friction is
carried down by the streams which flow from the melting glacier, so that
the water which in the morning may be pure, owing what little strength
it has chiefly to the rock springs, is in the afternoon not only
increased in volume, but whitened with dissolved dust of granite, in
proportion to the heat of the preceding hours of the day, and to the
power and size of the glacier which feeds it.

Sec. 2. The long drought which took place in the autumn of the year 1854,
sealing every source of waters except these perpetual ones, left the
torrent of which I am speaking, and such others, in a state peculiarly
favorable to observance of their _least_ action on the mountains from
which they descend. They were entirely limited to their own ice
fountains, and the quantity of powdered rock which they brought down
was, of course, at its minimum, being nearly unmingled with any earth
derived from the dissolution of softer soil, or vegetable mould, by
rains.

At three in the afternoon, on a warm day in September, when the torrent
had reached its average maximum strength for the day, I filled an
ordinary Bordeaux wine-flask with the water where it was least turbid.
From this quart of water I obtained twenty-four grains of sand and
sediment, more or less fine. I cannot estimate the quantity of water in
the stream; but the runlet of it at which I filled the flask was giving
about two hundred bottles a minute, or rather more, carrying down
therefore about three quarters of a pound of powdered granite every
minute. This would be forty-five pounds an hour; but allowing for the
inferior power of the stream in the cooler periods of the day, and
taking into consideration, on the other side, its increased power in
rain, we may, I think, estimate its average hour's work at twenty-eight
or thirty pounds, or a hundred weight every four hours. By this
insignificant runlet, therefore, some four inches wide and four inches
deep, rather more than two tons of the substance of the Mont Blanc are
displaced, and carried down a certain distance every week; and as it is
only for three or four months that the flow of the stream is checked by
frost, we may certainly allow eighty tons for the mass which it annually
moves.

Sec. 3. It is not worth while to enter into any calculation of the relation
borne by this runlet to the great torrents which descend from the chain
of Mont Blanc into the valley of Chamouni. To call it the thousandth
part of t