Effects of predators on reef fishes: separating cage artifacts from effects of predation

Predators are thought to play prominent roles in determining population abundance and dynamics of many species, but it is often difficult to document effects of predators in the field. For lack of any other effective means of manipulating predator densities, many studies use predator exclosures, however, the results of these studies are difficult to interpret because the exclosures may introduce artifacts that cannot be separated from predatory effects. Using a design common in tests of predatory effects, exclosure cages (? predators), "cage controls" = partial cages ( + predators), and no cages ( + predators), patterns of survival were found for two reef fishes that were consistent with both predator and cage effects. To separate cage artifacts from effects of predators, direct tests of cage effects in an environment free of predators were conducted, within a large predator exclosure. Cage artifacts had dramatic effects on one fish species but not the other. Moreover, the way the affected species responded to cages was exactly what would be expected if predators had strong negative effects on these fish: growth and survivorship were much lower on uncaged reefs than on caged reefs. To circumvent the cage effects, rather than compare uncaged reefs ( + predators) to reefs in exclosure cages (? predators), exclosures were compared to partial cages ( + predators). In the absence of predators, the partial cages did not differ from exclosure cages in their effects on the prey. Using the partial cages and exclosure cages in an area where predators were present, predator effects on the two fishes were tested. Survivorship of one species was greatly reduced by predators, but the other species was not affected. The lack of a predator effect on the one species was apparently an artifact of comparing completely caged reefs to partially caged reefs; another experiment documented strong predator effects on this species. Partial cages probably greatly reduced the ability of predators to consume the "unaffected" species. When cage artifacts affect prey, partial cages may be usefully employed to avoid confounding predator effects with cage effects at the cost of underestimating the magnitude of predator effects. Direct tests of cage effects are necessary to interpret the results of any study that uses cages to test for predatory effects.