The view that has now gained general acceptance is that the parallel roads of Glen Roy represent the shores of a lake "that came into being with the growth of the glaciers and vanished as these melted away." (517/13. Sir Archibald Geikie, loc. cit., page 269.)
Mr. Darwin became a convert to the glacier theory after the publication of Mr. Jamieson's paper. He speaks of his own paper as "a great failure"; he argued in favour of sea action as the cause of the terraces "because no other explanation was possible under our then state of knowledge." Convinced of his mistake, Darwin looked upon his error as "a good lesson never to trust in science to the principle of exclusion." (517/14. "Life and Letters," I., page 69.)
LETTER 517. TO C. LYELL. [March 9th, 1841.]
I have just received your note. It is the greatest pleasure to me to write or talk Geology with you...
I think I have thought over the whole case without prejudice, and remain firmly convinced they [the parallel roads] are marine beaches. My principal reason for doing so is what I have urged in my paper (517/15. "Observations on the Parallel Roads of Glen Roy, and of other parts of Lochaber in Scotland, with an attempt to prove that they are of Marine Origin." "Phil. Trans. R. Soc." 1839, page 39.), the buttress-like accumulations of stratified shingle on sides of valley, especially those just below the lowest shelf in Spean Valley.
2nd. I can hardly conceive the extension of the glaciers in front of the valley of Kilfinnin, where I found a new road--where the sides of Great Glen are not very lofty.
3rd. The flat watersheds which I describe in places where there are no roads, as well as those connected with "roads." These remain unexplained.
I might continue to add many other such reasons, all of which, however, I daresay would appear trifling to any one who had not visited the district. With respect to equable elevation, it cannot be a valid objection to any one who thinks of Scandinavia or the Pampas. With respect to the glacier theory, the greatest objection appears to me the following, though possibly not a sound one. The water has beyond doubt remained very long at the levels of each shelf--this is unequivocally shown by the depth of the notch or beach formed in many places in the hard mica-slate, and the large accumulations or buttresses of well-rounded pebbles at certain spots on the level of old beaches. (The time must have been immense, if formed by lakes without tides.) During the existence of the lakes their drainage must have been at the head of the valleys, and has given the flat appearance of the watersheds. All this is very clear for four of the shelves (viz., upper and lower in Glen Roy, the 800-foot one in Glen Spean, and the one in Kilfinnin), and explains the coincidence of "roads" with the watersheds more simply than my view, and as simply as the common lake theory. But how was the Glen Roy lake drained when the water stood at level of the middle "road"? It must (for there is no other exit whatever) have been drained over the glacier. Now this shelf is full as narrow in a vertical line and as deeply worn horizontally into the mountain side and with a large accumulation of shingle (I can give cases) as the other shelves. We must, therefore, on the glacier theory, suppose that the surface of the ice remained at exactly the same level, not being worn down by the running water, or the glacier moved by its own movement during the very long period absolutely necessary for a quiet lake to form such a beach as this shelf presents in its whole course. I do not know whether I have explained myself clearly. I should like to know what you think of this difficulty. I shall much like to talk over the Jura case with you. I am tired, so goodbye.
LETTER 518. TO L. HORNER. Down [1846].
(518/1. It was agreed at the British Association meeting held at Southampton in 1846 "That application be made to Her Majesty's Government to direct that during the progress of the Ordnance Trigonometrical Surveys in the North of Scotland, the so-called Parallel Roads of Glen Roy and the adjoining country be accurately surveyed, with the view of determining whether they are truly parallel and horizontal, the intervening distances, and their elevations above the present sea-level" ("British Association Report," 1846, page xix).