To compare the accuracy of interpretation of digitized radiographs with that of plain films, the authors prospectively evaluated the first 685 plain film cases (530 adult and 155 pediatric cases, each of which consisted of one or more images) transmitted from an outpatient center 18 miles (approximately 29 km) to a hospital radiology department by means of a high-speed teleradiology system. Plain films were digitized and transmitted via a T-1 (1.544 Mbit/sec) data link for display on high-resolution (2,560 x 2,048-pixel) workstations. Radiologists at the hospital used a copy of the radiology requisition that had been faxed from the remote center. Interpretation of the digital images was followed by review of the original plain radiographs within 1 working day. Discrepant interpretations occurred in 18 cases (2.6%) (sensitivity, 96%; specificity, 99%); an arbitration panel decided that they were associated with observer performance more than with the fidelity of the digital display. It is concluded that primary diagnosis without review of the original plain radiographs is feasible with state-of-the-art teleradiology systems.