The control of sweating in humans has been described quantitatively in terms of skin and core temperatures (Tsk and Tcore, respectively). However, the precision with which features of the relationship between sweat rate and Tcore at a given Tsk can be reproduced in the short term is not known. We focused on the threshold Tcore. We held Tsk at 38 degrees C until sweating began for two periods separated by a period of cooling with Tsk at 32 degrees C in six men and three women. The esophageal temperature (Tes) at which sweating began was invariably lower in the second period of heating (average difference 0.09 degree C; maximum 0.17 degree C). Also, the rate of rise in Tes was invariably higher (average 148%) during the second period of heating. Thus, although a threshold cannot be reproduced within the error of Tes measurement, the consistency and small magnitude of the downward shift recommend our protocol as a practical method for evaluating other influences on thermoregulation, provided that the effects are big enough to be seen against a background of an expected small decrease. From the fundamental point of view, the consistency of the downward displacement has provocative implications, e.g., the rate of change in Tcore influences sweating or thermosensitive units in slow-responding thermal compartments contribute to the Tcore input signal.