RESUMEN
Mercury (Hg) concentrations in aquatic biota, including fish and shellfish, were measured over the period 2006-2012 in the lower Penobscot River and upper estuary (Maine, USA). The Penobscot is a system contaminated with Hg by a chlor-alkali plant that operated from 1967 to 2000, discharging 6-12â¯tons of mercury into the river. Mercury levels in aquatic biota were highest at sites downstream of the chlor-alkali plant and spatial trends were similar to those of sediments. Mean total Hg concentrations in fish muscle (adjusted for size or age) in the most affected areas were 521 (480, 566; 95% CI) ng/g ww in American eels, 321 (261,395) in mummichog, 121 (104, 140) in rainbow smelt, 155 (142,169) in tomcod, 55.2 (42.7,71.4) in winter flounder, and 328 (259,413) in American lobster tail and 522 (488,557) ng/g dw in blue mussel. Levels exceeded the 50â¯ng/g ww considered protective for piscivorous predators and were of concern for human health, with American eels and American lobster exceeding Maine's mercury action level of 200â¯ng/g ww. Calculations of trophic position (using nitrogen isotopes) suggested that the spatial patterns observed in total Hg concentrations were not due to changes in feeding habits of the species. Fish feeding in benthic food webs, as defined by stomach content and stable carbon isotope analyses, showed no change in Hg concentrations over time. In contrast, declining trends in Hg were found in two species dependent on pelagic food webs. The absence of declines in Hg concentrations in the benthically-based food webs, despite the fact that most Hg was discharged into the system >40â¯years ago, is consistent with the long recovery predicted from dated sediment cores and from similar studies elsewhere.