compression of the water itself. When the pressure is judged to be
sufficient, the button, _n_, is turned, and the air compressed by the
pump finding an exit, the needle of the pressure gauge will be seen to
redescend and the level of the tube, _a b_, to rise.

If the glass of the bottle has undergone no permanent deformation, the
level will rise exactly to the zero mark, and denote that the bottle
has supported the test without any modification of its structure. But
if, on the contrary, the level does not return to the zero mark, the
limit of the glass's elasticity has been extended, its molecules have
taken on a new state of equilibrium, and its resistance has
diminished, and, even if it has not broken, it is absolutely certain
that it has lost its former resistance and that it presents no
particular guarantee of strength.

The vessel, A B, which must be always full of water, is designed to
keep the bottle at a constant temperature during the course of the
experiment. This is an essential condition, since the bottle thus
filled with water constitutes a genuine thermometer, of which _a b_ is
the graduated tube. It is therefore necessary to avoid attributing a
variation in level due to an expansion of the water produced by a
change in temperature, to a deformation of the bottle.

The test, then, that can be made with bottles by means of the
elasticimeter consists in compressing them to a pressure of ten
atmospheres when filled with water at a temperature of 25°, and in
finding out whether, under such a stress, they change their volume
permanently. In order that the elasticimeter may not be complicated by
a special heating apparatus, it suffices to determine once for all
what the pressure is that, at a mean temperature of 15°, acts upon