to stir. In this pebble-paved section, because the
stream could not attack the foundation rock, we find no gorges--in
fact, the whole of this upper section of the torrent system is
peculiarly conditioned by the fact that the streams are dealing not
with bed-rock, but with boulders or smaller loose fragments. If they
cut a little channel, the materials from either side slip the faster,
and soon repave the bed. But when the streams have by a junction
gained strength, and can keep their beds clear, they soon carve down a
gorge through which they descend from the upper mountain realm to the
larger valleys, where their conjoined waters take on a riverlike
aspect. It should be noted here that the cutting power of the water
moving in the torrent or in the wave, the capacity it has for abrading
rock, resides altogether in the bits of stone or cutting tools with
which it is armed. Pure water, because of its fluidity, may move over
or against firm-set stones for ages without wearing them; but in
proportion as it moves rocky particles of any size, the larger they
are, the more effective the work, it wears the rock over which it
flows. A capital instance of this may be found where a stream from a
hose is used in washing windows. If the water be pure, there is no
effect upon the glass; but if it be turbid, containing bits of sand,
in a little while the surface will appear cloudy from the multitude of
line scratches which the hard bits impelled by the water have
inflicted upon it. A somewhat similar case occurs where the wind bears
sand against window panes or a bottle which has long lain on the
shore. The glass will soon be deeply carved by the action, assuming
the appearance which we term "ground." This principle is made use of
in the arts. Glass vessels or sheets are prepared for carving by
pasting paper cut into figures on their surfaces. The material is then
exposed to a jet of air or steam-impelling sand grains; in a short
time all the surface which has not been protected by paper has its
polish destroyed and is no longer translucent.

The passage from the torrent to the river, though not in a
geographical way distinct, is indicated to the observant eye by a
simple feature--namely, the appearance of alluvial terraces, those
more or less level heaps of water-borne _debris_ which accumulate
along the banks of rivers, which, indeed, constitute the difference
between those streams and torrents. Where the mountain waters move
swift