FIG. 150.—The half-submerged volcano of Krakatoa in the Sunda Straits before and after the eruption of 1883 (after Verbeek).] In order to carry our dissecting process down to levels below the base of the volcanic mountain, it is usually necessary to inspect the results of erosion by running water. Here the plug or chimney, instead of being surrounded by tuff, is inclosed by the country rock of the region, which is commonly a sedimentary formation. Such exposed lower sections of volcanic chimneys are numerous along the northwestern shores of the British Isles. Where aligned upon a dislocation or noteworthy fissure in the rocks, the group of plugs has been referred to as a scar or _cicatrice_ (Fig. 151). Associated with the plugs of the cicatrice are not infrequently dikes, or, it may be, sheets of lava extended between layers of sediment and known as _sills_. [Illustration: FIG. 151.—The cicatrice of the Banat (after Suess).] If we are able to continue the dissection process to still greater depths, we encounter at last igneous rock having a texture known as granitic and indicating that the process of consolidation was not only exceedingly slow but also uninterrupted. This rock is found in masses of larger dimensions, and though generally of more or less irregular form, no one dimension is of a different order of magnitude from the others. Such masses are commonly described as _bosses_, or, if especially large, as _batholites_ (Fig. 152). Wherever the rock beds appear as though they had been forced up by the upward pressure of the igneous mass, the latter takes the form of a mushroom and has been described as a _laccolite_ (Figs. 479-481, pp. 441-442). Evidence seems, however, to accumulate that in the greater number of cases the molten rock has fused its way upward, in part assimilating and in part inclosing the rock which it encountered. This process of upward fusion has been likened to the progress of a red hot iron burning its way through a board. =The formation of lava reservoirs.=—The discarding of the earlier notion that the earth has a liquid interior makes it proper in discussing the subject of volcanoes to at least touch upon the origin of the molten rock material. As already pointed out, such reservoirs as exist must be local and temporary, or it would be difficult to see how the existing condition of earth rigidity could be maintained. From the rate at which rock temperatures rise, at increasing depths below the surface, it is clear that all rocks would be melted at very moderate depths only, if they were not kept in a solid state by the prodigious loads which they sustain. Any relief from this load should at once result in fusion of the rock. [Illustration: FIG. 152.—Diagram to illustrate a probable cause of formation of lava reservoirs, and to show the connection between such reservoirs and the volcanoes at the surface.]