The concentration of delta-9-tetrahydro-cannabinol in marijuana available in the United States has increased by 250% since investigations of the effects of marijuana on short-term memory first appeared in scientific journals. Moreover, previous investigations of short-term memory in long-term marijuana smokers involved adults only. We evaluated the auditory/verbal and visual/spatial memory of 10 cannabis-dependent adolescents and compared the results with performance of 17 subjects in two control groups. The control groups included 8 adolescent drug abusers who had not been long-term users of cannabis and another 9 adolescents who had never abused any drug. All three groups were matched on age, IQ, and absence of previous learning disabilities. Adolescents with a history of frequent alcohol or phencyclidine abuse were excluded from entering the study. A battery of seven neuropsychological tests was administered initially to all subjects and a parallel test battery was administered 6 weeks thereafter. Significant differences between the cannabis-dependent group and the two control groups were obtained initially on the Benton Visual Retention Test (F[2,24] = 6.07) and the Wechsler Memory Scale Prose Passages (F[2,23] = 7.04). After 6 weeks of supervised abstention from intoxicants, subjects in the cannabis-dependent group showed some significant improvement on the Wechsler Memory Prose Passages score and on the Benton Visual Retention Test; however, the improvement failed to achieve statistical significance. We concluded that cannabis-dependent adolescents have selective short-term memory deficits that continue for at least 6 weeks after the last use of marijuana.