Present address: Department of Agricultural & Environmental Science, University of Newcastle, Newcastle upon Tyne, NE1 7RU, UK.,
The energetics of Gentoo Penguins, Pygoscelis papua, during the breeding season
Article first published online: 19 APR 2002
Volume 16, Issue 2, pages 175–190, April 2002
How to Cite
Bevan, R. M., Butler, P. J., Woakes, A. J. and Boyd, I. L. (2002), The energetics of Gentoo Penguins, Pygoscelis papua, during the breeding season. Functional Ecology, 16: 175–190. doi: 10.1046/j.1365-2435.2002.00622.x
- Issue published online: 19 APR 2002
- Article first published online: 19 APR 2002
- Received 25 May 2000; revised 15 August 2001; accepted 31 August 2001
- heart rate;
- metabolic rate;
- 1The food consumption of an animal, both at the individual and the population level, is an essential component for assessing the impact of that animal on its ecosystem. As such, measurements of the energy requirements of marine top-predators are extremely valuable as they can be used to estimate these food requirements.
- 2The present study used heart rate to estimate the rate of energy expenditure of gentoo penguins during the breeding season. The average daily metabolic rate (ADMR) of penguins when one adult was necessarily present at the nest (incubating eggs or guarding small chicks; IG; 4·76 W kg−1) was significantly lower than that when both parents forage concurrently during the major period of chick growth (CR; 6·88 W kg−1).
- 3The ADMR of a bird was found to be dependent on a number of factors, including the day within the breeding season and the percentage time that the bird spent foraging during that day.
- 4When they were ashore, the estimated metabolic rate of IG birds (3·94 W kg−1) was significantly lower than that of CR birds (5·93 W kg−1). However, the estimated metabolic rates when the birds were at sea during these periods were essentially the same (8·58 W kg−1).
- 5The heart rate recorded when the penguins were submerged (128 beats min−1) was significantly higher than that recorded from resting animals when ashore (89 beats min−1). However, it was lower than that recorded from birds that were swimming in a water channel (177 beats min−1). This might indicate that, although primarily aerobic in nature, there was an anaerobic component to metabolism during diving. An alternative interpretation is that the metabolic requirement during diving was lower than when the birds were swimming with access to air.
- 6There was a significant decline in abdominal temperature, from 38·8 °C at the start of a diving bout to 36·2 °C at the end, which may indicate a reduction in overall metabolic rate during submersion. This in turn may explain the lowered heart rate.
- 7In the present study, we have shown that the metabolic rate of the gentoo penguin varies during the breeding season. The relatively constant metabolic rate of the birds when at sea could represent an upper physiological limit that the birds are unable to exceed. If so, it will only be possible for the birds to increase foraging effort by diving more frequently and/or for longer periods thus reducing their foraging efficiency (the energy gained during foraging vs. energy spent gaining that food). During years when food is scarce, this reduction in foraging efficiency may have a profound influence on the reproductive productivity of the gentoo penguin.
Penguins are one of the major consumers of the marine resources in the Antarctic region (Croxall 1984). As such, it is important that their energy expenditure and hence food consumption is known. This information can then be used in models of the energy flux within the Antarctic ecosystem (Huntley, Lopez & Karl 1991; Murphy et al. 1998; Croxall, Reid & Prince 1999) or be included within fisheries models (Croxall 1984). Unfortunately, these energy data can be difficult to obtain both at the population level, due to inaccuracies in population estimates, and at the individual level where the energy costs of different activities are unknown. The gentoo penguin, Pygoscelis papuaForster (1781), which feeds mainly over the continental shelf regions of sub-Antarctic islands, is potentially vulnerable to changes in local prey density. Accurate estimates of the daily energy requirements of this species across significant proportions of the life cycle are therefore particularly important.
The breeding biology and behaviour of the gentoo penguin at South Georgia have been investigated in a number of studies (Williams 1990; Williams & Rothery 1990; R.M. Bevan, P.J. Butler, A.J. Woakes & J.P. Croxall unpublished). Several studies have also estimated the energy requirements of breeding gentoo penguins using doubly labelled water (DLW) (Davis, Kooyman & Croxall 1983; Davis, Croxall & O’Connell 1989). These studies showed that there was no significant difference in the field metabolic rate (FMR) of birds foraging for small or large chicks and that the FMR of the adults remains relatively constant during the whole of the reproductive period (Davis et al. 1983, 1989). However, Gales et al. (1993), using a small data set (n = 3), suggest that there may be a substantial increase in metabolic rate (MR) during the crèche stage. This finding would agree with studies on other species that have shown that chick rearing is one of the most energetically costly periods, in terms of energy expenditure, for adult birds (Ricklefs 1983). This is because the adult birds have to provide increasingly more food for the chicks as they grow (Hamer & Thompson 1997; Granadeiro, Burns & Furness 1999). In order to provide sufficient food, the adults have to forage for longer and/or with increased effort and will consequently have a higher overall metabolic rate.
To determine the long-term, temporal changes in the rate of energy expenditure, a technique is required that has a longer period of monitoring than is possible with DLW. Heart rate ( fH) can be used to estimate the rate of energy expenditure, as it is linked to the aerobic metabolic rate of an animal as described by the Fick equation for the cardiovascular system (Butler 1993). Estimating energy expenditure from fH has been validated in the laboratory for a number of avian species (Bevan & Butler 1992; Nolet et al. 1992; Bevan et al. 1994) including the gentoo penguin (Bevan et al. 1995b). In the study on the gentoo penguin (Bevan et al. 1995b), the authors demonstrated that by continually monitoring heart rate, the average rate of oxygen consumption for 6 birds could be estimated to an accuracy of 1%. Furthermore, the monitoring period could be subdivided into shorter intervals so that the rate of energy expenditure required for specific activities could also be estimated. Electronic data loggers have been developed that enable fH to be recorded in free-ranging animals (Woakes, Butler & Bevan 1995) over extended periods of time (> 30 d) and with the required fine time resolution (15 s). These loggers have previously been used to estimate the rate of energy expenditure of free-ranging black-browed albatrosses, barnacle geese and fur seals (Bevan et al. 1995c; Butler, Woakes & Bishop 1998; Boyd et al. 1999). An additional benefit of recording heart rate is that the cardiovascular response of an animal to specific behaviours can also be monitored. This is particularly important when dealing with a diving animal, where changes in heart rate can indicate changes in the aerobic status of an animal (Butler & Jones 1997).
As such, the main aim of the present study was to record the heart rates of free-ranging gentoo penguins and to use these to estimate the rate of energy expenditure of the birds. Once estimated, a further aim was to determine any differences in energy expenditure between the reproductive phases. Other associated aims of the study were (a) to monitor the temporal changes in energy expenditure and (b) to determine the cardiac and thermoregulatory adjustments associated with diving in the gentoo penguin.
Materials and Methods
STUDY SITE AND BIRDS
This study was undertaken at the British Antarctic Survey (BAS) Station on Bird Island, South Georgia during the austral summers of 1991–92 and 1992–93. Gentoo penguins used in the study were breeding birds from the Johnson Cove colony on the NW of the island. This colony had been studied intensively by Williams (1990, 1991). Twenty-four penguins were caught and implanted with heart rate and temperature data loggers (HRTDL; Woakes et al. 1995); 13 in 1991–92 (mean mass ± SE = 6·4 ± 0·2 kg) and 11 in 1992–93 (mean mass = 6·7 ± 0·2 kg). Table 1 provides various details of the 14 birds from which data were obtained. The reproductive phases examined were (1) incubation/chick guarding (IG, both parents take it in turns to incubate one or two eggs or to remain with the chick/s, while the other one forages) and (2) chick rearing (CR; both parents forage while the chick/s form crèches). The implantation procedures that we used conformed to the stringent conditions of UK law [Animals (Scientific Procedures) Act (1986)].
|Mean ± SE||6·5 ± 0·2||6·1 ± 0·2||18·23 ± 1·62|
HEART RATE AND TEMPERATURE DATA LOGGERS
HRTDLs (55 mm × 24 mm × 6 mm, 24 g) were programmed to record fH over 15-s or 30-s periods and to take an instantaneous measurement of body temperature every min. At a sampling frequency of 15 s, the recording period was 18·2 d; at 30 s it was 30·3 d. The body of the data logger was encased in wax before it was encapsulated in biocompatible silicone rubber (for details see Woakes et al. 1995).
Experimental birds were manipulated in the following manner. At the start of each breeding season, incubating birds were marked on the back with a small spot of gloss paint. The colony was monitored to determine when marked birds exchanged with their partner and headed to sea. If a marked bird was ashore (incubating or guarding), it was monitored to determine when it exchanged nest duties with its partner. The bird was caught, using a hand held net, when it had left the vicinity of the nest. The bird was weighed with a spring balance (Pesola) and sexed using bill measurements (Williams 1990). It was then taken into a hide where it was anaesthetised with halothane in air (induction 3–4%, maintenance 1–2%). All surfaces of the hide used during the surgical procedures were covered in plastic-coated paper (Benchcote) and sterilised with chlorhexidine. Aseptic procedures were observed whenever possible. A HRTDL was sterilised in ethanol and then implanted into the abdomen of the penguins in the manner described by Stephenson, Butler & Woakes (1986) and Bevan et al. (1995c). Each logger incorporates a low power transmitter that emits a click on each QRS wave of the electrocardiogram. A receiver was used to confirm that the logger was recording the ECG. The loggers were previously tested and were shown to record heart rate accurately in the gentoo penguin (Bevan et al. 1997).
A distinctive pattern of small dots was painted on the breast feathers in order to identify the bird from a distance. A metal flipper tag was also used as a more permanent form of identification. On recovery from the anaesthesia (5–30 min after the halothane had been turned off), the birds were released into the colony. It was only possible to monitor the birds for a short period postanaesthesia, as most of the operations were performed during the evening when breeding pairs exchange incubation and guard duties. Low light levels therefore made it difficult to follow an individual bird. However, over the time that it was possible to monitor them, released birds would run to the edge of the colony where they would rest and preen. Birds were deemed to have recovered from the surgical procedures when body temperature and heart rates while ashore were not significantly different from those recorded at the end of the monitoring period. Using these criteria, all birds had recovered from the operating procedures within 2 days. The data from the first two days were excluded from any of the subsequent analyses.
The colony was monitored daily to determine the presence of implanted birds at the nest site. Nests were also monitored to determine the time of hatching and the number of chicks. Towards the end of the monitoring period (18–30 d depending on the sampling rate), observations were made of all birds returning from sea during the evening. Instrumented birds were recaptured before changing over with their partners and re-weighed. The loggers were removed, again under halothane anaesthesia, and the birds released. The heart rate and temperature data contained within the logger memory were downloaded onto a laptop computer. Initial analyses were performed using purpose written software that also visualised the data. Data from this initial analysis were then imported into a spreadsheet (MS Excel) and further analyses were performed.
ESTIMATING THE RATE OF ENERGY EXPENDITURE FROM HEART RATE
In Bevan et al. (1995b), the relationship between fH and the mass-specific rate of oxygen consumption (sV̇o2) was obtained from gentoo penguins that were walking on a treadmill or swimming in a water channel. Using the modified statistical technique described in Green et al. (2001), these original calibration data were reanalysed. This revised technique incorporates an error term that is due to the different animals being used. Using the original calibration data, a new predictive equation was derived using a general linear model (SPSS, SPSS Inc). In this procedure, the model used was: loge sV̇o2 = logefH + bird identity where loge sV̇o2 is the loge transformed mass-specific oxygen consumption and logefH is the loge transformed heart rate data from 9 treadmill calibration birds and 6 waterchannel birds (Bevan et al. 1995b). The term ‘bird identity’ is the identity of the individual used in the calibration procedure. ‘Bird identity’ was introduced as a random factor, as there were significant differences between the intercepts of the individual regression lines (see Bevan et al. 1995b). This random effect of the individual bird contributes an additional error component (d2) to the overall variance (sŶi)2 of the estimated rate of oxygen consumption (Zar 1984) and was quantified using the Variance Components procedure in SPSS. The standard error of the estimated rate of oxygen consumption was calculated by: (1) (Green et al. 2001) where (sŶi) is the standard deviation of the estimated values of sV̇o2, d2 is the variance due to the birds, e2 is the variance of the regression (the mean squares of the error of the regression), Xi is the value of heart rate that is being used to estimate metabolic rate, X̄ is the mean heart rate of all the calibration data, ∑x2 is the sum of squares of the heart rate values of all the calibration data , n1, n2, n3, n4 are the numbers of calibration animals, calibration data points, field animals and field data points, respectively. The derived equation using this procedure was: (2) sV̇o2 = 0·0163 × fH1·48 and the various components for estimating the SD were: d2 = 0·0885, e2 = 0·0311, ∑x2 = 12·67 (loge transformed data), X̄ = 4·77 (loge transformed data), n1 = 16, n2 = 154. Xi was calculated from the pooled heart rates from all the birds engaged in a particular activity (Green et al. 2001). It should be noted that n1, n2, n3 and n4 usually exceed the minimum identified by Green et al. (2001) for reducing the size of (sŶi).
DOUBLY LABELLED WATER
To compare the rates of energy expenditure estimated by the fH technique with those estimated by other techniques, six previously implanted penguins were also injected with DLW (3H2O and H218O, see Table 1 and Fig. 2). These birds were caught 11·9 ± 1·9 d after the implantation of the HRTDL. They were weighed, an initial blood sample for measuring background levels of the isotopes was taken and then 1·003 ± 0·089 mL 3H2O (200 µCi ml−1) and 1·658 ± 0·186 mL H218O (50·2% APE) were injected into the pectoral muscle as two separate injections. The isotopes were allowed to equilibrate with the water pool for 2 h (Davis et al. 1983) before a blood sample (~5 mL) was taken from the brachial vein and the birds were released. They were recaptured, on average 6·24 ± 0·79 d later, when a final blood sample (~5 mL) was taken again from the brachial vein. Blood samples were collected in heparinised syringes and centrifuged. Five to 10 samples of plasma (~10 µl each) were flame-sealed in haematocrit tubes for later analysis for 18O. The remaining plasma was frozen for tritium analysis. The plasma samples were analysed for 18O at the Scottish Universities Research and Reactor Centre, East Kilbride or at the Centruum voor Isotopen Onderzoek, Gröningen, Netherlands. Tritium analysis was performed on water distilled from the frozen plasma samples (Ortiz, Costa & Le Boeuf 1978). The rate of CO2 production was estimated using equation 36 of Lifson & McClintock (1966), as this was found in the validation study (Bevan et al. 1995b) to provide the best estimate of the rate of energy expenditure.
TIME DEPTH RECORDERS
The foraging behaviour of 6 of the penguins (mean mass = 6·5 ± 0·1 kg, see Table 1) was monitored using time depth recorders (TDR; MK IV, Wildlife Computers Inc., size = 100 mm × 35 mm × 16 mm, mass = 90 g). These were programmed to record depth every 5, 10 or 15 s and, via a saltwater-switch, to record the duration of periods when the TDR was dry. These dry periods were taken to be the times when the penguins were ashore. It was possible in some birds to confirm visually when they entered or left the sea and, in these cases, the TDRs accurately determined the times. It was therefore assumed that the TDRs accurately recorded the times at which all the birds equipped with TDRs entered or left the sea.
Data from the recovered TDRs were initially analysed using proprietary software (Dive Analysis, Wildlife Computers Inc.). A purpose written program also extracted the data from the downloaded files for later display with any simultaneously recorded heart rate data. For most birds, attaching the TDR took place several days after the birds had been implanted with HRTDLs to allow them to recover from the surgical procedures. Birds were again caught with a net after they had exchanged with their partner and were departing to sea. The birds were weighed, placed in a wooden restraint frame and a small bag was tied around the head to restrict their vision. The birds would then remain calm until they were released. The TDR was attached to the back feathers using quick-drying epoxy glue (RS Components). In order to reduce the drag caused by the externally attached instrument (Bannasch, Wilson & Culik 1995), it was shaped to increase hydrodynamic streamlining and placed as caudally as possible without infringing preening or movement of the tail feathers when standing. The TDRs were removed at the same time as the HRTDLs and the data were downloaded onto a computer using proprietary software.
During each season, food samples were obtained by water flushing (Wilson 1984) from nonexperimental gentoo penguins returning to the colony. Samples were obtained from 10 penguins once a week for four weeks during chick rearing (n = 40 for each season). Food samples were analysed for prey composition and it was assumed that the study birds would be feeding on similar prey and at a similar proportion. In the 1991–92 season, approximately 50% (by weight) of the food eaten by the gentoo penguins was krill and the rest was fish, whereas during the 1992–93 season the proportions were 84% krill and 16% fish. From the prey composition, it was estimated that in 1991–2 the energy equivalent per ml of O2 was 18·8 J and in 1992–3 it was 18·9 J (Crawford 1979; Clarke 1980; Davis et al. 1989). These values were used to convert the estimates of rates of oxygen consumption to rates of energy expenditure (W).
The salt-water switch in the TDRs indicated the time that each foraging trip started and ended. These times were used to calculate the duration of the foraging trip (‘at sea’) and the duration of the time spent ashore (‘ashore’). For those birds that did not have a TDR attached, the time of departure from and the arrival back at the colony was estimated from the abdominal temperature of the bird, as a very distinctive pattern was seen (Fig. 1). When at sea, there was an initial increase in abdominal temperature followed by a gradual decline that occurred when the bird was diving (cf. Butler et al. 1995; Handrich et al. 1997). Abdominal temperature quickly returned to preforaging levels when the bird arrived back at the colony. Using these temperature changes, it was possible to estimate the time of departure from and return to the colony and hence the duration of the foraging trips. To assess the accuracy of the estimates of foraging trip duration from the temperature data alone, the start and end times of several foraging trips of birds with both HRDLs and TDRs were compared with those obtained from TDR data. On average, the difference between the timings was 2·67 ± 9·12 min (mean ± SE) and was not significantly different from zero (paired t-test, t = 0·200, P = 0·422). For some trips, it was not possible to determine easily the start and end of a foraging trip from the temperature data and the data from these trips were not included in any analyses. Using these times, the periods ‘at sea’ and ‘ashore’ could be calculated. Where available, the pressure data from the TDR were used to subdivide the time while at sea into periods corresponding to the diving bouts (‘foraging’) and the trips to and from foraging areas (‘travel’). The time spent actively foraging was calculated as the time difference between the first and the last dives that were to depths > 10 m (Williams et al. 1992). For gentoo penguins, the time spent at sea is synonymous with foraging as only a very small proportion of their time is spent in the other activities (R.M. Bevan, P.J. Butler, A.J. Woakes & J.P. Croxall, unpublished).
Heart rate and abdominal temperature (Tab) data were analysed in conjunction with the depth data. The heart rates over specific periods of the dive were extracted as described in Bevan et al. (1997). These were the mean heart rate during submersion, the mean heart rate during the surface interval, the mean heart rate over the whole dive cycle and the minimum heart rate recorded during a dive. The reaction time of the device to changes in temperature was slow (time constant = 98 s), so only the Tab at the end of a dive was extracted.
There is considerable confusion in the literature between the use of the terms basal metabolic rate (BMR) and resting metabolic rate (RMR). In any wild situation, it is impossible to tell whether the conditions needed for measuring BMR (postabsorptive, inactive, adult birds under thermoneutral conditions) have been achieved. Hence, for the purposes of this paper, only RMR is used when describing measured data. Where BMR is mentioned, it has been calculated from the allometric formula of Ellis (1984): (3) BMR (kJ d−1) = 381·8 mass (kg)0·721 Likewise, in any comparisons with other studies, we have again estimated the BMR of the birds in the particular study using the formula of Ellis (1984).
Statistical testing was performed using the spreadsheet functions (MS Excel) or proprietary statistical software (SPSS, SPSS Inc or Minitab v12, Minitab Inc). Z-tests were used to compare pairs of estimates of metabolic rate. Wolff’s test for equality was used to compare more than two means. If any differences were detected using this procedure, multiple comparisons were made using z-tests and applying the appropriate Bonferroni correction. Multivariate analysis was performed using the general linear modelling function within SPSS or Minitab. For all statistical tests used, results were deemed significant if P < 0·05. The mean value of a variable is the overall mean of the average from each individual bird except for estimated metabolic rate which is derived from 2 and the SD of the estimate from 1. Where appropriate, the standard deviation (SD) of the mean is also provided. As the estimates of metabolic rate are derived from loge transformed data, the upper and lower SD are not equal. In these cases both the upper and lower SD are displayed.
Of the 24 gentoo penguins implanted with HRTDLs, 22 were recaptured. The other two birds deserted 7 d and 26 d after implantation. Heart rate data were not recovered from 10 birds due mainly to failure of the logger battery (n = 6). In the remaining 4 loggers that failed, temperature data had been recorded accurately but not heart rate, probably because of poor electrode placement. The data from these birds were not used in the present study. Simultaneous measurements of fH and Tab data were therefore successfully recorded from 12 free-ranging birds (Table 1). Time depth recorders were successfully deployed on six of these birds. Two birds were monitored during the transition from incubation to chick guarding and one during the transition from chick guarding to chick rearing. In the latter bird only the data from the CR period were used. Heart rate data were analysed from a total of 212 bird days. Figure 2 illustrates a 14-day period of fH recorded from a bird during incubation. Note that periods of human disturbance (injection of DLW and blood samples taken) are clearly discernible from the fH traces (points a and b).
HEART RATES DURING MONITORING PERIOD
The frequency distributions of fH when the birds were ashore differed between the reproductive phases (Fig. 3a). There was a clear peak in fH at 64 beats min−1 during IG (n = 8). During the CR period (n = 4), the heart rates of all the birds were clearly bimodal with peaks occurring at 80 beats min−1 and 128 beats min−1. Conversely, the frequency distributions of heart rates recorded during the ‘at sea’ phase were similar during the two reproductive phases, showing peaks at approximately 110 beats min−1 (Fig. 3b). The mean heart rates (± SEM) of individual birds over the recording periods ranged from 78 ± 0·1 beats min−1–136 ± 0·3 beats min−1. The overall mean heart rate of the free ranging penguins was 109·2 ± 5·6 beats min−1. Note that the mean heart rate for estimating metabolic rate differs from this, as it is derived from the pooled data from all the birds (Table 2).
|Bird||Heart ratea (beats min−1)||nb||MR (W kg−1)c||Prediction intervalsd|
The frequency distributions of heart rates ranged from 32 beats min−1–316 beats min−1 during IG and from 24 beats min−1–308 beats min−1 during CR. Over 99% of the heart rates monitored in birds in the wild were within the range of heart rates recorded during the calibrations (Bevan et al. 1995b).
FIELD METABOLIC RATE
The mean heart rate over the entire monitoring period was calculated for each bird and was used to estimate its rate of oxygen consumption (Table 2). The mean estimated rate of oxygen consumption of all the birds was 16·98 (–3·42, +4·28) ml min−1 kg−1 which is equivalent to a metabolic rate of 5·32 (–1·07, +1·34).
There was no significant difference between the mean heart rates of the two groups of birds when ‘ashore’ (z-test, zcrit= 2·50, z = 1·977, P > 0·05). The estimated metabolic rates during these periods were 3·94 (–0·29, + 0·32) W kg−1 and 5·93 (–0·43, + 0·47) W kg−1 during IG and CR, respectively (Fig. 4). Likewise, the estimated metabolic rates when ‘at sea’ were 8·58 (–0·98, + 1·10) W kg−1 and 8·58 (–1·32, + 1·56) W kg−1 during IG and CR, respectively, and again there was no significant difference between these (z-test, zcrit= 2·50, z = 0·002, P > 0·05) (Fig. 4). There was a significant difference between the estimated metabolic rates of the IG penguins when ashore compared with when they were at sea (z-test, zcrit = 2·50, z = 4·529, P < 0·05) but the same was not true for CR birds (z-test, zcrit= 2·50, z = 1·562, P > 0·05).
The average daily metabolic rate (ADMR) of each bird was calculated for the entire monitoring period by taking the mean of all the individual days. The ADMR of IG birds was 4·76 (– 0·58, + 0·67) W kg−1 which was significantly lower than that during chick rearing (6·88, –1·07, + 1·26 W kg−1; z-test, z = 1·973, P < 0·05) (Fig. 4).
The metabolic rate of the penguins changed with respect to the time of day (Fig. 5). Z-tests were used to compare the metabolic rates of the two groups at each separate hour. The estimated mean hourly metabolic rates throughout the evening and early morning (16 : 00–06 : 00) did not differ significantly between the two groups. However, the estimated mean hourly metabolic rate of the birds during CR was significantly higher (z-test) than those of IG birds at all hours between 07 : 00 and 15 : 00.
The resting metabolic rate (RMR) of the birds was determined when they were ashore during the hours of darkness and this was taken to be the period 00 : 00–04 : 00. The RMR of the IG birds was 4·14 (–0·51, + 0·58) W kg−1, which was not significantly different (z-test, zcrit= 2·50, z = 1·401, P > 0·05) from that of the CR birds (5·57, –0·86, + 0·101 W kg−1). For each individual bird, the minimum resting metabolic rate (RMRmin) was derived from the minimum value of hourly heart rate recorded over the entire monitoring period. Chick rearing birds had a RMRmin of 2·50 (–0·39, + 0·46) W kg−1 and this not was significantly higher (z-test, zcrit= 2·50, z = 1·77, P > 0·05) than that determined for IG birds (1·81, –0·23, +0·27 W kg−1). The maximum hourly metabolic rate was 9·76 (–1·20, + 1·36) W kg−1 and 13·58 (–2·12, + 2·51) W kg−1 during IG and CR, respectively. These maximum rates were attained while the birds were at sea and are equivalent to 5·4 × RMRmin for both IG and CR birds, respectively.
There was a tendency for ADMR to increase with time during the IG phase (Fig. 6) in most birds (n = 7) whereas there was no such correlation during the CR phase. Figure 6 shows how ADMR changed during the transition between IG and CR. Multivariate statistical analysis was used to determine the relationship between the metabolic rate and a number of variables, including individual, breeding phase, day within breeding season, sex, body mass and the percentage time spent at sea. When not a significant factor, each term was removed from the equation. There were significant differences between the metabolic rates of individual birds that masked any contribution from other factors. Consequently, the analysis was repeated for individual birds. For each individual bird during IG there was a very significant correlation between the estimated daily metabolic rate, the day within the breeding season and the percentage time that the bird spent foraging in that day (r2 = 0·88 ± 0·03). There was a lower, and for 2 of the birds an insignificant, correlation between these variables for CR birds (r2 = 0·60 ± 0·13).
DOUBLY LABELLED WATER
DLW experiments were carried out on six birds but only four of these were recaptured with sufficient isotope remaining in the water pool to enable metabolic rate to be estimated. Of these four, only two also had useable heart rate records. It was therefore impractical to make a direct comparison between the estimates of energy expenditure from the heart rate and DLW techniques within individual birds. All the DLW estimates were made on birds that were guarding chicks and these estimates of energy expenditure (8·26 ± 0·72 W kg−1; n = 4) were compared with those obtained by the heart rate technique but over the same period (5·35, –0·75, +0·84 W kg−1; n = 6). These estimates were significantly different (z-test, zcrit= 1·96, z = 2·194, P < 0·05).
METABOLIC RATE WHILE AT SEA
For the six birds with TDRs, it was possible to partition the time spent ‘at sea’ into the time spent travelling to and from the foraging area and the time spent actually foraging. The mean metabolic rates when travelling to the foraging areas, while foraging and when returning to the colony were 10·53 (–1·42, +1·64) W kg−1, 8·95 (–1·20, + 1·38) W kg−1 and 8·30 (–1·11, + 1·28) W kg−1, respectively. These rates were not significantly different from one another (Woolf’s test for equality, d.f. = 2, χ2 = 1·416, P > 0·05). It was also possible, in the present study, to estimate the cost to the birds of foraging with and without TDRs. The overall metabolic rate of birds carrying TDRs (n = 6) during the periods ‘at sea’ was 8·28 (–1·11, +1·28) W kg−1, which was not significantly different (z-test, zcrit= 1·96, z = 0·146, P > 0·05) from the MR of birds that did not have externally mounted TDRs (7·53, –1·00, +1·16 W kg−1, n = 6).
HEART RATE DURING DIVING
By synchronising the heart rate data with those from the TDR, it was possible to determine the heart rate at different phases of the dive cycle. This was only possible for a proportion of the dives made by the birds (1452 dives, 24% of total) as the sampling rate of the TDR was relatively long (see Boyd 1996; Bevan et al. 1997). It should therefore be noted that the interpretation of these data relates only to long dives (mean dive duration = 112 ± 6·6 s). The mean heart rate recorded from 6 TDR equipped birds during the various sections of a dive cycle were significantly different (repeated measures anova, F1,5 = 103·6, P < 0·001). The values (± SE) for the different phases were: 128 ± 14 beats min−1 during submersion, 150 ± 12 beats min−1 during the surface interval and 140 ± 11 beats min−1 over the entire dive cycle. The average minimum heart rate recorded during submersion was 90 ± 7 beats min−1. The mean heart rate of penguins while ashore was 89 ± 9 beats min−1. Heart rates of the penguins over the dive cycle were significantly higher than when they were at rest on land (Tukey’s HSD, P < 0·001) and the minimum heart rate recorded during submersion (Tukey’s HSD, P = 0·007). There was no significant difference between the heart rates over the dive cycle and the average heart rate during submersion (Tukey’s HSD, P = 0·295). The average heart rate during diving was significantly higher than the minimum heart rate during submersion (Tukey’s HSD, P = 0·012) but was not significantly different from the resting heart rate when on land (Tukey’s HSD, P = 0·099). There was no significant difference between the minimum heart rate recorded during submersion and the resting heart rate on land (Tukey’s HSD, P = 0·389). The heart rate recorded during the dive cycle was not significantly different from that recorded in penguins that were resting in water (144 ± 10 beats min−1; Tukey’s HSD, P = 0·998) but was lower than that recorded from penguins swimming in a water channel (177 ± 17 beats min−1, Bevan et al. 1995b; Tukey’s HSD, P = 0·023). Based on the heart rates measured over the dive cycle, the estimated metabolic rate over these periods was 7·66 (– 1·02, + 1·18) W kg−1.
The Tab of the penguins changed with respect to diving behaviour. While ashore, abdominal temperature averaged 38·4 ± 0·2 °C. After a dive bout, Tab was always depressed (see Fig. 1). The mean Tab at the start and end of a diving bout were 38·8 ± 0·1 °C and 36·2 ± 0·6 °C, respectively, while the mean minimum temperature attained during diving was 33·6 ± 1·5 °C. The difference between the abdominal temperature at the start and end of a diving bout was statistically significant (paired t-test, t5 = 3·861, P = 0·006). Abdominal temperature was found to decline throughout the dive bout (Fig. 7) and this decline in temperature was best described by a power curve for each individual bird. The mean body temperature over the whole of a dive bout of average duration (3·16 ± 00·42 hours) was 2·4 ± 0·6 °C lower than when the penguins were not diving.
EFFECT OF INSTRUMENTATION
In the present study, there was a recapture rate of over 90%. Of the two birds that did desert the nest, these did so 7 d and 26 d after implantation, which is the level of desertion that would have been expected in nonmanipulated birds (Williams 1990). All birds also resumed their nest duties after implantation, suggesting that the procedures used did not have an overall negative effect on the birds. Neither the implantation procedure of the HRDL or the attachment of the TDR, had any significant effect on the energy expenditure or behaviour (R.M. Bevan, P.J. Butler, A.J. Woakes & J.P. Croxall unpublished). In addition, although the chicks of the study birds were not followed to fledging, the birds implanted with HRDLs were raising at the end of the study a greater number of chicks than the rest of the colony. These data all suggest that instrumentation had no long-term effect on the birds.
HEART RATE DISTRIBUTIONS
The differences in the frequency distribution of heart rate seen in the ‘ashore’ phases were probably reflecting differences in behaviour exhibited during these phases. When incubating, the birds lay predominantly in a prone position, only occasionally standing to turn the eggs or when disturbed. Likewise, the birds were predominantly inactive when guarding small chicks. A low activity level over prolonged periods would be indicated by a low, relatively constant heart rate. They became more active during the exchange with the partner, after which they would often walk around the colony gathering stones to take to the nest before departing to sea. When the birds were rearing chicks, they were more active and, depending on the weather conditions, would often stand at the nest. Simply standing up from a prone position has been shown to cause an increase in heart rate and metabolic rate in a number of avian species (Bevan et al. 1994; Bevan et al. 1995b; Gabrielsen 1996; Hawkins, Butler & Woakes 1997). In addition, when the parents returned to the colony, they were chased by their chicks that were demanding to be fed. Having fed the chick(s), the adults departed back to sea during daylight hours but remained ashore and, presumably, inactive during the night. The overall frequency distributions therefore represent a composite of the frequency distributions of the individual behaviours.
ENERGY COSTS DURING INCUBATION AND BROODING
When a bird is at the nest, incubating or brooding, its energy needs should decrease due to a reduction in its activity level. It need only supply its maintenance metabolism, any thermoregulatory costs and the energy costs associated with maintaining the egg or chick at its optimum temperature. In order to replace the heat lost from the egg to the environment, the incubating adult may have to increase its overall metabolic rate above the resting level (Kendeigh 1963). However, most studies on large (> 0·5 kg) birds have indicated that FMR during incubation is similar to or even lower than that of non-incubating birds (Grant 1984; Gabrielsen, Mehlum & Nagy 1987; Pettit et al. 1988). On the other hand, some authors suggest that incubation is not a phase of low energy expenditure even in relatively large birds. Thompson, Furness & Monaghan (1998) found that in the kittiwake (Rissa tridactyla), the metabolic rate during the incubation phase was not significantly different from that estimated for birds that were raising chicks. These estimates, however, included the time at sea and might have reflected the high costs of foraging in this species rather than any high cost of incubation per se. Indeed, in a study by Gabrielsen et al. (1987), it was found that the metabolic rate of the kittiwake during incubation was only 1·9 times the RMR. In the present study, the energy expenditure of IG birds when ashore (3·94 W kg−1) was only 1·5 × estimated BMR and only 1·4 × RMR of birds in a metabolic chamber (Bevan et al. 1995b). However, the metabolic rate of the IG birds could be even lower since RMRmin during this period was equivalent to 73% of the predicted BMR. This indicates that other factors, not just the energy required by the egg, play a role in elevating the metabolic rate.
It is likely that the energy used above the BMR during IG will be due primarily to thermoregulation and the costs associated with nest attendance after exchanging with the partner. Adverse meteorological conditions such as low and high temperatures, wind or precipitation may increase the metabolic rate of the birds (Chappell et al. 1989). There will also be a certain amount of activity while at the nest (although not to the same extent as for CR birds). This activity, for example incubating birds are constantly interacting with conspecifics and with other species such as the Antarctic fur seal (Arctocephalus gazella), will again increase the metabolic rate. Croxall (1982), reviewing a number of penguin studies, estimated that the energy cost of incubation would be 1·4–1·7 × BMR (as calculated by Ellis 1984), the same as that found in the present study (1·5 × BMR). Davis et al. (1989) using DLW, estimated that the MR of incubating female gentoo penguins was between 3·4 W kg−1 and 4·2 W kg−1. Again, this is close to the estimates provided in the present study.
METABOLIC RATE DURING CHICK REARING
Parental energy demand is hypothesised to be greatest during the period of chick rearing as the adult birds have to forage to fulfil both their own energy requirements and those of their growing chicks (Ricklefs 1983). There are few studies that have measured the FMR of adult birds during the whole of the breeding cycle (Ricklefs, Roby & Williams 1986; Gabrielsen et al. 1987; Obst, Nagy & Ricklefs 1987; Gales & Green 1990; Montevecchi, Birt-Friesen & Cairns 1991; Gabrielsen 1994) but all of these studies have found that FMR is elevated when raising chicks compared with that during incubation. The MR of birds while at sea will be a composite of the energy requirements of various activities, e.g. maintenance metabolism, foraging and travel costs. Therefore, it might be expected that during the chick rearing period the mean ADMR over the whole of this period would be highest. However, earlier studies on the gentoo penguin failed to show any significant differences in the metabolic costs associated with the different reproductive phases (Davis et al. 1983; Davis et al. 1989; Gales et al. 1993). The present study shows that the energy costs of chick rearing are significantly higher, on average, than those of the earlier stages. This is almost certainly due to the increased time that the individual birds spend at sea (Fig. 6; R.M. Bevan, P.J. Butler, A.J. Woakes & J.P. Croxall unpublished) and expending energy at the higher ‘at sea’ rate.
TEMPORAL CHANGES IN METABOLISM
Although in the present study there were too few animals for statistical analyses of the temporal pattern, Fig. 6 shows the potential that heart rate has to show systematic changes in metabolic rate through time. The bird illustrated (Fig. 6) covered the transition from chick guarding to chick rearing (note that the IG data are used here simply for illustrative purposes, for all the analyses only the CR data were used from this bird). It can clearly be seen that the daily metabolic rate changes with time. These daily differences were primarily due to the length of time that the bird spent at sea in any particular day. There was also a tendency for the metabolic rate of the bird to increase gradually during chick rearing as the bird spent a greater proportion of time at sea.
The increase in ADMR to a peak just before the chicks became thermally independent is an interesting phenomenon. The present results indicate that the most energetically demanding phase of reproduction is when the adult birds were having to forage singly in order to supply their own energy requirements and the food demand of the growing chick(s) (Fig. 6). It is possible that the elevated level of energy expenditure is a factor that prompts the adult birds to leave the chicks. This enables both parents to forage at the same time, i.e. that the energy demands of the chick are too great for a single parent to satisfy at any one time. Culik (1994) came to the same conclusion from monitoring the amount of food that was brought back to the gentoo penguin chicks.
The increased costs in energy expenditure with time may also explain the apparent paradox between the present study and previous ones with regards to whether chick rearing is energetically expensive. In the previous studies, no indication was given as to when the actual measurements were made. If they had been made towards the end of chick guarding then they would have been measuring the increased MR associated with this period. Only by looking at the changes of MR with time is it possible to determine these gradual changes. This further emphasises the potential benefits of using the heart rate technique.
Energy costs of penguins while ‘at sea’
The metabolic rates of birds while ‘at sea’ did not differ significantly between the reproductive phases. This level of metabolism included all activities while the bird was at sea, such as the periods when the birds were travelling to and from the foraging sites and the periods of diving. It is possible that during their time at sea, they are unable to increase their metabolism, i.e. the birds were working at their maximum sustainable rate (Drent & Daan 1980). There is debate over what level this rate is in terms of multiples of BMR, but most evidence suggests a level that is 3–4 times BMR (Drent & Daan 1980; Daan, Maasman & Groenwold 1990; Gabrielsen 1994). Peterson, Nagy & Diamond (1990) reviewed a number of studies and found that the sustainable metabolic rate rarely ever increased above 5–6 × BMR and in fact most animals maintained a lower level of 2–3 × BMR. This is not to say that the animals are unable to increase their metabolic rate above this level, simply that if they do so for an extended period, then they will be in negative energy balance and will have to utilise energy stores. A limiting factor is likely to be the rate at which the digestive tract can process food (Diamond et al. 1986; Peterson et al. 1990).
In the present study, the estimated metabolic rate of birds while at sea was 3·3 × the estimated BMR (Ellis 1984). This was much lower than values reported in previous studies (Davis et al. 1989) where the ‘at sea’ metabolic rate was estimated to be 6·0 × BMR. The present estimate, however, is more consistent with the birds being able to maintain both their own requirements and those of the chicks. High ‘at sea’ metabolic rates were estimated for the chinstrap (5·6 × BMR; Moreno &Sanz 1996), Adélie (5·6 × BMR; Nagy & Obst 1992) and king penguins (4·7 × BMR; Kooyman et al. 1992). Results from penguins swimming in a water channel indicated that, when at their preferred swimming speed, the metabolic cost would be 2·9–4·3 × the BMR (Culik, Wilson & Bannasch 1994; Bevan et al. 1995b), which is a level of metabolic scope similar to that found in the present study. If these rates are more realistic, then the estimates of ‘at sea’ metabolic rate of the other studies may have been overestimated.
The studies with high FMR:BMR ratios all used the DLW technique and it may be that some of the assumptions regarding the technique have been violated (Nagy 1980). Although there have been many studies that have validated the DLW technique (see Speakman 1997), only those by Bevan, Speakman & Butler (1995a) and Boyd et al. (1995) have investigated animals that were exercising in water. These two studies showed that the DLW method might be less accurate when applied to animals exercising under these conditions. In a study of the Antarctic fur seal (Costa, Croxall & Duck 1989), DLW was used to estimate the metabolic rate of the seals when they were at sea. The average metabolic rate over the whole period that the fur seals spent at sea was estimated by the DLW technique to be as high as the maximum metabolic rate that could be maintained for only short periods by sea lions swimming in a water channel (Costa, Croxall & Duck 1989; Butler et al. 1992; Boyd et al. 1995). This strongly suggests that the DLW technique overestimated the metabolic rate of the fur seals at sea.
It should be pointed out that the discrepancy between the estimated ‘at sea’ metabolic rates of the different techniques might also be due to other factors such as prey availability and sea temperature. At the time that the present study was carried out, the gentoo penguins spent very little time travelling to their foraging area, indicating that food was readily available (R.M. Bevan, P.J. Butler, A.J. Woakes & J.P. Croxall unpublished). The birds may therefore have been working at a level below their maximum rate. In the studies that used DLW, the birds may have had to work much harder searching for food and then exploiting it or they may have been foraging in colder water which could, in part, explain the differences between the studies. These factors may explain the interstudy differences but do not explain the disparity between the estimates of metabolic rate derived by the two techniques used in the present study.
While it is possible that the DLW technique overestimates the MR of animals in the aquatic environment, there is also the possibility that the heart rate technique could be underestimating it. Penguins swimming in a swim channel had a measured MR of 11·7 W kg−1 (Bevan et al. 1995b). This is 1·4 times greater than the rate estimated by the fH method for free-ranging animals at sea (8·58 W kg−1 for both IG and CR birds). However, the metabolic rates of the birds in the swim channel might have been elevated due to the stresses of the experimental procedures in the swim channel (Culik et al. 1994). It has been shown that diving animals are not necessarily active during descent and ascent (Williams et al. 2000) but instead rely on changes in buoyancy. If the gentoo penguin also uses such a mechanism, then the energy cost of diving may be lower than that measured from swim channel experiments.
PHYSIOLOGY DURING DIVING
The mean heart rate of the diving penguins when they were submerged was 128 beats min−1. This is higher than that recorded from resting birds (89 beats min−1) but lower than that recorded from birds swimming in a water channel (177 beats min−1; Bevan et al. 1995b). This might indicate that, although the dives were primarily aerobic in nature, there might have been an anaerobic component to the overall diving metabolism (Carbone & Houston 1996). In the tufted duck, heart rate during diving is lower than that recorded during swimming at the surface, even when the rate of oxygen consumption is the same (Woakes & Butler 1983). The penguins may also be displaying a similar response. The fact that the heart rate over the entire dive cycle, at 140 beats min−1, was at the same level as that recorded in birds simply resting in water (and therefore that the metabolic rates are the same) suggests a number of other possible explanations. One is that the heat generated by activity was substituting some of the thermoregulatory costs when the animals were at sea which reduced their overall metabolic requirement (Butler & Jones 1982). Another potential explanation is that there was a reduction in the overall metabolism during diving (Handrich et al. 1997).
One of the intriguing problems of penguin biology is the apparent discrepancy between the length of dives that the penguins regularly make and the estimates of their aerobic dive limit (ADL). The calculated ADL (cADL) is the theoretical maximum duration of a dive that is fuelled entirely by aerobic metabolism (Butler & Jones 1997). It is calculated from the useable oxygen stores and the rate at which these stores are used during submersion. As has been illustrated above, the estimates of metabolic rates of penguins when at sea, and by implication diving metabolic rate, vary between 2 and 6 × BMR. Using estimates of diving metabolism of 4–6 × BMR and oxygen stores of 58 mL kg−1 (Butler & Jones 1997), the cADL would be approximately 80 s. If these assumptions are correct, then over 75% of the dives made by the gentoo penguin exceed the cADL (R.M. Bevan, P.J. Butler, A.J. Woakes & J.P. Croxall unpublished). Likewise, if the metabolic rate measured from gentoo penguins swimming in a water channel (10·9 W kg−1; Bevan et al. 1995b) is used then the cADL is only marginally increased to 100 s. The behavioural data suggest that the birds are not utilising anaerobic metabolism to any great extent (R.M. Bevan, P.J. Butler, A.J. Woakes & J.P. Croxall unpublished). This suggests that the metabolic rate during submersion must be reduced by some mechanism; be it physiological or behavioural. Butler & Woakes (1984) found that the rate of oxygen consumption of Humboldt penguins when diving underwater was not significantly different from that when resting in water (although it should be noted that the birds studied by these authors were not very active). The MR could be reduced through behavioural mechanisms, i.e. by simply not being active during descent and/or ascent (Williams et al. 2000). However, even if the metabolic rate of the gentoo penguin during diving is virtually the same as that when it is resting in water (7·1 W kg−1; Bevan et al. 1995b), which is supported by the estimated metabolic rate during long dives being 7·7 W kg−1, then almost 21% of the dives would still be longer than the cADL of 153 s.
A possible mechanism for reducing metabolism during diving is via a regulated regional hypothermia in which the penguins allow the temperature of certain parts of the body to fall or even enhance temperature loss. In the present study, abdominal temperature declined from 38·8 ± 0·1 °C at the start of a diving bout to 36·2 ± 0·6 °C at the end. This reduction in temperature is related to the diving behaviour of the birds (cf. Handrich et al. 1997; Ponganis et al. 2001) as the birds can maintain normothermia during nondiving periods at sea. The temperature reduction may help to reduce the overall metabolic rate during submersion and explain the relatively low heart rates recorded during submersion (Handrich et al. 1997; cf. Butler & Woakes 2001).
The potential benefit of allowing tissue temperature to fall can be illustrated as follows. If we assume a metabolic rate during submersion equivalent to that when resting in water of 7·1 W kg−1, an apparent Q10 of 3 (Heldmaier & Ruf 1992) and an average temperature drop of 2·4 °C of the whole body over a dive bout, then the cADL can be extended by 38 s and all dives would lie within the cADL. It is unlikely that all these assumptions are valid, however, the calculation does illustrate the potential benefit to the animal of allowing its body temperature to fall. Whether this reduction in regional temperature is actively controlled or simply due to the cooling of the outer body shell (Ponganis et al. 2001) is a matter to be resolved. In the context of energy saving, the issue is irrelevant as a drop in tissue temperature should be matched by a concomitant drop in metabolic rate regardless of its cause.
It has been suggested that the drop in core temperature could be a mechanism by which the penguin reduces the rate of digestion of the food contained in the stomach rather than a useful energy saving mechanism (Peters 1997). By reducing the rate of digestion, it can return to the chick with undigested food. While it is interesting to speculate how some penguins and other seabirds are able to spend extended periods at sea and not digest the food in the stomach, it is unlikely that this inhibition is responsible for the drop in body temperature or even that the fall in temperature plays any significant role in this proposed mechanism. There are a number of arguments to support this contention. Firstly, the decreases in body temperature also occur in birds that are incubating and are not having to return to the nest with any food. Secondly, the temperatures do not remain depressed but can increase between dives and between bouts. In fact, in the king and macaroni penguins body temperature returns to ‘normal’ levels while the birds are still at sea overnight (Handrich et al. 1997; J.A. Green, P.J. Butler, A.J. Woakes & I.L. Boyd, unpublished data). Thirdly, while it is true that there is a decrease in temperature and an increase in pH in the stomach (Peters 1997), in king penguins there is an even lower temperature in the abdominal cavity (Handrich et al. 1997) indicating that the temperature reduction is due to the cooling effect of the water surrounding the animal rather than to ingestion events (Culik et al. 1996; Handrich et al. 1997; Ponganis et al. 2001).
In the present study, we have used heart rate to determine the metabolic rate of the gentoo penguin and to study its temporal adjustment during the various phases of the breeding season. The metabolic rate of free-ranging penguins while at sea is, it would appear, independent of breeding phase and may represent the upper limit to sustainable metabolism in these birds. If so, it will only be possible for the birds to increase foraging effort by diving more frequently and/or for longer periods. During years when food is scarce due to natural causes or, potentially, to overfishing, this apparent inability of the birds to increase their metabolic rate could be a severe limiting factor to their reproductive productivity.
The authors would like to thank all the personnel at the British Antarctic Survey Base on Bird Island who helped with this study. The work was funded by NERC grant GR3/7508.
- 1995) Hydrodynamic aspects of design and attachment of a back-mounted device in penguins. Journal of Experimental Biology 16, 83–96. , & (
- 1997) Heart rates and abdominal temperatures of free-ranging South Georgian shags, Phalacrocorax georgianus. Journal of Experimental Biology 200, 661–675. , , , , & (
- 1992) The effects of temperature on the oxygen consumption, heart rate and deep body temperature during diving in the tufted duck Aythya fuligula. Journal of Experimental Biology 163, 139–151. & (
- 1995a) Estimating the energetics of tufted ducks, Aythya fuligula: a comparison between respirometry, and the doubly labelled water and heart rate techniques. Functional Ecology 9, 40–47. , & (
- 1994) The use of heart rate to estimate oxygen consumption of free-ranging black-browed albatrosses Diomedea melanophrys. Journal of Experimental Biology 193, 119–137. , , & (
- 1995b) Heart rates and oxygen consumptions of exercising gentoo penguins. Physiological Zoology 68, 855–877. , , & (
- 1995c) The energetics of free-living albatrosses. Philosophical Transactions of the Royal Society of London. B 350, 119–131. , , & (
- 1996) Temporal scales of foraging in a marine mammal. Ecology 77, 426–434. (
- 1999) Heart rate and behaviour of fur seals: implications for measurement of field energetics. American Journal of Physiology 276, H844–H857. , , & (
- 1995) Validation of heart rate and doubly-labelled water as measures of metabolic rate during swimming in California sea lions. Functional Ecology 9, 151–160. , , , & (
- 1993) To what extent can heart rate be used as an indicator of metabolic rate in free-living marine mammals. Symposia of the Zoological Society of London 66, 317–332. (
- 1982) Comparative physiology of diving in vertebrates. Advances in Physiology and Biochemistry, Vol. 8 (ed. by O. E.Lowenstein), pp. 179–364. Academic Press, New York. & (
- 1997) Physiology of diving of birds and mammals. Physiological Reviews 77, 837–899. & (
- 1984) Heart rate and aerobic metabolism in Humboldt penguins, Spheniscus humboldti, during voluntary dives. Journal of Experimental Biology 108, 419–428. & (
- 2001) Seasonal hypothermia in a large migrating bird: Saving energy for fat deposition? Journal of Experimental Biology 204, 1361–1367. & (
- 1998) Behaviour and physiology of Svalbard barnacle geese, Branta leucopsis, duing their autumn migration. Journal of Avian Biology 29, 536–545. , & (
- 1992) Relationship between heart rate and oxygen consumption during steady-sate swimming in California sea lions. Journal of Experimental Biology 170, 35–42. , , & (
- 1996) The optimal allocation of time over the dive cycle: an approach based on aerobic and anaerobic respiration. Animal Behaviour 51, 1247–1255. & (
- 1989) Convection and thermoregulation in 2 Antarctic seabirds. Journal of Comparative Physiology 159, 313–322. , , & (
- 1980) The biochemical composition of krill Euphausia superba from South Georgia. Journal of Experimental Marine Biology and Ecology 43, 221–236. (
- 1989) Foraging energetics of Antarctic fur seals in relation to changes in prey availability. Ecology 70, 596–606. , & (
- 1986) Energy requirements of free-ranging little penguin, Eudyptyla minor. Comparative Biochemistry and Physiology 85A, 135–138. , & (
- 1987) Foraging energetics of black-browed albatrosses Diomedea chrysostoma at Bird Island, South Georgia. Ibis 129, 149–158. & (
- 1979) Effect of starvation and experimental feeding on the proximate composition and calorific content of an Antarctic teleost. Notothenia Coriiceps Neglecta. Comparative Biochemistry and Physiology 85A, 135–138. (
- 1982) Energy costs of incubation and moult in petrels and penguins. Journal of Animal Ecology 51, 177–194. (
- 1984) Seabirds. Antarctic Ecology, Vol. 2 (ed. by R. M.Laws), pp. 533–619. Academic Press, London. (
- 1999) Diet, provisioning and productivity responses of predators to differences in availability of Antarctic krill. Marine Ecology Progress Series 177, 115–131. , & (
- 1994) Energetic costs of raising pygoscelid penguin chicks. Polar Biology 14, 205–210. (
- 1996) Core temperature variability in diving king penguins (Aptenodytes patagonicus): a preliminary analysis. Polar Record 16, 37–38. , , , , & (
- 1994) Underwater swimming at low energetic cost by pygoscelid penguins. Journal of Experimental Biology 197, 65–78. , & (
- 1990) Avian basal metabolic rates: their association with body composition and energy expenditure in nature. American Journal of Physiology 259, R333–R340. , & (
- 1989) The reproductive energetics of gentoo (Pygoscelis papua) and macaroni (Eudyptes chrysolophus) penguins at South Georgia. Journal of Animal Ecology 58, 59–74. , & (
- 1983) Water flux and estimated metabolism of free-ranging gentoo and macaroni penguins at South Georgia. Polar Biology 2, 41–46. , & (
- 1986) Digestive physiology is a determinant of foraging bout frequency in hummingbirds. Nature 320, 62–63. , , & (
- 1980) The prudent parent: energetic adjustments in avian breeding. Ardea 53, 99–160. & (
- 1984) Energetics of free-ranging seabirds. Seabird Energetics (ed. by G. C.Whittow & H.Rahn), pp. 203–234. Plenum Press, New York. (
- 1994) Energy expenditure in Arctic seabirds. PhD Thesis, University of Tromso, Norway. (
- 1996) Energy expenditure of breeding common murres. Canadian Wildlife Series Occasional Papers 91, 49–58. (
- 1987) Daily energy expenditure and energy utilization of free-ranging black-legged kittiwakes. Condor 89, 126–132. , & (
- 1990) The annual energetics cycle of little penguins Eudyptyla minor. Ecology 71, 2297–2312. & (
- 1993) Breeding energetics and food requirements of gentoo penguins (Pygoscelis papua) at Heard and Macquarie Islands. Journal of Zoology 231, 125–139. , , , & (
- 1999) Food provisioning to nestling shearwaters: why parental behaviour should be monitored? Animal Behaviour 57, 663–671. , & (
- 1984) Energy cost of incubation to the parent seabird. Seabird Energetics (ed. by G. C.Whittow & H.Rahn), pp. 59–71. Plenum Press, New York. (
- 2001) Heart rate and oxygen consumption of exercising Macaroni penguins. Journal of Experimental Biology 204, 673–684. , , , & (
- 1997) Provisioning and growth rates of nestling fulmars Fulmarus glacialis: Stochastic variation or regulation? Ibis 139, 31–39. & (
- 1997) Diving hypothermia in foraging king penguins. Nature 388, 64–67. , , , , , , & (
- 1997) Heat increment of feeding in Brunnich’s guillemot Uria lomvia. Journal of Experimental Biology 200, 1757–1763. , , & (
- 1992) Body temperature and metabolic rate during natural hypothermia in endotherms. Journal of Comparative Physiology B 162, 696–706. & (
- 1991) Top predators in the Southern Ocean – major leak in the biological carbon pump. Science 253, 64. , & (
- 1963) Thermodynamics of incubation in the house wren Troglodytes aedon. Proceedings of the Inernational Ornithological Congress 13, 884–904. (
- 1992) Diving behaviour and energetics during foraging cycles in king penguins. Ecological Monographs 62, 143–163. , , , , , & (
- 1966) Theory of use of the turnover rates of body water for measuring energy and material balance. Journal of Theoretical Biology 12, 46–74. & (
- 1991) Reproductive energetics and prey harvest in Leach’s storm petrels in the Northwest Atlantic. Ecology 73, 823–832. , & (
- 1997) The effects of hatching date and parental quality on chick growth and creching age in the Chinstrap Penguin (Pygoscelis antarctica): a field experiment. Auk 114, 47–54. , , & (
- 1996) Field metabolic rates of breeding chinstrap penguins (Pygoscelis antarctica) in the South Shetlands. Physiological Zoology 69, 586–598. & (
- 1998) Interannual variability of the South Georgia marine ecosystem: Biological and physical sources of variation in the abundance of krill. Fisheries Oceanography 7, 381–390. , , , , , , , , & (
- 1980) CO2 production in animals: analysis of potential errors in the doubly-labelled water method. American Journal of Physiology 238, R466–R473. (
- 1992) Food and energy requirements of Adélie penguins (Pygoscelis adeliae) on the Antarctic Peninsula. Physiological Zoology 65, 1271–1284. & (
- 1984) Energy Utilization by Free-Ranging. Jackass Penguins, Spheniscus demursus. Journal of Ecology 65, 1648–1655. , & (
- 1992) Estimation of daily energy expenditure from heart rate and doubly labeled water in exercising geese. Physiological Zoology 65, 1188–1216. , , & (
- 1987) Energy utilization by Wilson’s storm petrel (Oceanites oceanites). Physiological Zoology 60, 801–810. , & (
- 1978) Water and energy flux in elephant seal pups fasting under natural conditions. Physiological Zoology 51, 166–178. , & (
- 1997) A new device for monitoring gastric pH in free-ranging animals. American Journal of Physiology 36, G748–G753. (
- 1990) Sustained metabolic scope. Proceedings of the National Academy of Science 87, 2324–2328. , & (
- 1988) Incubation energetics of the laysan albatross. Oecologia 74, 546–550. , , & (
- 2001) Temperature regulation in emperor penguins foraging under sea ice. Comparative Biochemistry and Physiology A 129, 811–820. , , & (
- 1983) Some considerations on the reproductive energetics of pelagic seabirds. Studies in Avian Biology 8, 84–94. (
- 1986) Daily energy expenditure by adult Leach’s Storm-petrels during the nesting cycle. Physiological Zoology 59, 649–660. , & (
- 1997) Doubly Labelled Water: Theory and Practice. Chapman & Hall, London. (
- 1986) Diving behaviour and heart rate in tufted ducks (Aythya fuligula). Journal of Experimental Biology 126, 341–359. , & (
- 1998) Field metabolic rates of Kittiwakes Rissa tridactyla during incubation and chick rearing. Ardea 86, 169–175. , & (
- 1990) Annual variation in breeding biology of gentoo penguins Pygoscelis papua at Bird Island, South Georgia. Journal of Zoology 222, 247–258. (
- 1991) Foraging ecology and diet of gentoo penguins Pygoscelis papua at South Georgia during winter and an assessment of their winter prey consumption. Ibis 133, 3–13. (
- 1992) Diving pattern and performance in relation to foraging ecology in the gentoo penguin, Pygoscelis papua. Journal of Zoology 227, 211–230. , , , & (
- 2000) Sink or swim: Strategies for cost-efficient diving by marine mammals. Science 288, 133–136. , , , , , , & (
- 1990) Factors affecting variation in foraging and activity patterns of gentoo penguins (Pygoscelis papua) during the breeding season at Bird Island, South Georgia. Journal of Applied Ecology 72, 1042–1054. & (
- 1984) A new improved stomach pump for penguins and seabirds. Journal of Field Ornithology 55, 109–112. (
- 1983) Swimming and diving in tufted ducks, Aythya fuligula, with particular reference to heart rate and gas exchange. Journal of Experimental Biology 107, 311–329. & (
- 1995) An implantable data logger system for heart rate and body temperature: its application to the estimation of field metabolic rates in Antarctic predators. Medical and Biological Engineering and Computing 33, 145–151. , & (
- Zar (1984) Biostatistical Analysis, 2nd edn. Prentice Hall, Inc., New Jersey.