Genetic variances and heritabilities in a cloned Eucalyptus grandis breeding population of families derived from open pollinated selections were estimated and the results are presented. The genetic variance was partitioned into additive and non-additive genetic variance components that allowed the estimation of broad and narrow sense heritabilities. Predicted gains for breeding and production population options are discussed. The magnitude of the coefficient of relationship between sibs was shown to have a considerable impact on the estimate of variance components and the importance of understanding the level of relatedness in the population is highlighted. Growth traits (volume, diameter at breast height/DBH, height), stem from and disease tolerance were assessed at 38 and 66 months in each of the three separate trials established as sub-populations of the breeding population. The additive genetic variance was the largest proportion of genetic variance for the growth traits (84% for volume, 94% for height and 74% for DBH), whereas the proportion of non-additive genetic variance was notably higher for stem form and disease tolerance (37% and 46% respectively). The growth traits and stem from are, economically, the most important traits and a breeding strategy that exploits the additive genetic variation by selection to increase the frequency of the alleles causing the desirable genotypes is appropriate. The higher proportion of non-additive genetic variance for disease does, however, suggest higher gains (compared with the afore-mentioned strategy of selection for general combining ability) will be achieved by exploiting the non-additive variance by for example, selection for specific combining ability, using inbred lines, clones. The composition of the genetic variance was investigated separately in the F1 and F2 families to obtain an indication of whether or not there was a change in proportion of non-additive and additive genetic variance over the two generations. A notably larger proportion of non-additive variance was found for the growth traits and stem form among the F2 families. This is probably due to the reduction in additive variance through selection for these traits in the previous generations. No selection for disease took place in earlier generations and the proportion of non-additive genetic variance for this trait remains approximately the same over both generations. These results may indicate that with advanced generations of breeding in this population, that gains achieved through selection for additive variance will decline compared with that achieved in previous generations. A strategy for future generations that exploits non-additive variance may be appropriate. A high proportion of error variance was estimated and in situations such as these, cloning is particularly beneficial as is shown by the high clone mean heritabilities estimated in these trials. High mortality, resulting in fewer ramets per clone, erodes the benefit of cloning in these trials. The predicted gains showed the benefit of the cloned breeding population both in terms of breeding population gains and production population gains. Reducing the breeding cycle by bulking up clones faster will also increase gains per year. High gains in the production population were predicted, particularly for the selection of tested clones for deployment, which can be done at the same time as selections are made for the next generation. The benefit of the cloned population was therefore shown to be twofold, namely increasing the accuracy of within family selection and increasing the gains in the rapid deployment of tested clones and therefore facilitating the faster realization of predicted gain in the plantation.
Dissertation (MSc (Genetics))--University of Pretoria, 2007.