130 A NEW METHOD OF ESTIMATING STREAM-FLOW 
surface (if one neglects seiches, tides and barometric effects) becomes disturbed, 
first by ripples, then later by waves. As the waves increase in size, the wind is 
caused to strike the water with a component at right-angles to the water surface 
which is greater than that in the case of an evaporation pan in which at best only 
small ripples form. Such action, in which the wind is brought forcibly into contact 
with the sides of the waves, must act greatly to increase the rate of evaporation. 
Besides this effect, one might distinguish another. Suppose the wind velocity 
to increase beyond the point where white-caps begin to form (say beyond 11 miles 
per hour) . In the action of white-caps, when the top of the wave is caught by the 
wind and transported forward at a greater rate than the lower part of the wave, 
air is trapped under it, later on bubbling through. Moreover, bits of spray are 
caught and thrown high into the air and far over the leeward shore of the lake. 
Under such conditions the exposed area of water in contact with the air, and the 
increased effectiveness and intimacy of that contact, together with the water car- 
ried inland, must act greatly to augment the rate of increase of evaporation with 
increasing wind velocity. Viewed in this manner this rate of increase must neces- 
sarily be much greater than in the case in small evaporation pans. 
