Dec. i, 1920 Permanence of Differences in Experimental Plots 341 



Table II. — Mean yields of Die Huntley uniform cropping experiment 



Crop. 



191 1, sugar beets 



19 12, total alfalfa. . . . 



1913, alfalfa I 



1913, alfalfa II 



19 13, alfalfa I and II 



1914, alfalfa I 



1914, alfalfa II 



19 14, alfalfa I and II 

 1914, alfalfa III 



1914, alfalfa I to III . 



1915, ear corn 



1916, ear corn 



1917, oat grain 



1917, oat straw 



1917, total yield . . . . 



1918, silage corn 



1919, barley grain 



1919, barley straw. . 

 19 19, total yield. . . . 



Number of 

 pounds 

 per plot. 



4, 179- 

 356- 

 54i- 

 483. 



1, 024. 

 489. 

 499. 

 988 



47i 



1, 460. 



522. 



396. 



555' 

 5 21 



i>°77 



3.175' 

 358. 

 230. 



Number of 



tons or bushels 



per acre. 



12. 29 

 1. 04 

 !-59 



1. 



4- 



5 2 - 



41. 



102. 



3. 10 



9-34 



43-8 



.67 



i-73 



The data furnished by this series of records are of particular value, since 

 (a) they are based on irrigated plots and (b) it is possible to compare the 

 correlations between the same crop and different crops in the different 

 years. t 



The correlations between the yields of the various crops in the different 

 years may be considered in three series. 



(1) The first comprises the yields for the whole plots. In this series 

 we determine the correlation between the crop produced on the 46 plots 

 in one year and that produced on the same 46 plots in another year. 



(2) In the study of certain crops the plots were divided into two sub- 

 plots, and we may determine the relationship between yield of individual 

 subplots in different years. Then the number of observations is twice 

 what it was in the preceding correlation, that is, N = 92 instead of 46. 



(3) Finally, in a more limited series of cases the 46 original plots were 

 harvested in 4 subplots each, thus increasing the number of units which 

 may be entered in the correlation tables to 184. 



The data for determining the correlations between yields of various 

 crops for the 46 whole plots are given in Table III. The data for half 

 plots and quarter plots may be obtained from the diagrams in an earlier 

 paper by Harris (3) on the practical universality of field heterogeneity 

 as a factor affecting plot yields. The correlation coefficients and their 

 probable errors for whole plots are shown in Table IV. 



