38 E1CIITH0FEN NATURAL SYSTEM 



2d. Mineral composition, in respect to which each class represents a series of 

 gradations essentially dependent on the chemical composition, and therefore chiefly 

 coincident with the chemical series, but differing from it inasmuch as it is more articu- 

 late. Certain types, corresponding to certain steps in the chemical scale, and consist- 

 ing in distinct aggregations of a few minerals, are the centers for which the petro- 

 graphical names have been applied in the first place ; around them other members are 

 grouped which connect every two types by gradual passages, and are ordinarily com- 

 posed of an aggregation of all the minerals peculiar to either of them. Expansion of 

 the series, as it were, in a lateral sense, occasioned by the accession of minerals inferior 

 in importance, are not unfrequent, but do not affect in a great measure the definite 

 character of mineralogical gradation, as they are of local occurrence, and probably less 

 dependent upon any material differences in chemical composition, than upon certain 

 influences which acted upon the mass of the rock, either before its ejection, and 

 then by the admixture of matter differing from it in nature, or after its solidification, 

 and then by chemical metamorphism. 



3d. Specific gravity, which increases in the reverse proportion of silica. In this 

 respect, too, each class represents a series of infinite gradations from the lowest to the 

 highest value. 



Taking the three foregoing principles exclusively as the basis for classification, 

 all eruptive rocks would have to be united into one class. This union, made regard- 

 less of any other relations, may be traced as the leading feature of nearly all systemat- 

 ical arrangements proposed. There are, however, other points of view which must be 

 considered if a more perfect classification is aspired. They lead to the establishment 

 of further separations, by principles similar to those which we had to apply above in 

 defining the different orders of the compounds of hornblende and oligoclase among 

 volcanic rocks. These points of view are the following : 



4th. Mode of Texture. — Eruptive rocks exhibit, in regard to their modes of text- 

 ure, very peculiar differences, which are little capable of explanation with our present 

 state of knowledge. They are especially conspicuous with the most silicious rocks. 

 Free silica, in granite, has probably been solidified cotemporaueously with the other 

 component minerals ; but its solidification appears often to have been completed last 

 of all, as the quartz does envelop the aggregation of the other crystals. In quartzose 

 porphyry and rhyolite, on the other hand, free silica, at least the greater part of it, 

 has been solidified first, which is obvious from the fact that its perfect crystals are 

 imbedded in an ordinarily microciystalline aggregation of the other ingredients, which 

 however contains, in most cases, some surplus of free silica that had not entered into 

 the composition of the crystals. This fundamental difference from granite points to 

 some difference in the mode of origin. Quartzose porphyry and rhyolite differ in regard 

 to the texture of their paste, which has a compact aspect in the first, while it is more or 

 less vesicular throughout the mass in most varieties of the latter. This difference, like 

 the former, appears to indicate, that the molecular condition of the liquid mass, at the 

 time when it was ejected, was different in either rase. The variety of texture diminishes 

 with the decrease of silica, and the more basic rocks of the three classes bear a much 



(76) 



