352 Journal of Agricultural Research Vol. xx, no. s 



The average value for the nine pairs of correlations deduced from 

 the yields of whole plots is +0.159 for alfalfa and corn yield in 1915 

 but +0.708 for alfalfa and corn yield in 191 6. For the six pairs of 

 correlations which may be deduced for half plots the average of the 

 coefficients for the various yields of alfalfa in 191 3 and 1914 and the 

 yield of ear corn in 1915 is +0.181, whereas the average correlation of 

 the same yields of alfalfa with corn one year later is +0.729. Finally, 

 in the four cases in which it was possible to calculate correlations between 

 alfalfa and corn yields on the basis of data for quarter plots the average 

 for the correlations with ear corn in 191 5 is +0.159, whereas the con- 

 stants showing the relationship between the yield of alfalfa in 191 3 and 

 1914 and ear corn in 1916 give an average of +0.626. 



This more intimate relationship between the yields of alfalfa and the 

 second crop of ear corn does not necessarily mean that the corn crop of 

 1 91 6 was larger than that of 191 5 but merely that the variations in the 

 individual plot yields in 1916 are more dependent than those of 191 5 

 upon the yields of alfalfa during 191 2 to 191 4. As a matter of fact the 

 average yield in 191 5 was 522.6 pound per plot, while in 191 6 it was 

 396.2 pounds per plot. The greater yield in 191 5 may have been, and 

 probably was, due to factors other than soil conditions as such. 



It is of interest in this connection to turn back to the table of coeffi- 

 cients of variation of yield (p. 347) and to note that for whole plots, half 

 plots, and quarter plots the coefficients of variation of plot yield are 

 distinctly lower in 191 5 than in 191 6. This result is quite in line with 

 what one would expect if the fixed nitrogen of the varying growths of 

 alfalfa were not yet fully available in 191 5. 



There is also another possible explanation for the lower correlation 

 between the alfalfa yields and the yields of corn in 191 5. It is always 

 a difficult matter on the heavy soils at Huntley to break up alfalfa sod 

 and to get the soil into good tilth for the succeeding crop. It may be 

 that some of the plots in this field include heavier soil which ordinarily 

 ogives good yields but which was harder to get into good condition in 

 time for the 191 5 corn crop. If this were the case, these differences in 

 tilth might have been smoothed out by the season's cultivation so as 

 not to be expressed in the 191 6 crop yields. 



Some light may be thrown upon the problem of the residual influence 

 of alfalfa in the following manner. 



If the correlations between the plot yields of later crops be in a large 

 degree determined by differences in fertility referable to differences in 

 stand and yield of the preceding alfalfa crops, one might expect a closer 

 correlation between the yields of ear corn in 191 6 and oats in 191 7 than 

 between ear corn in 1916 and ear corn in 1915, since, as is shown above, 

 variations in the alfalfa yields have little influence until 191 6. This 

 will be true, provided there be a residual influence of the variations in 

 the yields of alfalfa such that these variations in fertility due to varia- 



