not so satisfactory. It is, however, not necessary to adopt this method in practice, because the solvent power of the liquid in the seal can be reduced by adding to it a saline body which simultaneously lowers its freezing-point and makes the apparatus more trustworthy in winter. FREEZING OF GASHOLDER SEAL.--The danger attendant upon the congelation of the seal in an acetylene holder is very real, not so much because of the fear that the apparatus may be burst, which is hardly to be expected, as because the bell will be firmly fixed in a certain position by the ice, and the whole establishment lighted by the gas will be left in darkness. In these circumstances, hurried and perhaps injudicious attempts may be made to thaw the seal by putting red-hot bars into it or by lighting fires under it, or the generator-house may be thoughtlessly entered with a naked light at a time when the apparatus is possibly in disorder through the loss of storage-room for the gas it is evolving. Should a seal ever freeze, it must be thawed only by the application of boiling water; and the plant-house must be entered, if daylight has passed, in perfect darkness or with the assistance of an outside lamp whining through a closed window. [Footnote: By "closed window" is to be understood one incapable of being opened, fitted with one or two thicknesses of stout glass well puttied in, and placed in a wall of the house as far as possible from the door.] There are two ways of preventing the seal from freezing. In all large installations the generator-house will be fitted with a warm-water heating apparatus to protect the portion of the plant where the carbide is decomposed, and if the holder is also inside the same building it will naturally be safe. If it is outside, one of the flow-pipes from the warming apparatus should be led into and round the lowest part of the seal, care being taken to watch for, or to provide automatic arrangements for making good, loss of water by evaporation. If the holder is at a distance from the generator-house, or if for any other reason it cannot easily be brought into the warming circuit, the seal can be protected in another way; for unlike the water in the generator, the water in the holder-seal will perform its functions equally well however much it be reduced in temperature, always providing it is maintained in the liquid condition. There are numerous substances which dissolve in, or mix with, water, and yield solutions or liquids that do not solidify until their temperature falls far below that of the natural freezing- point. Assuming that those substances in solution do not attack the acetylene, nor the metal of which the holder is built, and are not too expensive, choice may be made between them at will. Strictly speaking the cost of using them is small, because unless the tank is leaky they last indefinitely, not evaporating with the water as it is vaporised into the gas or into the air. The water-seal of a holder standing within the