Interactions between piscivorous coral reef fish and their prey

Predation appears to play an important role in the regulation of populations and communities of coral reef fish. Interactions between piscivorous coral reef fish and their prey have rarely been studied, however. The aim of this thesis was to examine how piscivorous coral reef fish respond to fluctuations in the abundance of their prey and to use this information to predict their impact on prey populations. In the first part of this thesis a broad suite of piscivorous species at Lizard Island on the Great Barrier Reef were studied, while in the second part I concentrated on the influence of prey abundance on the ecology of two common piscivores, the rock-cods, Cephalopholis cyanostigma and C. boenak (Serranidae). The distribution, abundance and community structure of piscivorous fish at Lizard Island was examined initially. Central to this was the trial of a new baited census technique used to bring cryptic species into view. Traditional census methods (strip transects) were found to underestimate the abundance of cryptic piscivores by approximately 50%. In contrast, the baited technique accounted for approximately 90% of cryptic fish. Strip transects were found to be more appropriate for conspicuous, mobile species, however, as the baited method attracted mobile fish from an unknown area. By combining the baited method for cryptic species with the transect method for mobile species a relatively accurate picture of the piscivorous fish community at Lizard Island was obtained. Abundance of piscivores varied at both a large scale (between different zones of exposure kilometres apart) and a local scale (between sites hundreds of metres apart). This variation may have implications for the impact of piscivores on prey. The abundance of piscivores and their prey was then monitored over an 18 month period in two different habitats, patch and contiguous reef. This allowed two hypotheses to be examined: (1) that piscivores would aggregate in areas of high prey abundance and (2) this aggregation would cause density-dependent mortality of prey. The abundance of both piscivores and prey was consistently higher on patch reefs than on contiguous reef. There were also strong positive relationships between prey and piscivore abundance within sites throughout the study. Mortality of prey was density-dependent over the 6 month period it was monitored. Mortality of prey was positively related to piscivore abundance but was also positively related to the ratio of prey to piscivores. This suggested piscivore abundance only partly explained patterns of prey mortality. Recruitment patterns did not explain the relationship between prey and piscivore abundance, suggesting post-settlement processes were responsible. Tagging of Cephalopholis cyanostigma and C. boenak was used to examine the contribution of movement. Movement between sites or habitats was rare for both species, however, suggesting it did not explain large-scale patterns of abundance. In contrast, at a local scale territory size was inversely proportional to prey density and a change in current direction caused a corresponding small-scale movement of both piscivores and prey. Hence, although the mechanisms could not be fully determined, the abundance of piscivores and their prey appeared to be strongly linked. At the same sites and times as above the influence of prey abundance on the dietary composition, prey selection and feeding rates of Cephalopholis cyanostigma and C. boenak was examined. Gut contents were collected mainly from regurgitated samples, to allow the same populations to be monitored over time. Both species were found to be over 90% piscivorous with prey fish of the families Apogonidae, Pomacentridae and Clupeidae dominating the diet. The interacting effect of fluctuations in prey abundance and patterns of prey selection caused diet to vary both temporally and spatially. Mid-water schooling prey belonging to the family Clupeidae were selected for over other families. In the absence of these prey apogonids were selected for over more reef-associated pomacentrids. Feeding rates of both rock-cods were much higher in summer than winter and in summer they concentrated on small, recruit sized fish. There was little variation in feeding rates between patch and contiguous reef, however, despite apparent differences in prey abundance. The combination of high densities, high feeding rates and selection for certain sizes and types of prey suggested the two rock-cod species were having a considerable impact on populations and communities of their prey. The life history characteristics and population structure of Cephalopholis cyanostigma and C. boenak on patch and contiguous reef were also examined over a 2 year period. Knowledge of these patterns can provide indirect evidence of different processes limiting populations and the mechanisms that determine these effects. Tetracycline injection of tagged fish allowed for validation of annual increments in the otoliths of both species. Estimates of growth rate could therefore be obtained from both size-at-age data and recaptured tagged fish. Both species were relatively long-lived (32 years for C. cyanostigma and 16 years for C. boenak) and slow growing. Size-at-age curves of both species were almost identical on the two reef types, although growth of tagged C. boenak was slightly higher on patch reefs. Growth of tagged C. cyanostigma was significantly higher in summer than spring but this may have been due to either water temperature or prey abundance. Age and size-at-maturity also showed no difference between patch and contiguous reef. Age and size structures were similar on the two reef types for C. cyanostigma but there were more small / young C. boenak on patch reefs. C. boenak may undergo an ontogenetic shift from patch to contiguous reef. Based on longevity, mortality of both species was also similar on the two reef types. Sex ratios of C. boenak indicated they were monogamous on patch reefs but polygamous on contiguous reef. This may have been due to increased predation pressure on patch reefs. Overall, despite apparent differences in prey abundance there were few differences in the life history and population structure of the rock-cods on patch and contiguous reef. Differences in prey abundance appeared to be compensated for by territory size and the abundance of competitors and predators of the rock-cods themselves. The behavioural and developmental responses of Cephalopholis boenak to variation in prey abundance were further examined in a field experiment. Two hypotheses were tested: (1) that C. boenak would move from areas of low to high prey density and (2) that feeding and growth rates of C. boenak would be higher in areas of high prey density. Small patch reefs, which were equivalent in terms of habitat type, isolation and competitor and predator density, were used in the experiment. Over the 6 months of the experiment 31% of tagged C. boenak moved between patch reefs, all from reefs of low to high prey density. Feeding rates were also higher on patch reefs of higher prey to piscivore ratio. Due to few recaptures of tagged fish on reefs of low prey density, growth rates could not be compared within the experiment. Both feeding and growth rates on the experimental reefs, however, were much higher than on natural patch reefs over the same period. These patterns corresponded with much higher prey densities on experimental reefs. Hence these results suggest that growth of C. boenak was food-limited on natural patch reefs studied and that C. boenak responds both behaviourally and developmentally to variation in prey abundance. In summary, this study provided considerable evidence that piscivorous coral reef fish are limited by the availability of prey and are therefore likely to be in competition for this resource. This food-limitation appeared to cause numerical, behavioural and developmental responses of piscivores to variation in prey abundance. These responses may also have a stabilising effect on prey populations and communities. Predation by piscivorous fish is therefore likely to have been at least partly responsible for recent observations of density-dependent mortality of coral reef fish. This study has therefore supplied further support for the regulatory role of predation in coral reef fish communities