alkali from which it was
separated by the acid; for Dr. _Hales_ has clearly proved, that alkaline
salts contain a large quantity of fixed air, which they emit in great
abundance when joined to a pure acid. In the present case, the alkali is
really joined to an acid, but without any visible emission of air; and
yet the air is not retained in it: for the neutral salt, into which it
is converted, is the same in quantity, and in every other respect, as if
the acid employed had not been previously saturated with _magnesia_, but
offered to the alkali in its pure state, and had driven the air out of
it in their conflict. It seems therefore evident, that the air was
forced from the alkali by the acid, and lodged itself in the _magnesia_.

These considerations led me to try a few experiments, whereby I might
know what quantity of air is expelled from an alkali, or from
_magnesia_, by acids.

Two drams of a pure fixed alkaline salt, and an ounce of water, were put
into a Florentine flask, which, together with its contents, weighed two
ounces and two drams. Some oil of vitriol diluted with water was dropt
in, until the salt was exactly saturated; which it was found to be, when
two drams, two scruples, and three grains of this acid had been added.
The vial with its contents now weighed two ounces, four drams, and
fifteen grains. One scruple, therefore, and eight grains were lost
during the ebullition, of which a trifling portion may be water, or
something of the same kind. The rest is air.

The celebrated _Homberg_ has attempted to estimate the quantity of solid
salt contained in a determined portion of the several acids. He
saturated equal quantities of an alkali with each of them; and,
observing the weight which the alkali had gained, after being perfectly
dryed, took this for the quantity of solid salt contained in that share
of the acid which performed the saturation. But we learn from the above
experiment, that his estimate was not accurate, because the alkali loses
weight as well as gains it.

Two drams of _magnesia_, treated exactly as the alkali in the last
experiment, were just dissolved by four drams, one scruple, and seven
grains of the same acid liquor, and lost one scruple and sixteen grains
by the ebullition.

Two drams of _magnesia_ were reduced, by the action of a violent fire,
to two scruples and twelve grains, with which the same process was
repeated, as in the two last experiments; four drams, one scruple,