THE FLOW OP WATER IN WOOD-STAVE PIPE. 45 



mean velocity of four or five batches of color being accepted as the mean velocity of 

 the water within the pipe for that particular run. Many of these color tests were 

 checked by current meter measm-ements made by the two-tenths and eight-tenths 

 depth method at the meter station shown in Plate III, figure 3. This station is in 

 the concrete flume about 70 feet below the outlet from the siphon. The agreement 

 between the two methods is shown in Table 1. As this pipe had been in use but a 

 few months the interior was probably in excellent condition. The maximum dis- 

 charge of the pipe necessitates a mean velocity of about 5.5 feet per second, so that it 

 is probably scoured quite clean and smooth at all times. The capacity of the pipe 

 was about 15 per cent greater than that computed by the new formula. 



No. 50, Expt. S-16, 78-inch Continuous- Stave Douglas Fir Pipe,* Mohawk 

 Hydro-Electric Co., Ephratah, N. T. — The power house of the Mohawk Hydro- 

 Electric Co., near Ephratah, is supplied with water by a trunk line of about 2J miles 

 of 78-inch stave pipe from the reservoir, Peck I^ake, to the surge tank. (PL IV, fig. 3.) 

 From the tank a stave pipe 96 inches in diameter extends to a point 1,460 feet distant, 

 where the pressure head is 160 feet. It here joins a steel pipe of the same diameter, 

 which completes the additional distance of a few hundred feet to the tiu^bines. The 

 wiiter conducted a series of tests on a reach of the 78-inch pipe 2,650 feet long. The 

 lower end of this reach was about 1,000 feet above the surge tank. The whole line 

 abounds in gentle curves, both horizontal and vertical. The pipe, built in 1910, was 5 

 years old at time of test. It is full of water throughout the year and is not protected 

 against freezing, being so placed that some portions are completely buried and some 

 completely exposed. Although extremely cold weather is experienced in this part 

 of New York, the wood appears to furnish suflicient insulation. The peak load 

 demands a velocity in this 78-inch pipe of less than 7 feet per second. This velocity 

 and the fact that water comes from a resei'voir that should act as a settling basin prob- 

 ably guarantees a pipe free from sediment. Several minor leaks were found on the 

 reach tested. These are mostly at ends of staves where no additional bands were 

 placed, and the pressure has bent outward the end of the stave farthest from the sup- 

 port of a band, the bend, of course, occurring under the last band. Whether the 

 elastic limit of the wood had been exceeded and the fiber torn could not be ascer- 

 tained, but the condition was such as to emphasize the desirability of confining all 

 the joints in a stave pipe to a zone a few feet in length and placing extra bands through- 

 out this zone. This of course does not apply to pipes under light pressiu:es, say, 

 30 or 40 foot heads. Velocities within the pipes were determined directly with 

 fluorescein for observations 1, 2, 3, 4, 7, 8, 9, and indirectly by comparison with the 

 rating curve of the concrete channel forming the taikace, for observations 5 and 6. 

 The tailrace was calibrated by means of six careful current-meter gaugings, and a 

 rating curve was plotted showing the comparison between gauge heights in the tail- 

 race and velocities in the 78-inch pipe. The comparison between color tests and 

 meter tests is shown in Table 1. The agreement between the two methods is closer 

 than is usually expected by experienced hydrographers. Mercury manometers were 

 iised at both gauges. Some trouble was experienced from freezing temperatures 

 (tests were made the first week in April, 1915), but no trouble occiu-red from air in 

 the pipe, as the intake is deeply submerged. The color was injected at gauge No. 1 

 and observed at a secondary tap in the pipe 1 foot downstream from gauge No. 2, the 

 water flowing into a white-lined pan which reflected greenish color. The same 

 general procedure was followed here as on the Altmar tests (No. 51). That is, simul- 

 taneous readings were made over a long period of time, on both manometers and on a 

 hook gauge in the taihace. When the records were brought together, periods of slight 

 fluctuation might be selected and each of these called an observation. Some such 

 method as this must be chosen when a power plant in commercial operation is tested, 

 as no one knows just when the changes in load, and consequent changes in velocity 



I Engin. Rec., Vol. 64, No. 22, Nov. 25, 1911, p. 627. 



