The use of organochlorine pesticides, including DDTs, chlordanes, and dieldrin, peaked in San Francisco Bay's watershed 30-40 years ago, yet residues of the pesticides remain high. Known as legacy pesticides for their persistence in the Bay decades after their uses ended, the compounds and their breakdown products occur at concentrations high enough to contribute to advisories against the consumption of sport fish from the Bay. Combined with other data sets, the long-term monitoring data collected by the San Francisco Estuary Regional Monitoring Program (RMP) for trace substances allow us to track recovery of the Bay from these inputs and predict its future improvement. Legacy pesticides enter the water and sediment of San Francisco Bay from a variety of sources, including runoff from California's Central Valley and local watersheds, municipal and industrial wastewater, atmospheric deposition, erosion of historically contaminated sediment deposits, and dredging and disposal of dredged material. Runoff from small-urbanized tributaries may contribute as much or more to the loads than runoff from the agricultural Central Valley, even though 90 percent of the freshwater flow comes from the Central Valley via the Sacramento and San Joaquin rivers. The fates of legacy pesticides in San Francisco Bay are controlled by their chemical properties, including their solubilities and partition coefficients. Degradation in the sediments, outflow through the Golden Gate, and volatilization-in that relative order-result in removal of pesticides from the Bay. A contaminant fate model was used to estimate recovery times of the Bay under various scenarios. For example, under a scenario in which no new legacy pesticides entered the Bay, model predictions suggested that concentrations of pesticides in the water and the active sediment layer would reach risk-reduction goals within one to three decades. Under scenarios of continued inputs to the Bay, recovery time would be considerably longer or not reached at all. Long-term tissue monitoring corroborates model predictions of slow declines in DDT and chlordane concentrations. Field-transplanted bivalve samples indicate declines since 1980, and lipid-weight concentrations of pesticides have declined in fishes, but the declines are slow. The critical management question for the Bay is whether there are feasible management actions that would decrease concentrations in sport fish significantly faster than the existing slow progress that has been observed.