The only cascade which we can point out as having really diminished in height, is that of Tungasca in Siberia. We do not, however, assert but that there may be others. So many causes different from those of erosion may concur to lower a cascade, or even make it disappear almost entirely, that we are rather astonished at the small number of examples mentioned, than embarrassed by the objections which these examples might present to the opinion which we are defending: for the fall of a part of the rock which forms the cliff from which the cascade is precipitated; an abundant accumulation of debris at the foot of the cliff; a real destruction of the softer deposits, forming part of the strata of the mountain from which they fall, are sufficient causes for changing the height of waterfalls. These causes must present themselves pretty frequently; but how different is their action from that of erosion? This, if it existed, would extend from the source of the river to its mouth, and would have a considerable influence upon the configuration of the earth’s surface. Those which we have mentioned have, on the contrary, an action so limited and so local, as to be scarcely appreciable. 3. Allowing, for the moment, that a river, possessed of a vast erosive or disaggregating power, may have scooped out the valley in the bottom of which it at present flows, in a state of feebleness very different from its original state, we must account for the disposal of a vast mass of earth and rock, which filled up the valley before the river had removed it. It is not possible to suppose that it has been transported into the sea, which is often more than a hundred leagues from the valley; for we know that when rivers, on reaching the plains, lose their rapidity, they allow the matters to be precipitated which they held in suspension. Besides, we have shown that many rivers, on leaving the mountains, traverse lakes, in which they deposit all the earthy matters suspended in their waters. This deposition is particularly striking in all the considerable rivers, which descend from the ridge of the Alps toward the north-west and south-east of that chain of mountains. These rivers meet, at the opening of the valleys they flow through, lakes, which they traverse, and which seem intended for their purification. Thus, on the northern side, we see the Rhone traversing the lake of Geneva; the Aar, the Lakes of Brientz and Thun; the Reuss, the Lake of the Four Cantons; the Linth, the Lake of Zurich; the Rhine, the Lake of Constance. On the south side, the Lac Majeur is traversed by the Tessin, the Lake of Como by the Adda, the Lake Disco by the Oglio; the Lake of Guarda by the Mincio, &c. Now, these lakes, which are only themselves deeper parts of the valley, would have been filled up by the debris conveyed to the valley, if this valley had the origin attributed to it. Proceeding from one hypothesis to another, it might