248 DAVIES, ON CRYSTALLIZATION 
2nd. Driving off by heat all or part of the water which 
is necessary to the formation of the crystal, and after- 
wards allowing it to reabsorb the same from the atmosphere, 
&c., as copper and magnesia (described in Vol. II, p. 128, of 
this Journal). 
3rd. Fusion, and again allowing to cool. 
From the second part of the first cause above mentioned 
it might appear that at a certain temperature the formation 
of crystals must inevitably take place; but this is not the 
case. Crystals do not readily form in any solution, even 
if “ supersaturated,” without some disturbance or inter- 
ference. This, perhaps, explains the difference in the shape 
of many crystals, as the same accidental causes which aid 
their formation necessarily act more or less upon the form 
finally assumed. But it must not be understood that the 
regulations required to ensure fine and well-shaped crystals 
are by any means arbitrary or useless. When crystalliza- 
tion commences at a high temperature, and the mother- 
liquor is allowed to cool slowly and uniformly, the crystals 
are better developed than those obtained by a different mode 
of proceeding. As an instance of this may be mentioned 
the following :—Mr. Jno. G. Dale has frequently shown me 
the vats in which he erystallizes certain of his salts. These 
are not only made warm before the solutions are allowed to 
flow into them, but are deeply imbedded in sawdust, which 
is an imperfect conductor of heat, and thereby the cooling 
is rendered much more gradual. Thus the “accumulation” 
of crystalline forms receives no more interference from 
brother crystals than is absolutely necessary. This accumu- 
lation is often distinctly visible in large crystals, particularly 
in alum. One which I have lately obtained from the vats 
of my friend Mr. Dale shows each superimposed plate very 
clearly, though the transparency is almost perfect. 
In cases where crystals are formed in masses their shapes 
are necessarily rendered invisible. Thick ice might be sup- 
posed to show no crystalline structure. It is, however, by no 
means “ structureless,” in proof of which may be quoted the 
authority of Professor Tyndall :—* Crystallization during 
freezing, takes place very similarly to snow crystals, even in 
thick blocks of ice from Norway or Wenham Lake, and the 
constitution of these masses is easily revealed by reversing 
the process which formed them. A large converging lens 
was placed in the sunbeams passing through a room, and 
the piece put into such a position that the poit of con- 
vergence fell within it. Along the course of sunlight the 
ice became studded with lustrous spots. On examining the 
