High- vs. Low-Stress Yield Test Environments for Selecting Superior Soybean Lines
- W. F. Whitehead and
- F. L. Allen
An important factor in cultivar development is the choice of yield testing sites that permit the best discrimination among breeding lines. The objectives of this study were to test the hypotheses that: (i) there are genotypes that have superior yield potential in both low- and high-stress edaphic conditions, and (ii) yield tests conducted exclusively in low-stress environments can be used effectively for selecting such genotypes. Forty-eight random F4-derived soybean [Glyciue max (L). Merr.] lines were evaluated for yield over 4 yr on a low soil acidity, high fertility site (low-stress) and on a high soil acidity, low-medium fertility site (high-stress) at each of two locations (Crossville and Milan, TN). In addition, low-stress yield tests were conducted at five other locations during each of the last 2 yr. The grand mean of low-stress tests was 2.25 Mg ha−1 compared to 0.62 Mg ha−1 for high-stress tests. The 16 lines (33% selection intensity), which had overall mean yields above the grand means of combined high- and combined low-stress tests were designated as the superior lines. Simulated selections revealed that 32 to 54% of the superior lines would have been selected using either type of nursery at either the Crossville or Milan location each year. At least one of the top three lines would have been chosen in every case, regardless of type(s) of nursery environment, location(s), or year(s) used. These results indicate that the low-stress environments commonly used in soybean breeding programs should provide high probabilities for selecting genotypes that have superior yield potential in both low- and high-stress edaphic conditions. Regression coefficients for line mean yield regressed on site mean were generally highest for the superior lines (b ≥ 1.0), but the deviations from regression were similar to the non superiors. The ranges for maturity and plant height were similar for both groups indicating that selection for superior lines would not have led to extremes for either trait.
Copyright © . .