CAGING EXPERIMENTS TO EXAMINE THE EFFECTS OF FISHES ON UNDERSTOREY SPECIES IN A SUBLITTORAL KELP COMMUNITY

The effects of predatory fish on the abundances of understorey species in a sublittoral Ecklonia radiata kelp assemblage were examined using exclusion cages. Fish were excluded from replicate areas for comparison with uncaged controls and appropriate treatments to test for caging artifacts. Experiments were done under the kelp canopy and in areas cleared of kelp (simulating damage due to storms). The understorey assemblage was sampled using quadrats for macroscopic species, an underwater microscope for attached microscopic organisms, and an underwater suction sampler for mobile microscopic fauna. Significant caging artifacts were found to act on certain species early in the experiments and on most species after 8 wk. Caging led to increases in abundances of many microscopic invertebrates, microscopic silt and layers of sediment. Caging led to decreases in the abundances of encrusting algae as experimental plots became dominated by layers of sediment. For some encrusting algae, Zonaria-Lobophora, Myxilla sp. and juvenile Ecklonia radiata, caging negated natural increases in abundances, whilst for sessile animals, caging negated a natural decrease. The species richness of the assemblage was greater in areas with fish than in areas excluded from fish. More species appeared to be affected by predation by fish under the canopy than in clearings due to the greater number of species living under the canopy. Some species showed significant effects (an increase, decrease or the negation of a natural change in abundance) at all times of the year when fish were excluded, whilst others only showed effects at times when they were in large abundances. After removing certain cages, fishes quickly returned the assemblage to one dominated by encrusting species. The results are discussed in terms of alternative models and hypotheses to account for the observed patterns (particularly effects of fish predation and caging artifacts), problems with using predator-exclusion cages in such experiments, and the significance of these experiments for future studies of these interactions.