The respiratory effects of caffeine and paraxanthine, two xanthine adenosine antagonists with phosphodiesterase (PDE) activity, CGS 15943, a non-xanthine adenosine antagonist lacking PDE inhibitory activity, and rolipram, a non-xanthine PDE inhibitor lacking adenosine antagonist activity, were characterized in unanesthetized, seated rhesus monkeys exposed to 10% O2 balanced in N2 (hypoxia). Ventilation was measured continuously by enclosing the monkey's head in a fitted helmet and using a pressure-displacement plethysmographic technique. Respiratory frequency (f) and minute volume (VE) increased during 15-minute periods of hypoxia, and intramuscular administration of caffeine (0.3 and 1.0 mg/kg), paraxanthine (0.3 and 1.0 mg/kg) and CGS 15943 (0.03 and 0.1 mg/kg) attenuated the ventilatory response to hypoxia. In contrast, rolipram (0.003-0.03 mg/kg) did not significantly alter the ventilatory response to hypoxia. Drug effects also were characterized in monkeys exposed to air (normoxia) or 3%, 4% and 5% CO2 balanced in air (hypercapnia). Doses of caffeine, paraxanthine or CGS 15943 that attenuated the ventilatory response to hypoxia had no significant effect on f or VE during conditions of normoxia or hypercapnia. The results indicate that adenosine may play a major role in the function of peripheral, O2-sensitive mechanisms during hypoxia.