Genetic architecture of yield and earliness components in maize (Zea mays L.)
Eight parents (maize inbred lines) derived from diverse base populations of maize were used to develop twenty eight single crosses excluding reciprocals in a diallel mating system. The twenty eight single crosses and their eight parents were tested in a randomized block design at Embu Agricultural Research station (UMZ) and Murinduko substation (UM4). The experiment was planted during the long rain season 1985. The normal agronomic practices were followed. The main objectives were to compute the relationship between grain yield and its components, estimation of heterosis, combining ability, heritability and identification of the nature of gene actions for yield and earliness components. The data were collected on ten randomly selected plants for ten characters namely; number of days to 50% pollen shed, days to 50% silking, number of leaves per plant, plant height, rows per ear, number of kernels per row, ear length, ear weight, 100 kernel weight an~ grain yield per plant. Statistical analysis was carried out according to Panse and Sukhatme (1964) procedure for randomized block design. (xiii) The genotypes were found to differ significantly CP~ O.OJ) for all characters studied. The hybrids Embu 206 x Embu 107 had the highest grain yield per plant (163.40 ~) followed by Embu 206 x Embu 108 (157.75 ~ ) at Embu Agricultural Research Station. At the Murinduko site, the hybrid with the highest grain yield per plant was Tx5855 x Embu 107 (148.30 g') followed by Bs l3(1a) x Embu 108 (147.10 ~ ) Ear length and 100 kernel weight were found to be suitable yield components to be utilized for selection for grain yield per plant improvement. This is due to their high correlations with grain yield per plant, high heritability and ease of measurement of these traits. The studies showed that earliness and grain yield per plant were non-significantly correlated and therefore, it was possible to select for earliness without adversely affecting grain yield per plant. The general and specific combining ability findings suggested that both additive and non-additive gene actions were responsible for the inheritance of grain yield per plant, 100 kernel weight, ear weight, ear length, kernels per row, leaves per plant, days to 50% pollen shed and days to 50% silking. General combining ability contribution for plant height expression was non-significant and for rows per ear (xiv) expression specific combining ability was non-significant. The general combining ability effects of parents were not reflected in the specific combining ability effects of their crosses. The study indicated that the .~ ~ performance of a parent cannot be used as a measure of its general combining ability. Five parents namely; Embu 107, Embu 206, Embu 204, Embu 108 and TX5855 were identified as promising parents for hybrid development. Embu 107 had high general combining ability at both sites (UM2 and UM4) , suggesting that it can be used as a parent in synthetic development. The number of leaves per plant was positively and significantly correlated (P = 0.05) with the days to 50% pollen shed and 50% silking. It has a high narrow sense heritability and it would be a suitable ( trait when selecting for earliness since its response to selection is expected to be higher than when selection is based on days to flowering. The experimental findings suggested that combining ability, particularly specific combining ability and broad sense heritability were location dependent and ~~J not be generalized.