| Popis: |
In eucalypts, inbreeding may occur under open-pollination (OP) or when effective population sizes are reduced by intensive selection. Breeding strategies employing inbreeding have also been suggested. This Thesis examines inbreeding in three widespread and commercially important temperate eucalypts: Eucalyptus globulus, E. nitens and E. regnans. Inbreeding depression was severe for growth traits in all species. Flowering was also depressed in E. nitens, whereas pilodyn penetration and relative bark thickness were unaffected. In a long term E. regnans study, the survival of inbred progenies was significantly lower than outcrosses. However, the mortality of selfs obscured an increase in inbreeding depression with age for growth traits. In most traits examined, inbreeding depression was associated with levels of quantitative dominance, however, some traits exhibited inbreeding depression but no dominance, suggesting the action of rare alleles with major effect. Inbred progenies were virtually absent from OP families from closed native E. globulus forests, however, outcrossing rates were lower in families from isolated parents. Biparental inbreeding may also be important under OP in native stands, as mating nearest neighbour depressed growth equivalent to half-sib inbreeding. However, the selection against selfs observed in E. regnans, suggested most inbreds are eliminated by reproductive maturity thereby reducing the level of inbreeding in natural populations and the ability to purge genetic load. The fitness of parents under OP did not appear to be related to additive differences but may be correlated with outcrossing rate. However, it is argued that mixed mating is evolutionary stable, favoured by the reproductive ecology of the species. Heritabilities estimated from OPs were greatly inflated compared to heritabilities estimated from outcross progenies in E. regnans and E. nitens, particularly for traits that exhibited inbreeding depression. A reduction in outcrossing rate from 1.0 to 0.8 in E. globulus was associated with lower breeding value estimates, suggesting variability among families in outcrossing rates may inflate heritabilities estimates and bias breeding values. In the E. regnans study, inflated heritabilities and additive variances estimated from OPs approximated outcross parameters as inbreds were selected against. However, OP heritabilities were also reduced as stand development enhanced of variation within OP families. In addition, OPs overestimated age-age correlations and poorly predicted outcross performance. Mixed model methods were developed to account for dominance and inbreeding, and were applied to self and outcross progenies of E. regnans. Ignoring inbreeding depression biased breeding values for diameter, but dominance was unimportant. Accounting for rare deleterious alleles improved the model, however, breeding values predicted under a simple model were almost perfectly correlated with those under a more complex model. |
