A fresh analysis turned up in my morning feed: they finally measured the body temperature of a Tyrannosaurus, it came out to about 36 °C — almost like a human. Two paragraphs in, I started recognizing the details. Isotopes in tooth enamel. A calculation based on how often heavy oxygen-18 ends up bonded to heavy carbon-13. Crocodiles from the same formation as a control. A caveat that a giant body might retain heat on its own.
All of this had happened before. In 2011, nearly the same people — Robert Eagle and Aradhna Tripati, then at Caltech — published a paper in Science using the same method with a similar result: 36–38 °C, but on sauropods. And in 1994, Reese Barrick and William Showers had already argued in Science that Tyrannosaurus was homeothermic; that was the first isotope thermometer ever applied to a dinosaur.
The hook turned out not to be the Tyrannosaurus's temperature. What's interesting is something else: science can measure the same thing for over thirty years, get converging numbers, and still not close the question. I went to figure out what exactly is being measured here, and why the same answer solves nothing.
Enamel is the only skeletal tissue that doesn't remodel. Bone and dentin dissolve and redeposit dozens of times over an animal's lifetime, while enamel forms once, at the edges of the tooth, and after that only wears down. For a paleontologist, this means a logger: the isotopic signature of the moment when the tooth was growing stays locked inside the mineral for tens of millions of years.
Here's how it works. In the carbonate-containing apatite of the tooth, the heavy isotopes ¹³C and ¹⁸O have a thermodynamic tendency to bond with each other. The lower the mineralization temperature, the more of these "clumped" pairs there are. The parameter measured is Δ₄₇: the sample is dissolved in phosphoric acid, the released CO₂ is captured, and the proportions of different isotope combinations are counted. After that — pure equilibrium arithmetic, converted into degrees.
The classical isotope thermometer based on δ¹⁸O worked differently: it required knowing the isotopic composition of the water the animal drank. We don't have Cretaceous water, and this was the main gap — the estimate had to be propped up with assumptions. Clumped isotopes give temperature without water. The method's accuracy based on calibrations is about 1 °C, reproducibility 1–2 °C.
The method didn't exist at all in the 1960s. The hypothesis that "dinosaurs were warm" was pulled out of morphology by John Ostrom: in 1964 he found a predator in Montana, described in 1969 as Deinonychus antirrhopus, and the bird-like skeleton with a sickle claw on its foot sat poorly with the image of a sluggish lizard. The frontispiece for his monograph was drawn by a student — Robert Bakker, who three years later turned this hunch into a research program.
Bakker's arguments weren't about chemistry, but about the accounting of ecosystems. First: the annual energy budget of a warm-blooded population is 10–30 times higher than that of cold-blooded animals of the same weight, and since for 140 million years large terrestrial roles were held by dinosaurs, not mammals, a cold-blooded animal simply couldn't sustain losing for that long. Second: the predator-to-prey ratio in fossil layers resembles what you see in communities with high metabolism, and doesn't match what modern ectotherms produce. Third: bone histology with high growth rates.
In 1972 this came out as a paper in Nature, in 1975 as a big cover story in Scientific American titled "Dinosaur Renaissance," where Bakker stated directly: dinosaurs didn't go fully extinct, one of their groups is alive, we call them birds. This was a public quarrel, and there was far more about departmental reputations in it than about isotopes.
The isotope answer came in 1994. Barrick and Showers took δ¹⁸O of phosphate in Tyrannosaurus bones and got what sounds fairly dry in the abstract: the animal maintained homeothermy with a temperature spread of less than 4 °C, and this requires a metabolism at the level of modern warm-blooded animals. It's worth looking closely at what exactly they measured: the stability of temperature, not its magnitude.
In 2011, Eagle and colleagues moved to clumped isotopes and to enamel: 11 sauropod teeth from Tanzania, Wyoming, and Oklahoma, Giraffatitan at about 38.2 °C, Camarasaurus at about 35.7 °C. The abstract in Science ended with a puzzle: the numbers came out 4–7 °C lower than what the model linking body temperature to mass had predicted. The giant should have overheated and died, but somehow it didn't overheat. The authors suggested that sauropods had heat-dumping mechanisms — air sacs, long necks and tails acting as radiators.
The new paper came out on September 16 in Science Advances. Three teeth: two from one skeleton — a young adult specimen LACM 150167 with the museum name "Thomas" and a living mass of over three tons — and one isolated tooth from another Tyrannosaurus. All from the Hell Creek Formation in Montana — the same one where Barnum Brown dug up the first Tyrannosaurus in 1902.
The individual teeth yielded 37.3, 35.9, and 34.7 °C, averaging 36.3 ± 2.5 °C. Crocodilians from the same formation, five teeth, averaged 30.9 ± 2.6 °C, and the difference between the Tyrannosaurus and them is statistically significant. For comparison: bivalve mollusks from the same location give 25.9 ± 1.2 °C as a proxy for environmental temperature, while a paleoclimate model for these latitudes gives about 17.8 °C mean annual at one CO₂ level and 20.7 °C at another. That is, the animal sat roughly 15 degrees above mean annual air temperature.
Next, the study takes a step that wasn't there in 2011: it tests whether these 36 °C can be written off to size. A model for an ectotherm weighing over three tons gives a temperature noticeably below what was observed — meaning heating from the sun alone doesn't explain it. This very model was the skeptics' main argument.
Plus, the climate reconstruction shows that Tyrannosaurus could have lived across nearly the entire North American continent, including high latitudes with below-freezing winter temperatures, and juvenile tyrannosaurids have indeed been found in Alaska at a paleolatitude around 85°.
Co-author Robert Eagle comments on the result with restraint: the temperature is about what he expected — higher than a reptile or a slow mammal like a sloth, lower than a bird's. Thirty-two years between the first isotope number and this one, and the answer matched the specialist's intuition. The point of the whole exercise isn't to learn the degrees: before, any "they were warm" ran into "that can be explained by size," and now the objection is addressed to a specific model.
Headlines said: Tyrannosaurus was warm-blooded. The paper itself says it more cautiously: body temperature is not an unambiguous indicator of physiology and metabolism, and the conclusion is limited to homeothermy and possible — "possibly" — tachymetabolic endothermy.
Behind this caution lies twenty years of dispute that 36 °C does nothing to settle. In 2014, Felisa Smith's group modeled dinosaur energetics through growth rates and compared them with living clusters. Dinosaurs fell between endotherms and ectotherms — closest to mesotherms: tuna, lamnid sharks, leatherback turtles, and the echidna. These are animals that raise their body temperature through metabolic heat, but don't defend a set point. The echidna holds its set point around 31 °C, but can shift it by 10 degrees in a day just following the weather. That paper came to a separate, biting conclusion: the very dichotomy of "warm-blooded or cold-blooded" is too crude a framework.
Then new lines of evidence followed. In 2020, eggshells were added: all three major dinosaur clades showed warm temperatures, meaning metabolically controlled thermoregulation is the ancestral condition for the entire group. In 2022, a Yale team working on biomolecules in bones announced that endothermy arose as early as the common ancestor of dinosaurs and pterosaurs. In 2023, paleontologists from Davis demolished this conclusion with three points: resting metabolism and active metabolism are different things (the Komodo dragon goes from a low resting metabolism to a very high one in motion), preservation of biomolecules in bones isn't calibrated, and the margins of error are too large to pin endothermy on the common ancestor; more likely it appeared later and independently.
The latest twist to date — 2024, when another group looked at 1,000 fossil finds, climate models, paleogeography, and phylogenetic trees. Theropods and ornithischians, according to this work, moved into cold latitudes in the Early Jurassic, about 183 million years ago, around the Jenkyns Event — and that's apparently where endothermy was acquired. Sauropods stayed in warm zones and grew into giants, which also fits the picture of physiological constraints. Thirty years, four independent research programs, none fully agreeing with the others.
Let me start with the main thing. Inside the crocodilian sample there's a tooth that yielded 38.2 ± 2.8 °C — warmer than the Tyrannosaurus average. The authors note this honestly and explain that the value falls within two standard errors of the physiologically normal range for a crocodile of 30–35 °C, and that a single outlier means nothing where you need a sample. But the entire "Tyrannosaurus vs. crocodile" difference is the difference of two means calculated from 13 measurements, each of which itself carries an error of 2–4 degrees. Thirty-six plus or minus two and a half — a range that encompasses not just an elephant, but a large leatherback turtle.
Next. The equivalent-size ectotherm model that the authors refute is someone else's work and is itself contested: in the second half of the last century a different calculation stood in its place, and in 2013 it was taken apart in a separate paper showing that it poorly predicts body temperature even in living birds and crocodiles. One model is pitted against another model, and both are about an animal no one has ever seen.
Another detail, small and therefore unpleasant. The fractionation equation used to reconstruct the isotopic composition of water from Tyrannosaurus tooth phosphate is taken from modern birds. The authors explicitly acknowledge that it may not be the perfect model for Tyrannosaurus. And the fractionation equation for crocodilians is from modern crocodiles.
And lastly. The reference list of the new paper contains not a single work on mesothermy and not the word "mesotherm." The 2014 dispute about whether Tyrannosaurus might have had tuna-like physiology simply isn't raised — the study plays on the field of "endotherm or ectotherm," and is silent about the existence of a third cell on that field. I won't claim this is an omission: perhaps the authors consider the question closed, and they have grounds for that. But it looks as though the answer was obtained to a question that was slightly reformulated ten years ago.
What's left when you boil it down. The temperature of Tyrannosaurus is now known about as well as the temperature of an elephant, and that's a real result: 36 in the third decade of repeats stops being a guess. But what this number means for its metabolism is still the subject of four incompatible research programs, and none of them turns on isotopes. We've built the thermometer. We're still arguing about what it reads.