PROCEEDINGS OF THE OHIO ACADEMY OF SCIENCE. 251 



show the general relation between two factors, but the exact 

 measure of the relation between two variables is best determined 

 by the correlation table as used by statisticians. This measure 

 is shown by the nearness of the correlation coefficient (r) to unity. 

 The values of r have been determined between the yield of corn 

 in Ohio and the rainfall for the different summer months and are 

 shown in table 2. 



Table 2. Correlation of the Rainfall with the Yield of Corn in 

 Ohio. 1854 to 1913. 



Correlation 



Coefficient Probable 



Period. (r) ± Error. 



June 0.12 ±0.09 



July 0.59 ±0.06 



August ...' 0.37 ±0.08 



June and July combined 0.48 ±0.07 



July and August combined . 67 ±0 . 05 



June, July, and August combined . 57 ±0 . 06 



This table makes plain that the rainfall for the month of 

 July has a far greater effect upon the yield of corn than either 

 the month of June or August, somewhat greater than for June 

 and July combined, and slightly greater than for June, July, and 

 August combined, but that the rainfall for July and August com- 

 bined has a greater favorable effect than for July alone. Similar 

 correlation tables made for both the four as well as the eight 

 largest corn States show the same dominating effect of the July 

 rainfall as compared with the other summer months. 



In order to ascertain what other or shorter periods than 

 calendar months might have a marked effect upon the corn yield 

 correlation tables have been worked out between the corn yield 

 in central Ohio and the rainfall for each period of 10, 20, 30, 40, 

 and 50 days from June i to August 31. The period giving the 

 highest value of r. in each case is given in table 3. 



