KERRERA. GEOLOGY. 1 19 



trap with which it is connected. With respect to com- 

 position I may remark, that in general, perhaps in all 

 cases, the conglomerates partake of the different older 

 rocks next to which they lie ; containing fragments of 

 all these, or at least of all such as have had sufficient 

 durability to survive the mechanical attrition they have 

 undergone before they were consolidated into one mass. 

 Thus a given line of conglomerate, however continuous 

 it may be as a bed, will vary in its composition as the 

 rocks on which it lies in different places, change their 

 characters ; and for the truth of this remark I may refer 

 to the different parts of the great conglomerate line of 

 Scotland, where differences of this nature actually occur. 

 In the case of the Kerrera rock, it is easy to conceive, 

 that the conglomerate has been formed during those revo- 

 lutions of the surface that produced the whole of this 

 substance wherever it is found, and in the vicinity of 

 those trap rocks which constitute so remarkable a feature 

 in this country. Hence it has necessarily partaken of 

 their nature ; while it includes, at the same time, frag- 

 ments of granite and quartz, of an origin more distant, 

 and more susceptible of transportation without injury. 

 One remark of an important nature, but similar to that 

 deduced from the structure of Canna, may be de- 

 rived from the nature of the rocks now described. 

 Two distinct formations of trap are indicated by its 

 composition and situation; one prior to the deposition 

 of the conglomerate, and one posterior ; the former being 

 the source of the fragments it contains, and the other 

 now overlying and intersecting the beds. The inde- 

 pendent veins hereafter described are, equally, proofs of 

 a third. 



With respect to the two latter classes of this trap, there 

 is nothing remarkable in this particular instance ; as ana- 

 logous proofs of two formations of trap have in various 

 cases come under review throughout this work. The 

 former deserves a separate consideration ; since the evi- 

 dence derived from the structure of Canna, does not 



