This paper reports the treatment progress of methadone maintenance clients who were discharged or withdrew from treatment and then were readmitted for a second episode of treatment. Thirty-nine clients in a contingency contract condition remained in treatment long enough (6 months) during both the initial and a second treatment episode, to be exposed to discharge sanctions that were part of the contingency contract. Of these clients 34 failed treatment during the initial treatment episode. Nine (26%) of these initial treatment failures improved their performance in the second episode compared to the first, and only one (20%) of five initial treatment successes who left treatment during their first treatment episode for non-contract reasons showed a poorer performance (failing the second after succeeding in the first episode). Of 17 clients in a condition that applied no contingencies for positive urines, three of 14 (21%) who failed during the initial treatment episode improved their performance, and two of three (67%) who succeeded during the initial treatment episode failed in the second episode. For a subset of clients the efficacy of contingency contracting may not be realized until it is reapplied during a subsequent admission.