has been followed by more limited and local ice-action. Even on the same rock-surface two intersecting sets of striae may be seen, the older set quite independent of the present outline of the country, the newer clearly determined by the existing mountain-valleys and glens. It is quite possible that, after the land emerged from the sea, ice may have again formed in the mountain-corries and flowed in vast rivers down the glens. There is, indeed, evidence that this has occurred, in at least the higher portions of the district. But we must ever be careful to avoid confusing these mere local and partial glaciers, very limited both in size and influence, with the far older and more universal ice-action of the first period. The earlier glaciers can have had no connexion with the origin of the lines ; the later may.
When examining the lines on various occasions, I have ever been on the search for some fact or facts that might serve as a criterion of the truth or falsehood of the rival theories. At one time I thought the character of the shingle composing them, and the mode of its deposition, might serve this purpose. But I could find no character of this kind at all satisfactory. The shingle on the shores of our Highland fresh- water lakes and that on those of the inland salt-water lochs, are too similar in most respects to admit of accurate discrimination. At length one character did occur to me of a testing and discriminating kind, one point in which the two views were essentially different. Mr. Milne long ago pointed out the remarkable fact that the three best-marked lines corresponded nearly with cols or gorges between the hills, and showed that these gorges must have formed the outlets for the fresh-water lakes to which he ascribed the cutting out of the lines. Now of this there can be no doubt. If from any cause the water of the Spean was prevented from flowing off to the west till it rose to the height of the first or lower line, it would form a lake flowing off to the Spey by the pass of Maccoul; so also a barrier raising the waters of the Roy to the second line would cause them to overflow by Glen Glaister ; and were this exit also shut, when they rose to the third or highest line they would then escape at the very top of Glen Roy into the Spey. How far the levels of the respective lines correspond exactly with these cols must be left to the Ordnance Survey to decide. The levels are either very near, or perhaps the lines in some cases a little higher; so that the supposed lakes would necessarily have overflowed at the points mentioned.
It is thus certain that, if the lines were formed by fresh-water lakes, each of these passes must have been the exit of a river of very considerable size and flowing in a narrow valley for a long period. If, on the other hand, they are of marine origin, then these same passes were sea-straits — narrow channels connecting one great bay with another. Here then was a marked difference in the two theories, a matter of fact which existing phenomena might enable us to decide. For this purpose I examined the various passes carefully, and found that whilst in none of them was there the slightest trace of an ancient river, in all there were distinct indications of the
