268 AGRICULTURAL INDICATORS. 



Akron and Cheyenne, while it was nearly normal at Williston and above 

 normal at Amarillo, 1910 being the driest year at both these places. The year 

 1914 was dry at Akron and Cheyenne, nearly normal at Amarillo, and above 

 normal at Williston. The regional variations in seasonal rainfall, both abso- 

 lute and relative, are even more marked. In 1915, the year of greatest sea- 

 sonal rainfall at Amarillo and Akron, they received 18 inches from April to 

 August inclusive, Cheyenne 10 inches, and Williston 6 inches. The sea- 

 sonal rainfall was respectively 66, 55, and 50 per cent of the annual for the 

 year. The year of greatest relative seasonal rainfall was that of 1910 at 

 Amarillo, when 90 per cent of the annual rainfall came during the growing 

 season. The corresponding values for Akron, Cheyenne, and Williston were 

 73^ 50, and 70 per cent respectively (figs. 16-19). 



Anticipation of cycles. Crop production makes much greater demands as 

 to the forecasting of rainfall than either grazing or forestry. These deal 

 primarily with perennials, and in the case of trees in particular the depen- 

 dence upon the summer rainfall is much less marked. As a consequence, a 

 knowledge of the probable occurrence of the wet and dry phases of the 22- 

 year and 1 1-year cycles or of the approximate total rainfall for any year is of 

 much value. With annual crops the case is very different. While there is a 

 general relation between annual and seasonal rainfall, the latter may vary 

 between 50 and 90 per cent of the annual, as at Amarillo. Moreover, the 

 distribution and timeliness of the seasonal rainfall are even more critical (Ball 

 and Rothgeb, 1918:24, 6). It must be frankly admitted that at present 

 there are almost no clues to either distribution or timeliness, but this is due 

 largely to the fact that their correlations have received almost no intensive 

 study. It seems not improbable that the same basic processes of action and 

 reaction and of compensating balance apply during the year and season as 

 during cycles, and that they must be considered with reference to spatial 

 variations as well. It is probable that the most important clue to the annual 

 and seasonal rainfall of a particular year lies in the excess-deficit cycle of 2 to 

 3 years, which Arctowski has noted in crops and Douglass in trees. The 

 assumption that a cycle of similar character may apply to the months receives 

 striking confirmation from the studies of Douglas (1919) on the relation of 

 weather to business. The general correlations of climate with production 

 and prices and the existence of economic cycles have been dealt with by Moore 

 (1914, 1917). All of these represent independent investigations and can hardly 

 fail to strengthen the view that both long-period and short-period cycles occur 

 in crop production (figs. 20 and 21). 



In the endeavor to definitize climatic and production cycles and to discover 

 a working basis for their prediction, investigations are under way to determine 

 the climates and subclimates of the West on a plant basis. It is hoped to 

 ascertain the response to the 22-year, 11-year, and 2 to 3 year cycles in t^rms 

 of tree growth, grass yield, and crop production for different regions, suggested 

 by the type or amount of rainfall. It is expected that the general correlations 

 between rainfall and production will serve to mark the climates proper, but 

 that the latter will show a series of subdivisions leading to restricted locaUties 

 as the units upon which the practical anticipation of rainfall must be based. 



