154 Professor J. W. Gregory — 



pp. 132-3). " From the higli ground of the interior," says Dr. Mort, 

 " glaciers diverged, which finally joined the main streams, and 

 moved south down the Firth " (Mort, 1914, p. 25). The flow to 

 the north was blocked by ice from the mainland, and the chief 

 discharge was southward. As granite erratics from Arran are 

 rarely found in Ayrshire its ice must have been deflected to the 

 south immediately after leaving the island by ice from the north. 



If the Arran valleys had been formed by the radially flowing 

 ice they should have been radial from the granite centre, and it has 

 been claimed that they " run approximately from the centre in 

 radial fashion " (Mort, 1914, p. 7). This is true of the minor valleys ; 

 but the most conspicuous feature in the arrangement of the major 

 valleys of Arran is that they are not radial. 



Dr. Mort remarks (op. cit., p. 71) that the Arran glens present 

 features which cannot be explained by normal stream erosion ; and 

 this view is certainly correct as regards their plan, which would 

 not have been adopted by streams flowing from the granite dome of 

 Northern Arran unless l3he denudation had been guided by pre- 

 existing lines of weakness. The same objection, however, would 

 apply to ice erosion ; for, unless guided by pre-Grlacial valleys, glacial 

 erosion should also have produced valleys radial from the high 

 ground. The major valleys were jorobably determined by structural 

 lines of weakness due to the uplift of the area. Glen Rosa and upper 

 Glen Sannox occur along a basaltic dyke, which is the longest dyke 

 in Arran. This dyke was forced into a north and south rift formed 

 before the end of the igneous period. The uplift which formed the 

 thousand-foot platform probably caused a series of north and south 

 fractures which did not produce many faults, but rifts or bands 

 of ruptured weakened rock (the shatter-belts of Marr). The streams 

 naturally followed these rifts and weak bands, and thus the major 

 valleys of North Arran were a series of three north and south valleys 

 across the granite block. That these three valleys were tectonic in 

 origin appears to be further indicated by their agreement with the 

 general structural lines of the district. The general course of the 

 eastern and western coasts has been determined by faults and 

 fractures. Kilbrennan Sound to the west and the main Firth of 

 Clyde to the east are probably both sunken blocks, which have 

 been let down by north and south trough faults ; and the block of 

 North Arran was doubtless rent by north and south clefts formed by 

 the same series of movements. 



3. Tectonic Clefts and Shatter-belts. 



The valleys, however, are not fault valleys, as no recognizable 

 displacement has taken place among them. They may be simple 

 clefts formed by the gaping of fractures in rocks undergoing tension 

 during uplift. The necessary formation of such tension clefts has 

 been repeatedly asserted, as e.g. by James Smith in 1862 (p. 24), 

 by Ruskin (1856, vol. iv, p. 229, i.e. pt. v, ch. xv, p. 31), and Bailey 



