FLUME TRACTION. 



207 



Next in order are two varieties of unplaned 

 wooden surface; the first being that of boards 

 or planks, paiallel to the current, retaining the 

 roughness left by the saw; the second a pave- 

 ment made by sawing planks of Oregon pine 

 into short equal blocks and setting them on 

 edge. Both these surfaces, as well as those 

 described below, are illustrated in Plate III. 

 These two varieties proved to have approxi- 

 mately the same properties in respect to trac- 

 tion, and the capacities associated with them 

 are 23 per cent less than those for planed 

 lumber. The range in ratio is not large for 

 the different experiments, and the value given 

 may be taken as a constant representing the 

 difference in efficiency between new unplaned 

 and planed wooden flumes. The difference 

 tends, however, to diminish with wear, the 

 unplaned lumber becoming smoother and the 

 planed rougher. 



The next grade of roughness was given by 

 coarse sand d6bris of grade (G) set in 

 cement, so as to constitute a pavement re- 

 sembling sandpaper. The only material run 

 over this was debris of the same grade, the 

 special purpose being to compare flume trac- 

 tion with stream traction the condition of 

 fixed bed with that of mobile bed when the 

 degree of roughness is the same. The experi- 

 ments gave the streams 50 per cent greater 

 capacity when sweeping the debris over the 

 fixed bed than when moving it at the same 

 slope by the method of stream traction. 



The sand pavement gives capacities half as 

 great, on the average, as the surface of planed 

 lumber, but the contrast is stronger for the 

 smaller discharge and lower slope and less 

 marked for the larger discharge and steeper 

 slope. 



The roughest surface used, a pavement of 

 pebbles prepared by setting in cement a mix- 

 ture of grades (H) and (I), gave still lower 

 capacities. These range from 20 to 62 per 

 cent of the corresponding capacities given by 

 planed lumber. The obstructing influence of 

 the rough bottom is most strongly manifested 

 when the material transported has a coarseness 

 corresponding to the texture of the pavement. 

 For finer material its roughness is mitigated by 

 the lodgment of debris, which has the effect of 

 establishing a pavement of the finer material. 



The word "debris" in the table indicates a 

 channel bed composed of loose debris, the 



debris in transit, and the associated process is 

 that of stream traction. The available data 

 afford comparison only for the four finer 

 grades, (B), (C), (E), and (G), the grades 

 which would be designated sand. Each com- 

 parison, with an apparent exception to be 

 considered immediately, shows stream trac- 

 tion to be less efficient than flume traction. 

 When stream traction is compared with flume 

 traction over a smooth surface, the observed 

 ratio of efficiency ranges from 19 to 88 per 

 cent, the smaller ratios being associated with 

 the coarser grades of debris. 



The exception occurs when capacity over a 

 bed of debris is compared with capacity over 

 a pavement of pebbles, the two capacities 

 being found to be the same. The cases which 

 afford this comparison are for grades (E) and 

 (G), and these fine materials, by filling the 

 hollows of the pavement, create a condition of 

 bed in which stream traction dominates. The 

 comparison is really between normal stream 

 traction and stream traction modified by the 

 appearance of crests of fixed pebbles in the 

 channel bed. In harmony with tliis interpreta- 

 tion is the fact that capacities for stream trac- 

 tion and for traction over the gravel pavement, 

 when compared severally with capacity for trac- 

 tion over smooth wood, both show contrasts 

 which increase with coarseness of the load. 



The important general facts brought out by 

 the comparisons are (1) that with a given dis- 

 charge, channel width, and slope, the process 

 of flume traction is able to transport more 

 debris than that of stream traction, and (2) that 

 a stream's capacity for flume traction varies 

 inversely with the roughness of the flume bed. 



The first of these principles serves to explain 

 certain phenomena of clogging. When there 

 is fed to a flume a load greater than its stream 

 is able to transport, a portion is deposited. 

 This changes the character of the bed in such 

 a way as to substitute stream traction for 

 flume traction. Stream traction, being less 

 efficient, can carry still less load, and a larger 

 fraction is deposited. If the conditions re- 

 main unchanged the bed is built up until its 

 slope becomes that necessary to carry the en- 

 tire load by stream traction. Unless the trough 

 is deep or short, overflow results. 



When clogging is initiated by a temporary 

 overloading, the stream loses power to carry its 

 normal fractional load, and deposition con- 



