Songbird feathers as indicators of mercury exposure: high variability and low predictive power suggest limitations

Abstract

Although feathers are commonly used to monitor mercury (Hg) in avian populations, their reliability as a sampling matrix has not been thoroughly assessed for many avian species, including most songbirds (Order Passeriformes). To better understand relationships between total Hg (THg) concentrations in feathers and other tissues for birds in the thrush and sparrow families, we (1) examined variation in THg concentrations among tissues, including feathers from six different tracts, nails, liver, and muscle; (2) tested relationships between THg concentrations in the various feather tracts and those in internal tissues from the same birds, to assess the predictive power of feather THg, and; (3) compared these relationships to those between THg concentrations in nails and internal tissues, to assess the viability of nails as a non-lethal sampling alternative. THg concentrations in all feather tracts and nails were consistently higher than those in the liver and muscle, and THg was higher in the thrushes than the sparrows. When comparing feather tracts, we observed high variation within some individuals, suggesting that estimates of Hg exposure could vary depending on which feather was sampled. Despite this variation, feather type had little effect on the predictive power of feather THg concentrations, which ranged from extremely weak in the sparrows (0.09 ≤ R2 ≤ 0.16) to moderate (0.29 ≤ R2 ≤ 0.42) in the thrushes. Alternatively, we found that nail samples better predicted internal tissue THg concentrations in both the thrushes (0.44 ≤ R2 ≤ 0.48) and sparrows (0.70 ≤ R2 ≤ 0.78). Nails have been used to monitor Hg in mammals and reptiles, but their reliability as a sampling matrix for monitoring Hg in avian populations has yet to be assessed for most taxa. While nails exhibit stronger relationships to internal tissue THg concentrations, they may not be an effective sampling option for all avian species because the collection of sizable nail samples could harm living birds, particularly small songbirds. However, this method may be reasonable for retrospective museum studies. Overall, our results suggest that, despite their current use in the literature, feathers are not a suitable sampling matrix for Hg monitoring in some songbird species.