434 



THE AMERICAN NATURALIST [Vol. XLIX 



Ultimately, I hope to present experimental data of my 

 own bearing on this problem. For the present, the method 

 is admirably illustrated by a number of published records. 



II. Illustrations of Method 

 A. Cases in which the Combination Plots are Equal in 

 Size 



Illustration 1. Influence of substratum heterogeneity 

 on finld of experimental plots of mangolds. 



TABLE I 



Yield of Combination Plots for Mangolds, Obtained by Combining the 

 Entries of Map A by Fours as Indicated by the Heavier Lines 



193 | 196 



number of contiguous p plots to form a number m of C p plots. The su 



tracting the square of the mean jrielda of the ultimate plots, p, and tl 

 remainder divided by the square of the standard deviation of yields c 

 ultimate plots, <r p z. The quotient is the correlation between the yields oi 



required. Thus the c 



ulation of the c 



