Ch. 3] BED-MATERIAL LOAD 67 



as surface creep in a very narrow layer above the bed, whereas all 

 the bed material moving in higher layers of flow is interpreted as 

 suspended load. 



The existing suspended load theory (Vanoni, 1946) permits calcu- 

 lation of the sediment concentration at any point throughout the 

 depth of a given flow, provided that the concentration at only one 

 point in the depth is known. It has been found that the surface creep 

 in the narrow layer which it occupies near the bed defines such a 

 reference concentration. Conditions are encountered, especially in 

 large streams with comparatively fine bed sediment, where the sedi- 

 ment goes into suspension easily; hence the rate of suspended-load 

 transportation is many times larger than that of surface creep of the 

 same bed particles. In these cases the material moved by surface 

 creep may be negligible in amount in itself, but it still retains its 

 large significance as a valve or trigger controlling all suspended-load 

 concentrations. 



All bed-load equations that are in existence have been derived from 

 flume measurements. Experimental flumes that have been used for 

 this purpose have ranged from a few inches to 7 feet in width and 

 depth and from 5 to 200 feet in length. Most equations were derived 

 originally from experiments with uniform sediment and then checked 

 for applicability to sediment mixtures. Most mixtures for which the 

 range of particle sizes is not excessive and which do not contain a 

 large percentage of particles finer than 0.2 millimeter have been found 

 to move as a unit. This finding indicates that the mechanical com- 

 position of the bed sediment is about the same as that of the material 

 being transported. If very small rates of transport are included in 

 the experiment, or if one of the two limitations given above is trans- 

 gressed, a considerable segregation of the particles becomes apparent, 

 and an overall appraisal of the total transport by means of a repre- 

 sentative grain diameter is impossible. 



Thus it appears to be possible that the normal bed-load equations 

 can be applied to the different components of a bed if the basic as- 

 sumption is made that every particle moves according to a law that 

 includes its own availability, its ability to move, and the capacity of 

 the flow to move it. It appears that this law does not depend to a 

 large degree on the presence of other sediment particles. This same 

 assumption of non-interference between sediment particles in motion 

 is also one of the basic assumptions of the existing suspended-load 

 theory (Vanoni, 1946). The composition of the bed, however, deter- 

 mines the bed roughness and, indirectly, the flow pattern. Publica- 



