There is a considerable body of empirical research that has identified adolescent peer relationships as a primary factor involved in adolescent cigarette smoking. Despite this large research base, many questions remain unanswered about the mechanisms by which peers affect youths' smoking behavior. Understanding these processes of influence is key to the development of prevention and intervention programs designed to address adolescent smoking as a significant public health concern. In this paper, theoretical frameworks and empirical findings are reviewed critically which inform the current state of knowledge regarding peer influences on teenage smoking. Specifically, social learning theory, primary socialization theory, social identity theory and social network theory are discussed. Empirical findings regarding peer influence and selection, as well as multiple reference points in adolescent friendships, including best friendships, romantic relationships, peer groups and social crowds, are also reviewed. Review of this work reveals the contribution that peers have in adolescents' use of tobacco, in some cases promoting use, and in other cases deterring it. This review also suggests that peer influences on smoking are more subtle than commonly thought and need to be examined more carefully, including consideration of larger social contexts, e.g. the family, neighborhood, and media. Recommendations for future investigations are made, as well as suggestions for specific methodological approaches that offer promise for advancing our knowledge of the contribution of peers on adolescent tobacco use.