122 A NEW METHOD OF ESTIMATING STREAM-FLOW 
the wind term and the other term is neglected, the computed evaporation on such 
a day will stand 1 chance in 2 of being in error less than 10 per cent, assuming no 
systematic or constant errors present, and no errors in the measured wind velocity 
and saturation deficit. 
COMPARISON OF FORMULA DEVELOPED IN THIS INVESTIGATION WITH OTHERS 
The presentation of the methods used and the results obtained in that part 
of this investigation dealing primarily with evaporation would be incomplete with- 
out comparing the results with those obtained by other investigators. Accord- 
ingly, the remaining paragraphs of this appendix are devoted to making such 
comparisons. No attempt is made to present an exhaustive comparison. The 
number of evaporation formulas extant in the scientific and engineering literature 
is very large. These formulas differ from the one derived in this investigation in 
three ways, (a) They are, so far as the writer knows, all based upon observations 
on small-scale apparatus, on evaporation pans a few square inches or a few square 
feet in area, placed either in a laboratory or out in the open in different positions 
with reference to the surface of the ground, water or vegetation. In these artificial 
conditions he the fundamental weakness of all such formulas, namely, the assump- 
tion that the laws of evaporation so obtained will be correct when applied to 
natural conditions, (b) In most of the evaporation formulas based on small-scale 
apparatus, the observations of the meteorological elements have been taken at or 
near the apparatus. The formulas developed, therefore, express primarily the 
relationship between the rate of evaporation in the pan and the meteorological 
elements so observed. Such formulas, then, are of limited application, even if the 
defect mentioned in (a) were not present. This follows because one can not use 
them with existing, standard observations of the weather elements as made by the 
Weather Bureau. In this sense the evaporation in any locality in the past can not 
be investigated, and special observations are required for estimating current 
evaporation, (c) As compared with the evaporation formula developed in this 
investigation, most other formulas are based on relatively few observations. 
Moreover, the precision with which the formulas express the stated relationships 
has not usually been determined in the sense that the probable errors of the pro- 
portionality constants have not usually been derived. 
In comparing and contrasting the results of this investigation with others it 
is important to bear in mind these three points of difference. The comparisons 
presented are probably typical. They serve to bring out the principal differences 
obtained by the two radically different lines of attack. 
COMPARISON WITH FREEMAN'S (AND MEYER'S) FORMULA 
The formula developed by John R. Freeman 1 is as follows: 
£. = 0.5 (V-v) +0.05 (V-v) w, in which (A) 
V= Saturation vapor-pressure corresponding to the temperature of the water, in inches 
of mercury; 
v = Saturation vapor-pressure corresponding to the dew-point temperature, in inches of 
mercury; 
w = Average velocity of wind, in miles per hour; 
E„= Evaporation in inches per day. 
1 Regulation of the Great Lakes, by John R. Freeman (1926), Appendix 2. 
