Bird evolution

A phylogenetic tree featuring birds (Neornithes), their ancestors (ancestral taxa such as Coelurosauria and Therapoda) and related groups. Abbreviations:
AS = Airsacs
ASW = Airsacs supported by immobile femur and a knee-driven walk
BL = Bellow lungs

We're all Homo here Evolution

Relevant Hominids

Charles Darwin Alfred Russel Wallace Gregor Mendel Richard Dawkins Jerry Coyne

A Gradual Science

Christian Schwabe Directed mutation Signal detection theory Biology

Plain Monkey Business

Young Earth Creationism Old Earth Creationism Intelligent Design Microevolution vs. Macroevolution v - t - e

The debate over the phylogeny and origins of avians (or bird evolution) has been lengthy and important to the study of evolution. At the root of the debate has been establishing from which taxonomic group of dinosaurs the first bird originated. One of the major parts of the debate has been whether birds originated from "thecodonts" (literally, "socket-teeth"), a now-obsolete term for a diverse group of dinosaur relatives that are now regarded as an unnatural grouping of dinosaur and crocodilian relatives, a so-called "wastebasket taxon".

Introduction

Science as a very human enterprise simultaneously exemplifies the best attributes of humanity and the most disappointing ones. One idea sees science as an endeavor in which scholastic restraint and respect reign triumphant, and where debate remains purely objective. Unfortunately, this is not always the case. Scientists wage bitter, vitriolic feuds, which transcend any evidence to become ideological wars. The famed rivalry between Sir Richard Owen, eminent anatomist and favorite of the U.K. Royal Family, and Thomas Huxley, "Darwin's Bulldog", provides just one example. The later rivalry between Edward Drinker Cope and Othniel C. Marsh is another.

Indeed, few disciplines produce such acrimony and contention as paleontology. Of the many debates in said field, none is more bitter, or more public, than the debate over the origin and evolution of birds. Ironically enough, we in the 21^st century are replaying a drama that was enacted over a century ago, in the England of Queen Victoria, which pitted Special Creation against Darwinian evolution. While today the contenders are modern dinosaur paleontology and classical paleornithology, the question remains very much the same.

Did birds evolve from theropod dinosaurs (a large and diverse group of bipedal dinosaurs that included the famous Tyrannosaurus rex), or do they descend from "thecodont" archosaurs? This article is intended as a more rigorous treatment of select aspects of the debate on avian origins and evolution, with specific emphasis on answering in some detail the arguments proposed by those who favor a "thecodont" ancestry for birds. By showing that key components of "thecodont" ancestry are demonstrably incorrect, it is hoped that the serious problems with this hypothesis can be further elucidated.

Key concepts

Here are some useful concepts that one should understand before proceeding.

Evolution

Phylogenetic terminology

• Cladistics — the taxonomical technique of grouping organisms into clades (or monophyletic groups) that have a common ancestor (common descent). Paleontologists usually have to rely on the morphology (form) of fossils rather than genetics because DNA generally decays and disappears within thousands of years unless it is preserved in amber.
• Convergent evolution — researchers have observed that many groups of organisms have evolved analogous traits to accomplish similar functions. For example insects, pterosaurs, birds and bats have all independently evolved traits that enable flight with wings.
• Homology — characteristics that are derived from a common ancestor, e.g. bat wings are homologous to human arms, horse hooves are homologous to human fingers and toes.
• Homoplasy or homoplastic ("same molding", i.e., having the same form) — encompasses both concepts of convergent evolution and parallel evolution.
• Parallel evolution — This differs from convergent evolution in that the groups of organisms have a common ancestor. For example marsupials and placental mammals have a common ancestor, and some animals in each group have evolved similar ecological functions.
• Phylogeny — the relationships between groups of animals according to their evolutionary heritage
• Phylogenetic bracketing — a method to estimate the likelihood of a trait developing over time based on species relations on a phylogenetic tree. It is especially useful for fossils where DNA is unavailable.
• Synapomorphy and apomorphy — in phylogenetics, these refer to genetically-derived traits derived from a common ancestor in clade. This is the opposite of homoplasy, in which traits are similar but not due to shared genetics. Making claims about this becomes more controversial for fossils because DNA is usually not available, and so claims must be based on morphology.

Phylogeny

See the main article on this topic: Phylogeny

Archaeopteryx lithographica fossil, cast of the London specimen (BMNH 37001)

The current scientific consensus is that this animal, while possessing numerous avian characteristics, is also a theropod dinosaur.

• Archaeopteryx ("old wing") — a fossil of this genus was first discovered in 1861, two years after the publication of Darwin's On the Origin of Species. It has non-avian dinosaur characteristics and is hence considered within the group avialae.
• Aves ("birds") — the phylogenetic class that includes all living birds, but is variously defined for ancient species. These are the only living dinosaurs.
• Avialae — non-avian feathered dinosaurs that could fly
• Deinonychus — The study of Deinonychus fossils, particularly the forelimbs, in the 1960s led to the modern consensus that birds descend from dinosaurs.
• Neornithes ("new fowls") — all modern birds as well as all bird ancestors up to the point where they diverged from flying feathered dinosaurs
• Proaves — a hypothetical extinct taxon that encompassed proto-birds
• Thecodonts — a now-obsolete term for a diverse group of non-dinosaurs that are now regarded as an unnatural grouping, and therefore do not form a clade (see the first paragraph, above)

Anatomy

A proposed evolutionary model of the furcula. Columba is a genus of pigeons.

• Clavicle — the collarbone in humans. In birds and some other theropod dinosaurs, it became fused to form the furcula.
• Furcula (plural: furculae) — The furcula is the "wishbone" found in birds and some other theropods. This bone strengthens the thorax structures to withstand rigors of movement during flight.
• Manus — the hand in humans, or homologous parts of anatomy in other animals (e.g. fore-hooves in horses, wings in bats and most birds, fore-flippers in cetaceans)

Historical review

The opposition to dinosaurian origin for birds dates from 1860. Reasons have varied. Richard Owen maintained that Archaeopteryx lithographica, despite basal attributes, was a bird without issue due to his opposition to Darwinian evolution. H.G. Seeley elucidated the first cogent homoplastic argument concerning the affinities of Archaeopteryx in 1866. Seeley's arguments were more objective than Owen's, who had ulterior motives, and the same can be said of those who picked up his conceptual legacy.

Despite the veracity of the work by both those supporting dinosaurian origin for birds (e.g., Huxley 1868, 1870, Baur 1883) and those advocating non-dinosaurian origin, the four decades following the initial description of Archaeopteryx lithographica were nebulous, with a plethora of theories for avian origin advanced. Putative avian ancestors ranged from therapsid mammals, to pterosaurs. Consensus in the paleontological community on the affinities of birds did not emerge until the early 20^th century.

It was the work of the enterprising Robert Broom on "thecodonts" from the Triassic redbeds of South Africa that paved the way for the ascendancy of the homoplastic arguments first formulated in the 1860s and 1870s. Broom's 1913 description of the fabulous Euparkeria capensis was to be the cornerstone upon which Gerhard Heilmann, a Danish artist and scientist, would build the seminal argument for the "pseudosuchian hypothesis" for the origin and evolution of birds. Heilmann's 1926 treatise, The Origin of Birds, collated data from previous research on Archaeopteryx, and undertook detailed comparative analysis of Dinosauria then known, and Triassic Archosauria. Heilmann concluded that due to the absence or reduction of clavicles ("collarbone") and therefore furculae in dinosaurs, the otherwise facile argument for dinosaurian ancestry for birds, fell apart. Heilmann also rejected the cursorial origin of flight, most cogently advanced by the eccentric and brilliant Franz Baron Nopsca.

Advanced "thecodonts" displaying prominent or at least generalized clavicles were presented as more likely candidates for avian ancestors. Heilmann's study differed notably from the work of his intellectual antecedents, such as Seeley, by proposing what we would now refer to as parallelism, as opposed to mere convergence, as the causal agent for morphological similarities between dinosaurs and birds. Heilmann's hypothesis was based in part of an overly strict interpretation of "Dollo's Law", and to a large extent was an artifact of the comparatively poor theropod fossil record in the 1920s. Nevertheless, Heilmann's work marked an upheaval in vertebrate paleontology, and the "thecodont hypothesis"; as elucidated in The Origin of Birds, became the paradigm for avian phylogeny. Alfred Sherwood Romer, one of the most influential vertebrate paleontologists of the modern era, championed it, as did others.

Deinonychus antirrhopus and Tenontosaurus tilletti fossils

The manuses of Deinonychus (left) and Archaeopteryx (right) compared.

Dinosaurian origin did not regain prominence until the 1964 discovery of Deinonychus antirrhopus by John H. Ostrom from Lower Cretaceous beds of the Cloverly Formation, in Montana. The subsequent description of D. antirrhopus led to an extensive reevaluation of the Archaeopteryx material from which Ostrom et al. concluded that small, cursorial theropod dinosaurs were paravian to a degree incompatible with convergence. In their 1974 analysis of dinosaur systematics, Bakker & Galton championed Ostrom's revision of dinosaurian origin, and subsequent analyses throughout the latter half of the 1970s and up to present have substantiated this conclusion. Nevertheless, there remains opposition to these conclusions. Chief among those who discount the theropod ancestry of birds are L. Martin, M. K. Hecht, S. F. Tarsitano, and most notably, Alan Feduccia — the most vocal, influential, and vitriolic of all "thecodont hypothesis" proponents. Feduccia is also the most publicly recognizable, and arguably serves as a spokesperson of "thecodont ancestry". His well-known books, The Age of Birds and The Origin and Evolution of Birds, are standard references for anyone seeking to discount dinosaurian origin. His study of flight capability in the Urvogel, critique of feathery integument in Maniraptoriformes, general reviews of dinosaurian origin, work on "avimorph thecodonts", studies on digital homology in Aves, and his contentious model of explosive radiation of Aves post-K-Pg boundary are the primary documents which elucidate his views on avian origin and evolution.

Several signature concepts emerge from this body of work:

1. Archaeopteryx was too avian to be closely linked to terminal Jurassic/early Cretaceous coelurosaurs
2. Lack of digital homology in theropods and birds, excludes theropods from the avian ancestral stock.
3. Any morphologic similarities between dinosaurs and birds are merely the result of convergence.
4. Birds arose from "avimorph thecodonts".
5. Aves suffered catastrophic extinctions at the K-Pg Boundary followed by explosive radiation in the early Tertiary from which Neornithes was derived.

These postulates are all inextricably linked to Feduccia's arguments against dinosaur origin for birds. There are serious problems with all of these assertions, and their validity must be critically reviewed, starting with the primordial bird itself.

The argument against the thecodont hypothesis, relies on showing that Archaeopteryx lies evolutionarily between Neornithes and Theropods, but that Archaeopteryx is more Theropods-like. Making this argument also relies more solidly on clades, whereas the thecodont hypothesis relies on the non-cladistic thecodont grouping. The argument for Archaeopteryx being a Thecodont ancestor to birds can be summarized as follows:

Anatomy	Thecodont	Archaeopteryx	Modern bird
Teeth?	Sharp	Reptilian	No
Furcula (wishbone)	No	Yes	Yes
Clawed hand	Yes	Yes	No
Theropod pubis	Yes	Yes	No
Large breastbone	No	No	Yes
Long bony tail	Yes	Yes	No
Reversed 1st toe	Yes	Yes	Yes
Fused metatarsals	No	No	Yes

The sham Archaeopteryx — the Urvogel according to Feduccia & company

A central tenet of the "thecodont hypothesis"; is that Archaeopteryx is more avian than has been claimed (and by default less dinosaurian).

A detailed review of purported avian apomorphies in the Urvogel is in order, starting with the skull follows:

Skull

Archaeopteryx lithographica, the Berlin specimen

Feduccia (1996) has offered the principal arguments for the avian characters of the skull in Archaeopteryx lithographica, based primarily on the restorations figured by Paul Buhler in his 1985 review of HMN 1880 (the Berlin specimen), JM 2257 (the Eichstatt specimen), BMNH 37001 (the London specimen). The Buhler restoration, later accepted without alteration by Feduccia, asserts five diagnostic avian traits for the Urvogel skull:

1. A propulsion joint between the quadrate and pterygoid
2. An orbital process of the quadrate
3. A prokinetic nasal flexion zone
4. The lack of a dorsal temporal bar
5. Lack of a complete postorbital bar caudal to the orbit

However, the fossils at hand do not support such restorations and the claim that Archaeopteryx displays modifications allowing streptostyly and prokinesis (types of cranial kinesis), is entirely without foundation.

In most birds, when the mouth is opened, the mandible and palate are pushed forward simultaneously and the beak rotates dorsally relative to the braincase, due to a weak flexion zone in the nasals. This prokinetic motion is made possible by several modifications of the cranial elements. Starting caudally and proceeding rostrally, these modifications include a double-headed condyle on the proximal articular surface of the quadrate, which renders the quadrate streptostylic, an orbital process of the quadrate, no descending process of the squamosal to inhibit forward rotation of the quadrate, a propulsion joint between the quadrate and pterygoid, the lack of a dorsal temporal bar bounding the lateral temporal fenestra, and reduction or loss of a complete postorbital bar caudal to the orbit. Comparative to most archosaurs, this suite of traits, and the biomechanical processes they facilitate, are apomorphic.

The presence of such a functional system in Archaeopteryx would significantly bolster arguments that the Urvogel was more avian than has been supposed. But, do the fossils actually display this apomorphic character state?

The quadrate in Archaeopteryx, restored as nearly vertical in support of a propulsion joint, is in fact strongly procumbent (contra Martin 1983, Buhler 1985, Feduccia 1996, 1999). While arguments that this orientation is a preservation artifact have been made, restoration of a strongly vertical quadrate is not mechanically possible without incorrectly articulating the temporal region as a whole and obliterating the infratemporal fenestra. Contrary to the impression offered by Haubitz et al. (1988), Feduccia (1996) and other researchers, the proximal articular surface of the quadrate is not bifurcated, and the quadrate is excluded from the braincase.

Based largely on Buhler's work, Feduccia (1996) argued that the rostromedial surface of the quadrate was distinctly enlarged, bearing a robust orbital expansion similar to that seen in most birds. However, such a reconstruction of the quadrate is entirely contradicted by the material in JM 2257 and BSP 1999, the Munich specimen. In these fossils a smooth, convex arc is observed, with no indication of the alleged orbital process.

In total then, the quadrate is seen to be lacking a double-headed articular surface with the squamosal and similarly lacking an orbital process on the rostromedial surface of the bone. It can be concluded that the quadrate displays none of the apomorphic conditions ascribed to it by Feduccia et al.

And how does the quadrate articulate with the pterygoid? In the diagnostic manner seen in ornithurines, forming a propulsion joint? Or do these elements articulate in a more dinosaurian manner? The pterygoids in Archaeopteryx articulate with the quadrate via a narrow, twisted, quadrate process, characterized by a reduced ectopterygoid process that fails to contact the quadrate at all. Combined with the procumbent orientation of the quadrate, such morphology results in not a trace of the propulsion joint claimed for the Urvogel by adherents of the so-called "thecodont hypothesis."

With such a preponderance of evidence refuting the presence of a propulsion joint and quadrate modified for streptostyly, addressing whether or not the dorsal temporal and postorbital bars were absent as claimed by Feduccia et al., seems almost redundant. However, as Feduccia and his colleagues have persisted in inaccurately claiming that both bars were missing in the Urvogel, as they are in advanced birds, the matter does need to be discussed.

Historically, there has been ambiguity about these features of the suspensorium. In all the Archaeopteryx specimens, as is the case with most archaic birds as a whole, the temporal region of the skull was damaged postmortem. As a result, few authors have restored the postorbital and its articulation with the squamosal and/or jugal in any detail, preferring instead to leave this matter as the great question mark of archaeopterygiform skeletal anatomy. However, as Paul summarizes, research in the past decade has greatly clarified the articulation and structure of the postorbital in the Urvogel. In turn, this has more-or-less solved the mystery of whether a dorsal temporal bar and postorbital bar was present.

Postorbital fragments are preserved in the JM 2257 specimen and more notably, in the HMN 1880 specimen. Caudal to the orbit in the famed Berlin specimen is a tri-radiate element, which corresponds in position, orientation and structure to a postorbital and it has been interpreted as such by several independent reviews of the Archaeopteryx material. The presence of a similarly positioned element, which can be unambiguously identified as the frontal process of the postorbital in JM 2257 bolsters this interpretation.

Together these postorbital remains indicate that this bone in the Urvogel was well developed, with a thin, but distinct caudal process articulating with the quadrate, and at least an incipient jugal process (contra Buhler 1985, Feduccia 1996, 1999, Martin & Zhou 1997). The squamosal in BSP 1999 has a distinct joint on the rostral process of the bone, which articulates with the caudal process of the postorbital, as preserved in HMN 1880. These data are unambiguous: Archaeopteryx at the very least had a dorsal temporal bar.

Two Confuciusornis sanctus fossils found in one slab

Phylogenetic bracketing is also of use in settling the matter, although some authors (e.g. Paul 1988, 2002) disagree. However, dorsal temporal bars are preserved in par-archaeopterygiform taxa such as the confuciusornithids (indeed, in Confuciusornis sanctus the bar is quite robust) and Cathayornis yandica (possibly synonymous with Sinornis santensis). Combined with the anatomical data at hand, such bracketing corroborates the presence of a dorsal temporal bar in Archaeopteryx.

The situation regarding the presence of a postorbital bar, or lack thereof, is not as clear. The available postorbitals in Archaeopteryx give only a limited picture. In HMN 1880 the extent of the jugal process of the postorbital is unclear, although at least part of the bone is directed ventrally. In JM 2257, there is no jugal or otherwise ventral portion of the postorbital left whatsoever. None of the other Urvogel fossils preserve the postorbital. The jugal itself in Archaeopteryx is equally difficult to interpret.

Phylogenetic bracketing is perhaps the most useful tool, to deciphering the status of the postorbital bar in Archaeopteryx. It is known from the confuciusornithids and some Enantiornithes, that a postorbital bar was retained in various basal Avialae. Protopteryx fengningensis and Confuciusornis sanctus are the most convincing of these comparative taxa, the former displaying a gracile postorbital bar and the latter a robust bar as previously mentioned.

It is possible that retention of a postorbital bar in such archaic birds is indicative of the presence of such a bar in the Urvogel. Considering the varying degrees of reduction of the bar seen in non-avialian Maniraptora (e.g., Bambiraptor feinbergei) there may have been limited contact between the postorbital and jugal in Archaeopteryx. Given the status of the bar as complete and sometimes robust in basal birds, it as possible that in Archaeopteryx, the postorbital bar did not approach the extreme reduction bordering on loss seen in some dromaeosaurs and other birds.

Review of purported apomorphies in the Urvogel skull can conclude with the claim by Feduccia (1996), following the work of Buhler, that the nasals in Archaeopteryx formed a mesethmoid dorsal to the antorbital fenestra, and rostral to the lacrimal, the final component of the alleged prokinetic system displayed in the Urvogel. However, JM 2257 and HMN 1880 show a strong ossification of the lacrimal and maxilla into a lacrimomaxillary bar, which combined with the lack of any evidence for highly kinetic articulations in the mid-skull roof in Archaeopteryx make it most unlikely that there was such a nasal flexion zone. Considering the lack of any of the accessory traits of a streptostylic system, the isolated presence of such a flexion zone would be illogical.

We can round out the discussion of the cranial morphology in Archaeopteryx by critically examining other contentious aspects of the Buhler & Feduccia skull reconstructions. Most notable among these, are inaccuracies in the position, size, and orientation of the fenestrae rostral to the orbit. Buhler & Feduccia restore the antorbital fenestra in Archaeopteryx as nearly rectangular, oriented parasagittally, whereas in reality the antorbital fenestra is triangular, and parallel to the dorsal rim of the maxilla. Buhler and subsequently Feduccia also fail to restore the well-developed antorbital fossa seen in the articulated skulls of JM 2257 and HMN 1880.

Rostral to the antorbital fenestra, Buhler restores two subellipitical, parasagittally oriented accessory fenestrae whose orientation is questionable at best. That at least one accessory fenestra is located within the antorbital fossa is well established (contra Heilmann 1926), but it is elongate, oblate, and not oriented perpendicular to the dorsal rim of the maxilla, as Buhler and Feduccia claim. The second accessory fenestra is nearer the external naris than the Buhler restoration portrays. Both are most likely external openings leading to accessory sinuses.

Considering the lack of the avian cranial apomorphies, which various authors have attributed to Archaeopteryx, it is logical to conclude that in reality, the skull of the oldest known bird was far more dinosaurian than recent attempts to demonstrate otherwise would indicate.

Feduccia et al. also claim apomorphic conditions for the post-cranial skeleton in Archaeopteryx, but in his overviews in 1996 and 1999, Feduccia concentrated on five character states specifically. Foremost among these is a purported degree of pubic retroversion in the Urvogel approaching that seen in ornithurine birds, and the implication (outright assertion in Tarsitano et al 2000) that Deinonychosauria lacked significant opisthopubic development. Following the work of Larry Martin (1991), Feduccia claims authoritatively a hypopubic cup on the caudodistal portion of the pubes. Additionally, Feduccia claims for Archaeopteryx a fully reversed hallux (180° relative to the cranially oriented digits), and Feduccia presents the orientation and articulation of the pectoral girdle and glenoid in the Urvogel, as derived comparative to homologues in Theropoda. Lastly, following Yalden (1985), Feduccia claims that the overlap of the manal digits (III and IV) seen in HMN 1880, JM 2257 and S5 (the Maxburg specimen) is indicative of the position and orientation of these digits in life.

The degree to which the pelvic girdle was opisthopubic in Archaeopteryx is not as easy to resolve as one would think, given the preservation quality of the available specimens. One fact is clear, the nearly 140° to 150° retroversion of the pubes as restored by Larry Martin et al. and supported by Feduccia, is observed only in BMNH 37001, where the pubis is no longer articulated with the ischium and is disassociated from the iliac blade. Postmortem rotation of the pubis is almost certain considering the far lesser degree of retroversion seen in the remaining Archaeopteryx material. Combined, these data suggest that the degree of retroversion seen in BMNH 37001 is a preservation artifact.

Pubic bones

A more substantiated figure for the degree of pubic retroversion is 110° at the least, to 140° at most, though the data suggests that in life, the degree of retroversion was significantly less than 140°. Feduccia and Tarsitano's assertion that similar degrees of retroversion were absent in Coelurosauria is demonstrably wrong, and in Deinonychosauria the degree of retroversion well exceeds 110°.

The claim by Martin & Feduccia that a hypopubic cup is present in the Urvogel is particularly difficult to account for, as no such structure is preserved in any of the Archaeopteryx materials. An ostensible hypopubic cup in BMNH 37001 is likely a preservation artifact, and the lack of any similar trait in other remains, argues strongly for the absence of such a caudodistal expansion of the pubes.

Shoulder

Feduccia presents the orientation of the glenoid and the structure of the pectoral girdle in Archaeopteryx as apomorphic comparative to dromeosaurs, whereas in fact, these theropods display avian style pectoral girdles including a horizontal, tapering scapula, cranially directed coracoids, with coracoids retroverted and sharply flexed relative to the scapular blades. The glenoid is directed laterally in both basal dromeosaurs and derived dromeosaurs, to a degree whereby the humeri could describe an arc of movement permitting an incipient flapping motion.

It is doubtful however, that the humerus could be elevated fully vertically, given a lack of sufficient dorsal orientation of the glenoid. Arguments presented by Feduccia et al that this is a significant limitation to any incipient flapping motions in coelurosaurs overlook the fact that this same limitation is applicable to the Urvogel itself.

Considering the additional presence of ossified and cartilaginous sternal plates, furculae, ossified uncinate processes, and sternal ribs, it is inexplicable why Feduccia presents the pectoral girdles of Deinonychosauria (and troodonts) as plesiomorphic relative to those of Aves, when indeed they are very similar.

Hallux (first digit) orientation

The true orientation of the hallux in Archaeopterygiformes is equally contentious, though one would think from the fine preservation of the Urvogel fossils it should not be. Contrary to Feduccia's assertions that dinosaur-origin proponents seek to restore the hallux in Archaeopteryx as non-reversed, it is made quite explicit in the most recent treatments (e.g., Paul 2002) that the hallux was well reversed, but not to the degree seen in modern birds, as Feduccia et al have claimed. Postmortem processes have arguably distorted the position of the hallux relative to the other pedal digits, and it is likely that the hallux was in life reversed less than 120°.

Unfortunately, phylogenetic bracketing is of limited use in determining the orientation of the hallux in the Urvogel. In Protarchaeopteryx robusta, arguably closest to Archaeopteryx, the halluxes have been repositioned postmortem, and are of little to no use in establishing the life-position of the first digit.

A concomitant claim of Feduccia et al. concerning a reversed hallux in Archaeopteryx is that reversion of the hallux is not observed in Coelurosauria. This is nearly presented as a truism by Feduccia, and at the very least Feduccia et al. fail to mention that there is a great degree of ambiguity concerning the orientation of the hallux in both Eumaniraptora, and more basal Maniraptoriformes. For instance, the alleged cranial orientation of the hallux in Compsognathus longipes may in fact represent postmortem distortion. More pertinent, the hallux in Sinornithosaurus millenii is disarticulated, and rotated along its lengthwise axis, making restoration of its life-position impossible at this time. Considering these data and recent fossil discoveries from the Yixian Beds, the matter of hallux orientation in coelurosaurs remains far too ambiguous to categorically declare, in the manner of Feduccia et al., that no dinosaur displayed a reversed hallux.

Other digits

In his treatment of Archaeopteryx, Feduccia asserts that manal digits III and IV, preserved crossed in three of the Urvogel remains, assumed this position in life. However, Feduccia overlooks the fact that the proximal digital elements have been disarticulated, and subjected to postmortem rotation. Paul (2002) summarizes Kemp & Unwin (1997) and the more likely process whereby the causal agent for disarticulation and subsequent overlap of the digits is found in the attachment of the remiges to digit III. Considering the available evidence, Feduccia presents no convincing argument that the unusual overlap was the life-condition in Archaeopteryx. Overall then, reviewing the osteology of Archaeopteryx fails to reveal an apomorphic character state comparative to Maniraptora, and arguments to the contrary are not borne out by the fossils at hand.

1,2,3 vs. 2,3,4 — the great digit debate, and the dynamism of evolution

A postulated relationship of digits among theropods.

The digital homology of theropods and birds has been a subject of the highest contention since before Archaeopteryx lithographica was unearthed from its calcium carbonate tomb. As early as 1836, Sir Richard Owen argued that the digits in the avian manus were II, III and IV. Owen's view remained unchallenged until the work of William K. Parker who in 1888 presented research arguing that the digits were in fact I, II and III. The matter has been debated since, often pitting paleontology against developmental biology.

Following the work of Ostrom et al., the identity of the avian digits as I, II and III has been widely accepted. However, the research of Tarsitano & Hecht (1980) and more so J. R. Hinchcliffe (1985) on embryology in birds, laid the foundation for contemporary studies on digital identity in birds, carried out by Feduccia & Burke (1997), Feduccia (1999), Feduccia & Nowicki (2002) and most recently Feduccia (2003), which suggest that the digits in birds are indeed II, III, and IV. This has been hailed as a lack of homology, which is fatal to a theropod origin for birds (most vociferously by Feduccia 1999, 2003).

However, the situation is just not that clear-cut. Wagner & Gauthier (1999) have proposed a homeotic frame shift mutation in which developmental processes acted on by evolutionary constraints, led to an identity shift such that digits I, II, and III equal digits II, III and IV. The frame shift hypothesis has gained a wide degree of support in the paleontological community, and is bolstered by possible examples of homeotic frame shifts in birds, as well as documented digital variation in various metazoans. Nevertheless, the lack of a clear-cut evolutionary impetus for a frame shift of the sort proposed by Wagner & Gauthier undermines this hypothesis (although it by no means invalidates it).

Considering this, alternative suggestions put forward by Galis et al (2003) that assuming digital identities I, II and III for coelurosaurs, may be fallacious, are worth further consideration. Galis et al. postulate that the pattern of postaxial reduction yielding a I, II, III hand in Staurikosaurus pricei and Herrerasaurus ischigualastensis is not applicable to Theropoda, and concomitantly, that herrerasaurs are not representative of the ancestral character state for Theropoda, but rather represent a plesiomorphic outgroup to Saurischia. If this is the case, acquisition of a symmetrical pattern of digital reduction yielding an II, III, IV hand in theropods is plausible.

Feduccia (2003) and Larrsson & Wagner (2003) have categorically rejected this postulate as unsubstantiated, and the latter authors firmly assert that Herrerasauria is a neotheropod outgroup. If this is the case then Galis et al.'s hypothesis is unlikely. However, do the fossils support so definitive a classification of these basal dinosaurs?

Contrary to the contemporary consensus stemming from the work of Paul Sereno (1997), Sereno et al. (1993) and Philip Currie (1997), the phylogenetic status of Staurikosaurus and Herrerasaurus remains uncertain. While Sereno et al. have accepted Herrerasauria as a monophyletic outgroup of Neotheropoda, representative of the ancestral character state most closely allied to that seen in neotheropods, this interpretation is untenable. "Herrerasauria" is delimited on tenuous anatomical grounds, and the similarities between this clade and Theropoda as a whole, are primarily symplesiomorphic, and of little to no phylogenetic relevance. Indeed Staurikosaurus and Herrerasaurus display plesiomorphic attributes comparative to Theropoda including a) a robust medial wall of the acetabulum, producing a semiperforate condition for the hip-socket, b) persistent development of metatarsals I and V, c) absence of epipophyses in the cervicals and d) elongation of cervicals three through four.

Interestingly enough, it is apparent that some Ceratosauria displayed a more robust metacarpal IV than that seen in "Herrerasauria", and given the difficulties in assuming digital recapitulation, said condition argues a lack of direct-ancestry for [Neotheropoda from "Herrerasauria". When combined with other data, it becomes likely that "Herrerasauria" constitutes an outgroup of "Saurischia", predating the splitting of this clade into "Sauropodomorpha" and "Theropoda". Needless to say, the pattern of postaxial digital reduction evidenced in staurikosaurs and herrerasaurs, can arguably be presented as reflective of an independent evolutionary trajectory from that pursued by Theropoda, characterized by the reduction of digits IV and V to yield a tridactyl manus, non-homologous to that seen in Theropoda. Considering the ambiguities of placing "Herrerasauria" as a neotheropod outgroup, and the evidence to suggest that component taxa thereof are not representative of the ancestral osteology of Theropoda, Galis et al.'s postulate is more than plausible.

Taken together, these hypotheses and the degree of similarity between avialians and coelurosaurs reduce the matter of digital identity so vociferously presented by Feduccia et al. to a non-issue, which, while illustrating a previously unknown component of avian evolution, is in no way fatal to theropod origin for birds. Rather, it is indicative of a highly plastic and dynamic evolutionary process in the primary radiation of Ornithodira, and suggestive of the flexibility of the dinosaur morphology in adapting to the conditions presented in the terminal Triassic — a trait no doubt crucial to the rise of these most spectacular of archosaurs.

The ever-ready alibi: convergence

Few ideas are as dogmatically and tenaciously held to by proponents of the "thecodont" hypothesis, as the assertion that any morphological similarities between dinosaurs and birds are ascribable to convergent evolution. As more and more data emerges from the fossil record of Dinosauria and Aves, convergence remains the ever-ready alibi of "thecodont" ancestry, ceaselessly pressed into service to categorically dismiss any data substantiating theropod origin for birds. In this fashion, synapomorphies have been portrayed as characters of dubious homology, and independent acquisition among taxa, which are only superficial confreres, has been advanced as inerrant truth. This sort of narrow-minded, automated rejection of data has become routine among the "thecodont" camp, in tactics which Paul (2002) and Dalton (2000) rightly compare to those employed by creationists.

This linchpin, upon which the arguments of Alan Feduccia and his colleagues so precariously rest, was once at the forefront of paleontological thought. It was as close as scientific postulates can be to canon, following Heilmann's work. And yet in the past three decades the case for convergence has been so seriously called into question that it can no longer credibly be advanced as a disproof of theropod ancestry. Flaws notwithstanding, the cherished myth that derived coelurosaurs only converged with birds in the later Cretaceous is still championed by Feduccia and company, and they argue for it with much voracity and little veracity. It has proven as injurious a concept as the "scales gone frayed" version of feather origins, and has outlived its welcome to the same extent, as has the Bohr model of the atom.

Before further considering the nature of these problems, a matter of semantics must be clarified. When discussing convergence in the case of avian origins and theropod evolution, we are in fact addressing two distinct though related processes of evolution: convergence itself, and parallelism.

Convergence is usually seen as synonymous with homoplasy, and for all intents purposes the terms can be considered equivalent without issue (though the Dean of the modern synthesis]] — the incomparable Ernst Mayr — might disagree, see Mayr 2001). The stance of Mayr aside, see the eloquent G. G. Simpson for a concise generalization of what is meant by the terms convergence and homoplasy, which incidentally shows that they can indeed be thought of as synonymous: acquisition of similar characters in two or more lineages which are not the result of common ancestry relevant to those characters.

Parallelism, or parallelophyly, on the other hand, results in characters independently acquired in lineages which are derived from a common ancestor, following different adaptive trajectories, as a result of the phenotypic potential of that shared ancestry. Thus defined, parallelism is a more robust process, which can account for a wider range of similar yet non-homologous characters, than can mere convergence.

Differences aside, it is clear that the two processes are similar; they graft superficial similarities onto the phenotype of organisms not particularly related. It is the very superficiality of the characters that indicates either the lack of common ancestry, or the independent evolutionary trajectories of the taxa in question. And it is for that reason that these characters lack phylogenetic relevance. As it were, convergence and parallelism are generalists, a pattern demonstrated time and again by the exemplar cases of these processes.

If these two forms of evolution yield general similarities, superficial in nature, then it can be logically concluded that there is a more or less inversely proportional relationship between the frequency and intricacy of characters displaying possible homology, and the likelihood of convergence or parallelism being the causal factors for those traits. In other words, it becomes statistically improbable that either convergence of parallelism can explain observed morphological similarities as the frequency and intricacy of those similarities increases. Probability dictates such a constraint on the process of convergent or parallel evolution, and the fossil record confirms it nowhere as unambiguously as in the case of theropod and avialian morphology.

As the homology of the characters advanced as synapomorphies uniting Eumaniraptora as a holophyletic clade remain valid despite attempts to demonstrate non-homology, proponents of a "thecodont" ancestry are themselves constrained to explain the presence of synapomorphies in Dinosauria and Aves in a context which does not on grounds of statistical implausibility, rule out some level of homoplasy or parallelism. Cataloguing the characters that must be accounted for in this fashion, it becomes clear that the case for convergence constitutes nothing less than special pleading.

Rarely has the fossil record presented a character list linking two taxa as formidable as that which can be generated for coelurosaurs and birds; the total number of traits runs into the hundreds. Convergence or parallelism, by the reasoning of Feduccia et al, must account for synapomorphies ranging from skeletal anatomy to physiology, behavior, and even so obscure a trait as eggshell microstructure. Yet this proposes a level of convergent evolution wholly without precedent in the vertebrate fossil record, so outlandish and so smothering in its entirety that it defies the statistical restraints that exist on the very process of homoplastic development. When one has a far simpler, and far more substantiated explanation at hand — that such massive convergence did not occur in the first place — the logical and parsimonious conclusion must favor the hypothesis, which has data to which it can turn for substantiation.

Perhaps in recognition of this tremendous difficulty with the convergence argument, Feduccia has presented in The Origin and Evolution of Birds, three cases of what he considers similar examples of "massive convergenc": hesperornithiform morphology comparative to gaviiform morphology, falconiform morphology comparative to strigiform morphology, and the relationship of Hupehsuchus to mosasaurs. Yet, the comparison is entirely specious. Feduccia presents these examples as cases of convergence to a degree forbidden by standard interpretations of the process of homoplastic evolution. On the contrary, all three are stereotypical of superficial similarities resulting from well-defined convergent processes, without need of outlandish scenarios for an explanation. Moreover, in none of the cases Feduccia isolates, can one find an adequate analogy for the sheer volume of synapomorphies in dinosaurs and birds, which must be explained away by convergence.

Probalistic woes aside, convergence is fatally undermined by the lack of any clear evolutionary impetus for such a tremendous degree of phenotypic mimicry, and this flaw is rightly seen as insurmountable. The fossils at hand beg the question: why should two lineages, derived from separate ancestors or at the very least following radically different evolutionary trajectories, be so stunningly similar to each other? None of the stimuli, which promulgate convergent evolution, are at large in the case of dinosaur and avian phylogeny. Aside from common ancestry, how is the prospective opponent of theropod origin to explain why, for example, Deinonychosauria and Aves should both possess diamond-shaped dorsal supraoccipitals? How does convergence account for a strongly twisted paroccipital process seen in these clades, or an infratemporal fenestra reduced to a foramen (or eliminated altogether)? What of auxiliary processes of the caudodorsal surface of the ischia, or maxillary sinuses with accessory maxillary fenestrae?

How does convergence plausibly account for the shared presence of a dorsal fossa]] of the ectopterygoid process of the pterygoid in Deinonychosauria and Aves? What was the reason for these two distantly related groups to both acquire such an obscure feature? What homoplastic mechanism accounts for virtually indistinguishable eggshell microstructure in birds and theropods, or the shared presence of feathery integument?

And the list goes on, ad infinitum. Indeed it is such a lengthy number of questions of this sort, which could be generated and posed to Feduccia and his colleagues, that it is impossible to present them all here. Refer to Paul (2002) who summarizes them quite thoroughly in his latest work. And yet to all these questions, begging of answers, Feduccia et al. remain silent. After all, birds are not dinosaurs and therefore the data must be made to fit their distorted phylogeny. While this sort of phylogenetic denial may be loudly voiced in some circles, it is not an example of "good" science, and its adherents jeopardize their own credibility by continuing to subscribe to so weak a postulate.

In the final analysis similarity so vast and so thorough in nature as that observed in the derived coelurosaurs and avialians precludes either convergence or parallelism in the search for a causal factor. The most logical and substantiated conclusion is that common ancestry is the agent at work. If one considers the astonishing degree to which the most avian of dinosaurs and the most archaic of birds (including the Urvogel itself) approach each other in their comparative morphologies, the rationality and indeed elegance of this standpoint is further underscored. No other potential avian ancestors are so similar to birds as Maniraptora; so similar that the remains of the two groups are often confused with each other.

The case for convergence is closed.

The phylogenetic sangreal—Thecodonts as "proaves"

Ultimately all attempts at phylogenetic reconstruction for birds must come to the identification of the stock ancestral thereto. It is this pivotal matter—finding an avian ancestor—around which the rise and fall of phylogenies has revolved. The "thecodont hypothesis", as advocated by Feduccia et al., is saddled with the need to isolate a paravian, yet generalized archosaur, as a more convincing "proavis"; than anything seen within Coelurosauria. Few things require a greater degree of special pleading in vertebrate paleontology—it is the ultimate phylogenetic Sangreal.

Inextricably linked with this attempt to find the "thecodont proavis" is the need to trivialize those synapomorphies linking Coelurosauria and Aves, by either dismissing them outright, or questioning the homology of the elements at hand. Feduccia has taken this to the level of criticizing the entire phylogenetic map of Theropoda, in an effort to display that said taxon is no less a hodgepodge of vaguely related forms, than "Thecodontia". In The Origin and Evolution of Birds this trend reaches its delusional zenith in calling into the question the holophyly of Dinosauria as a whole, via implicit assertion that there is a great deal of ambiguity as to what defines "dinosaur". These attempts to deconstruct the phylogeny of Dinosauria are wholly lacking in substantiating data.

"Thecodont proaves" of varying sort have been advanced, since Heilmann's seminal work. Heilmann considered Broom's marvelous Euparkeria capensis and the derived archosaur Ornithosuchus longidens, described by Huxley in 1877 from Upper Triassic beds of Carnian age, in Scotland, to be specimens closest to the "proavis]]". Since 1926, it has become apparent that these taxa are too basal or too specialized along their own evolutionary trajectories to credibly be considered as bird ancestors, and as a result few researchers continue to present them as hypothetical "proaves" (Welman being an exception, as he in 1995 argued for a close affinity between Euparkeria and Archaeopteryx).

Scenario-based models for the origin of avian flight, integral to "thecodont" ancestry and its rejection of phylogeny-based models to the contrary, make the concomitant claim that the only suitable ancestral type to birds, would be a highly arboreal, gracile "thecodont". This by default excludes still comparatively large and cursorial taxa such as Euparkeria, which at most was only faculatively bipedal for short periods of time, and Ornithosuchus.

Following this redefinition of the criteria for a "thecodont" proavis, the term "avimorph" has been applied to describe the character state of the specimens sought by Feduccia et al. The earliest specimen hailed as an "avimorph" was Cosesaurus aviceps, a prolacertiform archosauromorph from the Lower Triassic of Spain, described by Ellenberger & de Villalta (1974) and Ellenberger (1977). The holotype and singular specimen of Cosesaurus is a badly preserved skeletal imprint, which according to Ellenberger's 1977 treatment of the material, displayed feathery integument. Ellenberger presented Cosesaurus as a paravian par excellence and as its archosaurian affinities remained unclear following its initial description; Ellenberger hypothesized lepidosaurian ancestry for birds. Claimed synapomorphies uniting Cosesaurus and birds included feathers, a furcula, an inflated braincase, a triangular, abbreviated skull, and an opisthopubic pelvis.

Subsequent reexaminations of the holotype by Sanz & Lopez-Martinez (1984) called into question all of these synapomorphies. Purported feathers in Cosesaurus are vague preservation artifacts, which display no convincing homologies to any aerofoil. The ambitious restoration of the braincase as greatly inflated, as per the avian condition, is not substantiated by the poorly preserved and plesiomorphic caudal region of the skull. There is most likely no furcula in Cosesaurus although poor preservation of the pectoral region does not rule out well-developed coracoids at least. The opisthopuby of the pelvis is unsubstantiated, and the pelvic girdle is notably basal comparative to Aves. Lastly, the triangular, abbreviated skull, superficially birdlike, is hardly a convincing phylogenetic character. Such skulls, are found throughout Diapsida as a whole, and are of absolutely no use in delimiting ancestral stock for birds.

A more convincing case has been made for the avian affinities of Megalancosaurus preonensis, a drepanosaur from the Upper Triassic of Italy, which has been advanced as the paradigmatic "proavis" by proponents of "thecodont" ancestry. However, there are significant difficulties in this assessment of Megalancosaurus, not the least of which is the fact that there is not a single synapomorphy uniquely linking said taxon to Aves, and morphologic similarities between Megalancosaurus and birds are largely symplesiomorphic. If "Protoavis texensis" is considered a drepanosaur, sensu Paul (2002), then a case could be made for the presence of 15 synapomorphies between drepanosaurs and birds. However, given the dubious nature and dismal preservation of Chatterjee's Triassic "bird", any such conclusion as made by Paul, remains speculative.

Despite the paucity of synapomorphies to link Megalancosaurus to birds, Feduccia et al have claimed multiple derived traits indicative of avian affinity. Included in this list is a rhamphotheca, incipient or otherwise, a furcula, a semilunate carpal element, a horizontal scapula, and heterocoelus cervical vertebrae displaying paravian elongation, and lastly, the token triangular, abbreviated skull. None of these claims has been substantiated by further review of the specimen (Renesto 1994).

There is no evidence whatsoever of a vascular region underlying what would have been in life, a rhamphotheca, and the origin of said claim is unfathomable. The purported furcula is not existent. A horizontal orientation of the scapula, in the avian fashion, is supported by a singular specimen, which in turn is the most poorly preserved of the Megalancosaurus remains, and restoring the scapula in an avian manner, is unsupported by the remaining skeletons.

Presenting the vertebrae, and in particular the cervical and thoracic series as avian, overlooks the apomorphic nature of these elements, comparative to birds, most particularly in the caudodistal elongation of the thoracic vertebrae and the development of additional supraneural elements. These characters are not seen in even the most basal avian— Archaeopteryx, nor are they observed in any other birds, extinct or otherwise. The heterocoelus condition of the vertebrae, claimed by Feduccia et al., is incipient and of doubtful homology to the true heterocoelus character seen in birds.

Lastly, the claim that a semilunate carpal element was present in Megalancosaurus remains without substantiation. In Renesto's review of the megalancosaur material, no evidence for such an element could be found, and it is questionable how the claim originated in the first place. In all regards, the carpus (and more so the manus) of Megalancosaurus is plesiomorphic comparative to birds.

The most accurate conclusion, which can be drawn as to the paleobiology of this peculiar arboreal archosaur, is that it was a highly specialized taxon, which while displaying curious traits in its own right, is not a suitable ancestral type for Aves.

The last and most eccentric specimen that has recently been advanced as an "avimorph" is the tiny Longisquama insignis, from the Lower Triassic of Krygzystan, initially described by A. G. Sharov in 1970. The holotype and single specimen of Longisquama consists of a skull and scattered postcrania, in an overall poor state of preservation. Contrary to Feduccia (1996), the entire skeleton is not preserved. It is more than likely that Longisquama would have been relegated to historical obscurity in a curator's drawer, were it not for the preservation of integumentary structures which apparently functioned as airfoils, and have been referred to as "protofeathers" by Sharov (1970), Feduccia (1996, 1999) and Jones et al (2000, 2001). The claim led to considerable excitement among proponents of "thecodont" ancestry and subsequent examination led to a panoply of avian apomorphies claimed for Longisquama including a robust furcula, and yet again, the ubiquitous triangular, abbreviated skull.

The first and most contentious claim—the presence of protofeathers—is the most dubious, and has been subject to the most detailed analysis thus far. The integumentary structures are not actually articulated with the spinal column, and this has led to their association with the skeleton questioned by multiple authors. The recovery of isolated examples of these structures has prompted some authorities to argue that in fact, they represent vegetation, which in one case at least, was associated with a skeleton. While this argument is rather speculative, it underscores the very non-avian nature of these structures in Longisquama. The dorsal orientation of the structures, their conspicuous difference from feathers, proto or otherwise, and what may have been a superficial dermal attachment in life, has led most authors to restore the structures as a dorsal frill, or a gliding apparatus convergent on avian flight mechanisms.

The claim that a robust furcula is present in Longisquama is equally open to criticism, and the damage to the pectoral region furthers ambiguity. The evidence from high-resolution imaging of the holotype indicates that the "furcula" is actually two separate clavicles, which have been compressed and distorted postmortem, at best representing a homoplastic trait in an otherwise pseudo-avian. Nevertheless, Feduccia has claimed that Longisquama possesses a true furcula, and advances Longisquama as a veritable beau idol of avian ancestry. It is curious to note that Feduccia accepts the far more ambiguous data in this case, rather than accept the presence of far more clear-cut furculae in Maniraptora.

Longisquama is of further interest in that its archosaur affinities are themselves, open to contention. The only synapomorphy linking the taxon to Archosauromorpha is an antorbital fenestra, but due to poor preservation of the cranial region, the actual presence of this quintessential archosaur trait is questionable. Considering the arboreal specializations and superficially avian gliding apparatus, it is conceivable that Longisquama is close to the Lepidosauromorpha, which further damages attempts to present Longisquama as a "proavis".

It is perhaps recognition of the significant difficulties in presenting any of the species just discussed as a "thecodont proavis" that has led Feduccia et al to perambulations in dinosaur phylogeny. After all, it is a principal criticism of the "thecodont" hypothesis that the very term "thecodont" has no taxonomic meaning, and that systematically "thecodont" is not definable.

Feduccia presents two lines of reasoning as to why this is a non-issue. The first is the argument that basal forms should not display a suite of advanced characters by which they can be united into a homogenous assemblage, as per the necessities of accurate phylogenetic reconstruction. Inherent in Feduccia's work is the assertion that the early archosaurs are too generalized to permit any sort of detailed classification. However, the miscellany of archosaurs gathered into "Thecodontia" is in fact sufficiently apomorphic relative to each other, that it is possible to delimit clades therein. Even the most basal of "thecodonts", represented by the proterosuchids and erythrosuchids can be differentiated from ornithosuchids, for example. Remaining "thecodont" forms include the vast assemblage of Crurotarsi and outgroups thereto (namely Proterochampsidae), and the Ornithosuchia, from which Ornithodira and dinosaurs would eventually arise.

Considering these data, the argument that "Thecodontia" is valid, ad hoc or otherwise because no more concise classification can be formulated, is entirely incorrect. "Thecodontia" is invalid because its members are so distinct from each other, and cannot be readily grouped into a single massive assemblage. To do so would be to misrepresent the phylogeny of these marvelous reptiles.

Feduccia, sidestepping this issue, proceeds to assert that dinosaur systematics are no less ambiguous and strewn with taxonomic litter as it were, than those of "Thecodontia" His principal argument is that Maniraptora is a clade of the most tenuous nature, advanced by overzealous dinosaur paleontologists seeking to solidify the theropod ancestry of birds. Feduccia dismisses the synapomorphies presented in support of a holophyletic Maniraptora as trivial, and of dubious validity. This argument is seriously flawed, and the data suggesting holophyly of Maniraptora are formidable, including synapomorphic characters ranging throughout the cranial and postcranial skeleton. Some of these characters include (after Gauthier 1986, summarized in Dodson et al 1990):

1. Reduction or loss of the prefrontal
2. Prominent axial epipophyses
3. Specialized hypapophyses on ventral surfaces of cervical and thoracic vertebrae
4. Sub-rectangular coracoids
5. Elongate forelimb comparative to Tetanurae
6. Ulna bowed caudally
7. Semilunate carpal block
8. Slender metacarpal III bowed laterally
9. Low pubic process of the ilium
10. Reduced ischium
11. Lesser trochanter confluent with the head of the femur
12. Fourth trochanter absent
13. Pedal digit IV longer than digit II

Feduccia suggests that these features are plesiomorphic in Archosauria as a whole and thus their presence in Maniraptora is merely indicative of symplesiomorphies in some theropods, negating the validity of Maniraptora. However, there are no "thecodonts" which approach the maniraptoran condition.

Ornithosuchus, regardless of its size and general "dinosaur-like" appearance, is in fact quite unlike dinosaurs, and is clearly not classifiable as "maniraptoran" as Feduccia asserts. Feduccia's argument that the pelvic girdle could be called maniraptoran overlooks notable plesiomorphies comparative to Maniraptora, including an acetabulum in which the perforate condition was merely incipient, and claimed opisthopuby for the pelvis in Ornithosuchus is insignificant. A host of other traits in Ornithosuchus clearly marks this animal as a specialized archosaur, which is not close to Dinosauria (indeed, Ornithosuchidae belongs in Crurotarsi, indicative of an entirely different evolutionary trajectory than that followed by the ornithodiran line). Feduccia fails to illustrate any other example of a "thecodont" that displays allegedly invalid maniraptoran synapomorphies, and the reason is simple: there are none.

Feduccia carries his completely speculative deconstruction of Maniraptora to Dinosauria as a whole, implying that the clade is polyphyletic. Holophyly of Dinosauria is supported by multiple independent reviews, and is based on the presence of multiple autapomorphies distinguishing Dinosauria from other archosaurs. The argument that Dinosauria is polyphyletic is so discredited that I shall not review it any great detail here, and merely refer the reader to the appropriate references and even Sankar Chatterjee (1997), a longtime proponent of dinosaur polyphyly, have all defended the holophyly of Dinosauria.

Ultimately then, Feduccia's special pleading for the systematic validity of "Thecodontia" and his wholly unsubstantiated efforts to introduce lethal ambiguity into dinosaur phylogeny, are tacit recognitions of the very non-avian nature of those "thecodonts" advanced as the ancestral stock of birds. The Sangreal remains as elusive as ever.

A riddle wrapped in an enigma: the tempo of avian evolution

Aves is one of the most remarkable success story evolution has ever produced. Since the primordial bird took flight on some Solnhofen island, 145 million years ago, the conquest of the sky has been underway. Since their derivation, birds have become lords of the air, exquisite in form and function, representative of the final and most profound adaptive radiation of Dinosauria. Today, with more than double the number of bird species than mammal species, we are still very much living in an "age of dinosaurs".

In the past three decades a series of marvelous fossils have vastly improved our knowledge of the earliest and most crucial evolutionary phases in the history of Aves: the radiation of archaic birds in the Cretaceous, the origin of ornithurines, and the increasingly compelling case for the Cretaceous origin of at least some neornithine lineages. These data suggest that to hitherto unappreciated levels, the Cretaceous was a period in which the avian lineage burgeoned, and rapidly exploited the myriad niches made available through the mastery of flight.

Much of this renaissance in avian paleontology owes to the discovery and description from Lower Cretaceous shales, of a heretofore-unknown avifauna, the aptly named Enantiornithes. Ranging from Barremian deposits to the latest Maastrichtian, and known from North and South America, the Iberian Peninsula, and the Asian mainland, it is increasingly evident that this clade of archaic birds was an evolutionary powerhouse, arguably the dominant component of the terrestrial aviary throughout the Cretaceous.

Of equal importance has been the recovery and description of fossils representing a group of basal pygostylians predating the derivation of Ornithurae, from lacustrine Barremian deposits in China—the Confuciusornithidae. The spectacular fossils of Confuciusornis sanctus and Changchengornis hengdaoziensis have thus far provided the majority of the data about this group. The confuciusornids, while displaying some apomorphic traits—such as the first true rhamphotheca—are nevertheless markedly plesiomorphic comparative to Enantiornithes and even the stem-neornithines. For instance, all known confuciusornids display a manus with grasping, functional digits, seen elsewhere only in Archaeopterygiformes. These data suggest that Confuciusornithidae represented an early offshoot of the as yet unresolved base of the pygostylian node.

Recent work has also shed light on long-known taxa, such as Hesperornithiformes and Ichthyornithiformes, first described from the chalk beds of Kansas by Marsh in 1880. In the past several decades fossils attributable to these orders have shown that Hesperornithiformes and Ichthyornithiformes were among the most numerous and widespread of marine birds Drawing on this work, it has been postulated that ornithurines were largely restricted to marine and or near-shore environments. However, new discoveries of stem-ornithurines cast doubt on this assertion.

The discovery of Gansus yumenensis in 1981 and its subsequent description as a basal ornithurine by Hou & Liu in 1984 at first bolstered the assumption that ornithurines were primarily excluded from terrestrial environs, which were no doubt dominated by the spectacular Enantiornithes. However, Ambiortus dementjevi described in 1982 and 1985 by Evgeny Kurochkin is clearly a derived ornithurine displaying robust flight capacity via a modern pectoral arrangement as well as further refinements for flight including the fusion of the carpal and metacarpal elements into a carpometacarpus. The holotype of Ambiortus was recovered from Upper Cretaceous lacustrine shales and presents evidence that ornithurines did spread into some terrestrial environments.

Yet another taxon supporting this view is the marvelous Chaoyangia beishanensis, recovered and described in 1993 by Hou Lian-Hai and Zhang Jiangyong, which while displaying a mosaic of plesiomorphic traits is arguably a basal ornithurine, and likely the most generalized ornithurine yet described.

And yet despite the data yielded by these fossil discoveries, there remain significant gaps in our understanding of the early phylogenetic history of Aves. Of particular contention is both the time at which the stem-neornithine lineages were derived, and which lineages were in fact the font from which later neornithines were descended. Few matters beside avian origins are as controversial, and conflicting theories have been advanced to model the origins and early adaptive radiation of Neornithes. However, two basic variations on the theme exist: that the neornithines trace their origins to the Cretaceous, in either their entirety, or in some select groups, or, that the neornithines appeared post-Cretaceous, presumably during the Palaeocene-Eocene.

Alan Feduccia, in 1995, laid the groundwork for the theory that there was a rapid secondary adaptive radiation of Aves following the terminal Cretaceous extinctions, such that by the early Eocene, most if not all of the modern neornithine lineages were present. Feduccia speculates that a single ornithurine lineage emerged from the KT-Boundary, while the rest of Aves suffered catastrophic extinction. From this lonely survivor, the Neornithes arose, much like the proverbial phoenix from the ashes. Feduccia has likened this postulate to "bird evolution's big bang". In The Origin and Evolution of Birds, Feduccia cites that distribution of neornithine fossils as the primary data substantiating this postulate. Feduccia's work has since gained acceptance among some ornithologists, including Sankar Chatterjee, Storrs Olson, of the Smithsonian Instition and Daniel Janzen (1995), of the University of Pennsylvania. Feduccia's hypothesis is intriguing, and may well be correct, yet is it as unequivocally substantiated as its proponents have claimed?

To assess this matter, one must first address the principal assertions at work in the "phoenix hypothesis", that a) the terminal Cretaceous extinction was catastrophic, and b) that besides Charadriiformes—putative ancestral stock for the subsequent Tertiary neornithine radiation—no other Neornithes are known from the Cretaceous. As Feduccia explicitly stipulates these conditions as data substantiating his hypothesis, showing either incorrect or unfounded, concomitantly weakens that hypothesis.

While nothing is particularly unambiguous about the fateful events of the terminal Cretaceous, the patterns of extinction and survival across the KT-Boundary are well documented and subject to critical review. Feduccia maintains that the extinctions 65 million years ago were catastrophic, and that by all logic Aves should have been severely impacted. The geological and paleontological data for this standpoint is wanting, however. While the KT events have become synonymous in recent years to a "cosmic zap" which extirpated en masse a huge percentage of taxa, there is significant evidence for "stepwise" extinction events, and a highly selective pattern of survival across the KT-Boundary.

The work of J. David Archibald and Charles Officer is particularly germane to the discussion, and Archibald's extensive catalogue of taxon survival across the KT-Boundary utterly failed to corroborate arguments that the terminal Cretaceous witnessed mass-extinction events of multiple lineages. While it is clear that an impact event occurred in the terminal Cretaceous, the view that the KT extinction was anything other than a fairly gradual, "stepwise" series of extinctions, displaying high selectivity is not tenable. Regrettably, zeal for invoking a massive die-off as the only causal factor sufficiently capable of extirpating Dinosauria, has replaced more substantiated science, which refutes this assertion.

There is no data to suggest a pattern of mass-extinction among birds in the latest Maastrichtian, which would reduce Aves to the point that but a single lineage would emerge from the KT-Boundary unscathed. Indeed, the avian fossil record arguably supports a "stepwise", selective extinction event, as demonstrated by the gradual extinction of Hesperornithiformes and Ichthyornithiformes suggested by the fossil record of these groups. The diverse Enantiornithes were almost certainly the avian lineage that bore the brunt of the KT events, whatever those may have been, and their position as the dominant though by no means solitary component of the terrestrial aviary (as summarized earlier) makes this hypothesis only logical. The causal factor in the extinction of the Enantiornithes remains unclear, though their physiology and ontogenetic emphasis on highly precocial flight architecture at the expense of concomitant morphological development may have rendered the group vulnerable to climatic and environmental factors at large during the KT transition.

It is thus evident that there is no basis for Feduccia's contention that a catastrophic terminal Cretaceous die-off would have caused a bottleneck in avian evolution, whereby the neornithines only appeared in the Tertiary. However, what of the fossil record itself? Feduccia contends that with the exception of his putative charadriiform lineage, no neornithines were present in the Cretaceous. (Feduccia 1995, 1996, 1999). On the contrary, robust molecular analyses vigorously refute this assertion, and argue for the origin of most if not all neornithine lineages in the Lower to Upper Cretaceous.

It is however not as unambiguous as Feduccia maintains. The assertion that the charadriiform assemblage was the solitary Cretaceous representative of the future neornithine dynasty is incompatible with the fossil evidence available, which despite often-fragmentary nature is sufficient for phylogenetic placement. Feduccia on the other hand has categorically rejected a great deal of important material, which he feels to be wanting of merit. Feduccia's inclinations notwithstanding, at least two neornithine lineages are represented in the Cretaceous, namely the charadriiform assemblage and stem-Procellariiformes. The procellariiform]] material from the Upper Cretaceous, reviewed by Olson in 1985, consists mainly of fragmentary material usually attributed to the pectoral girdle, including some furculae. Feduccia presents no evidence, which would compel one to dismiss this material as insignificant, and does not address the data supporting their interpretation as procellariiform. Indeed, in The Origin and Evolution of Birds Feduccia himself comments that the derived status of procellariiform fossils recovered from the Lower Eocene London Clay, suggests a much earlier derivation of this order. Given the fossils, which bolster a Cretaceous origin for Procellariiformes, it is inexplicable why Feduccia argues instead, that this group appeared only post-KT.

In addition to Charadriiformes and Procellariiformes, it is increasingly likely that the Gaviiformes originated in the Cretaceous, as evidenced by Neogaeornis wetzeli, from Upper Cretaceous sediments of Chile. Initially considered allied to the Podicipediformes by Lambrecht in 1933, subsequent reevaluation of the Neogaeornis material in 1992 by Storrs Olson has made the gaviiform affinities of the specimen clear. Feduccia, in his 1996 review of the origins of Gaviiformes, offered no argument as to why Neogaeornis should not be considered the late Cretaceous gaviiform it is, instead appealing to its stratigraphic occurrence—which is incompatible with his "phoenix hypothesis."

Further tantalizing though as yet equivocal evidence suggesting that Gaviiformes originated in the Cretaceous, is found in the 1989 discovery of fragmentary material from Upper Cretaceous beds on Seymour Island of a putative loon, by Sankar Chatterjee. Chatterjee subsequently described "Polarornis" in 1997, but the material remains—like his more dubious "Protoavis texensis" wanting of rigorous analysis.

The data suggesting a Cretaceous origin for multiple neornithine lineages does not end with the Gaviiformes, however. As Feduccia himself concedes, there is evidence to suggest that the stem-paleognaths extend into the Upper Cretaceous, exemplified no doubt by par-lithornithiform morphology. The molecular research presented in Tuinen, Sibley, & Hedges (2000) bolsters arguments that cursorial, ratite-like progenitors of the paleognathous assemblage were indeed extant in the late Cretaceous.

The derivation of the Anseriformes, an exemplar to Feduccia of post-KT neornithine explosive radiation, arguably occurred sometime in the latter half of the Cretaceous, as evidenced by fossil material of Presbyornis spp., or species similar thereto, which have been recovered from Upper Cretaceous strata of Vega Island, and Mongolia. While there remains a distinct possibility that these materials are not actual specimens of Presbyornis, they are almost assuredly presbyornithids occupying outgroup status to Presbyornis itself. Considering the unresolved systematic affinities of the anseriform lineage, the presence of fossils from the Upper Cretaceous diagnosable as stem-taxa to modern Anseriformes, by default suggests that other related clades were simultaneously present at that time.

In summation, the fossil record offers compelling evidence for the Cretaceous origin of, in the most conservative analysis, five neornithine lineages: Charadriiformes, Procellariiformes, Gaviiformes, stem-paleognaths, and Anseriformes. This is incompatible with the strict hypothesis put forward by Alan Feduccia, in which only the charadriiform assemblage emerged from the Cretaceous extinctions. Molecular phylogeny, largely resting on molecular coalescence dating to determine the time at which a common ancestor for two taxa lived, has suggested that neornithines were in fact far more abundant in the Cretaceous than the fossil record suggests. Cooper & Penny (1997) have gone so far as to suggest that no less than 22 neornithine lineages originated in the Lower Cretaceous.

We thus have two extremes: Feduccia's solitary survivor, and molecular systematist's massive-neornithine survival across the KT-Boundary. The fossil record falls somewhere in the middle, providing evidence that while more neornithines originated in the Cretaceous than suspected by Feduccia, the clade as a whole was not as derived as molecular models would suggest. What is the most accurate conclusion, which can be drawn from the data at hand? By combining the power of molecular coalescence studies, which are robust and supported by multiple independent reviews, with the fossil record, bearing in mind preservation bias and the fact that fossils represent the minimum estimate for the age of a group, a parsimonious and more balanced model for the derivation and early evolution of Neornithes can be pieced together.

It could be argued that such a modest scenario may have been not unlike this: Originating somewhere in the middle Cretaceous, several stem-neornithine lineages subsequently underwent a modest adaptive radiation in the later Cretaceous, exploiting niches not yet dominated by the Enantiornithes and archaic ornithurines. With the decline of archaic ornithurines throughout the Upper Cretaceous, the evolutionary fortunes of the stem-neornithines increased. At the KT Boundary, the dominant terrestrial avifauna, the Enantiornithes, were extirpated, allowing the neornithines—waiting in the wings as it were—to undergo a major adaptive radiation throughout the Paleocene in which the remaining neornithine lineages were derived. This model combines the strengths of Feduccia's arguments—the reality of a rapid adaptive radiation of Neornithes—and the strengths of the molecular data, that the neornithine lineage is rooted in the Cretaceous, to yield a plausible scenario.

In time, the fossil record, which is the final external check on molecular phylogenies, will either corroborate, or refute both the modest scenario herein outlined, and the models to the contrary.

Conclusions

In summary, this narrative has concerned itself with addressing and critically reviewing several key aspects of the argument for "thecodont" ancestry of birds, starting with the osteology of the Urvogel, Archaeopteryx lithographica, advancing to matters of homology and convergence, proceeding through the nature of the so-called "avimorph thecodonts", and concluding with perambulations into the early phylogenetic history of Aves. In the process we have seen that key arguments, which are so frequently advanced as damning to theropod origin for birds are wanting in data and range from the confused, to the outright specious, in character. We have a wealth of data to substantiate the hypothesis that derived coelurosaurs gave rise to Aves somewhere in the latter half of the Jurassic, and that since then, Aves have been at the pinnacle of evolutionary success.

Lacking empirical support, proponents of "thecodont" ancestry have invariably proposed that contrary viewpoints arise from overzealous systematists seeking to promote cladistic methodology as a phylogenetic panacea, which neither addresses the data in support of theropod origins, nor constitutes real science; it is nothing less than conceptual ad hominem. The fossil evidence could not be more compelling in its case, and yet Alan Feduccia and his adherents continue to advance discredited notions of avian origins and evolution. Ultimately, the situation can be diagnosed as such: Alan Feduccia has labeled theropod origin "accommodation of the cladogram". In fact, theropod origin is nothing less than accommodation of reality.

Post-debate evidence

In 2018, Andrea Cau published a paper which used a phylogenetic analysis of Pan-Aves (the clade comprising all animals more closely related to birds than to crocodilians, which includes dinosaurs) to demonstrate that there is nothing resembling sharp delimitation at which avian ancestors make a junp from a fundamentally reptilian state to a fundamentally avian state. Instead, the evolution of bird ancestors into birds was a gradual accumulation of characteristics that took place over the course of 160 million years. There simply isn't any sort of convenient cutoff point between birds and dinosaurs that creationists and BANDits can leap upon to prove birds cannot have evolved from the basal dinosaurian condition. This conclusively proves that birds can be and ARE dinosaurs.

Cau presents an exhuastive list of characteristics and adaptations acquired during this process, and it is notable that Archaeopteryx does not occupy any kind of special position. The supposed "first bird" does not display any of the adaptations which the paper could unambiguously correlate with avian-grade powered flight, such as modifications to the sternum. This falsifies the second component of creationist and BANDit arguments: "Archaeopteryx is fully bird and not a dinosaur". In reality, Archaeopteryx only displays about two-thirds of the characteristics that distinguish birds from the bird-crocodilian common ancestor, while the rest of its traits are unambiguous synamorphies with theropods.

In 2021, a fossilized egg that had been discovered two decades previously in Ganzhou, China was examined more closely. The interior of the egg contained a fossilized full-formed oviraptorosaur, a type of theropod that is ancestral to birds. The oviraptorosaur was near the stage of hatching, and exhibited tucking behavior similar to modern birds that had never been seen before in non-avian dinosaurs.

Sources

• Archibald, J. D. 1996. Dinosaur Extinction and the End of an Era. Columbia University Press, New York.
• Bakker, R. T.]] 1986. The Dinosaur Heresies. New York, William Morrow.
• Cau, Andrea 2018. The assembly of the avian body plan: A 160-million-year long process. Bollettino della Societa Paleontologica Italiana
• Chatterjee, S. 1991. Cranial anatomy and relationships of a new Triassic bird from Texas. Philosophical Transactions of the Royal Society of London B 332: 277-346.
• Currie, P. J. 1995. New Information on the anatomy and relationships of Dromaeosaurus albertensis. Journal of Vertebrate Paleontology 7: 72-81.
• Elzanowski, A. & Wellnhofer, P. 1995.The skull of Archaeopteryx and the origin of birds. Archaeopteryx 13: 41-46.
• Elzanowski, A. & Wellnhofer, P. 1996. Cranial morphology of Archaeopteryx: Evidence from the seventh skeleton. Journal of Vertebrate Paleontology 16: 81-94.
• Elzanowski et al. 2000. The quadrate of an odontognathous bird from the Late Cretaceous Lance Formation of Wyoming. Journal of Vertebrate Paleontology 20: 712-719.
• Hecht et al. 1985. The Beginning of Birds: Proceedings of the International Archaeopteryx Conference, Eichstatt, 1984. Freunde des Jura-Museums Eichstatt, Eichstatt.
• Lofgren, D. F. 1995. The Bug Creek problem and the Cretaceous-Tertiary transition at McGuire Creek, Montana. University of California Publications in Geological Sciences 140: 1-200.
• Martin, L. D. 1991. Mesozoic birds and the origin of birds. In: Origins of the Higher Groups of Tetrapods, 466-477.
• Maryanska, T. & Osmolska, H. 1997. The quadrate of oviraptorid dinosaurs. Acta Paleontologica Polonica 42: 361-371.
• Owen, R. 1836. Aves. In: Todd's Cyclopedia in Anatomy and Physiology, 265-358.
• Parker, W. K. 1888. On the structure and development of the wing in the common fowl. Transactions of the Zoological Society of London 179: 385-395.
• Russell, D. A. & Dong, Z. M. 1993. A nearly complete skeleton of a new troodontid dinosaur from the Early Cretaceous of the Ordos Basin, Inner Mongolia, People's Republic of China. Canadian Journal of Earth Sciences 30: 2107-2127.
• Tarsitano et al. 2000. On the evolution of feathers from an aerodynamic and constructional viewpoint. American Zoologist 40 (4): 676-686.

See also

• Archaeopteryx
• Baraminology
• Geologic timeline