Molecular evidence for reticulate evolution in the Tasmanian eucalypts

In 1971, eucalypt biologists Pryor and Johnson proposed that natural hybridisation had played a significant role in the evolutionary history of the Australian eucalypts. Their proposal was based on the fact that interspecific F rhybrids and regions of intergradation between species are quite common in Eucalyptus today, and contribute to difficulties in species delimitation. The past movements of hybridising eucalypt taxa through areas of contact, in response to climatic and environmental changes, may therefore have enabled the interspecific exchange of genes and possibly adaptations. This work uses a variety of phylogeographic markers to investigate evolutionary processes, including interspecific gene flow, among species of Eucalyptus subgenus Symphyomyrtus from the Australian island of Tasmania. Since much of the evidence presented in this thesis relies on the rp12-trnHpsbA marker region of chloroplast (cp) DNA, heritability of this region was checked in 425 progeny of Eucalyptus, comprising 194 progeny of 25 different intraspecific crosses of Eucalyptus globulus, and 231 interspecific hybrids (Fi , F2and backcrosses) between E. globulus and E. nitens. The results supported the use of this marker to trace the maternal lineage. An overview of cpDNA variability in Tasmanian species of section Maidenaria was obtained by sequencing the rp12-trnH region for multiple samples of all 17 species across their natural geographic ranges in Tasmania. The results showed that in Tasmanian Maidenaria, cpDNA correlates with geography, far more than with taxonomy. Many widespread species demonstrate intraspecific cpDNA variability, and this is coupled with extensive haplotype sharing between species within regions. A history of gene flow between species, in addition to shared ancestry, is the most likely explanation. Overall haplotype diversity across all species is lowest in central areas of Tasmania formerly occupied by alpine vegetation during glacial intervals, and in northern regions that were periodically linked to continental Australia by land bridges. The eastern region of Tasmania is significantly richer in haplotype diversity, consistent with the former existence of one or more glacial refugia. This finding is in agreement with both climate modelling and patterns of endemism. To further investigate the hypothesis of gene flow among species, a detailed phylogeographic analysis of two species was undertaken using cpDNA, nuclear ribosomal DNA, and introns of the cinnamoyl coA reductase (CCR) gene. The widespread forest tree, Eucalyptus globulus, and its rare congener, Eucalyptus cordata, share a cpDNA lineage in their region of sympatry in southern Tasmania. To clarify the reasons for cpDNA sharing, mixed and allopatric populations of the two species were sampled across their full geographic ranges. To distinguish between hybridisation and lineage sorting, the analysis considered both the geographic distribution of shared cpDNA haplotypes and their positions in the haplotype network for each species. The results supported at least five occurrences of cpDNA introgression from E. cord ata to E. globulus. Introgression appeared not to be mediated by the co-occurring species Eucalyptus vimirtalis/dalrympleana/rubida. Unlike cpDNA haplotypes, nuclear ribosomal DNA sequences showed local differentiation between species. Data from the CCR gene were consistent with cpDNA data, although some differences were noted. Southern Tasmanian populations of E. globulus were differentiated from adjacent northern populations and had high levels of a CCR haplotype that grouped with the dominant haplotype in E. cord ata. Within southern Tasmania, levels of this haplotype were highest in individuals with the same cpDNA lineage as E. cordata. However, this result could not be unambiguously attributed to introgression, as opposed to lineage sorting. This is the first use of a single copy nuclear gene to investigate evolutionary processes in Eucalyptus, and highlights some advantages and problems inherent in the use of nuclear genes for phylogenetic and phylogeographic research. The overall picture emerging from this research is that a complex combination of historical factors has contributed to geographically structured genetic variation in Tasmanian Eucalyptus. Despite the complexity of the pattern, it has been possible to partially resolve it, and to show that hybridisation is one of the contributing factors.