Parallel Terraces. ?39 



of level between the terrace of Glen Gloy and those of Glen 

 Roy, and would obviate the necessity of imagining soulevemens 

 of the neighbouring valleys, which communicate in the same 

 manner with the ocean, and do not nevertheless exhibit any 

 trace of terraces. 



In following up these facts in all their variety, we are easily 

 enabled to explain the numerous terraces which we meet with 

 in Scotland, by supposing barriers of ice at the mouths of the 

 valleys ; whether it was that the lateral valleys closed them 

 by their glaciers, as at the Bridge of Roy, or that the waters 

 of the sea, by heaping up ice on the coasts, offered a tempo- 

 rary obstacle to the running off of the waters of the land, or 

 intercepted large sheets of salt water. The presence of an 

 Arctic fauna, in the deposits superior to the till, which might 

 be formed in these creeks of the sea, would thus present no- 

 thing but what is quite natural. 



In order the better to understand the explanation I have 

 given of the phenomena of Glen Roy and of Glen Spean, I 

 shall now endeavour to compare them with what would take 

 place in the valley of Chamounix, if the Glacier des Bosons, or 

 that of Taconay, by extending a little farther than it now 

 does, had barred the valley in such a manner as to interrupt the 

 passage of the waters of Arve, and force them to flow off by the 

 valley of Trient (Plate IV. Fig 2). A lake would thus be formed 

 above Chamounix, of which the level would leave some traces 

 on the flanks of the valley ; but, when the glacier was no longer 

 so elevated as the height of the Col des Montets^ the water would 

 flow to the west, as at present, forming terraces as often as it 

 was maintained for a certain time at a fixed height. To ren- 

 der this comparison still more striking, let us suppose the case 

 where a second glacier, as, for example, the glacier des Bois^ 

 also shut up the valley ; then, as in the instance of Glen Roy, 

 there would be between the two glaciers a lake, whose waters 

 would first of all descend to the east over the least elevated gla- 

 cier, and would not acquire their natural direction to the west 

 until the two glaciers disappeared, and the bottom of the valley 

 was free. Ice, cast on the coasts of Holland at the present day, 

 and interrupting the dispersion of the sand transported by the 

 Rhine and by the Scheldt, would every where reproduce the 



