During hibernation at ambient temperatures (Ta) above 0°C, rodents typically maintain body temperature (Tb) ∼1°C above Ta, reduce metabolic rate, and suspend or substantially reduce many physiological functions. We tested the extent to which the presence of an insulative pelage affects hibernation. Tb was recorded telemetrically in golden-mantled ground squirrels (Spermophilus lateralis) housed at a Ta of 5°C; food intake and body mass were measured at regular intervals throughout the hibernation season and after the terminal arousal. Animals were subjected to complete removal of the dorsal fur or a control procedure after they had been in hibernation for 3–4 wk. Shaved squirrels continued to hibernate with little or no change in minimum Tb, bout duration, duration of periodic normothermic bouts, and food intake during normothermia. Rates of rewarming from torpor were, however, significantly slower in shaved squirrels, and rates of body mass loss were significantly higher, indicating increased depletion of white adipose energy stores. An insulative pelage evidently conserves energy over the course of the hibernation season by decreasing body heat loss and reducing energy expenditure during periodic arousals from torpor and subsequent intervals of normothermia. This prolongs the hibernation season by several weeks, thereby eliminating the debilitating consequences associated with premature emergence from hibernation.
- body mass
- food intake
small mammals that contend with low ambient temperatures (Ta) and reduced food availability conserve energy by physiological and behavioral means (21, 34). During the autumn and winter months, the golden-mantled ground squirrel (Spermophilus lateralis) and other sciurid rodents conserve energy by reducing metabolism and maintaining body temperature (Tb) at ∼1°C above Ta (23, 59). Many physiological and cellular functions, including respiration, reproduction, cardiac function, digestion, renal metabolism, mitosis, RNA translation, and immune function, are suspended or substantially reduced during hibernation (1, 7, 18, 26, 45, 57, 61, 63). The metabolic rate of hibernating mammals is <6% of corresponding normothermic values, which ensures considerable energy savings (23, 34, 54). Golden-mantled ground squirrels consume little or no food during the hibernation season (e.g., see Ref. 44); they survive the winter primarily by deriving fuel from stored body fat (14, 16). Almost all of the decrease in body mass during hibernation reflects depletion of white adipose tissue stores (16, 28).
Golden-mantled ground squirrels, in common with all other mammalian hibernators, do not remain torpid continuously; they arouse repeatedly throughout the hibernation season and temporarily achieve normothermic Tb values (∼37°C) at intervals of ∼2–10 days, only to reenter hibernation less than a day later (22, 34, 48, 62). Up to 70% of a hibernating mammal's winter energy expenditure occurs during these periodic arousals (34, 58). The functional significance of repeated arousals from torpor remains unknown (5, 7, 11, 45, 60), but, as noted by Humphries et al. (27), “torpor arousals are unlikely to depend on a single limiting process… a suite of negative physiological consequences may necessitate periodic arousals and result in benefits.”
Mammalian pelages ameliorate the energetic challenges of winter. Fur insulation attenuates conductive heat loss (2, 33, 46, 51, 52), and seasonal increases in the density, thickness, and length of hairs significantly decrease thermal conductance in the winter (2, 10, 24, 39, 46, 51). Insulation provided by fur most likely plays a significant role in conserving energy during hibernation by retarding loss of physiologically produced heat during periodic arousals to normothermia (53) and reducing costs of maintaining normothermic Tb values during the several hours between successive torpor bouts.
The Siberian hamster (Phodopus sungorus) is the only species in which the influence of fur loss on heterothermy has been assessed. These animals display shallow daily torpor [minimum Tb (Tb min) ∼22°C] for up to 8 h/day, several times per week. The frequency of daily torpor, the rate of arousal from torpor, and the energy conserved by displaying torpor were all substantially reduced in shaved hamsters (31).
We surmised that the absence of an insulative pelage might be less limiting in hibernating ground squirrels than in Siberian hamsters because the former maintain a much lower Tb min during torpor; consequently, at a Ta of 5°C, the Ta to Tb gradient is a modest 1°C for hibernating squirrels but ∼17°C for torpid hamsters. Squirrels presumably lose much less heat to the environment during stable torpor than do hamsters. On the other hand, arousal from torpor presumably involves relatively greater energy expenditure in squirrels than hamsters, given that Tb increases by ∼30 and 15°C, respectively; hibernators may, therefore, rely on an insulative pelage to significantly decrease the energy expended during periodic arousals by reducing heat loss during rewarming from deep torpor (53). Given these considerations, in the absence of pelage insulation, hibernators may modify several characteristics of hibernation behavior to counteract the negative energy balance associated with fur loss; for example, furless hibernators may arouse less frequently from deep torpor or increase their food intake during the normothermic intervals between torpor bouts. Alternatively, furless squirrels could abandon hibernation altogether and resume ingesting large quantities of food to fuel increased energy expenditure in the cold.
The goal of the present study was to determine the extent to which energy savings conferred by hibernation are contingent on the presence of an insulative pelage. Specifically, we assessed 1) whether golden-mantled ground squirrels resume hibernation after sustaining fur removal, 2) how the lack of fur affects energy savings during hibernation, 3) the impact of fur loss on Tb min, torpor duration, and rates of entry into and arousal from torpor, and 4) the duration of the hibernation season in shaved squirrels.
MATERIALS AND METHODS
Golden-mantled ground squirrels born in the laboratory were housed from birth at a Ta of 23 ± 3°C in a light-dark cycle that provided 14 h light/day (lights on at 0800, Pacific Standard Time). Food (Simonsen rat chow, maintenance diet) and water were available ad libitum. At various times between December 2000 and July 2002, male and female squirrels 2–3 yr of age were transferred to a cold chamber (Ta of 5 ± 1°C) and provisioned with an amount of bedding adequate for sanitation but insufficient for constructing a full, well-insulated nest.
Telemetric recording of Tb.
Tb was recorded telemetrically with calibrated temperature-sensitive radiotransmitters (model VM-FH, ∼1.5 cm long and 1 cm wide; MiniMitter, Sunriver, OR) implanted intraperitoneally via a single midline incision in normothermic ground squirrels deeply anesthetized with ketamine (70 mg/kg body mass, administered ip). Transmitters were calibrated before surgical implantation using several predetermined temperature set points. After the incision was closed with sterile sutures, postsurgical analgesia was provided by adding acetaminophen plus codeine phosphate to the drinking water (1% solution) for 3 days. Animals were returned to the cold chamber 4–8 h after surgery. Tb data collected every 10 min via receiver boards under each animal's cage were transmitted to and stored in a computer for subsequent analysis by an automated computer program (Dataquest, St. Paul, MN).
Squirrels were anesthetized with isoflurane vapors during periodic normothermia, and fur was removed as described previously (30, 31). Briefly, animals were shaved close to the skin with animal clippers, after which a depilatory was applied for several minutes to remove remaining hair. Fur was removed from the entire dorsal integument except for the head and front and hindpaws, resulting in an estimated loss of >95% of the animal's dorsal pelage. Control animals were anesthetized and handled in a similar manner except that only an 8-cm2 patch of ventral fur was removed, estimated as <10% of the entire pelage. Removal of this amount of fur from the ventral surface of much smaller Siberian hamsters held at 5°C did not affect food intake (30) and is likely of little energetic consequence in torpid squirrels, which curl up into a ball and occlude the ventral surface during hibernation. Animals were allowed to recover at 22°C postoperatively for several hours before they were returned to the cold chamber. Hair regrowth was monitored and considered significant when a thin layer of fur covered >80% of the dorsal surface.
Food intake and body mass measurements.
Daily food intake, which correlates highly with daily energy expenditure in euthermic animals (25, 36 50), was measured throughout the experiment. Each squirrel was initially provided with ∼200 g food; pellets remaining in the food hopper and cage were collected and weighed (±0.1 g) each day. The total amount of food remaining was subtracted from the previous day's value to determine 24-h intake. Because food pellets absorb moisture at low Ta values, the food supply was kept at 5°C for ∼7 days before it was offered to the animals; at ∼7-day intervals, pellets in each cage were replaced with fresh food. Body mass was recorded weekly to the nearest ±0.1 g in nonhibernating (normothermic) squirrels and during each of the periodic arousals from torpor for hibernating squirrels. Food intake and body mass measures were recorded between 1100 and 1400.
The onset of a hibernation bout was defined as the time at which Tb decreased to <34°C and remained below this value for at least 12 consecutive hours. Hibernation characteristics measured included: 1) Tb min, defined as the lowest Tb attained during a torpor bout; 2) bout duration, defined as the number of hours Tb was <34°C; 3) the rate of rewarming from hibernation, calculated from the slopes of the Tb plot during the increase from Tb min to 34°C; and 4) the duration of the normothermic interval between successive hibernation bouts, measured as the number of hours Tb was >34°C before the next hibernation bout was initiated.
After ground squirrels had been acclimated to a Ta of 5 ± 1°C for several weeks, transmitters were implanted, and Tb was monitored for several weeks. Animals that displayed at least three hibernation bouts during this baseline period (n = 16) were assigned to one of two groups (designated as shaved and control groups, respectively), matched for mean body mass and hibernation bout duration. The entire dorsal integument was shaved in animals designated for complete dorsal fur removal (n = 10; 5 females and 5 males); control animals (n = 6; 4 females and 2 males) had an 8-cm2 patch of fur removed from the ventrum. Daily food intake, body mass, and Tb measurements resumed the following day and continued until several weeks after the termination of each squirrel's hibernation season.
None of the shaved animals regrew their fur during the hibernation season. To further investigate fur growth in this group, six of the shaved squirrels were monitored for regrowth for several months after hibernation ended. For purposes of comparison, two additional groups of nonhibernating squirrels housed at 5°C were shaved (n = 7; 3 females and 4 males) or subjected to the control procedure (n = 7; 4 females and 3 males); body mass, food intake, and the condition of the pelage of these nonhibernators were monitored for several months during the hibernating season. Experimental procedures were approved by the Animal Care and Use Committee, University of California, Berkeley.
Body mass, food intake, and torpor characteristics (Tb min, bout duration, arousal rate, etc.) for each group were determined for various intervals (preshave, postshave) and analyzed by repeated-measures ANOVA to determine changes over time (Statview 5.0; Abacus Concepts, Berkeley, CA). Where significant differences over time were established, control and shaved group means were compared within each time period and analyzed using ANOVA and Fisher's protected least significant difference test. One shaved hibernator died during the hibernation season; data from this animal were excluded from analyses of hibernation season duration and body mass increase after the terminal arousal. Differences were considered significant at P < 0.05 and are reported as such regardless of the actual P value. All values reported are means ± SE.
All hibernating animals that were shaved remained furless for the remainder of the hibernation season (∼10 wk), with no evidence of hair regrowth based on visual inspection. A subset of six of these animals monitored after the termination of hibernation first evidenced regrowth of moderate amounts of fur ∼22 wk later; at this time (∼31–34 wk after shaving), at least 80% of the dorsal surface was covered with at least a thin layer of hair. Hair regrowth was initiated first in the middle of the dorsum, resulting in denser and longer hairs than elsewhere in the regrown pelage. A group of nonhibernating squirrels that was shaved regrew comparable amounts of fur within a similar time interval (∼34 wk; P > 0.05).
Hibernation season duration.
Control and shaved squirrels manifested similar numbers of hibernation bouts before fur removal (3.5 ± 0.3 vs. 3.4 ± 0.2 bouts). After being shaved, allsquirrels displayed multiday torpor bouts interrupted by periodic arousals during which Tb values were >34°C for <24 h. Shaved and control animals continued to display hibernation postoperatively for 9.8 ± 1.5 and 13.5 ± 2.0 wk, respectively (P = 0.16).
Hibernation bout duration, Tb min, and normothermic intervals.
Before fur removal, individual hibernation bout durations did not differ between control and shaved groups (mean ∼4.5 ± 0.5 days for both groups; range from 1.6 to 7.9 days; Table 1). Both groups increased the duration of hibernation bouts by ∼50% during the three bouts before fur removal (from ∼3.4 days to ∼5.1 days). Mean bout length continued to increase from the beginning to the middle of the hibernation season (range 2.5–9.5 days; cf., Ref. 22), but it did not differ significantly between control and shaved groups after fur removal (P > 0.05; Table 1). Tb min values during hibernation ranged from 4.0 to 9.4°C; mean Tb min values did not change significantly over time for either group, nor did they differ between control and shaved groups pre- or postoperatively (P > 0.05; Table 1).
The duration of periodic normothermia (the number of hours that Tb was >34°C) ranged from 7.2 to 28.8 h; most values were in the 12- to 14-h range. Mean duration of normothermia (∼13 h) did not differ between control and shaved groups before or after fur removal; values did not change postoperatively for either group (Table 1).
Tb changes during entrance into and arousal from hibernation.
The rate of decline in Tb during entrance into hibernation was similar for control and shaved groups before hair removal (∼1.8°C/h) but increased by ∼19% to 2.1°C/h in hibernating squirrels subjected to complete dorsal fur removal (P < 0.05; Table 2).
Overall rewarming rates during periodic arousals from hibernation ranged from 7.4 to 15.0°C/h and did not differ between control and shaved animals preoperatively (∼11.2°C/h; Table 2). Fur removal did not significantly affect overall rewarming rates; the overall rate of rewarming during arousal decreased by ∼6.7% in shaved squirrels and increased by ∼7.4% in control animals, but the group difference fell short of statistical significance (P = 0.10; Table 2). The mean change from pre- to postoperative testing in the latency to rewarm from Tb min to 34°C was −8 min for control animals and +12 min for shaved squirrels (P < 0.05).
Rewarming rates were analyzed for three segments of the arousal process. Groups did not differ in rates of rewarming from Tb min to 15°C, but the Tb increase was significantly slower in shaved than control animals between 15 and 25°C and 25 and 34°C (P < 0.05; Table 2).
Food intake in hibernating and nonhibernating animals.
Nonhibernating squirrels housed at Ta of 5°C consumed 29.6 ± 2.0 g food/day (range 18–40 g). Removal of the dorsal pelage from one-half of the nonhibernating squirrels resulted in an ∼60% increase in daily food intake relative to preshave baseline values (mean: 43.6 ± 3.3 g; range 36–59 g; P < 0.05; Fig. 1); in contrast, daily food intake did not change postoperatively in control animals, in which 8 cm2 of ventral fur were removed (Fig. 1; P > 0.05). Food intake of nonhibernating shaved animals remained elevated throughout the >30-wk postoperative testing period and declined to baseline preshave levels after hair regrowth (Fig. 2).
Furred animals that hibernated ate very little food during the days in which they temporarily aroused from torpor (mean = 2.1 g and range = 0–6 g/normothermic interval). When corrected for the duration of the normothermic interval during which food consumption is possible (∼13 h), food intake was 13% of the 24-h values of nonhibernating furred squirrels. During intervals of normothermia, shaved hibernators did not eat more food than did furred controls (2.8 ± 0.3 vs. 2.3 ± 0.3 g; Fig. 3). Food intake of shaved and control animals did not differ throughout the hibernation season; only after hibernation ended did shaved animals increase their food consumption above that of furred controls (Fig. 4; P < 0.05). Groups' food intake differed significantly beginning on the 1st day after the terminal arousal (Fig. 4).
Body mass of hibernating and nonhibernating squirrels.
Shaved and furred hibernating groups were equated at the time of fur removal (207 ± 7 and 202 ± 7 g, respectively). The rate of body mass loss during the first several weeks of hibernation before shaving also did not differ between the two groups (5.8 ± 0.5 and 6.2 ± 0.4 g/wk, respectively; P > 0.05). Postoperatively, body mass of both groups continued to decrease for the remainder of the hibernation season but at a significantly accelerated rate in shaved squirrels; their rate of body mass decline during the first five postoperative weeks was more than double that of furred controls (5.9 vs. 2.6 g/wk; P < 0.05; Fig. 5). Mean body mass on the day of terminal arousal did not differ between the two groups (155 ± 5 and 160 ± 7 g, for the shaved and control groups, respectively; P > 0.05); note, however, that shaved animals terminated hibernation 3.7 wk in advance of control animals. Body mass values at the end of hibernation season represented ∼75–80% of the body mass at the time of fur removal. During the first 3 wk after the terminal arousal, body mass increased more slowly in shaved than control animals (19 vs. 31 g; P < 0.05; Fig. 6A); although absolute body masses of shaved squirrels were lower from weeks 3 to 5 after termination of hibernation, the rate of increase in body mass for the two groups was no longer significantly different during this time (P > 0.05; Fig. 6B).
The two groups of nonhibernating squirrels housed at 5°C weighed ∼237 ± 5 g before fur removal (range 199–297 g); body mass did not differ significantly between shaved and control nonhibernators during the first several weeks postoperatively (224 ± 7 and 232 ± 13 g, respectively; P > 0.05).
Ground squirrels that sustained near complete loss of the dorsal pelage continued to hibernate with frequencies that did not differ from those of control animals; Tb min values attained, the duration of individual hibernation bouts, the duration of the interbout normothermic intervals, and the amount of food consumed during episodes of normothermia were all unaffected by the loss of fur. A considerable portion of the energy saved by golden-mantled ground squirrels during the hibernation season was, however, contingent on an intact pelage. In shaved hibernators, rates of rewarming from torpor were reduced significantly, the time to complete the arousal process increased, and the rates of body mass loss, which reflect depletion of white adipose tissue (16), accelerated. The ground squirrel's dorsal fur evidently contributed to energy conservation over the course of the hibernation season and may thereby extend the hibernation season by several weeks. This conclusion must remain tentative pending comparison of control and shaved groups with equal proportions of males and females sustaining fur removal at the very onset of hibernation. The cost of premature termination of hibernation is high for animals that maintain elevated Tb values during the late winter months and may be life-threatening in the absence of food stores or above-ground sources of nutrition.
During the hibernation season, furred golden-mantled ground squirrels maintained a Tb min of ∼5°C for ∼5–6 days before periodically rewarming and sustaining normothermia for ∼13 h; these findings are in agreement with previous reports on this species (15, 35, 45, 48, 53). The lack of influence of fur loss on Tb min and hibernation bout duration parallels observations on daily torpor in shaved Siberian hamsters (31) and suggests that these features of the torpid state are tightly regulated in hibernators and daily heterotherms.
It is notable that the duration of the normothermic interval between successive hibernation bouts was unaffected by shaving. Shaved squirrels could save considerable amounts of energy by abbreviating the length of time they are normothermic because hibernating squirrels incur substantial heat loss during this interval (53). That they do not adopt this tactic suggests that the duration of normothermic intervals is resistant to deviations, and presumably functionally significant. Intermittent rewarming may restore or repair critical processes interdicted by extended periods at low Tb (7, 20, 34, 45). Norepinephrine-stimulated lipolysis of white and brown adipose tissue stores is compromised at typical hibernation Tb values (14). Arousal to normothermic Tb values may be a prerequisite for the lipolysis that provides energy to fuel the next hibernation bout (13, 14).
During immergence into hibernation, Tb decreased ∼17% faster in shaved than normal ground squirrels. The more rapid Tb decline in shaved animals is unlikely to conserve much energy, given that once metabolic heat production has been curtailed, entry into hibernation is largely a passive process (53). In contrast, arousal from hibernation is costly (34, 37, 55); as much as 70% of the winter energy budget is devoted to achieving and maintaining normothermia (11, 58, 59). The decreased rates of rewarming in shaved squirrels likely reflect increased conductive heat loss during rewarming. Shaved animals had significantly slower rates of rewarming during the intermediate (15–25°C) and high (25–35°C) phases of the arousal process but were unaffected while rewarming from Tb min to 15°C. A similar pattern was observed in shaved Siberian hamsters (31). Siberian hamsters normally require ∼1 h to complete the arousal process, whereas ground squirrels typically rewarm in ∼3 h. The longer duration of rewarming sustained by ground squirrels reflects a significantly lower Tb min, and may necessitate greater energy expenditure per arousal. The large Tb-to-Ta gradient during the later stages of the rewarming process is associated with the highest metabolic rates in golden-mantled ground squirrels and several other heterotherms (42, 49, 53); conductive heat loss is greatest, and insulation evidently is most critical, at this time.
The accelerated rate of body mass loss in shaved hibernating ground squirrels was not accompanied by compensatory increases in food intake. Food consumed during brief normothermic intervals likely remains largely undigested; at 5°C both intestinal absorptive capacity and secretory activity are severely compromised (6, 7). Fat-storing hibernators are aphagic or severely hypophagic during the hibernation season (8, 9, 17, 19, 20, 27, 34, 38, 41) and unresponsive to energy challenges that at other times of year promote increases in body mass (12). Thus food consumption increased markedly in nonhibernating squirrels after fur removal but remained unaffected in shaved squirrels undergoing repeated hibernation bouts. These findings are in contrast to those observed in furless Siberian hamsters that markedly increased their food intake during the normothermic intervals between successive torpor bouts (31); this may reflect differences in the mechanisms that control energy balance during hibernation and shallow daily torpor.
Body mass declined approximately two times as fast in shaved than furred hibernating animals during the first 5 wk after hair removal. Both groups terminated the hibernation season at a body mass of ∼155 g, but this occurred almost 4 wk earlier in shaved animals. Although males of some heterothermic species terminate hibernation earlier than females, this sex difference did not account for the shorter hibernation season in our shaved animals; both male and female shaved squirrels displayed a tendency to terminate hibernation 3–4 wk earlier than their furred same-sex counterparts. The ground squirrels in our study were shaved 2–4 wk after the first hibernation bout. We speculate that earlier fur loss, sustained at the outset of the hibernation season, would have resulted in substantially greater advances in the timing of the terminal arousal and an abbreviated hibernation season. In the field, early termination of hibernation exposes animals to above-ground Ta values far below 5°C (3, 4) and markedly increases the cost of thermoregulation. Morton and Sherman (40) documented the dire consequences of premature termination of hibernation in Belding's ground squirrels. An insulative pelage enables hibernators to retain sufficient white adipose tissue stores to extend the hibernation season and to meet postemergence energy demands.
Fur regrowth was surprisingly slow in ground squirrels; restoration of a moderate amount of hair was not achieved in less than ∼7 mo. By comparison, shaved Siberian hamsters regrew their dorsal hair within 3 wk (30). A circannual molt mechanism initiates hair replacement in the spring in ground squirrels (29, 56). The present data are in accord with these earlier observations: hair regrew beginning in the subjective spring, after animals had achieved their trough body mass.
Fur serves Siberian hamsters and golden-mantled ground squirrels in different ways. The dorsal pelage contributes substantially to energy conservation during torpor maintenance in hamsters housed at 5°C by diminishing heat loss, thereby mitigating the large Tb-to-Ta gradient during shallow torpor (31); in contrast, heat loss (and therefore insulation) is likely of lesser import for torpid golden-mantled ground squirrels than for Siberian hamsters because the former maintain a modest 1–3°C Tb-to-Ta gradient during hibernation. Fur does, however, confer greater benefits in golden-mantled ground squirrels during rewarming and normothermia; shaved ground squirrels must rewarm from a significantly lower Tb min and, unlike shaved hamsters, cannot increase food intake between successive torpor bouts to counter the additional energy expended on thermoregulation. The resulting depletion of energy stores in furless hibernators contributes to increased rates of body mass loss and subsequent premature termination of the hibernation season.
Ruben and Jones (47) speculate that a complete fur covering made its initial appearance in the earliest mammals and was preceded by the development of endothermy. Although fur likely evolved for several reasons unrelated to hibernation (e.g., insulation, physical protection, camouflage), its presence increases the short- and long-term energetic benefits derived from deep torpor; the development of a full pelage may, therefore, have been a prerequisite for or coincident with the development of mammalian heterothermy. Effectively hairless mammalian species are very few in number (43), and the fitness costs of fur loss in nature can be substantial; in California voles (Microtus californicus) studied in the field during winter, daily energy expenditure was increased and survival rate decreased after complete fur removal (32). The fitness costs of fur loss induced by fighting or disease in natural populations remain to be established, as does the effect of fur loss on hibernation in the wild.
This research was supported by National Institute of Child Health and Human Development Grant HD-14595.
We thank Chris Tuthill, Kim Pelz, Jin Ho Park, Matt Butler, and Dave Freeman for helpful technical assistance and John Dark and Ned Place for insightful comments on the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society