A NEW METHOD OF ESTIMATING STREAM-FLOW 179 
pressed in percentages, to Stream A is larger than that to Stream B for many days 
thereafter. In the case of flood-flows the rate of delivery to Stream B is larger 
than that to Stream A after about the third day after a large addition to storage. 
If the run-off curves are reduced to the same basis by taking account of the 
difference in drainage areas, they represent the comparative flows of the two streams 
as affected by changes in storage in their respective drainage areas. Since, how- 
ever, the difference in drainage areas is only about 10 per cent, the curves may, as 
they are, be taken for approximate comparative flow characteristics. 
EXAMPLES OF COMPUTATIONS USED IN THE DETERMINATION OF C, F, M AND T" 
Further comments upon the comparative flow characteristics of the two 
streams will be reserved for a later place. It is now proposed to present sample 
computations from Solution A A, Stream A, which was the final solution on that 
stream which served to fix final values of C, F, M and T". These specific illustra- 
tions will serve to clarify the theory presented in general terms on pages 140 to 
143, and represents an illustration of some multiple of step (6) on page 147. 
EXAMPLE OF COMPUTATION OF D n , SOLUTION AA, STREAM A, MARCH TO APRIL 1913 
In the computation of D„, equation (43) is used, substituting therein the best 
values of S e , R\, R\, . . . R\ determined to date, and the computed values of 
rii, n 2 , n 3 , . . . n [0 as computed from equation (40) in exactly the same manner 
as n, r 2 , r 3 , . . . r !0 are computed from equation (41) and as illustrated on pages 
154 and 155 except that in the computation of n h n 2 , n», . . . n 10 the net melting is 
not known and therefore does not enter the computation. The only precipitation 
which enters the computation of n x , n 2 , n», . . . w, is in the form of rain. In the 
computation of n,, n 2 , n 3 , . . . for Solution A A, the evaporation from land was 
Ei 
estimated from equation (34) in which — was taken as 2.6 instead of 2.3 used in 
E w 
Solution M, the value 2.6 being based upon computations made subsequent to 
Solution M, and probably more exact. Because of the similarity between the 
computation of n, r 2 , r 3 , . . . and Wi, n 2 , n 3 . . . , and since the former has already 
been illustrated, the latter will be assumed to be known in this illustration. 
The best values of S c , R\, R' 2 , . . . R'k, determined at the time Solution A A 
was made were those of equation (76). Using those values, the computation of 
D n is shown on page 180. 
The values in parentheses at the tops of the columns are the R"s of equation 
(76) except that R f 9 is multiplied by 10 because the unit in which n 9 was computed 
was 1.0 inch of depth instead of 0.1 inch of depth as in the case of r„. These values 
should be considered as repeated down the column. 
The other values except in the first, second and last columns are n h n 2 , . . . n 9 ; 
in the following units: rii to n t , 0.01 inch of depth; n & to n*, 0.1 inch of depth; w», 
1.0 inch of depth. 
The products, nR' , and S c and D n are in 0.001 c.f.s. 
In column 2 is S c , the constant part of the stream-flow. 
The sum of S c and the nR' — products gives D n . 
It is to be noted that the n's are all negative. The reason is that there was no 
addition to storage in the form of rain in the period shown, and none, or not enough 
as rain, in the earlier periods into which the late n's reached back to make them 
positive. 
