Nearly all ecosystems are contaminated with highly toxic methylmercury (MeHg), but the specific sources and pathways leading to the uptake of MeHg within and among food webs are not well understood. In this study, we report stable mercury (Hg) isotope compositions in food webs in a river and an adjacent forest in northern California and demonstrate the utility of Hg isotopes for studying MeHg sources and cross-habitat transfers. We observed large differences in both δ(202)Hg (mass-dependent fractionation) and Δ(199)Hg (mass-independent fractionation) within both food webs. The majority of isotopic variation within each food web could be accounted for by differing proportions of inorganic Hg [Hg(II)] and MeHg along food chains. We estimated mean isotope values of Hg(II) and MeHg in each habitat and found a large difference in δ(202)Hg between Hg(II) and MeHg (∼2.7‰) in the forest but not in the river (∼0.25‰). This is consistent with in situ Hg(II) methylation in the study river but suggests Hg(II) methylation may not be important in the forest. In fact, the similarity in δ(202)Hg between MeHg in forest food webs and Hg(II) in precipitation suggests that MeHg in forest food webs may be derived from atmospheric sources (e.g., rainfall, fog). Utilizing contrasting δ(202)Hg values between MeHg in river food webs (-1.0‰) and MeHg in forest food webs (+0.7‰), we estimate with a two-source mixing model that ∼55% of MeHg in two riparian spiders is derived from riverine sources while ∼45% of MeHg originates from terrestrial sources. Thus, stable Hg isotopes can provide new information on subtle differences in sources of MeHg and trace MeHg transfers within and among food webs in natural ecosystems.