Family Per Se Response to Selfing and Selection in Maize Based on Testcross Performance: A Simulation Study

theoretical response of yield in self-pollinated crops. Other researchers have addressed this relationship with Determining the appropriate number of lines to retain across generrespect to maize (Baker and Curnow, 1969; Rawlings, ations during inbred line development in maize (Zea mays L.) is interest during line development is performance of testcross progeny, Using a specific genetic model assuming additivity genetic information on lines per se adds to basic understanding of the within and among loci (no dominance, no epistasis) and genetic response of maize to inbreeding and selection. The objective of no linkage, Baker and Curnow (1969) simulated effects this research was to compare genetic responses of families per se to of population size on progress from selection in maize. varying selection intensities across early generations of selfing in In this simulation study, selection was aimed at improvmaize. A FORTRAN program simulating a stochastic genetic model ing the composite population as a source of inbred lines. with 30, 40, and 30% additive, dominant, and over-dominant loci, Rawlings (1979) discussed choices of effective popularespectively, and a random environmental effect was used to compare tion size considering short- and long-term objectives in varying combinations of selection intensities across generations. In the recurrent selection programs from the standpoint of simulation, individual plants were evaluated on the basis of testcross progeny performance using multiple testers and environments. Indi- gene fixation, maintenance of genetic variation, and vidual plants corresponding to top performing testcross progenies linkage equilibrium, and gave formulas for calculating were selected as parents for successive generations. Within each gener- appropriate population sizes and selection intensities. ation, phenotypic and genetic means of families per se increased with From a survey of U.S. maize breeders, Bauman (1981) increased selection intensities. Response to selection in any given found that, on average, U.S. maize breeders started with generation was not dependent on selection intensity used in the previ- approximately 500 S1 families and by the S4 generation, ous generation. Percent homozygosity and frequency of superior al- only 40 were retained. Bauman pointed out that breedleles increased as selection pressure intensified. Few original plants ers discarded 64% of the families in S1 generation based contributed to retained families with only approximately 5% of S0 on visual selection, indicating that both visual selection plants appearing in pedigrees of selected S4 families. Retaining few but larger families resulted in increased genetic values of families per and the breeder’s experience were important in selecse, compared with saving a large number of smaller families, even tion. Using procedures outlined by Robson et al. (1967), though selection was based on testcross performance. Johnson (1989) calculated the probability of correctly identifying S3 progenies on the basis of S2 testcross performance and found that critical decisions on allocation