Stability and Performance of Hard Red Spring Wheat Bulks for Grain Yield 1

Twenty‐eight hybrid populations resulting from crosses among eight high‐yielding parents of hard red spring wheat ( Triticum aestivum L. em Thell.) were evaluated as F 1 's in one environment, as F 2 bulks in six environments, and as F 3 bulks in five environments. The average yields of the F 2 and F 3 bulks were similar to midparent performance across environments. Neither the F 2 nor the F 3 bulk generations exhibited superior yield performance in stress environments when compared to their midparents. Performance of the bulks varied from 95.9 to 110.9% of midparent across environments. The yield rank of each generation bulk was generally consistent from F 2 to F 3 (r=0.87, significant at the 0.01 probability level) since the same seven crosses (top 25%) would have been chosen with either generation yield test. The F 2 bulk yields also were associated with midparent yields (r=0.73, 0.01 level) and F 1 yields (r=0.73, 0.01 level). Parents were included in each trial, however, the Fa bulks were tested together with F 2 bulks in only two of the environments. The F 2 bulk yields, averaged over environments, correlated as well with the F 1 yields as with midparent. Regression analyses of response to environments for yield indicated that bulks had the same range of b values as parents. Average deviations from regression were smaller for bulks than for parents, but the deviations did not differ significantly. Parental deviations from regression for yield were correlated with gca (general combining ability) effects of their progeny (r=0.72, 0.05 level), but gca effects for response (b) were not significantly associated (r=0.49). Deviations from regression apparently were inherited from the parents in a relatively more predictable manner than response (b) to environments.