238 Journal of Agricultural Research voi.x, No.s 



for purely scientific investigation; in most cases they were secured by 

 engineers for immediate use in the study of a water supply and not under 

 standard conditions, because there have been no standards. It is not 

 strange that under these conditions the meteorological records are lack- 

 ing. These evaporation depths undoubtedly served their purpose to 

 the extent at least of satisfying the investigator. For the 84 stations, 

 31 have accompanying temperature figures, 9 relative humidity per- 

 centages, and 7 the wind movement. As they stand the records can be 

 applied only under great disadvantage to other work, even in the same 

 section. In many cases Weather Bureau records will supply part of 

 the missing data which should accompany the evaporation figures 

 proper, but not as satisfactorily as records taken at the evaporation 

 station. It is obvious that Weather Bureau data, taken in a city, the 

 station located on a roof, represent conditions which may differ greatly 

 from those at a reservoir on which is floating an evaporation pan, though 

 this reservoir may be but a few miles from the city station. 



Factors which are at present believed to be the principal ones con- 

 trolling evaporation are outlined. The effects of sunshine and radia- 

 tion as measured by instruments for that purpose are understood even 

 less than those of the so-called principal factors. There is, however, no 

 proof that they are negligible, and a record of these may add to the 

 value of future evaporation research. 



TEMPERATURE 



If water temperatures taken at the surface were available to accom- 

 pany all evaporation measurements made, they would be of value. 

 Evaporation research has proved that approximate calculation of amounts 

 could be made from water temperatures alone. However, these data are 

 to be had in connection with evaporation records in but few cases, and 

 for estimating losses for districts where previous evaporation records 

 have not been made water temperatures are never available. The in- 

 creased value of the records justifies water-temperature measurements. 

 Water temperatures are dependent upon air temperatures. 



The long establishment and very complete records of the Weather 

 Bureau make it possible to secure data on air temperature for all parts 

 of the country, taken under standard and somewhat similar conditions 

 (27). The advantage of a method of basing evaporation estimates upon 

 these records is obvious. The publications of that Bureau present daily 

 records arranged in periods of one calendar month. Mean temperatures 

 are essential, and from the records means may be found for periods of 

 any length. A month is too long, since there may be such variations 

 during the period that the mean does not correctly define the condi- 

 tions in their application to evaporation. One day is too short, since 

 water-temperature variations lag behind those of air, and their effect 

 may be shown several hours later on the evaporation record. Periods 



