184 A NEW METHOD OF ESTIMATING STREAM-FLOW 
Example of computation of (ni—Gi), (rii—Gi), .... Solution A A, Stream A 
Date 
{t-T") 
(t-T")M 
C 
Esti- 
mated 
net 
melting 
n 
G 
r/i 
(«i-G.) 
fa-Gt) 
(/!} — Cj) 
(n t -G t ) 
(n h — Gi) 
1913 
Mar. 29 
30 
31 
Apr. 1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
13 
14 
15 
16 
+ 3 
+ 9 
+ 10 
+ 8 
+ 4 
- 6 
- 2 
+ 6 
+ 10 
+ 3 
- 3 
- 8 
- 7 
+ 1 
+ 5 
+ 9 
+ 10 
+ 12 
+ 12 
+ 15 
+45 
+50 
+40 
+ 20 
+ 4 
+4 
+4 
+4 
+4 
+ 19 
+49 
+54 
+44 
+24 
+ 13 
+38 
+42 
+37 
+ 17 
51 
42 
34 
23 
23 
+ 8 
+ 14 
-46 
-30 
-46 
-30 
-46 
-30 
-46 
-76 
-30 
-4.6 
-7.6 
-7.6 
-7.6 
-3.0 
-3.0 
-3.0 
-3.0 
-3.0 
+ 30 
+ 50 
+ 15 
+4 
+ 4 
+ 4 
+34 
+ 54 
+ 19 
+22 
+ 47 
+ 12 
31 
23 
22 
+ 14 
-30 
-30 
-30 
-30 
-30 
+ 5 
+25 
+45 
+50 
+60 
+ 60 
+4 
+4 
+4 
+4 
+4 
+4 
+ 9 
+ 29 
+49 
+54 
+64 
+64 

+ 17 
+36 
+37 
+44 
+47 
35 
36 
24 
22 
21 
21 
+ 12 
+ 15 
+ 23 
+ 26 
-37 
-39 
-41 
-38 
-37 
-39 
-41 
-37 
-39 
-37 
nection with the observed flow, D' '. These can be verified from Plate 10. Thus 
on March 30, D' was 138, which gives a value of 34 for G on March 31. Since the 
estimated r,, +42, for that date is greater than G, the melting was large enough to 
initiate a flood-flow. Hence March 31 becomes one date to use in the computation 
of D f , and the value of (n t — Gi) for that date, —46, is the sum of n x shown on page 
180, —12, and the value of — G estimated above, —34. This becomes the value 
of (n 2 — G 2 ) for the following date, (n 3 — G>) for the next following date, and enters 
into the values of {n l — G i ), (n 6 — G 5 ), . . . for the succeeding dates according to 
the grouping of days in the definitions of rii, n 2 , n s , . . . and r h r t , r», . . . . 
Each date on which the estimated value of r x is greater than G for that date is 
used in the computation of (rii — Gi), (n* — Gt), . . . which values are shown in the 
computation of D,, page 181. 
It is important to note that the assumed values of C, M and T" enter only 
indirectly into the computations of this Solution A A. They serve merely to fix 
certain dates to be used in that solution. If, after the solution is completed, the 
derived values of C, M and T" from it check the dates estimated as above, no error 
whatsoever is introduced into the derived C, F, M and T" from the solution as a 
result of the assumed values of C, F, M and T" . If they do not verify those dates, 
it becomes necessary to estimate the error produced by using the wrong dates, and, 
if large, to repeat the process until the desired degree of accuracy is attained. 
Up to this point in the illustrations of the computations used in Solution A A, 
there has been shown the method of computing — D, the absolute term of the 
observation equation (46). Illustrations of the method of computing the other 
known quantities x, y and z will now be given. 
EXAMPLE OF COMPUTATION OF x n SOLUTION AA, STREAM A 
The definition of x n is given by equation (50). It consists of the known quan- 
tities, the R"s and the time intervals involved in those R"s. Substituting in equa- 
