proved that such intimate previous union is necessary to precede the
chemical reaction causing explosion; indeed, some explosions in powder
works, before the mixture of the materials, or just at its commencement,
seem to point to the contrary. It is also certain that in the
manufacture of gunpowder the usual nitrate of potassium (saltpeter) can
be replaced by the nitrates of soda, baryta, and ammonia, also by the
chloride of potassium; charcoal by sawdust, tan, resin, and starch; and
though a substitute for sulphur is not easily found, the latter, or a
similar substance, is not an absolute necessity in the composition of
gunpowder.[2]

[Footnote 1: Encyclopaedia Britannica, new edition, viii, p. 806.]

[Footnote 2: _Vide_ Abel's Experiments in Gunpowder, as detailed in
Phil. Trans. Eoy. Soc, 1874.--_Vide_ also _Bull. Soc. d'Encouragement_,
Nov., 1880, p. 633, _Sur les Explosives_.]

The generally received theory of the chemical action which makes
gunpowder explosive is that it is due to the superior affinity of the
oxygen of the niter (KNO_3) for the carbon of the charcoal, and the
production of carbonic acid gas (CO_2) and carbonic oxide (CO) suddenly
and in great volume. The latter extinguishes flame as well as the
former, unless its own flammability is supported by the oxygen of the
atmosphere until the degree of oxygenation CO_2 is reached. Considering
that water (H_2O) is composed of two volumes of hydrogen and one of
oxygen, and that under an enormously high temperature and the excessive
affinity of oxygen gas for potassium or sodium (freed from nitrate
union), dissociation of the water may be possible, aided by its being in
the form of spray and steam, we would hesitate to deny that an explosive
union of suitable crude salts could occur during the burning of a