Non-lethal sampling of avian indicators reveals reliable geographic signals of mercury despite highly dynamic patterns of exposure in a large wetland

Abstract Mercury is a global contaminant with special relevance for aquatic food webs, where biomagnification can result in strong effects on apex predators. Non-lethal sampling of tissues such as blood and feathers is often used to assess mercury risk and spatiotemporal variability of mercury exposure on avian populations. However, the assumption that samples from individuals within a population are representative of local mercury exposure underpins those approaches. While this assumption may be justified, it is rarely expressed quantitatively. Further, the stability of the tissue/exposure relationship over time or space may depend on the sampling medium used, since some tissues and age classes may be better at reflecting local or short-term changes in exposure. Here, we present analyses of mercury concentrations from three tissues (albumen, blood and feathers) of the same individual great egret ( Ardea alba ) nestlings from breeding colonies in the Florida Everglades collected over three consecutive years. The interaction of year and colony location explained at least 50% of the observed variation in mercury concentration in all the sampled tissues. Annual colony-wide average Hg concentrations in any of the sampled tissues correlated with average Hg concentrations in the other two tissues from the same colony (R 2  > 0.53 in every case), while concentrations in albumen, blood and feathers from the same individual correlated poorly (R 2