Abstract
Comparisons were made between the genetic means and variances of a quantitative trait determined by 8 loci in simulated populations of lines derived by diploidizing haploids (DH) on the one hand and by single seed descent (SSD) on the other. In the absence of linkage no differences between the populations were observed, but when linkage was present, recombination was more frequent in the SSD populations as indicated by the relative differences in variance between these and the DH populations. In addition, differences in means between the populations derived by the two methods were observed when non-allelic interaction was present. The direction and magnitude of the differences in both means and variances depended upon the linkage phase, the recombination frequency and the presence or absence of interaction. The conclusion was drawn that the SSD method was to be preferred from theoretical considerations although in practice the choice of method will also depend upon practical and technical factors.