2. Geographical and geological setting

As noted by Bunzel (Reference Bunzel1871), the first vertebrate element discovered in the coal-bearing beds at Muthmannsdorf, west of Wiener Neustadt (Lower Austria), was a dinosaur tooth found by Ferdinand Stoliczka in 1859, during an excursion led by Eduard Suess. Subsequently, ‘Prof. Suess was so fortunate as to obtain the assistance of Bergverwalter Pawlowitsch in conducting excavations’ in the coal seam, which resulted in the discovery of ‘a perfect cemetery of the remains of Cretaceous reptiles’ (Seeley, Reference Seeley1881, p. 621).

The vertebrate-bearing beds at Muthmannsdorf, in the eastern part of the Calcareous Alps, belong to the Gosau Group, a succession of Upper Cretaceous to Palaeocene mainly marine sediments (see Summesberger, Machalski & Wagreich, Reference Summesberger, Machalski and Wagreich2007 for a recent synthetic section of the Gosau Group). Although Suess (Reference Suess1881) initially thought they were older than Turonian, it is now clear that the vertebrate-bearing beds are in fact Campanian (Plöchinger, Reference Plöchinger1961; Thenius, Reference Thenius1974; Summesberger, Machalski & Wagreich, Reference Summesberger, Machalski and Wagreich2007). In current stratigraphic nomenclature (Summesberger et al. Reference Summesberger, Wagreich, Tröger and Scholger2000), they are placed in the Grünbach Formation of the Lower Gosau Subgroup. The Grünbach Formation is referred to the early Campanian on the basis of microfossils (foraminifera, spores and pollen) and of marine fossils in the underlying and overlying formations (see Kvaček & Herman, Reference Kvaček and Herman2004 for a recent review). The vertebrate fauna from the Grünbach Formation includes chelonians, squamates, choristoderes, crocodilians, dinosaurs (ornithopods, ankylosaurs and theropods) and pterosaurs (Sachs & Hornung, Reference Sachs and Hornung2006).

3. Previous interpretations of the pterosaurian material from Muthmannsdorf

As noted by Seeley (Reference Seeley1881) and Wellnhofer (Reference Wellnhofer1980b), the pterosaur material from Muthmannsdorf consists of fragmentary phalanges (which are too incomplete to be of any systematic use), the proximal part of a humerus and the left articular region of a lower jaw.

Seeley, who consistently misspelled Bunzel's name as ‘Bünzel’, used the name Ornithocheirus Bünzeli for this material (and was followed in this by Sauvage, Reference Sauvage1882 and Newton, Reference Newton1888); according to the current rules of zoological nomenclature, the specific name should therefore be buenzeli. Wellnhofer (Reference Wellnhofer1978, Reference Wellnhofer1980b) suggested that it should be emended to Ornithocheirus bunzeli, but, as remarked by Sachs and Hornung (Reference Sachs and Hornung2006), this emendation is not acceptable under the ICZN. Wellnhofer (Reference Wellnhofer1980b) also indicated that the name proposed by Seeley (Reference Seeley1881, p. 701) only applies to the lower jaw fragment, presumably because Seeley considered it as ‘the only specimen of importance’ and specifically referred to the illustration provided by Bunzel (Bunzel, Reference Bunzel1871, pl. VI, figs 6, 7).

Seeley did not clearly explain why the pterosaur material from Muthmannsdorf was referred to the genus Ornithocheirus, and only mentioned that ‘this jaw seems to be well distinguished from the species already described’ (Seeley, Reference Seeley1881, p. 701). Nopcsa (Reference Nopcsa1926) merely confirmed the identification of the specimen as pterosaurian. Wellnhofer referred to the jaw fragment as ‘Ornithocheirus’ bunzeli and placed it in the Ornithocheiridae incertae sedis (Wellnhofer, Reference Wellnhofer1978) or ‘Ornithocheiridae von unsicherer systematischer Stellung’ (Wellnhofer, Reference Wellnhofer1980a). In his detailed revision of the material from Muthmannsdorf, Wellnhofer (Reference Wellnhofer1980b, p. 102) referred to the jaw fragment as Ornithocheirus bunzeli and to the humerus as Ornithocheirus sp., but remarked that the genus Ornithocheirus is poorly characterized and concluded that ‘aufgrund der oben gegebenen Charakterisierung ist eine eindeutige Zuordnung der vorliegenden Pterosaurierreste aus der Gosau zur Gattung Ornithocheirus nicht möglich’ (‘on the basis of the characters given above, an unequivocal attribution of these pterosaur remains from the Gosau to the genus Ornithocheirus is not possible’). Nevertheless, Wellnhofer (Reference Wellnhofer1991, p. 116) referred to the jaw fragment from Muthmannsdorf as Ornithocheirus bunzeli and to the humerus as Ornithocheirus. Jianu, Weishampel & Şticuă (Reference Jianu, Weishampel and Şticuă1997) noted that the humerus did not appear to belong to a pteranodontid, because it lacks a warped deltopectoral crest and a caudally directed ulnar crest, and considered that it may be a member of the Nyctosauridae on the basis of a supposedly hatchet-shaped deltopectoral crest.

Buffetaut (Reference Buffetaut2008, p. 254) briefly mentioned the pterosaur material from Muthmannsdorf and noted that ‘the shape of the proximal articular area and deltopectoral crest of the humerus from Muthmannsdorf would certainly not preclude an attribution to the Azhdarchidae’.

In our revision of the pterosaur material from Muthmannsdorf, we have considered the humerus and the jaw fragment separately. They are unlikely to be from the same individual, as the jaw fragment seems to be from a larger animal than the humerus. As noted by Seeley (Reference Seeley1881) and Wellnhofer (Reference Wellnhofer1980b), the phalanx fragments are too incomplete to be identified beyond Pterosauria and are not discussed here. Photographs of the humerus and jaw fragment are published here for the first time (Figs 1, 2); for excellent drawings of this material, see Wellnhofer (Reference Wellnhofer1980b).

Figure 1. Proximal part of a right azhdarchid humerus (UWPI 2349/102) from the Lower Campanian Grünbach Formation of Muthmannsdorf (eastern Austria). (a) anterior; (b) dorsal; (c) posterior; (d) posteroventral; (e) proximal views. Abbreviations: ch – caput humeri; dpc – deltopectoral crest; mp – medial process; pf ? – pneumatic foramen (?).

Figure 2. Pterodactyloidea indet. distal lower jaw fragment (UWPI 2349/101) from the Lower Campanian Grünbach Formation of Muthmannsdorf (eastern Austria). (a) dorsal; (b) medial; (c) lateral views. Abbreviations: dmdm – depression for musculus depressor mandibulae; f – foramen; gl – glenoid; lcg – lateral cotyle of the glenoid; mcg – medial cotyle of the glenoid; r – ridge; rp – retroarticular process.

4. The humerus

Contrary to Seeley's opinion, who thought that the humerus (UWPI 2349/102) was ‘of no importance’ to the anatomist (Seeley, Reference Seeley1881, p. 701), this specimen (Fig. 1), although it has undergone crushing and lacks the distal part, shows a number of characters that allow a relatively precise systematic assignment to the family level. Despite some deformation, the spatial arrangement of the caput humeri and deltopectoral crest is well preserved, as already noted by Wellnhofer (Reference Wellnhofer1980b). In proximal view the caput humeri is regularly rounded, with a roughly oval outline. Ventrally, the articular surface is clearly separated from the more distal part of the bone by a well-marked step-like sulcus. Dorsally, the edge of the articular surface strongly overhangs the shaft; although this has probably been exaggerated by compression, it is a characteristic feature of the proximal articular region. The deltopectoral crest lacks its tip, but its insertion on the shaft is well preserved and shows that it is broad and unwarped. More distally, it is gently curved, with more or less parallel sides, and there is no indication that it was ‘hatchet-shaped’ as in nyctosaurids, contrary to the claim of Jianu, Weishampel & Şticuă (Reference Jianu, Weishampel and Şticuă1997), which may have been based on Wellnhofer's (Reference Wellnhofer1980b) reconstruction that showed a hypothetical slightly expanded tip of the deltopectoral crest. The specimen actually does not support this interpretation and rather suggests that the tip of the crest had a rounded outline. The medial process is incomplete, but it can be seen that it was at right angles to the deltopectoral crest in proximal view. Although this whole area is poorly preserved, what is possibly a pneumatic foramen can be seen on the ventral surface of the bone, between the caput humeri and the proximal inception of the deltopectoral crest. Although a pneumatic foramen in such a position is frequently considered an azhdarchoid feature (Unwin, Reference Unwin, Buffetaut and Mazin2003; Witton, Martill & Green, Reference Witton, Martill and Green2009), inter- as well as intraspecific variability in pneumatic features is a well-known phenomenon among birds (O'Connor, Reference O'Connor2004) and has also been reported in Rhamphorhynchus (Ősi & Prondvai, Reference Ősi and Prondvai2009), so the significance of this character should perhaps not be overestimated.

As already noted by Wellnhofer (Reference Wellnhofer1980b), the humerus from Muthmannsdorf is clearly different from those of Pteranodon and Nyctosaurus. The deltopectoral crest of Nyctosaurus (and other nyctosaurids, Frey et al. Reference Frey, Buchy, Stinnesbeck, González González and Di Stefano2006) shows a characteristic hatchet-like shape, viz. it has concave proximal and distal margins (Williston, Reference Williston1903; Price, Reference Price1953), of which there is no indication in the Austrian specimen. In Pteranodon, the deltopectoral crest is notably warped (Bennett, Reference Bennett2001), unlike that of the Muthmannsdorf humerus. A warped deltopectoral crest is an ornithocheiroid character (Unwin, Reference Unwin, Buffetaut and Mazin2003). The unwarped crest of the Muthmannsdorf specimen thus shows that it cannot be placed among the ornithocheiroids. Wellnhofer (Reference Wellnhofer1980b) noted similarities between the specimen from Muthmannsdorf and a humerus from the Cambridge Greensand described and illustrated by Owen (Reference Owen1859, pl. 3, figs 14, 15). Contrary to Wellnhofer's (Reference Wellnhofer1980b) statement, Owen did not refer to this bone as Ornithocheirus sp. (in 1859, the name Ornithocheirus had not yet been erected by Seeley, Reference Seeley1869). Although Seeley later referred much (Seeley, Reference Seeley1869) and then practically all (Seeley, Reference Seeley1870) of the pterosaur material from the Cambridge Greensand to Ornithocheirus, a recent revision (Unwin, Reference Unwin2001) has shown that the pterosaur assemblage from the Cambridge Greensand is in fact much more diverse than previously recognized, and that taxa other than ornithocheirids (such as lonchodectids and pteranodontids) also occur. Although not mentioned by Unwin (Reference Unwin2001), azhdarchids may be present in the Cambridge Greensand (E. Buffetaut, unpub. data). To judge from the unwarped deltopectoral crest shown on Owen's figure, the humerus from the Cambridge Greensand mentioned by Wellnhofer, which does resemble the specimen from Austria, certainly does not belongs to an ornithocheirid.

The characters visible on the humerus from Muthmannsdorf are strongly reminiscent of those of azhdarchid humeri. Features in common include the unwarped deltopectoral crest and the thick, oval caput humeri, which overhangs the dorsal side of the shaft and is separated from its ventral side by a well-marked step-like sulcus. Although the unwarped crest appears to be the plesiomorphic condition in pterosaurs, the general morphology of the proximal part of the humerus in azhdarchids, including the caput humeri, the deltopectoral crest and the medial process, seems to be characteristic of that group of pterosaurs. It has been described or illustrated in Quetzalcoatlus from the Maastrichtian of Texas (Lawson, Reference Lawson1975), Hatzegopteryx thambema from the Maastrichtian of Romania (Buffetaut, Grigorescu & Csiki, Reference Buffetaut, Grigorescu, Csiki, Buffetaut and Mazin2003), two azhdarchids from the Late Cretaceous of Montana (an unnamed large form, Padian & Smith, Reference Padian and Smith1992; and Montanazhdarcho minor, McGowen et al. Reference McGowen, Padian, De Sosa and Harmon2002), an azhdarchid from the Campanian of Alberta (Godfrey & Currie, Reference Godfrey, Currie, Currie and Koppelhus2005) and an azhdarchid from the Late Cretaceous of Provence (Buffetaut, Mechin & Mechin-Salessy, Reference Buffetaut, Mechin, Mechin-Salessy and Csiki2006). On the basis of these similarities, it seems reasonable to refer the humerus from Muthmannsdorf to the Azhdarchidae. An attribution to the Tapejaridae, which are often considered as the sister-group of Azhdarchidae within Azhdarchoidea, although different interpretations have also been put forward (e.g. Witton, Martill & Green, Reference Witton, Martill and Green2009), could also be envisioned. Few details have been published about the morphology of the humerus in tapejarids, although Kellner (Reference Kellner2004) noted that the medial process is massive. According to Lü & Yuan (Reference Lü and Yuan2005), in Huaxiapterus jii, the deltopectoral crest is elongate with a long axis roughly perpendicular to the long axis of the shaft, a rounded extremity and more or less parallel margins; in these respects, the humerus is similar to that of azhdarchids. Our observations on tapejarid humeri, including a specimen identified as Tapejara wellnhoferi (SMNK – PAL 1137) and one belonging to an indeterminate tapejarid (SMNK – PAL 3856 (2./4)), show that the proximal part is indeed very similar to that of azhdarchids, notably in the shape and orientation of the deltopectoral crest. However, some differences can be seen in the morphology of the caput humeri, which in tapejarids is less expanded dorsoventrally, less prominent dorsally and not separated by a distinct step-like sulcus from the ventral surface of the bone. In these respects, the specimen from Muthmannsdorf is more reminiscent of azhdarchids than of tapejarids.

Comparison with the humeri of thalassodromids and chaoyangopterids, which have also been considered as closely related to azhdarchids (Witton, Martill & Green, Reference Witton, Martill and Green2009), is difficult because the humeri of these pterosaurs have not been described in great detail: most known chaoyangopterid specimens, from the Early Cretaceous of China, are crushed (Witton, Martill & Green, Reference Witton, Martill and Green2009). In the chaoyangopterid Shenzhoupterus chaoyangensis, the deltopectoral crest is described as markedly elongate, with a rounded distal termination (Lü et al. Reference Lü, Unwin, Xu and Zhang2008), which suggests some similarity with azhdarchids and tapejarids. The humerus of the thalassodromid Tupuxuara, as figured by Unwin (Reference Unwin, Buffetaut and Mazin2003) and Witton, Martill & Green (Reference Witton, Martill and Green2009), bears a deltopectoral crest that is less displaced distally, and a caput humeri that seems to be less ventrodorsally expanded, than in the Muthmannsdorf specimen, and in azhdarchids generally. Similarly, the ‘Wessex humerus’ from the Early Cretaceous of England, considered as belonging to a thalassodromid or a chaoyangopterid (Witton, Martill & Green, Reference Witton, Martill and Green2009), shows a less distally displaced deltopectoral crest and a less expanded caput humeri than azhdarchids and the Muthmannsdorf humerus. In these respects, the Austrian specimen appears to be closer to azhdarchids than to thalassodromids and chaoyangopterids.

It should also be mentioned that tapejarids, thalassodromids and chaoyangopterids have so far not been reported from the Late Cretaceous of Europe, whereas azhdarchids are known from many European localities of that age (in France, Spain, Hungary, Romania and Russia; see Barrett et al. Reference Barrett, Butler, Edwards and Milner2008).

5. The jaw fragment

The left articular region of a pterosaur jaw (UWPI 2349/101) from Muthmannsdorf, originally described as Lacerta by Bunzel (Reference Bunzel1871), and then as Ornithocheirus Bünzeli by Seeley (Reference Seeley1881), was redescribed in great detail by Wellnhofer (Reference Wellnhofer1980b), who interpreted it in functional terms. There is no need to provide here one more detailed description of this well preserved, although incomplete, specimen (Fig. 2). Instead, we shall focus on its distinctive characters, which may provide evidence concerning its systematic position. As noted by Wellnhofer (Reference Wellnhofer1980b), one of the important features of the specimen is the shape of the fossa articularis mandibulae, which is divided into two cotyles by an oblique ridge extending in a posterolateral to anteromedial direction. This peculiar structure has implications for the opening of the jaws and has been interpreted as resulting in a wide gape (Wellnhofer, Reference Wellnhofer1980b, but see Bennett, Reference Bennett2001 for a different interpretation). Its systematic significance is even more uncertain, because this oblique ridge is found in various, not closely related groups of pterosaurs. Wellnhofer (Reference Wellnhofer1980b) noted that it is present in Pteranodon (see also Bennett, Reference Bennett2001) and reported it in a jaw fragment referred to Ornithocheirus from the Cambridge Greensand (see also Wellnhofer, Reference Wellnhofer1985 for a description of a helical jaw joint in an ornithocheirid from Brazil). However, this type of ‘helical’ jaw joint is also known in Quetzalcoatlus (Kellner & Langston, Reference Kellner and Langston1996). In Hatzegopteryx thambema, from the Maastrichtian of Transylvania, the articular is not known, but the quadrate articulation shows two condyles separated by an oblique groove (Buffetaut, Grigorescu & Csiki, Reference Buffetaut, Grigorescu, Csiki, Buffetaut and Mazin2003), which indicates that the fossa articularis mandibulae must have been similar in structure to that of Quetzalcoatlus. Williston (Reference Williston1902) described a similar condition in Nyctosaurus. Since the helical jaw joint occurs in pteranodontids, ornithocheirids, nyctosaurids and azhdarchids, and in fact appears to be rather common in pterosaurs (Wellnhofer, Reference Wellnhofer1985), its systematic significance is limited. Moreover, not all azhdarchids show this type of articulation. In Bakonydraco galaczi, from the Santonian of Hungary, the morphology of the fossa articularis mandibulae is different from that of the Muthmannsdorf specimen. In the Hungarian form, the glenoid fossa is mediolaterally enlarged and S-shaped but is not separated into lateral and medial cotyles (Ősi, Weishampel & Jianu, Reference Ősi, Weishampel and Jianu2005), so the helical type of jaw articulation is not as developed as in Quetzalcoatlus and the Austrian specimen. An additional difference is that in Bakonydraco galaczi, the fossa depressoria, on the dorsal side of the retroarticular process, is not as elongate and deep as that of the fragment from Muthmannsdorf (as described by Wellnhofer, Reference Wellnhofer1980b). It thus appears that the jaw mechanics of the Hungarian and Austrian forms were different.

In conclusion, the morphology of the jaw articulation in the Muthmannsdorf specimen does not preclude its inclusion in the Azhdarchidae, nor does it demonstrate it. Other groups of pterosaurs, such as pteranodontids and nyctosaurids, show a similar kind of jaw articulation, and some azhdarchids, such as Bakonydraco, do not exhibit it.