The structure of the salinity field in northern San Francisco Bay and how it is affected by freshwater flow are discussed. Two datasets are examined: the first is 23 years of daily salinity data taken by the U.S. Bureau of Reclamation along the axis of northern San Francisco Bay; the second is a set of salinity transects taken by the U.S. Geological Survey between 1988 and 1993. Central to this paper is a measure of salinity intrusion, X₂: the distance from the Golden Gate Bridge to where the bottom salinity is 2 psu. Using X₂ to scale distance, the authors find that for most flow conditions, the mean salinity distribution of the estuary is nearly self-similar with a salinity gradient in the center 70% of the region between the Golden Gate and X₂ that is proportional to X⁻¹₂. Analysis of covariability of Q and X₂ showed a characteristic timescale of adjustment of the salinity field of approximately 2 weeks. The steady-state response deduced from the X₂ time series implies that X₂ is proportional to riverflow to the 1/7 power. This relation, which differs from the standard 1/3 power dependence that is derived theoretically assuming constant exchange coefficients, shows that the upstream salt flux associated with gravitational circulation is more sensitive to the longitudinal salinity gradient than theory supposes. This is attributed to the strengthening of stratification caused by the stronger longitudinal salinity gradient that accompanies larger river flows.