During winter, many temperate-zone animals must survive harsh environmental conditions of low ambient temperatures, long nights, and unpredictable food supplies. Two evolutionary forces are thought to affect an individual's survivorship during winter: risk of starvation and risk of predation (Lima 1986, McNamara and Houston 1990, Grubb and Pravosudov 1994a, Pravosudov and Grubb 1997). Foraging and vigilance for predators appear to be mutually exclusive forms of behavior that an animal must trade off (Lima 1986). To reduce the risk of starvation, an animal must maximize the rate of foraging and energy intake, but to reduce the risk of predation, it must maximize its vigilance. Social foraging has been thought to benefit all members of a group by allowing each member to increase its foraging efficiency while sharing vigilance with the rest of the group (Elgar 1989). Vigilance in social groups of birds may be directed at predators or dominant group members (Waite 1987a, b; Elgar 1989). Recent studies have suggested that vigilance toward dominant individuals could be directed at both conspecifics and heterospecifics in multispecies groups (Popp 1988, Carrascal and Moreno 1992, Sasvari 1992). Ekman (1987) reported that subordinate Willow Tits (Parus montanus) were more vigilant than dominant individuals. Hogstad (1988a), on the other hand, reported the opposite trend, with dominant Willow Tits being more vigilant than subordinates. The latter report, however, was equivocal because in some instances juvenile females, which always have the lowest dominance rank, had the highest vigilance rates (Hogstad 1988a). However, the absolute rate of vigilance might not be the best measure of benefits for group members. Both studies of Willow Tits examined vigilance of group members within groups but did not provide comparative records for solitary individuals. Because parids seldom (if ever) are solitary in nature, only a manipulative study can provide records for solitary birds under controlled circumstances. Although dominance-related vigilance within the same species occasionally has been studied in controlled experiments (Waite 1987a, b), the effect of heterospecific group members on vigilance has been