The defect of the book before us is its inconclusiveness: from the preface we are led to expect the solution of a theorem; after reading the book through, we find ourselves not indeed as far at sea as ever, but aided mainly by negations. The actual origin of gunpowder or Greek fire is not traced; many of the connecting links in the chain of pyrotechnic discovery are still deficient; and the conjectures, which stand in the place of conclusions, are frequently founded upon what appear to us insufficient data. On the other hand it must be admitted, that on a subject so involved in obscurity, inasmuch as proof is impossible, speculation is to a certain extent admissible as a link to render isolated facts intelligible. It may be well here if, before passing to the more immediate object of this paper--viz. a sketch of the probable progress of pyrotechny--we explain to those of our readers who are unacquainted with chemistry, the philosophy of explosive combustibles. Combustion is nothing else than rapid chemical union, taking place between two dissimilar substances, which have what is called an affinity for each other, _i.e._ a tendency to unite and form a new compound. When a candle or lamp is burned, it is carbon and hydrogen, the principal constituents of oil or fat, which combine with oxygen, one main ingredient of the atmosphere. As it requires a certain temperature for this union to take place, to prevent the cooling effect of mass, a wick is used which can be readily heated, and where, as soon as chemical action has once taken place, other portions of the oil or melted tallow are absorbed, which ascend just as water through the pores of a sponge, and supply the place of those burned. In this example, only a small ignited surface is exposed to the influence of the oxygen: if, however, this latter element could be obtained in a solid state, and mixed up with the combustible, each particle throughout the whole mass would have in contact with it a particle of oxygen; so that, if the whole were raised to the necessary temperature for combustion, combustion would be instantaneous--or if the temperature of a part were sufficiently elevated, the combustion of this portion would communicate an intense heat to the contiguous portions, and the whole would rapidly kindle as a fuse does. In this case also, the access of the air being immaterial, combustion might take place in a closed vessel, or even under water.