The records of 215 patients presenting with 218 penetrating or blunt injuries to the extremities from 1977 through 1983 have been reviewed. All patients presenting with pulsatile hemorrhage, expanding hematoma, or absent distal pulses were explored immediately. Patients with injuries in close proximity to a major artery but without classical signs of arterial injury were explored routinely from 1977 through 1980. Thereafter, similar patients were evaluated initially with contrast arteriography and explored only if arteriographic abnormalities suggested arterial injury. Routine exploration of proximity injuries has been compared with selective exploration based on contrast arteriography. Sixty-one patients (group 1) underwent routine exploration for proximity injuries. Vascular injuries were detected in ten (16%) patients, while 51 (84%) of the explorations were negative. Eighty-four patients (group 2) with proximity injuries were studied arteriographically. Ten patients (11.9%) were explored on the basis of abnormal arteriograms, and eight arterial injuries were confirmed. Two (2.4%) of the 84 patients in this group had negative explorations. The use of contrast arteriography enabled the authors to reduce their negative exploration rate from 84 to 2.4 per cent in the management of proximity injuries.