Presented is the final report of the South End Project of the Boston University Head Start Evaluation and Research Center. In a Head Start community a setting was created to study and service preschool children with emotional and social disturbances from low income homes. During the 3-year project, 112 children were seen, families interviewed, referrals made, special programs developed, and follow-ups initiated. Attention was especially paid to the language used to describe disturbed behavior, the use of alternatives as educational-treatment devices to involve parents and community, the roles of professionals and other workers, and the use of behavioral variation as a conceptual vehicle for facilitating understanding of disturbance. Evidence was presented to support the argument that disturbance has to be studied in particular settings (either created or chosen), and that the findings will be a function of that setting. It is emphasized that, because lower income settings are particularly foreign to middle class researchers, they are very susceptible to bias, and that only after setting specific connections have been made and understood can generalization be possible. One is warned to be wary of generalizing a problem when it is simply a manifestation of a particular setting. (Author/KW)