hour, maximum thermometers placed immediately over the carbide in the different cells gave readings of from 70° to 90° C., the average maximum temperature being about 80° C. Hence the Austrian code of rules issued in 1905 governing the construction of acetylene apparatus contains a clause to the effect that the temperature in the gas space of a generator must never exceed 80° C.; whereas the corresponding Italian code contains a similar stipulation, but quotes the maximum temperature as 100° C. (_vide_ Chapter IV.). It is now necessary to see why the production of an excessively high temperature in an acetylene generator has to be avoided. It must be avoided, because whenever the temperature in the immediate neighbourhood of a mass of calcium carbide which is evolving acetylene under the attack of water rises materially above the boiling-point of water, one or more of three several objectionable effects is produced--(_a_) upon the gas generated, (_b_) upon the carbide decomposed, and (_c_) upon the general chemical reaction taking place. It has been stated above that in moat generators when the action between the carbide and the water is proceeding smoothly, it occurs according to equation (2)-- (2) CaC_2 + 2H_2O = C_2H_2 + Ca(OH)_2 rather than in accordance with equation (1)-- (1) CaC_2 + H_2O = C_2H_2 + CaO. This is because calcium oxide, or quicklime, the by-product in (1), has considerable affinity for water, evolving a noteworthy quantity of heat when it combines with one molecule of water to form one molecule of calcium hydroxide, or slaked lime, the by-product in (2). If, then, a small amount of water is added to a large amount of calcium carbide, the corresponding quantity of acetylene may be liberated on the lines of equation (1), and there will remain behind a mixture of unaltered calcium carbide, together with a certain amount of calcium oxide. Inasmuch as both these substances possess an affinity for water (setting heat free when they combine with it), when a further limited amount of water is introduced into the mixture some of it will probably be attracted to the oxide instead of to the carbide present. It is well known that at ordinary temperatures quicklime absorbs moisture, or combines with water, to produce slaked lime; but it is equally well known that in a furnace, at about a red heat, slaked lime gives up water and changes into quicklime. The reaction, in fact, between calcium oxide and water is reversible, and whether those substances combine or dissociate is simply a question of temperature. In other words, as the temperature rises, the heat of hydration of calcium oxide diminishes, and calcium hydroxide becomes constantly a less stable material. If now it should happen that the affinity between calcium carbide and water should not diminish, or should diminish in a lower ratio than the affinity between calcium oxide and water as the temperature of the mass rises from one cause or other,