place at a temperature of 1,000° C. (1,800° Fahr.), approximately that of lava, and under a volume equivalent to that of the water in the liquid state whence the vapor is derived. Under these conditions, we must suppose the vaporization to be total, for the critical temperature, above which the liquefaction of vapor can not be realized, is, according to M. Clausius, 332° C. (629° F.). The pressure, of which it is also possible to make an approximate estimate, then becomes comparable to that of the most powerfully explosive gases, and is, consequently, capable of producing very considerable dynamic effects. These effects may also be produced at a much lower temperature than that of lavas at 500° C. (900° F.); for example, if we suppose that the volume imposed upon the vapor is so limited as to correspond to a density of 0·8 or 0·9. No doubt such conditions are realized in the lower regions of the globe, where water is confined within limited spaces, and as hot as the melted rocks which we see gushing out from the surface at a temperature of 1,000° C. (1,800° F.) or more. We shall see, however, that such depths and such a temperature are not necessary.
The vapor of water when superheated acquires a power of which the most terrible boiler-explosions could give no idea if we had not the result before our eyes. The tubes of the best quality of iron that I used in observing the action of superheated water in the formation of silicates had an inside diameter of twenty-one millimetres and were eleven millimetres thick. They sometimes exploded, and were projected into the air with a noise like that of the firing of a cannon. Before bursting, the tubes swelled out into bulbous forms, and rents were opened in the middle of the bulbs. If the iron had no flaws and according to the estimate that it would preserve up to 450° C. (810° F.), the temperature to which it was raised, the same tenacity it had when cold, such rents must have indicated a pressure of several thousand atmospheres. A few cubic centimetres of water were sufficient to produce an effect like that; and, considering the small dimensions of the inside of the tubes as compared with the volume of the water, the vapor must have reached a density of about 0·9. If we apply the data we possess to the depths of the globe, it is not difficult to conceive very simple dispositions in which the vapor of water, under the conditions we have just determined, will suddenly provoke shocks or series of shocks that will too often make themselves felt on the surface. Whatever conception we may form of the volcanic reservoirs, we must admit it to be very probable that solutions of continuity exist between the soft or fluid masses in fusion and the solid masses superposed over them. Moreover, cavities may also exist in the solid rocks themselves that lie over the soft masses. On the other hand, the incessant losses, which these internal reservoirs suffer in consequence of the enormous volumes of water in the condition of vapor which they disengage every day, are probably repaired by supplies from the surface.