[Footnote 163: Histoire du Danube, tom. i. tab. 34.] 317. The courses of many rivers retain marks that they once consisted of a series of lakes, which have been converted into dry ground, by the twofold operation of filling up the bottoms, and deepening the outlets. This happens, especially, when successive terraces of gravelly and flat land are found on the banks of a river, § 100. Such platforms, or _haughs_ as they are called in this country, are always proofs of the waste and _detritus_ produced by the river, and of the different levels on which it has run; but they sometimes lead us farther, and make it certain, that the great mass of gravel which forms the successive terraces on each side of the river, was deposited in the basin of a lake. If, from the level of the highest terrace, down to the present bed of the river, all is alluvial, and formed of sand and gravel, it is then evident, that the space as low as the river now runs must have been once occupied by water; at the same time, it is dear, that water must have stood, or flowed as high at least, as the uppermost surface of the meadow. It is impossible to reconcile these two facts, which are both undeniable, but by supposing a lake, or body of stagnant water, to have here occupied a great hollow, (which by us must be held as one of the original inequalities of the globe, because we can trace it no farther back,) and that this hollow, in the course of ages, has been filled up by the gravel and alluvial earth brought down by the river, which is now cutting its channel through materials of its own depositing. There is no great river that does not afford instances of this, both in the hilly part of its course, and where it descends first from thence into the plain. Were there room here for the minuter details of topographical description, this might be illustrated by innumerable examples. 318. It is said above, that the water must have run or stood, in former times, as low as the present bottom of the river; but there is often clear evidence, that it has run or stood much lower, because the alluvial land reaches far below the present level of the river. This is known to hold in very many instances, where it has happened that pits have been sunk to considerable depths on the banks of large rivers. By that means, the depth of the alluvial ground, under the present bed of the river, has been discovered to be great; and from this arises the difficulty, so generally experienced, of finding good foundations for bridges that are built over rivers in large valleys, or open plains, the ground being composed of travelled materials to an unknown depth, without any thing like the native or solid strata. In such cases, it is evident, that formerly the water must have been much lower, as well as much higher, than its present level, and this is only consistent with the notion, that the place was once occupied by a deep lake.