A consecutive series of 50 diagnostic cerebral biopsies carried out over a 12-year period (1960 to 1972) at The Neurological Institute of New York is reported. The goal of the biopsy was to acquire a definitive diagnosis to aid the patient, counsel the family, and advance scientific knowledge. General indications for biopsy were either chronic bilateral cerebral findings associated with progressive dementia or profound retardation or acute unilateral cerebral findings, usually associated with signs of meningoencephalitis. Four patients were not suitable for analysis, leaving 46 patients to be surveyed. Diagnostic neuropathological features were judged to be present in 17 specimens (37%). Abnormal but not diagnostic pathological changes were found in 22 specimens (48%). Thus, 39 specimens (85%) displayed some abnormality. No apparent neuropathological changes were present in 7 specimens (15%). These findings correspond favorably with the results of others and suggest that a certain “irreducible minimum” of normal biopsies occurs with the techniques utilized. Fourteen (30%) of the patients were known to be alive at the conclusion of the study. Twenty-two patients (48%) had died, but necropsy findings were available for only 10 of these. Except in 2 cases in which biopsies had been normal, no discrepancy was noted between the basic neuropathological changes found at necropsy and those of the biopsy specimen. One death could be directly attributed to the procedure, and total morbidity and mortality was approximately 13%. The authors recommend that any biopsy program should be organized at an institutional level to take into account the patient's special illness and the capabilities of that institution. In addition, suitable material must be forwarded to appropriate experts who have the knowledge and facility to make full use of biopsy material.