42 GEOLOGY OF ARRAN. 



elude that moving masses of ice were the tran spoiling agents, 

 la the passage of glacier ice adown the valleys, and the 

 buoyancy of floating bergs, forces of sufficient energy would 

 be lodged to carry the largest masses ; and this agency we 

 know is adequate to produce the various phenomena of trans- 

 port, grouping, and " perched blocks." 



20. We seem thus shut up to the conclusion that the agent 

 in this transport was ice in motion. Now, this agency may 

 have been brought into play in two ways. The northern 

 mountains may have formed a mass of ice-covered land, with 

 glaciers descending to the bold shores of the sea of that 

 period, while the southern plateau may have been under 

 water. From the extremities of the glaciers masses of ice, 

 of which some must have been considerable bergs, and the 

 sea therefore not very shallow, would float away, carrying 

 the granite blocks which had fallen on the ice towards the 

 heads of the valleys, and been borne along on the glacier. 

 Stranding or melting, these floating masses would throw 

 down their load of blocks, and thus the shores of the island 

 and the surface of the southern plateau may have become 

 encumbered with vast multitudes of granite blocks, chiefly 

 of that coarse-grained variety which constitutes all the 

 highest mountains. But Ih another way ice may have been 

 the agent of the transport in question. The whole island, 

 elevated in both divisions high above the waters, may have 

 been wrapped in sheets of ice, across which the granite 

 blocks, as they dropped from the high peaks and precipices 

 on which the snow could not rest, or were torn off by the 

 pressure of the ice from the sides of the glens, would be 

 carried onwards in all directions by the slowly descending 

 viscous mass. Existing glaciers do not require very steep 

 slopes for accomplishing the transport of large masses. A 

 moderate inclination is sufficient; and we need not therefore 

 suppose that there existed any very great difference in the 

 relative levels of the surface in Arran from that which now 

 obtains. All the largest blocks are found at the lowest 



