Both storm water event and nonevent flow contributed to the annual discharge from Graywood Gully, a small sub-watershed of Conesus Lake, New York USA, whose land use is 74% agriculture. While events contributed significant amounts of water in short periods of time, nonevents accounted for the majority of water on a yearly basis and could have flow rates matching those that occurred during events. Event storm water was elevated in materials associated with particulates such as total suspended solids, total Kjeldahl nitrogen, and total phosphorus. Water from high flow nonevents was elevated in soluble components such as sodium, nitrate, and soluble reactive phosphorus. As a result, events contributed the majority of particulates to the yearly loading from Graywood Gully whereas nonevents contributed the majority of soluble materials. The levels of total coliforms, Escherichia coli, Enterococcus, and total heterotrophic bacteria were elevated in storm water relative to nonevent flow, indicating that they acted as particulates. The median level of E. coli in nonevents was 200 CFU/100 mL whereas the median level during events was 3660 CFU/100 mL. Consequently, storm events accounted for 92% of all E. coli loading from Graywood Gully. Best Management Practices (BMPs) resulted in the mean, median, maximum and minimum levels of event-driven E. coli loading from Graywood Gully to decrease 10 fold over a 5-year period. The implementation of BMPs in the Graywood Gully watershed has improved the microbiology of event waters and consequently decreased the role that the watershed plays as a contributor of microbial pollution to Conesus Lake.