Phytogeography, endemism and diversity of the fynbos of the southern Langeberg

• Main Author: McDonald, David Jury
• Other Authors: Cowling, Richard M
• Format: Doctoral Thesis
• Language: English
• Published: University of Cape Town 2016
• Subjects: Botany
• Online Access: http://hdl.handle.net/11427/22502

Bibliography: pages 188-194. === The thesis presents a multi-faceted investigation of aspects of phytogeography, endemism and species diversity in the rich fynbos flora of the southern Lange berg, Cape, South Africa. The basis was a survey of the fynbos vegetation of the southern Langeberg carried out following the Braun-Blanquet or Zurich-Montpellier approach. Forty-six plant communities were described which were ultimately synthesized into 14 'community groups'. The core of the study consisted of three parts: (I) A multivariate analysis of vegetationenvironment relationships where questions concerning the distribution of fynbos communities and environmental controls were addressed. It was found that mean annual precipitation was the most important abiotic variable, apart from fir determining distribution of fynbos communities. A model was developed whereby distribution of floristically determined 'community groups' could be predicted from combinations of environmental variables. (2) The southern Langeberg is rich in endemic species and patterns of endemism were investigated using (a) two-way contingency table analysis and (b) generalised linear modelling. From the contingency table approach it was found that endemism is not randomly distributed taxonomically, in terms of biological attributes of species and with respect to the montane habitats where they occur. The majority of endemic species are restricted to 6.7% of the plant families represented and were significantly over-represented in high-altitude wet habitats. The use of generalised linear modelling was a new approach and extension of the contingency table method. It allowed numerical estimation of the probability that a species with a given set of attributes would be endemic. Seven inferences were made from the model but most importantly, an ant-dispersed, non-sprouting low shrub has the largest probability of being endemic. (3) Species diversity in the southern Langeberg flora was analysed firstly in terms of species richness (α diversity) and rarity and secondly with respect to turnover on ecological gradients (β diversity) and geographical gradients (γ diversity). The patterns of species diversity observed are related more to high turnover of sparsely distributed (rare) species on ecological and geographical gradients than to levels of a diversity.