ATMOMETRY AND THE ATMOMETER 23 



posed be altered in the same way for both pans, then both rates 

 of loss should change, but this change should be the same for 

 both instruments. 



A somewhat less obvious result follows if two different kinds 

 of water pan are exposed first to one external complex and 

 then to another. For such a case, let it be supposed that one 

 instrument loses twice as much water per hour as does the other, 

 under the first conditional complex. Under the second com- 

 plex of external conditions this relation between the two rates 

 of loss no longer holds; the first instrument does not then lose 

 just twice as much as does the second. In other words, if 

 one of these unlike pans be standardized to the other, the 

 coefficient of correction (by which the loss from one pan may 

 be multiplied, in order to give the loss from the other) would 

 be different for each new set of external conditions. 



It appears that the principle just dealt with is but vaguely 

 appreciated by many students of atmometry, a conclusion 

 reached by a perusal of the meteorological literature of evapor- 

 ation; in spite of frequent publications establishing the influence 

 of size, shape, etc., of pans used, workers in this field still per- 

 sist in employing a variety of pans and tanks, thus rendering 

 their measurements of but little lasting or generally compara- 

 tive value. If it might be remembered that the evaporation 

 rate depends upon the water surface and upon the surroundings 

 of this surface (both below and above), then this misapprehen- 

 sion should be avoided. 



From what precedes it is clear that no form of pan is any 

 more a natural standard for measuring evaporation than is 

 any other. When pans are to be employed in atmometry it is 

 only essential that all pans be alike, if the measurements thereby 

 obtained are to be compared. In any event, the pan used must 

 be described in detail if the results are to be of any value to 

 other workers. 



Another consideration that appears worthy of emphasis here 

 is this, that the exposure of several atmometers must be com- 

 parable if the readings are to be so. Thus, if one of two like 

 pans be placed at ground level and the other at some elevation 



