In response, a system for the surveillance of fatal and nonfatal gunshot injuries was developed and evaluated, and the resulting gunshot wound surveillance system was pilot tested. Reports from emergency departments were compared with emergency department logs, and reports from medical examiners and coroners were compared with death certificates. The pilot system was found to be simple and acceptable to data providers. The predictive value positive was 99 percent for reports from emergency departments and 100 for reports from medical examiners and coroners. The sensitivity of emergency department reporting was 76 percent, while the sensitivity of medical examiner and coroner reporting was 81 percent. The pilot test of the system demonstrated that States interested in developing a gunshot wound surveillance system should: (1) verify the existence of a statewide gunshot wound reporting requirement; (2) consider establishing such a requirement if one does not exist; (3) include emergency departments as a data source; (4) establish the capacity to link records if multiple data sources are used; (5) limit data reporting requirements so they fit on a one-page form; (6) periodically review data and reporting practices of data providers and be prepared to make changes if things are not working as planned; and (7) disseminate surveillance data and system evaluation findings on a regular basis. 10 references, 6 tables, and 2 figures