Accurate interpretation of antipyretic drug action may be confounded by the temperature acquisition site. As an initial test of this hypothesis, we compared intermittently measured tympanic membrane temperature to continuously recorded rectal and axillary temperatures. Twelve (12) children (7 males, 5 females), ranging in age from 8 months to 13 years, were studied for 4 to 8 hours. Seven (7) subjects were normothermic (rectal temperature < 100 ° F), 3 subjects had a low-grade fever (rectal temperature between 100 and 102° F), and 2 subjects had a high-grade fever (rectal temperature > 102° F). Subjects with high fever were given acetaminophen at a dose appropriate for their age. Rectal and axillary data were acquired continuously at 1 Hz using a Biolog Data Recorder (UFI, Morrobay, Calif.). Right and left tympanic membrane temperatures were measured at least hourly using a CoreCheck® Thermometer System (Model 2090, IVAC Corp., San Diego, Cal.). Differences between rectal temperature, and the corresponding axillary and tympanic membrane temperatures were averaged for each subject. Axillary temperature averaged 2.71 ± 0.34° F (mean ± SEM) below rectal temperature. Right and left tympanic membrane temperatures averaged 1.64 ± 0.65° F and 1.50 ± 0.71° F, respectively, below rectal temperature. There was no apparent correlation between surface area or age of the child and the difference between rectal and axillary temperature. In this pilot study, fever did not appear to alter temperature site differences. The results suggest that tympanic membrane temperature, obtained intermittently, approximates rectal temperature more closely than continuously recorded axillary temperature.