Fragments of Earth Lore: Sketches & Addresses Geological and Geographical

Of the extension of glacier-ice in the British Islands at the epoch in question I shall only say that the glaciation of Scotland was hardly, if at all, less extensive than during the climax of the Ice Age. Ireland, too, appears to have been almost as thickly mantled; but the ice-sheet that covered England and Wales did not extend so far south as that of the penultimate glacial epoch, a considerable area in East Anglia and the midland counties remaining apparently free from invasion. The Scandinavian and British ice-sheets, however, again coalesced upon the floor of the North Sea.

III.

The Results of Fluvio-glacial Action in Europe

The third question which I now proceed to consider is the result produced by the rivers and torrents of the Ice Age. This, I am aware, is a wide subject, and one upon which much has been written. But there are a few points which may be advantageously discussed for the purpose of bringing into prominent view the conditions which obtained in the river-valleys of Europe during the last great extension of glacier-ice.

A little consideration will serve to convince one that the intense glacial conditions that obtained in our Continent during the cold epochs of the glacial period were due to a low temperature, combined with excessive snow-fall. The winters, we can have no doubt, must have been prolonged and severe. But mere low temperature will not account for the enormous precipitation of snow. For this, great evaporation was required. And we are therefore forced to admit that the direct heat of the sun in summer must have been greater than it is in the same regions at the present day. Now, if this were really the case (and I do not see how otherwise the facts can be explained), then we ought to meet with evidence of swollen rivers, torrents, and widespread inundations everywhere outside of the glaciated areas. And this is precisely what we do find. Immense accumulations of coarse gravels are widely spread over all the valleys that head in regions which were formerly the sites of snow-fields and glaciers. These gravels are of such a character and are so distributed as to make it certain that they could not have been transported to and deposited in their present positions by rivers like those which now wind their way down the valleys of middle Europe. Still more remarkable are the enormous sheets of loam which are spread over much wider areas and reach to more considerable heights than the gravels. The origin of the gravels is sufficiently evident; they are simply the coarser detritus, swept along by the enormously flooded rivers of the glacial period, and meet with their analogues in the torrential gravels of modern glacier-valleys in the Alps and other elevated regions. The more widely-spread loams, according to the opinion of most glacialists, represent the finer mud and silt deposited from the muddy waters of the same period. But the height to which such gravels and loams ascend is so great that those who hold them to be of fluvio-glacial origin have found it difficult to maintain this view. Some writers, indeed, who have not sufficiently considered the weight of the evidence in its favour, have set it aside, and boldly suggested all kinds of wonderful hypotheses in its place. One imaginative author, for example, believes the wide-spread loams to be of volcanic origin, while another finds in the same deposits strong evidence of the Deluge. By a well-known and experienced observer, the famous löss of middle Europe is considered to be an Æolian accumulation – that is to say, a wind-blown deposit – the result of long-continued or frequently-repeated dust-storms. This is the opinion of Baron Richthofen, whose great work on China is so justly esteemed. He infers that at the time of the formation of our löss central Europe was a dry desiccated region, just as wide areas in central Asia are in our own day. He does not attempt to show us, however, how such climatic conditions could ever obtain in Europe. In point of fact, the geographical conditions of our Continent have not changed materially since Pleistocene times, and the presence of the wide Atlantic Ocean, that laves all our western shores, is of itself sufficient to preclude the possibility of such a climate having obtained in middle Europe. Richthofen’s theory likewise fails to account for the geographical distribution of the löss, and for many facts relating to its geology. Only one of these last shall I mention. The löss is intimately associated with accumulations, the glacial and fluvio-glacial origin of which cannot be doubted. It belongs, in fact, to the glacial series, and was laid down at a time when vast snow-fields and ice-sheets existed, and when it is quite impossible that a dry climate could have characterised any part of our Continent. In common with most geologists, I believe that the löss is simply an inundation-mud, deposited in temporary lakes and over flooded areas during the summer meltings of the snow- and ice-fields; and I shall now try to show how the occurrence at high levels of gravels and such loams as the löss may be accounted for without having recourse to volcanic action or to winds, or even to the Deluge. I shall invoke no agencies other than those which we are perfectly well assured were in full operation during the Ice Age.

Now, I ask you, in the first place, to bear in mind that while a glacial epoch continued, extreme conditions could not have been restricted to the areas undergoing glaciation. There is abundant evidence, indeed, to show that heavy, snows occasionally covered other regions, and that in such places severe frosts acted upon the rocks and soils even of the low-grounds. Need we wonder if at a time when the northern ice-sheet approached the fiftieth parallel of latitude in middle Europe, when almost every mountain-group of central and southern Europe had its snow-fields and glaciers – need we wonder if at such a time the climate of wide areas outside of the glaciated tracts was extremely ungenial? The more closely the superficial accumulations of such areas are studied, the more clearly do we perceive in them the evidence of cold and humid conditions. Try, then, to picture to yourselves the probable aspect of those regions during a glacial epoch. Immediately south of the northern ice-sheet deep snows must have buried large tracts of country, and such snows may have endured often for long years, notwithstanding the great melting that took place in summer. Even much further south, as in Spain and Italy, deep snows would cover the lesser hills and hill-ranges, while frost would act energetically in many a district where such action is now either inconsiderable or unknown. Such being the general conditions that must have obtained in the non-glaciated areas, let us very briefly consider what the results of such conditions must necessarily have been. Every one has noticed, during the more or less rapid melting of snow in winter and early spring, that our streams and rivers are then much muddier than when in summer and autumn they are swollen by heavy rains. This of course is due to the action of frost, by means of which rocks are disintegrated and soils are broken up and pulverised, so that when thaw supervenes, the superficial covering becomes soaked with moisture like a sponge. To such an extent does this take place, that one may often see the saturated soil creeping, slipping, and even flowing down the slopes. The effect of mere thaw is of course much intensified when the water derived from melting snows is present. Rills and tiny brooks then become converted into dark muddy torrents, and enormous quantities of fine-grained detritus are eventually swept into the rivers. The rivers rise in flood and inundate their plains, over the surface of which considerable deposits of loam and silt often accumulate. We cannot doubt that similar but much more intense action must have taken place over very wide regions in Europe during a glacial epoch. Such having necessarily been the case, we are not required to suppose that the löss and similar loams have been deposited entirely by rivers flowing from glaciers. It is doubtless true that most of the rivers headed in those days in glacier regions, and must in consequence have been highly discoloured with glacial mud, and probably a very large proportion of the loams in question consists of the fine flour of rocks – the result of glacial grinding. But the action of frost and thaw and melting snow upon the low-grounds, such as I have described, cannot be ignored, and seems to have played a more important rôle than has yet been recognised. I think it helps us better to explain the well-known fact that land-shells are more or less commonly distributed through the löss. One can readily understand, at all events, how snail-shells might be swept down the slopes of the land at the time of the spring thaws, and how large numbers might find their way eventually into the swollen glacial rivers. I have often observed, during the melting of snow and the thawing of soils, quantities of snail-shells in the very act of being swept into our brooks and rills. And we are all familiar with the fact that, after a spring-flood has subsided, snail-shells, along with vegetable débris, are often plentifully stranded upon the valley-slopes and flood-plains of our rivers.

Admitting, then, that the löss and similar accumulations are simply inundation-loams formed at a time when glaciers were discharging immense volumes of muddy water, and when the low-grounds were liable every summer to the denuding action of melting snows, and so forth, I have yet to account for the fact that these supposed inundation-loams sometimes occur at a height of 100 feet, or even of 300 feet, above the present levels of the rivers. Two theories have been advanced in explanation, each of which seems to me to contain an element of truth. It has, in the first place, been maintained, as by Prestwich, that the löss at the higher levels was probably deposited long before the rivers had excavated their channels to their present depths. Thus, during flood, they would be enabled to overflow tracts which they could not possibly have reached when they had deepened their valleys to a much greater degree. But while we must fully admit that the erosion effected by the rivers of the Pleistocene or Glacial period was excessive, yet we find it difficult or impossible to believe that great valleys, several miles in width, and two or three hundred feet in depth, were excavated in hard Devonian and other equally durable rocks by the swollen and active rivers of the Ice Age. And although it is extremely probable that the löss at the highest levels is older than the similar deposit at the lowest levels of such a valley as the Rhine, yet this does not get us out of our difficulty.

The other view to which I have alluded takes little or no account of river-erosion, but maintains that the floods of the Ice Age were sufficiently great to reach the highest levels at which river-gravels and loams occur. It is likely enough that, under present conditions, we can form but a very inadequate idea of the vast bodies of freshwater which formerly swept down our valleys, but we may be pardoned if we express our inability to conceive of our European rivers flowing with a breadth of many miles, and a depth of two or three hundred feet.

A few years before his death, Mr. Darwin made a suggestion to me, which I think gives us the true solution of the problem. He thought that during an Ice Age great beds of frozen snow might have accumulated over the low-grounds outside of the glaciated areas (in the manner I have already described), and that many valleys might have been filled to a considerable depth during a large part of the year with blown snow, afterwards congealed. In autumn, when the running water failed, the lines of drainage might in many cases be more or less choked, and it would be a mere chance whether the drainage, together with gravel, sand, and mud, would follow precisely the same lines during the next summer. Such action being repeated year after year, it might well happen that many river-valleys might become largely filled with rudely alternating layers of frozen snow and fluviatile detritus. And if this were so, the flooded rivers in summer would be enabled to overflow much wider and more elevated tracts than they could otherwise have reached. As the climate became less excessive, we can conceive of the frozen snows gradually melting, and of river-detritus being deposited at lower and lower levels in the valleys.

The probability of such frozen masses having choked up valleys and impeded the drainage during the Ice Age is not a mere plausible conjecture. In the far north of Alaska – in a region which was certainly not overflowed by the North American ice-cap – extensive sheets of ice occur, more or less deeply buried under thick soil. Nor can there be much doubt that these ice-masses date back to the Glacial period itself, seeing that in the soils which overlie them we meet with remains of the mammoth and other contemporaneous mammalian forms. Here, then, we have direct proof of the fact of frozen snow and ice having accumulated in the hollows of the land outside of the glaciated areas.[14 - I have given Mr. Darwin’s views, and discussed the origin of the Pleistocene fluvio-glacial deposits at some length in Prehistoric Europe, chaps, viii. and ix. To this work I refer for detailed geological evidence in support of the view advocated above.]

Now, if such conditions existed in the valleys of middle Europe, the widespread loss of those regions is readily accounted for. The occurrence of irregular sheets and shreds of gravel and loam at heights of more than a hundred feet above a valley-bottom offers no difficulty – it is in fact precisely the kind of phenomenon we might have expected. We are therefore not required to go out of our way to dream about impossible volcanic action, or to call upon the winds of heaven to help us, or upon the waters of the Deluge to float us out of our difficulties. But while I believe the views I have now advocated sufficiently account for the appearances presented by the ancient valley-gravels and loams of central Europe, there are two very considerable areas of löss which require some further explanation. The first of these is that broad belt of löss which extends from west to east across the plains of northern Germany, and the northern boundary of which coincides with the limits reached by the last great ice-sheet, from which it spreads south to the foot-hills of the Harz, and other mountains of middle Europe. Here we have a sheet of löss which bears no apparent relation to the valley-systems of the region in which it occurs. But the fact of its northern boundary being coincident with the terminal front of the last great northern ice-sheet at once suggests its origin. It is evident that this ice-sheet must have blocked the rivers flowing north, and dammed back their waters.[15 - The late Mr. Belt, as is well known, was of opinion that all the rivers flowing north in Europe and Asia were dammed back by a great Polar glacier, and that all the low-tracts in the northern portions of the two continents were thus covered by wide inland seas of freshwater. As I do not believe that such a Polar ice-cap existed during the Glacial period, I cannot agree with Mr. Belt that the alluvial plains of northern Siberia mark the sites of ice-dammed lakes.] A wide sheet of muddy water must therefore have extended east and west over the very area which is now covered by the belt of löss in question. This temporary lake would doubtless be subject to great alternations of level – a portion draining away perhaps under the ice-sheet – but the water would for the most part make its way westward, and eventually escape into the English Channel. From the waters of this great lake, fed by many large glacial rivers, abundant precipitation of loam and silt must have taken place.

The second and by far the most extensive sheet of löss in Europe is the so-called “black earth,” or “tchernozem,” with which such enormous tracts in southern Russia are covered. This widespread löss – for such it really is – I have elsewhere tried to show consists of the flood-loam and inundation-muds laid down by the water escaping along the margin of the northern ice-sheet, which discharged its drainage in the direction of the Black Sea, its black colour being due to the grinding down and pulverising of the black Jurassic shales which extend over such wide regions in middle Russia.

IV.

The Extent of Glaciation in North America

The various phenomena of glaciation which go to prove that a great ice-sheet formerly covered a wide region in northern Europe are developed on a still more extensive scale in North America. Smoothed and striated rock-surfaces, crushed and dislocated rock-masses, and enormous accumulations of morainic débris and fluvio-glacial detritus, all combine to tell the same tale. The morainic accumulations of North America have been distributed upon the same principles as the similar deposits of our own Continent. Boulder-clay of precisely the same character as that of Scotland and Scandinavia, of Switzerland and north Italy, covers vast tracts in the low-grounds of the British Possessions and the northern States of the Union, where it forms enormous sheets, varying in thickness from 30 or 50 up to 100 feet or more. In the rough Laurentian high-lands, however, it is more sparingly developed, and the same is the case in the hilly regions of New England. In short, it thickens out upon the low-grounds, and thins off upon the steeper slopes, while it attains its greatest thickness and forms the most continuous sheets in the country that lies south of the great lakes.

The southern limits of this deposit form a kind of rude semi-circle. From New York the boundary-line has been followed north-west through New Jersey and Pennsylvania to beyond the forty-second parallel, after which it turns to the south-west, passing down through Ohio to Cincinnati (39°); then, striking west and south-west through Indiana, it traverses the southern portion of Illinois. Its course after it reaches the valley of the Missouri has been only approximately determined, but it turns at last rather abruptly to the north-west, sweeping away in that direction through Kansas, Nebraska, Dakota, and Montana.

The general course followed by the ice-sheet underneath which this boulder-clay was formed has been well ascertained, partly by the evidence of the clay and its contents, and partly by that of roches moutonnées and striated rocks. The observations of geologists in Canada and the States leave it in no doubt that an enormous sheet of ice flowed south over all the tracts which are now covered with boulder-clay. During a recent visit to Canada and the States, I had opportunities of examining the glacial deposits at various points over a somewhat extensive area, and everywhere I found the exact counterparts of our own accumulations. In Minnesota, Wisconsin, Iowa, Illinois, Indiana, and Ohio, and again in New York, Connecticut, and Massachusetts, and the low-grounds of Canada, I recognised boulder-clay of precisely the same character as that with which we are familiar at home. The glacial phenomena of the Hudson valley and of the lower part of the Connecticut River were especially interesting. In those regions the evidence of a southward flow of the ice is most conspicuous, and the phenomena, down to the smallest details, exactly recalled those of many parts of Europe. Professor Dana, under whose guidance I visited the Connecticut valley, showed me, at a considerable height upon the valley-slope, an ancient water-course, charged with gravel and shingle, which could not possibly have been laid down under present conditions. It was, in fact, a sub-glacial water-course, and resembled the similar water-courses which are associated with boulder-clay in our own country.

If I met with only familiar glacial phenomena in the low-lying tracts traversed by me, I certainly saw nothing strange or abnormal in the hillier tracts. In passing over the dreary regions between the valley of the Red River and Lake Superior I was constantly reminded of the bleak tracts of Archæan gneiss in the north-west of Scotland, and of the similar rough broken uplands in many parts of Scandinavia and Finland. The whole of that wild land is moutonnée. Rough tors and crags are smoothed off, while boulder-clay nestles on the lee-side. In the hollows between the roches moutonnées are straggling lakes and pools and bogs innumerable. Frequently, too, one comes upon rounded cones and smooth banks of morainic gravel and sand, and heaps of coarse shingle and boulders, while erratics in thousands are scattered over the whole district. If you wish to have a fair notion of the geological aspect of the region I refer to, you will find samples of it in many parts of the Outer Hebrides and western Ross-shire and Sutherland. Cover those latter districts with scraggy pines, and their resemblance to the uplands of Canada will be complete.

From descriptions given by travellers it would appear that morainic detritus – mounds and sheets of stony clay, gravel and sand, shingle, boulders, and erratics – are more or less plentifully sprinkled over all the British Possessions and the islands of the Arctic Archipelago; so that we have every reason to believe that the ice-sheet which left its moraines at New York and Cincinnati extended northwards to the Arctic Ocean. Nor can there be much doubt that this same mer de glace became confluent in the west with the great glaciers that streamed outwards from the Rocky Mountains; while we know for a certainty that the southern portion of Alaska, together with British Columbia and Vancouver Island, were buried in ice that flowed outwards into the Pacific.

Along the eastern sea-board north of New York city there is no tract which has not been overflowed by ice. The islands in Boston Harbour are made up for the most part of tough boulder-clay; and boulder-clay and striated rocks occur also in Maine, New Brunswick, Nova Scotia, and Newfoundland.

Thus we may say that the ice-covered region of North America was bounded on the north by the Arctic, on the west by the Pacific, and on the east by the Atlantic Oceans. The Rocky Mountains, however, divided the great mer de glace that overflowed Canada and the States from the ice that streamed outwards to the Pacific. Measured from the base of the Rockies to the Atlantic, the mer de glace of Canada and the States must have exceeded 2500 miles in width, and it stretched from north to south over 40 degrees of latitude.

Outside of this vast region and the great mountain-ranges of the far west, there are few hilly areas in the States which reach any considerable elevation. South of the mers de glace of the north and west, no such mountain-groups as those of middle and southern Europe occur, and consequently we do not expect to meet with many traces of local glaciation. Nevertheless, these have been recognised in the Alleghany Mountains, West Virginia, and in the Unaka Mountains, between Tennessee and North Carolina. But the glaciers of those minor hill-ranges were of course mere pigmies in comparison with the enormous ice-streams that flowed down the valleys of the Rocky Mountains and the Sierra Nevada. Even as far south as the Sierra Madre of Mexico glaciers seem formerly to have existed; and Mr. Belt has described the occurrence of what he considered to be boulder-clays at a height of 2000 to 3000 feet in Nicaragua.

I have mentioned the fact that in Europe we have, outside of the glaciated areas, certain accumulations (such as the Gibraltar breccias) which could only have been formed under the influence of extreme cold. Similar accumulations occur in North Carolina, where they have been carefully studied by Mr. W. C. Kerr. According to Mr. Kerr, these deposits have crept down the declivities of the ground under the influence of successive freezings and thawings; and now that attention has been called to such phenomena, our American friends will doubtless detect similar appearances in many other places.

The facts which I have now briefly indicated suffice to show that during the climax of glaciation North America must have presented very much the same appearance as Europe. Each continent had its great northern ice-sheet, south of which local glaciers existed in hilly districts, many of which are now far below the limits of perennial snow. We may note, also, that in each continent the mers de glace attained their greatest development over those regions which at the present day have the largest rainfall. Following the southern limits of glaciation in Europe, we are led at first directly east, until we reach central Russia, when the line we follow trends rapidly away to the north-east. The like is the case with North America. Trace the southern boundary of the ice-sheet west of New York, and you find, when you reach the valley of the Missouri, that it bends away to the north-west. Now we can hardly doubt that one principal reason for the non-appearance of the mer de glace in the far east of Europe and the far west of America was simply a diminishing snow-fall. Those non-glaciated regions which lay north of the latitudes reached by the ice-sheets were dry regions in glacial times for the same reasons that they are dry still. The only differences between glacial Europe and America were differences due to geographical position and physical features. The glaciation of the Urals was comparatively unimportant, because those mountains, being flanked on either side by vast land-areas, could have had only a limited snow-fall; while the mountain-ranges of western North America, on the other hand, being situated near the Pacific, could not fail to be copiously supplied. For obvious reasons, also, the North American ice-sheet greatly exceeded that of Europe. In all other respects the conditions were similar in both continents.

V.

Changes of Climate in North America during the Ice Age

American geologists are now pretty well agreed that their “interglacial deposits” – the existence of which is not disputed – have precisely the same meaning as the similar deposits which occur in Europe. They tell of great climatic changes. At present, however, there is no certain evidence in the American deposits of more than one interglacial epoch; but the proofs of such an epoch having obtained are overwhelming. The occurrence again and again of fossiliferous beds intercalated between two separate and distinct sheets of boulder-clay and morainic accumulations, leaves us in no doubt that we are dealing with precisely the same phenomena which confront us in Europe. No mere partial recession and re-advance of the mer de glace will account for the facts. We have seen that during the culmination of the Glacial period the American ice-sheet overflowed Ohio, Indiana, and Illinois. Now interglacial deposits occur as far north as the Canadian shores of Lakes Ontario and Superior, so that all the country to the south must have been uncovered by ice before those interglacial deposits were laid down. But the evidence entitles us to say much more than this. The interglacial beds of Ohio, Indiana, Illinois, and other States, afford abundant evidence of a great forest-growth having covered the regions vacated by the ice of the penultimate glacial epoch. The trees of this forest-land included sycamore, beech, hickory, red-cedar, and others; and amongst the plants were grape vines of enormous growth, which, according to Professor Cox, “indicate perhaps the luxuriance of a warmer climate.” At all events, the climate that nourished such a forest-growth could not have been less genial than the present. And such being the case, we may reasonably infer that the vast regions to the north of the lakes were no more inhospitable then than they are now.

To this genial interglacial epoch succeeded the last glacial epoch, when a great ice-sheet once more enveloped a wide area. In the extreme east this latest mer de glace appears to have reached as far south as that of the earlier epoch; but as we follow its terminal moraines westward they lead us further and further away from the southern limits attained by the preceding ice-sheet. These great terminal moraines form an interesting study, and the general results obtained by American observers have been very carefully put together by Professor Chamberlin. I traversed wide regions of those moraines in Indiana, Illinois, Wisconsin, and Minnesota, and, so far as my observations went, I could only confirm the conclusions arrived at by Professor Chamberlin and others. The mounds, banks, cones, and ridges are unquestionably moraines – of enormous dimensions, no doubt, but in all their phenomena strictly analogous to similar gravelly moraines in our own country and the Continent. Many of the American moraines consist almost entirely of water-worn material – sand, gravel, shingle, and boulders, together with large angular and sub-angular erratics. These deposits are generally stratified, and frequently show diagonal or false-bedding. In this and other respects they exactly reproduce – but of course on a much larger scale – our Scottish kames, and the similar accumulations of north Germany and Finland, and the low-grounds of Italy opposite the mouths of the great Alpine lakes. The kames of Wisconsin again and again reminded me of the gravelly moraines that cover the ground for many miles round the lower end of Lake Garda. It is this gravelly and sandy aspect of the American moraines that is most conspicuous, water-assorted materials seeming everywhere to form their upper and outer portions. Now and again, however, a deep cutting discloses underneath and behind such water-worn detritus a mass of confused materials, consisting of clay, sand, gravel, shingle, and boulders, which are angular and sub-angular, often smoothed and striated, and of all shapes and sizes. According to Mr. Chamberlin, this unstratified material “is indistinguishable from true till, and is doubtless to be regarded as till pushed up into corrugations by the mechanical action of the ice.”

This grand series of moraines stretches from the peninsula of Cape Cod across the northern States, and passes in a north-westerly direction into the British Possessions, over which it has been followed for some 400 miles. The disposition of the moraines, forming as they do a series of great loops, shows that the ice-sheet terminated in a number of lobes or gigantic tongue-like processes. Nothing seen by me suggested any marine action; on the contrary, every appearance, as I have said, betokened the morainic origin of the mounds; and Mr. Chamberlin assured me that their peculiar distribution was everywhere suggestive of this origin. No one who has traversed the regions I refer to is at all likely to agree with Sir W. Dawson’s view, that the American mounds, etc., are the shore-accumulations of an ice-laden sea.

The morainic origin of these accumulations having been demonstrated by American geologists, we are now able to draw another parallel between the European and American glacial deposits. We have seen that in Europe the ice-sheet of the latest glacial epoch was by no means so extensive as that of the preceding glacial epoch. The same was the case in North America. Moreover, in America, just as in Europe, the latest occupant of the land was not the sea, but glacier-ice. In Scotland and Scandinavia the gradual disappearance of the latest ice-sheets was marked by a partial submergence, which in the former country did not greatly exceed 100 feet, and in the latter 700 feet. In America, in like manner, we find traces of a similar partial submergence. In Connecticut this did not exceed 40 or 50 feet, but increased to some 500 feet in the St. Lawrence, and to over 1000 feet in the Arctic regions. If there ever was during the Glacial period a greater submergence than this in North America it must have taken place in earlier glacial or interglacial times, but of such a submergence no trace has yet been recognised. In this respect the American record differs somewhat from our own, for in Britain we have evidence of a submergence of over 1000 feet, which supervened in times immediately preceding the latest great extension of continental ice.[16 - See footnote, p. 173.] But nowhere in middle Europe, and nowhere in North America, in the region south and west of the great lakes, is there any trace of a general marine submergence. The “Palæocrystic Sea” is as idle a dream for the northern States of America as it is for any part of Europe.

VI.

The Results of Fluvio-glacial Action in North America

The close analogies which obtain between the glacial and interglacial deposits of Europe and North America are equally characteristic of the fluvio-glacial accumulations of the two continents. As in Europe, so in America we meet with considerable sheets of gravel and shingle, sand, fine clay, and loam, which are evidently of freshwater origin. In the gently-undulating tracts of the northern States those deposits often spread continuously over wide regions; in the hillier districts, however, they are most characteristic of the valleys. They are very well represented, for example, in the Connecticut valley, where they have been carefully studied by Professor Dana. Like the similar deposits of our own Continent, they have been laid down by the torrents and swollen rivers of the Glacial period. The great range of moraines which marks the extreme limits reached by the latest ice-sheet is generally associated with sheets of gravel and sand, which one can see at a glance are of contemporaneous origin, having been spread out by the water escaping from the melting ice. Nor can one doubt that the vast sheets of löss in the Missouri and Mississippi valleys are strictly analogous in origin, as they are in structure and disposition, to the löss of Europe. I have spoken of the probable existence of a glacial lake formed by the damming back of the Rhine and other rivers by the European ice-sheet. Now, in North America we meet with evidence of the same phenomenon. When the last ice-sheet of that continent attained its maximum development, all the water escaping from its margin in the north States necessarily flowed south into the Mississippi and Missouri rivers. But in course of time the ice melted away beyond the drainage-area of those rivers, and disappeared from the valley of the Red River of the north, which, it will be remembered, empties itself northward into Lake Winnipeg. When the ice-front had retired so far it naturally impeded the drainage of the Red River basin, and thus formed a vast glacial lake, the limits of which have been approximately mapped out by Mr. Upham, by whom the ancient lake has been designated Lake Agassiz. The deposits laid down in this lake consist of finely laminated clays, etc., which resemble in every particular the similar unfossiliferous clays so frequently found associated with glacial accumulations in Europe. Had the drainage of the Red River valley been south instead of north, the clays and loams of the far north-west would not have been arrested and spread out where they now are, and Manitoba would have been covered for the most part with loose shingle, gravel, and sand.

Thus the final disappearance of the American ice-sheet was marked by the formation not only of moraines, but of flood-gravels and torrential- and inundation-deposits of the same character as those with which we are familiar at home. Wherever similar geographical conditions prevailed, there similar geological results followed.

VII.

Conclusion

There are many other points of resemblance between the glacial and fluvio-glacial accumulations of the two continents, but to these time forbids any reference. Indeed, I cannot recall any signal difference. Such differences as do occur are due simply to the varying conditions of the two continental areas. The glacial phenomena of North America are a repetition of those of Europe, but upon a much grander scale. The boulder-clays of the former continent, in their composition, structure, and distribution, exactly recall our own. Interglacial beds occur under similar circumstances in both continents; and the same is the case with the gravelly moraines and fluvio-glacial accumulations. We are driven, then, to the conclusion that the physical conditions of the Glacial period were practically the same in Europe and North America. What those conditions were I have already indicated, and have shown that the results arrived at by geologists are not vague dreams and speculations, but a logical induction from well-ascertained facts. Before we can believe that volcanic eruptions, a general deluge, or a Palæocrystic Sea have produced the many varied phenomena of our glacial formations, either in whole or in part, we must first shut our eyes and then erase from our minds all knowledge of the facts which have been so laboriously gathered by a long succession of competent observers.

VII.

The Intercrossing of Erratics in Glacial Deposits.[17 - The Scottish Naturalist, 1881.]

Among the many phenomena connected with the glacial deposits of this country which have puzzled geologists there is none more remarkable than the “intercrossing of erratics.” The fact that such wandered blocks have apparently crossed each other’s tracks in their journeys appears at first sight inexplicable on the assumption that their transport has been effected by land-ice. The phenomena in question, therefore, have always been appealed to by those who uphold the iceberg origin of our boulder-clays, etc., as evidence decisively in favour of their views. No one can deny that any degree and amount of intercrossing might take place in the case of icebergs. We can readily conceive how floating ice, detached from a long line of coast, might be compelled by shifting winds and changing currents to tack about again and again, so as to pursue the most devious course, and scatter their stony burdens in the most erratic manner over the sea-bottom; while, on the other hand, it is quite impossible to understand how a similar irregular distribution of erratics could take place under one and the same glacier flowing in a determinate direction. It is little wonder, then, that the curious phenomena of the intercrossing of erratics should have had much importance attached to it by the upholders of the iceberg theory, seeing that all the other proofs which have been adduced in favour of this theory have only served to demonstrate its insufficiency. Upon the facts connected with the intercrossing of erratics, the supporters of this time-honoured theory are now making what I must believe is their last stand. I purpose therefore, in this paper, to give a short outline of those facts, with the view of showing that so far from being antagonistic to the land-ice theory, they are in complete harmony with it; and indeed must be considered as affording an additional demonstration of its truth.

Some years ago I called attention to the fact that in the middle districts of Scotland the boulder-clay not infrequently contains a curious commingling of northern and southern erratics.[18 - Great Ice Age, 2nd edit., p. 609.] I showed that this was the case throughout a belt of country extending from the sea-coast near Ayr, north-east to the valley of the Irvine, and thence across the watershed into the Avon, and east to Lesmahagow, then down the valley of the Clyde to Carluke, stretching away to the east by Wilsontown, and thereafter continuing along the crest of the Pentlands and the northern slopes of the Lammermuir Hills, by Reston and Ayton, to the sea. “All along this line,” I remarked, “we have a ‘debatable ground’ of variable breadth, throughout which we find a commingling in the till of stones which have come from north and from south. South of it, characteristic Highland stones do not occur, and north of it stones derived from the south are similarly absent.” The explanation of these facts is obvious. The belt of ground referred to was evidently the meeting-place of the Highland and southern mers de glace. Here the two opposing ice-flows coalesced and became deflected by their mutual pressure to right and left – one great current going east and another west. It is evident that the line of junction between the two mers de glace could not be rigorously maintained in one and the same position during a period of glaciation, but would tend to oscillate backwards and forwards, according as one or the other ice-sheet prevailed. Sometimes the southern ice-sheet would be enabled to push back the northern mer de glace, while at other times the converse would take place. Nor is it necessary to suppose that the advance of one ice-sheet was general along the whole line. On the contrary, it is most likely that the movement was quite irregular – an ice-sheet advancing in some places, while at other points its line of junction with the opposing ice-sheet remained stationary, or even retrograded. Such movements would obviously give rise to oscillations in the sub-glacial débris of clay and stones; and thus we have a simple and natural explanation of those intercrossings of erratics which are so characteristic of that region which I have termed the “debatable ground.” And this conclusion is borne out by the fact that the glacial striæ of the same “debatable ground” afford like evidence of oscillation in the trend of the ice-flow.

Along the base of the Highland mountains in Forfarshire, etc., we meet with similar intercrossings of erratics. Thus we occasionally encounter in the boulder-clays overlying the Silurian regions erratics of Old Red Sandstone rocks which have come from the east or south-east; while the abundant presence of erratics of Silurian origin, on the other hand, bespeak an ice-flow from the west towards the low-grounds. In some places within the Silurian area we encounter a greyish-blue boulder-clay containing Silurian fragments only, while in other places within the same area the boulder-clay becomes reddish, and is charged with many boulders of Old Red Sandstone rocks. Now the greyish-blue till could only have been laid down by glacier-ice descending from the Silurian high-grounds to Strathmore, while the red boulder-clay points to a partial invasion of the Silurian regions by land-ice, which had previously traversed the lower-lying Old Red Sandstone areas. These apparently contradictory movements are readily accounted for by the former presence in the area of the North Sea of the great Scandinavian mer de glace. Dr. James Croll was the first to point out that the glacial phenomena of Caithness and the Shetlands could only be accounted for by the advance of the Scandinavian ice-sheet towards our coasts, where it encountered and deflected the Scottish ice-sheet out of its normal course – a sagacious induction, which the admirable and exhaustive researches of my colleagues, Messrs. B. N. Peach and J. Horne, have now firmly established. The lower blue boulder-clay was evidently accumulated at a time when the Scottish ice was able to flow more or less directly east or south-east towards what is now the coast-line; while the overlying red boulder-clay points to a subsequent period when the presence of the Scandinavian mer de glace was sufficiently great to compel the Scottish ice out of its normal course, and cause it to flow in a north-easterly direction. In doing so it now and again passed from tracts of Old Red Sandstone to invade the Silurian area, and thus an overlying red boulder-clay was here and there accumulated upon the surface of a greyish-blue till in which not a single fragment of any Old Red Sandstone rock occurs.

Recently Messrs. B. N. Peach and J. Horne, in a most instructive paper on the “Glaciation of Caithness,”[19 - Proceedings Royal Physical Society, Edinburgh, 1881.] have described some remarkable comminglings of material which occur in a region where the glacial striæ afford equally striking evidence of conflicting ice-movements. These phenomena are developed here and there along a line which indicates the meeting-place of two rival ice-streams, on each side of which the boulder-clay presents different characteristics – the one boulder-clay being the moraine profonde of the ice that flowed ENE. and NNE. towards the Caithness plain, while the other is an accumulation formed underneath the ice that streamed across that plain from SE. to NW. These phenomena are thus, as my colleagues remark, quite analogous to those met with in the middle districts of Scotland, as described by me, and referred to in a preceding paragraph. Now it is obvious that while these examples of “intercrossings” of erratics and “cross-hatching” of striæ all go strongly to support the land-ice theory of the glacial phenomena, they at the same time negative the notion of floating-ice having had anything to do with the production of the phenomena under review.

Before considering the evidence adduced by Mr. Mackintosh and others as to the intercrossings of erratics in the drift-deposits of England, I shall mention some of the more remarkable examples of the same phenomena which have been noticed by continental geologists. The first cases I shall cite are those which have been observed in the glacial accumulations of the Rhone valley in eastern France. The land-ice origin of these accumulations has never been called in question, and as the intercrossings of erratics in that region are not only more common, but much more striking and apparently inexplicable than any which have been noticed elsewhere, it will be admitted that they of themselves afford a strong presumption that the conflicting courses followed by the erratics in certain regions of our own country are the result rather of oscillations in the flow of land-ice than of the random and eccentric action of icebergs. The researches of Swiss and French glacialists have proved that during the climax of the Glacial period an enormous area in the low-grounds of eastern France was covered with a huge mer de glace, formed by the union of the great Rhone glacier with the glaciers descending from the mountains of Savoy and Dauphiny. A line drawn from Bourg by way of Chatillon, Villeneuve, Trévoux, and Lyons to Vienne, and thence south-east by Beaurepaire to the valley of the Isère, a few miles above St. Marcellin, indicates roughly the furthest limits reached by the mer de glace. Over all the low-grounds between that terminal line and the mountains are found widespread sheets of boulder-clay and sand and gravel, together with loose erratics. Now and again, too, well-marked terminal moraines make their appearance, while the rock-surfaces, when these are visible and capable of bearing and retaining glacial markings, present the usual aspect of roches moutonnées. The same kinds of morainic materials and ice-markings may of course be followed up into the valleys not only of the Alps properly so-called, but also into those of the hills of Bugey and the secondary mountain-chain of Savoy and Dauphiny. It has indeed long been known that local glaciers formerly occupied the mountain-valleys of Bugey. For example, a number of small glaciers have descended from the slopes of the mountains west of Belley (such as Bois de la Morgue, Bois de Lind, etc.) to the Rhone, and again from Mont du Chat to the north-west. These glaciers were quite independent of the greater ice-streams of the neighbouring Alps of Savoy, and the same was the case with the glaciers of that mountainous tract which extends from Nantua south to Culoz, between the valleys of the Ain and the Rhone. From this elevated region many local glaciers descended, such as that of the Valromey, which flowed for a distance of some twenty miles from north to south. Again, similar local glaciers have left abundant traces of their former presence throughout the mountainous belt of land that stretches between Chambery and Grenoble to the west of the valley of the Isère. The moraines of all those local glaciers, charged as they are with the débris of the neighbouring heights, clearly indicate that the local glaciers flowed each down its own particular valley. There are certain other appearances, however, which seem at first sight to contradict this view. Sometimes, for example, we encounter in the same valleys erratics which do not belong to the drainage-system within which they occur, but have without doubt been derived from the higher Alps of Switzerland and Savoy. And the course followed by these foreign erratics has crossed at all angles that which the local glaciers have certainly pursued – occasionally, indeed, the one set of erratics has travelled in a direction exactly opposed to the trend taken by the others. As examples, I may cite the case of the erratics which occur in Petit Bugey. In this district we encounter many locally-derived erratics which have come from Mont du Chat to the west of the Lac du Bourget – that is to say, they have travelled in a north-westerly direction. But in the same neighbourhood are found many erratics of Alpine origin which have been carried from north-east to south-west, or at right angles to the course followed by the local erratics. Again, in the valley of the Seran we have evidence in erratics and terminal moraines of a local glacier which flowed south as far as the Lyons and Geneva Railway, in the neighbourhood of which, a few miles to the west of Culoz, its terminal moraines may be observed. This is the extinct Glacier du Valromey of MM. Falsan and Chantre. Now it is especially worthy of note that in the same valley we have distinct evidence of an ice-flow from south to north —i. e., up the valley. Erratics and morainic materials which are unquestionably of Alpine origin have been followed a long way up the Seran valley – for two-thirds of its length at least. Before they could have entered that valley and approached the slopes of Romey, they must have travelled down the valley of the Rhone from the higher Alps of Savoy in a south-west and south direction until they rounded the Montagne du Grand Colombier. It was only after they had rounded this massive mountain-ridge that they could pursue their course up the valley of the Seran, in a direction precisely opposite to that which they had previously followed. These and many similar and even more remarkable examples of the “intercrossings” of streams of erratics are described by MM. Falsan and Chantre, and graphically portrayed in their beautiful and instructive work on the “Ancient Glaciers and Erratic Deposits of the Basin of the Rhone”; and the explanation of the phenomena given by them is extremely simple and convincing. The local erratics and moraines pertain partly to the commencement and partly to the closing stage of the Glacial period. Long before the south branch of the great glacier of the Rhone had united with the glacier of the Arve, and this last with the glaciers of Annecy and Beaufurt, and before these had become confluent with the glacier of the Isère, etc., the secondary mountain-ranges of Savoy and Dauphiny and the hills of Bugey were covered with very considerable snow-fields, from which local glaciers descended all the valleys to the low-ground. But when the vast ice-flows of Switzerland, Upper Savoy, etc., at last became confluent, they completely overflowed many of the hilly districts which had formerly supported independent snow-fields and glaciers, and deposited their bottom-moraines over the morainic débris of the local glaciers. In other cases, where the secondary hill-ranges were too lofty to be completely drowned in the great mer de glace, long tongues of ice dilated into the valleys, and compelled the local ice out of its course; sometimes, as in the case of the Valromey, forcing it backward up the valleys down which it formerly flowed. But when once more the mighty mer de glace was on the wane, then the local glaciers came again into existence, and reoccupied their old courses. And thus it is that in the hilly regions at the base of the higher Alps, and even out upon the low-grounds and plains, we encounter that remarkable commingling of erratics which has been described above. Not infrequently, indeed, we find one set of moraines superposed upon another, just as in the low-grounds of northern Germany, etc., we may observe one boulder-clay overlying another, the erratics in which give evidence of transport in different directions. The observations recorded by MM. Falsan and Chantre, and their colleagues, thus demonstrate that “intercrossings” of erratics of the most pronounced character have been brought about solely by the action of glaciers. In the case of the erratics and morainic accumulations of the basin of the Rhone, the action of icebergs is entirely precluded.

I may now mention some of the more remarkable examples of intercrossings of erratics which have been recorded from the glacial accumulations of north Germany, etc. An examination of the glacial striæ, roches moutonnées, and boulder-clays of Saxony leads to the conviction, according to Credner, Penck, Torell, Helland, and others, that the whole of that region has been invaded by the great Scandinavian mer de glace which flowed into Saxony from NNE. to SSW. Erratics from southern Sweden and Gothland occur in the boulder-clay, and the presence of these, taken in connection with the direction of the glaciation, leaves us no alternative but to agree with the conclusions arrived at by the Saxon geologists. But, apparently in direct contradiction of this conclusion, we have evidence to show that boulders of the same kinds of rock occur in Denmark and Holland, pointing to a former ice-flow from north-east to south-west and west. Thus boulders derived from Gothland occur at Gröningen in Holland, while fragments from the island of Öland are met with in Faxö; and erratics from the borders of the Gulf of Finland are encountered at Hamburg. Indeed, when geologists come to examine the erratics in north Germany and Poland generally, they find evidence of apparently two ice-flows – one of which went south-south-west, south, and south-east – spreading out, as it were, in a fan-shape towards the southern limits reached by the great “Northern Drift,” – while the other seems to have followed the course of the Baltic depression, overflowing the low-grounds of northern Prussia, Holland, etc., in a south-west and west direction. Now, it is quite evident that no one mer de glace could have followed these various directions at one and the same time. The explanation of the apparent anomaly, however, is not far to seek. It is reasonable to infer that long before the mer de glace had attained its maximum dimensions, when as yet it was confined to the basin of the Baltic and was only able to overflow the northern regions of Prussia, etc., its course would be determined by the contour of the pavement upon which it advanced. It would, therefore, be compelled to follow the Baltic depression, and for a long time it would carry erratics from Finland, the Baltic islands, and eastern Sweden in a south-west and west-south-west direction. And this would continue to be the direction even after a considerable portion of the low-grounds of Prussia, etc., had been overflowed. But when the ice-sheet was enabled to advance south into Saxony, Poland, and Lithuania, erratics from Finland, the Baltic islands, etc., would necessarily cease to travel towards the west, and hold on a south-south-east, south, and south-south-west course. Again, when the mer de glace was on the decline, a time would return when the ice, as before, would be controlled in its flow by the Baltic depression, and this would give rise to a further distribution of erratics in a prevalent west-by-south direction.[20 - For a fuller discussion of the distribution of erratics on the Continent, I may refer to Appendix, Note B, in Prehistoric Europe, where the reader will find references to the literature of this interesting subject. [Continental geologists now recognise a distinct stage of the Ice Age, during which their “Upper Diluvium” was deposited by a great glacier that occupied the basin of the Baltic. This “Great Baltic Glacier” appears to have been contemporaneous with the local ice-sheets and valley-glaciers of the Highlands and other mountain-tracts of our island. See Article X. 1892.]]

No one of late years has been more assiduous in the collection of facts relating to the intercrossing of erratics in the drift-deposits of England than Mr. D. Mackintosh.[21 - This enthusiastic geologist died in 1891.] He has written many instructive and interesting descriptions of the phenomena in question, which he justly thinks are of prime importance from a theoretical point of view. In a recent paper[22 - Quart. Journ. Geol. Soc., vol. xxxv. p. 425] he presents us with the results of a systematic survey of the direction and limits of dispersion of the erratics of the west of England and east of Wales, which he evidently is of opinion afford strong support to the iceberg theory, while at the same time they are directly opposed to the theory of transport by land-ice. I have attentively considered all the arguments advanced by Mr. Mackintosh in favour of his views – the one upon which he apparently lays most stress being that of the intercrossings of erratics observed by him – and I shall now proceed to point out how the phenomena described by him are most satisfactorily explained by the land-ice theory. They seem to me, indeed, to lend additional support to that theory, in the same manner as the intercrossings of boulders observed in Scotland, northern Germany, etc., and sub-alpine regions of France. Mr. Mackintosh calls attention to the fact that erratics of the well-known Criffel granite are found scattered over a large part of the plain of Cumberland, from which they extend south along the coast to near the mouth of the estuary of the Duddon. They reappear on the coast in the neighbourhood of Blackpool and Liverpool, and again at intervals on the coasts of north Wales from Flint to Colwyn Bay, and thence to Penmaenmawr and the neighbourhood of Beaumaris. They are dispersed over the peninsula of Wirral and the Cheshire plain, etc., and they have been followed south-east as far as the neighbourhood of Cardington, near Church Stretton, Burton, Wolverhampton, Stafford, Hare Castle, Macclesfield, and Manchester. This great stream of boulders, therefore, spreads out to south-east, south, and south-west: the erratics, to quote Mr. Mackintosh, “have radiated from an area much smaller than their terminal breadth.” The same is the case, I may remark in passing, with erratics in the boulder-clays of Scotland, Scandinavia, north Germany, etc., as also with those in the drift-deposits of the great Rhone glacier and other ancient glaciers both on the north and south side of the Alps. Now, the course followed by the Criffel erratics is crossed at an acute angle by the path pursued by many boulders of Eskdale granite, and various felspathic rocks derived from the Cumberland mountains. For example, Cumberland erratics of the kinds mentioned occur near St. Asaph and Moel-y-Tryfane and in Anglesey, and they have been followed over a wide district in Cheshire, etc., extending as far south as Church Stretton and Wolverhampton, and as far east as Rochdale. More than this, we find that numerous erratics of felstone, derived from the mountain of Great Arenig, in north Wales, have gone to north-east as far as Halkin Mountain, in Flintshire, Eryrys, near Llanarmon, and Chirk, from which last-named place they have been traced in a south-easterly direction to Birmingham, Bromsgrove, etc. A glance at the map of England will show that this south-easterly drift of erratics crosses at an acute angle the paths followed by the Criffel granite boulders and the erratics derived from Cumberland, so that we have now several intercrossings to account for. How can this be done by the land-ice theory?

The explanation seems to me obvious, for the phenomena are, after all, less striking than similar appearances which have been observed in Scotland, especially by my colleagues, Messrs. Peach and Horne, in Caithness and the Orkney and Shetland Islands; and they are certainly less intricate than the facts recorded by MM. Falsan and Chantre concerning the intercrossing, interosculation, and direct opposition of erratic paths in Savoy and Dauphiny. We have only to reflect that the great mer de glace– to which, as I believe, all the English phenomena are due – did not come into existence and attain its maximum dimensions in the twinkling of an eye, nor could it afterwards have disappeared in the same sudden manner. On the contrary, a period of local glaciation must have preceded the appearance of the great ice-sheet. At first, and for a long time, permanent snow would be confined to the higher elevations of the land, and glaciers would be limited to mountain-valleys; but as the temperature fell the snow-line would gradually descend, until at last, probably after a prolonged period, it reached what is now the sea-level. Thus the formation of névé and glacier-ice would eventually take place over what are now our low-grounds, and other tracts also, which are now submerged. It is quite impossible that the vast sheets of ice which can be demonstrated to have covered Scotland, a large part of England, Ireland, Scandinavia, and north Germany, and even the limited area of the Faröe Islands, could possibly have been fed by the snow-fields of mountain-heights only. The precipitation and accumulation of snow, and the formation of névé and glacier-ice, must have taken place over enormous regions in what are now the temperate latitudes of Europe.

It is obvious that the direction of ice-flow in the basin of the Irish Sea opposite the south of Scotland and the west of England, while preserving a general southerly trend, would vary at different periods. Before the mer de glace in that basin had attained its climax there must have been a time when the ice, streaming outwards from the high-grounds of Cumberland, was enabled to push its way far westward out into the basin of the Irish Sea. At that time it was still able to hold its own against the pressure exerted by the Scottish ice. But as the general mer de glace increased in thickness, the course of the Cumberland ice would be diverted ever further and further to the south-east, until, eventually, the Scottish ice came to hug the coast of Cumberland, and to overflow Lancashire in its progress towards the south-east. So gorged with ice did the basin of the Irish Sea become, that a portion of the Scottish ice was forced over the plain of Cumberland and up the valley of the Eden, where it coalesced with the ice coming north from the Shap district, and thereafter flowed in an easterly direction to join the great mer de glace of the North Sea basin.

Thus the intercrossings of the Criffel and Cumberland erratics described by Mr. Mackintosh receive a ready explanation by the land-ice theory. Nor do the intercrossings of the Welsh erratics with those derived from Scotland and Cumberland offer any difficulty. The ice coming from the Welsh mountains would naturally be deflected towards south-east by the mer de glace that streamed in that direction, and might quite well have carried its characteristic boulders as far as Birmingham before the general mer de glace had attained its greatest dimensions. But when that period of maximum glaciation arrived, the Welsh boulders would be unable to travel so far towards the east, and the Scottish and Cumberland boulders would then cross the path formerly followed by the felstone erratics from Great Arenig.

Again, it is evident that when the mer de glace was gradually decreasing similar oscillations of the ice-flow would take place, but in reverse order, and thus would give rise to a second series of intercrossings. Moreover, we must remember that the Glacial period was characterised by several great changes of climate. It was not one continuous and prolonged period of cold conditions, but consisted rather of a succession of arctic and genial climates; so that the same countries were overrun at different epochs by successive mers de glace, each of which would rework, denude, and redistribute to a large extent the morainic materials of its predecessor, and thus might well cause even greater complexity in the dispersion of erratics than has yet been recognised anywhere in these islands.