42 BULLETIN 852, U. S. DEPARTMENT OF AGRICULTURE. 
pipe is 11.54 square feet. The diameter of the pipe line was accepted 
as 46 inches. 
The velocities for the various runs were determined by timing the 
passage of fluorescein from the point of injection, at gauge 1, to its 
appearance at the outlet. From 2 to 5 batches of color were used for 
each run. For any one run the elapsed time of the colors agreed 
within one-third of 1 per cent. 
Reference to Plate VI shows that the curve for the 1915 experi- 
ments, when produced, passes through the midst of the points for 
Newell's tests made in 1911 and 1912. If this fact is truly indicative, 
then the capacity has not changed during the interim. This is rea- 
sonable, because the pipe is operated at a rather high velocity 
throughout the season and the feeder canal leading from the main 
canal to the intake is in a hard cemented gravel that would not corn- 
tribute ravelings or silt, while the water itself comes from Cold 
Springs reservoir, and so is free from silt, except such as it may pick up 
in transit from the reservoir to the pipe. The clean condition of the 
pipe interior, where examined at the outlet, would indicate a pipe 
comparatively free from slime. 
The mean value of the coefficient in the new formula, as determined 
from the Newell tests and our own, is 0.395, but it must be borne in 
mind that this pipe is practically straight, a rather unusual condition 
for a pipe nearly 2 miles in length. 
No. 33b, Experiment S-60. — 48-inch monolithic reinforced con- 
crete pipe, Churn Creek siphon, Anderson-Cottonwood irrigation dis- 
trict, California. — In order to serve a certain acreage of land lying east 
of the Sacramento River, the flow of Churn Creek lateral is conveyed 
for about 4,300 feet in a round, monolithic siphon pipe extending from 
an open earth canal to a steel flume across the main channel of the 
river. (PL IV, fig. 2.) It was the original intention to extend the 
siphon under the bed of the river as a cast-iron pipe, but this plan was 
changed, so that at the edge of the deep-water channel the concrete 
pipe is turned vertically upward and placed in the first pier of a bridge 
across the river. This bridge supports a steel flume and the pipe dis- 
charges directly into the flume. 
As shown in the plate, the siphon was constructed continuously, 
wood being used for both inside and outside forms. Though originally 
designed 4J feet in diameter, the pipe was constructed with a diameter 
of but 4 feet, owing to a reduction in the acreage to be served. 
A single experiment for loss of head was made on this pipe in the 
summer of 1919. Piezometer tubes of type A were thrust into the 
siphon at inlet and outlet and the pressure head was carried over the 
banks to glass-gauge water columns. Piezometer No. 1 was thrust 
about 6 diameters (24.7 feet) down the inlet end and No. 2 was held 
10 feet down the vertical outlet pipe located in the bridge pier. 
Between these piezometers there is a developed distance of 4,242.3 
feet, so that a very material loss of head occurs for the commercial 
velocities — between 3 and 4 feet per second. The reach tested is 
practically straight in both vertical and horizontal planes, with the 
exception of the90°bend as the pipe is turned upward in the bridge pier. 
The loss due to this bend is immaterial compared to the friction loss, 
and was ignored in the computations. The velocity of the water in 
the pipe was determined by injecting a solution of "Congo red" dye 
directly into the mouth of the siphon pipe, not into the pool in the 
