102 BURTON EDWARD LIVINGSTON 



in length in the undilated tube should be placed in the bulb 

 of each valve. 



The tests carried out by Harvey (loc. cit.) showed that it is 

 easily possible to arrange the mercury valves so that a cup 30 cm. 

 above the water level will need to absorb no more than about 

 0.01 cc. in order to stop water movement into the reservoir. 

 With such an instrument, each time the valves are operated, as 

 with the beginning of rain, 0.01 cc. of water is injected into the 

 reservoir and when evaporation recommences this amount must 

 be lost before the true record is taken up again. Thus, for every 

 complete reversal of the valves a minus error is introduced 

 amounting to only about 0.01 cc. This is quite negligible, and 

 it must seldom happen that enough reversals occur in a period of 

 operation to riiake the error considerable even in the aggregate. 



Another error, also negligible, may be introduced by amounts 

 of absorption too small to operate the valves. It may- be sup- 

 posed, for example, that a gentle shower, driven by wind, may 

 wet one side of the cup so as to produce some absorption, while 

 evaporation equalling or exceeding the amount of absorption may 

 still be in progress on the leeward side. In such a case the final 

 reading, or apparent loss for the period, would be less than the 

 actual loss by the amount of water actually absorbed by the cup. 

 If such conditions were long continued this kind of error might 

 amount to more than that usually resulting from several complete 

 reversals of the valves. For this difficulty no remedy can be sug- 

 gested, other than that several cups as operated simultaneously 

 on the same pair of valves and reservoir (to increase the actual 

 rate of water loss) and that tubes of extremely small bore be 

 employed in the valves By this procedure the possible error 

 may be reduced indefinitely. 



The non-absorbing mounting should always be used for out- 

 door work ; if observations are worth the trouble of making they 

 should surely just'fy the slightly increased cost of operating 

 this form of instrument. The futility of making records from the 

 absorbing form, when operated in times of much rain, is well 

 shown by the study carried out by Brown 24 in Jamaica. 



24 Brown, W. EL, Evaporation and plant habitats in Jamaica. Plant World 

 13: 268-272, 1910. 



