GLACIAL LAKES WHICH MARKED THE DECLINE OF THE LAST ICE AGE

Interference of glaciers with drainage.—Every advance and every retreat of a continental glacier has been marked by a complex series of episodes in the history of every river whose territory it has invaded. Whenever the valley was entered from the direction of its divide, the effect of the advancing ice front has generally been to swell the waters of the river into floods to which the present streams bear little resemblance (Fig. 346). Because of the excessive melting, this has been even more true of the ice retreat, but here when the ice front retired up the valley toward the divide. A sufficiently striking example is furnished by the Wabash, Kaskaskia, Illinois, and other streams to the southward of the divide which surrounds the basin of the Great Lakes (Fig. 347).

Wherever the relief was small there occurred in the immediate vicinity of the ice front a temporary diversion of the streams by the parallel moraines, so that the currents tended to parallel the ice front. This temporary diversion known as “border drainage” was brought to a close when the partially impounded waters had, by cutting their way through the moraines, established more permanent valleys (Fig. 348).

Temporary lakes due to ice blocking.—Whenever, on the contrary, the advancing ice front entered a valley from the direction of its mouth, or a retreating ice front retired down the valley, quite different results followed, since the waters were now impounded by the ice front serving as a dam. Though the histories of such blocking of rivers are often quite complex, the principles which underlie them are in reality simple enough. Of the lakes formed during advancing hemicycles of glaciation, and of all save the latest receding hemicycle, no satisfactory records are preserved, for the reason that the lake beaches and the lake deposits were later disturbed and buried by the overriding ice sheets. We have, however, every reason to suppose that the histories of each of these hemicycles were in every way as complex and interesting as that of the one which we are permitted to study.

As an introduction to the study of the ice-blocked lakes of North America, and to set forth as clearly as may be the fundamental principles upon which such lakes are dependent, we shall consider in some detail the late glacial history of certain of the Scottish glens, since their area is so small and the relief so strong that relationships are more easily seen; it is, so to speak, a pocket edition of the history of the more extended glacial lakes.

The “parallel roads” of the Scottish glens.—In a number of neighboring glens within the southern highlands of Scotland there are found faint terraces upon the glen walls which under the name of the “parallel roads” (Fig. 349) have offered a vexed problem to scientists. Of the many scientists who long attempted to explain them, though in vain, was Charles Darwin, the father of modern evolution. He offered it as his view that the “roads” were beaches formed at a time when the sea entered the glens and stood at these levels. When, however, Jamieson’s studies had discovered their true history, Darwin, with a frankness characteristic of some of the greatest scientists, admitted how far astray he had been in his reasoning. Let us, then, first examine the facts, and later their interpretation. The map of Fig. 350 will suffice to set forth with sufficient clearness the course of the several “roads.” These “roads” are found in a number of glens tributary to Loch Lochy, and of the three neighboring valleys, Glen Roy has three, Glen Glaster two, and Glen Spean one “road.” The facts of greatest significance in arriving at their interpretation relate to their elevations with reference to the passes at the valley heads, their abrupt terminations down-valleyward, and the morainic accumulations which are found where they terminate. The single “road” of Glen Spean is found at an elevation of 898 feet, a height which corresponds to that of the pass or col at the head of its valley and to the lowest of the “roads” in both Glens Glaster and Roy. Similarly the upper of the two “roads” in Glen Glaster is at the height of the pass at its head (1075 feet) and corresponds in elevation to the middle one of the three “roads” in Glen Roy. Lastly, the highest of the “roads” in Glen Roy is found at an elevation of 1151 feet, the height of the col at the head of the Glen. In the neighboring Glen Gloy is a still higher “road” corresponding likewise in elevation to that of the pass through which it connects with Glen Roy.

To come now to the explanation of the “roads”, it may be said at the outset that they are, as Darwin supposed, beach terraces cut by waves, not as he believed of the ocean, but of lakes which once filled portions of the glens when glaciers proceeding from Ben Nevis to the southwestward were blocking their lower portions. The several episodes of this lake history will be clear from a study of the three successive idealistic diagrams in Fig. 351.

To derive the principles underlying this history, it is at once seen that all changes are initiated by the retirement of the ice front to such a point that it unblocks for the waters of a lake an outlet that is lower than the one in service at the time. This is the principle which explains nearly all episodes of glacial lake history. Thus, when the ice front had retired so as to open direct connections between Glen Roy and Glen Glaster, the col at the head of Glen Roy was abandoned as an outlet, and the waters fell to the level fixed for Glen Glaster. A still further retirement at last opened direct connection between Glen Glaster and Glen Spean, so that the lake common to Glens Glaster and Roy fell to the level of the col which was the outlet of the Spean valley at the time. This stage continued until the ice front had retired so far that the waters drained naturally down the river Spean to Loch Lochy and thence to the ocean.

Only in their far grander scale and in the lesser relief of the land over which they formed, do the complex histories of the great ice-blocked lakes of North America differ from these little valley lakes whose beaches may be visited and the relationships worked out, thanks to Jamieson, in a single day’s strolling.

The glacial Lake Agassiz.—The grandest of the temporary lakes referable to blocking by the continental glaciers of the ice age must be looked for in the largest valleys that lay within the territory invaded and which normally drain toward the retiring ice front. In North America these rivers are the Red River of the North in Minnesota, the Dakotas, and Manitoba; and the St. Lawrence River system. To the ice dam which lay across the Red River valley we owe the fertility of that vast plain of lake deposits where is to-day the most intensive wheat farming of the northwest (Fig. 352). Lakes Winnipeg, Winnipegoosis, and Manitoba, and the Lake of the Woods, are all that now remain of this greatest of the glacial lakes, which in honor of the distinguished founder of the glacial theory has been called Lake Agassiz (Fig. 353). With their natural outlet blocked by the ice in northern Manitoba and Keewatin, the waters of the Red were swollen by melting from the retiring glacier and spread over a vast area before finding a southern outlet along the course of the present Lake Traverse and the valley of the Minnesota River. Along this route there flowed a mighty flood which carved out a broad valley many times too large for the Minnesota, its present occupant, and this giant prehistoric river has been called the Warren River (Fig. 354).

It is interesting to follow this ancient waterway and to discover that, like our normal, present-day streams, it was held up in narrows wherever outcroppings of harder rock had constricted its channel (Fig. 355). The upper end of the Warren River valley is now occupied by the long and relatively narrow Lakes Traverse and Big Stone, each the result of blocking by delta deposits where a tributary stream has emerged into the valley, but this gigantic channel continues down to and beyond Minneapolis, occupied as far as Fort Snelling by the Minnesota River—a mere pygmy compared to its predecessor. To the earnest student of glacial geology there can be few sights more impressive than are obtained by standing at Fort Snelling, just above the confluence of the Minnesota and the Mississippi rivers, and surveying first the steep and narrow valley of the Mississippi above the junction,—a stream fitted to its valley for the simple reason that it has carved it,—and then gazing up and down that broad valley in which the great Warren River once flowed majestically to the sea, now the bed of the Minnesota above the Fort and of the Mississippi below it (Fig. 356).

Just as the “parallel roads” of Glen Roy, roads in name only, are the beaches of earlier glacial lake stages, so in Lake Agassiz we have parallel beaches of the barrier type which are often roads in fact as well as in name, and which mark the stages of successive lakes within this vast basin. The Herman beach, corresponding to the highest level of the lake, is thus a sharp topographic boundary between lake deposits and morainal accumulations, and is further itself a well-marked topographic feature composed of wave-washed and hence well-drained materials (Fig. 357). Farmers of the district have been quick to realize that these level and slightly elevated ridges lack the clay which would render them muddy in the wet seasons, and are thus ideally adapted for roads. They have in many sections been thus used over long stretches and are known as the “ridge roads.”

Episodes of the glacial lake history within the St. Lawrence valley.—Within this great drainage basin it has apparently been possible to read the records of each stage in the latest lake history—complex as this has been. We have only to recall the lake stages cited from the Scottish glens and remember that each new stage was begun in a retirement of the glacier front which unblocked an outlet of lower level than the last. This sequence might, however, have been varied by a temporary readvance of the ice, as indeed once occurred in the Huron-Erie lobe of the great North American glacier.

The crescentic lakes of the earlier stages.—So long as the glacier covered the entire drainage basin of the St. Lawrence River system, all water was freely drained away by streams which flowed away from the ice front (Fig. 358). So soon, however, as at any point the front had retired behind the divide, impounding of the waters must locally have occurred. Lakes of this type are to-day to be seen in Greenland and in the southern Andes; and though upon a diminutive scale, some idea of their aspect may be obtained from the appearance of the MÄrjelen Lake of Switzerland, here blocked by a mountain glacier (Fig. 446, p. 411). Within all areas of small relief, such as the prairie country surrounding the present Laurentian lakes, the earlier and smaller stages of such ice-blocked lakes are generally crescentic in outline. This is because a moraine in most cases forms the land margin of the lake, and because the ice cliff upon the opposite border, although somewhat straightened, as a consequence of wave-cutting and iceberg formation, still retains the convex outlines characteristic of ice lobes (Fig. 359).

Within each of the Great Lake basins a crescentic lake early appeared at that end of the depression which was first uncovered by the glacier: Lake Duluth in the Superior basin, Lake Chicago in the Michigan basin, and Lake Maumee in the Huron-Erie basin (Fig. 360).

We may now, with profit, trace the successive episodes of the glacial lake history, considering for the earlier stages those changes which occurred within the Huron-Erie basin, since, these are in essential respects like those of the Michigan and Superior basins, although worked out in greater detail. Lake Chicago must, however, be brought into consideration, since in all save the earliest and the later stages, the waters from the Huron-Erie depression were discharged through the Grand River into this lake and thence by the so-called “Chicago outlet” into the Mississippi (plate 20 A).

The early Lake Maumee.—The area, outline, and outlet of this lake are indicated upon Fig. 360. Its ancient beaches have been traced, as well as the water-laid moraine beneath its former ice cliff; and no observant traveler who should take his way down the ancient outlet from Fort Wayne, Indiana, past the town of Huntington, could fail to be impressed by its size, suggesting as it does the great volume of water which must once have flowed along it. Now a channel a mile or more in width, its bed for the twenty-five miles between Fort Wayne and Huntington may be seen from the tracks of the Wabash Railway as a series of swamps merely, while at Huntington the Wabash river enters by a young V-shaped valley at the side, much as the Mississippi emerges into the old channel of the Warren River at Fort Snelling, Minnesota (see p. 327).

The Huron River of southern Michigan, which now discharges into Lake Erie, then found its lower course blocked by the glacier and was thus compelled to find a southerly directed channel now easily followed to the northern horn of the crescent of Lake Maumee.

The later Lake Maumee.—When the ice lobe had retired its front sufficiently, an outlet lower than that at Fort Wayne was uncovered past the city of Imlay, Michigan, into the Grand River, and thence through Lake Chicago and its outlet into the Mississippi. This old outlet south of Chicago follows the course of the present Drainage Canal and the line of the Chicago & Alton Railway. The traveler journeying southward by train from Chicago has thus the opportunity of observing first the beaches of the former lake, and then the several channels which were joined in the main outlet at the station of Sag (plate 20 A).

In this stage of our history Lake Maumee pushed a shrunk arm up past the site of Ypsilanti in Michigan (Fig. 361), the well-marked beach being found on Summit Street opposite the State Normal College. The Huron River, which in the first lake stage had followed the valley now occupied by the Raisin River southward into Indiana, now discharged directly into a bay upon this arm of Lake Maumee, and so formed a delta at Ann Arbor.

Lakes Arkona and Whittlesey.—The ice front in the Huron-Erie basin now retired so far that the impounded waters, instead of following the more direct “Imlay outlet” to the Grand, passed at a lower level completely around “the thumb” of Michigan into the Saginaw basin. Meanwhile a crescent-shaped lake had developed in that basin, so that now the waters of the Maumee basin were joined to those in the Saginaw basin as a common lake, just as the lowering of the waters in Glen Roy caused a union with those of Glen Glaster in the example cited for illustration. Our records of this third North American lake stage, referred to as Lake Arkona, are however most imperfect, for the reason that it was followed by a readvance of the ice front which closed the passage around “the thumb” and raised the level of the waters until an outlet was found past the town of Ubly at a lower level than the “Imlay outlet.” When the waters of a lake are thus rising, strong beach formations result, and those of this stage, which is known as the Lake Whittlesey stage, are much the strongest that are found within the Huron-Erie basin. Traced for some three hundred miles entirely around the southern and western margins of Lake Erie, this beach is for much of the distance the famous “ridge road” (Fig. 362).

Lake Warren.—As the ice advance which had produced Lake Whittlesey came to an end, the normal recession was resumed and a lake once more formed as a body common to the Saginaw and Erie basins. This lake, known as Lake Warren, extended a shrunk arm far eastward along the ice front into western New York, though it was still blocked from entering the great Mohawk valley (Fig. 363).

Lakes Iroquois and Algonquin.—It must be evident that toward the close of the Lake Warren stage a profound change was imminent—a transfer of the glacial waters from their course to the Mississippi and the Gulf to the trench which crosses New York State and enters the Atlantic. So soon as the ice front had retired sufficiently to lay bare the bed of the Mohawk, an outlet was found by this route and its continuation down the Hudson valley to the sea. The Lake Ontario basin now became occupied by a considerably larger water body known as Lake Iroquois, and the three upper lakes, then joined as Lake Algonquin, discharged their combined waters into Lake Iroquois at first through a great channel now strongly marked across Ontario in the course of the Trent River and Lake Simcoe, the so-called “Trent outlet.” At this time a smaller Lake Erie probably occupied the basin of that lake, and later the Trent outlet was abandoned for the Port Huron outlet (Fig. 364).

The Nipissing Great Lakes.—We have now followed the ice front step by step in its retreat across the valley of the St. Lawrence system. The successive unblocking of outlets offers but one further possibility—the opening of the French River-Nipissing Lake-Ottawa River, or “North Bay outlet.” Though not so to-day, the bed of this ancient channel was then much lower than that of the “Mohawk outlet”, and so soon as the glacier had in its retreat uncovered this northern channel, the waters of the upper lakes discharged through it past the site of Ottawa and into an arm of the sea which then occupied the lower St. Lawrence valley and has been called the Champlain Gulf or Sea (Fig. 365). The level of the waters was lowered and the area of the lakes correspondingly reduced.

The reader who has had no opportunity to observe these ancient channels which carried the swollen waters of the former glacier lakes, will find it interesting to consider that every one of them has been fixed upon by engineers for improvement as artificial waterways. Thus we have the Illinois Drainage Canal and projected ship canal along the “Chicago outlet”, the projected Mississippi-Lake Erie Canal along the “Fort Wayne outlet”, the Grand River canal project to connect Lake Michigan and Saginaw Bay along the course of the “Grand River outlet”, the Trent Canal along the “Trent outlet”, the Erie Canal along the “Mohawk outlet”, and, lastly, the proposed Georgian Bay ship canal to the ocean along the “North Bay” or “Nipissing outlet.”

Summary of lake stages.—We have omitted in this summary of late lake history in the Laurentian basin all the less important lake stages, including some of a transitional nature which were represented by beaches and outlets easily traced to-day. This is because it is an outline only which it seems best to present, and the episodes of this abridged history may be tabulated as follows:

EPISODES OF GLACIAL LAKE HISTORY

Mississippi Drainage

Lake Maumee (early), Fort Wayne outlet.

Lake Maumee (late), Imlay City outlet.

Lake Arkona, “thumb” outlet.

Lake Whittlesey (with readvance of glacier), Ubly outlet.

Lake Warren, “thumb” outlet.

Atlantic Drainage

Lakes Iroquois and Algonquin (early), Trent and Mohawk outlets.

Lakes Iroquois and Algonquin (late), Port Huron and Mohawk outlets.

Nipissing Great Lakes, North Bay outlet.

Permanent changes of drainage affected by the glacier.—While the lake history which we have sketched is made up of episodes which endured only while the ice front lay between certain stations upon its retreat, there were none the less brought about the profoundest of permanent modifications in the drainage of the region. It is possible to restore upon maps in part only the preglacial drainage of the north central states, but we know at least that it was as different as may be from that which we find to-day. The Missouri and the Ohio take their courses to-day along the margin of the glaciated area as an inheritance from the border drainage of the ice age. Within the glaciated regions rivers have in many cases been compelled by morainal obstructions to enter upon new courses, or even to travel in the opposite direction along their former channels. In districts of considerable relief these diversions have sometimes caused the streams to plunge over the walls of deep valleys, and it may truthfully be said that we owe much of our most beautiful scenery in part to the carving and molding of glaciers, but especially to the cascades and waterfalls directly due to their interference with drainage.

Many diversions or reversals of former drainage lines, through the influence of the continental glacier, are at once suggested by the abnormal stream courses, which appear upon our maps, and the correctness of these suggestions may often be confirmed by very simple observations made upon the ground. The map of Fig. 366 shows how different was the preglacial drainage of the upper Ohio region from that of to-day.

An interesting additional example is furnished by the Still River which in Connecticut is tributary to the Farmington, and is no less remarkable for its abnormal northerly course and sluggish current perpetuated in its name, than for the way in which it is joined to the Farmington system (Fig. 367 A). A careful study of the district has shown that the Still River was once a part of the Naugatuck and flowed southward toward Long Island Sound like other rivers of the district (Fig. 367 B). It possessed, however, an advantage in a narrow belt of softer rock along its course, and because of this advantage it captured a portion of one of the tributaries to the Farmington (Fig. 367 C). The continental glacier later covered the region, and on its retreat laid down morainal obstructions directly across this river and also at the head of the severed arm of the Farmington tributary (Fig. 367 D). The now impounded waters found their lowest outlet near Sandy Brook, and in waterfalls and cascades the now reversed river falls one hundred feet to the bed of that stream. With the aid of the excellent topographic maps which are now supplied by a generous government at a merely nominal price, such bits of recent history may be read at many places within the glaciated region.

Glacial Lake Ojibway in the Hudson Bay drainage basin.—When by passing over the “height of land” in northern Ontario the greatly reduced continental glacier had vacated the basin of St. Lawrence drainage, it was in a position to impound those waters which normally drained to Hudson Bay. The lake which then came into existence has been called Lake Ojibway and was the latest of the entire series. Though of but recent discovery in a country till lately a trackless wilderness, its extension seems to have been that of the clay beds suited for farming. The beaches and outlets remain to be mapped when the country has been made more easily accessible.

Reading References for Chapter XXIII

Parallel roads of Glen Roy:—

Charles Darwin. Observations on the Parallel Roads of Glen Roy and of Other Parts of Lochaber in Scotland, with an attempt to prove that they are of Marine Origin, Phil. Trans., vol. 8, 1839, pp. 39-82.

Louis Agassiz. Geological Sketches, Boston, 1876, vol. 2, pp. 32-76.

T. T. Jamieson. On the Parallel Roads of Glen Roy and their Place in the History of the Glacial Period, Quart. Jour. Geol. Soc. Lond., vol. 19, 1863, pp. 235-259.

Glacial Lake Agassiz:—

Warren Upham. The Glacial Lake Agassiz. Mon. 25, U. S. Geol. Surv., pp. 658, pls. 38.

F. W. Sardeson. Beginning and Recession of St. Anthony’s Falls, Bull. Geol. Soc. Am., vol. 19, 1908, pp. 29-36.

Glacial lakes in the St. Lawrence valley:—

Chamberlin and Salisbury. Geology, vol. 3, pp. 394-405.

Frank Leverett. Outline of the History of the Great Lakes (Presidential Address), 12th Rept. Mich. Acad. Sci., 1910, pp. 19-42. The Pleistocene Features and Deposits of the Chicago Area. Chicago, 1897, pp. 86, pls. 8 (Chicago Outlet).

H. L. Fairchild. Glacial Lakes in Western New York, Bull. Geol. Soc. Am., vol. 6, 1895, pp. 353-374, pls. 18-23; Glacial Waters in Central New York. Bull. 127, N. Y. State Mus., 1909, pp. 66, pls. 42, and maps in cover.

Early lakes in the Erie basin:—

Frank Leverett. On the Correlation of Moraines with Raised Beaches of Lake Erie, Am. Jour. Sci. (3), vol. 43, 1892, pp. 281-301.

F. B. Taylor. The Great Ice Dams of Lakes Maumee, Whittlesey, and Warren, Am. Geol., vol. 24, 1899, pp. 6-38, pls. 2-3; Relation of Lake Whittlesey to the Arkona Beaches, 7th Rept. Mich. Acad. Sci., 1905, pp. 30-36.

Frank Leverett. The Ann Arbor Folio, Folio No. 155, U. S. Geol. Surv., 1908, pp. 10-12.

---