THE TRAP ISLANDS. 77 



lively to the inferior strata, at which they were deposited. 

 The consequences resulting from the former are obvious. 

 From the latter it is apparent that they can bear no fixed 

 relations to the strata ; as, independently of the disturb- 

 ances which they have produced in these, many changes 

 of the surface, and extensive losses of substance, have 

 occurred before their deposition, or at periods intermediate 

 between their several depositions. But it is unnecessary 

 to pursue a subject of this nature. 



Under whatever periods of time the trap rocks of these 

 islands may have been formed, it is evident that they are 

 all mutually dependent; and that they are, at the same 

 time, concentrated within a certain limited space, with the 

 sole exception of St. Kilda. It is also to be remarked 

 that no trap occurs on that part of the continent which 

 adjoins the northern portion of the group ; the northern- 

 most mass on the mainland being that of Airdnamurchan. 

 By means of this district, and that of Morven, the con- 

 nexion between the southern and northern portions is 

 established, in this case, as in that of the secondary strata. 

 Mull, which forms the largest continuous mass in the 

 southern division, as Sky does in the northern, is not 

 only connected with Morven, but appears equally related 

 to a considerable tract of these rocks, fragments of which 

 will be found extensively scattered, both over the neigh- 

 bouring mainland, and among the islands on the coast 

 of Lorn which will hereafter come under review. The 

 nearest connexion of this nature is that already pointed 

 out on the Morven shore, where an extensive district 

 of trap is seen extending to Airdnamurchan in a northerly 

 direction, and terminating towards the south by an abrupt 

 juxtaposition to a mountain of gneiss, through which it 

 sends veins; the line which separates the one substance 

 from the other being almost perpendicular, while the 

 common height of the ridge at this junction is not much 

 less than 2000 feet. It offers an interesting example of 

 a mode of connexion between these two classes of 





