OF UNSTRATIFIED ROCKS AND VEINS. 159 



But the rock of St. Kilda is connected .with common 

 greenstone and augit rock; substances esteemed to 

 appertain to the Trap family. If these, instead of 

 being admitted among the traps, are referred to gra- 

 nite, St. Kilda will merely offer an instance to add to 

 the former parts of this analogy. 



In Sky, a quadruple compound of hornblende, 

 felspar, mica, and quartz, passes into a triple one in 

 which mica is absent, and at length, by a variety of 

 gradation, into a compound of felspar and quartz, 

 felspar and hornblende, claystone, porphyry, and 

 greenstone. The whole mass is continuous; and, in 

 some place or other, every member of it is found 

 lying above the secondary strata. Here then is a 

 mass of trap, containing a granite, undistinguishable, 

 not only from many of the varieties which occur in 

 Arran, but from those which are so often found be- 

 neath the primary strata. 



There is little doubt that whenever adequate ob- 

 servers shall choose to follow the course here pointed 

 out, and to pursue, unbiassed, that chain of observa- 

 tion which is only here for the first time indicated, 

 more instances of the same nature will be brought to 

 light, both among the traps and granites. It is es- 

 pecially necessary, however, to guard against being 

 misled by mineralogical terms. It is not in the minute 

 arrangements of a cabinet of specimens, that the great 

 features of nature are to be studied; nor can geology 

 ever rise to the rank of a science, if it is to be culti- 

 vated by grovelling among fragments and substituting 

 words for knowledge. 



I may now indeed add, that since this chapter has 

 been so far arranged as to render its alteration incon- 

 venient, I have been satisfied, by the examination, 

 both of collections and their collectors, that the same 



