410 
DR. A. E. H. TUTTON ON THE 
The approximate values obtained for a from the p -faced prism and for y from the 
g-faced prism, were as under :— 
Light. 
Approximate a. 
Approximate y. 
Li. 
1-5078 
1-5307 
C. 
1-5083 
1-5312 
Na.. 
1-5113 
1-5345 
T1. 
1-5145 
1-5381 
Cd. 
1-5163 
1-5400 
F. 
1-5182 
1-5422 
G. 
1-5242 
1-5484 
As the direction for which this approximate value for y was determined was only 
3|-° from the exact direction for the y axis of the ellipsoid, it is probable that the above 
value for y is practically correct. The approximate value for a was determined for a 
direction lOip removed from the true direction for the a (minimum) axis of the ellipsoid, 
and its value is therefore somewhat too high, probably by about O'OOl to 0‘002. 
Finally, in the fourth and last crop of crystals obtained there was a crystal large 
and transparent enough to suggest an attempt at grinding a prism to alford a and y 
directly. One was therefore ground in the correct orientation to alford these indices 
directly and absolutely. The refracting edge was parallel to the second median line, 
(the a axis of the ellipsoid), so that vibrations affording the signal image when the 
Nicol was at 90° corresponded to the a index. The two faces of the prism were 
symmetrical to (inclined 30° on each side of) the symmetry plane b {010}, so that the 
other rectangular direction of vibration within that plane, corresponding to the image 
transmitted when the Nicol was at 0°, was parallel to the y axis of the ellipsoid and 
afforded the index y. By working very rapidly and keeping the crystal covered with 
the lubricating oil after its goniometrical adjustment to the correct desired orientation, 
and as soon as the grinding and polishing were completed covering the prepared 
surfaces with balsam-cemented glass plates, the prism was completed before any 
serious loss of transparency had occurred ; an hour was permitted for the drying of 
the hard balsam in benzene, and the transparency was then still such as enabled the 
refractive indices a and y to be determined for all wave-lengths from Li red as far as 
greenish blue F of the spectrum, only the G-images being too weak for accurate 
allocation to the cross wires. The results are given below. 
Light. 
a. 
y- 
Li. 
1-5059 
1-5306 
C. 
1-5064 
1-5311 
Na. 
1-5095 
1-5345 
T1 ....... . 
1-5127 
1-5379 
Cd. 
1-5145 
1-5399 
F. 
1-5164 
1-5421 
