on many occasions. May not a "wet fire" be a way of saying "a molten, viscous mass of fire"? The masses would float and although some might become extinguished, some would probably burn on the surface of the water; also its use at sea would, with a range up to a hundred yards, be quite as easy as on land.
Secondly, "Its composition was such as could be kept secret at Constantinople." If, as Col. Hume says, saltpetre as such was unknown at the time, it was only as a separate kind of salt. It was undoubtedly known, but not distinguished from sea salt or nitrate of soda. Would not this fact render the concealment of the ingredients used more easy?
Thirdly, "It burned with much noise and smoke." Allowing for some slight exaggeration the first condition is fulfilled, as undoubtedly is the latter.
Fourthly, "It was necessarily connected in some way with syphons." As Col. Hume points out, there is ambiguity between the word syphon and tube, and if the latter word meets the facts it seems the more likely rendering.
The writer saw this effect produced during experiments with smoke-producing compositions, and it is probable that the mixture in question was not in the most effective proportions, but so striking was the result that there is little doubt that experiments on such lines would produce a terrible and effective weapon under the conditions of warfare then in existence.
The "Dictionnaire Mobilier Français" gives a diagram of a weapon of a somewhat similar nature stated to have been used by the Arabs in the fifteenth century. The illustration shows what is virtually a Roman candle, and appears plausible until one considers the facts. What is most probable is that the weapon, which was of an incendiary nature, was similar to that described above, which fulfils the requirements
