REFUGE AVAILABILITY STRUCTURES ASSEMBLAGES OF TROPICAL REEF FISHES

1. We examined assemblage structure of tropical fishes on small, artif icial reefs to determine if differences in refuge availability could m odify patterns of species richness and abundance of fishes established at recruitment. 2. Our artificial reefs were designed to provide prey refuge of two types (permanent and transient), Permanent refugia are those that physically exclude predators, i.e. small holes provide perm anent refuge from large-bodied predators. Transient refugia result whe n habitat complexity increases the probability that prey will elude pr edators. These refugia do not physically exclude predators and thus pr ovide no permanently safe sites. 3. We conducted this experiment at tw o widely separated locations on Australia's Great Barrier Reef. The sp ecies pools of reef fishes, recruitment rates and predator densities a ll differ between these locations. 4. At neither location were pattern s of recruitment influenced by the presence of either type of refuge. By the end of the experiment, however, there were more resident fishes on reefs with additional refugia. 5. Species richness of residents wa s positively related to total abundance of residents. Therefore, refug e availability indirectly affected species richness through its effect on abundance. There was no indication, however, that permanent refugi a provided any greater protection to prey species than did transient r efugia. 6. Our results therefore indicate that for these communities, patterns of species richness and abundance established at settlement c an be modified considerably over small spatial and temporal scales by differences in refuge availability. Furthermore, habitat complexity ne ed not provide permanently safe sites to affect patterns of coexistenc e.