The study used 10 radiographs in standard projections. Some were actual forensic cases and some were obtained randomly from the radiographic files at the Medical College of Virginia. The projections included frontal skull, lateral skull, frontal cervical spine, female pelvis, lateral foot, frontal hand, frontal lower leg, and frontal chests. The comparison films closely resembled the test cases both in size and overall shape to ensure that the identification would be challenging to all participants. The three groups of interpreters consisted of high school seniors participating in the medical division of the Virginia Governor's School for the Gifted, fourth-year medical students in the general radiology elective, and diagnostic radiology residents of all levels of experience. Results revealed that although all three groups identified a majority of cases correctly, none were completely accurate. The diagnostic radiology residents, the group with training in radiographic interpretation, performed significantly better than the other groups. Findings supported the need for trained interpreters in cases of forensic identification. Tables and 14 references (Author abstract modified)