Evaluation of Synthetic Varieties of Maize for Yield1

Abstract

Nine maize synthetic varieties were crossed in a diallel series and evaluated for yield performance per se and in crosses. The synthetic varieties and crosses were grown at 3 locations in each of 2 years, and, on the basis of the combined analysis over locations and years, the 6 experiments were considered as representative of 6 random enviromnents.A nonorthogonal partition of the entry sum of squares obtained estimates of average, variety, and specific heterosis. F‐tests indicated that average heterosis was significant at the 1% level in 5 of the 6 experiments and for the combined analysis over experiments. Of the total sum of squares due to heterosis, 73% was attributed to average heterosis. Variety heterosis was significant in the combined analysis, but the proportion of the heterosis sum of squares due to variety heterosis was only 11%. Specific heterosis was nonsignificant.Estimated and observed heterosis were obtained for each of the synthetic varieties. Average observed heterosis on the basis of constant, mid‐parent, and high parent was 12, 11, and 6%, respectively. Estimated heterosis, obtained as average heterosis (h̄) plus variety heterosis (hj) for each variety, ranged from 6% for C.B. Syn. #1 to 14% for Ind. Syn. A. Average estimated heterosis (h̄) was 11%.Estimates of synthetic variety effects (vj), obtained for each variety by least squares analysis, included the performance of a variety per se and in variety crosses. The estimates of variety effects ranged from −5.7 for Ind. Syn. A to 5.2 for L.E. Syn., and formed three distinct groups. The relatively large, positive vj group included C.B. Syn. #3, S.S. Syn., and L.E. Syn. Pa. Int. Syn. had essentially zero variety effects. C.B. Syn. #1 and #2 and Ind. Syn. A. and B had relatively large, negative effects.The correlation between the predicted (Ŷj = μv + vj) and observed yield performance of the nine synthetic varieties was significant (r = .92**), as was the correlation between predicted [Cjj = μv + 1/2 (vj + vj) + h̄] and observed variety cross performance (r = .74**).