Path analysis and expected response in indirect selection for grain yield in soybean

Abstract

This study aimed to determine the best auxiliary trait for indirect selection of soybean grain yield, through path analysis and in avoidance of the adverse effects of multicollinearity and expected response. Seventy-nine F5 soybean genotypes from the cross FT-Cometa x Bossier were used. The populations were distributed on the field was the families inserted with replicated controls. Primary and secondary traits of grain yield were evaluated in four phenotypically superior plants per family. The traits number of pods, height and number of nodes were considered as the most important, showing the best combination of direct effect and genotypic correlation. The number of pods achieved the highest expected gain through the estimation method based on the selection differential. On the other hand, plant height, by the method based on selection intensity, was not a good indicator of the most productive plants.