of similar phenomena in connection with some of the large tongues of ice which are protruded from the great "inland ice" of Greenland. He says: "The rubbing of the glacier (Bowdoin Glacier) against the shoulders of rock projecting from the side of the valley gave opportunity for observing some of the special phenomena of such situations. At one point the process of plucking' was well indicated (though not actually observed) on the lee-slope of a spur of gneissoid rock. Blocks ranging up to three or four feet in width and length, and one or two feet in thickness had been detached in considerable numbers. The process involved much breaking and bruising with relatively little wear. Corners and angles were broken off, and heavy bruise marks were observed both on the blocks and on the sides and edges of the cavities from which they had been removed. At some points considerable crushed rock was observed. On the other hand, systematic grooves and striæ were not abundant nor pronounced. The dynamic impression given was that of a forceful tearing out of blocks by the action of a relatively rigid agency, which did not press the blocks hard upon the leeslope after their removal."

It is clear, then, that under existing glaciers and ice-fields rocks are sometimes smoothed and polished, sometimes crushed and shattered. The pressure of the ice tends to disrupt rock-masses, which yield or resist according to their character and structure, and fragments detached must often serve as wedges to

dislocate and detach others. Nor can it be doubted that the rocky bed of a glacier is also attacked by frost. The constant outflow of water shows that infraglacial melting goes on all the year round. The temperature at the bottom of the ice oscillates about the freezing-point, and as a glacier flows on its way thawing and freezing must be continually taking place. In this way joints are no doubt opened, rock-masses loosened, and larger and smaller fragments become more readily plucked and dragged out of place.

We cannot, therefore, hesitate to conclude that ice in motion, whether in the form of glaciers or of icecaps, is a powerful agent of erosion. It not only abrades and smooths, but breaks up and quarries the rocks over which it flows, and the débris thus obtained constitutes the true ground-moraine.

2. Geological action of prehistoric glaciers. Geologists rightly insist upon the potency of river-erosion. The study of modern denudation has quite convinced them that valleys can be and have been excavated by running water. In proof of this they point not only to the present action of rivers-to the rate of transport of sediment-but to the immense accumulations formed by river-action in prehistoric times. The broad alluvial plains of river-valleys, the great deltas which encroach upon the sea, the wide stretches of flat lands occupying the sites of silted-up lakes, are all cited as evidence of the potency of running water as a producer and transporter of sediment. So in like manner the glacialist appeals to far-ex

tended accumulations of ground-moraines as proof of the efficiency of flowing ice as an agent of erosion and transport.

The study of modern glacial action is carried on under certain obvious disadvantages. The bed of a glacier is concealed from our view. Now and again we may get a peep under the ice; or, better still, we may have the opportunity of examining the ground from which a glacier has temporarily retired. But the portions of a glacier's bed thus at times exposed are not those where erosive action is most intense. A glacier thins away towards its extremity, and the rate of motion at the same time diminishes, so that pressure and erosion must decrease with the attenuation of the ice. To such an extent is this the case, that the snout of a glacier deploying upon a relatively flat surface often rests upon its terminal moraines, or even overrides the fluvio-glacial gravels spread out in front of it. Such facts have led some observers to conclude that glaciers do not erode at all, and did the facts referred to stand alone there would be some justification for that conclusion. should be remembered, however, that were observers of river-action to confine attention to the broad plaintrack to the region known as the "base-level of erosion "-they would no doubt readily come to the conclusion that running water transports and deposits sediment; but, by following the process of reasoning just alluded to, they might also infer that rivers are incapable of erosion. Were the beds of existing

It

glaciers as open to investigation as the channels of rivers, we should probably hear little about the feeble erosive action of ice. But although we cannot make direct observations underneath the central and thicker portions of a glacier, we can yet examine great valleys and broad lowland regions which have been formerly subjected to intense glaciation. And the evidence of effective glacial erosion there displayed is too clear to be wholly misunderstood. Let us then consider the general results which have been obtained by the careful investigation of certain well known glaciated regions-the Alpine lands of Central Europe.

At the climax of the Glacial Period the snow-line in the Alps appears to have been upon an average some 4700 feet lower than now. Viewed from the north, the mountains must at that time have presented the appearance of a great ice-field, broken here and there by Nunatakkr—the protruding peaks of the dominant elevations of the secondary ranges, and bounded on the south by the snow-clad ridges of the Central Chain. In a word, so thick was the ice. in the valleys that as the glaciers made their way to the low grounds they frequently coalesced or became confluent across intervening mountain-ridges. Under such conditions it is obvious that the formation and accumulation of superficial moraines must have been relatively limited. The area buried under névé and ice was greatly in excess of that which remained. uncovered. If it be true, therefore, that groundmoraines consist chiefly of rock-débris derived from

superficial sources, those of the Glacial Period should be of little importance. The very reverse, however, is the case. The ground-moraines assume an enormous development, their dimensions being in direct proportion to the size of the ice-flows. The larger the body of ice, the greater the mass of ground-moraine.

It must be admitted, therefore, that the materials. of the old ground-moraine cannot have been derived from superficial sources. Some have suggested, however, that the accumulations in question consist to a large extent of the products of weathering, of torrential and fluviatile action, which had gathered over the mountain-slopes and in the valleys before the advent of the Glacial Period. There is no reason to believe, however, that rock-rubbish throughout the Alpine lands attained a greater development at the beginning of the Ice Age than it does now. The old snow-fields and glaciers doubtless gradually extended as the temperature fell. As the depression of the snow-line continued, rock-rubbish would accumulate abundantly, just as at present, in every valley occupied by a glacier. For a long time, too, superficial moraines would assume a relatively great importance, so that large terminal moraines would mark every pause in the progress of the ice-flows. But as the glaciers thickened in the valleys, and more and more bare rock disappeared below the ice, the supply of detritus from above would become gradually limited, until in many places, as in the region of the secondary ranges, it practically ceased altogether. Were a glacial period