FLUME TRACTION. 



213 



occurs among the data for traction over a gravel 

 pavement it illustrates a feature of stream trac- 

 tion rather than flume traction. 



The two principles may be illustrated to- 

 gether by saying that if a stream is carrying its 

 full load of a grade narrowly limited in range 

 of fineness, and a different grade of debris is 

 added, the total load is thereby increased, but 

 this increase is accompanied by a diminution 

 of the quantity carried of the first-mentioned 

 grade. In contrast with this is the law found 

 for stream traction that the load of the ini- 

 tially transported grade is increased by the 

 moderate addition of other debris, provided 

 the added d6bris is relatively fine. 



The difference between the two cases is 

 thought to be connected with rolling. In 

 flume traction over a smooth bed the path for 

 rolling particles is roughened by the presence 

 of smaller particles. In stream traction the 

 pathway for larger particles is smoothed by the 

 presence of smaller particles and rolling is 

 promoted. 



In stream traction the capacity for a mixture 

 is determined chiefly by the capacity for its 

 finer components, and as mean fineness also 

 depends chiefly on the fineness of the finer 

 componejits, mean fineness is a serviceable 

 gage of capacity. In flume traction the rela- 

 tion is quite different. Because of the double 

 ascent of the curve of capacity and fineness, 

 it may readily occur that the capacity for a 

 mixture is most nearly related to that for the 

 coarsest component in fact, that is true of 

 the three mixtures tested in our experiments 

 and when that is the case there is no parallelism 

 between capacity and mean fineness. 



CAPACITY AND FORM RATIO. 



The data bearing on the relations of capacity 

 to the depth and width of current, and their 

 ratio, are meager. Most of the experiments 

 were conducted with a single trough width, 1 

 foot. The only other width used was 1.91 

 feet, and its use was associated with but four 

 grades of debris and a single character of 

 channel bed the smoothest. 



Depths of current were not in general meas- 

 ured during the passage of loads, because the 

 surfaces of load-bearing currents were usually 

 so rough as to make good determinations im- 

 possible. Good measurements were made of 

 unloaded streams, and the results are here 

 tabulated. Attempts to measure depths of 

 loaded streams yielded one result thought 

 worthy of record. With a discharge of 0.734 

 ft. 3 /sec., a width of 1 foot, a slope of 4 per cent, 

 and a full load of debris of grade (E 3 IT 2 I 3 J 2 ), 

 the depth was 9 per cent greater than for the 

 corresponding unloaded stream. 



Table 76 compares the capacities found for a 

 trough width of 1.91 feet with corresponding 

 capacities for a width of 1 foot. By aid of 

 Table 75 it brings capacities into relation also 

 with depths. 



TABLE 75. Depths and form ratios of unloaded streams, in 

 troughs of wood, planed and painted. 



TABLE 76. Capacities for flume traction in troughs of different wullhs. 



