STAFFA* GENERAL DESCRIPTION. 17 



twenty-two feet; preserving, as will be seen by these 

 measures, a considerable degree of regularity throughout. 

 The length is 227 feet.* The sides of this cave are, like 

 the front, columnar, -and in a general sense perpendicular : 

 though when accurately viewed, they are, in the same 

 way, far from possessing that geometric regularity which 

 accompanies all the views of it hitherto published. The 

 columns are frequently broken and irregularly grouped, 

 so as to catch a variety of direct and reflected tints 

 mixed with unexpected shadows, that produce a pictu- 

 resque effect which no regularity could have given. The 

 ceiling is various in different parts of the cave. It is 

 deeply channelled in the middle by a fissure parallel to 

 the sides and prolonged from the point of the exterior 

 arch to the end. That portion which lies on each side of 

 this fissure toward the outer part of the cave, is similar to 

 the upper incumbent bed, being formed of a minutely 

 fractured rock. In the middle it is composed of the 

 broken ends of columns, which produce an ornamental 

 and somewhat architectural effect; while at the end, a 

 portion of each kind of rock enters into its formation. 

 From attending only to one or other of these portions, 

 different observers have described the ceiling in a different 

 manner; and each party has accused the other of mis- 

 representation. The surfaces of the columns above, are 

 sometimes distinguished from each other by the infiltration 

 of carbonat of lime into their interstices. As the sea 

 never ebbs entirely out, it forms the only floor to this 

 cave ; but the broken range of columns which produces the 

 exterior causeway, is continued on each side within it. 

 This range is most perfect at the eastern side, and admits 

 of access over the broken summits to the further end, 



* These measures were all taken with great care. If there be any dif- 

 ferences between them and former measurements, they may have arisen 

 from the impossibility of different observers choosing the same points in 

 an irregular surface. 



VOL. II. C 



