CANNA. GEOLOGY. 457 



Appearances of disturbance and renewal which characterize 

 them are accompanied by no corresponding changes 

 in the later stratitied rocks. It appears equally to 

 follow, that the deposits of trap have a tendency to 

 recur in the same spot ; as in this, among many others, 

 there is found a succession of partial formations limited 

 to a narrow space. If it be argued that the trap has 

 formed one universal bed over the surface, which subse- 

 quent changes have partially destroyed, a more accurate 

 correspondence should be exhibited between the neigh- 

 bouring rocks of this class ; in the case for example 

 now under review, between those of Canna, Rum, Egg, 

 and Sky. 



I am aware that the opinions thus hazarded on the 

 partial formations of trap and on their repetition in 

 the same spot, may be supposed to give a colour to 

 the well known hypothesis of their recent volcanic 

 origin. But the other evidence of this is too deficient 

 to afford any just grounds of conclusion ; while it is 

 scarcely necessary to say that ancient volcanoes, some- 

 times evidently submarine, are capable of fulfilling all 

 the requisite conditions. Should these appearances be 

 even supposed the consequences of ordinary, but distant 

 volcanic eruptions, they are not difficult of explanation. 

 Posterior changes of the earth's surface, of which the 

 traces are so generally diffused, might easily have 

 divested these mountains of the peculiar substances 

 that characterize existing volcanoes, since these are 

 frequently of a nature to suffer from such revolutions ; 

 while the peculiarity of form which the places now 

 in question exhibit, may equally have resulted from 

 changes of such a nature as to leave little or no traces 

 of their original outline. Amid the alterations that 

 appear to have taken place on the surface, it is asking 

 for very little to suppose that these partial deposits of 

 trap are the remains of more extensive masses. 



