J". W. Judd—On Volcanos. 303 



it is questionable whether they ought not really to be assigned to 

 that class of rocks. 



Every attempt, like that of Gustav Eose, to give granite a purely 

 mineralogical definition, has failed, in consequence of the variation, 

 even in different parts of the same mass, in its constituent minerals. 

 The several felspars may replace one another in an almost infinite 

 number of ways, and different micas and hornblendes may be simi- 

 larly substituted for one another, while accessory minerals may so 

 increase in abundance as to become important constituents of the 

 rock, without its in any way forfeiting the title to be considered a 

 true granite. The texture of the rock, however, appears to afford 

 surer ground on which we may base a definition, than the exact 

 species of minerals which compose it. Normal granites consist of an 

 aggregate, in which distinct crystals of orthoclase, and often of some 

 plagioclastic felspar, with those of one or more species of mica or 

 hornblende, have separated, leaving a base composed of quartz, ex- 

 hibiting a greater or less tendency to form distinct crystals, and a 

 crystalline mass of felsitic matter enveloping the perfect and im- 

 perfect crystals, and representing the " mother liquor " out of which 

 these latter have been formed, portions of which are also entangled 

 in their cavities. There are, however, granites in which the quartz 

 appears to have more readily crystallized, and to have been among 

 the first minerals separated from the mass. 



Now the remarkable rock of the Ponza Islands has an ultimate 

 chemical composition identical with that of many granites ; its con- 

 stituent minerals — orthoclase albite or oligoclase, quartz and mica 

 or hornblende —are precisely those of ordinary granite ; and hence 

 it must be by its texture, if at all, that we must hope to be able to 

 separate it from that class of rocks. 



The study of this Ponza rock clearly proves that the minerals of 

 which it is composed have had four different modes of origin. 



I. They may have crystallized out from a liquefied magma, prob- 

 ably under great pressure, and long before it reached the surface. 

 This is, I believe, the origin of the large crystals of mica, hornblende, 

 felspar, and the smaller and less perfect ones of quartz, which are 

 found scattered, often in great abundance, alike through the most 

 vitreous and the most stony varieties of the rock. In proof of this 

 fact of the formation of large crystals in the magma before its 

 eruption I may cite the following facts. 



1. In the masses of volcanic sand blown from the throats of vol- 

 canos, crystals (usually of course broken and damaged, but of pre- 

 cisely similar character to those embedded in the lava) abundantly 

 occur. The perfect augite ciwstals ejected by Stromboli afford an 

 interesting illustration of this fact. 



2. Where the lava contains these large porphyritically embedded 

 crystals, the scoria? or pumice formed from it will be found to contain 

 the same crystals in a perfect condition, entangled in the meshes of 

 the distended rock; clearly proving that these crystals were floating 

 in the liquefied mass before its ejection. This fact is exemplified in 

 many of the pumices and scoria? of Ischia. 



