THE OBSERVATIONS. 



17 



out the influence of conditions on one another, 

 and the difficulty of measuring the detrital 

 loads of streams might be dwelt upon, but 

 enough has been said to warrant the statement 

 that an adequate analysis with quantitative 

 relations can not be achieved by the mere obser- 

 vation of streams in their natural condition. 

 It is necessary to supplement such observation 

 by experiments in which the conditions are 

 definitely controlled. 



OUTLINE OF COURSE OF EXPERIMENTATION. 



In the work of the Berkeley laboratory 

 capacity for hydraulic traction was compared 

 with discharge, with slope, depth, and width 

 of current, and with fineness of debris; and 

 minor attention was given to velocity and to 

 curvature of channel. For the principal ex- 

 periments a straight trough was used, the sides 

 being vertical and parallel, the ends open, the 

 bottom plane and horizontal. Through this a 

 stream of water was run, the discharge being 

 controlled and measured. Xear the head of 

 the trough sand was dropped into the water at 

 a uniform rate, the sand grains being of approxi- 

 mately uniform size. At the beginning of an 

 experiment the sand accumulated in the trough, 

 being shaped by the current into a deposit with 

 a gentle forward slope. The deposit gradually 

 extended to the outfall end of the trough, and 

 eventually accumulation ceased, the rate at 

 which sand escaped at the outfall having be- 

 come equal to the rate at which it was fed 

 above. The slope was thus automatically 

 adjusted and became just sufficient to enable 

 the particular discharge to transport the par- 

 ticular quantity of the particular kind of sand. 

 The slope was then measured. Measurement 

 was made also of the depth of the current ; and 

 the mean velocity was computed from the dis- 

 charge, width, and depth. 



In a second experiment, with the same dis- 

 charge, the sand was fed to the current at a 

 different rate, and the resulting slope and 

 depth were different. By a series of such ex- 

 periments was developed a law of relation 

 between the quantity of sand carried, or the 

 load, and the slope necessary to carry it, this 

 law pertaining to the particular discharge and 

 the particular grade of sand. The same ex- 

 periments showed also the relations of the 

 velocity of the current to slope and load. 

 20921 Xo. 80 14 2 



Another series of experiments, employing a 

 greater or a less discharge, gave a parallel set 

 of relations between slope, load, and velocity. 

 By multiplying such series the relations be- 

 tween discharge and slope, discharge and 

 load, and discharge and velocity were de- 

 veloped. 



Then a third condition was varied, the 

 width of channel; and finally the remaining 

 condition under control, the size of the sand 

 grains. Thus data were obtained for studying 

 the quantitative relations between load, slope, 

 discharge, width, and fineness, as well as the 

 relations of depth and mean velocity to all 

 others. In all, the range of conditions in- 

 cluded six discharges, six widths of channel, 

 and eight grades of sand and gravel, but not 

 all the possible combinations of these were 

 made. The actual number of combinations 

 was 130, and under each of these were a series 

 of measurements of load, slope, and depth. 

 There were also limited series of experiments 

 involving a greater number of discharges and 

 a greater number of widths. The separate 

 determinations of load and slope numbered 

 nearly 1,200, and those of depth about 900. 



SCOPE OF EXPERIMENTS. 



Before proceeding to a fuller description of 

 apparatus and experiments, let us consider to 

 what extent the conditions of the laboratory 

 were representative of the conditions which 

 exist in the natural stream. 



The sand used came from the beds of Ameri- 

 can and Sacramento rivers and was assumed to 

 be representative of river sand in general. 

 No attention was paid to the influence on 

 traction of the form and density of grains. 

 Each sample used was separated from the 

 natural mixture by means of two sieves and 

 was composed of grains which passed through 

 a certain mesh and were arrested by a mesh 

 slightly smaller. In the sand carried by a 

 river near its bed the range of size is much 

 wider. The limit of coarseness is found in 

 those particles which the current is barely 

 able to roll, the limit in fineness in those par- 

 ticles which the swirls of the current are not 

 quite able to lift into suspension; and the 

 limits vary from point to point of the channel 

 bed. This difference in condition was not 

 wholly ignored, but a short supplementary 



