88 
BULLETIN OF THE BUREAU OF FISHERIES 
Under primitive conditions the surface waters of a period of heavy rainfall might 
have required weeks to find its way through forests and over unbroken ground to the 
river channel, 2 while still further weeks must have been consumed as the swollen 
stream wound its way to the sea through the dense swamps that bordered the normal 
channel. A new flood period might, indeed, have begun even before the first had passed 
and before a stage of extreme subsidence had ensued. 
Under modern conditions a similar excess of rainfall finds not only the forests 
diminished but surface ditches and tile drains carefully laid out to expedite its removal 
from the lands to the natural channels of the rivers. Once in the rivers, it again finds 
its way in some measure cleared, so that it may be conveyed to the sea at a more 
rapid rate. The river, then, rises and overflows with relative dispatch and subsides 
the more quickly again to a low stage; the fluctuations of level are sharper and more 
extreme. 
The effect of rapid fluctuations upon fish life in the stream is too obvious to 
require lengthy comment. We know that many of our most valued fishes move out 
into the shallow waters for the purpose of reproduction, and we see too often that an 
untimely recession of the floods leaves not only mature fishes of various sizes but also 
a large proportion of an entire generation of young fishes to perish in isolated overflow 
ponds, except as they are rescued by artificial means. Furthermore, the conditions 
for the development and maintenance of a food supply for fish are obviously the less 
favorable, the greater the degree of fluctuation in level and in expanse of the water. 
One would never expect to secure the maximum production of fish in a pond that is 
to-day 10 acres in expanse and 20 feet deep and next week 3 acres in extent and 8 feet 
in depth. In any stream the fish and other aquatic animals are subject to natural 
vicissitudes under the best conditions, but the trend of the changes in our rivers 
wrought indirectly by man’s alteration of the face of the land has been in a direction 
generally unfavorable to the growth and multiplication of fish. 
Some of the changes wrought by man may be in some measure compensatory 
in effect, however. Of this nature we might expect, generally speaking, to be those 
developments that tend to control the stream flow or to diminish the degree of fluctua- 
tion of level. Among the factors contributing to the control of stream flow are 
improved methods of agriculture, providing for storage of water in the soil, and the 
impounding of waters under such conditions as to hold back the surface water in times of 
excess and to release it gradually in times of deficiency. Such impounded waters we 
may call reservoirs when the purpose is essentially the temporary storage of excessive 
rainfall. Another class of impounded water consists of the pools or lakes formed by 
dams constructed for purposes of power. With these the storage and the liberation 
of water may be governed largely by power demands, and their effect upon stream 
flow below the pool may be one thing or another, according to local conditions and 
according to the mode of operation of the plant. In any case, however, we are likely 
to find large expanses of water of relatively fixed level and affording favorable breed- 
ing and feeding grounds for many species of fish. 
In the case of impounded waters classed as reservoirs, where the water is held 
temporarily, we look for the beneficial effect upon fish life not in the reservoir itself 
so much as in the stream below, the flow of which is rendered more uniform. In the 
case of artificial pools and lakes we look for beneficial effects primarily in the lake itself 
and above it, if fish should thrive in the pool and wander from it. 
2 Kafter (1903) cites important references bearing on this question. 
