Bibliography: leaves 115-134. / This study investigated the biological characteristics (linked with daynight and migration aspects) of the stock, the behavioural patterns of sardinella spp (schoolbehaviour and dynamics) that may lead to under-estimation as well as bias. Biological and acoustic data collected during 22 pelagic surveys (1985-1999), as well as schooling records of three cruises (1996-1998), were used to study the biology, distribution, behaviour of the school dynamics of sardinella.

Includes bibliographical references (p. 161-170). / Some moths emit ultrasonic clicks in response to bat echolocation calls. These clicks are believed to serve as a defence mechanism against bat predation by means of jamming, aposematism or startle. By assessing the characteristics and variation of the ultrasonic moth clicks it was possible to define the most likely function of the ultrasonic clicks by moths from sites in South Africa. Additionally, this study used Schaefer based mathematical models informed by field data on moth diversity and abundance, moth click parameters and bat population numbers and diversity, from these same sites in South Africa, to gain insight into the conditions that are required for each of these functions to work. The Jamming hypothesis proposes that moth clicks function by interfering with the bat's perceptual system, which is believed to be most effective in the terminal phase of the bat's attack were the clicks have been shown to be the most similar to the bats echolocation calls. This would place functional constraints on the clicks, decreasing the variation in the clicks. This was shown not to be the case, as a high level of variation was found between species' clicking parameters (e.g. peak frequency, inter pulse interval and intensity). A high level of individual moth variation was also found, with numerous patterns being produced by the same moths. Clicking variation was created both by manipulation of individual tymbals (e.g. changes in modulation cycle rate and durations) and between tymbals (e.g. timing of tymbal alternation). Furthermore the moth clicks are too different to bat echolocation calls to allow them to be mistaken by the bat as echoes from its own calls. The aposematic hypothesis suggests that the clicks function as a warning signal to inform the predator of the prey's unpalatability. As in visual based aposematism, the cost of. predator learning on the prey population would exert selective pressure towards higher densities of clicking moths and convergence in warning signals. This would allow one species to benefit from the predator learning induced by another species. This strategy is thus density dependent and would require a minimum proportion of clicking moths to work. Both the data from the ultrasonic moth clicks and mathematical modelling does not support this hypothesis as there is no signs of convergence and the mathematical modelling suggests that the low proportion of clicking moths (2.3%) in the natural moth population does not allow bats to form the association between moth clicks and noxiousness of the moth. In the case of jamming, the low proportion of clicking moths also does not reflect the competitive advantage supposedly gained from jamming. The startle hypothesis proposes that moth clicks confuse the bat by altering the normal sequence of expected events during the bat's attack. However, bats will habituate to these sounds if they are encountered regularly. Both the low proportions of clicking moths and the extreme variability of the click patterns hinder habituation and are consistent with the startle hypothesis, which suggest that the most likely function of ultrasonic moth clicks in these South African sites is startle. On the other hand, the aposematic and startle models do suggest that the existence of both noxiousness and clicking create the initial conditions for the evolution of an aposematic function for moth clicks. If the clicking moth is noxious the startle effect of the clicking could allow the population to grow to high enough numbers to make the aposematic function of the clicks viable.

Includes bibliographic references (leaves 91-104). / The Cape serotine bat (Chiroptera: Vespertilionidae) is an endemic species of sub-Saharan Africa and occupies all biomes throughout its distribution. It roosts in anthropogenic structures in small colonies of up to ten individuals. Since its discovery in the early 1800's by Arthur Smith little more than a few aspects of its reproductive biology and diet have been documented. Almost nothing is known about philopatry, migration and dispersal patterns of the Cape serotine bat and therefore nothing is known about its population structure. In this study I use microsatellite and mitochondrial D-Ioop sequences to determine the genetic structure of the Cape serotine bat population within South Africa. I investigated the degree of genetic differentiation between subpopulations in different biomes, and among colonies within subpopulations.

Bibliography: leaves 22-24.

Word processed copy. / Includes bibliographical references (leaves 45-52) / Spatial variability in egg volume was investigated for anchovy, sardine and round herring in the southern Benguela in 1998, 2000 and 200 l. The relationship between sea surface temperature (SST) and egg volume was determined and the parental condition hypothesis was investigated through the relationships of adult size and condition with egg volume. General linear models for each species were used to determine the factors that were significantly related to egg volume at the scale of sample collection. Visual interpretation of the effects of adult size and condition on egg volumes was done on a large scale.

Bibliography: leaves 97-104. / The objective of this study was to determine the effects certain physical and biological factors (tidal height, wave exposure, biogeography, mussel species and mussel size) have on the community composition of the infauna associated with mussel beds along the coast southern Africa.

Includes bibliographical references. / The findings of this study support several ideas relevant to the interpretation of evolutionary patterns among cooperative species. First, the reproductive costs of delaying reproduction may not be as high as previously assumed in some cooperative species if, as in females of this species, birds delaying breeding tend to live longer than birds reproducing at maturity. Second, as demonstrated by inter-sexual difli:rences in the reaction of birds to varied levels of breeder saturation, behavioral responses to ecological constraints may depend on life-history pattem.

Bibliography: leaves 56-65. / The following study was undertaken as part of the Karoo Biome Project (KBP) and sets out to investigate an aspect of karoo ecology that will contribute to a predictive understanding of both biotic and abiotic processes in the region for the development of meaningful management guidelines (Cowling, 1986).

Includes bibliographical references (p. 98-108). / Despite the recent surge in interest in wetland invertebrates, given their immediate importance in overall wetland functioning, these studies have largely been conducted in the northern hemisphere and very little is known in South Africa. In order to fill this knowledge vacuum, this study was conducted to investigate biodiversity patterns of wetland macro invertebrates assemblages in the south-western Cape, South Africa. 140 wetlands were sampled in four geographic clusters, the Cape Flats, the West Coast, the Cederberg and the Agulhas Plain. Open-water, submerged vegetation and emergent vegetation habitats were sampled separately using a square-framed sweep net with a 23.5cm mouth and 80l1m mesh and physico-chemical attributes measured at each site. A total of 126 taxa emanating from more than 26 000 individuals representing 73 genera and 51 families was recorded. Currently, five species new to science were also discovered, three from Hydraenidae (Prosthetops sp. nov, Parhydraena sp. nov and Mesoceration repandum Perkins, 2009), Hydryphantidae (Hydryphanres sp. nov) and Streptocephalidae (Streptocephalus sp. nov). However, all of these new species but one were recorded from single location, suggesting that they might be rare or unique. Richness estimators predicted true taxon richness at about 20-47 taxon more than the observed taxon richness. The high number of uniques and singletons, as well as the Incidence-based Coverage Estimator, suggest that sampling was not adequate to capture the full biodiversity suite. However, I believe this is not necessarily true and this is discussed in detailed in the discussion. The sub-area Cederberg was found to be the biologically richest area, with the three of the new species recorded there. A significant difference was observed in relative abundance (RA) and frequency of occurrence between open-water habitat and vegetated habitats (submerged and emergent vegetation). On the other hand, a significant difference was not found between wetlands with different number of habitat types. No appreciable congruence was seen between different taxonomic groups (i.e. at order and family level) themselves, suggesting that neither can be used as a surrogate for the others. However, predators - which by the way had the highest frequency of occurrence and 2nd highest relative abundance after gatherers - were strongly correlated with overall assemblage and Hemiptera. In addition, together with Hemiptera they had ≥75% similarity with the overall assemblage, suggesting that they are fairly good surrogates for overall assemblage. These results are quite promising especially for biomonitoring programmes and warrant further research. Environmental variables contributed significantly higher in explaining assemblage composition, and the model containing pH, Phosphate, conductivity, turbidity and ammonium was found to explain significant variation in macro invertebrate assemblage.

Bibliography: leaves 128-138. / Limpets of the genus Patella are very common intertidal molluscs that have a widespread world-wide distribution. Despite this, the systematics of the genus is still in a state of confusion. However, in comparison to other regions, the southern African representatives of Patella are fairly well characterised. Nevertheless, some long-standing taxonomic uncertainties still exist for certain of the southern African species. This thesis combines both morphological and genetic techniques to solve these uncertainties.