66 THE PROBABLE SEAT OF VOLCANIC ACTION. VL] 



It may here be remarked, that if we regard the liquefaction 

 of heated rocks under great pressure, and in presence of water, 

 as a process of solution rather than of fusion, it would follow 

 that diminution of pressure, as supposed by Mr. Scrope, would 

 cause, not liquefaction, but the reverse. The mechanical press- 

 ure of great accumulations of sediment is to be regarded as co- 

 operating with heat to augment the solvent action of the water, 

 and as being thus one of the efficient causes of the liquefaction 

 of deeply buried sedimentary rocks. 



That water intervenes not only in the phenomena of volcanic 

 eruptions, but in the crystallization of the minerals of eruptive 

 rocks, which have been formed at temperatures far below that 

 of igneous fusion, is a fact not easily reconciled with either the 

 first or the second hypothesis of volcanic action, but is in per- 

 fect accordance with the one here maintained, which is also 

 strongly supported by the study of the chemical composition 

 of igneous rocks. These are generally referred to two great 

 divisions, corresponding to what have been designated the 

 trachytic and pyroxenic types, and to account for their origin, 

 a separation of a liquid igneous mass beneath the earth's crust 

 into two layers of acid and basic silicates was imagined by 

 Phillips, Durocher, and Bunsen. The latter, as is well known, 

 has calculated the normal composition of these supposed trachytic 

 and pyroxenic magmas, and conceives that from them, either 

 separately or by admixture, the various eruptive rocks are de- 

 rived ; so that the amounts of alumina, lime, magnesia, and 

 alkalies sustain a constant relation to the silica in the rock. It', 

 however, we examine the analyses of the eruptive rocks in 

 Hungary and Armenia, made by*Streng, and put forward in 

 support of this view, there will be found such -discrepancies 

 between the actual and the calculated results as to throw grave 

 doubts on Bunsen's hypothesis. 



Two things become apparent from a study of the chemical 

 nature of eruptive rocks : first, that their composition presents 

 such variations as are irreconcilable with the simple origin gen- 

 erally assigned to them ; and, second, that it is similar to that of 

 sedimentary rocks whose history and origin it is, in most cases, 



