THE VARIATION OF ANGLES OBSERVED IN CRYSTALS. 
473 
these are generally two in number, one flatter than tlie other, and replacing the 
upper and lower portions of the large apparent octahedron face (fig. 7). 
An apparent octahedron face of ahnn gives, therefore, in general, not one but three 
images. For the purpose of correct adjustment, and in order to distinguish easily 
between the various images, I have generally used a small s<|uare aperture as 
collimator signal in place of the customary Websky slit. 
Two apparent faces, o, w, of the octahedron (fig. 8), consisting in reality of the flat 
triakis-octahedron a, ^8, y, will then give the reflections represented in fig. 9. The 
three images a, y belonging to the face o are not necessarily separated from each 
other by the same intervals as those belonging to the fane o ; but the zone can be 
correctly adjusted by means of the images and < 0 ^. 
Examj)le. 
A very perfect octahedron, after 10 days' immersion in solution and various 
changes of temperature, was found on Novemlier 27, 1893, at 10 A.M., to yield three 
Fig. 10. 
Calling these faces o, 
follows:— 
&», o, M respectively, the readings and Images were as 
o, = 327° 57i', 
= 257° 401', 
(o^=:co^ = 257° 26^', 
o', 148° 61', 
o', = 147° 5G|', 
0 .',=: 77°39l', 
(Here and in the 
following pages o, 
denotes the reading 
for winch the imao’e 
o 
from a. is on tlie 
vertical wire.) 
Here none of the measured angles is the true octahedron angle, but if we assume 
that the three faces a, /3, y of the triakis-octahedron rejflacing any one octahedron 
VOL. CCII.—A. 3 p 
