OVERFLOW OF RED RIVER OF THE NORTH. 13 
The elevations of the United States Coast and Geodetic Survey- 
bench marks are the result of the latest (1912) adjustment of the 
precise leveling done by that survey in 1906. At that time the eleva- 
tions of some of the United States Geological Survey bench marks 
were redetermined by this precise leveling and the others in the 
vicinity were adjusted to agree. Since then, the leveling of the 
Geological Survey in the valley has been based upon the Coast and 
Geodetic Survey elevations, so that at the present time the elevations 
of all bench marks of both surveys, as used in this investigation, 
accord with the latest adjustment of the precise leveling results. 
Most of the leveling was done with wye levels. A small part 
along the Bois de Sioux was done with transits. All levels were re- 
quired to check within the usual limits of error for wye level work. 
Base-level lines were run for the entire length of the Bois de Sioux 
and Red Rivers within the United States, and elevations at frequent 
intervals were taken of the river bed and banks. From the site of 
the proposed detention dam northward as far as Wahpeton all land 
that is affected by floods was covered by the level parties, elevations 
for the most part being taken along section lines. These areas are 
shown on sheets 2, 3, and 4, figure 3. Over the remainder of the val- 
ley the leveling was confined to the immediate vicinity of the river 
channel and to ties to various bench marks and points of former 
surveys. 
CROSS SECTIONS. 
Cross sections of the Bois de Sioux and Red Rivers were taken 
along the whole length at intervals of approximately one mile, and 
at all bridges (see fig. 4). The valley was cross-sectioned at the 
sites of the proposed levee at Browns Valley and the detention dam 
near White Rock. Test borings were also made at the latter point. 
A few cross sections of the Otter Tail, North Dakota Wild Rice, 
Sheyenne, and Red Lake Rivers were taken. 
HIGH-WATER. DATA. 
Special effort was made to locate all reliable high-water marks, 
both those left by the water or ice on trees, bridges, etc., and those 
which were set by observers as permanent records. The field parties 
were constantly on the lookout for reliable information as to high 
water and from time to time special field trips were made by en- 
gineers for this purpose. The elevations of all authentic high-water 
marks found were taken, and are shown on figure 3, sheets 1 to 20. 
STREAM GAUGING. 
Staff stream gauges were established at Haggart and Valley City, 
N. Dak., on the Sheyenne River; at Wild Rice, N. Dak., on the 
Dakota Wild Rice; and near Fairmount, N. Dak., on the Bois de 
