skip to content

Darwin Correspondence Project


To Charles Lyell   8 [September 1847]1

Down Farnborough Kent

Wednesday 8th

My dear Lyell

Many thanks for the Paper.2 I do admire your zeal on a subject on which you are not immediately at work. I will give my opinion as briefly as I can & I have endeavoured my best to be honest. Poor Mrs Lyell will have, I foresee, a long letter to read aloud, but I will try to write better than usual.—3

Imprimis, it is provoking that Mr Milne has read my paper with little attention, for he makes me say several things, which I do not believe,—as that the water sunk suddenly p. 10!! that valley of Glen p. 13. & Spean4 was filled with detritus up to level of the lower shelf, against which I conceive there is good evidence &c,—but I suppose it is the consequence of my paper being most tediously written. He gives me a just snub for talking of demonstration;5 & he fights me in a very pleasant manner.

Now for business,—I utterly disbelieve in the barriers for his lakes & think he has left that point exactly where it was in the time of Macculloch & Dick.6 Indeed in showing that there is a passage at Glen Glaster at the level of the intermediate shelf he makes the difficulty to my mind greater. When I think of the gradual manner in which the 2 upper terraces die out at Glen Collarig, & at mouth of Glen Roy—the smooth rounded form of the hills there,—the lower shelf retaining its usual width where the immense barrier stood;7 I can deliberately repeat “that more convincing proofs of the non existence of the imaginary Loch Roy could scarcely have been invented with full play given to the imagination.” &c8 But I do not adhere to this remark with such strength, when applied to the Glacier lake theory:9 oddly I never was at all staggered by this theory, until now having read Mr Milne’s arguments against it; I now can hardly doubt that a great glacier did emerge from L. Treig, & this by the ice itself (not moraine) might have blocked up the 3 outlets from Glen Roy. I do not, however, yet believe in the glacier theory, for reasons which I will presently give.—

There are three chiefly hostile considerations in Mr Milne’s paper.— First the Glen luy shelf not coinciding in height with the upper one: four observations giving 12ft, 15ft, 29ft, 23ft;10 if the latter are correct, this terrace must be quite independent & the case is hostile,11 but Mr Milne shows that there is one in Glen Roy 14 ft below the upper one, & a second one again (which I observed) beneath this, & then we come to the proper second shelf. Hence there is no great improbability in an independent shelf having been formed in Glen Gluy.

This leads me to Mr Milnes second class of facts, (obvious to everyone) namely the non-extension of the three shelves beyond Glen Roy; but I abide by what I have written on that point,12 & repeat, that if in Glen Roy,, where circumstances have been so favourable for the preservation or formation of the terraces, a terrace could be formed quite plain for 34 of a mile with hardly a trace elsewhere, we cannot argue from the non-existence of shelves, that water did not stand at these same levels in other valleys. Feeling absolutely convinced that there was no barrier of detritus at the mouth of Glen Roy, & pretty well convinced there was none of ice, the manner in which the terraces die out when entering Glen Spean, which must have been a tide way, shows on what small circumstances the formation of these shelves depended. With respect to the non-existence of shelves in other parts of Scotland, Mr Milne shows that many others do exist, & their heights above the sea have not yet been carefully measured, nor have even those of Glen Roy, which I suspect are all 100 ft too high.13 Moreover according to Bravais,14 we must not feel sure that either the absolute heights, or the intermediate heights between the terraces, would be at all the same at distant points. In levelling the terraces in Lochaber, all I believe have been taken in Glen Roy neary N. & S; there shd be levels taken at right angles to this line, & to the Great Glen of Scotland, or chief line of elevation.

Thirdly: the nature of the outlets from the supposed lakes: this appears to me the best & newest part of the paper. If Sir James Clark15 would like to attend to any particular points, direct his attention to this; especially to follow Glen Glastig from Glen Roy to L. Laggan. Mr Milne describes this as an old & great river course with a fall of 212 ft. He states that the rocks are smooth on upper face & rough on lower, but he does not mention whether this character prevails throughout the whole 212 vertical feet,—a most important consideration. Nor does he state whether these rocks are polished or scratched, as might have happened even to a considerable depth beneath the water (Mem. great icebergs in narrow fiords of T. del Fuego)16 by the action of icebergs, for that icebergs transported boulders on to the terraces, I have no doubt. Mr Milne’s description of the outlets of his lakes, sound to me more like tidal channels nor does he give any arguments how such are to be distinguished from old river courses. I cannot believe in the body of fresh water, which must on the lake theory have flowed out of them. At the pass of Muckul, he states that the outlet is 70 feet wide & the rocky bottom 21 feet below the level of the shelf, & that the gorge expands to the eastwards, “into a broad channel of several hundred yards in width, divided in the middle by what has formerly been a rocky islet, against which the waters of this large river had chafed in issuing from the pass.”17 We know the size of the river at the present day which would flow out through this pass, & it seems to me (& in the other given cases) to be as inadequate: the whole seems to me far easier explained by a tideway, than by a formerly more humid climate.

With respect to the very remarkable coincidence between the shelves & the outlets (rendered more remarkable by Mr Milne’s discovery of the outlet to the intermediate shelf at Glen Glastig),18 Mr Milne gives only half of my explanation; he alludes to (& disputes) the smoothing & silting up action, which I still believe in. I state “If we consider what must take place during the gradual rise of a group of islands, we shall have the currents endeavouring to cut down & deepen some shallow parts in the channels, as they are successively brought near the surface, but tending from the opposition of tides to choke up others with littoral deposits. During a long interval of rest, from the length of time allowed to the above processes, the tendency would often prove effective, both in forming, by accumulation of matter, isthmuses, & in keeping open channels. Hence such isthmuses & channels just kept open, would oftener be formed at the level, which the waters held at the interval of rest, than at any one other p. 65”19

I look at the pass of Muckul (21 ft deep, Milne) as a channel just kept open, & the head of Glen Roy (where there is a great Bay silted up) & of Kilfinnin,20 (at both which places, there are level-topped mounds of detritus above the level of the terraces) as instances of channels filled up at the stationary levels. I have long thought it a probable conjecture, that when a rising surface becomes stationary, it becomes so not at once, but by the movements first becoming very slow: this would greatly favour the cutting down many gaps in the mountains to the level of the stationary periods.—

Glacial Theory.— If a glacialist admitted that the Sea before the formation of the terraces covered the country (which would account for “land-straits” above level of terraces) & that the land gradually emerged; & if he supposed his lakes were banked by ice alone, he would make out, in my opinion, the best case against the marine origin of the terraces. From the scattered boulders & till, you & I must look at it as certain that the sea did cover the whole country; & I abide quite by my arguments from the buttresses &c &c21 that water of some kind receded slowly from the valleys of Lochaber (I presume Mr Milne admits this). Now I do not believe in the ice-lake theory, from the following weak, but accumulated reasons because 1st the receding water must have been that of a lake in Glen Spean & of the sea in the other valleys of Scotland, where I saw similar buttresses at many levels. 2d because the outlets of the supposed lakes, as already stated, seem from Mr Milnes statements too much worn & too large. 3d When the lake stood at the 34 of mile shelf, the water from it must have flowed over ice itself for a very long time & kept at the same exact level: certainly this shelf required a long time for its formation 4th I cannot believe a glacier cd have blocked up the short, very wide, valley of Kilfinnin, the Great Glen of Scotland, also, being very low there. 5th The country at some places, where Mr Milne has described terraces, is not mountainous, & the number of ice-lakes appears to me very improbable. 6th I do not believe any lake could scoop the rocks22 so much as they are at the entrance of Loch Treig, or cut them off at the head of upper Glen Roy 7th the very gradual dying away of the terraces at the mouth of Glen Roy, does not look like a barrier of any kind.— 7th I shd have expected great terminal moraines across the mouth of Glen Roy, Glen Collarig & Glastig, at least at the bottoms of the valleys; such I feel pretty sure do not exist.23

I fear I must have wearied you with the length of this letter, which I have not had time to arrange properly.— I could argue at great length against Mr Milnes theory of barriers of detritus,24 though I could help him in one way, viz by the soundings which occur at the entrances of the deepest fiords in T. del Fuego.25 I do not think he gives the smallest satisfaction with respect to the successive & comparatively sudden breakage of his many lakes.

Well, I enjoyed my trip to Glen Roy very much, but it was time thrown away: I heartily wish you would go there: it should be some one who knows glacier & iceberg action & sea-action well: I wish the Queen would command you.— I had intended being in London tomorrow, but one of my painful plagues will, I believe, stop me; if I do I will assuredly call on you. I have not yet read Mr Milne on Elevation, so will keep his paper for a day or two.26

Farewell, my dear Lyell, ever most truly yours C. Darwin

P.S. As you cannot want this letter, I wish you would return it me, as it will serve as a memorandum for me—possibly I shall write to Mr Chambers,27 though I do not know whether he will care what I think on the subject. This letter is too long & ill-written for Sir J. Clark


The date is based on the assumption that the paper CD refers to in the letter is Milne 1847b. CD received a copy of Milne 1847b in September. See letter from Robert Chambers to David Milne, 7 September 1847, in which CD is referred to as being anxious to obtain a copy.
Milne 1847b. The page references in the letter correspond to those of the privately printed memoir on the parallel roads of Lochaber. An annotated copy is in DAR 139: 5. References in the notes below are to Milne 1849, the version published in the Transactions of the Royal Society of Edinburgh.
Because of Lyell’s poor eyesight, his wife Mary Elizabeth frequently read to him. This letter was apparently returned to CD, as requested in the postscript, but a copy of it was also made by an unknown amanuensis, presumably for Lyell to retain. The copy is preserved among the letters to Lyell at the American Philosophical Society (APS 68).
In the original letter CD wrote ‘Glen Roy’, with ‘p. 13.’ interlined after ‘Glen’ and ‘& Spean’ interlined after ‘Roy’. This text is followed in the copy. Later, apparently, CD deleted ‘Roy’ and the ampersand, so that the text then read ‘Glen p. 13. Spean’. However, Milne referred only to Glen Roy on page thirteen, so it appears that CD deleted ‘Roy’ and the ampersand by mistake. The deletion is ignored in the copy. In his discussion of Milne’s paper CD followed a list of rough notes and comments he had made as he read it. These notes, together with a large number of others, some of which are labelled ‘Glen Roy scraps’, are preserved in DAR 50 (ser. 3): 7–30.
In ‘Observations on the parallel roads of Glen Roy’ CD stated that the marine theory of the origin of the roads ‘appears to me demonstrated’ (Collected papers 1: 118). Milne, in his paper, regretted that CD ‘should have expressed himself in these very decided and confident terms, especially as his survey was incomplete’ (Milne 1849, p. 400).
MacCulloch 1817 and Dick 1823.
CD alludes to the lack of any remnants of earthen barriers.
Collected papers 1: 98.
According to Louis Agassiz and William Buckland, the parallel roads were formed as a result of a glacier damming up the valley and forming a glacial lake, of which the roads were former beaches. See Agassiz 1840b, 1840c, and 1842b, pp. 237–40. CD rejected this theory mainly because of the absence of a known outlet for the lake, except over the glacier itself. CD could not accept the possibility that a glacier could have existed at a particular level long enough to create a ‘road’ without having been worn down by the waterflow over it. See Correspondence vol. 2, letter to Charles Lyell, [9 March 1841].
The four figures were obtained by four different observers in Glen Gluoy and Glen Roy (see Milne 1849, p. 397). CD wrote ‘Glen luy’, but Milne’s text makes clear that ‘Gluoy’ was intended.
On the assumption that the roads were of marine origin, the height should have been the same. See Collected papers 1: 116 and Milne 1849, p. 401.
Collected papers 1: 111–14.
See Milne 1849, pp. 402, 414–17, and Collected papers 1: 111–13.
Bravais 1845, in which Auguste Bravais stated the need for accurately measured levels of ancient coast lines, and described a method of measuring them based on present mean sea-levels, in turn determined by the upper range of littoral algae. He discussed the problems of measurement at Glen Roy (pp. 542–4). See also Milne 1849, p. 396, which refers to Bravais 1845.
James Clark, physician to Queen Victoria and friend of Lyell, was probably accompanying the Royal family on their visit to a Highland lodge located on the shores of Loch Laggan, in the vicinity of Glen Roy and Glen Spean. The visit took place from 11 August to 17 September, see Annual Register 1847, Chronicle pp. 116–18.
CD argued that icebergs floating in a narrow strait were capable of scraping and polishing the sea bottom. See Journal of researches, pp. 617–25.
Milne 1849, p. 399; for his discussion of the river courses, see pp. 398–400. The Pass of Mukkull is at the eastern end of Glen Spean.
Milne had discovered a col in adjacent Glen Glaster (not ‘Glastig’, as CD has it) at the level of Glen Roy’s middle shelf. Earlier supporters of the lake hypothesis, as well as CD, had overlooked it. This discovery of a likely drainage point removed a powerful objection to the lake theory (see Milne 1849, p. 398).
Collected papers 1: 115. CD’s argument, which he hoped would render any coincidences of height between ‘roads’ and cols irrelevant, was that the cols were ‘land straits’, which emerged between branches of the sea as the land rose and which were maintained just at the level of the water through choking by littoral deposits.
CD described Kilfinnin in Collected papers 1: 92, 132 n. 10. See also CD’s map of the Glen Roy area, p. 87.
CD observed that these ‘buttresses’ had been formed at various levels on the sides of the valley. Some of them were associated with, and formed extensions to, the shelves, while others were smaller and unconnected with any shelf. The buttresses were found along the courses of streamlets descending the valley, which led CD to conclude that they had been formed by the deposition of detritus when the descending stream met the edge of the lake filling the valley. If the water-level in the valley had stayed constant for a long period, both shelves and buttresses would have been formed; if it had remained at the same level for a shorter period then buttresses alone would have been formed. See Collected papers 1: 102–4.
At this point (presumably after the letter was returned, see n. 3, above) CD added in pencil, ‘Have Lakes much more power with ice’.
Louis Agassiz reported many evidences of glaciers at Glen Roy and neighbouring valleys, including polished rocks and moraine deposits (Agassiz 1842b, pp. 222, 236–9).
Milne 1849, pp. 404–9.
See South America, p. 24 n. CD described these ‘fiords’ as much shallower at their mouths than further inland. He attributed this to the accumulation of sediment formed by the wearing of the rocks exposed to the open sea. Such ‘sea channels’ resemble CD’s conception of Glen Roy’s ancient waterways, and this resemblance allowed CD to attribute any accumulation of detrital matter, such as those described by Milne (1849, p. 405), to the action of the sea.
The last section of Milne 1847b is devoted to showing that Scotland ‘to the depth of at least 1200 feet, has been recently immersed in the ocean, and for a very long period’ (p. 25). Milne concluded, however, that if a survey revealed that raised beaches along the coasts of Scotland, Ireland, England, and France were continuous and horizontal, the cause of these and of the parallel roads would be changing levels of the ocean rather than of the land (p. 54).
Letter to Robert Chambers, 11 September 1847.

Letter details

Letter no.
Darwin, C. R.
Lyell, Charles
Sent from
Physical description
8pp & 6pp &


Discusses David Milne’s Glen Roy paper ["On the parallel roads of Lochaber", Trans. R. Soc. Edinburgh 16 (1849): 395–418]. Rejects Milne’s theory that outlet of Glen Roy is blocked by detritus. Impressed by Milne’s discovery of an outlet at the level of the second shelf. Believes this strengthens theory that lakes were formed by glacier blocking Glen Roy. Offers arguments against glacier theory.

Please cite as

Darwin Correspondence Project, “Letter no. 1116,” accessed on 13 February 2016,