A base rate of deviant behavior was obtained for the eight most disruptive children in a third grade adjustment class. In a token reinforcement program, the children received teacher's ratings, which were exchangeable for reinforcers such as candy and trinkets. With the introduction of the token reinforcement program, an abrupt reduction in deviant behavior occurred. Delay of reinforcement was gradually increased to 4 days without increase in deviant behavior. The program was equally successful for all children observed, and anecdotal evidence suggests that the children's appropriate behavior generalized to other school situations. (Author)