A test of the Menge-Sutherland model of community organization in a tropical rocky intertidal food web.

Menge and Sutherland (1976) predicted that in physically benign habitats: (1) community structure will be most strongly affected be predation, (2) the effect of predation will increase with a decrease in trophic position in the food web, (3) trophically intermediate species will be influenced by both predation and competition, and (4) competition will occur among prey species which successfully escape consumers. These predictions were tested in a tropical rocky intertidal community on the Pacific coast of Panama. The most abundant mobile species included fishes and crabs, which occupied the top trophic level, and predaceous gastropods and herbivorous molluscs, which occupied intermediate trophic levels. The most abundant sessile organisms were encrusting algae, foliose algae, barnacles, and bivalves. Diets were broad and overlapping, and 30.3% of the consumers were omnivorous. Each consumer group had strong effects on prey occurring at lower trophic levels: (1) Fishes and crabs reduced the abundance of predaceous snails, herbivorous molluscs, foliose algae, and sessile invertebrates. (2) Predaceous gastropods reduced the abundance of herbivorous molluscs and sessile invertebrates. (3) Herbivorous molluscs reduced the abundance of foliose algae and young stages of sessile invertebrates, and altered relative abundances of the encrusting algae. The encrusting algae, although normally the dominant space occupiers, proved to be inferior competitors for space with other sessile organisms when consumers were experimentally excluded. However, the crusts escaped consumers by virtue of superior anti-herbivore defenses and competed for space despite intense grazing. Observations do not support the hypothesis that the trophically intermediate species compete. Hence, with the exception of this last observation, the predictions of the Menge and Sutherland model were supported. Although further work is needed to evaluate other predictions of the model in this community, evidence from this study joins an increasing body of knowledge supporting the model. Contradictory evidence also exists, however, indicating that aspects of the model require revision.