OBJECTIVE This study compared the agreement between core temperature measurements obtained using an ingestible temperature pill telemetry system (Tpill) with those obtained from rectal (Tre) and esophageal (Tes) thermocouples under conditions of increasing and decreasing body temperature. METHODS Four men and five women (age 25+/-2 yr, BSA 1.81+/-0.05 m2, VO2 peak 3.1+/-0.4 L x min[-1]) participated in four 3-h trials: cold (18 degrees C) water rest (CWR), cold water exercise (CWE), warm (36 degrees C) water rest (WWR), and warm water exercise (WWE). Subjects were immersed to the neck for each trial. During resting trials, subjects sat quietly. During exercise trials, subjects completed three bouts of 15 min of rest, followed by 45 min of exercise on a cycle ergometer at 50% of peak oxygen uptake. The temperature pill was taken 10-12 h before testing, after which the subjects fasted. RESULTS The trials created conditions of constantly decreasing (CWR) or increasing (WWR) core temperature, as well as periods of oscillating core temperature (CWE and WWE). Root mean squared deviation (RMSD) was calculated for each pair of measurements (Tpill vs Tre, Tpill vs Tes, Tre vs Tes) for each trial. An RMSD of "0" indicates perfect agreement; as RMSD increases, agreement worsens. On CWR, the RMSD for Tpill-Tes (0.23+/-0.04) was lower (P < 0.05) than for Tpill-Tre (0.43+/-0.10) or Tre-Tes (0.46+/-0.09). There were no significant differences in RMSD between measurement pairs on any other trial (average RMSD = 0.26 degrees C). Telemetry pill temperature and response time tended to be intermediate between Tre and Tes. CONCLUSIONS These results suggest the telemetry pill system provides a valid measurement of core temperature during conditions of decreasing as well as increasing body temperature and during steady state.