Performance of S1 Progenies from a Sorghum Random‐Mating Population Sampled in Different Years1

Abstract

Samples of 200 S1's each taken from the same random‐mating sorghum [Sorghum bicolor (L.) Moench] population in 3 different years were evaluated as S1 progenies for agronomic and grain protein traits for 2 years in a blocks‐in‐repllcatlons design to determine the effect of different seasons and different personnel on sampling. Means, family and family ✕ year variances, heritabilities, predicted gains from selection, and genetic correlations were determined from the three samples. The results were related to expected plant breeding progress, particularly recurrent selection. Significant differences were obtained among the three sample means for each of seven traits, except tillerlng. The ranges of differences between sample means were small, however, and of little plant breeding importance. Family variance components differed little among the three samples for all traits except hcight and tillering. These two traits also had relatively large family ― year variance components compared to their respective family variance components. Variability in height was attributed to drought that suppressed height during years of sampling but which was expressed during more favorable years of evaluation. Genetic correlations among the same pair of traits tended to be similar in the three samples, but there were four nonhomogeneous sets. Broad‐sense heritabilities, calculated from plot means, and predicted gains from selection for each trait were similar in each of the three samples and indicated that S1's sampled any year from the same base population could be used for family recurrent sclectlon with expectation of essentially the same progress.