To characterize the role of CB1 receptors in mediating the acquisition of new behavior or learning, delta 9-THC (delta 9-tetrahydrocannabinol), WIN 55,212-2 (R-(+)-(2,3-dihydro-5-methyl-3-[(4-morpho-linyl)methyl]pyrol - (1,2,3-de]-1,4-benzoxazin-6-yl)(1-naphthalenyl)methanone monomethanesulfonate), SR141716A (N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlor- phenyl)-4-methyl-1H-pyrazole-3-carboxamide hydrochloride) and cannabidiol were administered to monkeys responding under a multiple schedule of repeated acquisition and performance of conditional discriminations. SR141716A, a putative antagonist at CB1 receptors, was also administered in combination with delta 9-THC. In one component of the multiple schedule, subjects acquired a different complex discrimination each session (acquisition component), whereas in the other component the discrimination remained the same each session (performance component). Correct responding in each component was maintained by food presentation under a variable-ratio (VR) schedule, whereas incorrect responding (errors) produced a time-out. Administered prior to the start of the session, delta 9-THC and WIN 55,212-2 dose-dependently decreased overall response rate in both the acquisition and performance components. Both drugs also selectively increased the percentage errors in the acquisition component, but only at higher doses. SR141716A and cannabidiol also dose-dependently decreased overall response rate in both schedule components, but neither drug increased the percentage of errors. Decreases in response rate were also observed 24 hours after administration of SR141716A at doses greater than 1 mg/kg. When lower doses of SR141716A (0.1-1 mg/kg) were administered in combination with delta 9-THC, there was a dose-dependent antagonism of the rate-decreasing and error-increasing effects of delta 9-THC (i.e. the dose-effect curves for delta 9-THC-induced disruptions in responding were shifted rightward). In summary, CB1-receptor agonists such as delta 9-THC and WIN 55,212-2 were more disruptive to the rate and accuracy of learning in old-world monkeys than the CB1-receptor antagonist SR141716A or cannabidiol.