Published October 7 Vol. 40: 2539, 1987 Epibenthic fish communities on Florida Bay banks: MARINE ECOLOGY - PROGRESS SERIES Mar. Ecol. Prog. Ser. relations with physical parameters and seagrass cover Susan M. Sogard*,George V. N. Powell, Jeff G. Holmquist8* National Audubon Society Research Department, 115 Indian Mound Trail, Tavernier, Florida 33070, USA ABSTRACT. Epibenthic fish comn~unities residing in seagrass beds on shallow (<0.5 m) mudbanks in Florida Bay. USA, were quantitatively sampled with a throw trap method. The overall average density of 11 fish m-2 was substantially higher than most previously reported densities for seagrass habitats. Four sites, representing 4 different subenvironments of Florida Bay, differed widely in species composition and densities of individual species; results of discriminant function analysis indicated that fish com- munities at the 4 sites were relatively distinct. Species composition at different sites is proposed to be a result of complex interactions between the deterministic influence of habitat quality and the stochastic influence of larval availability. Restricted water circulation, effected by the network of banks, and different sources of water mass exchange are proposed as constraints on larval avdability. Differences in species richness and fish densities across individual banks corresponded to gradients in depth, sediment structure, detrital loads, and various measures of seagrass structural complexity. The greater physical stress on top of a bank appeared to limit species richness, while fish densities across in&vidual banks were regulated by habitat gradients. Multiple regression analysis indicated that the standing crop of seagrasses and the accumulation of vegetation litter were important determinants of fish densities; physical factors, such as depth and sediment structure, were also influential. INTRODUCTION An extensive body of information has been published regarding ecological relations of fishes inhabiting sea- grass ecosystems (see Pollard 1984 for recent review). One area of particular interest concerns the role of the structural complexity of a seagrass bed in determining fauna1 densities. The functional means by which sea- grass structure enhances densities is thought to be a reduction in predation risk and/or increased food availability (e.g. Heck & Orth 1980, Orth & Heck 1980, Stoner 1983). While the importance of the architectural structure of the seagrass canopy has been examined for several infaunal and epifaunal invertebrates (Orth et al. 1984 and references therein), only limited information is available for fishes. Martin & Cooper (1981), Huh (1984), and Middleton et al. (1984) noted contrasts in Present address: Rutgers Marine Field Station, PO Box 278, Tuckerton, New Jersey 08087, USA ' ' Present address: Dept of Biological Sciences, Florida State University, Tallahassee, Florida 32306, USA C' Inter-Research/Printed in F. R. Germany the fish communities in meadows dominated by differ- ent seagrass species, but did not quantify differences in the physical structure of the grassbeds. Orth & Heck (1980) found an association of both species richness and total abundance with eelgrass biomass in Chesapeake Bay, but did not separate this effect from regular seasonal cycles in abundance. Stoner (1983) specifically examined different aspects of the seagrass canopy, and found fish density most correlated with seagrass biomass in one location, but with blade density in another location. Bell & Westoby (1987a) found fish densities associated with canopy height and seagrass density on a narrow scale, but the relationship weakened when examined on a broader scale (Bell & Westoby 1987b). In addition to mechanisms of predation and food availability (as reflected in correlations of fish density with seagrass architecture), the grassbed fish commun- ity is also structured by physical characteristics, includ- ing water temperature, salinity, and depth (Livingston 1982), the presence of additional algal and sponge microhabitats (Weinstein & Heck 1979, Heck & Orth 1980), and water circulation patterns (Adams 1976). In