flow from the surface almost entirely to cease, and allow the whole to be carried off by the springs. Whenever this was the case, any great and rapid excavation of the valleys from rain alone would be rendered almost impossible; and with no extreme reduction in the total amount of annual flow of the rivers, yet by their originating in perennial springs subject to but slight variations, and from their being no longer to any extent immediately connected with the surface drainage, there would cease to be that immense difference between their maximum and minimum volume, which must have formerly existed. The result of this comparatively uniform flow would be a great diminution in the tendency of any river to change its bed, and even if it occasionally received a great accession of water, it would find relief by overflowing into the wide valley due to its former more violent action. In the less inclined portions of its valley, the parts now almost deserted by the stream would be favourable for vegetation, such as would result in the formation of peat, and any occasional overflowing of the banks might, owing to the less torrential character of the inundations, have a tendency to fill up and level these marginal spaces rather than to excavate them deeper. The deposits of gravel, sand, and clay at the low levels would also be more continuous than those at the higher.
In tracing the effects of subaërial action in forming valleys, I have assumed the subsoil or rock in which they were formed to have been chalk, as it is principally in valleys in the Chalk that the gravels containing Palæolithic implements are known to occur. This is probably on account of the greater natural abundance of flints in such valleys, which of course led to implements being there chipped out in greater numbers, as well as to their being less cared for, from their being more easily replaced than they would be where flint was scarce. The effects on other soft and absorbent soils would not materially differ from those on chalk. On clay, the general amount of denudation would perhaps be greater, but the valleys broader, and with less inclined slopes on their sides. In a clay country we might, I think, expect to find the old river-gravels not unfrequently at greater distances from the existing streams than in a chalk-district.
It must, however, be borne in mind that in such a country the materials from which river-gravels can be formed are usually absent, and can only have been derived from older superficial beds, or brought from Chalk higher up the valley. In some
