ically suspended.
But you will ask whether filtering will not remove the suspended
matter? The grosser matter, undoubtedly, but not the more finely
divided matter. Water may be passed any number of times through
bibulous paper, it will continue laden with fine matter. Water passed
through Lipscomb's charcoal filter, or through the filters of the
Silicated Carbon Company, has its grosser matter removed, but it is
thick with fine matter. Nine-tenths of the light scattered by these
suspended particles is perfectly polarised in a direction at right
angles to the beam, and this release of the particles from the
ordinary law of polarisation is a demonstration of their smallness. I
should say by far the greater number of the particles concerned in
this scattering are wholly beyond the range of the microscope, and no
ordinary filter can intercept such particles. It is next to
impossible, by artificial means, to produce a pure water. Mr.
Hartley, for example, some time ago distilled water while surrounded
by hydrogen, but the water was not free from floating matter. It is
so hard to be clean in the midst of dirt. In water from the Lake of
Geneva, which has remained long without being stirred, we have an
approach to the pure liquid. I have a bottle of it here, which was
carefully filled for me by my distinguished friend Soret. The track
of the beam through it is of a delicate sky-blue; there is scarcely a
trace of grosser matter.

The purest water that I have seen--probably the purest which has been
seen hitherto--has been obtained from the fusion of selected specimens
of ice. But extraordinary precautions are required to obtain this
degree of purity. The following apparatus has been constructed for
this purpose: Through the plate of an air-pump passes the shank of a
large funnel, attached to which below the plate is a clean glass bulb.
In the funnel is placed a block of the most transparent ice, and over
the funnel a glass receiver. This is first exhausted and refilled
several times with air, filtered by its passage through cotton-wool,
the ice being thus surrounded by pure moteless air. But the ice has
previously been in contact with mote-filled air; it is therefore
necessary to let it wash its own surface, and also to wash the bulb
which is to receive the water of liquefaction. The ice is permitted
to melt, the bulb is filled and emptied several times, until finally
the large block dwindles to a small one. We