Mineralogy of Sky. 181 



dantly plain that the appearance of stratification in the trap is here 

 the result of the form of the rocks on which it is placed, or among 

 which it has intruded, in the former case surmounting them, and 

 in the latter appearing to alternate with them. The instances of 

 this apparent alternation are highly interesting, from their great 

 extent, as well as from the perfect conviction which they present of 

 the fallacious nature of this supposed connection. In many cases 

 the alternations of the trap are as regular, as decided, and as evenly 

 parallel, as those of the stratified rocks themselves, the sandstone 

 and limestone among which it lies. Yet in no instance does it 

 not happen, but that at some point or other the alternating bed of 

 trap will detach an intersecting vein, unite itself to the superin- 

 cumbent mass, or, quitting the interval between two given beds of 

 limestone or sandstone, make its way across the one immediately 

 above or below, and then proceed with a regularity as great, for 

 another long space, between some other pair of proximate strata. In 

 one or more instances I have observed this to happen after more 

 than a mile in extent, throughout all which space not the minutest 

 irregularity had appeared to indicate any thing else than a perfectly 

 conformable and alternating stratification. I have no doubt that, 

 could such extensive exposure be oftener procured, all the instances 

 of supposed alternation between the trap rocks and the stratified 

 ones would prove similar to these. 



With respect to the trap itself it is most generally amorphous. As 

 we approach however towards the northern end of the promontory 

 it becomes columnar, and this character prevails round the points of 

 Aird and Hunish beyond Duntulm, where it at length terminates. 

 Although the columns are formed on a large scale, and are indivi- 

 dually rude and imperfectly defined, yet their picturesque effect, 

 when seen from a point of view where they can be properly com- 



Vol. iv. 2 a 



