CHAPTER XVI

MODERN EARTH HISTORY

(_Cenozoic Era_)

Since the Cenozoic era is the last one of geologic time, it will be of particular interest to trace out the main events which have led up to the present day conditions, especially in North America. Both because of the recency of the time and the unusual accessibility of the rocks, which are mostly at or near the surface, our knowledge of the Cenozoic era is exceptionally detailed and accurate. It will, therefore, be more necessary than ever to select only the very significant features of this history for our brief discussion.

During the first half of the Tertiary period portions only of the Atlantic coastal plain were submerged under shallow water, but soon after the middle of the period (Miocene epoch) the sea spread over practically the whole Atlantic coastal plain area from Martha's Vineyard south to and including Florida. During the late Tertiary the marine waters had become greatly restricted, and by the close of the period the sea was entirely excluded from the Atlantic seaboard. The total thickness of these Tertiary strata is less than 1,000 feet, and they all tilt downward gently toward the sea. The strata consist mostly of unconsolidated sands, gravels, clays, marls, etc.

The Gulf coastal plain area from Florida through Texas and south through eastern Mexico was largely overspread by the sea during most of Tertiary time, except the latest. During early Tertiary time an arm of the Gulf reached north to the mouth of the Ohio River. Late in the period but little of the Gulf Plain was submerged, and at its close sea water was wholly excluded. On the Gulf Coast the Tertiary strata from 2,000 to 4,000 feet thick are also mainly sands, gravels, clays, and marls. They are commonly rich in fossils, and they show a gentle tilt downward toward the Gulf.

Throughout Tertiary time local portions of the Pacific border of the continent were submerged, this having been especially true of portions of California, Oregon, and Washington. In spite of the very restricted marine waters, the Tertiary strata of the Pacific Coast, especially in California, are remarkably thick, 10,000 to 20,000 feet being common, while the maximum thickness is fully 30,000 feet. Such great thicknesses are readily explained when we realize that erosion was notably speeded up by pronounced uplifts resulting from crustal disturbances toward the close of the preceding period, and again in the midst of the Tertiary period itself.

To summarize the Tertiary relations of sea and land for North America we may say that only local portions of the continental border ever became submerged, and that, by late Tertiary time, practically the whole continent was a land area. At the close of the period the continent was, as we shall see, even larger than now because the continental shelves of the ocean were then also largely above water.

The whole of the Cenozoic era, including both the Tertiary and Quaternary periods, has been a time of profound crustal disturbances throughout much of the continent, certain of these movements having continued right up to the present time, with positive evidence that some of them are still continuing. These great movements have included notable foldings of strata, uplifts without folding, faulting, and igneous activity, the whole effect having been to greatly increase the general altitude and ruggedness of the continent. In fact, North America is not known ever to have been at once higher, broader, and more rugged than it was very late in the Tertiary, or early in the Quaternary, period. Since that time the only notable change (barring the great Ice Age and its effects) has been a restriction of the area of the continent to its present size by spreading of sea waters over the borders of the continent, that is over the continental shelves.

We shall now rather systematically consider the more profound earth changes which have affected the continent, producing the existing major relief features, from west to east.

The "Coast Range Revolution" took place in the midst of the Tertiary period. Over the site of the Coast Ranges, strata had accumulated, especially during Cretaceous and earlier Tertiary times, to a thickness of thousands of feet. In middle Tertiary time these strata were subjected to a mountain-making force of compression and more or less folded, faulted (fractured), and uplifted into the Coast Range Mountains. Some portions of the range were intensely folded and faulted and upraised many thousands of feet, while other portions were only moderately folded and uplifted. It is an interesting fact that the great San Francisco earthquake rift or fault originated at this time. It was a renewed, sudden movement of a few feet along this fault which caused the disastrous earthquake of 1906. Still other considerable earth movements took place in the Coast Range region during late Tertiary and Quaternary times, as, for example, uplift without folding, as proved by distinct sea-cut terraces at altitudes of more than a thousand feet, like those north of San Francisco and south of Los Angeles. A moderate amount of still later subsidence has caused the development of San Francisco Bay. The large islands off the coast of southern California have in very recent geologic time (probably Quaternary) been cut off from the mainland by sinking of the land.

The Sierra Nevada Range, which originated by intense folding of rocks late in the Jurassic period, underwent profound erosion until about the middle of the Tertiary period, by which time it had been cut down to a range of hills or low mountains. Then the great fault (fracture) previously described began to develop along the eastern side. As a result of many sudden movements along this fault, which is hundreds of miles long, the vast earth block has been tilted westward with a very steep eastern face and a long, more gradual western slope, the crest of the fault block forming the summit of the range. The amount of nearly vertical displacement along this fault has been commonly from 10,000 to 20,000 feet, and, in spite of considerable erosion of the top of the fault block and accumulation of sediment at its eastern base, the modified fault face now usually stands out boldly from 2,000 to 10,000 feet high. As an evidence that this movement of faulting has not yet ceased we may cite the Inyo earthquake of 1872, when there was a sudden renewal of movement of ten to twenty-five feet along this fault for many miles. Since the great Sierra block began to tilt, the many mighty canyons, like Yosemite, Hetch-Hetchy, King's River, and Feather River, have been carved out by the action of streams, in some cases aided by former glaciers. King's River canyon has been sunk to a maximum depth of 6,900 feet in solid granite solely by the erosive action of the river!

The Cascade Mountains, too, were reduced to nearly a peneplain condition by late Tertiary time when they began to be rejuvenated by arching or bowing of the surface unaccompanied by great faulting or fracturing, and many canyons, like that of the Columbia River, have since been carved out.

Mention should now be made of the vigorous volcanic activity which took place in the Cascade and Sierra Nevada Ranges. Most of this activity occurred during Tertiary time (particularly in the latter part) and it has continued with diminishing force practically to the present time. In California streams of lava buried many gold-bearing river gravels which have yielded rich mines. Many well-known mountain peaks, such as Shasta, Lassen, Pitt, Hood, and Rainier, from northern California to Washington, are great volcanic cones which date from Tertiary time, and which are now mostly inactive. That this volcanic activity has not yet altogether ceased is shown by renewed eruptions of Mount Lassen (or Lassen Peak, altitude 10,437 feet) in northern California. Since the beginning of this renewed activity in 1914, several hundred outbursts have occurred. No molten rock has flowed out, but large quantities of rock fragments, dust and steam have been erupted, in many cases forming great clouds two or three miles high over the top of the mountain (Plate 10). At this writing (October, 1920), Mount Lassen is still showing vigorous activity. At Cinder Cone, only ten miles from Mount Lassen, there were two eruptions of cinders and a considerable outpouring of lava within the last 200 years. Still other very recent cinder cones occur in southeastern California and Arizona.

One of the greatest lava fields in the world forms the Columbian Plateau between western Wyoming (including the Yellowstone National Park) and the Cascade Mountains from northeastern California to northern Washington. It covers fully 200,000 square miles and is really considerably larger than shown on the map because the lava in parts of the plateau region are covered by very recent sedimentary materials.

The great lava fields of the Deccan, India, and of the plateau region of western Mexico are comparable in size to the Columbian field and these lava fields are all of the same age. In the Columbian Plateau most of the lava was poured out during later Tertiary time. Sheets of molten rock, averaging fifty to one hundred feet in thickness, spread out over various parts of the region and piled up by overlapping layers one over another until the lava plateau more than a mile high was built up. Many hills and low mountains were completely buried under the molten floods, and in other places the liquid rock masses flowed against the higher mountains. "For thousands of square miles the surface is a lava plain which meets the boundary mountains as a lake or sea meets a rugged and deeply indented coast.... The plateau was long in building. Between the layers are found in places old soil beds and forest grounds and the sediments of lakes.... So ancient are the latest floods in the Columbia River Basin that they have weathered to a residual yellow clay from thirty to sixty feet in depth, and marvelously rich in the mineral substances on which plants feed. In the Snake River Valley the latest lavas are much younger (Quaternary). Their surfaces are so fresh and undecayed that here the effusive eruptions may have continued to within the period of human history." (W. H. Norton.) Many of the lava layers are plainly visible where the Columbia River has cut its great gorge or canyon. The Snake River in places has sunk its channel several thousand feet into the lava plateau without reaching underlying rock.

Both north and south of the Columbian Plateau there was also much volcanic activity in the Rocky Mountain region during Tertiary time. A single formation in Colorado consists mostly of volcanic "ash" or dust over 2,000 feet thick. There was also much volcanic activity over the Colorado Plateau area of southern Utah, New Mexico, and Arizona. The volcanoes there exhibit all stages from those which are very recent and practically unaffected by erosion to others which have been completely cut away with the exception of the cores or "volcanic necks."

During the second half of the Tertiary period the whole region known as the Great Basin, between the Sierra Nevada Mountains of California and the Wasatch Mountains of Utah, began to be affected by profound faulting or fracturing and tilting of portions of the earth's crust. The two largest faults, one on the western side of the Wasatch Range and the other on the eastern side of the Sierra Range, are each hundreds of miles long. Each of these ranges owes most of its present altitude to the uptilting of great fault blocks, and most of the many nearly north-south Basin Ranges of Nevada and Utah are in reality recently tilted fault blocks.

Turning now to the Colorado Plateau, studies have shown that region to have been more or less periodically raised fully 20,000 feet since the beginning of Tertiary time, but because of profound erosion in the meantime its present altitude is only 6,000 to 9,000 feet. During late Tertiary time the land stood at a much lower level than to-day, so that, practically during the last period (Quaternary) of geologic time, the region has been elevated to its present position. As a direct result of this profound rejuvenation the Colorado River has had its erosive activity tremendously increased, and it has carved out the mightiest of all existing canyons--the Grand Canyon. The work of deepening and widening the canyon is still proceeding at a rapid geologic rate.

As we have learned, the Rocky Mountains and many of its subsidiary ranges were formed by folding and uplift of strata toward the close of the Mesozoic era (Cretaceous period). During much of Tertiary time the newly formed mountains had been considerably reduced by erosion. Then, late in the Tertiary period, much of the Rocky Mountain region, as well as much of the Great Plains area just east of the mountains, became rejuvenated by differential uplift without any notable folding of strata. We can tell that this general uplift amounted to at least several thousand feet because definite formations of relatively late Tertiary strata, originally horizontally deposited under inland bodies of water, gradually rise so that at the base of the Front Range of the Rockies they are fully 3,000 feet higher than they are 200 miles or more farther east. Thus, the original folding and faulting of the Rockies, Tertiary volcanic activity, late Tertiary rejuvenation, and subsequent erosion account for the present altitude and relief features of the great Rocky Mountain system.

Portions of the rejuvenated Great Plains region have been notably dissected by erosion since the late Tertiary, this being particularly true of the so-called "Bad Lands," especially in parts of Wyoming and South Dakota, where mostly relatively soft Tertiary strata have been cut to pieces.

Turning our attention now to the eastern half of the continent we find that all, or nearly all, of it was more or less raised toward the close of the Tertiary period. Practically the whole Mississippi Valley east of the Great Plains, as well as much of the country to the north in Canada, was elevated some hundreds of feet and the streams have since the late Tertiary uplift (except where the land was ice-covered during the Ice Age) been at work sinking their channels below the newly upraised surface.

As already pointed out, the lowlands of the Atlantic and Gulf Coastal Plains were mostly submerged under the sea during early middle Tertiary time. By the close of the period they had emerged practically to their present positions, and they have been only moderately affected by erosion.

We have still to explain the existing topography or relief of a large and important part of eastern North America, including the whole of the Appalachian Mountains, Allegheny Plateau, Piedmont Plateau, New York, New England, and the Canadian region to the north. As a starting point in this discussion we should recall the fact that, after the great Appalachian Mountain Revolution toward the close of the Paleozoic era, the predominant geologic process which affected the region under consideration was erosion throughout the succeeding Mesozoic era. By about the close of the Mesozoic (Cretaceous period) the whole region, with some local exceptions, has been worn down to a comparatively smooth plain (peneplain) not far above sea level. Local exceptions were mainly in the New York and New England region as, for example, some of the higher parts of the Adirondack and White Mountains, Mount Monadnock in southern New Hampshire, and Mount Greylock in western Massachusetts. These and other masses rose rather conspicuously above the general level of the great plain of erosion commonly called the "Cretaceous peneplain" because it is believed to have been well developed by the close of that period.

The uplift of the vast Cretaceous peneplain about the beginning of the Cenozoic era (Tertiary period) was an event of prime importance in the recent geological history of eastern North America because it was literally the initial step in bringing about nearly all of the existing major relief features of the Appalachian-New York-New England-St. Lawrence region. The amount of uplift (unaccompanied by folding) of the peneplain was commonly from a few hundred to a few thousand feet with the greatest amount in general along the main trend of the Appalachians. The fact should be emphasized that nearly all the principal topographic features of the great upraised region have been produced by dissection (erosion) of the uplifted peneplain surface. Thus nearly all the valleys, small and large, including those of the St. Lawrence, Hudson, Mohawk, Connecticut, and Susquehanna, have been carved out by streams since the uplift of the great peneplain.

The streams which flowed upon the old low-lying peneplain surface meandered sluggishly over deep alluvial or flood-plain deposits, and their courses were little if any determined by the character and structure of the underlying rocks, because, with few exceptions, all rocks were worn down to the general plain level. The uplift of the peneplain, however, caused great revival of activity of erosive power by the streams, the larger ones of which soon cut through the loose superficial alluvial deposits and then into the underlying bedrock. Thus the large, original streams had their courses well determined in the overlying deposits, and when the underlying rocks were reached the same courses had to be pursued entirely without reference to the underlying rock character and structure. Such streams are said to be "superimposed" because they have, so to speak, been let down upon and into the underlying rock masses. As Professor Berkey has well said: "The larger rivers, the great master streams, of the superimposed drainage system, in some cases were so efficient in the corrosion of their channels that the discovery of discordant structures (in the underlying rocks) has not been of sufficient influence to displace them, or reverse them, or even to shift them very far from their original direct course to the sea. They cut directly across mountain ridges because they flowed over the plain out of which these ridges have been carved, and because their own erosive and transporting power have exceeded those of any of their tributaries or neighbors."

Fine examples of such superimposed streams which are now entirely out of harmony with the structure of regions through which they flow are the Susquehanna, Delaware, and Hudson. Thus the Susquehanna cuts across a whole succession of Appalachian ridges while, in accordance with the same explanation, the Delaware cuts through the Kittatiny range or ridge at the famous Delaware Water Gap. The ridges are explained as follows: while the great master streams were cutting deep trenches or channels in hard and soft rock alike, numerous side streams (tributaries) came into existence and naturally mostly developed along belts of weak, easily eroded rock parallel to geologic (folded) structure. Thus the Appalachian valleys have been, and are being, formed, while the ridges represent the more resistant rock formations which have more effectually stood out against erosion. The lower Hudson River flows at a considerable angle across folded formations above the Highlands, after which it passes though a deep gorge which it has cut into the hard granite and other rocks of the Highlands. The simple explanation is that the Hudson had its course determined upon the surface of the upraised Cretaceous peneplain, and that it has been able to keep that course in spite of discordant structure and character of the underlying rocks. In a similar manner we may readily account for the passage of the Connecticut River through a great gap in the Holyoke ridge or range of hard lava in western Massachusetts.

Before leaving this part of our discussion we shall briefly present some evidence showing that the New York-New England-St. Lawrence region at least must have been considerably higher shortly before the Ice Age (Quaternary period). An old channel of the Hudson River has been traced about 100 miles eastward beyond the present mouth of the river and it forms a distinct trench under the shallow sea in the continental shelf. Even in the Hudson Valley, many miles above New York City, the bedrock bottom of the river lies hundreds of feet (near West Point, 800 feet) below sea level. Obviously this submerged channel must have been cut when the land in the general vicinity of New York City was fully 1,000 feet higher than at present. That the land thus stood higher late in the Tertiary and possibly early in the Quaternary periods is proved as follows: (1) because most of Tertiary time must have been needed for the river to erode such a deep valley after the initial uplift of the peneplain about the beginning of the period; and (2) because glacial deposits of Quaternary age filled the former channel to a considerable depth. The valleys of the coast of Maine, and the submerged lower St. Lawrence Valley (Gulf of St. Lawrence), in a similar way lead us to conclude that the region farther north was also notably higher just before the Ice Age.

In the eastern hemisphere early in the Tertiary period a great submergence set in and marine waters spread over much of western and southern Europe, northern Africa, and southern Asia. The sites of the Himalayas, Alps, Pyrenees, Apennines and other mountains were then mostly submerged. A very remarkable marine deposit, made up almost wholly of carbonate of lime shells of a single-celled animal called Nummulites, formed on the floor of this vastly expanded early Tertiary mediterranean. This rock attains a thickness of several thousand feet. It is doubtful if any other single formation made up almost entirely of the shells of but one species is at once so widespread and thick. In the Alps this remarkable marine deposit may be seen 10,000 feet above sea level, and in Tibet fully 20,000 feet. Much of the rock in the Egyptian pyramids was quarried from this formation.

Later in the Tertiary in Eurasia and Africa the marine waters gradually became very restricted, so that by the close of the period the relations of land and sea were not strikingly different from the present, although northwestern Europe, like northeastern North America, was notably higher just before the Ice Age than it is to-day.

Eurasia witnessed tremendous crustal disturbances during the middle and later Tertiary time when, due to intense folding and uplift of great zones, the Himalayas, Caucasus, Alps, Pyrenees, Apennines, and other great ranges were formed. The crustal disturbance was most remarkable in the region of the Alps, where the movement resulted in "elevating and folding the Tertiary and older strata into overturned, recumbent, and nearly horizontal folds, and pushing the southern or Lepontine Alps about sixty miles (over a low angle fault fracture) to the northward into the Helvetic region. Erosion has since carved up these overthrust sheets, leaving remnants lying on foundations which belong to a more northern portion of the ancient (early Tertiary) sea. Most noted of these residuals of overthrust masses is the Matterhorn, a mighty mountain without roots, a stranger in a foreign geologic environment." (C. Schuchert.)

The last period of geological time--the Quaternary--was ushered in by the spreading of vast sheets of ice over much of northern North America and northern Europe, and this ranks among the most interesting and remarkable events of known geological time. On first thought the former existence of such vast ice sheets seems unbelievable, but the Ice Age occurred so short a time ago that the records of the event are perfectly clear and conclusive. The fact of this great Ice Age was discovered by Louis Agassiz in 1837, and fully announced before the British Scientific Association in 1840. For some years the idea was opposed, especially by advocates of the so-called iceberg theory. Now, however, no important event of earth history is more firmly established, and no student of the subject ever questions the fact of the Quaternary Ice Age.

Some of the proofs of the former presence of the great ice sheet are as follows: (1) polished and striated rock surfaces which are precisely like those produced by existing glaciers, and which could not possibly have been produced by any other agency; (2) glacial bowlders or "erratics" which are often somewhat rounded and scratched, and which have often been transported many miles from their parent rock ledges (Plate 20); (3) true glacial moraines, especially terminal moraines, like that which extends the full length of Long Island and marks the southernmost limit of the great ice sheet; and (4) the generally widespread distribution over most of the glaciated area of heterogeneous glacial débris, both unstratified and stratified, which is clearly transported material and typically rests upon the bedrock by sharp contact.

The best known existing great ice sheets are those of Greenland and Antarctica, especially the former, which covers about 500,000 square miles. This glacier is so large and deep that only an occasional high rocky mountain projects above its surface, and the ice is known to be slowly moving outward in all directions from the interior to the margins of Greenland. Along the margins, where melting is more rapid, some land is exposed, and often the ice flows out into the ocean where it breaks off to form large icebergs.

The accompanying map shows the area of nearly 4,000,000 square miles of North America covered by ice at the time of maximum glaciation, and also the three great centers of accumulation and dispersal of the ice. The directions of flow from these centers have been determined by the study of the directions of many thousands of glacial scratches on rock ledges. The Labradorean (or Laurentide) glacier spread out 1,600 miles to the south to Long Island and near the mouth of the Ohio River. The vast Keewatin glacier sent a great lobe of ice nearly as far south, that is into northern Missouri. "One of the most marvelous features of the ice dispersion was the great extension of the Keewatin sheet from a low flat center westward and southward over what is now a semiarid plain, rising in the direction in which the ice moved, while the mountain glaciers on the west (Cordilleran region), where now known, pushed eastward but little beyond the foot-hills." (Chamberlin and Salisbury.)

The Labradorean and Keewatin ice sheets everywhere coalesced except in two places. One of these is an area of about 10,000 square miles mostly in southwestern Wisconsin. In spite of several ice invasions during the Ice Age, this area, hundreds of miles north of the southern limit of the ice sheets, was never ice-covered. There is a total absence of records of glaciation within this area, and so we here have an excellent sample of the kind of topography which prevailed over the northern Mississippi Valley just before the advent of the ice. A much smaller, nonglaciated area occurs in northeastern Missouri near the southern limit of ice extension.

The Cordilleran ice sheet was the smallest of the three, and it was probably not such a continuous mass of ice, the higher mountains projecting above its surface. A surprising fact is that neither this ice sheet nor any other overspread northern Alaska, which is well within the Arctic Circle, during the Ice Age. More than likely the temperature was low enough, but precipitation of snow was not sufficient to permit the building up of a great glacier.

At the same time that nearly 4,000,000 square miles of North America were ice-covered, about 600,000 square miles of northern Europe were buried under ice which spread from the one great center over Scandinavia southwest, south, and southeast over most of the British Isles, well into Germany, and well into Russia.

In both North America and Europe the high mountains, well south of the great glacier limits, especially the Sierras, Rockies, Alps, Pyrenees, and Caucasus, supported many large local glaciers in valleys which now contain none at all or only relatively small ones.

Records of glaciation, such as glacial scratches, bowlders, lakes, etc., occur high up in the White and Green Mountains, Adirondacks, Catskills, and the Berkshire Hills, thus proving that the ice must have been at least some thousands of feet thick over New England and New York. We have good reason to believe that even the highest summits, except possibly in the Catskills, from 4,000 to over 6,000 feet above sea level, were completely submerged under the ice. On top of a mountain of Archeozoic granite nearly 4,000 feet in altitude, facing the St. Lawrence Valley in northern New York, the writer has found many fragments of sandstone which were picked off by the ice in the low valley, moved southward a good many miles, and uphill several thousand feet to the top of the mountain. The reader may wonder how a great glacier at least a mile thick in northern New York could have thinned out to disappearance within the short distance to the southern border of the State, but observations on existing large glaciers show that it is quite the habit for them to thin out very rapidly near their margins, thus producing steep ice fronts.

The fact that glacial ice flows as though it were a viscous substance is well known from studies of valley glaciers in the Alps and Alaska, and the great ice sheet of Greenland. A common assumption, either that the land at one of the great centers of ice accumulation during the Ice Age must have been many thousands of feet higher, or that the ice must there have been immensely thick, in order to permit ice flowage so far out from the center, is not necessary. Viscous tar slowly poured upon a level surface will gradually flow out in all directions, and at no time need the tar at the center of accumulation be very much thicker than elsewhere. The movement of glacial ice from the great centers of dispersal during the Ice Age was much the same in principle, only in the case of the glaciers the accumulations of snow and ice were by no means confined to the immediate centers.

The fronts of the vast ice sheets, like those of ordinary valley glaciers, must have undergone many advances and retreats of greater or less consequence. In the northern Mississippi Valley, and also in Europe, there is positive proof for five or six important advances and retreats of the ice which gave rise to the true interglacial stages. The strongest evidence is the presence of successive layers of glacial (morainic) débris piled one upon another, a given layer often having been oxidized, eroded, and even covered with plant life before the next or overlying layer was deposited. Such is the condition of things throughout much of Iowa, where wells sunk into the glacial deposits commonly pass through layers of partly decomposed vegetable matter at depths of from 100 to 300 feet. Near Toronto, Canada, the finding of warm climate plants between two glacial deposits proves that the climate there during an interglacial stage was much like that of the southern States to-day. During the great interglacial stages the vast glaciers were notably restricted in size, and in some or possibly all, cases they may have wholly disappeared from the continent.

In former years there was a tendency to ascribe mighty erosive power to the vast slow-moving ice sheets, but to-day scarcely any geologist would hold that the ice really produced large valleys solely by ice erosion, or that mountains were notably cut down. Throughout the glaciated region, especially toward the north, the deep preglacial residual soils and rotten rocks were nearly all scoured off by the passage of the ice. That the ice, where properly shod with rock fragments, actually eroded to at least little depths into hard and fresh rocks is well known, but the evidence is clear and conclusive that the preglacial hills and mountains, and most of the valleys (including all the large ones), were rarely more than a little modified in shape and size.

One of the principal effects of the Ice Age is the widespread distribution of glacial deposits, and other deposits which were formed under water in direct association with the ice. Such materials have been described in the chapter on "Glaciers and Their Work."

As a direct result of the Ice Age, many thousands of lakes came into existence throughout the glaciated region where few, if any, previously existed. Many of these lasted only while the ice was present because their waters were held up by walls of ice acting as dams. Thousands of others still persist, most of these having their water levels maintained by dams of glacial débris left by the ice across valleys. Good examples of lakes of both types, including a summary of the remarkable history of the Great Lakes, are considered in the chapter on "A Study of Lakes."

Many drainage changes, gorges, and waterfalls have also directly resulted from the great Ice Age. In fact it is not too much to say that practically all true gorges and waterfalls of the glaciated region have originated as a direct result of the Ice Age. The most remarkable combination of waterfall and gorge thus produced is that of the world-famous Niagara, described in the chapter on "Stream Work." Not only are Niagara Falls and gorge of postglacial origin but there was no Niagara River as such before the Ice Age. In New York the well-known Ausable Chasm, Trenton Falls Gorge, and Watkins Glen are all excellent examples of gorges cut since the Ice Age by streams which, because their old valleys were filled with glacial débris, have been forced to take new courses. A gorge of very special interest is that at Little Falls in central New York. This gorge, two miles long, with its precipitous walls hundreds of feet high, is the most important gateway for traffic between the Atlantic border and the Great Lakes region. The bottom of this defile contains six tracks of the New York Central and West Shore Railroads, the Barge Canal, an important highway, and the Mohawk River. Before the Ice Age there was a stream divide instead of a gorge, several hundred feet above the present river level. During a late stage of the Ice Age, when the Great Lakes drained through the Mohawk Valley, a tremendous volume of water passed over the divide and cut it down to form nearly all of the gorge except the inner or bottom trench which has since been eroded by the Mohawk River.

Only a few of the numerous stream changes directly due to the Ice Age will be briefly referred to. Certain of the principles involved are exceptionally well illustrated in the general vicinity of Saratoga Springs and Lake George, New York. During the retreat of the great glacier a lobe of ice occupied the Lake George Valley and forced the Hudson River west over a divide at Stony Creek. Then, because of heavy glacial deposits near Corinth, the Hudson could not continue south through what had been the preglacial valley of Luzerne River, but it was forced eastward over a divide in a low mountain ridge to Glens Falls. The remarkable shift of the Sacandaga River from its preglacial channel was caused by the building up of a great morainic ridge across the valley in the vicinity of Broadalbin.

The drainage of the basin of the upper Ohio River has also been revolutionized as a result of the glaciation. All the drainage of western Pennsylvania passed northward into Lake Erie just before the Ice Age instead of southwestward through the Ohio River as at present.

Rivers as large as the Mississippi and the Missouri were also more or less locally deflected from their preglacial courses. Thus the Missouri, which in preglacial time followed the James River Valley of eastern South Dakota, was forced, by a great lobe of retreating ice, to find its present course many miles farther west.

How long ago did the Ice Age end? In seeking an answer to this question we should bear in mind not only the fact that the Ice Age ended at different times, according to latitude, the more southern districts having been first freed from ice, but also the fact that approximately 4,000,000 square miles of the polar regions are now ice-covered, so that in a real sense those portions of the earth are still in an Ice Age. Some of the best estimates of the length of postglacial time for a given place are based upon the rate of recession of Niagara Falls, the average of the estimates being about 25,000 years. The evidence for this conclusion is briefly set forth in Chapter III. A careful study of the rate of recession of St. Anthony Falls, Minnesota, has led to the conclusion that the last retreat of the ice occurred there from 10,000 to 16,000 years ago. Certain clays deposited under tidewater since the last withdrawal of ice in Sweden show a remarkable succession of alternating layers thought to represent seasonal changes. By counting the layers it has been estimated that Stockholm was freed from ice only 9,000 years ago.

Although the actual duration of the Ice Age is by no means accurately known, we can be quite sure that the time represented is far longer than that of postglacial time. That it must have lasted fully 500,000 years seems certain when due consideration is given to amount of time necessary to bring about the repeated changes of climate between the glacial and interglacial stages; the amount of plant accumulation during the interglacial stages; the amount of weathering and erosion of the various layers of glacial deposits. Some estimates run as high as 1,500,000 years for the duration of the Ice Age, and an average is about 1,000,000 years, which probably indicates, at least roughly, the order of magnitude of the time involved.

When it is considered not only that the fact of the great Ice Age was not even thought of until 1837, but also that many factors enter into the general problem of the climate of geologic time, it is not surprising that the cause (or causes) of the glacial climate is still not definitely known. A few of the various hypotheses which have been advocated to account for the glacial climate will now be very briefly referred to. One is that the increased cold (not more than 10 to 15 degrees for the yearly average) was brought about by the notably increased altitudes of late Tertiary and early Quaternary times in northern North America and Europe. In this connection it is interesting to note that the four times of real glaciation during geologic time (mid-Proterozoic, early Paleozoic, late Paleozoic, and early Cenozoic) did occur directly after great crustal disturbances and notable uplifts of land. According to this hypothesis the interglacial stages would have to be explained by a rather unreasonable assumption of repeated rising and sinking of the glaciated lands.

Another hypothesis, long held in favor, is based upon certain astronomical considerations. Thus we now have winter in the northern hemisphere when the earth is nearest the sun, but in about 10,500 years, due to wobbling of the earth on its axis, our winter will occur when the earth is farthest from the sun, thus making the winters longer and colder, and the summers shorter and hotter. After a much longer period of time the earth will be millions of miles farther from the sun in winter than in summer and this would still further accentuate the length and coldness of the winters. The interglacial stages represent the 10,500 year periods when the earth in winter (northern hemisphere) is nearest the sun. A difficulty in the way of accepting this hypothesis is that it is inconceivable that each glacial and interglacial stage lasted only 10,500 years. Another objection to the hypothesis as an explanation of Ice Ages is that it is directly opposed by the fact of widespread glaciation at low latitudes either side of the equator during the late Paleozoic Ice Age.

Another hypothesis is based upon variations in quantity of carbonic acid gas and water vapor in the air. Increase or decrease of these constituents causes increase or decrease of temperature because they have high capacities for absorbing heat. "The great elevation of the land at the close of the Tertiary seems to afford conditions favorable both for the consumption of carbon dioxide in large quantities (by weathering of rocks) and for the reduction of the water content of the air. Depletion of these heat-absorbing elements was equivalent to the thinning of the thermal blanket which they constitute. If it was thinned, the temperature was reduced.... By variations in the consumption of carbon dioxide, especially in its absorption and escape from the ocean, the hypothesis attempts to explain the periodicity of glaciation (i.e., glacial and interglacial stages)." (Chamberlin and Salisbury.)

Still another suggested explanation is based upon variability of amount of heat radiated by the sun. Slight variations are now known to take place, and possibly in the past during certain periods of time these variations may have been sufficiently great to cause a glacial climate with interglacial stages.

Here, as in the case of so many other great natural phenomena, a single, simple explanation does not seem sufficient to account for all the features of the several well-known glacial epochs of geologic time. Two or more hypotheses, or parts of hypotheses, must more than likely be combined to explain a particular Ice Age.