Previous studies of total sleep deprivation in the rat by the disk-over-water method had shown an initial rise in body temperature (T(b)), a later decline to below baseline levels, and a progressive rise in energy expenditure (EE). To evaluate the role of changes in temperature setpoint in these results, the ambient temperature choices (T(c)) of six totally sleep-derived (TSD) rats and their yoked control (TSC) rats were determined by the position in a thermally graded alley at which they chose to fall asleep. (Rats were removed from the alley and returned to the deprivation apparatus 1 min after sleep onset.) Sleep deprivation was continued until TSD rats were near death. T(b) and EE results were like those of the earlier studies. T(c) rose progressively from 26.8-degrees-C during baseline to 49.9-degrees-C near the end of the deprivation period in TSD rats, whereas TSC rats showed only modest, nonsignificant increases. The attempt by TSD rats to get warmer, in spite of an elevated T(b), indicates a raised setpoint for behavioral thermoregulation and suggests that the increase in EE was also, at least in part, an attempt to reach an elevated setpoint. The progressively raised T(c) also indicates that the late decline in T(b) was a decline below setpoint rather than a response to a lower setpoint. Because EE increased, T(b) decline must have resulted from excessive heat loss. Thus, TSD in the rat produced two thermoregulatory deficits-increased setpoint and excessive heat loss. Because increased T(b) in TSC rats was correlated with reductions in high electroencephalographic voltage nonrapid eye movement (NREM) sleep, it is attributable to the partial loss or interruption of NREM sleep processes. However, because T(c) did not increase in TSC rats, they must have been closer to setpoint than TSD rats.