avoured the
former hypothesis. I found, however, that when I admitted a much
smaller quantity of water, confined in a narrow glass tube, a part only
of the air disappeared, and that very slowly, and that more of it
vanished upon the admission of more water. This observation put it
beyond a doubt, that this air was properly _imbibed_ by the water,
which, being once fully saturated with it, was not capable of receiving
any more.

The water thus impregnated tasted very acid, even when it was much
diluted with other water, through which the tube containing it was
drawn. It even dissolved iron very fast, and generated inflammable air.
This last observation, together with another which immediately follows,
led me to the discovery of the true nature of this remarkable kind of
air.

Happening, at one time, to use a good deal of copper and a small
quantity of spirit of salt, in the generation of this kind of air, I was
surprized to find that air was produced long after, I could not but
think that the acid must have been saturated with the metal; and I also
found that the proportion of inflammable air to that which was absorbed
by the water continually diminished, till, instead of being one fourth
of the whole, as I had first observed, it was not so much as one
twentieth. Upon this, I concluded that this subtle air did not arise
from the copper, but from the spirit of salt; and presently making the
experiment with the acid only, without any copper, or metal of any kind,
this air was immediately produced in as great plenty as before; so that
this remarkable kind of air is, in fact, nothing more than the vapour,
or fumes of spirit of salt, which appear to be of such a nature, that
they are not liable to be condensed by cold, like the vapour of water,
and other fluids, and therefore may be very properly called an _acid
air_, or more restrictively, the _marine acid air_.

This elastic acid vapour, or acid air, extinguishes flame, and is much
heavier than common air; but how much heavier, will not be easy to
ascertain. A cylindrical glass vessel, about three fourths of an inch in
diameter, and four inches deep, being filled with it, and turned upside
down, a lighted candle may be let down into it more than twenty times
before it will burn at the bottom. It is pleasing to observe the colour
of the flame in this experiment; for both before the candle goes out,
and also when it is first lighted again, it burns with a beautiful
green,