200 



TRANSPORTATION OF DEBRIS BY RUNNING WATER. 



proper adjustment of the feed, and when this 

 condition existed the load delivered at the out- 

 fall was assumed to represent the capacity of 

 the stream. The outfall was then directed for 

 a measured interval of time to a reserved 

 division of the settling tank, and the d6bris 

 thus separately received was weighed. 



Besides the deposit connected with the feed- 

 ing of debris there were transitory deposits of 

 a rhythmic character, as described later. 



Five characters of channel bed were used, 

 namely, a planed and painted wood surface; a 

 rough-sawn, unplaned wood surface; a surface 

 of wood blocks, with grain vertical; a pavement 

 of sand grains, set in cement; and a pavement 

 of pebbles. (See PI. III.) 



PROCESSES OF FLUME TRACTION. 

 MOVEMENT OF INDIVIDUAL PARTICLES. 



Flume traction differs from stream traction 

 in its extensive substitution of rolling for sal- 

 tation and in the important place it gives to 

 sliding. The relative importance of these 

 modes of particle movement is determined (1) 

 by the texture of the bed surface in relation to 

 the size of the particle, (2) by the velocity of 

 the water in relation to the size of the particle, 

 and (3) by the shape of the particle. 



In stream traction the order of roughness of 

 the bed is given by the fineness or coarseness of 

 the material of the load, and this fact deter- 

 mines saltation as the dominant process. In 

 flume traction the bed may be much smoother. 

 On a smooth stream bed any particle with a 

 broad facet is apt to slide, and a well-rounded 

 particle to roll. Rolling is determined (rather 

 than sliding) not only by the fact that the pro- 

 pulsive force of the current and the resistance 

 given by the bed constitute a couple, but also 

 by the fact that the current applies a greater 

 force to the upper part of the particle than to 

 the lower. The less smooth the bed surface, 

 the greater its resistance and the more effective 

 the couple in causing the particle to roll. With 

 any particular texture of bed, the sizes of par- 

 ticles may determine their modes of progress 

 the largest sliding, those of smaller size rolling, 

 and the smallest leaping. Increase of velocity 

 tends to increase saltation at the expense of 

 rolling and to increase rolling at the expense 

 of sliding. 



A particle rolled slowly is in continuous con- 

 tact with the bed. A round particle may roll 



rapidly on a smooth bed without parting from 

 it. Roughness of the bed causes changes of 

 direction in the vertical plane, and such changes 

 combined with high velocity cause leaps. If 

 the particle is not round its rolling involves 

 rise and fall of the center of mass, and such 

 changes combined with high velocity cause 

 leaps. Shape of particle may thus be a deter- 

 minant between saltation and rolling, as well 

 as between rolling and sliding. 



A flattish particle, which may either slide or 

 roll, travels faster when rolling. This is due 

 partly to the fact that when it rolls it rolls on 

 edge and thus projects farther into the current, 

 and partly to the fact that the resistance at 

 contact with the bed is greater for sliding than 

 for rolling. It is also true that rolling par- 

 ticles as a class outstrip sliding particles as a 

 class, the difference in speed being marked. 



For particles of similar size those which domi- 

 nantly roll outstrip those which dominantly 

 leap. This is part of a more general fact that 

 the better-rounded particles travel faster than 

 the more angular. In the traction of mixed 

 debris, where rolling is characteristic of larger 

 particles and saltation of smaller, the larger 

 travel faster than the smaller. There are thus 

 two important ways in which rolling gives 

 greater speed than saltation. It was not 

 learned whether a particle which alternately 

 rolls and leaps travels faster in one way than 

 in the other. 



A suspended particle, having the same speed 

 as the water, outstrips all others. It is there- 

 fore possible that as saltation approaches the 

 borderland of suspension its speed exceeds that 

 of rolling. 



When samples of different grades are fed tc 

 the same current in succession, it is found that 

 the coarser travel the faster, whatever the mode 

 of progression (except suspension). In the fol- 

 lowing record of experiments the speeds of 

 grade (J) constitute an apparent exception, but 

 their slowness is ascribed to the fact that the 

 particles of that grade were relatively angular. 



TABLE 66. Relative speeds of coarse and fine debris in flume 

 traction. 



