retorts, into
which they were pushed; the covers of which were then luted with clay
and closely applied. The bottoms of the retorts being perforated,
permitted the escape of the vapors and gases into the furnaces beneath,
where inflaming, they supplied mainly the heat required in the
operation. In about two hours the slip cylinders were withdrawn from the
retorts and moved by the cranes over, and lowered into the cast iron
coolers beneath the floor; these had water from the canal circulating
around them; the covers being then put on to exclude the air, the mass
of charcoal was rapidly cooled. As soon as a slip cylinder was removed
from a retort a freshly charged one would take its place, and thus the
process was continued. The slip cylinders were taken out of the coolers
in succession by the cranes, and swung over a long and broad table upon
which their contents were dropped; here the sticks of charcoal were
separately examined and the imperfect rejected. The charcoal was then
placed in pulverizing barrels with bronze balls, which revolving by
machinery, soon reduced it more or less to a fine powder; it was then
bolted, and with the sulphur and saltpetre taken to the weighing house.
Here the three materials were arranged into sixty pounds charges, by
mingling forty-five pounds of saltpetre, nine pounds of charcoal and six
pounds of sulphur, which was then moistened and ready for incorporation.

Reflecting over the processes for making gunpowder, it suggested itself
that the chemical reactions would necessarily have the most favorable
conditions, when there should be the most intimate approximation of the
component molecules. That, as the charcoal by its combustion with the
oxygen of the saltpetre, supplied the expanded gases which produced the
explosive force, it was of the first consideration that there should be
the most perfect mixture practicable between these two ingredients.
Under the microscope a fine particle of charcoal was seen to be a mass
of carbon penetrated by numerous pores, hence it became necessary to
completely fill these minute pores with the saltpetre to have the best
condition. This might be accomplished by the usual processes, as the
charge is kept moistened when stamped or rolled, but as it will not
answer to have the mass wet during the incorporating operation, only
moist or damp, the completion of the process was necessarily delayed. If
this mass of material could be made into a semi-liquid condi