In the dose range of 4.0--32.0 mg/kg s.c., caffeine produced most of the signs which are commonly seen after the administration of naloxone (0.05 mg/kg s.c.) to morphine-dependent monkeys. The signs designated as lying on side or abdomen, avoiding contact, vocalizing, crawling or rolling, restlessness or pacing, tremors, retching, vomiting, coughing, vocalizing when abdomen palpated, rigid abdomen and salivation were noted. A randomized and blind experimental design, which included vehicle and positive (naloxone) controls was used. The significance of the differences between total scores for the whole syndrome was tested by the Mann-Whitney U-test. In preliminary studies in naive monkeys, caffeine was found to elicit some withdrawal signs but the results were equivocal. Na benzoate also elicited some withdrawal signs in morphine-dependent monkeys at 32.0 mg/kg s.c., but few signs were seen in naive monkeys. Caffeine was found to be approximately 10X more active than Na benzoate in inhibiting cAMP phosphodiesterase activity in a neuroblastoma cell whole homogenate assay. These results are consistent with the observations of Collier and Francis that morphine abstinence in rodents is associated with increased brain levels of cAMP.