Fecal bacterial contamination of nearshore waters has direct economic impacts to coastal communities through the loss of shellfisheries and restrictions of recreational uses. We conducted seasonal measurements of fecal coliform (FC) sources and transport pathways contributing to FC contamination of Buttermilk Bay, a shallow embayment adjacent to Buzzards Bay, MA. Typical of most coastal embayments, there were no direct sewage discharges (i.e., outfalls), and fecal bacteria from human, domestic animal, and wildlife pools entered open waters primarily through direct deposition or after transport through surface waters or groundwaters. Direct fecal coliform inputs to bay waters occurred primarily in winter (December-March) from waterfowl, ~33 x 1012 FC yr-1 or ~67% of the total annual loading. Effects of waterfowl inputs on bay FC densities were mitigated by their seasonality, wide distribution across the bay surface, and the apparent limited dispersal from fecal pellets. On-site disposal of sewage by septic systems was the single largest FC source in the watershed-embayment system, 460 x 1012 FC yr-1, but due to attenuation during subsurface transport only a minute fraction, < 0.006 x 1012 FC yr-1, reached bay waters (<0.01% of annual input to bay). Instead, surface water flows, via storm drains and natural streams under both wet- and dry-weather conditions, contributed the major terrestrial input, 12 x 1012 FC yr-1 (24% of annual input), all from animal sources. Since most of the surface water FC inputs were associated with periodic, short-duration rain events with discharge concentrated in nearshore zones, wet-weather flows were found to have a disproportionately high impact on nearshore FC levels. Elution of FC from shoreline deposits of decaying vegetation (wrack) comprised an additional coliform source. Both laboratory and field experiments suggest significant elution of bacteria from wrack, ~3 x 1012 FC yr-1 on a bay-wide basis (6% of annual input), primarily by periodic tidal flooding and possibly by major rain events. Release of coliforms during resuspension of subtidal sediments was estimated to be a minor source in this system (<1.5 x 1012 FC yr-1 or < 3% of annual input), primarily associated with large storm events in the fall and winter. Based upon the relative source strengths and the spatial and temporal patterns of FC input to Buttermilk Bay, it appears that management practices in similar settings should account for migratory waterfowl, but remediation efforts should focus on the redirection of stormwater runoff through the groundwater transport pathway.