ADJUSTMENT OF OBSERVATIONS. 



57 



outfall conditions extended indefinitely up- 

 stream. An expression of this influence was 

 found in the water profile, which was usually 

 convex in the lower part of the trough, the 

 degree of convexity diminishing upstream. 

 Its effect on the profile of the bed is not readily 

 analyzed, because that profile is adjusted 

 through the velocity of water at the bottom of 

 the current, and the bed velocity is not simply 

 related either to mean velocity or to depth. 



FIGURE 15. Diagrammatic longitudinal section of outfall end of experi- 

 ment trough, illustrating influence of sand arrester on water slope. 



A second factor at the outfall end was the 

 arrangement for separating the debris from the 

 current. This included a well, ABCD, figure 

 15, which was sunk below the trough bed and 

 into which the debris sank, while the current 

 passed on to the outfall at E. In part of the 

 work the space AD was entirely open; in 

 another part a coarse screen was stretched 

 across it. In either case the resistance of this 

 part of the channel bed differed from the 



D 



resistance along the debris slope and may have 

 been greater or less. From the well to the 

 outfall, DE, the frictional resistance was loss 

 than elsewhere. As the fixed part of the 

 channel bed, DE, was horizontal and the 

 debris portion, GA, was inclined, the profile 

 of the bed changed at A. Projected forward, 

 the slope GA passed below E to H, and when 

 the debris slope, down which the transporting 

 current flowed, was steep, the part DE was 

 related to it somewhat as a dam. The tend- 

 ency of the quasi-darn was to retard the cur- 

 rent near the outfall and make the water 

 profile concave, and in some of the experiments 

 the profile actually became concave. Other 

 outfall factors were recognized, but they are 

 not here mentioned because they are believed 

 to be of relatively small importance. 



In the reduction of observations on slope an 

 attempt was made to lessen the effect of intake 

 and outfall influences by omitting from the 

 calculations the profile data obtained near the 

 ends of the trough. The data from a con- 

 siderable number of experiments were finally 

 discarded altogether and do not appear in the 

 tables. To replace the discarded data experi- 

 ments were afterward made with a modified 

 apparatus. 



FIGUKE 16. Diagrammatic longitudinal section of d<!bris bed and stream, in a long trough. 



CHANGES IN APPARATUS. 



As the terminal influences of all kinds 

 diminish with distance from the trough ends, a 

 manifest mode of avoidance is to employ a 

 very long experimental trough and determine 

 slopes from observations in the medial portion 

 exclusively. A trough length of 150 feet was 

 tried and proved moderately successful for 

 very low slopes. For steeper slopes there were 

 practical difficulties the need either of a 

 trough with adjustable slope or of a very deep 

 trough and a correspondingly large stock of 

 assorted debris and these led to the consid- 

 eration of other plans. The one adopted was 

 to contract the channel at the outfall and thus 

 increase the resistance to flow at that point, 

 and with the contracted outfall to use a trough 

 of moderate length. The theory of this plan 

 may be illustrated by a diagram. In figure 

 16, AB represents in profile the bed of a long 



trough, CB the profile of debris, and DB' the 

 water profile. In the tract EF the water and 

 debris profiles are nearly parallel and depth and 

 velocities are therefore practically uniform. 

 From F to B' the water profile is notably con- 

 vex because the resistance to flow afforded by 

 the water itself steadily diminishes toward B'. 



FIGURE 17. Diagrammatic longitudinal section of outfall end of 

 trough, illustrating influence of contractor. 



The plan undertook to introduce at F, by con- 

 traction, a resistance equivalent to that 

 afforded by the water beyond F and then dis- 

 pense with the portion of the trough between 

 F and B. In the longitudinal section, figure 

 17, ADB is the bed of the trough, with the well 

 for catching debris, as already shown in figure 



