Early human ancestors evolved hunting in the midday heat on the dry African savannah and developed favourable biological adaptations that permit prolonged running in the heat. These physiological adaptations must have included the capacity to sweat profusely to maintain a low body temperature when running for 4-6 hours in dry heat, an absence of adverse consequences from developing mild to moderate fluid deficits caused by sweat losses during the hunt, a serum osmolality based thirst mechanism and the ability to 'outrun their thirst' (to resist the deleterious psychological and other effects of severe thirst). Until the early 1970s, the guidelines for fluid ingestion during exercise were not to drink and are consistent with this interpretation. By 1996, guidelines stated, "individuals should be encouraged to consume the maximal amount of fluids during exercise that can be tolerated without gastrointestinal discomfort up to a rate equal to that lost from sweating," and this was interpreted by some as "to drink as much as tolerable." This article argues that humans are designed to drink just enough to maintain plasma osmolality, not necessarily bodyweight, both at rest and during exercise. Drinking to maintain bodyweight may impair exercise performance by inducing a weight penalty and may increase the probability of exercise-associated hyponatraemia in slow marathon runners.