Field experiments withMercenaria mercenaria in a relatively high-energy environment demonstrated that clams on unvegetated sand flats failed to grow during autumn while those within seagrass beds grew substantially. Clam growth rates at the seagrass margin that first receives the faster-flowing, flood-tidal currents were about 25% less than at the opposite edge. In a second experiment, pruning, which reduced average blade length by 50-75%, was shown to enhance near-bottom current velocities and to reduce shell growth ofMercenaria during summer by about 50%. As in the first experiment, clams in the unvegetated sand flats exhibited no net growth. Clam mortality, caused mostly by predatory crabs and whelks, was much higher on sand flats than in seagrass beds and intermediate in clipped seagrass. Although consistent with some previous reports, these growth results are still surprising given that they contradict the generalization that suspension feeders grow faster under more rapid current regimes.Three types of indirect interactions might explain the observed effect of seagrass on growth of buried clams: (1) altering food supply; (2) changing the intensity of biological disturbance on feeding clams; and/or (3) affecting the physical stability of the sediments. Previous research on this question has focused almost exclusively on processes that alter food supply rates. In this study, food concentrations, as indicated by suspended chla, were 30% higher inside than outside one seagrass bed, whereas chla concentrations in two other beds were not different from those on adjacent sand flats. This result is sufficient to show that more intense food depletion was not induced by the reduction in flow velocities under the seagrass canopy. Nevertheless, the possible small difference in food concentrations between vegetated and unvegetated bottom seems insufficient to explain the absence of growth of sand-flat clams, especially given the virtual lack of food limitation among suspension feeders in this system. Two data sets demonstrated that the effects of biological disturbance agents cannot be ignored. An outdoor laboratory experiment showed that even in the absence of physical contact between predator and prey the presence of a whelk reduces the amount of time spent feeding byMercenaria. This result suggests that sand flats, where predation rates are higher, may be sites of lower clam growth than seagrass beds because of greater consumer interference with clam feeding. Furthermore, clam siphons are proportionately larger inside seagrass than on sand flats, implying that siphon nipping may not be as intense inside seagrass. This process, too, would reduce net growth of sand-flat clams. Finally, no explicit test was conducted of the hypothesis that enhanced sediment transport in the absence of flow baffling and root binding by seagrass inhibits net growth of clams on high-energy sand flats. Nevertheless, this is a reasonable explanation for the pattern of enhanced growth of seagrass clams, and could serve to explain the otherwise unexplained pattern of lower clam growth at the edge of the seagrass bed that experiences the faster flood-tidal current velocities. Each broad process, changing fluid dynamics, altering consumer access, and varying sediment stability, represents a mechanism whereby habitat structure, provided by the dominant plant, has an important indirect influence on the functional value of the habitat for resident animals.