1. An analysis is presented of 866 patients who, during the past twelve and a half years, had myelography as part of the diagnostic study by the orthopaedic surgeons at Barnes Hospital for a lumbar intervertebral disc lesion; surgical exploration of the lumbar area was then performed. 2. At operation, 625 positive and thirty-seven negative myelograms were confirmed, giving an accuracy of 76.4 per cent. 3. Major discrepancies were noted in 150 (17.3 per cent) cases. 4. Relatively unimportant or minor discrepancies were present in fifty-four cases (6.2 per cent). 5. Six extradural and three intradural neoplasms masquerading as ruptured intervertebral disc lesions occurred in this series of 866 patients, representing an incidence of 1 per cent. Even with myelograms only two intradural tumors out of the total of nine neoplasms were diagnosed definitely before operation. 6. In cases in which operation for a lumbar disc lesion is clearly indicated on the basis of the clinical picture and lack of response to conservative treatment, a negative myelogram should not prevent the surgeon from exploring the lumbar spinal canal. 7. Myelography can be a valuable aid in diagnosis and may be very helpful in borderline cases. We now believe that myelography should routinely precede surgery for lumbar disc lesions. 8. In our opinion a lumbar myelogram should not be done when a ruptured lumbar intervertebral disc is suspected unless surgical treatment is being considered.