Wednesday 26 August 2020

Unorthodox Ideas about Bird Origins

Fig. 1: Two versions of “Proavis”, a hypothetical proto-bird even more primitive than Archaeopteryx. The one on the left is Gerhard Heilmann’s version from the 1920s, a thecodont that uses scales to glide. The one on the right is Greg Paul’s from 2002, an arboreal dinosaur similar to a tree kangaroo that uses airfoils derived from insulation to soften its fall. Guess which vision won out.

Birds are dinosaurs. This is consensus now and has been at the very least since the 90s (which are 20-30 years away now). While this is a fact that to our modern minds seems obvious or even trivial, the road that led to this consensus was long, convoluted and lined with some quite bizarre side-tracks. The purpose of this post, similar to the one about pterosaurs, is to highlight some of these unorthodox ideas that have now fallen into obscurity (though note that not all will be ridiculed here, as some certainly have some merit). Before that however a little rundown on the history of research on bird origins: The idea that birds might descend from dinosaurs had already gained traction in the 1870s, as Thomas Henry Huxley and the prominent paleontologists Edward Drinker Cope and Othniel Charles Marsh endorsed this idea based on their research on dinosaurs and fossil birds. It became a debated issue with no clear consensus reached. A great shift came then in 1926 when Danish amateur-ornithologist Gerhard Heilmann released his The Origin of Birds. In it he is only one sentence away from reaching the conclusion that birds truly descend from coelurosaurian theropod dinosaurs, but then takes a turn in the last second. The reason for this is that in the dinosaurs known to him at the time the clavicle, a bone of the chest, was either underdeveloped or absent, while this bone was strongly developed into a furcula in birds. As Heilmann believed in Dollo’s Law (the idea that once a trait was lost it could not re-evolve) he therefore concluded that birds could not have descended from dinosaurs and must instead have been derived from earlier “thecodont” reptiles similar to Euparkeria, which still had a developed clavicle. Any similarity between dinosaurs and birds was therefore due to convergence, though he still thought they must share a close common ancestor. The problem with Heilmann’s thesis was that Oviraptor, a dinosaur with clearly developed clavicles, was already known at the time, but was unfortunately not studied well. Nonetheless, thanks to his writing and artwork, Heilmann’s book was seen as the last word on bird origins, became wide consensus and unintentionally shut down any further research on the topic (Paul 2002). The discovery of clavicles in Segisaurus in 1936 was blatantly ignored (Paul 2002), as were Percy Lowe’s (1935) and Nils Holmgren’s (1955) re-suggestions of a dinosaurian bird-origin. This only changed in the 1970s when John Ostrom compared the anatomy of Archaeopteryx with that of Deinonychus and many new fossil discoveries were made, which allowed paleontologists to robustly revive Huxley’s ideas of bird origins. An intense debate broke out in the 80s and early 90s between dinosaur-experts and paleo-ornithologists who still clung to the old Heilmann-idea. Cladistic methods however always favoured the dinosaur-hypothesis and soon enough fossils of clearly feathered dinosaurs were also discovered, basically annihilating any opposition. Today the thecodont-hypothesis and its variations are seen as untenable (in part also because thecodonts are not a thing anymore) and it is generally agreed that birds are maniraptoran theropods, with their closest relatives being deinonychosaurs such as Velociraptor and Stenonychosaurus.

Birds descend from fish

Fig. 2: If you ever wondered what a mix between fish and poultry would taste like, I recommend trying alligator meat. I always eat some when I visit Florida.

Of course, if you want to be a Smarty McSmartpants, you could say that this statement is true (at least depending on your definition of fish) in a broad sense. But what the title refers to here is the hypothesis proposed by Chinese professor Tao Hai that birds directly descend from flying fish. Yes, you heard that right. “What Victorian-era silliness is this?” you might think to yourself, until I tell you that this suggestion was made in the far and distant year of 1993. He came to this conclusion after examining fossils from the Xinjiang region which he called “proto-birds with fin-like wings”. According to Wolfgang Hassenpflug and Gerald Kopp the fossils do indeed have some sort of resemblance to decapitated bird-carcasses, but are in actuality normal fish which have become bloated and disarticulated due to decomposition. There is a possibility that they might be early flying fish, but they definitely are not bird ancestors. Apparently Hai even made reconstructions of his bird-fish, but I have not been able to find them. How exactly a teleost fish would be able to convergently evolve a tetrapod body-plan and an avian lung-system was never explained by him and frankly the idea seems so ridiculous that it would only be overtaken by me proposing that birds descend from flying insects.

Edit from December 2021: Thanks to my friend Bob Guan I have actually been able to find Tai’s reconstruction of his hypothetical creature, as well as the fossil he based it on. Bob’s Chinese, so he was able to ask other Chinese academics on social media if they knew about this paper, which has produced these results:

As Bob's research turned out, the fossil in question was not even that of a genuine flying fish (which are teleosts of the Beloniformes order) but of a paleonisciform, which are a now extinct order of early ray-finned fish. Seeing the fossil does make it somewhat more understandable how Tai saw bird-like affinities, as, especially at this low resolution, the splayed out pectoral fins of the decapitated fish do superficially resemble preserved bird-wings, like those of the famous Berlin-specimen of Archaeopteryx. Though you can probably achieve a similar effect on the fins of many actinopterygian fossils. I still would not recommend showing this to David Peters, as misinterpreting photos of fossils is pretty much his whole methodology, so he might get a wrong idea again. A whole other matter is Tai's reconstruction, which is downright hilarious as it looks like he filled out the missing gaps of the fish fossil by attaching parts of a rubber chicken from the toy store. With a little bit more painterly quality it could very well pass as a chimaeric creature from ancient mythology or folklore, a sort of aquatic relative of the cockatrice. On the maps of ancient cartographers there indeed could be found a creature called sea-chicken, though this was born out of the idea that every animal on land must have a counterpart living in the ocean (which is how we ended up with animal names such as sea-elephant, sea-cow, sea-lion and so on).

Birds are the closest relatives of mammals

Fig. 3: The stem-haemotherm, a hypothetical ancestor to a clade that consists of only mammals and birds.

In the nineteenth century Richard Owen grouped birds and mammals together in the group Haemothermia, suggesting that they both descend from a close common ancestor, as they were the only warm-blooded vertebrates around. The two groups share other superficial traits, such as insulation and four-chambered hearts, but these are all clearly not inherited from a common ancestor but examples of convergent evolution. Despite this some researchers have still tried to maintain the Haemothermia/Haematothermia hypothesis. The first to resurrect it was embryologist Søren Løvtrup (a critic of Darwinism who instead believes in occasional saltation-events) in 1977, followed by Brian Gardiner in the 80s. Both however largely based their conclusions on data obtained in 1866 and 1693 (!) and their cladistic models only worked by excluding vast amounts of unignorable fossil and anatomical data. Notorious David Peters’ recent suggestion that mammals are the sister-group to archosaurs also has clear shades of this. Despite of how untenable the haemotherm-hypothesis is, Phillippe Janvier produced a reconstruction in 1983 of what the stem-haemotherm, the common-ancestor of birds and mammals, should look like, which you can see here. Needless to say, such a bird-mammal has yet to show up in the fossil record and it never will. It is however nice fuel for speculative evolution.

Birds are crocodilian-descendants

Fig. 4: Effigia okeefeae as illustrated by Raul Martin. These shuvosaurids show remarkable convergence with theropod dinosaurs, so much that Shuvosaurus was once mistaken for an ornithomimid. In reality these animals were more closely related to crocodiles.

From the frankly ridiculous suggestions we now come to ones that at least have (or had) some merit. Birds, being archosaurs, are today the closest living relatives of crocodilians and share many anatomical characteristics with them. These, we know today, were given down by a distant common ancestor. For a brief time in the 1970s however it was suggested that birds could have directly descended from proto-crocodiles. This idea came from Alick Walker after his description of Sphenosuchus in 1972. Sphenosuchus and its relatives, the Triassic sphenosuchians, were a group of crocodylomorphs which are now thought to have been the immediate ancestors of the crocodyliforms, including our modern crocs. Despite this their anatomy was highly gracile, they had upright limbs, could probably walk bipedally at least on occasion and likely had a high metabolism. They in fact resembled theropod dinosaurs so much that some sphenosuchians like Hallopus and Saltoposuchus were once mistaken for early dinosaurs. Based on this and the assumption that sphenosuchians were more capable climbers than dinosaurs, Walker picked them as the bird-ancestors. At this point we should probably mention an important controversy in the debates around bird origin: Did birds evolve flight by taking off from the ground or by gliding from trees down? To some this had important implications around the question if birds descended from dinosaurs or from other archosaurs, as it was thought by some that dinosaurs were incapable climbers, while “thecodonts” were seen as at least partially arboreal. Some researchers thinking that bird-flight could have only evolved in an arboreal setting therefore automatically excluded dinosaurs as candidates, despite any anatomical similarities. This mutual exclusion became nil however after paleontologists like Gregory S. Paul and Sankar Chatterjee convincingly demonstrated that theropods like Sinornithosaurus or Ornitholestes would have been perfectly capable of climbing trees in the same way as cats or tree kangaroos do today. In fact, they would have been better at doing so than sphenosuchians, which actually lacked many of the arboreal qualities of earlier archosaurs such as Euparkeria. Some of the other characteristics Walker chose to argue for a sphenosuchian-bird link also turned out to not be convincing. While crocodilians and modern birds do share some interesting similarities in their skull-bones, like a double-headed, pneumatic quadrate, these are not present in early birds such as Archaeopteryx and therefore represent a case of convergent evolution (Paul 2002). Even more important is that crocodilians and birds share a completely different ankle-structure, so much so that this is the defining characteristic which separates croc-line from bird-line archosaurs. Walker himself has recanted his hypothesis in 1985.

Birds descend from pterosaurs

Fig. 5: Archaeopteryx and Rhamphorhynchus certainly would not enjoy a family reunion.

Birds and pterosaurs share many characteristics, such as the avian-style ankle, hollow bones, filamentous body-coverings, beaks and of course the capability of powered flight. That they share a direct ancestor-descendant-relationship is therefore an intuitive but wrong assumption made by pop-culture, laypeople and children (though I have witnessed moments in museums where even six-year-olds were able to correct their parents about this, which I can only hope to achieve one day when I have kids). In the pre-Heilmann era a close relationship between pterosaurs and birds (though not necessarily a direct ancestry) was however a very popular, legit hypothesis and was seen as the most viable alternative to Huxley’s dinosaur-bird link. It was championed by figures such as St. Georges Mivart and Harry Govier Seeley, the latter thought they were each other’s closest relatives with a common ancestor living somewhen in the late Paleozoic. An earlier even more unusual suggestion was made by Samuel Thomas von Soemerring, one of the original workers on Pterodactylus. In 1810 he proposed that pterosaurs represent an intermediate stage between birds and bats (though not necessarily in an evolutionary sense as Darwin had not published his theories yet). Any such notions were already carefully criticised by Huxley in the 1870s and thankfully whisked away by Heilmann in his 1926 work. It has since become obvious that most similarities shared by the two groups convergently evolved due to the adaptation to powered flight. In 1983 Larry Martin nonetheless suggested that Scleromochlus, an archosaur thought to be closely related to pterosaurs (at least until recently), was also an especially close relative of birds, but this has largely been ignored or dismissed for good reason. How many of the similarities between birds and pterosaurs are truly convergent is still an interesting question. Recent work done on pterosaur pycnofibers heavily suggests that these structures are homologous to the proto-feathers of dinosaurs and might be called such. The last common ancestor between dinosaurs and pterosaurs may therefore already have been quite a birdy animal.

Birds descend from ornithischian dinosaurs

Fig. 6: The fuzzy ornithischian dinosaur Kulindadromeus, drawn by Tom Parker.

The, albeit superficial, similarities between birds and some ornithischian dinosaurs, especially the aptly named ornithopods, is fascinating and something I feel is underappreciated. Both have a retroverted pubis, three-toed feet at the end of long bipedal legs, long necks and beaks. It is not hard to imagine that many smaller ornithopods had lifestyles and behaviours similar to those of modern ground-birds and landfowls. Until the 70s it was even thought that some ornithopods were arboreal and had a retroverted first toe like modern perching birds. The heterodontosaur Manidens may have been a genuine tree-dweller. That birds descend from the bird-hipped dinosaurs is therefore again an assumption that could be intuitively made, but will confuse a lot of laypeople, as birds counterintuitively descend from Saurischia, the lizard-hipped dinosaurs. The similarities to ornithischians are largely due to convergence or common ancestry. Nonetheless there have been serious suggestions, made by some quite prominent figures, that the former is the case. The first was made by Huxley after comparing Archaeopteryx with the ornithopod Hypsilophodon, mainly based on the shape of their pubis, though he saw a relationship to theropods such as Compsognathus as equally likely. This was further elaborated on by Othniel Charles Marsh in his 1896 monograph The Dinosaurs of North America where he writes about ornithopods:

The various dinosaurs thus briefly referred to under their respective genera have many other points of interest that can not be here discussed, but their resemblance to birds is worthy of some notice. This is apparent in all of them, but in the diminutive forms the similarity becomes most striking. In all the latter the tibia is longer than the femur, a strong avian character, and one seen in dinosaurs only in the small bird-like forms. In Nanosaurus nearly all, if not all, the bones preserved might have pertained to a bird, and the teeth are no evidence against this idea [Marsh had previously discovered the toothed birds Hesperornis and Ichthyornis]. In the absence of feathers an anatomist could hardly state positively whether this was a bird-like reptile or a reptilian bird.” (Marsh 1896 p. 201)

Especially the last sentence is interesting, as it might raise the question if Nanosaurus may have been feathered in life. While I am not sure if that is what Marsh wanted to insinuate, if he did, he would have predicted fossil discoveries made over 100 years in the future. Since 2010 ornithischian dinosaurs have been discovered with structures very similar to and likely homologous with feathers, most important among these being Kulindadromeus, a basal neornithischian just like Nanosaurus. The bird-ornithischian link was last revived in 1970 by Peter Galton. He however had to admit that ornithischians were far too derived for them to be the direct ancestors and instead just proposed a close common ancestor. He later recanted his idea. Ironically, this may not have been too far off the mark, given the recent Ornithoscelida hypothesis.

Birds originated in the Triassic

Fig. 7: Pretty much everything in this picture is wonky as hell. The anatomy of the Postosuchus is completely off (though at least they got the bipedalism right), Technosaurus is considered a dubious genus now, Shuvosaurus was a croc-line archosaur and Protoavis was not a real taxon, let alone a Triassic bird. The limestone also looks suspiciously like a brick wall.

When birds first started appearing was a controversial question for quite some time. In the early days of paleontology it was first generally thought that they only appeared in the Cenozoic, but then the discovery of Archaeopteryx extended the timeline all the way into the Late Jurassic. But were there also birds long before Archaeopteryx? The first such claim came from Edward Hitchcock in 1848, 12 years before the Urvogel was even discovered. He identified giant three-toed footprints of the Triassic New Red Sandstone as pertaining to gigantic moa-like waterfowl, but these later turned out to have been produced by basal sauropodomorphs. Some later paleontologists, who wanted to argue against a dinosaurian origin of birds, also used Triassic footprints as evidence for a pre-dinosaurian origin, but many of these footprints were as large as a human’s, if not larger, so they likely made the same mistake as Hitchcock. In the 70s Paul Ellenberger claimed that birds descend from the Triassic reptile Cosesaurus, but none of the subsequent studies were able to confirm his observations of bird-like features in this animal, including feathers, and it is now regarded as a protorosaur related to Tanystropheus. The most recent suggestion of Triassic birds came from Sankar Chatterjee, who in 1991 discovered some interesting bones in the Texan Dockum Beds (I have written more about this here). He called it Protoavis texensis and reconstructed it as a Triassic proto-bird even more derived than Archaeopteryx. Notably though, Chatterjee still argued in favour of birds descending from maniraptoran theropod dinosaurs, but in order to do so he had to move the origin of that clade all the way back into the Triassic. In the process he accidentally classified Shuvosaurus as a Triassic ornithomimosaur. It later turned out to have actually been a pseudosuchian crocodile-relative. After further examination by other workers, Chatterjee’s Protoavis turned out to not even be a genuine taxon and many of the bird-like characteristics he ascribed to it could not be found. Instead it was a fossil chimaera of one or more early theropods and a drepanosaur which died at the same spot likely due to a monsoonal flash flood. Modern consensus still is that the bird-lineage originated somewhen in the mid-to-late Jurassic, with crown-group birds only appearing in the Late Cretaceous.

Only some birds descend from dinosaurs

Fig. 8: Struthiomimus, illustrated by Gerhard Heilmann with some very ostrich-like behaviour. Take special note of the one scratching itself in the background.

At times the idea that all birds descend from the same ancestor has also been called into question. In the late nineteenth century Carl Vogt, St. George Mivart and Robert Wiedersheim for example believed that Archaeopteryx and modern flying birds descended from or were closely related to pterosaurs, while flightless ratites, such as ostriches, directly descended from dinosaurs (Paul 2002). A similar belief was also held by Percy Lowe and Baron Franz Nopsca, who believed ostriches and rheas directly descended from the ornithomimd theropods, such as Struthiomimus, Ornithomimus or Gallimimus, with Hesperornis serving as an intermediate stage. In this view, ratites did not secondarily lose the ability to fly, but never evolved flight in the first place and supposedly evolved their wings for different purposes. Needless to say, this view never caught on, but when I was a kid I remember having a German dinosaur book, Das Grosse Buch der Saurier, that actually still presented this idea in a serious manner. The book was originally written in 1989 by a certain Peter Klepsch, of whom I could not find more information, and was beautifully illustrated by Peter Thiemeyer, an accomplished fantasy-artist. The version I own was a reprint from 2000. I scanned and translated the anachronistic section below. What may have influenced this “anachronism” is that bird-polyphyly was briefly considered again by Rinchen Barsbold in his 1983 descriptions of Mongolian theropods, in which he observed that multiple theropod lineages went, in parallel, through a process of “ornithization”, making it possible that they could have given rise to birds more than once. However, Barsbold’s studies were quite obscure at the time, most being published in Russian.

Fig. 9: “For a long time animals with scales were regarded as reptiles and those with feathers as birds. But since in the meantime dinosaurs with feathers have been discovered, some scientists order animals into the reptiles or birds according to their ability to fly. Emus, cassowaries, kiwis, rheas and ostriches never flied though and do not descend from flying animals. They are according to the latest understanding either the closest relatives of the dinosaurs or still belong to them today.

Birds descend from theropods other than maniraptorans

Fig. 10: Climbing, feathered Ornitholestes, illustrated by Greg Paul in 1988. The idea that Ornitholestes had a nasal crest is discarded today and was based on a crushed skull.

Even among people who agreed that birds are a group of theropods it was not always universally agreed upon from which theropods they came from. Before it was realized that it was a chimaera, Greg Paul in 1988 briefly considered the possibility that Protoavis and all later birds descend from the herrerasaurs, but he saw this as unlikely. In 1985 Michael Raath also proposed that coelophysoids were especially close relatives of birds, but this was generally dismissed as these theropods had short arms and were still relatively archaic. Even after Thomas Huxley, Compsognathus still gained a lot of attention as a possible close relative of Archaeopteryx, mainly because they lived in the same place and had roughly similar skeletons (to the point where some Archaeopteryx preserved without feathers were originally identified as Compsognathus), but this dinosaur was too basal of a coelurosaur for that to be the case. A somewhat bizarre idea came from Andrzej Elzanowski and Peter Wellnhofer in 1992/3, who concluded that troodonts, spinosaurs and birds form a monophyletic clade. Birds and troodonts being close relatives is obvious, but the titanic Spinosaurus being an especially close relative of Archaeopteryx seems rather hard to believe. Though the way shoebill storks look at me at the zoo sometimes makes me doubt my assertions.

Some dinosaurs are actually secondarily flightless dino-birds

Fig. 11: Are dinosaurs like Tarbosaurus and Therizinosaurus actually secondarily flightless? It is more plausible than you think.

One of the most fascinating and plausible ideas discussed in this post comes from Gregory S. Paul. First in his 1988 book Predatory Dinosaurs of the World and later in more detail in his 2002 book Dinosaurs of the Air he proposed the idea that some of the dinosaurs we conventionally think of as non-avian are actually secondarily flightless and descend from flying dino-birds more derived than Archaeopteryx. Several lines of evidence make this idea attractive:
  • Archaeopteryx is, skeleton-wise, not really more derived than other maniraptoran dinosaurs, essentially being just a miniature version of Deinonychus. Some supposedly non-avian dinosaurs on the other hand even have skeletal features more similar to modern birds than Archaeopteryx.
  • Many of the most bird-like dinosaurs, including Deinonychus, start appearing in the Cretaceous, after Archaeopteryx.
  • The most basal dromaeosaurs and troodonts were forms such as Microraptor, which were small and probably capable of flying or gliding, and the lineages grew larger and less bird-like as time wore on. Even the oldest therizinosaurs and tyrannosaurs were small, bird-like animals.
  • These Cretaceous maniraptorans possess many flight-related features, such as wing-folding-mechanisms, asymmetrical feathers, large sterna, ossified sternal ribs, binocular vision and large braincases, that are hard to explain in a terrestrial context and differ from previous terrestrial theropods, but make more sense if these animals descended from an ancestor capable of flight.
  • Secondary flightlessness has evolved in birds on multiple occasions, even during the Mesozoic, examples being Gargantuavis and Hesperornis.
While Paul never declared any dinosaur-group with absolute certainty as neoflightless dino-birds, strong contenders were the dromaeosaurs, troodonts, oviraptorosaurs and alvarezsaurs, but possible candidates were also therizinosaurs (based on their reduced tail, posture and arm-folding-mechanism) and even tyrannosaurs (based on their reduced, ratite-like forelimbs and binocular vision). The problem with assessing these possibilities, Paul himself admitted, is that during the process of flight-loss traits are lost or reversed which would have given more clues about ancestor-descendant relationships. We know from modern ratites that once flight is lost, details of the pectoral girdle and limbs revert back to a dinosaurian state. An incomplete skeleton of an ostrich could under certain circumstances therefore be misclassified as a non-avian dinosaur. Something quite similar happened with Balaur bondoc, a flightless avialan from Hateg island that was at first identified as a velociraptorine. Even more uncertainties are introduced once we consider the possibilities that certain dinosaur groups may have evolved flight and neoflightlessness independently from each other. Most cladistic work does not strongly support Paul’s idea and recent finds have shown that troodonts and dromaeosaurs have existed in the Jurassic before Archaeopteryx, but unlike the other ideas discussed here, his hypothesis does have some significant following among respected paleontologists, such as Michael J. Benton. In his renowned textbook Vertebrate Paleontology Benton notably classifies oviraptorosaurs as flightless birds and there are many anatomical oddities among these dinosaurs, especially among forms like Caudipteryx, which are hard to explain without a secondary loss of flight. Even famous evolutionary biologist Stephen Jay Gould had some support for the idea. Depending on future fossil finds of small theropods from the Jurassic, it remains to be seen how well Paul’s hypothesis will do. Some analyses have recovered a slightly different scenario, which is that dromaeosaurs like Microraptor or Rahonavis evolved flight independently of birds. All three could nonetheless still have descended from a common gliding ancestor. One thing I have curiously never seen anyone, not even Paul, address is that the neoflightless hypothesis could explain one of the strangest anatomical oddities in therizinosaurs. Unlike any other non-avian theropod, these scythe-turkeys have a four-toed foot, with the first toe facing forward and touching the ground. I propose that, based on basal therizinosaurs like Beipiaosaurus, if these dinosaurs really do descend from flying or at the very least arboreal, tree-sloth like ancestors, they might have originally had a retroverted first toe like modern perching birds, which, as they grew in size, was re-adapted to add further support to the feet. The double sickle-claws of Balaur are proof that something like this can happen with secondarily flightless dino-birds. (Edit: I have been informed that the most basal therizinosaurs had ordinary feet, so this renders my pet theory into total bunk. Oh well, at least I tried).

Dinosaurs descend from birds

Fig. 12: Paul’s illustration of a feathered proto-dinosaur Lagosuchus bothering a proto-mammal.

Now what if we all got this backwards? Seemingly taking Greg Paul’s idea to an extreme, George Olshevsky argued in 1994 that all dinosaurs descend from Triassic small, feathered, bird-like animals capable of climbing or even flying that then became secondarily flightless. He essentially modified an idea had by Othenio Abel in 1911, who suggested that birds and dinosaurs shared a close, common arboreal ancestor (Paul 2002). In this view, even giants like Brachiosaurus and Stegosaurus would be flightless proto-birds. As possible ancestors of this weird bird-dino-clade he suggested Protoavis, which in turn would have evolved from bizarre reptiles such as Longisquama. Needless to say, this did not stand the test of time, as Protoavis was not a real taxon, Longisquama had nothing to do with bird evolution (as explained further below) and the earliest dinosaurs generally do not show signs of having been strongly arboreal or secondarily flightless. There is nonetheless still some merit to speculating that the ancestors of dinosaurs may have been partially arboreal and feathered. Their close relatives are the pterosaurs, which as mentioned earlier may have already possessed proto-feathers, and most would agree that these archosaurs evolved flight from arboreal ancestors. Even more intriguing are the discoveries of new dinosauromorphs such as Kongonaphon kely, which was described just this year. This lagerpetid stood just 10 centimeters tall and possessed an anatomy reminiscent of a jerboa. If it was also as metabolically active as one it surely would have needed some form of insulation. It heavily suggests that the ancestors of both pterosaurs and dinosaurs went through a size-squeeze at one point in their evolution, which led to the development of proto-feathers and other anatomical quirks in the two groups. Could the cause of this squeeze have been an adaptation to living in the trees? Early ornithodirans like Kongonaphon, Scleromochlus or Lagosuchus do not show any notable specializations for arboreality, though it is not inconceivable that they could have occasionally climbed trees in the manner of modern tree kangaroos. This would have helped them avoid competition with sphenosuchians and other croc-line archosaurs which were dominant on the ground throughout most of the Triassic.

Alan Feduccia in general

Fig. 13: Drepanosaurus, here illustrated by Mark Witton, is a fascinating animal with a bird-like head but an entirely un-bird-like body. It is about as well-suited for being a bird-ancestor as a tortoise. Those have beaks too, you know.

Those who have read the previous pterosaur post will notice the obvious parallel here, as I ended that one with a section on David Peters. It would however be unfair to directly compare Alan Feduccia to Peters, as the latter is basically a hack fraud from outside the main scientific community. Feduccia on the other hand used to be a respected authority in paleoornithology, but alas fell from grace due to his increasingly unscientific insistence to this day that birds could not possibly have descended from dinosaurs. In his 1980 book The Age of Birds his view on bird origins was still largely agnostic with being open to the dinosaur-hypothesis, but preferring Heilmann’s thecodont-hypothesis (a still respectable opinion at the time). This tone drastically changed in 1996 with his The Origin and Evolution of Birds, in which he called the dinosaur-hypothesis a “dogma” that other paleontologists just blindly follow (in reality it had become consensus due to robust cladistic studies). I will now list his main arguments and their respective refutations:
  • Birds could not have possibly evolved from theropod dinosaurs, as the theropod hand consists of digits I-III (corresponding with our thumb, index- and middle-finger) while the hand of birds consists of digits II-IV (corresponding to our index-, middle and ring-finger). The problem with this argument is that we actually do not know if the latter statement is true. It is merely an assumption born from embryology, as it is the norm for tetrapods to first lose their outermost digits during evolution. But just because this is the norm does not mean that this is always the case and no embryologist has ever said with confidence that the bird-hand truly consists of digits II-IV. Instead both embryological and fossil evidence heavily suggest that birds truly possess digits I-III and that this condition was inherited from theropods who first lost digits IV-V. Alternatively, evidence from dinosaurs such as Limusaurus might also suggest that the typical theropod hand could actually consist of digits II-IV.
  • Birds descend from Triassic reptiles such as Longisquama, Megalancosaurus and Drepanosaurus. The former because it possesses feather-like appendages on its back and a furcula-like breastbone, the latter because of their eerily bird-like heads. While Longisquama is certainly striking with its ornamentation we have actually no idea what these structures are and it seems very unlikely that they are related to feathers. Some have once suggested that they are not even part of the animal but actually vegetation that was preserved with it. If they are part of the animal, the resemblance to feathers is only superficial and very divergent in the finer detail and no living or fossil bird has comparable structures on its back. Most regard these today as extravagant display-scales. That they were used for gliding, like some have assumed, has also become untenable, as Longisquama had apparently only one singular row of these along the back. Its breastbone is not preserved well enough to determine if it was a furcula or just a pair of clavicles like earlier reptiles had, which is ironic as Feduccia does not consider the presence of very obvious furculae in theropods as evidence that they are related to birds (even though only birds and other theropods have this bone). Drepanosaurs like Megalancosaurus do have remarkably bird-like heads, but the rest of the body could not be more different. They possessed monkey-like prehensile tails with large hooks at the end and clamp-shaped hands like chameleons, which bear no resemblance to bird-hands at all. The forelimb of Drepanosaurus has been found to be so odd that its radius, ulna and a carpal formed a triangle, making it unable to swivel its forearm, let alone evolve a bird-like wing. Their heads are therefore a case of convergent evolution. Perhaps even more concerning is that none of these animals possess true archosaur-characteristics, while birds are by definition archosaurs. Instead, cladistic analyses find these three to either be related to protorosaurs like Tanystropheus or to even be basal diapsids. The latter assumption would make them less closely related to birds than even snakes.
  • The most bird-like dinosaurs start appearing only after Archaeopteryx evolves, therefore birds could not have descended from these dinosaurs. This “temporal-paradox” is essentially a strawman-argument used by Feduccia, as no paleontologist ever claimed that birds directly descend from Cretaceous forms like Deinonychus, but rather that these represent relics from an earlier time in evolution. The “paradox” has been largely resolved by Jurassic maniraptoran fossils even older than Archaeopteryx, such as Anchiornis and Scansoriopteryx. Feduccia curiously seems oblivious to the glaring temporal gap in his own hypothesis, which is that drepanosaurs (as well as any other non-dinosaurian candidate for bird-ancestry) went extinct at least 60 million years before Archaeopteryx showed up.
  • Birds evolved flight from climbing ancestors, while dinosaurs were purely terrestrial and therefore could not have been their ancestors. As mentioned earlier, while the former is likely true, the latter is not as many smaller dinosaurs would have been perfectly able to climb trees and lead scansorial lives. The maniraptoran group scansoriopterygidae is even named after their climbing abilities.
  • Any similarity between birds and dinosaurs is purely due to convergent evolution. This point is defeated by Feduccia’s aforementioned assertion that birds evolved from arboreal ancestors, while dinosaurs always lived on the ground and were unable to climb. For convergent evolution to take place, two unrelated animals generally need to have the same lifestyle and live in the same environments. How would purely terrestrial dinosaurs and purely arboreal birds then be able to converge on an almost identical anatomy? Feduccia never offered an explanation for this and also did not explain why dinosaurs and birds resemble each other even in the microstructure of their bones and eggshells, a feature that is nearly impossible to explain through convergence (Paul 2002).
Feduccia, being a paleoornithologist, was not an expert on dinosaurs, and many of his assertions seem to be arguments from ignorance. This is evident by his anachronistic view of dinosaurs being a polyphyletic group of unrelated, ectothermic terrible lizards and many anatomical errors he makes. He also still believed that feathers evolved from scales and purely in an aerodynamic context, not for insulation, something which was already disproven at the time by genetic and embryological studies (feathers are homologous with crocodilian scutes, not scales, and first evolved for insulation).While his work received praise from ornithologists at the time it was obviously panned by paleontologists. Ironically in the same year as The Origin and Evolution of BirdsSinosauropteryx, the first known non-avian dinosaur with proto-feathers preserved, was discovered. What was Feduccia’s response to that? First he called it the “Piltdown Dinosaur” (comparing it to the Piltdown Man and thereby insinuating that it was a fraud, a serious and unfounded allegation) and then claimed that its filaments are not actually feathers, but fossilized muscle-fibers or a collagen frill like a lizard’s (Paul 2002). While geochemical analyses could not be done on the Sinosauropteryx fossils, they were able to be carried out by Mary Schweitzer on identical structures in the alvarezsaur Shuvuuia in 1999. These studies found that the structures had a hollow central shaft and were keratinous in origin, meaning they were feathers and not collagen-fibres. A 2017 study carried out by Fiann Smithwick directly compared Sinosauropteryx’s filaments with the fossilized collagen-fibres on the ichthyosaur Stenopterygius and found no similarities. The supposed frill around Sinosauropteryx which Feduccia wanted to identify also turned out to be an artifact of fossil-preparation (Paul 2002). Despite all this, as far as I am aware, Feduccia still uses the collagen-fibre argument whenever proto-feathers are discovered on dinosaurs such as Yutyrannus or Beipiaosaurus. Once it became undeniable that dromaeosaurs and other maniraptoran dinosaurs possessed clearly avian pennaceous feathers he did a complete 180 degree turn and, after years of saying that these dinosaurs had absolutely no relation to birds, declared that the maniraptorans are actually not dinosaurs at all, but secondarily flightless birds descended from drepanosaurs. Theropods would therefore be polyphyletic and the likes of Velociraptor or Deinonychus would be non-dinosaurs that just convergently look like theropod dinosaurs. While he details this the most in his 2012 book Riddle of the Feathered Dragons, he came to this conclusion around 2002. This is curiously the same year in which Gregory S. Paul published his Dinosaurs of the Air. Indeed, it seems very much like Feduccia’s current position is a bastardization of Paul’s and Olshevsky’s neoflightless hypotheses, which he misappropriated to weasel his way out of admitting that he was wrong and that birds are dinosaurs. Funnily enough, Feduccia is so predictable that Paul already knew in 2002 that this would happen. Back then he wrote:

That argument that avepectoran dinosaurs [what Paul calls maniraptoriformes] were birds has not been presented but can nevertheless be anticipated. Let us assume that feathered Caudipteryx is an oviraptorosaur relative and, further, unambiguous contour feathers are found in the likes of dromaeosaurs, troodonts, oviraptorosaurs, or therizinosaurs [this has indeed come to pass at least for the first three]. According to the hypothesis that feathered archosaurs are birds with fingers II-IV rather than dinosaurs with digits I-III, the avepectorans were members of the bird clade, a clade that evolved independently from theropods that lacked ossified sternal plates! The problem with this notion is that avepectorans share far too many derived characteristics with other predatory dinosaurs for us to accept the extreme degree of convergence required to produce this version of dinosaur polyphyly. The closest known sister taxa to avepectorans clearly are less-derived averostran avetheropods such as Allosaurus, Ornitholestes and Coelurus. Consider that opponents of the dinosaur-bird link have long thought that the teeth of dromaeosaurs and troodonts are nonavian, so making these avepectorans post-urvogels only serves to reinforce the link. The presence of feathers in nonavepectoran dinosaurs also works against the possibility that dinosaurs were not monophyletic.”  (Paul 2002, p. 217-218)

Feduccia’s current BANDit-movement (short for Bird Are Not Dinosaurs) was therefore dead on arrival and has only become more untenable, with the discovery of feathers in even more compsognathids (which clearly were not secondarily flightless), the large Tyrannosaurus-relative Yutyrannus and ornithischians like Kulindadromeus and Psittacosaurus, as well as numerous cladistic analyses agreeing that Dinosauria is a monophyletic group (Feduccia’s ideas in fact only work if proper cladistic methods are completely disregarded). If we were then to follow his logic to the inevitable conclusion, Tyrannosaurus and Triceratops would not be dinosaurs at all, but secondarily flightless drepanosaur-descendants. In the early 2000s even the ornithological community turned on him for this nonsense. The harshest critique came from Richard Prum of the American Ornithologist’s Union, who argued that Feduccia’s methodology is downright pseudoscientific, as he has clearly already made a conclusion beforehand, which is that birds could not have possibly descended from those icky reptilian behemoths, and tries to bend reality to fit his conclusion. The way in which he is willing to constantly contradict himself and immediately move the goalposts once disproven, as well as the extreme denial, has even been compared to the strategies used by young-earth-creationists. His rhetoric is especially reminiscent of the "Teach the controversy"-attitude of intelligent design advocates. This is not helped by the fact that actual creationists have used some of Feduccia’s work to argue against birds evolving from reptiles. Feduccia’s last major argument in 2014 was that maniraptorans, through Scansoriopteryx, descend from the Triassic pterosaur-relative Scleromochlus (an idea he likely borrowed from Larry Martin, who unfortunately died in 2013 due to cancer). If somewhen in the future Feduccia will go further and argue that birds descend from pterosaurs, then even Gerhard Heilmann will be rolling in his grave.

If you liked this and other articles, please consider supporting me on Patreon. I am thankful for any amount, even just 1$, as it will help me at dedicating more time to this blog and related projects. Patrons also gain early access to the draft-versions of these posts.

Related Posts:
Literary Sources:
  • Bakker, Robert Thomas: The Dinosaur Heresies. New Theories Unlocking The Mystery of theDinosaurs and Their Extinction, New York 1986.
  • Benton, Michael James: Vertebrate Paleontology, 1990 (3. Edition from 2005, German translation).
  • Chatterjee, Sankar: The Rise of Birds. 225 Million years of Evolution, Baltimore 1997.
  • Desmond, Adrian: The Hot-Blooded Dinosaurs. A revolution in Paleontology, London 1975.
  • Heilmann, Gerhard: The Origin of Birds, London 1926.
  • Klepsch, Peter: Das grosse Buch der Saurier, Ravensburg 1989 (2000 reprint).
  • Knight, Charles Robert: Life through the Ages, New York 1946 (Commemorative Edition).
  • Marsh, Othniel Charles: The Dinosaurs of North America, Washington 1896.
  • Norman, David: TheIllustrated Encyclopedia of Dinosaurs, London 1985.
  • Paul, Gregory Scott: Predatory Dinosaurs of the World. A Complete Illustrated Guide, New York 1988.
  • Paul, Gregory Scott: Dinosaurs of the Air. The Evolution and Loss of Flight in Dinosaurs andBirds, Baltimore 2002.
  • Feduccia, Alan: The Age of Birds, Cambridge 1980.
  • Feduccia, Alan: The Origin and Evolution of Birds, New Haven 1996.
  • Feduccia, Alan: Riddle of the Feathered Dragons, New Haven 2012.
  • White, Steve: Dinosaur Art. The World's Greatest Paleoart, London 2012.
  • Witton, Mark: Recreating an Age of Reptiles, Marlborough 2017.

Online Sources/Further Reading:
Image Sources:
  • Fig. 1 left: Heilmann 1926, p. 200.
  • Fig. 1 right: Paul 2002, p. 129.
  • Fig. 2: Wikimedia.
  • Fig. 3: Janvier 1984.
  • Fig. 4: White 2012, p. 176.
  • Fig. 5: Knight 1946, p. 17.
  • Fig. 6: Wikimedia.
  • Fig. 7: Chatterjee 1997, p. 42.
  • Fig. 8: Heilmann 1926, p. 184.
  • Fig. 9: Klepsch 1989, p. 49
  • Fig. 10: Paul 1988, p. 211.
  • Fig. 11: Paul 2002, p. 291.
  • Fig. 12: Paul 1988, p. 241.
  • Fig. 13: Witton 2017, p. 78.


  1. Very detailed summary. A few comments-

    Archaeopteryx has since been shown to have a pneumatic quadrate and quadrate-braincase articulation (Alonso et al., 2004; Rauhut et al., 2018; Kundrat et al., 2019).

    Scleromochlus has been recently reinterpreted as a basal archosauriform (Bennett, 2020).

    Two Archaeopteryx specimens were originally identified as Compsognathus (Eichstatt and Solnhofen).

    I think Falcarius kills your therizinosaur foot hypothesis, since it's basically Ornitholestes-grade but has a normal theropod hallux. The enlarged first toe didn't appear until Alxasaurus at which point the animals were pretty massive. Retroverted halluces don't start appearing in birds until after Balaur branches off assuming it as Jeholornis-grade, with even Confuciusornis' hallux being only medially directed instead of posteriorly (Middleton, 2003).

    I believe the modern consensus is that bird manual digits are embryologically II-III-IV but that a frame-shift at some point caused this from the I-II-III inherited from maniraptoromorphs (e.g. Stewart et al., 2019).

    I don't think anybody nowadays thinks Longisquama's parafeathers are vegetation (e.g. Buchwitz and Voigt, 2012). Also, of course if birds were found to descend from prolacertiforms or simiosaurs, those groups would by definition become archosaurs. So it's not a valid objection to say they aren't archosaurs given the consensus placement of birds closer to crocs than to them, therefore birds can't be descended from them because birds are archosaurs.

    1. Thank you a lot for your helpful corrections. I based a lot of this article off Greg Paul's Dinosaurs of the Air (2002), so naturally some details have become outdated without me being aware of it.

      I did know about the recent reclassification of Scleromochlus, but decided to keep it conservative as that is a very new analysis and may therefore not become common consensus. I did want to mention it at least somewhere, but couldn't fit it into the flow of the text.

      I'll add a note to the therizinosaur section.

      Again I based the section around digit homology off Greg Paul's opinion. As far as I am aware most ideas about frame-shifts rely on Limusaurus, but its role in theropod hands is in itself controversial, so I was unsure to mention that hypothesis as well as to not make the text too complicated.

  2. This is some weirdass stuff people have proposed.

    "Birds and troodonts being close relatives is obvious, but the titanic Spinosaurus being an especially close relative of Archaeopteryx seems rather hard to believe."
    Hate to be pendatic, but I have to point out cladistics are cladistics. Whales being related to pangolins would seem bizarre, but lo and behold they are, in the clade Ferungulata (Simpson 1945) - although now that I think about it, a better comparsion would be whales and chevrotains, both being ruminants.

    1. Notice how I said "especially close relatives" because Wellenhofer's/Elzanowski's classification scheme suggests spinosaurids and birds to be almost direct sister-clades. That is a lot closer than whales are to pangolins or chevrotains, so none of your comparisons work. The better comparison would be the relationship between whales and hippos, but those share obvious similarities.

  3. The 'Haematothermia' concept did have a brief resurgence in the 1990s as early DNA phylogenies showed birds closer to mammals than reptiles. This error evaporated as molecular phylogenetic methods improved and a better understanding of the sources of bias in DNA evolution developed.

  4. "If somewhen in the future Feduccia will go further and argue that birds descend from pterosaurs, then even Gerhard Heilmann will be rolling in his grave."

    The only thing worse then that would be someone saying birds descend from arthropods.