Predicted Response to Recurrent Selection with Intra‐ and Inter‐Varietal Testers in Corn ( Zea mays L.) 1

' I 'HE type of gene action for yield and other quantitative •*• characters in corn has been the subject of considerable discussion among corn geneticists for a number of years. The various questions raised have important implications relative to choice of breeding procedures providing maximum gains from selection. If heterozygote superiority (over-dominance) is of most importance for yield, as suggested by Hull (4, 5), then selection for specific combining ability is likely to be the most effective procedure. If additive genetic variance is present in significant amounts in corn populations, as suggested by many recent experiments, intra-population selection and selection for general combining ability may be more effective, at least initially. Reciprocal recurrent selection, proposed by Comstock et al. (2), is a compromise scheme which, theoretically, is effective regardless of the type of gene action predominating. The experiment described in this paper was designed to provide information for an initial cycle of reciprocal recurrent selection. At the same time, information was desired which could lead to theoretical considerations of comparative intraand inter-varietal performance. The principal objective of this paper is to present predictions of performance of future populations reconstituted on the basis of such information and to illustrate the use of quantitative genetics in making plant breeding decisions when various alternative schemes are available.