The Protoavis controversy
Chatterjee’s Chimera: A Cold Look at the Protoavis Controversy
“The most remarkable thing about Protoavis is that, although it predates Archaeopteryx by 75 million years, it is considerably more advanced than Archaeopteryx...Protoavis is more closely related to modern birds than is Archaeopteryx.”
-Sankar Chatterjee The Rise of Birds, 1997
“Except for a few elements, the available material of Protoavis is extremely fragmentary. Chatterjee’s interpretations of certain bones are questionable, and even the association of elements into specimens and then into a single taxon seems difficult to support.”
-Luis Chiappe Nature, 1995
“Smushed and mashed and broken.”
-Jacques Gauthier informally describing the holotype quoted in Shipman, 1998
In the science of avian phylogenetics, already rife with acrimony and vitriol, few specimens have produced as savage a debate, as has a peculiar archosaur from Texas. Archosaurs are comparatively abundant in Texas, and have been recovered in some quantity since E. D. Cope worked the redbeds of the panhandle over a century ago. Yet, there have been some extraordinary claims advanced on the behalf of this specimen. Its discoverer, Sankar Chatterjee of Texas Tech University, has assigned the binomial Protoavis texensis ("first bird from Texas") to a small cache of bones, allegedly conspecific. Chatterjee, and a small cadre of paleornithologists (namely Kurochkin and Peters) have claimed that this material documents the Triassic origin of birds, and the presence of a bird more advanced than Archaeopteryx, some 75 million years before the urvogel took to the skies over Solnhofen.
However, there remain critical questions, which as of yet have not been satisfactorily answered. How accurate are the assertions that Chatterjee has made? Do the data on the pattern of avian evolution support a Triassic origin of birds? Is Protoavis as avian as claimed? The data at hand call into question the arguments which Chatterjee has built around his putative Triassic bird, and the case for the avian status of these fossils remains in need of further substantiation.
Putting Protoavis in its Stratigraphic and Paleoclimatic Context
The holotype of Protoavis (TTU P 9200), the paratype (TTU P 9201), and all referred materials, have been recovered from the Dockum Formation, from the panhandle of Texas. The Dockum dates from the Carnian through the early Norian, in the terminal Triassic and is composed of four units of decreasing age: the Santa Rose formation, the Tecovas formation, the Trujillo formation, and lastly the Cooper Canyon formation (Lehman 1994, Chatterjee 1997). The Protoavis material is from the Tecovas level of the Dockum strata (Chatterjee 1991, 1997). The type material was collected from mudstones in June 1973 and initially identified as a juvenile Coelophysis bauri (Chatterjee 1997). The level of the Dockum group from which the Protoavis material was recovered, was most likely deposited in a deltaic river system. The bone bed excavated by Chatterjee and his students, in which Protoavis was discovered, likely reflects an incident of mass mortality following a flash flood (Lehman, Chatterjee, and Schnable 1992, Chatterjee 1997).
The inferred paleoclimate of the Dockum group would have been subtropical and governed by a distinct dry/wet season pattern, with the latter marked by monsoonal rains (Chatterjee 1997). The botanical evidence indicates that the area was densely forested, and the abundance of both invertebrate and vertebrate material from the site suggests that the locale was in general richly populated by a wide variety of species (Ash 1972, Chatterjee 1997). Dinosaurs were still fairly rare in the Dockum group, and only some ceratosaurs and other basal forms are well documented (Chatterjee 1984, 1993, 1997). The principal carnivores of the locale would have been poposaurids such as Postosuchus, a species well represented in the Triassic redbeds of Texas (Chatterjee 1997). Other archaic archosaurs such as rhynchosaurs and aetosaurs were fairly common.
The Taphonomy of the Protoavis Material
Both the holotype and paratype were recovered from disparate locations, both disarticulated and unassociated. Consequently, spatial relationships are impossible to determine (Ostrom 1987, 1996, Feduccia 1999). No record of the original orientation of the material even as recovered, exists. Further material assigned to the taxon has been recovered in isolation with no apparent spatial relationships to each other, and more or less has been referred to Protoavis on a whim. Thus, the presentation of the holotype and paratype as coherent skeletons by Chatterjee (1991, 1997) is fallacious. Such representations are ad hoc conglomerations of bone whose status as conspecific is not apparent from their taphonomy.
The fossils themselves display significant postmortem damage, and are in some cases so badly crushed and distorted at the hand of geological processes, that accurate interpretation thereof is impossible (Paul 1988, Witmer 1991, Chiappe 1995, Currie 1995, Ostrom 1996, Padian & Chiappe 1998, Sanz et al 1998, Feduccia 1999, Paul 2002). Consequently, the lucid analyses offered by Chatterjee (1991, 1995, 1997, 1998, 1999) are in many cases prey more to imagination than factual reality.
In his definitive analysis of the material, The Rise of Birds (1997), Chatterjee failed to illustrate the Protoavis fossils via pictures or sketches of the fossils proper, and instead offers the reader truly gorgeous artistic reconstructions. For this, Chatterjee has been sharply criticized (see especially Feduccia 1999). Indeed, such an approach is entirely intolerable in that it idealizes the material at hand, and obscures the very fragmentary nature of the fossils, and their poor state of preservation.
Osteological Quirks and the Allegedly Derived Nature of Protoavis
The author will primarily review the anatomy of Protoavis in the manner in which Chatterjee has in his 1997 tome, starting with the cranium, and proceeding through the axial and appendicular skeletons. This is done not to imply association of the elements herein discussed, but to be congruent with the most comprehensive defense of their identity and association yet offered.
Chatterjee presents the skull of Protoavis as complete, although only the caudal aspect of the cranium is represented in the available fossils (Ostrom 1996, Feduccia 1999, Paul 2002). Chatterjee argues that the temporal region displays a streptostylic quadrate with orbital process for attachment of the M. protractor pterygoidei et quadrati, with associated confluence of the orbits with the temporal fenestrae, thus facilitating prokinesis. He further asserts ornithurine synapomorphies for the braincase of Protoavis including the structure of the otic capsule, the widespread pneumatization of the braincase elements, a full complement of tympanic recesses, and the presence of an epiotic.
Of this material, only the quadrate and orbital roof, in addition to limited portions of the braincase are preserved with enough fidelity to permit any definitive interpretation (Chiappe 1995, Paul 2002). The quadrates of TTU P 9200 and TTU P 9201 are not particularly alike; a fact not easily explained away if the material is conspecific, as Chatterjee insists (Paul 2002). There does not appear to be an orbital process present on either bone, and the modifications of the proximal condyle permitting wide range of motion against the squamosal, are not readily apparent. Furthermore, the quadratojugal and jugal appear far more robust in the Protoavis specimens themselves, than represented by Chatterjee (1997, 1999). The size and development of the quadratojugal seems to contradict Chatterjee’s assertion that this bone contacted the quadrate via a highly mobile pin joint (Chiappe 1995, Ostrom 1996, Paul 2002). These data render the assertion of prokinesis in the skull of Protoavis questionable at best, and it seems most parsimonious to conclude that the specimen displays a conventional opisthostylic quadrate.
The braincase is where Protoavis comes close to being as avian as Chatterjee has maintained. The otic capsule in allegedly organized in avian fashion, with three distinct foramina arranged as such: fenestra ovalis, fenestra pseudorotunda, and the caudal tympanic recess, with a bony metotic strut positioned between the fenestra pseudorotunda and caudal tympanic recess (Chatterjee 1991, 1997). The claim that the full complement of tympanic recesses seen in ornithurines, are similarly observed in Protoavis is questionable, as the preservation of the braincase is not adequate to permit concrete observations on the matter. Chatterjee omits in his 1997 account of the braincase, the presence of a substantial post-temporal fenestra, which in all Aves (including the urvogel), is reduced or absent altogether (Currie 1995, Paul 2002), and the lack of a pneumatic sinus on the paroccipital (Currie 1995). Furthermore, the braincase possesses multiple characters symplesiomorphic of Coelurosauria, including an expanded cerebellar auricular fossa, and a vagal canal opening into the occiput (Witmer 2002). These data suggest that in the braincase attributed to Protoavis we are in fact seeing the earliest record of a coelurosaurian theropod, as opposed to that of an ornithothoracine bird.
What is preserved of the preorbital skull curiously lacks apomorphic characters to be expected in a specimen, which is allegedly more closely allied to Ornithurae than is Archaeopteryx lithographica. Most telling is the complete absence of accessory fenestrae in the antorbital fossa, leading to maxillary sinuses (Paul 2002).
The post-crania are as badly preserved, or worse, than the cranial elements, and their interpretation by Chatterjee (esp. in Chatterjee 1997) are in many cases unsubstantiated or speculative. Of the postcranial skeleton, Chatterjee has isolated the axial skeleton as displaying a suite of avian characters, including heterocoelus centra, hypapophyses and reduction of the neural spines. The validity of these assertions is questionable. First and foremost, the preservation quality of the vertebrae is abysmal. While the centra are modified, they do not appear to be truly heterocoelus (Paul 2002). The presence of incipient hypapophyses in and of itself might be considered indicative of avian affinity, but their poor development and presence on vertebrae otherwise thoroughly non-avian, is most parsimoniously regarded as mild convergence until further material should be brought to light. The reduction of the neural spines is questionable.
Curiously, Paul (2002) has noted that the cervicals of Protoavis and drepanosaurs are astonishingly similar, such they are hardly distinguishable (Fig. 10.7Ba in Paul 2002) from one another. Considering the modification of the drepanosaur neck for the purposes of snap-action predation, it becomes more likely that superficial similarities in the cervicals of both taxa are in fact only convergent with Aves (Paul 2002). Chatterjee does not identify the remaining vertebrae as particularly avian in their osteology (Chatterjee 1997).
The pectoral girdle is advanced by Chatterjee as highly derived in Protoavis, displaying synapomorphies of Ornithothoraces, including the presence of a hypocleidium-bearing furcula, and a hypertrophied, carinate sternum. Chatterjee’s interpretation of the fossils identified as such in his reviews of the Protoavis material (see esp. Chatterjee 1997) are open to question due to the preservation quality of the elements (Chiappe 1995) and as of this time, it is not clear whether either character was in fact present in Protoavis. The glenoid appears to be oriented dorsolaterally permitting a wide range of humeral movement. Chatterjee implies that this is a highly derived trait which allies Protoavis to Aves (Chatterjee 1997), but why this should be so is not clear. In and of itself, the orientation of the glenoid is not a sufficient basis for placing Protoavis within Aves. The scapular blade is far broader than illustrated by Chatterjee in his 1997 account, and not particularly avian in its gross form (Paul 2002). The coracoid, identified by Chatterjee as strut-like and retroverted, is, like the supposed furcula and sternum, too poorly preserved to permit accurate identification. Moreover, the original spatial relationship of the alleged coracoid to the scapula is entirely unknown (Ostrom 1996, Paul 2002). Uncinate processes and sternal ribs are missing.
Chatterjee claims that the humerus of Protoavis is “remarkably avian” (1997, 53), but as in all matters with the fossils referred to this taxon, accurate identification of the elaborate trochanters, ridges, etc., attributed to the humerus by Chatterjee is impossible at this time. Interestingly, the expanded distal condyles, which appear to be present in the humerus of Protoavis and enlarged deltopectoral crest, are congruent with the morphology of ceratosaur humeri, as is the apparent presence of a distal brachial depression (Gauthier & Rowe 1990).
In his 1997 account, Chatterjee infers the presence of feathers from alleged quill knobs on the badly smashed ulna (though he does caution that this is uncertain). Considering how poorly preserved the ulna is, it is entirely premature to make any definitive conclusions as to the presence of quill knobs until such time as more adequate material becomes available. Upon further examination of the material, Witmer (in Witmer & Chiappe 2002) could not isolate any structures homologous to remigial papillae.
The manus and carpus are among the few areas of the Protoavis material, which are well preserved, and they are astonishingly non-avian. The distal carpals, while long, are in no way similar to those observed in the urvogel or other archaic birds. There is not semilunate element, and the structure of the radiale and ulnare would have limited the flexibility of the wrist joint (Chatterjee 1997, Paul 2002). The manus is not tridactyl, and metacarpal V is present. In even the most basal avialian, Archaeopteryx, there is no vestige of the fifth metacarpal and its presence in Protoavis; seems incongruous with the claim that it is a bird, let alone one more derived than the urvogel.
Chatterjee claims that quill knobs are present on metacarpals III and IV and infers the presence of remiges from such structures. As is the case with the alleged quill knobs on the ulna, the metacarpal structures appear to be attributable to post-mortem damage (Paul 2002). Moreover, the thumb, unlike the case in all birds, is not medially divergent.
Chatterjee asserts that the pelvic girdle is apomorphic comparative to archaic birds and displays a retroverted pubis, fusion of the ischium and ilium, an antitrochanter, and the presence of a renal fossa. The pubis does appear to display opisthopuby, although this has yet to be verified. The alleged fusion of the ischium and ilium into an ilioischiadic plate is currently not substantiated by the fossils at hand, despite Chatterjee’s auspicious illustration to the contrary in The Rise of Birds (Chiappe 1998, Feduccia 1999, Paul 2002). At this time the pelvic girdle is not sufficiently well preserved to ascertain whether or not a renal fossa was present, although as no known avian from the Mesozoic displays a renal fossa, it is not clear why Protoavis should, even if it is ornithothoracine (Witmer in Witmer & Chiappe 2002). Similarly, it is unclear if the alleged antitrochanter has been correctly identified as such.
The femur of Protoavis is astonishingly like that of a ceratosaur. The proximal femur displays a trochanteric shelf caudal to the lesser and greater trochanters, a feature synapomorphic of Ceratosauria (Rowe 1989, Gauthier & Rowe in Dodson et al 1990). Further similarities between the proximal humerus of Protoavis and that of ceratosaurs are found in the shared presence of an enlarged obturator ridge, whose morphology in Protoavis is again, uncannily like that observed in robust ceratosaurs, e.g., Syntarsus (Rowe 1989). The resemblance between the femur of Protoavis and that of a ceratosaur becomes ever more pronounced at the distal end of the bone. Both share a crista tibiofibularis groove, a synapomorphy of Ceratosauria separating the medial and lateral condyles (Rowe 1989, Gauthier & Rowe in Dodson et al 1990, Chatterjee 1997).
The tibia of Protoavis allegedly possesses both a lateral and cranial cnemial crest, though the validity of this claim is subject to question due to the preservation quality of the material. The fibula is continuous to the astragalocalcaneal unit.
A tibiotarsus is absent, unusual considering Chatterjee’s claims for the ornithurine affinity of Protoavis, as is a tarsometatarsus (Dingus & Rowe 1998, Paul 2002). The ascending process of the astragalus is reduced, a character entirely incongruous with a highly derived status for Protoavis. Curiously, such abbreviation of the ascending process is found in ceratosaurs, and in its general osteology, the Protoavis tarsus and pes, is quite similar to those of ceratosaurs. Chatterjee’s restoration of the hallux as reversed is nothing more than speculation, as the original spatial relationships of the pedal elements are impossible to ascertain at this time (Paul 2002).
Griphosaurus or Underdog?
The title beast of this section, Griphosaurus, refers to the derisive genus erected by one J. Andreas Wagner of the University of Munich in 1861, for the holotype of Archaeopteryx lithographica (BMNH 37001), and it reflects his conviction that the urvogel was nothing more than an aberrant reptilian (hence the generic name, which means “mythical lizard”). If Protoavis is a chimera then Griphosaurus is an apt (if invalid under the ICZN) moniker for Chatterjee’s “bird.” Or, are Chatterjee and his small band of supporters correct? Is Protoavis precisely what it is claimed to be: a bird from the Upper Triassic closer to ornithurines than the urvogel?
As poorly preserved as they may be, the bones are the only source to which we can turn in an attempt to answer so crucial a question. The issue has been contentious since day one, and Chatterjee has defended his viewpoint that the materials he and his students have collected from the Dockum beds are representative of not only a single species, but also two specimens. No more, no less. Chatterjee has principally invoked the non-duplication of parts to defend this argument, but there are significant problems with associating the fossils available into a single taxon, and then into coherent individual skeletons (Chiappe 1995, Ostrom 1996, Paul 2002).
Not only were the remains recovered disarticulated and unassociated, there are glaring morphometric differences in the various components of the holotype and paratype. For instance, the scapulae and coracoids are so reduced, that the association with the axial skeleton is extremely difficult to support. Juvenile ontogeny cannot be invoked credibly to explain this discrepancy (Paul 2002). Furthermore, the degree of morphometric variation in the holotype and paratype seems incongruent with the component material representing a conspecific assemblage of bones.
In the final analysis, the most damning evidence to suggest that Protoavis is a chimera, is the astonishing similarities seen in various elements of the cranial and postcranial skeleton, to both drepanosaurs, and theropods (Paul 1988, Gauthier & Rowe in Dodson et al 1990, Welman 1995, Dingus & Rowe 1998, Feduccia 1999, Paul 2002). Welman has argued that the quadrate of Protoavis displays synapomorphies of Theropoda. Paul has demonstrated the drepanosaur affinities of the cervical vertebrae. Gauthier & Rowe, and Dingus & Rowe have argued convincingly for identifying the hind limb of Protoavis as belonging to a ceratosaur. Feduccia has argued that Protoavis represents an arboreal “thecodont” (Feduccia 1996). Everywhere one turns; the very fossils ascribed thereto challenge the validity of Protoavis. The most parsimonious conclusion to be inferred from these data is that Chatterjee’s contentious find is nothing more than a chimera, a morass of long-dead archosaurs.
Pseudo-Revolution: The Phylogenetic Implications of Protoavis
Permit a brief excursion into a phylogenetic hypothetical. If Protoavis is a valid taxon, what are the implications for the two leading hypotheses with which the origin of birds is accounted for? It has been argued that if valid, Protoavis will represent the death knell to the theropod descent of birds (Martin 1988, Tarsitano 1991, Feduccia 1996, 1999, Bock 1997). Dinosaur paleontologists counter that if valid, Protoavis in no way falsifies the theropod origin of birds (Chatterjee 1997, Paul 2002). The very fact that Chatterjee used his putative bird to defend theropod origin seems to contradict the argument of Alan Feduccia that a true bird from the Triassic would bring about the collapse of the theropod“dogma.”
Arguments from authority aside, let us examine the central tenets upon which these arguments rest. It has been claimed that there are no suitably par-avian theropods known from the Upper Triassic, which might be allied to “Protoavis” (Feduccia 1996, 1999, Witmer in Chatterjee 1997) and therefore it is difficult to conceive of a plausible scenario in which birds might have been derived from theropod ancestors.
Yet how valid is this sweeping statement? The osteology of those marvelous early theropods, Ceratosauria, begs to differ. Outside of Maniraptoriformes, there is no other dinosaur clade as avian as Ceratosauria. The acquisition of avian apomorphies within Ceratosauria is astounding, and were it not for an even greater body of data to suggest the maniraptoran affinity of Aves, ceratosaurs would surely be linked with birds (Gauthier & Rowe in Dodson et al 1990). A list of synapomorphies convergent with Ornithurae observed in Ceratosauria includes such quintessential traits as: a) synsacrum present, b) pubis, ilium, ischium fused, c) astragalus and calcaneum fused, and confluent with tibia, d) tarsometatarsus present (Gilmore 1920, Rowe 1989, Gauthier 1984, 1986, Gauthier & Rowe 1990).
The earliest known Ceratosauria date from the Carnian horizon of the Triassic, and are fairly derived, indicating an extensive evolutionary history throughout the Upper Triassic and perhaps into the Lower Triassic. Combined with the osteological data, the temporal distribution of ceratosaurs arguably demonstrates that from the outset, Theropoda displayed distinct avian attributes.
Of greater interest, is that such traits do not appear to be limited to Ceratosauria. The opisthopubic pubis of Herrerasaurus ischigualastensis, either a basal saurischian or basal theropod, is again a convergent trait with Aves present in an otherwise thoroughly plesiomorphic bauplan.
If there are in fact theropods from the Triassic, which display a suite of par-avian characters, the assertion that conferring taxonomic validity to “Protoavis” scuttles theropod origin, is found to be lacking in realism. If “Protoavis” is actually avian, it is plausible that birds share a common ancestry with Ceratosauria, or are nested within Ceratosauria. If Protoavis is a chimera, and only contains avian material within this hodgepodge, the situation is less clear. Such scenarios however remain speculative until such time as more compelling evidence for the avian affinity of any of the Protoavis material is brought forth.
Reaching a Verdict on Protoavis
The osteological and taphonomic data, as well as information on temporal distribution, call into doubt every assertion Chatterjee has advanced on the behalf of his Triassic bird. In particular, Chatterjee’s assigning ornithothoracine status to Protoavis texensis seems to be nothing more than wishful speculation. A brief list of incongruities reflects the paucity of evidence to underwrite Chatterjee’s conclusions that Protoavis is a) a bird and b) more derived than the urvogel:
- The cranium appears to lack accessory fenestrae enclosed within the antorbital fossa, leading to auxiliary maxillary sinuses
- The braincase displays a large post-temporal foramen
- A pneumatopore leading to a pneumatic sinus is not present on the paroccipital process
- The quadrate lacks an orbital process, doubled proximal condyle, and pin-joint with the quadratojugal
- The vertebral centra are only mildly heterocoelus
- Uncinate processes and sternal ribs absent
- Carpus lacks a semilunate element
- Carpometacarpus absent
- Fifth metacarpal retained
- Manus not tridactyl
- Thumb is not medially divergent
- Tibiotarsus absent
- Pygostyle absent, tail elongate
Similarly, the characters which suggest that Protoavis is a chimera and not a valid taxon, include:
- Quadrates similar to those of Theropoda (Welman 1995)
- Cervicals of Protoavis closely resemble those of drepanosaurs
- Significant morphometric discrepancy between the holotype (TTU P 9200) and the paratype (TTU P 9201)
- Material attributed to the taxon was recovered disarticulated and unassociated
These data strongly suggest that Protoavis is both a chimerical taxon, and that the case for a Triassic origin of birds, is at best unsubstantiated. Alleged synapomorphies linking Protoavis to Aves are either of questionable validity, or found in Theropoda in addition to Protoavis, or both (Currie 1995, Paul 2002). Consequently, there are no synapomorphies uniquely allying Protoavis with Aves.
In the course of a succinct if sometimes oversimplified review of the Protoavis controversy, data refuting the claims made by Chatterjee on the behalf of his Triassic bird have been advanced in an effort to demonstrate how unlikely a controversy this matter should be. As there remains no compelling data to support the avian status of Protoavis or taxonomic validity thereof, it seems mystifying that the matter should be so contentious. The author very much agrees with Chiappe in arguing that at present, Protoavis is irrelevant to the phylogenetic reconstruction of Aves. While further material from the Dockum beds may vindicate this peculiar archosaur, for the time being, the case for Protoavis is non-existent.
- Ash, S. R. 1972. Upper Triassic Dockum flora of eastern New Mexico and Texas. In New Mexico Geological Society Guide Book, 124-128.
- Bock, W. 1997. Review of: The Origin and Evolution of Birds, by A. Feduccia. Auk 114: 531-534.
- Chatterjee, S. 1984. A new ornithischian dinosaur from the Triassic of North America. Naturewissenschaften 71: 630-631.
- Chatterjee, S. 1991. Cranial anatomy and relationships of a new Triassic bird from Texas. Philosophical Transactions of the Royal Society of London 332: 277-342.
- Chatterjee, S. 1993. Shuvosaurus, a new theropod. National Geographic Research and Exploration 9: 476-491.
- Chatterjee, S. 1995. The Triassic bird Protoavis. Archaeopteryx 13: 15-31.
- Chatterjee, S. 1997. The Rise of Birds: 225 Million Years of Evolution. Johns Hopkins University Press, Baltimore.
- Chatterjee, S. 1998. The avian status of Protoavis. Archaeopteryx 16: 99-122.
- Chatterjee, S. 1999. Protoavis and the early evolution of birds. Palaeontographica A 254: 1-100.
- Chiappe, L. 1995. The first 85 million years of avian evolution. Nature. 391: 147-152.
- Currie, P. 1995. New information on the anatomy and relationships of Dromaeosaurus albertensis. Journal of Vertebrate Paleontology 15: 576-591.
- Dingus, L. & Rowe, T. 1998. The Mistaken Extinction: Dinosaur Evolution and the Origin of Birds. W. H. Freeman & Company, New York.
- Feduccia, A. 1996. The Origin and Evolution of Birds, First Edition. Yale University Press, New Haven.
- Feduccia, A. 1999. The Origin and Evolution of Birds, Second Edition. Yale University Press, New Haven.
- Gauthier, J. & Rowe, T. 1990. Ceratosauria. In The Dinosauria, Dodson et al (eds.).
- Gilmore, C. 1920. Osteology of the carnivorous Dinosauria in the United States National Museum, with special reference to the genera Antrodemus (Allosaurus) and Ceratosaurus. Bulletin of the U. S. National Museum 110: 1-154.
- Kurochkin, E. N., 1995. Synopsis of Mesozoic birds and early evolution of Class Aves. Archaeopteryx. 13, 47-66.
- Lehman, T. 1994. The saga of the Dockum Group and the case of the Texas/New Mexico boundary fault. Summarized in The Rise of Birds: 225 Million Years of Evolution, Chatterjee 1997.
- Lehman, T., Chatterjee, S. & Schnable, J. 1992. The Cooper Canyon Formation (Late Triassic) of western Texas. Texas Journal of Science 44: 349-355.
- Martin, L. 1988. Review of: The Origin of Birds and the Evolution of Flight, K. Padian (ed). Auk 105(3): 596-597.
- Ostrom, J. 1987. Protoavis, a Triassic bird? Archaeopteryx 5: 113-114.
- Ostrom, J. 1996. The questionable validity of Protoavis. Archaeopteryx 14: 39-42.
- Padian, K. & Chiappe, L. 1998. The origin and evolution of birds. Biological Reviews 73: 1-42.
- Paul, G. S. 1988. Predatory Dinosaurs of the World. New York, Simon & Schuster.
- Paul, G. S. 2002. Dinosaurs of the Air: The Evolution and Loss of Flight in Dinosaurs and Birds. Johns Hopkins University Press, Baltimore.
- Rowe, T. 1989. A new species of the theropod dinosaur Syntarsus from the Early Jurassic Kayenta Formation of Arizona. Journal of Vertebrate Paleontology 9: 125-136.
- Sanz et al. 1998. Living with dinosaurs. Nature 393: 32-33.
- Shipman, P. 1998. Taking Wing: Archaeopteryx and the Evolution of Bird Flight. Simon & Schuster, New York.
- Tarsitano, S. 1991. Archaeopteryx: Quo vadis? In: Schultze, H. P. and Trueb, L. (eds.), Origins of the Higher Groups of Tetrapods, 319-332.
- Welles, S. P. 1984. Dilophosaurus wetherilli (Dinosauria, Theropoda). Osteology and comparisons. Palaeontographica A 185: 85-180.
- Welman, J. 1995. Euparkeria and the origin of birds. South African Journal of Science 91: 533-537.
- Witmer, L. 1991. Perspectives on avian origins. In: Origins of the Higher Groups of Tetrapods, H-P Schultze and L. Trueb (eds.), 427-466.
- Witmer, L. 1997. Introduction. In: The Rise of Birds: 225 Million Years of Evolution, Chatterjee.
- Witmer, L. 2002. The debate on avian ancestry. In: Witmer, L. and Chiappe, L. (eds)., Mesozoic Birds: Above the Heads of Dinosaurs", 3-30.