644 PROCEEDINGS OF SECTION H. 
the gorge of the Otira, where two bridges 20 feet high were 
buried in the shingle. 
This raised bed was all removed after a few years, and 
the usual slope restored, and if it were not that it left a big 
boulder perched on a point of rock, the coach-drivers 
would have difficulty in persuading tourists about the 
great shingle slide. 
But the character of the rocks of the mountains determines 
the nature of the sediments in the rivers. and this determineg 
the slope of the river-bed. Thus, the rivers of New Zealand 
may be sloped 60 to 100 feet per mile, while on the Colorado 
River in North America the rocks are ground down, and 
are carried away to such an extent that the river-beds are 
sloped only a few feet per mile, and are six to ten thousand 
feet deep. 
The above shows the difficulties that will be met with 
in trying to mitigate floods by deepening the river-bed. 
Nevertheless, it has been done successfully in parts of many 
rivers, and its effectiveness, it seems to me, depends on the 
character of the deposits as compared to the volume of the 
river; thus, where the river is so great, and- the deposits 
so fine, that the river is of great depth, the problem is 
solved naturally; where the deposits are heavy, and the 
river small, there it is impossible to permanently deepen 
the whole or even part of the bed; and in cases inter- 
mediate between these extremes the bed may be deepened 
permanently by carefully selecting the position and direction 
of the deepening, and assisting the operation by judicious 
training of the banks within stone embankments. Of 
course, if the bed is deepened, in a corresponding degree 
the effect of floods will be mitigated, because lowering the 
bed is equivalent to raising the banks. A notable example 
of relieving floods by lowering the river-bed and straighten- 
ing its channel was seen in the great flood in the Yarra, in - 
April 1901, which in consequence of this beneficial kind of 
treatment did no damage at all, although the flood was 
greater than any since 1863, each previous flood having 
caused great damage. ~But there is necessarily a limit to a 
partial improvement of a river-bottom—which is that the 
beneficial effect will only extend to the next important 
change of grade in the surface-inclination either above or 
below the improvement; thus, the part under treatment 
may have a general inclination of one in five thousand, but 
above it may be a long stretch of one in three thousand, 
and below it a long stretch of one in twelve thousand, and 
