Individual behavioural responses are key to predicting the long-term effects of environmental stressors on wildlife. Breeding individuals are particularly vulnerable as they are constrained to breeding areas and must balance self-maintenance with reproduction. This thesis examined two important stressors in incubating birds: predation risk and contamination by per- and polyfluoroalkyl substances (PFAS). I investigated links between these two stressors and incubation behaviour in the common goldeneye (Bucephala clangula) and the common eider (Somateria mollissima). Chapters I and II focused on anti-predator behaviour in female goldeneyes. Anti-predator responses showed both repeatability and within individual plasticity. Females became bolder with age and clutch size up to a nest parasitism threshold, beyond which older females with higher clutches became comparatively shyer. Although anti-predator behaviour was not associated with hatching success, females exhibiting exceptionally bold behaviour were more likely to be killed during predator attacks, indicating potential fitness costs of behavioural strategies. In Chapters III and IV, I examined nest attendance using artificial egg loggers recording egg-turning frequency and egg temperature. While egg-turning frequency did not differ between species, goldeneyes maintained higher egg temperature and showed greater behavioural consistency than eiders. Higher perfluorooctanesulfonic acid (PFOS) burdens were associated with lower egg temperature in eiders (Chapter III). Although in Chapter IV plasma prolactin and corticosterone levels were positively associated with nest attendance in eiders, I found no clear evidence that PFAS-hormone associations explain variation in nest attendance. Overall, this thesis demonstrates that behaviour can carry fitness costs and that species differences may modulate contaminant-behaviour associations. By integrating behavioural ecology, ecophysiology, and ecotoxicology, this work highlights behaviour as a meaningful endpoint in wildlife monitoring under multiple stressors.