The ability to identify persons in need of drug treatment quickly and accurately is an important function for the criminal justice system. A brief screening interview that can identify arrestees most likely to be drug dependent would save time and could improve how personnel and scarce drug treatment resources are used. Using data collected from the Baltimore (Maryland) Substance Abuse Need for Treatment among Arrestees (SANTA) study, the Center for Substance Abuse Research (CESAR) Arrestee Drug Screener (CADS) was developed to screen for cocaine and heroin dependence. The scales were then replicated among arrestee populations in three other SANTA sites. The CADS developed in this study uses 2 sets of 3 questions from the modified Diagnostic Interview Schedule (DIS) for the purpose of identifying cocaine and heroin dependence among arrestees that admitted using the drugs 11 or more times during the previous 18 months. The results indicated that the newly constructed cocaine and heroin scales identified 99 percent of arrestees that were diagnosed dependent using the full diagnostic interview. One of the most desirable characteristics a screener can possess is the ability to identify all subjects that are subsequently found to have the outcome of interest. Sensitivity was the most important criterion in evaluating potential screeners. Results show that the advantages of drug-specific screening scales within criminal populations are clear. It is critical that a key opportunity to identify arrestees in need of treatment is not lost. The successful identification of arrestees with the highest risk for dependence should subsequently permit their diagnosis and referral to appropriate treatment programs. Such referrals may, at least temporarily, prevent the continuation of drug use, hinder the progression of drug addiction, and reduce other public health concerns, such as the transmission of HIV in this high-risk population. 1 figure, 3 tables, 28 references