The Ontogeny of Metabolic Rate and Thermoregulatory Capabilities of Northern Fur Seal, Callorhinus Ursinus, Pups in Air and Water

Pinnipeds are unique among mammals in that they couple a primarily aquatic existence with terrestrial parturition. Because foraging occurs only at sea, the young must eventually enter the water to feed independently. Because the aquatic environment has a thermal conductivity 23 times greater than that of air (Schmidt-Nielsen, 1984), small mammals may be thermally stressed in water, primarily as a result of heat loss resulting from high surface area to volume ratios (Irving et al., 1935; Irving, 1973; Estes, 1989). Marine mammals may meet this challenge by maintaining elevated metabolic rates, increased insulation, lability of core body temperature, increased activity and specialized vascular complexes that act as heat exchangers (Scholander and Schevill, 1955; Hart and Irving, 1959; Kanwisher and Sundnes, 1966; Irving, 1969, 1973; Tarasoff and Fisher, 1970; Hampton et al., 1971; Hampton and Whittow, 1976; Kanwisher and Ridgeway, 1983). Juvenile mammals and birds generally display elevated metabolic rates relative to adults, complicating the interpretation of the relationship between metabolic rate and thermoregulation (Hissa, 1968; Dobler, 1976; Piekarzewska, 1977; Poczopko, 1979). There is some evidence that resting metabolic rates of marine mammals decline as animals mature (Davydov and Marakova, 1965; Matsuura and Whittow, 1973; Oritsland and Ronald, 1975; Miller and Irving, 1975; Ashwell-Erickson and Elsner, 1981; Thompson et al., 1987; Rea and Costa, 1992), supporting the claim that the historical use of young animals has confounded the interpretation of marine mammal metabolic rates (Lavigne et al., 1986).

One exception appears to be the work of Costa and Kooyman (1982), who examined the metabolic rates of sea otters Enhydra lutris. They demonstrated that elevated metabolic rates were present in adult animals. Mustelids in general seem to display elevated metabolic rates (Morrison et al., 1974; Iverson, 1972; Williams, 1986; Kruuk et al., 1994). That these relatively small animals entered the water recently, compared with pinnipeds, may partly explain these elevated metabolic rates (see Costa and Williams, 1999).

Elevated metabolic rates of juvenile animals are associated with growth of the young, termed the ‘work of growth’ by Brody (1945), reflecting heat produced in association with biochemical synthesis. In addition, the maturation of thermoregulatory capabilities must influence metabolic costs as pinniped young develop swimming skills and embark on a primarily aquatic existence. Physiological development may therefore constrain the age at which pinnipeds spend an increasing amount of time in the water, because energy must be allocated between maintenance, growth and activity. This pattern of energy allocation is diversified in the pinnipeds (the seals and sea lions). In the true seals, family Phocidae, pups grow quickly over a relatively short lactation period (Bowen et al., 1985; Iverson et al., 1992). Newborn fur seals and seal lions, family Otariidae, generally remain on land for periods of days to over a month before entering the water and typically couple this with an extended period of maternal dependence and low growth rates (Bowen, 1991; Renouf, 1991; Oftedal et al., 1987; Trillmich, 1986; see also Trillmich, 1996). Few studies have investigated the development of otariids from birth until weaning (Horning and Trillmich, 1997; Arnould et al., 1996a; Baker and Donohue, 2000) and none has examined the metabolic rates of individual animals over this period of development. Metabolic rates of very young (3–5 weeks old) California sea lion (Zalophus californianus) pups in air are similar to those reported for young mammals in general (Thompson et al., 1987), although the in-water metabolic rates of these pups were not determined.

Of the total population of northern fur seals Callorhinus ursinus, 72 % breed and pup on the Pribilof Islands, Alaska (Lander and Kajimura, 1982). Pups are born in a black natal pelage, penetrable by water, with no underfur. At approximately 1 month of age, the pups begin to molt the natal coat as adult-type underfur emerges with associated guard hairs (Scheffer, 1962). Pups first enter the water during the molting period, although diving behavior is minimal prior to weaning (Baker and Donohue, 2000). The aim of the present study was to examine the transition from a terrestrial to an aquatic existence by investigating the following variables in northern fur seal pups: (i) the ontogeny of resting metabolic rate (RMR) during pre-weaning development, (ii) how thermoregulatory constraints affect the relative cost of entering the water for pups at two stages of development at the water temperatures they are likely to encounter in the wild and (iii) body composition in relation to thermoregulatory capabilities.

This study was conducted on St Paul Island, Alaska, between 7 July and 25 October 1996. Newborn northern fur seal (Callorhinus ursinus L.) pups identified by a fresh umbilicus were marked within 1 day of birth with individual bleach marks (Clairol Corporation, Stamford, CT, USA; Le Boeuf and Peterson, 1969). To reduce potential variability due to maternal experience, all study pups were the offspring of females possessing white vibrissae, and thus mature and not likely to be primiparous (Scheffer, 1962). When possible, mothers of study pups were also marked with identifying bleach marks. Seventy-five pups were marked thus from 7 to 14 July, during the peak of pupping (Bartholomew and Hoel, 1953). As pups molted their natal coat (at approximately 2 months of age), bleach-marked pups were tagged on the trailing edge of each foreflipper with individually numbered plastic tags (AllFlex USA Inc., USA). The attendance pattern of mothers to marked pups was determined by observations of the rookery and surrounding area at least twice daily for the duration of the study. Metabolic measurements were conducted at two stages of development, before the post-natal molt (2–7 weeks of age) and post-molt (12–15 weeks of age).

Pups were captured when their mothers were at sea on regular foraging trips. To determine how long milk is retained in the stomach of pups, 54 individual pups were intubated 1–3 times at equivalent ages in 1995. All pups intubated during the first 3 days of a fast had some gastric milk present. No milk or other material was observed in the stomach of pups during normal or terminal fasts lasting 4–21 days. Metabolic rate measurements were obtained when pups had been fasting for a minimum of 3 days but no more than 5 days, as determined by female attendance patterns, to ensure that the pups were post-absorptive and at consistent fasting stages. Serum from blood samples collected before metabolic measurements was visually inspected for lipemia, which indicates whether recent lipid absorption has occurred. On the basis of fasting status and serum clarity, all pups were deemed post-absorptive during measurements. The metabolic rates of 18 pups in air pre-molt and of 16 pups in air post-molt were measured. Of these 16 post-molt pups, 12 had been sampled pre-molt. In-water metabolic rate determinations were obtained from 20 individual pups; 10 pre- and 10 post-molt. Each of these 20 pups was sampled at three water temperatures, as described below.

Metabolic rates were determined using indirect flow-through calorimetry. Pups were placed in a metabolic chamber that consisted of an opaque plastic tub (63 cm×46 cm×46 cm) fitted with a clear Plexiglas lid. An airtight seal between the chamber and lid was ensured by a foam-rubber gasket. Two small plastic fans were positioned on the upper portions of each end of the chamber to ensure the continuous mixing of gases. A temperature/humidity probe was affixed to the inside lid of the chamber from which values were recorded every 10 min. The chamber was indoors, with the intake port originating outdoors. Air temperature in the chamber was therefore influenced by ambient air temperature. A second identical tub, filled to within approximately 12 cm of the rim with fresh water, was used for in-water measurements. Water temperature was controlled by pumping water through a heat exchanger (Neslab, Portsmouth, NH, USA) regulating water temperature to within 0.5 °C.

Pups were held in portable mesh cages or ventilated wooden boxes until metabolic measurements were performed. Prior to placement in the chamber and immediately after each run, body temperature (°C) was recorded using a thermocouple thermometer (RET-1, Physiotemp, NJ, USA). The plastic-coated copper–constantan probe was inserted 16 cm into the rectum. Probes were calibrated before and after the study, with no drift recorded. Oxygen consumption was recorded for a minimum of 1 h in air and for 0.5 h at each water temperature. Metabolic rates were measured at air temperatures ranging from 19.3 to 24.5 °C and at three water temperatures, 5, 10 and 20 °C. Pups were removed from the chamber between water temperature measurements and allowed a minimum of 0.5 h between runs to ensure thermoneutrality at the onset of each run. The order in which the three water temperature measurements were made was randomized. For each pup, all in-water measurements were completed on the same day, and no pup was measured both in water and in air on the same day. Pups were generally captured, measurements completed and released the same day. Some pups were captured the day before measurements and held overnight on the rookery in ventilated boxes with mesh lids. No pups were held overnight indoors or for more than one night on the rookery. Some pups struggled initially when placed in the in-air chamber, but most became calm and some attempted to sleep minutes after placement. Pups that attempted to sleep were kept awake for a minimum of 1 h of measurements by gently tapping on the chamber, after which they were allowed to sleep for up to 0.5 h for additional measurements. Some pups also struggled when placed in the in-water chamber; however, most floated or swam calmly and some slept.

For all runs, pup activity was recorded every 10 min and whenever activity changed. Pups were scored as quiescent (awake, but lying or floating calmly), active (awake, moving about calmly), panicked (awake and agitated or struggling) or sleeping (lying or floating without movement, eyes closed).

Resting metabolic rates (RMRs) for in-air measurements were calculated from the lowest rate of oxygen consumption recorded during a consecutive 10 min period when the pup was in a quiescent state. In-water standard metabolic rates (SMRs) were calculated from the lowest rate of oxygen consumption recorded during a consecutive 10 min period at each temperature, regardless of activity. Because of the lag time of approximately 3 min before oxygen consumption was expressed as a steady value, the first 5 min of each trial was omitted before analysis.

Statistical analyses were performed using the software package SigmaStat (SigmaStat, Jandel Scientific, IRC, USA). Least-squares linear regression was used when indicated. With non-normal distributions (including percentage data) or those with unequal variances, a Mann–Whitney rank sum test or a Kruskal–Wallis one-way analysis of variance (ANOVA) on ranks was used. Repeated-measures analysis was employed when appropriate to account for the non-independence of longitudinal data. For repeated-measures distributions with unequal variances, a Friedman repeated-measures analysis of variance on ranks was employed. Means are reported with ±1 S.E.M. Statistical tests were considered significant at P⩽ 0.05.

The pups in the above study were one component of a larger investigation examining the energetics and activity of northern fur seals (Donohue, 1998; Baker and Donohue, 2000). To assess the effects of repeated capture, handling and isotopic dosing on study pups, additional pups were selected from the initial pool of marked pups for the present study and used as controls for experimental pups. Non-study pups included as mass controls were not handled, apart from brief marking and tagging captures, prior to obtaining measurements. Because any effects of repeated manipulation might be compounded over time and to increase the probability of detecting any such effects, the masses of control pups were measured in late October, very near the end of the study period, and compared with post-molt, late October masses of experimental pups. Because the female attendance of the mothers of all marked pups was monitored as detailed above, the percentage of time that mothers spent on shore with their pups was also evaluated for experimental and control pups during most of the lactation period. In addition, a concurrent investigation examining the development of swimming and diving in northern fur seals provided behavioral controls for study pups. The swimming and diving study included some pups initially marked for, as well as a subset of pups utilized in, the present study (Baker and Donohue, 2000). Pups in the swimming and diving study were fitted with small (3.3 cm×2.8 cm×2.3 cm, 24 g) time-wet recorders (TWRs) affixed to their dorsal pelage using marine epoxy resin. The TWRs recorded and stored the percentage of time that the pups spent in sea water. Data from TWRs were downloaded opportunistically when pups were captured for other measurements. As before, control pups had not been manipulated prior to obtaining control measurements apart from brief marking and tagging captures.

There were no significant differences between study and control pups in the time spent in sea water (31.5±1.1 % and 30.1±1.0 %, respectively; t=−0.89, d.f.=28, P>0.05; Baker and Donohue, 2000), late October mass (14.3±0.3 kg and 15.2±0.8 kg, respectively; t=0.39, d.f.=44, P>0.05) or maternal attendance during most of the lactation period (percentage of time mother spent on shore with pup, 27.0±0.9 % and 25.4±0.6 %, respectively; t=−1.64, d.f.=62, P>0.05).

During the study, sea surface temperatures measured near the study site ranged from 4.5 to 9.9 °C, while air temperatures ranged from −1.3 to 20.2 °C. Mean in-air RMRs of pups increased with mass both within measurement periods (pre- and post-molt) and over the duration of the study (Table 1; Fig. 1). Mean in-air RMR of 18 pre-molt pups (10 female, eight male) aged 4–7 weeks was 113±5 ml O₂ min⁻¹, a value approximately twice Kleiber’s (1961) predicted value of 45 ml O₂ min⁻¹ for adult animals of similar body size. Mean in-air RMRs of female (106±7 ml O₂ min⁻¹) and male (122±6 ml O₂ min⁻¹) pups were not significantly different (t=−1.64, d.f.=16, P>0.05) and were therefore combined for all subsequent analyses. Body mass of pre-molt pups ranged from 4.9 to 9.7 kg, with a mean of 7.2±0.3 kg. Pre-molt male pups (N=8) were significantly heavier than female pups (N=10; 8.2±0.4 kg and 6.8±0.3 kg, respectively; t=−4.22, d.f.=16, P<0.05). Pre-molt RMR scaled to M_b^0.85 rather than to M_b^0.75, as predicted by Kleiber (1961), or M_b^0.66 reported for intraspecific studies of metabolic rate (Heusner, 1982), but was not significantly different from either. Although scaling similarly, measured pre-molt RMRs were twice the level predicted by Kleiber (RMR=20.7M_b^0.85, r=0.78, d.f.=16, P<0.001).

When transformed to reflect oxygen consumption on a per kilogram basis, in-air metabolic rates of pups decreased with mass over the course of the study. The mean in-air mass-specific metabolic rate of pre-molt pups was 15.4±0.5 ml O₂ min⁻¹ kg⁻¹ (N=18). Pre-molt female (15.7±0.6 ml O₂ min⁻¹ kg⁻¹, N=10) and male (15.0±0.6 ml O₂ min⁻¹ kg⁻¹, N=8) pups did not have significantly different mass-specific metabolic rates (t=0.24, d.f.=16, P>0.05).

The mean in-air RMR of 16 post-molt pups was 160±4 ml O₂ min⁻¹ (Fig. 1). The mean in-air RMR of post-molt female pups was significantly lower than that of male pups (151±3 ml O₂ min⁻¹ and 173±5 ml O₂ min⁻¹, respectively; t=−3.34, d.f.=14, P<0.05). The mean body mass of post-molt pups was 13.3±0.5 kg (range 8.8–16.2 kg), with male pups being significantly heavier than females (14.5±0.7 kg and 12.3±0.5 kg respectively; t=−4.11, d.f.=14, P<0.05). Of the 16 pups measured post-molt, 12 had been measured pre-molt. The RMR of longitudinally sampled pups increased significantly from 116±7 ml O₂ min⁻¹ before molting to 159±5 ml O₂ min⁻¹ post-molt (repeated-measures ANOVA, F=36.30, d.f.=11, P<0.05). Post-molt RMR scaled to M_b^0.55, at a level four times that predicted by Kleiber (1961) for adult animals of similar body size (metabolic rate=39.2M_b^0.55, r=0.81, d.f.=14, P<0.001). The slopes and intercepts for pre- and post-molt log–log regressions of RMR on body mass did not differ significantly (P>0.10). The RMR for 22 pre- and post-molt pups combined (only one randomly selected measurement for pups sampled at both stages was used) scaled to M_b^0.70 at 2.6 times the level predicted by Kleiber (1961) for adult animals of similar body mass (metabolic rate=26.2M_b^0.70, r=0.95, d.f.=20, P<0.001; Fig. 2).

The mean in-air mass-specific metabolic rate of post-molt pups (N=16) was 12.2±0.3 ml O₂ min⁻¹ kg⁻¹. There was no significant difference between the mean in-air mass-specific metabolic rate of female (12.4±0.4 ml O₂ min⁻¹ kg⁻¹, N=9) and male (12.0±0.5 ml O₂ min⁻¹ kg⁻¹, N=7; t=0.71, d.f.=14, P>0.05) pups, and they were combined in subsequent analyses. For the animals sampled repeatedly, both pre- and post-molt, mean in-air mass-specific metabolic rate decreased significantly from 15.6±0.7 to 12.3±0.3 ml O₂ min⁻¹ kg⁻¹ (repeated-measures ANOVA, F=30.16, d.f.=11, P<0.05, N=12).

For all independent measurements pre- and post-molt, total body mass (M_b,t) was a better predictor of whole-body metabolic rate (RMR=49.5+8.32M_b,t, r²=0.89, d.f.=17, P<0.05) than lean body mass (M_b,l) (RMR=13.2+16.1M_b,l, r²=0.83, d.f.=17, P<0.05). No correlation was seen between metabolic rate and body lipid mass (r²=0.16, d.f.=17, P>0.05).

In-water standard metabolic rates for all pups measured at all water temperatures are presented in Table 2. In-water mean SMRs, as well mass-specific metabolic rates, decreased over the duration of the study, with mass-specific rates demonstrating the greatest relative decline (Figs 3, 4). Mean in-water SMRs of 10 pre-molt pups at 5, 10 and 20 °C were 311±25, 303±43 and 270±54 ml O₂ min⁻¹ respectively. Males (N=5) and females (N=5) did not have significantly different mean metabolic rates at any water temperature measured (two-way repeated-measures ANOVA, F=3.28, P=0.11) and values were combined for analysis. Mean pre-molt SMRs did not vary with water temperature with the exception of that measured at 5 °C, which was significantly greater than the 20 °C value (repeated-measures ANOVA, F=3.89, d.f.=18, P<0.05). SMRs of pre-molt pups in water were, on average, 2.6 times in-air RMRs (Fig. 3).

Mean mass-specific metabolic rates of pre-molt pups at 5, 10 and 20 °C were 37.2±1.9, 37.0±2.0 and 32.7±1.5 ml O₂ min⁻¹ kg⁻¹, respectively. No sex differences were detectable in the mass-specific metabolic rates of pre-molt pups at any water temperature measured (two-way repeated-measures ANOVA, F=0.24, P=0.64), and values for males and females were pooled for all subsequent analysis. Mean mass-specific metabolic rates of pre-molt pups did not differ significantly regardless of water temperature (repeated-measures ANOVA, F=3.41, d.f.=18, P>0.05). On average, in-water mass-specific metabolic rates of pre-molt pups were 2.8 times in-air mass-specific metabolic rates (Fig. 4).

Post-molt pups, which were on average approximately twice the mass of pre-molt pups, did not have significantly different SMRs from pre-molt pups at any water temperature measured (t-test, P>0.05 for all). Mean post-molt values at 5, 10 and 20 °C were 265±17, 281±25 and 235±30 ml O₂ min⁻¹, respectively. Males (N=5) and females (N=5) did not have significantly different mean SMRs at any water temperature measured (two-way repeated-measures ANOVA, F=0.18, P=0.69) and values were combined for subsequent analysis. Mean in-water SMRs of post-molt pups were on average 1.6 times in-air RMRs (Fig. 3).

When examined on a per kilogram basis, the mean in-water metabolic rates of post-molt pups at 5, 10 and 20 °C were 18.2±1.1, 19.4±1.7 and 16.1±1.7 ml O₂ min⁻¹ kg⁻¹, respectively. No sex differences were detectable in the mass-specific metabolic rates of post-molt pups at any water temperature measured (two-way repeated-measures ANOVA, F=3.28, P=0.10), and values for males and females were pooled for all subsequent analysis. Mass-specific metabolic rates were not significantly different from each other at any water temperature measured, and at 20 °C were not significantly different from in-air measurements (Tukey multiple pairwise comparison, P>0.05 for all; Fig. 4).

Sleeping pups in air exhibited the lowest metabolic rates measured in the study. Of the 18 pups measured pre-molt, in-air sleeping metabolic rates were obtained from 17 animals. The mean in-air whole-body metabolic rate of sleeping pre-molt pups was 95±4 ml O₂ min⁻¹, a 16 % decrease in the mean RMR of 113±5 ml O₂ min⁻¹ (repeated-measures ANOVA, F=46.70, d.f.=16, P<0.05). The depression of metabolic rate while sleeping in pre-molt pups was also significant on a mass-specific basis, declining by 15 % from 15.4±0.5 to 13.1±0.3 ml O₂ min⁻¹ kg⁻¹ (repeated-measures ANOVA, F=32.18, d.f.=16, P<0.05). Male (N=7) and female (N=10) pre-molt pups did not have significantly different sleeping metabolic rates either on a whole-body or a mass-specific basis (t-test, P>0.05 for both).

Of the 16 pups measured in-air post-molt, adequate sleeping records were obtained for 12 animals. After molting, pups continued to demonstrate a metabolic depression of 16 % while sleeping in air when examined on a whole-body or mass-specific basis. The mean whole-body metabolic rate of post-molt pups 161±5 ml O₂ min⁻¹ (N=12) decreased to 135±5 ml O₂ min⁻¹ while sleeping (repeated-measures ANOVA, F=100.70, d.f.=11, P<0.05). An equivalent metabolic depression was seen during sleep with respect to mass-specific metabolic rates (12.4±0.3 ml O₂ min⁻¹ kg⁻¹ awake, 10.3±0.2 ml O₂ min⁻¹ kg⁻¹ asleep; repeated-measures ANOVA, F=83.02, d.f.=11, P<0.05). Post-molt male (145±6 ml O₂ min⁻¹; N=5) pups had significantly greater whole-body sleeping metabolic rates than females (126±5 ml O₂ min⁻¹; N=7; t=−21.86, d.f.=10, P<0.05) in air. When examined on a mass-specific basis, however, there was no difference in the post-molt in-air sleeping metabolic rates of male and female pups (10.2±0.3 and 10.4±0.4 ml O₂ min⁻¹ kg⁻¹ respectively; t=0.46, d.f.=10, P>0.05).

In contrast to the metabolic depression observed for pups sleeping in air, the greatest metabolic rates measured during the study were those of pups sleeping in water. Five post-molt pups were observed, and metabolic measurements were obtained while they slept in water. Of these five, all but one animal was in a position commonly observed for northern fur seals resting or sleeping in water, the ‘jughandle’ position. In this position, one pectoral flipper is held between the rear flippers above water, leaving one pectoral flipper submerged. Pups observed sleeping in the water were of both sexes (females, N=2; males, N=3), and sleeping occurred at two of the three water temperatures (5 °C, N=2; 20 °C, N=3). The mean whole-body and mass-specific metabolic rates of post-molt pups sleeping in water in the ‘jughandle’ position were 366±18 ml O₂ min⁻¹ and 24.6±0.9 ml O₂ min⁻¹ kg⁻¹, respectively. These values were approximately twice the value for the single sleeping measurement (not in the ‘jughandle’ position) of a post-molt female in water at 20 °C, 184 ml O₂ min⁻¹ and 13.1 ml O₂ min⁻¹ kg⁻¹, respectively. The mean whole-body in-water metabolic rate measured for the four pups in the ‘jughandle’ position (asleep) was 18 % greater than the mean in-water (awake) measurement at equivalent temperatures for these same pups (366±18 and 310± 12 ml O₂ min⁻¹, respectively; repeated-measures ANOVA, F=9.46, d.f.=3, P=0.05). When evaluated as a proportion of body mass, the elevation of mean in-water metabolic rate while sleeping in the ‘jughandle’ position compared with awake in-water measurements was significant and reflected a 19 % increase from 20.8±0.8 to 24.6±0.9 ml O₂ min⁻¹ kg⁻¹ (repeated-measures ANOVA, F=11.47, d.f.=3, P<0.05). The mean in-water mass-specific metabolic rate of sleeping pups in the ‘jughandle’ position (24.6±0.9 ml O₂ min⁻¹ kg⁻¹) reflects a 146 % increase in metabolic rate compared with the measurements for these same post-molt pups sleeping in air, 10.0±0.3 ml O₂ min⁻¹ kg⁻¹ (repeated-measures ANOVA, F=457.43, d.f.=3, P<0.05). No pre-molt pups were observed sleeping in water.

Thermal conductance (C) in air was independent of temperature in pre-molt (r=−0.18, d.f.=18, P>0.05) and post-molt (r=0.57, d.f.=8, P>0.05) pups over the range of temperatures measured (19–25 °C). Mean body temperatures of pre-and post-molt pups in air were significantly different (37.8±0.1 and 37.4±0.1 °C, respectively; repeated-measures ANOVA, F=18.22, d.f.=9, P<0.05). Body temperature and mass-specific metabolic rate of both pre- and post-molt pups were independent of air temperatures (P>0.05 for all; Fig. 5). The mean thermal conductance at all air temperatures decreased significantly from 1.0±0.1 ml O₂ min⁻¹ kg⁻¹ °C⁻¹ pre-molt to 0.8±0 ml O₂ min⁻¹ kg⁻¹ °C⁻¹ post-molt (repeated-measures ANOVA, F=14.23, d.f.=9, P<0.05).

In-water thermal conductance increased approximately linearly with temperature in both pre-molt (r²=0.64, d.f.=25, P<0.05) and post-molt (r²=0.33, d.f.=28, P<0.05; Fig. 6) pups. The mean thermal conductances of pre-molt pups at 5, 10 and 20 °C water were 1.2±0.1, 1.4±0.1 and 1.9± 0.8 ml O₂ min⁻¹ kg⁻¹ °C⁻¹, respectively. The mean thermal conductances of post-molt pups at 5, 10 and 20 °C water were 0.6±0, 0.7±0.1 and 0.9±0.1 ml O₂ min⁻¹ kg⁻¹ °C⁻¹, respectively. For both pre- and post-molt pups, the thermal conductance at 5 °C was slightly less than that at 10 °C, but this difference was not significant (Tukey multiple pairwise comparison, P>0.05 for both). Thermal conductances at 5 and 10 °C were, however, significantly lower than the mean thermal conductance at 20 °C for both pre- and post-molt pups (Tukey multiple pairwise comparison, P<0.05 for both). At equivalent water temperatures, the mean thermal conductances of pre-molt pups were approximately twice that of post-molt pups, and this difference was significant (Tukey multiple pairwise comparison, P<0.05 for all). The body temperature of pre-molt pups was labile, drifting upwards significantly with increasing water temperature and decreasing in cooler water (r²=0.23, P<0.05) (Fig. 7). Post-molt pups maintained a mean body temperature of 37.4±0.1 °C regardless of experimental water temperature (r²=0.02, P>0.05; Fig. 7). Although the body temperature of pre-molt pups increased with water temperature, mass-specific metabolic rate did not differ significantly as a function of water temperature for either pre- or post-molt pups (P>0.05 for both; Fig. 4).

With the exception of post-molt pups sleeping in water in the ‘jughandle’ position, sleeping pups exhibited a decreased thermal conductance. Pre-molt pups in air had significantly lower mean thermal conductance when asleep (0.8±0 ml O₂ min⁻¹ kg⁻¹ °C⁻¹) than when awake and quiescent (1.0±0.1 ml O₂ min⁻¹ kg⁻¹ °C⁻¹; repeated-measures ANOVA, F=27.82, d.f.=14, P<0.05). Post-molt pups in air also demonstrated a significant decline in thermal conductance from 0.8±0 ml O₂ min⁻¹ kg⁻¹ °C⁻¹ awake to 0.6± 0 ml O₂ min⁻¹ kg⁻¹ °C⁻¹ asleep (repeated-measures ANOVA, F=20.86, d.f.=11, P<0.05). This reflects a 20 % and 25 % decrease in thermal conductance when asleep in air for pre- and post-molt pups respectively. The mean thermal conductance of sleeping post-molt pups in water did not differ from that of these same pups measured awake in water of equivalent temperature (repeated-measures ANOVA, F=6.78, d.f.=4, P>0.05).

The body composition of pre- and post-molt pups is shown in Fig. 8. Pre-molt pups averaged 64.2±0.5 % total body water (TBW). Transforming TBW, the lipid content of pre-molt pups averaged 14.7±0.9 % and lean tissue 85.3±1.0 %. With a mean body mass of 7.4±0.3 kg, mean lipid and lean (lipid-free) mass therefore accounted for 1.1±0.1 kg and 6.3±0.2 kg, respectively, in pre-molt pups. At 12–15 weeks of age, post-molt pups averaged 49.1±0.9 % TBW. Mean lipid and lean contents of post-molt pups were 33.2±1.2 % and 66.8±1.0 %, respectively. With a mean mass of 13.4±0.6 kg, post-molt pups averaged 4.4±0.2 kg of lipid and 8.9±0.4 kg of lean tissue.

There were no differences in body composition between female and male pups pre-molt with respect to %TBW, TBL or %TBL (t-test, P>0.05 for all). The percentage TBW of male (N=12) and female (N=13) pre-molt pups was 64.3±1.1 % and 64.0±0.6 %, respectively. This reflected a mean TBW of 4.1±0.1 l for females, which was significantly lower than the mean of 5.1±0.2 l for males (t=−4.51, d.f.=23, P<0.05). Mean TBLs of pre-molt females and males were 0.9±0.1 kg and 1.3±0.2 kg, respectively. This translates into mean values of 14.2±0.8 %TBL and 15.2±2.0 %TBL for pre-molt females and males, respectively.

Post-molt female (N=6) pups had less TBW in absolute terms (l) than male (N=7) pups (t=−2.75, d.f.=11, P<0.05), but this difference was not observed when mass was accounted for (Mann–Whitney rank sum test, T=95.00, P>0.05). The mean TBWs for female and male pups were 5.9±0.3 l and 7.1±0.3 l respectively. This represents a percentage TBW of 48.6±1.3 % for females and 49.6±1.4 % for males. Post-molt female pups had slightly lower absolute TBL (4.2±0.4 kg) than their male counterparts (4.7±0.4 kg), but this difference was not significant (t=−0.52, d.f.=11, P>0.05). As a percentage of body mass, post-molt female pups contained slightly more lipid (34.1±1.8 %) than males (32.5±1.9 %), but not significantly so (t=0.55, d.f.=11, P>0.05).

The overall finding that northern fur seal pups have in-air RMRs three times the level predicted by Kleiber (1975) for adult mammals of similar body size is consistent with the elevated metabolic rates of juvenile mammals. To account for body size, it is customary to compare energy measurements on the basis of metabolic size. RMR of pups scales to M_b^0.70, giving a scaling coefficient similar to the value of 0.66 often reported for intraspecific studies of metabolic rate (see Heusner, 1982). Very young, pre-molt pups have in-air RMRs twice the level predicted by Kleiber (1961) and yield a scaling coefficient of 0.85. The pre-molt scaling factor of 0.85 is similar to the scaling factor of 0.83 determined for suckling terrestrial neonates at peak lactation (Oftedal, 1981, 1984). Our values are similar to the single measurement of a neonatal northern fur seal made by Blix et al. (1979), who reported a value twice that predicted by Kleiber (1961). Older, post-molt pups have RMRs four times the level predicted by Kleiber (1975) for adult mammals with a scaling coefficient of 0.55. Older pups might be expected to have RMRs more similar to Kleiber’s predicted scaling coefficient of 0.75 than younger, smaller pups. In absolute terms, in-air whole-body RMR increased as pups gained mass and molted.

One would expect larger animals to have higher metabolic rates than smaller animals. Contrary to this expectation, in-water metabolic rates of post-molt pups were equivalent to those of pre-molt pups of half their mass at all water temperatures. This parity in energy expenditure between pre- and post-molt pups reflects the increased thermoregulatory abilities of post-molt pups. The independence of metabolic rate and water temperature in post-molt pups, coupled with stable body temperatures, suggests that post-molt pups were within their thermoneutral zone in water. Alternatively, the equivalence of metabolic rate at all water temperatures may have represented a maximum thermal response of pups to submergence in water. The latter seems plausible in pre-molt pups, which exhibited labile body temperatures in conjunction with relatively high metabolic rates. In any case, pre-molt pups experience greater relative costs when entering the water regardless of temperature than do larger post-molt pups. In-water pre-molt pups matched the energy expenditure of larger post-molt pups while still failing to maintain their body temperature (see Figs 3, 4).

The thermoregulatory capabilities of post-molt pups are further demonstrated by comparing the mass-specific metabolic rates of pre- and post-molt pups in water and in air (Fig. 4). Per kilogram of body mass, post-molt pups incur the same metabolic costs in water at 20 °C and in air at 20 °C. This observation, coupled with a small but significant increase in costs in water at 5 °C and 10 °C compared with in air at 20 °C, suggests that northern fur seal pups near weaning experience few additional metabolic costs when entering the water.

Costa and Gentry (1986) measured higher mass-specific metabolic rates in free-ranging female northern fur seal pups relative to males using isotope dilution methods. Male pups in the present study were larger than females and generally exhibited greater RMRs in air. However, we could detect no sex differences in mass-specific metabolic rates. Given our sample sizes, we could have detected 12 % and 15 % differences in mass-specific RMRs in air pre- and post-molt, respectively (at the 0.05 significance level with power 0.90). Observed sex differences in mean RMRs were only 4.6 % (pre-molt) and 3.6 % (post-molt). In contrast, low statistical power probably affected the outcome of comparisons of SMRs in water. While observed mean mass-specific SMRs of post-molt male pups in water were generally approximately 25 % lower than those of females, the measurements were quite variable among individuals and the power to detect those differences was low (0.28). If, indeed, gender differences in mass-specific SMRs exist, they may be due to the mass-specific savings of the males’ larger body size. Sex differences in metabolic rate during nutritional dependence have not been reported for Antarctic fur seals (Arctocephalus gazella) or northern elephant seals (Mirounga angustirostris) (Arnould et al., 1996a; Rea and Costa, 1992).

The independence of body temperature and mass-specific metabolic rate from ambient air temperature (Fig. 5) suggests that pups were within their thermoneutral zone during in-air measurements. This is consistent with the observed, though nonsignificant trend of decreasing thermal conductance with air temperature. An alternative hypothesis is that pups were at or above their upper critical temperature in air. Although Thompson et al. (1987) suggested a lower critical temperature of 19 °C for similarly aged, but larger, California sea lion pups, the propensity of our animals to sleep while in the in-air chamber may have been an attempt to reduce metabolic heat production below resting levels (see Hansen et al., 1995; Miller and Irving, 1975; Worthy, 1985). Pups sleeping in air reduced their metabolic expenditure by 16 %, less than the 24 % reduction seen in sleeping California sea lions (Matsuura and Whittow, 1973) and the 41 % reduction seen in sleeping harp (Phoca groenlandica) and grey (Halichoerus grypus) seal pups (Worthy, 1987). The tendency of northern fur seals to sleep while in a metabolic chamber was also noted by Miller (1978) for sub-adult males.

The linear increase in heat conductance with water temperature and the independence of mass-specific metabolic rate and water temperature suggest that both pre- and post-molt pups were within thermoneutrality during in-water measurements. That body temperature was constant in post-molt pups regardless of water temperature is consistent with the assumption of thermoneutrality. In contrast, the lability of body temperature with water temperature in pre-molt pups indicates that the ability of young pups to respond to changing ambient temperature is limited. The trend of increasing metabolic rate with decreasing water temperature in pre-molt pups may have been an effort to stabilize body temperature without the benefit of an insulating pelage (see below). This suggests that pre-molt pups may have been at or near their lower critical temperature in water between 10 and 20 °C, values higher than the 0–4 °C reported for newborn northern fur seals by Blix et al. (1979). Alternatively, the high metabolic rates measured, coupled with the lability of body temperature of pre-molt pups, may have represented a maximum thermal response. This is supported by a a 3.1 °C reduction in body temperature in a single 7.3 kg neonatal fur seal held in water at 10 °C for 30 min. This supports the hypothesis that very young pups, even those of large body size, do not have the thermoregulatory ability to tolerate submergence in water. The present data suggest that the upper critical temperature lies above 20 °C for northern fur seal pups in water.

Measurements from the five post-molt pups that were investigated both awake and asleep in water suggest that heat produced by waking activity may play a role in the overall aquatic thermal budgets of these animals. The increased metabolic rates and lack of depression in thermal conductance of the pups during in-water sleep may reflect a physiological response resulting from increased thermal stress as a result of inactivity. The characteristic ‘jughandle’ position commonly observed in northern fur seals sleeping at sea is probably a behavioral adaptation to decrease heat loss by holding hairless appendages above water. Miller (1978) did not detect a thermoneutral zone in water between 0 and 25 °C for sub-adult male northern fur seals, which appear lean relative to pups, particularly post-molt pups near weaning. The thinner insulating blubber layer in sub-adult males, coupled with small body size, may limit the oceanic regions available for the at-sea component of their life histories. Most northern fur seals remain at sea, apart from the reproductive season, and are thought to winter in the Pacific Ocean as far south as Southern California (Gentry and Holt, 1986). Some adult males remain in the Bering Sea or migrate only as far south as the Gulf of Alaska, presumably able to do so because of thermoregulatory benefits associated with large body size (Kenyon and Wilke, 1953; Kajimura, 1984). Recently, J. D. Baker and G. A. Antonelis (unpublished data) determined that newly weaned northern fur seal pups may remain in the Bering Sea for months after weaning. The present results show that newly weaned pups have the thermoregulatory capability to remain in the Bering Sea, as demonstrated by the decrease in thermal conductance in post-molt pups, partially accounted for by the insulative quality of the weanling fur. The pelage of the fur seal pup was first detailed by Scheffer (1962). Young pups, before the post-natal molt, possess a black coat with no underfur. At this stage, pups are rarely observed entering the water and are often observed shivering on land in inclement weather, particularly as rain parts the hair and soaks the skin.

By autumn, most pups have shed this pelage and produced a pelage characteristic of juveniles and adults. This coat has a silver sheen resulting from guard hairs that cover a thick underfur. The underfur of the molted pup can no longer be penetrated by water. Only well after the completion of the post-natal molt do the pups embark on their first migration. Although they begin to enter the water at 26 days old, they spend an increasing amount of time at sea as the molt progresses (Baker and Donohue, 2000). The decrease in thermal conductance of pups pre-to post-molt reflects an increase in the insulation of the pup. This insulation is the result of a combination of the adult-type pelage and the increased total body lipid (TBL) of post-molt pups. These two factors appear to offset metabolic costs associated with increased body mass, relative to pre-molt pups, as weaned pups embark on a primarily aquatic existence.

Northern fur seal pups store increasing amounts of blubber prior to the transition to an aquatic existence and nutritional independence. During the lactation period, pups increase TBL by 20 %. Our %TBL value for pre-molt pups aged 4–7 weeks (15 %) is slightly greater than the 9 % determined empirically by Arnould et al. (1996b) for four Antarctic fur seal pups 1–36 days old. However, three of the four pups measured by Arnould et al. (1996b) were 1–4 days old, and the fourth was 36 days old, and all were below 6 kg in mass. Consequently, we would expect younger, smaller pups to have lower TBL than older larger pups. Pre-molt northern fur seal %TBL (15 %) is within the range for 1-to 3-month-old California sea lion pups (1.5–19.3 %) of greater mass (7.8–15.0 kg) reported by Oftedal et al. (1987). Near-weaning northern fur seal %TBL values (33 %) are comparable with those of harbor seals (Phoca vitulina; 35 % TBL), but lower than those of phocids in general at weaning (40–50 % TBL; Bowen et al., 1992; Bryden and Stokes, 1969; Crocker, 1995; Ortiz et al., 1978; Oftedal et al., 1993; Rea and Costa, 1992; Reilly and Fedak, 1990; Stirling and McEwan, 1975; Tedman and Green, 1987; Worthy and Lavigne, 1983a,b). Much of this lipid is probably stored as a subcutaneous blubber layer. In one non-study post-molt pup that succumbed to a respiratory infection, the subcutaneous blubber layer was approximately 7 cm thick (J. D. Baker, personal observation). While this blubber layer serves both a thermoregulatory function and as an energy store, the relative importance of these roles is undefined. The thermoregulatory necessity of this blubber layer in addition to the pelage of the post-molt pups is anecdotally supported by the shivering observed in post-molt but emaciated pups abandoned by their mothers, which we have observed on the rookery. The present data do not permit the elucidation of the relative contributions of the adipose stores and pelage to thermoregulatory capabilities in post-molt pups. The combination of these factors, however, prepares post-molt pups for the challenges of their initial migration.

Lean mass in northern fur seal pups appears to be the primary contributor to total metabolic rate, with fat mass contributing little, if at all, to this value. The strong correlation between both total body mass (r=0.94) and lean body mass (r=0.91) with metabolic rate and the lack of a relationship between fat mass (r=0.40) and metabolic rate support this conclusion. Similarly, Rea and Costa (1992) demonstrated a correlation between lean body mass and metabolic rate in juvenile northern elephant seals (Mirounga angustirostris) but found no correlation between fat mass and metabolic rate. They suggest that fat mass indirectly affects metabolic rate by increasing thermoregulatory capability. Lean mass has also been strongly correlated with metabolic rate in human subjects (Cunningham, 1980; Roza and Shizgal, 1984; Elliot et al., 1989; Forbes and Brown, 1989). In additional studies with human subjects, lean body mass metabolic rates were similar in obese and non-obese individuals, suggesting that additional fat did not increase metabolic rate considerably (Felig et al., 1983; Ravussin et al., 1986; Jequier, 1989; Segal et al., 1989).

In conclusion, northern fur seal pups demonstrate elevated metabolic rates compared with predicted rates for adult animals of similar body size, but demonstrate rates comparable with those of terrestrial mammalian young. The decrease in mass-specific metabolic rates as pups mature reflects savings due to increasing body size and, as demonstrated by in-water measurements, increased thermoregulatory capabilities. Increased lipid stores and insulating post-molt pelage account for these increased capabilities. Consequently, post-molt pups enjoy a relatively lower metabolic cost when entering the water compared with pre-molt pups. In combination, the relative decline in thermal conductance and mass-specific metabolic rate pre-to post-molt suggests that younger seals may be less successful in maintaining homeostasis under changing environmental conditions at potentially higher energetic costs. These data provide a proximate explanation for the observation that young pre-molt northern fur seals are not observed entering the water.

Future studies might focus on further delineating the thermoneutral zone of northern fur seal pups and, ideally, juveniles and adults. Because the greatest mortality occurs at sea after weaning, the thermoregulatory limitations of the young at weaning may contribute to our understanding of the underlying causes of this mortality. In addition, measurements on compromised animals, such as abandoned or starving pups, would illuminate the metabolic consequences of food limitation resulting in poor body condition. Overall, a model of the energy budgets of these animals at sea could contribute to estimating prey requirements.

We thank J. Jolly and L. Pham for field assistance. We are indebted to J. Merculief and the City of St Paul for logistical support. The manuscript was greatly improved by the comments of B. J. Le Boeuf and T. Williams. The editorial comments of J. Kendig also improved the manuscript substantially. The National Marine Fisheries Service provided housing, laboratory facilities and transportation in the field. Funding was provided by grants from the National Science Foundation OPP-9500072 and Office of Naval Research N00014-94-1-1013 to D.P.C. Clairol corporation provided bleaching agents. All research was conducted with the approval of the Chancellor’s Animal Research Committee (UCSC).