Abstract: Predator odours and habitat structure are thought to influence the behaviour of small mammalian prey, which use them as cues to reduce risks of predation. We tested this general hypothesis for house mice, Mus domesticus, by manipulating fox odour density via addition of fox scats and habitat via patchy mowing of vegetation, for populations in 15 x 15-m field enclosures. Using giving-up densities (GUDs), the density of food remaining when an animal quits harvesting a patch, we measured foraging behaviours in response to these treatments. Mice consistently avoided open areas, leaving GUDs two to four times greater in these areas than in densely vegetated patches. However, mouse GUDs did not change in response to the addition of fox scats, even immediately after fresh scats were added. There was no interaction between fox odour and habitat use. We then tested whether habituation to fox odours had occurred, by comparing the individual responses to scats of eight mice born into enclosures with fox scats to those of eight mice born into scat-free enclosures and five wild mice. In smaller enclosures, GUDs of trays with scats did not differ from GUDs of trays without scats for any treatment. We conclude that exposure to high levels of fox odours did not alter the foraging behaviour of mice, but that mice did reduce foraging in areas where habitat was removed, perceiving predation risk to be greater in these areas than controls. We suggest further that studies using the `scat-at-trap' technique, which have shown avoidance of predator odours by mice and other small mammals, may overestimate the general avoidance of predator odours by free-living prey, which must forage with a constant background of predator odours.