intensified on the bed of the valley above the constriction, and a shallow basin will be ground out. On the disappearance of the glacier a lake will necessarily appear, and many such lakes occur in highly glaciated mountain tracts; frequently, however, lakes of this kind become silted up, and their former presence is then only indicated by flat sheets of alluvium. Again, it is well known that valley-basins of the normal type often show irregular depths, and it is not always easy to say how these have originated. Sometimes they are the result of valley constriction, sometimes of sudden changes in the direction of the valley, which have caused the ice to erode more energetically on one side than the other, for the line of most rapid motion in a glacier, as in a river, will shift from the centre to the side, or from side to side, with the windings of its course. Again, inequalities in the floor of a rock-basin may sometimes be due to the unequal resistance of the rocks. Nor must we forget that during its final melting a glacier might dump débris in a very confused fashion over its bed, while the subsequent deposition of alluvial matter swept into the lake at many different points by streams and torrents would similarly tend to produce inequalities. But all valley-lakes, it must be remembered, are not rock-basins. On the contrary, not a few Alpine lakes, and many which occur in similar positions in the mountains of other lands, are true barrier-basins, dammed up wholly by morainic or by fluvio-glacial detritus, or by both. Again, numerous small lakes and pools occur in the cup-shaped and irregular depressions of the paysage morainique at the base of a mountain region. The moraines of this region mark the limits reached by the larger valley-glaciers. One of the most typical localities for the development of small morainic lakes of the kind referred to is the dreary district of the Dombes, in the valley of the Rhone. There, however, many of the pools are of artificial origin, and used as fishponds by the inhabitants. But it is the morainic character of the ground that makes this possible. Thus the paths of the old valley-glaciers are frequently marked by the appearance of glacial lakes, large and small, and variously formed. Great valleybasins may be restricted to the mountains, or may extend for some distance into the Vorländer, or may occur wholly outside of the mountains. Most of these are rock-basins, but their depth has often been increased by accumulations of superficial materials. Other valley-basins are essentially barrierlakes. Lastly, beyond the lowest valley-basins, generally well out upon the low grounds, we encounter the numerous pools and lakelets of the paysage morainique. 3. Plateau and Lowland Basins. The glacial basins we have hitherto been considering are products of the action of individual glaciers, small or great as the case may have been. They occur, therefore, either within mountain-valleys, or in their proximity. But over the wide tracts formerly invaded by the "inland ice" of Northern Europe glacial lakes are not confined to mountain-valleys and the adjacent Vorländer, but are scattered broadcast over plateaux and lowlands. In those regions two areas of special lakedevelopment may be recognised: (1) An area in which glacial erosion has been in excess of glacial accumulation; and (2) an area in which, conversely, accumulation has been in excess of erosion. In the former tracts roches moutonnées abound; the surface is thus often rapidly undulating. Low-lying, roundbacked rocks extend on every side, while here and there the general monotony of the landscape is partly relieved by bare hills and now and again by bald mountain-heights, all scraped, bared, worn, and abraded by severe glacial action. In the countless dimples and irregular hollows of the surface lakes of all shapes and dimensions make their appearance, and the presence of innumerable bogs and marshes show further how many shallow sheets of water have been gradually obliterated. The most notable region of the kind in Europe is Finland, a land of lakes. But excellent examples occur in our own islands, such as the Outer Hebrides and the low-lying, rocky coast-lands of the tract lying between Loch Ewe and Loch Laxford. In North America the particular lake-lands of which we now speak are practically confined to and nearly co-extensive with the Dominion of Canada. Of the basins developed in those regions some have been excavated, while others are barrier-basins. in pre The distribution of the former cannot always be satisfactorily explained. We may suppose that under a general ice-sheet some rocks would yield more readily than others. Some geologists are of opinion that certain rock-basins may be of preglacial origin, and that all the ice did was to plough out the alluvia with which such basins had been filled. The hollows themselves, they think, may have been caused by the weathering and rotting of rock, and the subsequent removal of the disintegrated materials by wind or other superficial agency. According to others the depressions may be tectonic basins filled up glacial times and only re-excavated by glacial action. Some of the larger lakes, such as Lakes Lodoga, Onega, and others in Northern Europe, and the Great Lakes of North America, almost certainly occupy tectonic basins, modified no doubt by considerable glacial erosion and accumulation. But the far more numerous small rock-basins of the regions now under review are unquestionably hollows of erosion. Some appear to have been ground out in places where the rocks offered less resistance to erosion, but probably the position of a larger number has been determined by the form of the ground. This is seen in the frequent appearance of rock-basins in places where the glacial current suffered constriction or obstruction. Thus in broken, hilly ground the thickness of ice and the rate of flow would vary from place to place, and unequal erosion of its bed would follow as a natural course. Not infrequently prominent obstructions rose in its path, and in front of these deflectionbasins were eroded, which usually extend in a direction at right angles to the trend of the ice-flow. If the obstruction were an isolated hill or mountain the hollow often assumed a horse-shoe shape, encircling the base of the hill. Much morainic débris was usually accumulated in the rear by such an obstruction, so as to form a long, sloping "tail." Again, valleys which have chanced to coincide in direction with the ice-flow not infrequently show a succession of two or more constriction-basins. In flat lands of tolerably even surface, however, deflection- and constriction-basins are wanting, the great majority of the lakes being drawn out in the direction of ice-flow. Although, owing to the presence of glacial and other superficial accumulations, we cannot always be sure whether such lakes rest wholly in rock-basins or not, there can be no doubt that they owe their origin to glacial action, partly to erosion and partly to accumulation. In the low grounds of Lewis (Outer Hebrides) the multitudinous lakes almost invariably tend to assume a linear direction, and by far the larger number are arranged along one or other of two lines, which strike as nearly as may be N.W. and S.E., and N.E. and S.W. respectively. Not infrequently one and the same lake shows both lines of direction, one portion of the water trending at right angles to the other. Nearly all the longest and most considerable lakes range from S. E. to N.W. This is the direction of glaciation, and the lakes having this particular |