values of bi will be different in each equation. For
this reason, the increase obtained from the additional
term is not simply bkZxky.
 The value of the sources of variation for the ANOVA
(Table VI-3) in the rice example are calculated as
follows:

Linear: bExly = 0.5416(4.875)
 = 2.603

Addition of quadratic term:

 blExly + b2Ex2y bExlY
 = 1.225(4.875) 0.228(13.0875) 0.5416(4.875)
 = 5.972 2.984 2.6403
 = 0.3477

Total Ey2 = 3.2483

Error: total linear quadratic
 = 3.2483 2.6403 0.3477
 = 0.2603


TABLE VI-3.


Analysis of variance comparing
Quadratic equations.


Source of Degrees of Sum of Mean
Variation Freedom Squares Squares
Linear
equation k-1 = 1 2.6403

Addition of
the quadratic
term 1 0.3477 0.3477

Error or
variation of n-k-1 = 3 0.2603 0.0868
the quadratic
equation

 Total n-1 = 5 3.2483

The mean square values in Table VI-3 are obtained in
usual way, by dividing the sum of squares by


linear and