Quantitative observations of fever and its treatment before the advent of short clinical thermometers.

INTRODUCTION The diagnosis of fever syndromes had rested on both objective and subjective feelings of warmth since antiquity, but by the seventeenth century a fast pulse had become the single most common criterion for diagnosing any fever. For instance, in 1692, Thomas Willis was quite explicit about the importance of this diagnostic clue: "First, we consult the Pulse as a Thermometer constituted by Nature for measuring the heat kindled in a Fever."' Three years later Friedrich Hoffmann explained the febrile rapid pulse as a reflex that occurs when blood vessels become obstructed, and agreed that "a rapid or frequent pulse is present in all fevers".2 Hermann Boerhaave made it an article of faith that a rapid pulse is pathognomonic of fever.3 He, Hoffmann, and, later, William Cullen, who tried to shift the pathophysiological focus to the nervous system,4 proposed classifications of fever based on their own respective theories of its origin.5 However, their diverse hypotheses did not much affect the bedside reasoning that led clinicians to diagnose fevers in most patients with rapid pulses, even in some who did not feel ill. Moreover, fever could be diagnosed in patients with normal or slow pulse rates, if their other symptoms warranted it.