246 



TRANSPOKTATION OF DEBEIS BY KUNNING WATEE. 



substitution throws the maximum to the sur- 

 face of the water and makes a consistent series 

 of the three curves. 



In another series of observations the trough 

 remained horizontal while the discharge was 



.923 



Velocity 



FIGURE 79. Modification of vertical velocity curve when mean velocity 

 is increased by change of discharge. The corresponding discharges 

 are indicated in ft.'/sec. 



varied. The width was 0.66 foot, and the dis- 

 charges 0.182, 0.363, 0.545, 0.734, and 0.923 

 ft. 3 /sec. The resulting curves (fig. 79), with 

 exception of that for the smallest discharge, 

 form a consistent series, the level of maximum 

 velocity rising slowly with increase of discharge. 

 In this case also the discordant curve may be 

 brought into harmony by the rejection of a 



3=5J 



01234 

 Velocity 



1'iGUKE 80. Modification of vertical velocity curve by changes in the 

 roughness of the channel bed. Curve 1, paraffin; 2, roughness of debris 

 grade (A); 3, roughness of grade (D); 4, roughness of grade (F). 



single observation, and a substitute curve is 

 suggested in the diagram by a broken line. 1 



In another series the texture of the channel 

 bed was made to vary. Constant features were 

 the discharge, 0.734 ft. 3 /sec.; the width, 0.92 

 foot; the slope, 0.58 per cent; and the texture 

 of the channel sides, which were planed and 

 unpainted. The varieties of bed texture were 

 (1) paraffin, coating a smooth board; (2) a 



' In this instance, in the one before mentioned, and also in the case of 

 three curves in figure 80 the aberrant point records a measurement which 

 was made very near the surface of the water. In such positions the 

 constant of the Pitot-Darcy gage has a special value, and it is on the 

 whole probable that the apparent errors of observation are occasioned 

 by an error of the rating formula. The matter is discussed in 

 Appendix A. 



pavement of debris of grade (A); and (3 and 4) 

 similar pavements with grades (D) and (F). 

 The observed curves, given in figure 80, indi- 

 cate that the resistance occasioned by a rough 

 bed retards the whole current but retards the 

 lower parts in greater degree than the upper, 

 so that the level of maximum velocity is raised. 

 The amount of retardation is greater as the 

 texture of the bed is coarser. 



In figure 81 are two curves illustrating the 

 influence of load on the vertical distribution of 

 velocity. One curve was observed in a stream 

 without load, flowing over a smooth and hori- 

 zontal bed, the other in the same stream when 

 carrying a load of 17 gm./sec. on a self-adjusted 



.3 



!, 



Q 



Velocity 



FIGURE 81. Modification of ve r tical velocity curve by addition of load 

 to stream, with corresponding increase of slope. A, without load; B 

 with load. 



slope of 0.64 per cent. To judge from the data 

 shown in figure 78 (p. 245), the effect of slope 

 alone would be to double or nearly double the 

 mean velocity, but the actual increase was only 

 13 per cent. The acceleration due to slope was 

 almost wholly neutralized by the retardation 

 due to roughness of bed and to the work of 

 traction. The retardation had also the effect 

 of raising the level of maximum velocity. 



FIGUBE 82. Modification of vertical velocity curve by addition and 

 progressive increase of load. 



Similar contrasts are shown in figure 82, 

 where the curve for an unloaded stream is com- 

 pared with curves for the same stream when 

 transporting three different loads. The con- 

 stant factors in this case are: Width of channel, 



