Genetic improvement of soybean from populations developed by alternative recurrent selection strategies

Two aspects of a recurrent selection program that can be manipulated are the extensiveness of testing used to identify superior individuals and the number of intermating generations conducted among the superior individuals to form a population for the next cycle of selection. The objective of the research was to compare the performance of soybean Glycine max (L.) Merrill lines derived from populations developed by recurrent selection for seed yield with one or two stages of replicated yield testing to identify superior individuals and one or three generations of intermating among the selected parents. The base population, AP6, used for the alternative strategies of recurrent selection was derived from 40 high-yielding strains of Maturity Groups 0 to IV. For this study, the 40 original parents of AP6 and the lines used as the parents for the most advanced cycle of selection for each of the strategies were evaluated in four Iowa environments. The number of cycles of selection completed for the strategies under evaluation varied from two to five. Strategies with two stages of replicated yield testing to select superior individuals resulted in a larger number of high-yielding lines than when only one stage of testing was conducted. A larger number of high-yielding lines was obtained for strategies with one generation of intermating than when three generations were used. The results indicated that an appropriate strategy for recurrent selection of seed yield in soybean would involve two stages of replicated yield testing to identify superior individuals and one generation of intermating among the selected individuals to form a new population