Measuring Seed Yield in Soybean Populations Segregating for Male Sterility 1

Using genetic male sterility to facilitate intermating in a recurrent selection scheme requires that either the male‐sterile allele be removed from the population before yield testing or that the effects of male‐sterile plants be known and considered in evaluating seed yield. This study was conducted to determine the effect of male‐sterile plants on seed yield and to test methods of estimating the yield of fully fertile plots from data collected on plots segregating for male sterility in soybean [ Glycine max (L.) Merr.] populations. Nine F 2 populations were grown in both hill and row plots. Each population was tested with and without segregation for genetic male sterility. Plots were established, in a Flanagan silt loam (fine, montmorillonitic, mesic Aquic Argiudolls) in 1981 and a Drummer silty clay loam (fine‐silty, mixed, mesic Typic Haplaquolls) in 1982. Data were collected on maturity, seed yield, and number of fertile and sterile plants. The yield of plots segregating for male sterility was adjusted by four procedures: (i) mean yield of fertile plants ✕ total number of plants, (ii) plot yield divided by 0.75, (iii) least square means (LSM) calculated using total number of fertile plants as a covariate, and (iv) LSM calculated using total number of fertile plants and total number of plants as covariates. Seed set was very low on male‐sterile plants and the presence of male sterility significantly reduced yield. Methods of adjusting the yield of plots with male‐sterile plants were more successful with hill plots than with row plots. The presence of male sterility is a complicating factor in measuring seed yield and the statistical procedures tested did not successfully remove the detrimental effect of barren or near‐barren plants.