A single temperature measurement recorded on admission to the emergency department provides no information about temperature alterations occurring during the course of evaluation. Continuous monitoring of patients' temperatures in the ED, however, may alter management and decrease morbidity. Our study evaluated the reliability of liquid crystal thermometers (LCTs) and the clinical benefit of continuous temperature monitoring in the ED. Commercially available LCTs (corrected 4 degrees F to reflect core temperature) were applied to the foreheads of randomly selected patients. Serial oral electronic thermometer readings were compared to those obtained by LCT. Fever was defined as a temperature higher than 99.5 F orally or 100 F by LCT. One hundred two patients underwent simultaneous LCT and oral temperature measurements, with a correlation coefficient of 0.661. Hypothermia was not encountered. Eighty-four patients were afebrile, and 18 were febrile by oral measurement on admission. Of the afebrile patients, 13 (15.5%) became febrile while in the ED. The temperature course was identified correctly by LCT in 83.3% of cases. The LCT correctly identified all patients who were febrile on admission, as well as 92.3% of those who developed fever while in the ED. The latter fevers would have been missed by routine single-temperature determination on ED admission. Detection of fever stimulated more aggressive clinical evaluation of these patients. Eight of nine patients who defervesced in response to antipyretic therapy were identified correctly by LCT.(ABSTRACT TRUNCATED AT 250 WORDS)