The scientific question of which larger group of animalsbirds evolved within has traditionally been called the "origin of birds". The presentscientific consensus is thatbirds are a group ofmaniraptorantheropoddinosaurs that originated during theMesozoic era.

A close relationship between birds and dinosaurs was first proposed in the nineteenth century after the discovery of the primitive birdArchaeopteryx in Germany. Birds and extinct non-avian dinosaurs share many unique skeletal traits.^[1] Moreover,fossils of more than thirty species of non-avian dinosaur with preserved feathers have been collected. There are even very small dinosaurs, such asMicroraptor andAnchiornis, which have long,vaned arm and leg feathers forming wings. The Jurassic basalavialanPedopenna also shows these long foot feathers.PaleontologistLawrence Witmer concluded in 2009 that this evidence is sufficient to demonstrate that avian evolution went through a four-winged stage.^[2] Fossil evidence also demonstrates that birds and dinosaurs shared features such as hollow,pneumatized bones,gastroliths in thedigestive system,nest-building, andbrooding behaviors.

Although the origin of birds has historically been a contentious topic withinevolutionary biology, only a few scientists still dispute the dinosaurian origin of birds, suggesting descent from other types ofarchosaurianreptiles. Within the consensus that supports dinosaurian ancestry, the exact sequence of evolutionary events that gave rise to the early birds within maniraptoran theropods is disputed. The origin ofbird flight is a separate but related question for which there are alsoseveral proposed answers.

Scientific investigation into the origin of birds began shortly after the 1859 publication ofCharles Darwin'sOn the Origin of Species.^[3] In 1860, a fossilized feather was discovered inGermany'sLate JurassicSolnhofen limestone.Christian Erich Hermann von Meyer described this feather asArchaeopteryx lithographica the next year.^[4]Richard Owen described a nearly complete skeleton in 1863, recognizing it as a bird despite many features reminiscent ofreptiles, including clawed forelimbs and a long, bony tail.^[5]

BiologistThomas Henry Huxley, known as "Darwin's Bulldog" for his tenacious support of the new theory ofevolution by means ofnatural selection, almost immediately seized uponArchaeopteryx as atransitional fossil between birds and reptiles. Starting in 1868, and following earlier suggestions byCarl Gegenbaur,^[6] andEdward Drinker Cope,^[7] Huxley made detailed comparisons ofArchaeopteryx with various prehistoric reptiles and found that it was most similar to dinosaurs likeHypsilophodon andCompsognathus.^[8][9] The discovery in the late 1870s of the iconic "Berlin specimen" ofArchaeopteryx, complete with a set of reptilian teeth, provided further evidence. Like Cope, Huxley proposed an evolutionary relationship between birds and dinosaurs. Although Huxley was opposed by the very influential Owen, his conclusions were accepted by many biologists, including BaronFranz Nopcsa,^[10] while others, notablyHarry Seeley,^[11] argued that the similarities were due toconvergent evolution.

A turning point came in the early twentieth century with the writings ofGerhard Heilmann ofDenmark. An artist by trade, Heilmann had a scholarly interest in birds and from 1913 to 1916, expanding on earlier work byOthenio Abel,^[12] published the results of his research in several parts, dealing with the anatomy,embryology, behavior, paleontology, and evolution of birds.^[13] His work, originally written inDanish asVor Nuvaerende Viden om Fuglenes Afstamning, was compiled, translated into English, and published in 1926 asThe Origin of Birds.

Like Huxley, Heilmann comparedArchaeopteryx and other birds to an exhaustive list of prehistoric reptiles, and also came to the conclusion that theropod dinosaurs likeCompsognathus were the most similar. However, Heilmann noted that birds hadclavicles (collar bones) fused to form a bone called thefurcula ("wishbone"), and while clavicles were known in more primitive reptiles, they had not yet been recognized in dinosaurs. Since he was a firm believer in an interpretation ofDollo's law that stated that evolution was not "reversible", Heilmann could not accept that clavicles were lost in dinosaurs and re-evolved in birds. He was therefore forced to rule out dinosaurs as bird ancestors and ascribe all of their similarities toconvergent evolution. Heilmann stated that bird ancestors would instead be found among the more primitive "thecodont" grade of reptiles.^[14] Heilmann's extremely thorough approach ensured that his book became a classic in the field, and its conclusions on bird origins, as with most other topics, were accepted by nearly all evolutionary biologists for the next four decades.^[15]

Clavicles are relatively delicate bones and therefore in danger of being destroyed or at least damaged beyond recognition. Nevertheless, some fossil theropod clavicles had actually been excavated before Heilmann wrote his book, but these had been misidentified.^[16]The absence of clavicles in dinosaurs became the orthodox view despite the discovery of clavicles in the primitive theropodSegisaurus in 1936.^[17] The next report of clavicles in a dinosaur was in a Russian article in 1983,^[18] whereMongolian paleontologistRinchen Barsbold argued that one or several lineages of theropods may have evolved into birds.^[19]

Contrary to what Heilmann believed, paleontologists now accept that clavicles and, in most cases, furculae are a standard feature not just of theropods but ofsaurischian dinosaurs. Up to late 2007 ossified furculae (i.e. made of bone rather thancartilage) have been found in all types of theropods except the most basal ones,Eoraptor andHerrerasaurus.^[20] The original report of a furcula in the primitive theropodSegisaurus (1936) was confirmed by a re-examination in 2005.^[21] Joined, furcula-like clavicles have also been found inMassospondylus, an Early Jurassicsauropodomorph.^[22]

The tide began to turn against the 'thecodont' hypothesis after the 1964 discovery of a new theropod dinosaur inMontana. In 1969, this dinosaur was described and namedDeinonychus byJohn Ostrom ofYale University.^[23] The next year, Ostrom redescribed a specimen ofPterodactylus in theDutchTeylers Museum as another skeleton ofArchaeopteryx.^[24] The specimen consisted mainly of a single wing and its description made Ostrom aware of the similarities between the wrists ofArchaeopteryx andDeinonychus.^[25]

In 1972,British paleontologistAlick Walker hypothesized that birds arose not from 'thecodonts' but fromcrocodile ancestors likeSphenosuchus.^[26] Ostrom's work with both theropods and early birds led him to respond with a series of publications in the mid-1970s in which he laid out the many similarities between birds and theropod dinosaurs, resurrecting the ideas first put forth by Huxley over a century before.^[27][28][29] Ostrom's recognition of the dinosaurian ancestry of birds, along with other new ideas about dinosaur metabolism,^[30] activity levels, and parental care,^[31] began what is known as thedinosaur renaissance, which began in the 1960s and, according to some, continues to this day.^[32]

Ostrom's revelations also coincided with the increasing adoption of phylogenetic systematics (cladistics), which began in the 1960s with the work ofWilli Hennig.^[33] Cladistics is an exact method of arranging species based strictly on their evolutionary relationships, which are calculated by determining the evolutionary tree implying the least number of changes in their anatomical characteristics. In the 1980s, cladistic methodology was applied to dinosaur phylogeny for the first time byJacques Gauthier and others, showing unequivocally that birds were a derived group of theropod dinosaurs.^[34] Early analyses suggested that dromaeosaurid theropods likeDeinonychus were particularly closely related to birds, a result that has been corroborated many times since.^[35][36]

The early 1990s saw the discovery of spectacularly preserved bird fossils in severalEarly Cretaceousgeological formations in the northeastern Chinese province of Liaoning.^[37][38] In 1996, Chinese paleontologists describedSinosauropteryx as a new genus of bird from theYixian Formation,^[39] but this animal was quickly recognized as a more basal theropod dinosaur closely related toCompsognathus. Surprisingly, its body was covered by long filamentous structures. These were dubbed 'protofeathers' and consideredhomologous with the more advanced feathers of birds,^[40] although some scientists disagree with this assessment.^[41] Chinese andNorth American scientists describedCaudipteryx andProtarchaeopteryx soon after. Based on skeletal features, these animals were non-avian dinosaurs, but their remains bore fully formed feathers closely resembling those of birds.^[42] "Archaeoraptor", described withoutpeer review in a 1999 issue ofNational Geographic,^[43] turned out to be a smuggled forgery,^[44] but authentic remains continue to pour out of the Yixian, both legally and illegally. Feathers or "protofeathers" have been found on a wide variety of theropods in the Yixian.^[45][46] The morphological gap between non-avian theropods and birds is further closed by the discoveries of extremely bird-like non-avian dinosaurs,^[47] as well as non-avian dinosaur-like basal birds.^[48]

There is a debate betweenembryologists andpaleontologists whether the hands oftheropod dinosaurs and birds are essentially different, based onphalangeal counts—a count of the number of phalanges (finger bones) in the hand.

Embryologists and some paleontologists who oppose the bird-dinosaur link have long numbered the digits of birds II-III-IV based on multiple studies of the development in the egg.^[49] This is based on the fact that in mostamniotes, the first digit to form in a 5-fingered hand is digit IV, which develops a primary axis. Therefore, embryologists have identified the primary axis in birds as digit IV, and the surviving digits as II-III-IV. The fossils of advancedtheropod (Tetanurae) hands appear to have the digits I-II-III (some genera withinAvetheropoda also have a reduced digit IV^[50]). If this is true, then the II-III-IV development of digits in birds is an indication against theropod (dinosaur) ancestry. However, with noontogenical (developmental) basis to definitively state which digits are which on a theropod hand (because no non-avian theropods can be observed growing and developing today), the labelling of the theropod hand is not conclusive.^{[citation needed]}

Paleontologists have traditionally identified avian digits as I-II-III. They argue that the digits of birds' number I-II-III, just as those of theropod dinosaurs do, by the conserved phalangeal formula. The phalangeal count for archosaurs is 2-3-4-5-3; many archosaur lineages have a reduced number of digits, but have the samephalangeal formula in the digits that remain. In other words, paleontologists assert that archosaurs of different lineages tend to lose the same digits when digit loss occurs, from the outside to the inside. The three digits ofdromaeosaurs andArchaeopteryx have the same phalangeal formula of I-II-III as digits I-II-III ofbasal archosaurs. Therefore, the lost digits would be V and IV. If this is true, then modern birds would also possess digits I-II-III.^[49] Also, one 1999 publication proposed a frame-shift in the digits of the theropod line leading to birds (thus making digit I into digit II, II to III, and so forth).^[51][52]However, such frame shifts are rare in amniotes and—to be consistent with the theropod origin of birds—would have had to occur solely in the bird-theropod lineage forelimbs and not the hindlimbs (a condition unknown in any animal).^[53]This is calledLateral Digit Reduction (LDR) versusBilateral Digit Reduction (BDR) (see alsoLimusaurus).^[54]

A small minority, known by the acronymBAND (Birds Are Not Dinosaurs),^[55] including ornithologistsAlan Feduccia andLarry Martin,^[56] continues to assert that birds are more closely related to earlier reptiles, such asLongisquama orEuparkeria, than to dinosaurs.^[57][58]Embryological studies of birddevelopmental biology have raised questions about digit homology in bird and dinosaur forelimbs.^[59] However, due to the cogent evidence provided by comparative anatomy and phylogenetics, as well as the dramatic feathered dinosaur fossils from China, the idea that birds arederived dinosaurs, first championed by Huxley and later by Nopcsa and Ostrom, enjoys near-unanimous support among today's paleontologists.^[15]

An alternative to the frame-shift hypothesis is the axis-shift. According to this explanation, the primary limb axis in birds runs through digit III instead of IV.^[49][60] This idea is supported by palaeontological observations, which determine the phalangeal formula 2-3-4-1-X for the last common ancestor of ceratosaurs (includingLimusaurus) and tetanurans (including the tridactyl forms with the phalangeal formula 2-3-4-X-X).^[61]

Some later embryological data support the identification of bird digits as I, II, III (as in their theropod ancestors) due to the posterior finger's development outside theshh-expressing zone of polarizing activity.^[62]

A 2011 publication suggested that selection for the expansion ofskeletal muscle, rather than the evolution of flight, was the driving force for the emergence of this clade. Muscles became larger in prospectivelyendothermicsaurians, according to this hypothesis, as a response to the loss of thevertebratemitochondrialuncoupling protein,UCP1. Inmammals, UCP1 functions withinbrown adipose tissue, which isthermogenic to protect newborns againsthypothermia. In modern birds, skeletal muscle serves a similar function and is presumed to have done so in their ancestors. In this view,bipedality and other avianskeletal alterations were side effects of musclehyperplasia, with further evolutionary modifications of the forelimbs, including adaptations for flight or swimming, andvestigiality, being secondary consequences of two-leggedness.^[63][64][65]

Archaeopteryx has historically been considered the first bird, orUrvogel. Although newer fossil discoveries filled the gap between theropods andArchaeopteryx, as well as the gap betweenArchaeopteryx and modern birds,phylogenetic taxonomists, in keeping with tradition, almost always useArchaeopteryx as a specifier to help define Aves.^[67][68] Aves has more rarely been defined as acrown group consisting only of modern birds.^[34] Nearly all palaeontologists regard birds ascoelurosauriantheropoddinosaurs.^[15] Within Coelurosauria, multiplecladistic analyses have found support for aclade namedManiraptora, consisting oftherizinosauroids,oviraptorosaurs,troodontids,dromaeosaurids, and birds.^[35][36][69] Of these, dromaeosaurids and troodontids are usually united in the cladeDeinonychosauria, which is asister group to birds (together forming the node-cladeEumaniraptora) within the stem-cladeParaves.^[35][70]

Other studies have proposed alternative phylogenies, in which certain groups of dinosaurs usually considered non-avian may have evolved from avian ancestors. For example, a 2002 analysis found that oviraptorosaurs were basal avians.^[71]Alvarezsaurids, known fromAsia and theAmericas, have been variously classified asbasal maniraptorans,^{[35][36][72][73]} paravians,^[69] the sister taxon ofornithomimosaurs,^[74] as well as specialized early birds.^[75][76] The genusRahonavis, originally described as an early bird,^[77] has been identified as a non-avian dromaeosaurid in several studies.^[70][78] Dromaeosaurids and troodontids themselves have also been suggested to lie within Aves rather than just outside it.^[79][80]

Manyanatomical^[81] features are shared by birds and other theropod dinosaurs.

Archaeopteryx, the first good example of a "feathered dinosaur", was discovered in 1861. The first specimen was found in theSolnhofen limestone in southern Germany, which is alagerstätte, a rare and remarkablegeological formation known for its superbly detailed fossils.Archaeopteryx is atransitional fossil, with features clearly intermediate between those of non-avian theropod dinosaurs andbirds. Discovered just two years after Darwin's seminalOrigin of Species, its discovery spurred the nascent debate between proponents ofevolutionary biology andcreationism. This early bird is so dinosaur-like that, without a clear impression of feathers in the surrounding rock, at least onespecimen was mistaken forCompsognathus.^[82]

1. Blank out numbers in image

Vane

Rachis

Barb

Afterfeather

Hollow shaft
(calamus)

Parts of a feather

Since the 1990s, a number of additionalfeathered dinosaurs have been found, providing even stronger evidence of the close relationship between dinosaurs and modern birds. The first of these were initially described as simple filamentousprotofeathers, which were reported in dinosaur lineages as primitive ascompsognathids andtyrannosauroids.^[83] However, feathers indistinguishable from those of modern birds were soon after found in non-avialan dinosaurs as well.^[42]

A small minority of researchers have claimed that the simple filamentous "protofeather" structures are simply the result of the decomposition of collagen fiber under the dinosaurs' skin or in fins along their backs, and that species with unquestionable feathers, such asoviraptorosaurs anddromaeosaurs are not dinosaurs, but true birds unrelated to dinosaurs.^[84] However, a majority of studies have concluded that feathered dinosaurs are in fact dinosaurs, and that the simpler filaments of unquestionable theropods represent simple feathers. Some researchers have demonstrated the presence of color-bearingmelanin in the structures—which would be expected in feathers but not collagen fibers.^[85] Others have demonstrated, using studies of modern bird decomposition, that even advanced feathers appear filamentous when subjected to the crushing forces experienced during fossilization, and that the supposed "protofeathers" may have been more complex than previously thought.^[86] Detailed examination of the "protofeathers" ofSinosauropteryx prima showed that individual feathers consisted of a central quill (rachis) with thinnerbarbs branching off from it, similar to but more primitive in structure than modern bird feathers.^[87]

The 2022 description of branched feathers in the pterosaurTupandactylus provides strong evidence that "pycnofibers" are not actually a distinct integument unrelated to origin of feathers. The most parsimonious scenario is the presence of feathers in the last common ancestor of pterosaurs and dinosaurs already in the Early Triassic.Tupandactylus's melanosomes indicate visual signalling was an important factor in the evolution of feathers.^[88]

Because feathers are often associated with birds, feathered dinosaurs are often touted as the "missing link" between birds and other dinosaurs. However, the multiple skeletal features also shared by the two groups represent the more important proof forpaleontologists.

Comparisons of bird and dinosaur skeletons, as well ascladistic analysis, strengthens the case for the link, particularly for a branch of theropods calledManiraptora. Skeletal similarities include the skull, tooth build, neck, uncinate processes on the ribs, an open hip socket,^[89] a retroverted longpubis, flexiblewrist (semi-lunatecarpal), long arms, three-fingered hand, generalpectoral girdle,shoulder blade,furcula, andbreast bones. Almost all skeletal traits ofArchaeopteryx can be found in non-avian maniraptorans.

A study comparing embryonic, juvenile and adult archosaur skulls concluded that bird skulls are derived from those oftheropod dinosaurs byprogenesis, a type of paedomorphicheterochrony, which resulted in retention of juvenile characteristics of their ancestors.^[90]

Large meat-eating dinosaurs had a complex system of air sacs similar to those found in modern birds, according to an investigation led by Patrick M. O'Connor ofOhio University. In theropod dinosaurs (carnivores that walked on two legs and had birdlike feet) flexible soft tissue air sacs likely pumped air through the stiff lungs, as is the case in birds. "What was once formally considered unique to birds was present in some form in the ancestors of birds", O'Connor said.^[91][92]

Computed tomography (CT) scans conducted in 2000 of the chest cavity of a specimen of theornithopodThescelosaurus found the apparent remnants of a complex four-chambered heart, much like those found in today's mammals and birds.^[93] The idea is controversial within the scientific community, criticised for being bad anatomical science^[94] or simply wishful thinking,^[95] It is also not very surprising ascrocodilians also possess four-chambered hearts.

A study published in 2011 applied multiple lines of inquiry to the question of the object's identity, including more advanced CT scanning,histology,X-ray diffraction,X-ray photoelectron spectroscopy, andscanning electron microscopy. From these methods, the authors found that: the object's internal structure does not include chambers but is made up of three unconnected areas of lower density material, and is not comparable to the structure of anostrich's heart; the "walls" are composed ofsedimentary minerals not known to be produced in biological systems, such as goethite,feldspar minerals,quartz, andgypsum, as well as some plant fragments;carbon,nitrogen, andphosphorus,chemical elements important to life, were lacking in their samples; and cardiac cellular structures were absent. There was one possible patch with animal cellular structures. The authors found their data supported identification as a concretion of sand from the burial environment, not the heart, with the possibility that isolated areas of tissues were preserved.^[96]

The question of how this find reflects metabolic rate and dinosaur internal anatomy is moot, though, regardless of the object's identity.^[96] Both moderncrocodilians andbirds, the closest living relatives of dinosaurs, have four-chambered hearts (albeit modified in crocodilians), so dinosaurs probably had them as well; the structure is not necessarily tied to metabolic rate.^[97]

Fossils of thetroodontsMei andSinornithoides demonstrate that the dinosaurs slept like certain modern birds, with their heads tucked under their arms.^[98] This behavior, which may have helped to keep the head warm, is also characteristic of modern birds.

When laying eggs, female birds grow a special type of bone in their limbs. Thismedullary bone forms as a calcium-rich layer inside the hard outer bone, and is used as a calcium source to make eggshells. The presence of endosteally derived bone tissues lining the interior marrow cavities of portions of aTyrannosaurus rex specimen's hind limb suggested thatT. rex used similar reproductive strategies, and revealed that the specimen is female.^[99] Further research has found medullary bone in the theropodAllosaurus and ornithopodTenontosaurus. Because the line of dinosaurs that includesAllosaurus andTyrannosaurus diverged from the line that led toTenontosaurus very early in the evolution of dinosaurs, this suggests that dinosaurs in general produced medullary tissue.^[100]

SeveralCitipati specimens have been found resting over the eggs in its nest in a position most reminiscent ofbrooding.^[101]

Numerous dinosaur species, for exampleMaiasaura, have been found in herds mixing both very young and adult individuals, suggesting rich interactions between them.

A dinosaur embryo was found without teeth, which suggests some parental care was required to feed the young dinosaur, possibly the adult dinosaur regurgitated food into the young dinosaur's mouth (seealtricial). This behaviour is seen in numerous bird species; parent birds regurgitate food into the hatchling's mouth.

Both birds and dinosaurs usegizzard stones. These stones are swallowed by animals to aid digestion and break down food and hard fibres once they enter the stomach. When found in association withfossils, gizzard stones are calledgastroliths.^[102]Gizzard stones are also found in some fish (mullets,mud shad, and thegillaroo, a type of trout) and in crocodiles.

On several occasions, the extraction of DNA and proteins from Mesozoic dinosaurs fossils has been claimed, allowing for a comparison with birds. Severalproteins have putatively been detected in dinosaur fossils,^[103] includinghemoglobin.^[104] In 2023, beta-protein structures were reported from the feathers of the dinosaurSinornithosaurus and the early birdConfuciusornis.^[105] This confirms that ancient feathers had a composition similar to that of modern birds. Some fossil feathers were reported to have a composition rich in alpha proteins, but fossilization experiments demonstrate that this protein composition is simply an artefact of preservation, because beta-sheet protein structures are readily transformed to alpha-helices during thermal maturation.^[105]

In the March 2005 issue ofScience, Dr.Mary Higby Schweitzer and her team announced the discovery of flexible material resembling actual soft tissue inside a 68-million-year-oldTyrannosaurus rex legbone of specimen MOR 1125 from theHell Creek Formation inMontana. The sevencollagen types obtained from the bone fragments, compared to collagen data from living birds (specifically, achicken), suggest that older theropods and birds are closely related.^[106] The soft tissue allowed a molecular comparison ofcellular anatomy andprotein sequencing ofcollagen tissue published in 2007, both of which indicated thatT. rex and birds are more closely related to each other than either is toAlligator.^[107][108] A second molecular study robustly supported the relationship of birds to dinosaurs, though it did not place birds within Theropoda, as expected. This study utilized eight additional collagen sequences extracted from a femur of the "mummified"Brachylophosaurus canadensis specimen MOR 2598, ahadrosaur.^[109] However, these results have been very controversial. No otherpeptides of a Mesozoic age have been reported. In 2008, it was suggested that the presumed soft tissue was in fact a bacterial microfilm.^[110] In response, it was argued that these very microfilms protected the soft tissue.^[111] Another objection was that the results could have been caused by contamination.^[112] In 2015, under more controlled conditions safeguarding against contamination, the peptides were still identified.^[113] In 2017, a study found that a peptide was present in the bone of the modern ostrich that was identical to that found in theTyrannosaurus andBrachylophosaurus specimens, highlighting the danger of a cross-contamination.^[114]

The successful extraction of ancient DNA from dinosaur fossils has been reported on two separate occasions, but upon further inspection andpeer review, neither of these reports could be confirmed.^[115]

Debates about the origin of bird flight are almost as old as the idea that birds evolved fromdinosaurs, which arose soon after the discovery ofArchaeopteryx in 1862. Two theories have dominated most of the discussion since then: the cursorial ("from the ground up") theory proposes that birds evolved from small, fast predators that ran on the ground; the arboreal ("from the trees down") theory proposes that powered flight evolved from unpowered gliding by arboreal (tree-climbing) animals. A more recent theory, "wing-assisted incline running" (WAIR), is a variant of the cursorial theory and proposes that wings developed theiraerodynamic functions as a result of the need to run quickly up very steep slopes such as trees, which would help small feathered dinosaurs escape from predators.

In March 2018, scientists reported thatArchaeopteryx was likely capable offlight, but in a manner substantially different from that ofmodern birds.^[116][117]

The cursorial theory of the origin of flight was first proposed bySamuel Wendell Williston, and elaborated upon byBaron Nopcsa. This hypothesis proposes that some fast-running animals with long tails used their arms to keep their balance while running. Modern versions of this theory differ in many details from the Williston-Nopcsa version, mainly as a result of discoveries since Nopcsa's time.

Nopcsa theorized that increasing the surface area of the outstretched arms could have helped small cursorial predators keep their balance, and that the scales of the forearms elongated,evolving into feathers. The feathers could also have been used to trap insects or other prey. Progressively, the animals leapt for longer distances, helped by their evolving wings. Nopcsa also proposed three stages in the evolution of flight. First, animals developed passive flight, in which developing wing structures served as a sort ofparachute. Second, they achieved active flight by flapping the wings. He usedArchaeopteryx as an example of this second stage. Finally, birds gained the ability to soar.^[118]

While some authors had rejected the homology between feathers and scales due to their differentproteins,^[119] recent studies provide evidence that those structures do share a common origin.^[120][121] However, Nopcsa's theory assumes that feathers evolved as part of the evolution of flight, and recent discoveries show that feathers evolved millions of years before flight.^[120][88]

Feathers are very common incoelurosaurian dinosaurs (including the earlytyrannosauroidDilong).^[122] Modernbirds are classified as coelurosaurs by nearly all palaeontologists,^[123] though not by a fewornithologists.^{[124][125][126]} The modern version of the "from the ground up" hypothesis argues that birds' ancestors were small, feathered, ground-running predatory dinosaurs (rather likeroadrunners in their hunting style^[127]) that used their forelimbs for balance while pursuing prey, and that the forelimbs and feathers later evolved in ways that provided gliding and then powered flight. The most widely suggested original functions of feathers include thermal insulation and competitive displays, as in modern birds.^[128][129]

All of theArchaeopteryx fossils come from marine sediments, and it has been suggested that wings may have helped the birds run over water in the manner of theJesus Christ Lizard (common basilisk).^[130]

Most recent opposition to the "from the ground up" hypothesis attempt to refute the modern version's assumption that birds are modified coelurosaurian dinosaurs. The criticism is based onembryological analyses that suggest birds' wings are formed from digits 2, 3, and 4, (corresponding to the index, middle, and ring fingers in humans. The first of a bird's three digits forms thealula, which they use to avoidstalling in low-speed flight—for example, when landing). The hands of coelurosaurs, however, are formed by digits 1, 2, and 3 (thumb and first two fingers in humans).^[59] However, these embryological analyses were immediately challenged on the embryological grounds that the "hand" often develops differently inclades that have lost some digits in the course of their evolution, and that birds' "hands" do develop from digits 1, 2, and 3.^{[49][131][132]} For more information about this subject, see "Digit homology".

Fowleret al. (2011) proposed a model explaining how dromaeosaurids may have hunted. The animal would use its wing as stabilizers while standing on top of its prey eating it alive in the manner of an eagle or a hawk. The authors consider this an important addition to the topic of how flapping movements evolved, arguing they likely precede flight.^[133]

Thewing-assisted incline running (WAIR) hypothesis was prompted by observation of youngchukar chicks, and proposes that wings developed theiraerodynamic functions as a result of the need to run quickly up very steep slopes such as tree trunks, for example to escape from predators.^[134] This makes it a specialized type of cursorial ("from the ground up") theory. Note that in this scenario birds needdownforce to give their feet increased grip.^[135][136] But early birds, includingArchaeopteryx, lacked theshoulder mechanism by which modern birds' wings produce swift, powerful upstrokes. Since the downforce WAIR depends on is generated by upstrokes, it seems that early birds were incapable of WAIR.^[137] Because WAIR is a behavioural trait without osteological specializations, the phylogenetic placement of the flight stroke before the divergence of theNeornithes, the group which contains all extant birds, makes it impossible to determine if WAIR is ancestral to the avian flight stroke or derived from it.^[138]

Most versions of the arboreal hypothesis state that the ancestors of birds were very small dinosaurs that lived in trees, springing from branch to branch. This small dinosaur already had feathers, which were co-opted by evolution to produce longer, stiffer forms that were useful in aerodynamics, eventually producing wings. Wings would have then evolved and become increasingly refined as devices to give the leaper more control, to parachute, to glide, and to fly in stepwise fashion. The arboreal hypothesis also notes that, for arboreal animals, aerodynamics are far more energy efficient, since such animals simply fall to achieve minimum gliding speeds.^[139][140]

Several small dinosaurs from the Jurassic or Early Cretaceous, all with feathers, have been interpreted as possibly having arboreal and/or aerodynamic adaptations. These includeScansoriopteryx,Epidexipteryx,Microraptor,Pedopenna, andAnchiornis.Anchiornis is particularly important to this subject, as it lived at the beginning of the Late Jurassic, long beforeArchaeopteryx.^[141]

Analysis of the proportions of the toe bones of the most primitive birdsArchaeopteryx andConfuciusornis, compared to those of living species, suggest that the early species may have lived both on the ground and in trees.^[142]

One study suggested that the earliest birds and their immediate ancestors did not climb trees. This study determined that the amount of toe claw curvature of early birds was more like that seen in modern ground-foraging birds than in perching birds.^[143]

Archaeopteryx was the first and for a long time the only known featheredMesozoic animal. As a result, discussion of the evolution of birds and of bird flight centered onArchaeopteryx at least until the mid-1990s.

There has been debate about whetherArchaeopteryx could really fly. It appears thatArchaeopteryx had the brain structures and inner-ear balance sensors that birds use to control their flight.^[144]Archaeopteryx also had a wing feather arrangement like that of modern birds and similarly asymmetrical flight feathers on its wings and tail. ButArchaeopteryx lacked theshoulder mechanism by which modern birds' wings produce swift, powerful upstrokes (see diagram above of supracoracoideus pulley); this may mean that it and other early birds were incapable of flapping flight and could only glide.^[137]

But the discovery since the early 1990s of manyfeathered dinosaurs means thatArchaeopteryx is no longer the key figure in the evolution of bird flight. Other small feathered coelurosaurs from theCretaceous and LateJurassic show possible precursors of avian flight. These includeRahonavis, a ground-runner with aVelociraptor-like raised sickle claw on the second toe, that some paleontologists assume to have been better adapted for flight thanArchaeopteryx,^[145]Scansoriopteryx, an arboreal dinosaur that may support the "from the trees down" theory,^[146] andMicroraptor, an arboreal dinosaur possibly capable of powered flight but, if so, more like abiplane, as it had well-developed feathers on its legs.^[147] As early as 1915, some scientists argued that the evolution of bird flight may have gone through a four-winged (ortetrapteryx) stage.^[148][149]Hartmanet al. (2019) found that, because of how basal flying paravians are phylogenetically distributed, flight most likely evolved five times among paravians instead of only once.Yi,Archaeopteryx,Rahonavis andMicroraptor were thus considered examples of convergent evolution instead of precursors of bird flight.^[150]

A minority hypothesis, credited to the booksPredatory Dinosaurs of the World (1988) andDinosaurs of the Air (2002) by scientific illustratorGregory Paul, suggests that some groups of non-flying carnivorous dinosaurs — especiallydeinonychosaurs, but perhaps others such asoviraptorosaurs,therizinosaurs,alvarezsaurids andornithomimosaurs — actually descend from birds or other flighted maniraptorans. Paul also proposed that the ancestors of these groups were more advanced in their flight adaptations thanArchaeopteryx. The hypothesis would mean thatArchaeopteryx is less closely related to extant birds than these dinosaurs are.^[151] In 2016, Paul suggested thatomnivoropterygid avialans were closely related to oviraptorosaurs and thatjeholornithid avialans were closely related to therizinosaurs; he considered them to not be avians but suggested that they shared a flighted ancestor.^[152]

Mayret al. (2005) analyzed a new, tenth specimen ofArchaeopteryx, and concluded thatArchaeopteryx was the sister clade to the Deinonychosauria, but that the more advanced birdConfuciusornis was within the Dromaeosauridae.^[80][153] This paper, however, excluded all other birds and thus did not sample their character distributions. The paper was criticized by Corfe and Butler (2006) who found the authors could not support their conclusions statistically. Mayret al. agreed that the statistical support for the authors' earlier paper was weak but stated that it is also weak for the alternative scenarios.^[154]

Most subsequentcladistic analyses do not support Paul's hypothesis about the position ofArchaeopteryx. Instead, they indicate thatArchaeopteryx is closer to birds, within the cladeAvialae, than it is to deinonychosaurs or oviraptorosaurs.Microraptor,Pedopenna, andAnchiornis all have winged feet, share many features, and lie close to the base of the cladeParaves. This suggests that the ancestral paravian may have been a four-winged glider.^[2]Deinonychus may also display partial volancy, with the young being capable of flight or gliding and the adults being flightless.^[155] In 2018, a study concluded that the last common ancestor of thePennaraptora had joint surfaces on the fingers, and between the metatarsus and the wrist, that were optimised to stabilise the hand in flight. This was seen as an indication for secondary flightlessness in heavy basal members of that group.^[156] Hartman et al. (2019) found it likely thatArchaeopteryx was more closely related to dromaeosaurs than to birds, but the distribution of flying taxa among flightless ones suggests multiple origins of flight among maniraptoran theropods instead of only one.^[150]

InEuornithes, the earliest unequivocal example of secondary flightlessness isPatagopteryx.^[157]

Dinosaurs portal

Wikimedia Commons has media related to
Aves fossils.

• 'Dinosaurs Among Us' Retraces an Evolutionary Path from Dinosaurs to Birds, NY Times, March 28, 2016
• DinoBuzz A popular-level discussion of the dinosaur-bird hypothesis
• Archaeopteryx - FAQs from the Usenet newsgroup talk.origins.
• Dinosaurs among usArticle and VideoAmerican Museum of Natural History exhibit of dinosaur evolution leading to birds

• Evolution of birds
• Prehistoric birds
• Dinosaurs
• Origins
• Theropoda

• CS1 errors: ISBN date
• CS1 German-language sources (de)
• CS1 Russian-language sources (ru)
• CS1 maint: multiple names: authors list
• Webarchive template wayback links
• CS1: unfit URL
• Articles needing more detailed references
• Articles with short description
• Short description matches Wikidata
• All articles with unsourced statements
• Articles with unsourced statements from May 2023
• CS1 maint: DOI inactive as of August 2025
• Commons category link is locally defined