Accumulations of plant macrofossils in lake sediments provide useful tools for identifying past local presence of plant species, community composition, and vegetation dynamics. However, plant macrofossils have rarely been used in the reconstruction of Lateglacial-Holocene vegetation changes in the Southern Hemisphere. Furthermore, such macrofossil studies are hampered by a scarcity of studies of taphonomic biases that may affect the final representation of plant species in sediments. This thesis is the first systematic study of macrofossils covering an almost continuous stratigraphic sequence from the Lateglacial to the Holocene period (last 15,000 years) in southern Australia. Being located in a transitional climatic and vegetation zone near the treeline in south-central Tasmania, Lake Dobson is a lake formed in a glacial cirque which provides an ideal system to investigate both the taphonomy of Australian subalpine plants, and changes in vegetation and climate extending back to the Lateglacial. The first component of the thesis is a taphonomic study comparing recently deposited leaf types from modern sediments of Lake Dobson with the vegetation surrounding the Lake. This study suggests that proximity of the plant community to the depositional site is critical for the representation of species in sediments. Additionally, the floristic composition of plant assemblages in sediments strongly agrees with the composition of species in vegetation near the lake. However, the number of leaves in sediments does not predict the same abundance of the species in the standing vegetation, mainly because of the large differences in leaf size and rate of leaf production that exist among tree and shrub species. Overall, the results suggest that a correction factor accounting for proximity of the plant community to the site of deposition, differential leaf size, and number of leaves per ground area of vegetation produced by individual species can allow for better reconstruction of the original forest community. The second component of the thesis analysed the floristic composition of plant macrofossils from a core drilled in the deepest part of Lake Dobson in south-central Tasmania, Australia. This study provides evidence of the role of local glacial survival versus the postglacial colonisation of plant species in the assembly of treeline communities during a period of acute environmental changes: the transition of the last glaciation to the modern Holocene environments. In particular, the presence of high subalpine/alpine plant species in the oldest macrofossil-bearing sediments in the Lake Dobson core strongly support the local survival ‚ÄövÑvÆthrough at least the last part of the glacial‚ÄövÑvÆ of poorly dispersed plant species such as species of alpine conifers, and the cool temperate tree Nothofagus cunninghamii. Additionally, this fossil record also provides evidence of the time lag in migration of Eucalyptus around the early middle- Holocene boundary. Thus, the current flora is best explained as having been assembled from a mixture of species that survived the glacial locally and species that migrated upslope after the climate amelioration in the Holocene. This study represents the first quantitative study of macrofossils in Australia for this significant period of environmental change. The final chapter of the thesis explores the potential for anatomical aspects of leaf macrofossils to be used as indices of forest structure (in particular, closed forest versus open vegetation). In particular, it investigates the morphological and anatomical leaf variation of both fossil and contemporary leaves (i.e. canopy and litter leaves) of the cool temperate tree species Nothofagus cunninghamii from Lake Dobson. Overall, the results of this study suggest that allowing for the effects of leaf size and stomatal density in comparisons of vein density is a useful tool for reconstructing the structure of the vegetation from fossil leaves. In particular, it was found that closed canopy leaves have lower vein densities relative to stomatal density compared to open canopy leaves once differences in leaf size are taken into account. The fact that the relationship is dependent on leaf size and stomatal density is important because it means that canopy structure may be inferred from fossil leaves even when those leaves have undergone shrinkage in the process of fossilisation. However, the use of these parameters in the differentiation of open and closed litter it is not straightforward, and further research is required to clarify the higher upper to low canopy leaf representativity in the litter. This approach, however, represents an improvement on prior methods that directly employ vein density and have no capacity to allow for the effects of shrinkage. The overall findings of this thesis have direct application to complement vegetation reconstructions by means of pollen analysis in southern temperate Australia. In particular, the study provides better understanding of taphonomic factors affecting the fossilisation of plant remains, and important implications to understand plant responses from a period of climate variability that is not yet well understood in the Southern Hemisphere: The transition from the Last Glacial to the modern Holocene environments.