ause of change,
since we know no reason why one species of animal should not continue
to be reproduced as long as another, but for the occurrence of
physical changes of a prejudicial character.

We have still remaining the changes which have taken place in the
erosion of valleys since the caverns were occupied. Dupont informs us
that the openings of some of the caverns once flooded by rivers are
now in limestone cliffs two hundred feet above the water, while no
appreciable lowering of the bottoms of the ravines is taking place
now. This would in some contingencies put back the period of filling
of the caves to an indefinite antiquity. But then the questions
occur--Was there once more water in the rivers or more obstruction at
their outlets, or was the erosive power greater at one time than now,
or were the river valleys excavated in still more ancient time, and
partly filled with mud when the water entered the caves, and may this
mud have been since swept away? So, in like manner, the waters flowing
in the channels near Brixham Cave and Kent's Hole were apparently
about seventy feet higher in times of flood than at present, but the
time involved is subject to the same doubts as in the case of the
Belgian caves. Hughes has well remarked that elevations of the land,
by causing rivers to form waterfalls and cascades, which they cut
back, may greatly accelerate the rate of erosion. Farther, there is
the best reason to believe that in the glacial period many old valleys
were filled with clay, and that the modern cutting consisted merely in
the removal of this clay. Belt has shown in a recent paper[139] good
reason to believe that this is the case with the Falls of Niagara, and
that the cutting actually effected through rock within the later
Pleistocene and modern period has been that only of the new gorge from
the whirlpool to Queenstown, the main part of the ravine being of
older date and merely re-excavated. This would greatly reduce the
ordinary estimate of time based on the cutting of the Niagara gorge.

This leads us next to consider the occurrence of human remains and
objects of art in the river-gravels themselves, and the amount of
excavation and deposit involved in the deposition of these gravels.
In the river-gravels of the Somme, and of many other rivers in France
and Southern England, chipped flints and rude flint implements are
found in so great quantity as to imply that the beds and banks of
these streams were