362 
MR. R. T. GLAZEBROOK OR PLANE WAVES IN A BIAXAL CRYSTAL. 
-V) cos (^“#')• 
A 1 " A^c V/A? A 1 ® / 
We have seen that $ is decreased by the change. 
(9' is at first decreased, though by less than d, afterwards it is increased. 
Therefore, for any point on the curve L P the value of 6 — O' is decreased by this 
alteration. 
Therefore cos (6—O') is increased. 
~ —V) cos (6—0') is increased, though by less than 
cos (6—6') is increased. 
j^c fJ'a” 
Therefore — 7 ;+A ,— 1 — 
/V \f^c 
f^a 
Therefore /x., is decreased. 
Or the variation in [i 2 is in the same direction as in /x^ 
But since 6—6' is <0-}-6' -cos (6 — O') is >(— % -—cos (6-\-&). 
\H'c' l^a ) XM'C / 
And the whole change in 
h+—.-A—b) cos (6-ff) 
AV /W \JJ>c AW/ 
is much less than the whole change in 
- 2 +w 
—8— K) cos (6+6'). 
j^c /v 
Or the decrease in /x 3 is small compared with that in J u, 1 . 
Again, as the variation in 6 — O' varies from zero at L to a maximum about K, while 
cos (0—6') also varies considerably, the variation in /x^ will be considerably different 
for different points between L and Iv. 
After passing K, however, 6—0' is nearly constant all along the curve. 
The change in the value of jx 2 , therefore, will not differ much for different points, 
that is, for different waves. 
The whole effect of the alteration, therefore, will be to decrease throughout the 
theoretical value of /jl 2 , this decrease being greatest about Table V., line 23. 
In the greater part of Table V. the result will be to produce a closer agreement 
between theory and experiment. 
But in the rest of the work, since the theoretical value of /x 3 is less than the experi¬ 
mental, the effect will be to increase the difference between the two and so widen the 
discrepancy between theory and experiment. 
Thus taking both sheets we cannot produce on the whole a closer agreement- 
bet ween theory and experiment by decreasing the value of /x c . Neither can we by 
increasing it. 
For though we might produce closer agreement for the outer sheet by thus increasing 
