EvoWiki is now a project of the
We are moving all content to RationalWiki.
See the for details! .
Learning to Fly: How Birds Took to the Air
Learning to Fly: How Birds took to the Air
“When they began to fly, birds either lifted themselves up from the ground or glided down from high places; it is difficult to imagine an alternative.”
-Alan Feduccia, 1996
“Some of these diminutive Dinosaurs were perhaps arboreal in habit, and the difference between them and the birds that lived with them may have been at first mainly one of feathers.”
-O. C. Marsh, 1881
“It is my conviction that Archaeopteryx was still learning to fly—from the ground up—and that avian flight began in a running, leaping, ground-dwelling biped…”
-J. H. Ostrom, 1985
The origin of bird flight has long vexed naturalists and has posed as intractable a problem as deciphering the evolutionary origin of the class Aves itself. Flight is a seemingly magical thing to creatures rooted to the Earth, and dreams of cutting loose our bonds and taking to the skies permeate every aspect of human mythology. Thus there has been no lack of curiosity, from the classical age to the Renaissance and through to the present as to how birds have managed to do what we can only crudely imitate.
Unfortunately, answers have not been as readily forthcoming as our admiration for our feathered friends. When it really comes down to it, flight is a rather complicated business, and though we have a decent understanding of how modern birds fly, it has been extremely difficult to work backwards in time and discover the earliest stages of such flight. Complicating the issue is the amalgamation of biophysical, biological, ecological, and ethological factors, which are cumulatively implicated in the flight process. This has led to a plethora of alleged “solutions” to the problem of avian flight that have in turn just cluttered the literature (e.g., Rietschel’s “hopping proavis”).
While a detailed discussion of the origin of avian flight involves complicated and mildly esoteric kinematics, the goal of a general primer on this critical topic in avian phylogenetics is to review the principal arguments concerning how birds took to the skies. An analysis of closely related topics (e.g., the origin of feathers) can be found elsewhere.
The first model for the origin of flight in flying vertebrates was postulated by Charles Darwin in On the Origin of Species, in which he argued for a gliding precursor to modern bats. As specifically applied to birds, Samuel Wendell Williston formulated the first “cursorial theory” for the origin of flight (1879), though it was fairly vague and not particularly well thought out. Predictably, it was generally rejected. Williston’s colleague and sometimes employer, Othniel C. Marsh proposed a far more comprehensive arboreal theory for the origin of flight in 1880 through a gliding stage, subsequently elaborated on a year later. Marsh’s model heralded the historical narrative-explanations later invoked by Bock (e.g., 1965) by proposing intermediate adaptive steps with extant analogs in this arboreal model, something Williston had neglected.
In 1907 the eccentric, but undeniably brilliant Baron Nopsca expanded Williston’s rudimentary cursorial theory for the origin of flight. Nopsca suggested that avian flight started in a faculatively bipedal cursor, later adopting obligatory bipedalism and employing a rudimentary pelage of elongate scales to “oar” its way along with incipient flight strokes of the arms. Under this model the “proto-wing” would function as a thrust generating propeller to augment that gained by the action of the hind-limbs, eventually creating a Bernoulli Effect over the incipient airfoil and generating sufficient lift to propel the proavis into the air.
Though clearly superior to the original framework elucidated by Williston, Nopsca’s revised cursorial theory was not widely accepted due to the significant aerodynamic difficulties, which plagued this model. Thus a fundamental aspect of the so-called “modernist” consensus on the origin of birds was that avian flight originated in an arboreal context via a gliding intermediate stage. This position is almost entirely the work of a single study in avian phylogeny, arguably the most influential phylogenetic treatise ever authored (and certainly among the most beautiful), The Origin of Birds, by the Danish paleontologist and artist Gerhard Heilmann. Published in 1926 Heilmann argued so convincingly for an arboreal origin of bird flight that this viewpoint became canon in vertebrate paleontology until the latter half of the 20th Century.
Walter Bock of Columbia University reviewed the arboreal theory for the origin of flight, as Heilmann had enumerated it, in a magisterial analysis first presented in 1965 and subsequently elaborated on in 1969 and 1985. Bock concentrated on presenting a detailed, incremental historical narrative explanation for the origin of avian flight in an arboreal context using fully adaptive intermediate stages with known contemporary analogs. This model offered an explicit framework and set of predictions against which data on the earliest birds and proavians could be tested, thus making it by far the most robust model for the origin of avian flight yet offered (Feduccia 1980, 1996, Chatterjee 1997). Bock hypothesized that powered flight was derived through a bipedal cursor taking to the trees, parachuting, then gliding, and then increasing the glide path via flapping. Grafted onto this basic conceptual skeleton were other significant alterations of the basic reptilian bauplan associated with the transition to birds (e.g., the advent of homeothermy and feathers).
Following his description of Deinonychus antirrhopus (1969) and extensive reviews of the morphology and phylogenetic status of Archaeopteryx lithographica (e.g., 1976a, b), John H. Ostrom of the Yale Peabody Museum advocated as a corollary of his resurrected theropod origin of birds, a modified cursorial origin for bird flight. In a series of papers Ostrom postulated that the general similarity between dromaeosaurs—apparently obligate cursors—and Archaeopteryx, the earliest known bird, suggested that bird flight arose in a strictly terrestrial context (1974, 1976b, 1979). This viewpoint immediately gained significant support in paleontological circles, and in an astonishingly short period of time had displaced the intuitively appealing arboreal theory for the origin of flight as the leading solution to this problem (e.g., Padian 1982, Gauthier & Padian 1985).
In 1983, Caple et al. presented the most comprehensive cursorial model for the origin of flight, which remains with little alteration the sine qua non of the “ground-up” derivation of flight in proavians. Further elaborated by Balda et al. (1985) the central point of these pivotal biophysical analyses is that it is allegedly physically impossible for powered flight to arise in a gliding form, as incipient flapping motions during a glide would degrade the glide path as opposed to generating lift. From this conclusion, Caple et al. (1983) and later Balda et al. (1985) asserted that in light of the biophysical complications of an arboreal context for the origin of flight through a gliding intermediary stage, a cursorial origin of flight was the most logical alternative. These authors postulated that a strict cursor using its arms and incipient airfoil for balance and for course-correction during ballistic leaps was an ideal model for a proavian.
Following Ostrom’s lead, the terrestrial and arboreal models for the origin of bird flight have typically been associated with a theropod and “thecodont” ancestry for birds, respectively. Yet this distinction has been one of convention more than fact. Paul (1988) was among the first to synthesize the theropod origin of birds with the arboreal origin of flight, though it was Chatterjee (1997) who presented the first comprehensive amalgamation of these two models into one coherent theoretical framework. Witmer (1997) concurred with this assessment. Recently Paul (2002) has presented a compelling case for the arboreal ecology of the most paravian theropods and lent strong support through this analysis to the arboreal origin of flight. Most significantly, recent fossil discoveries of unambiguously arboreal theropods close to the origin of birds (e.g., Microraptor gui, Epidendrosaurus ninchingensis) have robustly corroborated the synthesis of theropod ancestry and arboreal origin for flight (Zhang et al. 2002, Xu et al. 2003). Proponents of this synthetic theory have made only one substantial alteration to the basic framework of the arboreal model outlined by Bock (1965), in arguing that the avian ancestor was by necessity an obligate biped before shifting to an arboreal habitus.
From the “Ground-Up” or From the “Trees-Down”
The debate between these conflicting scenarios is often characterized as a dichotomy of some sort: theropod ancestry versus “thecodont” ancestry, phylogeny versus historical narrative explanations, and so on. Unfortunately none of these are correct, although the persistent association of a cursorial origin of flight with a theropod ancestry of birds is the legacy of three decades of futile attempts by dinosaur paleontologists to demonstrate that flight evolved form the ground-up in Aves. Long plagued by a paucity of basal bird and proavian archosaur fossils, discoveries in the last 30 years have revolutionized our understanding of how flight arose in birds. In light of this influx of data critical evaluation of the leading hypotheses accounting for the origin of bird flight has increased our ability to discriminate between these hypotheses and isolate the most plausible model.
As it has garnered much-fanfare over the past several decades, the cursorial theory for the origin of bird flight is perhaps that which most readily comes to mind, at least amongst those who are not students of avian phylogenetics. As elaborated by Ostrom (1974, 1976b, 1979) and subsequently revised by later researchers (e.g., Caple et al. 1983, Balda et al. 1985), the cursorial model is a hypothesis generated around one basic observation: bird ancestors were obligate bipeds and cursors. From this, it is concluded that bird flight by necessity arose in a cursorial context. This approach has been broadly characterized as “phylogeny-based,” in that it extrapolates hypotheses on the origin of flight in birds from the phylogenetic history thereof. While this is correct the concomitant assertion that alternative models to the cursorial hypothesis are not grounded in phylogeny, is without substantiation.
Ostrom (1974, 1976b, 1979) concentrated on the osteology of Archaeopteryx lithographica in his assessment of the origin of bird flight. Via comparison with his newly described (1969) Deinonychus antirrhopus, presumed to be an obligate cursor, Ostrom concluded that there was little in the anatomy of the urvogel to support the classical assumption that it was an arboreal animal. Logically, if flight arose in a cursorial context the earliest birds and their closest ancestors will display anatomy indicative of their status as cursors, and thus other researchers followed Ostrom’s basic premise (e.g., Padian & Chiappe 1998).
Other researchers sought to bolster the cursorial model through quantitative analysis of the biomechanics and physics involved in bird flight in an attempt to discriminate on definitive grounds between the competing theories for the origin thereof. Caple et al. (1983) and Balda et al. (1985) drew a number of conclusions from aerodynamic studies based on an Archaeopteryx-like glider. These authors argued that incremental flapping could not increase the distance of the glide path and would prove aerodynamically deleterious, degrading the lift force generated by the airfoil. Furthermore, Balda et al. (1985) posited that attempts to graft incipient flapping onto a stable glide path would require sophisticated muscular coordination, which could not be reasonably expected from primitive birds such as Archaeopteryx. From these data, Caple et al. (1983) and Balda et al. (1985) asserted that a non-adaptive transition existed between a gliding stage and an early powered flight stage, thus weakening a fundamental aspect of the arboreal theory for the origin of bird flight.
These studies went on to suggest that a more plausible proavian would be a small, broadly insectivorous cursor, which would have utilized the forelimbs in an extended attitude for stability. Motion of the arms similar to the flight stroke would augment thrust achieved by the pelvic limb and contribute to lift over the incipient airfoil. Combined with ballistic leaping after insect prey, Caple et al. (1983) argued that this was an ideal context in which to derive powered flight.
In contrast, the arboreal theory for the origin of bird flight offers far greater resolution. Under this model (following Bock 1965, 1985 and Chatterjee 1997), the avian ancestor was a bipedal form, which moved from a cursorial habitus to an arboreal habitus. Arboreal ecology offers a number of distinct advantages to an organism which can successfully exploit this niche, including increased foraging range and latitude in diet, protection from predators, and so on. The invasion of the air via parachuting or gliding almost assuredly arose via a conjunction of two separate stimuli, outlined below.
Part and parcel of living in trees is falling out of them. For instance, Schlesinger et al. (1993) demonstrated in a study of litter-fall in California oak-forest that the ubiquitous fence lizard Sceloporus occidentalis was highly prone to falling from branches for a variety of reasons, frequently associated with prey pursuit and mating displays. Thus there is strong selective pressure to provide some sort of control to an otherwise wild fall, which may result in injury upon striking the ground. Many arboreal animals modify their behavior upon falling from a branch or perch, extending the limbs and flattening the body presenting maximal surface area to incident airflow (Bock 1985, Tarsitano 1985, Feduccia 1996). Examples include various arboreal reptiles (e.g., Draco, Chrysopelea) and a variety of arboreal mammals, which frequently parachute or glide between trees (e.g., Acrobates). Assuming such a posture automatically increases drag and thus slows the animal’s rate of descent, cushioning its impact with the ground.
Arboreal organisms face challenges in foraging for food and other resources which are not met by cursors, which can proceed in a rather linear fashion over terrain to find the resources needed. In an arboreal animal, climbing down a tree, running to another, and then climbing back up it, is energetically expensive and wasteful. Finding a way to proceed from one tree directly to the next is preferable as it maximizes net energy gain during foraging (Norberg 1990). Thus there is powerful selective pressure in arboreal animals to adopt an energy saving mode of locomotion. Just such a strategy is to leap or parachute between trees (Feduccia 1996).
Once this stage has been achieved any modifications of the phenotype which will increase the horizontal component of the animal’s trajectory will be selected for. Behavioral and morphological changes acting in a positive feedback loop will decrease the glide angle, theta, below 45 degrees at which point the animal is effectively gliding. This is perhaps the most crucial stage in the arboreal model for the origin of flight.
Enhancement of gliding and therefore further decreasing the glide angle would be energetically favorable for an arboreal animal as it would effectively mitigate the need to climb out of trees and proceed to the next. Morphological and behavioral modifications of existing traits would be required to eventually decrease theta to zero yielding a horizontal flight path. Such modifications would include the refinement of the pectoral girdle, the expansion of the pectoral musculature, and the elongation and ramification of the integumentary appendages. Control over the glide path in addition to increased lift would be affected by flapping of the wings. At this point the animal is no longer gliding but flying under its own power.
Mirror, Mirror…which model is the fairest of them all
In their skeletal form we have thus enumerated the principal theories for the origin of bird flight (for a more detailed account than this primer can allow I refer the reader to Caple et al. 1983, and Bock 1965). The question remains, which of these models can most accurately accommodate the data at hand and thus, form the most robust predictive model for understanding how birds took to the skies?
Both the cursorial and arboreal theories for the origin of bird flight are historical narrative explanations and thus are not strictly deductive; only an inductive process in which favorable data is advanced in their defense can test them. Falsifying one of the nomological deductive explanations of which the parent historical narrative explanation is composed is usually extremely difficult. Nevertheless various proposals have been advanced as examples in which a basic nomological deductive explanation underlying a historical narrative is rendered false, therefore casting doubt on the entire historical narrative itself.
Central to the arboreal model is the transition from gliding to flapping flight. Caple et al. (1983) and Balda et al. (1985) argued that there is a biophysical barrier preventing the derivation of flapping flight from the gliding stage in that incipient flapping motions would only inhibit lift and therefore decrease the length of the glide path. They postulated a non-adaptive transition zone and thus concluded that the arboreal model for the origin of bird flight was effectively falsified.
There are serious problems with this assertion, however. Norberg (1985) cast doubt onto Caple et al’s (1983) and Balda et al’s (1985) conclusions by demonstrating with multiple lines of evidence that their central argument was in fact faulty. Norberg documented using quasi-stationary aerodynamic models that low velocity flapping movements expected in incipient flight-strokes would nonetheless generate a horizontal thrust component increasing the length of the glide path, while the vertical lift component remains stationary. Crucial to this is the increase of the lift/drag ratio, in other words, the lift vector is increased at the expense of the drag vector thus decreasing the angle, theta prime, between the resultant of the lift and drag vectors. This process is readily accommodated within the framework of the classic arboreal model (Bock 1965). Further increase of the L/D ratio via higher aspect ratio of the wing combined with greater and more precise wing flexion and coordination of movement in the forelimb would produce a vertical lift component and thus lead to standard horizontal flight.
It has further been argued that the arboreal theory for the origin of avian flight is deficient as it is at odds with the inferred ecology of the earliest birds and their presumed theropod ancestors. This argument stems from Ostrom (1974, 1976b, 1979) who argued that both the urvogel and dromaeosaurs were strictly cursorial animals with no appreciable adaptations for an arboreal lifestyle. Though never a convincing argument, this objection can be regarded as effectively refuted.
The evidence indicating an arboreal ecology for both Archaeopteryx and other archaic Cretaceous birds is overwhelming, and has been summarized elsewhere (see Archaeopteryx and Urvogel Take Two: Confuciusornithidae and the Early Evolution of Birds). Furthermore, with new discoveries of basal members of Deinonychosauria (e.g., Microraptor gui) that are unquestionably indicative of arboreal ecology as basal to the clade from which birds are derived, the argument that theropods were strictly cursorial is also highly dubious at this point in time.
What then can be said of the cursorial origin of bird flight? There are unfortunately significant problems with the entire model. The aerodynamic absurdity of a terrestrial origin of bird flight in a swift-running cursor is worth consideration. In such an animal thrust is generated by the hind limbs and thus, as postulated in this scenario, a ballistic leap would immediately remove the source of propulsion and the animal would be quickly returned to the ground by the force of gravity. The elongation of integumentary appendages, such as feather precursors, would impede airflow over the body of the cursor and generate drag, which would merely hinder the forward progress of the animal and degrade its ballistic flight path (Tarsitano 1985, Feduccia 1996).
Recognizing these difficulties, Ostrom (1974, 1976b, 1979) and later Caple et al. (1983) proposed that the hypothesized cursorial proavis would have used incipient flight strokes of the forearm to assist in catching prey during its frenetic leaps against gravity, and from such activity, powered flight would arise. Neither intuitively facile nor biomechanically sound, such a scenario requires first the perfection of derived, slow-speed flight at the very earliest stages in the evolution of birds. It is worth considering the mechanics of such flight to illustrate the absurdity of such an argument. In slow flight the downstroke is used for both thrust and lift and flowing behind the airfoil incident air forms a turbulent wake of chain-ring vortices (Rayner 1985). This vortex ring gait is not a particularly simple means of effecting a flight-stroke and is restricted to slow-flight speeds and birds, which display a low aspect ratio. Mechanically, an alternative flight gait in which the upstroke is loaded (i.e., lift generating) and a continuous vortex is produced in the wake of the airfoil, is far simpler than and energetically preferable to the vortex ring gait.
Lift is sufficient at high enough speeds (intrinsic to a continuous vortex) to mitigate the need for a dorsal elevator, or at least a robust humeral adductor such as the M. supracoracoideus (Rayner 1985). Thus such an embryologically and morphogenically expensive character is not required in the earliest stages of the evolution of bird flight, under this model. In contrast, because dinosaurian cursors could probably not generate the sort of speeds required, a cursorial origin of flight would have to rely on a vortex ring gait preceding a continuous vortex, fast-flight context for the origin of flapping, requiring the presense of some sort of elevation musculature. Solokoff et al. (2001) attempted to demonstrate that in Sturnus the supracoracoideus tendons may be cut and ground-take off possible, but as noted by Olson (2002) such a conclusion applied to the earliest birds is dubious at best. Even if the recovery stroke were affected by the deltoid group, it would be unclear why the derived M. supracoracoideus system, which serves the same function, would have evolved (Rayner 2001). It is worth noting that this postulates a directly adaptive role for the M. supracoracoideus system, when all available evidence indicates that the dorsal migration of the biceps tubercle and formation of the acrocoracoid process automatically deflected the tendon of insertion of M. supracoracoideus. Thus there is no reason to think, given this process, that even with a deltoid group operating as a dorsal elevator the formation of the distinctive supracoracoideus pulley system of birds would have been precluded.
Furthermore, the neurological and muscular coordination required in vortex ring gait flight far exceeds that of a fast-flight continuous vortex gait, contrary to the assertion of Balda et al. (1985), and such features are derived in birds. In the earliest birds, such as Archaeopteryx, the carpal architecture, and inferred cranial anatomy do not lend credence to the suggestion that in any way Archaeopteryx possessed the modifications necessitated by the vortex ring gait flight profile. Though the brain of Archaeopteryx was apparently divided in typically avian fashion between the cerebrum, optic lobes and cerebellum, the cerebral hemispheres are not especially expanded and lack the visual Wulst as in more derived birds. Modifications of the cerebellum including the presence of a prominent floccular lobe are not observed in Archaeopteryx and there is no compelling reason to believe they were present (Chatterjee 1997).
Using encephalization quotient data as an index for intelligence, Archaeopteryx falls within the lower range of the avian regression, plotting out at an EQ value of 0.34 (Jerison 1973, Hopson 1980, Chatterjee 1991). In comparison, birds such as Corvus and Columba display higher EQ values and concomitant modifications of brain anatomy that reflect the modification and enhancement of flight in these forms. Tellingly, these data are all congruent in indicating that Archaeopteryx, the earliest known bird and closest specimen we have to an idealized proavis, did not possess the neurological modifications associated with slow-speed, vortex ring gait flight.
Of further interest is that in biophysical terms, continuous vortex fast-flight is extremely similar to gliding flight, as would be expected in an arboreal proavian (Rayner 1985). Rayner extensively compared the aerodynamic forces and geometrical configuration of the wings in each flight mode and argued convincingly that the latter is readily derived from the former. This meshes nicely with an arboreal context for the origin of bird flight via a gliding intermediary stage.
Perhaps the greatest strength of the arboreal model, however, is that it can produce a fully coherent pseudophylogeny with fully adaptive transitional stages between one ecological context and the next, without invoking saltations, while producing extant analogues for each stage in the pseudophylogeny (Bock 1965, 1985, Feduccia 1980, 1996). Thus, the arboreal model offers greater resolution and a more robust predictive framework for understanding the origin of bird flight.
The prevailing models for the origin of bird flight—cursorial and arboreal—can be examined critically in light of the totality of data available from the fossil record, ecology, and biophysics. The cursorial origin of bird flight has been rigorously scrutinized and has come up deficient. There are serious problems with this postulate and we have a superior model for the derivation of flapping flight in bird ancestors in the arboreal hypothesis.
Balda, R. P., Caple, G. & Willis, W. R. 1985. Comparison of the gliding to flapping sequence with the flapping to gliding sequence. In: Hecht, M. K., Ostrom, J. H., Viohl, G. & Wellnhofer, P. (eds.), The Beginnings of Birds : Proceedings of the International Archaeopteryx Conference Eichstatt 1984, 267-277.
Bock, W. 1965. The role of adaptive mechanisms in the origin of higher levels of organization. Systematic Zoology 14: 272-278.
Bock, W. 1969. The origin and radiation of birds. Annals of the New York Academy of Sciences 167: 147-155.
Bock, W. 1985. The arboreal theory for the origin of birds. In: Hecht, M. K., Ostrom, J. H., Viohl, G. & Wellnhofer, P. (eds.), The Beginnings of Birds : Proceedings of the International Archaeopteryx Conference Eichstatt 1984, 199-207.
Caple, G., Balda, R. P., Willis, W. R. 1983. The physics of leaping animals and the evolution of preflight. American Naturalist 121: 455-476.
Chatterjee, S. 1991. Cranial anatomy and relationships of a new Triassic bird from Texas. Philosophical Transactions of the Royal Society of London, Series B 332: 277-342.
Chatterjee, S. 1997. The Rise of Birds: 225 Million Years of Evolution. Johns Hopkins University Press, Baltimore.
Dial, K. 2004. What use is half a wing and WAIR? - Journal of Vertebrate Paleontology Vol 24, Abstracts of Papers (Supplement to No. 3): 52A.
Feduccia, A. 1980. The Age of Birds. Harvard University Press, Cambridge.
Feduccia, A. 1996. The Origin and Evolution of Birds, First Edition. Yale University Press, New Haven.
Gauthier, J. & Padian, K. 1985. Phylogenetic, functional, and aerodynamic analyses of the origin of birds and their flight. In: Hecht, M. K., Ostrom, J. H., Viohl, G. & Wellnhofer, P. (eds.), The Beginnings of Birds : Proceedings of the International Archaeopteryx Conference Eichstatt 1984, 185-197.
Heilmann, G. 1926. The Origin of Birds. Witherby, London.
Hopson, J. A. 1980. Relative brain size in dinosaurs: implications for dinosaurian endothermy. In: Thomas, R. D. K. & Olson, E. C. (eds.), A Cold Look at the Warm-blooded Dinosaurs, 287-310.
Jerison, H. 1973. Evolution of the Brain and Intelligence. Academic Press, New York.
Marsh, O. C. 1880. Odontornithes: A monograph on the extinct toothed birds of North America. Rep. Geol. Expl. Fortieth Parallel 7: 1-201.
Marsh, O. C. 1881. Discovery of a fossil bird in the Jurassic of Wyoming. American Journal of Science, 3rd ser., 21: 341-342.
Nopsca, F. von. 1907. Ideas on the origin of flight. Proceedings of the Zoological Society of London 1907: 223-236.
Norberg, U. 1985. Evolution of flight in birds: Aerodynamic, mechanical and ecological aspects. In: Hecht, M. K., Ostrom, J. H., Viohl, G. & Wellnhofer, P. (eds.), The Beginnings of Birds : Proceedings of the International Archaeopteryx Conference Eichstatt 1984, 293-302.
Norberg, U. 1990. Vertebrate Flight. Springer-Verlag, Berlin.
Olson, S. L. 2002. Review: New Perspectives on the Origin and Early Evolution of Birds: Proceedings of the International Symposium in Honor of John H. Ostrom. The Auk 119(4): 1202-1205.
Ostrom, J. H. 1969. Osteology of Deinonychus antirrhopus, an unusual theropod from the Lower Cretaceous of Montana. Bulletin of the Peabody Museum of Natural History 30: 1-165.
Ostrom, J. H. 1974. Archaeopteryx and the origin of flight. Quarterly Review of Biology 49: 27-47.
Ostrom, J. H. 1976a. Archaeopteryx and the origin of birds. Biological Journal of the Linnean Society 8: 91-182.
Ostrom, J. H. 1976b. Some hypothetical anatomical stages in the evolution of avian flight. Smithsonian Contributions to Paleobiology 27: 1-21.
Ostrom, J. H. 1979. Bird flight: how did it begin? American Scientist 67: 46-56.
Ostrom, J. H. 1985. The meaning of Archaeopteryx. In: Hecht, M. K., Ostrom, J. H., Viohl, G. & Wellnhofer, P. (eds.), The Beginnings of Birds : Proceedings of the International Archaeopteryx Conference Eichstatt 1984, 161-176.
Padian, K. 1982. Macroevolution and the origin of major adaptations: vertebrate flight as a paradigm for the analysis of patterns. Proceedings of the Third North American Paleontological Convention 2: 387-392.
Padian, K. & Chiappe, L. 1998. The origin of birds and their flight. Scientific American 278(2): 38-47.
Paul, G. 1988. Predatory Dinosaurs of the World. Simon Schuster, New York.
Paul, G. 2002. Dinosaurs of the Air: The Evolution and Loss of Flight in Dinosaurs and Birds. Johns Hopkins University Press, Baltimore.
Rayner, J.M. 1985. Mechanical and ecological constraints on flight evolution. In: Hecht, M. K., Ostrom, J. H., Viohl, G. & Wellnhofer, P. (eds.), The Beginnings of Birds : Proceedings of the International Archaeopteryx Conference Eichstatt 1984, 279-288.
Rayner, J.M. 2001. On the origin and evolution of flapping flight aerodynamics in birds. In: Gauthier, J. & Gall, L. F. (eds.), New Prespectives on the Origin and Early Evolution of Birds: Proceedings of the International Symposium in Honor of John H. Ostrom, 363-381.
Schlesinger, W. H., Knops, J. M. H., & Nash, III, T. H. 1993. Arboreal sprint failure: lizardfall in a California oak woodland. Ecology 74: 2465-2467.
Sokoloff, A. J., Gray-Chickering, J., Harry, J. D. Poore, S. O. & Goslow, G. E. 2001. The function of the supracoracoideus muscle during takeoff in the European starling (Sternus vulgaris): Maxheinz Sy revisited. In: Gauthier, J. & Gall, L. F. (eds.), New Prespectives on the Origin and Early Evolution of Birds: Proceedings of the International Symposium in Honor of John H. Ostrom, 319-332.
Tarsitano, S. 1985. The morphological and aerodynamic constraints on the origin of avian flight. In: Hecht, M. K., Ostrom, J. H., Viohl, G. & Wellnhofer, P. (eds.), The Beginnings of Birds : Proceedings of the International Archaeopteryx Conference Eichstatt 1984, 319-332.
Williston, S. W. 1879. Are birds derived from dinosaurs? Kansas City Review of Science 3: 457-460.
Witmer, L. 1997. Foreword; pp. vii-xii in S. Chatterjee, The Rise of Birds: 225 Million Years of Evolution. Johns Hopkins University Press, Baltimore.
Xing Xu, Zhonghe Zhou, Xiaolin Wang, Xuewen Kuang, Fucheng Zhang, & Xiangke Du. 2003. Four-winged dinosaurs from China. Nature 421: 335-339.
Fucheng Zhang, Zhonghe Zhou, Xing Xu, Xiaolin Wang. 2002. A juvenile coelurosaurian theropod from China indicates arboreal habits. Naturwissenschaften 89: 394-398.