The iconic ‘Urvogel’ Archaeopteryx has long been the ‘yardstick’ for bird evolution1. However, in the past 15 years our knowledge on early bird evolution has increased significantly, mainly because of the discovery of bird-like theropods and early avialans with feather preservation from the Middle–Late Jurassic and Early Cretaceous periods of China2,3,4,5,6. This has provided a wealth of new data on the evolution of feathers and their functional context, which has challenged many traditional ideas3,7,13. Although Archaeopteryx still plays a pivotal role in discussions of avialan origins and basal relationships5,6,14, information on the plumage of this crucial taxon has so far mainly been limited to the London and Berlin specimens, with some additional information from the recently described Thermopolis specimen.

The new, excellently preserved 11th skeletal specimen of Archaeopteryx shows extensive feather preservation, not only along the wings and tail (as in the London, Berlin and Thermopolis specimens), but also in other parts of the body (Fig. 1), and thus helps to clarify controversial issues about the plumage of basal avialans. Most of the skeleton is preserved in articulation, but the skull is incomplete and removed from the neck, and the right pectoral girdle and wing have been displaced to lie below the left hindlimb; the left wing is missing (Fig. 1, Extended Data Figs 1 and 2 and Extended Data Table 1).

Figure 1: Overview photograph of the 11th skeletal specimen of Archaeopteryx. Abbreviations: bp, body plumage; ft, feather ‘trousers’; hf, hackle feathers; lh, left hindlimb; n, neck; rf, right forelimb; rh, right hindlimb; rw, right wing; s, skull remains; t, tail; tf, tail feathers. Scale bar, 5 cm. Full size image Download PowerPoint slide

The plumage of the new specimen is preserved mainly as imprints, with a few patches showing possible preservation of organic material. Remains of the body plumage are preserved along the presacral axial skeleton and the pelvis (Fig. 1 and Extended Data Fig. 3). All body feathers show a high degree of overlapping, but because of excellent preservation the identification of feather structures is possible. The body plumage of Archaeopteryx consists of bilaterally symmetrical pennaceous feathers, possessing long, thin and slightly curved rhachides with regularly oriented barbs, forming narrow vanes (Fig. 2d). Parts of the neck plumage had separated from the cervical vertebrae with the skull and are preserved between the vertebral column and the skull remains (Fig. 1 and Extended Data Fig. 3); they show that pennaceous feathers also covered the neck up to the head.

Figure 2: Details of the plumage of the 11th specimen of Archaeopteryx. a, Right wing from dorsal view. b, Leg feathers of the right hindlimb. c, Detail of leg feathers of right hindlimb. d, Detail of body plumage from the belly region. e, Overview of tail feathers. f, Detail of asymmetrical feathers at the lateral side of the tail. Arrow in a indicates gap in the wing caused by the overlapping left foot. Scale bars, 1 cm. Full size image Download PowerPoint slide

Long hindlimb feathers are preserved along the femur and tibiotarsus (Fig. 2b, c). The length of these feathers is about 40–45 mm (51–58% of tibiotarsus length), but decreases from the middle to the distal end of the tibiotarsus to about 29 mm. More than 18 feathers are preserved along the right tibiotarsus; they are arranged mainly in parallel and preserved with their long axes oriented almost perpendicularly towards the tibiotarsus. The narrow vanes are fully symmetrical in shape, and seem to expand slightly towards their distal end, although this might be an artefact due to overlapping in the more basal parts. As in the body feathers, the rhachides are slightly curved. In the region of the right ankle and proximal metatarsus, a row of very short (about 25% of the length of the feathers on the mid-length of the tibiotarsus), narrow-vaned, symmetrical feathers is preserved (Fig. 2b, c and Extended Data Fig. 4d), unlike the elongate pennaceous feathers in other paravians3,4,15.

The disarticulated wing shows the dorsal aspect of the plumage and possesses 12 primary remiges (Fig. 2a and Extended Data Figs 4 and 5). A visible gap between the second and third primaries is artificial and caused by a disruption to the wing by the left foot during burial. The distal ends of the rhachides are extremely thin, but expand abruptly approximately 20–25% of the feather length towards their proximal parts. Dorsal coverts are preserved above the primary feathers and measure approximately half their length, as in extant birds16. The actual number of secondaries cannot be estimated owing to overlapping with coverts. These finds contradict an interpretation of the plumage of Archaeopteryx based on the Berlin specimen, according to which the number of primaries is reduced and the coverts are extremely elongated9. Thus, the wing morphology of Archaeopteryx conforms to that of modern birds, as previously suggested1 (see Extended Data Fig. 5 and Supplementary Information).

The remiges of the new specimen of Archaeopteryx show robust rhachides, indicating that recent studies assuming limited flight ability in Archaeopteryx and basal birds based on the relative rhachis diameter8 might be in error owing to the poorer preservation quality of the feathers measured (see Supplementary Information and Extended Data Fig. 6).

For the first time the distal end of the tail is completely preserved in Archaeopteryx (Fig. 2e). The distal rectrices are symmetrical in shape and extremely long, measuring about 99–114 mm (up to 61% of the length of the bony tail). The rhachis morphology resembles that of the wing feathers. The preserved shape of the tail plumage is probably narrower than originally, because some rectrices are completely covered by others. The distal end of the tail is bifurcated (Fig. 2e), but it cannot be ruled out that this might be caused by a moulting process of the terminal feather, which would indicate a centrifugal moulting pattern as in recent birds16. Lateral rectrices are slightly shorter than distal ones and clearly asymmetrical in shape (Fig. 2f). This may suggest that the tail of Archaeopteryx possessed an additional aerodynamic function increasing the total lift of the animal17,18.

To trace the evolution of pennaceous feathers across the origin of birds, we mapped their occurrences onto a phylogeny of advanced maniraptoran theropods (Fig. 3, Extended Data Fig. 7 and Supplementary Table 2). Our analysis places Archaeopteryx at the base of Avialae, but more derived than a clade formed by Pedopenna, Eosinopteryx, Anchiornis and Xiaotingia. Furthermore, we found Troodontidae as the sister group of Avialae, making Deinonychosauria (as traditionally conceived to include Dromaeosauridae and Troodontidae) paraphyletic (Extended Data Fig. 7), as recently hypothesized6.

Figure 3: Evolution of pennaceous feathers in maniraptoran theropods. Several character states referring to the presence/absence of pennaceous feathers on different parts of the body mapped onto a phylogenetic hypothesis of maniraptoran theropods. For details of phylogenetic analysis see Supplementary Information. Full size image Download PowerPoint slide

Our results indicate that the distribution of pennaceous feathers is highly variable even before the advent of flight, indicating that different functional aspects were involved in the early evolution of these structures. Symmetrical pennaceous body feathers were present in the ancestor of the clade containing Oviraptorosauria and Paraves (Pennaraptora; see Supplementary Information). However, this might not be the ‘true’ origin of this feather type, because the morphology of body feathers in basal coelurosaurs is often uncertain owing to poor preservation or strong overlap of single feathers within the plumage19. Elongated, symmetrically shaped pennaceous remiges first appear in the stem species of Pennaraptora, whereas asymmetrical aerodynamically shaped remiges were convergently developed in the dromaeosaurid Microraptor3 and the common ancestor of Archaeopteryx and Pygostylia10,20. Alula feathers allowing high manoeuvrability during low-speed flight in extant birds21 were convergently developed in Microraptor and Ornithothoraces3,22.

Elongated pennaceous feathers along the tibia are found in Microraptor3, Sinornithosaurus23 and many basal avialans4,5,15, indicating their presence in the last common ancestor of Dromaeosauridae, Troodontidae and Avialae. In contrast, elongated pennaceous feathers on the metatarsus are only known for Microraptor3, Xiaotingia5, Anchiornis4, Pedopenna24 and Sapeornis15. Based on our tree topology this feature did not originate singularly within Paraves, but evolved three times independently, in Dromaeosauridae, Sapeornis and the clade containing Anchiornis, Pedopenna and Xiaotingia.

Pennaceous rectrices were also present in the common ancestor of Pennaraptora, but the original shape of the tail plumage and its evolution are ambiguous. In Protarchaeopteryx, Caudipteryx and Microraptor, elongated pennaceous rectrices are present in the distal half of the tail2,12, whereas Jeholornis seems to be unique in showing a proximal and a distal tail fan25. The tail of Similicaudipteryx, Jinfengopteryx, Anchiornis and Archaeopteryx is completely covered with such feathers4,20,26, whereas Eosinopteryx lacks them completely27. Within short-tailed avialans, elongated rectrices are absent in some Enantiornithes, Yanornis and many Confuciusornis specimens15,28. In contrast, tail plumages with a pair of extremely elongated pennaceous feathers are known for Confuciusornis, some Enantiornithes and Epidexipteryx28,29. A fan-shaped tail plumage as present in modern birds is convergently developed in Sapeornis and several Ornithuromorpha15,30.

Assuming a single origin of pennaceous feathers within theropods, the present data indicate that this feather type evolved in the common ancestor of Pennaraptora, but rapidly differentiated into various subtypes within different body regions, leading to a considerable plumage diversity compared with more basal theropods. Based on these results, the evolution of pennaceous feathers is generally decoupled from the origin of flight and might be related to other biological roles. Like the filamentous plumages of more basal theropods, pennaceous body feathers were certainly important for body insulation. However, these feathers may have further been used for camouflage and display7,11. Furthermore, elongated remiges on the forelimb were probably used for balance during locomotion and breeding10. Although some taxa possess prominent feather trousers, an initial aerodynamic function, as previously hypothesized3,13, can be rejected because the presence of this character is extremely variable within Paraves, and elongated hindlimb feathers were usually restricted to the tibia and are symmetrical in shape. Thus, probably non-volant taxa such as Anchiornis or Xiaotingia might have used these feathers for display, breeding or other functions. The display function might be supported by the finding of complex colour or iridescent patterns in hindlimb feathers11,12. Archaeopteryx might have used its ‘trouser’ additionally as a vertical plane during landing, similar to recent raptors. Only for Microraptor does an improved aerodynamic adaptation of the trousers seem plausible, as indicated by the triangular shape of the trousers and the asymmetrical shape of the feathers. These results contradict the hypothesis that the flapping flight of modern birds was preceded by a four-winged gliding stage3,13, and indicates that flight ability in Avialae and Microraptor evolved convergently and was functionally different.

The huge diversity and homoplasy of plumage morphology of the tail within Pennaraptora indicate that this character may initially have evolved for display10,12,29. The asymmetrical shape of the lateral rectrices in Archaeopteryx and ornithurines indicates a secondary aerodynamic function for the tail.

The current results show that pennaceous feathers were present in the common ancestor of Pennaraptora. As in extant birds, this feather type probably fulfilled several biological roles related to body insulation, manoeuvrability, brooding, camouflage and display7,16. However, given the great diversity of pennaceous feathers found within different body regions and across the phylogeny, it seems plausible that the evolution of this feather type (especially in the wing, hindlimbs and tail) was primarily driven by display functions. These structures then represented an exaptation that was repeatedly, and probably convergently, recruited for aerodynamic functions. This indicates that the origin and evolution of flight in theropod dinosaurs were more complex than previously thought, and (as already shown for other anatomical complexes, for example forelimb myology and breathing apparatus) could draw on structures that evolved in different functional contexts.