The reduction of body core temperature (Tcore) after salt loading has been reported. In this study, we tested the hypothesis that, during a cold exposure in rats, (1) salt loading would decrease metabolic rate (MR), reducing Tcore, but (2) Tcore would be maintained when cold-escape/warm-seeking behaviour is available. In the first experiment (n = 7), MR and Tcore were measured by indirect calorimetry and telemetry, respectively, during 26, 20 and 10 degrees C exposure for 1 h each, in that order. In the second experiment (n = 7), each rat was placed in an operant system during the same exposure protocol as in the first experiment, where it could trigger a 40 degrees C air reward for 30 s at 20 and 10 degrees C by moving into specific areas (operant behaviour). In each experiment, rats repeated the same protocol twice with a subcutaneous injection (10 ml kg-1) of either isotonic saline (154 mM) or hypertonic saline (2500 mM). In the first experiment, MR in the isotonic-saline trial increased (P < 0.05) at 20 and 10 degrees C compared with that at 26 degrees C by 21 +/- 5 and 48 +/- 6 %, respectively (means +/- S.E.M.), with Tcore unchanged. However, values for MR and Tcore in the hypertonic-saline trial were lower (P < 0.05) than those in the isotonic-saline trial in any ambient temperature. In the second experiment, Tcore was also lower (P < 0.05) in the hypertonic-saline trial than in the isotonic-saline trial. The counts of the operant behaviour in the hypertonic-saline trial remained unchanged in each exposure period, but those in the isotonic-saline trial increased (P < 0.05) at 10 degrees C. These results may suggest that salt loading attenuates both metabolic and behavioural thermoregulatory responses to the cold.