Accuracy and simultaneous selection gains for grain yield and earliness in tropical maize lines

Winter maize is sown between January and March in Brazil. Although this maize is sown in unfavorable weather conditions, many farmers are successful, and winter maize has become an important crop. The sowing of early hybrids is a strategy to reduce the effects of stress on yield; however, low yields may result from earliness. Thus, the objectives in this study were to investigate tropical maize lines for the possibility of simultaneous selection for yield and earliness and to compare the differences among the simultaneous selection methods. Therefore, 64 lines were evaluated in two locations for grain yield, days to female flowering and grain moisture at harvest. The genotypic values for these traits were predicted using Restricted Maximum Likelihood/Best Linear Unbiased Predictor (REML/BLUP) single-trait (univariate) and multi-trait (multivariate) methods. Using three simultaneous selection methods (i.e., Additive index, Mulamba-Mock index and Independent culling levels) with two methods of prediction for genotypic values (single-trait and multi-trait), six simultaneous selection scenarios were considered and then compared for selection gains and accuracy. Because of the low correlation between these traits, the predictions of genotypic values were similar for single-trait and multi-trait methods. Thus, single-trait analysis should be prioritized because of its practicality. The Additive index obtained the highest selection gain for grain yield and simultaneously achieved good gains for days to female flowering and grain moisture at harvest. Therefore, the Additive index, using the single-trait prediction method, is the best simultaneous selection method for yield and earliness in tropical maize lines.