of the metal, and this air should be made
to pass through a pipe of clay kept continually red hot by the fire of
the furnace. By this means the inside of the muffle will never be
coolled, and processes will be finished in a few minutes, which at
present require a considerable space of time.

Mr Sage remedies these inconveniencies in a different manner; he places
the cuppel containing lead, alloyed with gold or silver, amongst the
charcoal of an ordinary furnace, and covered by a small porcelain
muffle; when the whole is sufficiently heated, he directs the blast of a
common pair of hand-bellows upon the surface of the metal, and completes
the cuppellation in this way with great ease and exactness.

SECT. III.

_Of increasing the Action of Fire, by using Oxygen Gas instead of
Atmospheric Air._

By means of large burning glasses, such as those of Tchirnausen and Mr
de Trudaine, a degree of heat is obtained somewhat greater than has
hitherto been produced in chemical furnaces, or even in the ovens of
furnaces used for baking hard porcelain. But these instruments are
extremely expensive, and do not even produce heat sufficient to melt
crude platina; so that their advantages are by no means sufficient to
compensate for the difficulty of procuring, and even of using them.
Concave mirrors produce somewhat more effect than burning glasses of the
same diameter, as is proved by the experiments of Messrs Macquer and
Beaume with the speculum of the Abbe Bouriot; but, as the direction of
the reflected rays is necessarily from below upwards, the substance to
be operated upon must be placed in the air without any support, which
renders most chemical experiments impossible to be performed with this
instrument.

For these reasons, I first endeavoured to employ oxygen gas for
combustion, by filling large bladders with it, and making it pass
through a tube capable of being shut by a stop-cock; and in this way I
succeeded in causing it to support the combustion of lighted charcoal.
The intensity of the heat produced, even in my first attempt, was so
great as readily to melt a small quantity of crude platina. To the
success of this attempt is owing the idea of the gazometer, described p.
308. _et seq._ which I substituted instead of the bladders; and, as we
can give the oxygen gas any necessary degree of pressure, we can with
this instrument keep up a continued stream, and give it even a very
considerable force.

The only apparatus