the continental glacier buried this land beneath ice, which moved with especial freedom through the north-south valleys, scouring them and leaving them both broader and deeper. The amount of deepening by this ice invasion was at least five hundred feet, and probably much more, the exact amount being impossible of determination at present, since much of the evidence was erased by later ice erosion. Moreover, the first ice advance may have left lakes in the valleys, whose surfaces acted as temporary base-levels, below which the interglacial gorges could not be cut.
With the recession of this ice sheet, the upland tributaries were left hanging five hundred feet or more above the overdeepened main-valley bottom, and the streams descending the steepened main-valley slope began to cut gorges in it. This condition lasted through interglacial times and resulted in the production of fairly broad and deep gorges. Then came a readvance of the ice, which again broadened and deepened the valleys and on its recession left the interglacial gorges, partly buried and erased, hanging high above the newly made bottom of the main valley. Since the retreat of the Wisconsin ice sheet the streams have been engaged again in gorge cutting on the steepened slope, in some places along the lines of the older gorges, but more commonly partly or completely independent of them.
There are some facts which indicate possible greater complexity of ice erosion, for in some of the valleys there is apparently more than one buried gorge; but the evidence on this point is not as yet convincing, and for the present we can point with certainty to no greater complexity than that of two periods, one the Wisconsin, the other of some one of the earlier ice advances with which the work of the glacial geologists of the Mississippi valley have made us familiar.
The various glens of the Cayuga and Seneca Lake valleys, whose general cause is as above stated, differ greatly in detail. They are all wild and picturesque, and they are all narrow gorges with many cascades and waterfalls. Their variations depend upon the varying combinations of effects from several causes. One of these is the influence of the buried gorges. Wherever the postglacial stream enters one of these its valley abruptly broadens. Where the postglacial course coincides with the buried gorge the expansion is continuous; but where it merely crosses the older gorge, the narrow rock-walled and rock-bottomed postglacial gorge is replaced by an expansion, forming an 'amphitheater' with drift walls and bottom. The valley again contracts where the stream leaves the buried gorge and has cut a postglacial glen in the rock of the other wall of the older gorge.
A second cause for differences in the gorges is the influence of the variation in resistance to erosion of the nearly horizontal strata of Devonian shales and sandstones in which the gorges arc cut. The
