195 



occasionally imbedded in rocks as their natural re- 

 positories. It is sufficient to quote as examples, gar- 

 net, which is sometimes abundant in micaceous schist, 

 or sparingly dispersed, or altogether absent, without 

 affecting its essential characters, and spodumene or 

 corundum, which may thus exist in granite. 



If we consider the great number of minerals in 

 nature, thus generally distinguished into essential and 

 unessential, or if even we limit our views to those 

 which may be considered as most essential, it Js in- 

 teresting to observe how few are the rocks which are 

 produced from them. If the varieties are most nu- 

 merous in the primary or older series, they are still 

 confined, and, within certain limits of variation, very 

 constant. In the later rocks, they are still more limited. 



When we reflect on the circumstances under which 

 the primary rocks at least have been produced, they 

 are confined to a much less number than would have 

 been anticipated. As most of the minerals of ordinary 

 occurrence are formed, for example, of the earths 

 which exist in granite and gneiss, we might have ex- 

 pected to find garnet, corundum, or andalusite, in 

 every mass of these; instead of being, as they are, 

 limited to a few occasional specimens. Nor is it 

 always very easy to account for those distinctions be- 

 tween gneiss, micaceous schist, quartz rock, or other 

 substances, which occur in the same antient series; 

 distinctions which, on the great scale, are really steady 

 and definite, notwithstanding the occasional interfe- 

 rences of character that occur in particular instances. 

 That these have been regulated, however, partly by 

 mechanical and partly by chemical laws, is certain; 

 and though we cannot perhaps explain every case, it 

 will immediately be seen that we are in possession of 

 general principles applicable to the solution of the 



o 2 



