CLIMATIC CONDITIONS OF THE UNITED STATES. 295 



summed, and to this sum have been added two quantities, representing 

 the evaporation for the two fractional parts of a month at the beginning 

 and at the end of the frostless season, respectively. These added 

 amounts have been derived, in each case, by dividing Russell's evapora- 

 tion-rate for the month in question by the number of days in that 

 month, and then multiplying the resulting quotient by the number of 

 days of the same month comprised in the period of the average frostless 

 season. 



The total evaporation for the season thus obtained is next divided 

 by the number of days in the period of the average frostless season. 

 Thus, for example, if the season extends from April 24 to October 3, 

 we sum Russell's monthly rates for May to September, inclusive, and 

 add to this six-thirtieths of Russell's rate for April and three thirty- 

 firsts of his rate for October. We then divide this total by 162, the 

 number of days from April 24 to October 3. The seasonal totals and 

 the approximate average daily evaporation-rates (1887-88) for the 

 period of the average frostless season, obtained as just described, for 

 133 stations, are given in columns 6 and 7 of table 11, and the latter 

 data are spread on the chart of plate 53, where the positions of the 

 stations employed are again shown by small circles. It should be 

 emphasized that Russell's data for June, and earher, refer to 1888, 

 while those for July, and later, refer to 1887. We have merely made 

 the best possible use of the available data. 



The isoatmic lines of this chart are drawn at intervals of 20 thou- 

 sandths of an inch of average daily depth of loss from some hypothe- 

 tical pan of water, Russell's measurements being in such terms. The 

 total range of values is from 52 (Tatoosh Island, Washington) to 349 

 (Independence, California). 



On this chart of the approximate average daily intensities of atmos- 

 pheric evaporating power for the period of the average frostless season, 

 from data of 1887 and 1888, it appears that the isoatmic lines in the 

 vicinity of the oceans have a very pronounced north-south trend. 

 They also have a north-south direction in the plains region. Little 

 relation to latitude, or to temperature, is here discoverable; during the 

 frostless season temperature is not usually a prime condition in the 

 determination of differences between different stations in the evaporat- 

 ing power of the air. 



The lines for values 120, 160, and 240 are shown on plate 53 as dis- 

 tinct from the others, and these may be taken as dividing the country 

 into four evaporation provinces. Following our usage in the case of 

 the precipitation indices, these provinces will be termed arid, semi- 

 arid, humid, and semihumid. They are shown also in figure 14, 

 which is derived from plate 53. This zonation is different from that 

 shown for precipitation in several important respects. Here the 

 humid province (values above 120) appears again as a western and an 



