er in inflammable air,
but there was no reason for my fear; for it exploded quite freely in
this air, leaving it, in all respects, just as it was before.

In order to make this experiment, and indeed almost all the experiments
of firing gunpowder in different kinds of air, I placed the powder upon
a convenient stand within my receiver, and having carefully exhausted it
by a pump of Mr. Smeaton's construction, I filled the receiver with any
kind of air by the apparatus described, p. 19, fig. 14, taking the
greatest care that the tubes, &c. which conveyed the air should contain
little or no common air. In the experiment with inflammable air a
considerable mixture of common air would have been exceedingly
hazardous: for, by that assistance, the inflammable air might have
exploded in such a manner, as to have been dangerous to the operator.
Indeed, I believe I should not have ventured to have made the experiment
at all with any other pump besides Mr. Smeaton's.

Sometimes, I filled a glass vessel with quicksilver, and introduced the
air to it, when it was inverted in a bason of quicksilver. By this means
I intirely avoided any mixture of common air; but then it was not easy
to convey the gunpowder into it, in the exact quantity that was
requisite for my purpose. This, however, was the only method by which I
could contrive to fire gunpowder in acid or alkaline air, in which it
exploded just as it did in nitrous or fixed air.

I burned a considerable quantity of gunpowder in an exhausted receiver
(for it is well known that it will not explode in it) but the air I got
from it was very inconsiderable, and in these circumstances was
necessarily mixed with common air. A candle would not burn in it.

SECTION VIII.

_QUERIES, SPECULATIONS, and HINTS._

I begin to be apprehensive lest, after being considered as a _dry
experimenter_, I should pass, with many of my readers, into the opposite
character of a _visionary theorist_. A good deal of theory has been
interspersed in the course of this work, but, not content with this, I
am now entering upon a long section, which contains nothing else.

The conjectures that I have ventured to advance in the body of the work
will, I hope, be found to be pretty well supported by facts; but the
present section will, I acknowledge, contain many _random thoughts_. I
have, however, thrown them together by themselves, that readers of less
imagination, and who care not to advance beyond