THE COMBINED EFFECT OF AFFILIATION AND GROUP RESPONSIBILITY ON ETHICAL RISK TAKING IS EXAMINED. SUBJECTS WERE 150 MALE COLLEGE STUDENTS RANDOMLY ASSIGNED TO THREE LEVELS OF AFFILIATION. THE TASK CONSISTED OF TRACING A LINE BETWEEN TWO CONCENTRIC CIRCLES WITHOUT TOUCHING EITHER CIRCLE. SUBJECTS REPORTED THEIR OWN "SUCCESSES" ON THE TASK, WHICH WAS IMPOSSIBLE. GROUP RESPONSIBILITY WAS VARIED BY HAVING TWO PAYOFF TYPES. FOR LOW RESPONSIBILITY, THE SUBJECT WAS PAID ON THE BASIS OF HIS PERFORMANCE WHILE FOR HIGH RESPONSIBILITY, THE SUBJECT WAS PAID ON THE BASIS OF HIS GROUP'S PERFORMANCE. AFFILIATION WAS VARIED BY HAVING THE TASK PERFORMED IN THE PRESENCE OF OTHER GROUP MEMBERS OR IN ISOLATION. ETHICAL RISK TAKING WAS FOUND TO VARY DIRECTLY WITH AFFILIATION WHEN GROUP RESPONSIBILITY WAS PRESENT, AND INVERSELY WHEN IT WAS NOT PRESENT. GROUP RESPONSIBILITY ALONE DID NOT PRODUCE A SIGNIFICANT EFFECT. SHARING RESPONSIBILITY WITH KNOWN GROUP MEMBERS WHO ARE EXPECTED TO COLLABORATE YIELDS HIGH ETHICAL RISK TAKING. A ONE-YEAR FOLLOW-UP STUDY OF 52 SUBJECTS IS DISCUSSED. THIS PAPER WAS DELIVERED AT THE AMERICAN PSYCHOLOGICAL ASSOCIATION CONVENTION, WASHINGTON, D.C., SEPTEMBER, 1967. (SK)