Sites selected for study were Baltimore, Md.; Washington, D.C.; Dade County, Fla.; Jefferson County, Ky.; Pima County, Ariz.; Santa Cruz County, Calif.; and Santa Clara County, Calif. The delivery system for pretrial release decisions was studied in each jurisdiction. The analysis identified the major steps in the pretrial release process and the most important organizations and individuals involved in that process. Further, a sample of defendants was studied from point of arrest to final case disposition and sentencing. The total sample for all sites was 3,488 defendants. The eight sites used a variety of pretrial release practices; moreover, the pretrial release programs used many different methods of identifying and processing arrested defendants. The trend toward releasing more defendants pending trial, as documented in a study of the 1962-71 period, continued through this study period (1977). The releasing of more defendants on nonfinancial conditions also continued. Program recommendations were strongly related to release outcomes and release decisions. Most released defendants (84 percent) were not arrested during the pretrial period, and most (87 percent) appeared for all court dates. No reliable predictors of failure-to-appear or pretrial arrest could be identified. Among the recommendations are that (1) jurisdictions should seek more ways to release more defendants pending trial; (2) programs should revise their release recommendation policies so that specific recommendations are made for all interviewed defendants; (3) courts should implement systematic followup procedures to identify fugitives who have not returned to court after a certain period; and (4) efforts to promote speedier trials should be continued. Tabular data are provided, and appended are summaries of delivery system assessments and defendant outcomes analyses and a description of the research methodology. A bibliography has 84 listings. (Author summary modified)