stern extension of this valley, damming back the waters which
filled Glen Roy and the central part of Glen Spean.

Evidently the glacier descending from Loch Treig was the first to yield,
for, by the time the Glen Roy lake had sunk to the level of the lowest
terrace, the entrance to the eastern extension of the valley must have
been free, otherwise the water could not have spread throughout that
basin as we find it did; but it would seem that by the time the western
barrier, or the glacier from Ben Nevis, was removed, the sheet of water
was too far reduced to have left permanent marks of its outflow into the
Great Glen, except by disturbing and remodelling the large moraines of
the older Glen Spean glacier. There are faint indications of other
terraces in Glen Roy, even at a higher level than the uppermost parallel
road, owing their origin probably to the short duration of a higher
level of the glacier-lake, when the great general glacier had not yet
been lowered to a more permanent level determined by a limited
circumscription within the walls of the valleys. There are other
terraces in neighboring valleys at still different levels,--in Glen
Gloy, for instance, where the one horizontal road was no doubt formed in
consequence of the damming of the valley by a glacier from Loch Arkeig.
Mr. Darwin has seen another in Glen Kinfillen, which I would explain by
the presence of a glacier in the Great Glen, the marks of which are
particularly distinct about the eastern end of Glen Garry.

The evidence of the ancient presence of glaciers is no less striking in
other parts of the Scotch Highlands. Between the southeastern range of
the Grampian Hills, in Forfarshire and Perthshire, and the opposite
ridge of Sidlaw Hills, stretches the broad valley of Strathmore. At the
time when Glen Spean received the masses of ice from the slopes of the
western Grampian range, the glaciers descended from the valleys on the
southern slope of the southeastern range and from those on the northern
slope of Sidlaw Hills into the capacious bed of the valley which divides
them. The glacial phenomena of this region present a striking
resemblance in their general relations to those of the Alps and the
Jura. The Grampian range on the northern side of Strathmore valley
occupies the same position in reference to that of the Sidlaw Hills
opposite, as does the range of the Alps to that of the Jura, while the
intervening valley may be compared to the plain of