200 A NEW METHOD OF ESTIMATING STREAM-FLOW 
observed flow is a solid line and of computed flow it is a dashed line. Each of these 
curves is the mean of the four 1 curves shown on Plates 9 to 17 inclusive, of the kind 
of^flow'on the stream to which it pertains. 
It is desirable to discuss in more detail the evidence as presented in graphic 
form on Plates 9 to 12 on Stream A and on Plates 13 to 17 on Stream B. In either 
case, a cursory examination shows that a large percentage of the discrepancies 
between the two curves, computed and observed flow, are under 20 per cent. On 
Stream A the principal discrepancy between the two curves is that the summit and 
center of gravity of the computed spring flow comes before that of the observed 
flood in every case. It is about 13 days before in 1911 (not shown on plates), 12 
days before in 1912, 10 days before in 1913, 18 days before in 1914, and 23 days 
before in 1915. 
This displacement in time on the Stream A spring floods is the principal dis- 
crepancy as a rule between the computed and observed flood-flows. 2 
The flood summits, observed and computed, are nearly the same height on both 
streams in 1911, 1913, 1914 and 1915. In 1912 the observed summit of the spring 
flood is about three times as high as the computed, on both streams. In 1912, 
the spring flood was the largest that occurred within the period 1911-1915. 
On Stream B the discrepancy in time between the computed and observed 
spring flood peak is small. In this connection it should be remembered that the 
flood coefficients used, R'/i, R' ti , R' /t , . . . R 1 ' n , as well as the values of G, were 
derived from observations in the summer months only. The use of these constants 
to compute the stream-flow observations in the winter months, with the degree of 
agreement shown, is a strong indication that assumption No. 10, especially, on 
which the laws of freezing and melting mostly depend, is not far from the truth. 
In this connection it is interesting to note that Horton also reached a conclusion 
(The Melting of Snow) that would justify assumption No. 10; viz, that under suit- 
able conditions snow behaves like any other permeable medium, such as porous 
soil, as regards the percolation of water through it and capillary retention of water 
in the interstices of the medium. 
A comparison of the extreme values on Stream A is as follows: 
The lowest daily discharge observed was 0.041 c.f.s. on December 16, 1914. 
The computed discharge on that day was 0.089 c.f.s., in error by 117 per cent or 
0.048 c.f.s. 
The lowest daily discharge computed was 0.048 c.f.s. on September 22-23, 
1915, in error by 38 per cent, the observed discharge being 0.078 c.f.s. and differing 
by only 0.007 c.f.s. or 17 per cent from the lowest observed discharge. 
The highest daily discharge observed was 1.645 c.f.s. on May 20, 1912. The 
computed discharge on that day was 0.350 c.f.s., in error by 79 per cent. 
The highest daily discharge computed was 0.845 c.f.s. on April 28, 1911. The 
observed discharge on that day was 0.328 c.f.s. and the computed discharge was 
therefore in error by 0.517 c.f.s. or 160 per cent. The highest computed discharge 
was (1.645 — 0.845 = ) 0.800 c.f.s. less than the highest observed discharge, or 49 
per cent less. 
The greatest absolute error occurred on May 20, 1912, when the computed 
discharge was only 0.353 c.f.s., and the observed was 1.642 c.f.s. 
1 Five curves for Stream B. 
2 This displacement was largely removed on Stream B as a result of improvement in the rule for determining 
the last day of full net melting in the spring. 
