Genetic gain for soybean seed yield by three recurrent selection methods

This study compared genetic gain for soybean [Glycine max (L.) Merr.] seed yield associated with three strategies of recurrent selection in soybean population AP6. For the strategy of a two-stage test and one intermating generation (2ST-1IG), 100 F4-derived lines from 45 single-cross populations were evaluated in each of three maturity classes, and the 10 highest yielding lines were intercrossed to begin the next cycle. The conventional breeding strategy (CB-2ST-1IG) consisted of testing 100 FA-derived lines from 10 single-cross populations in each maturity class and intermating 10 selected lines to initiate the next cycle. For the strategy of a two-stage test and three intermating generations (2ST-31G), 100 F4-derived lines were tested in the midseason maturity class, and 10 selected lines were intermated three generations to form the population for the next cycle. Composites of equal numbers of seeds of the parents for six cycles of 2ST-1IG and CB-2ST-1IG, four cycles of 2ST-31G, and the individual parents of the most advanced cycles were grown at three Iowa locations during 2 yr. The genetic gain per cycle for seed yield was significantly greater for 2ST-1IG(53.8 kg ha-1) with one generation of intermating between cycles than for 2ST-31G (3.5 kg ha-1)with three intermating generations. The genetic gain per cycle was greater for 2ST-1IG than for CB-2ST-1IG, although the differences were only significant for the early maturity class. The superior genetic gain for 2ST-1IG indicated that selection from the largest number of single-cross matings possible should maximize yield improvement from recurrent selection in soybean