Recurrent Selection for Grain Yield in a Maize Population 1

Mass selection for improved grain yield was conducted in the ‘Krug’ (BSK) maize ( Zea mays L.) population. Effectiveness of 14 cycles of mass selection for grain yield in BSK was measured in nine environments using the original (CO) and the even‐numbered cycles of selection (Cn) as populations per se and two sets of testcrosses (Cn ✕ CO and Cn ✕ B73). Unselected S 1 lines derived from the original, the 14th cycle of mass selection, and the eighth cycles of half‐sib and S 1 progeny selection were evaluated in four environments to estimate the effect of the three recurrent selection procedures on genetic variability and on grain yield improvement of S 1 progenies per se. Linear response per cycle of mass selection for grain yield was obtained for the populations per se (0.024 ± 0.009 Mg ha −1 ) and testcrosses with the CO population (0.026 ± 0.010 Mg ha −1 ) and B73 (0.028 ± 0.010 Mg ha −1 ). Average yield response to mass selection in BSK per se was 0.49% per cycle or 6.9% after 14 cycles of mass selection. Increased yield was accompanied by later flowering, increased root and stalk lodging, increased grain moisture at harvest, and higher ear placement. The mass selection methods we used for improving grain yield in BSK apparently require modification to include selection for other agronomic traits, but such modifications would increase the complexity of mass selection and perhaps reduce gain for yield. Estimates of genetic variability among S 1 progenies for grain yield showed a decrease in genetic variance for the S 1 and half‐sib derived populations for most traits and no change for the mass‐selected population.