according to equation (2), and this action will again take place without delay; but unless the temperature of the residual carbide has been kept well below 400° C., a further evolution of gas will occur which will not cease for an indeterminate period of time, and which, by strict theory, given the necessary conditions, might continue until a second volume of acetylene equal to that liberated at first had been set free. In practice this phenomenon of a secondary production of gas, which is known as "after-generation," is regularly met with in all generators where the carbide is in excess of the water added; but the amount of acetylene so evolved rarely exceeds one-quarter or one-third of the main make. The actual amount evolved and the rate of evolution depend, not merely upon the quantity of undecomposed carbide still remaining in contact with the damp lime, but also upon the rapidity with which carbide naturally decomposes in presence of liquid water, and the size of the lumps. Where "after-generation" is caused by the ascent of water vapour round lumps of carbide, the volume of gas produced in a given interval of time is largely governed by the temperature prevailing and the shape of the apparatus. It is evident that even copious "after-generation" is a matter of no consequence in any generator provided with a holder to store the gas, assuming that by some trustworthy device the addition of water is stopped by the time that the holder is two-thirds or three-quarters full. In the absence of a holder, or if the holder fitted is too small to serve its proper purpose, "aftergeneration" is extremely troublesome and sometimes dangerous, but a full discussion of this subject must be postponed to the next chapter.