Wild impala display a highly reciprocal allogrooming system that, by virtue of its frequency and high degree of reciprocity, is unique among ungulates. A herd of 35 free-ranging captive impala provided opportunity to examine the degree of reciprocity of allogrooming exchanges and the influence of relatedness, dominance, age and association on partner preferences and distribution of grooming between allogrooming partners. As in wild impala, the exchange of allogrooming bouts in the captive impala was highly reciprocal regardless of partners. Kinship and dominance had no influence on partner preference or distribution of grooming between partners. Although mothers showed a significant preference to allogroom with their unweaned offspring, this preference practically disappeared with older offspring. Age-mates (no greater than 6 mo apart) tended to associate with one another and spatial proximity was positively correlated with grooming partner preference. It was not clear whether impala actively sought out age-mates for grooming, or randomly chose grooming partners from nearby age-mates. The failure to find a role for kinship and dominance is counter to what has generally been found in most Old World terrestrial primate studies. The absence of pronounced social influences, coupled with the known effectiveness of grooming in removing ectoparasites, suggest that a utilitarian role, especially removal of ticks, is an important function of the impala reciprocal allogrooming system.