THE FLOW OF WATER IN CONCRETE PIPE. 85 
with fine gravel not exceeding 1 inch in diameter. The sections were joined together 
in the ordinary way with cement mortar mixed with 1 part cement to 2 parts fine 
sand. The line has tangents ranging in length from about 100 feet to 500 or 600 feet, 
and curves, none of which have a less radius than 15 feet. The slope was made uni- 
form at the rate of 10.56 feet per mile. 
This pipe line delivered at its lower end, when filled to its full capacity, 9.2 second- 
feet without any pressure at the intake, which was provided with an enlarged section 
for accelerating the flow of water in the pipe. It was observed that the pipe did not 
run full below the first 2,000 feet. Subsequently the upper 100 feet of the pipe were 
raised to accelerate the water a little more, after which numerous tests Were made to 
determine the greatest volume of water that could be passed through the pipe. From 
this I found that the pipe would carry the most Water when filled within 1 inch of 
the top, or when carrying a depth of 21 inches of water. Under this condition the pipe 
delivered 9.8 second-feet of water. The discharge of the pipe was measured by means 
of a rectangular, fully contracted Francis type of Weir in one-eighth inch steel plate, 
and head taken with hook gauge. The depth of water in the pipe, from which the 
velocity and hydraulic radius are computed, was measured by means of a straightedge 
and hook gauge at the manholes, which are located every 500 feet along the line. 
Instead of using a trowel to make the joints, as is ordinarily done, I made a brass band 
ring 10 inches wide, having spokes equipped with turnbuckles on the inside to vary 
its diameters, the ring being flexible and lapped to permit reducing or enlarging 
its circumference. 
In_ making the joints the pipes were carefully laid and evenly joined, and mortar 
applied in the groove all around . Then the brass ring, reduced in diameter to slip into 
place, was introduced and centered over the joint. By working the turnbuckles the 
ring was then expanded to fit the diameter of the pipe and its inside circumference, 
thus squeezing the mortar into the crack and the surplus out to the edges of the ring. 
Grasping the spokes, the ring was then turned around slowly about five times to 
give the joint a smooth surface, after which it was loosened and removed, and all 
surplus mortar removed with a trowel, in such a manner as not to mar the surface left 
by the ring. 
This method made the joint smoother than any other portion of the pipe, and, so far 
as the eye could detect, made the pipe continuous on the inside. Had the same 
method been used for the 31-inch pipe for Mill Creek No. 3 line (No. 51) and for the 
36-inch pipe for Lytle Creek (No. 53) I have no doubt these would have given about 
the same value for n in Kutter's formula. 
The nominal slope was used in computation. 
As shown in Table 11 the above test indicates the value of n to be 
about 0.0116 for this pipe. If truly representing the inner surface of 
the pipe, the joints were most carefully made and the pipe in excellent 
condition. (See discussion No. 51 following.) 
No. 51, Experiment FF-2. — 31-inch jointed cement pipe of South- 
ern California Edison Co. Mill Creek power plant, No. 3, Cali- 
fornia. — The description of the experiment upon this line, as 
taken from correspondence with Mr. Finkle, reads : 
The diameter of this pipe was 31 inches and it was given a slope of 10.56 feet per 
mile. Its length was between 5 and 6 miles, with a few interruptions in the line, 
where steel siphons were used to cross ravines. Having profited by my experience 
regarding accelerations at the intake, showing that the pipe would be full at the upper 
end and only partly filled at the lower end, I gave this line additional fall in its upper 
part, calculated by a formula for that purpose. 
This pipe was manufactured in the same way as the 22-inch pipe (No. 50, p. 84), 
except that 1 part of cement was used and 4 parts of sand, containing considerable 
gravel, the largest of which was 1 J inches in diameter. It was made in 2-foot sections, 
and laid in the same manner as the 22-inch pipe. The result was that the maximum 
capacity of the pipe occurred when it was filled to within 1^ inches of the top. or 
when it only carried a depth of water of 29^ inches. The carrying capacity of the pipe 
under these conditions was 19.72 second-feet. 
This line was laid in 1901 and tested the same year. There are numerous curves 
between the tangents of various lengths, but none of the curves has a radius of less than 
20 feet. The discharge was determined by using a Francis type rectangular, fully 
contracted Weir in three-sixteenths-inch steel plate, measuring head with a hook gauge. 
(See PI. X, fig. 1.) 
The depths of water in the pipe, from which the velocity^ and hydraulic radius are 
computed, were measured by placing straightedge along inside of upper invert on 
