The medical records and roentgenograms of 2,016 patients who were operated on from 1961 through 1974 using Harrington spinal instrumentation were reviewed to determine the incidence, clinical significance, and management of broken distraction and compression rods. The cases were divided into two study groups. Group A includes 1,128 patients operated on from 1961 through 1968, when no autogenous iliac-bone graft material was used, and Group B includes 888 patients operated on from 1969 through 1974, when autogenous bone was used. The incidence of broken distraction rods was 12.5 per cent (141 patients) in Group A and 2.1 per cent (nineteen patients) in Group B. The age of the patient at operation was not found to be a significant factor when comparing patients with fractured rods and those with intact rods; however, preoperative curve magnitude was found to influence the incidence of rod fractures. Reinstrumentation of distraction rods was required in twenty-three patients from Group A, but no patients in Group B required reinstrumentation. Eleven patients from Group A required removal of the rods. The compression rod fractured in forty patients (3.5 per cent) in Group A and in one patient in Group B; none required reinstrumentation or rod removal. The clinical management of rod fractures must be individualized for each patient. Reinstrumentation and fusion may be indicated in patients with early rod fracture, total loss of correction, or overlapping of the rod, but not in patients experiencing little or no loss of correction and no associated symptoms.