When rats, acclimated to an ambient temperature (T(a)) of 29 degrees C, are exposed to 10% O(2) for 63 h, the circadian rhythms of body temperature (T(b)) and level of activity (L(a)) are abolished, T(b) falls to a hypothermic nadir followed by a climb to a hyperthermic peak, L(a) remains depressed (Bishop B, Silva G, Krasney J, Salloum A, Roberts A, Nakano H, Shucard D, Rifkin D, and Farkas G. Am J Physiol Regulatory Integrative Comp Physiol 279: R1378-R1389, 2000), and overt brain pathology is detected (Krasney JA, Farkas G, Shucard DW, Salloum AC, Silva G, Roberts A, Rifkin D, Bishop B, and Rubio A. Soc Neurosci Abstr 25: 581, 1999). To determine the role of T(a) in these hypoxic-induced responses, T(b) and L(a) data were detected by telemetry every 15 min for 48 h on air, followed by 63 h on 10% O(2) from rats acclimated to 25 or 21 degrees C. Magnitudes and rates of decline in T(b) after onset of hypoxia were inversely proportional to T(a), whereas magnitudes and rates of T(b) climb after the hypothermic nadir were directly proportional to T(a). No hyperthermia, so prominent at 29 degrees C, occurred at 25 or 21 degrees C. The hypoxic depression of L(a) was least at 21 degrees C and persisted throughout the hypoxia. In contrast, T(a) was a strong determinant of the magnitudes and time courses of the initial fall and subsequent rise in T(b). We propose that the absence of hyperthermia at 21 and 25 degrees C as well as a persisting hypothermia may protect the brain from overt pathology.