To the Editor.
—The recent study of normal body temperature by Mackowiak et al1 should help to discredit the special significance of 37.0°C (98.6°F)—surely the best known number in all of medicine, but one without any particular clinical relevance. As they suggest, no single value of mean body temperature has much clinical importance because of individual and population variability; it is much more useful to know the upper limits of normal for morning and evening, the trough and peak of the diurnal variation. For several reasons, however, the limits found in their study may underestimate the true upper limits of normal. First, women comprised only 17% of their study population, but probably have slightly higher upper limits of normal than do men. Second, because they recorded only the 6 PM and midnight temperatures of their subjects, they may have missed the evening maximum temperature, which a small but thorough