152 



TRANSPORTATION OF DEBRIS BY RUNNING WATE&. 



There are, however, important discordances 

 among the data. Considered as errors, the 

 discordances constitute a group which were not 

 detected in the adjustments of capacities to 

 slopes and to discharges, but which escaped 

 those tests because related peculiarly to the 

 grades of debris. From the residuals of the 

 present readjustment the average probable 

 error of capacities before readjustment is esti- 

 mated at 5.4 per cent, whereas the average 

 probable error of the body of once-adjusted 

 capacities from which these were selected was 

 estimated at 2.5 per cent. On the assumption 

 that the estimate for the whole body of values 

 applies to the selected group, the share of error 

 associated with the grades is estimated as 



V5.4 2 -2.5 2 = 4.8 per cent. 



Inspection of the logarithmic plots suggested 

 that part of the discordance of the data is syste- 

 matic. To bring out the systematic element 

 the original values of capacity in Table 44 were 



divided by the readjusted values, and means 

 were taken of the quotients. The means are 

 listed below, and in figure 51 they are plotted 

 logarithmically in relation to fineness. The 

 plotted points conspicuously out of line are 

 those for grades (B) and (D), the capacities 

 determined for grade (B) being relatively too 

 small and those for grade (D) too large. The 

 same result was obtained from a canvass of a 

 wide range of data. 



Ratios oj original values of capacity to adjusted values. 



It is surmised that these errors arise in part 

 from variations of experimental method, and 

 this suspicion attaches especially to grade (B), 

 which was the first to be treated in the labora- 

 tory. It attaches much less to grade (D), for 



(G) 



(E) 



CD) 



(C) CB) (A) 



FIGURE 51. Average departures of original values of capacity from the system of values readjusted in relation to fineness of d<Sbris. Thehort 

 lontal line represents the readjusted system. The broken lines, above and below, correspond to departures of 10 per cent. 



which the experimental method was about the 

 same as for grades (A), (E), and (F). So far as 

 the quality of the experimental work is gaged 

 by the probable error, that on grade (D) would 

 appear to be considerably below the average. 



It is surmised that systematic error may also 

 be connected with properties of the grades 

 other than that of fineness. Range of fineness 

 has already been appealed to as an explanation 

 of apparent anomalies as to competent slope. 

 From the experiments with mixtures (Chapter 

 IX) we know that great range within a grade 

 would tend to increase capacity for traction. 

 As the range is small for grade (A), regularly 

 increases to grade (E), and is again small from 

 (F) to (H), the influence of this factor tends to 

 enhance the capacity of grade (E) ; but the 

 effect of that influence is not apparent in the 

 diagram. 



The separation of grades, being effected by 

 sieves, was a gaging of grains by certain dimen- 

 sions. The grains may have differed also as to 

 shapes and densities, and each of these proper- 



ties would affect capacity. The debris first 

 obtained for the laboratory had been washed 

 from the bed of Sacramento River when a flood 

 broke the levee below the mouth of American 

 River. This furnished material for the finer 

 grades. Other lots of debris were taken from 

 the bed of American River, and these furnished 

 material chiefly for the coarser grades. So far 

 as the separations from the different lots coin- 

 cided, they were used indiscriminately. It is 

 possible that grade (D) was composed chiefly 

 of the coarser particles of a fine alluvium, while 

 grade (E) was composed chiefly of the finer 

 particles of a relatively coarse alluvium. As 

 any natural body of tractional debris is the 

 result of a sorting process in which tractionable 

 particles are separated from the immovable on 

 one side and from the suspendible on the other, 

 it is likely to include among its coarser grains 

 many which are tractionable only because of 

 low density or favorable shape, and among its 

 finer grams many which escape suspension 

 because of high density or unfavorable shape. 



