riculture.

The levee is the crowning work of bank revetment, although not
necessarily in immediate connection. It may be set back a short distance
from the revetted bank; but it is, in effect, the requisite parapet.
The flood river and the low river cannot be brought into register, and
compelled to unite in the excavation of a single permanent channel,
without a complete control of all the stages; and even the abnormal rise
must be provided against, because this would endanger the levee, and
once in force behind the works of revetment would tear them also away.

Under the general principle that the local slope of a river is the
result and measure of the resistance of its bed, it is evident that
a narrow and deep stream should have less slope, because it has less
frictional surface in proportion to capacity; i.e., less perimeter in
proportion to area of cross section. The ultimate effect of levees and
revetments confining the floods and bringing all the stages of the river
into register is to deepen the channel and let down the slope. The first
effect of the levees is to raise the surface; but this, by inducing
greater velocity of flow, inevitably causes an enlargement of section,
and if this enlargement is prevented from being made at the expense of
the banks, the bottom must give way and the form of the waterway be so
improved as to admit this flow with less rise. The actual experience
with levees upon the Mississippi River, with no attempt to hold the
banks, has been favorable, and no one can doubt, upon the evidence
furnished in the reports of the commission, that if the earliest levees
had been accompanied by revetment of banks, and made complete, we should
have to-day a river navigable at low water, and an adjacent country safe
from inundation.

Of course it would be illogical to conclude that the constrained river
can ever lower its flood slope so as to make levees unnecessary, but it
is believed that, by this lateral constraint, the river as a conduit may
be so improved in form that even those rare floods which result from the
coincident rising of many tributaries will find vent without destroying
levees of ordinary height. That the actual capacity of a channel through
alluvium depends upon its service during floods has been often shown,
but this capacity does not include anomalous, but recurrent, floods.

It is hardly worth while to consider the projects for relieving
the Mississippi River floods by creating