ance at the surface, and in fact may have parted company with the lava at very considerable depths.
Or the latter may have been forced up almost simultaneously with the vapor, and poured out over the edge of the primitive cone. This, however, is not the general rule, for an escape of much gaseous material nearly always precedes for a variable period the appearance of the lava. In fact, when a volcanic outburst has forced a convenient passage for the vapor, the exit of liquid rock seems of secondary importance, for generally the terrific explosions, earthquakes, and subterranean thunder that accompany the first stage of eruption are more or less absent, or at least much diminished during the welling up of the fluid rock. If, as in the latter case, a cone of some considerable size has been formed, the lava will rise and occupy the whole of the crater-cavity. Two things may happen: If the cone which now forms, as it were, an embankment around the lava is of sufficient strength to withstand the pressure of the fluid mass contained within it and the continual explosive vibrations, the liquid rock pours out over the edge of the crater down the side of the cone, and may continue its course for variable distances from its starting-point; or if, on the other hand, the cone is too weak to support the strain, it may break away and give free passage to the lava through the breach. This condition is well illustrated in many of the Puys of central France. There is another series of events, that is to say, the formation of dikes, about which we shall have more to say anon.
The lava may form a series of little streams over the cone sides, changing their situation according to the point at which the crater is lowest. Here it will cool, forming a buttress of rock on the slopes of the cone. These masses will be covered again by lapilli, other buttresses formed in the same manner, and thus the cone built up higher and made stronger. If we see it in section, as in the diagram, it will present a stratification of alternate beds of rock and cinders. This, however, is misleading. The lava-streams do not form a continuous sheet surrounding the cone—see diagram, where they are seen cut through in transverse section. When a mountain of some height has been formed, it then becomes liable to fracturing, and the formation of so-called volcanic dikes. Mr. Mallet, in a communication to the Geological Society,[1] thoroughly explained this condition of things. As we have seen, the cone may form an embankment around the column of lava occupying the chimney and crater, consequently there is an enormous pressure put upon the supporting wall of loose material. Let us begin by taking the pressure of a column of water thirty-two feet high, then let us say another four thousand feet, roughly the altitude of Vesuvius, and compare that with a column of molten lava, whose specific gravity is two or three times that of water. This would be an interesting calculation: given the specific gravity of
- ↑ "Proc. Geol. Soc," London, vol. xxxii, part iv, p. 478.
