Sixty patients with suspected lumbar herniated disk and/or canal stenosis were studied prospectively with surface coil MRI, CT, and/or myelography, and the results were compared with the surgically confirmed abnormality. Forty-eight patients had lumbar surgery at 62 levels. There were no negative explorations. Thirty-nine patients had a myelogram and CT. Thirty of the CTs were performed following the injection of metrizamide for myelography. Nine patients had a CT without intrathecal contrast material 1 to several days before the myelogram. Six patients had myelography only, and three patients had CT only. All studies were evaluated for the location and type of disease in a forced choice fashion. Independent of the surgically correlated levels, there was 86.8% agreement between the MR and CT studies in all patients at 151 levels and 87.2% agreement between MR and myelography at 218 levels. At the operative levels, there was 82.6% agreement between MR and surgical findings for both type and location of disease; 83% agreement between CT and surgical findings; and 71.8% agreement between myelography and surgical findings. There was 92.5% agreement when MR and CT were used jointly, and 89.4% agreement when CT and myelography were used jointly. The results of this study indicate that a technically adequate MR examination was equivalent to CT and myelography in the diagnosis of lumbar canal stenosis and herniated disk disease. CT and MR can be complementary studies, and surface coil MR can be viewed as an alternative to myelography.