THE FLOW OF WATER IN CONCRETE PIPE. 81 
average, because observations were not of a character permittting reduction for indi- 
vidual pipes. These pipe measurements included the chamber and transition- 
section losses, of all kinds, at both ends of each pipe because it was not considered 
advisable at that time to provide piezometers, or other devices, necessary to the 
direct measurement of the pipe losses separately. These observations were reduced 
by taking advantage fo the fact that the pipes were of various lengths and the same 
type of construction, including the chambers, so that the assumption could be reason- 
ably made that the wetted surfaces are practically identical in character: hydraulically, 
and that chamber losses are the same in all chambers for any given flow conditions. 
Losses in the pipes were computed, using assumed pipe friction-loss coefficients, 
which computed pipe losses were subtracted from the measured losses. These com- 
puted residual losses were than compared and the set most nearly consistent selected. 
The coefficient used in obtaining the selected set was assumed to be an average for 
the several pipes and an approximately correct average because of the method used 
and the probability that not all pipes were in the same condition as to foulness, 
hydraulically considered. The value selected is believed to quite closely approxi- 
mate the truth, because some of the siphon pipes are so short that the total friction 
loss in them is very small as compared to the chamber losses. 
While the observations are not such that definite conclusions there- 
from would be warranted, still the indications are that a steel pipe, 
built up of relatively long sections, well jointed and lined with a 
smooth cement coating, gives a very efficient surface. Of course the 
primary object of this coat is to prevent the corrosion of the metal 
interior but is gives an added satisfaction to know that this prevention 
has been attained without sacrifice of capacity, for a given sectional 
area, If anything 7 the fractional loss is less than it would have been 
in a new metal pipe and the latter material would have continually 
lessened in capacity while the fining will probably remain about the 
same, after the first slime coat is acquired. 
No. 39, Experiment M-2. — Rondout pressure tunnel, 1 Catskill 
Aqueduct, X. Y. — At the Rondout River crossing the Catskill 
Aqueduct takes the form of a circular pressure tunnel, excavated 
in solid rock and lined with concrete. From the standpoint of 
capacity the tunnel is, in essentials, a circular pipe, constructed in 
place. (See PL IX, fig. 1.) 
The tunnel consists of vertical downtake and uptake shafts joined 
by an approximately horizontal tunnel. The developed length of 
the tunnel is 24,880 feet. About 16,000 feet from the intake a vertical 
drainage shafts extends from the ground line down to the tunnel. 
As the ground line at Rondout River is 300 feet below the hydraulic 
gradient, this shaft is, of course, sealed. 
In August, 1915, F. F. Moore, designing engineer of the board of 
water supply, conducted a series of experiments to determine the 
friction losses in this tunnel, from the drainage shaft to the outlet, a 
distance of 9,102 feet. The pressure head at the drainage shaft was 
measured with a mercury manometer of the pot-and-column type. 
Mercury readings were corrected for temperature of air and of water 
in the tunnel. The temperature of the water in the pipes to which 
the manometers were attached was assumed to be controlled by the 
ground and therefore unchanging. The mercury column was also 
read with no flow in the tunnel, thus establishing the levels between 
gauges. 
The elevation of the water surface at the outlet was determined 
by steel- tape measurements from a bench on the floor over the 
1 Eng. Rec, Mar. 11, vol. 63, 1911, p. 27P; Eng. News, June 1, vol. 65, 1911, p. 654; Water Works Hand- 
book, by Flynn, Weston, and Bogert, New York, 1916, p. 284. 
164725°— 20— Bull. 852 6 
