OVERFLOW OF RED RIVER OF TILE NORTH. 19 
thence out of the watershed. Precipitation is the source of run-off, 
but many different factors affect the amount and rate of run-off 
from a given watershed. A portion of the water is absorbed by the 
soil ; some of this passes through the soil into drainage channels. A 
part is evaporated from the soil or is transpired by plants and trees 
or is evaporated from the surface of the soil. A small amount sinks 
deep into the ground, and the remainder passes over the surface of 
the ground to a watercourse and through it out of the watershed. 
The rate and amount of run-off from a watershed are affected 
principally by the amount and distribution of the precipitation and 
its form — whether rain or snow. 
After precipitation the princip_al factors which affect the rate of 
run-off from a watershed are topography, size and shape of the 
watershed, prevailing temperature, character of vegetation on the 
watershed, and artificial drainage. In addition to these and some- 
what dependent upon them are evaporation, transpiration of plants, 
absorption by the soil, and storage, all of which tend to regulate the 
amount of water reaching drainage channels. The effect of these 
several factors upon the amount and rate of run-off is not constant 
but is continually changing. It is exceedingly difficult, if not im- 
possible, to measure the effect of the various factors upon the run-off, 
and in the absence of measurements from a given watershed the 
determination of the probable rate of run-off becomes a complex 
problem. 
The best bases for calculating future run-off are measurements of 
stream flow extending over a long period of time for the particular 
watershed under consideration. If such records are available, the 
amount and rate of run-off can be estimated closely unless the 
physical features which affect run-off from the watershed are ma- 
terially changed. If complete data for a long period are not avail- 
able, then records for a shorter time on the watershed under con- 
sideration, or on more or less similar watersheds, are most essential. 
For the watershed of the Red River, and especially for those of its 
tributaries, the run-off records are meager because settlement and 
development in the valley are only comparatively recent. 
Daily records of the river stage at Grand Forks, from which the- 
run-off at that point can be computed, are available as far back as 
1882. From 1882 to 1901 the gauge readings were recorded by the 
Corps of Engineers,- United States Army, and from 1902 to the 
present time by the United States Geological Survey. Similar 
records are available for the Reel River at Fargo from 1902 to the 
present time, as are also records for 12 stations on tributary streams 
for periods of from 1 "to 17 years, since 1901. The daily flow of the 
river at Emerson, Manitoba, just north of the international boundary, 
