After given 10 days of saline injections, rats in Group I were run 15 trials per day on a shuttle avoidance task to a 13/15 criterion plus 10 additional training days all under 20 mg/kg of chlordiazepoxide (D). From the following day, they were injected saline (N), being trained to the same criterion , after which they were trained only a day under D. Group II rats were similarly treated except they were not given 10 additional training days and Group III rats were treated like Group II except they were given D instead of N prior to training.All three groups behaved identically during the original training days under D and showed response decrements with the drug state change from D to N but the decrements were least marked in Group I. When the drug state was shifted back to the original drug state, Group I performed better than on the preceding criterion day under N; this finding was not observed in the other two groups. The obtained overtraining effects can not be ascribed simply to overmedication since Group III which was overmedicated but not overtrained failed to show the same effects. In explaining the results, the drug state was assumed as contextual cues which facilitate or inhibit the retrieval of the learned response.Our previous study (IN\ ahara, 1971) showed that chlordiazepoxide (CDP) at 20 mg/kg produced a marked dissociative effect upon shuttle avoidance learning such that avoidance responses of rats were impaired with drug-state changes especially from drug to placebo.This statedependency tended to decrease with overtraining prior to drug-state shifts but this tendency failed to achieve the statistical significance.The purpose of the present study is to investigate whether the same, but statistically significant effect can be obtained for a group of rats which are considerably overtrained before drug-state shifting in comparison with a control group of animals which are not overtrained.As overtrained rats are necessarily given longer days of medication than the second group of control rats, a third group is added which is administered drug injections prior to training for the same number of days as the first overtrained group is given additional days of training. METHOD Subjects. A total of 27 male rats of the Wistar-Imamichi strain were used as Ss; they were about 9 weeks old and weighed 221 g in average and 44 g in SD at the start of experimentation.At their home cages, food and water were always available.Apparatus. The apparatus used was an ordinary 2-compartment Miller-Morwer shuttle box, whose inside was painted grey. Both the grid floor and the hurdle in the center were electrically activated by passing a 600 V, 50 Hz currency if rats failed to shuttle to the next safe compartment within 5 see after the onset of a 80-phon buzzer sound.Procedure. Rats were classified into one of 3 groups, being equated in body weight, as shown in Table I. Groups I and II were given intraperitoneally (i. p.) 2 ml/kg of phyWe wish to give special thanks to Mr. Ken Sakurai and Mr. Hideaki Takagi for their ai...