be a gradual
cessation of the effects of the rays.

The result of our study of the halo may be summed up in the
statement that in nearly every particular we have the phenomena,
which have been measured and observed in the gas, reproduced on a
minute scale in the halo. Initial recombination seems, however,
to be absent or diminished in effectiveness; probably because of
the new stability instantly assumed by the ionised atoms.

One of the most interesting points about the halo remains to be
referred to. The halo is always uniformly darkened all round its
circumference and is perfectly spherical. Sections, whether taken
in the plane of cleavage of the mica or across it, show the same
exactly circular form, and the same radius. Of course, if there
was any appreciable increase of range along or across the
cleavage the form of the halo on the section across the cleavage
should be elliptical. The fact that there is no measurable
ellipticity is, I think, one which on first consideration would
not be expected.

233

For what are the conditions attending the passage of the ray in a
medium such as mica? According to crystallographic conceptions we
have here an orderly arrangement of molecules, the units
composing the crystal being alike in mass, geometrically spaced,
and polarised as regards the attractions they exert one upon
another. Mica, more especially, has the cleavage phenomenon
developed to a degree which transcends its development in any
other known substance. We can cleave it and again cleave it till
its flakes float in the air, and we may yet go on cleaving it by
special means till the flakes no longer reflect visible light.
And not less remarkable is the uniplanar nature of its cleavage.
There is little cleavage in any plane but the one, although it is
easy to show that the molecules in the plane of the flake are in
orderly arrangement and are more easily parted in some directions
than in others. In such a medium beyond all others we must look
with surprise upon the perfect sphere struck out by the alpha
rays, because it seems certain that the cleavage is due to lesser
attraction, and, probably, further spacing of the molecules, in a
direction perpendicular to the cleavage.

It may turn out that the spacing of the molecules will influence
but little the average number per unit distance encountered by
rays moving in divergent paths. If this is so, we seem left to
conclude that, in spite of its unequal and pola