Measurements of core temperature (Tc) at different sites produce on some occasions different cooling curves in cold-exposed humans, suggesting that the corresponding thermometric heat debts (HD) could be equally different when calculated by conventional methods [via the change in either Tc or mean body temperature (Tb)]. The present study also compared these thermometric HD values with the calorimetric HD obtained by partitional calorimetry (S). Nine subjects who showed similar initial but different final Tc [rectal (Tre) and auditory canal temperatures (Tac)] during nude cold exposure (2 h at 1 degrees C at rest) were used. Tc-derived HD corresponded to a heat gain of 12 +/- 21 kJ and an HD of 78 +/- 20 kJ with use of Tre and Tac, respectively, whereas the Tb-derived HD varied from 266 +/- 35 to less than or equal to 1,479 +/- 71 kJ with the use of various well-known Tb weighing coefficients. In contrast, S corresponded to 504 +/- 79 kJ, a level that could have been obtained only if the thermoneutral/cold Tb weighing coefficients had been 0.818/0.818 for Tre and 0.865/0.865 for Tac. The results demonstrate that calculation by conventional methods can markedly overestimate or underestimate HD. These differences could not be explained by the site chosen to represent Tc, inasmuch as about the same effect was observed with use of either Tre or Tac. It is concluded that the thermometric value of HD in the cold is not, at least under the present conditions, as accurate and reliable as S.