Analysis of responses to recurrent selection in the BS11 maize population

Abstract

Recurrent selection methods have been effectively used by maize (Zea mays L.) breeders to improve the performance of maize populations for quantitatively inherited trails. Although theoretical comparisons of such methods can be easily made, direct comparisons of the efficiencies of methods are time consuming and laborious. Because of these limitations, empirical data comparing multiple recurrent selection methods in the same base population are lacking for maize. Our study was designed to compare the response to selection for seven different (six intraand one inter-population) methods in the BSll maize population. minimum of four cycles of selection were conducted for each of the following methods: mass, modified ear-to-row. half-sib with inbred tester, full-sib, S^-progeny, S,-progeny, and reciprocal full-sib selection. Selections for all programs except mass and reciprocal full-sib were based on an index composed of grain yield, grain moisture, stalk lodging, and root lodging. Each trait in the selection index was weighted by its heritability. The populations per se, populations selfed, and testcrosses of the populations to the Cycle 0 population and to inbred B79 were evaluated in our study. Response to selection was measured for grain yield, grain moisture, stalk lodging, and root lodging along with other agronomic traits. All selection methods were successful in significantly improving the population per se performance for grain yield. S,-progeny selection had the greatest response for grain yield of 4.5% cycle' and mass selection had the lowest response (0.6 % cycle'). All selection programs in which index selection was practiced, except for modified ear-to-row, were successful in improving the populations per se for ail four traits simultaneously. Unlike some previous studies, inbred-progeny selection methods (S,, S,) performed well in BSl 1 in comparison with other selection methods.