106 



TRANSPORTATION OF DEBRIS BY RUNNING WATER. 



slopes. With the gentler slopes they give 

 indications of a minimum, but not with the 

 steeper. The character of the discrepancy is 

 such as to suggest that the values of the index 

 computed for width 0.23 foot vary too rapidly 

 with slope; and this result might be brought 

 about by assigning too large a value to a. A 

 critical review of the data, however, failed to 

 find warrant for any material change in that 

 constant. 



It is believed that a group of discrepancies 

 which this instance illustrates are connected 



with the history of the experimental work. 

 The first grade to be investigated was (B), and 

 the methods of manipulation were subject to 

 various minor changes, which were not always 

 recorded; but the range of conditions was large. 

 Grade (C) was next taken up, and again the 

 range of conditions was large. Other grades 

 followed, with less elaborate range of condi- 

 tions; but the work on grade (G) was somewhat 

 expanded, in order to learn the influence of 

 coarser debris on various factors. The work 

 on (G) also had the advantage of the fullest 



.66 



1.32 



1.00 

 Width 



FIGURE 32. Variations of ii in relation to width of channel. 



1.96 



development of experimental method as well as 

 that of uniformity of method. Because of this 

 history it is believed that the results for grades 

 (C) and (G) are of higher authority than those 

 of other grades; and the belief is strengthened 

 by the general symmetry and internal con- 

 sistency of the (C) and (G) results. The infer- 

 ences, given in preceding paragraphs, from 

 data of grades (C) and (G) are therefore ac- 

 cepted, and the discordance of data for grade 

 (B), while not specifically explained, is ascribed 

 in a general way to unrecorded differences in 

 laboratory methods. 



The curves of the fourth group of figure 32 

 all pertain to grade (G) and discharge 0.734 

 ft. 3 /sec. but differ in respect to slope. Com- 

 pared with the first and second groups they are 

 seen to be consistent with the inference as to a 

 minimum value of %, but the minimum falls 

 outside the range of widths for which data were 

 obtained. With grade (C) and discharge 0.734 

 ft. 3 /sec. the minimum falls between widths 1 

 foot and 1.32 feet, but nearer to the former. 

 With grade (G) and the same discharge it ap- 

 parently falls with some width less than 0.66 

 foot. This indicates that 



