The reading of the gauge board above the dam cannot be considered very accurate, as the smallest subdivision was 3 ins. It was taken to show by the observed entry head (Column 14) what error, if any, there might be in the heights of Piezometer No. 1.

This would indicate that the error in the height of Piezometer No. 1 was very slight, if any; probably less than could be observed on the gauge board, as a part of this resistance must have been due to resistance at the entrance.

Column 19 shows the sum of Columns 17 and 18. Column 20 shows the quotient of Column 19 divided by the length given in the heading and multiplied by 1 000. Columns 21 to 24, inclusive, give the average velocity in each of the three rings and the center of the opening at the well. Column 25 gives the average velocity through the average section of the pipe, which is always larger than the average velocity at the opening, as the opening where the current-meter measurements were taken is larger than the average section of the pipe.

Columns 26, 27, 28 and 29 show the discharge, in cubic feet per second, of the respective rings of the opening, and Column 30 the total discharge, which is the sum of Columns 26, 27, 28 and 29.

Table No. 2 shows the principal data from the 44-in. pipe experiments. In these experiments there were only two sections of pipe: Section 4, from the well to Piezometer No. 2, and Section 5, from Piezometer No. 2 to Piezometer No. 3.

In Observations 1 to 7, inclusive, the sluice-gate at the entrance to the 44-in. pipe was open only 9 ins., causing the excessive resistance shown in Column 13.

During Observations 8 to 11, inclusive, the sluice-gate was wide open, showing the resistance at the entrance to be very slight, as this also includes the resistance in the basin and in the conduit from the well to the basin.

The current-meter measurements were made at the same place and under the same system as for the 54-in. pipe, except that from the total discharge is taken the waste water from the settling basin.

Friction Head.

Columns 27 and 28 give the final results of velocity and loss of

head.

Reliability of Results.-In conducting these experiments, the writer has endeavored to eliminate, so far as circumstances would allow, all sources of error, and confine the probable error within the limit that this method of measurement ought to give. He has, in some cases, perhaps, gone to more pains than necessary in order that the reader

0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0 34 0.36 0.38

Velocity Head. H.
FIG. 10.

may understand clearly the conditions under which the experiments were conducted. If there has been any considerable error for which he has not allowed, it is to be hoped that it will be brought out in the discussion. The current-meter measurements are undoubtedly within 2% of the correct value, and the observations for loss of head are within 1 per cent. The size of the pipe is obtained from so many

separate measurements, 48 in the 54-in. pipe and 100 in the 44-in. pipe, that the probable error in the average size of the pipe must be very small.

Care was taken to guard against the accumulation of air in the connections between the piezometer tubes and the main pipe. That there was no considerable error in this respect is evident from the conformity of the individual observations to one line (see Fig. 10), with the possible exception of Observation 11 on the 54-in. pipe.

All observations of velocity are within 1% of this line, except Observation 11, which is within 2 per cent.

The very slight difference between the calculated and observed velocity head at the entrance of the 54 and 44-in. pipes proves the accuracy of the piezometer heights at Piezometers Nos. 1 and 2. The height in the well is checked by the height at Piezometer A.

Comparison with Other Experiments.—So far as the writer has been able to learn, all the available data regarding the loss of head in wood pipes are as follows:

1.-By James D. Schuyler, M. Am. Soc. C. E., on a 30-in. pipe line at Denver, Colo., the details of which are not available. The value of n, in Kutter's formula, is given as 0.0096.

2.-By Mr. Fred B. Gutillius, on the 24-in. pipe of the Butte City Water Company, in which he confirms the use of 0.01 as the value of n, in Kutter's formula.

Only one

3.-By Arthur L. Adams, M. Am. Soc. C. E., on an 18-in. pipe for the City of Astoria, Ore. The details of this experiment are given. The results show a value of n of 0.01, and c = 132.9. observation was taken at a velocity of 3.605 ft. per second. 4.-By Arthur L. Adams, M. Am. Soc. C. E., on a 14-in. pipe for the West Los Angeles Water Company. In all, there were seven observations, the velocity varying from 0.691 to 1.531 ft. per second; value of c, from 99 to 113; value of n, about 0.011. 5.-By Messrs. Marx, Wing and Hoskins, on the 72-in. pipe of the Pioneer Power Plant, at Ogden, Utah. This is by far the most complete and satisfactory set of experiments yet conducted to determine the loss of head in wood pipe. The results have been fully written about and discussed.*

6. By the writer, on 44-in. pipe, as described in this paper. 7.—By the writer, on 54-in. pipe, as described in this paper. * Transactions, Am. Soc. C. E., Vols. xl and xliv.

In Experiments Nos. 1 and 2, of the foregoing list, no detailed information is available as to the circumstances and conditions under which the experiments were conducted, and in No. 3 no range of velocities was experimented upon.

The only data (so far as the engineering public is aware) which can be taken as criteria for the loss of head in wood pipe are those contained in Experiments Nos. 4, 5, 6 and 7, on 14-in., 44-in., 54-in. and 72-in. pipes, respectively. The range of velocities experimented upon in the 14-in. and 44-in. pipes is not as much as would be desired, and there is a wide gap from the 14-in. to the 44-in. pipe, and from the 54-in. to the 72-in. pipe. It cannot be said, so far as this information goes, that there is any very close conformity in the results, except in the two larger sizes, which agree fairly well; but it is difficult to see wherein they conform with the results on the 44-in. pipe; at least, no general law can be formulated that would seem to apply to all sizes, without the use of a coefficient having arbitrary values for the different sizes of pipe.

Fig. 11, showing the relation between velocity head and friction head, includes only the 1897 experiments made on the 72-in. wood pipe at Ogden. To avoid unnecessary complication, the other two experiments are not given. This experiment was chosen because at that time the age of the pipe was more nearly comparable with the age of the other pipes experimented upon.

It will be seen from Fig. 10 that in all these experiments the velocity head, as compared with the friction head, conforms very closely to a straight line. The evidence is very strong that a formula based upon this hypothesis is correct, as it most nearly fits the experiments so far conducted. Starting with this assumption, the equations for the different sizes of pipe become:

On the diagram, Fig. 11, the curves for these several equations (1, 2, 3 and 4), of the relation between V and H, have been plotted. The curves, of course, show the same conformity to the individual observations as in Fig. 10, and will give correct results for the individual sizes of pipe, within the range of the experiments on which they are based, and within the limit of accuracy of these experiments.