Mr. Coleman deduces from experiment the following inferences, namely: that the ingredients of gunpowder only pulverized and mixed have but a very small explosive force: that gunpowder granulated after having been only a short time on the mill, has acquired only a very small portion of its strength, so that its perfection absolutely depends on very long-continued and accurate mixture and incorporation of the ingredients: that the strength of gunpowder does not depend on granulation, the dust that separates during this process being as strong as the clean grains: that powder undried, is weaker in every step of the manufacture than when dried: and lastly, that charcoal made in iron cylinders in the way already mentioned, makes much stronger powder than common charcoal. This last circumstance is of so much consequence, and is so fully confirmed by experience, that the charges of powder now used for cannon of all kinds have been reduced one-third in quantity, when this kind of powder is employed. In barrelling powder, particular care must be taken to avoid moisture, and this business is also generally reserved for dry weather. When powder is only a little damp, it may be restored to its former goodness merely by stoving; but if it has been thoroughly wetted, the nitre (the only one of the ingredients soluble in water) separates more or less from the sulphur and charcoal, and by again crystallizing, cakes together the powder in whitish masses, which are a loose aggregate of grains covered on the surface with minute efflorescences of nitre. In this case the spoiled powder is put into warm water merely to extract the nitre, and the other two ingredients are separated by straining and thrown away. The specific gravity of gunpowder is estimated by Count Rumford to be about 1.868. The strength and goodness of powder is judged of in several ways; namely, by the colour and feel, by the flame when a small pinch is fired, and by measuring the actual projectile force by the _eprouvette_, and by the distance to which a given weight will project a ball of given dimensions under circumstances in all cases exactly similar. When powder rubbed between the fingers easily breaks down into an impalpable dust, it is a mark of containing too much charcoal, and the same if it readily soils white paper when gently drawn over it. The colour should not be absolutely black, but is preferred to be more of a dark blue with a little cast of red. The trial by firing is thus managed; lay two or three small heaps of about a dram each on clean writing paper, about three or four inches asunder, and fire one of them by a red-hot iron wire: if the flame ascends quickly with a good report, sending up a ring of white smoke, leaving the paper free from white specks and not burnt into holes, and if no sparks fly off from it, setting fire to the contiguous heaps, the powder is judged to be very good, but if otherwise, either the ingredients are badly mixed, or impure.