PRIORITY EFFECTS IN CORAL REEF FISH COMMUNITIES

Demographically open communities are often viewed as stochastically structured assemblages because most colonizing juveniles arrive via unpredictable dispersal mechanisms. However, interactions between established residents and incoming juveniles may affect juvenile persistence in species-specific ways and could therefore impose a degree of determinism on future community structure. Using 16 spatially isolated communities of coral reef fishes, I conducted two experiments to determine how prior residency by two guilds of fishes affected juvenile recruitment. Each experiment factorially manipulated the presence and absence of two guilds: resident piscivores (groupers and moray eels) and interference competitors (territorial damselfishes). In the first experiment, guilds were manipulated via selective removals, and subsequent recruitment (larval settlement minus mortality) was monitored for 44 days. In the second experiment, guilds were placed within large cages to prevent direct resident–juvenile interactions, while allowing for any cues produced by enclosed fishes, thereby testing whether incoming larvae used resident-derived cues to select or reject settlement sites. Colonizing juveniles were collected from each reef over 42 days to prevent confounding resident- and recruit-derived cues. In the first experiment, piscivores inhibited recruitment of a damselfish (Pomacentridae) and a surgeonfish (Acanthuridae), and enhanced recruitment of a wrasse (Labridae). In contrast, territorial damselfishes inhibited recruitment of the damselfish and the wrasse, and enhanced recruitment of the surgeonfish. Observations of early recruitment patterns suggested that recruitment differences were established rapidly during the night or dawn periods shortly after settlement and before each daily census. In the second experiment, there was no evidence that larvae used resident-derived cues to select settlement sites, suggesting that recruitment differences in the first experiment resulted from differential mortality caused by direct resident–recruit interactions rather than differential larval settlement. These results demonstrate that interactions between established residents and newly arrived juveniles can have a strong influence on juvenile persistence, and that such interactions appear to be strongest within hours of larval settlement. Furthermore, because resident effects were species specific, the present composition of these communities may impose a previously undocumented degree of determinism on their future structure.