Path analyses identify indirect selection criteria for yield of late-planted soybean

Research aimed at increasing yields of late-planted soybean [Glycine mar (L.) Merr.] in the southeastern USA would be facilitated by identification of yield components that indicate plot yield (combine-harvested yield). Such yield components could be used as criteria in environmental and genetic research where plot yield cannot be obtained. Thus, our objective was to apply path analysis to previous studies to evaluate yield components for potential use as selection criteria (yield indicators) for future late-planted studies. Path analysis was applied to plot yield (response variable) and various yield components (predictor variables) for five late-planted studies conducted across 1987 to 1993 near Baton Rouge, LA (30 degrees N Lat). Studies were grouped into four data sets based on treatment factors: row spacing-plant population; row spacing; partial defoliation treatments during the reproductive period; and row spacing-cultivar. All studies involving row spacing, plant population, and partial defoliation showed that seed number was the best indirect selection criterion for predicting plot yield. However, in a study involving 12 cultivars, genotypic path analysis indicated pod per reproductive node as the best indirect selection criterion. Because of the greater inferential base of this study and the positive influence of pod per reproductive node at both phenotypic (path coefficient = 1.61) and genotypic levels (path coefficient = 4.31), greater attention should be given to this yield component for genetic improvement of soybean at late plantings.