produced as the by-product when calcium carbide is decomposed by water occupies considerably more space than the original carbide--usually two or three times as much space, the exact figures depending upon the conditions in which it is formed--and therefore a carbide container cannot advisedly be charged with more than one-third the quantity of solid which it is apparently capable of holding. The remaining two-thirds of the space is naturally full of air when the container is first put into the generator, but the air is displaced by acetylene as soon as gas production begins. Whether that space, however, is occupied by air, by acetylene, or by a gradually growing loose mass of slaked lime, each separate lump of hot carbide is isolated from its neighbours by a material which is also a very bad heat conductor; and the heat has but little opportunity of distributing itself evenly. Moreover, although iron or steel is a notably better conductor of heat than any of the other substances present in the carbide vessel, it is, as a metal, only a poor conductor, being considerably inferior in this respect to copper. If heat dissipation were the only point to be studied in the construction of an acetylene apparatus, far better results might be obtained by the employment of copper for the walls of the carbide container; and possibly in that case a generator of considerable size, fitted with a water- jacketed decomposing vessel, might be free from the trouble of overheating. Nevertheless it will be seen in Chapter VI. that the use of copper is not permissible for such purposes, its advantages as a good conductor of heat being neutralised by its more important defects.