for a considerable time
after the air is removed from it. This I have often observed with
respect to iron.

Inflammable air, made by a violent effervescence, I have observed to be
much more inflammable than that which is made by a weak effervescence,
whether the water or the oil of vitriol prevailed in the mixture. Also
the offensive smell was much stronger in the former case than in the
latter. The greater degree of inflammability appeared by the greater
number of successive explosions, when a candle was presented to the neck
of a phial filled with it.[4] It is possible, however, that this
diminution of inflammability may, in some measure, arise from the air
continuing so much longer in the bladder when it is made very slowly;
though I think the difference is too great for this cause to have
produced the whole of it. It may, perhaps, deserve to be tried by a
different process, without a bladder.

Inflammable air is not thought to be miscible with water, and when kept
many months, seems, in general, to be as inflammable as ever. Indeed,
when it is extracted from vegetable or animal substances, a part of it
will be imbibed by the water in which it stands; but it may be presumed,
that in this case, there was a mixture of fixed air extracted from the
substance along with it. I have indisputable evidence, however, that
inflammable air, standing long in water, has actually lost all its
inflammability, and even come to extinguish flame much more than that
air in which candles have burned out. After this change it appears to be
greatly diminished in quantity, and it still continues to kill animals
the moment they are put into it.

This very remarkable fact first occurred to my observation on the
twenty-fifth of May 1771, when I was examining a quantity of inflammable
air, which had been made from zinc, near three years before. Upon this,
I immediately set by a common quart-bottle filled with inflammable air
from iron, and another equal quantity from zinc; and examining them in
the beginning of December following, that from the iron was reduced near
one half in quantity, if I be not greatly mistaken; for I found the
bottle half full of water, and I am pretty clear that it was full of air
when it was set by. That which had been produced from zinc was not
altered, and filled the bottle as at first.

Another instance of this kind occurred to my observation on the 19th of
June 1772, when a quantity of air, half of which had bee