moutonnées, and boulders lifted above the parent-rocks indicate a northern direction for the great ice-stream from Loch Treig to the Spean, and then an eastern course on one hand up Loch Laggan, and a western, on the other, down the Spean. Up Glen Roy the ice had apparently passed north-eastwardly, over the watershed towards the Spey. In Knapdale, Argyllshire, similar evidence is obtained of a great ice-stream passing over hill and dale; here falling into the Sound of Jura. The author referred to Rink's and Sutherland's observations on the continental ice of Greenland as affording a probable solution of these phenomena; and, objecting to the hypothesis either of floating ice and of debacles being sufficient to account for the conditions observed, he thought that land-ice, moving from central plateaux downwards and outwards, has effected the extensive erosions referred to, both in Scotland and other northern regions, at a time when the land was at a much higher level than at present. This must have been followed by a deep submergence, to account for the stratified and shell-bearing drift-beds.
March 5.—"On the Glacial Origin of certain Lakes in Switzerland, Scotland, Sweden, and North America." By A. C. Ramsay, F.R.S., President of the Geological Society. The author first stated that in this memoir he proposed to extend his theory of the glacier-origin of the smaller mountain-lakes of Wales and Switzerland (published in 'The Old Glaciers of North Wales') to those greater lakes of Switzerland, which, like the tarns above alluded to, lie in true rock-basins. He then explained a map, compiled from those of Charpentier, Morlot, and Mortillet, showing the ancient extension of the great Alpine glaciers across the Lowlands of Switzerland to the Jura, also over the area that surrounds the Lake of Constance, and on the South into the plains of Piedmont and Lombardy. All the great lakes of Switzerland, and the lakes of Como, Lugano, and Maggiore, lie directly in the course of one or other of these great glaciers; and, as shown by the soundings, and the levels of the rocks at their mouths or in the river-beds below, each of these lakes, like the smaller tarns of the Todten Sea and the lake at the Grimsel, was shown to lie in a true rock-basin. He then considered the question of the denudation of the Alpine and Miocene areas of Switzerland, and showed that none of the lakes lie in aboriginal undenuded synclinal hollows. Next, that they do not lie in areas of mere watery erosion. Neither running water nor the still water of lakes can scoop large hollow basins like those of the lakes, bounded on all sides by rocks. Running water may fill them up, but cannot excavate them. He next contended that they do not lie in lines of gaping fracture. A glance shows this with respect to such lakes as those of Geneva, Neuchâtel, and Constance; and, reasoning on the nature of the contortion of the strata of the Alps, he contended that, though fractures of the rocks must be common, they need not be gaping fractures. To produce such a mountain-chain, the strata are not upheaved and stretched so as to produce open cracks; on the contrary they are compressed laterally and crumpled up into smaller space, and the uppermost strata, that pressed heavily on the crumpled rocks now visible, would prevent the formation of wide open fractures below, these upper strata, as in North Wales, having, over a great part of the area, been mostly or altogether removed by denudation. Next, lakes of the rock-basin kind do not lie each in an area of special subsidence. If so, for instance, we should require one for the Todten See, one for the Grimsel, one for the ancient lake of the Kirchet, several at the foot of the Siedelhorn. many hundreds close together in Sutherlandshire, and thousands in North America.
If then the lake-basins were formed by none of the above-named causes,
