36 GEOLOGY OP ARRAN. 



towards the openings of these valleys, we should expect the 

 effects of glacial action to be most distinctly traceable in the 

 striation and polishing of the subjacent rocks and transported 

 masses, and in the formation of lateral and terminal moraines. 

 Many broad surfaces of the natural rock are exposed both 

 on the sides and in the bottoms of these valleys, favourably 

 placed for receiving such impressions under the grinding 

 action of a descending mass ; yet the cases of striation and 

 polishing are not numerous. A granite surface is, how- 

 ever, very unfavourable for the preservation of such mark- 

 ings, especially the Arran granite, which is generally of such 

 structure as to be subject to rapid disintegration. The slate 

 is better fitted to retain impressions of this kind, its tough- 

 ness and fine-grained structure rendering it less liable to 

 decomposition; but it is seldom exposed in favourable situa- 

 tions, and is rarely found striated. Granite bosses in the 

 glens, and on many of the lower ridges, have that peculiar 

 rounded character, due to the action of ice, to which the 

 term "roches moutonn^es" has been applied; but perhaps 

 few of the cases ought to be decidedly referred to glacial 

 action, on account of the peculiar spherical structure so often 

 assumed here by granite on the large scale. On the slate 

 ridges, however, beyond the granite border, on the old red 

 sandstone, and on the trap of the southern plateau some well- 

 marked cases do occur. Other evidences, scarcely less satis- 

 factory, are of frequent occurrence. These are the terraces 

 and mounds of transported materials on the sides and at the 

 openings of the glens, the boulder clay deposits, and the 

 dispersed blocks in every part of the island. 



The terraces and mounds consist of earth and rounded 

 masses of rock of each particular glen, irregularly mixed, with- 

 out reference to weight, and in such situations that they 

 could not have been brought together by existing river action, 

 being much above the level of the streams which now tra- 

 verse the valleys. They are most probably referable to 

 glacier moraines of two classes the lateral and terminal 



