Operation of the Porous Cup Atmometer, 113 



the second cup the first cup is restandardized, and is ready to 

 operate durin,^ the third period, the second cup being then re- 

 standardized for use in the fourth j eriod, etc. T!ius, for a sea- 

 son's work in the field, it is necessary either to be almost con- 

 tinually at work with standardizations, or to provide at the 

 start a sufiicient number of cujjS so that a newly standardized 

 one mav be exposed at each station every month or oftener. 

 By the latter plan, each used cup, when taken from the instru- 

 ment, is allowed to dry naturally and then wrapped in paper 

 and preserved till circumstances allow all to be standardized 

 together. 



The value of the observations depends so largely upon the 

 standardization that this second test can never be safely omitted, 

 unless, indeed, the cups have been only a few days in operation. 

 Cups which are tested at short intervals for several months, at 

 the Desert Laboratory, showed that the change in the correction 

 coefficient progresses nearly continuously and nearly uniformly 

 and it is this fact which makes the method of repeated testing 

 practicable. Experience has shown that for habitat and gen- 

 eral climatic studies, the errors introduced are practically 

 negligible if the cups (otherwise, of course, properly treated) 

 are tested each month. The weekly or daily readings for the 

 month are corrected to standard by the average of the two 

 coefficients obtained before and after the exposure of the cups, 

 unless the cups show a great change (over ten or fifteen per cent.), 

 in which case the interpolated coefficients may be calculated on 

 the assumption that the variation has progressed uniformly 

 throughout the n-onth. Thus, if the original coefficient of 

 standardization for a given cup be c and the coefiicient obtained 

 after four weeks' operation be c, c and c not differing by more 

 than, say, 0.10, then the corrected reading, E, for each week, is' 



c+c^ 



obtained by multiplying the actual reading, R, by 



c+c' 



or E=R . When c-c^ is greater than 0.10, we may cor- 



c'-c 



rect for the first week by Uic following formula: E=^R(c+ ), 



n 



