cushion with air from the lungs. Experience shows what pressure is
necessary with any given paper, and the gauge enables one to know that
the pressure is neither deficient nor in excess of that which is safe
for the glass.

[Illustration: PLAN. COTTON FLANNEL REMOVED.]

[Illustration: SECTION AT CO.]

_The Construction of the Air-Cushion._--The expense of such an
air-cushion seemed at first likely to prevent its being used; but a
method of construction suggested itself, the expense of which proved to
be very slight. The wooden back-board, as constructed, is made in one
piece containing no wide cracks. It has laid upon it some thick brown
Manila paper, the upper surface of which has been previously shellacked
to make it entirely air-tight. Upon this shellacked surface is laid a
single thickness of thin paper of any kind; even newspaper will answer.
Its object is simply to prevent the sheet rubber, which forms the top of
the air-cushion, from sticking to the shellacked paper. The heat of the
sun is often sufficient to bring the shellac to a sticky state. It would
probably answer as well to shellac the under side of the paper, and to
use but one sheet, but I have not tried this plan. Around the periphery
of the pad, there is laid a piece of rubber gasket about one and a half
inches wide, and about one-eighth of an inch thick. In order that the
gasket may not be too expensive, it is cut from two strips about three
inches wide. One of them is as long as the outside length of the frame,
and the other is as long as the outside width of the frame. Each of
these strips is cut into two L-shaped pieces, an inch and a half in
width, with the shorter leg of each L three inches long. When the four
pieces are put together a scarf joint is made near each corner, having