[In the diagram, a series of light and heavy beds are represented in section by open and shaded spaces. A lava stream free to move upward or laterally will intrude itself at some point (_c_) so placed that every combination of superior beds (_a_), which includes the lowest, shall have a less average density; and every combination of inferior strata (_b_), which includes the highest, shall have a greater average density than that of the lava.] The first case is that of a volcano; the second is that of a laccolite; and the third is the general case, including the others and applying to all volcanoes and laccolites. Conversely we may say that, given a series of strata of diverse and alternating density, a very light lava will traverse it to the top and be extruded; a heavier will intrude itself at some lower level; and a series of dissimilar lavas may select an equal number of distinct levels. It is easy to imagine such a balancing of conditions that a slight change in a lava will determine a great change in its level of intrusion. Having seen the general application of the hydrostatic law, it is time to recall the condition which we laid aside at the start. Cohesion, or rigidity, is never absent and must affect every phase of vulcanism. It certainly opposes the free circulation of lavas, and it cannot but modify their obedience to the hydrostatic law. But granting this, and believing that a full comprehension of the subject must include this condition, I am at a loss to tell in what way it influences the selection by a lava flood of a subaërial or a subterranean bourne. Whether it will on the whole oppose upward progress more than lateral, or _vice versa_, is not clear. If it resists lateral intrusion the more strongly, it favors the formation of volcanoes; if it resists upward penetration the more strongly, it favors the formation of laccolites; and in either case the working of the hydrostatic law is modified. But in neither case is the working of the law more than modified. The law is not abrogated, and in obedience to it light lavas still _tend_ to rise higher than heavy however much the rising of all lavas may be hindered or favored. In brief, since lavas are fluids they are subject to the law of fluid equilibrium, and their behavior is conditioned by the relations of their densities to the densities of the solids which they penetrate; and since the latter solids are rigid and coherent, it is further conditioned by the resistance which is opposed to their penetration. When the resistance to penetration is the same in all directions, the relation of densities determines the stopping place of the rising lava; but when the vertical and lateral resistances are unequal, _their_ relation may be the determining condition. If we can decide whether the determinative condition in the Plateau region was that of densities or that of penetrability, we shall have solved our problem.