Thirty patients undergoing craniotomies were given infusions of fentanyl 1 microgram X kg-1 X hr-1, sufentanil 0.1 microgram X kg-1 X hr-1, or normal saline in a double-blind study of cerebral relaxation. Significantly better relaxation scores were achieved in patients given a narcotic infusion, but there was no difference between the scores with the two narcotics. Infusions of narcotics at these low rates did not delay recovery or alter the requirement for other anesthetic agents. Narcotic infusion rates that do not delay recovery or alter depth of anesthesia significantly improve cerebral relaxation.