It is still a debated question, whether there was a single Glacial age, or epoch, during which the ice-sheet, though having many episodes of advance, never entirely disappeared, or whether there were several distinct Glacial ages, when the snow accumulated to form an ice-cap which spread out widely from its centres of dispersal, separated by Interglacial ages, when the ice-cap completely melted away. Among students of these problems the present tendency is to accept the multiple character of the Glacial and Interglacial stages, one of the strongest arguments for which is the evidence of fossils showing the return of mild and even warm conditions in some of the Interglacial ages. At the same time, there is much difference of opinion regarding the number of these disappearances and reappearances of the ice.

Obviously, the problem is one of much difficulty, because each advance of the ice would tend to remove the older drift, or to bury it out of sight under new accumulations, when erosion was insufficient to remove it. Only on the margins of the successive ice-sheets, where they but partially coincided, should we expect to find the evidence preserved. A series of such advances and retreats of the ice must have produced an exceedingly complex succession of stratified and unstratified drift, and it is not surprising that the interpretations of such obscure phenomena should differ. If the superposed sheets of drift, one over the other (an arrangement which is not questioned), were separated by long, truly interglacial times, then each drift-sheet in turn must have been exposed to the denuding agencies for corresponding lengths of time and should exhibit the various stages of chemical and mechanical disintegration proportioned to the length of exposure. Between the earlier and later drifts there should be manifest differences in this respect.

Further, to complete the evidence, interglacial deposits, with testimony of climatic amelioration from the fossils, should be observed.

The following comparative table gives the views on this subject of Professors Chamberlin and Salisbury (I) for the Mississippi valley, of Professor James Geikie (II) for Europe, and of the Prussian Geological Survey (III) for North Germany.

In the Alps, Professor Penck has determined three Glacial stages, and Huntington has found evidence of five in the mountains of Turkestan.

The table must not be understood as attempting to correlate the events in Europe and America, as that would be premature. At any rate, the events in the two continents did not correspond in the way which the table seems to imply. For example, in the last of Geikie's Glacial stages, the Upper and Lower Turbarian, the glaciers are described as being restricted to the high lands and mountains, not forming a general ice-sheet. The Mecklenburgian and Polandian more nearly correspond to the Wisconsin.

American

At the time of greatest expansion the ice-sheets covered nearly all of North America down to lat. 400 N., anticipating the conditions of modern Greenland, though on a vastly larger scale. Three distinct centres or areas of maximum accumulation of the ice have been identified in northern Canada, from which the great ice-sheets flowed outward in all directions, though each one of the sheets had its own episodes of advance and retreat, so that the same region of country was overflowed, now by extensions from one sheet, and again by those from another. One oi these centres of accumulation and distribution lay to the north of the St. Lawrence River, and on the highlands of Labrador, sending its ice-mantle southward over the Maritime Provinces, New England, and the Middle States, as far west as the Mississippi, River. This is called the Laurentide, or Labradorean Ice-sheet or Glacier. A second centre was near the west coast of Hudson's Bay, and from this area the ice streamed outward in all directions westward toward the Rocky Mountains, northward to the Arctic Ocean, eastward into Hudson's Bay, southward through Manitoba into the Dakotas, Minnesota, and Iowa. This great ice-sheet has been named the Keewatin Glacier, from the Canadian district of that name.

A third centre was formed by the Cordillera of British Columbia, which for a distance of 1200 miles was buried under a great ice-mantle that flowed both to the northwestward and southeastward. To these large and well-defined centres should probably be added a fourth, Newfoundland, from which, there is reason to think, came the ice which crossed Cape Cod and extended over Nantucket Sound to the island.

In addition to the great northern ice-cap, large local glaciers accumulated in all the western mountains ranges: the Rocky Mountains, as far south as New Mexico, the Uinta, Wasatch, and Bighorn ranges, and the Sierras and the Cascades, even the San Francisco Mountains of northern Arizona, and the other ranges of the western Cordillera, all bore thousands of glaciers. In these mountains almost every valley shows the evidences of former glaciation, in cirques at the head, in the smoothed and striated walls and bed, and in the moraines at the foot. The mountains of Alaska were heavily glaciated, but not the lowlands.

In the Mississippi valley the Pleistocene sequence is best displayed. The first known advance of the ice (sub-Aftonian) is registered in much disintegrated drift, which is exposed by denudation in Iowa. A similar sheet of very old and much worn drift which extends from beneath much later drift in New Jersey and Pennsylvania may be of the same date.

A great retreat of the ice, if not its entire disappearance, brought about interglacial conditions at least in the Mississippi valley (AfIonian stage). The surface exposed by the retiring ice was occupied by vegetation, which in many places in Iowa formed accumulations of peat, sometimes to the depth of 25 feet. The Kansan stage represents the greatest extension southwestward of the ice-sheet, when the glacier descended from the north (perhaps the Keewatin glacier) nearly to the mouth of the Ohio River, and spread across Iowa and Missouri far into Kansas and Nebraska. East of the Mississippi the Kansan drift has not been recognized. Again came a time of retreat, when soil was formed, and the Kansan drift was eroded and deeply decomposed, and peat deposited upon it ( Yarmouth, or Buchanan stage). A renewed extension of the ice laid down the Illinois till-sheet, which is found not only in that State, but in Iowa also, overlapping the Kansan drift, and it extends to Wisconsin, eastward into Ohio and Indiana, and passes under later till-sheets to the northeast.

This Illinoian drift appears to be derived from the Laurentide glacier.