We have reviewed the evidence in favor of a prostaglandin mediator of the thermal responses in fever and found that PGE injected into the hypothalamus does not always cause fever, that cerebrospinal fluid concentrations of PGE are not reliable reflections of hypothalamic events, and that antipyretic drugs may act in ways other than inhibiting PGE synthesis. Fever is not blocked by prostaglandin antagonists, nor by ablation of PGE-sensitive areas of the brain. There is poor correlation between the effects of pyrogens and of PGE on cerebral neurons. There is evidence that at least one prostanoid other than prostaglandin is a mediator of fever, but the prostanoid has not been identified yet. We conclude that PGE may contribute to the neural responses in fever but is not essential.