University of Michigan Museum of Zoology

Skip to main content


The Evolutionary History of Birds

A resource of the UMMZ Bird Division
by Robert B. Payne

Look at fossil birds, and other fossils that had bird-like features. These are of interest in determining the age of birds, the ancestors of birds, the series of changes that birds underwent in evolution, whether modern birds are related to the radiation of fossil birds in the Mesozoic, and what makes a bird a bird. We shall consider whether feathers make the bird, or whether another definition of birds is more appropriate.

Evolution: Darwin's idea -- evolution proceeds when a new heritable trait emerges in some organism and is passed genetically to its descendants. The idea indicates that two groups of animals that share a set of these new or "derived" traits are more closely related to each other than they are to groups that display only the original traits but not the derived ones. Of course these traits can be lost subsequently, and similar traits can evolve more than once, but tracking these traits is the best way to work out the phylogenetic relationships among organisms. We may track more than one trait, such as both feathers and skeletons. Results of these analytical tracks is a "cladogram," a treelike diagram that describe the order in which new traits and new creature evolves. Each branching point reflects the emergence of a ancestor that founded a group having derived characters that were not present in the group from which it evolved. The ancestor and its descendants constitute a branch in the tree; root word "clade" - branch. This style of reasoning is cladistics. The groups are hierarchial in a historical sense, a branch includes all animals that shared a recent common ancestor. "What is a bird" is a historical question.

An early fossil bird from the Jurassic, ca 150 mya: Archaeopteryx, Greek for "ancient wing," the link between birds and reptiles, looked like a small dinosaur with feathers preserved in fine-grained limestone. Fossils were found in Germany. It was crow-sized, had a snout with teeth like reptile, and also wings and a tail with long feathers. The feathers were much like modern feathers in having barbs, present on wings and tail as flight feathers, as well as on the body. The feathers were asymmetrical as are the flight feathers of flying birds. As the bases of the vane feathers were downy, Archaeopteryx probably had down feathers as well as vane feathers. The most famous fossils are in the BM(NH) in London and in the Berlin Museum, sites of popular displays. A cast of the fossil is on display on second floor of the Exhibit Museum at UM. This form was contemporary in the Jurassic with dinosaurs.

Six skeletal specimens of Archaeopteryx are known. The first two were discovered in 1861, shortly after Darwin's Origin of Species was published in 1859, bearing out Darwin's prediction of intermediate evolutionary links. The first specimen was brought to London, the second to the Berlin museum. The most recent one, from the Solnhofen limestone, was discovered in 1987 in the private fossil collection of the mayor of Solnhofen. It is the largest and apparently most adult (most ossified) skeleton. Another fossil, discovered in 1956, was exhibited in a museum near Solnhofen until 1974, when removed by an eccentric. Since then the specimen disappeared and was advertised for sale. A greedy seller wanted millions of dollars, and the scientific loss is sad. A seventh specimen is a stray feather imprint.

Archaeopteryx is an odd mix of reptilian (ancestral, or plesiomorphic - shared with ancestral groups) and bird-like (derived) traits. It provides a good starting point for the evolution of birds. It is very like modern birds in its feathers, furculum? or fused wishbone, fused sclerotic ring, wing digits 1-3 with phalangeal formula 2,3,4; and part of the way towards modern birds with its partly fused tarsometatarsus, and in having 4 toes, with 2-3-4-5 phalanges.

Archaeopteryx is unlike modern birds in retaining some primitive conditions shared with dinosaurs and other reptiles: teeth in jaw, curved backward; snout rather than a bill; small braincase (intermediate in relative body size between modern reptiles and modern birds) with large olfactory lobes; abdominal ribs or gastralia. The caudal vertebrae are unfused, n=23 in London specimen, each with a pair of feathers. This represents a reduction in tail size (as in birds); there were twice that many in some dinosaurs. The scapula and coracoid are fused in London specimen, but not in Solnhofen specimen. The sternum shows no keel (as opposed to modern flying birds). There is also no foramen triosseum between the coracoid, sternum and clavicles, so no pulley for breast muscles to pull up the wing. It has unfused carpals and metacarpals. No active flight was possible, as there are no feather attachment bumps on the ulna. The pubis has a footplate; the position of pubis (anterior, ventral, or posterior as in modern birds) is questionable (loosely attached, perhaps bent posteriorly after death; is somewhat ventral in fossils, the angle varies). Nearly all these characters are the same as in coelosaurian dinosaurs, which were apparently the ancestors of birds. A second hypothesis is a crocodile origin, which is not now as likely as the dinosaur hypothesis of ancestry. The UM Exhibits museum has it evolved from thecodont reptiles, the common ancestor of dinosaurs and birds.

Other non-bird-like characters of Archaeopteryx are long grasping forelimbs; horny and bony claws on end of digits, which were perhaps used in climbing trees (independently derived in modern birds: hoatzin); and the long balancing tail (shorter than in most dinosaurs).

Fossil Archaeopteryx differ in size and proportions. Some have been named as separate species or genera. The variation may be due to age. The variation suggests indeterminate growth as in reptiles, rather than no growth after adult as in birds. Rapid growth to adult size is characteristic of modern birds, not of reptiles or dinosaurs which continue to grow for years; the series of sizes in Archaeopteryx suggests slow indeterminate growth in these flying animals. Morphological analysis by Houck et al. (Science 1990) indicates variation was allometric, the proportions of wings and tail length scale to size of leg bones changing with overall (body) size. In larger animals, the tibia and toe bones are relatively shorter. It would have been harder for a large bird to perch(?), which suggests a terrestrial rather than arboreal perching life. In contrast, the teeth become relatively smaller with increased size of animal, perhaps due to wear and abrasion of teeth with age. The degree of fusion of foot bones also varies, apparently with age.

Behavior of Archaeopteryx: It was capable of flapping flight, derived either from gliding from the trees ("trees down"), or cursorial from the ground ("ground up")? The more popular hypothesis at present is "ground up", for two main reasons:. 1) Original reconstructions of Archaeopteryx show it gliding down from trees, but recent fossil work with plants in its area (Solnhofen) show no trees, but rather plants only to 3 m high. 2) The mechanics of flight suggest that a gliding bird would be unable to land on a perch, but a terrestrial vertebrate with long legs could run fast enough to generate lift for flight. According to the cursorial hypothesis, then, first they ran, then they jumped and leaped, and then they flew; development of feathers and wing would be useful to this behavior at each stage. The wing had claws but not obviously a wing to both climb and glide. Dinosaurs had long legs and could run too. Ancestors of birds had feathers before they could fly.

The dinosaur-bird connection was first noted by Huxley, Darwin's champion bulldog, noted similarity, and small dino named Compsognathus in same fossil beds as Archaeopteryx. He thought they were close relatives. This has now again become the main view.

Evolution of feathers.
Feathers may have evolved for thermal insulation, for flight, or for display. Early fossil birds had layers of bones that suggest annual growth rings. These suggest that birds were not constant high-temperature animals, not endothermic. Endothermy did not necessarily occur together with feathers. The earliest birds look like they could fly (Archaeopteryx). This line of reasoning suggests that feathers evolved for flight rather than for thermoregulation. A third possibility is that they evolved for displays (sex and aggression). We have less evidence on this idea, but cannot rule it out, for there is plenty of comparative evidence in some living reptiles with large flashy scales.

According to the fossil record, early birds had feathers, but some fossil birds left no impressions of feathers. Also, some dinosaurs that were not birds had feathers. They had vane feathers, feathers with a shaft, barbs and vane. These led to the larger flight feathers of modern birds, which are asymmetric around shaft with the barbs and vane of leading edge shorter. These were used for balance, or warmth, or courtship display. Other pre-bird dinosaurs had insulating feathers (down, with barbs separate not linked together to form an aerodynamic surface). Some dinosaurs had both types of feathers, vane and insulating.

When is a bird not a bird? Do feathers make an animal a bird?
Down feathers in a Jurassic fossil from China, Sinosauropteryx, exist in a row along its back and on its flanks. It has a skeleton like the coelosaur Compsognathus, a chicken-sized, short-armed form in the time of Archaeopteryx. Among other dinosaurs, both down feathers and vane feathers occur in Protarchaeopteryx (Eumaniraptors) and in Caudipteryx (Eumaniraptora). Compared with Archaeopteryx they had long teeth and short arms. Down occurs in other dinosaurs at lower or more basal levels. These are younger fossils than Archaeopteryx, however, by tens of millions of years, and they had a short tail and ossified sternum more birdlike than Archaeopteryx; some biologists think they're old flightless birds. They may have been "living fossils" for many millions of years after they first evolved. Some fossil dinosaurs had other features like birds (wishbone, breastbone, and loss of 4th and 5th fingers, hollow long bones, long arms and hands, sideways flexing wrists, nesting behavior, and rapid growth rates: none of these are diagnostic of birds, however). Why aren't these feathered reptiles "birds"? Birds have arms at least as long as legs, longer and flightworthy feathers (though these features have been lost in some fossil and some modern flightless birds). Fossils indicate that birds evolved from small coelosaurian dinosaurs. Photos of these fossils can be seen in the July 1998 issue of National Geographic.

Another, cladistic, definition of a bird: Archaeopteryx and other modern birds and all the descendants of their most recent common ancestor. This type of definition is contrasted with "diagnosis", a list of unique features that only they possess. Feathers are not a diagnostic feature according to the cladistic definition since these occurred in some fossils that were outside the ancestral club of Archaeopteryx and living birds. Some dinosaurs had feathers.

Another contender for status as the earliest bird:
A fossil from Texas was discovered in 1983 by dinosaur paleontologist Chatterjee. It dates to 225 mya, late Triassic (before Jurassic), earlier than Archaeopteryx. It is Protoavis texensis, "first bird from Texas". In the Phil. Trans. Roy. Soc. London, 1991, the skull was described. Later, in 1995, the rest of the fossil was described. It is earlier than dinosaurs, so dinosaurs would not be ancestral to birds; rather, crocodiles came in the fossil record in time to be the sister group of birds. Most paleontologists question whether Protoavis is a bird, however. According to Chatterjee, it has a large braincase (but interpretation was based on very incomplete skull), large orbit (like birds, but based on incomplete parts), sclerotic ring (but glued together), furculum (wishbone, based on incomplete parts). Also, the cervical vertebrae and dorsal vertebrae have large vertebral canals (both are avian). With no feathers, but all bones, it is considered a small Triassic reptile of unknown affinity.

Other early fossil birds:
Another old fossil bird is Confuciusornis, from China, late Jurassic or early Cretaceous, over 120 million years old. It is chicken-sized, with feathers, a toothless beak, ability to fly well, lighter bones and shorter tail than Archaeopteryx, with claws on the wings. The male had long tail feathers, the female a short tail and slightly smaller body (Nat. Geogr). It was found in the same fossil beds in China as the pre-birds. The status of this fossil not well known.

From the early Cretaceous: ca 135-100 mya, Sinornis was found, a sparrow-sized bird from an old lake bed in China, described in 1992 Science. Primitive traits were a short snout; teeth; a flexible manus (finger bones were separate); coracoid buttress to sternum which would resist compressive forces by the flight muscles and provide skeletal support for primaries and secondaries; footed pubis; stomach ribs; and free elements of the pelvic girdle. In contrast to Archaeopteryx it had a broad sternum, a wing-folding mechanism with a modified wrist bone for distal wing bones to fold into, a pygostyle, and large reversed hallux. It represents a transition from the primitive wing of Archaeopteryx to a specialized wing more like that of modern birds, well on its way to flight. All these traits are related to flight or perching, as well as fewer dorsal vertebrae (11 vs 14 as in Archaeopteryx), tail short for 8 free vertebrae and pygostyle for the tail fan. Modern avian flight and perching therefore evolved in small-bodied birds in inland habitats not long after Archaeopteryx.

Ambyortus from Mongolia, also incomplete, is about the same age (Kurochkin). This is the first bird that looks like a bird.

Fossils have been found in Las Hoyas, Spain of Iberomesornis, also from the early Cretaceous. It is a sparrow-sized bird with incomplete strut-like coracoids like a bird, but the pelvic girdle is primitive. Compared to birds, the ilium, ischium, and pubis all are parallel and are directed backward.

Eoalulavis, from Spain, possesses the earliest known "alula," a tuft of feathers attached to the thumb, as in modern birds. The alula gives good control at low flying speeds, in takeoffs and landings. 115 mya.

In early Cretaceous birds, flight has taken over. Fossil birds were small, sparrow- to pigeon-sized, they could fly, they looked like birds, with feathers, sustained flight ability, and could perch (reversible first toe or hallux).

From South America, a large group Enantiornithes represents the late Cretaceous in Argentina. Several genera and species have been found. They have a birdlike wing skeleton, with a carpometacarpus. Metatarsals are different from modern birds in that they are not united proximally to distally. Another bird Patagopteryx was found here, with a more modern tarsometatarsus.

From the late Cretaeous, in Madagascar, we have Rahona, 65-70 mya. It has feathered wings, long tail, and a sickle claw like a meat-eating theropod dinosaur. The pelvic and pubic bones resemble Archaeopteryx. No feathers are preserved. Science 1998, 279:1851, 1915-1919. Other Cretaceous bird fossils from Madagascar are described in Nature 382:532 (1996).

An estimate of phylogenetic relationships (Chiappe 1995):

1-Aves, <26 caudal vertebrae, caudals with short prezygapophyses, reversed hallux, teeth with unserrated crowns and crown-base constrictions (Archaeopteryx, modern birds, and other descendants of their common ancestor);

2-carpometacarpus, other traits (Mononychus, Alvarezsausus; and Neornithes);

3-pygostyle, strut-like coracoid, sharp caudal end of scapula, radius:ulna ratio smaller than 0.7 (Las Hoyas bird + later birds);

4-synsacrum with more than 8 vertebrae, heterocoelous cervicals, distal tarsals fused to metatarsals (Enantiornithes + later birds);

5-quadrate with 3 distal condyles forming a triangle, caudal prezygapophyses absent or much reduced, absence of pubic foot, tarsometatarsal vascular distal foramen (Patagopteryx & later birds);

6-Ornithurae: sharp and pointed quadrate orbital process, < 11 dorsals, procoracoid process, small acetabulum, pubis parallel to ischium and ilium, femur with large patellar groove (Hesperornis and later birds);

7- head of humerus globe-shaped, with cranio-caudal convex head. Carinatae. Ichthyornis and modern birds and all the descendants from their common ancestor (Neornithes).

The trend throughout this sequence is a movement from hip extension (with long tail, forelimb modifications, and tail coupled to leg movements, and slow growth rates) to knee flexion (short tail skeleton, hindlimb modifications and fast growth rates). Flight has been lost in several groups, including living ratites (ostrich, kiwi, etc.).

This sequence is not the same as the earliest age when these groups are known from the fossil record. Only Archaeopteryx fits this criterion, so Archaeopteryx is the least controversial among the early fossil birds.

Mid-Cretaceous fossils Hesperornis and Ichthyornis were marine, toothed, large birds. Hesperornis includes 13 species of diving seabird ranging in size from a chicken to a large penguin, foot-propelled, superficially resembled modern loons, flightless. No keel on sternum, large cnemial crest on tarsometatarsus for swimming muscles like loon & grebe but different origin, vertebrae were heterocoelous like modern birds. Ichthyornis was a flying, tern-like bird, keel on sternum, same time as plesiosaurs in oceans. Apparently off the main lineage of bird evolution, vertebrae amphicoelous like primitive reptile, they left no living descendants. Both are closely related to modern birds, in Ornithurae.


1) The Cretaceous radiation of birds, as opposed to the K-T (Cretaceous Mesozoic-Tertiary Cenozoic) extinction and recent radiation. The same debate occurs with respect to mammals. Which model is accurate? Few fossils before K-T are clearly the same as modern orders, the most accepted being waterfowl. On the other hand, there are fossils of many orders by the Oligocene, and several in the Eocene. These are minimal ages, and so the birds could be more ancient than K-T. Did Cretaceous birds go extinct at the K-T boundary (meteor shower)? Some did, but some went extinct in middle Cretaceous so K-T would not explain their extinction.

1a) Molecular genetics (Hedges et al., 1996, Science 381:227). Fossils are used to date times of divergence of major groups, then the rate is estimated at which nucleotides have diverged between these groups. A bird-mammal reference time of lineage separation based on fossils (outside the bird lineages) indicates 310 mya (synapsid reptiles, as ancestors of mammals; and diapsid reptiles, as ancestors of birds). The first appearance of amniotes is 335 mya, of tetrapods is 370 mya, so the split between birds and mammals can't be much earlier than 310 mya. Using genes that diverge at consistent rates, and this fossil which is external to birds (vs internal) percent difference (allowing for multiple substitutions at saturated sites) divergence times for bird orders are indicated that are older than the earliest known fossils. The age for bird orders estimated from molecular data is 100 mya (in Cretaceous), while the age for the earliest fossils is post-K-T, ca 50-60 mya (in Tertiary). Mammals produce the same story, where the ages estimated by molecular genetics are about 50% larger than ages estimated from oldest fossils. This suggests we can't take fossil ages at their face value, only as minimal ages. These orders occurred by the ages determined but may have arisen considerably earlier.

These ages agree with times of separation of continental plates, in earth's plate tectonic movements, more closely than with K-T phenomena such as loss of reptile lineages that were competitors with early birds and mammals.

1b) Cooper & Penny, Science 275:1109(1997), illustrate another approach, and produce the same results in terms of an ancient pre-K-T origin of orders, but with earlier estimates of dates of evolutionary divergence, using different models of interpreting genetic divergence from genetic sequence data, suggest 140 mya for orders. That's back at Archaeopteryx time! Molecular biologists are telling us that many groups evolved much earlier than suggested by the fossil records. Are they right, or should they be more concerned about fossil evidence of times of origins of groups?

We are in a remarkable time of discovery of fossils. What we know now is but a stage in our ultimate understanding of the origin of birds and the origin of flight. May you live in interesting times!

60 mya to 2 mya. Eocene, 40-60 mya. Several modern orders of birds are represented, including waterbirds such as loons, cranes, ducks, petrels. Songbirds are reported from Eocene. The Eocene had large flightless big-billed birds Diatryma in W. North America and Europe, which were perhaps predators, scavengers or herbivores (Anders 1992). Also during this epoch existed big wading birds with duck-like heads Presbyornis. Large flightless carnivores were in South America, the Phorarhacids, which are large flightless carnivores where no mammals had this niche (lower Oligocene through Pliocene). One species of this group was found in Florida. How did it get there?

The largest sea bird was found at airport site in South Carolina in 1987, with bony teeth, a wingspan of 18-19 feet, weight ca 90 pounds (vs albatross, wingspan 11 feet, 20 pounds, and some pterosaurs had 40-foot wingspans). It must have depended on winds for long gliding flight. about 30 million years ago, in the Oligocene. Old World vultures have been found in the New World, and New World vultures in the Old World; not all groups were as restricted in distribution as they are now.

Huge vultures were found from the ice age: Teratornis meriami had a 4m wingspan in southern California, La Brea tarpits have >85,000 bird fossils of 133 species. Another vulture species T. incredibilis in Nevada had a wing span of 5-6 m, and fed on the megafauna of the time (mammoth). The California Condor is similarly adapted for an earlier time.

Relationships among modern groups of birds:
1) Paleognathous birds (large vomer connects to maxilla, pterygoids and palatines in straight line, basipterygoid articulation with pterygoid, as in ostrich and other large flightless birds, also S American tinamous),

2) ducks and chickens, and

3) others. There is not agreement among all ornithologists about these groups (either group monophyly, or branch position). One molecular systematics study (Mindell) suggests songbirds are basal to all the other birds.

How many species of birds have existed?
This is uncertain. One paleontologist Alex Wetmore estimated that 25% of modern birds could be sorted correctly by size and shape alone, so our estimates of fossil birds are minimal. 900 modern species have been found from Pleistocene, with estimated N =10,600 (Brodkorb 1971), so probably most bird species evolved then, in contrast ca 13 found from Pliocene. The estimated total is 150,000 species in history, based on estimate from numbers of contemporaneous families, species, and faunal turnovers.

How do species form?
There are two models of how bird populations evolve into species. 1) Dispersal and colonization of islands, or isolated habitats. Populations invade from mainland, or move from island to island as in the Galapagos. On each isolated site they evolve adaptations to local conditions, so they split and differentiate in their genes in different pathways until they are distinct species. In this model the birds disperse from one area to another. 2) Fragmentation of continental habitats that once were continuous. This is a question of vicariance (from "vicar" = local representative deputy of the bishop or pope). Regional splits of habitats on a large scale lead to differentiation of the populations in them, as when the Amazon basin was split into different areas in former dry period of earth's history. In this model, the birds stay in one place and the land forms or habitats change over time. In both processes, one ancestral population can split into two derived populations that differ enough to call them species.


National Geographic, July 1998

Chiappe, L. 1995. "The first 85 million years of avian evolution," Nature 378:349-355.

Hedges, S.B., Parker, P.H., Sibley, C.G. & Kumar, S. 1996. "Continental breakup and the ordinal diversification of birds and mammals." Nature 381:226-229.

Padian, K. & Chiappe, L. 1998. "The origin of birds and their flight." Scientific American 278, Feb. 1998:38-47.

Padian, K. & Chiappe, L.M. 1998. "The origin and early evolution of birds." Biological Reviews.

to larger selected bibliography on avian origins and early evolution; to a summary of extinct birds

To top of Evolutionary History of Birds
To UMMZ Bird Division Home Page