part of the ascent to the higher portion, and of the descent from it, is over a vast field of sloping or gently undulating snow.1 The contours of the buried region are perceptible only in the more immediate neighbourhood of the coast, and probably cannot be detected beyond about a hundred miles from it. Here the "Nunataks" begin to emerge from the ice and to rise above its surface like islands from a sea. On approaching these from the interior, the undulations. of the snowfields assume a more definite plan, and their slopes converge towards a broad shallow basinlike depression, which gradually deepens till the ice at the lower end runs like a broad causeway rather below the general level of the district. Here then the great central reservoir of ice is tapped by a definite drainage channel. This is the birthplace of one of the great glaciers. The ice no longer presents a smooth unbroken surface; crevasses become more frequent; travelling over it is at once more difficult and more dangerous. The scenery in one of these basins, though on a far grander scale, recalls that of the gathering-ground for one of the larger Alpine glaciers; such, for instance, as the névé of the Great Aletsch or the snow plateau which is the source of both the Gorner and the Findelen glaciers. In regions like the Alps, where the mountains rise more. steeply than in Greenland, lateral glaciers form among the rocky ridges on either side of the great

Here the breadth of the continent is about 350 miles.

trunk-stream; these sometimes become its tributaries, though often, when their own gathering-ground is small, they melt away before reaching it. If so, their terminal moraines are spread out on the steep slopes, but occasionally may be confused with the lateral moraine of the trunk-stream. Such glaciers are less conspicuous features in Greenland scenery, owing to the greater depth of the ice and the more monotonous contours of the spurs separating the valleys; but occasionally an offshoot from the main mass of the inland ice may pass through the depression between two "Nunataks" in the direction of the great effluent, the surface of which is already lying at a lower level. Thus the offshoot becomes relatively a tributary glacier, though from a common source, and the result of this is curious. No moraines can be seen on the inland ice, not even a solitary erratic spotting its pure white surface. These only make their appearance in the region of "Nunataks" and of valley glaciers. Even here they are generally insignificant compared with those on the glaciers of the Alps. The reason is obvious. Where no rock rises above the ice, no fragments can fall upon its surface; the possibility of a moraine ceases with the last "Nunataks," and as these are not large, the stream of blocks which they originate must be correspondingly small. Still, in a few cases, piled-up heaps of stone occur on the ice farther inland than might have been expected. One such

was observed during Nansen's expedition in the neighbourhood of the "Nunataks" which bear his name. The moraine was full 4000 yards in length; apparently it was about 500 feet high, but this magnitude proved to be illusory, for the débris was little more than a veneer to a hidden mound of ice, which had been screened by the material and thus kept from melting. Still, not only was this moraine. larger than might have been expected from the size and character of the adjacent "Nunataks," but also the blocks on it were all subangular and polished. In other words, the material of which it was composed had travelled not upon, but beneath, a glacier. Its presence may be thus explained: an offshoot from a higher part of the main ice-sheet had passed between two of the "Nunataks toward the surface of the trunk-stream below, and on this surface, which it had just succeeded in reaching, it had deposited its terminal moraine; this, however, as few blocks, if any, came from the adjacent "Nunataks," consisted of débris which had travelled beneath the ice, and thus bore the usual indications of such a passage.1

So far as can be ascertained, the quantity of this. "ground moraine," even in Greenland, usually is not great. In one respect the estimation of it should be comparatively easy, because the configuration of the surface is not favourable to the accumulation

1 From what source this may have been derived will be discussed later.

of much superglacial material, even when the icesheet terminates in valley glaciers which are bounded by steep and rocky slopes. Thus in Greenland moraines are comparatively inconspicuous about the smaller glaciers of the mountainous border-land, since it, like Norway, is commonly much less rugged than Switzerland. In the case of the great trunk-streams of ice, the exact quantity of subglacial material is difficult to estimate, because they almost invariably end in fjords, their steep ice-cliff rising for some hundreds of feet above the water and from considerable depths below it. But whenever one of the smaller offshoots from the inland ice is melted away before reaching the sea, as sometimes happens, it appears not to leave behind any large amount of ground moraine. In some cases the glaciers have retreated, leaving, not a thick sheet of clay mixed with boulders, but merely a few scanty patches on a bed of bare rock: the tract differs from that deserted by a Swiss glacier only in the greater flatness of the ice-worn surfaces, and the general absence of angular débris. Nevertheless a considerable quantity of mud and a number of stones must be constantly travelling beneath the Greenland ice, and these, under certain circumstances, may accumulate: the most favourable, apparently, will be when a glacier, which ends either on flattish ground above the sea-level or in very quiet water, remains for a considerable time almost stationary, and then very

slowly retreats.

Thus the material of its ground

moraine, though inconsiderable in amount at any one moment, accumulates to form a kind of terminal moraine, and this is drawn very gradually, like a coverlet, over the bed of the valley. In the same way lateral moraines of a similar character may be occasionally constructed of material extruded from beneath the ice.1

A fact which may prove to be of some importance may be noticed in passing. These huge ice-streams, though fed from the vast mass of inland ice, seldom or never protrude seawards. The valley which they occupy ends as a fjord, often several miles in length, down which the bergs detached from the great terminal cliff of the glacier drift slowly to the open sea. Even in Eastern Greenland, where the climate is more rigorous, the same rule holds as to the absence of ice from the coast. For instance, the great fjord called Scoresby Sound (about latitude 71°) stretches far inland, and even branches out into tributary watervalleys, like the Sogne Fjord in Norway, the heads of them being closed by a tongue of the inland ice

66

1 Nordenskiold (" Arctic Voyages," p. 169) describes a district near Disco Bay which had been recently abandoned by the land-ice as curiously like the woodless gneiss districts in Sweden and Finland. 'Everywhere occur rounded, but seldom scratched, hills of gneiss with erratic blocks in the most unstable positions of equilibrium, separated by valleys with small mountain-lakes and scratched rock-surfaces. On the other hand, no real moraines were discoverable. These, indeed, seem to be commonly absent in Scandinavia, and are, generally speaking, more characteristic of small glaciers than of real

inland ice."