The Great Escarpment forms a semi-continuous mountain system 5 000 km long, stretching from Angola in the north-west, south through Namibia, and into western, southern and eastern South Africa, including Lesotho and Swaziland. It is composed of a wide variety of geological suites but is unified in representing the edge of the African plateau and the passive Gondwanan continental margin. The Great Escarpment falls into all major climatic zones on the subcontinent, is a repository of palaeo- and neo-endemics, hosts more than half of southern Africa‟s centres of plant endemism, and has a rich suite of endemic fauna. In addition, the Great Escarpment is believed to be both a refugium and corridor for biological diversity. Despite the biological richness of the Great Escarpment, research to date has been fragmented and many sections of the Great Escarpment have not been studied. The aim of this study is to contribute to research on the Great Escarpment by undertaking a detailed floristic study of the southern Great Escarpment (the Sneeuberg, Nuweveldberge and Roggeveldberge). Together these mountains comprise approximately 1 000 km (one fifth) of the Great Escarpment, and occupy a transition zone between the summer rainfall zone in the east and the winter rainfall zone in the west. They are also the sections of Great Escarpment most closely situated to the Cape Floristic Region (CFR) and would thus be involved in hypothesised migration routes for lineages that also occur further north through the Drakensberg Alpine Centre (DAC) to the East African mountain chain. Detailed fieldwork of the southern Great Escarpment was undertaken over a period of four years in all seasons. Approximately 8 000 specimens were collected. Particular emphasis was placed on areas that may represent refugia, i.e. the highest plateaux and peaks, mesic areas and cliff-lines. An overview of each mountain range, together with their endemic plant species and phytogeography, is provided. Approximately ten new species have been discovered during this study, two of which have been described to date. Numerous endemics only known from their types have also been rediscovered. The Sneeuberg is defined as a new centre of plant endemism on the Great Escarpment (endemism of 2.3%), and the role of the Boschberg and Groot-Bruintjieshoogde (part of the Sneeuberg) as a nexus for floristic migration routes is discussed. The Nuweveldberge is shown to have low endemism despite a floristic tally similar to the Sneeuberg, while the Roggeveldberge are confirmed to be the most endemic-rich section of the southern Great Escarpment. The field data collected was augmented by available data in taxonomic revisions, and floras for the Sneeuberg, Nuweveldberge and Roggeveldberge were compiled. In order to floristically compare the southern Great Escarpment with other sections of the Great Escarpment and the CFR, a database of some 12 000 taxa was created using available floristic data for the CFR, DAC and Great WinterbergAmatolas, together with the data collated for the Sneeuberg, Nuweveldberge and Roggeveldberge. These data were analysed using phenetic methods and Parsimony Analysis of Endemicity (PAE). The results indicate stronger linkages in the east, particularly between the Sneeuberg and Nuweveldberge, and between the Sneeuberg and the Great WinterbergAmatolas. The relationship of the Roggeveldberge with the rest of the southern Great Escarpment remains ambiguous. In order to refine notions of connectivity and migration routes, 19 well-sampled phylogenies were assessed for sister-taxon disjunctions to explore CFRGreat Escarpment connections. Palaeo-connectivity between the CFR and southern Great Escarpment is most strongly supported for the south-eastern (SE) connection, and less so for the north-western (NW) and Matjiesfontein connections. There is support for the current (or recent) use of these three connections from numerous species that occur on both sides of the connections. Results of these analyses indicate that the southern Great Escarpment is a palaeo-corridor, the functioning of which has been broken by the aridification of the Nuweveldberge since the Last Glacial Maximum (LGM). Floristic connectivity is strongest in the east, from the Nuweveldberge to the DAC, and is less so in the west between the Nuweveldberge and the Roggeveldberge a finding attributed to the transition from a reliable winter rainfall regime on the Roggeveldberge to an unpredictable moisture regime on the Nuweveldberge. The mountains of the southern Great Escarpment are thus a series of refugia from a previous moister, cooler climate and are a corridor between the eastern and western components of the Great Escarpment. The SE connection is the primary link between the CFR and the eastern Great Escarpment Afromontane region in southern Africa. The implications of this research are that accurate conservation assessments and Red Data listings for many of the previously poorly-known endemics can now be made, and appropriate conservation measures implemented. Climate change remains the primary threat to these endemics and montane taxa in general, while degradation of wetlands is the primary threat to the water catchment service provided by the southern Great Escarpment. Future detailed research on the Great WinterbergAmatolas and Stormberg and a comprehensive flora of the HantamRoggeveldberge will further enhance our understanding of the floristics of the southern Great Escarpment, and provide the necessary data for comprehensive GIS-based models of proposed climate change scenarios for local, regional and national conservation planning.

Common to the temperate floras throughout sub-Saharan Africa is a group of taxa with strong ties to the Cape Floristic Region (CFR) (≈ Cape elements). Their distribution is limited to the eastern escarpment of Africa (e.g. the Drakensberg Alpine Centre - DAC), on nutrient-rich humic soils, as well as on isolated sandstone outcrops of low elevation, on nutrient-poor soils (e.g. the Pondoland Centre - PC), suggesting that intrinsic soil fertility is not the primary determinant of their distribution. The principal aim of this study was to determine which aspect of the edaphic environment of the DAC is most influenced by temperature, that may indirectly render it nutrient-poor and therefore provide suitable niches for Cape elements, as in the PC. A multidisciplinary approach involving aspects of plant biogeography, plant ecology, plant ecophysiology and soil chemistry was therefore adopted. The study regions were the DAC, PC and the KwaZulu-Natal Midlands. The flora of the DAC was resurveyed for this study, and is richer than previously thought: 2818 native taxa, most of which (2520) are angiosperms. The phytogeography of the DAC and PC is discussed, and comparisons are made with the floras of KwaZulu-Natal and the CFR. Their climatic environments, as well as those for the CFR and Sneeuberge, were compared using rainfall and temperature data from a range of sources. These climatic regimes were correlated with the floristic patterns of Cape elements for the high-altitude regions of South Africa and Lesotho. Altitude and rainfall increased, and temperature decreased, as the number of Cape elements increased towards the DAC. This study provided a contemporary inventory of the Cape elements of the DAC and PC. A total of 89 genera are recognised as Cape elements, of which 60 (c. 67%) are shared between the two regions. The highest number of Cape elements recorded for the eastern escarpment was the DAC (72 genera), with the highest number from all sites analysed being the PC (77 genera). The most Cape elements are contributed by the Asteraceae, Scrophulariaceae, Iridaceae, Fabaceae, Orchidaceae and Restionaceae, partly due to the success of annual aerial parts and their geophytic growth forms, which are convergent in these families. Further compartmentalisation into life and growth forms shows that most Cape elements of the DAC and PC are either ericoid (and sclerophyllous) or mesic herbs and shrubs. The ecological and ecophysiological aspects of this study involved the use of reciprocal pot experiments established along a gradient of altitude from coastal hinterland to mountain, that investigated the interactions between altitude, temperature and substrate on plant productivity in sites known either to support or to exclude Cape elements. Three soils were used at each site, representative of the DAC, PC and KwaZulu-Natal Midlands. The interactions between 'soil' and 'site' (≈ the climatic environment) were quantified using a temperate test taxon (Diascia) that has a strong Cape-centred distribution. Plant characters relating to morphology and nutrient content, and soil characters relating to fertility, were used as the basis for comparing treatment effects (soil-site interactions). Soil nitrogen availability was assayed using pot experiments with Eragrostis curvula (Schrad.) Nees. Wheat pot experiments revealed no Al³⁺ toxicity in 'Drakensberg' soil. Non-metric multidimensional scaling (NMDS) and redundancy analysis (RDA) indicated that all soil-site interactions were significant contributors to biomass differences, and that the Cape taxon performed poorly in the nutrient-rich Drakensberg soil at low altitude. Soil samples indicated that Drakensberg soil was the most nutrient-rich, and Pondoland soil the most nutrient-poor. Although total nitrogen in Drakensberg soil was six times higher than Pondoland soil, both soils mineralised similar low levels of nitrogen at their respective spring temperatures. The result for Drakensberg soil (simulated so as to include the effect of altitude) meant that only 1.7% of its total nitrogen was mineralisable at 12°C (its mean spring temperature). These findings suggest that nitrogen mineralisation rate is a key growth-limiting factor in the DAC, exacerbated by a number of complex interactions with soil pH and organic matter. It is hypothesized that Cape elements are preadapted to high-altitude habitats. These habitats are nutrient-deprived due to low temperatures, which reduce metabolic rates and the movement of ions in cold soils. This constraint imposes nutrient-related stresses similar to those of the CFR and PC. Taxa that are adapted to the nutrient-poor soils of the CFR are preadapted to the temperature-induced 'nutrient-poor' soils of the DAC and vice versa. This 'compatibility' has allowed the reciprocal exchange of taxa between regions, as suggested by cladistic biogeographical analyses using Cliffortia, Disa, Moraea and Pterygodium. The strong overlap of Cape elements between the CFR and PC is a product of similar nutritional niches and ancient floristic continuity. The result therefore is a high number of Cape elements common to the DAC and PC. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2004.

Analysis of the flora of the Pondoland Centre of Endemism (PCE) recorded 2253 species in the combined checklist of four sites (Port St. Johns, Mkambati, Umtamvuna and Oribi Gorge). Of these 1 % species are endemic to Pondoland, representing 8.7% of the Species, 15% of the genera and 26% of the families of the combined flora. Forty-four percent ofthe combined flora was only recorded from one locality (between 17% and 26% of each flora) and only 12% of the flora was present in all four localities. Of the endemics only sixteen (8%) occurred in all four sites thus each site had its own complement of unique endemics and 21 % endemics were not recorded from any of the four sites. At species level the floras of Mkambati and Umtamvuna were the most similar, followed by that of Umtamvuna and Oribi Gorge. Port St Johns had the least in common with any of the othersites, but more in common with non-neighbours Umtamvuna and Oribi Gorge than with its nearest neighbour Mkambati. Mkambati and Umtamvuna had the largest proportion of PCE endemics and Port St Johns had the lowest. The four sites are quite similar at family level, sharing thirteen families in the top ten family list between them, but much less similar at generic level.