The effects of caffeine and several selective phosphodiesterase (PDE) inhibitors on ventilation and on schedule-controlled behavior were studied in rhesus monkeys. In seated, unanesthetized monkeys prepared with a head plethysmograph, ventilation during exposure to air (normocapnia) and to elevated levels of CO2 (3, 4 and 5%) mixed in air (hypercapnia) was measured after cumulative doses of each drug. In other monkeys, behavioral effects were studied by administering cumulative doses preceding sequential periods of fixed-ratio or fixed-interval responding. The nonselective PDE inhibitors, caffeine and 3-isobutyl-1-methylxanthine, and the type IV-selective PDE inhibitors, rolipram and Ro 20-1724, had pronounced respiratory-stimulant effects during conditions of normocapnia and hypercapnia, and their potencies in increasing ventilation corresponded with their potencies as PDE inhibitors. The type III-selective PDE inhibitor, CI-930, had only modest respiratory-stimulant effects at the highest dose studied, and the type V-selective PDE inhibitor, zaprinast, had no respiratory effect. CGS 15943, a selective adenosine antagonist lacking PDE-inhibitory effects, also had only modest respiratory-stimulant effects at the highest dose studied. In contrast to their relative potencies and efficacies in stimulating respiration, caffeine and 3-isobutyl-1-methylxanthine were less efficacious than CGS 15943 in increasing fixed-interval responding, and CI-930, rolipram and Ro 20-1724 only decreased fixed-interval responding. Zaprinast had little or no behavioral effect. The results support the interpretation that inhibition of type IV PDE plays a prominent role in the respiratory-stimulant effects of xanthines, whereas the behavioral-stimulant effects are more closely related to antagonism of adenosine receptors.