Dinosaur Body Temperatures Determined from Isotopic (13C-18O) Ordering in Fossil Biominerals

See allHide authors and affiliations

Science  22 Jul 2011:
Vol. 333, Issue 6041, pp. 443-445
DOI: 10.1126/science.1206196


The nature of the physiology and thermal regulation of the nonavian dinosaurs is the subject of debate. Previously, arguments have been made for both endothermic and ectothermic metabolisms on the basis of differing methodologies. We used clumped isotope thermometry to determine body temperatures from the fossilized teeth of large Jurassic sauropods. Our data indicate body temperatures of 36° to 38°C, which are similar to those of most modern mammals. This temperature range is 4° to 7°C lower than predicted by a model that showed scaling of dinosaur body temperature with mass, which could indicate that sauropods had mechanisms to prevent excessively high body temperatures being reached because of their gigantic size.

For the majority of the time since dinosaurs were first named in 1842, it was assumed that their metabolism was similar to ectothermic “cold-blooded” reptiles that derive the heat they need to function from the environment, rather than endothermic “warm-blooded” mammals and birds, which have higher and more stable body temperatures regulated by internal metabolic heat production. However, in the 1960s and 1970s, evidence began emerging that endothermy could be more consistent with observations on the behavior, paleogeographic distribution (for example, polar dinosaurs), and anatomy of nonavian dinosaurs (13). The initial case for dinosaur endothermy was largely made on the basis of interpretations of the inferred physical performance and behavior of dinosaurs, such as estimating running speeds from preserved tracks and predator/prey ratios determined by comparing biomass estimates from the fossil record to those ratios in modern ecosystems (35). These methods have been extensively debated (5) and have sparked several decades of study on dinosaur thermoregulation by using biophysical and behavioral modeling (610), bone histology and growth rate analysis (11), anatomical observations (12), and oxygen isotope paleothermometry (1315). Despite progress in these methods, a consensus opinion has not been reached (16, 17).

Sauropod dinosaurs are the largest terrestrial animals that have ever lived, and therefore understanding their physiology poses a particular challenge (18). Perhaps the most convincing argument in favor of endothermy in sauropod dinosaurs comes from the analysis of bone histology, which suggests very high growth rates that could not be sustained by a low basal metabolic rate (11, 1822). Conversely, the case for ectothermy in sauropods has been made by modeling heat exchange by animals with the environment, suggesting that endothermic sauropods would have severe problems with overheating (6, 2325). Recently, Gillooly et al. presented a biophysical model that is based on allometric scaling laws and dinosaur growth rate analysis, predicting that dinosaur body temperatures would increase as their mass increased, reaching over 40°C for the largest sauropods (9). Such models imply that dinosaurs were ectotherms, but that some dinosaurs would achieve high body temperatures because of their large mass. This phenomenon has been termed “gigantothermy” or “inertial homeothermy” and can be observed in some modern ectotherms, such as leatherback turtles (23, 24, 26). However, this modeling result was not supported by a study on smaller dinosaur taxa that used enamel phosphate oxygen isotope isotopes to reconstruct dinosaur body temperatures of around 33° to 38°C from Cretaceous dinosaurs with body weights spanning the range of 10 to 9000 kg (15).

We applied a different approach to this problem, using clumped isotope thermometry to determine the body temperatures of large Jurassic sauropods by analyzing material from six sites (figs. S1 and S2) (27). This technique is founded on the thermodynamic preference of rare heavy isotopes of carbon (13C) and oxygen (18O) to bond with each other (13C-18O), or “clump,” in carbonate-containing minerals (28, 29). Unlike the well-established oxygen isotope thermometer, application of clumped isotope thermometry is not dependent on knowing or assuming the oxygen isotope composition of the water from which a mineral grew (28). The parameter measured in this approach is the ∆47 value [supporting online material (SOM) text] of CO2 liberated from the carbonate component of tooth bioapatite [generalized as Ca5(PO4, CO3)3(OH, CO3, F, Cl)] (28). Bioapatite ∆47 values follow a temperature dependence indistinguishable from inorganic calcite (CaCO3) over the range of 22° to 37°C, and a theoretical model predicts that this should be the case even over a greater range of temperatures (29). This approach is capable of reconstructing the expected body temperatures of modern and fossil mammals and ectotherms with an accuracy of ~1°C and a precision (1 SE) of 1° to 2°C. Comparison with instrumental measurements of reptile body temperatures indicates that clumped isotope measurements of teeth probably reflect average body temperatures in ectotherms, rather than peak body temperatures (which could be similar to those seen in mammals in some cases) (7, 29, 30).

Fossils of Brachiosaurus brancai [recently assigned to a new genus, Giraffatitan (SOM text)] and a sauropod tentatively identified as being from the subfamily Diplodocinae from the Tendaguru Beds in Tanzania were analyzed, as were fossils of Camarasaurus sp. from five sites in the Morrison Formation (SOM text). A crucial component of our approach was the examination of whether the fossil material preserves primary physiological information or instead reflects isotopic modification during burial alteration. It has been shown that 13C-18O bond ordering can be preserved in the carbonate mineral lattice of calcite much older than Jurassic in age, providing that isotope exchange does not occur with the environment through dissolution/reprecipitation reactions or diffusion (the latter is a factor only at temperatures greater than ~200°C) (31). We focus our interpretations on tooth enamel as the most desirable record of primary growth temperature on the basis of multiple lines of evidence suggesting that the large and closely packed apatite crystals in enamel, as well as its organic-poor nature, can permit the preservation of geochemical signatures even over long time scales (32, 33). We have adopted several approaches to establish the state of preservation of each sample. First, we analyzed dentin from the same teeth, bone, and sparry calcites from each site in order to define diagenetic end-members, with the expectation that well-preserved enamel should yield distinct δ13C and δ18O values when compared with materials that are known to be altered or are secondary precipitates (fig. S3). Because the conditions associated with alteration will likely differ from primary (living body) conditions, the calculated Δ47-based temperatures and water δ18O compositions should also be distinct, barring a fortuitous similarity (fig. S3). Second, we determined phosphate δ18O (δ18OPO4) for each specimen analyzed and compared these values with δ18O values of carbonate in apatite (fig. S4). Because the oxygen isotope composition of phosphate groups is thought to be especially well preserved over geologic time scales, deviations from this offset can be used as an indicator of alteration of the carbonate component because of reactions with diagenetic fluids (fig. S4). Thin sections of representative tooth samples were also examined petrographically (fig. S5 and SOM text).

On the basis of these parameters, we were able to identify samples that were clearly altered and some samples that appear to be well preserved (table S1 and SOM text). From apparently well-preserved tooth enamel material from the Tendaguru Beds, we measured average body temperatures of 38.2 ± 1.0°C (1 SE) from three Brachiosaurus teeth and 33.6 ± 4.0°C from two sauropod teeth of a Diplodocinae (Table 1). Three of four Camarasaurus teeth analyzed from the Morrison Formation in Cimarron County Oklahoma appeared to be well preserved and yielded an average temperature of 36.9 ± 1.0°C—a value that decreases slightly to 35.7 ± 1.3°C when a well-preserved Camarasaurus tooth from the Howe Ranch Quarry in Wyoming is included (Table 1). All material from the Utah Quarries and Nail Quarry, Wyoming, appeared to be altered (SOM text). Because their precise taxonomic identification was uncertain and temperature determinations were also relatively uncertain owing to the few measurements that were possible (Table 1), we did not consider data from Diplodocinae in the physiological interpretations of our data (Fig. 1) and instead focused on the more precisely constrained data from Brachiosaurus brancai and Camarasaurus sp.

Table 1

Δ47 derived body temperature determinations on well-preserved dinosaur tooth enamel. Number of analyses represent the total number of Δ47 measurements made on distinct extractions of CO2 gas from tooth enamel material from each locality. Values for individual measurements and averages for each individual tooth specimen are given in tables S4 to S9. When average Δ47 values are calculated for each species, they are from the average of each tooth specimen and not of each individual analysis on distinct CO2 extractions. Errors are ±1 SE of the average Δ47 and the propagated error in temperature calculations.

View this table:
Fig. 1

Comparison of measured dinosaur body temperatures to a previously published model calling for scaling of body temperatures with body mass (9). Crocodile data was derived from modern species. Dinosaur body temperature determination from modeling were for the following taxa (in descending mass in this plot); Sauroposeidon proteles, Apatosaurus excelsius, Tyrannosaurus rex, Daspletosaurus torosus, Gorgosaurus libratus, Albertosaurus sarcophagus, Massospondylus carinatus, and Psittacosaurus mongoliensis. S. proteles body temperature was calculated by Gillooly et al. from extrapolation of their model rather than directly from growth rates (9). Camarasaurus sp. and Brachiosaurus brancai (green square and red square, respectively) body temperature determinations from this study are also presented. We have plotted the average of body mass estimates from the literature versus clumped isotope–derived body temperatures for each sauropod. Body mass estimates are taken from the compilation of Sander et al. (18). Error bars in the horizontal axis represent the range of estimates of body mass reported in the literature. Vertical error bars represent 2 SE of the clumped isotope temperature determinations.

In general, body temperatures of vertebrates reflect the combined influences of metabolism, size, environmental temperature, and in some cases, specialized physiological strategies for heat regulation. Therefore, our data taken in isolation are not unambiguous indicators of endothermy versus ectothermy. However, they do place quantitative constraints on sauropod physiology, limiting the range of possible thermoregulatory strategies.

Clumped isotope body temperature determinations from Brachiosaurus brancai and Camarasaurus sp. are 5° to 12°C higher than those previously measured from modern and fossil crocodiles and alligators, are within error to those measured from modern and fossil mammals (7, 29, 30), but are lower than most modern birds, which often have body temperatures in excess of 40°C (34). This observation should not be taken as a simple indication that sauropods were endothermic, however; our measured temperatures for both Camarasaurus sp. and Brachiosaurus brancai are lower than those predicted according to the model of Gillooly et al., even when lower estimates of body mass from the literature and the 2 SE range in the clumped isotope measurement are used (Fig. 1) (9). This result cannot easily be explained as an artifact of burial alteration because diagenetically altered enamel material examined in this study consistently records hotter temperatures than that of material that is inferred to be well preserved (fig. S4 and SOM text). Our data probably reflect temperatures of tooth formation. Further work with models and observations will be required to determine whether large thermal gradients could have existed across the large-bodied sauropods, although this is not observed in our clumped isotope data from modern taxa, in which isotopically determined body temperatures from teeth were in good agreement with expected body temperatures (29). If our temperature determinations are accurate, then in order for sauropod body core temperatures to be similar to those predicted by the model of Gillooly et al. there would have to be a 4° to 7°C temperature gradient between the site of tooth formation and the body.

Therefore, the largest sauropods had body temperatures cooler than the Gillooly model suggests. One possible explanation of this result is that adult sauropods had mechanisms to prevent excessively high body temperatures being reached and so could regulate their body temperatures to some extent. For example, they may have had a tracheal surface and air sac system that served as an internal cooling system (35, 36), and their long necks and tails could have facilitated heat dissipation by increasing their surface area (37). Overall, our data are most consistent with the hypothesis that sauropods sustained high metabolic rates during ontogeny to reach their gigantic size so rapidly, but that in maturity a combination of physiological and behavioral adaptations and/or a slowing of metabolic rate prevented problems with overheating and avoided excessively high body temperatures (18, 36). An unresolved question is whether such adaptations could have compensated for the high internal heat production associated with endothermy, or whether large adult sauropods must have had both heat-dissipating adaptations and a low basal metabolism to maintain body temperatures in the 36° to 38°C range that we have measured.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S6

Tables S1 to S10

References (38–103)

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. Acknowledgments: R.A.E. and J.M.E. are funded by National Science Foundation grant EAR-1024929. T.T. is funded by Deutsche Forschungsgemeinschaft (DFG) grants TU 148/2-1 and TU 148/4-1. A.K.T. is supported by the Division of Physical Sciences at UCLA and National Science Foundation grant EAR-0949191. We thank the Utah Museum of Natural History, the Oklahoma Museum of Natural History, the Berlin Museum für Naturkunde, the Aathal Dinosaurier Museum, D. Chure at Dinosaur National Monument, and the Wyoming Dinosaur International Society for the provision of specimens for analysis. We thank D. Schwarz-Wings for specimen identification and D. Goodreau for curatorial assistance. We also thank the stable isotope laboratories at Tübingen and Lyon for conducting phosphate δ18O measurements. Lastly, thanks to J. Gillooly for provision of published data. This is contribution number 113 of the DFG Research Unit 533 “Biology of the Sauropod Dinosaurs: The Evolution of Gigantism.” R.A.E. designed the project and experiments, performed the experiments, analyzed the data, and wrote the manuscript. J.M.E. and A.K.T. assisted with experimental design and data interpretation. Clumped isotope analyses were conducted in the lab of J.M.E. T.T. provided samples for the study, conducted phosphate oxygen isotope analysis, and assisted with data interpretation. T.S.M. assisted with isotope measurements and petrographic analyses. H.C.F., M.C., and R.L.C. provided samples for analysis and insights into the provenance of samples and data interpretation.
View Abstract

Stay Connected to Science

Navigate This Article