This research investigated the effects of employing fixed, variable, and extended token exchange periods for back-ups on the completion and accuracy of daily assignments for a total fifth and sixth-grade class. The results indicated that, in general, a higher percentage of assignments was completed when the number of days between point exchanges was variable, and that performance was maintained when the number of school days between point exchanges was gradually increased. The overall accuracy on assignments was high, but variable and did not appear to be under the systematic control of the experimental procedures. Data for two selected pupils indicated that there were individual differences as to the effects of the various token (point) exchange procedures. Thus, the variable sequencing of the number of days should be an important consideration in any effort to maintain the positive effects of token-reinforcement systems within classroom settings. (Author/ED)