trend rest sometimes in true rock-basins, sometimes in hollows between parallel banks formed wholly of glacial deposits, or partly of these and solid rock. The north-east and south-west lakes, on the other hand, are drawn out more or less at right angles to the path of the old ice-flow. They follow precisely the line of "strike" (or general direction of the outcropping ledges or reefs of gneiss); when this direction changes there is a corresponding change in the trend of the lakes. Thus in places where the strike is east and west we have east and west lakes, which wheel round to south-west as soon as the strike shifts to that direction. In preglacial times the low-lying tracts of Lewis were in many places traversed by a series of rough ridges and interrupted escarpments, with intervening hollows corresponding to outcrops of the harder and the less resisting beds of gneiss. The dip of the rocks being generally south-east, the escarpments naturally faced the north-west. inland ice, which subsequently overflowed this region from south-east to north-west, then advanced against the dip-slopes of the gneissose rocks, which were ground bare, while bottom-moraine was here and there deposited in front of the cliffs, knolls, and rocky ledges and ridges formed by the outcrops of the harder beds. Hence, when the ice finally disappeared, the hollows lying between parallel rock-ridges and escarpments were unequally coated with bottom-moraine, and an abundant series of longer and shorter troughs were thus prepared for the reception of water. The

The

In many

north-east and south-west lakes are consequently barrier-lakes, dammed up wholly by boulder-clay or with rock and boulder-clay together. The manner in which the two groups of lakes now described frequently unite offers no difficulty. In many places the old strike-ridges have been cut across by the ice at right angles, and a new system of ridges and hollows has resulted. And it is not surprising, therefore, to find that not only lakes but also streams exhibit both directions, now trending north-west and south-east, and then turning sharply off at right angles to the course previously followed.

When we leave the highly abraded and ice-worn regions of roches moutonnées-the lands of multitudinous lakes and lakelets-we eventually enter upon tracts over which glacial accumulation has been in excess of erosion. Here lakes become much less numerous, and are met with only at intervals. Most of them extend over shallow depressions in the surface of the old ground-moraines, but a few occupy rockhollows ground out in front of prominent obstructions. Lakes of the former kind were formerly much more plentiful, but owing to their limited depths many have been silted up, and are now replaced by alluvial flats. The deeper deflection-basins, on the other hand, have been more persistent as lakes, but they are comparatively few in number. Passing still farther outwards, and leaving behind the gently undulating and rolling plains, throughout which groundmoraine forms the dominant deposit at the surface,

we reach at last the paysage morainique, with its tumultuous hills, knolls, ridges, and embankments, and find ourselves once more in a region of lakes, or rather of lakelets, pools, and marshes. Among the most conspicuous examples of such a region is the paysage morainique of the last great Baltic glacier, extending from west to east through East Holstein, Mecklenburg-Strelitz, Uckermark, Neumark, Southern Pomerania, and the higher parts of West and East Prussia. Another well known region of similar character is the corresponding lake-dappled paysage morainique of North America, which embraces such vast tracts in the Northern States of the Union.

4. Ice-Barrier Basins. In existing glacier-regions ice-dammed lakes now and again appear. Of these the Märjelen See on the Aletsch glacier may be taken as an example. Their origin is simple enough. When a glacier advances across the mouth of a tributary valley, the stream flowing in the latter is dammed back, and a lake comes into existence. In the Alps lakes of this kind have formed from time to time, the sudden bursting of the ice-dams occasionally causing enormous devastation. In our own and other formerly glaciated countries the relics of such lakes -some of which must have persisted for long periods -are of not infrequent occurrence. The well known "Parallel-Roads" of Glen Roy are simply the beaches of an ice-barrier lake.

5. Submarine Basins. Here we are not concerned with the large and small basins that mark the

floor of the great oceanic troughs, all of which are doubtless tectonic. The hollows to which we would now refer are certain relatively smaller basins occurring in immediate proximity to the shores of recently depressed lands. The regions in which they appear, although submerged, form, nevertheless, a continuation of the continental plateau. The true border of the European continent, for example, extends in the north-west as far seaward as the 100-fathom line at least, and there is good ground for believing that within geologically recent times a large part, if not the whole, of that now depressed region existed as dry land. The sea-lochs of Scotland and the fiords of Norway simply occupy old mountain-valleys, while the numerous islets lying off those coasts and the British Islands themselves were all at one time connected and joined to the mainland of Europe. The basins to which we now call attention form two more or less well marked groups. One of these is practically confined to the fiords, the other is developed chiefly in front of islands that face the fiords.

Although it is not possible to go into much detail, it is nevertheless necessary to indicate the characteristic features of a typical fiord region. Norway, as we have already learned, is an ancient plateau, deeply incised and cut up, as it were, into irregular segments. These segments vary much in extent and form-sometimes the surface of the fjeld is flat and undulating, elsewhere it is scarped and worn into irregular groups and masses of variously shaped mountains and ridges

without any determinate arrangement. The orography is everywhere in strong contrast to that of the Alps, with their extended parallel chains and longitudinal valleys. Not less strong is the contrast between the fiord-valleys of Norway and the valleys of the Alpine chain. The latter in cross-section are commonly V-shaped, while the former are U-shaped. Again, fiord-valleys have relatively few lateral branches, the opposite being the case with the great valleys of the Alps, which are joined by numerous tributaries. Were the Alpine lands to be so submerged as to convert such valleys as the Rhone or the Inn into arms of the sea, it is obvious that numerous broad and long inlets would ramify right and left from these arms into the mountains. The fiordvalleys of Norway do not branch after that fashion; the hydrographic system of the country, as Professor Richter well observes, is imperfectly developed. The principal channels of erosion are the deep, trenchlike fiord-valleys, the tributaries which reach these from the fjelds or plateaux being relatively insignificant. The main stream, flowing through a deep mountain-valley, has cut its way down to the level of the sea, which it enters at the head of a fiord. Below this point, however, few or no side valleys, as a rule, break the continuity of the fiord-walls. Numerous tributary waters, some of which are hardly less important than the head-stream, do indeed pour into the fiord, but they have not yet eroded for themselves deep trenches. After winding through the plateau