338 
Though this contradiction is not satisfactory, our astonishment 
inust not be too great, to my opinion, as in the light of our present 
knowledge, the theory from which the equation has been derived 
is so very imperfect. There the assumption bas been made that 
equal molecules were placed at the points of the crystal lattice and 
that in each of these an electric moment is excited as will be the 
case when the molecule contains a quasi-elastically and moreover 
isotropically bound electron. According to the present opinion 
however the sodium and chlorine nuclei are placed alternately along 
each edge of the lattice, while round these nuclei and perhaps also 
round the lines of connection electrons are circulating. When this 
circulation takes place in planes, the position of these planes may 
give rise to an anisotropy. 
Perhaps the only thing that can make plausible the old theory is, 
that an anisotropy may be expected which like that determined by 
2 
J 
(1) is proportional with —. | have not tried a calculation based upon 
ZR? 
the new points of view. First we shall have to be further in the 
general treatment of light vibrations. 
That the old theory is imperfect ‘in several respect may be seen 
from the following. The difference in phase determined by (2) strongly 
increases with diminishing wavelength and when working with white 
light, we should therefore see the field distinctly coloured. This is 
however not at all the case. 
As to the value of the difference in phase, it has hardly been 
possible to determine it because of the imperfectness of the extinctions. 
For the crystals used it could not be measured with the compen- 
sator of Bapinet. The only thing that could be done was to deter- 
mine with this means the difference in phase of the compressed 
glass plate by which the double refraction of the rock-salt was 
compensated rather satisfactorily. In this way it was found, that 
the difference in phase was a small fraction, about >; or qo of a 
wavelength. 
The idea suggests itself to work also with crystals, the side-faces 
of which are perpendicular to an edge of the lattice, in which case 
a double refrection as has been described above cannot exist. 
To my astonishment even now we can often distinguish two 
mutually perpendicular positions in which the intensity is a mini- 
mum, so that we get the impression that also for the mean of the 
accidental double refraction taken over the cross-section of the crystal, 
we can speak of two principal directions. The phenomena however 
were doubtlessly less regular than for the erystals with which the 
