st necessarily act in the same way with
ordinary elastic fluids, and overturn every thing that opposes its
passage. This must, at least in part, take place when gun-powder is set
on fire in a cannon; as, although the metal is permeable to caloric, the
quantity disengaged at once is too large to find its way through the
pores of the metal, it must therefore make an effort to escape on every
side; and, as the resistance all around, excepting towards the muzzle,
is too great to be overcome, this effort is employed for expelling the
bullet.

The caloric produces a second effect, by means of the repulsive force
exerted between its particles; it causes the gasses, disengaged at the
moment of deflagration, to expand with a degree of force proportioned to
the temperature produced.

It is very probable that water is decomposed during the deflagration of
gun-powder, and that part of the oxygen furnished to the nascent
carbonic acid gas is produced from it. If so, a considerable quantity of
hydrogen gas must be disengaged in the instant of deflagration, which
expands, and contributes to the force of the explosion. It may readily
be conceived how greatly this circumstance must increase the effect of
powder, if we consider that a pint of hydrogen gas weighs only one
grain and two thirds; hence a very small quantity in weight must occupy
a very large space, and it must exert a prodigious expansive force in
passing from the liquid to the aeriform state of existence.

In the last place, as a portion of undecomposed water is reduced to
vapour during the deflagration of gun-powder, and as water, in the state
of gas, occupies seventeen or eighteen hundred times more space than in
its liquid state, this circumstance must likewise contribute largely to
the explosive force of the powder.

I have already made a considerable series of experiments upon the nature
of the elastic fluids disengaged during the deflagration of nitre with
charcoal and sulphur; and have made some, likewise, with the oxygenated
muriat of potash. This method of investigation leads to tollerably
accurate conclusions with respect to the constituent elements of these
salts. Some of the principal results of these experiments, and of the
consequences drawn from them respecting the analysis of nitric acid, are
reported in the collection of memoirs presented to the Academy by
foreign philosophers, vol. xi. p. 625. Since then I have procured more
convenient instruments, and