Regulation of Body Temperature by Some Mesozoic Marine Reptiles

See allHide authors and affiliations

Science  11 Jun 2010:
Vol. 328, Issue 5984, pp. 1379-1382
DOI: 10.1126/science.1187443


What the body temperature and thermoregulation processes of extinct vertebrates were are central questions for understanding their ecology and evolution. The thermophysiologic status of the great marine reptiles is still unknown, even though some studies have suggested that thermoregulation may have contributed to their exceptional evolutionary success as apex predators of Mesozoic aquatic ecosystems. We tested the thermal status of ichthyosaurs, plesiosaurs, and mosasaurs by comparing the oxygen isotope compositions of their tooth phosphate to those of coexisting fish. Data distribution reveals that these large marine reptiles were able to maintain a constant and high body temperature in oceanic environments ranging from tropical to cold temperate. Their estimated body temperatures, in the range from 35° ± 2°C to 39° ± 2°C, suggest high metabolic rates required for predation and fast swimming over large distances offshore.

The metabolic status of extinct vertebrates is a key to understand their feeding strategy, which was critical for satisfying their daily energy requirements, as well as their potential to exploit cold environments. Phylogeny and ecology most likely had a large influence on the thermophysiology of past vertebrates. High metabolic rates mean the need to access large amounts of high-quality food, which may be satisfied by adopting predatory behavior, as shared by many carnivorous mammals, except scavengers. Endothermy is the ability to generate and retain enough heat to elevate body temperature to a high but stable level, whereas homeothermy is the maintenance of a constant body temperature in different thermal environments (1, 2). Such internal production of heat is not restricted to mammals and birds. Heat generation can have several origins: digestive organs in mammals and birds (1) or muscles in endothermic lamniform sharks (3). Paladino et al. (4) proposed that some marine reptiles such as leatherback turtles display endothermy instead of inertial homeothermy, thus helping them to feed in cold waters. However, Lutcavage et al. (5) showed that the studied gravid female specimens raised their metabolic rates because of egg laying, thus biasing the evaluation of their true metabolic status. Most biologists agree that full or incomplete endothermy arose several times during species evolution and developed independently in several lineages. For example, partial endothermy is known in sharks, tunas, and even in some insects and flowers (68). The origin and spreading of endothermy are still a matter of great debate (911); its oldest occurrence could be as early as the Permian, with the appearance and radiation of the Synapsida. Among archosaurs, mass homeothermy or even endothermy have been proposed for dinosaurs (12) and pterosaurs (13) and suggested for the ancestors of crocodilians, because of the existence of a four-chambered heart, which modern crocodiles share with mammals and birds (11).

Large marine reptiles, including ichthyosaurs, plesiosaurs, and mosasaurs, inhabited the oceans from the Triassic to the Cretaceous. They represent three different lineages that became secondarily adapted to a marine mode of life. Ichthyosaurs evolved from basal neodiapsid reptiles, with the most obvious aquatic adaptations: a dolphin-like streamlined body without a neck, paddles, and a fish-like tail. Plesiosaurs are derived diapsids, which belong to the Sauropterygia, the sister group of the Lepidosauria (lizards and snakes). They are highly adapted for submarine locomotion, with powerful paddle-like limbs and heavily reinforced limb girdles. Motani (14) already discussed the possibility that plesiosaurs could not have had a typical reptilian physiology, thus indicating high metabolic activity. Mosasaurs constitute a family of Late Cretaceous varanoid anguimorphs highly adapted to marine life; they are derived lepidosaurs with an elongate body, deep tail, and paddle-like limbs (15). Both tooth morphology and the stomach contents of these three groups of marine reptiles indicate predatory behavior. Their anatomy could afford high cruising speeds and a basal metabolic rate similar to that of modern tunas (16). Moreover, the bone structure of adult plesiosaurs and mosasaurs corresponds to that of large pelagic marine predators designed for long cruises in open waters (17). The stomach contents of these marine reptiles have revealed that they were predators feeding on highly diverse foods, including other marine reptiles, fish, cephalopods, and crinoids (18). The metabolic status of these aquatic reptiles was also investigated through histological bone studies. It has been shown that Jurassic ichthyosaurs had rapid postnatal growth, followed by intense bone remodelling that could be related to a sustained metabolic rate close to that of marine endotherms (9).

Adaptation to cold marine waters was also revealed by the fossil reptile assemblage discovered in the Aptian southern high-latitude deposits of the White Cliffs in southeast Australia (19). The specimens were attributed to at least three families of plesiosaurs and at least one of ichthyosaurs. Paleoclimatic proxies indicate cold to near-freezing conditions at the seasonal scale, a climate mode that is not tolerated by modern ectothermic reptiles such as turtles or crocodiles. This observation suggests that some Mesozoic marine reptile taxa were able to cope with low-temperature marine environments (19). In this study, we investigated the metabolic status of ichthyosaurs, plesiosaurs, and mosasaurs using the oxygen isotope compositions (δ18O) of their phosphatic tissues.

The δ18O value of vertebrate phosphate depends on both body temperature and the composition of ingested water (20). In the case of the studied reptiles, the estimated body temperatures recorded in the δ18O value reflect that of the blood during tooth development; this should be also the case for some plesiosaurs that had protruding teeth. Within each studied reptile group, the taxonomic resolution is at the family level. Assuming that both reptiles and fish lived in the same water mass, differences in their δ18O values would reflect differences in body temperature. To estimate the dependence of reptile body temperature on that of ambient water, we reported the difference in δ18O value between coexisting marine reptiles and fish (belonging to the same sedimentary bed) as a function of the fish δ18O value, which has been proven to be a valuable proxy of seawater temperature (21). A compilation was performed by combining 53 new (22) and 27 published (2327) (table S1) δ18O values of coexisting marine fish and reptile tooth remains recovered from worldwide sedimentary deposits of Triassic to Cretaceous ages. Figure 1 illustrates the principles of the relationship between coexisting reptile and fish temperature differences and seawater temperature. If reptile body temperatures mimic those of fish, δ18O pairs are expected to plot along the horizontal null line. In the hypothetical scenario where reptile body temperature is constant and nearly independent of ambient water temperature, the isotopic pairs should lie on or close to a line with a slope of –1. Paired δ18O data (Table 1) are compatible with linear distributions whose slopes for ichthyosaurs, plesiosaurs, and mosasaurs are –1.38 ± 0.20 [coefficient of determination (R2) = 0.937], –1.08 ± 0.27 (R2 = 0.480), and –0.70 ± 0.20 (R2 = 0.588), respectively. Significant scattering in paired data is observed in Fig. 2, and it exceeds uncertainties associated with analytical measurements. Diagenetic alteration cannot be excluded for some samples, even though tooth enamel was favored because of its remarkable resistance to postdepositional alteration and recrystallization (28). Reptiles and fish collected from the same sedimentary bed may not be strictly contemporaneous, depending on how much time was condensed in the sedimentary layer. Contemporaneous reptiles and fish could have recorded distinct sea surface temperatures or water δ18O, because they lived at various depths or migrated seasonally for hunting or reproduction. However, the observed linear correlations are robust enough to indicate that reptile body temperature does not vary significantly with seawater temperature, except for mosasaurs, whose slope could suggest that body temperature could slightly decrease with decreasing ambient water temperature. Indeed, the range in δ18O of fish close to 3 per mil (‰) means a temperature variation of 13°C, according to the slope of the oxygen isotope fractionation equation between fish phosphate and water [temperature (°C) = 113.3 – 4.38 (δ18Ophosphate – δ18Owater)] that was determined by Kolodny et al. (21). This temperature range is valid only if the δ18O value of surface marine waters was constant at various latitudes and for distinct water masses.

Fig. 1

Model variation of the differences in the δ18O of tooth phosphate between marine reptiles and fish against the variation of the δ18O of fish teeth, assuming (1) an ectothermic and poikilothermic reptile [body water δ18O and body temperature (T) equal seawater δ18O and seawater temperature]; (2) an endothermic reptile with body temperature ranging from 35°C (solid black line) to 39°C (dashed black line) and body water 2‰ enriched relative to a seawater value of 0‰; and (3) an endothermic reptile with body temperature ranging from 35°C (dashed gray line) to 39°C (solid gray line) and body water 2‰ enriched relative to a seawater value of –1‰. For comparison, ichthyosaur (circles), plesiosaur (triangles), and mosasaur (squares) values are reported. V-SMOW, Vienna SMOW values.

Table 1

Mean δ18O values of tooth phosphate from worldwide Mesozoic ichthyosaurs, mosasaurs, and plesiosaurs, as well as coexisting marine fish.

View this table:
Fig. 2

Differences in the δ18O of tooth phosphate between the three marine reptiles [(A) ichthyosaurs, (B) plesiosaurs, and (C) mosasaurs] and fish are reported against the δ18O of fish teeth, approximating the body temperature differences between coexisting reptiles and fish and the seawater temperature where they lived (upper axis). The following reduced major axis regression lines with their 95% confidence limits are drawn: y = –1.38 (±0.20) × +26.53 (±3.07), R2 = 0.94 (ichthyosaurs); y = –1.08 (±0.27) × +20.92 (±5.02), R2 = 0.48 (plesiosaurs); and y = –0.70 (± 0.20) × +12.89 (±3.37), R2 = 0.59 (mosasaurs). Numbers refer to localities given in Table 1. Ichthyosaur and fish samples 11 from Monte San Giorgio should be considered cautiously for the following reasons: (i) a poor knowledge of the water salinity and consequently of the δ18O value of ambient water. (ii) Triassic ichthyosaurs were not tuna-shaped yet, so their thermoregulation was most likely not well developed and difficult to compare with that of other Jurassic specimens. However, removing these data does not significantly change the slope value of the regression line (–1.336 instead of –1.377).

It is known that budgets of evaporation and precipitation over the oceans are responsible for various trends between δ18O values of seawater and salinity (29). These relationships are difficult to apply to the past, especially for geological periods as old as the Mesozoic. At first order, we can consider that there is an 18O enrichment of water by at least 1‰ relative to the mean ocean composition under low latitudes. Seawater tends to be 18O-depleted by at least 1‰ when reaching latitudes above 50° as a consequence of precipitation dominating over evaporation; the δ18O value can be even lower in the presence of continental masses with fluvial discharge, as observed in the present-day North Atlantic ocean off Canada, Greenland, and Norway. Consequently, the range in Mesozoic fish δ18O values must be corrected by at least 2‰ because of the δ18O-salinity latitudinal gradient. In other words, the observed range of fish δ18O values is translated into a temperature range of at least 25°C. Lécuyer et al. (26) have shown that the δ18O of the global ocean most likely ranged from –1 to 0‰ [standard mean ocean water (SMOW) values] throughout the Jurassic and Cretaceous. Accordingly, the lowest temperatures of 12° ± 2°C correspond to the highest fish δ18O values approaching 22‰, and the highest temperatures of 36° ± 2°C correspond to the lowest fish δ18O values close to 18‰ (Fig. 2). Ichthyosaurs and plesiosaurs have δ18O values similar to those of fish at a corresponding temperature range of 26° ± 2°C estimated from fish and seawater δ18O values of 19.5‰ and –1 to 0‰, respectively (Fig. 2). The regression line for mosasaurs intercepts the horizontal null line at a lower δ18O value of 18.5‰, thus indicating a possible higher body temperature of 30° ± 2°C. These temperature ranges are the minimal values that can be considered for ichthyosaur, plesiosaur, and mosasaur body temperatures if the δ18O of their body equaled that of ambient seawater.

However, aquatic breathing vertebrates have body waters that are slightly 18O-enriched relative to ambient water in the absence of transcutaneous evapotranspiration, and published data for modern aquatic reptiles reveal that this isotopic enrichment does not exceed 2‰ (30, 31). Consequently, the body temperatures of studied ichthyosaurs and plesiosaurs could have been as high as 35° ± 2°C and even close to 39° ± 2°C for mosasaurs, according to Kolodny et al.’s equation (21). Both slope values of linear regressions and estimates of body temperatures are in good agreement with the swimming performances that were modeled for these three groups of marine reptiles. Massare (32, 33) and Motani (14) suggested that ichthyosaurs were pursuit predators, whereas most mosasaurs were ambush predators, not requiring high metabolic rates all the time. Plesiosaurs were considered to have been cruisers, although slower than ichthyosaurs in sustained speed.

The δ18O values of Mesozoic ichthyosaurs and plesiosaurs support the hypothesis that these large predators were able to regulate their body temperature independently of the surrounding water temperature even when it was as low as about 12° ± 2°C. In the case of mosasaurs, we cannot exclude the possibility that their body temperature was partly influenced by the temperature of ambient water. In any case, estimated body temperatures in the range from 35° ± 2°C to 39° ± 2°C encompass those of modern cetaceans (34) and suggest a high metabolic rate required for predation and fast swimming over large distances, especially in cold waters. δ18O data from tooth phosphate reveal the existence of homeothermy for ichthyosaurs and plesiosaurs, and of at least partial homeothermy for mosasaurs, with in all cases a taxonomic resolution that does not exceed the family or infraclass. These three distinct phylogenetic groups of large marine reptiles were able to maintain a body temperature substantially higher than that of ambient marine waters, indicating that some kind of endothermy operated as an internal source of heat.

Supporting Online Material

Materials and Methods

Table S1

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. The authors thank M. Sander, B. McNab, and R. Seymour for preliminary discussions of these data during the workshop dedicated to “Sauropod gigantism” that was held in Bonn, Germany, in November 2008. We are also grateful to D. Brinkman and J. Gardner (Royal Tyrrell Museum), S. Etches and J. Clarke, G. Suan, J. Lindgren (Lund University), B. Kear (La Trobe University), and A. Schulp (Maastricht University) for providing samples. L. Simon and G. Escarguel helped us in the statistical treatment of data. This study was funded by both CNRS and the Institut Universitaire de France.
View Abstract

Stay Connected to Science

Navigate This Article