OF VOLCANIC ROCKS. 63 



nection with any particular volcanic vent. It is no less obvious from the fact that two 

 neighboring volcanoes may not only eject, contemporaneously, different kinds of lava, 

 but also be, to some extent, independent of each other in their manifestations. 



We may base further conclusions in regard to the origin of volcanoes upon the 

 following premises, which we repeat from the foregoing pages : First, several facts 

 appear to indicate that the source of volcanic action is at a comparatively limited 

 depth ; second, all volcanoes, whether active or extinct, are intimately connected with 

 massive eruptions ; third, this connection is of such a character as to establish, in the 

 majority of cases, the close similarity and chemical identity of the mineral matter 

 ejected by the volcano in its first epoch of activity with that of neighboring hills or 

 of its own foundation, which had been accumulated by massive eruptions ; while in 

 other cases these neighboring hills or the foundation of the volcano are composed of 

 volcanic rocks not identical with the lava, and then the latter will belong lithologically to 

 a kind of rock which, according to the order of succession of massive eruptions, would be 

 of a more recent origin than the former ; that is, trachytic or rhyolitic volcanoes were 

 frequently opened at those places where only andesite had been accumulated before, 

 and basalt where either andesite or trachyte had preceded ; but the reversed order 

 appears not to occur, no basaltic volcano having been succeeded, in its own neighbor- 

 hood, by massive eruptions of trachyte or andesite ; fourth, many of those volcanoes 

 which have been active through a long period, have undergone a periodical change 

 in regard to the character of the mineral matter ejected by them, and this change is 

 in general (though with exceptions) conformable to the order of succession observed 

 in regard to massive eruptions. 



From these facts may be inferred the complete dependency of volcanoes upon 

 massive eruptions. The latter, as we attempted to show, were due, firstly, to the open- 

 ing of systems of fissures which extended throughout the solid crust ; and secondly, to a 

 quiet outflow, which was caused by the expansion attending the change of aggregation 

 of solid or highly viscous matter around the lowest part of the fissures into that of 

 aqueous fusion. The relations of volcanic activity to massive eruptions are indicative 

 of a process by which the elongated and extensive vents of the latter were gradually 

 differentiated into isolated and narrow channels feeding isolated orifices on the surface. 

 It has been observed that cinder eruptions mark generally the last stage of volcanic 

 action. We may go a step further back, and say that volcanic action is the last stage 

 of massive eruptions. 



In order to arrive at a conception of the manner in which the change from one 

 mode of action to the other could be effected, let us suppose that a main fissure was 

 filled with matter from below, and mountains of volcanic rocks accumulated above it 

 by the long continued overflow. Solidification would at once set in, and proceed 

 downward whenever a cessation of the extrusive action occurred, independent of 

 the question whether it would not simultaneously proceed upwards from the depth. 

 Its progress would not be equal in all parts of the fissure, since this must be 

 wider at some places, and more contracted at others. In this condition we should 

 have one of the causes for the isolation of centers of action, for the length of time 

 during which different portions of the matter would remain in a fluid condition, must, of 



(101) 



