OBJECTIVE To determine the behavior of a new and smaller model of infrared thermometer cone and to assess whether it is an appropriate alternative to determine fever in children under 14-years-old. METHODS We performed a cross-sectional, descriptive study comparing the temperatures taken with a mercury thermometer with those taken by an infrared thermometer in children under 14-years-old. The researchers followed the manufacturer's instructions for use for each thermometer. The overall sample of 400 children was divided into two age groups. Group I consisted of children aged less than 2 years and group II of 2-14-years-old. In group I the rectal mercury thermometer was placed in the rectum for 3 minutes and the tympanic thermometer was used in rectal mode and placed in the right ear. In group II, the axillary mercury thermometer was placed for 8 minutes in the right axilla and the tympanic thermometer was used in the axillary mode. To analyze intraobserver bias, 50 patients were selected from the sample and their temperatures were simultaneously taken by two nurses from the team. The intraclass correlation coefficient (ICC) was used both to measure the reliability of the tympanic thermometer and to analyze intraobserver bias. RESULTS Temperature measurements with both instruments showed an ICC of 0.91 (95% confidence interval [CI], 0.88-0.94) for group I and 0.90 (95% CI, 0.87-0.92) for group II. The reproducibility of the measurements taken by the two nurses in 50 patients showed an ICC of 0.97 (95% CI, 0.95-0.98) for the tympanic thermometer. CONCLUSIONS The infrared thermometer is an appropriate device for rapidly measuring temperature in the emergency department. However, the measurements taken should be confirmed by another method when clinical decisions are based on temperature values.