52 FORMATION OF SHELLS OF ANIMALS, ETC., 
which, being packed together in different ways, will 
produce all the varied forms peculiar to the several classes 
of crystals. Now, as the molecules or atoms composing 
the primitive crystals are finite magnitudes, the same 
observation will apply to the apparently smooth crystals 
as was applied to the apparently spherical calculi; namely, 
that as the latter were only polygons of a number of 
sides corresponding to that of the molecules contained in 
the most superficial lamine, so the former, however sharp 
their angles or smooth their surfaces may appear, neither 
their corners are geometrically angular nor their surfaces 
and edges mathematically plane, but both will present in- 
equalities of the same shape and size as their ultimate 
molecules. This explanation of the process of crystalliza- 
tion is not at all inconsistent with the fact of the expan- 
sion of water prior to and at the instant of congelation, as 
it is not improbable that, after a certain degree of approxi- 
mation of the molecules of water, electricity is evolved, 
causing at first expansion, the obvious result of molecular 
repulsion, and afterwards rectilmear molecular arrange- 
ment, the consequence of impulsion, this latter taking 
place suddenly. The force of expansion is well known to 
be very great, its total effect being probably equal to the 
sum of the impulses of all the moleeules of the freezing 
liquid. It is very probable that this might easily be de- 
termined by experiment. I may observe that I consider 
this observation as altogether unconnected with what has 
been before advanced, and merely regard it as a suggestion 
quite as likely to be true as any other explanation which 
may have been given of this remarkable fact. 
Having now discussed fully the subject of crystalliza- 
tion, I shall proceed to apply the facts elicited by the dis- 
cussion to the explanation of the cause of the radiated 
