Comparison of Three Methods of Recurrent Selection in Maize1

Abstract

Concurrent recurrent selection experiments were conducted for five cycles to compare three methods of selection for grain yield in maize (Zea mays L.): (1) combining ability with an inbred line (Inbred Tester Method), (2) combining ability with the broad‐base parental population (Parental Tester Method), and (3) yield S2 lines per se (S2 Progeny Method). The 15 selected populations were then crossed with the parental population and with another broad‐base synthetic to evaluate changes in general combining ability, and also were tested as randommated and selfed populations per se.A significant linear increase in general combining ability over cycles was obtained for all methods, but the Inbred Tester Method was significantly more effective than the other two methods (4.4% gain per cycle, compared with 2.4 and 2.0% for the Parental Tester and S2 Progeny Methods, respectively). There was no significant difference among methods for yield of random‐mated selected populations when adjustments were made for differences in rate of inbreeding.The results indicate that the inbred tester used was homozygous recessive at many important loci. This would result in larger testcross variances and more successful selection of dominant favorable alleles than is possible with a broad‐base tester, which probably has intermediate gene frequencies at most loci. The slow progress in combining ability improvement with the S2 Progeny Method suggests that this method also was not effective in fixing dominant favorable alleles.