Structure and impact of the guild of epibenthic predators foraging on infaunal communities were measured in the Rhode River, a small mesohaline subestuary of Chesapeake Bay. Measures of long-term variation in guild structure (species composition, abundance, and size), patterns of prey utilization (stomach contents), and predator exclusion experiments assessed the interaction of guild structure and function. Monthly otter trawls from 1981 to 1988 caught 38 species, w hich utilized the subestuary on a markedly seasonal cycle with peak abundances during summer months. Species composition was significantly consistent among years, with the blue crab Callinectes sapidus, a sciaenid fish Leiostomus xanthurus or, in one year, Micropogonias undulatus, and the sole Trinectes maculatus comprising the dominant members of the guild. However, abundances of all 4 dominant species fluctuated significantly among years and, except for M. undulatus, among stations. Seasonally consistent patterns of population size structure showed that: L. xanthurus was composed of only 1 yearclass recruiting in May; M. undulatus had 1 year-class with recruitment in November; T. maculatus was composed of 3 year-classes with recruitment in October; and C. sapidus Iiad a 2 year-class population with recruitment in late fall and spring. Despite marked juvenile growth over the season, ontogenetic changes in diets of the 4 species occur at sizes smaller than those comprising the populations. All predators consumed a diversity of infaunal prey early in the summer, but their diets varied significantly during the season as prey availability changed. L. xanthurus and T. maculatus consumed mainly amphipods and polychaetes early in the season, and took increasing frequencies of clam siphons later; while M. undulatus primarily consumed amphipods. C. sapidus consumed mainly whole clams, other blue crabs and fish, but shifted from amphipods early in the season to increased frequencies of clams late in the season. Dietary breadth and overlap within the guild reflected the generalized, overlapping diets of the 4 species, although dietary breadth varied with predator species and season. For all species stomach contents did not vary with sediment type. Natural densities of Macoma balthica, a major dietary component for C. sapidus, responded to annual fluctuations in crab abundance as well as to variation in clam recruitment. Experimental exclusion of predators from clams (M. balthica) placed in buckets in the subestuary significantly increased clam survival. Experimental exclusion of epibenthic predators at 5 stations in the subestuary resulted in significantly higher densities for 12 of 16 infaunal species as well as for total infaunal organisms and higher ratios of infaunal predators to prey. However, most infaunal species exhibited significant variation among stations, reflecting spatial variation in both predation intensity and prey abundances . Experimental exclusion of the epibenthic predators from plots of dyed sand showed that the guild caused maj...