THE FLOW OF WATER IN WOOD-STAVE PIPE. 83 
Lowell Hydraulic Experiments, and Clemens Herschel on Venturi Water Meter, pub- 
lished in Trans. Amer. Soc. Civ. Engin., vol. 26, p. 452). This form of entrance and 
exit also has the advantage of furnishing an accurate and convenient means of meas- 
uring the quantity of flow and of keeping track of the gauge height which will not be 
affected by the condition of the ditch. 
Page 5, line 15. In small pipe, from 4 to 12 inches, used for irrigation and water- 
works purposes, valves are always used at least in one place, and often a large number 
of fittings and valves whose effect on the flow is similar to the ordinary gate valve. 
In such tests as the writer has made he has found no appreciable loss in any single 
valve or fitting, but where there are many, as in waterworks systems, the loss on this 
account is a matter for serious consideration. There does not seem to be much infor- 
mation published that would throw light on this subject. 
Page 3, line 5, and page 48, line 35. Attention is directed to the fact that the 
writer in 1909 first devised this type of formula from suggestions of Messrs. Saph and 
Schoder. This formula is as follows [Formula 16, p. 48]: 
Q=1.28D 2 - 58 H. - 585 
Since that date he has been using this in all his calculations as to the flow in wood 
pipe. It is in the same form as the one devised by Mr. Moritz and the author. This 
formula gives results from to 15 per cent less than the author's formula, being about 
the same for smaller sizes of pipe and 15 per cent less discharge for the larger sizes of 
pipe. It was the writer's intention in devising this formula to so select the coeffi- 
cients of D and H that the calculated flow would more nearly approach the quantities 
of flow determined from tests that were lowest instead of those that were the average. 
Page 15, line 27. It is not at all impossible that in a number of the tests on which 
the formula is based the real average diameter is different from the nominal diameter 
assumed, due to the following causes: 
(a) Swelling of the wood by saturation. 
(6) Distortion due to imperfect backfilling or settlement where the pipe may have 
been laid on more or less of a fill, or where there has been more or less leakage, 
(c) Inaccuracy of manufacture. The writer in his examination of a 54-inch pipe 
found an average of one-half inch larger than the nominal diameter, making 
a difference of 1.8 per cent increase in the area. This pipe was measured every 
100 feet throughout its entire length, from one manometer to the other. In 
making these measurements considerable distortion was found to exist, and the 
writer is not certain that these measurements revealed the exact diameter. 
Page 46, line 42. It is the experience of most engineers who have had much to do 
with current-meter measurements that they can not be depended upon to give 
satisfactory results where the water is at all turbulent or where the cross section of any 
stretch of the channel is uneven, thus causing considerable turbulence. The greatest 
care in rating a meter will not help this very large source of error. The error due to 
turbulence is greater with the Price meter than with the Haskell meter, which latter 
was used in the writer's experiments on the 54-inch and 44-inch pipe-line tests. The 
measurements were taken by inserting the meter into the exit end of the 54-inch 
pipe, being held in exact position by a templet and a pin fastened through the upper 
end of the meter rod. 
Page 47, line 37. The line of maximum velocity within the pipe would be shifted 
from the center of the pipe to a line close to that portion of the outside of the pipe on 
which the convex side of the curve occurs. This line of maximum velocity would 
retain its position for a long distance from the curve and would occur within the 
length of the pipe tested. 
The writer doubts if this abnormal condition would affect the results, particularly 
if the manometers were attached in the neutral zone of velocity, as the author states 
was done. 
