It is now proper to describe the manner in which the waters of a lake act in producing shores of this nature, as being the same in which the assumed waters of Glen Roy must have acted in producing its lines. This action con sists, first, in the powers of the water in checking the fur ther descent of the alluvial matters which are constantly descending from above. As these become immersed in the lake, they lose a very large portion of their weight by immersion, which, in winter, being often rendered still more buoyant from being entangled in ice, they are thrown back against the face of the hills by the incessant action of the waves. Thus they become spread along its sides, and often distributed in a very equable manner, pro ducing an inclined shore, which is proportioned in breadth and declivity to the several circumstances already men tioned, and to the time during which this action has con tinued. This process may he witnessed in innumerable Highland lakes ; and wherever there is a prevailing lee shore in consequence of the course of the winds, it is ren dered particularly conspicuous, from the superior action of the waves on that part. We might point out examples in so many places, that it is unnecessary to enumerate any, as they cannot be missed by those who will turn their at tention to this subject. But we may here add, that Loch Rannoch, which is espesially exposed to westerly winds, displays an excellent example of the kind of lee shore just mentioned, at its eastern extremity.
The connection of the terraces of Glen Roy with an assumed lake, once occupying that valley, cannot now be a subject of much difficulty. The delta which we just noticed, as attending the entrance of a principal stream into a lake, is produced by the same action of the waters as that which forms the shores, with the sole difference of its being exerted on a larger quantity of materials, of which also there is a perpetual supply. The joint action of the river and the waves continues to level these, until, by the growth of plants, and by detaining sand and clay in the time of inundation, they rise above the level of the i water. Now, in Glen Roy, if we assume the lake to have stood at the level of the lowest line, or at its last state, it is easy to see that the present terraces at its upper ex temity which coincide with that line, are the ancient deltas of the Roy, formed at the head of what may be called Loch Roy. The lateral ones, at similar elevations, arc, in the same manner, those which accompanied the entrance of the lateral streams into the lake.
But we must explain the cause also of their present abrupt declivities. Were a lake to be suddenly drained, its deltas would be found producing very gentle slopes, and far prolonged at their bases. But, after this opera tion, the river would then flow on the alluvial bottom, and would thus act on these deltas, so as to undermine and cut down their sides, just as they produce these effects on the alluvial bottoms of valleys, where they shift their position as circumstances change. It is the present course of the Roy, in like manner, since the drainage of the lake, which has worn down the original deltas to their present forms ; the various subsequent changes it has since undergone having in the same way generated the successions of smaller terraces that are found in the course of the valley. Thus have they not only acquired their present forms and dispositions, while in certain cases, where particularly ex posed, they have entirely disappeared.
Thus much for the lowest state of the water in Glen Roy, as it relates to this question. Nor is it difficult to account for the absence of similar terraces at the higher levels, while the explanation itself tends to confirm the present views. The actual condition of the valley, it must be remarked, is essentially different from what it was after the two first subsidences took place which left the lines high impressed above the water. In both these cases, the bottom of the valley continued full of water, or was a lake, whereas after the last subsidence, it was entirely drained, and became dry land. It is evident that, under such circumstances, had any deltas existed at the two up per levels, they must have gradually been worn away by the action of the waters, for want of a sufficient resistance; whereas, on the final drainage of the whole, they were left in their original integrity, and exposed only to the gradual corrosive action of the river, which also, by gra dually lowering and changing its bed, must have shortly deserted them. Nor from the relative state of the lake, the hills, and the entering rivers, could any considerable deltas have been formed at the two upper levels. Still, though no conspicuous terraces are found at these higher levels, there are sufficient indications of them in many places ; while, in all, they lie near the entrances of the torrents or rivers, so as to confirm the explanation which has here been given of their origin. By this we can also explain, as was formerly hinted, those appearances of supernumerary lines which occur in different places among the principal ones. These are the remains of such terraces, or deltas, undermined and demolished during the subsidences of the lake, from causes now sufficiently ob vious.
Now, though we consider this theory as explaining not only the appearances in Glen Roy, but, with the ne cessary modifications, all those in the adjoining valleys, we are aware that there are some difficulties unexplained ; and, as is but just, we shall here point them out. The chief of these are the deficiencies that occur in the lines in some places, when none of the more obvious causes of their absence, formerly noticed, can be traced. A few of them, indeed, as has already been shown, assist in proving the truth of this hypothesis ; and it is not unlikely that, had we access to other phenomena of a similar nature, the whole of them might equally be explained. In Upper Glen Roy, as an instance of a difficulty, there is only the part of one line to be seen, though there is nothing ap parently in the nature or the form of the ground to prevent it from having been impressed everywhere. A consider able deficiency also of all the Imes occurs towards the bot tom of the lower valley, together with many partial ones in other places. The rocky nature of the ground, or the peculiarities of the slope, will not account for the whole of these ; nor is it easy to trace marks of recent waste, by which they could have been obliterated. Similar difficul ties occur in Glen Gloy and in Glen Spean. But these, however inexplicable they may yet be, are by no means sufficient to destroy, scarcely indeed to enfeeble a theory supported by so large and distinct a body of evidence ; and we shall now therefore pass to another, and by far the most difficult part of this subject, to that, namely, which relates to the manner and place in which these lakes were retained during the periods at which they ex isted.