The acquisition of imitative responses without reinforcement was investigated with infants by eliminating contingent reinforcement through the use of videotaped models. Twenty-nine male and female infants were randomly assigned to one of two groups, a Rhythmic Vocalization Group or a Conversation Control Group. Infants in the first group were shown a videotape of an adult female model repeating a novel phoneme in a sequence of sounds and pauses. Infants in the second group were shown a 10-minute videotape of a conversation in a daytime television drama. Trained observers behind a one-way vision mirror scored the frequency and duration of six types of behavior, including vocalization of the novel phoneme. Two types of data analyses were performed: a 2 (Treatment Group) X 2 (Sex of subject) X 3 (Repeated Sessions) analysis of variance, and an event sequential analysis to determine whether observed vocalizations at any lag significantly exceeded expectation. The general findings were: (1) no infant in the study imitated the novel phoneme or repeated the precise phoneme pattern; (2) data from the sequential analysis indicated evidence for imitation of the rhythmic component of the phoneme pattern; and (3) the type of stimulus presented had differential effects on males and females. Although there were no differences at pretest, following treatment males increased their vocalizations in response to the conversation videotape while females decreased their vocalizations. (Author/SS)