Genetic Components of Grain Yield Stability in Maize Diallel Crosses

In sub-Saharan Africa, maize (Zea mays L.) is produced in very diverse environments, which often results in complex genotype-by-environment (G x E) interactions. The aim of the study was to identify genotypes with broad and specific adaptation and high-performance environments that offer the best discrimination for selection and assess parents' contribution to hybrid stability. Sixty-six F-1 maize hybrids developed from a 12 x 12 half-diallel were evaluated at four locations in Kenya across two seasons to study the G x E patterns for grain yield and other agronomic traits. The genotype and genotype-by environment interaction (GGE) biplot method was used for graphical display of data. Stability of combining ability effects was examined using rank variance and Wricke's ecovalence. Environments contributed moderately (44%) to yield variability while the effects of G x E interaction (23%) and genotypes (21%) were almost equal. The eight environments were divided into two mega-environments, the first representing mediumtransitional locations, whereas the second mega-environment had medium-late locations. Hybrid 6 (Z419 x MUL114) and hybrid 22 (MUL71 x Osu23i) were the best genotypes for the first and second mega-environments, respectively. Hybrid 47 (CML509 x C92) was the most stable and highest yielding hybrid across environments and can be recommended for cultivation in both mega-environments. Inbred CML539 displayed the most stable general combining ability. Crossover interaction was found to exist, which signified the need to breed for both broad and specific adaptations.