Two baselines involving a repeated acquisition task were used to assess the effects of bretazenil, imidazenil, and triazolam. The first baseline was a multiple schedule of repeated acquisition and performance of conditional discriminations. In the first component, the subject acquired a four-response chain by responding sequentially on three keys in the presence of different combinations of colors and geometric forms displayed on a center key. Acquisition of the discrimination was defined by a decrease in errors as the session progressed. In the performance component, the four-response chain was the same each session. Incorrect responses in either component produced a 5s time out during which responding had no programmed consequence. The second procedure, which has been used to evaluate the effects of drugs on memory, involved the acquisition of a discrimination, followed by a 1h delay and a retest of the same discrimination to assess retention. Triazolam (0.32 and 0.56mg/kg) administered alone, produced dose-related decreases in response rate in each component. In addition, triazolam also produced a dose-related increase in percentage errors in acquisition with no effect in performance. Triazolam (0.32mg/kg) eliminated retention (0 percent savings) in the memory task. Bretazenil (0.1-5.6mg/kg) or imidazenil (0.1-1.8mg/kg) administered alone had little or no effect on either rate of responding or accuracy in either component. Furthermore, bretazenil but not imidazenil disrupted retention at the higher doses tested. The combination of imidazenil or bretazenil with triazolam produced dose-related attenuation of the disruptive effects of triazolam on both behavioral baselines. These data suggest that the disruptive effects of benzodiazepines on learning and memory may be a function of the intrinsic efficacy of these compounds at different GABA(A) receptor subtypes.