18 



TRANSPORTATION OF DEBRIS BY RUNNING WATER. 



series of experiments was made with definite 

 mixtures of sand of various sizes, as well as 

 with a natural mixture. 



The straight channel of the laboratory differs 

 materially from the curved channels of nature. 

 It gives comparatively uniform depths and 

 velocities from side to side and from point to 

 point in the direction of the flow, while in a 

 curved channel the depths and velocities vary 

 greatly both across and along the channel. 

 This difference in condition also received some 

 attention, a short series of experiments being 

 made with crooked and curved channels. 



The vertical sides of the troughs did not well 

 represent the sloping banks of rivers, and no 

 attempt was made to measure the qualification 

 due to this difference. The cross section of 

 the laboratory current was essentially a rec- 

 tangle and the ability of the current to trans- 

 port was found to be definitely related to the 

 ratio between depth and width; but satisfactory 

 connection was not made between this relation 

 and the forms of cross section in rivers. 



The thalweg of a river channel traverses an 

 alternation of deeps and shoals, the deeps being 

 characterized by a different system of velocities 

 and by a different line of separation between 

 the grades of debris carried severally by sus- 

 pension and traction; but these contrasts were 

 touched only in a qualitative way in the work 

 of the laboratory. 



Each experiment dealt with a slope in ad- 

 justment with a particular discharge and a par- 

 ticular grade of sand. In a natural stream the 

 discharge is subject to variation, and its 

 changes cause changes in the fineness of the 

 material carried along the bottom. Load and 

 the local slopes are ever in process of adjust- 

 ment to the temporary conditions of discharge 

 and fineness, but the adjustment is never com- 

 plete. The general or average slope is adjusted 

 to an indeterminate discharge which is neither 

 the smallest nor the greatest. For this phase 

 of disparity allowance is not easily made. 



One of the conditions affecting velocities is 

 friction on bed and sides of channel. Friction 

 on the bed depends partly on the roughness of 

 the bed and partly on the consumption of 

 energy by traction. Its laws are the same in 

 laboratory and in river. Friction on the sides 

 depends on the character of the channel wall 

 and must be materially greater on river banks 

 than on the smooth sides of the experiment 



trough. The magnitude of the difference was 

 not determined. Velocities are affected by the 

 viscosity of the water, variations in this factor 

 being caused by differences hi temperature and 

 by impurities in solution and in suspension. 

 The transportation of small particles is affected 

 by adhesion, a property varying with the min- 

 eral character of the particles and with the 

 impurities of the water. These factors were, 

 ignored but are probably negligible in com- 

 parison with the factors tested. It may be 

 mentioned, however, that the water of the 

 laboratory was practically free from sus- 

 pended material, whereas that of rivers is usu- 

 ally highly charged at the time of most active 

 traction. 



These comparisons serve to show that the 

 investigation treats of a group of important 

 factors of the general problem of stream 

 traction but by no means comprehends all. 

 Its results constitute only a contribution to 

 the subject. 



ACCESSORY STUDIES. 



Incidental and accessory to the main in- 

 quiry were a number of minor inquiries. One 

 pertained to the Pitot tube, a second to other 

 methods of measuring velocity near the bot- 

 tom, a third to the relation between the mean 

 velocity of a loadless stream and the rough- 

 ness of its channel bed, and a fourth to the 

 . mechanical process of hydraulic traction. 



FLUME TRACTION. 



In the experiments on flume traction the bed 

 of the channel was not composed of loose 

 d4bris but was the unyielding bottom of the 

 trough. The same apparatus was used, with 

 appropriate modifications. In each experi- 

 ment slope of channel was predetermined, the 

 trough being placed with definite inclination. 

 The bed of the channel was given a definite 

 quality of roughness or smoothness, and the 

 material of the load was of a particular fine- 

 ness or of a definite mixture of sizes. With a 

 definite discharge flowing through the trough, 

 debris was fed to the current at a definite rate, 

 and the rate was gradually increased until 

 clogging occurred. The rate of feed just be- 

 fore clogging was then recorded as the maxi- 

 mum load under the particular conditions. 

 The series of experiments used two widths of 



