INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ALTHOUGH THE INSULATION provided by fur is generally assumed to be adaptive, particularly for survival in the cold (10, 11, 35), few studies have directly tested this assumption in living animals. Most analyses have been restricted to the anatomic and physical characteristics of fur, but the adaptive role of fur in relation to behavior, energetics, and physiology has, for the most part, been neglected (22).

Fur is seemingly well suited for dealing with thermal challenges posed by winter conditions. In the cold, a high temperature gradient exists between a mammal's body temperature (T_b) and the ambient temperature (T_a). Because conductive heat loss to the environment is directly proportional to the temperature gradient, heat loss from animals is greater at lower values of T_a (4, 10). Mammals maintain relatively high constant T_b when challenged by low T_a by increasing endogenous heat production and decreasing conductive heat loss (1, 33, 35). Heightened heat generation is achieved by increased shivering and nonshivering thermogenesis, the latter primarily originating in brown adipose tissue (BAT; Refs. 14, 32, 37). Increased energy expenditure in the cold in the form of heat production is fueled by increased food intake; rodents exposed to cold challenges maintain their body mass while increasing their daily food intake (26). Heat loss in the cold is decreased behaviorally by modification of body posture, increased huddling, and nest building (13, 22, 38); physiologically by vasoconstriction that limits blood flow to the body surface; and anatomically by the insulation provided by fur (33, 35, 36). Fur insulation significantly reduces heat dissipation to the environment via conductive heat transfer (10, 16), and increases in the density, thickness, and length of the hairs of small mammal pelts significantly reduce thermal conductance (14, 29, 33, 35). Furthermore, many small mammals undergo photoperiod-dependent changes in pelage that result in denser, thicker fur in the winter (8, 25) and a corresponding reduction in heat loss compared with animals with thinner summer pelages (11, 33, 36). Such changes in fur properties were estimated to be more effective than changes in body size and body shape in minimizing heat loss at low T_a (11, 33, 35).

If fur is an adaptive trait, then it should directly affect correlates of fitness such as reproduction, energy expenditure, or mortality rates (20, 23). In harvest mice, Reithrodontomys megalotis, experimental removal of fur resulted in a 35% increase in metabolic oxygen consumption at three different T_as, although conclusions must be interpreted with caution as only one animal was tested (31). In the California vole, Microtus californicus, energy expenditure in the field was increased by 10% and survival rate was decreased as a consequence of complete fur removal (22). The degree to which fur conserved energy in that study may, however, be underestimated because of behavioral adjustments of the voles, which in winter tend to share nests with four or five other individuals (22); group huddling therefore likely decreased the effect of fur removal on thermogenesis and daily energy use in shaved voles.

The pelage of Siberian hamsters (Phodopus sungorus) changes seasonally under the proximate control of variations in day length (8, 18). The summer pelage is brownish-gray, of shorter length, and noticeably less dense than the whiter, longer, thicker fur of winter (8, 15, 26). Siberian hamsters encounter extremely low air temperatures during winter, with monthly above-ground averages around 24°C and daily temperatures as low as 45°C (40, 41). Siberian hamsters do not use deep hibernation as a winter survival tactic (42), and winter foraging is costly as hamsters must search for food on a frozen, bare surface that offers little protection from extreme cold and wind (40). Summer conditions are only slightly more favorable, with mean T_a in the low teens, and the thermal summer, defined by a mean daily T_a > 15°C, lasting for only 22 days (41). Because Siberian hamsters must contend with T_b-to-T_a gradients of 20°C in the summer and 60°C in the winter, the insulative properties of fur are likely to be of major significance for their overall energy balance. Although the seasonal changes of the Siberian hamster pelage are well documented (8, 15, 18, 26), the physiological impact of fur on daily energy expenditure has not been described in this species.

We sought to determine the contribution of fur to daily energy intake of Siberian hamsters housed under simulated winter and summer conditions. Because of seasonal changes in pelage insulation, fur removal might be predicted to have a greater impact on energy balance under winter than summer conditions. Furthermore, because insulation and heat loss vary with T_a, fur removal might have a greater impact in animals held at lower rather than higher T_a. This study therefore addressed the role of fur on daily food consumption of hamsters with summer or winter pelages housed at either moderate (23°C) or low (5°C) T_a. With the exception of the naked mole rat, adult rodents are never completely furless in nature, but animals can experience partial fur loss from intraspecific fighting, encounters with predators, or disease. Consequently, in addition to studying the impact of complete fur loss, we examined the effect of partial fur loss on daily energy intake. The minimum fur loss necessary to affect food consumption, as well as the correlation between amount of fur loss and energy intake, were also determined.

Fur thickness and density, as well as skin and body temperature, vary regionally on a mammal's body (2, 19, 21). The absence of insulation at hot spots (e.g., areas overlying BAT) may therefore result in greater heat loss than absence of fur at cooler body parts with smaller T_b-to-T_a gradients. This conjecture was evaluated by varying the location of partial fur removal on the dorsal surface: would fur loss above a hot spot such as the interscapular brown adipose tissue (iBAT) region influence daily energy intake to the same extent as fur removal from remote cooler regions? Because energy intake and expenditure are strong correlates of fitness and survival (6, 22, 44), measurement of daily food intake may be a valid indicator of the adaptiveness of mammalian fur.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals. Female Siberian hamsters (Phodopus sungorus) were born and maintained in a long-day photoperiod (LD; 16 h light/day, lights on at 0800 PST) at a T_a of 23 ± 2°C. Food (Purina rodent chow 5015) and water were available ad libitum. Animals were housed individually and provisioned with a reduced amount of wood shavings that was adequate for sanitation purposes but not for building a well-insulated nest.

Food intake and body weight measurements. Each hamster was provided with ~100 g of food at the beginning of the test period. Pellets remaining in the food hopper were weighed (±0.1 g) each day. Broken or cached food pellets in the cage were added to the food being weighed, and the total amount was subtracted from the previous day's value to determine daily food intake. Food measurements were obtained at the same time each day (see below for exact times). Because the food pellets absorb moisture at low T_a, all food was acclimated to each experimental condition for ~7 days before being offered to the hamsters; old food was replaced with fresh, acclimated food every 9-12 days. Body weights were recorded twice weekly to the nearest ±0.1 g in all experiments.

Fur removal. To remove fur, hamsters were anesthetized with ketamine (70 mg/kg body wt ip) and then shaved close to the skin using animal clippers, with care being taken not to damage the skin. The shaved area was then treated with a depilatory (Surgi-Cream) for 5 min to remove the remaining hairs. Animals that were rendered "furless" had fur removed from their entire body, except for the front of the face and the feet. Selective fur removal was achieved by shaving only a portion of the fur (patches) as measured by calipers (length and width of the exposed skin patch). As a general control procedure, some hamsters were anesthetized with ketamine and experimentally handled in a manner similar to those that were shaved, except that no fur was removed, nor any depilatory applied.

Pelage scoring. Pelages were rated on a scale from 1 (gray-brown, summer phenotype) to 4 (white, winter phenotype) as described by Duncan and Goldman (8). A short-day pelage was defined by a pelage score of 2.0.

Statistical analysis. Mean differences (±SE) in daily food intake and body weight were analyzed by ANOVA for repeated measures (Statview 4.1; Berkeley, CA) to test for the effect of treatment over time. Where differences over time existed, factorial ANOVA was used to determine on which days differences between groups were significant and when food intake returned to baseline levels. Factorial ANOVA was also used to compare differences in the magnitude of increase in food intake. Criterion for statistical significance was set at P < 0.05, using two-tailed tests.

Experiment 1: effect of complete fur removal on food intake under simulated summer and winter conditions. At 2-3 mo of age, animals were transferred to one of four conditions: short-day photoperiod (SD; 8 h light/day, lights on at 0800) at T_a 5°C (5.4 ± 1°C), SD at T_a 23°C, LD at T_a 5°C, and LD at T_a 23°C. All T_as were below the thermoneutral zone for this species, which ends at ~26°C (15). Hamsters were acclimated to these conditions for 12 wk, after which baseline daily food intake was recorded for seven consecutive days (days 7 through 1). The animals in each condition were then divided into two groups (designated shave group and control group) matched for mean body weight and pelage rating. There were 10 hamsters in each of the four LD groups and 8 hamsters in each of the SD groups, except for the shave group in SD at 5°C, which contained 10 animals. On day 0, animals in the shave groups were completely shaved, whereas those in the control groups were subjected to the control anesthetization and handling procedures. Complete fur removal resulted in an estimated loss of >90% of the entire animal's pelage. Daily food intake measurements resumed on day 1 and continued until day 35. Body weights were obtained twice per week, and pelage scores were obtained weekly for the remainder of the study. For the SD groups, only photoresponsive hamsters with a pelage score of 2.0 were included in the data analysis in this and subsequent experiments. Food intake, body weight, and pelage measurements were recorded between 0900 and 1000.

Experiment 2: effect of partial fur removal from the dorsal surface on food intake. At the end of experiment 1, the 20 hamsters from the LD-5°C group, along with 3 additional hamsters housed under identical conditions for the same amount of time, were further acclimated to LD-5°C for 3 wk. At the end of this interval, which was 21 wk after transfer to T_a 5°C, baseline daily food intake was measured for 7 days (day 7 through day 1); the animals were then divided into 4 groups matched for mean body weight and daily food intake: a control group that was not shaved (n = 6) and three patch groups that had either a small (4 cm²), medium (8 cm²), or large (16 cm²) patch of fur removed from the dorsal surface. The small, medium, and large patch groups resulted in an estimated loss of approximately 12-15, 30, and 60% of the total dorsal pelage, respectively. The small, medium, and large patch groups contained six, six, and five animals, respectively. Shaving and pseudoshaving procedures occurred on day 0. Daily food intake measurements were resumed on day 1 and continued until day 32. All food intake and body weight measurements were recorded between 0900 and 1000.

Experiment 3: effect of partial fur removal from above dorsal hot spots. At 1-2 mo of age, animals were transferred to a cold chamber maintained at T_a 5°C with an SD photoperiod (8 h light/day; lights on at 0800). Hamsters were acclimated to these conditions for 12 wk before daily food intake measurements were initiated. Baseline food intake was measured for seven consecutive days (day 7 through 1), after which the animals were divided into five groups matched for body weight, pelage rating, and food intake (n = 8 per group). One group of control hamsters was not shaved, and the other groups had either a small (4 cm²) or medium (8 cm²) patch of fur removed from either the caudal or rostral dorsal region; the latter area overlies both the iBAT pad and the dorsal cervical BAT pad. On day 0, animals were shaved or subjected to the control procedures. Food intake measurements resumed on day 1 and continued until day 30. Pelage scores were recorded weekly throughout the study; only data from animals with a pelage score of 2.0 were used. All measurements were recorded between 1030 and 1130.

Experiment 4: effect of multiple partial fur losses on food intake. At 1-2 mo of age, hamsters were transferred to a cold chamber maintained at 5°C in an SD photoperiod (8 h light/day; lights on at 0800). Animals were acclimated for 12 wk, after which baseline daily food intake was measured for seven consecutive days (days 7 through 1). At the end of this interval, the animals were divided into groups matched for body weight, pelage rating, and daily food intake (n = 10 per group) that included a control group which was not shaved; small-lower and small-upper patch groups, in which a 4 cm² patch of fur was removed from the caudal or rostral dorsal region, respectively; a medium patch group that had an 8-cm² patch of fur removed from the middorsal surface; and a split-medium group that had two 4-cm² patches of fur removed from the caudal and rostral dorsal regions (Fig. 1). The two patches of the last group were separated by 2.5-3.0 cm of furred tissue; the location of the two patches corresponded to that of the small-upper and small-lower patch groups. The combined area of the two small patches equaled the total area of the medium patch group (8 cm²). Hamsters in the patch groups were shaved on day 0; food intake measurements resumed on day 1 and continued until day 30. Pelage scores were recorded weekly throughout the study; only animals with a pelage score of 2.0 were used. All measurements were recorded between 1100 and 1200. 

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Schematic illustration of small (4 cm2), medium (8 cm2), or split-medium (4 cm2 × 2) patches of fur (shaded regions) removed from the dorsal surface of hamsters. A distance of 2.5-3.0 cm separated the 2 patches of the split-medium group.

Experiment 5: effect of ventral fur removal on food intake. Upon completion of experiment 4, 36 of the hamsters were acclimated to the SD-5°C condition for an additional 3 wk. At this point, 19 wk after transfer to short days, baseline food intakes were obtained for seven consecutive days (day 7 through day 1), after which three groups were constituted matched for body weight, pelage rating, and food intake (n = 12 per group): an unshaved control group, and dorsal-medium and ventral-medium patch groups, each of which had 8 cm² of fur removed from the middorsal or midventral surfaces, respectively. Animals in each patch group were shaved on day 0; fur removed from eight of the animals in each group was weighed (±0.001 g). Food intake measurements resumed on day 1 and continued until day 26. Pelage scores were recorded weekly; all animals displayed a pelage score of 2.0 throughout the experiment and were still maintaining the winter pelage when the experiment was terminated. All measurements were obtained between 1100 and 1200.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Body weight and pelage scores. In all five experiments, mean body weights and pelage scores did not change significantly over time for any group, and shaved and control groups housed in the same condition did not differ significantly on either measure (P > 0.05 for all comparisons). Mean body weights ranged from 36.4 ± 0.5 to 43.0 ± 0.5 g in LD groups and 28.5 ± 0.5 to 32.7 ± 0.5 g in SD groups. All LD animals had pelage scores of 1, whereas mean SD pelage scores ranged from 2.7 ± 0.1 to 3.1 ± 0.1.

Experiment 1. Daily food consumption did not differ between control and experimental groups before shaving (P > 0.05) but increased significantly in all groups subsequently subjected to complete fur removal (P < 0.05 for all conditions; Fig. 2). Food intake increased significantly within 2 days of fur removal and returned to control levels 20 days later except for the SD-23°C group, which returned to baseline levels after 17 days. The magnitude of increase in food intake during the first 2 wk after shaving was significantly greater in SD than LD groups at each T_a and was also greater at 5°C than at 23°C (P < 0.05; Fig. 3A). There was no statistical interaction between photoperiod and T_a in influencing food intake (P > 0.05); the two variables were additive in their effects.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Mean (±SE) daily food intake of hamsters before and after complete fur removal (designated by dashed vertical line). Animals were housed in either long-day photoperiod (LD; 16 h light/day) at 23°C ambient temperature (A; LD-23°C), LD-5°C (B), short-day photoperiod (SD; 8 h light/day)-23°C (C), or SD-5°C (D). , Control animals with no fur removed; , shaved animals. Food intake increased significantly over the first 2.5 wk after fur removal in all 4 conditions, as indicated by the horizontal bar above the abscissa.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Mean (±SE) %increase in daily food intake relative to baseline control values during the first 2 wk after shaving. A: increases after complete fur removal in 4 different conditions. B: increases in LD animals housed at 5°C that had either a small, medium, or large patch of fur removed from the dorsal surface. The data from the completely shaved group of B are replotted from the LD-5°C group of A. Groups with different letters differ significantly from each other (P < 0.05).

Experiment 2. Removal of a medium (8 cm²) or large (16 cm²) patch of fur resulted in significant increases in daily food intake in LD hamsters maintained at 5°C (P < 0.05; Fig. 4A), but removal of a small patch (4 cm²) in hamsters housed under the same conditions had no effect on food intake relative to that of unshaved controls (P > 0.05, Fig. 4B). Removal of a small patch of fur is therefore the most relevant control procedure for more extensive fur removal. The magnitude of the increase in food intake during the first 2 wk after shaving was similar for the medium and large patch groups (20.2 and 19.5%, respectively; P > 0.05) but was significantly less than that of completely shaved animals housed in the same conditions in experiment 1 (P < 0.05; Fig. 3B). Food intake increased in the medium patch and large patch groups within 2 days of shaving and returned to baseline control levels 18-19 days later.

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Mean (±SE) daily food intake of LD hamsters before and after removal of medium (8 cm2; ) or large (16 cm2; ) patches of fur (A) or small patches (4 cm2; ) of fur (B). On A and B,  represent control animals with no fur removed. The dashed vertical line designates when fur removal occurred. Food intake increased significantly over the first 2.5 wk after medium patch and large patch removal but did not change after small patch removal (convention as in Fig. 2).

Experiment 3. Food intake of hamsters housed in SD-5°C increased significantly during the first 15 days after removal of a medium (8 cm²) patch of fur, with no significant differences attributable to patch location relative to BAT location on the dorsal surface (P > 0.05; Fig. 5). Small patch (4 cm²) removal from either the caudal or rostral dorsal surface, the latter of which overlies the interscapular and cervical BAT pads, did not induce significant increases in food intake relative to that of unshaved controls (P > 0.05, data not illustrated). Food intake during the first 2 wk after shaving increased by 25.5 and 26.1% in the caudal-medium and rostral-medium patch groups, respectively; this was a significantly greater increase than that elicited by similar treatment of LD hamsters housed at 5°C (P < 0.05; Fig. 6). As in experiment 2, increases in food intake after medium patch removal were significantly lower than those manifested by completely shaved hamsters in the same condition (P < 0.05; Fig. 6).

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Mean (±SE) daily food intake of SD hamsters housed at 5°C before and after removal of a medium (8 cm2) patch of fur from above brown adipose tissue (BAT) () or non-BAT regions (). , Control animals with no fur removed. The dashed vertical line designates when fur removal occurred. Food intake increased significantly over the first 2.5 wk after medium patch removal regardless of location (convention as in Fig. 2).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Mean (±SE) %increase in daily food intake during the first 2 wk after removal of either medium BAT or medium non-BAT patches of fur. All groups were housed in SD-5°C except for the LD-medium group, whose data are replotted from Fig. 3B. Data for the SD-shaved group (complete fur removal) are also replotted from Fig. 3. Groups with different letters differ significantly.

Experiment 4. The removal of two small fur patches (each 4 cm²), neither of which alone affected food intake (experiment 3), produced a significant increase in food intake equivalent to that induced by the removal of a single medium (8 cm²) patch (P < 0.05; Fig. 7A). As in experiment 3, a single small patch of fur removed from either the caudal or rostral region of the dorsal skin surface did not affect food intake (P > 0.05; data not illustrated). The percent increase in food intake during the first 2 wk postshaving was slightly greater in the medium than the split-medium patch group (23.0 and 18.8%, respectively), but this difference was not statistically significant (P > 0.1). Both groups increased food intake within 2 days of fur removal and returned to control levels by day 16 for the medium group and day 13 for the split-medium group.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Mean (±SE) daily food intake of SD hamsters housed at 5°C before and after partial fur removal (designated by the dashed vertical lines). A: medium (8 cm2) patches of fur () or two small (4 cm2) patches of fur () were removed from the dorsal surface. B: medium (8 cm2) patches of fur were removed from either the dorsal () or ventral () surface. , Control animals with no fur removed. Food intake increased significantly over the first 2.5 wk after medium patch and split-medium patch removal from the dorsal surface but not after medium patch removal from the ventral surface (convention as in Fig. 2).

Experiment 5. Removal of a medium (8 cm²) patch of fur from the dorsal but not the ventral surface resulted in a significant increase in food intake relative to control values (P < 0.05; Fig. 7B). Increased food intake in the dorsal-medium group was manifested within 1 day of fur removal and averaged 23.2% during the first 2 wk after shaving, after which consumption returned to baseline control values. The mean weight of the fur removed from the dorsal surface (0.19 ± 0.02 g) was significantly greater (P < 0.001) than that of the fur removed from the ventral surface (0.07 ± 0.01 g).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Loss of fur significantly increased food consumption of Siberian hamsters housed in either long or short days at moderate (23°C) or low (5°C) T_a. Because food intake is a reliable index of daily energy expenditure (7, 34), and daily energy expenditure correlates with fitness (6, 44), these findings bear on the role of fur as an energetic adaptation. The significantly lower energy intake of animals that maintained a full pelage vs. those that were furless confirms that fur saves energy, in both summer and winter conditions. These results support the hypothesis that fur in mammals might have coevolved with endothermy as an energy-saving structure (5, 28).

Daily food intake of unshaved animals was greater at 5°C than 23°C; this confirms results of previous studies of this species (15, 26) and reflects an increased energy expenditure required to maintain thermoregulation in the cold. Daily food consumption of unshaved animals was also higher in long than short days, most likely due to the increased body weight and increased total metabolic rate of Siberian hamsters in long photoperiods (15, 39). Resting metabolic oxygen consumption is not, however, higher in long than short days when expressed per gram of body mass (15, 43).

Increases in food consumption after complete fur removal were greatest in short-day, winter-acclimated hamsters (44%) and least in long-day, summer-acclimated individuals (20%). These results are compatible with findings in other species. Shaved harvest mice held in short days at 18°C and 24°C increased oxygen metabolism by 35% relative to furred animals (31). Daily energy expenditure in shaved California voles studied in the field increased by 10% (22), a value that most likely underestimates the direct impact of fur removal on energy use: frequent huddling and availability of nest insulation probably countered the excess heat loss associated with furlessness. The lower values reported by Kenagy and Pearson (22) may also reflect the smaller amounts of fur removed (78% vs. 90% in the present study) and the measurement of daily energy expenditure 1-2 days after fur removal. We found that food intake was not always elevated the day after shaving, and maximum food consumption was usually recorded 4-7 days after fur removal. Kenagy and Pearson (22) indicate that two of their animals decreased energy consumption after shaving; for animals that increased food intake, the average increase of 29% was in close agreement with values obtained in our study, in which all animals increased food consumption.

Among genetically hairless mice, thermal conductance increased by 45% over that of furred mice (30), and BAT activity and nonshivering thermogenesis were 38% higher than in furred animals (12). The increased food intake of furless hamsters therefore most likely reflects a compensatory response to increased energy expenditure devoted to thermoregulation. With less insulation, shaved animals presumably endured greater heat loss and therefore expended more energy to maintain energy balance, which in turn necessitated an increase in food consumption. The increase in energy intake is not attributable to increases in body weight, which was unaffected by fur loss, and presumably not to increased white adipose tissue stores, which may also provide insulation. The effects of fur loss on locomotor activity are unknown; with reduced nesting material available, animals may increase activity in an effort to stay warm or become more sedentary to conserve energy.

The magnitude of daily food intake after fur removal varied with photoperiod and was ~10% greater in short- than long-day hamsters (44 vs. 32% at 5°C and 29 vs. 20% at 23°C). This suggests that the short-day pelage is a more effective insulator than the long-day pelage, as previously documented (15). Siberian hamster pelages are longer and thicker in winter than summer (15, 26); winter pelages, which are denser and longer, more effectively reduce conductive heat transfer than do thinner summer pelages (4, 11, 16, 36). In white-tailed deer, thermal resistance was significantly greater in the winter than summer pelage (21), and heat loss in lemmings was significantly greater through summer than winter fur at T_as ranging from 4°C to 24°C (33). On these grounds, fur loss would be expected to have a greater impact in short- than long-day animals maintained at moderate and low temperatures, as shown in our study.

The impact of fur removal on energy intake varied as a function of T_a and was 12-15% greater at 5°C than 23°C in both long and short photoperiods (Fig. 4). Fur provides more insulation at lower T_a than at higher T_a, which was expected given that fur insulation is directly proportional to the temperature gradient between T_a and T_b (3, 10, 16). Studies of Syrian hamsters (17) and laboratory mice (1) are consistent with this generalization in showing ~9% greater decreases in insulation after fur removal at lower than moderate T_as. In snow-shoe hares, foxes, and dogs, insulation increases ~1% per 1°C decrease in T_a (10). On this basis, insulation should be ~18% higher at 5°C than 23°C, which accords with the 12-15% differences in food intake of Siberian hamsters at the two T_as in this study. Additionally, metabolic oxygen consumption (15, 43) and daily food intake (26) of Siberian hamsters are both higher in the cold, suggesting that removal of insulation at lower T_as should have a greater impact on energy balance simply because animals are producing more heat. In the present study, both summer and winter pelages provided significantly greater insulation at lower than moderate T_as.

In most previous studies, metabolic rate or daily energy use was measured once, usually within 24-48 h after shaving. The present study documents changes in energy expenditure that persist for several weeks after fur removal. Food intake returned to preshaving baseline values within 17-20 days of complete fur removal, which corresponds to the recovery of a thin layer of fur approximately 13-16 days after shaving. Growth of the pelage of juvenile white-footed mice, Peromyscus leucopus, kept at low T_a was almost complete by 16 days of age, at which time it accounted for 66% of the total insulation achieved in adulthood (24). By 18-20 days of age, thermal conductance of juveniles was at a minimum and total insulation at a maximum. In two species of mice (P. leucopus and P. maniculatus), fur regrowth was completed 20-25 days after shaving (36), and in shaved voles released into the wild, significant fur covering was evident 3 wk later (22). The rate of fur regrowth correlated well with the onset of reduced food intake in the present study; a partial layer of fur was sufficient to return food intake levels to baseline values. We estimate that shaved animals regained ~75% of their full pelage at the time food consumption returned to baseline levels.

The rate of fur regrowth and the latency to reduce food intake to baseline values after fur removal were independent of day length and T_a. In deer mice and white-footed mice, fur regrowth was independent of T_as between 6°C and 32°C and began ~4 days after shaving; it was completed within 20-25 days (36). In Siberian hamsters the effects of day length on seasonal traits, including pelage, are transduced by variations in patterns of nocturnal melatonin secretion (9). Because the rates of fur regrowth were not affected by changes in day length, the increased duration of nocturnal melatonin secretion associated with short day lengths appears to influence the quality of Siberian hamster fur but not its rate of growth.

Removal of either a medium (8 cm²) or large (16 cm²) patch of fur caused equivalent increases in daily food intake in long-day hamsters at 5°C that were less marked than those effected by complete shaving in the same condition (20% vs. 32%). Conversely, loss of a small patch of fur (4 cm²) had no effect on food intake. The threshold area of fur loss necessary to elicit compensatory increases in energy intake therefore falls between 4 and 8 cm². We estimate that a 4-cm² patch represents approximately 12-15% of the total dorsal fur, whereas a medium patch represents approximately 25-30%. Thus, at 5°C, Siberian hamsters can lose 12-15% of their dorsal fur without incurring an energy deficit. As with complete fur removal, a medium patch of fur removal elicited greater increases in food intake in short- than long-day hamsters, further evidence that the short-day pelage more effectively insulated hamsters than did the long-day pelage. Siberian hamsters in the wild are exposed to T_as much lower than the 5°C cold challenges of the present experiments (40); 4-cm² patches that were ineffective at increasing food intake at 5°C might pose significant energetic challenges at T_as of 30°C or 40°C.

We found no evidence for regional differences in the insulative capacity of fur on the dorsal surface of Siberian hamsters. BAT is a significant source of nonshivering thermogenesis, and rodent skin temperature is at least 2°C higher above BAT regions than elsewhere (2, 19), but removal of a small patch of fur above BAT did not affect food consumption; removal of a medium patch of fur elicited similar increases in energy intake whether it was above BAT or elsewhere on the dorsal surface. Amount of fur loss sustained, rather than its location on the dorsum, influenced energy balance: medium patches always induced increases in food intake, and small patches never did, regardless of their dorsal position. This is surprising, considering that BAT mass and activity are 38-48% higher in hairless mice exposed to the cold (12). Although fur length and diameter do not vary on different regions of the dorsal surface (21, 27), skin thickness or vasoconstriction may vary regionally, and perhaps contribute to the absence of difference between BAT and non-BAT fur loss.

Fur loss from the ventral surface did not affect food intake, suggesting that the insulative role of ventral fur is minimal. The almost threefold greater weight of an 8-cm² patch of fur on the dorsal than ventral surface supports this conjecture. The ventral surface is in closer contact with the frozen ground in winter, but the dorsal surface has greater exposure to wind, which decreases the effective T_a considerably and increases conductive heat loss (4, 29). Findings in other species are compatible with ours: fur depth and length are greater in middorsal than midventral regions of deer mice (4), whereas the dorsal fur of cotton rats is also thicker than the ventral fur (27).

Although a single small patch of dorsal fur loss (4 cm²) did not influence energy intake, removal of two separated 4-cm² patches from the dorsum resulted in increased food intake not significantly different from that caused by removal of a continuous 8-cm² patch of fur (18.8 vs. 23.0%). Heat loss from multiple isolated regions apparently is summated and yields compensatory increases in food intake nearly equivalent to those of a much larger area of continuous fur loss. Rate of recovery of baseline levels of food intake was accelerated in animals that lost two small vs. one medium patch of fur, each totaling 8 cm² (13 days vs. 16 days). This may reflect differences in rates of fur regrowth, less total heat loss associated with the removal of two small patches (edge effect), or more effective redirection of blood flow away from small than medium patches of exposed skin.

In summary, energy intake of Siberian hamsters was altered for approximately 2-3 wk after complete fur removal or removal of ~30% of fur on the dorsal surface. Both winter and summer pelages contribute to energy savings at moderate T_a (23°C) as well as at cooler (5°C) T_a, with greatest savings observed in animals acclimated to short day lengths and lower temperatures. Additional research will be necessary to determine the role of fur on locomotor and foraging activity, as well as on thermoregulatory behaviors such as huddling, torpor, and nest building. Such studies would be most meaningful if conducted in winter-adapted animals.