Recently, we have shown that specific, transient carotid chemoreceptor (CC) inhibition in exercising dogs causes vasodilatation in limb muscle. The purpose of the present investigation was to determine if CC suppression reduces muscle sympathetic nerve activity (MSNA) in exercising humans. Healthy subjects (N = 7) breathed hyperoxic gas (F(IO(2)) approximately 1.0) for 60 s at rest and during rhythmic handgrip exercise (50% maximal voluntary contraction, 20 r.p.m.). Microneurography was used to record MSNA in the peroneal nerve. End-tidal P(CO(2)) was maintained at resting eupnoeic levels throughout and breathing rate was voluntarily fixed. Exercise increased heart rate (67 versus 77 beats min(-1)), mean blood pressure (81 versus 97 mmHg), MSNA burst frequency (28 versus 37 bursts min(-1)) and MSNA total minute activity (5.7 versus 9.3 units), but did not change blood lactate (0.7 versus 0.7 mm). Transient hyperoxia had no significant effect on MSNA at rest. In contrast, during exercise both MSNA burst frequency and total minute activity were significantly reduced with hyperoxia. MSNA burst frequency was reduced within 9-23 s of end-tidal P(O(2)) exceeding 250 mmHg. The average nadir in MSNA burst frequency and total minute activity was -28 +/- 2% and -39 +/- 7%, respectively, below steady state normoxic values. Blood pressure was unchanged with hyperoxia at rest or during exercise. CC stimulation with transient hypoxia increased MSNA with a similar time delay to that obtained with CC inhibition via hyperoxia. Consistent with previous animal work, these data indicate that the CC contributes to exercise-induced increases in sympathetic vasoconstrictor outflow.