Background/Question/Methods   Coral reef fisheries tend to target large-bodied, predatory species, providing the potential for profound alterations of the trophic structure and ecology of reef ecosystems. For example, uninhabited islands of the Pacific support high total fish biomass with a large proportion (frequently >50%) concentrated in large size classes and higher trophic levels. In contrast large-bodied predators are rare or absent at densely populated islands, accounting for only a small fraction (<10%) of total fish biomass. Despite clear patterns of fishery extractions, it remains unknown how the removal of marine predators affects the dynamics of lower-trophic level species. We used quantitative data from over 50 U.S. and Kiribati Pacific Islands spanning across a gradient of anthropogenic disturbance to explore the relationship among predators and lower-trophic level fishes. Our analysis focuses on identifying shifts in species composition and size-frequency distributions across gradients of predator abundance and fisheries extraction.

Results/Conclusions   Our results revealed little evidence of well-defined trophic cascades within the fish assemblage. As predator biomass is reduced, the biomass of other trophic guilds showed no consistent shifts. However, we did find clear evidence that the species composition and size-structure of the prey assemblage shifts depending on the abundance and biomass of top-level predators. Across gradients of decreasing predator biomass (from uninhabited, unfished islands to increasingly fished/inhabited islands), we found consistent evidence of reductions in size distributions across the fish assemblage. The reduction in size distribution is driven principally by the replacement of large-bodied species by smaller-bodied species within each trophic level. Surprisingly, we found an opposite pattern within fish species – along a gradient of decreasing predator biomass we found species-specific size distributions to shift to larger sizes. Taken together, these data suggest that predators significantly reduce survivorship of all species, but disproportionately reduce survivorship of smaller relative to larger species. We consider the functional implications of these indirect shifts associated with reduction of predator density. In particular, we suggest that predators may indirectly increase the productivity of the fishery by amplifying growth rates of the prey fish assemblage, specifically by favoring large-bodied species that are far from their somatic maximum size.