We measured the effects of filtration by the Atlantic ribbed mussel Geukensia demissa (Dillwyn, 1817) on microbial abundance in water exposed to the salt-marsh surface. Various groups of living and non-living particles (large and small autotrophs, large and small heterotrophs, cyanobacteria, bacteria, non-living clay-organic floc) were removed with greatly differing effectiveness, ranging from 25 to 95 O/o of initial abundance per hour Phytoplankton would contribute an estimated 72 "10 of livlng microbial carbon removed by field populations of mussels (47 O/O of living microbial carbon contnbuted by small autotrophs). Although an equal biomass of sn~all mussels was more effective at removing particles than large mussels, medium to large mussels would account for over 90 % of mussel grazing in the field because of their greater abundance. Particle removal was not solely dependent on particle size, as large heterotrophs and bacteria were removed with low (25 to 5 6 % h-') efficiency, whereas ~ntermediate-sized living particles and non-living clay-organic particles (ranging widely in size) were removed with high efficiency (up to 9S0/0 h-'). D~fferential removal of small heterotrophs (high efficiency) versus bacteria (low efficiency) may perturb the balance between bacterial production and microbial bacterivory in salt-marsh systems. The presence of G. den~issa enhanced bacterial production rates slightly; this could not be solely attributed to nitrogen excretion by the mussels. Mussel filtration was sufficient to balance bacterial production during high tide excursions of water onto the marsh, and may account for the net importation of bacteria from tidal creeks to the intertidal marsh reported in other studies. Present data indicate that ingestion of microbial carbon is not sufficient to meet the carbon requirements of G. demissa on an annual basis.
INTRODUCTIONThe feeding activity of suspension-feeding bivalve molluscs can have a profound influence on the abundance of water-colun~n microbiota in shallow water (Wright et al. 1982), and may be an important mechanism for coupling pelagic and benthic processes (Dame et al. 1980). Cloern (1982) suggested that the abundance of phytoplankton within the southern portion of San Francisco Bay is controlled by the effects of grazing by benthic bivalves, which could filter a water volume equivalent to the entire water column each day. Officer et al. (1982) considered the criteria which could lead to such control in south San Francisco Bay (shallow water, limited hydrodynamic exchange, dense benthic filter-feeding community) and identified 2 other estuarine systems which were also likely to exhibit benthic control of water-column microbial abundance. Peterson & Black (1987) concluded that benthic bivalves could potentially filter a large fraction of incoming tidal volume on sandflats of a subtropical embayment (Shark Bay, Australia). Sherr et al. (1986) found that in the Duplin h v e r estuary (Sapelo Island, Georgia, USA) much of bacterioplankton production was no...