Paleontology Page >> Dinosaur Myths >> Origin of Birds

Updated: 30 Dec 2005

Origin of Birds

Abstract

This page explores some of the debate surrounding the origin of birds; about how much we actually know about it, and about how well the hard evidence supports some of the popular accounts.

Keywords: bird, theropod, thecodont, dinosaur, Triassic, Jurassic, Cretaceous, bird origins, cursorial, Archeopteryx

Introduction

The popular press already has all the answers: birds evolved from dinosaurs – a hypothesis which many, including some professional paleontologists who really ought to know better, have adopted as an article of faith. In fact, the evidence is contradictory and difficult to interpret.

Central to any consideration of the origin of birds is the "original" bird, one of the most famous fossils known: Archaeopteryx. This animal is known from several fossils, approximately 150 million years old, and is still the oldest universally accepted bird fossil known. The earliest known maniraptorans and the so-called "feathered dinosaurs" post-date Archaeopteryx by at least 15 million years, and must be interpreted in that light. This is not a mere vagary of the fossil record, but the central hard fact of bird evolution which any theory of their origins must successfully explain. Commentators who are dismissive of the difficulties posed by the known fossil record are simply irresponsible.

Unfortunately, such commentary is quite characteristic of a debate which has been intemperate and, frankly, unworthy of the participants. The following is an attempt to present the data and diverse interpretations, in such a manner that the reader may draw their own conclusion. I will, however, confess up front to my own bias: I do not believe that birds are derived dinosaurs.

Defining a Bird

Lest we get lost in semantics, it is important to have a good idea of what we mean by the term bird.

The Fossil Record

Archaeopteryx

Despite intensive searching, the earliest known bird that everyone can agree upon is still the famous Archaeopteryx, known from only seven skeletons and an isolated feather, all recovered from the Tithonian (Late Jurassic, ~152 Ma) Solnhofen Limestone of Germany.

Recent evidence from a "new" specimen (recovered from one private collection and now contraversially housed in another, at the Wyoming Dinosaur Center) "shows that the osteology of [Archaeopteryx] is similar to that of non-avian theropod dinosaurs" (Mayr et al. 2005, p. 1483). Some paleontologists believe that the closest dinosaur relatives to Archaeopteryx, are the troodontids, typified by Troodon and Saurornitholestes from the Late Cretaceous (Carpenter? ### verify ###).

Archaeopteryx does not provide much information about the origin of feathers, however, because its feathers are almost identical to those of modern birds.

For some reason which remains inexplicable, the English astronomer Fred Hoyle became convinced that the Archeopteryx fossils were faked, publishing his views in 1986 (Hoyle & Wickramasinghe 1986). But Hoyle's attack on the authenticity and interpretation of Archeopteryx is just another example of the ignorant hubris which possesses some physical scientists – the same sort of vainglory which prompted Ernest Rutherford to famously declaim that all science was either physics or stamp-collecting. The Archeopteryx forgery claim was not his only nutty theory, and the less said about Fred Hoyle in this context, the better. He is better remembered for such achievements as predicting the resonant energy level in the carbon nucleus.

The small theropod dinosaur, Compsognathus, has also been recovered from the Solnhofen. This fossil record presents difficult problems for advocates of the theropod hypothesis: birds (specifically Archaeopteryx) are supposed to be most closely related to the dromaeosaurids, which do not appear in the fossil record until Albian times (mid Cretaceous, about 110 Ma). Yet Compsognathus, which is believed to have diverged from the theropod lineage long before the evolution of the dromaeosaurids, occurs alongside Archaeopteryx 40 million years earlier. At present, only the vagaries of the fossil reord can be invoked to ‘explain’ the stratigraphic disjunction; our present understanding is unsatisfactory.

Other Earliest Birds

Besides Archaeopteryx, the earliest fossil record of birds also provides us with Confuciusornis and Protopteryx, both from the early Cretaceous.

Confuciusornis dates from the 120-150 Ma (?Barremian) Yixian Formation of Sihetun, Liaoning Province, China. This creature is the earliest known bird to have a true beak. Although toothless, it is probably independent of the two major radiations of toothless birds, enantiornitheans and neornitheans. Confuciusornis had three functional fingers in the hand: the thumb, index and middle fingers. "Birds retain these three fingers, although they support flight rather than grasping. We have long wondered how dinosaurs made the transition from a grasping to a flying hand, and Confuciusornis gives us new insight into that problem. Confuciusornis still has fully functional raptorial claws on its thumb and middle fingers, but its index finger – the finger that supports the flight feathers – is composed of broad, flat bones and a reduced claw. As with other basal maniraptors [note the presumption of a theropod ancestry - CC], the thumb and middle fingers converge on one another while grasping in Confuciusornis, enabling its hand to support flight while still retaining some grasping ability" (Peabody Museum web page).

Protopteryx comes from Fengning County, Hebei Province, China, also from Yixian Formation, and is considered by its authors to represent the most primitive enantiornithine. "Protopteryx preserved some interesting types of feathers with characters between those of scales and modern feathers, thus providing fossil avian evidence for the origin of feathers from elongated scales in reptiles" (Zhang & Zhou 2000, p. 1955). To date there is only one species, Protopteryx fengningensis, which possesses teeth in the premaxilla and the dentary, as in most Mesozoic birds, and is about the size of a starling.

Feathered Dinosaurs

What are the so-called "feathered dinosaurs" and what is their significance?

The first question is easy......

Sinosauropteryx prima from Liaoning, China, closely related to Compsognathus

In the last few years, many fossils of feathered dinosaurs have been found near Yianxin, in Liaoning Province, China. Two new Chinese feathered dinosaurs dating from between 145 and 125 million years ago (during the late Jurassic and early Cretaceous periods) have been found, Protarchaeopteryx robusta and Caudipteryx zoui. Their features are more dinosaur-like than bird-like, and they are considered to be theropod dinosaurs. Their feathers were symmetrical, which indicate that they could not fly (flightless birds have symmetrical feathers while those that fly have asymmetrical ones).

The second question, to interpret the significance of the feathered dinosaurs, is more problematic. These finds have been advanced as evidence to reinforce the theory that birds are descended from dinosaurs.

But many of these arguments are fatally flawed. Consider, for example, the statement "Sinosauropteryx had a coat of downy, feather-like fibers that are perhaps the forerunner of feathers" (from the Enchanted Learning site).

Archaeopteryx already had REAL feathers at ~150 Ma, a good 15 or 20 million years before Sinosauropteryx’s supposed "forerunner" structures.

Presumably this kind of over-zealous interpretation is being advanced by lay people; one sincerely hopes that the professional researchers graduating from our universities today would not make such elementary errors of logic.

"Modern feathers probably evolved through the following stages: (i) elongation of scales, (ii) appearance of a central shaft, (iii) differentiation of vanes into barbs, and (iv) appearance of barbules and barbicel" (Zhang & Zhou 2000, p. 1957).

Insert a list of them, and the key features. Order by age if there is enough info to do so.

145-125 Ma - Caudipteryx becomes the fourth “feathered” animal to be found near the tiny village of Sihetun in Liaoning Province. It joins Protarchaeopteryx and Confuciusornis — a creature with relatively short, clawed wings that was probably one of the first birds to fly well — and Sinosauropteryx, one of the most important dinosaur finds of the 20th century.

130-135 Ma - Sinosauropteryx prima from Liaoning, China, closely related to Compsognathus

88 Ma - There are other dinosaurs with bird-like features. Unenlagia, from 88 million year old beds of Patagonia folded its arms in the same way that birds do.

Don’t know where (if anywhere) to put this - Oviraptor from Mongolia, once thought to be an egg thief, is now known from several specimens crouching over nests of their own eggs in exactly the same pose as brooding emus. And a reassessment of other theropods reveals such bird-like features as hollow bones and a foot with three functional toes, bird-like features that appeared over 50 million years before the first feeble flying flaps flung Archaeopteryx into the air.

Sinosauropteryx - see Chen, P.; Dong, Z.; Zhen, S. 1998: Nature, 391: 147
Beipiaosaurus - see Xu, X.; Tang, Z.; Wang, X. 1999: Nature 399: 350
Sinornithosaurus - see Xu, X.; Wang, X.; Wu, X. 1999: Nature 401: 262
"two controversial feathered dinosaurs" (??? Protarchaeopteryx robusta and Caudipteryx zoui ???) - see Feduccia, A. 1999: The Origin and Evolution of Birds. Yale University Press, 405 pp.; Ji, Q.; Currie, P.J.; Norell, M.A. 1998: Nature 393: 753; Zhou, Z.; Wang, X. 2000: Vertebr. Palasiat. 38: 111

This is what we can say: If the structures are indeed feathers, and if we believe that feathers have only ever evolved once, then we have a compelling case for an apomorphy.

Here again, though, it is the age of the fossils which presents the greatest theoretical difficulties because, Sinosauropteryx being the earliest of them, all of the so-called "feathered dinosaur" discoveries post-date Archaeopteryx by tens of Ma. The only possible relevance their supposed feathered condition can have is as a candidate derived character (a synapomorphy or symplesiomorphy) shared with birds. Neither interpretation lends more than circumstantial support to the theropod hypothesis, nor sheds any light on the development of feathers, or birds themselves, for that matter.

"The association of the hairlike structures of small theropod dinosaurs with feather origins is based on three known theropods, including Sinosauropteryx, Beipiaosaurus, and Sinornithosaurus (Chen et al. 1998; Xu, Tang & Wang 1999; Xu, Wang & Wu 1999), yet there is no convincing evidence that they are branched. More work needs to be done to reveal their implications for the evolution of the origin of feathers in birds. The recent discoveries of true avian feathers in Caudipteryx and Protarchaeopteryx have been regarded by many as the strong evidence for the presence of feathered dinosaurs (Ji et al. 1998; Zhou & Wang 2000); however, some believe that Caudipteryx could be a secondarily flightless bird, and therefore the feathers in Caudipteryx were also secondarily reduced (Feduccia 1999; Jones et al. 2000). ... [It is] uncertain whether the feathers in Caudipteryx were independently developed or were the primitive type of feathers defining birds and their immediate ancestors" (Zhang & Zhou 2000, p. 1957).

Longisquama

Longisquama insignis was a small archosaur known from fossils occurring in Norian (Late Triassic, ~220 Ma) lacustrine deposits of Kyrgyzstan, central Asia, and possessing highly unusual integumentary appendages, which have been interpreted as feathers:

"Along its dorsal axis Longisquama bore a series of paired integumentary appendages that resembled avian feathers in many details, especially in the anatomy of the basal region. The latter is sufficiently similar to the calamus of modern feathers that each probably represents the culmination of virtually identical morphogenetic processes. The exact relationship of Longisquama to birds is uncertain. Nevertheless, we interpret Longisquama’s elongate integumentary appendages as nonavian feathers and suggest that they are probably homologous with avian feathers" (Jones et al. 2000, p. 2202, Abstract).

If correct, this represents an occurrence of feathers almost 75 Ma before those of Archaeopteryx, the first known bird from the Late Jurassic of Germany.

"The similarities between the non-avian ‘feathers’ in Longisquama and modern avian feathers suggest that the earliest stages of the evolution of avian feathers were more complex than we had expected" (Zhang & Zhou 2000, p. 1957).

Protoavis

A key fossil in the debate may be Protoavis Chatterjee 1991 (date doesn't seem correct; cf. 1987 publication) from the Triassic of Texas, which Chatterjee interprets as a bird (Chatterjee 1997).

Note from http://www.abc.net.au/science/slab/dinobird/story.htm:

Protoavis is known from a handful of bony fragments collected from 200 million-year-old rocks in Texas. At this time the dinosaurs were just getting started and it’s 50 million years before Archaeopteryx.

If correct, the theropod view is untenable and the basal archosaur hypothesis would carry the day.

Its partly toothless jaw and keel-like breast bone were like those of birds. It also had a tail, dinosaur-like rear legs, and hollow bones. There is some dispute about whether this animal was a bird or a dinosaur; the answer depends partly on whether the Protoavis fossil belongs to one or two different genera.

However, the available material is very fragmentary, and both the interpretation of some of the bones and the association of the various elements into a skeleton has been difficult to support. The skull has been re-constructed as very avian, but most of the skull is either absent or poorly preserved.

(Read more. Also see.)

Trace Evidence

The oldest reported bird-like footprints were been recovered from the Norian-Rhaetian (Upper Triassic) Santo Domingo Formation of northwest Argentina, occurring 80 metres below a horizon dated (⁴⁰Ar/³⁹Ar) at 212.5 ± 7.0 Ma (Melchor et al. 2002, p. 936). Key features of the footprints include (p. 937) a greater width than length (excluding the hallux), slender digits, a wide angle between digits II and IV, a posterior or posteromedial hallux, slender claws curving away from the foot axis, a sole impression where digits II-IV converge in some prints, as well as some behavioural hints, such as high footprint density without a preferred orientation.

Although the authors cautiously conclude that these fossils "can only be attributed to an unknown group of theropods showing some avian characters" (p. 937), no Triassic theropod body fossil is known with the "avian" reversed hallux. A possible interpretation is that the Aves had already diverged from the theropod lineage as early as the Late Triassic, a view which is difficult to reconcile with the standard Maniraptoran hypothesis for the origins of birds.

where does the reversed hallux appear in the theropod body-fossil record?

Origin Hypotheses

Today, three possibilities for the origins of birds are recognised: The first is that they evolved from some unknown group of basal archosaurs, probably in the Triassic Period (see fig. 1, line B). Second, is that they are a sister group to the Crocodylians, perhaps arising from within the sphenosuchian crocodylomorphs in the Early Jurassic (fig. 1, line A). But probably the most widely held view among paleontologists – though unfortunately misrepresented as the only credible modern view by the popular media – is for an ancestry among the theropod dinosaurs, specifically the Maniraptora, in the Middle to early Late Jurassic (fig. 1, line C).

The most decisive resolution to the debate would be the discovery of a fossil bird or bird-ancestor from Triassic rocks. However, as we have seen, the Triassic evidence is problematic and poorly documented; the oldest universally accepted bird fossil remains the Late Jurassic Archeopteryx, which is certainly inconvenient for the theropod camp, though not a fully conclusive death-blow to this theory, which we shall examine first.

The Theropod Hypothesis

As a generalisation, the theropod hypothesis tends to be favoured by paleontologists, whereas a good many ornithologists bridle at the implication of a cursorial ("ground-up") origin of avian flight (see below).

The main support for the theropod (or dinosaurian) hypothesis comes from the very close similarity of the avian skeleton to that of certain bipedal theropod dinosaurs, specifically the Maniraptora. The similarities are most pronounced in the structure of the pelvis and, depending upon one's interpretation, the posterior limbs (the legs). As long ago as the 1860s, T.H. Huxley drew "attention to the remarkable similarity of the avian skeleton to that of certain reptiles and concluded that the birds had descended from dinosaurs. ... [R]ecently the dinosaur origin has been proclaimed by the cladists with such vigor that at present it seems to be the most widely accepted explanation of the origin of birds" (Mayr 2001, pp. 226-227).

Fig. 1: A simple cladogram, having no implied time dimension, showing three possible insertion points for the bird lineage. Although diagrams like this clearly illustrate the relationships between different taxa, it is precisely their (typical) failure to incorporate a scaled time dimension which hides a multitude of sins - see fig. 2.

Ornithologist Alan Feduccia (Feduccia 1999, p. vii) criticises the vertebrate paleontologists who "have accepted a formal, rigid cladistic methodology as the sine qua non for reconstructing phylogenies and, using this scheme in an almost religious manner, have [ironically!] discarded geological time as a tool in deciphering evolution...." One need only read a year's worth of articles in Nature or Science, or peruse a "techno-popular" compilation such as the Encyclopedia of Dinosaurs to corroborate this observation, particularly the absence of a quantitative time dimension to all this discussion. (For a particularly ridiculous example, see the cladogram on pages 78-79 of Prum & Brush 2004. Also see fig. 2). Of course none of this is the real professionals' literature; nevertheless, Feduccia's criticism is well-deserved.

However, there are numerous theoretical difficulties to the theropod hypothesis also. The most challenging are:

Fossil ages - Archeopteryx is not only inconveniently old itself, but it is clearly a highly derived animal, which had evolved "along avian lines" for many millions of years already.
Evolutionary considerations - Deriving a bird from a large, ground dwelling dinosaur requires some explaining. Mostly, the explanations we have to date are unconvincing.
Manus homologies - Despite some startling anatomical similarities, there are other respects in which bird skeletons do not resemble theropod dinosaurs. Chief among these are the three digits of the hand, which comprise fingers 1, 2, and 3 in dinosaurs, but 2, 3, and 4 in birds. "It is quite impossible to derive the avian digits from those of dinosaurs" (Mayr 2001, p. 68).

We examine these points in turn.

Fossil Ages

To get a little more technical, the real problem here is not that dromaeosaurid fossils appear so late; it is the temporal coincidence of a stem group organism, Compsognathus, with a highly derived crown member of the same lineage, Archaeopteryx. This problem requires more than a glib appeal to sampling inadequacies.

cgm2: = cgm1 + ‘A’ + ‘C’

Fig. 2: Another cladogram, where the ‘A’ and ‘C’ represent Archaeopteryx and Compsognathus, respectively. Now cladograms like this one show us the (hypothesised) order of branching but they are not biostratigraphic range charts; one cannot assume that a vertical line drawn on a cladogram will throw up a list of contemporaneous taxa. But this diagram does illustrate the accommodations which must be made to explain the co-occurrence of Archaeopteryx and Compsognathus in the same beds: Either ....

Discussion

Feathers

In some respects, the origin of birds comes down to the question ‘what is a feather?’ so, before looking at the fossil record more closely, let us digress briefly to consider this question.

"Feathers are highly ordered, hierarchical branched structures that confer birds with the ability of flight. Discoveries of fossilized dinosaurs in China bearing ‘feather-like’ structures have prompted interest in the origin and evolution of feathers. However, there is uncertainty about whether the irregularly branched integumentary fibres on dinosaurs such as Sinornithosaurus are truly feathers, and whether an integumentary appendage with a major central shaft and notched edges is a non-avian feather or a proto-feather. Here, we use a developmental approach to analyse molecular mechanisms in feather-branching morphogenesis. We have used the replication-competent avian sarcoma retrovirus to deliver exogenous genes to regenerating flight feather follicles of chickens. We show that the antagonistic balance between noggin and bone morphogenetic protein 4 (BMP4) has a critical role in feather branching, with BMP4 promoting rachis formation and barb fusion, and noggin enhancing rachis and barb branching. Furthermore, we show that sonic hedgehog (Shh) is essential for inducing apoptosis of the marginal plate epithelia, which results in spaces between barbs. Our analyses identify the molecular pathways underlying the topological transformation of feathers from cylindrical epithelia to the hierarchical branched structures, and provide insights on the possible developmental mechanisms in the evolution of feather forms" (Yu et al. 2002, p. 308, Abstract).

Convergence

"Convergent evolution is a central theme of vertebrate history, and it is a predominant thread throughout the avian ranks. Everywhere we look birds from different evolutionary backgrounds have come to look alike.... [C]onvergence is often so complete, so elusive and subtle, that clues can be difficult to ascertain. ... In early classifications many convergent pairs were placed together, for example swifts and swallows, and hawks and owls, illustrating that convergence is an insidious and treacherous trap, baited and waiting for the unsuspecting worker.

"Nowhere has the trap been more successful than in luring paleontologists to the theropod dinosaurian origin of birds" (Feduccia1999, pp. vii-viii).

Conclusion

Here again, the evidence is complex and appears, in light of the present state of our knowledge, contradictory. The point to draw is not that the "popular" hypothesis is wrong, but that the jury is still out. Claims that birds arose from the Maniraptora are just Bad Science: we simply do not know.

References

Carpenter, Kenneth 2002 (in press): Dinosauria (Dinosaurs). In Encyclopedia of Life Sciences. Nature Publishing Group, Macmillan.

Chatterjee, S. 1997: The Rise of Birds: 225 Million Years of Evolution. Johns Hopkins University Press, 312 pp.

Chen, P.; Dong, Z.; Zhen, S. 1998: Nature, 391: 147

Feduccia, A. 1999: The Origin and Evolution of Birds. Yale University Press, 405 pp.

Fisher et al. (2000): Science 288: 503.

Hoyle, F.; Wickramasinghe, C. 1986: Archaeopteryx, the primordial bird; a case of fossil forgery. Christopher Davies, Swansea.

Hutchinson, John R.; Garcia, Mariano (2002): Tyrannosaurus Was Not a Fast Runner. Nature 415, 1018 - 1021.

Ji, Q.; Currie, P.J.; Norell, M.A. 1998: Nature 393: 753; Zhou, Z.

Jones, T.D.; Farlow, J.O.; Ruben, J.A.; Henderson, D.H.; Hillenius, W.J. 2000: Nature 406: 716

Jones, Terry D.; Ruben, John A.;Martin, Larry D.; Kurochkin, Evgeny N.; Feduccia, Alan; Maderson, Paul F. A.; Hillenius, Willem J.; Geist, Nicholas R.; Alifanov, Vladimir 2000: Nonavian Feathers in a Late Triassic Archosaur. Science 288: 2202-2205.

Mayr, Gerald; Pohl, Burkhard; Peters, D. Stefan 2005: A well-preserved Archaeopteryx specimen with theropod features. Science 310: 1483-1486.

Melchor, Ricardo N.; de Valais, Silvina; Genise, Jorge F. 2002: Bird-Like Fossil Footprints from the Late Triassic. Nature 417: 936-937.

Norell, Mark A.; Gaffney, Eugene S.; Dingus, Lowell 1995: Discovering Dinosaurs. Little, Brown and Company, 204 pp.

Ostrom, J. H. 1996: The Questionable validity of Protoavis. Archaeopteryx 14: 39-42.

Prum, Richard O.; Brush, Alan H. 2004: Hair, scales, fur, feathers. Scientific American 14 (2): 74-81.

Xu, X.; Tang, Z.; Wang, X. 1999: Nature 399: 350.

Xu, X.; Wang, X.; Wu, X. 1999: Nature 401: 262

Yu, Mingke; Wu, Ping; Widelitz, Randall B.; Chuong, Cheng-Ming 2002: The Morphogenesis of Feathers. Nature 420, 308 - 312

Zhang, Fucheng; Zhou, Zhonghe 2000: A Primitive Enantiornithine Bird and the Origin of Feathers. Science 290: 1955-1959.

Zhou, Z.; Wang, X. 2000: Vertebr. Palasiat. 38: 111

Peripatus Home Page