Guidelines to establish safe environmental exercise conditions are partly based on thermal prescriptive zones. Yet there are reports of self-paced human athletic performances in extreme heat. Eighteen participants undertook a 25-km route march in a dry bulb temperature reaching 44.3°C. The mean (± s) age of the participants was 26.0 ± 3.7 years. Their mean ad libitum water intake was 1264 ± 229 mL · h(-1). Predicted sweat rate was 1789 ± 267 mL · h(-1). Despite an average body mass loss of 2.73 ± 0.98 kg, plasma osmolality and serum sodium concentration did not change significantly during exercise. Total body water fell 1.47 kg during exercise. However, change in body mass did not accurately predict changes in total body water as a 1:1 ratio. There was a significant relationship (negative slope) between post-exercise serum sodium concentration and changes in both body mass and percent total body water. There was no relationship between percent body mass loss and peak exercise core temperature (39 ± 0.9°C) or exercise time. We conclude that participants maintained plasma osmolality, serum sodium concentration, and safe core temperatures by (1) adopting a pacing strategy, (2) high rates of ad libitum water intake, and (3) by a small reduction in total body water to maintain serum sodium concentration. Our findings support the hypothesis that humans are the mammals with the greatest capacity for exercising in extreme heat.