362. Longitudinal valleys, with the water bursting out transversely from their sides, like the preceding, are by no means confined to mountains of the first order. We have a very good example, though on a small scale, of a valley of this sort, within a few miles of Edinburgh. The Pentland Hills form a double ridge, separated by a small longitudinal valley, that runs from N. E. to S. W., the water of which issues from an opening almost in the middle, and directed towards the south. This, therefore, is not the work of any great torrent, which overwhelmed the country; for no one direction, which it is possible to assign to such a torrent, will afford an explanation, both of the valley and its outlet.[183] [Footnote 183: In Scotland there is one valley, of a kind that I believe is extremely rare in any part of the world, in accounting for which, the hypothesis of a torrent or _debacle_ might, if any where, be employed to advantage. This is the valley which extends across the island, from Inverness to Fort William, or from sea to sea, being open at both ends, and very little elevated in the middle. It is nearly straight, and of a very uniform breadth, except that towards each end it widens considerably. The bottom, reckoning transversely, is flat, without any gradual slope from the sides towards the middle. From the sides the mountains rise immediately, and form two continued ridges of great height, like ramparts or embankments on each side of a large fossé. A great part of the bottom of this singular valley is occupied by lakes, namely, Loch Ness, Loch Oich, and Loch Lochy. Its length is about sixty-two miles, and the point of partition from which the waters run different ways, viz. north-east to the German Ocean, and south-west to the Atlantic, is between Loch Oich and Loch Lochy; and, by the estimation of the eye, I should hardly think that it is elevated more than ten or fifteen feet above the surface of either lake. The country on both sides is rugged and mountainous, and the streams which descend from thence into the valley, either fall directly into the lakes, or turn off almost at right angles when they enter the valley. Though the bottom of this valley, therefore, is every where alluvial, with the exception, perhaps, of a few rocks which appear at the surface, it is certainly not excavated by the rivers which now flow in it. The direction of the valley, it is to be observed, is the same with that of the vertical strata which compose the mountain on either side.