Discriminated escape behavior of mice in a T-maze was evaluated as a model for investigation of drug effects upon memory. All studies involved discriminated escape training trials on which footshock offset occurred following a correct goal-arm choice. Memory was assessed in retention tests that involved either reversal or relearning of the previously trained goal choice. Analyses of discriminated escape acquisition and training-retention interval were performed to identify parameters yielding either optimal or poor retention. Parameters yielding poor retention in untreated mice were used in studies involving drugs expected to improve retention. Pretraining treatment of mice with either physostigmine (0.08 mglkg) or Hydergine'" (0.64 mglkg) yielded improved retention scores when training was minimal (a criterion of two consecutive correct choices) and the retention interval was long (168 hr). The effects of scopolamine (0.64 mglkg) upon memory were assessed in the context of a short retention interval (24 hr) and more complete training (three consecutive correct choices), parameters yielding optimal retention in saline-treated mice. Scopolamine yielded poorer 24-hr retention performance when present during training, retention, or at both times. Overall, the results indicate that the discriminated escape procedures described are potentially useful for detection and analysis of pharmacologic modifications of memory processes.