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.