As industrial development, aerial deposition of pollutants, and chemical releases from thawing permafrost increase in the Arctic, there is need to monitor exposure to contaminants in a range of species. Surrogate or bioindicator species are often used to represent groups of taxa and are typically selected for their wide geographic ranges and ease of sampling. However, it is seldom known how well surrogates exemplify different species, ages, or breeding stages in different areas. We examined multiple trace elements in incubating females of five sea duck species (long-tailed ducks Clangula hyemalis, Steller’s eiders Polysticta stelleri, spectacled eiders Somateria fischeri, king eiders S. spectabilis, and common eiders S. mollissima), in pre-laying spectacled eiders, and in brood-rearing female spectacled and king eiders and their ducklings, among four areas on the Arctic Coastal Plain of Alaska, USA. For Se, Cd, and Cu that are probably acquired during the nonbreeding period, Se and Cd in the blood of hens varied appreciably among species, whereas all species appeared to regulate Cu at similar concentrations. Levels of blood Hg and Pb sampled during incubation were inconsistent among species, apparently reflecting variable exposure on breeding areas. During incubation, species of similar body mass had similar patterns of elements. Se, Cu, and Pb differed substantially across the breeding season in spectacled eiders, as well as between ducklings and hens of spectacled and king eiders. Although analysis of feathers avoids invasive sampling, patterns of element concentrations in feathers differed from those in blood. Species of comparable body mass and life history traits sampled at a given site generally had similar concentrations of trace elements in blood, suggesting some utility as surrogates, but still showed regional variation. Our results suggest that for sea ducks, surrogate species of varying body mass and life history traits may not consistently reflect levels of exposure across taxa, or even among congeners or sex-age classes in the same areas. Biomonitoring for contaminants should use species of comparable mass as indicators at small spatial scales, but at the regional scale, should focus on specific taxa of concern.