evaporation is promoted, but by evaporation the water only is removed.
The plate is then placed in a solar microscope, and an image of the
film is thrown upon a white screen. The warmth of the illuminating
beam adds itself to that already imparted to the glass plate, so that
after a moment or two the dissolved salt can no longer exist in the
liquid condition. Molecule then closes with molecule, and you have a
most impressive display of crystallizing energy overspreading the
whole screen. You may produce something similar if you breathe upon
the frost ferns which overspread your window-panes in winter, and then
observe through a pocket lens the subsequent recongelation of the
film.

In this case the crystallizing force is hampered by the adhesion of
the film to the glass; nevertheless, the play of power is strikingly
beautiful. Sometimes the crystals start from the edge of the film and
run through it from that edge; for, the crystallization being once
started, the molecules throw themselves by preference on the crystals
already formed. Sometimes the crystals start from definite nuclei in
the centre of the film, every small crystalline particle which rests
in the film furnishing a starting-point. Throughout the process you
notice one feature which is perfectly unalterable, and that is,
angular magnitude. The spiculæ branch from the trunk, and from these
branches others shoot; but the angles enclosed by the spiculæ are
unalterable. In like manner you may find alum-crystals,
quartz-crystals, and all other crystals, distorted in shape. They are
thus far at the mercy of the accidents of crystallization; but in one
particular they assert their superiority over all such
accidents--_angular magnitude_ is always rigidly preserved.

My second example of the action of crystallizing force is this: By