No abstract

Eight parental lines, 16 F1 hybrids, 16 2‐component blends of parental lines, and 16 2‐component hybrid blends of grain sorghum (Sorghum bicolor L. Moench) were grown in 9 environments over 2 years in Iowa.Standard analyses of variance and analyses for estimating stability parameters were made for grain yield and yield component data. A stable entry or population was defined as one with a regression coefficient close to 1.0 and deviations from regression near zero. A high mean yield also was deemed an essential attribute for an entry or population.The parameters for grain yield indicated that hybrid blends were the most productive and stable population type, although none of the populations were distinctly superior for all parameters. Hybrid blends had the highest mean yield over all environments and were second among the population types in terms of low deviations from regression and an average regression coefficient near 1.0. Data for the components of yield tended collectively to support the conclusion that hybrid blends were the most stable population type.A hybrid was the highest yielding individual entry in each environment, except one, where a hybrid blend was the most productive. Also, a hybrid had the highest mean yield among individual entries over all environments. Across all environments, 22 of the 32 blended populations yielded more than the pure‐stand mean of then: components. Yields for twelve of the 16 hybrid blends exceeded their mid‐component value. Collectively, the heterogeneous populations yielded 102% of the mean of their homogeneous components. Six of the blends exceeded the mean pure stand yield of their more productive component.

Comparative expressions of heterosis for 13 vegetative and mature plant characters were examined from 24 F1 hybrid populations of grain sorghum. Significant average heterosis was expressed by the hybrids for grain yield, mature plant height, stalk diameter and for measures of height and leaf number taken at a vegetative growth stage. Hybrids also exhibited longer and wider leaves giving more area per leaf, earlier blooming, more seeds per head, increased numbers of heads per plant and larger seed, but significance was not shown for the hybrid advantage in comparison with the parental means.Significant variation among hybrids for percentage heterosis was expressed for all characters except stalk diameter, seeds per head and heads per plant. Greatest heterotic response was observed for grain yield where hybrids averaged 122% of the mid‐parent values, with a range of 106 to 147% among individual hybrids.Variance analyses indicated that general combining ability effects were more often significant than were specific combining ability effects. However, significance for the s.c.a, component was shown for 6 of the 13 characters, and the mean square for parents vs. hybrids was significant for 5 of the characters, indicating that non‐additive gene effects were important in the expression of these traits.Inter‐character correlations among hybrids indicated that seeds per head was the character most highly associated with grain yield. Correlations between the mid‐parent mean and hybrid performance were significant for each character, but the coefficients for many of the characters were not large enough to indicate that mid‐parent values would be highly effective indicators of hybrid performance.

The performance of 40 F1 hybrid populations of grain sorghum, obtained by crossing each of 5 male‐sterile lines to 8 pollen‐fertility restoring lines, was evaluated from tests conducted at 3 locations over a 2‐year period. Measurements for grain yield, number of heads per plant, weight of 100 seeds, and number of seeds per head were analyzed to obtain estimates of the relative magnitude of general and specific combining effects. Significant differences among general combining ability effects of the lines were obtained for all characters, while significant differences among specific combining ability effects were expressed only for 100‐seed weight. An evaluation of variance components revealed that the variances for general combining ability for grain yield were three times larger than the component for specific effects. A similar ratio for the two effects was obtained for both number of heads per plant and weight of 100 seeds. For number of seeds per head, the ratio of general to specific effects was markedly larger than the ratio observed for the other characters. Specific combining ability effects were usually more stable than general effects over the environments of this experiment, particularly for grain yield and seeds per head.Correlations for the association of grain yield with its components disclosed that number of seeds per head was the component most highly related to yield, and that the expressions for yield were not affected appreciably by 100‐seed weight. While the ranges observed for number of heads per plant were relatively small in this experiment, significant negative associations for this trait with grain yield were observed for both the hybrids and parental lines.