Mehrabian (1971) has summarized evidence that nonverbal "immediacy" behaviors such as interpersonal distance, gaze, and shoulder orientation are important in communicating interpersonal attraction. Several authors (Duke & Nowicki, 1972; Efran, 1968; Efran & Broughton, 1966) have conceptualized nonverbal immediacy differences between friends and strangers within the context of Social Learning Theory (SLT). Duke and Nowicki, for example, proposed that interpersonal distancing is mediated by specific expectancies in interactions among friends, but by generalized expectancies, such as locus of control (Rotter, 1972), when interactions involve strangers. Some support for this hypothesis is provided by a study showing that locus of control orientation was a better predictor of how far subjects would place themselves (representationally) from a hypothetical stranger than from a hypothetical friend. Using a live-interaction situation, the present study examined distance, gaze, and body orientation as a function of degree of acquaintance (friend, stranger) and internal versus external locus of control. From the immediacy formulation (Mehrabian, 1971), it was hypothesized that friends would be more nonverbally immediate than strangers on all measured nonverbal dimensions. Following SLT, it was also predicted (a) that internals should be more immediate than externals among friends, but not strangers and (b) that immediacy measures should show more change over time in interactions among strangers than friends.Female "primary" subjects (n = 48), selected on the basis of extreme scores on Rotter's (1972) I-E Scale (internal-external locus of control), interacted in standing positions for 12 minutes with same-sex "secondary" subjects—either friends (n = 24) or strangers (n. = 24). Immediacy measures including frequency and duration of gaze, approach distance, and directness of body orientation were monitored by three trained observers during each one-half of the interaction; thus adding a third, within-subjects dimension to the overall design. Duration of speech, which is closely related to gaze behavior (Exline, Gray, & Schuette, 1965; Kendon, 1967; Kendon & Cook, 1969), was also recorded.The main hypotheses were not supported: friends were not more immediate than strangers; locus of control orientation was not more important in interactions involving strangers; and immediacy measures did not show more change over time in interactions among strangers. Several aspects of the data seem to contradict the hypothesis that the so-called immediacy behaviors are important in communicating interpersonal attitudes or attraction (Mehrabian, 1971). First, strangers tended to gaze longer, gaze more often, and assume more directly confronting shoulder orientations than friends. Second, immediacy measures were essentially unrelated to subjects' ratings of liking for their partners, both within friend and stranger groups and across the groups combined. Finally, the immediacy measures themselves were not consistently interrelated as might be expected if the various nonverbal behaviors were elements of a common construct. These indications that gaze, distance, and orientation do not always communicate attraction or liking should be considered in future research involving these nonverbal behaviors, as well as in research on "intimacy-equilibriums" (Argyle & Dean, 1965) or "compensation" (Patterson, 1973) in dyadic interactions. The failure to confirm predictions from SLT may be related to conceptual ambiguities (e.g., the role of reinforcement value) which limit the usefulness of SLT concepts in understanding proxemic and other nonverbal behaviors.