THE SCHISTOSE ISLES. 299 



direction of the shores is oblique to that of the rocks, 

 and the same disposition will be found to pervade the 

 southern part of the peninsula of Cantyre. 



If the lines of direction above mentioned are connected 

 throughout, it will be seen that they are curved ; and 

 the reality of the curvature is proved by its taking place 

 not in the interval between the southern and northern 

 islands, but in the middle of Luing, as was shown in 

 the description of that island. This curvature is however 

 neither great nor permanent, nor is it such as to impli- 

 cate all the parallel strata ; as it is not found in the 

 Garveloch isles which lie immediately to the west of 

 Luing, and as the strata again approximate to their ori- 

 ginal direction in Kerrera, and in Lismore ; this island 

 lying parallel to the long lines of the Linnhe Loch, 

 beyond which I shall not at present attempt to trace 

 these connexions. 



Although the directions of the strata are thus constant, 

 with certain trifling exceptions which must be rather 

 viewed as temporary undulations than as serious de- 

 viations from their general course, the dip is in different 

 places, not only various in quantity, but reverse in 

 position. In Gigha it is to the westward, while it is to 

 the eastward, not only in Isla and Jura, but in all the 

 northern islands. The point of change cannot be traced 

 in the islands, since it must exist somewhere in the strait 

 by which Gigha is separated from Isla and Jura. But 

 it is easily discovered on the mainland, where it has 

 already been pointed out in the preceding examination 

 of the Chlorite series, so distinctly as to render any repe- 

 tition here unnecessary. It must have been observed by 

 the reader, that to whatever quarter the dip lies, the 

 strata of these islands are rarely elevated to less than 

 twenty degrees, and as rarely exceed sixty: we may 

 therefore assume forty as a medium quantity, that being 

 at the same time, in round numbers, perhaps the most 

 common angle of general occurrence. It is obvious also 



