The methodology of diallel crosses is extended to provide procedures f
or the analysis of variance and estimation of effects when the recipro
cal crosses are included. The model of Gardner and Eberhart (Biometric
s 22: 439-452, 1966) was used for the analysis of variety dialiel cros
ses when reciprocal crosses are included. The least square procedure w
as used to derive the formulas for estimation of effects (components o
f means) and for the analysis of variance. It was found that the estim
ates of effects is the same if the analysis is performed according to
the original model, with varieties and the average of reciprocal cross
es to represent any hybrid combination. The standard errors of the est
imates of effects, however, change in the proposed procedure. The sums
of squares in the analysis of variance are different from those obtai
ned according to the original model. Some comparisons are made and an
example is given for illustration.