650 PROCEEDINGS OF SECTION H. 
If we observe what Nature does in its treatment of a river 
and its floods, we shall notice, first of all, that the treat- 
ment varies greatly according to the soil, the character of 
the surface, the rainfall, and so on. Every river may be 
divided into two parts. In one part, namely, the upper part, 
the river is. slowly cuttinz ut and lowering its bed; and in 
this part the river is seldom obstructed by its own sediments. 
The other part is the lower part, and here the sediments 
are generally increasing, and slowly encroaching on the 
upper part; and it is commonly in the lower part that 
floods are troublesome. But Nature has one working ‘ule 
which is used more than any other; in fact, she uses it 
wherever and whenever she can make it apply. 
The object of this working rule is, first of all, to reduce 
the surface of the river and the surface of the land to a 
uniform inclination; and next to raise the land above the 
river, so that the floods can be’ contained within the banks, 
To be entirely successful it is necessary that the river should 
have to deal with material of more or less uniform consistency. 
But, as I mentioned above, the material of the river 
‘sediments continually varies, even though it may be ages 
before the change is perceptible, so that the river may never 
succeed in fixing a limit at which it can control its own 
floods. Yet everywhere one finds approximation to a com- 
plete solution of the problem; and then one ‘observes that 
the floods have silted up the meadow lands until the greatest 
floods overflow only a very little; and this very little is 
more or less completely protected by the banks on the 
immediate edge of the river, which are always a few feet 
higher than the meadow land at the back. 
Tt js evident that man cannot imitate this operation. 
To raise all the land above flood-level is out of the question ; 
to raise the banks only in the form of levees is most com- 
monly resorted to, but is often very dangerous. 
But the natural operation just described above is in 
effect to raise the banks, and thus deepen the river, whereby 
the whole of the floods are contained within the banks, and 
the depth and impetus of the water is thus made sufficient 
to scour away the sediments as long as they do not change 
in weight and size. 
How then are we to imitate this operation, and secure the 
same results? Evidently to deepen the bed is the safest 
course; but in most cases, this must be combined with 
raising the banks; and the nature of the bed and other 
local conditions must determine how much of the one and 
of the other should be done. 
