Recurrent Selection for Yield in Soybeans

Abstract

Procedures have been developed for yield testing of soybeans (Glycine max (L.) Merr.) that permit S₁ testing with 1 year per cycle and S₄ testing with 2 years per cycle. The principal factors in the program include efficient procedures for obtaining crossed seed by hand pollination, crossing and generation advance in Puerto Rico, and yield tests in Iowa by use of hill plots. Details for each of the three factors are discussed. With the facilities available, expected genetic gain per year for yield was calculated for S₁ testing without male sterility (S₁), S₁ testing with male sterility S₁(_ST), S₄ testing without male sterility (S₄), and half-sib family selection with male sterility (HF). The greatest expected gain per year was obtained with S₁ testing. There is more genetic variability among S₄ lines than among S₁ lines due to inbreeding, but expected genetic gain is less for S₄ testing because 2 years are required per cycle. S₁(_ST) had less expected gain per year than S₁ or S₄ testing because 2 years are required per cycle if homozygous male-fertile genotypes are evaluated. Expected genetic gain was least for HF because less genetic variability is expressed among lines than for the other three methods.