The purpose of the present study was to investigate whether the cannabinoid and cholinergic systems impair working memory through a common mechanism. This hypothesis was tested by examining whether the cannabinoid antagonist SR141716A would ameliorate radial-arm performance deficits caused by either the naturally occurring cannabinoid, delta 9-THC, or scopolamine, a muscarinic antagonist. In addition, we evaluated whether the cholinesterase inhibitor, physostigmine, would prevent delta 9-THC-induced impairment of spatial memory. Finally, because the locomotor suppressive effects of cannabinoids may decrease radial arm choice accuracy independent of a direct effect on memory, we examined the impact of increasing the intertrial error on radial arm choice accuracy. As previously reported, delta 9-THC impaired maze performance (ED50 = 3.0 mg/kg). Increasing the intertrial interval from 5 s to 30 s resulted in a three-fold increase in the amount of time required to complete the maze without affecting choice accuracy. Importantly, SR141716A prevented delta 9-THC-induced deficits in radial-arm choice accuracy in a dose-dependent manner (AD50 = 2.4 mg/kg); however, the cannabinoid antagonist failed to improve the disruptive effects of scopolamine. Conversely, physostigmine failed to improve performance deficits produced by delta 9-THC. These data provide strong evidence that delta 9-THC impairs working memory through direct action at cannabinoid receptors. Moreover, these results suggest that scopolamine and delta 9-THC do not impair spatial memory in a common serial pathway, though they may converge on a third neurochemical system.