Evolutionary History, Genetic Diversity And Conservation Implications Of Selected Afro-Alpine Taxa

ABSTRACT

The tropical East African and Ethiopian mountains are famous for their exceptionally unique and high biodiversity. The flora on these mountains offers good examples of distinct adaptations to different altitudes as well as evolutionary differentiation, hence an ideal natural laboratory for studies on the dynamics of biodiversity. In this study the genetic diversity, evolutionary history and conservation implications of selected taxa occurring on these mountains were assessed. The objectives were to: 1) determine the level of intraspecific and interspecific genetic diversity of selected afro-alpine plant species, 2) explore the potential of AFLP markers for delimiting species and reconstructing the evolutionary relationships among some of the selected afro-alpine plants by comparing the results with those of previous morphological and molecular studies and 3) reconstruct the phylogeographic structure of the selected afro-alpine plant species.

Data for this study were collected from 1 ha (10000 m2) plots on 14 mountains around the region. Amplified Fragment Length Polymorphism (AFLP) markers were used to characterize the genetic patterns of the selected taxa. Six hundred eighty nine individuals from 154 populations (13 species, a total of 1168 AFLP markers with 97.9% reproducibility) of giant lobelias, 33 individuals, nine populations (two species, 172 AFLP markers, 97.86% reproducibility) of Deschampsia, and 153 individuals 36 populations (458 AFLP markers, 97.4% reproducibility) of Koeleria capensis were successfully analyzed.

Mean within-species (HT) and within-population (HS) genetic diversities were generally low across all species. Among the thirteen species of giant lobelias, the least diversity

iii

was observed in the most widely distributed species, L. giberroa (HT = 0.0751), followed by L. rhynchopetalum (HT = 0.0832). On the contrary, most diversity was observed in the narrow endemics L. bequaertii (HT = 0.2522) and L. thuliniana (HT = 0.2118). The low genetic diversity among L. giberroa populations may be attributed to bottlenecks following reduction of its montane forest habitat by human activities, which may have been less influential in the high-alpine Ruwenzori habitat of the local endemic L. bequaertii. There was however no correlation between the age of mountains and levels of genetic diversity, suggesting that the current populations on the older mountains originated from colonization episodes taking place long after their formation.

Except for Deschampsia spp. populations, the rest of the molecular and/or morphology-based recognized species were found to be genetically distinct. In Deschampsia, the individuals identified as the endemic D. angusta were not genetically distinct from those of D. caespitosa sampled in the same mountain, Ruwenzori, suggesting that the characters used to distinguish these species may reflect phenotypic plasticity rather than taxonomically significant variation. For giant lobelias, the relationships among species inferred from the primarily nuclear AFLP data were, with some notable exceptions, consistent with relationships earlier inferred from morphology and/or plastid DNA restriction site polymorphisms. High-altitude-restricted Lobelia species were intermixed with species occurring in the forest zone in the AFLP-based tree, supporting a main scenario of initial expansions of ancestral forest populations followed by parallel high-altitude adaptation and speciation in different mountain groups. However, the results did not support the proposed instances of hybrid speciation in this group while suggesting the most distinct intermountain vicarious patterns among the giant lobelias to be primarily high-alpine. For Koeleria capensis, there was neither distinct geographic structuring of the genetic variation nor support for recognition of infraspecific taxa. The results

iv

suggested that the afro-alpine populations of Koeleria capensis might have arisen by long-distance dispersal through Ethiopian mountains followed by intermountain dispersal into the tropical East African Mountains.

Given the current genetic structure and patterns, monitoring the most diverse and genetically most distinct populations of each species in situ and gemplasm tests for ex situ conservation are suggested in order to increase the probability for long-term survival of the studied plants and afro-alpine ecosystem at large. The study highlights that different afro-alpine species may have experienced very different phylogeographic histories and that long-distance dispersal among the isolated afro-alpine 'sky islands' can be more frequent than traditionally thought. Generally, the study demonstrates the need for further taxonomic exploration of the afro-alpine flora, in particular of taxa described as endemic.

This thesis is an outstanding contribution to knowledge as it provides for a refined evolutionary history and taxonomy of the previous morphology and molecular-based studies. For example the discovery that earlier proposed hybrid species L. bequaertii and L. bambuseti are actually not hybrids; the fact that Deschampsia angusta did not separate from D. caespitosa provides for a new idea that the previously known endemic D. angusta only from Ruwenzori mountains can actually be D. caespitosa; the knowledge about little genetic diversity within and among most of the studied species is a crucial contribution to conservationists for improved conservation strategies and; identification of areas that need further research such as the phylogenetic position of Lobelia thuliniana helps increase interests of natural biologists to work more on the studied taxa.