The purine nucleoside, adenosine, has been implicated as a neuromodulator in central respiratory depression during prolonged exposure to hypoxia. It may also be a mediator of hypoxic hyperpnoea, acting on the carotid bodies. As there may be adenosine-sensitive mechanisms of hypoxic respiratory control, we sought to determine whether adenosine might be involved as a respiratory modulator in another central but non-oxygen-related control mechanism, the ventilatory response to hyperoxic hypercapnia.Twelve normal subjects were studied following 3 days of oral treatment with placebo, dipyridamole (which potentiates adenosine effects by inhibiting cellular uptake), and theophylline (a specific adenosine antagonist of cell surface receptors). The drugs were given in a random order, double-blind fashion. Resting endtidal carbon dioxide tension (PET,CO 2 ) and the maximum rate of isometric inspiratory pressure change at the mouth ((dP/dt )max), an index of respiratory drive, were determined in all subjects on each treatment. Hyperoxic, hypercapnic ventilatory responses were determined in seven of these subjects using a rebreathing technique. For each hypercapnic response, minute ventilation (V 'E) and (dP/dt )max were plotted against PET,CO 2 breath-by-breath.Resting PET,CO 2 breathing room air was lower with theophylline (5.47 (SD 0.21) kPa) than with placebo (5.74 (0.26) kPa) or dipyridamole (5.86 (0.34) kPa), with no significant drug differences in resting (dP/dt )max. However, neither the slope nor the PET,CO 2 intercept of the relationship between ventilation or respiratory drive and PET,CO 2 were altered by the study drugs under hyperoxic conditions. We conclude that endogenous adenosine-related mechanisms are unlikely to be involved in determining either the sensitivity or the threshold of the ventilatory response to carbon dioxide under hyperoxic conditions. However, in normoxia, a centrally-acting, tonic, adenosine-mediated, respiratory modulation is not ruled out.