210 A NEW METHOD OF ESTIMATING STREAM-FLOW 
The whole of the error between the observed precipitation, 20.71 inches, and 
the computed precipitation, 28.10 inches, however, should not be attributable to 
the constants C, F, M and T" . Part of the error may be due to the errors in the 
decision as to when the freezing-melting period begins in the fall. It appears that 
the beginning of the freezing-melting period was taken too early rather than too 
late. In such a case it can be shown that the C-term would tend to cause too large 
an estimate of the computed net melting, not necessarily because C is itself in error, 
but primarily because it was entered into the estimate on too many days. 
The derivation of the constants C, F, M and T" was made from observations 
in two winters only, 1911-12 and 1912-13, in order that they could be used to esti- 
mate the net melting in the winters 1913-14 and 1914-15, not used in their deriva- 
tion, as a test of their validity. These two winters 1911-12 and 1912-13 happened 
to be winters of extremely heavy and extremely light snowfall, respectively. Using 
these extreme observations was ideal in the sense that the constants derived from 
such observations would represent the whole range of observations used. On the 
other hand, the number of observations used were so few, the decision as to which 
observations to reject in the spring of 1912 — the largest spring flood in the period 
1911-15 — makes it desirable to supplement the derivation of these values by other 
studies based upon more observations, particularly in years which are more repre- 
sentative of average conditions. 
The following comments are pertinent with reference to Table 59b : 
The computed net melting, computed as mentioned above, 
(1) had a minimum value in December, January and February of each year; 
(2) had a lowest monthly value of —0.81 inch in December 1913, or —0.026 
inch per day, which means that there was a subtraction from storage by water 
turning into ice at that average equivalent rate in that month; 
(3) had a maximum value in April or May of each year; 
(4) had a highest monthly value of 10.87 inches in April 1913, or 0.362 inch 
per day; 
(5) had an average annual value of 16.48 inches. 
The Observed Run-off — 
(1) had a minimum value in February, August or September of each year; 
(2) had a lowest monthly value in February 1914 and September 1915 of 
0.25 inch, or 0.0083 inch per day; 
(3) had a maximum value in May of each year; 
(4) had a highest monthly value of 2.94 inches in May 1912, or 0.095 inch per 
day; 
(5) had an average value of 0.017 inch per day. 
The Evaporation From Land, computed with Ei/E w = 2.3, 
(1) had a minimum value in December, January and February of each year; 
(2) had a lowest monthly value of 0.48 inch in December 1914, or 0.016 inch 
per day; 
(3) had a maximum value in June, July and August of each year; 
(4) had a highest monthly value of 4.74 inches in June 1915, or 0.158 inch 
per day; 
(5) had an average value of 0.06 inch per day. 
