the glacier makes its first advance, it may possibly sweep onwards loose material which has accumulated on the bed of the valley. This mixture of mud and stones, transported between the ice and the rock, is called ground moraine, or moraine profonde. As to its amount and importance' there is much dispute, some holding that a very large quantity of materials travels in this way, while others think it to be comparatively small.

Boulders and grit, as already stated, are engulfed in crevasses, and in some cases ultimately become embedded in the ice like currants in a cake. Their amount obviously depends partly on the quantity of surface moraine, partly on the number and depth of the crevasses. A smooth glacier may be almost buried under débris and yet swallow up but little; while one which is rent by crevasses may have its moraines comparatively small, because it engulfs them almost wholly. So also the amount of moraine stuff which makes its way from the top to the bottom of a glacier in some cases may be insignificant, while in others it may become large. But débris, varying in size from fine mud to huge boulders, is transported in greater or less proportions in these three ways: on the ice, in the ice, and under the ice.

Streams, fed by snow-beds or by hanging glaciers, occasionally descend the slopes on either side and plunge beneath the ice, bearing with them much débris and many partially rounded stones, which go

to augment either the ground moraine or the material which is swept along the main channel of subglacial drainage. Boulders which travel under the ice and help in scoring the subjacent rocks must themselves suffer a like treatment. It is often "diamond cut diamond." The burnisher is burnished; the scratcher is scratched in its turn. So while rocks which have travelled on or in the ice remain remarkably angular, those that have been dragged beneath it are smooth and worn, on some sides rather than on others, according as they are more easily retained in the grip of the ice. This causes them to assume a peculiar subangular form, the broader surfaces being marked by striations, which not seldom are irregular in

direction.

Yet one other mark is made by a glacier. Its surface is melted by the sun; the water trickles into rivulets, and the rivulets collect into streams, which quickly furrow a path for themselves; the ice is sculptured by a group of shallow channels like a model of a river basin on the land. But the course of the main stream of this drainage system is often cut short. If a crevasse opens across its path, the water plunges into the abyss and strikes on the rocks beneath. As the glacier advances, the friction of the falling water notches the ice in the opposite direction, so that the plunging action of the stream is limited to a comparatively small surface of the underlying rock; for if the ice travel forward more rapidly

than the stream can cut backward, a new chasm in all probability will open before long in the original position, and the cascade will begin to work afresh

over its former path.1

Boulders and débris from

time to time slip into the stream and are swept down

1 These wells in the ice are called moulins.

the waterfall; they are hurled against the rock beneath, and whirled round and round before they are carried onward along the subglacial trench which the torrent has worn, so that at last a hollow begins to be formed. This, when once it can retain some of the fallen blocks, is quickly deepened. As in the bed of any mountain torrent, so here a pot-hole is ground out, but the peculiar mode of excavation usually makes the hollow beneath the glacier deeper in proportion to its breadth than in the other case. It is frequently several feet, occasionally some yards, in depth. During the retreat of the ice these hollows, called "giants' kettles," often become filled up with mud and stones.

Such, then, are the effects of a glacier: namely, perched blocks and moraines, terminal, lateral, and subglacial; roches moutonnées; and surfaces smoothed or polished and striated or channelled, pitted occasionally with giants' kettles. Where these occur, it may be inferred that a glacier formerly existed, even though at the present time the snow vanishes from every recess in the surrounding mountains long before the summer reaches its end. It must be remembered, of course, that rocks and stones may be worn or striated by other causes than the action of a glacier. A mound need not always be a moraine; a perched block may possibly be due to some other agency; but if due care be taken to avoid mistakes, the aforesaid marks indicate the path of a glacier as certainly

as our footprints on the sand show that some one has been walking on the sea-shore.

We must now proceed to examine the valleys below the ends of the glaciers. In such a mountainchain as the Alps, the evidence, with many variations in detail owing to local circumstances, is everywhere substantially the same. Take, for instance, the valley of the Aar in the neighbourhood of the Grimsel hospice. The little tarn beneath the precipices of the Nägelisgrätli is surrounded by billowy hummocks of granitic rock, "like the backs of plunging dolphins." Far above the bed of the valley, up the steep crags leading to the "Dead Man's Lake," up the great mountain buttress round which the river bends sharply to the north, the rocks everywhere are rounded and smoothed and striated. (See Frontispiece.) All down the deep and narrow glen for some miles the story is the same: ice-worn, ice-scored rocks, perched blocks and scattered débris, prove that a vast glacier once welled up high against the precipitous slopes on either hand; for the marks extend almost from the present level of the stream to some two or three thousand feet above it, and further examination shows that many a lateral glen, in which now perhaps only a small snow-bed lingers in some sheltered corrie, once gave birth to a tributary glacier.

Another thing, however, is obvious; the dominant outlines of the valley are those indicative of the action of water, for it is V-shaped in section.

It