id bodies, even the metallic ones, and that these particles
constitute the most permanent part of the atmosphere, as Sir Isaac
Newton supposed, does not appear to me to be at all probable.

My readers will have observed, that not only is common air liable to be
diminished by a mixture of nitrous air, but likewise air originally
produced from inflammable air, and even from nitrous air itself, which
never contained any fixed air. From this it may be inferred, that the
whole of the diminution of common air by phlogiston is not owing to the
precipitation of fixed air, but from a real contraction of its
dimensions, in consequence of its union with phlogiston. Perhaps an
accurate attention to the specific gravity of air procured from these
different materials, and in these different states, may determine this
matter, and assist us in investigating the nature of phlogiston.

In what _manner_ air is diminished by phlogiston, independent of the
precipitation of any of its constituent parts, is not easy to conceive;
unless air thus diminished be heavier than air not diminished, which I
did not find to be the case. It deserves, however, to be tried with more
attention. That phlogiston should communicate absolute _levity_ to the
bodies with which it is combined, is a supposition that I am not willing
to have recourse to, though it would afford an easy solution of this
difficulty.

I have likewise observed, that a mouse will live almost as long in
inflammable air, when it has been agitated in water, and even before it
has been deprived of all its inflammability, as in common air; and yet
that in this state it is not, perhaps, so much diminished by nitrous air
as common air is. In this case, therefore, the diminution seems to have
been occasioned by a contraction of dimensions, and not by a loss of any
constituent part; so that the air is really better, that is, more fit
for respiration, than, by the test of nitrous air, it would seem to be.

If this be the case (for it is not easy to judge with accuracy by
experiments with small animals) nitrous air will be an accurate test of
the goodness of _common air_ only, that is, air containing a
considerable proportion of fixed air. But this is the most valuable
purpose for which a test of the goodness of air can be wanted. It will
still, indeed, serve for a measure of the goodness of air that does not
contain fixed air; but, a smaller degree of diminution in this case,
must be admitted t