The recent discovery of arachidonylethanolamide (anandamide), an endogenous ligand for cannabinoid receptors, and the synthesis of SR141716A, a cannabinoid antagonist selective for brain cannabinoid (CB1) receptors, have provided new tools to explore the mechanisms underlying cannabis abuse and dependence. Drug discrimination is the animal model with the most predictive validity and specificity for investigation of the psychoactive effects of cannabinoids related to their abuse potential, because, unlike many other drugs of abuse, delta9-tetrahydrocannabinol (delta9-THC), the major psychoactive ingredient of marijuana, is not self-administered by animals. Results of delta9-THC discrimination studies have revealed that the subjective effects of cannabis intoxication are pharmacologically selective for centrally active cannabinoid compounds, and that cannabis action at CB1 receptors is involved in medication of these effects. Less clear is the role of endogenous cannabinoid system(s) in cannabis intoxication. Anandamide, named for a Sanskrit word for "internal bliss," unreliably substitutes for delta9-THC. Further, substitution, when it is observed, occurs only at doses that also significantly decrease response rates. In contrast, delta9-THC and other structurally diverse cannabinoids fully substitute for delta9-THC at doses that do not substantially affect response rates. Attempts to train animals to discriminate anandamide (or SR141716A) have so far been unsuccessful. Preliminary evidence from drug discrimination studies with more metabolically stable anandamide analogs have suggested that these differences in the discriminative stimulus effects of delta9-THC and anandamide-like cannabinoids are not entirely due to pharmacokinetic factors, but the exact role of "internal bliss" in cannabis intoxication and dependence is still not completely understood.