Evolutionary patterns in Thamnochortus (Restionaceae) : a study of specification in the Cape floristic region

Master Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title

University of Cape Town

Patterns of speciation and potential evolutionary pressures and constraints were investigated in the genus Thamnochortus. Phenetic methods were used to define boundaries of species prior to cladistic analyses. Comparative techniques were employed to investigate aspects of dispersal biology and fire survival habit. Methods of historical biogeography were used to evaluate vicariance and dispersal hypotheses. The broader understanding of species evolution gained in such a comparative study is important in conservation of species or areas, forming a basis for further ecological and genetic predictions. The majority of Thamnochortus species have well-defined species limits; however, those of T. comptonii, T. platypteris and T. scabridus are more diffuse. For this species complex a matrix of 94 specimens, nine quantitative and sixteen qualitative characters was investigated, using cluster and ordination analyses, to define species boundaries. Thirty-four species of Thamnochortus, with three species of Rhodocoma as the outgroup, were used in the cladistic analysis. There were forty-three qualitative characters and ten quantitative characters. The number of species, height, reproductive output and geographic area were compared between sister lineages of seeding and resprouting species. In species classified as resprouters individuals survive fire by resprouting from the rhizome. In a post-fire environment seeding species recruit from seed and not by resprouting. Resprouters were significantly taller than seeders and covered a significantly larger distribution area. There was no significant difference in the amount of seed produced by seeding and resprouting lineages or in the geographic area covered by winged and keeled lineages. Correlated evolution tests indicated that wings of seeds evolved independently of the seeding condition, although the probability of wings evolving randomly was low. The evolution of keels was significantly associated with a switch to resprouting. There are few distinct ecological differences between the seeding and resprouting habits in soil type or rainfall; however, the inference is that resprouters do occupy habitats in higher rainfall areas than the sister seeders. Biogeographic analysis of species distributions, using cluster methods with a Jacard similarity coefficient, defined four phytogeographic areas which were considered to be areas of endemism. A concentric ring method recognised narrow areas of endemism and illustrated the overlap of species distributions between areas. The defined areas of endemism and similarity were used in general area cladograms to determine area relationships. The primary differentiations on the general area cladogram of areas of similarity distinguished a summer rainfall region (south coast) from a winter rainfall region (south Western Cape extending up the west coast). Within the winter rainfall region there is separation into a mesic (Cape Peninsula and south western mountain range) and an arid region (Cedarberg and Koue Bokkeveld). This analysis of Thamnochortus gives the first indication that the primary differentiation was between summer and winter rainfall, followed by the differentiation of the winter rainfall region into mesic and arid areas. Comparison within clades of distribution and habitat profiles indicated that, where distributions of closely related species overlap, there is niche differentiation in flowering time and substrate texture. Fire survival habit does not appear to have influenced speciation in Thamnochortus. There is, however, an evolutionary relationship between fire survival habit and female outer tepal specialization. Evidence from the general area cladogram indicates that speciation patterns in Thamnochortus may have been influenced by changes in rainfall in the Miocene. Habitat profiles of sister species indicate that alterations in flowering time and substrate texture are key factors in ecological differentiation of species.

Bibliography: pages 85-92.