Modelling the biological aspects of local endemism in South African Fynbos
- Cite this article as:
- McDonald, D.J., Juritz, J.M., Cowling, R.M. et al. Pl Syst Evol (1995) 195: 137. doi:10.1007/BF00989290
- 72 Downloads
The biological attributes, dispersal mode, growth form, and regeneration strategy were determined for the endemic and non-endemic flora of the southern Langeberg, Cape Province, South Africa.—Logistic regression analysis was used to assess the simultaneous effects and interactions between these biological attributes on the occurrence of endemism. The model allowed numerical estimation of the probability that a species with a given set of attributes would be endemic.—This approach extends a contingency table analysis of the data, which merely indicated the association between individual biological traits and endemism. Furthermore, the logistic model allows scope for the analysis of the influence of biological traits in determining endemism in other floras, and also tentative prediction of the probability of endemism in species with combinations of biological traits not yet observed in the flora of the southern Langeberg.