256 PROCEEDINGS OF THE OHIO ACADEMY OF SCIENCE. 



has been above the normal twice and below the normal 14 times. 

 This is a probability of 88% that the yield will be below the 

 normal if July is very warm. When the temperature has been 

 above the normal and the rainfall more than one inch either 

 greater or less than the normal the yield has always been below 

 the normal. When it has been cool and wet the yield has been 

 above the normal 12 times and below only three times. A cool 

 and dry July has just as many yields above as below the normal. 



A similar dot chart made by combining July and August 

 shows that warm and dry weather for the two months is decidedly 

 unfavorable, while cool and wet weather for the two months is 

 decidedly favorable. 



Table S shows the results of a correlation of the })otato yield 

 in central Ohio with the rainfall and temperature for each ten 

 days from .June i to August 31. This indicates that the period 

 from July i to 10 is the most important ten days, for potatoes, 

 as regards both temperature and rainfall. 



Table 8. Correlation of the Yield of Potatoes in Central Ohio 



WITH Temperature and Rainfall for each Ten Days. 



1891 to 1910. 



Correlation. Coefficient (r) 



Period. Temperature. Rainfall. 



June 1 to 10 —0.12 0.29 



June 11 to 20 —0.17 0.32 



June 21 to 30 —0.28 0.16 



July 1 to 10 —0.44 0.48 



July 11 to 20 —0.33 —0.29 



July 21 to 31 —0.33 —0.12 



August 1 to 10 —0.23 0.06 



August 11 to 20 —0.36 0.87 



August 21 to 31 —0.38 —0.26 



While most of the rainfall values of r are not high, it is 

 interesting to note that part of them are negative and part 

 positive, showing that part of the season rain is needed and part 

 of the time that it is a detriment. This condition persists through- 

 out similar correlations for 20, 30, 40, and 50 day i)eriod cor- 

 relations. 



A correlation between the yield of potatoes with the tern- 



