Organizations today place great emphasis on the selection of their managerial and supervisory personnel. Consequently, the need to develop strong predictors of leadership for use in selection tests is immense. Theoretically and historically, intelligence, dominance, masculinity-femininity, and self-monitoring have been strong indicators of leadership. Accordingly, biographical questionnaires and personality inventories were administered to 1137 Air Force officers attending Squadron Officer School to determine their levels of intelligence, dominance, femininity, and self-monitoring. These traits and individual difference variables were then used in a discriminant analysis to predict the subject's classification into one of four leadership categories. Analyses showed differences between leaders and nonleaders for intelligence, dominance, and femininity, but not for self-monitoring. However, a discriminant analysis and associated cross validation resulted in no predictive capability. A possible reason for the lack of classification capability was the uniqueness of the military sample. Since military individuals seem to share very common experiences and beliefs, the sample appeared to be quite homogeneous making differences between leaders and nonleaders extremely difficult to detect. Future studies should address this potential problem / Ph. D.