OB' VOLCANIC ROCKS. 65 



portion, where the least resistance was offered, and new fractures would be opened 

 branching off at that depth from the main fissure. The liquid matter would ascend 

 through these secondary fractures, and, if these were of sufficient dimensions, give rise 

 first to the formation of parallel ranges of volcanic rocks by massive eruptions, and 

 then only to a gradual isolation of channels of volcanic action, as in the case first ex- 

 plained ; or, if the fracture consisted of a series of smaller ruptures, cause at once the 

 formation of a series of volcanoes. The activity in these secondary fissures could con- 

 tinue long after any manifestations had ceased over the main fissure, and even after the 

 consolidation of the matter contained in this, with the exception of the volcanic hearths, 

 had proceeded into far greater depth. The formation of secondary fractures, branch- 

 ing off from the main fissure, might be repeated at different depths, and fissures of a 

 third order be formed, branching off from those of the second. Subterranean reser- 

 voirs of liquid matter, which may either be isolated or connected, would thus be formed 

 at different depths, and be arranged after a similar plan below ground, as we notice 

 among the active and extinct volcanic orifices above ground. The hypothesis of 

 the existence of such subterranean seas of melted matter, as they have been called, has 

 also, though in a very different meaning, been arrived at by the adherents of the theory 

 of a metamorphic origin of volcanic rocks. This coincidence increases the degree of its 

 probability. But unless the nature and distribution of those reservoirs is made de- 

 pendent on grander phenomena having connection with the interior of the globe, they 

 will not be capable of explaining the harmony prevailing either in one volcanic region 

 or between all these regions. We have been led, by arguing on our suppositions, to 

 the same conclusion at which we arrived before by induction from observed facts, namely, 

 that the seat of volcanic action must be at a comparatively limited depth. Yet it ap- 

 pears that this depth is in all cases below the shell of sedimentary rocks. Among the 

 reasons supporting this assumption, we mention only one. This relates to the chemical 

 composition of lava. The volcanic is typically the era of andesitic ami basaltic com- 

 pounds. The occurrence of trachyte and rhyolite among the ejected rocks goes to 

 show, that the seat of action receded at certain places from the andesitic regions to 

 those of the compounds corresponding in chemical composition to the two kinds of rock 

 named. But if it had been partially above those regions where matter is still in its 

 primaeval position, then we should expect that there would be volcanoes the lavas of 

 which, being derived from sedimentary rocks, would, as a whole, deviate in composition 

 from the law of Bunsen. No such volcanoes are known, and it is, therefore, not prob- 

 able that the seat of any of those the lava of which has been analyzed is within the 

 shell of sedimentary rocks. It is true that subordinate deviations from the composition 

 as required by theory occur; but it lias been found sufficient to ascribe them, as in the 

 ease of Vesuvius, to a mechanical destruction of the rocks surrounding the channel of 

 ejection by the ascending lava. 



Active volcanoes themselves furnish an illustration in evidence of their own 

 origin as here advocated. It is well known that small cones are frequently met with 

 on the slopes of larger volcanoes. If they occur in Larger number, as on Mount Etna, 

 they are usually situated in lines which radiate from the crater. Each of them is built 

 up of layers of scoria and ashes sloping away from the center, where a crater is im- 



(103) 



