A NEW EVAPORATION FORMULA 
113 
final values used, for Solution F,, were obtained. The evidence from the Bret 
studies on the two Colorado streams indicated that even the values finally used in 
Solution Vi are probably too large, and vindicates the evidence from that solution 
itself. The first approximate values of the R's obtained on Streams A and B, 
expressed as percentages, as in (30) are 
Value of— 
For Stream A 
For Stream B 
Mean for 
Streams .4 and li 
ffi 
Hi 
Rt 
i?4 
R» 
lit 
+0.71 
+ .41 
+ .42 
+ .28 
+ .17 
+ .15 
+0.08 
+ .19 
+ .10 
+ -11 
+ .002 
- .039 
+0.40 
+ .30 
+ .29 
+ .20 
+ .09 
+ .00 
It should be stated at this point that in equation (28), the symbol R with a 
subscript corresponds with the symbol R', with the corresponding subscript, used 
in the stream-flow studies. 
Compare the mean R's shown in the above tabulation with (30), page 106. 
These first approximate values from the stream-flow studies indicate that 0.40 per 
cent of the change in storage in the ground on the current day is delivered to the 
stream on that day. In Solution V h it was assumed to be 1 per cent. Similarly, 
from the above tabulation, 0.30 per cent of the change in storage in the ground on 
the current day reaches the stream the following day, if the addition to the storage 
in the ground on the following day is assumed to be zero. In solution F, it was 
assumed that 7 per cent (or 23 times as much) of the change in storage in the ground 
on the current day is delivered to the lake on the following day, under the same 
assumption. Similarly, a comparison of the other values shows that in Solution 
Vi the assumed percentages varied from 59 times larger in the case of R 3 to 5 times 
larger in the case of R t , than the first estimates of the actually derived percentages 
on the two Colorado streams. Clearly it would be unreasonable to assume that 
actually such a large difference does exist. The underground drainage system of 
the Great Lakes can not be as different from the underground drainage system of 
the Rocky Mountain region in the vicinity of Streams A and B as the difference 
in these figures indicate. In fact, the character of the underground drainage sys- 
tem of the Great Lakes is probably much nearer like that of the drainage areas of 
Streams A and B, as represented by the percentages shown in the preceding tabula- 
tion, than it is like the drainage system which would correspond to the percentages 
of delivery shown in (30). 
On the basis of this reasoning, the assumed R's for Lake Superior were made 
like the means shown in the preceding tabulation, except that they were slightly 
modified according to the general considerations stated in the following lettered 
paragraphs : 
(a) The maximum run-off was assumed to come on the second day after the 
change in storage takes place, on Lake Superior, instead of on the first day, 
as on the Colorado streams, according to the mean values shown above, be- 
cause the streams emptying into Lake Superior are much longer and the drainage 
area is much larger than those studied in Colorado. 
