land; and that during its reëmergence the arrangement of the materials which cover it was modified by exposure to the distributing and stratifying action of the waves, tides, and currents. The extent and immense number of modern icebergs seem to prove their capacity to reproduce upon the shoals and over the bottom of the Atlantic nearly all the phenomena of the drift formation. Measured as they are by miles in length, and rising at times more than 300 ft. in height, with only one fifth of their bulk then visible, they may well float off and distribute along their track the largest bowlders which they have abstracted from the rocky cliffs down which they moved as glaciers into the sea. Of late years, however, the theory of Agassiz, that the phenomena of the drift are due not to the submarine action of floating ice, but to terrestrial glaciers, has found much favor. The vast accumulations of ice which have been so well studied in the Alps are seen in their slow and irresistible motion down the valleys to score and groove the surfaces of the rocks over which they pass, rending masses of rock from the cliffs, moving the fragments forward, and finally leaving them rolled in the shape of bowlders, and grooved and scratched by the rubbing to which they were subjected when fixed in the ice. In the Alpine regions of Europe the effects thus produced are so remarkable, and spread over such extensive districts, that many geologists who have studied them are disposed to refer all the phenomena of the unstratified drift to the action of glaciers; and in this view they are confirmed by finding unmistakable evidence that the action of the glaciers formerly extended far beyond their present limits. But the unstratified drift is found to extend over vast regions where it is difficult to conceive that mountain glaciers could ever have found their way; and Agassiz, to account for this universal glaciation of circumpolar regions, has been led to maintain the existence of a great continental glacier, or ice cap, extending over the arctic and a great part of the temperate zone, moving downward from the polar region, and of such immense thickness as to surround and overflow the summits of our highest hills, which he supposes may have required in eastern North America a vertical thickness of two or three miles of solid ice. A similar continental glacier, according to this view, must have existed in the southern hemisphere. Prof. Dana, while adopting this notion of the origin of the glacial drift, regards the hypothesis of a central and common glacier source for each hemisphere as untenable, but supposes the existence of distinct glaciers of great magnitude. Such a one, according to him, had its origin along the watershed between the St. Lawrence and Hudson bay; but recognizing the necessity of an elevated source to give motion to the glacier, he supposes that this region, which is now not more than 1,500 ft. above the sea, was then raised many thousand feet above its present level.—In these theories of land glaciation a great depression of the surface is supposed to have succeeded the glacial period, affecting in the one case the great mountain plateau to the northward, and submerging the glaciated region so as to permit the deposition above its surface of the stratified clays and sands which overlie the bowlder clay. But, as we have seen, this in many parts of its distribution is clearly of marine origin; and a careful study of the whole of the phenomena of the drift period in eastern North America has led Dawson to regard the operation of land glaciers in this region as of very limited extent and importance, and to maintain that the widespread glacial drift is essentially submarine in its origin. This earlier view, which as set forth by Lyell has been partially explained above, endeavors to account for the phenomena in question by causes now in operation, rather than by supposing a condition of things which it is at once difficult to conceive and to explain. As expounded by Dawson, it maintains that at the beginning of the glacial time eastern North America was already under water, and was slowly rising, though with minor oscillations of level, from the ocean, the more western portion first. Along the eastern border of the rising land, over its still submerged plains, and through its valleys, then flowed the arctic current, as it now does along the coast of Labrador and the shores of Newfoundland, bearing great quantities of floating ice, by the action of which, combined with the current, the rocky strata were eroded, and the valleys and lake basins excavated. At an early period in this order of things, the great arctic stream, pursuing, in obedience to the force impressed upon it by the earth's rotation, a southwestern course, passed over the region of the great lakes and excavated their basins; while at a later time, diverted further eastward by the emergence of the Laurentides, it would pass along the present St. Lawrence valley, and thence southwestward to the Mississippi. To quote the language of Dawson: “The prominent southwestern striation, and the cutting of the upper lakes, demanded an outlet to the west for the arctic current. But both during the depression and the elevation of the land there must have been a time when this outlet was obstructed, and when the lower levels of New York, New England, and Canada were still under water. Then the valley of the Ottawa, that of the Mohawk, and the low countries between Lakes Ontario and Huron, and the valleys of Lake Champlain and the Connecticut, would be straits or arms of the sea; and the current, with its icebergs, obstructed in its direct flow, would set principally among these, and act on the rocks in north and south and northwest and southeast directions. To this portion of the process I would attribute the northwest and southeast striation.” As the process of elevation proceeded, and the northern current found its passage across the east-
