This empirical study examined the social provisions that adults felt they received from friendship as well as what they desired from friendship and explored discrepancies between what was received and desired. The social provisions were distinguished as: attachment, social integration, reliable alliance, guidance, reassurance of worth, and opportunity for nurturance. The study sought to provide a comprehensive picture of adult friendship by examining multiple predictor variables. The independent variables were subdivided into three categories: selfperception (personality type, gender orientation, selfesteem, and psychological well-being), family intimacy (intimacy with parents, siblings, and spouse) and personal attributes (age, sex, and marital status). The investigation was conducted in Blacksburg, Virginia. Respondents were drawn from a random sampling of faculty and staff members of the Virginia Polytechnic Institute and State University. The Total Design Method for mail surveys was applied to meet the objectives of the research (Dillman, 1978). A total of 449 questionnaires were completed and returned, representing a response rate of 65%. Regression analysis was used to address the research questions. The results of the study indicate that individual characteristics do indeed affect an individual's perceptions of the meaning and adequacy of friend relationships. The findings underscored the importance of using multiple explanatory variables in a study of this nature. However, the self-perception variables proved to be more predictive than the family intimacy or personal attributes variables. The extraverted personality type was the most significant independent variable having the greatest relative influence in ten of the eighteen regressions. The feeling personality type was also very significant. Because friendship is a voluntary association, expressions of personality characteristics may be more evident than in other relationships. The fact that the personality variables were such strong predictors in this study suggests that further research is needed in this area. Also, the results of the study indicate the need to include both gender orientation and sex as variables in future research because both were significant in this study. / Ph. D.