For the past 9 years, we experimentally flooded a wetland complex (peatland surrounding an open water pond) at the Experimental Lakes Area (ELA), northwestern Ontario, Canada, to examine the biogeochemical cycling of methyl mercury (MeHg) in reservoirs. Using input-output budgets, we found that prior to flooding, the wetland complex was a net source of approximately 1.7 mg MeHg ha(-1) yr(-1) to downstream ecosystems. In the first year of flooding, net yields of MeHg from the reservoir increased 40-fold to approximately 70 mg MeHg ha(-1) yr(-1). Subsequently, annual net yields of MeHg from the reservoir declined (10-50 mg MeHg ha(-1) yr(-1)) but have remained well above natural levels. The magnitude and timing of Hg methylation in the flooded peat portion of the wetland reservoir were very different than in the open water region of the reservoir. In terms of magnitude, net Hg methylation rates in the peat in the first 2 years of flooding were 2700 mg ha(-1) yr(-1), constituting over 97% of the MeHg produced at the whole-ecosystem level. But in the following 3 years, there was a large decrease in the mass of MeHg in the flooded peat due to microbial demethylation. In contrast, concentrations of MeHg in the open water region and in zooplankton, and body burdens of Hg in cyprinid fish, remained high for the full 9 years of this study. Microbial activity in the open water region also remained high, as evidenced by continued high concentrations of dissolved CO2 and CH4. Thus, the large short-term accumulation of MeHg mass in the peat appeared to have only a small influence on concentrations of MeHg in the biota; rather MeHg accumulation in biota was sustained by the comparatively small ongoing net methylation of Hg in the flooded pond where microbial activity remained high. In large reservoirs, where the effects of wind and fetch are greater than in the small experimental reservoir we constructed, differences can occur in the timing and extent of peat and soil erosion, effecting either transport of MeHg to the food chain or the fueling of microbial activity in open water sediments, both of which could have important long-term implications for MeHg concentrations in predatory fish.