Metabolic heat production, heat loss and the circadian rhythm of body temperature in the rat

Transcription

1 Metabolic heat production, heat loss and the circadian rhythm of body temperature in the rat Circadian rhythmicity is ubiquitous in mammalian physiology. A circadian pacemaker located in the suprachiasmatic nuclei of the hypothalamus is responsible for an oscillatory process that is expressed in practically every function of the organism (van Esseveldt et al. 2000; Moore & Leak, 2001). Variables whose circadian properties have been extensively studied include body temperature, locomotor activity, blood pressure and secretion of the hormone melatonin (Refinetti, 2000; Czeisler & Dijk, 2001). Several studies have examined the daily oscillation in metabolic rate of birds (Aschoff & Pohl, 1970; Graf, 1980; Rashotte et al. 1995) and mammals (Heusner, 1956; Morrison, 1968; Aschoff & Pohl, 1970; Fuller et al. 1985; Haim et al. 1988; Henken et al. 1993; Haim & Zisapel, 1995; Brown & Refinetti, 1996; Robinson & Fuller, 1999). In all cases, a robust daily oscillation in metabolism was observed. Metabolism and body temperature followed a very similar time course with a peak during the activity phase of the rest activity cycle (i.e. during the day in diurnal animals and during the night in nocturnal animals). However, with few exceptions, these studies did not measure heat loss simultaneously with heat production. Measurement of heat loss is essential for evaluation of the variations in heat balance responsible for the generation of the body temperature rhythm. Roberto Refinetti Circadian Rhythm Laboratory, University of South Carolina, PO Box 1337, Walterboro, SC 29488, USA (Manuscript received 5 November 2002; accepted 27 February 2003) Metabolic heat production (calculated from oxygen consumption), dry heat loss (measured in a calorimeter) and body temperature (measured by telemetry) were recorded simultaneously at 6 min intervals over five consecutive days in rats maintained in constant darkness. Robust circadian rhythmicity (confirmed by chi square periodogram analysis) was observed in all three variables. The rhythm of heat production was phaseadvanced by about half an hour in relation to the body temperature rhythm, whereas the rhythm of heat loss was phase-delayed by about half an hour. The balance of heat production and heat loss exhibited a daily oscillation 180 deg out of phase with the oscillation in body temperature. Computations indicated that the amount of heat associated with the generation of the body temperature rhythm (1.6 kj) corresponds to less than 1 % of the total daily energy budget (172 kj) in this species. Because of the small magnitude of the fraction of heat balance associated with the body temperature rhythm, it is likely that the daily oscillation in heat balance has a very slow effect on body temperature, thus accounting for the 180 deg phase difference between the rhythms of heat balance and body temperature. Experimental Physiology (2003) 88.3, Studies in humans exposed to long periods of bed rest have clearly established that the body temperature rhythm is not a mere collateral effect of the increased heat production associated with locomotion during the activity phase of the rest activity cycle (Murray et al. 1958; Smith, 1969; Marotte & Timbal, 1981; Gander et al. 1986). In rodents, voluntary bed rest is unattainable, but it has been shown that any specific level of locomotor activity is associated with higher levels of body temperature during the animal s activity phase than during the rest phase of the rest activity cycle (Bolles et al. 1968; Honma & Hiroshige, 1978; Refinetti, 1994, 1999). Thus, for instance, the metabolic rate of a nocturnal rodent at rest at night is higher than the metabolic rate of the same animal at rest during the day, which indicates that the body temperature rhythm is produced by a thermogenic or thermolytic process distinct from the one associated with the rhythm of locomotor activity. Simultaneous recording of metabolic heat production, locomotor activity and body temperature in golden hamsters showed that a given level of locomotor activity is associated not only with higher levels of body temperature but also with higher levels of metabolic heat production during the animal s activity phase (Brown & Refinetti, 1996). However, the measured day night increment in heat production was much greater than that needed to account for the variation in body temperature, Publication of The Physiological Society refinetti@sc.edu 2521

2 424 R. Refinetti Exp Physiol 88.3 which points to the necessity of simultaneous recording of heat loss. The large reduction in body temperature associated with torpor and hibernation (often greater than 30 C) results from a thermoregulated reduction in metabolic heat production, even though the fall in body temperature itself helps further reduce metabolic rate (Heller & Hammel, 1972; Heldmaier & Ruf, 1992; Song et al. 1995; Zimmer & Milsom, 2001). The mechanism responsible for the much smaller variation in body temperature associated with circadian rhythmicity (rarely greater than 4 C) has not been investigated in detail. It has been suggested that the circadian rhythm of body temperature results more from oscillations of heat loss (thermal conductance) than from oscillations of heat production (Aschoff, 1970; Refinetti & Menaker, 1992) but experimental support for this proposition has not been produced. To help fill this gap, simultaneous long-term recording of metabolic heat production, heat loss and body temperature in rats was conducted in the present study. METHODS Six male Long-Evans rats ( g) were used as subjects. Prior to experimentation, they were kept in individual plastic cages (25 cm w 46 cm w 20cm) lined with wood shavings under a h light dark cycle at 24 C. Rodent chow (Lab Diet 5001, Purina LabDiet, St Louis, MO, USA) and water were available at all times. Animals were cared for and experimented upon in accordance with the guidelines set out in the Guide for the Care and Use of Laboratory Animals by the US National Institutes of Health. The experimental setup included equipment for the measurement of metabolic heat production, dry heat loss and body temperature. All equipment was controlled by a personal computer via analogto-digital input boards and digital output boards (A-Bus System, Alpha Products, Darien, CT, USA). Dry heat loss was measured with a gradient-layer calorimeter (Model SEC-1201, Thermonetics Corp., San Diego, CA, USA). The internal dimensions of the calorimeter were 15 cm w 30cm w 15 cm. The animals were housed individually in a wire-mesh cage fitting tightly inside the calorimeter. Water at 24 C was circulated through the outer jacket of the calorimeter to ensure adequate heat flow through the wall of the calorimeter and to provide a constant-temperature environment for the animal. Air at 24 C was passed through the inner chamber of the calorimeter at the rate of 4 l min _1. Calibration of the calorimeter (using an electrical heater with known output) was conducted with the air flow on. An aliquot of the air passing through the calorimeter (1 l min _1 ) was dried in desiccant columns (Silica Gel Blue, Fluka Chemika, Buchs, Switzerland) and fed into a metal-air-battery oxygen analyser (O 2 -ECO, Columbus Instruments, Columbus, OH, USA) for the determination of oxygen consumption. To avoid drifts in the calibration of the oxygen analyser over time, the analyser was re-calibrated every 6 min with room air (20.9 % oxygen) bypassing the calorimeter by means of solenoid valves controlled by the computer. Metabolic heat production was calculated from the measurements of oxygen consumption (STPD) using a caloric equivalent of 20.3 J ml _1 (Schmidt-Nielsen, 1983). Body core temperature was measured by telemetry. Radio sensor-transmitters for the monitoring of body temperature (Model VM-FH, Mini-Mitter Co., Bend, OR, USA) were implanted intraperitoneally at least a week before the beginning of the study. The transmitters, which weighed 3 g (i.e. less than 1 % of the rat s body mass), generated temperature-dependent signals with accuracy greater than 0.1 C. The animals were anaesthetized with sodium pentobarbital (60mg kg _1 I.P.), and a 2 cm incision was made in the abdominal wall under sterile conditions. The transmitter was placed inside the abdominal cavity, and the muscle wall was sutured. The skin incision was closed with wound clips. An antibiotic ointment was applied to the wound and the animals allowed to recover. A loop antenna was wound around the wire-mesh cage and attached to a customized radio-receiver board (Model RTA-500, Mini Mitter Co.). The receiver board generated an analog signal proportional to the temperature of the transmitter. This analog signal was fed into the computer interface boards. Each of the animals was placed in the apparatus for five consecutive days, during which time measurements were made of heat production, dry heat loss and body temperature every 6 min (i.e. 0.1 h). Food and water were available ad libitum in small containers. Urine and faeces passed through the bottom of the wire-mesh cage and were absorbed by a thin layer of wood bedding. The inside of the calorimeter was constantly dark. Constant darkness allows the circadian pacemaker to free-run and, therefore, to exhibit a period slightly different from 24 h. Since the free-running period of the rat is between 24.0 and 24.2 h, a maximum phase delay of 12 min takes place each day. At the end of the experiments the animals were killed by CO 2 inhalation. Because the evaporative and convective heat loss components associated with the flow of air through the calorimeter were not measured, the values of heat loss obtained in this study were always smaller than the values of heat production. Since ambient temperature was constant (at 24 C) throughout each session, it is assumed that dry heat loss was a constant fraction of total heat loss. The time course of variations in heat loss was considered more important than the absolute amount of heat lost. Data were recorded on disk during the experimental sessions and analysed off-line later. To eliminate the potentially confounding influence of ultradian oscillations, all data were smoothed by a 4 h moving-averages procedure before further analysis. This procedure preserves the circadian time structure of the data while eliminating high-frequency oscillations (Refinetti, 1992). Because the first few hours of each data set had to be discarded during adaptation of the animals to the chamber, and because the data sets were shortened further by the moving-averages procedure, data analysis was restricted to four full days (960 data points) for each variable for each animal. To determine the presence of statistically significant rhythmicity in the data set for each animal, the x 2 periodogram procedure was used (Sokolove & Bushell, 1978). This procedure also provides a numerical index of the robustness of a rhythm, Q P. The Q P statistic was calculated for each individual using the 4 days of data (960 data points). For data sets of this size, a perfect wave (such as a mathematically generated cosine wave) produces a Q P value of 960. Thus all Q P values were expressed as a percentage of 960, which represents a perfectly stable wave form. Q P values above 23 % of maximal robustness are statistically significant at the 0.05 level (29 % at the 0.01 level).

3 Exp Physiol 88.3 Circadian rhythm of body temperature 425 Three other parameters of the rhythms were analysed: mean level, range of oscillation and acrophase (i.e. the time of the daily peak). The mean level of the rhythms was computed as the arithmetic mean of the 240 daily measurements (at 0.1 h intervals) for each individual. The mean for each animal over the 4 days was then calculated. Likewise, the range of excursion was computed as the difference between the highest and the lowest temperature each day for each individual, and the mean for each animal was calculated. The acrophase of the rhythm was calculated by the method of the single cosinor (Nelson et al. 1979) for each day for each animal, and the mean for each animal was then computed. Since the animals were free-running, absolute acrophase values are meaningless (because they are expected to drift over time). For each animal, on each 24 h cycle, the acrophases for heat loss and heat production were expressed as deviations from the acrophase for body temperature. For the computation of a heat balance index (heat production minus heat loss), the raw scores for metabolic heat production and dry heat loss were first converted to z scores each day for each animal using the following equation: z i =(x i _ m)/s, where m is the mean of the 240 daily raw scores (x i ) and s is the standard deviation of the raw scores. The z scores for heat loss were then subtracted from the z scores for heat production in each of the 960 time points for each animal to produce dimensionless scores of heat balance. Additional statistical analysis involved comparison of group means by analysis of variance (ANOVA), post hoc pairwise comparisons Table 1. Mean level, range of excursion and robustness of the rhythms of body temperature, heat production and heat loss Mean level Range Robustness Body temperature 37.4 ± 0.4 C 1.2 ± 0.2 C 60 ± 7 % Heat production 1.99 ± 0.11 W 0.70 ± 0.04 W 64 ± 6 % Heat loss 0.75 ± 0.05 W 0.50 ± 0.03 W 87 ± 5 % Values shown are means ± S.E.M. (n = 6). by Tukey s HSD test and computation of correlation coefficients by the principle of the least squares (Hays, 1988). RESULTS Records of body temperature, heat production and heat loss over 4 days for a representative rat are shown in Fig. 1. Clear circadian rhythmicity was observed in all three variables. Both heat production and heat loss rose in synchrony with body temperature during subjective night (and fell during subjective day) each day. The mean results for the three parameters of the rhythms of the six rats are shown in Table 1. Because the evaporative and convective heat loss components associated with the flow of air through the calorimeter were not measured, the measured Figure 1 Records of the rhythms of body temperature, metabolic heat production and dry heat loss in a representative rat over 4 days. The data were collected and are plotted in 6 min intervals. All data sets were smoothed by a 4 h movingaverages filter to eliminate high-frequency oscillations.

4 426 R. Refinetti Exp Physiol 88.3 Figure 2 Records of body temperature, metabolic heat production and dry heat loss in a representative rat over a 24 h period. Thin continuous lines indicate actual measurements (smoothed by a 4 h moving-averages filter). Thick continuous lines indicate the best-fit cosine waves used to calculate the acrophase of the rhythms. Vertical dashed lines indicate the acrophases. heat loss accounted for only 38 % of heat production. However, the rhythm of heat loss was quite robust, reaching 87 % of maximal robustness. Rhythm robustness was significantly different between the three variables (F (2,10) = 15.37, P < 0.01) because the robustness of heat loss was greater than the robustness of either heat production or body temperature (Tukey s HSD post hoc test). Robustness of all rhythms was above the 0.01 level of significance (Q P > 29 %). The mean free-running period was indistinguishable from 24.0h in all animals. Body mass did not change substantially (< 2 g) during the short duration of the study. Figure 3 Mean acrophases of the rhythms of heat production and heat loss in six rats over four consecutive days. The acrophases are expressed as deviations from the acrophase of the body temperature rhythm for each animal ( phase difference ). Negative phase differences indicate that the acrophase precedes that of the body temperature rhythm, whereas positive phase differences indicate that the acrophase is preceded by the acrophase of the body temperature rhythm.

5 Exp Physiol 88.3 Circadian rhythm of body temperature 427 To allow better visualization of the temporal relationship between the variables, records for a single day for a representative rat are shown in Fig. 2. The figure also shows the cosine waves used to compute the acrophase of the rhythms. As indicated by the vertical dashed lines, the acrophase for heat production preceded the acrophase for body temperature, whereas the acrophase for heat loss succeeded the acrophase for body temperature by about an hour. The mean phase differences between the rhythm of heat production and the rhythm of body temperature, and between the rhythm of heat loss and the rhythm of body temperature, for each of the 4 days are shown in Fig. 3. Although there was considerable day-to-day variability, the rhythm of heat production consistently preceded the rhythm of body temperature, whereas the rhythm of heat loss consistently succeeded it. Analysis of variance revealed a significant effect of variable type (heat production or heat loss, F(1,35) = 8.97, P < 0.01) but no effect of days (F(3,35) = 1.68, P > 0.10) or of the interaction between variable type and days (F(3,35) = 0.58, P > 0.10). For a closer analysis of the effects of heat balance oscillation on the oscillation of body temperature, standardized deviation scores were computed. Examples for four animals are shown in Fig. 4. Visual inspection suggests that the balance between heat production and heat loss is positive when body temperature is low and negative when body temperature is high. Significant (P < 0.01) but moderate negative correlations (r = _0.18 to r = _0.22) were found between heat balance and body temperature for the six rats. DISCUSSION Telemetric recording of the body temperature rhythm of rats provided a mean level of 37.4 C, a range of excursion of 1.2 C and a robustness of 60% of maximal robustness. These results are consistent with those of previous studies in other laboratories, which indicated a mean level of C and a range of excursion of C (Roussel et al. 1976; Tanaka et al. 1990; Shiromani et al. 1991; Yoda et al. 2000; Stephenson et al. 2001). In a previous study on 10 different species in this laboratory, the robustness of the body temperature rhythm was found to range from 50 to 80 % of maximal robustness (Refinetti, 1998). Figure 4 Heat balance (i.e. heat production minus heat loss) and body temperature records on four successive days in four representative rats. Values are expressed as normalized deviations from the mean (i.e. as z scores).

6 428 R. Refinetti Exp Physiol 88.3 The metabolic rate (i.e. the mean level of the rhythm of heat production) of the rats in this study was 1.99 W (or 5.1 W kg _1 ). This figure is comparable to previous measurements of resting metabolic rate of rats of similar body size under similar environmental conditions (Swift & Forbes, 1939; Bramante, 1961; Refinetti, 1989; Even et al. 2001). The robust circadian oscillation in metabolic heat production observed in this study (64 % of maximal robustness) confirms and expands previous observations in various mammalian species (Heusner, 1956; Morrison, 1968; Aschoff & Pohl, 1970; Fuller et al. 1985; Haim et al. 1988; Henken et al. 1993; Haim & Zisapel, 1995; Brown & Refinetti, 1996; Robinson & Fuller, 1999). The measurement of dry heat loss in a gradient-layer calorimeter revealed a very robust rhythm of heat loss in the rat (Q P = 87 % of maximal robustness). Few studies have previously measured heat production and heat loss simultaneously over several circadian cycles. In two studies, one in squirrel monkeys (Fuller et al. 1985) and one in pigeons (Graf, 1980), heat loss was estimated by changes in skin temperature while heat production was measured by indirect calorimetry. In two other studies, one in squirrel monkeys (Robinson & Fuller, 1999) and one in rats (Shido et al. 1986), heat loss was measured by direct calorimetry and heat production by indirect calorimetry. In all four of these studies, the gross daily patterns of oscillation of heat production, heat loss and body temperature were very similar, as in the present study (e.g. Fig. 1). However, there was no clear evidence of how the balance between heat production and heat loss generates the daily rhythm of body temperature. In the present study, the daily peak of the rhythm of body temperature was found to consistently follow the daily peak of the rhythm of heat production and to precede the daily peak of the rhythm of heat loss (Fig. 3), which is consistent with some of the previous studies. The novelty of the present study resides in the realization that the bulk of the oscillation in heat loss is synchronous with the oscillation in heat production, so that the oscillation in heat balance has a very small amplitude and is 180deg out of phase with the oscillation in body temperature (Fig. 4). The finding that the daily rhythm of heat balance is 180deg out of phase with the rhythm of body temperature might seem paradoxical, as the laws of thermodynamics require that changes in body temperature be the result of changes in heat balance. However, the circadian rhythm of body temperature is only one of many physiological processes that require metabolic activity. Daily oscillations in the levels of arousal, locomotor activity, diet-induced thermogenesis and other processes are responsible for most of the daily oscillation in heat production (and associated heat loss). The data from this study can be used for empirical thermodynamic computations. Assuming the specific heat of the body to be 3.47 kj kg _1 C _1 (c), and knowing that the body mass of the rats was 0.39 kg (m) and that their body temperature oscillated by 1.2 C each day (DT), the amount of heat required by the body temperature rhythm (Q = mcdt) can be estimated at 1.6 kj. The amount of heat required for homeothermy alone (i.e. raising body temperature from 24 to 37 C, DT = 13 C) is 17.6 kj, so that the heat associated with the circadian oscillation of body temperature is only 9 % of the heat associated with homeothermy. The total daily heat production is 172 kj (1.99 W over 24 h), so that the heat associated with the circadian oscillation of body temperature is less than 1 % of the daily energy budget. Because of the small magnitude of the fraction of heat balance associated with the body temperature rhythm, it is likely that the daily oscillation in heat balance has a very slow effect on body temperature. This could account for the 180deg phase difference between the rhythms of heat balance and body temperature observed in this study. The small positive balance would slowly raise body temperature during half of the day, and the small negative balance would lower body temperature during the other half of the day. Further studies, particularly in larger animals with greater thermal inertia, are necessary to confirm this hypothesis. Aschoff J (1970). Circadian rhythm of activity and of body temperature. In Physiological and Behavioral Temperature Regulation, ed. Hardy JD, Gagge AP & Stolwijk JAJ, pp Charles C. Thomas, Springfield, IL, USA. Aschoff J & Pohl H (1970). Rhythmic variations in energy metabolism. Fed Proc 29, Bolles RC, Duncan PM, Grossen NE & Matter CF (1968). Relationship between activity level and body temperature in the rat. Psychol Rep 23, Bramante PO (1961). Quantitation of oxygen consumption and spontaneous muscular activity of the rat. J Appl Physiol 16, Brown CM & Refinetti R (1996). Daily rhythms of heat production, body temperature, and locomotor activity in golden hamsters. J Therm Biol 21, Czeisler CA & Dijk DJ (2001). Human circadian physiology and sleep-wake regulation. In Circadian Clocks, ed. Takahashi JS, Turek FW & Moore RY, pp Kluwer, New York. Even PC, Rolland V, Roseau S, Bouthegourd JC & Tomé D (2001). Prediction of basal metabolism from organ size in the rat: relationship to strain, feeding, age, and obesity. Am J Physiol Regul Integr Comp Physiol 280, R Fuller CA, Sulzman FM & Moore-Ede MC (1985). Role of heat loss and heat production in generation of the circadian temperature rhythm of the squirrel monkey. Physiol Behav 34, Gander PH, Connell LJ & Graeber RC (1986). Masking of the circadian rhythms of heart rate and core temperature by restactivity cycle in man. J Biol Rhythms 1, Graf R (1980). Diurnal changes of thermoregulatory functions in pigeons. I. Effector mechanisms. Pflugers Arch 386, Haim A, Ellison GTH & Skinner JD (1988). Thermoregulatory circadian rhythms in the pouched mouse (Saccostomus campestris). Comp Biochem Physiol 91A,

7 Exp Physiol 88.3 Circadian rhythm of body temperature 429 Haim A & Zisapel N (1995). Oxygen consumption and body temperature rhythms in the golden spiny mouse: responses to changes in day length. Physiol Behav 58, Hays WL (1988). Statistics. Holt Rinehart & Winston, New York. Heldmaier G & Ruf T (1992). Body temperature and metabolic rate during natural hypothermia in endotherms. J Comp Physiol B 162, Heller HG & Hammel HT (1972). CNS control of body temperature during hibernation. Comp Biochem Physiol 41A, Henken AM, Brandsma HA, van der Hel W & Verstegen MWA (1993). Circadian rhythm in heat production of limited-fed growing pigs of several breeds kept at and below thermal neutrality. J Anim Sci 71, Heusner A (1956). Mise en évidence d une variation nycthémérale de la calorification indépendante du cycle de l activité chez le rat. C R Soc Biol Strasbourg 150, Honma K & Hiroshige T (1978). Simultaneous determination of circadian rhythms of locomotor activity and body temperature in the rat. Jpn J Physiol 28, Marotte H & Timbal J (1981). Circadian rhythm of temperature in man: comparative study with two experiment protocols. Chronobiologia 8, Moore RY & Leak RK (2001). Suprachiasmatic nucleus. In Circadian Clocks, ed. Takahashi JS, Turek FW & Moore RY, pp Kluwer, New York. Morrison SD (1968). The constancy of energy expended by rats on spontaneous activity, and the distribution of activity between feeding and non-feeding. J Physiol 197, Murray EJ, Williams HL & Lubin A (1958). Body temperature and psychological ratings during sleep deprivation. J Exp Psychol 56, Nelson W, Tong YL, Lee JK & Halberg F (1979). Methods for cosinor-rhythmometry. Chronobiologia 6, Rashotte ME, Basco PS & Henderson RP (1995). Daily cycles in body temperature, metabolic rate, and substrate utilization in pigeons: influence of amount and timing of food consumption. Physiol Behav 57, Refinetti R (1989). Body size and metabolic rate in the laboratory rat. Exp Biol 48, Refinetti R (1992). Non-parametric procedures for the determination of phase markers of circadian rhythms. Int J Biomed Comput 30, Refinetti R (1994). The contribution of locomotor activity to the generation of the daily rhythm of body temperature in golden hamsters. Physiol Behav 56, Refinetti R (1998). Homeostatic and circadian control of body temperature in the fat-tailed gerbil. Comp Biochem Physiol 119A, Refinetti R (1999). Relationship between the daily rhythms of locomotor activity and body temperature in eight mammalian species. Am J Physiol 277, R Refinetti R (2000). Circadian Physiology. CRC Press, Boca Raton, FL, USA. Refinetti R & Menaker M (1992). The circadian rhythm of body temperature. Physiol Behav 51, Robinson EL & Fuller CA (1999). Endogenous thermoregulatory rhythms of squirrel monkeys in thermoneutrality and cold. Am J Physiol 276, R Roussel B, Chouvet G & Debilly G (1976). Rythmes circadiens des températures internes et ambiance thermique chez le rat. Pflugers Arch 365, Schmidt-Nielsen K (1983). Animal Physiology: Adaptation and Environment. Cambridge University Press, New York. Shido O, Sugano Y & Nagasaka T (1986). Circadian change of heat loss in response to change in core temperature in rats. J Therm Biol 11, Shiromani PJ, Klemfuss H, Lucero S & Overstreet DH (1991). Diurnal rhythm of core body temperature is phase advanced in a rodent model of depression. Biol Psychiatry 29, Smith RE (1969). Circadian variations in human thermoregulatory responses. J Appl Physiol 26, Sokolove PG & Bushell WN (1978). The chi square periodogram: its utility for analysis of circadian rhythms. J Theor Biol 72, Song X, Körtner G & Geiser F (1995). Reduction of metabolic rate and thermoregulation during daily torpor. J Comp Physiol B 165, Stephenson R, Liao KS, Hamrahi H & Horner RL (2001). Circadian rhythms and sleep have additive effects on respiration in the rat. J Physiol 536, Swift RW & Forbes RM (1939). The heat production of the fasting rat in relation to the environmental temperature. J Nutr 18, Tanaka H, Yanase M, Kanosue K & Nakayama T (1990). Circadian variation of thermoregulatory responses during exercise in rats. Am J Physiol 258, R van Esseveldt LE, Lehman MN & Boer GJ (2000). The suprachiasmatic nucleus and the circadian time-keeping system revisited. Brain Res Brain Res Rev 33, Yoda T, Crawshaw LI, Yoshida K, Su L, Hosono T, Shido O, Sakurada S, Fukuda Y & Kanosue K (2000). Effects of food deprivation on daily changes in body temperature and behavioral thermoregulation in rats. Am J Physiol Regul Integr Comp Physiol 278, R Zimmer MB & Milsom WK (2001). Effects of changing ambient temperature on metabolic, heart, and ventilation rates during steady state hibernation in golden-mantled ground squirrels (Spermophilus lateralis). Physiol Biochem Zool 74, Acknowledgements This work was partially supported by a research grant from the US National Institutes of Health (MH-66826). Timothy Mosehauer s assistance with data collection is gratefully acknowledged.