The same eminent author proposes series of observations on the level of natural lakes, especially on those without outlet, as a means of determining the increase or diminution of precipitation in their basins, and, of course, of measuring the effect of clearing when such operations take place within those basins. But it must be observed that lakes without a visible outlet are of very rare occurrence, and besides, where no superficial conduit for the discharge of lacustrine waters exists, we can seldom or never be sure that nature has not provided subterranean channels for their escape. Indeed, when we consider that most earths, and even some rocks under great hydrostatic pressure, are freely permeable by water, and that fissures are frequent in almost all rocky strata, it is evident that we cannot know in what proportion the depression of the level of a lake is to be ascribed to infiltration, to percolation, or to evaporation.[189] Further, we are, in general, as little able to affirm that a given lake derives all its water from the fall of rain within its geographical basin, or that it receives all the water that falls in that basin except what evaporates from the ground, as we are to show that all its superfluous water is carried off by visible channels and by evaporation. Suppose the strata of the mountains on two sides of a lake, east and west, to be tilted in the same direction, and that those of the hill on the east side incline toward the lake, those of that on the west side from it. In this case a large proportion of the rain which falls on the eastern slope of the eastern hill may find its way between the strata to the lake, and an equally large proportion of the precipitation upon the eastern slope of the western ridge may escape out of the basin by similar channels. In such case the clearing of the _outer_ slopes of either or both mountains, while the forests of the _inner_ declivities remained intact, might affect the quantity of water received by the lake, and it would always be impossible to know to what territorial extent influences thus affecting the level of a lake might reach. Boussingault admits that extensive clearing _below_ an alpine lake, even at a considerable distance, might affect the level of its waters. How it would produce this influence he does not inform us, but, as he says nothing of the natural subterranean drainage of surface waters, it is to be presumed that he refers to the supposed diminution of the quantity of rain from the removal of the forest, which might manifest itself at a point more elevated than the cause which occasioned it. The elevation or depression of the level of natural lakes, then, cannot be relied upon as a proof, still less as a measure of an increase or diminution in the fall of rain within their geographical basins, resulting from the felling of the woods which covered them; though such phenomena afford very strong presumptive evidence that the supply of water is somehow