which comes to the same thing. Putting the matter in another shape, it may be said that the reaction between calcium carbide and water is exothermic, evolving either 14.0 or 29.1 calories according as the byproduct is calcium oxide or solid calcium hydroxide; and therefore either reaction proceeds without external assistance in the cold. The reaction between carbide and slaked lime, however, is endothermic, absorbing 1.1 calories; and therefore it requires external assistance (presence of an elevated temperature) to start it, or continuous introduction of heat (as from the reaction between the rest of the carbide present and the water) to cause it to proceed. Of itself, and were it not for the disadvantages attending the production of a temperature remotely approaching 400° C. in an acetylene generator, which disadvantages will be explained in the following paragraphs, there is no particular reason why reaction (3) should not be permitted to occur, for it involves (theoretically) no loss of acetylene, and no waste of calcium carbide. Only one specific feature of the reaction has to be remembered, and due practical allowance made for it. The reaction represented by equation (2) proceeds almost instantaneously when the calcium carbide is of ordinarily good quality, and the acetylene resulting therefrom is wholly generated within a very few minutes. Equation (3), on the contrary, consumes much time for its completion, and the gas corresponding with it is evolved at a gradually diminishing speed which may cause the reaction to continue for hours--a circumstance that may be highly inconvenient or quite immaterial according to the design of the apparatus. When, however, it is desired to construct an automatic acetylene generator, _i.e._, an apparatus in which the quantity of gas liberated has to be controlled to suit the requirements of any indefinite number of burners in use on different occasions, equation (3) becomes a very important factor in the case. To determine the normal reaction (No. 2) of an acetylene generator, 64 parts by weight of calcium carbide must react with 36 parts of water to yield 26 parts by weight of acetylene, and apparently both carbide and water are entirely consumed; but if opportunity is given for the occurrence of reaction (3), another 64 parts by weight of carbide may be attacked, without the addition of any more water, producing, inevitably, another 26 parts of acetylene. If, then, water is in chemical excess in the generator, all the calcium carbide present will be decomposed according to equation (2), and the action will take place without delay; after a few minutes' interval the whole of the acetylene capable of liberation will have been evolved, and nothing further can possibly happen until another charge of carbide is inserted in the apparatus. If, on the other hand, calcium carbide is in chemical excess in the generator, all the water run in will be consumed