Includes bibliographical references (leaves 151-173). / A root-feeding, Mexican flea beetle, Longitarsus bethae Savini & Escalona 2005 (Coleoptera: Chrysomelidae: Alticinae), was evaluated in quarantine as a candidate biological control agent for the noxious weed, Lantana camara L. (Verbenaceae). The premise was that L. bethae would only be released if it could be ascertained that it was: (i) safe for non-target plants; (ii) likely to inflict significant damage on the target weed; and (ii) capable of surviving under the various (abiotic and biotic) environmental conditions in its new range.

Bibliography: leaves 201-232. / The aim of this study was to investigate the niche separation between these two species, using zooplankton net samples collected during biannual hydroacoustic surveys of pelagic fish between 1988 and 2000, as well as results from ship- and laboratory-based egg production and feeding experiments. C. carinatus is most abundant on the productive central West Coast, but also occurs at relatively low abundance on the Agulhas Bank, mainly on the outer shelf.

Includes bibliographical references (p. 125-140). / Freshwater ecosystems underpin the fabric of society and the environment, providing essential ecosystem services such as water and food, upon which all human beings depend. In order to secure these vital services requires the sustainable management of freshwater ecosystems. At present however, freshwater biodiversity is under severe threat from anthropogenic disturbances, and the situation is expected to worsen due to population growth and global change. Coupled with the threats to freshwater biota are the limited resources available to secure their protection. There is a need to therefore prioritise freshwater ecosystems in a comprehensive, adequate and representative manner to maximise the outcomes of conservation effort. The concept of systematic conservation planning was developed to address this challenge. It offers a suitable framework for achieving conservation goals in the face of other competing land uses. The principles of systematic conservation planning are being widely applied in terrestrial and marine ecosystems, but their application in freshwater ecosystems is still relatively limited. Freshwater ecosystems provide challenges to conservation planning that are unique from those of terrestrial and marine ecosystems, such as the longitudinal nature of river systems and the associated connectivity, and catchment divides that constrain some obligate species. As a result freshwater ecosystems require conservation planning tools and approaches that are specifically geared towards addressing these unique challenges. Progress has recently been achieved in addressing some of these challenges, but there are still other outstanding issues that have not been comprehensively addressed. The aim of this thesis was to develop new frameworks, and test approaches for the application of systematic conservation planning principles in the conservation of freshwater ecosystems in South Africa. The thesis addressed a range of issues along the systematic conservation planning continuum from biodiversity assessment to implementation. I first carried out a focused review of systematic conservation literature between 1987- 2006, to gauge the extent to which freshwater ecosystems have being integrated in conservation assessments. Most of the focus was found to be on terrestrial ecosystems with minimal incorporation of freshwater biodiversity. Wetlands for example, were in most cases incorporated into conservation assessments without taking their diversity into account. This was partly attributed to the difficulty of classifying wetlands. I therefore developed and tested a hierarchical GIS framework for automating wetland classification as a strategy for incorporating wetland biodiversity, functions and benefits into broad scale conservation planning.

Includes abstract. / Includes bibliographical references (p. 146-168). / This study investigates whether anthropogenic changes in the Cape Peninsula, South Africa, have affected gastrointestinal parasite infections in a free-living population of chacma baboons (Papio ursinus). Data on parasite diversity and prevalence were obtained from 616 faecal samples collected from over 350 individuals in eight troops (six from the Cape Peninsula, one from Pringle Bay, and one from Wildcliff Nature Reserve) between July 2006 and August 2007.

Includes bibliography. / This study investigated the potential usefulness of aerial seabirds as indicators of the distribution of peculiar oceanic biotopes and prey populations in the African sector of the Southern ocean. The hypothesis examined was that the distribution and abundance of seabirds are non-random and predictable with respect to the availability of prey at the sea-surface. The distribution of seabirds was correlated with prey density, in cases where the appropriate information was available. More often than not, however, data on prey distributions were unavailable, so that seabird distribution was correlated with the abiotic indicators of hydrodynamic processes which order the distribution of potential prey. The distribution of seabirds was determined by means of shipboard observations. The trophic structure of seabird assemblages was assessed according to the diversity, biomass and abundance of 35 seabird species according to four principal diet-classes. The ecological structure of seabird populations was defined in terms of relationships between the trophic structure of seabird assemblages and Southern Ocean oceanography, meteorology and biogeography. The predictability of seabird distribution and abundance with respect to environmental variation was assessed using linear and nonlinear regression procedures. A deterministic model was developed in which analytical techniques are standardized and which can be applied to other ocean areas. There are predictable associations between the trophic structure of seabird assemblages and the structure of the surface of the sea. Assemblages of putative prey feature squid and fish in low latitudes, and plankton predominate in higher latitudes. The ecological structure of seabird populations reflects this trend on a macro-scale level. On a meso-scale level, the abundance of seabirds by diet-class correlates with the relative abundance of putative prey. Moreover, seabird abundance and biomass correlate positively with abiotic indicators of frontal zones and sea-surface mixing, where prey diversity and abundance are greatest. Regressions of seabird abundance with sea-surface temperature explain important aspects of the variation in the ecological structure of selected seabird populations. This study showed that it may be possible to track the distribution of prey stocks through concurrent observations of seabirds and hydrography. Consequently, the descriptive model developed in this study contributes towards the objective of circumventing the need for direct sampling of prey stocks. The predictability of seabird distribution and abundance previously obtained has been surpassed by the results of this study through progressive reduction of the spatio-temporal scale of data collection and analysis, and by treating biotic-abiotic relationships with non-linear regression models where appropriate. The study concludes that sufficient advances have been made to justify further research into the use of selected species of aerial seabirds as indicators of prey distribution and abundance at sea.

Bibliography: leaves [148]-164. / This field study set out to identify the key ecological factors influencing the population dynamics of the Namaqua Sandgrouse Pterocles namaqua, through an investigation of diet and the nutritional demands during different stages of the annual cycle, the timing of breeding seasons and movements in relation to patterns of rainfall-dependent food availability, breeding success and the factors limiting productivity. In addition, the nesting habits and success of 11 coexisting arid-zone bird species were examined to test a variety of hypotheses regarding the relationship between nest-predation rate and nest site, nest density, predator-avoidance behaviour, stage of the nesting cycle and season, and degree of residency. Furthermore, the importance of rainfall as a breeding stimulus and its effects on clutch size were investigated for several species. The Namaqua Sandgrouse is an obligate granivore at all times, feeding on the seeds of annual plants, primarily of the family Fabaceae. Even while breeding, energy is the first-limiting nutrient in the foods of adults. Growing chicks have a proportionally greater protein demand, and are more dependent than adults on protein-rich legume seeds to satisfy first-limiting amino acid requirements. The chick growth phase was identified as the most nutritionally demanding stage in the annual cycle. The breeding season was found to be unexpectedly variable, and not consistently correlated with periods of peak food availability.

This thesis investigates ecosystem effects of bottom-trawling on demersal fish assemblages (1990-2006) off Namibia and their relationship to environmental variables.

Includes bibliographical references. / In the past decade African penguin (Spheniscus demersus) populations have experienced declining population sizes of > 60% in some instances. This has primarily been attributed to shifting prey availability and competition with regional purse-seine fisheries; however, possible novel threats exacerbated by diminishing population sizes and increased stress, may also be important contributors to the regional persistence of the species. These threats include the impacts of arthropod-borne blood parasites on the health of penguin populations. While parasitism of wild penguins has been poorly studied, susceptibility to infection with blood protozoa is well established in a wide range of penguin species held in captivity. This thesis assesses the prevalence of blood parasites in wild African penguins throughout the greater Agulhas-Benguela ecosystem. Using PCR-based techniques, 317 individuals were screened for the presence of known haemoparasite species of Plasmodium, Haemoproteus and Babesia across 12 breeding colonies. Babesia infection was confirmed for 60% of wild birds sampled, but methods used did not indicate infection with either Haemoproteus or Plasmodium species. Generalised linear modelling of ecological and life history parameters suggests that Babesia prevalence is primarily driven by a colony’s distance from the mainland, decreasing significantly as distance from the mainland increases. Captive birds held at the SANCOBB rehabilitation facility in Cape Town present with both Plasmodium and Babesia species, providing the positive controls for the study. The relative scarcity of Plasmodium and Haemoproteus infection in the wild may be the result of an absence of suitable vectors and/or high post mortality in the host. Alternatively, the sensitivity of the method used requires further investigation as Plasmodium infection has been confirmed previously in penguin populations using PCR-based approaches. This study provides the first baseline estimate of blood parasitism in African penguins across their breeding range, and raises the need for further research and monitoring. The results suggest that high Babesia prevalence in African penguins may be related to (1) an ecological system of chronicexposure to infection reservoirs, such as the co-occurring cormorant and gannet populations and (2) increasing anthropogenic impacts, especially in mainland colonies. Data on blood parasitism in co-occurring seabird species is required to fully elucidate their role in Babesia infection dynamics in the region. To improve understanding and facilitate timely detection of changes in blood parasite exposure, standardised methodologies are advocated to better inform the conservation management of this iconic species.

Bibliography: p. 436-482. / Eight new species of digenea were found in a survey of Aulacomya ater, Choromytilus meridiana/is, Perna perna and Mytilus galloprovincialis. Metacercaria notobucephala sp. nov., a bucephalid from Choromytilus has locality-specific prevalences ranging from 0.23% to 4.46%. It is a severe reproductive stress and reduces the sperm count by a factor of 1000. Metacercaria notobucephala also causes Choromytilus to lose ninhydrin positive substances. The gymnophallid Metacercaria perchorupis sp. nov. occurs in 90.7% to 100% of Choroinytilus and in 74.83% of Perna. Metacercaria A sp. nov., a renicolid, occurs in 78.86'yo to 90.33% of Choromytilus and 44.74% of Perna. Metacercaria B sp. nov., a lepocreadid, infects from 18.3% to 43.17% of Choromytilus and 3.51% of Perna. The combined effect of these three digenea causes 10% of the variation in emersion survival time of male Choromytilus (P = 0.02). Longer surviving mussels had fewer cysts; this suggests that these parasites constitute a somatic stress. Metacercaria columbinesis sp. nov., a zoogonid, was found in 28.6% of Mytilus. The fellodistomid Metacercaria maculatopsis sp. nov. occurred in 0.4% of Choromytilus. The gymnophallid, Metacercaria ater sp. nov. was found in 0.125% of Aulacomya. A rapid technique of semi-permanent staining and fixing with acetic orcein was developed to aid the description of these worms. Juvenile pycnogonids, Nymphonella sp., were found in 0.15% to 7.5%. ofChoromytilus. Nemerteans, nematodes, copepods, polychaetes and other minor symbionts are quantified in these mussels. Mastigocoleus sp., a shell boring cyanophyte, colonises Mytilus with prevalences from 1% to 94% and can reduce shell strength by 43%. The relationship between stress (amount of alga on the shell) and strain (degree of damage) is demonstrated: 61.8% of variation in damage is attributable to the extent of infection (P = 0.001). Stress is defined as an agent that reduces the fitness of the organism. The degree of reduction is termed strain. Fitness is an ontological statement about the organism. Schemes are proposed for rendering these concepts operational. Individual and ecosystem concepts of stress are not homologous and therefore cannot be substantively integrated. In contrast, psychological and physiological stress are deemed integrable: a scheme for this is proposed. Physical (salinity), chemical (ammonia & phenol) and biological agents (scavenging whelks) are shown to be substantively integrable as stresses since they all inhibit shell gaping and byssus production. Mussels demonstrate clear and appropriate closure responses when in the presence of salinity, ammonia levels and Burnupena: all agents that are likely to be of historical selective significance. In contrast, phenol, an unfamiliar agent, elicits an inconsistent response. It is hypothesised that this difference may help distinguish natural stresses from synthetic pollutants. As predicted, the valve opening dynamics of heat stressed mussels were qualitatively and invariably different from those of non-lethal voluntary movements. Dying dynamics have increasing velocity; voluntary dynamics show maximum velocity initially. Analysis of stimuli effects, often interpreted as eustresses, show that assays covering distal as well as proximal effects and those that tend towards whole body integrations are more likely to detect the agent as deleterious. Thus any notion of positive deflections or eustress must be treated with great caution and subject to longer term tests and with more complete integrations if it is to be accepted.

Bibliography: p. 147-160. / Protective gillnets (shark nets) have been successful in reducing the frequency of shark attacks on the coast of KwaZulu-Natal (KZN), South Africa. This is achieved primarily through a reduction in numbers of large sharks. The nets also take a by-catch of dolphins, sea turtles, batoids and teleosts.Catch rates of most shark species declined initially but have shown no trend since the mid-1970s. Turtle and teleost stocks do not appear to be threatened by net mortalities, but there is concern about the sustainability of catches of the humpback dolphin. Certain batoids may have declined despite a high release rate. A published contention that shark netting has resulted in a proliferation of small sharks through reduced predation is re-examined and considered to be exaggerated. Reduced predation on dolphins, as a result of shark netting, is estimated. Considerably less fishing effort is applied in the shark control programs of New South Wales and Queensland, Australia, than in that of KZN. On the basis of a comparison of factors such as the nearshore physical environments and trends in shark catch and catch rate, it is concluded that the number of nets used in KZN could be reduced. To test whether a 70 cm mesh would continue to capture potentially dangerous sharks, while at the same time reducing by-catch, a gamma distribution model was used to determine length-specific selectivities in 50.8 cm and 70 cm mesh nets respectively. A reduction in relative selectivity from 81 to 25 for a shark of 1.6 m PCL would result from an increase in mesh size from 50.8 to 70 cm. Despite a probable reduction in catch of dolphins .and certain other by-catch species, the introduction of the larger mesh would constitute an unacceptable reduction in levels of bather safety. Baited lines, or drumlines, were tested as possible alternatives to gillnets. They demonstrated greater species selectivity for sharks, including a higher catch of two of the target species, Carcharhinus leucas and Galeocerdo cuvier, and also a reduced by-catch of nonshark animals. The probability of the bait being scavenged, or a shark being caught, was modelled in relation to a number of physical factors. Although there were insufficient data for a quantitative comparison of catch rates between nets and drumlines, the results indicated that an optimal solution may be to deploy a combination of nets, using the existing 50.8 cm mesh, and drumlines, using 14/0 shark hooks.