252 



TRANSPORTATION OF DEBRIS BY RUNNING WATER. 



the comparator were closed, and the heights 

 of the columns were then read. 



The use of the gage to measure velocities 

 close to the bed of debris proved impracticable 

 because the presence of the receiver modified 

 the movement of the water and thereby modi- 

 fied the shape of the bed. (See pp. 26, 155.) 

 This effect could have been reduced by using a 

 different form of receiver. Darcy and Bazin * 

 bent the tubes at the bottom in such a way 

 that one or both openings met the water at 

 some distance upstream from the vertical part 

 of the tubes, and it is probable that the adop- 

 tion of their design would have diminished the 

 difficulty, although it could not have removed 

 it. With such a design, however, the practi- 

 cable forms for the second opening relate it to 

 the piezometer, and the advantage of the down- 

 stream opening is lost. 



That part of the design of the comparator 

 which consists in the inclination of tubes and 

 scale is not to be recommended. It refines by 

 magnifying the reading, but it introduces pos- 

 sibilities of error in other ways. If the tubes 

 are not straight or are not equally inclined, an 

 error is occasioned which does not enter if they 

 and the scale are vertical. Evidence of such 

 error was found in the fact that the still-water 

 or zero-velocity readings of the two columns 

 were not always identical, but no ready means 

 of correction was discovered. 



Another source of error was detected in ine- 

 qualities of sectional area of the glass tubes of 

 the comparator. To show the nature of this 

 error, let us assume that the pressure of the 

 current at the upstream opening of the receiver 

 is exactly equal to the negative pressure, or 

 suction, at the downstream opening. If the 

 glass tubes are of uniform and equal bore, one 

 column moves upward just as much as the other 

 moves downward, and the volume of air above 

 the columns is unchanged. Now, assume that 

 the tube containing the rising column has the 

 greater diameter. It is evident that equal 

 movement of the two columns will displace 

 more air in the one tube than it will provide 

 space for in the other, and the pressure of the 

 confined air will be thereby increased. The 

 effect of the increased pressure will be to lower 



i Bazin, F. A., Recherches experimentales sur 1'ecoulement de 1'eau 

 dans les canaux decouverts: Acad. sci. Paris Mem. math, et phys., 

 vol. 19, p. 49, PI. IV, 18fi5. This memoir was published also as part of 

 "Recherches hydrauliques," by H. Darcy and F. A. Bazin. 



both columns. There is also a secondary effect 

 of small amount connected with the fact that 

 the pressure of the confined air plus the head 

 of water between the tops of the columns and 

 the surface of the stream, on the one hand, and 

 the atmospheric pressure, on the other, are in 

 equilibrium, but into this we need not here 

 enter. 



As a means for the discussion of these errors 

 the tubes weie calibrated, by Prof. J. N. Le 

 Conte, in the following manner: The tubes 

 being closed at the bottom, a weighed quantity 

 of water was introduced into one and the height 

 of its column was read. By lepeated additions 

 of water and repeated readings the volumes of 

 divisions of each tube were thus measured, the 

 divisions being approximately 1.5 inches in 

 length. In similar manner the volume was 

 measured of the space above the straight tubes 

 to the pet cock. The average sectional area of 

 one tube was found to be 2.5 per cent greater 

 than that of the other. The sectional area in 

 the larger tube was found to vary through a 

 range of at least 2.6 per cent, and the range for 

 the smaller tube was 4.5 per cent. 



A table of corrections to readings was 

 computed from the data of calibration, and 

 this table was practically applied. In com- 

 paring the rise of one column with the asso- 

 ciated fall of the other the greatest correction 

 applied amounted to a little less than 1 per 

 cent. In the determination of velocity the 

 largest correction applicable for this reason was 

 about 0.3 per cent. These corrections, how- 

 ever, pertain only to the discussions of the in- 

 strument in the following pages. In the 

 ordinary work of the gage they were not ap- 

 plied, because it was found that the errors of 

 this class were practically eliminated when the 

 same methods were employed in the prepara- 

 tion and in the use of rating formulas. The 

 matter is mentioned here chiefly because errors 

 from unequal tube caliber, which may some- 

 times prove important, appear not to have 

 been allowed for in the discussions of Pitot- 

 Darcy gages. 



BATING METHODS. 



The first and second gages were rated by the 

 method of floats; the third was twice rated by 

 the running-water method and several times by 

 the still-water method. The floats used weie 



