er cork closed, by pressing the bladder against it, it
may be carried to any place, and if the tube be carefully wiped, the air
may be conveyed quite free from moisture through a body of quicksilver,
or any thing else. A little practice will make this very useful
manoeuvre perfectly easy and accurate.

In order to impregnate fluids with any kind of air, as water with fixed
air, I fill a phial with the fluid larger or less as I have occasion (as
_a_ fig. 10;) and then inverting it, place it with its mouth downwards,
in a bowl _b_, containing a quantity of the same fluid; and having
filled the bladder, fig. 9, with the air, I throw as much of it as I
think proper into the phial, in the manner described above. To
accelerate the impregnation, I lay my hand on the top of the phial, and
shake it as much as I think proper.

If, without having any air previously generated, I would convey it into
the fluid immediately as it arises from the proper materials, I keep the
same bladder in connection with a phial _c_ fig. 10, containing the same
materials (as chalk, salt of tartar, or pearl ashes in diluted oil of
vitriol, for the generation of fixed air) and taking care, lest, in the
act of effervescence, any of the materials in the phial _c_ should get
into the vessel _a_, to place this phial on a stand lower than that on
which the bason was placed, I press out the newly generated air, and
make it ascend directly into the fluid. For this purpose, and that I may
more conveniently shake the phial _c_, which is necessary in some
processes, especially with chalk and oil of vitriol, I sometimes make
use of a flexible leathern tube _d_, and sometimes only a glass tube.
For if the bladder be of a sufficient length, it will give room for the
agitation of the phial; or if not, it is easy to connect two bladders
together by means of a perforated cork, to which they may both be
fastened.

When I want to try whether any kind of air will admit a candle to burn
in it, I make use of a cylindrical glass vessel, fig. 11. and a bit of
wax candle _a_ fig. 12, fastened to the end of a wire _b_, and turned
up, in such a manner as to be let down into the vessel with the flame
upwards. The vessel should be kept carefully covered till the moment
that the candle is admitted. In this manner I have frequently
extinguished a candle more than twenty times successively, in a vessel
of this kind, though it is impossible to dip the candle into it without
giving