[[p. 616]] It is little more than fifty years ago that one of the most potent agents in modifying the surface features of our country was first recognised. Before 1840, when Agassiz accompanied Buckland to Scotland, the Lake District, and Wales, discovering everywhere the same indications of the former presence of glaciers as are to be found so abundantly in Switzerland, no geologist had conceived the possibility of a recent glacial epoch in the temperate portion of the northern hemisphere. From that year, however, a new science came into existence, and it was recognised that only by a careful study of existing glaciers, of the nature of the work they now do, and of the indications of the work they have done in past ages, could we explain many curious phenomena that had hitherto been vaguely regarded as indications of diluvial agency. One of the first fruits of the new science was the conversion of the author of Reliquiæ Diluvianæ--Dr. Buckland, who, having studied the work of glaciers in Switzerland in company with Agassiz, became convinced that numerous phenomena he had observed in this country could only be due to the very same causes. In November, 1840, he read a paper before the Geological Society on the "Evidences of Glaciers in Scotland and the North of England," and from that time to the present the study of glaciers and of their work has been systematically pursued with a large amount of success. One after another crude theories have been abandoned, facts have steadily accumulated, and their logical though cautious interpretation has led to a considerable body of well-supported inductions on which the new science is becoming firmly established. Some of the most important and far-reaching of these inductions are, however, still denied by writers who have a wide acquaintance with modern glaciers; and as several works have recently appeared on both sides of the controversy, the time seems appropriate for a popular sketch of the progress of the glacial theory, together with a more detailed discussion of some of the most disputed points as to which it seems to the present writer that sound reasoning is even more required than the further accumulation of facts.¹

    [[p. 617]] In the last century, Swedenborg, Linnæus, Pallas, De Luc, and many other eminent writers took notice of the remarkable fact that in Scandinavia, Russia, Germany, and Switzerland detached rocks or boulders were found, often in great abundance and of immense size, and of a kind that did not exist in situ in the same district, but which were often only to be discovered in remote localities, sometimes hundreds of miles away. Those who ventured to speculate on the origin of these travelled rocks usually had recourse to water-power to account for their removal; and as their large size and often elevated position required some unusual force to carry them, there arose the idea of enormous floods sweeping over whole continents; and for a long time this diluvial theory was the only one that appeared to be available, although the difficulties of its application to explain all the phenomena became greater the more closely those phenomena were studied. Still, there was apparently no other known or conceivable means of accounting for them, and for the enormous mounds of gravel or clay intermixed with boulders which often accompanied them; and the efforts of geologists were therefore directed to the discovery of how the water-power had acted, and by what means the supposed floods could have been produced.

    There were not wanting men who saw that no action of water alone could account for the facts. Sir James Hall pointed this out with regard to erratics on the Jura, whose source was undoubtedly in the far-distant Alps; and Mr. Grainger, in America, described some of the parallel grooves and flutings running for nearly a mile in Ohio, strongly arguing that no action of running water could have produced them, but that an agent was required the direction of whose movement was fixed and unalterable for long distances and for a great length of time. No light was, however, thrown on the problem till 1822, when Venetz, a Swiss engineer, finding that existing glaciers varied in extent from year to year and that historical records showed them to have considerably increased during the last eight centuries, was further led to observe that, long before the historical era the glaciers had been immensely more extensive, as shown by the smooth and rounded rocks, by longitudinal scratches and grooves pointing down the valleys, and by numbers of old moraines exactly similar in form and materials to those deposited by existing glaciers. He read a paper before the Helvetic Society of Natural History, and urged that glaciers once stretched down the Rhone valley as far as the Jura, and there deposited the erratic blocks which had so puzzled the diluvialists to explain.

    Other writers soon followed the clue thus given. In 1835 Charpentier, after a close study of the erratic blocks and of their sources, adopted the views of Venetz. Agassiz followed, and by his strenuous advocacy did much to spread correct views as to the former [[p. 618]] extension of the Alpine glaciers, and their capability of explaining the numerous superficial phenomena which in all northern countries had been thought to afford proofs of enormous floods and of the submergence of a large part of Europe under a deep sea. He has, therefore, gained the reputation of being the originator of the modern school of glacialists, which undoubtedly owes much to his energy, research, and powers of exposition, though all the more important facts, as well as the logical conclusions to be drawn from them, had been pointed out by previous writers.

    Before proceeding further, it will be well to give a brief outline of the phenomena which lead to the conclusion that glaciers have formerly existed in districts and countries where even perpetual snow on the mountain tops is now unknown. These may be briefly classed as--(1) Moraines and drifts; (2) Rounded, smoothed, or planed rocks; (3) Striæ, grooves, and furrows on rock-surfaces; (4) Erratics and perched blocks.

    (1) Moraines are those heaps or ridges of rock and other debris which are deposited on the surface of a glacier from the precipices or mountain slopes which border it, and which form what are termed lateral and medial moraines while upon it, and terminal moraines when, being gradually discharged at its end, either from above or from beneath it, they form great heaps of rock and gravel corresponding in outline and extent to that of the terminal ice-cliff. Such moraines can be seen on and near all existing glaciers, and their mode of formation and characteristics are perfectly well known. If the glacier is continuously retreating, then the terminal moraine will form more or less irregular heaps over the surface the glacier has formerly covered; but when, as is usually the case, the glacier remains stationary for a considerable period, then the terminal moraine will have a definite form, and will often stretch quite across the valley, but presenting one or more openings through which the glacier stream has cut its way. Such moraines form steep mounds, usually curved and often very regular, seeming from a little distance to block up the valley like an artificial earthwork. Among hundreds that might be enumerated good examples may be seen in Glen Isla (Forfarshire), in the Troutbeck valley near Windermere, and in Cwm Glas, on the north side of Snowdon, this latter being so regularly curved, evenly sloped, and level-topped as to look from below exactly like an ancient fortification. The characteristic features of moraines are their position in valleys where there are other indications of glacial action, their steep slopes and often level tops, but especially their composition of earth, stones, and gravel, with large fragments of rock irregularly scattered through them from top to bottom without any sign of stratification, while usually one or more large blocks rest upon their summits in [[p. 619]] positions where they could only have been left by the retreat of the glacier, or possibly stranded from floating ice. Where extensive glaciers have covered large areas of nearly level ground the moraines form great sheets extending for many miles, often concealing the original contours of the country, and then receive the general name of drift. The composition of drift is usually the same as that of well-marked moraines, large blocks of stone being distributed throughout its mass. It is this which mainly distinguishes drift from alluvial or shore deposits, in which the materials are always more or less assorted and stratified; but the angular forms of many of the contained blocks and the striated surfaces of others are also characteristic. Besides the terminal moraines of extinct glaciers, lateral moraines are also left along the slopes of open valleys from which glaciers have retreated. As a whole, moraines are well distinguished from all accumulations formed by water, and it has not been shown that any other agency than glaciers is capable of forming them. In all recently glaciated countries they are to be found more or less frequently, and thus afford an excellent first indication of the former existence of glaciers.

    (2) Smoothed and rounded rocks, called in Switzerland "roches moutonnées," from their supposed resemblance at a distance to sheep lying down, are perhaps the most general of all the indications of glacial action. Every glacier carries with it, imbedded in its under surface, numbers of rocks and stones, which, during the slow but unceasing motion over its bed, crush and grind down all rocky projections, producing in the end gently rounded or almost flat surfaces even on the hardest and toughest rocks. In many of the valleys of Wales, the Lake District, and Scotland every exposed rock has acquired this characteristic outline, and the same feature can be traced on all the rocky slopes, and often on the summits of the lesser heights; and the explanation of how these forms have been produced is not a theory only, but has been observed in actual operation in the accessible portions of many glaciers. Rocks and stones are to be seen embedded in the ice and actually scratching, grooving, and grinding the rock beneath in their slow but irresistible onward motion. The rocky islets in Windermere, Ullswater, and other lakes, as well as the Thousand Islands of the St. Lawrence, are thus ice-ground; and the amount of the grinding can often be seen to be proportional to the pressure and motion of the advancing glacier. I recently noticed in the marshy alluvial plain above Derwentwater a projecting rock which has been ground down to so regular a curve as to look like a portion of an enormous globe buried in the earth. By rough measurement and estimate this rock was about 250 feet across, and 20 or 30 feet high. It was formed of hard slate, with numerous quartzite veins, the whole ground down to a uniform [[p. 620]] spherical surface. It had evidently once been an island in the lake, having a much broader base now hidden by the alluvium, and may originally have been one of those abrupt craggy rocks a few hundred feet high, which, owing to their superior hardness or tenacity, resisted ordinary denudation, and which, when above the old ice-level, form those numerous "pikes" which add so much to the wild and picturesque scenery of the district. Looking at such rocks as this, with outlines so utterly unlike any that are produced in similar formations by sub-aerial denudation--and they are to be seen by scores in all glaciated regions--we cannot but conclude that the ice-tool has done more than merely rub off the angles and minor prominences, and that it has really ground away rocky hills to an unknown but very considerable extent; and this conclusion is, as we shall see, supported by a very large amount of confirmatory evidence. It may be noted that ice-ground rocks usually show the direction in which the ice has moved, by the side opposed to the motion being more completely smoothed than the lee side, which often retains some of its ruggedness, having been protected partly by the ice overriding it and partly by the accumulation of its own debris. Where such rocks occur in the higher parts of valleys the smooth side always looks up the valley from which the glacier has descended. In the more open parts of valleys, or in high coombs or cirques, where two or more small ravines meet and where the ice may have been embayed and have acquired a somewhat rotary motion, the rocks are seen to be ground down on all sides into smooth mammillated mounds or hummocks, showing that the ice has been forced into all the irregularities of the surface. An example on a small scale is to be seen in Cwm Glas, on the north side of Snowdon, above the fine moraine already mentioned, and in many other places around the same mountain. On the whole, considering their abundance in all glaciated regions, and the amount of information they give as to the direction and grinding power of ice, these rounded rocks afford one of the most instructive indications of the former presence of glaciers; and we must also agree with the conclusion of Darwin (in a paper written after studying the phenomena of ice-action in North Wales, and while fresh from his observations of glaciers and icebergs in the southern hemisphere) that "one of the best criterions between the effects produced by the passage of glaciers and of icebergs is boss or dome-shaped rocks."

    (3) Striated, grooved, and fluted rocks, though closely connected with the preceding, form a distinct kind of evidence of the greatest value. Most of the bosses of rock just described have been exposed to the action of the atmosphere, perhaps since the ice left them, and have thus become more or less roughened or even disintegrated; but where the rocks have been protected by a covering of drift, or even [[p. 621]] of turf, and have been recently exposed, they often exhibit numerous parallel striæ, varying from the finest scratches to deep furrows a foot or more in diameter. Fine examples are to be seen near the lakes of Llanberis, and they occur more or less frequently in every glaciated country. Perhaps none of the effects of ice so clearly demonstrate the action of glaciers as opposed to that of icebergs, owing to the general constancy of the direction of the striæ, and the long distances they may be traced up and down slopes, with a steadiness of motion and evenness of cutting power which no floating mass could possibly exert. Sir A. Geikie tells us that in Gareloch, Bute, and Cantyre the striations on the rocks run up and over the ridges, and are as clearly shown on the hill-tops as in the valleys. Mr. D. Mackintosh states (in his paper on the "Ice-sheet of the Lake District and of North Wales") that in the valley above Windermere the striæ cross Rydal Fell, Loughrigg Fell, and Orrest Head, ascending and descending their slopes, often obliquely. But it is in the United States that the most remarkable rock-groovings are to be found, extending over a large portion of the North-eastern States. In his report on "The Rock-scorings of the Great Ice Invasions" Mr. T. C. Chamberlin gives many fine illustrations, from photographs, showing striæ and grooves along sloping, curved, or vertical surfaces, the striæ following the changes of curve, so that the grinding material must have been slowly forced into close contact with the irregular surface. Of one of these examples Mr. Chamberlin says:--

    "The climax of adaptability is reached in the striation of warped and twisted surfaces, and of tortuous valleys. One of the most remarkable known instances of this within the limits of photographic illustration is furnished by the great glacial grooves at Kelly's Island (Fig. 17). These exhibit not only the pliancy of the ice, but at the same time its strong hold upon the armature with which it did its work of abrasion, grooving, and striation. For, while these grooves can scarcely be supposed to have been originated de novo by the gouging action of the ice, they are, nevertheless, ploughed with deep furrows, the symmetry, continuity, and peculiar form of some of which are only intelligible on the supposition that they were cut by a single graving tool, held with sufficient tenacity by the ice to execute by a single movement a deep, sharply-defined groove. There is, perhaps, no finer illustration of the pliancy with which the ice yielded to its encompassing barriers, the tenacity with which it held its armature, and withal the pressure that both forced it into compliance with its tortuous channel, and pressed it relentlessly forward."²

    Kelly's Island is at the western end of Lake Erie, and in the direction of the striæ to the north-east there is no high ground for about 400 miles. Looking at these facts, I cannot give any weight to the opinions of these who, from observations of existing glaciers, [[p. 622]] declare positively that ice cannot go up-hill, and can exert no grinding power on level ground.

    (4) Erratic blocks were among the phenomena that first attracted the attention of men of science. Large masses of granite and hard metamorphic rock, which can be traced to Scandinavia, are found scattered over the plains of Denmark, Prussia, and Northern Germany, where they rest either on drift or on quite different formations of the Secondary or Tertiary periods. One of these blocks, estimated at 1,500 tons weight, lay in a marshy plain near St. Petersburg, and a portion of it was used for the pedestal of the statue of Peter the Great. In parts of North Germany they are so abundant as to hide the surface of the ground, being piled up in irregular masses forming hills of granite boulders, which are often covered with forests of pine, birch, and juniper. Far south, at Fürstenwalde south-east of Berlin, there was a huge block of Swedish red granite, from one half of which the gigantic basin was wrought which stands before the New Museum in that city. In Holstein there is a block of granite 20 feet in diameter; and it was noticed by De Luc that the largest blocks were often found at the greatest distance from the parent rock, and that this fact was conclusive against their having been brought to their present position by the action of floods.

    It is, however, in Switzerland that we find erratic blocks which furnish us with the most conclusive testimony to the former enormous extension of glaciers: and as these have been examined with the greatest care, and the facts, as well as the main inductions from the facts, are generally admitted by all modern writers, it will be well to consider them somewhat in detail. It will be found that they give us most valuable information both as to the depth and extension of ancient glaciers, and also as to the possibilities of motion in extensive ice-sheets.

    The most important of these facts relate to the erratic blocks from the higher Alps, which are found on the flanks of the Jura Mountains wholly formed of limestone, on which it is therefore easy to recognise the granites, slates, and old metamorphic rocks of the Alpine chain. These erratic blocks extend along the Jura range for a distance of 100 miles, and up to a height of 2,015 feet above the Lake of Neufchatel. The first important point to notice is that this highest elevation is attained at a spot exactly opposite, and in the same direction as, the Rhone valley, between Martigny and the head of the Lake of Geneva, while north or south of this point they gradually decline in elevation to about 500 feet above the lake. The blocks at the highest elevation and central point can be traced to the eastern shoulder of Mont Blanc. All those to the south-west come from the left-hand side of the lower Rhone valley, while those [[p. 623]] to the north-east are all from the left side of the upper Rhone valley and its tributaries. Other rocks coming from the right-hand side of the upper Rhone valley are found on the right-hand or Bernese side of the great valley between the Jura and the Bernese Alps.³

    Now, this peculiar and definite distribution, which has been worked out with the greatest care by numerous Swiss geologists, is a necessary consequence of well-known laws of glacier motion. The debris from the two sides of the main valley form lateral moraines which, however much the glacier may afterwards be contracted or spread out, keep their relative position unchanged. Each important tributary glacier brings in other lateral moraines, and thus when the combined glacier ultimately spreads out in a great lowland valley the several moraines will also spread out, while keeping their relative position, and never crossing over to mingle with each other. So soon as this definite position of the erratics was worked out it became evident that the first explanation--of a great submergence during which the lower Swiss valleys were arms of the sea and the Rhone glacier broke off in icebergs which carried the erratics across to the Jura--was altogether untenable, and that the original explanation of Venetz and Charpentier was the true one. Sir Charles Lyell, who had first adopted the iceberg theory, gave it up on examining the country in 1857 and ascertaining that the facts were correctly stated by the Swiss geologists; and there is at the present day no writer of the least importance who denies this. Sir Henry Howorth, who is one of the strongest opponents of what he considers the extreme views of modern glacialists, gives a full summary of the facts as to the old Rhone glacier from Charpentier. He states that between Martigny and St. Maurice the moraine debris on each side of the valley shows the glacier to have reached a height of 3,000 feet above the river; farther on, where the valley widens over the Lake of Geneva, it sank to 2,600 feet, while on the Jura itself it seems to have been again raised to 3,000 feet at its highest point;⁴ and he quotes Charpentier's general conclusion:--

    "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 debris 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."

And then, after quoting the observations of Agassiz on the same phenomena and of those of North America, he gives his own conclusions in the following words:--

    [[p. 624]] "It is plain to those who would look without prejudice that the rounded and mammillated surfaces, the scratched, polished, and grooved rocks, and a great number of the phenomena which accompanied the distribution of the boulders and the drift, are consistent only with the fact that in the last geological age there was an immense development of glaciers which occupied not only the high ranges of the Alps and the Dovrefelds, but the secondary ranges and lower heights of the continents of Europe and North America. This conclusion seems supported by every form of converging evidence, and is apparently beyond the reach of cavil. So far there is no question at issue."⁵

We may take it, therefore, that the views of Charpentier, Agassiz, and Sir Charles Lyell as to the extent and thickness of the great Rhone glacier are admitted to be correct, or, at least, not to be exaggerated, by the most strenuous opponents of the extreme glacialists. We may, therefore, use this as a fixed datum in our further investigations, and I think it will be found to lead us irresistibly to conclusions which in other cases these writers declare to be inadmissible.

    We must now consider briefly the distribution of erratics in North America, because they present some peculiar features and teach us much concerning the possibilities of glacier motion.

    An immense area of the North-eastern States, extending south to New York, and then westward in an irregular line to Cincinnati and St. Louis, is almost wholly covered with a deposit of drift material, in which rocks of various sizes are embedded, while other rocks, often of enormous size, lie upon the surface. These blocks have been carefully studied by the American geologists, and they present us with some very interesting facts. Not only are the distances from which they have been transported very great, but in very many cases they are found at a greater elevation than the place from which they must have come. Professor G. F. Wright found an enormous accumulation of boulders on a sandstone plateau in Monroe County, Pennsylvania. Many of these boulders were granite, and must have come either from the Adirondack Mountains 200 miles to the north, or from the Canadian Highlands still farther away. This accumulation of boulders was 70 or 80 feet high, and it extended many miles, descending into a deep valley 1,000 feet below the plateau in a nearly continuous line forming part of the southern moraine of the great American ice-sheet.

    On the Kentucky hills, about twelve miles south of Cincinnati, conglomerate boulders containing pebbles of red jasper can be traced to a limited outcrop of the same rock in Canada to the north of Lake Huron, more than 600 miles distant, and similar boulders have been found at intervals over the whole intervening country. In both these cases the blocks must have passed over intervening valleys and hills, the latter as high or nearly as high as the source [[p. 625]] from whence the rocks were derived. Even more remarkable are numerous boulders of Helderberg limestone on the summit of the Blue Ridge in Pennsylvania, which must have been brought from ledges at least 500 feet lower than the places upon which they now lie. The Blue Ridge itself shows remarkable signs of glacial abrasion, in a well-defined shoulder marking the southern limit of the ice (as indicated also by heaps of drift and erratics), so that Mr. Wright concludes that several hundred feet of the ridge have been worn away by the ice.

    The crowning example of boulder transportation is, however, afforded by the blocks of light grey gneiss discovered by Professor Hitchcock on the summit of Mount Washington, over 6,000 feet above sea-level, and identified with Bethlehem gneiss, whose nearest outcrop is in Jefferson, several miles to the north-west, and 3,000 or 4,000 feet lower than Mount Washington.

    These varied phenomena of erratic blocks and rock striations, together with the enormous quantity of boulder-clay and glacial drift spread over the whole of the Eastern States, terminating southward in a more or less abrupt line of mounds having all the characteristics of an enormous moraine, have led American geologists to certain definite conclusions in which they all practically agree. It may be well first to give a notion of the enormous amount of the glacial debris under which a large part of the Eastern States is buried. In New England these deposits are of less thickness than farther south, averaging from 10 to 20 feet over the whole area. In Pennsylvania and New York east of the Alleghanies, the deposits are very irregular, often 60 or 70 feet thick and sometimes more. West of the Alleghanies, in New York, Pennsylvania and Ohio the thickness is much greater, being often 150 or 200 feet in the wide valleys, and 40 or 50 feet on many of the uplands. Professor Newberry calculates that in Ohio it averages 60 feet deep over an area of 25,000 square miles.

    The direction of the striæ and of the travelled boulders together with the form of the great terminal moraines show that there must have been two main centres of outflow for the ice-sheet, one over Labrador, the other over the Laurentian Highlands north of Lake Superior. The southern margin of the drift may be roughly represented by portions of circles drawn from these two points as centres. The erratics on the summit of Mount Washington show that the ice-sheet must have been a mile thick in its neighbourhood, and much thicker at the centres of dispersion, while the masses of drift and erratics on plateaus 2,000 feet high near its southern boundary indicate a great thickness at the termination. The Laurentian plateau is now about 2,000 feet above the sea-level, but there are numerous indications from buried river channels, filled with drift and [[p. 626]] far below the sea, which lead to the conclusion that during the Ice Age the land was much higher. That snow can accumulate to an enormous extent over land of moderate height when the conditions are favourable for such an accumulation is shown by the case of Greenland, the greater part of whose surface is a vast plateau of ice flowing outward by numerous glaciers into the sea. The centre of this plateau where Dr. Nansen crossed it was over 9,000 feet above sea-level, and it may be very much higher farther north. It, therefore, seems probable that the great American ice-sheet was, at least, as high, and perhaps much higher, and this would give sufficient slope for the flow to the southern border. Of course, during the successive stages of the glaciation there may have been numerous local centres from which glaciers radiated, and during the passing away of the Ice Age these local glaciers would have left striæ and other indications of their presence. But so much of the area covered by the drift--all, in fact, south of the New England mountains and the Great Lakes--is undulating ground, hill, valley, and plateau of moderate height that here all the phenomena seem to be due to the great confluent ice-sheet during the various phases of its advance and its passing away.

    Sir Henry Howorth, in his very instructive work already quoted, denies the existence and even the possibility of such ice-sheets as those here indicated as having occurred in North America and Europe. He maintains that ice of the requisite thickness could not exist, as it would be crushed or liquefied by its own weight; and further, that if it existed it could not possibly move over hundreds of miles of generally level country, passing over hills and valleys and carrying with it, either on its surface or in its lower strata, the enormous quantity of boulders, gravel, and clay which we find everywhere overlying the present surface of the ground. No doubt the difficulty does seem an enormous one, but I think that it can be shown to be not so great as it seems; and it is certainly by no means so insuperable as that of the apocryphal floods, or "waves of translation" as they have been called, to which he imputes the phenomena. He asks us to believe in one or more gigantic waves sweeping over Eastern North America, carrying boulders to the summit of Mount Washington, nearly 6,000 feet high, scattering others over an area which is roughly 1,000 miles from east to west and 600 from north to south, and in its course producing those wonderful striæ, grooves, and furrows in the rocks photographed in the American reports, and the enormous extent of smoothed and rounded rock surfaces that is found over this wide area. Besides these there are two other phenomena absolutely inconsistent with a diluvial agency. One is the enormous deposits of fine compact clay bearing rounded and scratched stones thickly scattered through it, utterly unlike any [[p. 627]] deposit produced by water, which would necessarily leave the stones hundreds of miles behind the place to which the fine mud would be carried. The other is the existence of well-defined heaps, mounds, and ridges of gravel and boulders, forming the terminal moraine of the ice-sheet. This is exactly similar in general form and structure to the moraines left by the old Alpine or North British glaciers, and if the former could have been produced by a flood so could the latter. But the American terminal moraine runs across the country almost irrespective of its contour, and is often as well marked on plateaus as in valleys and on the intermediate slopes. Moreover, this moraine often lies on the southern slope of the hills, draining towards the Mississippi valley; and we are asked to believe that a flood vast enough to carry gravel and rocks for hundreds of miles to such a position, left them all stranded on a slope down which it must have been rushing with increased velocity and without hindrance towards the Gulf of Mexico! So far as I know, Sir Henry Howorth is absolutely alone among living writers in his diluvial theories, and I only give this brief statement of their overwhelming impossibilities because his book is so interesting, and his assertions that his theory explains all the facts are so confident and so often repeated, that they are likely to confuse the judgment of readers who have not paid special attention to the subject.

    Returning to the main question, of the possibility of glaciers or ice-sheets moving over long distances of generally level ground with intervening hills and valleys, there is an important piece of evidence, the bearing of which appears to have been overlooked by objectors. The former existence of the great Rhone glacier carrying erratics to the slopes of the Jura from beyond Geneva on the south-west to Soleure on the north-east, is universally admitted. This glacier passed out of the gorge between the Dent du Midi and the Dent de Morcles, and a little below St. Maurice enters on the alluvial plain which extends to the lake. From this point to Geneva, a distance of about 60 miles, may be considered a level plain, the descent into the lake being balanced by the ascent out of it. Yet it is admitted that the glacier did move over this distance, since erratics which can be traced to their source on the left of the valley below Martigny are found near that city. But the main part of the glacier curved round to the right across the Lake of Neufchatel, and extended at least as far as Soleure, a distance of about 90 miles. To do this it must have ascended 500 or 600 feet to the country around Fribourg, and before reaching Soleure must have passed over a hill 300 or 400 feet higher. Yet on the flanks of the Jura above Soleure there are erratics which have been carried on the surface of the glacier from the east side of the valley below Martigny, and close to Soleure itself there are [[p. 628]] remains of a terminal sub-glacial moraine of compact boulder-clay. Sir Charles Lyell describes this as--

"an unstratified mass of clay or mud, through which a variety of angular and rubbed stones were scattered, and a marked proportion of the whole were polished and scratched, and the clay rendered so compact, as if by the incumbent pressure of a great mass of ice, that it has been found necessary to blow it up with gunpowder in making railway cuttings through part of it. A marble rock, of the age of our Portland stone, on which this old moraine rests has its surface polished like a looking-glass, displaying beautiful sections of fossil shells, while occasionally, besides finer striæ, there are deep rectilinear grooves, agreeing in direction with the course in which the extinct glacier moved according to the theory of M. Guyot before explained."⁶

    It is evident that, to have produced such effects as are here described, the glacier must have extended much beyond Soleure, and have been very thick even there. It thus proves to demonstration that a glacier can travel for 100 miles over a generally level country, that it can pass over hills and valleys, and that, even near its termination, it can groove, and grind, and polish rocks, and deposit large masses of hard boulder-clay. And all this was done by a single glacier issuing from a comparatively narrow valley, and then spreading out over an area many times greater than that of its whole previous course. In this case it is clear that such a vast mass of ice, constituting a veritable ice-sheet on a small scale, could not have derived its motion solely from the push given to it by the parent glacier at St. Maurice. Neither could gravitation derived from the slope of the ground have affected it, for it passed mostly over level ground or up slopes, and its termination at Soleure is actually nearly 200 feet higher than its starting-point at the mouth of the valley below St. Moritz! There remains as a cause of motion only the slope of the upper surface of the glacier, the ice slowly flowing downward, and, by means of its tenacity and its viscosity on a large scale, dragging its lower portion still more slowly over the uneven or upward-sloping surface. This mode of motion will be discussed later when dealing with the origin of lake-basins.

    No doubt at this epoch of maximum glaciation the ice-sheet extended over the whole country between the Bernese Alps and the Jura, and the downward flow of the lateral glaciers along the valley of the Sarine, Aare, and other rivers flowing towards Soleure greatly assisted the general onward motion. But the fact remains, and it cannot be too strongly insisted on, that here we have a veritable ice-sheet moving over hill and valley, carrying on its surface quantities of erratic blocks, rounding, striating, and polishing the rocks over which it passed, and with the material thus crushed and ground away forming great deposits of boulder-clay, much of which still remains, although enormous quantities must have been carried away [[p. 629]] by the rivers to the lowlands of Europe and to the sea. The fact is therefore demonstrated, and is implicitly admitted by the most conservative of glacialists, that in this case an ice-sheet has moved onward over a hilly plateau for nearly 100 miles, even when its terminal moraine is at a higher level than its exit from the mountain valley where it had its origin.

    It will now be well briefly to sketch the distribution of erratic blocks in Great Britain, and the conclusions to be drawn from them as to the former existence of an ice-sheet under which the greater part of our islands was buried.

    Every mountain group north of the Bristol Channel was a centre from which, in the earlier and later phases of the Ice Age, glaciers radiated; but many facts prove that during its maximum development these separate glacier systems became confluent, and formed extensive ice-sheets which overflowed into the Atlantic Ocean on the west, and spread far over the English lowlands on the east and south. This is indicated partly by the great height at which glacial striæ are found, reaching to 2,500 feet in the Lake District and in Ireland, somewhat higher in North Wales, and in Scotland to nearly 3,500 feet; but also by the extraordinary distribution of erratic blocks, many of which can be traced to localities whence they could only have been brought across the sea. The direction of the glacial striæ and of the smoothed side of ice-worn rocks also indicate that the shallow seas were all filled up by ice. The Outer Hebrides, for example, are all ice-ground from the south-east and east, showing that the deep channel of the Minch was filled up, and that the Scotch ice-sheet flowed completely over the islands. On all sides of Ireland, except the southern coast, the ice flowed outward, but on the north-east the flow was diverted southward, and on the extreme north, westward, by the pressure of the overflowing ice-sheet of Scotland which here encountered it. In like manner, the ice-marks on the east coast of Ireland and the west coast of Wales are diverted southward by the mutual pressure of their ice-sheets, which, together with that of the west of Scotland, filled up St. George's Channel. That such was the case is further proved by the fact that the Isle of Man is ice-ground in a general direction from north to south, and to the summit of its loftiest mountains which rise to a height of over 2,000 feet. This could only have been done by an ice-sheet flowing over it, and this view is further supported by some most remarkable facts in the dispersal of local erratics. These are always found to the south of the places where they occur in situ, never to the north; and, what is still more noteworthy, they are often found far above the native rock. Thus, boulders of the peculiar Foxdale granite are found about 1,400 feet higher than the highest point where there is an out-crop of this rock.

    [[p. 630]] The Scotch ice-sheet flowed outwards on all sides, but on the east it was met by the southward extension of the great Scandinavian ice-sheet. On the extreme north the meeting of these two ice-sheets resulted in a flow to the north-west which glaciated the Orkney Islands, while the Shetlands, much farther north, received the full impact of the Scandinavian ice alone, and are therefore glaciated from the north-east. The dividing-line of the Scotch and Scandinavian ice-sheets was in the North Sea, not far from the east coast of Scotland; but farther south, at Flamborough Head and Holderness, the latter impinged on our coast, bringing with it enormous quantities of Scandinavian rocks. Many years ago Professor Sedgwick described the cliffs of boulder-clay at Holderness as containing "an incredible number of smooth round blocks of granite, gneiss, greenstone, mica slate, &c., &c., resembling none of the rocks of England, but resembling specimens derived from various parts of the great Scandinavian chain." These are mixed, however, with a number of British rocks from the north and west, indicating the meeting ground of the two conflicting ice-sheets. Similar blocks occur all along the coast as far as the cliffs of Cromer in Norfolk. Across the peninsula of Flamborough about two miles west of the lighthouse there is a moraine ridge containing a few Scandinavian boulders, but mainly composed of British rocks. These latter consist of numerous carboniferous rocks from the north and north-west, together with many of Shap granite--a peculiar rock found only on Shap Fell in the eastern side of the Lake District, together with a few of Galloway granite. These facts, it will be seen, add further confirmation to the theory of great confluent ice-sheets indicated by the ice-markings upon the various groups of mountains, while it is hopelessly impossible to explain them on any theory of local glaciers, even with the aid of submergence and of floating ice.

    The study of our British erratics has been assiduously pursued for many years past by a committee of the British Association; and by means of a map showing the chief facts collected up to this date, kindly furnished me by Mr. Percy F. Kendal, secretary of the committee, I am able to give a brief sketch of the more important of the phenomena, and their bearing on the extent and motion of the British ice-sheet. The general reader may be informed that great numbers of rocks are so local and so characteristic, often being confined to a very limited district or to a single mountain, that the origin of a considerable portion of the erratics can be ascertained with the greatest certainty.

    Taking first the Shap granite, which has already been mentioned as occurring at Flamborough Head, we find that it has been carried northwards as far as the Solway Frith, and eastward to the Eden valley in great quantity and over a wide area. Thence can be [[p. 631]] traced a line of boulders of this rock over the high plateau of Stainmoor into the valley of the Tees, and onward round the coast by Scarborough to Holderness, while a branch descends southward along the valley of the Ouse to York. Coming back to its source on Shap Fell, a train of boulders of the same rock has been traced south-ward in a curving line, passing the east side of Morecambe Bay near Lancaster, and thence sparingly south-eastward to near Whalley. Along the same line are found boulders of peculiar granites from Eskdale and Buttermere, marking the line of junction of the northern ice-sheet with that which filled up the Irish Sea and pressed inward between the glaciers of Cumberland and North Wales. This is indicated by the fact that south of this line are scattered immense quantities of erratics, both from the south-west of Scotland and the Lake District, spreading over the whole of the low country as far as Bridgnorth and Wolverhampton, and eastward to the Derbyshire highlands. These same erratics are found round the north coasts of Wales and part of Anglesea, showing how the iceflows divided on either side of the mountain mass of North Wales.

    The centre of the great glacier sheet of North Wales appears to have been over the Arenig Mountains, whence erratics of a peculiar volcanic rock have been traced to the north and east, mingling with the last-described group; while a distinct train of these Welsh erratics stretches south-eastward to the country west of Birmingham.

    In the Isle of Man are found many erratics from Galloway and a few from the Lake District. But the most remarkable are those of a very peculiar rock found only on Ailsa Craig, a small island in the Frith of Clyde, and a single boulder of a peculiar pitchstone found only in the Isle of Arran. The Ailsa Craig rock has also been found at Moel Tryfaen on the west side of Snowdon, and more recently at Killiney, co. Dublin, on the seashore.⁷

    The case of the boulders in the Isle of Man, which have been carried nearly 800 feet above their source, has already been mentioned, but there are many other examples of this phenomenon in our islands; and as they are of great importance in regard to the general theory of glacial motion a few of them may be noted here. So early as 1818 Mr. Weaver described a granite block on the top of Cronebane, a slate hill in Ireland, and several hundred feet higher than any place where similar granite was to be found in situ; and he also noticed several deposits of limestone gravel in places from 300 to 400 feet higher than the beds of limestone rock which are from two to ten miles off. Debris of red sandstone is also found much higher than the parent rock. Boulders of Shap granite, Mr. Kendal tells us, have passed over Stainmoor by tens of thousands, and in doing so have been carried about 200 feet above their source; and the curious Permian rock, "Brockram," has been carried in the [[p. 632]] same direction no less than 1,000 feet higher than its highest point of origin.⁸ In Scandinavia there are still more striking examples, erratic blocks having been found at an elevation of 4,500 feet which could not possibly have come from any place higher than 1,800 feet.⁹ We thus find clear and absolute demonstration of glacier ice moving up-hill and dragging with it rocks from lower levels to elevations varying from 200 to 2,700 feet above their origin. In Switzerland we have proof of the same general fact in the terminal moraine of the northern branch of the Rhone glacier being about 200 feet higher than the Lake of Geneva, with very much higher intervening ground. As it is universally admitted that the glacier of the Rhone did extend to beyond Soleure all the a priori objections to the various cases of rocks carried much higher than their origin, in America, the British Isles, and Scandinavia, fall to the ground. We must either deny the existence of the ice-sheet in the great Swiss valley, and find some other means of accounting for the travelled blocks on the Jura between Geneva and Soleure, or admit that the lower strata of a great glacier can travel up-hill and over hill and valley, and that the ice-sheets of the British Isles, of Scandinavia, and of North America merely exhibit the very same characteristics as those of Switzerland, but sometimes on a larger scale. We may not be yet able to explain fully how it thus moves, or what slope of the upper surface is required in order that the bottom of the ice may move up a given ascent, but the fact of such motion cannot any longer be denied.

    The facts thus established render it more easy for us to accept one of the latest conclusions of British glacialists. A great submergence of a large portion of the British Isles during the glacial period or in the interval between successive phases of the glacial period, has long been accepted by geologists, and maps have been often published showing the small group of islands to which our country was then reduced, the supposed subsidence being about 1,400 feet. The evidence for this is the occurrence, at a few spots, of glacial gravels containing marine shells in tolerable abundance, the most celebrated being at Moel Tryfaen, on the west side of Snowdon, at a height of more than 1,300 feet. Shell-bearing drifts have also been found near Macclesfield at a height of over 1,100 feet, and to the east of Manchester at between 500 and 600 feet elevation. Others have since been found on Gloppa, a hill near Oswestry. The fact that the shell-bearing gravels of Moel Tryfaen are nearly 40 feet thick shows that, if they are due to submergence, the land must have remained stationary at that level for a considerable period of time, and there would probably be other stationary periods at lower levels. Yet nowhere in the valleys or on the hill slopes of Wales, [[p. 633]] or the Lake District, or in the English lowlands are there any of the old beaches or sea cliffs, or marine deposits of any kind, that must have been formed during such a subsidence and which can hardly have been everywhere cleared away by subsequent glaciation. Another difficulty is that the shells of these drifts are such as could not have lived together on one spot, some being northern species others southern, some frequenting sandy others muddy bottoms, some which live only below tidal water while others are shore species. And, lastly, they are very fragmentary, only a small percentage of entire shells being found.

    In consequence of these various difficulties it was suggested by the late Mr. Belt that the great Irish Sea ice-sheet had carried up a portion of the sea-bottom embedded in its substance, perhaps containing deposits of shells of various periods and thus explaining the intermixture of species as well as their fragmentary condition. The fact that boulders and pebbles from Scotland, Ailsa Craig, and Cumberland have been found in the Moel Tryfaen beds almost amounts to a proof that they were so uplifted; and a recent search has shown that in the other localities where marine shells have been found in drift at great elevations similar foreign rocks occur, rendering it almost certain that the same ice-sheets which have distributed foreign erratics so widely over our country, and which in doing so must have passed over the sea-bottom, have in a few cases carried with them a portion of that sea-bottom, and deposited it with the erratics in the places where both are now found. A full discussion of this point, with replies to various objections, by Mr. P. F. Kendal, will be found in the volume already quoted; and he has recently adduced a fresh argument against "the great submergence" in the fact that, if it ever occurred, our lowlands must for a long time have formed the bottom of a sea 200 fathoms deep, yet not a single shell characteristic of that depth has yet been discovered in the drift.¹⁰ The cumulative evidence against the submergence is now almost, if not quite, conclusive.

    In the brief outline now given of the facts of glacial geology bearing upon the former existence, the thickness, extent, and motion of ice-sheets, it has only been possible to treat the subject very broadly, omitting all those details and minor difficulties which cannot be discussed within the limits of a popular article. My object has been to explain the nature and amount of the converging evidence demonstrating the existence of enormous ice-sheets in the northern hemisphere, to serve as a basis for the discussion of the glacial origin of lake-basins, which will form the subject of another article.

Notes Appearing in the Original Work

1. The works referred to are:--Do Glaciers Excavate? by Prof. T. G. Bonney, F.R.S. (The Geographical Journal, vol. i., No. 6); The Glacial Nightmare and the Flood, by Sir H. H. Howorth, M.P., F.R.S.; Fragments of Earth Lore, by Prof. James Geikie, F.R.S.; Man and the Glacial Period, by Prof. G. F. Wright, F.G.S.A.; La Période Glaciaire, by A. Falsan; and the Glacialists' Magazine, edited by Percy F. Kendall, F.G.S.; from which works, and from those of Lyell, Ramsay, Geikie, and the American geologists, most of the facts referred to in the present article are derived. [[on p. 616]]

2. Seventh Annual Report of the United States Geological Survey, p. 179. Arrangements have now been made for the preservation of these remarkable examples of ice-work. [[on p. 621]]

3. A map showing the lines of dispersal of these erratics is given in Lyell's Antiquity of Man, p. 344, and is reproduced in my Island Life, p. 111. [[on p. 623]]

4. These figures are almost certainly incorrect, as the upper surface of the glacier must have had a considerable downward slope to produce motion. The recent work of M. Falsan, La Période Glaciaire, gives the thickness as about 3,800 feet at the head of the lake and 3,250 feet at Geneva. [[on p. 623]]

5. The Glacial Nightmare and the Flood, p. 208. [[on p. 624]]

6. The Antiquity of Man, 4th edition, p. 349. [[on p. 628]]

7. Nature, vol. xlvii., p. 464. [[on p. 631]]

8. Wright's Man and the Glacial Period, p. 154. [[on p. 632]]

9. James Geikie's Great Ice Age, 2nd ed., p. 404. [[on p. 632]]

10. Wright's Man and the Glacial Period, pp. 167-175. Also Geological Magazine, November, 1892, pp. 491-500. [[on p. 633]]