Objective To investigate the effects of propofol and inhalation anesthetics on the implicit memory under general anesthesia and determine the critical value of auditory evoked potential index (AEI) at which implicit memory no longer exists. Methods Sixty ASA I-II patients of both sexes scheduled for elective abdominal surgery or surgery on lower limb were enrolled in the study. They averaged (41.25?9.25) years and (63.8?9.57) kg. Patients with psychoneural or memory disorders were excluded. The patients were randomly divided into five groups of 12 patients in each group: isoflurane, sevoflurane, desflurane, propofol and control group. Each inhalation anesthetic group was further divided into two subgroups based on their end-tidal concentration: 0.8 or 1 .0 MAC. Propofol group was divided into two groups based on the infusion rate: 8 mg.kg. h-1 or 10mg.kg.h-1 . In control group operation was performed under epidural anesthesia. The patients were unpremedicated. Anesthesia was induced with propofot 1 .5-2.0 mg.kg and vecuronium 0. 　-0.2mg.kg-1, and maintained with inhalation anesthetic or propofol infusion supplemented with intermittent boluses of fentanyl and vecuronium. AEI was continuously monitored and recorded before anesthesia (T0), when the patient was sleepy (T1 ) and did not respond to verbal command (T2 ), after intubation (T3), after skin incision (T4), when peritoneum was being closed (or 30 min before the surgery) ( T5 ) and during emergence from anesthesia ( T6 ) . Eight hours after surgery muddy identified hearing rate was determined. Hemodynamic parameters were continuously monitored during operation. Results There was significant difference in AEI between the two subgroups before and after induction (P