The epifauna on the shells of a subtidal population of the bivalve Pinna bicolor is composed mainly of encrusting bryozoans and serpulid polychaetes, with tunicates and sponges relatively uncommon. Only a small subset of the observed interactions between sessile organisms were capable of changing the patterns of abundance on individual Pinna shells. Didemnid tunicates were capable of excluding most other species, but the most common competitive interactions, those between encrusting bryozoans, had variable outcomes and frequently did not result in the exclusion of the loser from a patch (shell). Sponges were also capable of excluding bryozoans and serpulids, although at slower rates than tunicates. Serpulids were overgrown by all other species, although overgrowth usually did not result in their exclusion from patches. Predators were, in general, not capable of influencing the abundance of any species greatly. The exception involved two species of monacanthid fish, juveniles of which ate newly metamorphosed didemnid ascidians. These fish thus could prevent monopolization of some patches by tunicates. The effects of predation by juvenile fish and overgrowth by tunicates were variable in space and time. The interaction between variable tunicate recruitment and predation was important. Didemnids recruited in low numbers and, even in a year of high recruitment, did not colonize a large number of patches; in addition, most recruits were eaten by fish. Predation by fish played no role in another year, because few tunicates recruited. It is likely that the abundance of fish was also variable, since their recruitment is also likely to be variable in time. The combination of these factors has resulted in few patches being colonized by tunicates. In the 2 yr of the study, none of the fifty patches under observation was colonized successfully by didemnids, and observations on other Pinna shells suggest that no substantial establishment occurred in the 2—3 yr preceding this study. Thus, in most patches, predation and competition did not strongly influence the abundance of sessile organisms. The patterns of occupation of individual patches, and the overall abundance of sessile species in the community, could be explained adequately by the observed patterns of recruitment. This situation contrasts with the occupation of patches on the pilings of a nearby pier, where competitive interactions between occupants of a patch are the most important factor. The results of these experiments provide a mechanism to explain the previously noted difference in the abundance of sessile invertebrates on the shells of Pinna bicolor and on the pilings of the pier.