THE VAPvIATION OF ANGLES OBSERVED IN CRYSTALS. 
489 
icositetrahedron inclined to o at O'’ 3|-'. It is, no doubt, the development of such 
faces which converts the triangular plates of growth into hexagonal plates, as 
described above. 
These and other instances indicate that the first effect of slow solution is to 
produce prerosion faces on the edges of the crystal; on the octahedron edge these 
belong to triakis-octahedra ; on the edges of the vicinal faces they belong to 
icositetrahedra. The latter are inclined to the octahedron at about the same ancle 
o 
as the triakis-octahedron, and they certainly do not truncate its edges. The next 
effect of the process of solution is to produce etched triangular pits bounded by flat 
triakis-octahedra. 
(9.) Some Possible Causes op the Variations. 
With all the substances examined, as with alum, vicinal faces are the rule; they 
change their position as the crystal grows; they continue to deviate more widely 
from simple forms, at any rate during the initial growth ; the reflections move or 
oscillate generally in well defined zones, and in accordance with the symmetry of the 
crystal; their movement is not gradual but per scdtum; their deviation from the 
simple forms which they replace generally amounts to from 3 to 20 minutes. 
In making the observations great care must be exercised lest deviations due to 
refraction through layers or currents of more or less concentrated liquid be mistaken 
for those due to the change of inclinations in the vicinal faces. Suclr currents are 
always streaming round a growing crystal, and do produce visible displacements of 
the images; Imt these displacements are rather of the nature of irregular oscillations, 
and may be detected with a little practice. 
When beginning the study of alum, I endeavoured to eliminate the action of tliese 
currents with the object of ascertaining whether the production of the vicinal faces 
might not be actually due to tliein. By means of a suitable clockwork arrangement 
a wooden vane with four arms was kept revolving in a beaker full of concentrated 
alum solution during several hours, so as to keeji the liquid continually stirred. 
Small alum crystals were allowed to crystallise on the vane, and were examined at 
intervals. I came to the conclusion that the crystals were equally perfect, and 
possessed vicinal faces of the same sort, whether they crystallised in a still solution 
or in one which was kept continually stirred. In the same way the stirring of the 
liquid in the goniometer trough seemed to produce no appreciable effect upon the 
nature of the vicinal faces on a crystal of alum which had been growing in the still 
solution; the introduction of a stirrer into the trough did not affect the smoothness 
or the angles of the faces so far as could be observed ; the experiments are, however, 
difficult to carry out, and it is of course impossible to make accurate measurements 
while the liquid is in violent motion, although the crystal itself is not disturbed.'^ 
* Other and more precise experiments of this nature are described below on p. 518. 
VOL. CCII.—A. 3 R 
