This study assessed the effects of head temperature on subjects under hyperthermic conditions. Six volunteers were trained to plateau on a manikin task that tested cognitive performance and reaction time. A subject's core temperature was driven with a full-length, liquid-conditioned garment perfused with water at 30 degrees or 43 degrees C, and head temperature was controlled by a cap with an inlet temperature of 8 degrees or 43 degrees C. Heavy insulation was worn overall; ambient temperature was 30 degrees C. Subjects were heated and cooled twice during each experiment to include all possible cap/suit temperature combinations. Each subject did one control and four stress experiments. Comfort and performance were measured as the subject's esophageal temperature (Tes) rose and fell through the range 37.5-39.0 degrees C. Cap temperature did not affect rectal temperature (Tre) but significantly altered the Tes rate of change; a cool head sometimes truncated the peak Tes value. Although the cap covered only 3-4% of the body surface, the head was a major determinant of subjective comfort. Body heating tended to shorten reaction time and diminish performance accuracy, while head cooling largely reversed these trends. Possible mechanisms include 1) change in overall heat balance; 2) countercurrent exchanges in the neck, and 3) change in sensory output. Head cooling deserves serious consideration for machinery operators where whole-body thermoneutrality is impractical.