46 Dr. Mac CuUoch on the 



Among these is found that compound of felspar and horn- 

 blende, with excess of the former mineral, to which the term 

 Syenite has been applied, and to which, together with the analo- 

 gous rocks of the same family, it is here exclusively limited. On 

 one side, this rock passes into porphyry in the usual manner ; 

 or, by the loss of its hornblende only, into a simple rock, which 

 in the same imperceptible manner, graduates into claystone. 

 But, in another part, quartz, and subsequently quartz and mica 

 both, are superadded to the compound of hornblende and fel- 

 spar ; and thus there is produced a rock, in no way differing 

 from many varieties of ordinary granite, and, in particular, 

 strongly resembling some of those which occur in Arran. The 

 connexion of this granite, or rather syenite of a granitic cha- 

 racter, with the adjoining ordinary trap-rocks, is such as to admit 

 of no doubt respecting its identity of origin ; and it is unne- 

 cessary to say that it is thus proved, even if more distinct 

 evidence of that circumstance were not accessible, to be super- 

 incumbent on conchiferous limestone. The instance already 

 mentioned as described by Mr. Von Buch, must doubtless be 

 considered as of the same kind ; and even those who would 

 otherwise be inclined to withhold their assent from this view of 

 its nature, will probably choose to adopt this conclusion, rather 

 than to admit of a granite more recent than the latest secondary 

 strata, or of one which occupies that superincumbent position, 

 the existence of which has been refused to those who argue in 

 favour of its igneous origin. 



In St. Kilda, there is found a mass of trap, consisting chiefly 

 or entirely of that variety which I have called augit rock, 

 connected with a syenitic rock in which hornblende and felspar 

 form the chief ingredients, but which also contains quartz. 

 Although no stratified rocks are found in this island, it may 

 be concluded, from the mineral characters of these rocks, and 

 more particularly from the presence of augit, which exists as 

 an essential contituent only in the trap family and in the volcanic 

 rocks, that St. Kilda belongs to the family of trap, and not to 

 that of granite. 



In this syenite cavities are of frequent occurrence, containing 



