THE RESEARCH WAS DESIGNED TO TEST THE UTILITY OF USING ETHOLOGICAL METHODS IN STUDYING SEXROLE BEHAVIOR AND TO TEST THE HYPOTHESIS THAT VARIATIONS IN FEMININITY AND MASCULINITY AMONG FEMALE PRISONERS ARE ASSOCIATED WITH VARIABLES SUCH AS LEADERSHIP IN ALL FEMALE POPULATIONS, TYPE OF OFFENSE, AND HOMOSEXUAL BEHAVIOR. FOUR CHARACTERISTICALLY FEMALE BEHAVIOR PATTERNS WERE DELINEATED: SUBMISSIVE-DEPENDENT BEHAVIORS, NURTURANCE, COQUETTISH-PASSIVE BEHAVIORS, AND ORIENTATIONS TOWARD OTHERS. BEHAVIORS DEFINED AS MASCULINE INCLUDED ARGUING AGRESSIVELY, VOCAL THREATENING, HITTING, TAKING OBJECTS AWAY FROM OTHERS, CURSING, INDICATING MECHANICAL INTEREST, TALKING IN A LOW VOICE, AND 'ROUGHHOUSING'IN PLAYFUL PUSHING, SHOVING AND HITTING. THERE WAS ALSO A NEUTRAL CATEGORY FOR BEHAVIOR NOT ASCRIBED TO EITHER GENDER. THE INTERACTION AMONG PRISON INMATE GROUPS WAS CODED SYSTEMATICALLY IN THREE CLASSROOM SETTINGS. INDIVIDUALS WERE RANKED ON THE BASIS OF THE DOMINANT PERCENTAGE OF MASCULINE OR FEMININE BEHAVIORS. THE PERCENTAGES OF MASCULINE OR FEMININE BEHAVIORS VARIED WIDELY AMONG THE WOMEN. FOR SOME, 100 PERCENT OF THEIR ACTIVITY WAS FEMININE, WHILE FOR OTHERS, ALL THEIR ACTIVITY WAS CHARACTERISTICALLY MASCULINE. INMATES ENGAGING IN ROUGHHOUSE BEHAVIOR OR HOMOSEXUAL ACTIVITY TENDED TO BE THOSE CONVICTED OF VIOLENT CRIMES. INMATES INVOLVED IN HOMOSEXUALITY TENDED TO SCORE HIGH ON MASCULINITY SCORES, AND THOSE WITH HIGH MASCULINITY SCORES TENDED TO INITIATE THE MOST INTERACTION IN THE CLASSROOM. A SIGNIFICANT RELATIONSHIP WAS FOUND BETWEEN LEADERSHIP IN COMPLAINING GROUPS AND VIOLENCE OF CRIME COMMITTED. INMATES WITH HIGHER EDUCATIONAL LEVELS TENDED TO BE RANKED HIGH ON LEADERSHIP ABILITIES, AND HOMOSEXUALLY ACTIVE FEMALES SCORED HIGHER ON GOAL ACHIEVEMENT AND LEADERSHIP IN COMPLAINT GROUPS. IT IS CONCLUDED THAT THESE FINDINGS SHOULD BE QUALIFIED BY FURTHER STUDIES IN VARIOUS TYPES OF INTERPERSONAL SETTINGS. APPENDIXES CONTAIN STUDY DATA AND REFERENCES. (DAG)