337 
we have at every point of the image rays that have crossed the 
crystal at different points of the section. We may say, that of that 
double refraction we observe the mean over the whole section. 
It may be remarked, that we can also speak of such a mean 
with respect to the succeeding layers of the crystal and that the 
great thickness which we use is in a certain respect an advantage. 
When namely the disturbing double refraction changes irregularly 
along the ray we may assume the intensity caused by it to be 
proportional with the thickness, while a regular double refraction 
gives an intensity proportional with the second power of the thickness 
as long as the difference in phase is small. 
By these disturbing anomalies a sharp focussing at the miniwum 
is impossible. In order to form an estimation of the accuracy of the 
experiment and therefore of the regularity of the phenomenon we 
read the positions of the crystal on a divided circle after having focussed 
at a minimum as accurately as possible. When one of the crystals 
was rotated continually we obtained the following readings in degrees: 
17 107 202 294 
14 105 196 293 
19 108 196 291 
When the numbers of the second column are diminished by 90, 
those of the third one by 180 and those of the fourth one by 270 
and when we then take the mean of the 12 numbers, we find 18, 
so that the principal positions would be 
18 108 198 288 
With the other crystals we obtained similar results, sometimes a 
little better, sometimes a little worse. 
8. The positions of extinction were always those, for which the 
directions indicated with AR, and &, coincide with the directions of 
vibration of the. Nicols. As to this, the theoretical expectation is 
therefore confirmed. 
In order to see whether also the sign of the double refraction 
agrees with (1) we used a glass plate compressed in one direction 
which was placed in the way of the rays and which was rotated 
in its plane. It was found then, that for all ten crystals that were 
investigated and which were cut from different pieces of rock-salt, 
the effect could be compensated by the glass plate, when the direction 
of compression coincided with &,. Taking into consideration § 4 
we come to the conclusion that the sign is opposite to that given 
by form (1). 
