28 BULLETIN 852, TT. S. DEPARTMENT OF AGRICULTURE. 
easily whenever the current carried it against the periphery of the 
pipe~ rather than lurch back into the current, as would a more 
irregular float. The maximum variation in time of any one float 
from the mean time of all of the floats was about 5 per cent. In 
connection with tests on pipe No. 7, floats of this type agreed with 
the velocity as determined by a well-made, contracted, rectangular 
weir 2.005 feet long within 1.2 per cent. At the intake a hook 
gauge in a stilling box determined the pressure head at the piezo- 
meter tube (type A) located 10 feet down the pipe. At the outlet 
a similar tube was held 1.6 feet in the pipe against the current. The 
pressure head at this tube was read in a water column beside the 
riser pipe. This pipe was laid in 1885 and appeared to be per- 
fectly clean. As it is in an open field there is no chance of roots 
choking the interior. The nominal diameter was accepted in com- 
puting area and quantity of water (see No. 1, p. 25). The value of 
C s varies with the velocity, from 0.244 to 0.278. 
No. 6, Experiment $-33. — 12-inch jointed cement pipe, San An- 
tonio Water Co., Ontario, Calif. — This pipe lies between two division 
boxes and extends under a wash located north of Twenty-fourth Street 
and west of Euclid Avenue near Ontario. Though too short for experi- 
mental purposes, if low velocities are to be considered the 227.5 feet 
between the boxes is ample when the commercial velocity, nearly 
6 feet per second, is obtainable. Water columns were used at both 
gauges. These were attached to piezometer tubes of type A, No. 1 
being placed 7.7 feet in the pipe from the inlet chamber and No. 2 
being held 2 feet up in the pipe from the outlet chamber. Velocities 
were computed from the mean time of four batches of fluorescein, 
timed from the instant of injection at the inlet to the mean between 
the first and last appearance at the outlet. Just four seconds (by 
test) was required for a given body of water to pass through the 
length of hose under a head of 1 foot. This head is of course the 
difference in elevation between the surface of the water in the 
chamber and the outlet of the hose. The nominal diameter was 
accepted in computing area and quantity (see No. 1, p. 25). 
This pipe line was laid in 1911. The units are 2 feet long, made 
with a dry mix tamped into steel molds and afterwards washed on 
the inside with cement grout. The line is buried 2 or 3 feet and is 
straight but for a slight sag under the wash. The value of C s , 0.322, 
shows the marked improvement in joints in pipe line laid in recent 
years. 
No. 7, Experiment S-27. — 12-inch jointed cement pipe, San An- 
tonio Water Co., Ontario, Calif. — A straight, level reach of lateral pipe 
on Fourth Street, Ontario, was chosen for this test. Water enters the 
pipe over a division wall in a sand box and leaves it at the bottom of a 
similar sand box on the corner of Fourth and Euclid Streets. Piezo- 
meter tubes of type A were used at both ends of a reach 869.4 feet 
long. The pressure head at gauge 1 was read in a water column 
beside the sand box, while that at the outlet piezometer was read by a 
hook gauge in a stilling box attached to the piezometer by pressure 
tubing. Piezometer No. 1 was 4.7 feet from the inlet and No. 2 was 
1.9 feet from the outlet. The mean time of six weighed orange floats 
like those described under No. 5 was used in computing the velocity 
of the water in the pipe. Accepting the mean area of samples of pipe 
of this size in the pipe yards of this company and comparing the ve- 
