leaving the rest. To be quite sure of this fact, I agitated a
quantity of common air in boiled water, and when I had reduced it from
eleven ounce measures to seven, I found that it extinguished a candle,
but a mouse lived in it very well. At another time a candle barely went
out when the air was diminished one third, and at other times I have
found this effect lake place at other very different degrees of
diminution.

This difference I attribute to the differences in the state of the water
with respect to the air contained in it; for sometimes it had stood
longer than at other times before I made use of it. I also used
distilled-water, rain-water, and water out of which the air had been
pumped, promiscuously with rain water. I even doubt, not but that, in a
certain state of the water, there might be no sensible difference in
the bulk of the agitated air, and yet at the end of the process it would
extinguish a candle, air being supplied from the water in the place of
that part of the common air which had been absorbed.

It is certainly a little extraordinary that the very same process should
so far mend putrid air, as to reduce it to the standard of air in which
candles have burned out; and yet that it should so far injure common and
wholesome air as to reduce it to about the same standard: but so the
fact certainly is. If air extinguish flame in consequence of its being
previously saturated with phlogiston, it must, in this case, have been
transferred from the water to the air, and it is by no means
inconsistent with this hypothesis to suppose, that, if the air be over
saturated with phlogiston, the water will imbibe it, till it be reduced
to the same proportion that agitation in water would have communicated
to it.

To a quantity of common air, thus diminished by agitation in water, till
it extinguished a candle, I put a plant, but it did not so far restore
it as that a candle would burn in it again; which to me appeared not a
little extraordinary, as it did not seem to be in a worse state than air
in which candles had burned out, and which had never failed to be
restored by the same means.

I had no better success with a quantity of permanent air which I had
collected from my pump-water. Indeed these experiments were begun before
I was acquainted with that property of nitrous air, which makes it so
accurate a measure of the goodness of other kinds of air; and it might
perhaps be rather too late in the year when I