In Malawi, fish contribute about 60-80% to the country's animal protein supply. The greater proportion (> 50%) comes from Lake Malawi. Bathyclarias nyasensis and other clariid catfish contribute up to > 20% of the total catches. Catches of Bathyclarias nyasensis in the inshore area of the south-east arm of Lake Malawi are declining and a management plan for the fishery is essentially lacking. There is paucity of biological data that precludes the use of any option to manage the species. The principal aim of the thesis was to define the ecological role B. nyasensis, the most abundant and common of the Bathyclarias species. By examining life history characteristics within a food web context, it was hypothesized that the study would provide an insight into the interrelationships between species, and, hence form the basis for the development of a rational exploitation strategy for the species. The study was undertaken in the south-east arm of Lake Malawi (9° 30'S, 14° 30'S). The principal objectives of the study were to investigate the feeding ecology of B. nyasensis by examining morphological characters and structures associated with feeding, diet of B. nyasensis, food assimilated in the species using carbon (∂¹³C) isotope analysis, daily food consumption rate for B. nyasensis; and to relate the feeding ecology to life history traits such as age, growth, and some aspects of the reproductive biology of B. nyasensis. The suitability of sectioned pectoral spines and sagittal otoliths to age B. nyasensis was assessed. Due to reabsorption of growth zones with increasing spine lumen diameter with fish size, and the relatively low number of spines that could be aged reliably, only otoliths were used. The maximum age for B. nyasensis was estimated at 14 vears. Growth was best was described by the four parameter Schnute mc: lt ={42+(81¹·⁸ - 42¹·⁸)x1-e⁻°·°⁵⁽t⁻¹⁾}¹/¹·⁸ over 1-e⁻⁽⁻°·°⁵⁾⁽¹¹⁾ for female, lt={41+(98¹·² - 41¹·²)x 1-e⁻°·°²⁽t⁻¹⁾}¹/¹·² over 1-e⁻⁽⁻°·°²⁾⁽¹³⁾ and for male fish. Age-at-50% maturity for females and males were estimated at 7 years and 4 years, respectively. Typically, fish grew rapidly in the first year, but slower during subsequent years. Smaller fish were found inshore while larger fish were found in offshore regions. It was hypothesised that the rapid growth in the first year and slower growth later is a consequence of change in diet from high quality and abundant food source to a more dilute food and that this may be associated with a shift in habitat. Morphological characters associated with feeding were used to predict the food and feeding behaviour of B. nyasensis. The size of premaxillary, vomerine, pharyngeal dental and palatine teeth and premaxillary and vomerine tooth plates suggested the capability of B. nyasensis to handle both large and small prey, with a propensity towards smaller prey in composition to C. gariepinus. The molariform teeth on the vomerine tooth plate suggested that molluscs form part of the diet. The relative gut length (1.27±0.24) suggested omnivory, with an ability to switch between planktivory and piscivory. Buccal cavity volume and filtering area changed with fish size at 500-600 mm TL upon which it was hypothesised that the fish diet changed to planktivory at this size. Detailed diet analysis provided information upon which the above hypotheses could be accepted. Percent Index of Relative Importance (%IRI) and a multi-way contingency table analysis based on log-linear models were used to analyse diet data. Results showed that B. nyasensis is omnivorous, but with a distinct ontogenetic dietary shift from piscivory to zooplanktivory at 500 - 600 mm TL. The increased buccal cavity volume at the same fish size therefore, suggests that B. nyasensis is well adapted to filter the dilute zooplankton resource. Increased foraging costs of feeding on zooplankton explained the slower growth of larger fish. The dietary shift was finally corroborated by results of the ∂¹³C isotope analysis. A polynomial equation described the change in carbon ratios with fish size: ∂¹³C = - 33.188 + 0.4997L - 0.0045 (total length)² (r² = 0.598, n = 12, p=0.022). The ontogenetic shift in diet was synchronised with a habitat shift postulated in life history studies. In the inshore region, B. nyasensis were predominantly piscivorous (apex predators), and were zooplanktivorous in the offshore region, thereby forming part of the pelagic food web in the latter region. After examining "bottom-up" and trophic cascade theories, it was postulated that perturbations of the B. nyasensis stock would be discernible both at the top and lower trophic levels. As a piscivore and therefore apex predator, effects of overfishing B. nyasensis in the inshore region could cascade to unpredictable ecological changes in inshore areas and, due to the ontogenetic habitat shift, in the offshore regions. Examples of trophic cascade phenomena are provided. On the basis of the feeding study, it was possible to reconstruct the pelagic food web of Lake Malawi. Apart from the lakefly Chaoborus edulis, B. nyasensis is the other predator that preys heavily on zooplankton in the pelagic zone. Perturbations of the B. nyasensis stock could affect size composition of zooplankton which in tum, could affect production of C. edulis, a resource for the top predators in the food web. The findings of the present study contributed to the ongoing debate of introducing a zooplanktivore into the pelagic zone of Lake Malawi. Proponents for the introductions have argued that zooplankton predation by fish is inferior to that of C. edulis. Introduction of a clupeid zooplankton was proposed as a strategy to boost fish production in the lake. The zooplanktivore would either out-compete or prey on C. edulis to extinction. Opponents to this view argued that zooplankton biomass in the pelagic region was too low to support introductions and that the fish biomass in the pelagic region may have been underestimated. Results from the present study suggest that planktivorous fish (including B. nyasensis) might not be inferior to C. edulis in utilising the zooplankton resource; B. nyasensis is well adapted to utilise the dilute zooplankton resource, and by omitting B. nyasensis from previous studies, overall zooplankton predation by fish may have been underestimated by between 7 - 33%. On the basis of the theoretical migratory life history cycle of B. nyasensis, it is recommended that the current interest in increasing fishing effort in offshore areas should proceed with caution. Ecological changes that may have occurred in the inshore areas due to overfishing have probably not been noticed: as the offshore zone has never been fished. The latter zone may have acted as a stock refuge area. Higher fishing intensity in the offshore areas could lead to serious ecological imbalances and instability. The study has shown that life history characteristics studied in the context of the food web, and in the absence of other fisheries information and/or data, strongly advocates the precautionary principle to managing changes in exploitation patterns.

Behaviour and space utilisation of rock-dwelling cichlids were observed at Thumbi East Island, Lake Malawi. 1. Males of five species of the mbuna complex held long-term territories. Pseudotropheus elongatus "aggressive" vigorously defended a feeding area and sometimes a spawning site interspecifically, but did not feed in the peripheral part of their territory. Spawning sites of Pseudotropheus zebra, Pseudotropheus tropheops "orange chest", Labeotropheus fuelleborni, and Petrotilapia nigra were interspecifically-defended, while larger mating territories were defended against conspecific neighbours. Feeding areas were shared with many fish and often extended beyond the defended area. There was considerable variation in behaviour and space use within and between species and between times of day. 2. Non-territorial P. zebra used larger ranges than territorial conspecifics, and fed more on plankton, but individuals had preferred benthic feeding areas, often in conspecific territories. These 'floaters' were often aggressive. Both size and relative brightness independently predicted the outcome of aggressive interactions between floaters, and a site-specific dominance hierarchy was suggested, with some individuals appearing to be semi-territorial. 3. Males and females of 21 and 13 species respectively were found to establish temporary breeding territories. Overall breeding seasonality was bimodal, but reproductive timing and territory characteristics differed among species. Temporary territories had a considerable impact on the behaviour and habitat use of all resident mbuna species, even causing abandonment of territories. 4. Non-breeding Protomelas taeniolatus had limited home ranges, and showed little aggression. During the highly-synchronised reproductive season, males defended spawning sites and females fry-guarding territories. Most chases were directed towards the commonest fish, but predators were chased further and faster. Female behaviour changed over the guarding period. Females generally continued territorial defence after the brood had disappeared. Most broods contained fry of different sizes and species. Significant benefits were found for guarding females with clustered territories, but females did not appear to choose sites adjacent to conspecific parental females. 5. Territoriality in fish is taxonomically widespread and may serve several functions according to species, sex and developmental stage. It also varies according to genotype- and phenotype-limited strategies and short term costs and benefits. Territories may be simultaneously multifunctional.