North and South Harris yield unequivocal evidence of having been overflowed by ice which did not stream out of the mountain-valleys, but crossed the island from the Minch to the Atlantic. A number of mountain-glens, coming down from the Forest of Harris, open out upon West Loch Tarbert, and these we see have been crossed at right angles by the ice--the mountains between them being strongly abraded from south-east to north-west. It is the same all over South Harris, which affords the geologist every evidence of having been literally smothered in ice, which has moved in the same persistent direction. The rock-faces that look towards the Minch are all excessively naked; they have been terribly ground down and scraped, and the same holds good with every part of the island exposed to the south-east. Now, the mode in which the rocks have been so ground, scraped, rounded, and smoothed betokens very clearly the action of land-ice, and not of floating-ice or icebergs. The abrading agent has accommodated itself to all the sinuosities of the ground, sliding into hollows and creeping out of them, moulding itself over projecting rocks, so as eventually to grind away all their asperities, and convert rugged tors and peaks into round-backed, dome-shaped masses. It has carried away the sharp edges of escarpments and ridges, and has deepened the intervening hollows in a somewhat irregular way, so that now these catch the drainage of the land and form lakes. Steep rocks facing the Minch have been bevelled off and rounded atop, while in their rear the ice-plough, not being able to act with effect, has not succeeded in removing the primeval ruggedness of the weathered strata. I have said that the movement of the ice was from south-east to north-west. But a close examination of the ice-markings will show that the flow was very frequently influenced by the form of the ground. Minor features it was able to disregard, but some prominent projecting rock-masses succeeded in deflecting the ice that flowed against them. For example, if we study the rocks in North Harris, we shall find that the Langa and the Clisham have served as a wedge to divide the ice, part of which flowed away into Lewis, while the other current or stream crept out to sea by West Loch Tarbert. The Langa and the Clisham, indeed, raised their heads above the glacier mass--they were islets in a sea of ice. It is for this reason that they and the Tarcull ridge in South Harris have not been smoothed and abraded, but still preserve their weathered outline. All surfaces below a height of 1600 feet which are exposed to the south-east, and which have not been in recent times broken up by the action of rain and frost, exhibit strongly-marked glaciation. But above that level no signs of ancient ice-work can be recognised.