A NEW EVAPORATION FORMULA 99 
In Solution BB, the form of the evaporation equation used was 
^,+[e(m-^)y+[e(m- 2A J] E -+'= < 25 > 
Dividing this through by e and placing the absolute term on the left-hand side 
of the equation there is obtained 
-4-*+Gjb -")*+(&-")'* ■ < 26) 
which is similar in form to (24) ; that is, y = , x = ( jkk — 2.4 j, a = E 1} b = E t and 
c = E t . 
In Solution BB,, for each value of ( tkx — 2.4), the corresponding values of 
e and J were tabulated and their means computed. The ratio of the mean I to 
the mean e gave the left-hand member of (26). The first and second differences of 
this ratio were then calculated. They are shown plotted in the lower half of 
Plate 6, and the computations are shown in Table 34. The curves of first and 
second differences, considering the curves as a whole, indicate that the first differ- 
ences are more constant than the second differences, or that n= 1, in equation (24). 
That is, for the curve as a whole, the facts can be represented better by a straight- 
line than by a parabola. However, for the values of ( r^r — 2.4] between —0.7 
and 0.0 there appears to be about equal constancy in the first and second differ- 
ences, or either a straight-line or parabola would represent the facts about equally 
well in that region. 
This evidence favors the adopted form of the evaporation formula only in 
respect of indicating evaporation to be best evaluated by two terms. It says 
nothing about the ventilating effects of low winds. 
EVAPORATION IS A LINEAR FUNCTION OF THE WIND VELOCITY ABOVE 
10.8 MILES PER HOUR 
Evidence similar to the above from Solution BB, on Lake Superior, indicating 
that the exponent of the wind term is practically unity, was verified earlier in the 
investigation by other analyses on both lakes — analyses of a different kind than 
the one just presented. It has already been stated that on Lake Michigan-Huron 
no departure whatever from a linear relationship could be detected, but that on 
Lake Superior faint evidence was obtained that there might be curvature in the 
evaporation curve for wind velocities below 10 miles per hour, and that this evi- 
dence was used as a basis for setting up Solution BB,. It is now proposed to set 
forth this evidence from both lakes. 
The evidence is from Solution £/,, Lake Superior, and Solution 7\, Lake Michi- 
gan-Huron. Each of these solutions contained six months of observations, May to 
October 1910. These solutions represent early attempts to determine simul- 
taneously the laws of evaporation from the lake surface and of run-off into the lake. 
The form of observation equation used in each is as follows : 
e tfi+lViSk -2 - 4 Yl ^i-r.fli-r,fl,-r I fl l -r«fl 4 -r l «,-r.fl.-r-/ = t; (27) 
[<&"")] 
