M.Sc. / Invasive species are a considerable threat to ecosystems globally, especially on islands where species diversity can be relatively low. Understanding the drivers of invasion is the first step towards an adequate management plan. Although Darwin’s naturalisation hypothesis has fuelled our understanding in this regard, several studies provided mixed results, suggesting that invasion success might be context-dependent. The main objectives of this study are two-fold: (1) testing Darwin hypothesis on Robben Island, and (2) investigating the relative role of invasive alien plants on phylogenetic diversity (PD) loss in native community. I sampled extensively the flora of the island, and using a Bayesian analysis, I reconstructed its phylogeny based on two plastid DNA loci, rbcLa and matK. I also surveyed a total of 127 plots of 50 x 50 m (i.e. local communities) where species presence/absence was recorded. Analysing phylogenetic patterns of the native and invasive floras at both regional (phylogeny level) and smaller scales (plots level), I found that invasive species are, on average, more distantly related to the native communities, giving strong support to the hypothesis tested. Furthermore I found that native communities have accumulated lower PD than alien communities; and that local communities are more overdispersed than expected. These findings suggest that competitive interactions might be the major ecological forces shaping plant communities, with the possibility of alien being higher competitors than native, and therefore decreasing native plant diversity. The implications of these findings for the recovery of native plants are also discussed. Key words: Invasion biology - Darwin’s naturalisation hypothesis - Phylogenetic diversity - Community structure - Conservation - Robben Island, South Africa.