Extreme sensitivity of biological function to temperature in Antarctic marine species


‡Author to whom correspondence should be addressed: L.S. Peck, NERC British Antarctic Survey, High Cross, Madingley Road, Cambridge, CB3 0ET, UK, E-mail: l.peck@bas.ac.uk.


  • 1Biological capacities to respond to changing environments dictate success or failure of populations and species over time. The major environmental feature in this context is often temperature, and organisms across the planet vary widely in their capacity to cope with temperature variation. With very few exceptions, Antarctic marine species are more sensitive to temperature variation than marine groups elsewhere, having survivable temperature envelopes between 5 °C and 12 °C above the minimum sea temperature of −2 °C.
  • 2Our findings show that in biological functions important to long-term survival these animals are even more tightly constrained. The Antarctic bivalve mollusc Laternula elliptica and limpet Nacella concinna both survive a few days in experiments at 9–10 °C, but suffer 50% failure in essential biological activities at 2–3 °C and complete loss at 5 °C. The Antarctic scallop Adamussium colbecki is even more sensitive, and loses the ability to swim as temperature approaches 2 °C.
  • 3These failures of activity are caused by a loss of aerobic capacity, and the animals investigated are so sensitive that a 2 °C rise in sea temperature could cause population or species removal from the Southern Ocean.


Organism responses to changing environments are complex, and species survival is usually determined by changes in the balance of ecological factors, rather than direct effects of alterations to the physical environment (Davis et al. 1998; Convey et al. 2002). Recent models mainly focus on higher scale metrics, such as habitat loss, that encompass multiple environmental characters (Travis 2003). The relative importance of ecological factors vs. direct effects of changed physical factors differs from site to site, and with evolutionary history.

A large proportion of the species on earth live either in the deep-sea (Grassle & Maciolek 1992; Snelgrove & Smith 2002), mid to deep ocean depths, or in the polar oceans (Clarke & Johnston 2003), where temperatures are below 5 °C, are very stable, and have been so for long evolutionary periods. Adaptation to these conditions has resulted in reduced rates of growth, development, and metabolism (Peck 2002). In these environments, changes in physical environmental characters may be more important than elsewhere.

Antarctic marine species are stenothermal, being much less able to survive elevated temperatures than species elsewhere (Peck & Conway 2000). Thus polar fish generally die at temperatures around, or slightly above 5 °C (Somero & DeVries 1967), and most cold-water marine invertebrates die in experiments at temperatures of 5–10 °C. These temperature envelopes are 2–4 times smaller than those for lower latitude species (Peck & Conway 2000). Metabolic rates in ectotherms vary with ambient temperature. Recent studies in Antarctic marine invertebrates have shown that as temperature rises oxygen consumption and heartbeat rate increase while blood oxygen content declines (Peck 1989; Pörtner et al. 1999; Peck, Pörtner & Hardewig 2002; Pörtner 2002). As demand increases a point is reached where oxygen supply cannot meet the total requirements of the animal, anaerobic pathways are entered and anaerobic metabolic end products accumulate (Pörtner et al. 1998; Peck et al. 2002). This point is termed the critical temperature, and indicates the long-term physiological limit, as anaerobiosis cannot be sustained indefinitely. Above the critical temperature a point is eventually reached where the short-term oxygen debt is beyond the capacity of the animal, and death occurs at the experimental upper lethal temperature. The temperature where this is reached depends on time of exposure to elevated temperature and species tolerance of hypoxia.

Frederich & Pörtner (2000) set the physiological responses of temperate species into a wider framework, identifying optimum, suboptimum (pejus) and nonsurvivable (pessimum) temperatures, while explaining physiological responses of the spider crab Maja squinado. They related these changes to Shelford's law of tolerance (Shelford 1931), which states ‘the survival of an organism depends on the completeness of a complex of conditions’, here termed the survival envelope. This law was important in the development of the niche and ecological hypervolume concept. Pörtner et al. (1998) suggested that translocating the survival curves to lower temperatures truncated them, and the decline in survival and aerobic capacity at the lower end of survivable temperatures is lost in polar species, because it falls below the freezing point of seawater.

Recently, the data showing that Antarctic ectotherms have very low temperature limits, and that this is associated with transfer of tissues to anaerobic metabolism, has been postulated to be a function of declining aerobic scope with elevated temperature (Mark, Bock & Pörtner 2002; Peck et al. 2002; Pörtner 2002). Thus, as temperature rises the difference between maintenance metabolism and maximum aerobic metabolic rate decreases because maintenance costs outstrip the ability to raise oxygen supply.

In any changing environment, the important criteria for survival are not the conditions that can be endured, but the conditions where essential biological functions cannot be maintained, and the balance of these conditions over time. Thus, the temperature where tissues turn to anaerobic pathways is not important in this context, but the temperatures where functions such as feeding, locomotion or reproduction are compromised are critical. These functional limits will set the survival envelope for species, either by setting immediate limits, or altering the balance of ecological interactions. Where the physical environment dominates survival factors, or organisms have insufficient physiological flexibility (as for polar benthos), physical limits may rise in importance relative to biological interactions.

We tested the hypothesis that the stenothermal nature of Antarctic marine species is caused by limited aerobic scopes. To do this we measured the animals’ abilities to perform activity, and predicted that the loss of essential biological functions would occur progressively with very small temperature elevations. In our experiments we investigated critical activity functions in the Antarctic limpet Nacella concinna, the infaunal bivalve Laternula elliptica and the scallop Adamussium colbecki. In the limpet we evaluated the ability to right itself when turned over. If limpets are incapable of righting after disturbance their survival is compromised. In L. elliptica we measured its ability to rebury in sediment. L. elliptica is a large infaunal bivalve, living deeply buried in sediments. Specimens move vertically in the sediment during normal activity cycles, and also need to rebury when ploughed from the sediment by ice disturbance. In A. colbecki we assessed the proportion of animals that swam in response to stimuli.

Materials and methods

Experimental animals were collected by scuba divers from 8 to 15 m depth in Hangar Cove, Rothera Point, Adelaide Island (67°34′20″ S, 68°07′50″ W). Specimens of L. elliptica and N. concinna were held for 24 h in aquaria at ambient temperature before being used in experiments. Constant low-light levels were maintained, to mimic Antarctic summer conditions. In studies at ambient temperature animals were used immediately after the 24 h acclimation period. For elevated temperatures animals were held in jacketed water baths and temperatures raised at 0·1 °C h−1, until required temperatures were reached. Video recordings were made to determine burrowing or turning rate and times to completion. Data were collected using a Panasonic AG6124HB 24 h time-lapse video recorder, and subsequently analysed using a JVCBR-S610E video analysis machine. At each temperature for each species 18–26 animals were evaluated. A total of 145 L. elliptica were used, and measurements made only in 2002. For N. concinna, measurements were made in both 2001 and 2002 and a total of 190 specimens were used from the same population.

Specimens used to evaluate swimming in A. colbecki were collected from Hangar Cove, Rothera Point, from 5 to 25 m depth. Animals were stimulated initially at the ambient water temperature. Tank temperatures were then varied over a −1·9 °C to +1·9 °C range to compare acute responses. Temperature changes were made in 0·5 °C increments at a maximum rate of change of 1 °C day−1 (Bailey 2001). Light conditions were the same as previous. Prior to experiments animals were moved into a swim tank and allowed to rest for a minimum of 6 h before the first escape response was stimulated (6 h was the maximum time to 90% recovery recorded in this species for exhaustive exercise by Bailey et al. 2003). Handling was minimised and animals transported between tanks in water. The water in the original tank was slowly mixed with the water in the new tank to minimise any small differences in temperature or salinity. Escape responses were then stimulated using freshwater within 1 °C of tank temperature. Water was introduced to the rear of the animal, directly beneath the hinge, and typically 10 cm3 was injected at approximately 2 cm3 s−1. Freshwater is known to stimulate swimming in wild A. colbecki (Berkman 1988). There was minimal disturbance of the water around the animal and no force was exerted on the body of the scallop itself. The area of reduced salinity visible around the animal typically dispersed before adduction.


For N. concinna between 95 and 100% of individuals are capable of righting themselves within 24 h at 0 °C (Fig. 1a). Righting ability falls progressively with temperature, until no limpets can perform this activity at 5 °C. Fifty per cent of individuals are incapable of righting between 2 °C and 2·5 °C. For L. elliptica at 0 °C, approximately 90% rebury in 24 h, whereas none rebury at 5 °C, and 50% of the population lose the ability at 2 °C (Fig. 1b). In both species half the population have lost the ability to perform critical biological functions at, or very slightly above, 2 °C.

Figure 1.

(a) Righting responses in the Antarctic limpet N. concinna with temperature. Data shown are the proportion of limpets righting in 24 h. Data are from experiments in 2001 and 2002. For each point n = 20–31. All regressions were made following square root and arcsin transforms of percentage data (arcsin(√%righting) = 1·20–0·180T °C; r2 = 0·90, F = 77·9, P < 0·001, 9 df). (b) Reburying in the bivalve mollusc L. elliptica with temperature. Data show the proportion of animals reburying in 24 h (n = 18–26). Regression line: arcsin(√%burying) = 0·95–0·173T °C (r2 = 0·85, F = 22·4, P = 0·009, 5 df). A total of 190 N. concinna and 145 L. elliptica were used. In both figures dotted lines indicate 95% confidence intervals for regressions. For both plots, lines and confidence intervals shown were plotted following sine and square back transforms.

Swimming in the scallop A. colbecki also declined with rising temperature, but the rate of decline was more rapid (Fig. 2). Below 0 °C the proportion of scallops swimming when stimulated varied between 28% and 50%. Above 0 °C the proportion swimming declined monotonically. Above 2 °C no scallops could be induced to swim.

Figure 2.

The proportion of Antarctic scallops, A. colbecki, swimming in response to freshwater stimulation. Each point is the proportion swimming at that temperature. The number tested is given above each point and varied between 57 and 175. The total number of animals used in the experiments was 858. A regression was fitted to data for temperatures above −0·3 °C, where a clear temperature effect was apparent. This regression was fitted to square root and arcsin transformed percentage values. Regression Line: arcsin(√%swimming) = 0·682–0·230T °C (r2 = 0·93, F = 51·5, P = 0·006, 4 df).

Times required for limpets to right themselves are difficult to evaluate accurately, as the process involves a period of probing the substratum, followed by attachment and then turning. The turning element takes a short time, of the order of 5–20 s. Burrowing in L. elliptica varied with size, and was completed at 0 °C in around 1 h in small individuals and up to 16 h in large specimens.

Between 1997 and 2000, sea temperatures at 15 m depth at Rothera point ranged between −1·9 °C and +1·4 °C (Fig. 3). The stability of winter temperatures varies with sea ice extent and duration, and the 1997 and 1998 winters had more sea-ice than the 1999 winter. In each summer, temperatures rose above 0 °C for periods between 2 weeks and approximately 3 months (1999). Temperatures only rose above 1·0 °C in the 1997/98 and 1998/99 summers, and then only for periods of approximately 1–2 weeks.

Figure 3.

Temperature recordings from 15 m depth in Ryder Bay, Adelaide Island. Recordings were made 800 m from the site of collection of the experimental animals used, and were continuously logged throughout. Minimum temperatures in winter were around −1·8 °C, but in summer rose to values between 0·2 °C and 1·4 °C.


Populations of both L. elliptica and N. concinna at Rothera exhibit high levels of ability to perform critical biological functions at 0 °C, with nearly 100% of individuals reburying (L. elliptica) or righting (N. concinna). These activities are rapidly lost with rising temperature, such that half the population have lost these abilities at +2·0 to +2·5 °C, and all individuals lose the ability to bury or right at +5·0 °C. For the scallops it was not possible to elicit swimming responses from more than 50% of the animals studied at any temperature. This may have been because swimming is a more energetically costly activity than righting in limpets or burrowing in Laternula, and other factors such as feeding or reproductive condition interfered with swimming ability. Brokordt, Himmelman & Guderley (2000) showed that swimming was markedly reduced in animals near to or post spawning in Chlamys islandica. Experiments were conducted on A. colbecki during the Antarctic summer, when animals would be feeding on the intense phytoplankton bloom, and would also be in post-spawning condition (Tyler et al. 2003). The variability in swimming at temperatures below 0 °C may therefore reflect natural variation in individual ability to swim at this time of the year. It is also possible that stronger stimuli may have elicited more swimming. Bailey et al. (2003) obtained higher swimming responses from A. colbecki that had been returned to the UK and Germany prior to experimentation and were not recently collected. However, their experiments were conducted in June, when A. colbecki gonads are immature.

The scallops in this study rapidly lost the ability to swim, and none swam at 2 °C and above. A. colbecki has a lower upper lethal temperature than the other two species. Bailey (2001) was unable to acclimate specimens to temperatures above +3 °C because at 4 °C animals rapidly lost condition and 50% mortality occurred in 19 days. L. elliptica and N. concinna both have experimental upper temperatures of 9–10 °C (Peck 1989; Peck et al. 2002).

The ability to perform work in animals is part of the aerobic scope of individuals in that population. Aerobic scope is used for many functions, including growth, and processes associated with feeding, as well as activity. Each individual at any time will have a range of factors reducing or enhancing aerobic scope, and individuals will vary in this respect. Thus, evaluations of population capacities are a measure of average capacity in those conditions.

Previous studies showing that upper experimental temperatures and transfers to anaerobic metabolism set the various short, medium and long-term temperature limits for marine species living in polar conditions, fit a model where the underlying control is aerobic scope (Fig. 4), as suggested by Pörtner (2002). As temperature increases for ectotherms, underlying metabolic costs rise. Usually maximum aerobic rate either rises slightly, or stays the same. Thus, with rising temperature aerobic scope, the difference between maintenance cost and maximum rate, decreases. At some point costs equal the maximum sustainable rate, tissues begin to turn to anaerobic metabolism, and the critical long-term physiological limit is reached (Pörtner et al. 1998; Peck et al. 2002; Pörtner 2002). With even higher temperatures the anaerobic deficit (magnitude of excess of metabolic costs over oxygen supply) increases until the short-term anaerobic limit is reached and the animals die. The final limits are dictated by anaerobic tolerances, and death occurs by processes similar to asphyxiation. Prior to the critical physiological temperature being reached reductions in scope dictate that progressively more biological functions are compromised, as indicated by the hatched zone in Fig. 4. Our experiments are a test of this hypothesis. If rising temperature progressively reduces aerobic scope, then the ability to perform work should decline progressively as upper limits are approached. These limits to work should also be at lower temperatures than the experimental or physiological upper limits. The monotonic decreases in the capacity to perform various behaviours requiring work in limpets, scallops and burrowing bivalve molluscs, combined with cessation of these activities at significantly lower temperatures than previously identified upper temperatures for survival all support the aerobic scope model. It should be noted here that swimming in scallops is an anaerobic activity. However, recovery from exercise is aerobic, and the ability of an individual to swim will depend on its aerobic condition when acclimated to any given temperature. It should also be noted that isolated muscle fibres from A. colbecki adductor muscle continue to contract normally at least to temperatures above 4 °C (Bailey 2001).

Figure 4.

Schematic of the effects of rising temperature on aerobic capacity in Antarctic marine stenotherms. Aerobic scope declines with temperature and biological functions are lost up to a point where aerobic balance becomes negative and tissues enter an anaerobic state (the critical temperature). Eventually hypoxic levels become too great and the short-term temperature limit is reached.

Sea temperatures at Rothera point, where this study was conducted are typical of the maritime Antarctic, ranging from −1·8 °C in winter to maxima around 1 °C in summer. Our data would indicate that ectothermic species inhabiting the maritime Antarctic are close to their upper temperatures for the initiation of significant loss of biological function in summer. A 1 °C rise in summer sea temperatures would take most of them to 50% loss of biological functions, and this during the short summer bloom period, that only lasts for 2–3 months (Clarke, Holmes & White 1988), when most herbivores need to be active to exploit their major food supply. The most sensitive species, like A. colbecki, would completely lose their ability to swim in summer with a 1 °C temperature rise. Predictions of future sea temperatures around Antarctica are difficult, because of the complexity added by sea–ice interactions, but most models predict that global sea temperature will rise by 2 °C or more in the next 100 years (Mitchell, Johns & Senior 1998). Clearly, summer sea temperatures of 2–3 °C would pose extreme, and possibly insurmountable problems for many Antarctic marine species.

Times to complete burial at 0 °C in L. elliptica were very similar to those reported by Peck et al. (2004) at the same temperature, and were around × 10 slower than rates for temperate burrowing bivalve molluscs (Peck et al. 2004). When Antarctic and temperate marine species were compared at their normal habitat temperatures, activity rates were × 3 to × 10 slower for 8 comparisons of activity (Peck et al. 2004). The only exception was sustained swimming in fish, which is compensated because of markedly increased mitochondrial densities in red pectoral muscles of cold-water fish (Johnston et al. 1998). The lack of compensation of activity elsewhere suggests the increase of mitochondrial density seen in fish red muscle might not be widespread, leading to poor abilities to compensate work rate for temperature.

All of the 12 polar marine species so far evaluated have upper experimental temperatures below, or close to 10 °C (Peck & Conway 2000; Peck et al. 2002), and the most stenothermal species, the bivalve mollusc Limopsis marionensis, the brachiopod Liothyrella uva and the scallop Adamussium colbecki all die in experiments at approximately 4 °C (Peck 1989; Pörtner et al. 1999; Bailey 2001). The species studied here therefore represent the broad range of responses to elevated temperature identified in polar marine ectotherms. All species so far studied transfer to anaerobic metabolism 2–4 °C below their upper experimental temperatures, and here we show functions are lost 2–8 °C below upper experimental temperature limits. Any future sea temperature rise is unlikely to be consistent around Antarctica. However, the data here would suggest that the majority of Antarctica's in excess of 4000 marine benthic species so far described (Clarke & Johnston 2003) would be at risk of at least population level losses from only a 1–2 °C increase in local summer sea temperatures.

Individual to population level responses to environmental change that enhance survival fall into three main categories: (1) using inherent physiological capacities or scopes; (2) adapting to new conditions; or (3) migrating to areas consistent with survival. We have shown that the Antarctic marine fauna has a very poor ability to cope physiologically, possibly the poorest so far described. The Antarctic marine fauna is characterised by slow growth, increased longevity and deferred maturity (Peck 2002). Of the species studied, several live in excess of 40–50 years (Peck & Bullough 1993; Arntz, Brey & Gallardo 1994; Brey et al. 1995; Peck & Brey 1996; Peck, Brockington & Brey 1997; Peck 2002). Generation times are typically 10–20 years. Abilities to adapt to changing conditions are therefore poor. Finally, most continents have long coastlines covering large latitudinal ranges that, at least theoretically, allow migration to more hospitable conditions in a warming scenario. The outline of Antarctica covers less degrees of latitude than any other. The animals on the Southern Ocean seabed have less scope to migrate away from poor conditions than faunas elsewhere, irrespective of dispersal capabilities. The combination of very poor functional scopes, with slow rates of adaptation and restricted available dispersal ranges make Antarctic marine species amongst the most fragile to environmental change on earth.


We thank the support staff at Rothera station, and especially in the Bonner laboratory for assistance with animal collections. Work on scallops was carried out as part of a NERC CASE studentship to D.M.B. All experiments complied with relevant UK laws.