Precise quantity of fever was determined in 191 cases of childhood bacterial meningitis by calculating the areas between the line indicating 37.8 degrees C or 39.5 degrees C temperature and the line connecting all individual temperature values. Temperature measurements were performed rectally one to four times a day throughout the hospitalization. The obtained areas under the curves (AUC), expressed as degree-hours, proved to be a sensitive index for delineating each individual fever pattern and reflected the magnitude of fever more precisely than the traditional fever curves. Children under five had significantly (p less than 0.05) greater AUC than those at five to 15 years; similarly, patients with Haemophilus influenzae meningitis showed greater AUC (i.e., had more fever) than those with meningococcal disease (p less than 0.05). The overall rates of secondary (14%), persistent (16%), and prolonged fever (8%) were virtually identical to previous reports; no drug fever was reported in this study. In cases with prolonged fever, a significantly higher rate (40%) of neurological complications was found compared to those who became afebrile earlier. This method is potentially utilizable in other diseases and conditions where precise measurement of fever is of clinical or scientific relevance.