Estimates of Quantitative Genetic Parameters m IAP2B(M) Random-Mating Sorghum Population

Random-mating populations of sorghum [Sorghum bicolor (L.) Moench] are used extensively as germ plasm in the development of parental lines for hybrids. Populations improved by recurrent selection provide a useful source for the selection of superior parental lines. Our objective was to determine the effectiveness of mass selection for threshed-panicle weight by comparing the performance of lines (S1 families) chosen randomly from CO and C4 of sorghum population IAP2B. Sixty S1 families from each cycle were grown in five Iowa environments. Four cycles of selection for panicle weight shifted the mean for grain yield favorably, and the range among families was extended. Increased values in C4 also were observed for seeds/panicle, 100-seed weight, days to midbloom, and plant height, but panicles/plant decreased. Genotypic variance among S1 families increased from CO to C4 for grain yield and days to midbloom. Heritability estimates for grain yield also increased from CO to C4. Expected gain/year for grain yield by using S1 family selection increased from 2.0% of the mean for CO data to 3.2% for C4. Correlation coefficients indicated that selection for individual-panicle grain weight also influenced other characters. Correlations changed substantially in a positive direction from CO to C4 for grain yield vs. seeds/panicle and grain yield vs. panicles/plant, but they changed in a negative manner for grain yield vs. 100-seed weight. IAP2B(M) should prove useful for developing female parents with a wide array of gene combinations. Higher estimates of genotypic variance and heritability for grain yield in C4 compared to CO, as well as greater expected gain in C4 from S1 family selection, indicated that further improvement of the population should be expected with additional cycles of selection