Evaluating consequences of habitat selection is an important step in understanding life history strategies and behavioural decisions of animals. Kilpi and Lindström (1997) found that incubating common eiders Somateria mollissima on exposed, treeless islands lost weight faster than females nesting on wooded islands and proposed that this difference was due to adverse incubation conditions at exposed nests. Therefore, we tested whether common eiders gained an advantage when nesting in sheltered habitats by placing artificial shelters over randomly-selected females after the onset of incubation within an eider colony in arctic Canada. We predicted that sheltered females would be heavier on completion of incubation than control hens lacking shelters. Females nesting in artificial shelters experienced a more moderate thermal environment at both cold and warm temperature extremes. Eiders nesting in shelters were heavier than control females during mid incubation, consistent with habitat-specific rates of weight loss reported by Kilpi and Lindström (1997). Natural overhead cover was available at potential nests but few eiders used those sites. We suspect that microclimatic advantages offered by sheltered sites may be offset by costs of increased female vulnerability to predators. Further work is needed to test this hypothesis, and to determine mechanisms responsible for lower weight loss in eiders attending well concealed nests.