Genetic diversity and adaptation in Eucalyptus pauciflora

Restoration of degraded land to combat habitat degradation and deforestation requires understanding on adaptive potential of the species. Local adaptation and the geographic scale over which the local adaptation occurs raise issue on how well the existing genotypes will succeed in the face of increasing pressures from climate change and anthropogenic disturbances leading to new environment. This thesis examines genetic factors affecting the success of restoration plantings of the Eucalyptus pauciflora subsp. pauciflora on the island of Tasmania. Open-pollinated seed and DNA samples were collected from 281 trees from 37 native Tasmanian populations across the distribution and environmental range of the species and used to provide a quantitative and molecular genetics framework to understand local adaptation and guide future environmental planting decisions. It specifically aims to: i) determine the mating system parameters of E. pauciflora, and to explore whether population variation is related to the degree of forest fragmentation or altitude; ii) assess the spatial pattern of genetic diversity in chloroplast and nuclear molecular markers, to understand historical and contemporary barriers to gene flow; iii) explore climate adaptation of the species, through assessing quantitative genetic variation in seedling morphology and growth in a glasshouse trial; iv) determine the effects of inbreeding, local climate and translocation from mainland Australia on genetic variation in performance in Tasmanian field trials up to age 3 years; and vi) provide the seed collection guideline based on the above observations. Molecular research showed that Tasmanian E. pauciflora has a high outcrossing rate (tm = 0.90). Outcrossing rates differed among populations, but this variation was not correlated with the degree of forest fragmentation nor with altitude. Nevertheless, fragmentation did affect early reproductive output by reducing the number of germinants per gram of capsule content. Chloroplast haplotypes showed clear geographic structure suggesting three low-altitude glacial refugia and recent colonization of high altitude areas. There was little population differentiation in neutral nuclear markers, but populations within 27 km were more similar than average. Similar significant quantitative genetic structure was also detected in the glasshouse trial, suggesting an operational limit for the definition of a 'local' population. Population genetic variation was found for 24 of the 25 seedling traits studied. In several cases this population differentiation exceeded neutral expectations arguing for the action of disruptive selection and that local adaptation has over-ridden historical and contemporary gene flow. This is supported by significant correlations with population altitude and climate variables, with many seedling traits best related to the maximum temperature of the warmest month at the site of origin. Integrating mating system parameters into the analysis of the two field trials revealed inbreeding depression for growth at the family level, but at the population level outcrossing rate did not affect performance. However, population differentiation was evident for early-age growth, survival, and susceptibility to drought and herbivory. Population differences in early performance appeared to reflect a trade-off between fast growth and herbivore susceptibility, with low altitude populations initially growing faster but rapidly losing their advantage through increased herbivory. Drought and high temperatures at one trial reshaped the fitness profile of the planting, selecting against populations from more moist areas. At both trials the Tasmanian populations outperformed those from the mainland, arguing against the need for seed translocations from mainland Australia.