Quantitative Genetic Studies of the NP3R Random‐mating Grain Sorghum Population1

Abstract

Listen

Translate article

The need for broadening the germplasm base in breeding populations to avoid genetic vulnerability and the availability of the genetic male‐sterile genes have prompted the development of random‐mating grain sorghum [Sorghum bicolor (L.) Moench] populations and the use of recurrent selection systems to improve those populations. The objectives of this study were to investigate the inheritance of grain yield and other quantitative traits in one such population and to utilize the information to make decisions as to which breeding systems are the most likely to be successful in achieving population improvement.Nine quantitative traits were measured in 196 half‐sib, 196 full‐sib, and 196 S1 randomly developed families derived from the NP3R random‐mating grain sorghum population and grown in intermixed 14 ✕ 14 simple lattice designs at Mead and Lincoln, Neb. Genetic and environmental components of variance, heritabilities, phenotypic correlations, and expected gains from selection were estimated for each trait in each type of family.Dominance variance () exceeded additive genetic variance () for grain yield per plant and per unit area and for kernels per plant, but the reverse was true for other traits. The trend in grain yield with full‐sib family means > half‐sib family means > S1 family means indicates that heterosis and inbreeding depression are important for this trait. Heterosis is also indicated for plant height and number of kernels per plant. In total genetic variance, S1 families tended to exceed full‐sib families which in turn tended to exceed half‐sib families. The S1 family component of variance does not appear to be suitable for estimating additive genetic variance in the NP3R population. Heterogeneous half‐sib families tended to be most stable over environments, and the S1 families were least. Heritability estimates on an individual plant basis were high for days to flower (0.88) and plant height (0.71), very low for grain yield (0.09), and quite for heads per plant (0.16) and kernels per plant (0.15).Predicted response from single trait selection was highest for S1 family selection for most traits, particularly grain yield. Highly heritable traits, days to flower and plant height, could easily be improved by mass selection. If selection is for yield alone, correlations indicate that the population would become later maturing and taller; therefore, selection procedures must take expected correlated responses into account.