l dug
along the shore, or by means of tight walls, according to the lay of the
place. The second process, which is preferable to the first, is in fact
impossible when the river runs, as is often the case, in a narrow, abrupt,
walled channel. These works are sometimes very important. In 1881, the
Acaba Mundo flume was 140 meters in length and 5.2 m. wide, and, with a
velocity of 2.25 m., discharged 4,500 liters per second; still longer ones
might be cited that discharged as much as 8,000 liters.

In the dry part of the river the extraction of the sand, stones, and
cascalho is done solely by hand. The men carry the sand upon their heads
in small wooden bowls called _carumbes,_ which hold about 15 kilogrammes,
and throw it somewhere where the deposit will not interfere with the
exploitation. Almost all of these men are negroes, who run with their load
upon their head over the white sand, singing some song of their country.
It, is very picturesque, but it is doubtful whether it is economical.

Since the century and a half that these rivers have been dug and redug, it
may be admitted that wherever the cascalho has been easy of access it has
been removed; and that wherever it has not been, little attempt has been
made to work it. How have these attempts, which have doubtless been made
at several periods, come out? This would at present be very difficult to
ascertain. The exploitations have been too numerous to allow us now to
estimate the value of a bed from the data furnished by geology, and local
tradition is too uncertain or exaggerated to allow us to place much
confidence in it.

We can, at the very most, say that if some points still remain intact it
must be because the exploitation of them was too difficult with the
processes that were employed, and this should be a reason, were it desired
to attempt new operations, for having recourse to entirely different modes
of work.

It would seem rational, as regards this, to try to put to profit the
hydraulic power that the flumes and canals render disposable for
mechanically extracting the sand. The field to be worked being naturally
long and narrow, it would be the proper thing to employ a series of
inclined planes distributed along the banks, actuated by water wheels, and
corresponding to so many small working points. The river often flows
through a genuine canon with nearly vertical walls, where space would be
absolutely wanting for installing wheels elsewhere than at the