To improve our understanding of how environmental mercury (Hg) concentrations influence Hg in fish, I conducted a field experiment to quantify the relative contributions of dietary and aqueous exposure to Hg levels in fish. To further assess the importance of water as a source of Hg to fish, a long-term dataset from the Mercury Experiment to Assess Atmospheric Loading in Canada and the United States was used as input data for two Hg bioaccumulation models: OneFish (accumulation from food and water) and Wisconsin (accumulation from food). Both approaches used enriched stable isotopes of Hg. Yellow perch accumulated 10-21% of their Hg directly from water. Wisconsin model predictions were significantly lower than observed fish Hg concentrations and OneFish predictions. These results suggest that waterborne Hg is an important contributor to Hg in fish and that the exclusion of water in bioaccumulation models may produce underestimates of fish Hg concentrations.