Our objective was to determine the factors that affect (i) the rectal-axillary temperature measurement difference (RATD) for digital and mercury-in-glass thermometers and (ii) the difference between mercury-in-glass and digital thermometer readings taken by the rectal and axillary routes. Children (between the ages of 2 and 48 months) who were admitted to Hacettepe University Children Hospital and Corum Children's Hospital in 2 months period were included in this study. Two measurements were taken from each patient within 2 days at two different occasions including day/night and sleep/awake by the same physician. A total of 135 patients were enrolled. RATD was 0.61 + 0.54 degrees C in the mercury-in-glass thermometer, and 0.81 +/- 0.57 degrees C in the digital thermometer. The mean differences between digital and mercury-in-glass thermometers were 0.16 degrees C for rectal routes and -0.02 degrees C for axillary routes. Children at older age with hypoalbuminemia and children exposed to higher ambient temperatures had lower values of RATD than others. Temperature recordings from different sites might change with ambient temperature and individual characteristics of children.