cial débris extruded at the lower end of the ice-flow bears, usually, a very small proportion to the supply of rock-rubbish travelling at the surface. This, however, is not invariably the case, even in the Alps. Not infrequently small "summit glaciers," lying upon mountain-slopes, bear no superficial detritus, while infraglacial débris, nevertheless, is constantly being extruded at their lower ends. Thus the small Stampflkees Glacier (Zillerthal), overlooked by hardly any exposed rock-surfaces, and consequently carrying little or no superficial rock-rubbish, yet exhibits at its terminal front a bottom- or ground-moraine some ten or fifteen feet thick. But that which is the exception. in Alpine lands is the rule in Arctic regions. The tongues of ice protruding from the vast mer de glace of Greenland are almost entirely free from the superficial débris, and yet they eject ground-moraine in abundance. The same, as we shall see presently, is the case with most of the Norwegian glaciers. It is obvious, therefore, that the relative importance of ground-moraine, as a product of glacial action, is really greater than a glance at the phenomena of any ordinary Alpine glacier would at first lead one to suppose.

The general nature of Alpine ground-moraine is well known. It consists simply of an aggregate of rock-fragments, grit, sand, and mud or clay, often frozen or pressed together, and so included in the lower or basal portion of the glacier. Many of the stones are subangular and blunted, and striated, smoothed, or polished on one or more sides. No

one doubts that this material has travelled underneath, and partly enclosed in the ice-flow, and that the rock-surface over which it has been carried is abraded, smoothed, and polished by its filing action. Everyone, in short, admits that some degree of erosion is the result of glacial action. Were that action entirely confined to mere abrasion and smoothing of rock-surfaces, it yet could hardly be considered insignificant. The fine powder or flour of rock which renders all glacial rivers turbid, shows that glacial grinding is really of great importance. It has been computed, for example, that the river extending from Aar Glacier carries away daily 280 tons of solid matter in suspension. Again, the Justedal Glacier, draining an ice-field 820 square miles in extent, discharges in a summer day 1968 tons of sediment. This is in excess of the average daily discharge during the year, which Helland estimates at 180 million kilogrammes. To this should be added the mineral matter carried in solution, amounting to 13 million kilogrammes, so that solid and dissolved materials taken together come up to 189,950 tons. This would form a mass equal to 90,252 cubic yards. According to the same geologist, the Vatnajökull (Iceland), draining an ice-field ten times larger than that of the Justedal, discharges annually 14,763,000 tons of sediment-an amount equal to 7,194,000 cubic yards of rock. Thus, even if a glacier does no more than abrade and smooth its bed, the amount of rock ground into powder is neither insignificant nor unimportant.

But is this all the erosion that a glacier accomplishes? What about the débris of its ground-moraine -whence is that derived? Professor Heim and others maintain that in the case of a large number of glaciers (Alps, Himalaya, New Zealand) infraglacial detritus comes chiefly from superficial sources. Overlying morainic rubbish, it is supposed, finds its way through crevasses to the bottom of the ice. Now there can be no doubt that surface-moraines are frequently engulfed in crevasses; but then the rockrubbish engulfed in this way sooner or later reappears at the surface of the glacier further down the valley. Obviously in such cases the débris does not descend to the bottom of the glacier, but is simply engorged at some distance from the surface, and again becomes exposed, owing to the curving upwards of the lines or planes of flow and the ablation of the surface. If crevasses penetrated the whole thickness of a glacier, doubtless débris plunging into them might well reach the bottom of the ice, and be included as ground-moraine. But the plasticity of ice necessarily limits the depths to which a crevasse can extend. The larger The larger glaciers, according to Heim, are never penetrated to the bottom by crevasses, which when not kept open and deepened by ablation do not exceed a depth of 100-150 metres. Superficially carried rock-rubbish, therefore, can reach the bottom of a moderately thick glacier only along the margin, where the crevasses open to the rock-head. Here and there, perhaps, débris may occasionally descend

by moulins; but as a rule the bed of such a glacier can receive only a very meagre supply of rock-fragments from above. And if this be the case with the relatively small glaciers of the Alps, it must be the same in a more marked degree with those of high northern and Arctic lands.

Reference has already been made to the fact that even in the Alps certain summit-glaciers are so placed that no debris is showered upon them, and yet these glaciers extrude more or less conspicuous groundmoraines. In a word, the existence of groundmoraines does not depend upon the presence of superficial moraines. The latter are not infrequently wanting; the former, on the contrary, never are. This is well seen in the case of the Norwegian glaciers, which, as compared with those of the Alps, might be described as almost devoid of surface-débris. Nevertheless, ground-moraines are always in evidence, appearing not only under the tongue-like glaciers which protrude from the plateau ice-fields, but at the base of the more or less steep walls in which those ice-fields usually terminate.

The great development of superficial moraines in the Alps as contrasted with their meagre appearance in Scandinavia is easily explained. In the former region we have a complicated series of mountaingroups and chains, the crests of which overlook profound cirque-like depressions. It is in these broad and deep troughs and basins that snows accumulate to form the reservoirs from which glaciers flow.

Even at its very source, therefore, an Alpine glacier has rock-débris supplied to it from above, and as it passes down its mountain-valley frost and avalanches keep up a constant bombardment, so that the farther it flows the greater becomes the amount of detritus eventually piled up in its terminal moraines. Norway, on the other hand, is a lofty plateau, more or less deeply trenched by fiords and valleys. The snows, therefore, accumulate upon a wide and relatively flat or undulating surface, not dominated by peaks or ridges. In the central part of a Norwegian snow-field the surface is more or less continuous, and seldom interrupted by crevasses. Now and again, however, these are encountered, and their walls show stratified névé above graduating downwards into compact ice. Towards the margin of such an ice-field crevasses become more frequent, and in these snow and névé are seen gradually thinning-off as the terminal wall is approached, until at last the blue ice is wholly exposed. In short, the Scandinavian plateau supports true ice-sheets, comparable in all respects, save as regards their extent, to the great "inland ice" of Greenland. In places, longer or shorter tongues of ice project from the ice-sheet into valleys; in other places, where no valleys are present, the sheet simply terminates in a continuous ice-wall.

Such being the character of the Scandinavian icefields, we need not wonder at the absence of superficial moraines. No mountains overlook the plateaux ; it is only when the ice creeps outwards into valleys