Effects of Differential Fish Grazing on the Community Structure of an Intertidal Reef Flat at Enewetak Atoll, Marshall Islands·

The high and middle intertidal zones of the northeastern portion ("barren") of the limestone bench on the windward side of Enewetak Island, Enewetak Atoll, Marshall Islands, have a significantly higher number of herbivorous fishes grazing at high tide than the same intertidal zones of the reef flat 300 m to the southwest ("algae-covered"). This portion of the reef flat in the barren, heavily grazed area has a significantly lower coverage by erect, macroscopic algae and a lower algal biomass than the same portion of the reefflat in the algae-covered area. The removal of part of the limestone substratum by the grazing fishes as they feed and the reduced coverage by erect, macroscopic algae result in a lower topographic relief in the barren area than found in the algaecovered area. The heavily grazed area has a significantly lower number of mobile epifaunal invertebrate species and individuals per square meter than the lightly grazed area. Differences in infauna (sipunculans, polychaete worms, and tanaid crustaceans) are not so clear. When portions of the barren area were excluded from fish grazing activity for three months, the substratum under the exclosures had 100 %coverage by an algal mat; the density ofmobile invertebrate epifauna was an order ofmagnitude higher than in quadrats outside the exclosures. Although the high and middle intertidal community is subjected to apparently severe physical stresses (desiccation, insolation, wave shock, ultraviolet radiation, and osmotic stress from evaporation in the tide pools and rainfall), it appears to be principally structured by the grazing activities of herbivorous fishes. The high level of grazing in the barren area results in coverage by filamentous blue-green algae and a diatom-bacterial film, which may be a nutritionally more important food source to the fishes than the coralline algae in the algae-covered area. ECOLOGICAL STUDIES ofcommunities in recent years have gone beyond purely descriptive approaches to examine the formation and maintenance of community structure; that is, the number of species, their relative abundances, their roles in the community, and their responses to biological and physical events. Much of this effort has been with rocky intertidal communities in both temperate areas 1 The field work at Enewetak Atoll was supported by ERDA grants to the Mid-Pacific Marine Laboratory and funds from California State University, Long Beach. Manuscript accepted 9 June 1982. 2 California State University, Long Beach, Department of Biology, Long Beach, California 90840. (e.g., Connell 1961, 1970, Dayton 1971, Menge 1976, 1978, Osman 1977, Paine 1966, 1971, Paine and Vadas 1969) and tropical areas (e.g., Bakus 1967, Bernstein 1974, Brander, McLeod, and Humphreys 1971, Brock 1979, Connell 1978, Kohn 1959, 1967, Kohn and Leviten 1976, Leviten 1974, Menge and Lubchenco 1981, Miller 1974, Stephenson and Searles 1960, Vine 1974). From these and other community studies, Connell (1975) and Menge and Sutherland (1976) developed a general model that predicts in part that predation (including herbivory) is the most important biological interaction that structures physically benign communities. In communities influenced by