Post-partum care of offspring by male parents (paternal care), although rare among mammals generally, occurs commonly within the Carnivora, Primates, and Rodentia (Kleiman, 1977). Whereas lactation constrains female mammals to care for young, males are not similarly constrained (Trivers, 1972). Therefore, occurrence of paternal care in certain mammalian taxa suggests either strong benefits to males (male-choice hypothesis-Trivers, 1972; Gubernick et al., 1993) or benefits to females of sufficient magnitude to promote sexual selection for this trait in males (femalechoice hypothesis-Orians, 1969; Wittenberger and Tilson, 1980). In the former case, paternal care primarily should enhance current reproductive success (i.e., of both sexes), whereas in the latter case, paternal care primarily should affect future reproductive success of the female (i.e., through increased survival or future fecundity). Among mammalian families, paternal care is most pronounced in the Canidae, in which males typically share in all pup-rearing behaviors except nursing (e.g., Kleiman, 1977). The relative benefits of these contributions to current and future reproductive success of females (or of males) have not been empirically investigated in this family. Benefits of paternal care to current and future reproductive success of mates can be investigated by observing consequences to the female of male removal (Mock and Fujioka, .1990). In the absence of paternal care, one of 3 possible strategies for the female canid is predicted from life-history theory (Williams, 1966; Stearns, 1992). First, the female could abandon or kill her litter, thereby sacrificing all current reproductive potential to maximize future reproductive potential (Tait, 1980). Second, she could provide the same amount of care as she would have in the presence of paternal care, presumably sacrificing some of her current reproductive potential, but with no change in her future reproductive potential. Third, she could provide a greater than normal amount of care to compensate for the absence of paternal care, thus sacrificing no current reproductive potential, but reducing her future reproductive potential. The first strategy is most consistent with the male-choice hypothesis and the third strategy is most consistent with the female-choice hypothesis. We used radiotelemetry and observations from a natural experiment of removals of male coyotes (Canis latrans) to investigate these strategies. Coyotes in our study area were routinely killed (regardless of whether they were radiocollared) by control personnel to reduce predation on domestic sheep (Sacks et al., 1999); this resulted in removal during pup-rearing of several paternal males and eventually some of their mates and pups, which were provided to us for post-mortem examination. We compared females with mates present to females