Timing and magnitude of net methylmercury effects on waterbird reproductive output are dependent on food availability

Mercury (Hg) is a globally distributed pollutant. Its sub-lethal effects on reproduction of birds have been used as indi-cators of contamination and of potential demographic effects. However, studies typically used single endpoints that might not be representative of entire reproductive cycle. To estimate timing and net cumulative effects of Hg exposure under field conditions, we used observational data over 11 years from >1200 nests of great egrets breeding under tem-porally and spatially varying food availability and Hg exposures in the Florida Everglades. We collected measures of fish biomass and availability (>100 locations annually) and used four avian reproductive endpoints that represented the entire breeding cycle. We calculated net reproductive loss by adding estimated Hg effects on failures prior to egg laying, clutch size, hatching success and nestling survival in response to food availability and Hg exposure. To val-idate and assess results of the observational egret study, we ran the same analyses with data of captive breeding white ibises experimentally exposed to Hg with ad libitum food over 3 years. We found large (>50 %) reductions in great egret offspring with high Hg exposure (18 mu g/g dw THg nestling feather,-0.7 mu g/g ww whole egg THg) and high food avail-ability, and even larger reductions (up to 100 %) with high Hg exposure and low food. Timing and the relative contri-bution of different endpoints to overall reproductive failure varied with food availability. Failures prior to egg laying were relevant at all food availabilities and proportionally most important during high food availability (-70 % of total losses). Under high food, post-hatching failures increased moderately with increasing exposure (-10 % of total losses), and under low food, hatching failures became dominant (similar to 50 % of total losses). Patterns of failure of captive white ibis fed ad libitum resembled those of great egrets under high food availability but differed in total magnitude. We suggest that, a) net reproductive effects of Hg in free-ranging animals are probably much higher than generally reported in studies using single endpoints, b) Hg effect sizes vary considerably among different endpoints and c) food availability is a strong driver of timing and net effects of Hg exposure.