Bibliography: leaves 25-31. / The Red-tailed Group of francolins falls within the class Aves, order Galliformes, family Phasianidae and genus Peliperdix. These are small francolins consisting of three putative species (Peliperdix coqui, P. albogularis and P. schlegelii) that are largely allopatric. Two species (P. coqui and P. albogularis) are considered to be polytypic species, but is a large discrepancy in the number of subspecies attributed to P. coqui and P. albogularis. The mitochondrial cytochrome b gene, organismal characters and combined data were used to identify diagnosable taxa, test the monophyly and reconstruct the phylogeny of the Red-tailed Group. Sixteen exemplars of the Red-tailed Group (representing all species and nearly all subspecies currently recognised in the genus Peliperdix) from different geographical localities were studied. Maximum likelihood (cyt b), maximum parsimony (cyt b, organismal, combined) , distance analysis (cyt b) were performed to resolve the phylogenetic relationships among them. The trees obtained from DNA sequence, organismal and combined data were incongruent in respect to the position of some taxa. The monophyly of the Red-tailed Group seems to be well supported, but the identity of, and interrelationships between the subspecies and species are less well resolved. The morphological and combined tree probably reflects the taxon phylogeny better than the tree based on mitochondrial DNA. only. The most remarkable result that is strongly supported by both organismal and combined trees was that they bring clear resolution between P. albogularis and P. schlegelii as two separate species from P. coqui. Due to the short number of DNA sequences obtained, one cannot make a decision as to whether the subspecies should be elevated to species. Finally, it is suggested that more systematic studies must be done based on multiple independent data sets in order to obtain a robust taxonomy and phylogeny for this group.

Bibliography: leaves 23-28. / The potential for using a combination of molecular and whole-organismal data has opened up new avenues for avian taxonomy, phylogenetics and biogeography. Such a multifaceted approach is used here to identify diagnosable taxa within the Orange River Francolin Scleroptila levaillanloides species complex and resolve evolutionary relationships between these taxa and other mono-and polytypic forms within the Red-winged Group of francolins (= genus Scleroplila sensli lalo). Mitochondrial cytochrome-b DNA sequence data (±250 b.p.) from 50 individuals and 19 morphological characters extracted from reports in published literature were employed to achieve these aims. These characters were analysed separately and also in combination using maximum parsimony (DNA sequences and organismal data), maximum likelihood (DNA sequences) and distance (DNA sequences) analyses. Monophyly of the Red-winged Group plus the Ring-necked Francolin Dendroperdix slreptophorus was supported by all the analyses (bootstrap support ranged from 50%-94%) except distance analysis. The Orange River Francolin complex was found to be non-monophyletic. Two distinct clades were identified, one comprising taxa from southwestern and the other from northeastern Africa. Morphological analysis yielded a distinct clade of the southwestern Orange River Francolin. The other polytypic species and assemblages thereof show poor resolution. The results of this study clearly demonstrate a need for further assessment of the taxonomic status of Scleroptila spp. and their phylogenetic relationships.

Bibliography: leaves 84-100. / A phylogeny was produced for the eight taxa comprising the largely South African genus Melianthus L. based on two plastid markers (trnL-F and psbA-trnH) and one nuclear marker (ITS). Topological comparisons with a tree based on an existing morphological data set revealed significant incongruence leading to a loss of resolution upon combination. Ultimately, the combined three-gene data tree was selected as the strongest phylogenetic estimate for Melianthus based on its better resolution and greater support levels. This tree confirms the monophyly of Melianthus with M. major being resolved as sister to the remainder of the genus. Within the remaining clade, M. villosus is resolved sister to a clade comprising two morphologically distinct subclades, one of these being noted for a western distribution (comprising M. elongatus and the M. pectinatus complex), the other being centred farther east (comprising M. cocomosus and the M dregeanus complex). A molecular clock analysis was employed to date the emergence of specific taxa and clades, while ancestral range and habitat reconstructions were performed to determine historical conditions under which these groups and their morphological novelties arose. A scenario depicting the evolution and diversification of Melianthus is developed against a backdrop of the paleo-history of southern Africa. The genus appears to have originated in the eastern part of South Africa during a mild, mesic Oligocene with subsequent westward expansion into drier habitats of Miocene and Pliocene origins. Observations of bird visitors to a range of Melianthus species indicate that the genus employs a generalist pollination syndrome, thus refuting suggestions of a co-exclusive relationship with sunbirds (Nectariniidae). Nevertheless, the genus is undeniably ornithophilous. Nectar studies reveal interspecific variation in both the volume and concentration of nectar produced as well as in nectar colourition.

Word processed copy. / Includes bibliographical references. / Jamesbrittenia is a genus of 84 largely perennial sub-shrubs and herbs with a wide distribution in southern Africa (except J. dissecta in Sudan, Egypt and India). Plastid (rps16 and psbA-trnH) and nuclear (GScp) sequences were obtained for 42 species, mostly from the arid winter-rainfall west and southern regions of southern Africa. Applying both parsimony and Bayesian inference to combined molecular and morphological data sets, a phylogenetic hypothesis which is robust at the deeper nodes, was produced. This supports the monophyly of Jamesbrittenia, and also indicates the early divergence of J. ramosissima and three major clades.

The Response of mobile intertidal invertebrates to subsistence harvesting in Northern KwaZulu Natal was investigated. Spatial scale comparisons were made between as well as within harvested and non-harvested locations. The difference in three population variables was investigated: i) abundance of species groups. ii) size structure of each organism and iii) community structure of mobile organisms. Abundance analysis revealed no significant difference between harvested and no harvested regimes for most species' groups, excepting Snails and Chitons. Snails were more abundant und chitons less abundant at harvested sites. A significant difference.... between locations was however noted, for all species groups, suggesting that variation in abundance within is more prominent than between harvesting regimes. Size stmcturc analysis revealed significantly larger sizes for most spec1cs at non harvested locations, with only Morula gm1111/u111 and Scutellastm exusw showing u significantly larger size structure at non-harvested locations. Community structure analysis revealed no clear distinction between or wi1hin harvested locations. A decrease in size structure with no co1Tespondi11g density effect may be a function of the preference of harvesters for larger individuals, thereby favouringjuveaile populations. The possibility also exists that density effects have been masked due to the use of size instead of biomass data. Converting to size data to biomass, using wet mass versus size regressions, may be a more appropriate analysis method. The observed differences in community structure between as well ru; within loi:-ations indicates that the intertidal communities are inl1erintly hcterogenous in this area. ln order 10 determine the irnpacr of exploitation. spatial scale comparisons between harvesting regimes thus ideall y need to be conducted to each location. [n this study there was also a lack of representative control sites, us unharvested locatio11s are impacted by fisherman and tourists. A possible strategy would be to demarcate "110-go" areas in harvested and non-harvested areas, to serve as both controls and provide broodi11g stocks for adjacent ledges. It is however recommended that more research emphasis be placed on U1e user in evaluating the effects ofhurmtn tmpact on intertidal resource.

Abiotic and top down control hypotheses do not adequately explain the fine scale distribution patterns of shrub and tree Colophospermum mopane (Caesalpinioideae). Genetic distinctiveness between growth forms is investigated as an alternative hypothesis. Tree and shrub C. mopane from the riparian and inland savanna zones were sampled at four sites in the northern Kruger National Park. Molecular DNA sequences were obtained for four plastid and one nuclear region, and the inter simple sequence repeat (ISSR) technique used to fingerprint individuals. Very low levels of sequence divergence were observed. The ISSR technique revealed no genetic structure between plants when grouped by growth form or by habitat in an analysis of molecular variance (AMOV A). Soil profile and xylem pressure potential data also did not explain the distribution of growth forms. A principle component analysis and a discriminant analysis of five leaf and branching characters identified a significant difference in the shape of shrub and tree C. mopane leaves. It is concluded that the C. mopane growth forms are not genetically distinct and that their fine scale distributions may be due to top down controls such as large mammals and fire.

The effectiveness of the structural defences of Acacias vary within and between species as a result of differences in thorn type (hooks versus spines) and branching architecture. This study aims to develop methods for quantifying the effectiveness of defences, thereby allowing for comparisons between individuals and localities. Measurements of a range of thorn and branching parameters of eight Acacia species were used to calculate three indices (stripping, biting and branching) describing different components of Acacia structural defences. These were combined to form an overall defence index that was tested against the results of three feeding simulation tests as well as actual animal feeding rates obtained from trials using nyala held in bomas. The stripping, biting and branching indices were shown to provide a good measure of the different components of a structural defence, although the scaling of their contribution to an overall defence index is complicated by variation in the suite of browsers and the resources available at a locality. The indices developed in this study provide a useful tool for assessing the role of large mammal herbivory in different environments when attempting to understand variation in the life history strategies of Acacias.

A one-year experimental fishery for P. delagoae was established in April 2004 to determine the frequency and magnitude at which pulse fishing may be sustainable (Government Gazette 2004). Determining the extent of recovery after fishing is one aim of the experimental fishery. Assessments of the growth rate, size at sex maturity and natural mortality of P. delagoae were performed in 2000 (Groeneveld et al) and can be supplemented by the results of this study, which assess the reproductive biology of the East Coast spiny lobster in order to address this aim. A recent study (Groeneveld in press) assessed fecundity, egg loss during gestation, relative reproductive potential and lifetime egg production per recruit for P. gilchristii from three areas along the South African south coast. A general east-west trend of increasing fecundity, size at sexual maturity, and lifetime egg production per recruit was found for P. gilchristii. Examining egg-loss through the first four ( of six) developmental stages of gestation showed significant loss of around 15 % by stage four, irrespective of lobster size (Groeneveld in press). The study of fecundity in P. gilchristii ( Groeneveld in press) was used as a framework for assessing the fecundity of P. delagoae in this study.

Zandvlei is a system that has been highly modified by the urban environment of Cape Town. Salinity is a major driver determining the physical environment of estuaries. The macroalgal community of estuaries forms an important part of the ecology of these systems as primary producers. The salinity of this environment is an important parameter determining the composition, abundance and diversity of these communities. Patterns of salinity fluctuations in the long term (1978-2003) and annually are described in order to establish how communities may vary. Zandvlei is in a Mediterranean climate and salinity was found to vary seasonally and monthly with fluctuations in rainfall. Historical records of macroalgae identified in the estuary were summarized and no record of the Polysiphonia sp., now dominant in the estuary, was found. Dominant macroalgae in the estuary were identified and grown under a range of salinities (0, 1, 5, 10, 20, 29 ppt). Dominant macroalgae included Cladophora sp., Polysiphonia sp. and Enteromorpha prolifera. None of these algae survived at O ppt. In 1 ppt E. prolifera survived but growth was retarded, Polysiphonia sp. survived but did not grow and Cladophora sp. grew successfully at this salinity. All three species have different optimum ranges for growth with E. prolifera preferring higher salinities, Cladophora sp. grew similarly across the range from 1 to 29 ppt and Polysiphonia sp. grew most rapidly at 5 ppt. Depending on the salinity range in the estuary different macroalgae will be dominant. This makes the macroalgae community quite resilient to fluctuating salinities. There should therefore always be estuarine macroalgae present in the estuary provided salinity does not drop below 1 ppt for an extended period. If salinities did drop below 1 ppt there could be an increase in freshwater algae.

The Zandvlei Nature Reserve is to be expanded to form the Greater Zandvlei Estuary Nature Reserve and will encorporate a much greater area (250 more hectares). The Greater Zandvlei area was classified by means of Zurich Montpellier (Braun Blanquet) procedures. 60 relevés were used to create a hierarchical phytosociology using TWINSPAN and DCA-ordination analyses. 3 major community types were identified, which were then subdivided to reveal 6 community types important to future management: Moist Grassland/Wetland Disturbed Parkland, Dune Asteraceous Fynbos, Thicket/Shrub Mosaic, Rhus-Euclea Dune Thicket and Homogenous Patches. CCA-ordination indicates that the abiotic soil factors sampled (texture composition, pH, carbon content and visible disturbance) explain very little of the variation between the communities (r²=0.002). This phytosociology is provided as a tool for future research in the area, and it is hoped to assist in future management decisions regarding the newly founded GZENR.