To further predict the possible activity on memory disorders, the effect of MCI-2016 (bifemelane hydrochloride) was examined using the passive avoidance (PAR) failure technique as an experimental model of amnesia. The amnesia was produced either by post training treatments of electroconvulsive shock (ECS), scopolamine (mice) and cycloheximide or by pre-test injection of scopolamine (rats). In ECS-PAR failure model, the retention test was carried out 3 hr (3 hr experiment) or 24 hr (24 hr experiment) after ECS. MCI-2016 showed a significant improvement when administered just after ECS (3 hr experiment, 30 mg/kg, i.p.) or 0.5 hr before the retention test (24 hr experiment, 10-30 mg/kg, i.p.). Cahopantenate was only active in the 3 hr experiment (500 mg/kg, i.p.), and piracetam was rather active in the 24 hr experiment (60 mg/kg, i.p.). MCI-2016 (30 mg/kg, i.p.) prevented the scopolamine-induced PAR-failure. In this model, physostigmine (0.3 mg/kg, i.p.) exhibited a tendency to improve the failure. In another scopolamine-induced PAR failure model in mice, all of the test drugs showed a significant improvement at different dose levels. The effect of MCI-2016 (25-100 mg/kg, p.o.) was superior to those of piracetam, aniracetam and choline chloride. Higher doses of MCI-2016 were required to improve the cycloheximide-induced PAR failure. Considering the experimental conditions and results, it may be suggested that MCI-2016 ameliorates the amnesia possibly through its influence on memory consolidation and retrieval processes.