VOL. XIX. (I) THE CONTROL OF RIVER CHANNELS 
33 
In Figs. I and 2 the river is shown as having been pushed 
into the desired line : in Fig. 3 it is shown as having been 
drawn into the same line of curve as that in Fig. 2. This is 
done by the influence of tributary streams, of which I have 
spoken as having an important diverting influence. They are 
a serious cause of trouble in the management of rivers ; but, 
on the other hand, if by their eigency, rivers are brought to 
points beyond which it is impossible to go, the cause of trouble 
may become an important advantage, because it is a means of 
lorming those fixed points, which, as I contend, are all that 
is necessary in the fixing of the low-water channels. Piers 
at the points of junction, as shown in Fig. 3, may or may not 
be necessary, according to the character of the soil or other 
circumstances ; and if the current in the river be such as to 
cut a channel across the curve, a short groyne to deflect the 
stream toward the intended line may be necessary. I am not 
suggesting that curves such as those shown in Figs. 2 and 3, 
would ever be constructed, unless it were for purposes of 
experiment or of ornament. 
Here the question arises, which is the more useful, a river 
flowing in a straight line or one that has a winding course ? 
I prefer the latter — a river that swings from one side to the 
other in wide curves, as shown in Fig. 4. This, as compared 
to a straight line, slightly increases the length of the stream, but 
it has important advantages. There is no need to have a back- 
water on both sides ; the river bank is drained, as it always 
should be, in a direction away from the river, so as to keep the 
bank unbroken, excepting only at the fixed points where the 
tributaries join the river on the convexities of the curves. 
Here they fall in at a tangent with the river line, thus avoiding 
disturbance of the stream at the point of junction. There is 
no need to guide the lateral stream to a middle line, where 
fixed points would have to be constructed. Landing stages 
can be placed at the fixed points, so as to afford means of 
communication between the two sides of the river, with beacons 
on the land to indicate the line of the channel when the river 
area is covered with water. If the river be used for boating, 
the oblique lines are better adapted for tacking than a straight 
one. Finally, there is a reason which speciall}^ commends 
