The programs observed were the Intensive Treatment Program in Norwalk, Calif.; the Tri-Agency Program in Illinois; the Medfield Regional Adolescent Program in Massachusetts; the Protective Component Unit in Minnesota; and the Juvenile Forensic Unit for boys in Pennsylvania. Although the programs' core designs are similar, their differences range from the general (such as regional location, physical plant) to the highly complex (such as admission procedures, discharge criteria). Only Massachusetts has a coed admissions policy. All units have waiting lists and are located on the grounds of State mental hospitals, except the California program. Most program clients are serious repeat offenders who have experienced a long history of unsuccessful placements. All programs sought to adhere to a selection process based primarily on evidence of severe psychopathology. Each program had a formal treatment plan for each resident. The California program has seven treatment teams which formulate and evaluate the plan at monthly intervals, and Pennsylvania's two treatment teams must submit individual treatment plans. Clinical impressions and early observation form the basis of a preliminary intervention in Massachusetts, followed by a more indepth assessment of personality developed against an elaborate set of diagnostic protocols. The Minnesota unit's program was based on the concepts of milieu treatment as was the program in Pennsylvania, in which group treatment and family therapy formed the two therapeutic interventions. The units' staff problems are noted, and information on discharge planning is included. Finally, a discussion comments on the successes and failures of the programs in general and on their impact on clients. A few footnotes and about 35 references are supplied.