In a two phase design, an attempt was made to differentiate the effect of marijuana on the storage and retrieval of prose material. In the first phase, 40 male subjects were administered a single 500 mg marijuana cigarette containing 2.1%delta9-THC or a placebo cigarette. Fifteen minutes after smoking, they listened to and at the same time read a narrative passage of approximately 200 words in length. Subsequently, an immediate free recall test was given in which subjects were required to write down as much of the story as they could remember. The second phase was conducted 24h later. Marijuana and placebo subjects were randomly subdivided into four groups with half of the subjects participating in the same drug condition as occurred on day one while the others switched drug state. Fifteen minutes after smoking, all subjects recalled the passage presented on day one and then were given 24 questions concerning facts and events in the story which could be answered in a few words. These questions served as retrieval cues. Following this, a new passage was presented in the same manner as occurred on day one. After an immediate free recall test, another cued recall test was administered. Results indicated that marijuana reduced immediate recall under both cued and uncued conditions incomparison to placebo. No relative cued recall advantage was found in the marijuana groups for the old or new story and marijuana produced only a moderate decrement in recall of the old story on day two. However, marijuana given in the second phase significantly reduced memory for items recalled in the initial phase irrespective of drug or cueing condition in phase one, suggesting that retrieval was also affected. Some decrement in recall of the new story did occur as a function of drug state change in group M-P. This effect was related to the serial position of input items. Serial position did not interact with drug state under any other recall condition.