Reviews — Transport by Running Water. 133 



laboratory cannot be made to conform with the meanderings, variable 

 discbarge, and forms of cross-section of natural alluvial streams, and 

 the author is obliged to admit that the primary purpose of the 

 investigation, which was to determine for rivers the relation of the 

 load swept along the bed to the more important controlling factors, 

 was not accomplished. Although the gap between the river and the 

 laboratory has not yet been bridged, much valuable information was 

 obtained which is directly applicable to hydraulic transport and river 

 engineering. 



Streams carry debris in various ways which can be grouped 

 together under the headings of hydraulic suspension and traction, the 

 latter referring to what is sometimes called the bottom load. Each 

 of these two modes of progression may be further divided according 

 to the nature of the bottom. In natural streams the bed is usually 

 composed of debris which is similar to the material of the load. In 

 artificial channels, like flumes and pipes, the bed is rigid. The 

 Berkeley experiments were designed specifically to investigate stream 

 traction and flume traction, since there is a marked contrast between 

 the laws controlling each case. Generally, the measurement of the 

 suspended load in a river is only a matter of routine and patience, 

 but the measurement of the tractional or bottom load is difficult. 

 Knowledge of the average ratio between the suspended and the 

 tractional load would be of great value to geologists and engineers. 

 Humphreys and Abbot estimated the tractional load in the case of 

 the Mississippi to be 11 per cent of the suspended load. Guerard in 

 a study of the Rhone found the suspended load to be less than 

 25 per cent of the whole. Gilbert took measurements of the two 

 loads in the Yuba River and came to the conclusion that they were 

 approximately equal. These data are probably the best on record, 

 but their wide variability is a serious reflection on the value of 

 denudational studies in relation to the wearing down of the lands 

 and to the measurement of geological time. 



The factors which control the movement of the traction load are 

 numerous and complex, for the law of control in the case of each 

 factor is qualified by all the other factors. The slope and discharge 

 of the stream, the form of its channel, the fineness of debris, and the 

 velocity are all controlling factors. Of these factors most attention 

 has been devoted to velocity, and it has been generally assumed, 

 though erroneously, that the load carried varies with the sixth power 

 of the velocity. This deductive law strictly refers only to the 

 maximum size of the grains or pebbles which a given current is 

 competent to move. The actual law is too complex for analysis and 

 generalization. In the Berkeley experiments an attempt was made to 

 compare capacity for traction load with the mean velocity V m of the 

 stream. The following empirical average, results were obtained: 

 (1) slope constant; V m varies with discharge ; capacity oc V m 3 ' 2 . (2) 

 discharge constant ; V m varies with slope ; capacity oc F m 4 . (3) depth 

 constant ; V m varies with slope and discharge ; capacity oc F~ TO 3 ' 7 . 

 In each case the size of the debris carried provides a further interfering 

 factor. If fine material be added to coarse the total load is increased, 

 and even the load of coarse material is greater than before. 



