We tested the hypothesis that, in humans, hypohydration attenuates hyperthermic hyperpnea during exercise in the heat. On two separate occasions, thirteen male subjects performed a fluid replacement (FR) and a no-fluid replacement (NFR) trial in random order. The subjects performed two bouts of cycle exercise (Ex1 and Ex2, 30-60 min) at 50% peak oxygen uptake (Vo2 peak) in 35 degrees C separated by a 70- to 80-min rest period, during which they drank water containing 25 mosmol/l sodium in the FR trial but not the NFR trial. The drinking in the FR trial nearly restored the body fluid to the euhydrated condition, so that the body fluid status differed between the trials before Ex2 (the difference in plasma osmolality before Ex2 was 9.4 mosmol/kgH2O; plasma volume was 7.6%, and body weight was 2.5%). The slopes of the linear relationships between ventilatory variables (minute ventilation, ventilatory equivalents for oxygen uptake and carbon dioxide output, tidal volume, respiratory frequency, and end-tidal CO2 pressure) and esophageal temperature (Tes) did not significantly differ between Ex1 and Ex2, or between the FR and NFR trials. On the other hand, during Ex2 in the NFR trial, the Tes threshold for the onset of increased forearm vascular conductance (FVC) was higher, and the slope and peak values of the relationship between FVC and Tes were lower than during Ex1 in the NFR trial and during Ex2 in the FR trial. These findings suggest that hypohydration does not affect the hyperthermic hyperpnea during exercise, although it markedly attenuates the cutaneous vasodilatory response.