w it from
its connection with caloric, and to decompose air or oxygen gas. This
combustible body has itself so great affinity with caloric, that, unless
when engaged in a combination with some other body, it always subsists
in the aeriform or gasseous state, in the usual temperature and pressure
of our atmosphere. In this state of gas it is about 1/13 of the weight
of an equal bulk of atmospheric air; it is not absorbed by water, though
it is capable of holding a small quantity of that fluid in solution, and
it is incapable of being used for respiration.

As the property this gas possesses, in common with all other combustible
bodies, is nothing more than the power of decomposing air, and carrying
off its oxygen from the caloric with which it was combined, it is easily
understood that it cannot burn, unless in contact with air or oxygen
gas. Hence, when we set fire to a bottle full of this gas, it burns
gently, first at the neck of the bottle, and then in the inside of it,
in proportion as the external air gets in: This combustion is slow and
successive, and only takes place at the surface of contact between the
two gasses. It is quite different when the two gasses are mixed before
they are set on fire: If, for instance, after having introduced one part
of oxygen gas into a narrow mouthed bottle, we fill it up with two
parts of hydrogen gas, and bring a lighted taper, or other burning body,
to the mouth of the bottle, the combustion of the two gasses takes place
instantaneously with a violent explosion. This experiment ought only to
be made in a bottle of very strong green glass, holding not more than a
pint, and wrapped round with twine, otherwise the operator will be
exposed to great danger from the rupture of the bottle, of which the
fragments will be thrown about with great force.

If all that has been related above, concerning the decomposition of
water, be exactly conformable to truth;--if, as I have endeavoured to
prove, that substance be really composed of hydrogen, as its proper
constituent element, combined with oxygen, it ought to follow, that, by
reuniting these two elements together, we should recompose water; and
that this actually happens may be judged of by the following experiment.

_Experiment Fourth._

I took a large cristal baloon, A, Pl. iv. fig. 5. holding about 30
pints, having a large opening, to which was cemented the plate of copper
BC, pierced with four holes, in which four tubes terminat