Diallel Analysis of Four Maize Traits and a Modified Heterosis Hypothesis

Diallel analysis of quantitative traits has contributed greatly towards improving crops and understanding of heterosis. Nevertheless, the mechanisms of heterosis are still unclear. The objectives of this study were to: (i) investigate whether reciprocal crosses and environments affect heterosis of quantitative traits, such as maize (Zea mays L.) grain yield (GY), kernel number per row (KR), kernel rows per ear (RE), and 100-kernel weight (KW); (ii) determine whether heterosis of a composite quantitative trait (CQT), such as GY, is impacted by heterosis of its inherent component traits (ICTs; i.e., KR, RE, and KW); and (iii) evaluate if current hypotheses and/or theories of heterosis need to be modified. Data were obtained from a 12-parent diallel cross (F 1 and reciprocal F 1 ). Heterosis varied among crosses, between reciprocal crosses, and across environments. Further, heterosis of GY, a CQT, was greatly impacted by heterosis of its ICTs. Nonsignificant variances for females × males × year and specific combining ability (SCA) × year interactions for heterosis strongly suggested that the gene expression in both parents and their crosses might be similar or that the same genes from two parental lines might complement each other in crosses and offer additional choices for allele selection in hybrids, which could make hybrid performance stable across environments. Study results helped us propose a modification of Goff’s ‘energy-efficient allele selection’ (EEAS) model. We discuss how the modified model not only explains what previous heterosis theories can explain, but additionally how it can explain certain aspects that previous heterosis theories are unable to explain.