In pursuit of Irving and Scholander: a review of oxygen store management in seals and penguins

In a landmark paper in diving physiology, Kooyman and co-workers demonstrated that most dives of Weddell seals (Leptonychotes weddellii) were aerobic in nature and did not result in elevated post-dive blood lactate concentration (Kooyman et al., 1980). The dive duration at which post-dive blood lactate concentration became significantly elevated was named the aerobic dive limit (ADL). Over the past 30 years, the proliferation of electronic dive recorders has led to documentation of dive profiles in many species. The interpretation of the diving behavior and foraging ecology of these marine mammals and sea birds has been primarily influenced by this concept of aerobic diving (Butler and Jones, 1997; Kooyman and Ponganis, 1998; Ponganis, 2011). In order to make such interpretations, there have been many studies of the magnitude of O₂ stores in marine mammals and sea birds. In fact, in the first 3 months of 2010, there were two such studies on phocid seals in this journal (Burns et al., 2010; Hassrick et al., 2010). In contrast to the large number of dive behavior and O₂ store studies, there have been few investigations of the pattern, rate and magnitude of O₂ store depletion in free-diving animals. Yet, it is these very parameters that ultimately determine the limits to dive performance and foraging success. In essence, it is still necessary to pursue the pioneering work of Irving and Scholander (Irving, 1934; Irving, 1939; Scholander, 1940), but in free-diving animals.

Therefore, in this review, we first examine the magnitude of O₂ stores in representative species. This includes discussion of the potential sources of error in the calculation of those stores because of the frequent and almost routine estimation of O₂ stores in the diving physiology literature. Second, we review different options in the strategy of O₂ store management through examination of past studies of O₂ store depletion in marine mammals and diving birds. Third, we focus on our recent investigations of O₂ store utilization during sleep apnea and dives of elephant seals (Mirounga angustirostris) and during dives of emperor penguins (Aptenodytes forsteri). Lastly, we conclude with the implications of these findings for (a) the physiological responses underlying O₂ store utilization, (b) the physiological basis of the ADL and (c) the value of extreme hypoxemic tolerance, and the significance and relevance of the obligate avoidance of re-perfusion injury in these animals.

Oxygen stores are located in the respiratory system, blood and muscle. The magnitude and distribution of those stores vary in different species, and are dependent on the respiratory air volume during a dive, blood volume, hemoglobin (Hb) concentration, myoglobin (Mb) concentration and muscle mass (Table 1). Estimation of the O₂ stores from these parameters is dependent on a variety of assumptions.

Calculations of the respiratory O₂ store usually assume a 15% O₂ extraction and diving air volumes of about 50% and 100% total lung capacity in pinnipeds and cetaceans, respectively (Kooyman, 1989). These assumptions are based on the few data available from inflations of excised lungs, studies during free dives and simulated dives, and the fact that cetaceans appear to dive on inspiration (Kooyman et al., 1999; Ponganis et al., 2003a; Ridgway, 1986). In the past, respiratory data for penguins had only been obtained from simulated dives in pressure chambers (Kooyman et al., 1973c; Ponganis et al., 1999). These included a 15% O₂ extraction, and diving air volumes of 165 and 69 ml kg^–1 in Adélie (Pygoscelis adeliae) and king (Aptenodytes patagonicus) penguins, respectively. More recently, air volume in free-diving animals has also been estimated on the basis of buoyancy, body angles and predicted/observed swim velocities during gliding ascents (Miller et al., 2004; Sato et al., 2002; Sato et al., 2011). Use of the air volume calculated during the gliding ascent for estimation of the respiratory O₂ store is dependent on the assumption that the animal has not exhaled any air earlier during the dive prior to ascent. Notably, the air volumes estimated in free-diving penguins are greater than those determined during simulated dives; this results in a 10–20% increase in estimated total body O₂ stores (Sato et al., 2002). Air volume differences between free dives and simulated dives of penguins and other birds (Stephenson, 1995) as well as the relative paucity of respiratory air volume data in birds argue the need for further investigation. Indeed, even the allometric equation for respiratory air volume and body mass in birds is only based on five samples from different studies (Lasiewski and Calder, 1971).

Calculations of the blood O₂ store typically include the assumption that one-third of the blood volume is arterial and two-thirds are venous, that arterial Hb saturation declines from 95% to 20%, and that initial venous O₂ content is 5 ml dl^–1 less than initial arterial O₂ content and that all venous O₂ is consumed (Kooyman, 1989; Lenfant et al., 1970). In muscle O₂ store calculations, it is assumed that 100% saturated Mb is completely usable during the dive (Kooyman, 1989). O₂ content is calculated on the basis of 1.34 ml O₂ g^–1 of either Hb or Mb at 100% saturation.

Potential sources of variation or error in the estimation of O₂ stores include (a) frequent assumptions as to the size of the respiratory air volume (diving air volume vs total lung capacity) and the percentage extractable O₂, (b) the accuracy of blood volume measurements and Hb determinations in animals such as seals with large blood volumes, large spleens and, consequently, fluctuating hematocrit (packed red cell volume, Hct) and Hb concentration (Hurford et al., 1996; Persson et al., 1973; Ponganis et al., 1993a), (c) the percentage distribution of blood between the arterial and venous systems in animals with large blood volumes, (d) the frequent lack of accurate muscle mass measurements and (e) the frequent assumption that Mb content is the same in all muscles. As an example, because of splenic expansion and a decrease in Hct during anesthesia of seals, the size of the calculated blood O₂ store of the Weddell seal would be underestimated by 50% if the Hct of an anesthetized seal was used in the calculation of the blood volume and its O₂ store (Ponganis et al., 1993a).

The calculation of the blood O₂ store is also dependent on the initial and final Hb saturation. The initial 95% arterial Hb saturation is reasonable given that there may be a pulmonary shunt and/or an altered ventilation–perfusion ratio (mismatch of ventilation to perfusion) in the lung, and, at least in seals, there may be some carboxyhemoglobin (carbon monoxy–Hb) present. Pugh reported a sixfold elevation in blood carbon monoxide gas content in Weddell seals, which was attributed to a high Hb content and the metabolic breakdown of Hb to bilirubin (Pugh, 1959). In regard to initial venous Hb saturation, Scholander observed that venous blood samples often appeared arterialized at the start of forced submersions (Scholander, 1940). If this occurs prior to diving, the venous portion of the blood O₂ store may be much larger (i.e. 5 ml O₂ dl^–1 venous blood greater) than frequently assumed. The end-of-dive limits on arterial and venous Hb saturations are supported by forced submersion studies, in which arterial and venous P_O2 values were 10 and 2 mmHg, respectively, at the electroencephalographic threshold for hypoxemic brain damage (Elsner et al., 1970; Kerem and Elsner, 1973).

The magnitude of total body O₂ stores and the distribution of O₂ in the respiratory, blood and muscle compartment are listed for several representative species of cetaceans, pinnipeds and penguins in Table 2 and Fig. 1. The calculations are based on the parameters in Table 1 and the general assumptions described above. For comparison, with use of human data (Schmidt-Nielsen, 1997) and the same assumptions, human total body O₂ stores are estimated at 24 ml kg^–1, with 42% in the lungs, 44% in the blood and 14% in muscle. Specific values in Table 2 vary from past estimations (Kooyman and Ponganis, 1998) because of the use of more recent data now available in the literature (Table 1). In addition, two calculations of O₂ stores are provided for both emperor penguins and elephant seals in order to demonstrate how our most recent data change both the magnitude and distribution of the total body O₂ store in each of these species. Despite such changes in specific estimates, the general trends in both the magnitude and distribution of O₂ stores in shallow and deep divers remain the same. Deep-diving marine mammals such as the sperm whale and Weddell seal have larger mass-specific O₂ stores than their shallow-diving counterparts (Snyder, 1983). In contrast, the magnitude of the total body O₂ store does not vary that greatly between shallow and deep-diving penguins (Table 2), although these values are somewhat higher than in other diving birds (Butler, 2000; Butler, 2004; Green et al., 2005). In the thick-billed murre (Uria lomvia), the total O₂ store is 45 ml O₂ kg^–1, while in the mallard duck (Anas platyrhynchos), it is only 29 ml O₂ kg^–1 (Croll et al., 1992; Keijer and Butler, 1982).

The distribution of the total O₂ store also varies between deep and shallow divers. In animals that dive deeper and longer, the contribution of the respiratory O₂ store is less, while the proportion of O₂ in the blood and muscle compartments is greater. This pattern even occurs to some degree in penguins, which have much larger mass-specific respiratory O₂ stores in comparison to cetaceans and pinnipeds (Table 2, Fig. 1). Less reliance on the respiratory O₂ store in deep divers decreases the need for gas exchange at depth. Smaller diving air volumes in deep divers also represent a smaller nitrogen reservoir and, in mammals, promote alveolar collapse and pulmonary shunting (Fahlman et al., 2009; Kooyman, 1989; Kooyman et al., 1999). These factors should minimize the risk of excess nitrogen absorption during deep dives.

To date, the rate, pattern and magnitude of O₂ store depletion have been most comprehensively studied in the forced submersion experiments of Scholander, Irving and colleagues on young harbor seals (Phoca vitulina), gray seals (Halichoerus grypus) and hooded seals (Cystophora cristata) (Irving et al., 1941; Scholander, 1940; Scholander et al., 1942). The O₂ content of ‘alveolar air’ declined to 3% (∼21 mmHg or 2.8 kPa) (Scholander, 1940). Under conditions of severe bradycardia (slow heart rate, i.e. less than resting) and peripheral vasoconstriction in seals, the isolated muscle O₂ store was almost completely depleted in about 10 min, while the blood (and respiratory) O₂ store lasted about 20–25 min (Fig. 2A). In comparison, arterial Hb saturations can decrease to as low as 50–80% in less than 2 min during breath holds and dives of humans, and in less than 1 min during human sleep apnea (Dempsey et al., 2010; Lindholm and Lundgren, 2009; Qvist et al., 1993; Yumino and Bradley, 2008). These slow rates of blood O₂ depletion during the bradycardia and vasoconstriction of forced submersions conserve O₂ for the brain and heart because of the now well-demonstrated redistribution of blood flow away from O₂-consuming peripheral tissues (Blix et al., 1983; Elsner et al., 1966; Irving, 1939; Scholander, 1940; Zapol et al., 1979).

The muscle O₂ depletion rate during forced submersions, 10 ml O₂ kg^–1 muscle min^–1, was considered elevated at about 2.5 times the resting rate as a result of struggling (Scholander et al., 1942). Under these conditions, muscle lactate began to accumulate at a Mb saturation of 10–20% (Fig. 2B). Again, 70 years since their publication, these remain the only studies in marine mammals in which muscle lactate and O₂ content have been measured simultaneously.

Further studies by Elsner and colleagues established that blood O₂ in seals could be depleted to arterial and venous P_O2 values as low as 10 and 2 mmHg, respectively (1.3 and 0.03 kPa) (Elsner et al., 1970; Kerem and Elsner, 1973). In forced submersions of Weddell seals conducted by Zapol, Hochachka and colleagues, arterial P_O2 averaged 32 mmHg (4.5 kPa) at 8–12 min of submersion (Zapol et al., 1979); arterial and mixed venous P_O2 reached 20 mmHg (<3 kPa) during a 46 min forced submersion (Hochachka et al., 1977). In addition, as demonstrated by values for hepatic sinus O₂ content that were greater than arterial values near the end of forced submersions, blood O₂ uptake from the lungs did not occur during the later portions of forced submersions of elephant seals (Elsner et al., 1964).

Comparison of the blood O₂ depletion data in the forcibly submerged harbor seal (P. vitulina) with those during asphyxia in the anesthetized, paralyzed dog revealed that the arterial blood O₂ content decreased at a rate 2–4 times faster in the dog (Kerem and Elsner, 1973). Declines in arterial and venous O₂ content of seals during forced submersions were about 1–2 ml O₂ dl^–1 min^–1 (Elsner, 1969; Elsner et al., 1964; Irving et al., 1941; Kerem and Elsner, 1973; Scholander, 1940). Arterial P_O2 at the asphyxial end point (hypoxemic EEG threshold) was 14 mmHg (about 2 kPa) in the dog vs 10 mmHg (1.3 kPa) in the seal. However, that asphyxial end point was reached at 4.25 min in the dog and 18.5 min in the seal. Arterial blood O₂ content at the asphyxial end points were similar in the two species.

Investigations of diving physiology and O₂ store depletion in free-diving seals have been primarily and most extensively conducted with the isolated dive hole technique developed by Kooyman on the sea ice of McMurdo Sound (Kooyman, 1968; Kooyman, 1985; Kooyman, 2006; Kooyman and Kooyman, 2009). In a study conducted by Zapol and colleagues, arterial blood O₂ depletion rates in free-diving Weddell seals were about 0.8 ml O₂ dl^–1 min^–1 (Qvist et al., 1986), in the same range as described above during forced submersions of much smaller seals. In comparison, arterial O₂ content declined at 6 ml O₂ dl^–1 min^–1 in Korean diving ama (Qvist et al., 1993).

The lowest arterial P_O2 reported near the end of a dive of a Weddell seal was 18 mmHg (2.4 kPa), corresponding to 28% Hb saturation (Qvist et al., 1986). End-tidal P_O2 values were in a similar range (Kooyman et al., 1973a; Ponganis et al., 1993a). After dives of 26 and 34 min duration, end-tidal P_O2 values were 13–14 mmHg (<2 kPa). These arterial and end-tidal values, which were much greater than those at the end of extreme forced submersions, also raised the question as to how far the blood O₂ store was depleted during routine diving. This question was also raised by findings during spontaneous sleep apneas of Weddell seals, in which, arterial P_O2 declined at variable rates to only about 25 mmHg (about 3.5 kPa) (Kooyman et al., 1980).

In contrast to the situation during forced submersions of seals, muscle Mb desaturation was incomplete in free-diving Weddell seals (Guyton et al., 1995). Combined with the relatively mild bradycardias of free-diving Weddell seals (Hill et al., 1987), this finding suggested that muscle blood flow persisted to some degree during free dives and that the blood O₂ store was not isolated from muscle. This contrasted with the severe bradycardia and peripheral vasoconstriction during forced submersions. Muscle O₂ depletion rates in free-diving Weddell seals averaged about 7 ml O₂ kg^–1 muscle min^–1 (5.1% change in Mb saturation min^–1) during dives of less than 17 min (near the ADL), and 3.4 ml O₂ kg^–1 muscle min^–1 (2.5% change in Mb saturation min^–1) in longer dives (Guyton et al., 1995). These values were lower than those during Scholander's forced submersion experiments, but greater than the 2 ml O₂ kg^–1 muscle min^–1 of tourniqueted human muscle at rest (Blei et al., 1993; Tran et al., 1999). The low Mb desaturation rates in swimming Weddell seals could be consistent with blood O₂ supplementation of muscle metabolism and/or a lower muscle metabolic rate due to hydrodynamics, prolonged gliding and a low cost of swimming (Williams et al., 2000; Williams et al., 2004). In support of the concept of such blood O₂ supplementation of muscle, it is notable that maintenance of some muscle blood flow during restrained submersions of trained vs naïve seals was accompanied by a slower rate of Mb desaturation in the trained seals (Jobsis et al., 2001).

Limited data are also available for cetaceans during trained dives and stationary breath holds. This research, pioneered by Ridgway, and more recently extended by Williams and colleagues, has largely been conducted by or in association with the US Navy Marine Mammal Program. After dives of a trained bottlenose dolphin (Tursiops truncatus), the O₂ concentration of exhaled air was as low as 3% or 22 mmHg (3 kPa) (Ridgway et al., 1969). During stationary breath holds of dolphins, venous P_O2 in blood vessels of the tail fluke reached values as low as 18–20 mmHg (about 2–3 kPa) (Williams et al., 1999). In beluga whales (Delphinapterus leucas), fluke venous P_O2 values were near 20 mmHg (about 3 kPa) after stationary breath holds between 10 and 17 min duration (Shaffer et al., 1997).

In birds, O₂ store depletion has been determined in both penguins and ducks. In his 1940 monograph, Scholander examined blood and muscle O₂ depletion in macaroni (Eudyptes chrysolophus) and gentoo (Pygoscelis papua) penguins (Scholander, 1940). Arterial blood O₂ content declined quickly from 21 to 3 ml O₂ dl^–1 within 5 min, yielding an overall blood O₂ depletion rate of 3.6 ml O₂ dl^–1 min^–1. We are unaware of comparable data from an avian non-diver. Muscle O₂ content declined from 40 to 0 ml O₂ kg^–1 muscle by 5 min, yielding a minimum muscle O₂ depletion rate of 8 ml O₂ kg^–1 muscle min^–1. This muscle O₂ depletion rate during forced submersion of the penguin is in the same range as resting muscle O₂ consumption (11 O₂ kg^–1 muscle min^–1) reported in the pekin duck (A. platyrhynchos) (Grubb, 1981).

During forced submersion studies of pekin ducks by Jones and colleagues, air sac P_O2 was near 30 mmHg (4 kPa, 25% of the initial value), and arterial and venous P_O2 values were 30 and 23 mmHg, respectively (4 and 3 kPa), at the point of ‘imminent cardiovascular collapse’ (Hudson and Jones, 1986). Because of the P₅₀ (P_O2 at 50% Hb saturation) of duck Hb and the Bohr effect, blood O₂ content was zero at those blood P_O2 values. In comparison to Scholander's penguins, both blood and muscle O₂ depletion rates were lower in the smaller pekin duck. Arterial and venous O₂ content depletion rates during forced submersions in ducks were 2.2 and 1.4 ml O₂ dl^–1 min^–1, respectively (Stephenson and Jones, 1992). The muscle O₂ store was estimated to be depleted within 45 s; this would result in a minimum O₂ depletion rate of about 5 ml O₂ kg^–1 muscle min^–1 (Stephenson and Jones, 1992). As muscle O₂ depletion was not directly measured in this study, the actual depletion rate may have been faster, i.e. near the previously cited 11 ml O₂ kg^–1 muscle min^–1 resting value (Grubb, 1981).

During simulated dives of 5 min duration in Adélie and gentoo (P. papua) penguins in a pressure chamber, air sac O₂ concentration decreased at a rate of 2.2% min^–1 to minimum values near 2% (about 15 mmHg or 2 kPa) (Kooyman et al., 1973c). Arterial P_O2 declined from 80 mmHg (10.2 kPa) to 20–30 mmHg (2.6–4.3 kPa). These end-of-dive values were similar to those above in pekin ducks at the point of ‘imminent cardiovascular collapse’. However, the extreme limit to hypoxemic tolerance in penguins was not determined. Blood O₂ content and muscle O₂ depletion were not examined in this study.

Because of the scarcity of data on the rate and magnitude of O₂ store depletion during dives, we have attempted to examine O₂ store depletion patterns in three situations. Serial blood gas sampling and ¹H nuclear magnetic resonance (NMR) spectroscopy have allowed analyses of both blood and muscle O₂ store depletion during spontaneous sleep apneas (breath holds) of the northern elephant seal (M. angustirostris) (Ponganis et al., 2002; Ponganis et al., 2008; Stockard et al., 2007). The refinement of catheterization techniques for the elephant seal in the sleep apnea studies and the development of a backpack P_O2 recorder have also allowed investigation of blood O₂ depletion in translocated, free-diving elephant seals (Meir et al., 2009). Lastly, application of the P_O2 recorder and a backpack near-infrared Mb saturation recorder have allowed examination of O₂ store depletion in the air sacs, blood and muscle of emperor penguins diving at an isolated dive hole in the sea ice of Antarctica (Meir and Ponganis, 2009; Ponganis et al., 2009; Ponganis et al., 2007; Stockard et al., 2005; Williams et al., 2011).

The spontaneous breath holds that occur during sleep in seals represent a convenient model to investigate O₂ store depletion during an apneic period. Close access to the sleeping seal allows collection of blood samples as well as the opportunity to conduct ¹H NMR spectroscopy to investigate Mb desaturation. During these apneic periods, heart rate is near 50 beats min^–1, cardiac output is maintained at resting levels and muscle blood flow declines but persists at an average value near 50% the eupneic (during a breath) level (Ponganis et al., 2006). Hence, the muscle O₂ store is not completely isolated from the circulation during these breath holds. The lack of change in blood lactate concentration during and after the apneas is also consistent with adequate organ perfusion and O₂ delivery during the breath hold (Castellini et al., 1986).

Serial blood gas analyses during sleep apneas in young elephant seals revealed that arterial and venous P_O2 values quickly equilibrated within the first minute, indicative of minimum gas exchange in the lung after the first minute (Stockard et al., 2007). The minimal role of the lung as an O₂ store is consistent with the exhalation of air observed at the start of the apnea, the low lung volumes determined in simulated dives (Kooyman et al., 1973b), prior observations of the equilibration of arterial and venous O₂ contents during forced submersions (Elsner et al., 1964), and Elsner's proposal that the large hepatic sinus–vena cava of the seal acts as a venous blood O₂ reservoir (Elsner et al., 1964).

Fig. 3 demonstrates the decline in arterial and venous P_O2 and O₂ content during sleep apneas as long as 10.9 min (Stockard et al., 2007). It is apparent that arterial P_O2 and the corresponding Hb saturation during most of the breath hold (i.e. beyond 1 min) is in a range that would be considered critical [<60 mmHg (7.8 kPa) and 90% saturation] and, indeed, severe [<50 mmHg (6.5 kPa) and 80% saturation] in most human patients (Mason et al., 2005; Nunn, 1977). The lowest arterial P_O2 values during sleep apnea are less than the mean value (25 mmHg, 3.3 kPa) of humans breathing ambient air at 8400 m near the summit of Mount Everest (Grocott et al., 2009).

Blood O₂ content declined by about 2 ml O₂ dl^–1 min^–1 during sleep apnea. This was about twice the rate previously observed during forced submersions and was consistent with the maintenance of cardiac output during the breath hold. For a typical 7 min apnea in these young seals, about 56% of the blood O₂ store was consumed (Stockard et al., 2007). That portion alone of the blood O₂ store was sufficient to provide enough O₂ for a metabolic rate of 4.2 ml O₂ kg^–1 min^–1 during the apnea. This estimation of the blood O₂ store was based on the previously reviewed O₂ store assumptions and on a blood volume of 196 ml kg^–1 and Hb concentration of 23.5 g dl^–1 in these young seals (Stockard et al., 2007).

The development of ¹H NMR spectroscopy techniques to analyze Mb saturation by Jue and associates allowed investigation of Mb desaturation in elephant seals sleeping inside a magnetic resonance imaging scanner (Ponganis et al., 2008). Mb desaturated rapidly with the onset of apnea (Fig. 4B), and settled at a steady-state level near 80% saturation by 4 min into the apneic period. In these young seals, that initial desaturation rate corresponded to a muscle O₂ depletion rate of 1–2.3 ml O₂ kg^–1 muscle min^–1 during the first 4 min of the apnea. The Mb saturation remained at the 80–85% level until the end of apnea, after which it quickly re-saturated to >95% saturation. Low muscle blood flow and blood-to-muscle O₂ transfer during the apnea allowed maintenance of the steady-state 80% saturation throughout the apnea. Assuming all muscle in the body desaturated to that level, the O₂ from muscle, in addition to that provided by the blood, would yield a metabolic rate of 4.7 ml O₂ kg^–1 min^–1 during a 7 min apnea in these young elephant seals. This O₂ store depletion rate of 4.7 ml O₂ kg^–1 min^–1 is 26% greater than the metabolic rate predicted at rest for a mammal of this size (Kleiber, 1975), and is consistent with the maintenance of heart rate and cardiac output during the breath-hold period.

In summary, it is the blood O₂ store that is primarily utilized during sleep apnea in elephant seals. The pattern of O₂ store utilization contrasts strikingly with that during forced submersion (Fig. 4A). Blood O₂ depletion rates are greater and muscle O₂ depletion rates slower during sleep apnea than during forced submersion (Fig. 4). Higher heart rates and tissue perfusion result in greater blood O₂ extraction. The lower rate of muscle O₂ depletion is secondary to both maintenance of some blood O₂ delivery and a lower muscle metabolic rate during sleep than during the stress of a forced submersion. Under these conditions, Mb is able to function not only as an O₂ store but also as the primary intracellular O₂ transporter because of its translational diffusion coefficient and high concentration in seal muscle (Gros et al., 2010; Ponganis et al., 2008).

For typical 7 min apneas of these young elephant seals, the blood O₂ store is far from depleted; more than 40% of the initial blood O₂ still remains at the end of the breath hold. In addition, the muscle O₂ store is only 20% depleted during sleep apnea. Thus, although heart rate, cardiac output and metabolic rate are maintained near resting levels, O₂ store depletion does not limit the breath-hold duration during sleep apneas.

Blood O₂ store depletion during diving was evaluated with the use of indwelling P_O2 electrodes in translocated, juvenile elephant seals, and conversion of the P_O2 profiles to Hb saturation profiles with use of the elephant seal O₂–Hb dissociation curve (Meir et al., 2009). This translocation model was developed by Oliver and LeBoeuf (Oliver et al., 1998). Mean dive heart rates of such translocated seals ranged from 31 to 48 beats min^–1 (Andrews et al., 1997). The typical 2–3 day return trips from release sites to the rookery provided routine dives of 10–20 min duration and 100–200 m depth, and occasional dives as long as 44 min and as deep as 700 m (Meir et al., 2009). During these trips, dives were spontaneous, voluntary and, in contrast to Weddell seals diving under sea ice at an isolated dive hole (Kooyman, 1968), unrestricted in access to the surface. Blood P_O2 was recorded at three sites – the aorta, the hepatic sinus and the extradural vein (Fig. 5).

Arterial P_O2 values were as low as 12–23 mmHg at the end of dives, corresponding to routine Hb saturations of 8–26%, again demonstrating exceptional hypoxemic tolerance. The lowest arterial P_O2 measured in this study (12 mmHg) is nearly as low as the ‘critical P_O2’ of harbor seals and Weddell seals (10 mmHg) (Elsner et al., 1970; Kerem and Elsner, 1973). Routine venous P_O2 values and Hb saturations were as low as 2–10 mmHg (0–4%), equivalent to or even lower than those measured in forced submersion studies, and even lower than the well-documented hypoxemic extremes of horses performing strenuous exercise (Bayly et al., 1989; Manohar et al., 2001). These low end-of-dive P_O2 values result in near-complete depletion of blood O₂ stores during routine dives, with net O₂ content depletion values up to 91% and 100% in the arterial and venous stores, respectively.

It is notable that venous P_O2 continued to increase during the early phase of the dive, sometimes reflecting arterial values (Fig. 5). Such ‘arterialization’ of venous blood (P_O2 values greater than those typically found in venous blood) is not consistent with blood extraction by the tissues or blood flow to muscle in this period, and is suggestive of an arterio-venous (a–v) shunt. There was also substantial overlap between end-of-dive arterial and hepatic sinus Hb saturation values (Fig. 6). This is consistent with previous forced submersion and sleep apnea studies, and is expected in a breath-hold diver with collapsed lungs, and supports the concept of the hepatic sinus–venae cavae as a significant O₂ reservoir.

The rate of decrease in venous P_O2 (and often arterial P_O2) and venous Hb saturation often became steeper concurrent with the ascent of the dive, usually most pronounced in the final 15–45 s (see Fig. 5A,B and the declines in saturation during ascent in Fig. 7). This period is coincident with both the ascent or ‘anticipatory’ tachycardia (increased heart rate) (Andrews et al., 1997) and the intense stroking that occur in this species (Williams et al., 2000). These P_O2 data support the hypotheses that: (a) the ascent tachycardia serves to increase blood flow and O₂ delivery to depleted tissues at the end of the dive, maximizing the gradient for O₂ uptake at the surface, and (b) some blood flow to muscle occurs in this period (Thompson and Fedak, 1993). Some muscle blood flow during this period is also consistent with the lack of elevated muscle temperature during dives of seals and the slight elevations in blood lactate near the end of long dives (Guppy et al., 1986; Ponganis et al., 1993b). Increased muscle blood flow during increased exertion is also predicted by a numerical model of blood O₂ transport in which the duration of aerobic metabolism during a dive is optimized by coupling muscle blood flow to muscle O₂ demand (Davis and Kanatous, 1999).

The arterialization of venous blood early in a dive does not increase the magnitude of the total body O₂ stores as this increase in blood O₂ is due to O₂ transfer from the lung store (already included in the O₂ store calculation (Kooyman et al., 1999; Meir et al., 2009). However, if depletion of arterial O₂ to below 20% Hb saturation (typical end-of-dive arterial Hb saturation value used for such calculations) occurs during dives (as documented in this study with values as low as 8% S_a,O2), the available O₂ store would increase by about 3 ml O₂ kg^–1 to 88 ml O₂ kg^–1 (with parameters specified in Table 1).

Meir et al.'s results also indicate that at the whole animal level, juvenile elephant seals are not ‘hypometabolic’ during diving, and that they do not require any significant anaerobic metabolism during routine dives (Meir et al., 2009). For example, the contribution of the venous O₂ store alone to metabolic rate is >100% of the allometrically predicted basal metabolic rate, even for routine dives >10 min.

Oxygen store depletion has been investigated in emperor penguins diving freely at an isolated dive hole in McMurdo Sound. This approach, again pioneered by Kooyman (Kooyman et al., 1971), has allowed examination of heart rate, swim speed, stroke frequency and temperature during diving (Kooyman et al., 1992; Meir et al., 2008; Ponganis et al., 2001; Ponganis et al., 2004; Ponganis et al., 2003b; van Dam et al., 2002). Most importantly, an ADL of 5.6 min has been documented by blood lactate determinations in birds diving at the isolated dive hole (Ponganis et al., 1997b). During these dives, the penguins travel as far as 1.2 km from the dive hole, and primarily feed on the sub-ice fish Pagothenia borchgrevinki (Ponganis et al., 2000; Shiomi et al., 2008). Because of the availability of this prey item, most dives are less than 100 m – relatively shallow for emperor penguins.

Air sac and blood O₂ depletion during dives have been examined with the use of a P_O2 electrode and custom-built recorder (Ponganis et al., 2009; Ponganis et al., 2007; Stockard et al., 2005). As in research with the elephant seal, P_O2 profiles have been converted to Hb saturation profiles by determination of O₂–Hb dissociation curves of emperor penguins, and application of the curve to the P_O2 data (Meir and Ponganis, 2009). Investigation of pectoral muscle O₂ depletion has been conducted with the development of a backpack near-infrared recorder and probe (Williams et al., 2011). In these studies, the birds are typically instrumented under general anesthesia in the evening and, after overnight recovery, are allowed to dive for 1–2 days at the dive hole before recapture, anesthesia and removal of the recorder.

Air sac P_O2 profiles in 5–75 m deep dives of up to 11 min duration revealed a compression hyperoxia followed by a decrease in P_O2 as the respiratory O₂ fraction declined secondary to O₂ consumption and as the ambient pressure decreased during ascent (Fig. 8) (Stockard et al., 2005). Final P_O2 declined exponentially and reached values as low as 0 mmHg; it was less than 20 mmHg (2.7 kPa) in 42% of dives. By comparison, the inspiratory P_O2 for a bar-headed goose (Anser indicus) at a simulated altitude of 11,580 m was 23 mmHg (3.1 kPa), and air sac P_O2 of an Adélie penguin at the end of a simulated dive was 15 mmHg (2.0 kPa) (Black and Tenney, 1980; Kooyman et al., 1973c). End-tidal P_O2 values of a climber breathing ambient air on Mount Everest (35 mmHg, 4.7 kPa) and human shallow-water black-out thresholds (25 mmHg, 3.3 kPa) were also greater than many of the end-of-dive P_O2 values in the air sacs of emperor penguins (Ferretti et al., 1991; Ferrigno and Lundgren, 2003; West et al., 1983).

The low end-of-dive air sac P_O2 values in emperor penguins contrast with the air sac P_O2 values near 30 mmHg (4 kPa) of ducks at the point of imminent cardiovascular collapse (Hudson and Jones, 1986). Such hypoxic tolerance is afforded, in part, by a left shift of the O₂–Hb dissociation curve of the emperor penguin (and other penguins as well as high altitude flying birds) in comparison to that of the duck (Meir and Ponganis, 2009). At a P_O2 of 20 mmHg (2.7 kPa), the Hb of the duck would be devoid of O₂ while the Hb of the emperor penguin would still be 27% saturated.

Initial air sac P_O2 values during dives indicated that initial O₂ fractions could be as high as 19%, which is greater than the 15% value observed in simulated dives in a pressure chamber (Kooyman et al., 1973c). Complete consumption of the 19% O₂ fraction resulted in a greater respiratory O₂ store than that typically calculated with the pressure chamber results (Kooyman, 1989; Ponganis et al., 2010). The overall rate of change in the respiratory O₂ fraction during dives of up to 11 min duration ranged between 5 and 2% min^–1. By comparison, the respiratory O₂ fraction changed about 2% min^–1 during simulated dives of penguins (Kooyman et al., 1973c).

Arterial P_O2 profiles of diving emperor penguins were characterized by a compression hyperoxia followed by a decline to values as low as 26–30 mmHg during dives of up to 12 min duration (Ponganis et al., 2009; Ponganis et al., 2007) (see Fig. 8A). The pattern was similar to that observed in the air sac, but it was unclear why these end-of-dive arterial values were generally greater than air sac values for dives of similar duration. The magnitude of air sac P_O2 depletion was quite variable, and it is possible that differences in physiological responses and depth profiles of individual dives may have at least partially contributed to differences between the air sac and arterial studies. Regardless of the mechanisms, it is notable that arterial Hb saturation of emperor penguins can be maintained near 100% during almost the entire dive (Fig. 9), even during dives as long as 12 min (Meir and Ponganis, 2009).

Arterial Hb saturation declined primarily during the ascent, and reached a minimum value of only 47%. These final arterial P_O2 values and Hb saturations were greater than those of the bar-headed goose at a simulated altitude of 11,580 m (22 mmHg, 28% saturation) and similar to those of humans breathing ambient air near the summit of Mount Everest (25 mmHg, 54% saturation) (Black and Tenney, 1980; Grocott et al., 2009). Whether lower arterial P_O2 values and Hb saturation occur in other dives of emperor penguins is unknown and awaits further investigation. During the ascent, the arterial Hb desaturation rate was about 25% min^–1, although the average value over a dive of 9 min duration was much lower, about 5–6% min^–1 (Meir and Ponganis, 2009).

The ability to maintain arterial Hb saturation near 100% during most of the duration of a dive in emperor penguins contrasts with that in elephant seals, where arterial Hb saturation is usually between 70% and 30% during the majority of the dive (Fig. 9). This highlights the size and role of the respiratory O₂ store in penguins and emphasizes the significance of lung-to-blood O₂ transfer during dives of penguins.

Venous P_O2 and Hb saturation profiles during dives of emperor penguins were remarkable for the variability in the rate and pattern of decline throughout the dive (Fig. 8B, Fig. 10). Venous P_O2 could transiently rise during the early portion of some dives while it declined in others. Initial or peak P_O2 values were often consistent with early, even pre-dive arterialization of venous blood (i.e. >90% Hb saturation). P_O2 and Hb saturation declined at variable rates and, at times, with fluctuations during the later portions of dives (Fig. 8B, Fig. 10). The transient elevations in P_O2 and Hb saturations during dives are not consistent with muscle blood flow and blood O₂ extraction by muscle during such dives. Rather, these increases in P_O2 emphasize (a) the potential role of a–v shunts in arterialization of the venous O₂ store and, again, (b) the significance of net lung-to-blood O₂ transfer during dives of penguins. In contrast, muscle blood flow may occur in those dives in which venous P_O2 and Hb saturation decline significantly during the early portion of the dive. This plasticity in the peripheral blood flow is also suggested by the range of venous P_O2 values and Hb saturations for a dive of given duration. For example, near 6 min dive duration, final venous Hb saturations ranged from 3% to 75%. End-of-dive venous Hb saturations indicated that the entire venous O₂ store could be consumed; 15% of dives had a final venous Hb saturation ≤5%.

Mb desaturation profiles of diving emperor penguins revealed two distinct patterns (Fig. 11): a monotonic decline (type A) and a mid-dive plateau pattern (type B) (Williams et al., 2011). In type A dives, Mb saturation generally followed a monotonic decline throughout the dive and, in dives near the ADL, the final Mb saturation was near 0%. Mean desaturation rate in type A dives was 14.4±3.8% min^–1. Type A dives are consistent with isolation of muscle from the circulation during dives. The desaturation rate is similar to rapid Mb desaturation observed in forced submersion studies (Scholander et al., 1942) and is higher than desaturation rates observed in Weddell seals, where muscle perfusion was suspected during dives (Guyton et al., 1995) (see Fig. 12). Mean muscle O₂ consumption (12.4±3.3 ml O₂ kg^–1 muscle min^–1) based on the Mb desaturation rate in diving emperor penguins is low, less than one-tenth the pectoralis–supracoracoideus muscle O₂ consumption calculated from emperor penguins swimming maximally in a flume (160 ml O₂ kg^–1 muscle min^–1) (Kooyman and Ponganis, 1994; Ponganis et al., 1997a), demonstrating the efficiency of underwater locomotion in diving emperor penguins.

In type B dives, Mb desaturation rate during the initial descent was either rapid, as in type A dives, or moderate. Desaturation rate then slowed significantly, often leveling off completely until the ascent phase when desaturation rate again increased (Fig. 11B and Fig. 12). Given the continuous, almost constant, stroking pattern throughout these dives, the minimum change in Mb saturation implies the muscle O₂ store was supplemented by the circulation during these periods. Such intermittent muscle blood flow is consistent with the variable venous P_O2 profiles discussed above. The increase in Mb desaturation rate during the ascent phase implies that muscle blood flow was reduced during this portion of the dive. The type B plateau period of Mb desaturation in diving emperor penguins is similar to that during sleep apnea in elephant seals, during some dives of Weddell seals and during exercise in humans (Guyton et al., 1995; Ponganis et al., 2008; Richardson et al., 1995). Saturation values in type B dives of 8–10 min duration often declined to below 10% (Fig. 11B and Fig. 12); mean desaturation rate in type B dives was 9.8±2.4% min^–1. Despite the potential supplementation of the Mb O₂ stores during dives, the near-complete O₂ depletion of Mb in longer type B dives (Fig. 11) is supportive of the concept that the onset of post-dive lactate accumulation is secondary to muscle O₂ depletion.

The results of these studies of O₂ store depletion in emperor penguins emphasize the significance of the relatively large respiratory O₂ store in penguins as well as the apparent plasticity in peripheral vascular responses. In contrast to seals, this allows for maintenance of the penguin's arterial Hb saturation during much of the dive (Fig. 9). We hypothesize that not only are the transient elevations in heart rate during the early segments of dives of penguins important for lung-to-blood O₂ transfer but also the increased cardiac output can be utilized either to enhance the venous O₂ store through the use of a–v shunts or to supplement the muscle O₂ store through maintenance of muscle blood flow (Figs 10 and 11).

As recently reviewed (Ponganis et al., 2010; Sato et al., 2011), the findings in these latest studies have increased the estimated total body O₂ stores in emperor penguins to 68 ml O₂ kg^–1, well above prior estimates of 56 ml O₂ kg^–1. This is primarily due to the large diving air volumes, greater respiratory O₂ extraction and the pre-dive arterialization of venous blood measured in recent studies. The respiratory system contains 33% of the total O₂, while the blood and muscle compartments hold 31% and 36%, respectively. Note that pre-dive arterialization of venous blood can occur in emperor penguins whereas in elephant seals, arterialization of venous blood appears to occur primarily during early descent (Meir et al., 2009; Meir and Ponganis, 2009). Therefore, in contrast to the elephant seal, the venous O₂ store of the emperor penguin can be optimized to arterial levels before the dive.

The average depletion rates of all the body O₂ stores of the emperor penguin contributed 6.8 ml O₂ kg^–1 min^–1 to whole-body metabolic rate for dives of about 10 min duration. The muscle O₂ store contributed most (53%), while the respiratory and blood compartments contributed 31% and 16%, respectively (Williams et al., 2011). This average total O₂ store depletion rate during diving demonstrates the low metabolic cost of diving in emperor penguins as it is similar to measured and predicted resting metabolic rates. It must be emphasized that O₂ depletion rates are highly variable, and that there is probably a wide range of diving metabolic rates dependent on the activity and duration of an individual dive. In addition, the actual metabolic rate during dives, especially dives longer than the ADL, may be greater because of phosphocreatine breakdown and glycolysis.

These recent investigations of O₂ store depletion in elephant seals and emperor penguins highlight differences in O₂ store management associated with (a) the magnitude and distribution of O₂ stores, (b) the intensity of cardiovascular responses during a breath hold or dive and (c) locomotory muscle workload. The depletion patterns of these O₂ stores during dives also provide insight into the processes underlying the ADL, i.e. the duration of aerobic metabolism and the onset of post-dive blood lactate accumulation. In addition, effective utilization of the entire O₂ store demonstrates the importance of extreme hypoxemic (low arterial O₂) tolerance and avoidance of re-perfusion injury.

Recent findings demonstrate that assumptions previously made in the calculation of O₂ stores may underestimate the true magnitude of the respiratory and blood O₂ stores (Ponganis et al., 2010). First, the volume of respiratory air during spontaneous breath holds, especially in penguins, may be greater than the commonly assumed values determined during simulated dives (Sato et al., 2002; Sato et al., 2011). Second, again in penguins, the start-of-dive respiratory O₂ fraction and the net extraction of respiratory O₂ may be greater than those obtained during simulated dives and forced submersion. Third, in emperor penguins, arterialization of the venous blood prior to the dive suggests that the venous portion of the blood O₂ store can be larger than that typically assumed in the O₂ store calculation. And lastly, especially as evidenced by our data in elephant seals, end-of-dive arterial saturations may well be less than the assumed 20% value, again increasing the size of the calculated O₂ store.

The difference in arterial Hb saturation profiles between emperor penguins and elephant seals emphasizes the role of the large respiratory O₂ store in emperor penguins (Fig. 9). In dives of emperor penguins, arterial Hb saturation can be maintained near 100% during much of the dive, even during dives as long as 12 min. Although lower arterial Hb saturations may occur as suggested by low end-of-dive air sac P_O2 values on some dives, 12 min is remarkable because more than 99% of dives of emperor penguins at sea are less than that duration (Kooyman and Kooyman, 1995; Wienecke et al., 2007). In contrast, seals, which exhale prior to diving, reach arterial Hb saturations of 60–70% within a few minutes after the start of a dive, and within a minute after initiation of a sleep apnea breath hold (Fig. 9). Elephant seals routinely experience arterial P_O2 values and Hb saturations that would be considered critical in humans.

We also conclude that heart rate, peripheral blood flow distribution and muscle workload are the primary determinants of the rate and pattern of O₂ store utilization. This is exemplified by sleep apnea in elephant seals, during which higher heart rate, higher cardiac output and presumed organ perfusion, and maintenance of muscle blood flow result in a faster rate of blood O₂ depletion and a slower muscle O₂ depletion rate than during forced submersions (Figs 3 and 4). In seals, sleep apnea and forced submersion can be considered to represent opposite ends of the spectrum of physiological responses and O₂ store management during a breath hold. Diving fits in between those two extremes with its exact position in the spectrum determined by the nature and circumstances of a given dive. During sleep apnea, average muscle blood flow is only one-half the eupneic level. In that regard, it is notable that heart rates during dives of elephant seals are even lower than those during sleep apnea. Therefore, we further hypothesize that muscle blood flow, at least until the start of increased heart rate during ascent (the ‘ascent tachycardia’), is low to nil during dives of seals. In contrast, in emperor penguins, highly variable venous P_O2 profiles and two distinct patterns of muscle Mb desaturation suggest that regulation of blood flow to muscle and to peripheral a–v shunts is quite plastic both within dives and between different types of dives (Figs 10 and 11). The variability in the muscle blood flow response in emperor penguins is probably associated with the relatively larger magnitude of their respiratory O₂ store and with the transient elevations in heart rate observed in the initial portions of their dives (Meir et al., 2008). In penguins, in contrast to seals, a high heart rate early in the dive and a large respiratory O₂ store allow for increased blood O₂ uptake from the lungs, and either enhancement of the blood O₂ store through a–v shunting (arterialization of venous blood) or supplementation of aerobic muscle metabolism through muscle blood flow.

The degree of muscle perfusion and, indeed, the general tissue distribution of blood flow during a dive remain key issues in understanding the rate and pattern of O₂ store depletion. Differences in renal and hepatic clearances between short, aerobic dives and long, exploratory dives of Weddell seals presumably reflect differences in heart rate, renal/hepatic perfusion and blood O₂ depletion rates during those types of dive (Davis et al., 1983; Guppy et al., 1986). Numerical modeling of blood flow distribution and tissue O₂ consumption during dives suggests that heart rate and muscle blood flow should be tailored to muscle workload in order to optimize the duration of aerobic metabolism (Davis and Kanatous, 1999). However, studies of California sea lions, Steller sea lions (Eumetopias jubatus), harbor seals, gray seals, bottlenose dolphins and emperor penguins have all reported that, in general, diving heart rate does not correlate with locomotory effort (Fedak et al., 1988; Hindle et al., 2010; Thompson and Fedak, 1993; Williams et al., 1999; Williams et al., 1991; Meir et al., 2008). Finer scale analyses of heart rate–stroke rate relationships would be useful. And, unfortunately, muscle blood flow has yet to be measured during diving. Further resolution of this topic awaits future research.

We conclude that the most plausible physiological basis of the ADL is the depletion of the primary locomotory muscle O₂ store and the net accumulation of intramuscular lactate. We suspect, as demonstrated by Scholander and colleagues' findings in Fig. 2B, that such lactate accumulation will begin when Mb saturation declines to about 10–20%. After that threshold, lactate accumulation and phosphocreatine breakdown will increase.

We also hypothesize that a low rate of muscle O₂ consumption (locomotory effort) and possible maintenance of muscle perfusion and blood-to-muscle O₂ transfer may delay the onset of lactate accumulation and prolong the ADL in individual dives (Kooyman and Ponganis, 1998; Williams et al., 2011; Williams et al., 2000). In emperor penguins, we think that vascular responses and the degree of muscle ischemia are quite plastic during different dives, consistent with our observed heart rate, venous P_O2 and Mb desaturation profiles. In seals, some muscle blood flow may also occur, especially during the ascent tachycardia. This is consistent with minor elevations of blood lactate concentration late in long dives (Guppy et al., 1986), a lack of elevation of muscle temperature during dives (Ponganis et al., 1993b) and steeper declines in extradural vein Hb saturation profiles during late ascent (Meir et al., 2009).

Our hypothesis that muscle O₂ depletion underlies the ADL is not consistent with the significant muscle Mb saturation remaining in Weddell seals even during dives beyond the ADL (Guyton et al., 1995) (see Fig. 12). We suspect that this difference is due to the specific muscle (latissimus dorsi) examined in the Weddell seal study. The longissimus dorsi–iliocostalis muscle complex has been considered the primary locomotory muscle for the hindlimb propulsion of phocid seals (Howell, 1930; Kanatous et al., 1999). On an anatomical basis, the latissimus dorsi muscle participates in movement of the forelimb, not the hindlimb (Howell, 1930). Therefore, we suspect that the slow muscle desaturation rate observed in the Weddell seal was due, in part, to the low metabolic rate of a muscle not primarily used in phocid propulsion. We suggest that the longissimus dorsi muscle would have a higher rate of Mb desaturation. Again, we emphasize that more studies are needed to unravel this important aspect of O₂ store utilization in diving animals.

To fully deplete and take advantage of their large O₂ stores, diving mammals and birds should have extreme hypoxemic tolerance. Although such tolerance had been demonstrated during forced submersions of seals, extreme respiratory and blood O₂ depletion have now been recorded during free dives of both emperor penguins and elephant seals. As already reviewed, these extremes are below inspiratory values found in bar-headed geese at simulated altitude as well as below end-tidal and blood levels in humans breathing ambient air near the summit of Mount Everest. This tolerance is best exemplified in the elephant seal. This seal functions during most of its dive under what would be considered hypoxic conditions in human patients. Indeed, human breath-hold divers experience cardiac arrhythmias and neurological complications with arterial saturations that, while low for humans, are far higher than those routinely observed in seals (Andersson et al., 2009a; Andersson et al., 2009b; Lindholm and Lundgren, 2009; Liner and Andersson, 2009). And, sleep apnea patients develop pulmonary hypertension and right ventricular failure with chronic episodic exposures to hypoxia (Dempsey et al., 2010). Again, in contrast, chronic hypoxia has no apparent sequelae in the elephant seal, even after months-long trips to sea, during which it spends 80 to 90% of its time diving (LeBoeuf et al., 1988).

The mechanisms underlying such tolerance as well as avoidance of re-perfusion injury remain to be fully determined. Higher brain capillary densities, shorter O₂ diffusion distances, mild hypothermia, intrinsic neuronal tolerance, neuroglobin and cytogoblin function, greater glycolytic and buffering capacities or other biochemical adaptations may all play significant roles (Blix et al., 2010; Castellini and Somero, 1981; Folkow et al., 2008; Fuson et al., 2003; Halasz et al., 1974; Kanatous et al., 2002; Kerem and Elsner, 1973; Kerem et al., 1973; Odden et al., 1999; Ramirez et al., 2007; Williams et al., 2008). For example, the shift in the O₂–Hb dissociation curve of penguins is an advantage in comparison to other birds (Meir and Ponganis, 2009). As regards hypothermic protective mechanisms, although temperature fluctuations and regional heterothermy occur during dives, it should be noted that core hypothermia, at least in emperor penguins and elephant seals, has not been observed (Meir and Ponganis, 2010; Ponganis et al., 2004). However, re-perfusion injury may be avoided with enhanced O₂ free radical scavenging, especially through elevated glutathione levels and increased activities of enzymes involved in glutathione recycling (Elsner et al., 1998; Vázquez-Medina et al., 2006; Vázquez-Medina et al., 2007; Zenteno-Savin et al., 2010). And perhaps, in seals, a slight elevation in carboxyhemoglobin levels may even contribute to protection against re-perfusion injury (Nakaoa et al., 2005). Investigation of these processes is required and is relevant not only to the biology of diving but also to a better understanding and potential treatment of human ischemic (decreased perfusion) events (i.e. stroke, myocardial ischemia, organ preservation for transplantation).