Alfred Russel Wallace : Alfred Wallace : A. R. Wallace :
Russel Wallace : Alfred Russell Wallace (sic)

Letters to "Nature" on the Glacial Origin
of Alpine Lakes (1893-1894): S462, S472,
S482, S484, S486, S487, and S489

Editor Charles H. Smith's Note: A series of letters to the Editor of the journal Nature responding to various discussions concerning the possible glacial origin of alpine lakes. Original paginations indicated within double brackets. To link directly to this page, connect with:

The Glacier Theory of Alpine Lakes (S462: 1893)

    [[p. 437]] The letter of the Duke of Argyll against the theory of the formation of alpine lakes by glacial action shows such an amount of misconception of the theory itself, and so completely ignores the great weight of evidence in its favour, that a few words on the other side seem desirable.

    The Duke says that glaciers "do not dig out," do not "act like a ploughshare," but, when moving down a slight incline do "scoop," as well as rub down and abrade. No observer of glaciers has ever stated, so far as I know, that they do "dig out," and it is equally erroneous to say that they "scoop," for that implies that it is the end of the glacier that acts. But the end is its weakest point, where it is melting above and below, and where consequently it can do practically nothing. The whole action of a glacier is a grinding action, and its grinding power is greatest where it is thickest, and where, consequently, it presses on the rocks with the greatest weight. The result of this grinding is seen in the muddy stream issuing from all existing glaciers; while the well-known "till" is the product of the rock grinding mill of ancient glaciers and ice-sheets.

    Notwithstanding the Duke's disbelief in ice-sheets I venture to think that their former existence has been demonstrated both in Scotland and Ireland; but leaving this point, I wish to make a few remarks on the extreme inadequacy of the earth-movement theory to account for the facts. In the first place it is certain that no alpine lake can possibly have a long life, geologically speaking. In the course of a few thousands of years, certainly in less than a hundred thousand, all alpine lakes would be filled up by the sediment brought into them. It follows that all the existing lakes must have been formed about the same period, and that, geologically, a very recent one, and corresponding approximately with that of the well-known glacial epoch. But if these lakes were all formed by earth movements, either just before the glacial epoch came on, or during its continuance, or afterwards we have to explain the remarkable fact that such movements only occurred within the limits of glaciation, never beyond those limits. In Wales, Cumberland, and Scotland, in the Alps, in Scandinavia, in Finland, in the northern United States and Canada, in Mongolia and Thibet, in Tasmania and New Zealand, we have thousands of rock-basin lakes, amid palpable signs of glaciation. But the moment we pass beyond the glaciated districts, mountain lakes abruptly cease. There are hardly any in Spain, none in the Great Atlas, none in Sardinia or southern Italy, except in the volcanic areas and away from the mountains, none in any of the West Indian islands with their fine mountain-ranges, none in the peninsula of India or in Brazil. And there is exactly the same distribution of fiords. We have them in Norway, in West Scotland, in Alaska, in South-West America, and in New Zealand, all characterised by deeper water within than at their outlets, and all in glaciated countries, but nowhere else in the world.

    Now it is simply impossible to believe that at a very recent period there should have been earth-movements of such a character as to produce lakes, but always in glaciated districts and never beyond them, unless the movements were a result of the glaciation. This has not, I believe, been yet suggested; but, in view of the modern theory that any considerable loading of the surface produces subsidence, it is at least a possible explanation. But there are some important facts that seem more in favour of the grinding out of the lake-basins by the enormous weight of ice accumulated over their sites during the height of the ice-age. Looking at a geological map of the Alps it will be seen that most of the lakes are more or less bordered by tertiary or secondary rocks. Lakes Annecy and Bourget are in miocene [[p. 438]] and eocene; the lake of Geneva on the north side is miocene or jurassic; the lake of Neuchatel, miocene; lakes Thun and Brienz, eocene or jurassic; lake Lucerne, eocene and miocene; lakes Zug and Zurich in miocene; lake Constance miocene; lake Maggiore is mostly in gneiss, but it is very suggestive that it is here comparatively shallow, but becomes suddenly deeper and reaches its maximum depth in its lower portion where it is bordered on the east by the jurassic beds; lake Como also has its greatest depth in triassic rocks, the upper portion, where gneiss prevails, deepening gradually southward as in a submerged valley. Equally suggestive is the fact that in the eastern Alps of Tyrol and Carinthia, where gneiss, porphyry, and the older stratified rocks prevail, and where glaciers are not now so extensive, there are hardly any lakes, except on the northern borders, where a considerable number occur in eocene, cretaceous, jurassic, or triassic formations.

    These various facts as to the distribution of alpine lakes--their almost total absence in all parts of the world outside of glaciated districts, and within glaciated districts their prevalence in the newer and more easily denuded rocks--are what have to be explained by the advocates of the theory of earth-movements, and this, so far as I am aware, they have never attempted to do. Equally important, and equally difficult to explain on the earth-movement theory, is the fact that alpine lakes are almost always situated just at those spots where, by means of converging valleys, the glaciers would become heaped up and attain their maximum thickness, or where there is good evidence that they have been very thick; and it is the grinding power of this enormous weight of ice, acting differentially as regards the softer and harder rocks, that has worn out hollows in pre-existing valleys now occupied by lakes. In almost every case, too, it will be seen that there is a constriction or narrowing of the valley towards or beyond the lower end of the lake, which, by preventing the free escape of the ice, has increased its thickness and grinding power.

    In the presence of such important series of facts as those here referred to, mere opinions, or even small and detailed cases of difficulty, can have no weight; but there is yet another consideration, which most geologists will admit is antagonistic to the earth-movement theory. The whole tendency of geological observation is in favour of the usually very slow rate of earth-movements, while it is equally in favour of the comparatively rapid action of denudation by running water. But in order that earth-movement could form a lake, it would be necessary that the rate of elevation or depression should be so great that the river could not keep pace with it by cutting down its channel; and, considering that all the rivers in question are rapid mountain streams carrying great quantities of sediment, this will be admitted to be a very improbable supposition. But when we add to this the still greater improbability that such rapid earth movements have occurred in scores and hundreds of cases, all at about the same time, geologically speaking, and all just in those spots where it can be shown that during the glacial period ice must have accumulated, and where the rocks were of such a character as to admit of being ground away; and yet further, that no similar earth movements producing similar results have recently occurred in any part of the globe beyond the limits of glaciation, the whole assumption becomes so hugely improbable as to render the theory of lake-formation by ice-grinding easy in comparison.

    Sir Charles Lyell considered that the gravest objection to the glacial-erosion theory was the entire absence of lakes where they ought apparently to exist; and he instanced the valley of Aosta and the Dora Baltea, the glacier of which produced the enormous moraines of Ivrea. The valley of the Rhone above Martigny may be adduced as another example of the absence of lakes where they might be expected. But this kind of difficulty will apply to many other valleys, and can only be answered by general considerations. In both these cases the valleys are comparatively broad and open, and have a rather rapid descent. It is probable, therefore, that the ancient glacier in both was of a nearly uniform thickness, so that its wearing action on the floor of the valley would be tolerably uniform. To produce a lake we require essentially a differential action. There must be much more rapid degradation in one part than in another, due either to greater ice-accumulation or to softer rocks in one part than in another. In both the valleys referred to there is much uniformity in the rock-formations throughout, and even if some lakes or chains of lakes had been formed, the enormous amount of debris still brought down may well have filled up and altogether obliterated them. The absence of lakes in certain valleys cannot be considered an argument of any value until it is ascertained by borings that none have been formed and filled up again. It must also be shown that the whole conditions are such as to produce that amount of differential grinding down, without which no lake can be expected to have been formed.

    It certainly seems to me that all the facts, all the probabilities, all the converging lines of evidence, are in favour of the glacial theory, to which the only serious objection is the assumption that glaciers cannot move uphill. But that they can do so, and have done so, is now admitted by most students of glacier-motion. Mr. Jamieson, and other Scotch geologists, have proved that glaciers, over 2000 feet thick, have travelled up lateral valleys, and up the slopes of many hills and mountains; and when we consider that the Rhone glacier was 5000 feet thick just above the lake of Geneva, and more than 2000 feet thick where it abutted against the Jura, we can have no difficulty in admitting that it might have travelled up the very gentle slope of the lake bottom, which appears to be less than 100 feet in a mile in its steepest parts.

*                *                *                *                *

The Glacier Theory of Alpine Lakes (S472: 1893)

    [[p. 198]] The Editor having given me the opportunity of reading Mr. Graham Officer's interesting letter, I will make a few remarks upon it.

    It seems to me that, without further information as to the nature of the search for drift, erratics, or ice-worn surfaces, and judging from the statement that the plateau studded with lakes and tarns was only looked down upon from an adjacent mountain summit, we can hardly give much weight to the positive statements, "I am confident that evidences of glaciation do not exist," and--"as I have shown, the glacier theory will not account for by far the greater number of the Alpine lakes on the great central greenstone plateau." Some light may perhaps be thrown on the matter by the consideration that the undoubted marks of glaciation in many parts of Australia are believed to have been caused by, comparatively, very ancient glaciers, since some of the glaciated surfaces are overlain by pliocene deposits, while others are believed to be of palæozoic age. If the Tasmanian glaciation was also of pliocene age, most of the superficial indications may have been destroyed by denudation, or, if preserved, may be hidden by vegetation or by alluvial deposits. We must therefore wait for a much more thorough examination of the district and of other parts of Alpine Tasmania before it can be positively stated that no evidences of glaciation exist.

*                *                *                *                *

The Recent Glaciation of Tasmania (S482: 1893)

    [[p. 3]] In a paper read before the Royal Society of Tasmania in June last, Mr. R. M. Johnston, F. L. S., gives a sketch of what is known of the glaciation of the island, or rather of the western portion of it, for no indications of glaciers appear to have been discovered in the eastern half. This difference is supposed to be due to the fact that on the western side of the island the rainfall is from 50 to 76 inches annually, while in the central valley it is but little over 20 inches. Indications of glaciation among the western mountains were noticed by Mr. Charles Gould, Government geologist, about forty years ago, and from information received from him through the late Chief Secretary of Tasmania, the Hon. J. R. Scott, Mr. Johnston took up the inquiry, and for many years has made explorations in the western plateaus and mountains. Mr. C. P. Sprent was another explorer who published some account of the glacial phenomena in 1886, while more recently Mr. T. B. Moore and Mr. Dunn have recorded similar observations. Mr. A. Montgomery, the present Government geologist, has also just published a paper on the same subject.

    Mr. Johnston tells us that he has personally explored the whole of the western mountains, from the Picton and Craycroft Rivers, southern branches of the Huon, in the extreme south, along the mountain ranges forming the western border of the central plateau, quite through to Emu Bay on the north coast; and that he has found the clearest evidences of glaciation in almost every valley throughout this great extent of country. From the Arthur Range in the south to Mount Bischoff in the north, are numerous moraines, roches moutonnées, tarns and lakes in great abundance, polished and striated rock-surfaces, and numbers of true erratics. Near the sources of the Franklin River, under Mount Hugel, and only six or seven miles west of Lake St. Clair, are Lakes Dixon and Undine, of which Mr. Johnston writes:--"The valley of Lake Dixon is par excellence, the ideal of a perfect glacier valley. No one, however ignorant of glacial action, could in this neighbourhood gaze upon these beautiful scooped, or rather abraded lakes or tarns, the snow-white, polished, billowy, and cascade-like roches moutonnées, composed of quartzites, on the upper margin of Lake Dixon, together with the tumbled moraines and large erratics on the lower banks--at a level of about 2000 feet--without being impressed with the idea that its singularly characteristic features must have been produced by the slow rasping flow of an ancient river of ice."

    Further north, the Murchison, Macintosh and Huskisson rivers, all branches of the Pieman River, contain similar glacial markings; and Mr. Dunn has recently described others of the same character about Lake Dora, nearer to the west coast. The latter observer lays special stress on the rounded planed and scored rocks, on hard quartzite and conglomerate rocks rounded and polished, on numerous tarns in rock-basins, on moraines covering hundreds of acres, and on numerous huge erratics and perched blocks. (See Annual Report of the Secretary for Mines, Victoria, 1893, p. 21.)

    Mr. T. B. Moore states that he found the rocks polished and striated within 25 feet of the top of Mount Tyndall, or 3850 feet above the sea, a sufficient indication that the great central plateau at an average elevation of nearly 4000 feet must have been buried in ice or névé to a considerable depth, and have formed the feeding ground for the glaciers, whose effects are so visible in the adjacent western valleys. The Tasmanian geologists are united in the belief that the glaciers never reached the coast or descended much below the 2000 feet level, and that the ice did not extend to the central valley or the eastern side of the island. They therefore speak of it as a glacier, not a glacial period, the conditions being somewhat similar to those of the Alps at the present time; but, owing to the great difference in the rainfall, there was a more marked contrast between the western and eastern districts, while the lofty central plateau afforded a much more extensive snow-field than Switzerland now possesses.

    The facts here stated on the authority of Mr. Johnston, [[p. 4]] supported by those of three other observers, two of them being the Government geologists, render more singular the statements of Messrs. Officer and Spencer (Nature, June 29, p. 198) as to their not finding any traces of glaciation in the country around Lake St. Clair, which they explored for a month. Lake Dixon, which Mr. Johnston describes as presenting all the evidences of glaciation in their fullest development, appears to be less than ten miles from the lower end of Lake St. Clair, according to the best map I can refer to; while Lake Petrarch, which Mr. Officer describes as seeing from the top of Mount Olympus, lies between the two in the Cuvier valley, and is also mentioned by Mr. Johnston as being within the highly-glaciated region. It is quite possible that the lakes on the great plateau may be due to damming up, owing to movements of the superficial gravels and clays by the ice or névé sheet; but there are evidently an abundance of small valley-lakes and tarns in the western valleys so surrounded by all the marks of extensive glaciation as to render it almost certain that they are true ice-eroded rock basins. It is much to be wished that a more detailed account of this interesting district, with a good map showing all the mountains, lakes, and valleys referred to, would be given us by one of the local geologists.

*                *                *                *                *

Sir Henry H. Howorth on "Geology in Nubibus" (S484: 1893)

    [[p. 52]] Having given my views on glacial geology in the current issue of the Fortnightly Review, to be followed by one dealing at some length with the ice-origin of lake-basins, I should not have thought any reply to Sir Henry Howorth's "Appeal" necessary except for the consideration that my articles may not be seen by many readers of Nature. And first, I would remark, that the mental attitude which Sir H. Howorth imputes to extreme glacialists I have myself been unable to detect in their writings. In fact, I was under the impression that the "scoffing" and "jeering" was chiefly from the other side; but it seems I was mistaken, and I must apologise for my ignorance. Those who read my articles will see that I make no appeal to "transcendental ice," but judge of its powers and properties by its admitted effects. Sir H. Howorth says that "ice is known to crush under moderate pressure," implying that a glacier a mile or perhaps half a mile thick is impossible. But will he or anyone else tell us what happens to the ice after it is crushed, and the pressure that crushed it is continued and slowly increased? Will it not suffer re-gelation and become denser ice; and if by sudden increase of pressure it is again crushed, will it not by still further pressure again suffer re-gelation? He stops at the first "crushing," as if that were the end of all things so far as a glacier is concerned. All this, however, is beside the question from my point of view. The work of ice on the rocks is as clear as that of palæolithic man on the flints; all the difficulties that may be suggested as to how he lived, or how he shaped the flints do not in the slightest degree affect our conclusion that the palæolithic flint implements are the work of man; and there is equally clear evidence that ice did march a hundred miles, mostly uphill, from the head of Lake Geneva to Soleure, whatever transcendental qualities it must have possessed to do so.

    As to "perhaps the largest and most remarkable collection of rock-basins in the world"--the largest being of 50 acres and the deepest 30 feet deep--I must really decline to occupy your space in showing how simply these may have been produced by ordinary denuding agencies, or in denying that any glacialist, even of "the most extreme and aggressive school," would claim them as proofs of glaciation. As regards the question of Tasmanian glaciation, my last communication to Nature (Nov. 2) seems to me to render any further observations unnecessary. No doubt the conclusions of the various writers will be fully harmonised by a more complete study of the whole region.

     The last point touched on by Sir H. Howorth--whether the advocates of the ice-origin of certain groups of lakes are "extravagant" in their views, following the methods of Aristotle rather than those of Bacon, and founding their beliefs on "purely hypothetical properties of matter and forces of nature"--I will leave to the judgment of those who do me the honour of reading my forthcoming article in the Fortnightly Review.

*                *                *                *                *

"Geology in Nubibus" (S486: 1893)

    [[p. 101]] Sir Henry Howorth wishes to continue the discussion of glaciation in the pages of Nature, but I find in his last letter very good reason why this cannot be done. No discussion can lead to definite results unless the parties to it accept as data what they themselves have recently and deliberately admitted. But when I stated that the Rhone glacier did reach the Jura, and deposit on it erratic blocks between Geneva and Soleure, I did so because it was one of the data already admitted by Sir H. Howorth. In his "Glacial Nightmare," pp. 169-173, he gives a full summary of Charpentier's first memoir on the erratic blocks of Switzerland, describing the glacial phenomena exhibited along the whole course of the old glaciers from the Alps to the Jura, and showing that they "even climbed that range and went over to the other side of it." Sir H. Howorth then says: "I have quoted at considerable length from this excellent memoir, because I look upon it as having definitely applied inductive methods to this question with results which are for the most part sound and unanswerable." (Italics mine.) In the same chapter (pp. 195-202) Charpentier's second memoir is summarised still more fully, and his general conclusion is thus quoted: "It goes without saying that not only all the valleys of the Valais were filled with ice up to a certain height, but that all lower Switzerland, in which we find the erratic débris of the Rhone valley, must have been covered by the same glacier. Consequently all the country between the Alps and the Jura, and between the environs of Geneva and those of Soleure has been the bed of a glacier." Agassiz and other writers are quoted as giving further evidence of the same kind. Nowhere in the whole of this chapter can I find a single objection to the conclusions of the chief writers quoted, and the concluding paragraph, at p. 208, frankly accepts them. It declares that they are supported by "every form of converging evidence," and that--"So far there is no question at issue." Yet, when I take these same conclusions of Charpentier as admitted data, Sir H. Howorth says: "This form of dogmatic argument is assuredly incomprehensible!" Charpentier's proof that the Rhone glacier reached Soleure, was, a year ago, "sound and unanswerable," and was an example of "definitely applied inductive methods"; but when I accept these same results as something to reason upon, I am told that I am making use of "hypotheses outside the laws of nature." I have now justified my opening statement that a discussion carried on in this manner can serve no useful purpose.

*                *                *                *                *

The Origin of Lake Basins (S487: 1893)

    [[p. 197]] In his last communication Sir Henry Howorth makes two statements which are so erroneous and so misleading that I cannot allow them to pass without correction. The first is, that Mr. Deeley "repudiates Dr. Wallace's notion that regelation can in some way act as a compensating element when crushing supervenes in ice." Here is a double misstatement. Mr. Deeley "repudiated" no notion of mine, or he would, I am sure, have said so plainly, and he said nothing whatever about "crushing." Neither did I say a word about regelation acting as a "compensating element," for I do not believe in the crushing of glaciers by their own pressure. I asked Sir Henry what would happen to the ice after it was crushed, the pressure continuing; and I get no reply but the above double misstatement.

    Then, further on, Sir Henry says: "Mr. Wallace confesses he does not like to face these mechanical issues." This is simply untrue. I "confessed" nothing of the kind, and I challenge Sir Henry Howorth to quote any words of mine which will bear such a meaning. I maintain that his "mechanical issues" are pure theories, and are beside the question of the actual facts of glacier motion. Lastly, he attempts to evade the real issue between us, which is, that he himself accepted Charpentier's conclusions as to the extent of the Rhone glacier, but refuses to allow me to use these same conclusions as a datum in the discussion.

    I have now shown ample reason why further discussion of this matter with Sir Henry Howorth must be unprofitable.

*                *                *                *                *

The Origin of Lake Basins (S489: 1894)

    [[p. 220]] I welcome the criticism of my article on the glacial origin of a certain class of lakes by an experienced geologist like Mr. Oldham, because it probably embodies the strongest argument that can be adduced on the other side--at all events as regards [[p. 221]] the one aspect of the problem which he alone touches upon. He urges that my paper contains a fallacy and a misrepresentation. The alleged fallacy is, that because the lakes in question are found in glaciated and not in otherwise similar non-glaciated regions, "therefore the rock-basins in which the lakes lie were excavated by glaciers." But this is not my argument, and therefore not my fallacy. What I say is--"there must be some causal connection between glaciation and these special types of lakes. What the connection is we shall enquire later on." That there is a "causal connection" Mr. Oldham asserts as strongly as I do myself, though it is a different, and as I have endeavoured to show, an untenable one.

    This brings us to the alleged misrepresentation, which is, that I have imputed to the opponents of the ice-erosion theory, the view that the earth movements which, as they allege, produced the lakes, occurred in the period just before the ice-age came on. Mr. Oldham says, this is an unreasonable and unfounded limitation, since the movements in question probably occurred throughout the glacial period itself. I quite admit the validity of this criticism, and that I should have added, "or during the glacial period itself," to, "immediately before" it. I certainly had this probability in my mind, and the reason I did not express it was twofold. In the first place, all the advocates of the earth-movement theory appeared to assume, either directly or implicitly, the preglacial origin of the lakes; and secondly, this assumption gave them the strongest argument against my views, and I therefore gave them the benefit of it. Mr. Oldham appears to have overlooked this. Yet it is clear that the shorter you make the time since the formation of lake basins by earth-movements the more difficulty there is in explaining the total absence of valley-lakes from all the non-glaciated mountain regions of the world, since there is less time for them to have been all silted up. When arguing this point I said--in the passage evidently referred to by Mr. Oldham--"The only way to get over the difficulty is to suppose that earth-movements of this nature occurred only at that one period, just before the ice-age came on, and the lakes produced by them in all other regions have since been filled up." I thus gave my opponents the benefit of an extreme supposition which was all against myself; while the more reasonable view, that earth-movements are just as likely to have occurred during and since the glacial epoch as before it, renders my argument from the geographical distribution of lakes much stronger, since it is impossible to believe that, if lake basins as large and as deep as those of Geneva, Maggiore, Como, Constance, and Garda, were formed in non-glaciated regions as recently as the middle or latter part of the glacial epoch, a considerable number of them would not be still in existence.

    Of course, if it can be shown that filled up lake-basins exist in tropical and subtropical regions, corresponding in number, position, size, and depth, with those of glaciated areas, the argument from geographical distribution will break down. At present I am not aware of any evidence that such is the case. But even if it were so, there remains the singular correlation between the size and depth of lake basins and the known size of the glaciers that occupied these valleys; together with the surface and bottom contours of the lakes themselves, so strongly opposed to their production by any form of valley-subsidence or earth-movements.

    A friend has pointed out an unsound argument in my article on the above subject in the Fortnightly Review, and I therefore ask to be allowed to state what it is, and thus avoid its being possibly made the subject of discussion in the pages of Nature. As a proof of the very great erosive power of ice I have adduced Dr. Helland's estimate of the quantity of Scandinavian débris in Northern Europe. But it is evident that this only proves the great carrying power of the ice, since the rock and gravel would be mostly of sub-aerial origin. It, however, indicates a very long period during which the ice-sheet was at work, while the clayey element in it would be due to erosion. The larger part of this, however, would certainly have been carried away into the North Sea during the passage of the ice-sheet across the Baltic. The enormous quantity of boulder-clay in North America, which I have also referred to, is a better indication of true ice-erosion.

*                *                *                *                *

Return to Home