56 



TRANSPORTATION OF DEBRIS BY RUNNING WATER. 



values of capacity and slope by a system of 

 harmonious or adjusted values. 



ERRORS. 



As a first step in the treatment of the errors 

 of the data they were studied with a view to 

 the discrimination of the systematic and the 

 accidental. 



The three modes of traction the dune, the 

 smooth, the antidune although intergrading, 

 are mechanically different. It was surmised 

 that they might differ in efficiency, so that the 

 capacity-slope curve might show a step in 



400 



200 



100 

 5-80 



| 60 



40 



20 



7 



.4 .6 .8 I 23 



Slope 



FIGURE 14. Logarithmic plot of a series of observations on capacity 

 and slope. Compare figure 13. 



passing from one to another; and it was also 

 surmised that the law connecting capacity 

 with slope might not be the same for the 

 several modes. A suggestive observation had 

 shown that on very low slopes slopes so low 

 that capacity is minute the current changes 

 an artificially smoothed bed of debris to a 

 system of dunes, and that with the develop- 

 ment of dunes the load is notably increased 

 without any change in general slope. To test 

 the surmises all the series were plotted on log- 

 arithmic section paper. Figure 14 shows a 

 logarithmic plot of the same data which appear 

 in figure 13; and it will be observed that the 

 line suggested by the points has much less 



curvature in the logarithmic plot. Its ap- 

 proximation to a straight line makes the study 

 of its local peculiarities comparatively easy. 

 The examination of the plots, while not dis- 

 proving the surmises, showed that whatever 

 diverse influences may be exerted by the modes 

 of traction, they are too small to be discrim- 

 inated from the irregularities due to other 

 causes. 



Other sources of systematic error are con- 

 nected with the methods of experimentation. 



INTAKE INFLUENCES. 



As the water entered the experiment trough 

 from the stilling tank it was accelerated, the 

 gain in mean velocity being associated with a 

 quick descent in the surface profile. Beyond 

 this descent the profile usually rose somewhat, 

 and there was commonly a moderate develop- 

 ment of fixed waves. This development was 

 modified and the waves were on the whole 

 reduced by the addition of the debris. As it 

 fell into the water the debris had no forward 

 momentum, and it therefore tended to retard 

 the current. But the d6bris also accumu- 

 lated on the bottom, reducing the depth of the 

 water at that point, and this reduction neces- 

 sitated an increase in mean velocity. In the 

 immediate neighborhood of the place where 

 debris was fed the slope of the water was 

 affected by an abnormality distinguishable 

 from the intake abnormality proper, and the 

 joint abnormality faded gradually downstream. 

 The nature of these features varied with the 

 discharge and load, with the gradual develop- 

 ment of the adjusted slope, and also with the 

 mode of feeding. During the greater part of 

 the experimental work the feeding was either 

 automatic and continuous or else manipulated 

 by hand in such way as to make it nearly con- 

 tinuous, but for a minor part the feeding was 

 intermittent, a measureful of debris being 

 dumped into the water at regular intervals. 



OUTFALL INFLUENCES. 



In all the earlier work the trough had the 

 same cross section at the lower end as else- 

 where, and the water fell freely from its open 

 end to the settling tank. As the resistance to 

 its forward motion was less at the outfall than 

 within the trough, the water flowed faster 

 there. Its faster flow diminished the resist- 

 ance just above, and thus the influence of 



