In an ongoing study of adolescent maternal-infant interaction, we investigated the accuracy with which judges could assign the age of two groups of females, based on videotaped facial cues alone. The two groups were eight "adolescents," whose ages were less than or equal to 17 years, and 20 "adults," whose ages were greater than 20 years. Twenty-seven judges, whose average age was 28.7 years with a range from 27 through 44 years of age, were asked to judge the ages of the subjects in the two groups. Overall, the judges were successful in identifying adolescents 41.8% of the time. There was no relationship between the ages of the judges and the accuracy of their judgments. On the basis of facial cues alone, our results clearly suggest that the process of assigning an individual to an adolescent or adult group is essentially random, approaching a 50/50 breakdown. This study supports the concept that observed differences between adolescents and adults interacting with their infants is due to actual differences in behavior and not the cultural bias of the judge.