This article describes a research program to develop an operant treatment for cocaine addiction in low-income, treatment-resistant methadone patients. The treatment's central feature is an abstinence reinforcement contingency in which patients earn monetary reinforcement for providing cocaine-free urine samples. Success and failure of this contingency appear to be an orderly function of familiar parameters of operant conditioning. Increasing reinforcement magnitude and duration can increase effectiveness, and sustaining the contingency can prevent relapse. Initial development of a potentially practical application of this technology suggests that it may be possible to integrate abstinence reinforcement into employment settings using salary for work to reinforce drug abstinence. This research illustrates the potential utility and current limitations of an operant approach to the treatment of drug addiction. Similar research programs are needed to explore the limits of the operant approach and to develop practical applications that can be used widely in society for the treatment of drug addiction. (Contains 4 figures.)