VOL. XIX. (i) THE CONTROL OF RIVER CHANNELS 
31 
recognised, but I may call attention to the fact that if two 
streams flow down a smooth surface, as of glass, gently and 
evenly inclined longitudinally but perfectly level from side to 
side, they will flow in parallel lines, in accordance with the force 
of gravity. If, however, one of the two streams be diverted so 
as to touch the other, there will be a momentary hesitation 
before the two coalesce : the surface-tension must be overcome ; 
but when they have coalesced they will flow on together. 
The cohesion of the water is such that they will not separate 
although the force of gravity is in operation just as before. 
The diverted stream flows on in the line of the one which it has 
joined. So, too, if the main stream of a river come in contact 
with a smaller one or tributary as, at flood time, it will do, 
the larger stream may adopt the channel of the smaller one 
and the two flow on together, the channel of the larger one 
remaining only as a backwater. In fact, tributaries are often 
seen to be flowing into a river on the convexity of a curve 
while the remains of an old channel can be traced along the 
line of the arc of the curve. 
If a river have been diverted from a straight line and the 
object be to compel it to return, and to remain in that line, 
continuous guiding or training walls are not necessary ; there 
is no need for treating the river as if it were a herd of cattle, 
from which one or more of them might at any point go astray. 
A number of fixed points which the river must pass will be 
sufficient. In Fig. i a series of low-water groynes are in- 
dicated, each of them coming to an imaginary line on either 
side of the river. How near together these must be will 
depend not only on the number and the character of the divert- 
ing influences, but, also, on the size of the river. One which 
required a large space in which to turn would, of course, be 
unable to deviate from its line unless the interval between 
the fixed points were wider than would be sufficient in the 
case of a smaller one. Thus fewer groynes would be needed. 
If the object be to compel the river to follow a curved 
line, as shown in Fig. 2, similar groynes will suffice, the number 
corresponding to the number of curves. Here the river turns 
aside before it comes to any obstruction at all : it flows for a 
short distance in a direction transverse to the downward line. 
