LONCHODECTIDAE, Unwin, 2001

Averianov, A. O., 2020, Taxonomy of the Lonchodectidae (Pterosauria, Pterodactyloidea), Proceedings of the Zoological Institute of the Russian Academy of Sciences 324 (1), pp. 41-55 : 50-53

LONCHODECTIDAE

Serradraco sagittirostris ( Owen, 1874) . This taxon is known from a single specimen (NHMUK PV R 1823) from the Fig. 4. Mandibular fragments of Ikrandraco machaerorhynchus ( Seeley, 1870) . Upper Tunbridge Wells Sand Formation at A, B – NHMUK R2269, an anterior fragment of the mandibular symphysis, St Leonards on Sea, East Sussex, England. in dorsal (A) and lateral (B) views (image in B is flipped horizontally). C, D –

CAMSM B54855, holotype, a middle fragment of the mandibular symphysis, in The specimen, holotype of Pterodactylus dorsal (C) and lateral (D) views. Modified from Unwin (2001: figs 11E, F, 12D, sagittirostris Owen, 1874 , consists of two E). Scale bar = 10 mm. fragmentary mandibular rami posterior to the mandibular symphysis ( Owen 1874: Referred specimens. CAMSM B54.486, ros- pl. 2). This taxon was subsequently referred to as trum fragment (holotype of Ornithocheirus microdon Ornithocheirus ( Seeley 1901; Wellnhofer 1978), Seele, 1870); CAMSM B54.439, rostrum fragment Lonchodectes ( Hooley 1914; Unwin 2001), or its (holotype of Ornithocheirus oweni Seeley, 1870 ); own genus Serradraco ( Rigal et al. 2017) . Rodrigues NHMUK R2268 and R2269, fragments of the man- and Kellner (2013) considered Pterodactylus sagdibular symphysis; all from the type locality. ittirostris a nomen dubium. As was noted by these Diagnosis. Differs from I. avatar by a relatively authors, it cannot be directly compared with most shallower mandibular crest and approximately 4.5 of the lonchodectid taxa, which are based on jaw alveoli per 3 cm of alveolar margin. tips. The most characteristic feature of this taxon is Comments. According to Wang et al. (2014), a asymmetric borders of the raised alveoli, which slope combination of a low and elongate crestless skull with gently anteriorly and steeply posteriorly, giving to a lower jaw that has a well developed mandibular the alveolar border a peculiar saw-like appearance crest with a hook-shaped posterior process found in ( Rigal et al. 2017). It should be noted, however, that I. avatar has no parallel among the known pterosaur this particular pattern of alveolar border is confined species. However, at least two of these features are only to the anterior preserved part of the dentary, found in a pterosaur species from Cambridge Green- in more posterior parts the alveoli are little raised sand: a low crestless skull ( Ornithocheirus microdon ) and have symmetric borders. S. sagittirostris is also and a developed mandibular crest ( Ornithocheirus peculiar in having a distinct groove on the external machaerorhynchus ). The presence of a hook-shaped surface of the mandible ventral to the alveolar border, posterior process of dentary cannot be established which is connected with branching grooves for blood for the Cambridge Greensand species because of its vessels covering most of the mandibular external incompleteness. Ornithocheirus microdon is based on surface. None of the lonchodectid taxa has a similar rostrum fragment and O. machaerorhynchus – on a alveolar edge and external mandibular groove. Serdentary fragment. These fragments are referred here radraco sagittirostris does not preserve any character to a single species, which is referred to as Ikrandra- considered here to be diagnostic for the Lonchodecco because of a combination of crestless skull and tidae. The alveoli are raised, but in the Lonchodectidae , they are raised above the surface of the palate or mandibular symphysis that cannot be checked for S. sagittirostris . Therefore, this taxon is not included in the Lonchodectidae .

BEXHM 2015.18. It is a fragmentary pterosaur skeleton from the Upper Tunbridge Wells Sand Formation or Lower Weald Clay Formation (upper Valanginian or lower Hauterivian) at Foreshore at Bexhill, East Sussex, England. It includes a jaw fragment with three teeth, six articulated dorsal vertebrae, distal ulna fragment, proximal syncarpal, and fragments of wing phalanges ( Rigal et al. 2017). The jaw is likely part of the mandibular ramus posterior to the symphysis. The teeth are small with conical crowns, evenly spaced and located in alveoli with prominent, dorsally raised borders. The dorsal vertebrae are not fused. According to the original description, the anterior-most complete vertebra has an articular facet for the rib that is significantly larger than that of all of the more posterior vertebrae. Rigal et al. (2017) think that the enlarged size of the parapophysis of the anterior-most complete vertebra suggests that it is a free dorsal vertebra. I assume that the authors mean the height of the transverse process rather than the size of the rib articulation facet. The rib articular surfaces (parapophysis and diapophysis) appear to be missing in all vertebrae. The preserved cross-section of the transverse process of the first vertebra is not significantly larger than in other vertebrae. The free vertebrae have a high transverse process but do not have an enlarged parapophyses (articular facet for the rib capitulum). They have a joined articular facet for a single-headed rib, and this facet is distinctly smaller than rib articular facets in anterior dorsals. In free dorsals, the height of the transverse process is more or less uniform. In BEXHM 2015.18, the transverse process of the first complete dorsal vertebra is, however, higher than in other vertebrae. In the two succeeding vertebrae, the height of the transverse process decreases and then slightly increases in the more posterior vertebrae. In Pteranodon , the dorsal vertebrae with a similarly high transverse process is the fifth dorsal, which is followed by dorsals with the decreased height of the transverse process ( Bennett 2001: fig. 46). The fifth and sixth dorsals are incorporated in the notarium in Pteranodon . The lower neural spines of dorsal vertebrae in BEXHM 2015.18 also suggest that they are anterior dorsals, not posterior dorsals, as was assumed by the Rigal et al. (2017). It is either this specimen is immature with unfused notarium, and missing supraneural plate or notarium was shorter in this taxon.

Rigal et al. (2017) identified in BEXHM 2015.18 a distal fragment of the left ulna and right proximal syncarpal. However, the ulna is the right one according to its morphology. It is more likely to find the association of the bones from a wing of the same side, not a mixture of right and left wing elements, especially for the distal ulna and proximal syncarpal that are articulated in life and closely approximated postmortem in BEXHM 2015.18. The exposed flattened surface of the ulna indeed accommodated the radius, but this is the anterior ulna surface, not posterior surface, as indicated in Rigal et al. (2017: fig. 4b). The proximal syncarpal retains a suture between its ulnar and radial parts. This is in line with likely not ossified notarium in BEXHM 2015.18 and suggests a juvenile age for this specimen.

Rigal et al. (2017) also indicated fragments of the second, third, and fourth phalanges of the left wing digit. However, these fragments are poorly preserved and do not preserve any characters that may suggest an attribution to the left or right side wing. As was noted above, all wing elements in BEXHM 2015.18 appear to be from the right side. Also, I cannot follow the attribution of these fragments to the second, third, and fourth phalanges. Some of these fragments could be pieces of the first wing phalanx. The idea that a subtriangular cross-section of the fourth wing phalanx may be a diagnostic character for the Lonchodectidae is unfounded because identification of this piece of bone as a fourth wing phalanx is dubious and attribution of BEXHM 2015.18 to the Lonchodectidae is poorly supported.

BEXHM 2015.18 was referred to the Lonchodectidae because of flattened, simple cone-shaped tooth crowns in alveoli with raised borders ( Rigal et al. 2017). The latter character was considered an autapomorphy of the Lonchodectidae by Unwin (2001). Rigal et al. (2017) differentiated BEXHM 2015.18 from Serradraco sagittirostris by having vertically directed teeth and raised alveolae with symmetric borders. According to these authors, BEXHM 2015.18 agrees well in these features with Lonchodraco giganteus and for this reason this specimen was referred to Lonchodraco sp. However, the tooth inclination may vary within the jaw, with more vertical teeth anteriorly. The only known specimen of S. sagittirostris (NHMUK R1823) comprises posterior dentary fragments. I cannot confirm that the

tid Ikrandraco avatar ( Wang et al. 2014:

fig. 3e) has a similar plesiomorphic, not

“warped” deltopectoral crest with parallel

margins (contra the original description).

Still, this crest is distinctly shorter than

in “ Palaeornis .” The morphology of

NHMUK 2353 and 2353a was reviewed

by Averianov (2012), who considered it as

belonging to a basal azhdarchoid.

The “Moon Goddess.” A well pre-

served pterosaur skeleton from the Ap-

tian Jiufotang Formation of Liaoning

Province, China has been attributed to

still undescribed lonchodectid taxon, in-

formally known as the “Moon Goddess,”

or “Chang-e” ( Unwin et al. 2008; Witton

2013). Lonchodectids are present in the

Jiufotang Formation, as shown by Ikran-

draco avatar .

Unwindia trigonus Martill, 2011 .

This taxon is based on a partial rostrum

SMNK PAL 6597 from the Albian Fig. 5. Dentary fragments BEXHM 2015.18 (A) and NHMUK R1823, holotype of Serradraco sagittirostris ( Owen, 1874) (B) (after Rigal et al., 2017: figs Romualdo Formation in Santana do Cari- 5a and 7c). Arrows indicate asymmetric alveolar borders. Scale bar = 10 cm. ri region, Ceará Province, Brazil ( Martill

2011). The genus is characterized by teeth in BEXHM 2015.18 have raised alveoli with reduced dentition with only seven tooth pairs antesymmetric borders. This interpretation is based on rior to the nasoantorbital fenestra. The teeth are of a poorly preserved part of the jaw with most of the similar size, in contrast with the heterodont dentition bone is missing. Moreover, the better preserved al- of ornithocheiroid taxa. Originally, the taxon was veolar margin posterior to the preserved teeth shows referred to the Ctenochasmatoidea ( Martill 2011). the asymmetric alveolar borders, so characteristic Witton (2013: 211) cited personal communication for S. sagittirostris (Fig. 5). Serradraco sagittirostris from D.M. Unwin that “several features of its jaw and comes the Upper Tunbridge Wells Sand Formation, dental morphology are consistent with a lonchodectid from where BEXHM 2015.18 may also come. The identity.” Indeed, the long and low rostrum resembles latter specimen is more likely belong to S. sagitti- that of Lonchodectes and Ikrandraco , and the variarostris and thus excluded from the Lonchodectidae . tion of the tooth size is not significant. Nevertheless, Palaeornis cliftii Mantell, 1844 . This species the tooth size variation is more pronounced than in is based on an isolated humerus broken in two parts lonchodectids: the terminal teeth are somewhat big- (NHMUK 2353 and 2353a) from the Hastings Beds ger and separated from the lateral teeth by a small Group (probably from the Valanginian Upper Tun- tooth. This condition is more reminiscent of the orbridge Wells Formation) at Cuckfield, West Sussex, nithocheiroid dentition. Unwindia could represent an England ( Witton et al. 2009: figs 3–4). Witton et al. ancestral stage in teeth reduction leading to the Lon- (2009) referred this specimen to the Lonchodectidae chodectidae, but cannot be included in that group. indet. based on its similarity with isolated humerus Prejanopterus curvirostris Fuentes Vidarte et CAMSM B 54.081 from Cambridge Greensand, Meijide Calvo, 2010. Prejanopterus is known from which Unwin (2003) referred to Lonchodectes sp. disarticulated cranial and postcranial elements of However, there are no grounds for referral of the several individuals from the Aptian Leza Formation latter humerus to Lonchodectes , and reference of at Fuente Amarga, Préjano, La Rioja Province, Spain the “ Palaeornis ” humerus to the Lonchodectidae is ( Fuentes Vidarte and Meijide Calvo 2010). Prejansimilarly groundless. The humerus of the lonchodec- opterus was originally referred to Pterodactyloidea

incertae sedis, but then assigned to the Pterodactylidae ( Pereda Suberbiola et al. 2012) . According to Witton (2013), “the long, low nature of its jaw and unusual tooth socket morphology tentatively suggest it might represent a lonchodectid.” Similarities with the Lonchodectidae include a low elongated rostrum lacking the premaxillary crest (present in Lonchodraco but not in Lonchodectes and Ikrandraco ) and small homodont teeth. However, Prejanopterus lacks raised alveolar margins, palatal ridge, and mandibular crest. These characters do not allow the attribution of Prejanopterus to the Lonchodectidae .