12 BULLETIN 194, U. S. DEPARTMENT OF AGRICULTURE. 
as it would if the lower nail were set first. Enough cross sections 
were taken to determine the mean water area and length of wetted 
perimeter throughout the reach. Nails were set with their tops flush 
with the water surface at each end of the reach. Levels were then 
run between these nails as bench marks. Temperatures of air and 
water were taken. After all these data had been taken the chief of 
party was familiar enough with the conditions to write an intelligent 
description covering the features influencing the flow of water 
throughout the reach tested. ■ 
Throughout the descriptive matter in this publication station is 
at the upper end of the reach, and integral stations represent 100-foot 
intervals below station 0. 
Canal reach. — Actual conditions encountered in the operation of 
irrigation systems and under which canals must carry water were 
those desired, as it was believed the designer must take these into 
consideration in developing the hydraulic elements of his proposed 
canal. Other things being equal, a reach of canal 1,000 feet long 
was chosen with quite uniform flow, not only throughout the reach 
but also up and down stream so far as conditions might influence 
results. These reaches were chosen on tangents, on curves, and on 
both, in order to determine the influence of curves on the retardation 
factor. 
Discharge measurements. — Small discharges were, when practica- 
ble, measured with a Cipolletti weir, under standard conditions. 
Discharges of more than 4 second-feet were measured with the current 
meter. The surface widths of ditches less than 10 feet in width were 
divided by vertical lines one-half foot apart. Those of canals more 
than 10 feet in width were divided into approximately 20 verticals, 
the sections thus formed near the banks being narrower than those 
toward the middle, for the reason that the velocity changes more 
rapidly near the banks. The average velocity in the verticals was 
determined, as a rule, by the multiple-point method, interpreted 
through vertical velocity curves. In any one vertical the meter was 
held at points 0.2, 0.4, 0.6, and 0.8 of the total depth below the surface 
of the water. In addition to these points, at approximately every 
other vertical, the meter was held at points 0.1 or 0.2 foot (depending 
on the surface velocity and consequent roughness) below the surface, 
and just clearing the bottom. These points gave the necessary 
information to plat correctly the vertical velocity curves between 
the surface of the water and the point 0.2 of the depth below the 
surface and between the bottom of the channel and the point 0.8 of the 
depth below the surface. In addition to all of these points, in many 
verticals the meter was held at a point 0.3 of the depth below the 
surface for the purpose of developing the point representing approxi- 
mately the maximum velocity in the vertical. The velocity at this 
