Dissolved and particulate Ag concentrations (Ag(D) and Ag(P), respectively) were measured in surface water and suspended particulate matter (SPM) along the salinity gradient of the Gironde Estuary, South West France, during three cruises (2008-2009) covering contrasting hydrological conditions, i.e. two cruises during intermediate and one during high freshwater discharge (~740 and ~2,300 m(3)/s). Silver distribution reflected non-conservative behaviour with 60-70 % of Ag(P) in freshwater particles being desorbed by chlorocomplexation. The amount of Ag(P) desorbed was similar to the so-called reactive, potentially bioavailable Ag(P) fraction (60 ± 4 %) extracted from river SPM by 1 M HCl. Both Ag(P) (0.22 ± 0.05 mg/kg) and Ag(P)/Th(P) (0.025-0.028) in the residual fraction of fluvial and estuarine SPM were similar to those in SPM from the estuary mouth and in coastal sediments from the shelf off the Gironde Estuary, indicating that chlorocomplexation desorbs the reactive Ag(P). The data show that desorption of reactive Ag(P) mainly occurs inside the estuary during low and intermediate discharge, whereas expulsion of partially Ag(P)-depleted SPM (Ag(P)/Th(P) ~0.040) during the flood implies ongoing desorption in the coastal ocean, e.g. in the nearby oyster production areas (Marennes-Oléron Bay). The highest Ag(D) levels (6-8 ng/L) occurred in the mid-salinity range (15-20) of the Gironde Estuary and were decoupled from freshwater discharge. In the maximum turbidity zone, Ag(D) were at minimum, showing that high SPM concentrations (a) induce Ag(D) adsorption in estuarine freshwater and (b) counterbalance Ag(P) desorption in the low salinity range (1-3). Accordingly, Ag behaviour in turbid estuaries appears to be controlled by the balance between salinity and SPM levels. The first estimates of daily Ag(D) net fluxes for the Gironde Estuary (Boyle's method) showed relatively stable theoretical Ag(D) at zero salinity (Ag (D) (0) = 25-30 ng/L) for the contrasting hydrological situations. Accordingly, Ag(D) net fluxes were very similar for the situations with intermediate discharge (1.7 and 1.6 g/day) and clearly higher during the flood (5.0 g/day) despite incomplete desorption. Applying Ag (D) (0) to the annual freshwater inputs provided an annual net Ag(D) flux (0.64-0.89 t/year in 2008 and 0.56-0.77 t/year in 2009) that was 12-50 times greater than the Ag(D) gross flux. This estimate was consistent with net Ag(D) flux estimates obtained from gross Ag(P) flux considering 60 % desorption in the estuarine salinity gradient.