A token reinforcement system was introduced into a second grade classroom, and seven (out of 21) class members and the teacher were observed for 8 months. Each of the eight phases of the study lasted from 2 to 5 weeks. The first four phases established a base period and determined the separate effects of classroom rules, half hour academic lessons, and teacher praise. In the Token I phase each child in the afternoon classes received points for desirable behavior, which were exchanged for back-up reinforcers including comic books and dolls. The percentage of disruptive behavior significantly decreased. Tokens and back-up reinforcers were then withdrawn, and disruptive behavior increased, but not significantly. Tokens and back-up reinforcers were again reinstated and withdrawn and then replaced by the use of more usual classroom reinforcers by the teacher. At this time disruptive behavior was 37% compared to 53% during the base period. Results indicated student gains on the California Achievement Test, higher class attendance, and the teacher's increased use of praise and decreased use of criticism and threats. (DR)