without difficulty, cause it to rise six or seven inches.

I next took an iron wire, (MN, Plate IV. Fig. 16.) properly bent for the
purpose, and making it red hot in the fire, passed it through the
mercury into the receiver, and brought it in contact with the small
piece of phosphorus attached to the tinder. The phosphorus instantly
takes fire, which communicates to the tinder, and from that to the iron.
When the pieces have been properly arranged, the whole iron burns, even
to the last particle, throwing out a white brilliant light similar to
that of Chinese fireworks. The great heat produced by this combustion
melts the iron into round globules of different sizes, most of which
fall into the China cup; but some are thrown out of it, and swim upon
the surface of the mercury. At the beginning of the combustion, there is
a slight augmentation in the volume of the air in the bell-glass, from
the dilatation caused by the heat; but, presently afterwards, a rapid
diminution of the air takes place, and the mercury rises in the glass;
insomuch that, when the quantity of iron is sufficient, and the air
operated upon is very pure, almost the whole air employed is absorbed.

It is proper to remark in this place, that, unless in making experiments
for the purpose of discovery, it is better to be contented with burning
a moderate quantity of iron; for, when this experiment is pushed too
far, so as to absorb much of the air, the cup D, which floats upon the
quicksilver, approaches too near the bottom of the bell-glass; and the
great heat produced, which is followed by a very sudden cooling,
occasioned by the contact of the cold mercury, is apt to break the
glass. In which case, the sudden fall of the column of mercury, which
happens the moment the least flaw is produced in the glass, causes such
a wave, as throws a great part of the quicksilver from the bason. To
avoid this inconvenience, and to ensure success to the experiment, one
gross and a half of iron is sufficient to burn in a bell-glass, which
holds about eight pints of air. The glass ought likewise to be strong,
that it may be able to bear the weight of the column of mercury which it
has to support.

By this experiment, it is not possible to determine, at one time, both
the additional weight acquired by the iron, and the changes which have
taken place in the air. If it is wished to ascertain what additional
weight has been gained by the iron, and the proportion between th