increase in the flow of the lava twice daily at intervals of about twelve hours' duration, and with a little less than an hour's retardation from one day to the next, as observed in the ocean-tides. The eruption commenced on the 1st of May, and the periodic swelling of the lava-stream was observed from the 5th to the 19th. Such observations could easily be made in the island of Hawaii, on the borders of the lava-lake of Kilauea.
It must not be forgotten that these subterranean tides do not by any means demonstrate the liquidity of the nucleus, but simply the existence of a liquid mass of a certain depth. It will be shown how astronomic phenomena furnish data for the solution of this question, but we will first glance at some purely physical considerations that have been advanced in elucidation of the matter.
Mr. James Thompson was the first to point out that compression would have the effect of lowering the point of fusion, and consequently of retarding the congelation of those liquids that expand in solidifying. This has been shown to be the case with water, and it would probably be found to be the same with iron. On the other hand, in the case of the far more numerous substances that contract in solidifying, compression facilitates congelation by cooling the mass. It ought, therefore, to raise the point of fusion, and this is known to occur with many bodies. Thus the melting-point of sulphur, which shrinks materially in solidifying, is raised from 107° to 140° under a pressure of eight hundred atmospheres. Now, according to the experiments of Bischof, the greater part of the rocks expand by fusion and contract in solidifying. Granite, the schists, and trachyte shrink about one fifth in solidifying. This tends to confirm the supposition, says Sir W. Thomson, that the earth's nucleus has long been solidified.
Let us conceive the earth as primarily wholly liquid. There would be established in the mass an equilibrium of temperature corresponding to a given pressure. As the mass cooled, solidification would commence, either on the surface or at the center. The question is a very complex one and can not be fully solved without a better knowledge than we have of the properties of the liquid under consideration. But assuming that solidification commences at the surface, a thin skin will first be formed, and this being by the hypothesis heavier than the liquid it covers—since its volume shrinks in solidifying—it follows that it must be broken up and the fragments be carried to the bottom or center, forming there a solid nucleus. Thus, in any event, solidification will occur at the center, and, when the entire mass has acquired a temperature near the point of solidification, a solid carapace will gradually cover the surface, beneath which here and there masses of liquid will still exist.
This argument, however, is open to question in several respects. M. Mallet's experiments, for instance, with the scoriæ of smelting furnaces, show that certain silicates contract only six per cent. The
