Modern birds’ upper beaks are made up mostly from skull bones called the premaxilla, but the snouts of their earlier non-avian dinosaur ancestors were instead formed by large maxilla bones.

And Falcatakely forsterae here had a very unusual combination of these features.

Living in Madagascar during the Late Cretaceous, about 70-66 million years ago, it was around 40cm long (1’4″) and was part of a diverse lineage of Mesozoic birds known as enantiornitheans. These birds had claws on their wings and usually had toothy snouts instead of beaks, and many species also had ribbon-like display feathers on their tails instead of lift-generating fans.

Falcatakely had a long tall snout very similar in shape to a modern toucan, unlike any other known Mesozoic bird, with the surface texture of the bones indicating it was also covered by a keratinous beak. But despite this very “modern” face shape the bone arrangement was still much more similar to other enantiornitheans – there was a huge toothless maxilla making up the majority of the beak, with a small tooth-bearing premaxilla at the tip.

This suggests that there was more than one potential way for early birds to evolve modern-style beaks, and there may have been much more diversity in these animals’ facial structures than previously thought.

Retro vs Modern #21: Deinonychus antirrhopus

Deinonychus antirrhopus was one of the most significant dinosaur discoveries of the 20th century, kicking off the Dinosaur Renaissance and the recognition of the evolutionary link between maniraptoran theropods and modern birds.


The first remains of this species were discovered in North America in the 1930s, but at the time the fossils weren’t officially described or named. It wasn’t until the 1960s that more specimens were found in Montana, representing at least three preserved individuals, and paleontologist John Ostrom recognized that there was something very special about this dinosaur.

In contrast to the prevailing view at the time that theropods were all upright tail-dragging “sluggish lizards” this was clearly a highly specialized and active predator, with a huge sickle-shaped claw on each foot and a long stiff tail for balance – inspiring its scientific name’s meaning of “counterbalancing terrible claw”.

And while the very first reconstruction of Deinonychus might seem retro to modern eyes, at the time it was revolutionary and it went on to become an iconic representation of the species for the next couple of decades. Drawn by Robert Bakker, who was Ostrom’s student at the time, it depicted a lizard-like creature with its body held in a horizontal pose and its tail held out straight behind it. Its head was portrayed as more domed than we now know Deinonychus’ skull to have been, and its neck was up in an alert posture while the animal ran at full sprint, with its sickle-claws held up away from the ground to keep them sharp.

A few years later further discoveries showed a highly bird-like pelvis and hands very similar to those of Archaeopteryx, triggering the Dinosaur Renaissance reinterpretation of dinosaurs as active warm-blooded animals, and the revival of the 19th century idea that they were the ancestors of birds.


As the “birds are dinosaurs” idea began to gain acceptance with increasing amounts of anatomical evidence, some paleontologists in the 1980s began to also suspect that highly bird-like dromaeosaurids like Deinonychus might have also been feathered. Some reconstructions during this time showed this to varying degrees, particularly those drawn by Bakker and by Gregory Paul – but it didn’t really catch on more widely at first, for one very big reason:

Jurassic Park happened.

Dromaeosaurs hadn’t been well-known dinosaurs to the general public before that point, but the 1993 JP “raptors” were an instant hit in pop culture. Physically based much more on Deinonychus than on Velociraptor, and exemplifying the renaissance view of dinosaurs in major media for the first time, the movie’s fully scaled and oversized version of these animals dominated popular depictions for years afterwards. Even the most rigorous and anatomically accurate artwork showcasing their bird-like features still usually kept them completely naked to retain that familiar reptilian appearance.

Most 1990s attempts at any feathering tended towards being as sparse as possible – often along with the shrinkwrapping typical for the era – at best being decent for the time but what we’d now deride as “half-assed”, and at worst being “a few token quills on the back of the head”.

Deinonychus fossils found in association with Tenontosaurus were also interpreted as being evidence of cooperative pack hunting behavior during this time, and it became a common paleoart meme to depict the large herbivore being constantly swarmed by ravenous raptors.


The mid-1990s discovery of fully-feathered dinosaurs like Sinosauropteryx in China, followed a few years later by raptors with wing-feathered arms like Sinornithosaurus, gradually began to put the fluff back onto dinosaurs like Deinonychus.

(…At least in reasonably scientific paleoart. The much much stronger and ongoing resistance from popular culture is far too big a subject to get into here. But maybe, just maybe, we’re finally hitting a turning point there?)

Early attempts at properly feathering dinosaurs were a bit awkward, usually looking rather like a bunch of scruffy greasy hair glued onto a scaly raptor, a dinosaur wearing fuzzy pajamas, or like the old “bird-lizard” depictions of Archaeopteryx. Even into the early 2010s some paleoart memes were still common in depictions of dromaeosaurs, but increasingly better understanding of their anatomy and plumage arrangements over the last decade or so has brought us to a much more birdlike interpretation of these animals – with paleoartists like Emily Willoughby being especially influential in popularizing the modern view of dinosaurs like Deinonychus.

We now know Deinonychus lived during the Early Cretaceous, about 115-108 million years ago, in what is now the Mountain West and South Central United States. Up to around 3.4m long (11′), it stood about 1m tall (3’3″), similar in size to a large dog.

It had blade-like teeth in its jaws, and forward-facing eyes with stereoscopic vision. Its three-fingered arms would have been covered by wing-like feathers, and its tail probably had feathers all the way along its length and was stiffened but not totally inflexible.

It may have used the sickle-claws on its feet to pin down struggling prey, eating it alive while flapping its wings and waving its tail for balance. And while often depicted as an extremely fast-runner, its leg proportions and foot anatomy suggest it was actually built more for walking and had an especially strong grip strength in its feet, trading speed for power and probably being more of an ambush predator – often being compared to a “giant ground-hawk”.

Pack hunting has been called into question recently, too, arguing that the Tenontosaurus sites may actually represent crocodile-like or Komodo dragon-like behavior with mobs of scavenging individuals congregating at a carcass. But other evidence from trackways and Utahraptor does offer potential support for pack behavior in raptors, so it’s still open to interpretation.

Retro vs Modern #14: Therizinosaurus cheloniformis

Therizinosaurs were some of of the most unique theropod dinosaurs. It’s only in the last few decades that we’ve started to understand much about them, and they’re still somewhat enigmatic even today.


The first known therizinosaur fossil discovery was Therizinosaurus cheloniformis, found in Southern Mongolia during the late 1940s and described and named in the mid-1950s based on a few fragments that included some unusually large and elongated claws.

These remains were interpreted as belonging to a giant turtle-like reptile that used its scythe-like claws to harvest aquatic plants – inspiring both parts of its scientific name, with Therizinosaurus meaning “reaper lizard” and cheloniformis meaning “turtle-shaped”.


The turtle interpretation began to be questioned during the 1970s, and the discovery of some slightly better (but still fragmentary) specimens reclassified Therizinosaurus as an unusual theropod dinosaur.

In the late 1970s and early 1980s fossils of another group of dinosaurs known as “segnosaurs” were also starting to be discovered, with a confusing mixture of anatomical features that seemed to link them to multiple different dinosaur lineages. As a result opinions about their evolutionary relationships varied during the 1980s, sometimes considering them to be theropods, sometimes late-surviving “relic” prosauropods, and sometimes a whole new major lineage of rare and weird saurischians.

Similarities between segnosaurs and the known material of Therizinosaurus were soon noted, and the discovery of the fairly complete Alxasaurus in the early 1990s confirmed that they were all part of the same group of bizarre herbivorous theropods – and the name “segnosaurs” was dropped in favor of “therizinosaurs”, since older names usually get priority in taxonomy.

Reconstructions of Therizinosaurus during this time tended to look rather weird and awkward. As with the majority of dinosaurs during this period it was depicted as entirely scaly and reptilian, and was often shown with a stiff hunched downcurving neck and an oddly tiny-looking tail.

Further discoveries during the 1990s finally began to clarify therizinosaurs’ evolutionary affinities, eventually placing them as an early branch of bird-like maniraptoran theropods, closely related to both oviraptorosaurs and the alvarezsaurs – and in 1999 the discovery of the small early therizinosaur Beipiaosaurus helped to confirm this relationship, revealing impressions of an extensive coat of filamentous feathers and longer stiffer quill-like structures.


Over the next couple of decades more and more therizinosaur fossils were found in both Asia and North America, and details about these dinosaurs’ appearance, biology, and ecology gradually became better understood. While there’s still a lot we don’t know about them, we do now at least have some fairly complete examples like Nothronychus, fossilized therizinosaur footprints, more feathers, an idea of Beipiaosaurus’ coloration (it was brown!), and even eggs and potential colonial nesting sites.

And while Therizinosaurus itself is still only represented by fragmentary and incomplete material, we now have a much better idea of what it was probably like. It lived in what is now the Gobi Desert during the Late Cretaceous, about 70 million years ago, and would have been both the largest known therizinosaur and the largest known maniraptoran dinosaur, estimated to have been as much as 10m long (33′) based on the proportions of its relatives.

It would have had a tiny head with a toothless beak at the front of its jaws, a long neck, and a wide bulky “pot-bellied” body housing its huge plant-fermenting gut. With its especially large body size it probably wasn’t as extensively feathered as its smaller relatives, but it may have still been sparsely fuzzy across parts of its body.

Unlike most other theropods it walked on all four toes of its feet, with the dewclaw enlarged into an extra weight-bearing digit. Recent analysis of footprints has also suggested that the larger therizinosaurs like Therizinosaurus may actually have been plantigrade, walking with their feet completely flat on the ground. This might turn out to just be an artifact of how the tracks were preserved, but therizinosaurs are certainly already weird enough that it could be a plausible interpretation.

But Therizinosaurus’ most distinctive feature was its hands, with extremely long narrow claws each at least 50cm long (1’8″). Unlike the strongly curved claws seen in other therizinosaurs, these ones were fairly straight for most of their length, only curving more sharply towards their tips.

While in the past these claws have been proposed as being weapons or digging adaptations, they were actually relatively delicate and were probably mainly used for pulling clumps of vegetation closer in a convergently similar manner to the later mammalian chalicotheres and ground sloths.

Therizinosaurus would have been a heavy slow-moving animal, and probably spent a lot of time sitting on its haunches supported by its especially robust hip bones while it browsed on large amounts of vegetation. It likely relied on its pure bulk and intimidation to deter potential predators, possibly even making aggressive displays with its claws when threatened – essentially it may have been a giant goose-sloth.

Retro vs Modern #04: Archaeopteryx lithographica

Archaeopteryx lithographica was first discovered in the 1860s, still in the early days of our understanding of dinosaurs, and was a timely example of the sort of transitional form first proposed by Charles Darwin only a couple of years earlier. For over a century it was a famous icon of evolution, and has been part of a lot of weird drama over the years – it’s been central to arguments about bird origins, was accused of being a fake, and one specimen even vanished under mysterious circumstances.


At the time of its discovery Archaeopteryx was actually fairly quickly accepted as demonstrating an evolutionary link between dinosaurs and birds… but sadly this view wasn’t to last.

In the early 20th century opinion shifted towards birds not being dinosaurs but instead descended from “thecodont” reptiles (what we’d now call early archosaurs and pseudosuchians). And so for a long time Archaeopteryx ended up being depicted as simply the “first bird”, a half-reptile half-avian curiosity.

Reconstructions of it from this time period varied from very good to kind of awkward depending on how much the artist was trying to emphasize its reptilian ancestry, commonly featuring wonky-fingered wings and a scaly lizard-like face. It was also frequently depicted with bright gaudy parrot-like coloration, with a specific yellow-and-blue color scheme becoming a “paleoart meme” so prolific that it would eventually inspire the design of a Pokémon.


After decades of stagnation the dinosaur-bird link was resurrected in the early 1970s, with the discovery of the bird-like Deinonychus kicking off the Dinosaur Renaissance. Along with the explosion of spectacularly feathered dinosaur fossils from China in the mid-1990s, Archaeopteryx finally began to be properly presented as a feathered dinosaur again.

Continued study of the known Archaeopteryx specimens in the last couple of decades has vastly improved our knowledge of what this animal would have looked like, revealing previously unknown features like the exact plumage arrangement on its wings and legs, and even potentially some details about its coloration.

Living in southern Germany during the Late Jurassic, about 150-148 million years ago, Archaeopteryx inhabited what was then an island archipelago in a shallow tropical sea. It grew to around 50cm long (~1’8″) and was almost entirely covered with pennaceous feathers, externally probably just looking like a long-tailed bird.

It had broad wings, with asymmetrical flight feathers similar to those of modern birds but with more extensive coverts, some of which were probably a matte black color. Its legs also sported long “feather trousers” and a “raptor“-like hyperextensible second toe, and there was a slight forked shape to the tip of its tail.

Arguments have gone back and forth about how well it was actually able to fly, with current thinking being that it made short bursts of active flapping flight a little like a modern pheasant – but since its shoulder joints were less mobile than those of modern birds it must have used a different sort of flight stroke to generate lift.

It’s no longer always considered to have been the “first bird”, or even to have been the direct ancestor of any modern birds. Instead it represents an offshoot lineage of early birds (or very-bird-like dinosaurs) that was just one part of a still-expanding flock of feathery fossil discoveries.


66 million years ago, the end-Cretaceous mass extinction wiped out all dinosaurs except for the avian bird lineage.

…Or did it?

But I’m not talking about the dubious claims of non-avian dinosaur fossils found in places they shouldn’t be. This is about something else entirely: an unassuming little bird known as Qinornis paleocenica.

Living in Northwest China during the mid-Paleocene, about 61 million years ago, Qinornis was roughly pigeon-sized at around 30cm long (12″). It’s known only from a few bones from its legs and feet, but those bones are unusual enough to hint that it might have been something very special.

Uniquely for a Cenozoic bird, some of its foot bones weren’t fully fused together. This sort of incomplete fusion is seen in both juvenile modern birds and in adults of non-avian ornithurine birds from the Cretaceous – and the Qinornis specimen seems to have come from an adult animal.

If it was fully grown with unfused feet, then that would suggest it was actually part of a “relic” lineage living 5 million years after the mass extinction, surviving for quite some time longer than previously thought.

The last known non-avian dinosaur.


About 72 million years ago, in the Late Cretaceous of what is now Mongolia, a dead dromaeosaurid dinosaur lost its head.

30 years ago, in 1991, its headless fossilized remains were discovered during a joint Mongolian Academy of Sciences / American Museum of Natural History expedition in the Gobi Desert.

For a long time the specimen was known only by the nickname of “Ichabodcraniosaurus”, in reference to a character haunted by a headless ghost in the story The Legend of Sleepy Hollow – but now it’s finally been given a full scientific description and a proper name.

Say hello to the first new non-avian dinosaur of 2021, Shri devi!

Named after a buddhist deity, this little dinosaur was around 2m long (6’6″), roughly the size of a modern peacock or wild turkey. It was a very close relative of Velociraptor, but lived in a slightly different part of the ancient Gobi than its famous cousin, giving us a glimpse of how dromaeosaurid species varied across that region.

A map of the Gobi region of Mongolia and China, showing locations where various dromaeosaurid dinosaur fossils have been found.
[ From fig 28 in Turner, A. H. et al (2021). A New Dromaeosaurid from the Late Cretaceous Khulsan Locality of Mongolia. American Museum Novitates. https://doi.org/10.1206/3965.1 ]


Named after the mythological bird-like Anzû – and also nicknamed “the chicken from hell” – Anzu wyliei was one of the larger known oviraptorosaurs, measuring about 3m long (9’10”).

Its fossils are some of the most complete for a North American member of this dinosaur group, with four different specimens representing about 80% of the whole skeleton.

Living right at the end of the Cretaceous, about 66 million years ago in North Dakota and South Dakota, USA, Anzu inhabited the ancient floodplains of Hell Creek and appears to have been a fairly fast-moving omnivorous generalist. It had a large crest on its head made of rather fragile thin-walled bone, which may have been used for display or sound amplification similar to the casque of modern cassowaries.

Some of the fossil specimens also show evidence of healed injuries, including a broken rib and an arthritic toe.

Glowing Dinosaurs

Many modern birds are capable of seeing into the ultraviolet regions of the electromagnetic spectrum, and some of their non-avian dinosaur ancestors might have had the same sort of vision. And much like their living relatives, that means various parts of their bodies and plumage may also have been UV-reflective and UV-fluorescent.

So here’s a Velociraptor with some speculative UV coloration – although this is just what it would look like to human eyes under a blacklight. What it would actually look like to a creature that can see extra colors is impossible to depict on a screen designed for trichromatic vision!

Island Weirdness #08 – Balaur bondoc

When Balaur was described in 2010 it was initially thought to be a dromaesaurid closely related to Asian forms like Velociraptor. With its particularly stocky legs built for strength rather than speed, two-fingered hands, and two large sickle claws on each foot, it was interpreted as a weird highly specialized predator terrorizing the other Hațeg Island species at the end of the Cretaceous. Although only 1.8m long (5’10”), it was hypothesized to have taken down prey much larger than itself with powerful slashing kicks.

But later analyses cast doubt on this interpretation.

A lot of the anatomical features of Balaur’s skeleton were odd for a dromaeosaurid, but matched those of avialans – a group of close evolutionary “cousins” to the dromaeosaurids, containing Archaeopteryx and the common ancestors of all modern birds. And, by 2015, multiple studies had confirmed Balaur wasn’t really a “raptor” but instead a little further along on the bird lineage.

So now our picture of this dinosaur is very different: a chunky-bodied island bird, grown large and secondarily flightless sort of like a Cretaceous equivalent to the dodo. Its double sickle claws were probably adaptations for climbing and perching in trees, and based on similar avialans it was likely a herbivore rather than a hypercarnivore.


Erlikosaurus andrewsi, a therizinosaur from the Late Cretaceous of Mongolia (~90 mya).

Named after Erlik, the Turko-Mongolian god of death, it’s only known from partial remains – but it was the first therizinosaur ever found with a preserved skull, helping to fill in some of our knowledge of these oddball dinosaurs’ anatomy.

It was closely related to Therizinosaurus, but was only about half the size, estimated to have measured around 4-5m long (13′-16’4″). It would have had a toothless beak at the front of its jaws, an adaption for a herbivorous diet, along with long claws on its hands and a coat of fluffy down-like feathers. I’ve also given it some longer quill-like feathers here, similar to those known in Beipiaosaurus.