DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 433 
The reservoirs already constructed and being operated at the head- mr. McMath, 
waters of the Mississippi afford a very useful object lesson, if studied 
and the lesson applied intelligently. So far as the writer knows no 
such study and application have been made.* 
The formal perfunctory Annual Reports to the Chief of Engineers 
drop a fact here and there, but are chary of giving interpretations. 
The following explicit statement is made in the Report for 1905, Vol. 
VI, page 1675: 
“Without the reservoirs the discharge of the river at St. Paul is 
liable to fall to 1700 cubic feet per second, as it has in the past, while 
with them the discharge can be readily held throughout the navigation 
season at or above 6 000 cubic feet per second.” 
That is, the present reservoirs are entitled to credit for a contribution 
of 4300 cu. ft. per sec. to the Mississippi from below St. Paul to the 
Gulf of Mexico. For it must be borne in mind that a contribution to 
maintain a flow is subject to loss from absorption or evaporation only 
in the proportion that the augment increases the surface, which is very 
small. If delivery of the augment is delayed, and it be expected to 
fill a depleted bed, the losses by absorption and evaporation will easily 
consume the augment. 
For this reason, the writer credits the reservoirs with the full 
amount of their contribution from St. Paul to the Gulf. <A further 
contribution from the Great Lakes Reservoirs, via the Chicago Drainage 
Canal, is added to the Middle Mississippi at Grafton. The minimum 
discharge at St. Louis has been said to be about 40000 cu. ft. per sec. 
A 10% increment from St. Paul and 20% more from Chicago is a 
material increase, and certainly assists in the formation and mainte- 
nance of the low-stage channel; but, from the writer’s point of view, 
a more important result is the general condition of the river bed due 
to low-stage scour. Nothing will be said about bank erosions because 
they are not perceptibly affected by contributions. 
Suppose this prepared condition, which ordinarily will last and 
progress during the winter months, is followed by a heavily silt-laden 
flood from the Missouri River. The water flood-wave will outrun the 
silt load which drops on and elevates bars; not only is the low-water 
channel obliterated, but the elevated bars are obstructions left by the 
flood-wave to impede the flow of subsequent flood-waves, which, to pass 
equal volumes, must rise higher on the gauge. In other words, flood- 
waves impair the efficiency of the regular river bed. Mere lies the 
weakness of levees as a protection against extreme floods. There is not 
only no way of arriving at the volume that may come as a result of 
combinations of discharge from numerous tributaries, but in case of a 
* The author in a foot-note refers to a report of a Board as containing ‘‘ exhaustive 
data upon the system and its operation’; the writer has not seen the report. 
