Halothane in humans depresses the ventilatory response to hypoxemia in a manner that suggests a selective action on one or more components of the peripheral chemoreflex arc. To test the hypothesis that this action is at the carotid bodies themselves, the authors studied the ventilatory response to subanesthetic concentrations of halothane (0.15–0.30% inspired) in six fit volunteers maintained in a steady state of isocapnic hypoxemia (PEo250 mmHg). Upon exposure to halothane, hypoxemia-driven ventilation decreased promptly and progressively (from 7.5 · 1.2 1 · min−1· m−2in the control state to 5.9 · 0.9 and 4.8 · 0.7 1 · min−1· m−2at 30 s and 60 s of inhalation respectively, means · SEM). The relationship of hypoxemia-driven ventilation to end-tidal halothane tensions at 30 and 60 s of halothane wash-in (PEHal0.4 and 0.6 mmHg, respectively) approached the relationship observed in near steady states of halothane inhalation. The results are interpreted as indicating that the site of selective action is at a tissue that accumulates halothane very rapidly during the first minute of inhalation. To make possible such pharmacokinetics, that tissue would require 1) a location having a brief circulatory transit time from the lungs, and 2) an extremely high rate of perfusion in relation to its capacity for uptake of halothane. The only tissue of the peripheral chemoreflex pathway that can satisfy these requirements is that of the carotid bodies.