The relative accuracy and precision of the two non-invasive thermometry systems, FirstTemp (Intelligent Medical Systems, USA), an infrared tympanic thermometer, and CTM-205 (Terumo, Japan), a newly developed deep body thermometry system, were evaluated in 32 patients undergoing various surgeries under general anesthesia using esophageal temperature as the reference value. The "limits of agreement (mean difference +/- 2SD)" of the "core" temperature measured by FirstTemp and esophageal temperature was 0.4 +/- 0.5 degrees C, and was larger (P < 0.01) than those between rectal temperature and esophageal temperature (0.2 +/- 0.7 degrees C), between tympanic membrane temperature and esophageal temperature (-0.1 +/- 0.4 degrees C), and between "forehead deep body temperature" measured by CTM-205 and esophageal temperature (-0.2 +/- 0.7 degrees C). The "limits of agreement" between "forehead deep body temperature" and esophageal temperature did not differ from that between tympanic membrane temperature and esophageal temperature. The repeatability of the measurement by FirstTemp was good; the difference between paired measurement values was 0 +/- 0.2 degrees C (mean +/- 2SD). We conclude that the relative accuracy and precision of the two systems are still not sufficient for monitoring body temperature during general anesthesia.