The more recent research has used different data sources, but it has not invalidated the findings of the research of the 1960's. Thus, gangs are normally territorially bounded, have members from age 11 to the early 20's, are homogeneous in gender and ethnicity, are made up of clusters of subgroups that are based on age, and have three categories of members ranging from core members to fringe members. In recent years they have included more older members, have become increasingly involved in schools, and have developed in many smaller cities in the United States. Intervention efforts in the 1960's reflected the rehabilitative model and focused on individual change and value transformation. Subsequent evaluations showed these approaches to be ineffective. In recent years the programs have rested on the surveillance/deterrence model and continue on the basis of faith and conviction, because they have not been rigorously and independently evaluated. These programs need to overcome the difficulties of intervening too late to be effective and should avoid viewing natural maturation and attrition processes as proof of success. Future programs should reflect awareness that some approaches actually make the gang problem worse and should consist of area-based rather than gang-based programs. Tables, notes, and 65 references.