they remain to me as improbable as ever. But the oddest result of your
paper on me (and I assure you, as far as I know myself, it is not
perversity) is that I am very much staggered in favour of the ice-lake
theory of Agassiz and Buckland (521/4. Agassiz and Buckland believed that
the lakes which formed the "roads" were confined by glaciers or moraines.
See "The Glacial Theory and its Recent Progress," by Louis Agassiz, "Edinb.
New Phil. Journ." Volume XXXIII., page 217, 1842 (with map).): until I
read your important discovery of the outlet in Glen Glaster I never thought
this theory at all tenable. (521/5. Mr. Milne discovered that the middle
shelf of Glen Roy, which Mr. Darwin stated was "not on a level with any
watershed" (Darwin, loc. cit., page 43), exactly coincided with a watershed
at the head of Glen Glaster (Milne, loc. cit., page 398).) Now it appears
to me that a very good case can be made in its favour. I am not, however,
as yet a believer in the ice-lake theory, but I tremble for the result. I
have had a good deal of talk with Mr. Lyell on the subject, and from his
advice I am going to send a letter to the "Scotsman," in which I give
briefly my present impression (though there is not space to argue with you
on such points as I think I could argue), and indicate what points strike
me as requiring further investigation with respect, chiefly, to the ice-
lake theory, so that you will not care about it...

P.S.--Some facts mentioned in my "Geology of S. America," page 24 (521/6.
The creeks which penetrate the western shores of Tierra del Fuego are
described as "almost invariably much shallower close to the open sea at
their mouths than inland...This shoalness of the sea-channels near their
entrances probably results from the quantity of sediment formed by the wear
and tear of the outer rocks exposed to the full force of the open sea. I
have no doubt that many lakes--for instance, in Scotland--which are very
deep within, and are separated from the sea apparently only by a tract of
detritus, were originally sea-channels, with banks of this nature near