A cold strain index (CSI) based on rectal (T(re)) and mean skin temperatures ((sk)) using data from seminude resting subjects has been proposed (Moran DS, Castellani JW, O'Brien C, Young AJ, and Pandolf KB. Am J Physiol Regulatory Integrative Comp Physiol 277: R556-R564, 1999). The current study determined whether CSI could provide meaningful data for clothed subjects exercising in the cold with compromised insulation. Ten men exercised in cold-wet conditions (CW) for 6 h before (D0) and after 3 days of exhaustive exercise (D3). Each hour of CW consisted of 10 min of standing in rain (5.4 cm/h, 5 degrees C air) followed by 45 min of walking (1.34 m/s, 5.4 m/s wind, 5 degrees C air). The change in T(re) across time was greater (P < 0.05) on D3 than on D0, and the change in (sk) was less (P < 0.05) on D3 than on D0. Although CSI increased across time, the index at the end of both trials (D3 = 4.6 +/- 0.6; D0 = 4.2 +/- 0.8) was similar (P > 0.05). Thus, while (sk) was 1.3 degrees C higher (P < 0.05) and T(re) was 0.3 degrees C lower (P < 0.05) on D3 than on D0, CSI did not discriminate the greater heat loss that occurred on D3. These findings indicate that when vasoconstrictor responses to cold are altered, such as after exhaustive exercise, CSI does not adequately quantify the different physiological strain between treatments. CSI may be useful for indicating increased strain across time, but its utility as a marker of strain between different treatments or studies is uncertain because no independent measure of strain has been used to determine to what extent CSI is a valid and reliable measure of strain.