Tuesday, 31 May 2022

The mad, mad, mad stilt-legged tyrannosaurid hatchling (and other efforts at reconstructing baby tyrants)

By now many readers will have seen the excellent, excellent BBC/Apple TV+ documentary Prehistoric Planet, a show being widely celebrated as the most significant televisual portrayal of dinosaurs since 1999’s seminal Walking with Dinosaurs. There’s a huge amount to say about the reaction to Prehistoric Planet as well as how it was put together. I’m proud to say I was involved as a concept artist and consultant but, for now, those of us involved with the show have been (understandably) asked not to give too much away. We can, nevertheless, use the programme as a jumping-off point to discuss some cool dinosaur science, and that’s what I want to briefly do today.

One of the most memorable scenes in Prehistoric Planet is the opener of the first episode, and you might be familiar with it even if you haven’t seen the full series. It’s the sequence featuring the swimming Tyrannosaurus and his juveniles crossing a small body of water to engage in some beach-related foraging. Part of this scene featured in a promotional teaser that dropped so many jaws earlier this year, where Tyrannosaurus juveniles hunted turtles on a beach. If you've somehow missed this, here it is.

The official sneak-peek for Prehistoric Planet, from the Apple TV+ YouTube channel.

It’s easy to buy these superbly rendered, fluffy, long-legged and big-eyed baby tyrants as 100% accurate takes on such animals because they look so real. But under their wireframes and digital rigging, these juvenile dinosaurs are based, in fact, on hypothetical reconstructions, not fossil skeletons. Despite the intense interest in Late Cretaceous North American fossil sites, hatchling tyrant dinosaurs are virtually unrepresented by fossils and the first genuine remains of such animals were only described in 2021, by Greg Funston et al. These important finds constitute a piece of lower jaw, a tooth, and a foot claw from embryonic individuals of Daspletosaurus and Albertosaurus. Each sheds a little light on the size and proportions of baby tyrannosaurs, but we’re still far, far away from a complete picture of their entire osteology. Thus, to date, the only way to restore the life appearance of baby tyrants with some degree of scientific rigour has been to extrapolate their proportions from older, larger individuals. We’ve been attempting this activity for at least 50 years and, even now, we struggle to make accurate, or at least statistically sound, models of baby tyrant appearance.

The first effort at predicting hatchling tyrant anatomy was published in Dale Russell’s 1970 paper on the tyrant dinosaurs of Canada. Russell predicted the proportions of a hatchling “Albertosauruslibratus (=Gorgosaurus libratus today) by calculating scaling trajectories of larger tyrants and projecting them to an individual with a 100 mm long femur. The resultant image is one that might be familiar to many readers, as it has appeared in subsequent dinosaur books and papers. I certainly remember my own first encounter with it in David Norman’s influential 1988 Normanopedia Illustrated Encyclopaedia of Dinosaurs, as it looked very different to the pop-culture plush toy-like baby dinosaurs I was familiar with (hello, The Land Before Time et al.).

Dale Russell's hypothetical Gorgosaurus libratus hatchling, from Russell (1970).

Russel’s prediction shows baby tyrants as long-legged, long-tailed and small-bodied animals that were very different to their parents, a consequence of the pronounced allometry that shaped tyrannosaurs as they grew. The overall animal, he calculated, was dog sized: just over 75 cm from the snout to the end of the tail. Some aspects of Russell’s model, including the shapes of the bones themselves, were created by “juvenilising” the bone shapes of adult albertosaurines and, even today, it's a good-looking skeletal. It certainly matches what we see in other non-avian dinosaur embryos and hatchlings, which are generally leggy animals adapted for precocial, relatively independent early lives. It’s probably this skeleton that a lot of palaeoart and dinosaur documentaries have based their baby tyrants on as, all these years later, it's still one of the only published hatchling tyrannosaur skeletals available.

In 2003 another attempt to predict tyrannosaur hatchling proportions was made by Phil Currie. Armed with more specimens and thus more metric data, Currie was able to provide more informed tyrant scaling regimes and, in theory, give us a superior prediction of hatchling proportions than Russell’s effort of three decades prior. As before, his scaling was to an individual with a 100 mm long femur, and the result was, again, dog-sized. Currie didn't specify what tyrannosaur he was modelling specifically, so his data can be viewed as a generic tyrannosaurid, although albertosaurine scaling metrics were prioritised where they differed significantly from those of tyrannosaurines. Because Currie (2003) only presented the results in a data table and did not illustrate the resultant hypothetical hatchling, I took the liberty to restore the animal with his suggested proportions myself*. And, holy moly…

The stilt-legged tyrant hatchling, restored after the dimensions given in Currie (2003). As explored below, the data behind this reconstruction were identified as unbelievable in Currie's original paper: this is not what Phil Currie (or anyone else) seriously thinks juvenile tyrannosaurus looked like! But it is what models of tyrannosaur ontogeny, as understood in 2003, predicted for a tyrant of hatchling dimensions.

…those are some l-o-n-g legs. They’re slightly maddening to look at, in fact. Like it’s been teleported in from the Nightmare Before Christmas meets Prehistoric Planet Super Crossover Spectacular. While the rest of the animal isn’t too dissimilar to Russell’s baby gorgosaur, the tibia is 60% longer and the metatarsals 220% longer, creating a tiny dinosaur on stilts. The result is a theropod with leg proportions outside of anything known within Dinosauria: even long-legged birds, like stilts, secretary birds and seriemas, are being put to shame. Needless to say, if anyone seriously thought that this is what baby tyrant dinosaurs looked like we’d be considering all sorts of ecological and biomechanical implications. Were baby tyrants waders? Specialists of prairies and open habitats? Did they make money on the side at Mesozoic carnivals as stilt walkers? What the heck is going on?

*I had to take lengths of the vertebral column and manus from Russell (1970), as these were not modelled in Currie (2003).

But, of course, no-one does think that this is what hatchling tyrannosaurs looked like. The outlandish hindlimb proportions of this model were dismissed as unbelievable by Currie (2003) as quickly as they were introduced, and the whole exercise ended with a blunt cautionary note: “there are limitations to what can be done in extrapolating this data.” (Currie 2003, p .663). And before anyone asks, it's worth stressing that Currie's values were not errors or the outputs of wrongful methodology: they're simply what fell out of legitimate, perfectly normal scaling equations. The "limitations" Currie referred to are fundamental difficulties with scaling the proportions of extinct animals outside of observed ranges. Such issues include low sample sizes, our scaling curves being skewed by outliers and overall low confidence intervals, but it's unusual to talk about these as problems for scaling tiny animals. We're generally much more used to these frustrations when trying to scale smaller animals to the size of giant adults. 

And strange as it seems, I admit to finding Currie’s stilt-limbed tyrant baby a more profound example of these issues than our usual subjects of scaling woes, like titanic sauropods or giant pterosaurs, because we know so much more about small dinosaur proportions. It's easy to spot these data as (probably) “wrong” because they're so out of the ordinary, but concerning proportions are harder to intuitively identify in animals beyond the size limits of anything ever witnessed by a person. It's a good reminder that any extrapolation is viewed as shaky from a mathematical perspective, and recalls Knut Schmidt-Nielsen's comments in his classic 1984 book Scaling: why is animal size so important: "unless we fully understand all the pertinent factors involved (and that is not likely to happen in a biological system), going beyond the limits of observation is not simply chancy, but outright perilous” (p. 25). It’s perhaps only when we see relatable, but obviously wrong-looking reconstructions like the one above that we’re confronted with this reality. And while we basically have no choice but to extrapolate and go beyond data limits when dealing with certain unknowns in the fossil record, it's good to come back to Earth from time to time and realise the unreliable nature of the mathematics we're using. So many extrapolations we're familiar and comfortable with might be just as crazy as that lanky-limbed hatchling, we just don't see them as obviously "wrong". 

All this said, that stilt-legged tyrant baby still has some utility. While it probably doesn't genuinely reflect what baby tyrants looked like, it does serve as a great visualisation of the allometry expressed through tyrannosaur ontogeny. If we compare the stilt-legged model with a regular-old T. rex, for instance...

The mad stilt-legged hatchling vs. a large adult Tyrannosaurus, based on "Sue".
...we get a good sense of how tyrant bodies grew into their legs as they progressed from hatchling to adults. We can also see that the head length remained about proportionate to femur length and that the arms grew relatively slowly, appearing to "shrink" a bit en route to adulthood (Currie 2003). Of course, in detail, there's actually a mosaic of different allometries occurring through growth: the head doesn't lengthen proportionately, for instance, but it does deepen. The arms and legs don't really "shrink" as one unit, either, but actually at varying rates in different parts of each limb (Currie 2003). Because the stilt-legged hatchling is our understanding of tyrannosaur scaling regimes expressed to an absurd degree, the reality of tyrannosaur ontogeny probably wasn't quite as extreme as the graphic above suggests, but it might not be too far off. Both Russell's (1970) and Currie's (2003) data point to long-legged, shallow bodied tyrant hatchlings and that's the model we should generally follow, even for chunky, robust animals like T. rex and Tarbosaurus. As noted by Currie in the same 2003 paper, "allometric differences among mature specimens of different [tyrannosaurid] species are shown to be trivial when compared with the allometric differences associated with growth". The stilt-legged tyrant (and more recent work and studies on this same topic) makes that crystal clear.

Currie's paper is now almost 20 years old so, to wrap this up, one last question: have we made any progress in being able to reconstruct baby tyrants reliably in recent years? To be frank, no, we don't yet have the data needed to clear the problems outlined above. Things have moved ahead thanks to Greg Funston and colleagues (2021) who, finally, have been able to factor some embryonic tyrant fossil material into their scaling equations. These have generally pointed to animals of the same size and shape as predicted by Currie and Russell, but their 95% confidence intervals (CI) for many predictions are still enormous (which means, in plain terms, that the true proportions may have been very different to the predicted ones). For instance, Funston and colleagues scaled their Albertosaurus embryo dentary to a body length 715 mm, but with a CI range of 496–897 mm. Their (seemingly) larger Daspletosaurus specimen, represented by a toe claw, scaled to 1101 mm long, with a CI of 250–5954(!) mm. Estimates for their femoral lengths were also not well constrained, at least in Daspletosaurus, with CI values of 76–256 mm. Some of the upper ranges of these CIs are consistent with larger tyrant specimens that are already several years old, and they're thus far from the measurements we'd expect from embryos or hatchlings (Funston et al. 2020). Our most charitable assessment is that we're consistently getting the same basic signals about tyrant hatchling form, so we're probably on the right track, but we have a long way to go before we'll have reliable body proportion predictions for tiny tyrants. The discovery of more small tyrannosaurids would go a long way to helping this but, of course, a skeleton of an embryo or hatchling would be the best and most direct way to resolve any uncertainty. 

Enjoyed this post? Support this blog for $1 a month and get free stuff!

This blog is sponsored through Patreon, the site where you can help artists and authors make a living. If you enjoy my content, please consider donating as little as $1 a month to help fund my work and, in return, you'll get access to my exclusive Patreon content: regular updates on upcoming books, papers, paintings and exhibitions. Plus, you get free stuff - prints, high-quality images for printing, books, competitions - as my way of thanking you for your support. As always, huge thanks to everyone who already sponsors my work!


  • Currie, P. J. (2003). Allometric growth in tyrannosaurids (Dinosauria: Theropoda) from the upper Cretaceous of North America and Asia. Canadian Journal of Earth Sciences, 40(4), 651-665.
  • Funston, G. F., Powers, M. J., Whitebone, S. A., Brusatte, S. L., Scannella, J. B., Horner, J. R., & Currie, P. J. (2021). Baby tyrannosaurid bones and teeth from the Late Cretaceous of western North America1. Canadian Journal of Earth Sciences, 58(9), 756-777.
  • Russell, D. A. (1970). Tyrannosaurs from the Late Cretaceous of western Canada. National Museum of Natural Sciences, Publications, in Paleontology, 1, 1-34.
  • Schmidt-Nielsen, K., & Knut, S. N. (1984). Scaling: why is animal size so important?. Cambridge university press.


  1. This comment has been removed by the author.

    1. There's also MPC-D 107/7 - the juvenile Tarbosaurus, which at 2-3 years old is a pretty handy data point.

    2. Indeed, thanks Scott. This is one of the "several years old" specimens I alluded to in the final paragraph after Funston et al. used it as a sanity check for their own predictions. I could have made that clearer with my wording!

  2. I suspect Currie just extrapolated juvenile proportions assuming a sort of logaritmic curve for his data, but it is more plausible that the complete growth curve was closer to a logistic (with the dramatic changes concentrated in the intermediate ages, and earliest stages changing in a less steepy way).

  3. I'm sure you're correct with this: we're probably just not getting the specimens to plot that shallower slope of early allometry.

    1. Unfortunately, Currie (2003) does not include the measurement data, because with those it would be quite easy to check and eventually re-infer the juvenile proportion changing the type of curve. I'd be curious to see what results setting a logistic equation.

  4. I'm familiar with Russell's hatchling because the diagram was displayed in the old fossil hall of my local museum: the Canadian Museum of Nature, where Russell used to hang out. It was never obvious to me that it was hypothetical, but I was also a child at the time.

    Currie's hatchling looks like Sharovipteryx.

  5. Thank you for the post! Highly informative and fun to read too.
    I wonder if the information from Funston et. al. paper was on any use to the Prehistoric Planet T. rex juvies? Or was their appearance finalized before the paper came out?

  6. If you had anything to do with the dancing Carnotaurus you deserve a medal! My son and I have really enjoyed the Prehistoric World series.