RASAY. GEOLOGY. 247 



conglomerate, apparently as partial, formed of dark red 

 felspar and hornblende schist equally mixed in distinct 

 fragments, and producing, from the contrast of colour, 

 a very peculiar appearance.*' 



But the first conglomerate that forms a part of the 

 sandstone beds, is found immediately above these ; pre- 

 senting some detached and elevated rocks of a very 

 remarkable character, on one of which is situated the 

 very whimsical and picturesque structure, Brochel Castle, 

 the ancient seat of Macleod of Rasay. This building is 

 so contrived as to cover the whole summit of the sharp 

 eminence on which it stands ; its walls being continuous 

 with the precipitous faces of the rock. The projections 

 of these have been so contrived as to form parts of the 

 building ; and they are at the same time so like in appear- 

 ance to the masonry of which it is constructed, that it 

 is often difficult to distinguish between the artificial and 

 the natural wall.f This rock is composed of the usual 

 fragments that enter into the conglomerate in other places, 

 but it appears to be very partial ; as in many parts of the 

 line of junction the fine sandstone is found in immediate 

 contact with the gneiss without any intervening substance. 

 At the same time it bears no marks of stratification ; 

 while its permanence, standing highly elevated as it does 

 in pinnacles far above the surrounding rocks, shows, that 

 with a texture apparently feeble and loose, it must possess 

 much greater durability than the surrounding rocks, of the 

 destruction of which it now remains a monument.J 



* An extensive examination of the western shore of the mainland 

 has since convinced me that these conglomerates appertain to the red 

 sandstone, as they occur in most places where the contact of that sub- 

 stance with the gneiss can be fairly traced. 1818. 



f Vide Plate II. 



* The spiry form assumed by the conglomerate rocks on wasting 

 is a very remarkable circumstance. I need scarcely remind the reader 

 of Montserrat, and of the instances pointed out by Dr. Holland in 

 Greece. 



