Heterosis and combining ability of transitional highland maize (Zea mays L.) inbred lines

Twenty-three genotypes (5 lines, 3 testers and 15 crosses) were evaluated using randomized complete block design with two replications during 2003 cropping season at Ambo and Kulumsa. One standard check, BH660, was also included in the study. The objective of the study was to determine heterosis, combining ability and heterotic patterns in the crosses of maize inbred lines and testers using line by tester analysis. Data for grain yield, number of kernel rows per ear, number of kernels per row, ear length, ear diameter, number of ears per plant, 1000 kernel weight, days to tasselling, days to silking, days to maturity, plant height, ear height, number of nodes, inter-node length, stalk diameter, leaf area, total biomass yield and shelling percentage were recorded. Significant differences between the mean of parents and crosses were observed in all traits, except in days to maturity, leaf area and shelling percentage at both sites and across locations, in stalk diameter at both sites and days to silking and biomass yield at Kulumsa. Across locations, L sub(4)x T sub(1) (9258.8 kg/ha) followed by L sub(1)x T sub(1)(8635.7 kg/ha) and L sub(2)x T sub(2) (8261.6 kg/ha) were the top-yielding crosses which could be utilized for future maize breeding activities. Most of the crosses out-yielded the standard check (BH660) at both locations and across locations. Relatively, earlier tasselling and silking were recorded for crosses as compared with the parents, while among the parents, lines were earlier in maturity than testers but testers had higher mean value than that of lines in plant and ear height, number of kernels per row, internode length and 1000 kernel weight. Ecuador tester (T sub(1)) performed better in most of agronomic traits while Kitale tester (T sub(3)) was earlier in maturity traits. L sub(3), L sub(1) T sub(3) were good general combiners for plant and ear height reduction. The mean square due to specific combining ability (LxT) was significant for grain yield and leaf area at Kulumsa. Maximum SCA effect for grain yield was obtained from L sub(4)x T sub(1) and L sub(1)x T sub(1). Generally, the magnitude of mean squares due to GCA was higher than that of the SCA for yield and other agronomic traits indicating the predominance of additive gene actions in controlling the inheritance of these traits. Proportional contribution of GCA of lines to total variance was higher in yield and other agronomic traits. Thus, additive genetic variance was more important than non-additive gene action in controlling the inheritance of most of the characters in this study.