of the visible spectrum. We can omit the term for the ultrared 
altogether in the sum S for the constant of rotation, as for this 
term we may put 4” =O, so that then we tind a one-term form 
for the constant of rotation, namely : 
2 vy? (n + 2)? D 
Den (wrr?) 
oO == . 
This form has been examined for the constants of rotation of the 
first series of observations (preceding communication p. 103). 
Observations of the index of refraction for different wavelengths 
have been made by H. BecQqvereL *). He found: 
C D E F G 
n= 1,5948 1,6043 1,6171 1,6293 1,6557. 
By the aid of this an interpolation formula has been calculated, 
and with this the values of 7 for the wavelengths occurring in the 
two tables with results have been determined. Then the values of 
D and yr, have been determined by the method of the least squares 
for the first series. The agreement appeared unsatisfactory. 
If it is further tried to represent BECQUEREL's values for the index 
of refraction with the found free frequency (»,* = 40,381.10°®) by 
a form 
n°—l A B 
=A = lg laa ara 
n° +2 vy? Vr 
it is found that this is only possible when a negative value is taken 
for A. This result must be rejected, as A=} 6” must always 
be positive. 
We shall therefore introduce an ultrared and two ultraviolet 
free frequencies, namely one far in the ultraviolet, so that »,* can 
be substituted for »,°—r° in the corresponding terms, and one at a 
moderate distance from the visible spectrum. ‘Two terms must then 
occur in the sum in the expression for the constant of rotation, and 
we must therefore put, 
Ip? (n2-1-2)2 D 
Agel p? (n?4 * (Be b ) 
I 9 
5 Yen (vy? —v?)? 
whereas for 7 a form of three terms must be taken, viz.: 
n?—l Cy Ce 
n° 42 viv Vv 
The values Da, Dy, and », of the first form have been calculated 
1) H. BecQvEREL. Ann. d. Ch. et d. Ph. (5) 12 p. 82 (1877). 
