The series consisted of 200 patients on whom lumbar myelography was performed for sciatica. After myelography, a disc operation was carried out on 95 patients. The episode of sciatica was the first for 90 patients. Objective neurologic signs were present in 185 patients, while 15 had only subjective symptoms. Definite or possible disc herniation was revealed by myelography in 66%. Most positive findings were located at the L4-L5 interspace. The clinical diagnostic accuracy rate was assessed from the patients' histories. As confirmed by operation, the accuracy of the clinical diagnostics was 52%, and the accuracy rate of myelography was 90%. The rate of false positive findings in myelography was 4%, that of false negative findings 6%. When the clinical or myelographic diagnosis was definitive, a disc herniation or protrusion was always found at operation. Almost one fourth of the clinically diagnosed definite herniations were not treated surgically because myelography proved negative. One fifth of those patients in whom myelography revealed an unequivocal disc herniation were not operated on because these patients had clinically improved before being admitted to myelography. The results of this study justify the following conclusions: the clinical level diagnostics of a disc herniation is rather unreliable, and myelography is therefore always indicated before operation; myelography should only be performed in those cases in which there is a clear clinical indication for surgery; myelography ought to be performed within 1 week; an unequivocal positive finding in myelography predicts a good operative result.