88 
Journal of Agricultural Research 
Vol. XXVII, No. 
Adjusting yields to 3-row plats of which the rows to be adjusted are 
components may seem objectionable. In the experiment under consid¬ 
eration, however, inspection of figure 1 shows that the individual 
variation of A is of relatively little importance in determining T. It is 
the variation in A concomitant with that of the row on each side that is 
important, and T is superior to B because more of the independent 
y7££0S O/ 7 {Tj 
Fig. 2.—Correlation table for the percentage yields of individual rows (A) and 3-row plats ( T); the line of 
regression of A on T; and the mean values of A for each corresponding value of T. 
variation of the individual rows is equalized when three rows are averaged 
than when two are averaged. T is superior to C as a check because it 
responds more quickly, and more nearly approximates the extremes of A . 
the method of adjusting 
In adjusting the yields, each actual yield is divided by the correspond¬ 
ing predicted percentage yield. The latter are obtained from the regres¬ 
sion equation, in this case, a= i .02 jt, in which a and t are the deviations 
