THE FLOW OF WATER IN WOOD-STAVE PIPE. 65 
Moritz suggests the extensive use of air valves in some form. The 
best air outlet is probably a "chimney" rising above the hydraulic 
gradient. There should be several connections to the siphon, either 
normal to the pipe or, better, pointing slightly upstream. These 
may be independent of each other or all may connect to a common 
pipe extending along the top of the siphon, the upper end passing 
through the head wall of the intake chamber so that any water that 
blows out with the air will fall back into the pipe. The last connec- 
tion should be at a lower elevation than the water surface in the out- 
let chamber so that, with a small discharge, the air entrained will be 
collected and passed off. As a rule low discharges entrain more air 
than does a full discharge, since the water rushes down the initial 
reaches of pipe in a turbulent condition. On the other hand, the 
upper air vents are necessary to care for air entrained and compressed 
during discharges approaching maximum capacity. These vents 
may be from 1 to 12 inches in diameter, depending on the size of the 
wood pipe, and should be so assembled that they may be taken 
apart, as debris collects in such vents and must be periodically 
removed. If excessive air troubles are present, a collecting chamber 
may be attached to the siphon at each vent, the air pipe being at- 
tached to the top of the chamber rather than directly to the siphon 
pipe. 
FRICTION LOSSES. 
The loss of head necessary to overcome internal resistances within 
the pipe is proportional to the length of the pipe but is independent 
of the static pressure in the pipe. That is, the loss necessary in the 
conveyance of a given quantity of water through a siphon pipe will 
be the same whether the low point is, say, 10 feet or 150 feet, below 
the hydraulic grade line, the other factors remaining unchanged. The 
influence of temperature upon the frictional resistances was found 
by Saph and Schoder to be considerable in small brass pipes but has 
not been studied in connection with tests on large wood pipes. It is 
doubtful whether the influence of temperature could be differen- 
tiated from that of friction alone in tests on large pipes in commercial 
service. 
In order to determine the size of pipe and the loss of head neces- 
sary to overcome the frictional resistances in the conveyance of a 
given quantity of water, two estimate diagrams and a table have 
been prepared. Two examples of typical pipe problems are given. 
In these the use of the diagrams only is explained, as the table is 
considered self-explanatory. The factors of safety given below 
should be considered in each problem, as a study of Plate VII shows 
that an averaging formula, accepted literally, can not assure the 
desired discharge for a given loss of head. 
68796°— Bull. 376—25 5 
