Fossils explained 53

Fossils explained 53 Titans of the skies: azhdarchid pterosaurs Pterosaurs, the flying reptiles of the Mesozoic, often play second fiddle in popularity to their contemporaries, the dinosaurs. Such treatment conceals the remarkable diversity and biology of this group: not only were pterosaurs the first vertebrates to achieve powered flight, but they also existed for 160 million years longer than any other flying vertebrates. Named after the Uzbek mythical dragon azhdarkho, the Azhdarchidae are among the most enigmatic of all pterosaurs. As with most pterosaurs, azhdarchid remains are rare, and their fossil record is largely represented by isolated bones or incomplete skeletons. Despite the collection of azhdarchid fossils over the last 100 years, recognition of these pterosaurs as a distinct group was not achieved until relatively recently. It is now clear that the azhdarchids were a highly successful group that probably first appeared in the Early Cretaceous, gradually spreading across the globe until the latest Cretaceous when they, as one of the last remaining groups of pterosaurs, became extinct. Although most notable for achieving wingspans comparable with light aircraft, other aspects of azhdarchid morphology and ecology make them not just aberrant animals but also unusual pterosaurs. Mark Witton Palaeobiology Research Group, School of Earth and Environmental Sciences, University of Portsmouth, Portstmouth PO1 3QL, UK. mark.witton@port.ac.uk Azhdarchid origins Some controversy surrounds the ancestry of pterosaurs, but most agree that they should be included within Archosauria, the reptilian group that includes crocodiles, dinosaurs and birds. Exactly where they fall in this category is still debated because of the highly specialised nature of even the most basal pterosaurs: their highly modified skeletons leave few clues to their ancestry. The first pterosaur fossils are found in the late Triassic, and these basal groups are typically characterized by long tails and toothed jaws. These groups dominated pterosaur evolution until the late Jurassic but ultimately yielded control of the skies to a diverse group of derived, tailless pterosaurs: the Pterodactyloidea. This group diversified in the Early Cretaceous and includes the Azhdarchidae as the last and most derived family to evolve (Fig. 1). Cretaceous forms such as Tapejara and Tupuxuara are perhaps the closest related taxa to the azhdarchids, sharing similarities in their toothless jaws, relatively short wings and subequal limbs. Fig. 1. Relationships of the Pterodactyloidea (based on Unwin 2003 see Suggested reading). Basal pterosaurs is used here to refer to all pterosaur taxa basal to Pterodactyloidea. Although pterosaur systematics are controversial, the Azhdarchidae are frequently suggested to be the most derived clade within Pterodactyloidea. Basal pterosaurs are represented by the skull of Rhamphorhynchus, Ornithocheiroidea by Coloborhynchus, Ctenochasmatoidea by Gnathostoma, Dsungaripteridae by Germanodactylus, Tapejaridae by Tapejara, and the Azhdarchidae by Quetzalcoatlus. 33

Fig. 2. Reported occurrences of azhdarchid fossils across the globe. Fig. 3. Restoration of the skeleton of Quetzalcoatlus, a large azhdarchid from Texas. The earliest occurrences of azhdarchid fossils occur in Lower Cretaceous deposits of China and Brazil, although controversial material from Tanzania may extend azhdarchid origins down into the Upper Jurassic of Africa. Because the first well-preserved azhdarchid material occurs in geographically distant regions, their location of origin is unclear. Azhdarchids appear to have become more widespread in the Northern Hemisphere than the Southern, with numerous finds from Upper Cretaceous rocks across Asia, Europe and North America (Fig. 2), but remains in Australia and New Zealand indicate that azhdarchids were also present in southern regions by Late Cretaceous times. Skeletal anatomy Azhdarchid skeletons possess many features that clearly distinguish them from other pterosaurs (Fig. 3). Their jaws are straight and elongate (over 2 m long in some species) but possess no teeth. Instead, foramina lining the jaws of some taxa indicate that a bird-like horny beak was present. The skull is lightened by a large pneumatized foramen known as the nasoantorbital fenestra (within which the nostrils were placed) and reduced jaw muscles. The neck in some taxa is almost twice as long as the head (making them the longest necks of any pterosaurs); a highly diagnostic feature achieved through hyperelongation of the mid-series neck vertebrae. Unlike most other pterosaurs, the neck of azhdarchids articulates with the underside of the cranium and orientates the long axis of the skull, almost 90 to the shaft of the neck in some species. Structures on the middle-neck vertebrae suggest a stiff neck with very limited articulation, whilst the cranial condyle is reduced to a poorly developed, hemispherical structure allowing restricted movement of the skull. As with all pterosaurs, the torso is short (perhaps less than half the length of the skull) but strongly fused to anchor the flight muscles of the arms and legs. The wings comprise a short but powerfully built humerus, with the radius and ulna leading to a long, robust metacarpal which in turn supports a relatively reduced wing finger. This finger supported the main flight membrane; a thin but responsive organ also attached to the forearm, torso and leg. In all pterosaurs the wing area is increased through additional membranes supported between the pteroid (a long bony shaft projecting from the wrist) and shoulder, with a third set of membranes between the tail and hindlimbs. When grounded, azhdarchids walked on the distal end of their wing metacarpal (further supported by three small digits) and, because of unusually long hindlimbs, may have held their 34

Fig. 4. Size ranges of azhdarchids. The fragmentary remains of many giant azhdarchids (such as Arambourgiania) mean that their sizes can only be estimated based on more complete material from other azhdarchid taxa. In the case of Arambourgiania, wingspans of 11 13 m are predicted; the largest estimate is figured here. body almost parallel with the ground whilst standing and walking. This contrasts with other large pterosaurs that had disproportionately long forelimbs, forcing the body into a more erect posture whilst walking and perhaps limiting their terrestrial capability. Such limb disproportions are not seen in azhdarchids however, suggesting a terrestrial prowess better than most pterosaurs. The largest flying animals ever? Many pterosaurs grew to sizes far in excess of any living birds, but no other pterosaur group obtained gigantic size as frequently as the azhdarchids. Early forms held modest wingspans of 2.5 m, but their size increased steadily throughout the Cretaceous until they had the largest wingspans of any flying vertebrates (Fig. 4). Their size was the cause of some confusion in the early twentieth century: an extremely elongate neck-vertebra discovered in Jordan in the 1920s (now known to belong to the azhdarchid Arambourgiania) was originally thought to be a pterosaur wing element, such was its size and peculiar features. It was not until the 1970s and the discovery of more complete azhdarchid material in Upper Cretaceous deposits of Texas that the true size of the azhdarchids was appreciated. These sediments yielded three azhdarchid individuals belonging to the genus Quetzalcoatlus, with one, despite its fragmentary nature, clearly enormous. Initial estimates gave this giant a wingspan of somewhere between 11 and 21 m, but an improved understanding of azhdarchid anatomy now favours the lower estimate. When standing, the shoulder of Quetzalcoatlus would have stood some 2.5 m off the ground a height comparable with a mature Asian elephant. The discovery of Quetzalcoatlus prompted some aeronautical engineers to state that the genus represented the largest flying creature possible. However, and very likely to the dissatisfaction of aeronautical engineers everywhere, new discoveries in Europe and a reappraisal of fossil material of the genus Arambourgiania hint at forms with wingspans of 12 m or more. Equally, the probability of fossilization indicates that the azhdarchid individuals known to palaeontologists are very likely to be of average sizes: truly enormous, freakishly big individuals have almost no chance of entering the fossil record but are very likely to have existed in the same way that giant individuals are seen in modern animal populations. Being so much larger than all other known flying animals, it seems quite plausible that the azhdarchids were indeed the largest fliers of all time. Despite their enormous size, azhdarchids possess typical pterosaur features of hollow and extremely thin-walled bones. An extensive pneumatic system and thin bone histology makes azhdarchid skeletons deceptively lightweight and, although establishing pterosaur mass is controversial to the point where some workers claim it impossible, most workers agree that even the largest azhdarchids would had weighed relatively little. Estimates for Quetzalcoatlus with a 10 11 m wingspan range from 70 to 250 kg, but it is Fig. 5. Azhdarchid wing configuration (demonstrated by Quetzalcoatlus bottom left) compared to Pteranodon (bottom right). The shorter, broader wings of the azhdarchid are more akin to the statically soaring Andean condor (top left) than the dynamically soaring wandering albatross (top right). 35

Fig. 6. Shoulder and wing structures in pterosaurs (based on Frey et al. 2003 see Suggested reading). The stable top-decker configuration with the centre of gravity (indicated by crosses) below the wings (top) seen in ornithocheiroid pterosaurs; the characteristic azhdarchid middle-decker configuration with the centre of gravity between the wings; and the unstable bottom-decker configuration seen in tapejarid pterosaurs with the centre of gravity above the wings. probable that the latter figure is far too high. A figure between 70 and 85 kg conforms more to our understanding of pterosaur anatomy and flight dynamics, but more data regarding pterosaur soft tissues is needed before these figures can be verified. In the air As seems to be the case with many fossil groups, palaeontologists did not portray an attractive picture of pterosaurs for much of the twentieth century. Pterosaur aerodynamics were particularly criticised, with flight modelling of large pterosaurs during the 1970s casting doubts over their flight capabilities. It was thought that take-off was particularly difficult, with either a headwind or a long drop from a cliff edge required to become airborne. It was even suggested that if conditions were too adverse the pterosaurs would have to remain grounded as their clumsy flight capability would not have withstood the flight stresses of blustery conditions. This attitude has now changed to one more understanding of this highly successful and diverse group: evidence from numerous pterosaur fossils indicates that many species had a hair-like integument covering their bodies, necks and skulls, perhaps suggesting that pterosaurs controlled their core body temperatures in a manner similar to all modern actively flying animals (birds, bats and flying insects possess methods of temperature regulation). This possibility has led to thoughts that pterosaurs were active, powerfully flying creatures able to become airborne regardless of weather or topographical conditions. Because of their size many azhdarchids probably required a brief run-up to assist with take-off, as seen in larger species of modern birds. Such locomotion was probably possible: pterosaur trackways indicate that pterosaurs may have been competent terrestrial animals, with both quadrupedal galloping and bipedal running possibly used during take-off. Although headwinds, slopes and cliffs may have assisted azhdarchids in becoming airborne, it seems unlikely that they would be mandatory for take-off. Many large pterosaurs are thought to have been dynamic soarers akin to modern gulls and albatross, manipulating air currents to travel with minimal expenditure of energy on flapping, but several features of azhdarchid skeletons indicate that their method of flight may have been different. Because the pterosaur wing incorporates both sets of limbs in its construction, an idea of wing shape or aspect ratio (AR = wingspan 2 /wing area) can be ascertained through relative limb lengths (Fig. 5). Most large pterosaurs had disproportionately long forelimbs compared to their hindlimbs and consequently have long, narrow wings (high aspect ratio). However, azhdarchids have relatively long hind limbs and truncated distal components of the wing finger, resulting in relatively short and broad wings (low aspect ratio). This low aspect ratio has implications for other basic flight calculations. Wing loading (WL = mass/ wing area) is indicative of flying style: high wing loading is typical of faster flight whereas low wing loading suggests a gentler pace. Generally, estimates of pterosaur wing loading are lower than flying vertebrates of equivalent dimensions, although there are no modern flying animals large enough to compare the biggest azhdarchids against. Because of their broad wings azhdarchids have lower wing loading values than other large pterosaurs, suggesting that azhdarchids may have flown relatively slowly. Although smaller forms may have been capable of powered, flapping flight, the energy requirements for flapping flight in larger azhdarchids probably excluded the largest forms from this activity. The position of the shoulder joint (glenoid) is also unusual in azhdarchids. It is neither in the topdecker configuration seen in ornithocheiroid pterosaurs or the bottom-decker condition known from tapejarid pterosaurs, but instead forms a middle decker construction with the glenoid positioned halfway up the shoulder girdle (Fig. 6). Top-decker pterosaurs carried their centre of gravity beneath the wings for stable flight (akin to a modern cargo plane) whereas bottom-deckers acted more like fighter planes with the wings below the centre of gravity, making for a less stable but more manoeuvrable configuration. As such, the unique middle-decker system of azhdarchids may have been a compromise 36

between stability and manoeuvrability. Although many of these characters make azhdarchids unique amongst pterosaurs, they do conform to some types of extant birds. Azhdarchids may have not been dynamic soarers like gulls, but their wing loading and aspect ratios indicate that they may have been static soarers akin to modern vultures and storks. These birds manipulate thermal updrafts to gain high altitude, then slowly glide to the next thermal column. The increased manoeuvrability offered by the middle-decker shoulder configuration may have enabled large azhdarchids to control their position within thermals more efficiently. As such updrafts are not generated readily on open water, it seems unlikely that large azhdarchids were capable of long-distance migration over bodies of water as suggested by some authors if azhdarchids did perform long, trans-continental migrations, they probably did so over land. The unusual abundance of azhdarchid remains in continental deposits may support this notion; most large pterosaur remains are preserved in open marine or coastal sediments, implying a different environmental preference for azhdarchids. Because of their lower masses and lower energy requirements to achieve flapping flight, smaller azhdarchids may have been capable of flapping flight as well as soaring. Feeding habits The majority of pterosaur fossils are found in marine or coastal sediments, ranging from near-shore lagoons to shallow seaways. Although some azhdarchids are found in such settings, most are found in lucustrine deposits and, when found in marine deposits, azhdarchids are often associated with material derived from continental settings such as plants or dinosaur fossils. This strongly suggests a more land-based ecology for azhdarchids than other pterosaurs: from palaeoenvironmental analysis of azhdarchid bearing sediments, it seems that azhdarchids like Quetzalcoatlus inhabited arid environments with few large bodies of water, whilst others such as Zhejiangopterus occupied wooded settings crossed with small rivers and streams. Some controversy has surrounded the ecological significance of the apparent azhdarchid preference for terrestrial environments, particularly concerning their feeding habits. Because of its unusual habitat and in situ association with numerous sauropod dinosaur remains, Quetzalcoatlus has been suggested to be a scavenger of large dinosaur carcasses. Other workers have argued that, because of the richness of trace fossils and burrows in the Quetzalcoatlus horizon, large azhdarchids probed for invertebrate infauna. Both these suggestions ignore the long, rigid neck of azhdarchids; modern avian scavengers and probers have flexible necks able to probe deep into carcasses or investigate narrow burrows, suggesting that these lifestyles are not viable for azhdarchids. Most authors currently cite aerial fishing as the most likely method of azhdarchid feeding, plucking fish and other pelagic organisms from bodies of water by either dip feeding or skimming. However, the structure of the neck also makes these interpretations problematic. Modern birds that partake in dip feeding have short, flexible necks capable of curving ventrally to minimize drag and resistance when submerging their jaws into the water during feeding. The neck also articulates with the skull at an angle slightly below horizontal, facing the long-axis of the skull essentially forwards. The long, stiff necks and strongly downturned skulls of azhdarchids bear little resemblance to their analogues in modern dip-feeders and would surely impair their ability to feed in this manner. Skim-feeding like that seen in the modern bird Rhyncops is also unlikely; although laterally compressed, the mandibles of azhdarchids do not show the extreme degree of lateral compression seen in modern skimmers; nor do they possess the horizontally orientated skulls or short, strong neck of skimming animals. A model for azhdarchid feeding that complies more with their anatomy is that azhdarchids acted in a similar manner to some modern storks, obtaining most of their food when foraging terrestrially or wading (Fig. 7). Such a lifestyle explains many otherwise problematic anatomical features of azhdarchid skeletons. The relatively long hindlimbs position the torso sub-horizontally, extending the neck forward along a horizontal plane, holding the skull close to the ground. With a sub-horizontally held neck and ventrally-orientated occipital condyle the head is naturally directed downwards, a useful Fig. 7. A Quetzalcoatlus seizes an unwary crocodile in a Cretaceous swamp. Whether azhdarchids fed in this manner is controversial, but many aspects of their skeleton imply that wading was certainly a potential feeding option. 37

adaptation for large animals feeding at ground level. The elongate skull positions the jaw tips close to feeding level, whilst the peculiar stiffness of the neck may have served to minimize the energy requirements to hold the head in feeding position. The length of the neck may have allowed azhdarchids to extend their feeding range into areas where they could otherwise not reach such as the deeper waters of rivers or lakes. It is not inconceivable to imagine azhdarchids standing in shallow water or along riverbanks with open jaws suspended in the water column, snagging fish and other forms as they swam between the jaws or perhaps even collecting prey whilst walking across plains and woodlands. The straight nature of azhdarchid beak morphology suggests that they did not specialize in one food item and, as with storks, a vast range of small vertebrates may have been eaten. More work is needed before any of these hypotheses can be verified but, on a theoretical level at least, the latter idea seems most satisfactory. A dying breed? The azhdarchids are among some of the last pterosaurs known. It seems that pterosaur diversity declined towards the end of the Mesozoic and, by the latest Cretaceous, only the azhdarchids remained in any abundance (a few other pterosaur lineages may have been present at this time, but their remains are rare). It has been suggested that the apparent decline of pterosaurs may merely be a taphonomic artefact, perhaps instigated by a shift of pterosaur habit from marine environments towards continental settings. These environments are less likely to be preserved in the geological record and do not provide many settings conducive to fossilization, possibly hiding true Late Cretaceous pterosaur diversity. The disappearance of azhdarchids from the fossil record at the end of the Mesozoic is often attributed to their size: larger animals are more prone to extinction through slower rates of reproduction and higher energy requirements. The extinction of the azhdarchids and pterosaurs may be attributed to competition from birds, a group which appear to have gradually replaced pterosaurs in a variety of roles throughout the Cretaceous. As the continental environments that the azhdarchids had apparently become well adapted-to changed at the termination of the Mesozoic it appears that birds and not azhdarchids possessed the necessary diversity and numbers to progress into the Tertiary. No flying animals before or since the azhdarchids have managed to capture their size or grandeur; although we now only have fossils as evidence of their presence, they clearly demonstrate that azhdarchids must have been fantastic sights to behold. Suggestions for further reading Chatterjee, S. & Templin, R.J. 2003. Posture, Locomotion and Palaeoecology of the Pterosaurs. Geological Society of America Special Publication, v.376, pp.1 64. Kellner, A.W.A. & Langston, W. Jr. 1996. Cranial remains of Quetzalcoatlus (Pterosauria, Azhdarchidae) from Late Cretaceous sediments of Big Bend National Park. Journal of Vertebrate Paleontology, v.16, pp.222 231. Langston, W. Jr. 1981. Pterosaurs. Scientific American, v.244, pp.92 102. Frey, E., Buchy, M.-C. & Martill, D.M. 2003. Middleand bottom-decker Cretaceous pterosaurs: unique designs in active flying vertebrates. In: Buffetaut, E. & Mazin, J.-M. (eds), Evolution and Palaeobiology of Pterosaurs. Geological Society Special Publication, v.217, pp.267 274. Martill, D.M. & Frey, E. 1998. Discovery of the holotype of the giant pterosaur Titanopteryx philadelphiae Arambourg 1959, and the status of Arambourgiania and Quetzalcoatlus. Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen, v.207, pp.57 76. Ösi, A., Weishampel, D.B. & Jianu, C.M. 2005. First evidence of azhdarchid pterosaurs from the Late Cretaceous of Hungary. Acta Palaeontologica Polonica, v.50, pp.777 787. Unwin, D.M. 2003. On the phylogeny and evolutionary history of pterosaurs. In: Buffetaut, E. and Mazin, J. M. (eds), Evolution and Palaeobiology of Pterosaurs, Geological Society Special Publication, v.217, pp.139 190. Unwin, D.M. 2005. The Pterosaurs from Deep Time. Pi Press, New York. 38