Molecular systematic study of Southern African Oxalis (Oxalidaceae)
The genus Oxalis forms a major part of the flora of southern Africa, in particular the Cape Floristic Region (CFR) at the southwestern tip of the continent, but the current taxonomy is outdated and ecological knowledge of the lineage is sadly incomplete. In this thesis I set out to address several aspects of Oxalis systematics that urgently require attention. Firstly, the current macro-morphological taxonomy requires phylogenetic testing, as it is acknowledged to be incomplete and artificial. I address this need by providing a DNA sequence-based phylogeny of three markers, using three different inference methods, for nearly three quarters of the indigenous species. This phylogeny confirmed both the monophyly of the southern African taxa, and the artificiality of the current classification system. It is congruent with previous sequence-based reconstructions of smaller groups of southern African Oxalis species, and with the palynological classification proposed for the genus. Secondly, previous phylogenetic work on the southern African members could not resolve basal relationships within the southern African clade. I attempt to address this problem by sequencing three extra chloroplast markers for a select group of taxa, followed by separate and combined (total evidence) molecular phylogenetic analyses. This approach did increase resolution at the base of the southern African lineage, but many clades still showed poor resolution and support despite the use of more than 7 000 bases of sequence data. Resolving these clades within the southern African Oxalis phylogeny remains a challenge, and should prove a fertile field for future research. Thirdly, the ages (and thus duration of presence) of many Cape plant lineages within the CFR are of major interest, given that the CFR represents a global biodiversity hotspot. The age of the genus in the Cape is estimated by analyzing combined sequence data for all sampled taxa under both a Bayesian Relaxed Clock and a semiparametric Penalised Likelihood method, using calibration points inferred from Relaxed Clock analyses of the entire order Oxalidales, for which fossil data are available. In an attempt to account for known problems with divergence time estimation, I explored the potential bias introduced by method used, marker genome source and different calibrations on the root. The results indicate substantial variation in the age of crown southern African Oxalis over a nearly twenty million year period, varying according to source data, calibration estimate and methodology employed in the reconstruction. Despite this major variability, all average estimates are older than iv 18 million years, which agrees with a growing body of evidence that there has been a gradual accumulation of floristic diversity in the CFR, rather than a rapid, recent burst of speciation. Fourthly, as the produced phylogenies conclusively show the artificial nature of the current taxonomy, I propose a new, almost completely different classification for southern African Oxalis taxa. Although a significant improvement, this classification is considered informal due to the complete disagreement between the old and proposed new taxonomies, poor resolution in some of the proposed lineages, and a need to confirm proposed groups (clades) with the identification of morphological synapomorphies. Potential synapomorphies for various clades are proposed and discussed, which should guide future research. Fifthly, the presence of bulbs in this genus is of great interest as a potential preadaptation for seasonally arid climates. The evolution of the bulbous habit in Oxalis is here explored for the first time. I address the sequence of major morphological character state changes leading to the suite of characters corresponding to the bulbous habit. The homology of basal leaf petioles, fleshy leaf scales and tunics is discussed, and it is shown that many bulb characters present in the southern African lineage are also found in the close relatives of this lineage, and are thus older than this lineage. The ecological and evolutionary implications of bulb geophytism in the CFR Oxalis are also discussed. Finally, I address certain taxonomic issues that arose during the course of this study. Co-authors and I describe the new species O. saltusbelli and O. ericifolia. We also clarify issues surrounding the tremendously variable group species O. flava and propose some nomenclatural changes and synonyms for related taxa. We also address the taxonomic position of the rare species O. purpurata, which was located too late in the course of this study to include in the main analyses.