AND CIRQUES IN NORWAY AND GREENLAND. 173
quartzose slates ; further down there is gneiss and gabbro ; then at the mouth a conglomerate. The fjords in Greenland cut through both the basalt and the Azoic rocks. The large Waigat Fjord, which is a continuation of the present ice-fjord of Torsukatak, is surrounded by rocks of Cretaceous and Miocene age, which are capped by an enormous sheet of basalt ; and the structure of the two sides of the fjord is symmetrical. Thus the erosion of the fjord must be posterior to the Miocene period ; for, at Atanekerdluk, on the very side of the fjord, occur the well-known Miocene fossils ; and as the great ejections of basalt are still later than these, but older than the fjords, the latter must be of very recent date.
Fjords are always numerous along the same coast-line. This might be deduced from the theory of their glacial formation. A great ice-sheet covering the interior of Norway or Greenland could not discharge its ice by one or a few glaciers only, but would require many of these, and along the whole coast-line. Hence the fjords are numerous along the coast, and so also are the lakes. The glaciers which formed them must have been thick enough to reach the bottoms of the lakes and, unless the land lay higher than now, the bottoms of the fjords as well ; that this was the case has already been shown. If, then, the glaciers during the Glacial epoch were able to form fjords and lakes, those which now exist must also be able to scoop out the ground beneath them ; and if they have been at work long enough, the depressions formed by them must be found. We have already pointed out that cirques occur near modern glaciers, being recesses formed by them. As fjords and lakes are only associated with the old glaciers, so the cirques are confined to small isolated modern glaciers. When we regard the great effects produced by these small glaciers, the enormous erosion of these ancient and thick glaciers becomes less surprising. The Norway glaciers proceeded from an extensive inland ice-sheet ; if, then, the valleys and the fjords result from the erosion of ice or of water, it follows that they must start from the highest part of the country, and on the whole increase in breadth and depth in proportion to the increase of the glaciers ; that is, we must be able to trace up a fjord through branch-fjords, fjord -valleys, and branch-valleys to remote glens in the mountains. The Norway fjords can be shown by numerous examples to satisfy this requirement. The Sogne Fjord is an excellent instance ; it branches off, as may be easily seen, into six large tributary fjords, every one of which is continued by a fjord- valley ; this is formed by other valleys into which little tributary glens debouch. Further, if these fjords and valleys are formed by the erosion of ice or water, their breadth and depth must be in proportion to the districts which once fed their glaciers and now feed their rivers. On comparing the limits of different fjords and valleys along the watershed, it is, on the whole, remarkable how their dimensions show a dependence on their districts of rain- fall. As these increase, so do the fjord-valleys. I do not mean, indeed, that the area of a transverse section through a valley can be connected by an empirical formula with the area of the district ; for
