Popular Science Monthly/Volume 49/July 1896/Causes, Stages, and the Time of the Ice Age
By WARREN UPHAM.
IF we could see the entire earth at once, by some grand extension of our range of vision, as we might walk around a geographic globe a hundred feet in diameter, and examine it fully, with comparison of all portions of its area, probably no other features of the great terrestrial panorama would be so impressive as the wonderful diversity of climatic conditions. At the same time with perpetual summer on the equator and throughout nearly all of the intertropical zone, a wintry covering of snow and ice would be seen on all lands in high latitudes about one or the other pole. While every bounty of luxuriant plant and animal life is present to attract the traveler and furnish him sustenance in the central zone, the rigorous climate which is gradually encountered in approaching the poles, and the general decrease and limitation of both flora and , have opposed insuperable obstacles to the most eager and courageous explorers. About four hundred and fifty miles at the north, and about eight hundred and fifty miles at the south, lie beyond the farthest limits of exploration; and more than double these distances must be crossed, respectively, if one would pass, according to Nansen's hope and plan, from one side to the other of the hitherto untraversed circumpolar areas.
During the Ice age, or Glacial period of geology, very extensive and thick sheets of land ice, like those now enveloping the Antarctic continent and the interior of Greenland, overspread the northern half of North America (excepting the greater part of Alaska) and northern Europe, with nearly the whole of the British Isles. The southern boundary of the North American ice sheet crossed Nantucket and Martha's Vineyard, Block Island, Long Island, and Staten Island. On the mainland it extended through northern New Jersey and northeastern and northwestern Pennsylvania, being indented by a great angle, whose apex was at Salamanca in southwestern New York. Thence it reached southwest and west across southern Ohio, Indiana, and Illinois, and through central Missouri, into northeastern Kansas; and beyond, it curved far northward, crossing eastern Nebraska and South and North Dakota. From near Bismarck it again trended westward through Montana, Idaho, and Washington, to the Pacific Ocean not far south of Puget Sound. North of this line an area of about four million square miles, stretching to the Arctic archipelago, was covered with ice hundreds and thousands of feet deep.
The comparatively small present ice sheet of Greenland covers five hundred and seventy-five thousand square miles, and rises with average slopes of one hundred feet or more per mile to a central height, along its axial portion, of eight to ten thousand feet, or almost two miles measured vertically, above the sea level. The ancient ice sheets had a similar altitude and thickness. From the directions of outflow of the North American ice fields as shown by the transportation of the glacial drift, and from the observed upper limits of glaciation on high mountains, Prof. James D. Dana estimated the thickness of the ice formerly accu- mulated above the Laurentide highlands, between the St. Law- rence River and Hudson Bay, to be fully two miles. It probably varied in thickness from one to two miles across Labrador, the Laurentide highlands, James Bay, Lake Winnipeg, Reindeer and Athabasca lakes, to the Rocky Mountains, in the region of the Peace River, where their summits, lower than southward, were probably buried beneath the ice expanse. In British Columbia, according to Dr. George M. Dawson's observations of glacial strise and drift on mountains, the ice sheet exceeded a mile in depth.
In all directions from its thick central areas the vast conti- nental glacier flowed outward, carrying its drift from Hudson Strait, Labrador, and Newfoundland easterly beyond the present coast line; from the provinces of Quebec and Ontario south- easterly across New England, and southerly and southwesterly across the basins of the Laurentian lakes; from Manitoba and the Saskatchewan region southerly into Minnesota, Iowa, the Dakotas, and Montana; from British Columbia into Idaho and Washington on the south, into the edge of the Pacific Ocean on the west, and down the Yukon Valley on the north; and from the great northern Barren Grounds northerly down the Macken- zie and across the islands of the Arctic Sea.
Northern Europe and the present basins of the Irish, North, Baltic, and White Seas were covered by an ice sheet which at- tained an extent of two million square miles, being half as large as that of North America; and its maximum depth above Sweden and the beds of the Baltic Sea and Gulf of Bothnia was one mile, or more probably two miles. The high, much eroded, and chan- neled Scandinavian plateau even now has numerous local ice fields, varying in size up to five hundred square miles, which are doubtless remnants of a continuous glaciation through all the centuries since the vast European ice field of the Glacial period
erick Wright and Warren Upham (D. Appleton & Co., 1896). From my chapters in this book some later paragraphs of the present paper are derived, with condensation and re- arrangement. flowed outward on all sides from this great plateau. Bowlders from its rock formations were then borne by the slow glacial cur- rents eastward to the head waters of the Volga, southward to the Dnieper and the Rhine, and southwestward to the northeastern shore of England, where the confluent current of the ice flowing away from the Scottish Highlands warded off the Scandinavian ice after its passage over the bed of the shallow North Sea. The European ice sheet extended south only to the latitude of 50", while that of our continent reached to 38° in southern Illinois; but the difference was similar to the present contrast of the mean annual temperature and isothermal lines of the two continents.
To-day the Greenland ice sheet, the Malaspina ice sheet be- tween Mount St. Elias and the ocean, many glaciers southward along the Cordilleran mountain belt, and the ice fields and glaciers of Norway and the Alps, may be regarded as lingering repre- sentatives of the conditions of the Glacial period, which not long ago, geologically speaking, spread a white pall of snow and unin- habitable desolation over large parts of the earth that are now temperate, fruitful, and populous. The returning mild and habit- able conditions, with luxuriant plant and animal life, are like the average of long geologic eras which preceded the Ice age, and were of far greater and indeed almost inconceivable duration. The severely cold and snowy Glacial climate of extensive land areas was wholly unlike their mild or even hot climates during the very long Tertiary and Mesozoic eras, of which we find testi- mony in their fossil floras and faunas. Palms allied to those of the tropics, and sequoias closely related to the big redwood trees of California, grew during Tertiary times in Greenland, Spitz- bergen, and the New Siberia Islands. Baron Nordenskjold, after examining thousands of miles of arctic shore lines, with frequent clearly exposed geologic sections belonging to the periods extend- ing back from the Ice age through the Tertiary or Cenozoic, the Mesozoic, and the Palaeozoic eras, affirmed that he nowhere dis- covered any evidence of glaciation previous to the Pleistocene period, which followed the Tertiary and introduced the Quater- nary era.
Latest in the great series of periods made known by the geo- logic record, the Pleistocene or Glacial period stands alone and unique, unless we must also recognize a general prevalence of gla- cial conditions at or near the end of the Palseozoic era. Bowlder- bearing deposits which can be explained only as glacial drift, and striation of the underlying rock which testifies unmistakably of the action of great glaciers or sheets of land ice, are found in the Carboniferous or the Permian series, closing the Palseozoic sys- tem, in Britain, France, and Germany, Natal, India, and south- eastern Australia. In Natal the striated glacier floor is in latitude 30° south, and in India only 20° north, of the equator. Such evi- dences of late Palaeozoic glaciation are also very clearly exhibited on the Varanger Fiord, in the extreme northern part of Norway, beyond the Arctic Circle. During all the earth's history before the Ice age of Pleistocene time, no other such distinct indications of general or interrupted and alternating glaciation have been found. Geologic exploration reveals only these two glacial peri- ods, and they are separated in time by the vast Mesozoic and Ter- tiary eras, together estimated by Dana and others to comprise some ten to fifteen or twenty million years.
It is especially suggestive, in our inquiry concerning the causes of the Ice age, that both the Palaeozoic and the Quaternary glacial periods were characterized by very unusual and exceptional oscil- lations in the height of continental areas and by the formation or renewed uplifting of great mountain ranges. Epochs in which certain mountain belts came into existence, or, after being partly or chiefly worn away, were restored by great uplifts, have alter- nated with far longer periods and eras of comparative repose. Between the epochs of mountain-building, the slow wearing and gnawing of rain, frost, and chemical decay have striven to carry away the mountains to the plains and the sea. At two times of the birth or rejuvenation of the grandest mountain chains of the world, with the most remarkable upward and downward move- ments of continents, the accumulation of glaciers and ice sheets has been closely associated.
Each of these periods of mountain formation, continental up- lifts, and widespread glaciation was geologically short; but they were separated by a lapse of time so long that it can be adequate- ly imagined only through the aid of a mathematical or geometric illustration on an almost infinitely reduced scale. Let the dura- tion of a lifetime of seventy years be represented by a span, or nine inches. A century on this scale will be denoted by a foot, a thousand years by ten feet, and a million years by about two miles. The whole duration of the earth's existence since the be- ginning of life upon its surface, if between fifty and a hundred million years, as estimated by Dana, Walcott, and others, would then be represented by a distance of about one hundred, one hun- dred and fifty, or two hundred miles. In accordance with the probable ratios of the several great eras of geology, which are de- termined through comparisons of their thicknesses of sedimentary rocks and their progress in evolutionary changes of faunas and floras, we may place the Palaeozoic glacial period at a distance of twenty to forty miles back from the present day, corresponding to some ten to twenty million years. That glacial period may have been no longer than the Ice age recently ended—that is, twenty-five thousand to fifty thousand years, more or less. The ice ages, therefore, would be marked by a length of some two hundred and fifty to five hundred feet for each on the assumed scale, and they would be separated by an interval approximately two hundred or four hundred times as long, according to the range in the estimates of the length of the intervening Mesozoic and Tertiary eras.
The chief astronomic theories of the causes of glaciation, pro- posed by Dr. James Croll and General A. W. Drayson, would re- quire the frequent recurrence of glacial epochs during all the vast interval dividing the two times of actual widely extended glacia- tion of which geology bears record. It seems quite certain, there- fore, that we must look rather to unusual conditions of the earth itself than to its astronomic relations as the causes of the Ice age.
Another theory, which supposes changes ih the earth's atti- tude toward the sun, is the suggestion, first made in 1866 by Sir John Evans, that, while the earth's axis probably remained un- changed in its direction, a comparatively thin ^crust of the earth may have gradually slipped as a whole upon the much larger nu- cleal mass, so that the locations of the poles upon the crust have been changed, and that the Glacial period may have been due to such a slipping or transfer by which the regions that became ice- covered were brought very near to the poles. The same or a very similar view has been recently advocated by Dr. Fridtjof Nansen, who writes:
The easiest method of explaining a Glacial epoch, as well as the occur- rence of warmer climates in one latitude or another, is to imagine a slight change in the geographical position of the earth's axis. If, for instance, we could move the north pole down to some point near the west coast of Greenland, between 60° and 65° north latitude, we could, no doubt, produce a Glacial period both in Europe and America.
Very small changes of latitude which had been detected at astronomical observatories in England, Germany, Russia, and the United States, seemed to give some foundation for this theory, which in 1891 was regarded by a few American glacialists as worthy of attention and of special investigation by astronomers, with temporary establishment of new observatories for this pur- pose on a longitude about 180° from Greenwich or from Wash- ington. During the year 1892, however, the brilliant discoveries by Dr. S. C. Chandler of the periods and amounts of the observed variations of latitude, showing them to be in two cycles respect- ively of twelve and fourteen months, with no appreciable secular change, forbade reliance on this condition as a cause or even as an element among the causes of the Ice age. This theory is now entirely out of the field. Sir Robert S. Ball, after reviewing Dr.
since the Glacial period, and from even earlier geologic times, has heen without greater changes of position than would lie inside the area of a block or square inclosed by the intersecting streets of a city.
We come now to the wholly terrestrial or geologic theory of the causes of the Ice age, which in terms varying with increasing knowledge has been successively advocated by Lyell, Dana, Le Conte, Wright, and the present writer. According to this ex- planation, the accumulation of the ice sheets was due to uplifts of the land as extensive high plateaus receiving snowfall through- out the year. Geology has received from Gilbert, in his mono- graph on Lake Bonneville for the United States Geological Sur- vey, the terms epeirogeny and epeirogenic (continent-producing), to designate the broad movements of uplift and subsidence which affect the whole or large portions of continental areas or of the oceanic basins. This view, accounting for glaciation by high altitude, may therefore be very properly named the epeirogenic theory. It is adversely criticised by Prof. James Geikie, who calls it "the earth-movement hypothesis."
So early as 1830 Lyell pointed out the intimate dependence of climate upon the distribution of areas of land and water and upon the altitude of the land. In 1855 Dana, reasoning from the prevalence of fiords in all glaciated regions, and showing that these are valleys eroded by streams during a formerly greater elevation of the land previous to glaciation, and from the marine beds of the St. Lawrence Valley and basin of Lake Champlain belonging to the time immediately following the glaciation, an- nounced that the formation of the drift in North America was attended by three great continental movements: the first upward, during which the ice sheet was accumulated on the land; the second downward, when the ice sheet was melted away; and the third, within recent time, a re-elevation, bringing the land to its present height. But with the moderate depth of the fiords and submarine valleys then known, the amount of preglacial eleva- tion which could be thus affirmed was evidently too little to be an adequate cause for the cold and snowy climate producing the ice sheet. The belief that this uplift was three thousand feet or more, giving sufficiently cool climate, as Prof. T. G. Bonney has shown, to cause the ice accumulation, has been only reached within the past ten years through the discovery, by soundings of the United States Coast Survey, that on both the Atlantic and Pacific coasts of the United States submarine valleys evidently eroded in late Tertiary and Quaternary time reached to profound depths, two thousand to three thousand feet below the present sea level. The continuation of the Hudson River Valley has been traced by detailed hydrographic surveys to the edge of the steep conti- nental slope at a distance of about one hundred and five miles from Sandy Hook. Its outermost twenty-five miles are a sub- marine fiord three miles wide and from 900 to 2,250 feet in verti- cal depth measured from the crest of its banks, which with the adjoining flat area decline from three hundred to six hundred feet below the present sea level. The deepest sounding in this fiord is 2,844 feet. An unfinished survey by soundings off the mouth of Delaware Bay finds a similar valley submerged nearly twelve hundred feet, but not yet traced to the margin of the continental plateau. Again, the United States Coast Survey and British Admiralty charts, as Spencer states, record submerged fiord out- lets from the Gulf of Maine, the Gulf of St. Lawrence, and Hud- son Bay, respectively 2,664 feet, 3,666 feet, and 2,040 feet below sea level. The bed of the old Laurentian River, as the preglacial St. Lawrence is named by Spencer, from the outer boundary of the Fishing Banks to the mouth of the Saguenay, a distance of more than eight hundred miles, is reached by soundings 1,878 to 1,104 feet in depth. Advancing inland, the sublime Saguenay fiord along an extent of about fifty miles ranges from three hundred to eight hundred and forty feet in depth below the sea level, while in some places its bordering cliffs, one to one and a half miles apart, rise abruptly fifteen hundred feet above the water.
On the Pacific coast of the United States Prof. Joseph Le Conte has shown that the islands south of Santa Barbara and Los Angeles, now separated from the mainland and from each other by channels twenty to thirty miles wide and six hundred to one thousand feet deep, were still a part of the mainland during the late Pliocene and early Quaternary periods. In northern Cali- fornia Prof. George Davidson, of the Coast Survey, reports three submarine valleys about twenty-five, twelve, and six miles south of Cape Mendocino, sinking respectively to 2,400, 3,120, and 2,700 feet below the sea level, where they cross the hundred-fathom line of the marginal plateau. If the land there were to rise one thousand feet, these valleys would be fiords, with sides towering high above the water, but still descending beneath it to profound depths. Le Conte has correlated the great epeirogenic uplifts of North America, known by these deeply submerged valleys on both the eastern and western coasts, with the latest time of oro- genic disturbance by faulting and upheaval of the Sierra Nevada and Coast Range in California during the closing stage of the Ter- tiary and the early part of the Quaternary era, culminating in the Glacial period. In the Mississippi basin, from the evidence of river currents much stronger than now, transporting Archsean pebbles from near the sources of the Mississippi to the shore of the Gulf of Mexico, Prof. E. W. Hilgard thinks that the pregla- dal uplift, inaugurating the Ice age, was four thousand or five thousand feet more in the central part of the continent than at this river's mouth.
Although the adequacy of the preglacial epeirogenic elevation of this continent to produce its Pleistocene ice sheet was tardily- recognized, it was distinctly claimed by Dana in 1870 that the Champlain subsidence of the land beneath its ice load, supposing it to have been previously at a high altitude, must have brought climatic conditions under which the ice would very rapidly dis- appear. The depression would be like coming from Greenland to southern Canada and New England. In Prof. Dana's words: "Such an extended change of climate over the glacier area was equivalent in effect to a transfer from a cold, icy region to that of a temperate climate and melting sun. The melting would there- fore have gone forward over vast surfaces at once, wide in lati- tude as well as longitude."
Such explanations as these, accounting for the gradual accu- mulation and comparatively rapid dissolution of the North American ice sheet, are also found to be applicable to the ice sheets of other regions. The fiords of the northern portions of the British Isles and of Scandinavia show that the drift-bearing northwestern part of Europe stood in preglacial time one thou- sand to four thousand feet higher than now; while, on the other hand, late glacial marine beds and strand lines of sea erosion testify that when the ice disappeared the land on which it had lain was depressed one hundred to six hundred feet below its present height, or nearly to the same amount as the Champlain depression in North America. Mr. T. F. Jamieson appears to have been the first in Great Britain or Europe to attribute the ice accumulation to altitude of the land, and to hold the view (which I receive from him) that the submergence of glaciated lands, when they were loaded with ice, was caused directly by this load pressing down the earth's crust upon its fused inte- rior, and that the subsequent re-elevation was a hydrostatic up- lifting of the crust by underflow of the inner mass when the ice was melted away. Just the same evidences of abundant and deep fiords and of marine beds overlying the glacial drift to heights of several hundred feet above the sea are found in Patagonia, as described by Darwin and Agassiz. On these three continental areas the widely separated chief drift-bearing regions of the earth are found to have experienced in connection with their glaciation in each case three great epeirogenic movements of similar char- acter and sequence—first, a comparatively long-continued uplift, which in its culmination appears to have given a high plateau climate with abundant snowfall, forming an ice sheet whose dura- tion extended until the land sank somewhat lower than now, lead- ing to amelioration of the climate and the departure of the ice, followed by re-elevation to the present level. The coincidence of these great earth movements with glaciation naturally leads to the conviction that they were the direct and sufficient cause of the ice sheets and of their disappearance; and this conclusion is confirmed by the insufficiency and failure of the other theories which have been advanced to account for the Ice age.
The end of the Tertiary era and the subsequent Glacial period were exceptionally characterized by many great oscillations of continental and insular land areas. Where movements of land elevation took place in high latitudes, either north or south, which received abundant precipitation of moisture, ice sheets were formed; and the weight of these ice sheets seems to have been a chief cause, and often probably the only cause, of the subsidence of these lands and the disappearance of their ice.
The general contemporaneousness of the Glacial period on the opposite sides of the North Atlantic Ocean had been long accepted as probable, but its demonstration and the identification of the corresponding parts of the Ice age, having the same se- quence on the two continents, were first made known less than two years ago by the studies of Geikie and Chamberlin in the new third edition of The Great Ice Age, and by their later papers in the Journal of Geology. According to the subdivision recog- nized by these authors, the time of principal accumulation of marginal moraines is regarded as an epoch distinct from the pre- vious portions of the Ice age; and Chamberlin has named the earlier divisions of this period, when the North American ice sheet reached its culmination, the Kansan and lowan stages, while the later moraine-forming time is called the Wisconsin stage, from the magnificent development of the moraines in east- ern Wisconsin. Between these glacial stages, which appear well recognizable and synchronous in North America and Europe, these authors suppose that there were prolonged interglacial epochs, when the ice sheets were in large part or wholly melted away. To the most important of the warm intervals, separating the Kansan and lowan stages of ice accumulation and advance, the name Af tonian is given by Chamberlin, from Afton in Iowa, where a thick bed of peat, formed during that time, lies between deposits of glacial drift.
Instead of this view of distinct epochs of glaciation, the Ice age seems to me, while accepting the successive stages here noted, to have been still essentially a single and continuous glacial pe- riod, with moderate fluctuations of the ice borders during boththe growth and wane of the ice sheet. The marginal moraines I
Stages of the Ice Age in the United States and Canada. Unglaciated land areas are dotted.
from its lowan stage, with no important general readvance di- viding the lowan from the Wisconsin or moraine-forming stage.
Not only are the Kansan and lowan stages of culmination of the ice sheets closely alike for North America and Europe, but also the land depression of the Champlain epoch in both these widely separated great areas brought marine submergence of coastal tracts, and caused rapid disappearance of the ice sheets, with the formation of their drumlins and marginal moraines. These two continents were included in the portion of the earth's crust which twice experienced far-extended epeirogenic move- ments, first of high uplift, bringing the cold climate and snow and ice accumulation of the Glacial period, and afterward of de- pression somewhat lower than now, whereby the vast ice fields were melted away.
The accompanying maps show the area of the North Amer- ican and European ice sheets in their maximum extension, and at definite times in their recession, as known by their areas of drift and belts of marginal moraines, and by the beaches of glacial lakes formed between the present watersheds and the northwardly retreating ice border. These maps give the boundaries of the Kansan, lowan, and Wisconsin formations, adopting these names, according to the law of priority, for both continents, and add for the northeastern United States and Canada the subsequent War- ren, Toronto, Iroquois, and St. Lawrence stages in the glacial re- treat.
The culmination of the great epeirogenic uplift, which had been in progress through the preceding Lafayette period, raised the glaciated areas, both in North America and Europe, to so high altitudes that they received snow throughout the year, and be- came deeply ice-enveloped. Submerged valleys and fiords show that this elevation was one thousand to four thousand feet above the present height. The accumulation of the ice sheets, due to snowfall upon their entire areas, was attended by fluctuations of their gradually extending boundaries, giving the Scanian and Norfolkian stages in Europe, and an early glacial recession and readvance in the region of the Moose and Albany Rivers, south- west of James Bay.
During the Kansan stage the ice sheet attained its farthest ex- tent in the Missouri and Mississippi River basins and in northern New Jersey, this being probably at the same time with the Saxo- nian stage, as later named by Geikie, of maximum glaciation in Europe.
In the Aftonian stage the ice sheet receded from its Kansan
Stages of the Ice Age in Europe. Unglaciated land areas and sea borders of the ice sheet are dotted.
Minn., in the Red River Valley, and two hundred and fifty miles or more in Illinois, according to Leverett; but probably little be- tween the Scioto River, in Ohio, and the Atlantic coast, the maxi- mum retreat of that portion being twenty-five miles or more in New Jersey. A cool temperate climate and coniferous forests ex- tended up to the retreating ice border in the upper Mississippi region. This great glacial recession was attended with much ero- sion of the early drift. A corresponding interruption of the severity of the Ice age in Europe is named by Geikie the Helve- tian stage or epoch. The greater part of the drift area in Russia was then permanently relinquished by the much diminished ice sheet, which also retreated considerably on all its sides. During this stage the two continents probably retained mainly a large part of thfeir preglacial altitude. The decrease of the ice sheet may have been caused by the astronomic cycle which brought our winters of the northern hemisphere in perihelion between twenty- five thousand and fifteen thousand years ago.
In the lowan stage renewed ice accumulation covered the Aftonian forest beds, so that the continental glacier extended again into Iowa, to a distance of three hundred and fifty miles or more from its most northern indentation by the Aftonian retreat, and in Illinois it readvanced about one hundred and fifty miles, while its boundary eastward from Ohio probably remained with little change. At the same time, apparently, was the Polandian stage of renewed growth of the European ice sheet, probably advancing its boundaries in some portions hundreds of miles from the Helvetian retreat.
These foregoing stages belong to the early and longer part of the Glacial period, which may be called pre-eminently the Glacial epoch, including the times of mainly very cold and snowy climate which tended to the formation and preservation of the ice sheet. The lowan stage was terminated by a depression of the ice-bur- dened area mostly somewhat below its present height, as shown by fossiliferous marine beds overlying the glacial drift up to three hundred feet above the sea in Maine, five hundred and sixty feet at Montreal, three hundred to four hundred feet from south to north in the basin of Lake Champlain, three hundred to five hundred feet southwest of Hudson and James Bays, and similar or less altitudes on the coasts of British Columbia, the British Isles, Germany, Scandinavia, and Spitzbergen. Glacial recession from the lowan boundaries was rapid under the tem- perate (and in summers warm or hot) climate belonging to the more southern parts of the drift-bearing areas when reduced from their great preglacial elevation to their present height or lower. The finer portion of the englacial drift, swept down from the ice fields by the abundant waters of their melting and of rains, was spread on the lower lands and along valleys in front of the departing ice as the loess of the Missouri, the Mississippi, and the Rhine. Marine beds reaching a maximum height of about three hundred and seventy-five feet at Neudeck, in western Prussia, give the name of this Neudeckian stage.
A moderate re- elevation of the land, to approximately its present height, advanced in the northern United States and Canada as a permanent wave from south to north and northeast, keeping nearly equal pace with the continued retreat of the ice along most of its extent. Throughout all the distance from the Atlantic to the Rocky Mountains the mainly retreating but often fluctuating ice margin formed many belts of knolly and hilly drift, called marginal moraines. It is also to be noted that the river basins which slope northward or northeastward were ob- structed by the waning ice sheet, so that they were temporarily filled by great glacial lakes, as Lake Agassiz, in the basin of the Red River of the North and of Lake Winnipeg, and a very remark- able series of lakes in the basin of the St. Lawrence, the glacial precursors of the present five great lakes from Superior to Ontario. The very grand development of the marginal moraines in Wisconsin (scarcely, however, surpassing Minnesota) led to the application of the name Wisconsin to this stage of the Ice age and to its drift. In Europe this is named by Geikie the Meck- lenburgian stage. Conspicuous moraine accumulations were formed in Sweden, Denmark, Germany, and Finland, on the southern and eastern margins of the great Baltic glacier.
During the maximum extent of the glacial Lake Warren, held on its northeast side by the retreating ice border, one expanse of water, as mapped by Spencer, Lawson, Taylor, Gilbert, and others, appears to have reached from Lake Superior over Lakes Michigan, Huron, and Erie, to the southwestern part of Lake Ontario. Its latest southern beach, traced east by Gilbert to Crittenden, New York, is correlated by Leverett with the Lockport moraine. This and later American stages, all of minor importance and duration in comparison with the preceding, can not probably be shown to be equivalent with Geikie's European divisions belong- ing to the same time.
In the next ensuing Toronto stage, slight glacial oscillations, with temperate climate nearly as now at Toronto and Scarborough, Ontario, are indicated by interbedded deposits of till and fossilif- erous stratified gravel, sand, and clay. Although the waning ice sheet still occupied a vast area on the northeast, and twice read- vanced with deposition of much till during the formation of the Scarborough f ossilif erous drift series, the climate then, determined by the Champlain low altitude of the land, by the proximity of the large glacial Lake Algonquin, succeeding the larger Lake Warren, and by the eastward and northeastward surface atmos- pheric currents and courses of all storms, was not less mild than now. The trees whose wood is found in the interglacial Toronto beds now have their most northern limits in the same region.
Somewhat later came the full expansion of the glacial Lake Iroquois, in the basin of the present Lake Ontario and northward, outflowing at Rome, N. Y., to the Mohawk and Hudson rivers. Gradual re-elevation of the Rome outlet from the Champlain subsidence had lifted the surface of Lake Iroquois in its western part from near the level of the present lake at Toronto to a height there of about two hundred feet, finally holding this height during many years, with the formation of the well-developed Iroquois beach.
The final stage in the departure of the ice sheet which we are able to determine from the history of the Laurentian lakes and St. Lawrence Valley, was when the glacial lake St. Lawrence, out- flowing through the Champlain basin to the Hudson, stretched from a strait originally one hundred and fifty feet deep over the Thousand Islands, at the mouth of Lake Ontario, and from the vicinity of Pembroke, on the Ottawa River, easterly to Quebec or beyond. As soon as the ice barrier was melted through, the sea entered these depressed St. Lawrence, Champlain, and Ottawa valleys; and subsequent epeirogenic uplifting has raised them to their present slight altitude above the sea level.
Further stages of the glacier recession are doubtless recogniz- able by moraines and other evidences, the North American ice sheet becoming at last, as it probably also had been in its begin- nings, divided into three parts—one upon Labrador, another northwest of Hudson Bay, as shown by Tyrrell's observations, and a third upon the northern part of British Columbia. From my studies of the glacial lake Agassiz, whose duration was prob- ably only about one thousand years, the whole Champlain epoch of land depression, the departure of the ice sheet because of the warm climate so -restored, and most of the re-elevation of the un- burdened lands, appear to have required only a few (perhaps four or five) thousand years, ending about five thousand years ago. These late divisions of the Glacial period were far shorter than its Kansan, Aftonian, and lowan stages; and the ratio of the Glacial and Champlain epochs may have been approximately as ten to one. The term Champlain conveniently designates the short, final part of the Ice age, when the land depression caused rapid though wavering retreat of the ice border, with more vigor- ous glacial currents on account of the marginal melting and increased steepness of the ice front, favoring the accumulation of many retreatal moraines of knolly and bowldery drift.