Behavioral ecology of coral reef fishes at spawning aggregation sites

Abstract

Coral reef fishes aggregate to spawn on certain locations of reefs. This thesis is an extensive investigation of the behavioral and ecological relations between spawning reef fishes, predators and their environment at a spawning aggregation site. Many hypotheses have been proposed regarding the adaptive nature of different diel reproductive cycles observed in coral reef fishes. This study quantified the spawning patterns of eleven different reef fish species at one location (Johnston Atoll, Central Pacific), while making simultaneous measurements of the environmental factors likely to affect the spawning behavior of reef fishes. The environmental variables measured included time of day, tides, current velocity, current speed and abundance of piscivorous predators, and were correlated with observed spawning outputs through multifactorial analyses. High interspecific variability in spawning patterns was found among the eleven monitored species. The majority of species spawned at a specific time of the day, in agreement with the timing of spawning described at other locations, indicative of a fixed general response by fishes across distribution areas. Spawning of most fishes with daytime spawning peaks was correlated with local changes in current direction and predatory risks, showing responses designed to reduce the mortality of propagules and adults. Dusk-spawning species generally did not respond to changes in flow direction and predator abundance, most likely due to their short spawning periods and the reduced predatory pressures that occurred at dusk. Tides did not seem to be used exclusively as synchronizing cues to adult fishes for spawning. The influence of current speed in determining diel timing of spawning varied among species, with some species showing responses to current speed while others showed no response. Predation is a selective force hypothesized to influence the spawning behavior of coral reef fishes. This study describes and quantifies the predatory activities of two piscivorous and three planktivorous species at a coral reef fish spawning aggregation site in Johnston Atoll (Central Pacific). To characterize predator-prey relations, the spawning behavior of prey species was quantified simultaneously with measurements of predatory activity, current speed and substrate topography. Diel activity patterns and predator-prey relations varied among the predatory species analyzed. The activity patterns of piscivores, measured both as abundance and attack rates, were high during the daytime, decreased during the late afternoon hours and reached a minimum at dusk. The abundance of piscivores was significantly correlated with the abundance of prey for only one (Caranx melampygus) of the two piscivorous species, while the other species (Aphareus furca) did not respond to prey abundance. The selection of certain prey species by piscivores was consistent with two different hypotheses: the satiation of predators and the differences in spawning behaviors among prey species. Two of the three planktivorous species fed most actively at dusk, and selected as prey those species of reef fishes that produced eggs of large size. The third planktivorous species fed at all times of the day. Spawning prey fishes were more abundant over substrates with complex topography where refuges from piscivores were abundant than over smooth substrates. Overall attack rates by piscivores on adult spawning fishes were higher than by planktivores feeding on recently released eggs. The diel spawning patterns displayed by reef fishes at the study site seem to be influenced by the diel activity and prey selection patterns of piscivores previously described. The highest diversity of prey species occurred at dusk, when piscivores were least abundant and overall abundance of prey fishes was lowest. The behavioral strategies used by the piscivore Caranx melampygus (Carangidae) while feeding on spawning aggregations of coral reef fishes were studied for two years at Johnston Atoll (Central Pacific). Visual behavioral observations revealed the existence of two different hunting behaviors employed by this predator. A 'midwater' hunting behavior, which consisted of midwater high speed attacks on spawning fishes, is typical of large sized transient predators and yielded a low capture success rate (2%). An 'ambush' hunting behavior consisted of attacks on spawning fishes from hiding locations in the substrate, and yielded a much higher capture success rate (17%). While ambushing their prey, C. melampygus displayed territorial aggressive behaviors toward other intruding conspecifics, defending a specific section of the reef. This specialized ambushing behavior is atypical of fast swimming carangids, but illustrates the behavioral flexibility of this predator. I suggest that the use of these two hunting behaviors by C. melampygus can potentially cause density-dependent mortality rates in prey communities, a demographic consequence previously attributed to the simultaneous action of various guilds of predatory species. Two species of trunkfishes (Ostraciidae) were observed spawning above a coral reef at Johnston Atoll (Central Pacific). This study analyzed the potential causes determining the difference in spawning ascent height in Ostracion meleagris (3.3 m average) and 0. whitleyi (1.5 m average). One hypothesis proposes that the risk of predation by piscivores influences how far each species can swim from the substrate, and that predation risk is greater for 0. whitleyi than 0. meleagris. Trunkfishes have an armoured exoskeleton and secrete an ichthyotoxic mucous under stress conditions, two defenses against predation. Because the two species used the same spawning grounds and spawned at approximately the same time, their size and toxicity levels were analyzed to assess their susceptibility to predation. Toxins were extracted from wild fishes and tested using a mosquitofish assay. Ostracion whitleyi was more toxic than 0. meleagris, refuting the predation-risk hypothesis. A second hypothesis proposes that long ascents are a way for spawning pairs to avoid disturbances by other male conspecifics. Observations of the spawning behaviours of the two species showed that male 0. meleagris were frequently involved in fighting episodes and showed high rates of male streaking (intruding nonpaired males attempting to fertilize eggs from spawning paired females), while none of these behaviours were observed in 0. whitleyi. The larger spawning height from the substrate may be an attempt by pairs of 0. meleagris to reduce the possibility of interference by other male conspecifics. Thus, the height of spawning ascents corresponds to the expectation from the male disturbance hypothesis, but not to the expectation of the predatory risk hypothesis.