetic resolution of a value amounting to about a quarter 
of the normal value. As was already remarked in the first commu- 
nication this negative magnetic resolution can be accounted for by 
assuming couplings between the electrons. 
: Ala Ne 
In Lorentz’s theory the quantity 6 is an abbreviation for wi 5 
in which A is closely connected with the mass of the vibrating 
particles. The fact that for 2 4,, connected with the ultrared wave- 
length, a much smaller value is found than for the other wavelengths 
is in keeping with the view that the ultrared vibrations are executed 
by vibrating molecules, so that much larger masses come into play. 
Taking everything together it may be said that the dispersion of 
TiCl, can be explained by the theory of Lorenrz. 
Physical and electrotechnical laboratory 
Delft. of the Technical University. 
1) Lorentz, loc. cit. p. 229. 
