much discussed by the young men about the works, and they
frequently had occasion to refer to Foureroy's well-known book for the
purpose of determining various questions of difference which arose
among them in the course of their inquiries. The book was, however, in
many respects indecisive and unsatisfactory; and, in 1793, when a
reduction took place in the Company's staff, and David Mushet was left
nearly the sole occupant of the office, he determined to study the
subject for himself experimentally, and in the first place to acquire a
thorough knowledge of assaying, as the true key to the whole art of
iron-making.

He first set up his crucible upon the bridge of the reverberatory
furnace used for melting pig-iron, and filled it with a mixture
carefully compounded according to the formula of the books; but,
notwithstanding the shelter of a brick, placed before it to break the
action of the flame, the crucible generally split in two, and not
unfrequently melted and disappeared altogether. To obtain better
results if possible, he next had recourse to the ordinary smith's fire,
carrying on his experiments in the evenings after office-hours. He set
his crucible upon the fire on a piece of fire brick, opposite the
nozzle of the bellows; covering the whole with coke, and then exciting
the flame by blowing. This mode of operating produced somewhat better
results, but still neither the iron nor the cinder obtained resembled
the pig or scoria of the blast-furnace, which it was his ambition to
imitate. From the irregularity of the results, and the frequent
failure of the crucibles, he came to the conclusion that either his
furnace, or his mode of fluxing, was in fault, and he looked about him
for a more convenient means of pursuing his experiments. A small
square furnace had been erected in the works for the purpose of heating
the rivets used for the repair of steam-engine boilers; the furnace had
for its chimney a cast-iron pipe six or seven inches in diameter and
nine feet long. After a few trials with it, he raised the heat to such
an extent that the lower end of the pipe was melted off, without
producing any very satisfactory results on the experimental crucible,
and his operations were again brought to a standstill. A chimney of
brick having been substituted for the cast-iron pipe, he was, however,
enabled to proceed with his trials.

He continued to pursue his experiments in assaying for about two years,
during w