RUM. GEOLOGY. 487 



as the cliffs, since no boat can approach, and it is ac- 

 cessible only in one place from the shore. At the point 

 of Bridianoch this syenite is elevated to an angle of 

 about sixty degrees, being disposed in distinct beds. 

 Proceeding eastward, these become insensibly bent, and 

 at length assume a horizontal direction; till, near the 

 shores of Harris, they lose all form, and terminate in an 

 irregular mass. It has a tendency to a columnar fracture 

 at right angles to the beds, which, where the bed is 

 horizontally disposed, puts on in some places the ap- 

 pearance of regular columns.* In the interior, this rock 

 extends far to the north-east, forming the entire mass of 

 the long irregular mountain Oreval, and its dependency 

 Fiunacra, of nearly equal height but of far less extent. 

 These hills have a round smooth outline; unlike Halival 

 and Haiskeval, which are very acute and rough. Their 

 declivities are overwhelmed with fragments, of which 

 the progress downwards has formed the alluvial terrace 

 already mentioned as existing at Harris. 



It is now necessary to inquire into the relation which 

 the syenite bears to the stratified rocks before described. 

 Where its junction with the sandstone ought to take 

 place, the land is absolutely inaccessible, since it lies 

 in the high cliffs about Bridianoch ; while it is also at too 

 great a distance, considering the similarity in the colour 

 and disposition of the two substances, to admit of being 

 perceived from below. The spectator indeed passes 

 from the one rock to the other, in skirting the coast 

 in a boat, without perceiving the transition, and does 

 not readily discover that he has quitted the sandstone 

 and is contemplating the syenite. Nor could I in any 

 other part of the island find out this junction, since 



1 Viewing it from a vessel as I sailed past in a dark and stormy 

 day, I was misled by this circumstance to pronounce it a basalt, in 

 a slight notice of this island formerly inserted in the Geological 

 Transactions ; an error which I am glad thus to correct. 



