Summary and Conclusions 141 



of spring wheat from the yield of winter wheat; and, 

 secondly, since the growth seasons and critical periods 

 of these two varieties of wheat are different, it seemed 

 unwise to attempt to connect the rainfall of any season 

 with the yield per acre of wheat in which the figures for 

 the yield referred to spring and winter wheat taken 

 together. For these reasons the representative crops 

 were limited to corn, hay, oats, and potatoes; and the 

 yield per acre of these several crops throughout a long 

 period of time, together with the rainfall of their 

 respective critical seasons, form the numerical data of 

 the investigation. 



The method of determining the critical seasons was to 

 find, by the use of the statistical theory of correlation, 

 the month or months, in the lifetime of the several 

 crops, the rainfall of which gave the highest correlation 

 with the ultimate yield. This preliminary inquiry 

 afforded a partial answer to our general question as to 

 the relation between rainfall and the crops. We found 

 that in case of each of the crops the yield per acre is 

 directly connected with the rainfall of some critical 

 period, and in all of the crops except oats the connection 

 is very close. It seemed probable, therefore, that since 

 the rainfall passes through definite cycles, and since 

 the yield per acre of the crops is intimately related with 

 the rainfall of their respective critical seasons, the yield 

 per acre of the crops should likewise pass through the 

 double cycle described by the rainfall of the critical 

 seasons. 



The investigation of the relation of the cycles of the 



