OF VOLCANIC ROCKS. 67 



The foregoing considerations may explain what appears to have been the most 

 probable way in which volcanic activity was developed at and near the places of mass- 

 ive eruptions. They apply without difficulty to those cases where volcanoes have un- 

 dergone no change in regard to the character of their lava. We have still to consider 

 those more intricate cases where the latter has periodically changed. There are volca- 

 noes which exhibit a regular succession of andesitic, trachytic, and rhyolitic lavas, fol- 

 lowed in a later period by the outpouring of basaltic lava, either through the same, or 

 through other vents in the immediate vicinity. There are others in which only a part 

 of this series can be observed, such as the succession of basalt to trachyte or rhyolite, 

 which appears to be of the most frequent occurrence, or of rhyolite only to andesite. 

 As the order of succession is generally the same as that exhibited by massive eruptions, 

 it would appear that it must be due to the same causes in both cases. If we take it 

 for granted, that those extensive reservoirs of melted matter, from which was either 

 continued its quiet protrusion, or volcanoes were fed, have had their seat in those con- 

 centric layers of the crust composed of silicious masses which correspond to trachyte 

 or rhyolite in composition, then there is little difficulty in explaining the protrusion 

 of portions of them. For there must be a limit to the expansion of a substance such 

 as a given mass of andesite, by aqueous fusion, and thereby a limit to its ejection. 

 The eruptive activity might then either come to rest or continue. As it is very im- 

 probable that water takes originally a part in the composition of those masses below 

 the shell of sediments which have crystallized from a molten state, its access to them 

 at places contiguous to a source of heat, such as must be given by a fissure filled with 

 molten matter from below, must be attended by a powerful influence on them. It 

 would exert itself in aqueous fusion and expansion. But the viscidity peculiar to these 

 highly silicious substances would not allow their extrusion until after the more liquid 

 audesitic masses had been ejected. Supposing the reservoir in which this first 

 change was effected to have been in the trachytic region, further action from the same 

 could be cut off by the cessation of the ingress of water to it. Another reservoir, sit- 

 uated in the rhyolitic region, might then be isolated within the educting channel, filled 

 now with trachytic matter, and the same process repeated, as before, ending with the 

 change of trachytic into rhyolitic rocks. As regards the succession of basalt to these 

 silicious rocks, we refer to the fact established before, that the fissures which gave vent 

 to basalt, have all been formed at a much later epoch than those through which andesite 

 had ascended, and that they were only partly coinciding with them. It would appear 

 that subterranean reservoirs of liquid matter, connected with the surface by channels of 

 ejection, should have offered, in many cases, the places of least resistance. It may, 

 therefore, be inferred, that basalt, the great comparative liquidity of which is a well- 

 known fact, would enter many of those reservoirs, and be emitted through the same, 

 or through newly-formed channels, in preference to any matter of a more viscous con- 

 sistency ; and it will not be difficult to understand why basalt should, in many instances, 

 have again been followed by, or alternated during long epochs with, lavas of a rhyo- 

 litic composition. 



These considerations, which may be equally applied to the order of succession 

 of massive eruptions and to that of volcanic lavas, are not given with a view of cxplain- 



(105) 



