IRRIGATION EXPERIMENTS IN IDAHO, A] 
SEEPAGE INVESTIGATION. 
The losses to which the waters of an irrigation project are subject 
in passing from the point of diversion to the land to be irrigated 
have not always been given sufficient consideration in the designing 
of projects. It became evident early in the investigation that the 
losses to which canals in different types of soil are subjected should be 
measured in order to determine how much water should be diverted 
to deliver the quantity required. During the years 1912 and 1913, 
118 sections of Idaho canals were measured for the determination of 
the seepage losses. These canals varied in discharge from 0.07 cubic 
foot per second to over 3,190 cubic feet per second, and in cross 
section from 0.117 square foot to 984 square feet. A total of 287.31 
linear miles was included, covering all of the common soil types of 
Idaho, and it is believed sufficient data have been secured to furnish 
a substantial basis for estimating the losses to which the canals of 
this State will be subject. 
METHOD OF INVESTIGATION. 
All canals included in the investigation with a discharge of 3 
second-feet and less were measured with Cipolletti weirs installed 
at the upper and lower ends of each section. The head on these weirs 
was read to the nearest 0.001 foot with small inexpensive hook 
gages that were especially designed for the purpose. All canals with 
a greater discharge than 3 second-feet were measured with current 
meters. The 0.2, 0.6, and 0.8, or 3-point, method of measurement 
was used in all cases where the canal exceeded 1 foot in depth, and 
either the 0.6, or integration, method with the shallower canals. In 
_ order to eliminate errors in computation, individual characteristics 
of the meters, and personal equation of the men, every section meas- 
ured with current. meters was measured by at least two hydrog- 
raphers, each with a different meter, thus furnishing a double check 
on all results. It was found that fluctuations of gage height and 
discharge were the most troublesome factor in connection with 
Seepage determinations, and the following method was used for the 
elimination of the effect of the fluctuations: 
Canals to be included in the investigation were thoroughly looked 
over before any ratings were made, and gages which could be read 
to the nearest 0.02 foot were installed approximately every 2 miles 
in the main canal and at the head of all diversions, care being used 
to pick good sections in the middle or in the lower third of as long 
a tangent as possible. Discharge curves were plotted for each sta- 
_ tion by using the bottom of the canal as the zero or point of no flow, 
and the point at which water stood when the measurement was made 
as the only other point that was determined for each curve, care 
