Bins 
The true value of d was then taken from the first estimations 
of van per Waats. Nowadays however we know the absolute 
dimensions of tke molecules and also the distance Jd. Moreover, 
thanks to the investigations on the interference of the Röntgen rays 
we can trace the structure of the crystals in detail. This made it 
desirable to take the problem at hand once more. While the wave- 
length of the Röntgen rays is comparable with d, the question is, 
whether already for light rays the molecular discontinuity is of 
influence, whether there are any indications that d/à may not be 
quite neglected. 
In my former calculations | made use of the theory of Maxwe.n 
in the form given to it by Hermnoutz. | followed this way because 
I had not yet penetrated deeply enough into the ideas of Maxwerr. 
In the first place, therefore, the calculations had to be repeated and 
to be based upon the theory of Maxwe.u and the theory of electrons. 
The new calculations gave the same results as the first ones. 
3. It will suffice to consider a cubical arrangement of the molecules. 
The equations for the light motion were derived on the supposition 
that equal particles are placed at the points of a cubic lattice. 
Further it has been assumed that in each molecule an electric force 
is excited and a corresponding electric moment in the direction of 
that force. 
From considerations on the symmetry of the crystal we may 
easily deduce that for some definite directions of propagation com- 
parable with the axis of monoaxial crystals we have only one velocity 
of propagation; these directions are those of the edges of the cubic 
lattice and of the diagonals of the elementary cube. By the diagonals 
of the side-faces of this cube however those directions of propagation 
ars given for which we may expect (and this is confirmed by the 
experiments) the anisotropy in question to be felt most strongly. 
Further on we shall always assume the direction of propagation to 
coincide with such a diagonal of a side-face of the cube. Then the 
principal directions of propagation A, and &, may be indicated 
immediately. 
The first one is that of the edge of the cube perpendicular to that 
side-face, the other one that of the second diagonal of that face. 
The velocities of propagation belonging to these directions of vibration 
will be indicated by v, and v,, while the corresponding values of 
: . C C \ : 
the refraction index — and — (c the velocity in vacuum) will be 
1 V5 
represented by mu, and wu. For the difference between these last 
