they could steal round the ridge and so escape outwards to the Atlantic. This being the case, we have a very simple and obvious explanation of the absence of Skye rocks in the till of the Long Island. One sees readily enough that the sub-glacial débris dragged across the Minch would naturally be carried away to south-west and north-east by the "under-tow" or deflected ice. It is quite impossible that any Skye fragments or bits of rock from the mainland could travel over the bed of the Minch, and then be pushed up the precipitous rock wall of the Long Island. There is only one place in all the Outer Hebrides where we might expect to meet with extraneous boulders in the till, and that is in the north of Lewis, where the land shelves gently into the sea, and the great rocky ridge terminates. Here the under-strata of the ice would begin to steal up upon the land, favoured by its gentle inclination, and in that very place accordingly we meet with a deposit of till in which are found many boulders of a hard red sandstone, and some of various porphyries which are quite alien to the Long Island. Moreover, the till itself in that locality is much more of a clay than the usual sub-glacial débris in other parts of Lewis, and contains numerous fragments of sea-shells. All this is quite in keeping with the other evidence. The extreme north end of Lewis was overflowed by the under-current that crept up the bed of the Minch, hugging the Hebridean ridge, and dragging along with it a muddy mass interspersed with the shells and other marine exuviæ that lay in its path, and numerous stones, some of which may have come from Skye, while others were derived from the mainland. I have already said enough, perhaps, about the abrasion of the Hebrides, but I may add a few words upon the origin of the freshwater lakes. Many of these rest in complete rock-basins; others, again, seem to lie partly upon solid rock and partly upon till; while yet others appear to occupy mere shallow depressions in the surface of the till. All of them thus owe their origin to the action of the ice-sheet. As one might have expected, the great majority lie along the outcrop of the gneissic strata, which, as a rule, corresponds pretty closely to the flow of the ice. Hence the general trend of the lakes is from south-east to north-west. In many cases in fashioning these rock-basins the ice has merely deepened in an irregular manner previously existing hollows, which are now, of course, filled with water. In not a few places, however, the lakes are drawn out in other directions--this being due usually to changes in the strike or outcrop of the strata. For example, over a considerable district in the south of Lewis many lake-hollows extend from south-west to north-east, or at right angles to the direction of the ice-flow. Such lakes are usually dammed up at one or both extremities by glacial débris.