Abstract
The genetical expectations of the means, variances and covariances of populations of doubled haploid lines derived from f1, F2, F3 and intermated F2 (S3) generations are presented. These expectations are identical, regardless of genetical architecture, providing there is no linkage disequilibrium. In the presence of linkage disequilibrium differences will occur whose magnitude and direction will depend on the degree of disequilibrium, recombination frequency and the presence or absence of epistasis. Data from an experiment to detect linkage disequilibrium in a cross between two spring barley varieties are presented. This involved a comparison of means, variances and covariances of doubled haploid populations derived from the F1 and F2 generations using the H. bulbosum system. Linkage disequilibrium was detected for important agronomic characters and the effect of this disequilibrium on the choice of generation for doubled haploid production is discussed.