Emperor Penguin Oxygen Consumption, Heart Rate and Plasma Lactate Levels During Graded Swimming Exercise

Maximum oxygen consumption has not been measured in birds, but is about 10–13 times the resting rate during horizontal flight (LeFebre, 1964; Tucker, 1968; Butler et al. 1977) and is over 15 times the resting rate in hovering hummingbirds (Lasiewski, 1963). In contrast, ducks swimming in a watermill can only increase their about 3.5–4 times above out-of-water resting rates (Prange and Schmidt-Nielsen, 1970; Woakes and Butler, 1983; Baudinette and Gill, 1985; Butler, 1991). The low increases for swimming ducks are due to the small amount of leg muscle required for propulsion, compared with the larger flight muscles (Prange and Schmidt-Nielsen, 1970; Butler, 1991). Since penguins ‘fly’ under water, they use a much greater amount of their muscle mass than do swimming ducks and may, therefore, have a metabolic scope closer to that of flying birds. However, it would seem that diving birds, such as the emperor penguin, would be metabolically adapted to O₂ conservation and would consequently have a moderate aerobic scope, similar to that measured in seals (Ponganis et al. 1990a) and sea lions (Ponganis et al. 1991; Williams et al. 1991). Furthermore, the exceptionally low drag of their fusiform body shape (Nachtigall and Bilo, 1980) suggests that penguins may swim at low energetic costs even at high speeds.

We selected emperor penguins, the largest of all aquatic birds, so that we would be able to make comparisons, at high swimming effort, with marine mammals of similar mass that have already been studied within the same type of exercise protocol. This comparison was, in part, prompted by the diving physiology and behaviour of emperor penguins. During dives, heart rate responses of emperor penguins are similar to those observed in phocid seals (Kooyman et al. 1992b). However, for their size, birds appear to be better divers if duration, depth and speed are used as the evaluating criteria. Emperor penguins cruise at a speed about 50% faster than those of seals and sea lions (Ponganis et al. 1990b; Castellini et al. 1992; Kooyman et al. 1992b). King penguins (Aptenodytes patagonicus) often dive for longer than 5 min, to depths greater than 200 m, and at speeds of slightly more than 2 m s⁻¹ (Kooyman et al. 1992a); emperor penguin diving capacities are about twice those of king penguins (G. L. Kooyman, unpublished observations). In an analysis of several fur seal species, none matched any of these diving standards even though all were at least two times greater in body mass (Gentry and Kooyman, 1986).

We therefore sought to measure the metabolic scope, exercise heart-rate response and alteration in plasma lactate concentration in emperor penguins (Aptenodytes forsteri) during graded exercise. These variables provided us with the information needed to determine whether the birds were near or at ⁠. The results show a close correlation between average heart rate and average ⁠. The reluctance or inability of the birds to swim beyond the maximum induced work load and the rapid rise in plasma lactate concentration at this inflection point indicate that an aerobic/anaerobic threshold had been reached. This level of exercise was similar to that at the highest measured previously in seals, sea lions and dolphins (Ponganis et al. 1990a, 1991; Williams et al. 1991, 1993).

On 12 November 1988, four 4-month-old emperor penguin chicks (CW, M, T, N) weighing between 2.5 and 4.5 kg were collected at Cape Washington (74°39′S, 167°22′E), Antarctica. They were flown by twin-otter aircraft to a sea-ice laboratory near McMurdo Station. After 6 days at the station, they were flown directly to San Diego, California, in a refrigerated US Air Force Starlifter. Brief stops were made at Christchurch, New Zealand, and Honolulu, Hawaii. Within 24 h, the birds were placed in the Sea World polar bird quarantine facility (Mission Bay, San Diego, California). In this environment, they were hand-fed until they were transferred to the Sea World public display facility. The air temperature in the display was 2°C and the water temperature was about 5.5°C.

Resting was measured in three birds when they were 1 year old. One bird (N) did not moult in the normal cycle, so it was not used in the heart-rate experiments. The exercise experiments were conducted when the birds were between 1.5 and 2.5 years old. During that period, the birds weighed between 19.9 and 26.7 kg.

For each experiment, the birds were transported to Scripps Institution of Oceanography (SIO), University of California, San Diego (UCSD), in upright, padded boxes in a refrigerated van. They remained in the van except when an experiment was conducted. Total time away from Sea World was between 36 and 48 h for each trip.

At the SIO Hydraulics Laboratory, UCSD, the birds were placed individually in the water channel previously described by Davis et al. (1985). All procedures for measuring ⁠, increasing the workload (swim effort) and measuring heart rate were the same as formerly described (Davis et al. 1985; Ponganis et al. 1990a; Williams et al. 1991). Very briefly, water in the water channel (1.1 m square × 16 m long) was chilled to 2–4°C the day before the experiment. This water temperature was about the same as that at the Sea World facility. Water flow rate was held constant at 1 m ⁻¹ while the birds swam in a 2.5 m long test section partitioned by two grates from the rest of the channel.

Water velocity was measured with an electromagnetic flow sensor placed 30 cm above the floor of the flume. A vertical profile of water velocity varied by less than 0.1 m s⁻¹ (Davis et al. 1985). The velocity was correlated to the rotation rate of the propellers, which in normal operation was used to set the water speed. Water velocity in the test section was the same as in the rest of the channel.

Resting was obtained from three birds held for several hours in a small water bath, temperature regulated at 1.5–6.1°C. During exercise in the water channel, the swim procedure and measurements were as follows. Before releasing a bird into the flume, a narrow epoxy-glue strip painted on the mid-line of the back and about two-thirds of the body length from the head was used as an anchor for attachment of a 5 cm long, 1.5 cm diameter tube to the feathers with nylon ties. The tube had a swivel to prevent an attached load line from becoming twisted when the bird occasionally made turns. The 6 mm diameter, braided nylon line passed through the rear grate in the centre of the channel and made an approximately 45° turn through a pulley towards the surface. About 1 m above the surface, the line turned vertically through another pulley hanging at the edge of the channel. About 2 m below this second pulley, the desired weight was hung onto the line. A weight of at least 0.25 kg was always applied, to prevent slack in the line. This arrangement produced a horizontal pull on the penguin when it swam in the centre of the channel. Increased swim effort was produced by adding loads ranging from 0.25 to 4 kg.

Once the bird had been released in the test section, the water surface was covered with an acrylic dome mounted on a sheet of 6 mm plywood. The birds breathed under this dome, and air was drawn through the dome at a suitable rate to maintain the O₂ concentration above 20%. Air flow, O₂ concentration, calibration of the system and calculation of were accomplished with an on-line computer system similar to that described by Davis et al. (1985).

Experiments in which the heaviest loads (3–4 kg) were added lasted for about 3 min, and was determined every 10 s. Experiments at these high loads ended when the birds showed signs of fatigue. A plateau of values within the range ±15% of average was short, or not found. Experiments with lighter loads lasted for about 30 min and soon after they began, reached a plateau and remained stable until the end. Average values were determined from the sum of all values from all runs at a specific load. During heart rate and lactate experiments, average values were calculated from the measurements during the specific run in which the heart rate or lactate determination was made.

Prior to f_H and lactate measurements, the birds were anaesthetized. The procedure used was similar to that described previously (Kooyman et al. 1992b). In brief, birds were masked with a 1 l plastic bottle in which a neoprene neck ring formed a seal. Induction was at 4% isoflurane, which was reduced to 1% during the procedure.

For placing heart-rate electrodes, anaesthesia lasted about 30 min, and after about 5 h of recovery the bird was released into the flume. As in previous studies (Kooyman et al. 1992b), the two electrodes were placed on the dorsal midline; one was placed between the wings and another about two-thirds of the body length behind the head. Heart rates were monitored on a Hewlett-Packard chart recorder. The total heart beat count of each surface and submersion interval was divided by the duration of the interval to obtain f_H. The average f_Hfor a specific was calculated by summing all beats over the entire exercise cycle and dividing by the total time. Heart rate analysis was performed only for periods in which was stable. Surface and submersion times were manually recorded by an observer and stored on the computer as was measured.

The placement of catheters required 30–60 min under anaesthesia. An 18 gauge (1.3 mm diameter, 3 cm length) catheter was placed 2.5 cm into the vein of the middle toe. The catheter was held in place with Vetwrap coated with epoxy glue to form a rigid cast that allowed the foot enough mobility for normal positioning while swimming. A 2 m extension was placed on the catheter, the dead space of which was 1.6 ml. This length was necessary to raise the sampling stopcock above the water for sampling during swimming. Sampling duration ranged from 30 s to 10 min and was fastest when the bird was swimming and slowest between runs, when the bird rested in the water and was usually shivering. Samples were analyzed immediately after collection on a YSI model 2300 Stat lactate and glucose analyzer. Heart-rate and blood-sampling experiments were carried out separately to avoid catheter, electrode and load lines becoming tangled when the birds made turns in the flume.

All statistics and tests of significance were obtained using the Statistix commercial software package (St Paul, MN). Unless stated otherwise, confidence intervals are expressed as one standard deviation of the mean.

Resting values were measured during quiet periods in the water-bath metabolic chamber (Fig. 1, Table 1). The lowest was 6.2 ml O₂ kg⁻¹ min⁻¹ for individual CW when the water temperature was 1.5°C.

During exercise, was closely correlated with load, and the coefficients of determination (r²) ranged from 0.93 to 0.99. There was no significant difference among the birds either in the slope or the intercept (analysis of covariance, multiple-comparison test, P<0.05; Zar, 1984). The pooled analysis is shown in Fig. 1.

The maximum mean when loaded was 47.3 ml O₂ kg⁻¹ min⁻¹ for M while swimming and pulling a 4 kg load. Maximum values for both CW and T were close to this, at 44.6 ml O₂ kg⁻¹ min⁻¹ and 44.7 ml O₂ kg⁻¹ min⁻¹ respectively, also with a 4 kg load (Fig. 1). The maximum average value for a single run was 52 ml O₂ kg⁻¹ min⁻¹ for M (Figs 2 and 3).

Regression lines for the relationship between f_Hand during surface swimming and submersion are summarized in Fig. 2 and Table 2. The r² values ranged from 0.07 for M when surfaced to 0.39 for CW and T when submerged. In all three birds, examination of surface versus submerged f_Hvalues showed that mean f_Hvalues at the surface and while submerged were significantly different (P<0.001, paired sample t-test, d.f.=23). These mean surface and submerged f_Hvalues are given in Fig. 2. Submersion times ranged from 5 to 15 s; the lowest values corresponded to the highest values. Surface times ranged from 2 to 6 s. The average heart rate values at specific values for each bird are plotted in Fig. 2 for each bird and in Fig. 3 for pooled results.

The highest lactate levels were 7 and 9.5 mmol l⁻¹ in birds CW and M, respectively. Fig. 4 shows the relationship between [lactate] and swimming

The 10-to 13-fold increase in from resting levels during flight in birds previously cited should not be assumed to represent the entire aerobic scope of birds since was not measured. Similarly, the highest measured in this study (52 ml O₂ kg⁻¹ min⁻¹) may not represent for the emperor penguin, since no plateau in occurred. However, the observers’ impressions of the swimming effort of the birds when pulling a 4 kg load, as well as the high plasma lactate concentration (Fig. 4) at these loads, suggested that the birds were near their maximum limit. The lowest measured resting of 6.2 ml O₂ kg⁻¹ min⁻¹ is similar to previous values obtained for emperor penguins resting in air in their thermoneutral zone (6.2–7.3 ml O₂ kg⁻¹ min⁻¹; Pinshow et al. 1976; Le Maho et al. 1976; Deswasmes et al. 1980). This is greater than the resting rate of 5.7 ml O₂ kg⁻¹ min⁻¹ for a 20.8 kg bird predicted from the equation given by Aschoff and Pohl (1970). The metabolic scope of the emperor penguin, from our values, is 7.8, and it is 9.1 using the predicted resting ⁠. The maximum increase in measured in swimming emperor penguins is less than that measured in birds during flapping flight. However, all flying birds are considerably smaller than emperor penguins (Butler, 1991), so that direct comparisons are difficult. The emu (Dromaius novaehollandiae) is of similar body mass to the emperor penguin, and the highest measured for this bird (48 ml O₂ kg⁻¹ min⁻¹) and the resting (4.2 ml O₂ kg⁻¹ min⁻¹) gave a metabolic scope of 11.4 (Grubb et al. 1983). As was not measured, the actual aerobic capacity of the emu is therefore probably much greater than that of the emperor penguin.

The aerobic capacity of the emperor penguin measured in this study is, therefore, lower than that for the running emu or for flying birds. Patak and Baldwin (1993) recently reported that the relative mass and oxidative capacity of the pelvic limb muscle of emus were similar to those of the flight muscles of birds. Although the emperor penguin pectoralis/supracoracoideus muscle mass constitutes a similar percentage of the body mass (P. J. Ponganis, unpublished results) to that of the pelvic limb muscle of the emu (Patak and Baldwin, 1993), muscle oxidative capacity may be less, as previous studies have suggested that emperor penguin muscle is ‘best adapted for anaerobic work’ (Baldwin, 1988). Its lower aerobic capacity in comparison with the emu and with flying birds may, therefore, be related more to differences in mitochondrial density than to relative muscle mass.

The measured exercise capacity of the emperor penguin is also less than the 30- and 12-fold aerobic scopes of dogs and goats, terrestrial mammals of similar body mass to emperor penguins (Taylor et al. 1987). Those values were based on measured data and predicted resting values. The ninefold increase in emperor penguin above predicted resting levels is, however, similar to that measured in three marine mammals. Maximum measured in the bottlenose dolphin (Tursiops truncatus) was 7–11 times the predicted resting rate (Williams et al. 1993). In harbour seals (Phoca vitulina) and sea lions (Zalophus californianus), both of similar mass to the emperor penguin, the maximal measured exercise capacity was 9–10 times both the predicted and measured resting (Ponganis et al. 1990a, 1991). Although we did not obtain a well-defined ⁠, the measured aerobic capacity of the emperor penguin appears to be most similar to the aerobic capacities measured in diving mammals.

The surface and submersion heart rates during surface swimming of emperor penguins (Fig. 2) were much higher than those during diving (Kooyman et al. 1992b), and both increased with workload in a manner similar to that found in sea lions (Williams et al. 1991). Mean f_Halso increased with (Fig. 3), as in both harbour seals and sea lions (Williams et al. 1991; Butler et al. 1992). The f_Hof emperor penguins is similar to that of seals and sea lions when the is at the same level (Williams et al. 1991; Butler et al. 1992). However, the highest levels reported for these mammals was 20–25 ml O₂ kg⁻¹ min⁻¹. Seals and sea lions show a significant drop in f_Hbetween surface breathing and when submerged; in the seal, submersion f_His about one-third of the surface rate (Ponganis et al. 1990a). The surface and submersion f_Hpattern of the emperor penguin is similar to that of the seal, although absolute f_Hvalues were higher.

It was not possible to estimate swimming speeds from the methods used in this study. Flow was turbulent because of the grate, and thrust generation by the wings while swimming in natural conditions may be different from thrust generation while swimming against flowing water in a channel or against flowing water with a load applied. Thus, direct extrapolation of diving in the wild may be limited. Comparisons between video tapes of penguins swimming in the wild and in the channel showed that the amplitude and frequency of the wing stroke appeared to be lower under natural conditions than for the captive birds swimming with the lightest load. Wing-beat frequency of emperor penguins swimming in the wild was about 45 beats min⁻¹ (G. L. Kooyman unpublished), and an average of 62–70 beats min⁻¹ for captive birds swimming with the lightest loads. Similarly, heart rates during natural dives (Kooyman et al. 1992b) are often less than the submerged heart rates observed at the lowest workloads in this study. Therefore, it seems reasonable to assume that of diving birds in the wild would be less than the lowest submerged swimming values obtained in this study, 20 ml O₂ kg⁻¹ min⁻¹ (Fig. 1). If so, then the emperor penguin diving metabolic rate would be less than three times the resting levels we measured, which is less than the swimming of six times the resting metabolic rate measured in the Adelie penguin (Pygoscelis adeliae) surface swimming in a 20 m water channel at 2.3 m s⁻¹ (Culik and Wilson, 1991). Such a diving metabolic rate would be consistent with the moderate aerobic capacity measured in this study and may help to explain the emperor penguin’s remarkable diving breath-hold capacity of 15–20 min (Kooyman et al. 1971).

We thank L. Starke, L. Winter, P. Jobsis and S. Eckert for assistance in these studies. This research was supported by grants NSF DPP 87-15863 and USPHS HL 17731.