A simple, catchment-scale, cascade model was used to assess the importance of sinks and sources of methylmercury (MeHg) in a boreal catchment that contains a forested upland, a lowland peatland and a small lake. The three compartment model was run using realistic flow rates and atmospheric loading of MeHg, and the model was constrained by observed concentrations of MeHg in each compartment. Assuming no internal sinks and sources of MeHg, modelled catchment yields showed reasonable agreement with field observation, but the predicted internal MeHg concentrations in each compartment were implausible. Only when sources and sinks of MeHg are added to the three compartments do MeHg-pool concentrations fall into the range of those measured in the field. To maintain both catchment-scale and compartment-scale continuity, the upland and peatland were net sources of MeHg (0.0007 and 0.1065 mg ha(-1) d(-1) respectively), and the lake a net sink (-0.2215 mg ha(-1) d(-1)). These source/sink rates are 1.73, 259 and -539 times the input of MeHg via wet precipitation input for the modelled ice-free season. Sensitivity analysis revealed that the volume of runoff delivered to the peatland by the upland area, peatland size and porewater MeHg concentration in the peatland an important controls on catchment MeHg yield, and that contemporary atmospheric deposition of MeHg is insignificant compared to the sources of MeHg within the catchment.