Reciprocal Recurrent Selection and Modifications in Simulated Populations¹

Abstract

Computer simulation was conducted by the Monte Carlo method for a bisexual organism with two allels at each of 40 independently segregating loci. Reciprocal recurrent selection and two modifications were performed for the completely dominant and purely overdominant models at a number of starting gene frequencies. Genetic divergence is an unsound basis for selecting starting populations when the goals of selection are long term. A short term increase in the means of the recurrent populations cannot be used as evidence for a predominance of partially or completely cominant gene effects. Two modifications are important for the improvement of genetic potential and rate of progress for types of recurrent selection using progeny testing. 1) A synthetic variety should be made from all material entered into a moderate to long term recurrent selection program. 2) One generation of selling (or other inbreeding) should precede the test crosses in each cycle, where the progress per unit time indicates this is more efficient.