Paläontologische Zeitschrift

, Volume 87, Issue 2, pp 269–289

New information on body size and cranial display structures of Pterodactylus antiquus, with a revision of the genus

Research Paper

DOI: 10.1007/s12542-012-0159-8

Cite this article as:
Bennett, S.C. Paläontol Z (2013) 87: 269. doi:10.1007/s12542-012-0159-8

Abstract

Pterodactylus antiquus has long been thought to have been quite small (~50 cm wingspan) and to have differed from P. longicollum, Ctenochasma, Germanodactylus, and Gnathosaurus in lacking a bony cranial crest, though a soft tissue crest and occipital lappet have been described. This article describes a new specimen of P. antiquus larger than all previously known specimens, which demonstrates that the species exceeded 1 m in wingspan and had a low bony cranial crest. A smaller, incipient crest was identified on the holotype specimen. Additional specimens, including the counterpart of Wellnhofer’s original occipital lappet specimen, provide evidence of the occipital lappet and the soft tissue crest extending upward above the naso-antorbital fenestra and orbit. In order to provide a proper taxonomic context for the findings, the recent synonymization of the species Pterodactylus antiquus and P. kochi on the basis of shared correlation of tooth number and skull length despite perceived differences in dentition and skull, neck, and trunk proportions is reviewed. A measurement error that had made it appear that P. antiquus differed significantly from P. kochi in proportions is documented, and after correction of the measurement error and reevaluation of the dental evidence there are no significant differences between the two nominal species. Thus, the synonymization of P. antiquus and P. kochi was appropriate, and a revised diagnosis is presented. In addition, the species P. longicollum and P. micronyx, which for some years have been viewed as not congeneric with P. antiquus, are placed in a new genus and transferred to Aurorazhdarcho, respectively.

Keywords

Pterodactylus Pterodactylidae Pterosauria Jurassic Solnhofen Limestone 

Kurzfassung

Für eine lange Zeit Pterodactylus antiquus schien recht klein (~50 cm Spannweite) zu sein, und unterschied es sich von P. longicollum, Ctenochasma, Germanodactylus und Gnathosaurus im Fehlen einer knöchernen Sagitalcrista, obwohl ein Weichteil-Kamm und Hinterhauptslappen beschrieben worden sind. Dieser Beitrag beschreibt ein neues Exemplar von P. antiquus, größer als alle bisher bekannten Exemplare, und zeigt, daß die Arten 1 m Spannweite überschritten und hatte einen geringen knöchernen Sagitalcrista. Eine kleinere, beginnende Sagitalcrista wurde auf der Holotypus identifiziert. Andere Exemplare, darunter die Gegenplatte des ursprünglichen Hinterhauptslappen von Wellnhofer, zeigen Spuren der Hinterhauptslappen und das Weichteil-Kamm nach oben verläuft oberhalb des Nasopraeorbitalöffnung und Orbit. Um eine korrekte taxonomische Kontext für den Erkenntnissen, das Synonymization der Arten Pterodactylus antiquus und P. kochi auf der Grundlage eine gemeinsamer Korrelation von Zahn Anzahl und Schädellänge trotz der Unterschiede in Gebiss und Schädel-Hals-Rumpf Proportionen wird neu bewertet. Ein Messfehler machte es scheinen, daß die Proportionen von P. antiquus signifikant unterschieden sich von denen von P. kochi, und nach der Korrektur des Messfehlers und Neubewertung der Form und Anzahl der Zähne gibt es keine signifikanten Unterschiede zwischen den beiden nominellen Arten. Deshalb war das Synonymization von P. antiquus und P. kochi richtig, und eine neue Diagnose wird vorgestellt. Seit einigen Jahren haben sich die Arten P. longicollum und P. micronyx nicht als artverwandte mit P. antiquus angesehen worden. Deshalb wird P. longicollum in einer neuen Gattung gesetzt, und P. micronyx wird zu Aurorazhdarcho übertragen.

Schlüsselwörter

Pterodactylus Pterodactylidae Pterosaurier Jura Solnhofener Plattenkalk 

Introduction

The first pterosaur known to science was a complete skeleton that is now the holotype of Pterodactylus antiquus, discovered in the Upper Jurassic Solnhofen Limestone near Eichstätt, Germany. When first described by Collini (1784) it was not recognized as a flying animal or a reptile, and it fell to Cuvier (1801) to properly interpret it as a reptile and a flier, and to name it a ‘pterodactyle.’ Over the years, many more pterosaur specimens of multiple genera have been found in the Solnhofen Limestones (Wellnhofer 1970, 1975), but P. antiquus and the similar P. kochi became the archetypal Jurassic pterodactyloid pterosaurs, small and uncrested.

Wellnhofer (1970) presented the first thorough taxonomic revision of the Solnhofen pterodactyloid fauna, interpreting 14 Pterodactylus species names as junior synonyms and concluding that there were six valid species of Pterodactylus (P. antiquus, P. kochi, P. elegans, P. micronyx, P. longicollum, and P. suevicus) as well as four other valid species in the genera Germanodactylus, Gnathosaurus, and Ctenochasma represented by large specimens with skulls. Wellnhofer (1970) identified 5 specimens of P. antiquus and 23 of P. kochi, and distinguished between the two species on the basis of tooth number and shape, and skull, neck, and trunk proportions. Bennett (1995, 1996a) described the presence of year classes in pterosaurs from the Solnhofen Limestone and argued that the immature specimens assigned to the small species P.antiquus, P. kochi, P. micronyx, and P. elegans were juveniles of species represented by the large mature specimens assigned to P. longicollum, Germanodactylus, Gnathosaurus, and Ctenochasma. I attempted to match up the small and large species and later predicted that mature specimens of all Solnhofen pterodactyloid species would have bony cranial crests (Bennett 2002), but I put off formal taxonomic revision pending a thorough restudy of the pertinent specimens.

Subsequently, my suggestion that P. elegans, Ctenochasma gracile, and C. porocristata pertained to a single species, C. elegans, was implemented by Jouve (2004; see also Bennett 2007) based on his demonstration that tooth number was proportional to skull length in Solnhofen pterodactyloids, and juveniles of Germanodactylus cristatus were identified in the samples of P. kochi and P. micronyx, which prompted a revision of Germanodactylus (Bennett 2006). I also noted in an abstract (Bennett 2003) that measurement errors had produced the appearance of proportional differences between P. antiquus and P. kochi, that specimens assigned to P. kochi pertained to P. antiquus, and that P. longicollum was not conspecific with P. antiquus and needed a new genus name.

Jouve (2004) demonstrated that specimens assigned to P. antiquus and P. kochi fall along the same regression line when tooth number is plotted against skull length and on that basis implemented my suggestion that P. kochi was a junior synonym of P. antiquus. However, the synonymization seemed premature in that he did not address the apparent differences in skull, neck, and trunk proportions between the two nominal species and so left open the possibility that the tooth number to skull length relationship characterized the genus Pterodactylus, whereas P. antiquus and P. kochi were distinct species characterized by differing skull, neck, and trunk proportions. Perhaps because of this or because of ignorance of Jouve’s (2004) revision, some have continued to view P. antiquus and P. kochi as distinct species (e.g., Wang et al. 2008).

This article started out as a short note to document the measurement error that had made it appear that P. antiquus and P. kochi differed in skull, neck, and trunk proportions and by so doing to dispel any possible confusion surrounding Jouve’s (2004) synonymization of P. antiquus and P. kochi. However, in the course of my restudy of Solnhofen pterodactyloids, new information about body size and cranial display structures of Pterodactylus was found. A large isolated skull demonstrates that Pterodactylus antiquus reached larger sizes than previously thought and had a low bony cranial crest; other specimens provide additional information about the soft tissue crest and occipital lappet of P. antiquus. Given the new information, a revised diagnosis of the species is warranted. Therefore, this article presents the new information about body size and cranial display structures, reviews the synonymy of P. antiquus and P. kochi, and then presents a revision of the genus with a revised diagnosis of P. antiquus and transfers P. micronyx and P. longicollum to other genera.

Institutional Abbreviations: AFGM, Fryxell Geology Museum, Augustana College, Rock Island, IL; AMNH, American Museum of Natural History, New York; BMMS, Burgermeister Müller Museum, Solnhofen; BSP, Bayerische Staatssammlung für Paläontologie und Geologie, Munich; CMNH, Carnegie Museum of Natural History, Pittsburgh, PA; JME-SOS, JuraMuseum (Solnhofen Sammlung), Eichstätt; MB.R., Humboldt Museum, Berlin; MCZ, Museum of Comparative Zoology, Harvard University, Cambridge, MA; NHMUK, Natural History Museum, London, UK (formerly BMNH); NHMW, Naturhistorisches Museum Wien; NMB, Natural History Museum Basel, Basel, Switzerland; PMU, Palaeontological collections, Museum of Evolution, Uppsala University, Uppsala, Sweden; PTH, Philosophisch-Theologische Hochschule, Eichstätt (its collections were taken over by the JuraMuseum); RM, Rijksmuseum van Geologie en Mineralogie, Leiden; SM, Senckenberg Museum, Frankfurt; SMNS, Staatliches Museum für Naturkunde, Stuttgart; SNSD-MMG, Senckenberg Naturhistorische Sammlungen Dresden, Museum für Mineralogie und Geologie; and TM, Teyler Museum, Haarlem.

Body size and cranial display structures of Pterodactylus antiquus

A new isolated skull, BMMS 7, is the largest known specimen of Pterodactylus antiquus and preserves a bony sagittal crest, which validates Bennett’s (2002) prediction of a bony crest in the species. A smaller incipient bony crest is also described from the holotype, BSP AS I 739. The previously undescribed counterpart of Wellnhofer’s (1970) original occipital lappet specimen, MCZ 1505, provides additional information about the extent of the soft tissue crest atop the skull and the occipital lappet, as does BSP 1929 I 18. These specimens are described as needed below.

BMMS 7

This specimen is a large incomplete skull preserved in the right lateral view (Figs. 1, 2) with the atlas-axis complex and cervical vertebra 3, recently collected from the Hienheim Beds, which are stratigraphically equivalent to the Mörnsheimer Limestone and so Malm Zeta 3 (Röper 2005). The skull is missing only the anterior ends of the upper and lower jaws. The preserved length from the broken tip of the premaxillae to the posterior margin of the occiput is 142 mm. The preserved length of the mandible from the broken tip of the dentary to the anterior margin of the jaw articulation is 119 mm, and although the mandible is preserved depressed at an angle of 45° and pulled 1–2 mm anteriorly such that the jaw articulation is slightly dislocated, if it were properly articulated and lying adjacent to the upper jaw, the preserved length of the skull from the broken tip of the dentary to the posterior margin of the occiput would be 149 mm. Thus, the skull as preserved exceeds the reconstructed skull length of the largest previously known specimen (~140 mm; Hofker, 1922; Wellnhofer, 1970) and is the largest known skull of Pterodactylus antiquus.
Fig. 1

Photograph of Pterodactylus antiquus, BMMS 7, in right lateral view. Scale bar 3 cm

Fig. 2

Line drawing of Pterodactylus antiquus, BMMS 7, in right lateral view with the missing tips of the upper and lower jaws reconstructed. Scale bar 3 cm. lac lacrimal, lram left ramus of mandible, man mandible, max maxilla, naof naso-antorbital fenestra, nas nasal, od oval depression, or orbit, pmx premaxilla, and sc sagittal crest

The skull is less crushed than is usual for Solnhofen pterodactyloids, probably because of its large size, and there has been minimal preparation of the specimen other than to prepare around its outlines. There are no visible sutures between skull bones, and the skull seems to have been fully fused. Thus, there is no evidence of immaturity, and the specimen was mature and probably at or approaching the maximum size for the species. The skull tapers evenly from the braincase region to its anterior end and has a nearly straight superior margin; the very slight concavity may result from lateral crushing of the skull, but does not approach that seen in specimens assigned to P. longicollum. The naso-antorbital fenestra (NAOF) is 59 mm long, and the nasal process is not visible within it, presumably because the matrix in the fenestra has not been excavated down to the midline of the skull. The anterior margin of the NAOF is crushed inward, and 10 mm anterior to it there is an shallow oval depression 17 mm long with a calcite infilling. There is no evidence of any opening within the depression, and it probably reflects the internal pneumatization of the rostrum by a paranasal sinus made apparent by compression. Parts of the premaxillae, nasal, lacrimal, and anterior ramus of the jugal are broken open, as are parts of the braincase region, revealing calcitic infillings. The quadrate and squamosal are badly damaged because of crushing. There is no trace of a sclerotic ring.

A low sagittal crest extends along the midline of the premaxillae above the posterior part of the NAOF and the orbit. It is 47.5 mm long and has a maximum height of 0.9 mm above the orbit (Fig. 3). The lateral surface of the crest has a smooth texture unlike the striated crests of P. longicollum, Ctenochasma, Germanodactylus, and Gnathosaurus, and although it is not possible to measure the crest’s thickness, it appears to be no more than 0.2 mm thick. There are no visible traces of soft tissue preservation associated with the crest, and examination of the specimen under long wave UV light failed to find any traces of soft tissues.
Fig. 3

Photograph and interpretive line drawing of part of the bony sagittal crest atop the orbital region of Pterodactylus antiquus, BMMS 7. In the drawing the crest is indicated by stippling, other cortical bone is indicated by gray shading, and places where the cortical bone is broken away are indicated by hatching. Scale bar 3 mm. fr frontal, la lacrimal, or orbit, pmx premaxilla, prf prefrontal, and sc sagittal crest

The upper jaw preserves evidence of 16 teeth, which are lettered A–P from front to back so as prevent confusion with actual tooth numbers. The first preserved tooth, Tooth A, lies immediately posterior to the broken end of the premaxillae and is preserved as only a root in the alveolus and an impression in the matrix. Tooth B, has a maximum width of 2.2 mm with a somewhat constricted base and is 3.4 mm tall. The subsequent teeth are progressively smaller, and Tooth H has a maximum width of 1.0 mm and a height of 2.0 mm. Teeth L–P are broken off in their alveoli; the last, Tooth P, is under the middle of the NAOF and 33.4 mm anterior to the jaw articulation.

The preserved portion of the mandible includes what seems to be a short section of the symphysis and presumably all of both rami although most of the left ramus is buried in matrix. What is visible is well preserved. The anterior tip of the mandible is broken at the level of the posterior wall of an alveolus on the right side, and there are 17 alveoli along the right ramus, each with the tooth broken off flush with the surface of the mandible. The last tooth is 32.6 mm anterior to the jaw articulation and so would occlude slightly behind maxillary Tooth P. The left side of the mandible preserves two unbroken teeth near its anterior end, but no other teeth or alveoli are visible. In both the upper and lower jaws there is no evidence of active tooth replacement, either in the form of replacement teeth forming behind a functional tooth or as an erupting functional tooth.

The atlas-axis complex and cervical vertebra 3 are poorly preserved. However, the length of cervical 3 (i.e., >23 mm) indicates that the pterosaur had an elongated neck.

BSP AS I 739

This specimen is the holotype of Pterodactylus antiquus (Wellnhofer 1970). Although unrecognized until BMMS 7 provided an indication of what to look for, this specimen has a small bony sagittal crest formed by the premaxillae above the orbit (Fig. 4). The maximum height of the crest is only 0.7 mm, and its length probably does not exceed 12 mm or 11 % of skull length, but anteriorly it blends into the rounded apex of the premaxilla, and so it is difficult to tell where it ends and the rounded apex begins. As in BMMS 7, the crest has a smooth lateral surface unlike the striated crests of P. longicollum, Ctenochasma, Germanodactylus, and Gnathosaurus, and it is probably was no more than 0.2 mm thick.
Fig. 4

Photograph and interpretive line drawing of the incipient bony sagittal crest atop the orbital region of Pterodactylus antiquus, BSP AS I 739. In the drawing the crest is indicated by stippling; other bone is indicated by light gray shading except for the median nasal process that is indicated by darker gray shading. Note that the crest grades into the rest of the premaxillae anterior to it, and there is no clear dividing line, hence the interfingering. Scale bar 3 mm. cb ceratobranchial, fr frontal, la lacrimal, na nasal, np nasal process, or orbit, pmx premaxilla, prf prefrontal, and sc sagittal crest

MCZ 1505

This specimen is the counterpart of BSP 1883 XVI 1, the specimen from which Wellnhofer (1970) first described the occipital lappet of Pterodactylus. MCZ 1505 was purchased from E. Häberlein in early 1884 by L. Agassiz, and although it has been on display in the MCZ for many years, it seems to have escaped the notice of pterosaurologists. MCZ 1505 and BSP 1883 XVI 1 exhibit so-called Sockel-Erhaltung (Barthel 1978), and based on MCZ 1505’s pedestals it is the lower slab, whereas BSP 1883 XVI 1’s sockets show it is the upper slab. Wellnhofer (1970) provided a description of BSP 1883 XVI 1.

MCZ 1505 preserves the specimen in right lateral view with the head and neck bent dorsally in an opisthotonic pose (Fig. 5), the fore and hindlimbs flexed more or less as if in terrestrial locomotion, the right hindlimb extended posteriorly, and the left flexed anteriorly. The right pes is missing as are the third and fourth phalanges of the right wing. The skull and many of the bones seem to have split longitudinally when the part and counterpart slabs were separated from one another, for the most part unequally such that the majority of the bone remained with BSP 1883 XVI 1. However, where the bone is absent MCZ 1505 preserves impressions of the bones. It also preserves impressions (atop pedestals) and traces of soft tissues, some traces as tan to reddish iron-oxide films visible under daylight illumination and others that fluoresce under long wave UV illumination. Note that BSP 1883 XVI 1 also preserves impressions and visible traces of soft tissues, but examination under UV illumination did not reveal any fluorescent traces.
Fig. 5

Photograph of Pterodactylus antiquus, MCZ 1505. Scale bar 5 cm

The impressions of the soft tissues associated with the skull consist of an impression of the occipital lappet extending posteriorly from the back of the skull and impressions of the neck tapering downward from the occiput and the area of the retroarticular process of the mandibular (Fig. 5). The impression of the neck is more pronounced behind the cervical series than that anterior to it, probably because it was more substantial: epaxial muscles versus trachea and esophagus. The ceratobranchials are preserved some distance below the mandible, seemingly in normal position, but there is no trace of a gular sac. The impression of the occipital lappet is ~16 mm long or ~14 % of skull length, and its width is irregular, ranging from ~2.6 mm proximally to ~3.5 mm over most of the remainder of its length (Fig. 6). The impression does not preserve any discernible structure.
Fig. 6

Photographs under visible light (above) and UV light (middle), and interpretive line drawing of the skull and soft tissue traces of Pterodactylus antiquus, MCZ 1505 in right lateral view. In the drawing bone is indicated by dark shading, the fluorescent aura is indicated by light shading, the occipital lappet is indicated by hatching, and the soft tissue crest is indicated by cross-hatching. Scale bar 3 cm. cb ceratobranchial, cm cervical musculature, fa fluorescent aura, naof naso-antorbital fenestra, pmx premaxilla, ol occipital lappet, sr sclerotic ring, and tc soft tissue crest

The skull is surrounded by manganese dendrites and fluorescent deposits. The dendrites are similar to those that are found on many fossils from the Solnhofen Limestone. The fluorescent deposits are of two different types. The first type, which might be termed a fluorescent aura, is fluorescence surrounding most of the skull that is diffuse and rather faint in most areas, but brighter along its irregular outlines. This aura seems to reflect organic matter that leached out the skull and seeped into surrounding sediments along cracks and bedding planes, and the brighter outlines reflect concentration of the fluorescent material by drying. The aura is above and below the rostrum and also around the occipital lappet. Above the lappet and in one place below the mandible there are intermediate bright margins suggesting that the aura resulted from multiple instances of seepage. The irregularity of the outlines indicates that they do not provide useful information about the morphology of the soft tissues of the head.

The second type of fluorescence is restricted to the roughly triangular area above the posterior half of the NAOF and the orbit. The fluorescence results from a somewhat diffuse mineral deposit similar to that seen in the preservation of the soft tissue crest of Germanodactylus (Bennett 2002). The deposit extends along ~44 mm or ~38 % of skull length and reaches a maximum height of 9.5 mm above the posterior sixth of the NAOF. If the margins anterior and posterior to the apex continued straight until reaching the top of the skull, then the crest would have been ~51 mm long or ~45 % of skull length.

BSP 1929 I 18

This specimen is the holotype of Döderlein’s (1929) Pterodactylus cormoranus and consists of an articulated skull, neck, and anteriormost part of the torso in left lateral view along with an isolated pes. The head and neck are bent dorsally in an opisthotonic pose and are accompanied by traces of soft tissues that occur as reddish brown iron-oxide staining on the limestone that did not fluoresce when examined under long wave UV illumination.

Comparison of Döderlein’s photograph with the specimen as it is today suggests that soft tissue traces have faded in the intervening years, or perhaps the specimen was originally photographed with filters that accentuated the traces. Döderlein (1929: 68) described soft tissue traces of the gular sac and neck, but also noted that there was a “fleischigen, hahnenkammartigen Auswuchs” (i.e., soft tissue crest) atop the skull above the orbits. The iron-oxide staining of this soft tissue crest was weaker than that of the gular sac and neck, but was bounded by a weak furrow and was readily apparent in the published photograph (Fig. 7).
Fig. 7

Photo of Pterodactylus antiquus, BSP 1929 I 18, reproduced from Döderlein (1929, Tafel 3). The arrows indicate the superior margin of the soft tissue crest before it was prepared away

The left side of the skull roof above the orbits was crushed dorsally presumably as a result of compression during preservation such that the lacrimal and nasal that formed the anterosuperior corner of the left orbit project above the superior margin of the midline of the skull anterior to the orbit and the frontal and parietal behind the orbit. The lacrimal and nasal thus project into the middle of the subtriangular trace of the soft tissue crest somewhat anterior to its apex (Fig. 7). Unfortunately, sometime after Döderlein’s photograph was taken someone chipped away the limestone matrix along the superior margin of the skull above the NAOF, orbit, and braincase, and in doing so destroyed most of the trace of the soft tissue crest. Only the apex of the trace of the crest is still preserved, although it is quite faint (Fig. 8a). The greatest height of the soft tissue crest above the skull roof was 3.1 mm, and the apex was slightly behind the midpoint of the orbit. The greatest length of the soft tissue crest based on Döderlein’s photograph is 20.6 mm, which represents ~25 % of the skull length.
Fig. 8

Photographs and interpretive line drawings of a the posterior skull and neck of Pterodactylus antiquus, BSP 1929 I 18, and b detail of the occipital lappet in left lateral view. In the drawing of a bone is indicated by dark shading, soft tissue impressions and red-brown iron-oxide deposits are indicated by light shading, and the short arrows indicate the anterior and posterior ends of the small apical region of the soft tissue crest that was not prepared away. The extent of the soft tissue crest from Döderlein’s photo (Fig. 7) is indicated by cross-hatching, and the occipital lappet is indicated by hatching. Scale bars 3 cm and 1 mm, respectively. Note that in b the hue and saturation of a color photograph were adjusted to improve the clarity of the filamentous structures before the image was converted to grayscale. asc anterosuperior corner of orbit crushed upward, cb ceratobranchial, cv cervical vertebrae, fi filaments, gs trace of gular sac; lighter trace surrounding filaments, naof naso-antorbital fenestra; occiput of skull, ol occipital lappet, or orbit, stc soft tissue crest, stn soft tissues of neck, tm tool mark, and tr trunk

The specimen also preserves a trace of an occipital lappet extending posteriorly and somewhat upward from the back of the skull near the contact of the parietals and supraoccipital. As with the soft tissue crest, preparation around the margin of the skull damaged part of the trace, and there is a deep gouge across what would have been the proximal part of the trace (Fig. 8b). What remains of the trace shows that it was ~11 mm long or ~13 % of skull length and ~1 mm wide. Within the trace are slender darker bands that seem to have been part of a loose bundle of perhaps 4–6 spiralling filaments. Between the filaments the trace is lighter. Presumably as a result of the spiralling filaments, the trace is not of consistent width or taper, but rather has wider and narrower parts and an irregular tip.

Synonymy of Pterodactylus antiquus and Pterodactylus kochi

Wellnhofer (1970) separated Pterodactylus antiquus and P. kochi on the basis of tooth number (P. antiquus, 20–25; P. kochi, 15–20), tooth shape (P. antiquus, flat conical with narrow base; P. kochi, flat conical with broader base), and the proportions of the skull, neck, and trunk (P. antiquus, skull and neck long relative to trunk; P. kochi, skull and neck shorter relative to trunk than in P. antiquus, but longer than in P. micronyx and P. elegans). Wellnhofer (1970) considered only five specimens to pertain to P. antiquus: the holotype, three other incomplete specimens equivalent in size to the largest specimens assigned to P. kochi, and one small specimen, TM 10341, half the size of the smallest large specimen. The teeth of TM 10341 are too poorly preserved to assess tooth number or shape, so the teeth of the specimens assigned to P. antiquus were only known for large specimens. In contrast, the teeth of the specimens assigned to P. kochi were known primarily from small specimens, and some of the information from large specimens assigned to P. kochi was unreliable. For example, Wellnhofer (1970) suggested that the largest specimen assigned to P. kochi (BSP 1883 XVI 1) had only ~16 teeth per jaw side; however, its skull was split down the middle when the part and counterpart slabs were separated, and the matrix was not prepared away down to the lateral side of the maxilla and dentary. As a consequence the small teeth at the posterior end of the tooth row could be seen, and so the specimen probably had more than ~16 teeth. The fact that tooth numbers of specimens assigned to P. antiquus could only be counted in large specimens whereas tooth numbers of P. kochi were known primarily from small specimens may explain the difference in tooth number noted by Wellnhofer (1970). Regardless, Jouve (2004) showed that specimens assigned to P. antiquus and P. kochi fall along the same regression line when tooth number is plotted against skull length; therefore, the suggestion that the nominal species differed in tooth number can be rejected. As for tooth shape, I have not been able to confirm any significant differences in the width of tooth bases (Bennett 1996a), but again it may be significant that tooth shape in specimens assigned to P. antiquus could only be assessed in large specimens whereas tooth shape in specimens assigned to P. kochi was known primarily from small specimens, so perhaps the shape differences noted by Wellnhofer (1970) are size-related.

Turning to the skull, neck, and trunk proportions, when I examined TM 10341 my measurements of the skull and most postcranial elements were within ~0.5 mm of Wellnhofer’s (1970) measurements. However, my measurements of neck and praecaudale Rumpfwirbelsäule (PCRW; = combined dorsal and sacral vertebrae) length were markedly different (Table 1) despite the facts that the neck and trunk are preserved more or less in a straight line in TM 10341 and I repeated the measurements multiple times. Although I have argued elsewhere that all pterosaurs had nine cervical vertebrae (Bennett 2004) whereas Wellnhofer (1970, 1978) viewed Pterodactylus as having seven, the differing measurements cannot be ascribed to differences in the identification of the transition point between the cervical and dorsal series because the sum of Wellnhofer’s lengths of the cervical vertebrae and PCRW is 5 mm less than the straight-line distance between the anterior end of the atlas and the posterior end of the sacrum. I must assume that Wellnhofer made an error in recording or transcribing the measurements.
Table 1

Comparison of measurements (in mm) of Pterodactylus antiquus, TM 10341

 

Skull length

Neck length

PCRW length

Neck + PCRW length

Wellnhofer (1970)

44

29.5

24.5

54.0

This paper

44.5

25.9

33.3

59.2

PCRW combined length of dorsal and sacral vertebrae

Wellnhofer’s (1970) measurements suggested that the skull of TM 10341 was 75 % longer and the neck 20 % longer than the PCRW, comparing favorably with the proportions of the holotype of P. antiquus, in which the skull is 71 % longer and the neck is 27 % longer than the PCRW, and differing markedly from small specimens assigned to P. kochi in which the skull is 20–40 % longer and the neck is 10–20 % shorter than the PCRW. Thus, the skull to neck to PCRW proportions may have appeared to support a referral of TM 10341 to P. antiquus. However, the corrected measurements show that the skull is only 34 % longer and the neck 22 % shorter than the PCRW, comparing well with the proportions of the similarly sized specimens assigned to P. kochi; this does not support the suggestion that the skull to neck to PCRW proportions of specimens assigned to P. antiquus are markedly different from those of similarly sized specimens assigned to P. kochi.

Pterodactylus length measurement data were plotted and analyzed to look for evidence that P. antiquus and P. kochi are not conspecific. Length measurements of the skull, neck, PCRW, humerus, radius, metacarpal IV, wing phalanges (WP) 1–4, femur, and tibia (from Wellnhofer 1970 except for the corrected neck and PCRW measures of TM 10341) for all specimens assigned to P. antiquus and P. kochi [Wellnhofer’s Exemplare 1–8, 10–28; Exemplar 9 was excluded as Germanodactylus cristatus (Bennett 2006)] are plotted in modified Nopcsa curves (Bennett 2006, 2007; see also Wiman 1925; Nopcsa 1923) based on the absolute measures of skeletal elements and not spaced out vertically on the graph (Fig. 9). Essentially the same data set (with neck and WP4 lengths excluded because of missing data) was subjected to principal components analysis (PCA) using the statistical software package SPSS 10.5 (SPSS Inc., Chicago, IL, USA), and the resulting components are plotted in Fig. 10.
Fig. 9

Modified Nopcsa curves of skeletal proportions of Pterodactylus antiquus based on measurement data from Wellnhofer (1970: Exemplare 1–8, 10–28) and Frey and Tischlinger (2000), but with corrected neck and PCRW measurements of TM 10341. The upper heavy line represents the holotype of P. antiquus (BSP AS I 739), the lower heavy line represents TM 10341 including the corrected neck and PCRW measurements, and the broken heavy line represents Wellnhofer’s (1970) neck and PCRW measurements. The letters and numbers on the X axis stand for skull, neck, PCRW (i.e., praecaudale Rumpfwirbelsäule = dorsal and sacral vertebrae), humerus, radius, metacarpal IV, wing phalanges 1–4, femur, and tibia length

Fig. 10

Plots of factors 1 versus 2 (a) and 2 versus 3 (b) from the principal components analysis of Pterodactylus antiquus measurement data from Wellnhofer (1970) with corrected measurements for TM 10341. The holotypes of P. antiquus (BSP AS I 739) and P. kochi (SM R404/BSP AS XIX 3) and the largest specimen (BSP 1883 XVI 1) referred to P. kochi by Wellnhofer (1970) are indicated

In the plot of modified Nopcsa curves (Fig. 9) similarly sized specimens clump together, and the plot permits comparison of the lengths of individual elements and the proportions of different specimens, and evaluation of the isometry or allometry of adjacent elements within a sample of a specimens. No significant differences in proportions, such as previously have been seen between specimens assigned to P. kochi, P. micronyx, and Germanodactylus cristatus (Bennett 2006), are present in the plot of P. antiquus and P. kochi, and thus there is no evidence that the two are not conspecific. Although the uncorrected neck and trunk measurements of TM 10341 produce a curve in which the pattern of the skull to neck to PCRW portion of the curve is quite unlike those of similarly sized specimens assigned to P. kochi and resembling those of large specimens assigned to P. antiquus, the pattern of the corrected measurements is similar to the patterns of the similarly sized specimens assigned to P. kochi.

The plot of modified Nopcsa curves reveals an allometric trend toward increasing skull and neck length relative to PCRW length with increasing size. In addition, they show neck length is more variable than other length measurements. Three possible explanations for this result are: (1) there may be considerable difficulty in correctly identifying the cervical-dorsal transition point (C7-D1 in Wellnhofer’s scheme) on a specimen because of the arrangement of the skeleton, deposition of calcite, etc.; (2) neck length is difficult to measure accurately in specimens in which the cervical series of strongly curved (e.g., BSP AS I 739, 1883 XVI 1); (3) it is possible that neck length was not as tightly constrained as the lengths of forelimb segments involved in the wing, and so it may have been more variable within and between Pterodactylus populations.

The PCA does not provide any evidence that more than one species is present in the sample. In the plot of factors 1 versus 2, two of the three specimens assigned to P. antiquus fall within the cloud of points of specimens assigned to P. kochi, which also includes the holotype of P. kochi (BSP AS XIX 3/SM R404). The holotype of P. antiquus (BSP AS I 739) is an outlier, but that presumably is a result of its large size and the fact that factor 1 largely reflects absolute size. The largest specimen assigned to P. kochi (BSP 1883 XVI 1) also is an outlier in the plot of factor 1 versus 2, and BSP AS I 739 again is an outlier in the plot of factors 2 versus 3, but as demonstrated by the analysis of Alligator (Bennett 2008) particularly large specimens may fall as outliers in a sample of a single species, and so they provide no evidence that more than one species in the sample. Note that in the case of BSP AS I 739 in the plot of factors 2 versus 3, the placement of the data point probably results from a humerus that is a little shorter than average, as adjusting the humerus length to a slightly longer value and repeating the analysis moves the BSP AS I 739 data point upward into the main point cloud.

Because the holotype of Pterodactylus kochi comes from the Mörnsheimer Limestone (Malm Zeta 3) of Kelheim, it is stratigraphically younger than the specimens from the Solnhofen Limestone (Malm Zeta 2) of Solnhofen, Eichstätt, and adjacent localities that Wellnhofer (1970) and others referred to P. kochi. Therefore, it is conceivable that it is not conspecific with the specimens from the Solnhofen Limestone referred to P. kochi, but instead represents a distinct species. However, given that no one has questioned the referral of most Solnhofen Pterodactylus specimens to P. kochi and that no differences in dentition or skeletal proportions can be demonstrated between the Solnhofen specimens referred to P. antiquus and those referred to P. kochi, there is no reason not to accept Jouve’s (2004) synonymization of P. kochi with P. antiquus. In the end, I suspect that the separation of P. antiquus from P. kochi was due in large part to the unusual attributes of the holotype of P. antiquus: its large size, unusual body position, and unusual quality of preservation.

Discussion

The new large specimen, BMMS 7, pertains to Pterodactylus antiquus because the skull exhibits an even taper from the orbit to its anterior end, straight upper and lower jaw margins, flat conical teeth that are progressively smaller posteriorly, more than 18 teeth per jaw side, and the posterior end of the tooth row extending under the middle of the NAOF, and because the elongate cervical vertebra 3 indicates that it had a long neck. It cannot pertain to P. longicollum because that taxon has a markedly concave superior margin of the skull, a maximum of ~15 teeth per jaw side, the posterior teeth are roughly equivalent in size to the anterior teeth, and the posterior end of the tooth row does not extend under the NAOF. It cannot pertain to Germanodactylus cristatus or G. rhamphastinus because those taxa have a maximum of ~16 teeth per jaw side, the posterior teeth larger and more robust than the anterior teeth, and a relatively short neck.

The skull length of BMMS 7 can be estimated by scaling up the proportions of complete large skulls. The largest known complete skull of P. antiquus is that of BSP 1883 XVI 1, in which the NAOF equals 26.6 % of the 113.5-mm skull length and the anterior margin of the NAOF lies at 46.5 % of skull length, respectively. In the next largest, the holotype BSP AS I 739, the NAOF equals 29.8 % of the 108 mm skull length and the anterior margin of the NAOF was at 46.0 % of skull length, respectively. If the 59 mm long NAOF of BMMS 7 equaled 26.6 % of skull length, then the estimated skull length would be 222 mm, whereas it if equaled 29.8 % of skull length then the estimated skull length would be 198.0 mm. Similarly, if the 90.1 mm length of BMMS 7 from the anterior margin of the NAOF to the posterior margin of the occiput equaled 53.5–54.0 % of skull length, then the estimated skull length would be 193.8–195.9 mm.

Alternatively, the skull length of BMMS 7 can be estimated by reconstructing it with increasing numbers of teeth and comparing the reconstructed tooth counts and skull lengths to Jouve’s (2004) regression equation of tooth count versus skull length. In BMMS 7 Teeth A–D are spaced 6.7 mm apart, so it can be assumed that each additional tooth would add 6.7 mm to the reconstructed skull length. The lengths and tooth counts of various possible reconstructions can be expressed in the equations:
$$ \begin{aligned} L_{\text{R}} & = L_{\text{P}} + (T_{\text{A}} \times S) \\ T_{\text{T}} & = T_{\text{P}} + T_{\text{A}} \\ \end{aligned} $$
where LR equals reconstructed length, LP equals preserved skull length (i.e., 142 mm), TA equals the number of additional teeth reconstructed, S equals the spacing of the reconstructed teeth (i.e., 6.7 mm), TT equals the total number of teeth per jaw side, and TP equals the number of teeth in the preserved portion of the upper jaw per side. The upper jaw preserves evidence of 16 teeth; however, the fact that the mandible preserves an alveolus that would have held a tooth that occluded behind maxillary Tooth P combined with the fact that in Pterodactylus the maxillary tooth row usually extends behind the dentary tooth row suggest that there was one more tooth in the maxilla behind Tooth P. Therefore, it is assumed that there were 17 teeth in the preserved portion of the upper jaw. If the upper jaw is reconstructed by adding only one terminal tooth, then TT = 18 and LR = 148.7 mm, whereas if TT = 24, then LR = 188.9 mm, if TT = 25, then LR = 195.6 mm, and if TT = 26, then LR = 202.3 mm. These data can be compared to skull lengths for the same number of teeth calculated using Jouve’s (2004) regression equation of tooth number versus skull length (y = 0.1732x + 15.695, where x = skull length and y = combined tooth count in the upper and lower jaws per side). According to the equation, a skull with 24 teeth per jaw side would be 186.5 mm long, whereas with 25 teeth the length would be 198.0 mm and with 26 teeth would be 209.6 mm.

The consilience of the estimates for skull lengths based on: (1) adding 8 teeth per jaw side to the preserved skull (i.e., 195.6 mm), (2) Jouve’s regression equation for 25 teeth per jaw side (i.e., 198.0 mm), (3) the relative position of the anterior margin of the NAOF in BSP 1883 XVI 1 and BSP AS I 739 (i.e., 193.8–195.9 mm), and (4) the relative length of the NAOF in BSP AS I 739 (i.e., 198.0), suggests that the skull of BMMS 7 was ~197 mm long and had 25 teeth.

Maximum body size

In order to reconstruct the size and skeletal proportions of BMMS 7, length measurements in mm of various skeletal elements from Wellnhofer’s (1970) Exemplare 1–8 and 10–28 (including the corrected measurements of TM 10341 discussed above) were used to calculate simple regression equations for skeletal elements versus humerus length (Table 2). Starting with the estimated skull length of 197 mm, the skeletal proportions were reconstructed using the regression equations, and the estimated lengths in mm are: neck, 160.1; PCRW, 135.4; humerus, 64.1; radius, 88.8; McIV, 68.4; WP1-4, 91.9, 86.5, 73.0, and 58.2, respectively; femur, 68.8; and tibia, 93.7. Using my standard method for estimating wingspan with normal flexures of the elbow, wrist, etc. (Bennett 2001), the estimated wingspan in life would be 106.2 cm. The regression equations reveal positive allometric growth in the length of the radius, proximal wing phalanges, and tibia relative to humerus length, and especially strong positive allometric growth of the skull, neck, and PCRW relative to humerus length, all of which can be noted in the modified Nopcsa curves in Fig. 9.
Table 2

Linear regression equations and correlation coefficients (R2) for length measurements (in mm) of selected skeletal elements versus humerus length for Pterodactylus antiquus from the Solnhofen Limestone of southern Germany (data from Wellnhofer 1970, but with corrected neck and PCRW measurements of TM 10341)

Linear regression equation (y = a + bx)

R2 value

Skull = −11.0 + 3.25 humerus

0.95

Neck = −18.7 + 2.79 humerus

0.96

PCRW = −2.7 + 2.16 humerus

0.97

Radius = −1.4 + 1.41 humerus

0.97

McIV = −3.2 + 1.17 humerus

0.97

WP1 = −3.9 + 1.50 humerus

0.96

WP2 = −3.9 + 1.41 humerus

0.98

WP3 = −1.0 + 1.15 humerus

0.96

WP4 = 0.4 + 0.90 humerus

0.96

Femur = −3.1 + 1.12 humerus

0.97

Tibia = −4.3 + 1.52 humerus

0.97

PCRW combined length of dorsal and sacral vertebrae, Mc metacarpal, WP wing phalanx

Pterodactylus antiquus had been thought to be small with 50 cm wingspans in adults (Wellnhofer 1991; Unwin 2006), plus a few ‘big, old individuals’ that Unwin characterized as ‘half again as large’ (presumably based on RM St. 18184 with an estimated wingspan in life of 70.5 cm). However, BMMS 7 demonstrates that P. antiquus reached estimated wingspans of at least 1.06 m, which is more comparable to other taxa in the Solnhofen pterosaur fauna; among the pterodactyloids, Aurorazhdarcho reached 1.06 m (Frey et al., 2011), P.longicollum reached 1.45 m (see below), Ctenochasma reached 1.9 m (Bennett 2007), and Cycnorhamphus reached ~2 m (Bennett 2013), whereas the rhamphorhynchoid Rhamphorhynchus reached 1.8 m (Bennett 1995).

Bony crest

The presence of the low bony crest in BMMS 7 and the holotype BSP AS I 739 demonstrates that Pterodactylus antiquus had a bony crest as predicted by Bennett (2002). The small size of the crest in BSP AS I 739, which though a rather large specimen is still immature with unfused cranial elements, scapula and coracoid, pelvic girdle, etc. (Bennett 1996a), indicates that the bony crest only began to develop as the animal approached adult size. BMMS 7 does not exhibit any signs of skeletal immaturity and so probably was near its maximum size. The crest extends along ~24 % of the reconstructed skull length from a short distance behind the middle of the NAOF to the posterior margin of the orbit, whereas the crest of Germanodactylus rhamphastinus (Bennett 2002) extends along ~55 % of the skull length from a short distance behind the anterior end of the NAOF to the posterior end of the premaxillae a short distance behind the posterior margin of the orbit, and the crest of Ctenochasma elegans (Bennett 2007) extends along ~33 % of the skull length from the posterior third of the rostrum to the anterior two-fifths of the NAOF. It is probable that the bony crest of P. antiquus would have increased in length and height as the animal grew older.

In both crested specimens of P. antiquus, the lateral surface of the bony crest is smooth, unlike the striated crests of P. longicollum, Ctenochasma, and Gnathosaurus. This might be phylogenetically significant; however, in one specimen of Germanodactylus the lateral surface of the posterior part of its otherwise striated crest was smooth (Bennett 2002). It is possible that bony crests were initially smooth, and only became striated in older individuals as a result of interaction with the large and thick keratinous crest. Thus, it is possible that striae would have appeared on P. antiquus crests as the animals grew older, in which case the difference would be of ontogenetic rather than phylogenetic significance.

Soft tissue crest

The bony cranial crest of P. antiquus was presumably extended upward by soft tissues as in other Solnhofen pterodactyloids, but unfortunately evidence of bony crests has only been found in specimens that do not preserve soft tissue traces. BSP 1929 I 18 with a skull length of 84 mm preserves traces of a small soft tissue crest ~25 % of skull length, MCZ 1505 with a skull length of 113.5 mm preserves traces of a larger crest perhaps 38–45 % of skull length, and Frey and Tischlinger (2000) described a privately held specimen with a skull length of 137 mm that preserved traces of a taller crest at least ~33 % of skull length. In each of these specimens the soft tissue traces are above the posterior half of the NAOF and the orbit and in the same region as the bony crest of BMMS 7, and the relative height of the soft tissue crest was greater in large specimens. Together these specimens confirm the presence of a soft tissue crest in large individuals of Pterodactylus and document its progressive development.

Frey et al. (2003), Frey and Tischlinger (2010), and Tischlinger and Frey (2007) reconstructed the soft tissue crest of Pterodactylus as not only above the posterior NAOF and orbit, but also extending backward behind the skull and along the superior margin of the occipital lappet such that the lappet supported or was incorporated into the crest. However, they have not published evidence to support the interpretation, and in the privately held specimen (Frey and Tischlinger 2000), the soft tissue crest extends to the back of the skull, but does not seem to extend any further.

Frey and Tischlinger (2010) suggested that the bony crest in Solnhofen pterodactyloids grew by the mineralization of a ‘fibrous’ soft tissue crest. It is not clear what sort of fibers they think formed the soft tissue crest, and they make no distinction between the ‘fibrous’ crest and the fibers that can be seen in occipital lappets, but Frey et al. (2003) made it clear that they do not think that the soft tissue crest was formed of heavily cornified epidermis. I disagree with their interpretation. There is no evidence of gradual mineralization of soft tissues producing the crest rather than normal ossification. The pattern of vertical striations seen on the sides of the bony crests of various pterosaurs (though not Pterodactylus as yet) is consistent with thick cornified epidermis being pushed upward by new deposition around the base of the crest. In addition, the thick layer of calcite external to the bony crest in a large Ctenochasma elegans (JME-SOS 2179, Bennett 2007) is consistent with a thick cornified soft tissue crest. Furthermore, the marked faintness of the iron-oxide trace of the soft tissue crest of BSP 1929 I 18 compared to the prominent trace of the gular sac indicates that the crest was formed of a different tissue or tissues than the skin, muscle, and fibrous connective tissues of the gular sac. Finally, the fluorescent traces of soft tissue crests seen in MCZ 1505 and Germanodactylus (MCZ 1886, Bennett 2002) suggest that the crest was formed of thick and/or resistant soft tissue; MCZ 1505 preserves the trace of the crest but no traces of the gular sac, the position of which is clearly shown by the ceratobranchials preserved in situ.

Occipital lappet

Wellnhofer (1970) first described the occipital lappet of Pterodactylus in BSP 1883 XVI 1, terming it in German a Hinterhauptslappen and suggesting that it was a sexual display structure. Frey and Tischlinger (2000) described one based on UV photography of a privately held specimen of Pterodactylus and also referred to it in German as a Hinterhauptslappen. They subsequently referred to the structure in English as an occipital cone (Frey et al. 2003; Tischlinger and Frey 2007; Frey and Tischlinger 2010). However, it is not clear that the structure was conical; although their example does look somewhat conical, those of MCZ 1505 and BSP 1929 I 18 do not. Because there is as yet no evidence that it was connected to the soft tissue crest, the term occipital lappet, which is a better translation of Wellnhofer’s term, is appropriate [q.v., Hunter and Morris (1897, 3: 2849) define a lappet as “a little lap or loose part of a dress, especially part of a head-dress hanging loose”]. Moreover, if it were connected to the soft tissue crest, it seems unlikely that it would have a different composition than the rest of the crest, and it probably would not appear to be a discrete structure.

BSP 1929 I 18 confirms that the occipital lappet consisted of spiral arranged filaments (Frey and Tischlinger 2000; Frey et al. 2003) seemingly embedded in a thinner material. Interestingly, along the margins of the soft tissue impressions of the neck BSP 1929 I 18 also preserves traces of the short filamentous body covering like those described by Frey and Martill (1998), fig. 5). The filaments in the occipital lappet and those along the neck are of comparable appearance and thickness. Because it is unlikely that epidermal hair-like structures would be embedded and spiraling within soft tissues ~3.5 mm thick, it is probable that the lappet was formed of a tuft of hair-like structures, longer than those on the skin of the neck, and perhaps modified with thin lateral flanges or a pennate structure, and spiraling around one another in a more or less compact bundle.

Lifestyle

The reconstruction of Pterodactylus antiquus presented here and based on BMMS 7 and other specimens is of an animal with a skull 197 mm long with 25 teeth per jaw half and a wingspan of 1.06 m. In terrestrial locomotion the forelimb would be somewhat longer than the hind, 221 mm versus 161 mm, but given the marked flexion of the elbow and the fact the forelimb was extended forward more than in typical quadrupeds, a stable stance and easy locomotion would result. Such an animal would be well suited to feed on invertebrates while walking on beaches and mudflats, though the dentition suggests it was primarily a piscivore.

It is generally assumed that Pterodactylus fed on the abundant small fishes of the Solnhofen lagoons such as Leptolepides sprattiformis. With a skull length of 20 cm consumption of Leptolepides, which are typically 5–8 cm long, would be easy. Feeding might be accomplished while floating on the water in the manner of ducks and dipping for fishes and other foods, or flying low over the water and snatching fishes near the surface without alighting on the water, or by diving into the water in the manner of gannets and pelicans. The gular sac, however, was not in any way comparable to that of pelicans, and though it might have held fishes as they were transported, it could not have been used is trapping fishes. If reproduction of Pterodactylus was synchronized with breeding of fishes such as Leptolepides, then the precocial hatchlings would have an abundant supply of tiny fishes upon which to feed until their modest growth rate and increasing skills enabled them to prey on larger fishes.

Systematic paleontology

  • Order PTEROSAURIA Kaup, 1834

  • Suborder PTERODACTYLOIDEA Plieninger, 1901

  • Family PTERODACTYLIDAE Bonaparte, 1838

  • GenusPTERODACTYLUSCuvier, 1809

Type species

Ornithocephalus antiquus Sömmerring, 1812.

Included species

Pterodactylus antiquus (Sömmerring, 1812).

Distribution

Lower Tithonian, Malm Zeta 2 and 3, of Solnhofen, Eichstätt, Kelheim, etc., Bavaria, Germany.

Diagnosis

As for the type and only species.

Discussion

Given that Ornithocephalus or Pterodactylus was the first pterosaur genus named, many other species and specimens were referred to it. Wellnhofer’s (1970) revision lumped many species together and left only the four small species, P. antiquus, P. kochi, P. elegans, and P. micronyx, and the large P. longicollum and P. suevicus in the genus. Two additional species that were named on the basis of fragmentary large specimens (“P.grandis and “P.grandipelvis) and a large wing referred to as “P.” sp. (PTH 1963) were excluded from the genus and viewed as Pterodactyloidea incertae sedis. Wellnhofer (1978, 1980) subsequently repeated much of that taxonomic scheme but transferred P. suevicus to Gallodactylus and excluded additional species (“P.cerinensis and “P.suprajurensis from France, “P.manseli and “P.pleydelli from England, and “P.arningi, “P.brancai, and “P.maximus from Tanzania) and also the teeth from the Guimarota mine of Portugal referred to Pterodactylus sp. (Wiechmann and Gloy 2000), viewing them all as Pterodactyloidea incertae sedis.

As asserted by Jouve (2004) and confirmed above, P. kochi is a junior synonym of P. antiquus. In addition: specimens assigned to P. elegans are juveniles of Ctenochasma elegans (Jouve 2004; Bennett 2007); P. suevicus is properly referred to as Cycnorhamphus suevicus (Bennett 1996b, 2013); specimens assigned to P. micronyx are not congeneric with P. antiquus and almost certainly are juveniles of Gnathosaurus subulatus though we still lack a large specimen with associated skull and postcranials to confirm the association (Bennett 1996a; see below); specimens assigned to P. longicollum are not congeneric with P. antiquus and will be placed in a new genus (see below); and species based on large fragmentary materials from the Solnhofen Limestone that Wellnhofer (1970) viewed as Pterodactyloidea incertae sedis are best viewed as nomina dubia because the materials are insufficient to characterize a species. Thus, P. antiquus is the only species of Pterodactylus known from the Solnhofen Limestone and the adjacent, roughly contemporaneous formations.

PTERODACTYLUS ANTIQUUS (Sömmerring, 1812)

  • Ornithocephalus antiquus Sömmerring, 1812: 89.

  • Pterodactylus longirostris Cuvier, 1819: 1126.

  • Ornithocephalus kochii Wagner, 1837: 165.

  • Pterodactylus kochi (Wagner), von Meyer, 1859: 35.

  • Pterodactylus antiquus (Sömmerring), Lydekker, 1888: 5.

  • Pterodactylus antiquus (Sömmerring), Wellnhofer, 1970: 15.

  • Pterodactylus kochi (Wagner), Wellnhofer, 1970: 22.

  • Pterodactylus antiquus (Sömmerring), Jouve, 2004: 549.

Holotype

BSP AS I 739.

Horizon and locality

Lower Tithonian, Solnhofen Limestone, Malm Zeta 2, of Eichstätt, Bavaria, Germany.

Distribution

Solnhofen Limestone, Malm Zeta 2, Solnhofen, Eichstätt, and adjacent localities, Germany; and Mörnsheimer Limestone, Malm Zeta 3, Kelheim and adjacent localities, Germany, and the stratigraphically equivalent Hienheim Beds. In addition, a specimen that appeared to be referable to P. antiquus, which was recently sold by a commercial collector was described as having come from Painten where the limestones are somewhat stratigraphically lower (Malm Zeta 1–2) than those of Solnhofen and Eichstätt; therefore it may be that P. antiquus was present during Malm Zeta 1 time.

Diagnosis

Upper Jurassic pterodactyloid with elongate skull, a slender and elongated rostrum, and the superior margin of the skull straight to only slightly concave upward; naso-antorbital fenestra length ~20–25 % of skull length in large individuals; tooth number proportional to skull length with up to ~25 teeth per jaw side in large individuals; teeth flat conical, large anteriorly and tooth size decreasing in size posteriorly; tooth row length ~75 % of jaw length, and upper tooth row extending posterior to the anterior margin of the naso-antorbital fenestra; a low sagittal bony crest over the naso-antorbital fenestra and orbit in large individuals, which apparently lacks the striations seen in Ctenochasma, Germanodactylus, and Gnathosaurus; a soft tissue crest extending upward from the bony crest in large individuals; and an occipital lappet of soft tissues extending posteriorly from the occipital region. Cervical vertebrae 3–7 elongate and neck relatively longer that in Cycnorhamphus, Ctenochasma, and Aurorazhdarcho (shared with “P.” longicollum; see below). In small specimens WP2 is 93–96 % of WP1 length, in large specimens ~91 %, and WP1-4 lengths typically exhibit a convex upward curve in modified Nopcsa curves (Fig. 9; shared with Germanodactylus). In the pes, MtII is greater than or equal to MtI in length and the proximal phalanges of digit I–III show progressive reduction in length whereas those of digits I and IV are subequal (shared with Germanodactylus).

Referred materials

Note that when the part and counterpart of a specimen are in separate museums, the catalog number of the other slab is preceded by a slash and placed in parentheses after the catalog number of the one slab: AFGM V-822 (/BMMS 1; = Wellnhofer’s (1970) Exemplar [Ex.] 27); AMNH 1942; BM 42736; BMMS 1 (/AFGM V-822; = Ex. 27), 2 (= Ex. 17), 7; BSP 1878 VI 1, 1883 XVI 1 (/MCZ 1505), 1924 V 1, 1929 I 18, 1937 I 18, 1967 I 276, 1968 I 95, 1986 XV 131, AS XIX 3 (/SM R404), AS V 29; MB.R. 3530 (= Ex. 14), 3533 (= Ex. 3); MCZ 1503, 1505 (/BSP 1883 XVI 1); NHMW 1975/1756/0000; PMU 24791 (old # R68/R440, = Ex. 24); RM St. 18184; SM R404 (/BSP AS XIX 3), R4072, R4074; SMNS 81775 (Bennett 2006); JME-SOS 4008 (old # PTH 1953/80, = Ex. 26), 4588, 4592 (old # PTH 1962/148, = Ex. 16); TM 10341, 13105; and two privately held specimens described by Frey and Martill (1998) and Frey and Tischlinger (2000). Note that Wellnhofer’s Ex. 9 that was referred to P. kochi (old # PTH 29.III.1950, now JME-SOS 4593) was reinterpreted as a juvenile of Germanodactylus cristatus (Bennett 2006). Two specimens included in Wellnhofer’s (1970) revision are missing: Ex. 18 described by Zittel (1882) was destroyed in World War II, and Ex. 19 was in the private collection of E. Schöpfel of Obereichstätt, which apparently was sold off at some time in the past (H. Tischlinger, personal communication). The counterpart slab of Ex. 19 was sold at auction in August 2008 and then offered at auction in January 2009 and did not sell, but its present whereabouts and those of the part slab are unknown.

Discussion

The synonymy listed above is abbreviated: for a complete synonymy of names applied before 1970 see Wellnhofer (1970: 15, 22–23). Note that the diagnoses presented here include a suite of characters that together distinguish one taxon from other taxa, not a list of autapomorphies that each distinguish the taxon from all others.
  • GenusARDEADACTYLUS gen. nov.

Type species

Pterodactylus longicollum von Meyer, 1854.

Included species

Ardeadactylus longicollum (von Meyer, 1854).

Distribution

Lower Tithonian, Malm Zeta 1–3, of Nusplingen, Württemberg, and Eichstätt, Schamhaupten, and Daiting, Bavaria, Germany.

Diagnosis

As for the type and only species.

Etymology

From Ardea meaning “heron” and the name of a genus of herons + -dactylus meaning “finger,” a common component of pterosaur names. The name is intended to refer to the presumed heron-like (i.e., long-necked long-legged piscivore) character of this pterosaur.

Discussion

It has been evident to many for some years that this species did not belong in the modern conception of the genus Pterodactylus. Some have suggested that the species should be placed in the genus Diopecephalus, which was erected by Seeley (1871) to include P. kochi, P. longicollum, and P. rhamphastinus; however, Seeley (1901) designated P. kochi as the type species of Diopecephalus, and so Diopecephalus is a junior synonym of Pterodactylus and cannot be applied to P. longicollum (see Bennett 2006 for a thorough discussion of this issue). Therefore, a new name was needed.

ARDEADACTYLUS LONGICOLLUM (von Meyer, 1854)

  • Pterodactylus longicollum von Meyer, 1854: 52.

  • Pterodactylus (Ornithocephalus) vulturinus Wagner, 1857: 174.

  • Pterodactylus longicollis Wagner, 1858: 456.

  • Diopecephalus longicollum (von Meyer), Seeley, 1871: 55.

  • Pterodactylus suevicus (Quenstedt), Fraas, 1878: 25.

  • Pterodactylus longicollum (von Meyer), Wellnhofer, 1970: 56–57.

Neotype

SMNS 56603. The original holotype specimen described by von Meyer (1854) was destroyed during World War II, and Wellnhofer (1970) designated this specimen as the neotype.

Horizon and locality

Nusplingen Limestone, Malm Zeta 1, of Nusplingen, Württemberg, Germany.

Distribution

Nusplingen Limestone, Malm Zeta 1, of Nusplingen, Württemberg, and Solnhofen Limestone, Malm Zeta 2, Eichstätt and Schamhaupten, Bavaria, and Mörnsheimer Limestone, Malm Zeta 3, of Daiting, Bavaria, Germany.

Diagnosis

Upper Jurassic pterodactyloid with elongate skull, a slender and elongated rostrum and the superior margin of the anterior skull markedly concave; tooth number ~15 teeth per jaw side in mature adults; teeth strong, gently curving, and somewhat anteriorly inclined; posterior teeth equivalent in size to anterior teeth; tooth row length ~50 % of jaw length, and upper tooth row not extending under the naso-antorbital fenestra. Cervical vertebrae 3-7 elongate, and neck relatively longer that in Cycnorhamphus, Ctenochasma, and Aurorazhdarcho (shared with Pterodactylus). In known specimens WP2 is 65–71 % of WP1 length, and WP1-4 lengths exhibit a concave upward curve in modified Nopcsa curves (Fig. 11).
Fig. 11

Modified Nopcsa curves of skeletal proportions of Aurorazhdarcho micronyx based on measurement data from Wellnhofer (1970: Exemplare 30, 32, 35, 36, and 38–44) and NMB Sh 110. The dashed line is the holotype specimen, the heavy lines are group B, and all other specimens are group A (see text for explanation). Format and vertical scale are the same as Fig. 9

Referred materials

JME-SOS 2428 (= Ex. 57).

Remarks

Wellnhofer (1970) considered the sample of P. longicollum to consist of six specimens: von Meyer’s (1859) holotype and referred specimen (= Ex. 55 and 60), Münster’s (1839) holotype of P. longipes (MB.R. 3532; = Ex. 56), Wagner’s (1858) holotype of P. vulturinus, Fraas’s (1878) specimen described as P. suevicus though it is not conspecific with Cycnorhamphus suevicus (Quenstedt) (SMNS 56603; = Ex. 58), and a previously undescribed specimen at the JuraMuseum (JME-SOS 2428; = Ex. 57). Wellnhofer’s referral of Münster’s (1839) holotype of P. longipes was tentative because the specimen consisted of only a left femur and incomplete left tibia, but I think that the specimen does not provide enough evidence to refer it with confidence to this species rather than to Cycnorhamphus, Gnathosaurus, or Aurorazhdarcho, and so P. longipes should be considered a nomen dubium. Of the remaining five specimens, von Meyer’s and Wagner’s specimens were lost during World War II, so there are only two surviving specimens.

The species Pterodactylus longicollum was diagnosed by Wellnhofer (1970) as follows [translated from the original German, with numbering of characters added]:

[1] Large Pterodactylus with the following features: [2] continuous naso-antorbital fenestra, [3] jaws toothed from the tips to almost the midpoint of the skull with about 15 strong, anteriorly curved teeth. [4] Cervical vertebra greatly prolonged; neck longer than the skull. [5] Scapula and coracoid fused. [6] Sternal plate with angular margin. [7] Metacarpals significantly longer than the forearm. [8] Significant decrease in length of wingfinger phalanges from first to fourth. [9] Sacrum firmly adherent to the ilia. [10] Ilium, pubis and ischium bound by sutures.

Most of these characters do not adequately distinguish this species from other Solnhofen pterodactyloids. Taken more or less in order: 1—Despite the size of the specimens (the estimated wingspan of the neotype using my usual method is 1.45 m and the referred specimen is ~10 % larger), large size is not informative because it is now known that P. antiquus reached estimated wingspans of 1.06 m (see above), Aurorazhdarcho reached 1.06 m (Frey et al. 2011), Ctenochasma reached 1.9 m (Bennett 2007), and Cycnorhamphus reached ~2 m (Bennett 2013). 2—This probably refers to the fact that the median nasal process is not visible in the available specimens, however, that is a consequence of the preservation and preparation of specimens rather than a taxonomically informative character. 4, 7, and 8—These are ontogenetic characters in that the growth of the skeleton in Solnhofen pterodactyloids has been shown to exhibit significant allometry (e.g., large specimens of Pterodactylus antiquus and Ctenochasma elegans have significantly longer necks and skulls than small specimens and also show marked changes in the proportions of wing elements with the metacarpus and proximal wing phalanges increasing in length much more than the antebrachium and distal wing phalanges), thus the features merely reflect the skeletal maturity of specimens referred to A. longicollum. 5, 9, and 10—These are ontogenetic characters that reflect the fact that the available specimens are large and have attained skeletal maturity, whereas smaller, immature individuals of the species would not exhibit the features. 6—This is an ontogenetic character reflecting the complete ossification of the sternal plate, whereas smaller immature specimens would have smaller, more rounded sternal plates.

Ardeadactylus longicollum seems to be a rare component of the Solnhofen pterodactyloid fauna; however, it is possible, perhaps even probable, that juveniles of the taxon are mixed in the sample of Pterodactylus antiquus and yet to be identified. It was similar to Pterodactylus antiquus in general body form with an elongate skull and long neck, but seems to have been larger with fewer, relatively larger teeth that suggest that it preyed on larger fishes than Pterodactylus.
  • Family CTENOCHASMATIDAE Nopcsa, 1928

  • GenusAURORAZHDARCHOFrey, Meyer & Tischlinger,2011

Type species

Aurorazhdarcho primordius Frey, Meyer & Tischlinger, 2011.

Included species

Aurorazhdarcho micronyx (von Meyer, 1854).

Distribution

Lower Tithonian, Malm Zeta 2, Solnhofen Limestone of Solnhofen, Eichstätt, and adjacent localities, Bavaria, Germany.

Diagnosis

As for the type and only species.

AURORAZHDARCHO MICRONYX (von Meyer, 1856)

  • Ornithocephalus Redenbacheri Wagner, 1851: 17.

  • Pterodactylus micronyx von Meyer, 1856: 826.

  • Pterodactylus longirostris von Meyer, 1859: 31.

  • Pterodactylus brevirostris von Meyer, 1859: 55.

  • Pterodactylus pulchellus von Meyer, 1861: 470.

  • Pterodactylus micronyx von Meyer, Wellnhofer, 1970: 35–36.

  • Aurorazhdarcho primordius Frey, Meyer, & Tischlinger, 2011: S40.

Holotype

ELTE V 256 (=Wellnhofer’s Ex. 39).

Horizon and locality

Lower Tithonian, Solnhofen Limestone, Malm Zeta 2, of Eichstätt, Bavaria, Germany.

Distribution

Lower Tithonian, Malm Zeta 2, Solnhofen Limestone of Solnhofen, Eichstätt, and adjacent localities, Bavaria, Germany.

Diagnosis

Upper Jurassic pterodactyloid with elongate skull, slender and elongated rostrum, and superior margin of the skull markedly concave upward; naso-antorbital fenestra length ~20–25 % of skull length in small individuals; jaws of small individuals bearing up to 18 closely-spaced, long, gently curving teeth per jaw side, anterior teeth angled anteriorly, posterior teeth shorter and more upright, and the upper and lower teeth interlocking to form a basket for sieving food items from water, but teeth shorter and stouter than in Ctenochasma. The skull is unknown in large individuals. Cervical vertebrae 3–7 of moderate length, and neck relatively shorter than in Pterodactylus and Ardeadactylus (shared with Ctenochasma and Cycnorhamphus). McIV exhibits marked positive allometry relative to the radius: in small specimens the radius length exceeds McIV length but the opposite is the case in middling to large specimens; in small to middling specimens WP2 is ~62–74 % of WP1 length, in the large specimen 59 %, and WP1-4 lengths typically exhibit a concave upward curve in modified Nopcsa curves (Fig. 11). In the pes, MtI is greater than MtII in length and the proximal phalanges of digit I–IV are subequal in length (shared with Ctenochasma).

Referred materials

BSP 1936 I 50, 1964 XXIII 100; CM 11425, 11426; MB.R. 3531 (= Ex. 40); NMB Sh 110; TM 13104; RM St. 18183; NHMUK 42735; SNSD-MMG BaJ 21 (= Ex. 41); JME-SOS 4594 (old # PTH 1983/2649); and PMU 24792 (old #R 67/R 439, = Ex. 43). Note that Wellnhofer’s Ex. 31 referred to P. micronyx (JME-SOS 4006, old # PTH 1957/52) was reinterpreted as a juvenile of Germanodactylus cristatus (Bennett 2006).

Discussion

Note that the synonymy listed above is abbreviated: for a complete synonymy of names applied before 1970 see Wellnhofer (1970: 35–36). Wellnhofer (1970) thought the holotype specimen was lost, and so designated another specimen described by von Meyer (1856), BSP 1911 I 31 (= Ex. 42), as the neotype. However, Ősi et al. (2010) rediscovered the holotype in the collections of Eötvös Loránd University, Budapest, Hungary.

Wellnhofer referred four small specimens to this taxon: his Ex. 29 at the Descartes-Gymnasium in Neuberg an der Donau, Germany; Ex. 33 described by Řikovský (1925) and according to Wellnhofer in Brunn, is in the collections of Masaryk University in Brno, Czech Republic (N. Doláková, personal communication); TM 13104 (= Ex. 34); and SM 405/NHMW 2012/0118/0001 (= Ex. 37). However, those specimens exhibit limb proportions that differ from those of small specimens of A. micronyx and compare well with those of similarly sized specimens of Ctenochasma elegans (Bennett 2007), and so they are referred to the latter taxon. Ex. 29 and 33 are two of the smallest specimens of Ctenochasma or Aurorazhdarcho, and although the smallest specimens of both taxa are quite similar, the proportions of their wing fingers and their femur to tibia seem to better match those of Ctenochasma. TM 13104 and SM 405/NHMW 2012/0118/0001 are somewhat larger and also exhibit wing finger and femur to tibia proportions that better match those of Ctenochasma, but also have radius to McIV proportions that are markedly different from those of similarly sized specimens of A. micronyx.

After removal of the four Ctenochasma specimens from the sample, examination of the modified Nopcsa Curves (Fig. 11) suggests that the remaining specimens can be separated into two groups: Group A consisting of the holotype ELTE V 256, BSP 1936 I 50 and 1964 XXIII 100, CM 11425, NHMUK 42735, MB.R. 3531, RM St. 18183, SNSD-MMG BaJ 21, and PMU 24792 with McIV lengths less than the average of the radius plus WP1 lengths and with a moderate proportional decrease in the lengths of WP1 through WP4; and group B consisting of BSP 1911 I 31 and CM 11426 with McIV lengths greater than the average of the radius plus WP1 lengths and with a great proportional decrease in the lengths of WP1 through WP4. If the two groups represent distinct species, the specific epithet micronyx will apply to group A and a new specific epithet will be needed for group B because both BSP 1911 I 31 and CM 11426 have only been referred to as P. micronyx; however, the status of A. micronyx groups A and B is beyond the scope of this article that was intended to revise the genus Pterodactylus, and I will take the matter up elsewhere.

NMB Sh 110, the holotype of Aurorazhdarcho primordius, belongs in group A, and the specific epithet primordius is a junior synonym of the epithet micronyx. When I first saw NMB Sh 110 in 2002, I was told that the then owner thought the specimen pertained to P. micronyx. After further analysis, I have concluded that NMB Sh 110 belongs to group A of this taxon because its short neck and pedal morphology indicate that it is a ctenochasmatid, and because its limb proportions match those of group A specimens and differ from those of Ctenochasma elegans. In particular, its radius to McIV ratio is markedly greater than would be the case in a similarly sized Ctenochasma. Linear regression equations of radius length to McIV length can be used to predict that a Ctenochasma elegans [McIV = 1.02 × radius − 3.3, R2 = 0.99; based on Ex. 29, 33, 34, 37, 45-52, 65 and the Schöpfl specimen (Bennett 2007)] with a radius the size that NMB Sh 110 has (i.e., 81 mm) would have a McIV 79.3 mm long, whereas a group A A. micronyx (McIV = 1.34 × radius − 9.5, R2 = 0.99) would have a McIV 99.4 mm long. The McIV of NMB Sh 110 is 95 mm, thus only 4 % shorter than predicted.

Frey et al. (2011) stated that they compared NMB Sh 110 to many different pterodactyloid taxa and concluded that it differed in morphology and limb proportions, but for the most part did not comment on specific comparisons. They stated (Frey et al. 2011: S50):

Aurorazhdarcho is the only known Late Jurassic pterodactyloid, in which the length of the metacarpal is equal to the length of the radius and ulna. Also unique is that the first wing finger phalanx is the longest element in the wing.

The suggestion that NMB Sh 110 is the only Late Jurassic pterodactyloid with McIV length greater than ulna length was incorrect because specimens assigned to P. micronyx all exhibit long McIV (Fig. 11). The suggestion that NMB Sh 110 is the only Late Jurassic pterodactyloid in which WP1 is the longest wing element was incorrect because the same is true of Ctenochasma elegans and specimens assigned to P. micronyx. On the same page, they stated that NMB Sh 110 was “very close” to Ctenochasma elegans, but they seem to have concluded that it was distinct from Ctenochasma because “no specimen of Ctenochasma elegans preserves the shoulder girdle in a way that the wing articulation can be reconstructed with reliability.” The fact the Frey et al. could not make a comparison between NMB Sh 110 and Ctenochasma in no way supports their interpretation that NMB Sh 110 represented a distinct previously unknown taxon. Furthermore, the fact that NMB Sh 110 exhibits complex morphologies that have not been noted in other Solnhofen pterodactyloids does not contradict the synonymy of micronyx and primordius because the bones of immature individuals usually lack complex shapes and processes (Johnson 1977; Brinkman 1988; Bennett 1995, 1996a). In the end, Frey et al. (2011) provided no evidence that NMB Sh 110 was not conspecific with P. micronyx, whereas the ctenochasmatid pes and limb proportions indicate that they are conspecific.

Whereas the species retains the epithet micronyx, it is transferred to the genus Aurorazhdarcho because it has been known for several years that the species is not congeneric with Pterodactylus antiquus. However, the taxonomy is complicated by the fact that Bennett (1996a) argued that the specimens assigned to P. micronyx represented juveniles of Gnathosaurus subulatus, which is represented by the holotype incomplete mandible BSP AS VII 369 and the isolated skull JME-SOS 4580 (= Ex. 70, old # PTH 1951/84), and noted that the latter name would have priority. Which group does Gnathosaurus subulatus belong to? At present, we cannot tell. It is a shame that NMB Sh 110 lacks a skull because an associated skull might help clear up some of the present uncertainly, but in the absence of a large specimen with an associated skull and skeleton to confirm the identification, it is necessary to use the genus Aurorazhdarcho for the species micronyx at least until such a specimen is discovered.

The species Pterodactylus micronyx was diagnosed by Wellnhofer (1970) as follows [translated from the original German, with numbering of characters added]:

[1] Medium sized species with slender limb bones. [2] Skull and neck relatively shorter and the metacarpus longer than in P. kochi and P. antiquus. [3] Naso-antorbital fenestra not longer than orbit. [4] Antorbital fenestra incompletely separated from the naris by a posteriorly directed nasal process. [5] Quadrate inclined. [6] The dentition in adults occupies only the anterior third of the skull. [7] Tooth type: slender, shorter than in P. elegans, anteriorly closely-spaced and pointing forward, posteriorly short and straight. [8] Number of teeth per jaw half: 15-18. [9] Relatively long metacarpus with positive allometry, metacarpal IV reaches the length of the radius in animals of ~40 mm PCRW and surpasses it in larger animals. [10] The ossification of the foot skeleton is completed with a phalangeal formula of 2.3.4.5.1 at about 40 mm PCRW.

Characters 2 and 7–9 were incorporated into the present diagnosis, but the other characters do not adequately distinguish this species from other Solnhofen pterodactyloids. Taken more or less in order: 1—The size reflects the fact that the sample of specimens available to Wellnhofer consisted entirely of immature individuals. 3, 6, and 10—These are ontogenetic characters that reflect the fact that all the available specimens are small, immature individuals, whereas larger individuals would probably not exhibit them. 4—This refers to the fact that the median nasal process, which was incorrectly thought to partially separate the naris from antorbital fenestra, is visible in some of the available specimens. 5—The statement that the quadrate is inclined does not specify how much, and the angle may have changed during ontogeny, might be altered by crushing, and does not seem to be much different than in other Solnhofen pterodactyloids. 10—This is an ontogenetic character reflecting the ossification of the foot skeleton despite the fact that all specimens known at the time were immature.

Acknowledgments

I must thank many: M. Röper of the BMMS, H. Mayr and O. Rauhut of the BSP, J. Leloux of the TM, U. Göhlich of the NHMW, F. Jenkins, Jr. and C. Schaff of the MCZ, G. Viohl and M. Koebl-Ebert of the JME, and R. Wild of the SMNS provided access to specimens under their care and assistance while studying them. O. Hampe of the HM, J.O.R. Ebbestad of the PMU, M. Wilmsen of the SNSD-MMG, J. Šušolová of the Moravian Museum, Brno, Czech Republic, and N. Dolákov of the UB provided information about specimens in, or not in, their collections. H. Tischlinger provided invaluable advice and information, B. Creisler consulted on the Latin, and an anonymous reviewer provided constructive criticism. This research was supported in part by a Deutscher Akademischer Austauschdienst grant (#A/02/15049) and an FHSU Internationalization grant.

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  1. 1.Department of Biological SciencesFort Hays State UniversityHaysUSA

Personalised recommendations