The curious phenomenon of similar levels of methylmercury (MeHg) accumulation in fish from contaminated and pristine environments may be explained by the observation that the proportion of total mercury (HgT) present as MeHg is inversely related to HgT in natural waters. We hypothesize that this "MeHg accumulation paradox" is explained by the quantitative induction of bacterial enzymes that are encoded by the mercury resistance (mer) operon, organomercury lyase (MerB), and mercuric reductase (MerA) by inorganic Hg (Hg[II]). We tested this hypothesis in two ecosystems in New Jersey: Berry's Creek in the Meadowlands (ML) and Pine Barren (PB) lakes. Across all sites, an inverse correlation (r2 = 0.80) between the concentration of HgT (ML, 113-4220 ng L(-1); PB, 0.3-5.4 ng L(-1)) and the proportion of HgT as MeHg (MeHg in ML and PB ranged from 0.08 to 1.6 and from 0.03 to 0.34 ng L(-1), respectively) was observed. The planktonic microbial community in Meadowlands surface waters exhibited adaptation to mercury, the presence of mer genes and mRNA transcripts, and high rates of reductive demethylation (k(deg) = 0.19 day(-1)). In contrast, the microbial community of PB was not adapted to mercury and demonstrated low rates of oxidative demethylation (k(deg) = 0.01 day(-1)). These results strongly support our hypothesis and show that the degradation of MeHg by mer-encoded enzymes by the water column microbiota of contaminated environments can significantly affect the amount of MeHg that is available for entry into the aquatic food web.