MR. .R. T. GLAZEBROOK OIST PLANE WAVES IN A BIAXAL CRYSTAL. 
319 
Such a variation would produce an alteration in the angle between the optic axes, 
and therefore in the values of 0, 0' for any wave normal. 
If, however, we take a wave normal at some distance from the optic axes, the values 
of 0, 0' will be altered by nearly equal amounts in opposite directions, and 0+0' will 
be nearly constant. 
Let us then find the value for a in order that the experimental value in line 29 of 
Table XXV. may agree with theory. 
We have 
2'y 1 ~ = cr+c 2 — (ft 2 — c~) cos (0+0') 
=« 2 {1 — cos (0+0')} 
+ c 2 {1 + cos (0+ 6 ')} 
o o o 0 + 0 
v x ~ — <r cos - "—-— 
• \ ft": 
sur 
0 + 0 ' 
v, = - 
0-1 
/x, = 1-67274 
0+0' = 3l° 45' 40' 
c 2 ='351870 
From these we get 
a 2 ='447553 
as against '427117, and the value of fi c is 
^ — 1-49478 
instead of 1'53013 
A reference to the tables for the second prism shows at once the impossibility of 
such a change. 
It is true that any decrease in fi c would produce a decrease in the angle between 
the optic axes. This would increase 0' and decrease 0 by nearly equal amounts, and 
on the whole produce a small decrease in 0+0'. 
This change would decrease v x and therefore increase /x l5 and hence to produce the 
required change in fiy—i.e., to make /x 1 equal to the experimental value in line 29—we 
must take a value for /i r larger, but only slightly larger, than that given above. 
There remains now for discussion the effect of a variation in the position of the 
plane of the prism with reference to the crystallographic axes. 
We shall consider this by discussing separately the effect of a rotation about eacli 
of three lines through the centre approximately at right angles. These lines are 
