A new lizard (Squamata) was the last meal of Compsognathus (Theropoda: Dinosauria) and is a holotype in a holotype

A new lizard (Squamata) was the last meal of Compsognathus (Theropoda: Dinosauria) and is a... Abstract A lizard preserved in the abdominal cavity of the Compsognathus longipes holotype traditionally has been referred to Bavarisaurus macrodactylus Hoffstetter. New observations on that specimen reveal that it differs from B. macrodactylus in the shape of the frontoparietal suture, the presence of paired parietals and the orientation of the parietal supratemporal processes. Phylogenetic analysis of 220 saurians and 839 morphological characters demonstrate the paraphyly of Ardeosaurus as traditionally used and the distinctiveness of three major squamate radiations in the Solnhofen Formation, and suggests rhynchocephalian affinities for the early lepidosaur Bharatagama rebbanensis. The gut content specimen is a diagnostic new species – a squamate holotype in a dinosaur holotype. It is an ardeosaurid based on the presence of a posteroventral process of the jugal, absence of a lateral parietal flange at the frontoparietal suture and a posteriorly broadened mandibular retroarticular process. Ardeosaurus, Bavarisaurus, Eichstaettisaurus, Palaeolacerta, Squamata INTRODUCTION Squamata is known from nearly 10 000 extant named species (Uetz, 2016) inhabiting essentially every non-polar environment (Pianka & Vitt, 2003). Squamata diverged from Rhynchocephalia by the Middle Triassic (Jones et al., 2013), but has a known unequivocal fossil record extending only to the Middle Jurassic (Evans, 1994a; Caldwell et al., 2015). Although earlier fossils have been suggested as representing squamates (Carroll, 1982; Evans, Prasad & Manhas, 2002; Datta & Ray, 2006), these suggestions have either been refuted (Evans, 2001; Hutchinson, Skinner & Lee, 2012) or are untested (Bharatagama rebbanensis Evans et al., 2002; see below). Currently, the earliest known, definitive, squamate is from the Middle Jurassic of Britain (Evans, 1998; Caldwell et al., 2015). The earliest known complete squamates are from the Late Jurassic of China (Evans & Wang, 2007, 2009; Conrad et al., 2013) and Europe (Meyer, 1855; Nopcsa, 1908; Grier, 1914; Huene, 1956; Kuhn, 1958; Cocude-Michel, 1961; Hoffstetter, 1964, 1967; Ostrom, 1978; Mateer, 1982; Estes, 1983; Evans, 1994a, b; Evans & Barbadillo, 1998; Evans, 2003; Conrad, 2008). Germany hosts the richest record of complete Jurassic squamates so far known, and these occur within the Solnhofen Formation from the Late Jurassic (Fig. 1). Five squamate species have been named from these limestones. Ardeosaurus brevipes Meyer, 1855, Ardeosaurus digitatellus Grier, 1914, Bavarisaurus macrodactylus Hoffstetter, 1953, Eichstaettisaurus schroederi Kuhn, 1958 and Palaeolacerta bavarica Cocude-Michel, 1961 are each known from complete or partial remains of articulated specimens from Solnhofen (see Estes, 1983; Evans, 1994b). Eichstaettisaurus schroederi and P. bavarica are each represented only by one specimen (Cocude-Michel, 1961; Cocude-Michel, 1963; Estes, 1983; Evans & Barbadillo, 1998). Figure 1. View largeDownload slide World palaeomap with modern pullout of Deutschland, highlighting the area from which SNSB-BSPG AS I 563b is recovered. The palaeomap is a reconstruction of the Late Jurassic and is drawn after Blakey (2011). Figure 1. View largeDownload slide World palaeomap with modern pullout of Deutschland, highlighting the area from which SNSB-BSPG AS I 563b is recovered. The palaeomap is a reconstruction of the Late Jurassic and is drawn after Blakey (2011). Ardeosaurus is a problematic taxon with two named species. Ardeosaurus brevipes is known by a cast of the holotype (NHM PR 38006) and a referred specimen (PMU.R58; Meyer, 1855; Camp, 1923; Estes, 1983). Ardeosaurus digitatellus is known only from the holotype (CM 4026; Cocude-Michel, 1963; Estes, 1983), but a second specimen (SNSB-BSPG 1923 I 501) has been referred to A. cf. digitatellus by Estes (1983). The relationships of these four specimens have not been cladistically tested. Bavarisaurus macrodactylus is known from the holotype, but a second specimen has been referred to that species (Huene, 1956; Evans, 1994b) or to Ba. cf. macrodactylus (Ostrom, 1978; Estes, 1983). The referred specimen of Bavarisaurus is preserved inside the abdomen of the holotype of the theropod dinosaur Compsognathus longipes Wagner, 1861 (Figs 2, 3; see Ostrom, 1978; Estes, 1983; Evans, 1994b). This specimen is of uncertain provenance and may come from the Solnhofen Formation, but might be from the Painten, Torleite or Rögling Formations with a likely age of either latest Kimmeridgian or earliest Tithonian (Viohl, 2000; Rauhut et al., 2012; Reisdorf & Wuttke, 2012). Figure 2. View largeDownload slide Body of Compsognathus longipes (holotype; SNSB-BSPG AS I 563) with the skeleton of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b) inside it. See also Figure 3. Figure 2. View largeDownload slide Body of Compsognathus longipes (holotype; SNSB-BSPG AS I 563) with the skeleton of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b) inside it. See also Figure 3. When the morphology of Ba. macrodactylus was reviewed in 1994, the gut content specimen was considered to be representative of Ba. macrodactylus (Evans, 1994b). Even so, in observing the holotype (Fig. 4) and the referred specimen (Figs 2, 3), several morphological differences became apparent. As a result, I revisited the morphology and/or the published data regarding the morphology of all of the available Solnhofen Jurassic squamates and have undertaken a comparative study of these specimens. The current paper has four major goals. Figure 3. View largeDownload slide Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b) as it lies in the body of Compsognathus longipes (SNSB-BSPG AS I 563). A, restored body silhouette of the dinosaur constructed and restored body silhouette showing its hypothesized orientation inside the gut based on the preserved skeletal associations. B, drawing of the lizard skeleton and the associated bones of the dinosaur. Note that the right femur is not currently preserved, but is represented by a cast of the specimen made before further preparation led to the femur’s removal. Figure 3. View largeDownload slide Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b) as it lies in the body of Compsognathus longipes (SNSB-BSPG AS I 563). A, restored body silhouette of the dinosaur constructed and restored body silhouette showing its hypothesized orientation inside the gut based on the preserved skeletal associations. B, drawing of the lizard skeleton and the associated bones of the dinosaur. Note that the right femur is not currently preserved, but is represented by a cast of the specimen made before further preparation led to the femur’s removal. Figure 4. View largeDownload slide Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501). Figure 4. View largeDownload slide Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501). Identify the differences between Ba. macrodactylus and the squamate preserved within the abdomen of the C. longipes holotype (SNSB-BSPG AS I 563b). Establish the identity of SNSB-BSPG AS I 563b. Present a phylogenetic analysis of Jurassic lizards with a focus on the parts of the tree pertaining to the Solnhofen specimens, including all of the putative Ardeosaurus species. MATERIAL AND METHODS Institutional abbreviations AMNH, American Museum of Natural History, New York City, NY; SNSB-BSPG, Staatliche naturwissenschaftliche Sammlungen Bayerns-Bayerische Staatssamlung für Paläontologie und Geologie, Munich, Germany; CM, Carnegie Museum of Natural History, Pittsburgh, PA; FMNH, Field Museum, Chicago, IL; GM, Geiseltal Museum, Martin-Luther University, Halle/Saale, Germany; IGM, Institute of Geology, Mongolian Academy of Sciences – American Museum of Natural History Expeditions, field numbers; IVPP, Institute of Vertebrate Paleontology and Paleoanthropology, Beijing, People’s Republic of China; KNM, Kenyan National Museums, Nairobi, Kenya; LACM, Los Angeles County Museum, Los Angeles, CA; MB, Museum für Natkurkunde, Berlin, Germany; MCZ, Museum of Comparative Zoology, Cambridge, MA; MINS, Missouri Institute of Natural Science, Springfield, MO; MNHN, Muséum national d’Histoire naturelle, Paris, France; MPN, Museo di Paleontologia, Napoli, Italy; NHM, The Natural History Museum, London, UK; NMNH, National Museum of Natural History, Washington, D.C.; PMU, Paleontological Museum of Uppsala University, Uppsala, Sweden; SDSMT, South Dakota School of Mines and Technology, Rapid City, SD; UCL, University College London, London, UK; UCMP, University of California Museum of Paleontology, Berkeley, CA; UF, Florida State Museum (University of Florida), Gainesville, FL; USNM, National Museum of Natural History, Smithsonian Institution, Washington, D.C.; YPM, Yale Peabody Museum, New Haven, CT; ZPAL, Zakład Paleobiologii, Polska Akademia Nauk (Paleobiological Institute, Polish Academy of Sciences), Warsaw, Poland. Abbreviations a, angular; as, astragalus; asca, astragalocalcaneum; bc, basioccipital c, coronoid; CA, caudal vertebra; colf, columellar fossa; d, dentary; Dr, dorsal rib; e, epipterygoid; ec, ectopterygoid; ecf, ectepicondylar foramen; ecp, ectepicondyle; ent, entepicondyle; f, frontal; fe, femur; fi, fibula; h, humerus; inc, intercentrum; is, ischium; j, jugal; L, left; l, lachrymal; n, nasal; m, maxilla; ma, manus; mn, mandible; mt. metatarsal; nea, neural arch; of, obturator foramen; oo, opisthotic; or, orbit; p, parietal; pa, palatine; pbs, parabasisphenoid; pe, pes; pf, postfrontal; pif, pineal foramen; plr, palatine ramus; pm, premaxilla; po, postorbital; pra, prearticular; prf, prefrontal; pt, pterygoid; pu, pubis; q, quadrate; qpr, quadrate process; R, right; r, radius; rac, radial condyle; rap, retroarticular process; S1, first sacral vertebra; S2, second sacral vertebra; Sr, sacral rib; sa, surangular; sk, skull; so, supraoccipital; sq, squamosal; st, supratemporal; stf, supratemporal fenestra; ti, tibia; tvp, transverse process; ul, ulna; ulc, ulnar condyle. Specimen reconstructions Skull and whole-body reconstructions were made from direct observations, new photographs and/or published photos and interpretive drawings. For those reconstructed taxa that were directly observed (BMNH PR 38006, SNSB-BSPG 1923 I 501, CM 4026, SNSB-BSPG 1873 III 501 and SNSB-BSPG 1937 Ia, b), digital photographs and notes were taken, and physical and digital line drawings were made with the specimen present. Where useful, I used published illustrations and photographs to help interpret these specimens. When specimens were not directly observed (i.e. PMU.R58 and MPN 539), published photos and interpretive drawings were used and the reconstructions are less detailed and more tentative. Digital line drawings were made in Adobe Photoshop CS6 (Adobe Systems, 2010). The midline was identified for each specimen and the alignments were modified accordingly for reconstruction. When one side of the specimen was more complete or better preserved, it was copied and reversed as a separate Photoshop layer to help reconstruct both sides of the skull. Individual bones were treated similarly and aligned, often as semi-transparent layers, to help reconstruct the specimen. Each illustration is based on a single specimen. SYSTEMATICS Squamata Oppel, 1811 Eichstaettisauridae Kuhn, 1958 Schoenesmahl dyspepsia gen. et sp. nov. Type species:Schoenesmahl dyspepsia sp. nov. Diagnosis: As for the type and only known species. Etymology: Schoenesmahl (Deutsch: schöne Mahl ‘beautiful meal’ where ‘beautiful’ is used here as a synonym of elegance), referring to the fact that it was the last meal of the holotype of C. longipes Wagner, 1861 whose generic name means ‘elegant jaw’ or ‘comely jaw’ (Greek). Dyspepsia (Greek: ‘difficult digestion’), referring to the undigested nature of this last meal of C. longipes. Holotype: SNSB-BSPG AS I 563b, an incomplete skeleton lacking nasals, vomers, palatines, postorbitals, quadrates, anterior presacral vertebrae, pectoral girdles, most of the radii and ulnae, manus, ilium, ischium, tarsals and the distal pedal phalanges from digits I, II, IV and V. Type locality and horizon: Kelheim (Land Bayern; Fig. 1), Solnhofen Formation (Tithonian) of Germany (Wagner, 1861). Known distribution: Known only from the type locality and horizon. Diagnosis: Schoenesmahl dyspepsia differs from Ba. macrodactylus in the possession of a U-shaped (rather than a W-shaped) frontoparietal suture, paired parietals and anteroposteriorly (rather than mediolaterally oriented) parietal supratemporal processes. Schoenesmahl dyspepsia differs from E. schroederi in possessing a weakly inclined nasal process and an anteroposteriorly elongate prefrontal, frontals with parallel-sided interorbital margins (rather than an hourglass-shaped frontal). Schoenesmahl dyspepsia differs from Eichstaettisaurus gouldi Evans, Raia & Barbera, 2004 in possessing posteriorly broadened retroarticular process of the articular. Schoenesmahl dyspepsia differs from E. schroederi and E. gouldi in lacking well-developed subolfactory processes of the frontal. DESCRIPTION Skull form Cranial form Although the skull is mostly disarticulated and slightly spread apart (Fig. 5), many details may be reconstructed (Fig. 6). The snout is elongate relative to the orbital and temporal areas (sensuMontero & Gans, 1999). This differs from the recent rhynchocephalian Sphenodon punctatus Gray, 1842, which possesses a relatively short snout compared to the post-snout part of the skull; however, basal rhynchocephalians have relatively elongate snouts (Evans, 1980; Fraser, 1982), suggesting that this may be the more plesiomorphic lepidosaur condition. Figure 5. View largeDownload slide Skull elements of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b). A, photograph. B, interpretive drawing. Dark grey areas in (B) indicate areas preserved by faint impressions and/or outlines. Lighter grey areas indicate areas where the bone surfaces have been lost, but some bone is preserved. White areas indicate bones or clear bone impressions. Figure 5. View largeDownload slide Skull elements of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b). A, photograph. B, interpretive drawing. Dark grey areas in (B) indicate areas preserved by faint impressions and/or outlines. Lighter grey areas indicate areas where the bone surfaces have been lost, but some bone is preserved. White areas indicate bones or clear bone impressions. Figure 6. View largeDownload slide Reconstruction of the skull of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b). A, left lateral view. B, dorsal view. Semi-opaque shadows indicate elements or bone surfaces not preserved. White areas indicate bones or clear bone impressions. C, only the preserved bones in left lateral view and D, dorsal view. Figure 6. View largeDownload slide Reconstruction of the skull of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b). A, left lateral view. B, dorsal view. Semi-opaque shadows indicate elements or bone surfaces not preserved. White areas indicate bones or clear bone impressions. C, only the preserved bones in left lateral view and D, dorsal view. The exact shape of the external naris remains uncertain due to the disarticulated nature of the skull bones and lack of preserved nasal bones. The external naris appears to have been relatively elongate as compared to the many lepidosaurs (e.g. Sp. punctatus, Abronia mixteca Bogert & Porter, 1967, Xenosaurus grandis Gray, 1856 and many others). This is suggested by the gently sloping anterior margin of the maxillary nasal process and the elongate nasal process of the premaxilla (Figs 5–7). The orbits are not particularly large and their greatest reconstructed diameter is slightly less than the probable length of the antorbital snout (Fig. 6). Based on the length of the supratemporal processes of the parietal and the parietal itself (Fig. 5), the supratemporal fenestrae may have been significantly smaller than the orbits (Fig. 6). The elements of the palate and their contacts are not sufficiently known to confidently reconstruct the size and shape of the suborbital fenestrae or the interpterygoid vacuity. Figure 7. View largeDownload slide Detail of the area of the skull of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) showing the impression of the left maxilla and some of the surrounding bones. Figure 7. View largeDownload slide Detail of the area of the skull of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) showing the impression of the left maxilla and some of the surrounding bones. Mandibular form The robust lower jaw is preserved as a combination of bones and impressions of bone (Figs 3, 5). The long axis of the mandible is straight (Fig. 6). The dentary constitutes more than one-half the length of the mandible. No data are available for the state of Meckel’s canal. Dermal skull roof Premaxilla Parts of each premaxilla are preserved (Fig. 5). The right is preserved as an impression, but the left is partially preserved in lateral view. The preserved parts of the left premaxilla include a robust body, two teeth, and part of the elongate, posterodorsally oriented, nasal process. The posterior part of the nasal process is preserved only as an impression. The paired premaxillae form the anterior margin of the external naris. The premaxillary body is narrow. The right premaxillary impression includes three tooth positions, but the original number of tooth positions remains uncertain. The preserved parts indicate that three or four were probably present on each premaxilla. The breadth of the complete left dental margin appears to be too small for any more than four tooth positions. The nasal process tapers distally. Its posterior surface has a distinct and deep ventral ridge. The posterior margin of this is vertical to a level just dorsal to the level of the main premaxillary body. Distal to that level, the nasal process consistently tapers toward its tip. The body of the premaxilla is joined with the posterior nasal process ridge by a dorsomedially oriented flange (Fig. 5). Anterior to this flange is a minor depression that would have lain in the anteroventral edge of the bony external naris. Based on the shape of the anterolateral edge of the body of the premaxilla, the premaxilla narrowly overlaid the premaxillary process of the maxilla anterolaterally. The presence and nature of the anterior ethmoidal foramina is unknown. Bavarisaurus macrodactylus possesses paired premaxilla bearing very narrow nasal processes (Fig. 8). The premaxilla is preserved as a combination of bone and impression. The body of the premaxilla is, apparently, similar in width relative to the snout tip as it is in S. dyspepsia. Two tooth positions are preserved as infilled impressions from the eroded premaxillary body, but there probably three or four tooth positions per premaxilla. Figure 8. View largeDownload slide The skull of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501). The skull is preserved in ventral view, meaning that the bone impressions preserved on the matrix represent dorsal views. A, photograph of the skull and partial mandible as preserved. B, reconstructed skull. White areas indicate parts preserved in bone. Semi-opaque grey areas overlapping white (e.g. premaxilla) indicate areas represented by bone, but on surfaces that are not visible in the reconstructed dorsal view. Black lines indicate visible suture patterns and/or bone margins. Grey areas without outline indicate unknown elements tentatively restored here. Figure 8. View largeDownload slide The skull of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501). The skull is preserved in ventral view, meaning that the bone impressions preserved on the matrix represent dorsal views. A, photograph of the skull and partial mandible as preserved. B, reconstructed skull. White areas indicate parts preserved in bone. Semi-opaque grey areas overlapping white (e.g. premaxilla) indicate areas represented by bone, but on surfaces that are not visible in the reconstructed dorsal view. Black lines indicate visible suture patterns and/or bone margins. Grey areas without outline indicate unknown elements tentatively restored here. Eichstaettisaurus schroederi possesses a short and broad premaxillary nasal processes on its paired premaxillae (Fig. 9). A proximomedial emargination that appears to be bilaterally preserved in SNSB-BSPG 1937 I 1 may be an emargination associated with an ethmoidal nerve. Gauthier et al. (2012) did not code this structure as if it were an ethmoidal branch and the nature of this condition remains uncertain. Figure 9. View largeDownload slide The skull of Eichstaettisaurus schroederi (holotype; SNSB-BSPG 1937 I 1a) in dorsal view. A, photograph of skull. B, reconstructed skull. Semi-opaque grey layer indicates missing parts restored here. Figure 9. View largeDownload slide The skull of Eichstaettisaurus schroederi (holotype; SNSB-BSPG 1937 I 1a) in dorsal view. A, photograph of skull. B, reconstructed skull. Semi-opaque grey layer indicates missing parts restored here. The premaxillae are paired in all Ardeosaurus specimens for which they are preserved (CM 4026, SNSB-BSPG 1923 I 501 and NHM PR 38006). As a unit, the premaxillae have a broad body and a moderately broad nasal process as indicated by the preserved part of the broken base and the distal part overlying the nasals. Premaxillary teeth are visible in none of the Ardeosaurus specimens observed here, and they were hidden in the specimen described by Mateer (1982). Palaeolacerta bavarica appears to possess an unpaired premaxilla with an elongate nasal process (Hoffstetter, 1964; Estes, 1983). Because of damage to the specimen and the nature of the published figures, the breadth of the premaxillary body is uncertain. Maxilla Only the left maxilla is known and it is preserved only as an impression on the rock (Figs 5, 7). This impression retains its contacts with the left lacrimal and jugal and has been moved slightly out of position with the ectopterygoid contact. The lithographic limestone preserves exquisite details of the lateral (external) surfaces of the maxilla, including five of its teeth. The overlying postfrontal only slightly obscures details of the posterodorsal margin of the maxillary nasal process. The maxilla is subtriangular with a tall nasal process, an elongate premaxillary ramus and somewhat shorter suborbital ramus (Figs 5–7). The gently and continuously tapering premaxillary ramus is attenuated at its anterior end and was overlapped by the body of the premaxilla at the level of the dental margin. Its dorsal margin describes a gentle posterodorsal curve without a clear distinction between the subnarial surface and the anterior margin of the nasal process. There is, however, a small and weakly developed spur at a level that probably indicates the anterior limit of the articulation with the nasal bone (Fig. 7). The suborbital ramus apparently did not reach the level of the middle of the orbit. No dermal sculpturing is apparent on the lateral surface of the maxilla. Seven superior labial foramina are present near the dental margin of the maxilla. The posteriormost of these is posteroventrally oriented, whereas the other six are more anteromedially oriented. The most anterior of these foramina is located dorsal to the level of the other six. Additionally, two other small foramina are present more dorsally, on the nasal process of the maxilla. The more anterior of the two is located just dorsal to the third superior labial foramen; the more posterior of them lies slightly more dorsal than the anterior one and at a level between the third and fourth superior labial foramina. The maxillary suborbital ramus is more obtusely attenuated than the premaxillary ramus. As preserved, a small sliver of the ectopterygoid is visible ventral to the level of the jugal. However, it is unclear whether the ectopterygoid abuts the posterior tip of the maxillary suborbital ramus, or clasps it. Impressions of five tooth positions are associated with the left maxilla. The most anterior preserved tooth impression is present ventral to and slightly anterior to the anteriormost maxillary ethmoidal foramen. The second tooth impression is present immediately posterior to the first. The other three tooth impressions are located below the level of the last four maxillary ethmoidal foramina. Few details are available regarding the maxilla of Ba. macrodactylus. Eichstaettisaurus schroederi has well-preserved maxillae, but details of the dental margin are hidden as preserved. Eichstaettisaurus schroederi differs from S. dyspepsia in that the nasal process is steeply inclined and the apex of the nasal process lies anterior to the midpoint of the maxilla (Fig. 9). The posterior margin of the nasal process also has a stronger overlap of the anterior part of the lacrimal in E. schroederi. Few details are available for the maxilla in Ardeosaurus. The holotype of A. brevipes offers the clearest view of the morphology. The nasal process is dorsally tall and possessed what appears to be a steep posterior margin to the external naris. Between eight and ten superior labial foramina are probably present, but the exact number is uncertain. The maxilla is mostly ventral to (rather than lateral to) the suborbital process of the jugal. The relationships between the maxilla and the prefrontal and frontal remain unclear. Prefrontal Although preserved with only its ventral aspect visible (Figs 5, 7), the prefrontal is well preserved. It is a robust bone with a mediolaterally broad palatine process. The nasal cavity area is concave and generally similar to that of other non-snake squamates (e.g. Oelrich, 1956; Conrad, 2004). The orbital process of the prefrontal is elongate and extends to approximately the level of the middle of the orbit, fitting into a shallow groove on the ventrolateral aspect of the frontal. The orbital process of the prefrontal is similarly long in Ba. macrodactylus (Fig. 8), but is much shorter in E. schroederi (Fig. 9). As illustrated by Evans et al. (2004), the Catalonian Eichstaettisaurus sp. has prefrontals bearing orbital processes that are longer than those of E. schroederi, but that do not reach the midpoint of the orbit. All of the observed Ardeosaurus specimens and the specimen described by Mateer (1982) are similar in that the orbital processes of their prefrontals extending nearly to the midpoint of the orbit. The specimens SNSB-BSPG 1923 I 501 and NHM PR 38006 are similar in having subtriangular orbital processes and, apparently, ovate anterior processes. By contrast, CM 4026 has a laterally concave margin such that the orbital processes become very narrow posteriorly. The orbital process of PMU.R58 is subtriangular (see Mateer, 1982), but seems less continuously anterolaterally expanded than in SNSB-BSPG 1923 I 501 and NHM PR 38006. Although the prefrontals are incompletely preserved in P. bavarica, the orbital processes seem to be mostly preserved and complete (Hoffstetter, 1964). The orbital process is more similar to that known for PMU.R58 than other known Solnhofen lizards. Lacrimal The left lacrimal is preserved, and only as a clear impression of a bone articulated with the jugal and maxilla (Figs 5–7). The jugal contact with the lacrimal is gently bowed posteriorly. The maxillary contact with the lacrimal demonstrates that the maxilla did not reach the orbital margin. A small lateral ridge is present at the orbital rim and is continuous with the same on the anterior part of the jugal. Although part of the right lacrimal is preserved in Ba. macrodactylus (Fig. 8), no details are apparent. The bilaterally preserved lacrimals of E. schroederi are small and simple. They are only slightly medially emarginated for contribution to the lacrimal foramen. Although the maxilla probably did not reach the orbital margin in E. schroederi, it did more extensively overlie the lacrimal. The E. schroederi lacrimal apparently lacks the small orbital ridge (Fig. 9) seen in the lacrimal of S. dyspepsia. It is impossible to determine definitive presence or absence of a lacrimal in any of the known Ardeosaurus based on specimens or Mateer (1982). Jugal An incomplete left jugal is preserved in articulation with the lacrimal and maxilla in S. dyspepsia (Figs 5, 7). Most of the preserved part is visible in medial view, but the anterior bit is preserved as an impression of the lateral surface. The part preserved as bone is eroded and lacks the natural bone surfaces. Its posterior part is narrowly obscured by the overlying left pterygoid and ectopterygoid. Because the postorbital process is preserved in medial view, there is no way to tell whether it possesses dermal sculpturing. The posterior and ventral margins of the jugal describe an obtuse angle. The suborbital part is curved somewhat anterodorsally and its dorsal and ventral margins are continuous with those of the lacrimal. Based on the impression of the lateral aspect of the jugal, the suborbital process of the jugal lies mostly dorsal to the maxilla (Figs 5, 6), as opposed to medial to it as in many squamates (e.g. most iguanians, most scincids and most gekkotans). At its anterior end, the jugal shows a continuation of the orbital ridge from the lacrimal, but this quickly disappears posteriorly. A small posteroventral process is preserved. It is directed posteriorly. The postorbital bar is similarly broad to the suborbital process. Eichstaettisaurus schroederi and A. brevipes share with S. dyspepsia the condition the jugal lying mostly dorsal to the maxilla. This differs from the condition present in the putative basal gekkonomorph Norellius nyctisaurops Conrad & Daza, 2015, and all of the modern gekkotans observed here. Presence or absence of a posteroventral process of the jugal cannot be confirmed in E. schroederi, but one is apparently absent in E. gouldi (Evans et al., 2004). This process is present in Ardeosaurus specimens for which it may be scored (Mateer, 1982; NHM PR 38006, SNSB-BSPG 1923 I 501). Only a small piece of the postorbital ramus of the jugal appears to be preserved in P. bavarica (Hoffstetter, 1964), but this may be a fragment of the mandible. Postfrontal and postorbital The left postfrontal is preserved slightly out of normal articulation with the frontal and parietal. It is triradiate and each side is nearly equal in length (Figs 5, 6). The frontal process tapers anteriorly and forms a slightly obtuse angle with the parietal process. The postfrontal spans the frontoparietal contact. The parietal process ends more bluntly than the frontal process. The postorbital process is attenuated. Presumably, the space between the postorbital and parietal processes originally bore a facet receiving the postorbital bone, but the state of preservation does not allow that to be confirmed. No postorbital is preserved. Small bits of the postorbital and postfrontal are preserved in Ba. macrodactylus and many additional details are visible as impressions on the matrix (Fig. 8). The postorbitofrontal bar is a robust and anteroposteriorly elongate structure. As in S. dyspepsia, the postfrontal of Ba. macrodactylus clasped the frontoparietal suture. However, the frontal process is much more robust than the parietal process in Ba. macrodactylus and the postorbital process is much broader and distally rounded. The postorbital of Ba. macrodactylus also is robust and its main body is anteroposteriorly and mediolaterally broad. Exclusive of its posterior (squamosal) process, it is nearly subequal in anteroposterior and mediolateral lengths. The squamosal process is attenuated and shorter than the main body of the postorbital. The postfrontal and postorbital are well preserved in E. schroederi (Fig. 9). The postfrontal clasps the frontoparietal contact with a relatively small frontal process and an elongate parietal process. The lateral margin of the postfrontal curved in dorsal view receives the concave medial surface of the postorbital. The posterior part of the postfrontal and the posteromedial surface of the postorbital body are slightly depressed, forming a shallow supratemporal fossa at the anterior end of the supratemporal fenestra. The robust postfrontal is mediolaterally and anteroposteriorly broad and possesses a robust squamosal process that tapers posteromedially. The specimen of Ardeosaurus described by Mateer (1982) possesses a postfrontal with a weakly forked medial surface that spans the frontoparietal suture. The frontal narrowly overlaps the anterior part of the postfrontal. The frontal process is approximately one-third the length of the parietal process. The postorbital is robust and elongate. The postorbital–postfrontal contact is elongate and is equivalent to more than one-half the length of the supratemporal fenestra. An apparent suture remains between the postfrontal and postorbital in NHM PR 38006 (Fig. 10A), but it is poorly defined. A similar suture is faintly visible in SNSB-BSPG 1923 I 501 (Fig. 10B). The area is damaged in CM 4026. Two small pieces of bone visible in the photograph in Hoffstetter (1964) appears to represent the left postfrontal and a fragment of the postorbital. The postfrontal is anteroposteriorly elongate and spans the frontoparietal suture. If these elements are correctly identified, they demonstrate that the two bones are unfused. Figure 10. View largeDownload slide A, the skull of NHM PR 38006 (cast of the holotype of Ardeosaurus brevipes) in dorsal view. B, the skull of SNSB-BSPG 1923 I 501 (labelled A. cf. brevipes) in dorsal view. Figure 10. View largeDownload slide A, the skull of NHM PR 38006 (cast of the holotype of Ardeosaurus brevipes) in dorsal view. B, the skull of SNSB-BSPG 1923 I 501 (labelled A. cf. brevipes) in dorsal view. Squamosal The posterior part of a squamosal, probably the right, is reserved near the parietal (Figs 5, 6). It shows the characteristic posteroventral downcurve that is common to most squamates. There is no indication of a dorsal process. This fragment is similar to the more complete squamosals present in Ba. macrodactylus (Fig. 8) and E. schroederi (Fig. 9). The squamosals of the Ardeosaurus specimens NHM PR 38006 and SNSB-BSPG 1923 I 501 are robust, but poorly preserved (Fig. 10). The posterior parts of the left squamosal in NHM PR 38006 and the right in SNSB-BSPG 1923 I 501 are the best preserved and each seems to demonstrate the absence of a dorsal process. The squamosals are well preserved in CM 4026 and clearly lack dorsal processes. By contrast, the specimen described and illustrated in Mateer (1982) demonstrates the presence of a robust dorsal process. Frontal The unpaired frontal is preserved in ventral view (Figs 5, 7). The anterior part is preserved and exposed, as is the left posterolateral part. The middle of the prefrontal is obscured by overlying pieces of the right mandible and a piece of bone that is of unknown identity, but which may represent the nasal process of the right maxilla. Because the frontal is preserved in ventral view, it is impossible to tell if there was dermal sculpturing on its dorsal surface. The frontal is hourglass-shaped with expanded anterior and posterior ends (Figs 5–7). The posterior end is broadest, but the antorbital part is also expanded and more than 1.8 times the width of the narrowest part of the bone in the posterior part of the orbit. The nasofrontal suture is W-shaped as indicated by the impression of the dorsal surface of the anterior part of the frontal (Figs 6, 7). A midline anterior projection anterior projection of the frontal would have laid deep to the nasal(s) in the articulated specimen. Short anterolateral processes are also present and extend to about the same anterior level as the midline projection. These processes are parallel-sided and anterolaterally oriented. They may partly underlie or invade the contact between the prefrontal and nasal. Their lateral surfaces have prefrontal facets. Dorsal to the level of the prefrontal facet, the frontal has a thin but well-developed superficial lamina that would have partly overlain the frontal process of the prefrontal in natural articulation. There is no development of the crista cranii of the frontal (Fig. 7). Posterior to the midpoint of the orbit, the frontal is expanded laterally. The frontoparietal suture is broad and somewhat anteriorly curved. Similar to S. dyspepsia, E. schroederi possesses an anteriorly arched frontoparietal suture (Figs 9, 11), albeit a less arched one. By contrast, Ba. macrodactylus possesses a more sinuous frontoparietal suture wherein the parasagittal anterior extensions of the parietal extend far forward of the posteromedial extension of the frontal (Fig. 8). Eichstaettisaurus gouldi seems to possess a less well-developed version of this type of suture (see Evans et al., 2004), such that it is somewhat W-shaped. The frontoparietal suture is transverse in A. brevipes (Mateer, 1982). Figure 11. View largeDownload slide Part of the mandible, braincase, palate and skull roof of Eichstaettisaurus schroederi (holotype; SNSB-BSPG 1937 I 1a) visible in ventral view. The specimen is extensively visible in dorsal view (see Fig. 9), and has been partly prepared from ventral view to show some relevant elements. Figure 11. View largeDownload slide Part of the mandible, braincase, palate and skull roof of Eichstaettisaurus schroederi (holotype; SNSB-BSPG 1937 I 1a) visible in ventral view. The specimen is extensively visible in dorsal view (see Fig. 9), and has been partly prepared from ventral view to show some relevant elements. The frontals are paired and possess a transverse frontoparietal suture in all of the Ardeosaurus specimens considered here. Parietals and supratemporals The parietals are preserved as a combination of small bits of bone and bone impressions with the left parietal being much more completely represented than the right (Fig. 5). They remain in articulation with the frontal. A small piece of what may be a left palatine partly obscures the lateral margin of the left parietal. The right parietal apparently passes out of the exposed plane on the matrix and is largely lost beyond the level of the rim of the pineal foramen. The paired parietals are joined at the midline by a straight suture. Together, the two parietal bodies form a skull table that is subequal in length and width (Figs 5, 6). A large pineal foramen is present within the parietal. Its anterior margin lays approximately one-third of the parietal body length from the middle of the frontoparietal suture. The frontoparietal suture is a gentle anterior curve, lacking frontal tabs. The visible parts of the lateral parietal margin indicate that it was anteroposteriorly straight and unconstricted. Only the left supratemporal process is preserved, but it demonstrates the anteroposterior (rather than mediolateral or posterolateral) orientation. The supratemporal process was relatively short and was approximately one-third the length of the body of the parietal. A transverse posterior margin of the parietal is present between the anteromedial bases of the supratemporal processes. Apparently, the jaw adductor musculature attached ventrally on the parietal. No discernible supratemporal is preserved in S. dyspepsia. Bavarisaurus macrodactylus possesses a broad parietal table similar to that of S. dyspepsia. It has been suggested that the parietal was fused as a single element (Estes, 1983). A later study disagreed, identifying a faint midline suture (Evans, 1994b). A faint feature that may represent a suture is present, but only anteriorly and just to the right of the midline. Because it is not as well defined in its appearance as the other sutures visible in the skull and because it is visible only at one point, I posit that this short furrow represents something other than a suture and that the parietals form a single, fused, element as suggested by Estes (1983). Eichstaettisaurus schroederi has been suggested as possessing incompletely fused parietal bones (Estes, 1983; Evans & Barbadillo, 1998, 1999). Inspection of the holotype suggests that this is not the case. The parietals are fused (Figs 9A, 11). The apparent incompletely fused suture is actually a small break or crack along the midline of the bone caused by diagenesis as the middle part of the parietal was pushed ventrally and the lateral sides pushed dorsomedially. This is confirmed by the ventral view of the specimen where the anterior part of the parietal is visible (Fig. 11). The parietal supratemporal processes are directed almost completely mediolaterally in Ba. macrodactylus (Fig. 8), contrasting the condition present in S. dyspepsia (Figs 5, 6), E. schroederi (Fig. 9) and Ardeosaurus (e.g. Fig. 10; Hoffstetter, 1964; Mateer, 1982). The left supratemporal process is the only part of the parietal preserved as bone in Ba. macrodactylus. The supratemporal process remains in contact with the supratemporal. The supratemporal bone is elongate and retains short contacts with the posterior part of the squamosal. Bavarisaurus macrodactylus is generally considered to lack a pineal foramen (Cocude-Michel, 1961; Hoffstetter, 1964; Estes, 1983). One study suggested that the pineal foramen is present (Evans, 1994b), but this was based on the presence of pineal foramen in the S. dyspepsia specimen which was, at the time, considered to be a specimen of Ba. macrodactylus. Re-examination of the holotype of Ba. macrodactylus reveals no indication of a pineal foramen on the parietal impression (Fig. 8). All of the Ardeosaurus considered here possess elongate, fused, parietals with posteriorly oriented supratemporal processes that are more than one-half the length of the parietal table. These parietals lack anterolateral expansions at the frontoparietal suture and a small pineal foramen located at about the midpoint of the parietal table. They also preserve vermiculate dermal sculpturing on their dorsal surfaces. The specimen described by Mateer (1982) preserves a small and weakly developed posteromedial process of the parietal. This may also be present in NHM PR 38006, but it appears to be absent in PMU.R58. The badly crushed parietal of P. bavarica is clearly azygous (see Hoffstetter, 1964; Estes, 1983). Unlike other known Solnhofen lizards, a large parietal foramen is present at the frontoparietal suture (Hoffstetter, 1964; Estes, 1983). Although the posterior part of the parietal is more damaged than the anterior part, elongate supratemporal processes are visible mostly by their bony outlines, and extend posterolaterally (see Hoffstetter, 1964; Estes, 1983). Palate Pterygoid Both pterygoids are preserved in ventral view in S. dyspepsia (Figs 5, 12). The right pterygoid (Fig. 12) is exposed a short distance away from (anteroventral to) the maxilla and (posterodorsal to) the left premaxilla. The left pterygoid is preserved and lies between the parietal and the jugal. Both pterygoids retain their contacts with the ectopterygoids. The right pterygoid is incomplete anteriorly with its palatine process. The left pterygoid is more complete, but its quadrate process has been eroded. Figure 12. View largeDownload slide Right pterygoid of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in ventral view. Anterior is toward the bottom and slightly offset to the left. The presumed columellar fossa has partly collapsed, creating the small circle of broken bone, and the quadrate process is somewhat damaged. Note the presence of a few small pterygoid teeth anteromedial of the underside of the columellar fossa, at the base of the vomerine process. Figure 12. View largeDownload slide Right pterygoid of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in ventral view. Anterior is toward the bottom and slightly offset to the left. The presumed columellar fossa has partly collapsed, creating the small circle of broken bone, and the quadrate process is somewhat damaged. Note the presence of a few small pterygoid teeth anteromedial of the underside of the columellar fossa, at the base of the vomerine process. The triradiate pterygoid possesses an elongate quadrate process, a robust transverse process that ends in an ectopterygoid articulation, and a narrow palatine process. A short ventral ridge is preserved on each pterygoid near the confluence of the transverse and palatine processes. This ridge is best preserved on the right pterygoid wherein the bases of three small teeth and a pit associated with a fourth tooth position are preserved (Figs 5, 12). The robust transverse process extends anterolaterally and is medially joined with the palatine process by a broad anterior flange – a ‘web’ of bone that unites those two processes. As a result of this anterior flange, the posterior margin of the suborbital fenestra would have been broadly rounded in the articulated skull. The lateral margin of the transverse process extends posteromedially toward the more anteroposteriorly oriented medial border of the pterygoid. Thus, the lateral, medial and suborbital fossa margins of the pterygoid together form an isosceles triangle wherein the medial and lateral margins of the bone describe the most acute angle. The columellar fossa is not directly observable in either of the ventrally exposed pterygoids. Even so, each pterygoid shows a circular area of collapsed bone within the main body of the pterygoid, near the confluence of the three major processes. This area of collapsed bone may represent an area of weakness created by the overlying columellar fossa. The quadrate processes are eroded on their exposed surfaces in each pterygoid, but they clearly were elongate and mediolaterally compressed. The quadrate process is approximately 135% the length of the transverse process as measured from the presumed columellar fossa. Bavarisaurus macrodactylus preserves no part of the pterygoid. The pterygoids are visible through the orbits in E. schroederi (Fig. 9) and demonstrate a relatively broad and transversely oriented parietal–palatine contact. Additionally, the suborbital fenestra is relatively narrower at its posterior terminus in E. schroederi than in S. dyspepsia. Pterygoids are in the Ardeosaurus specimens PMU.R58 (Mateer, 1982). The transverse processes are relatively narrow and the anterior contact with the palatine is, apparently, transversely broad (Mateer, 1982). The quadrate process is not preserved or visible. Ectopterygoid The short and robust ectopterygoids are exposed in ventral view and each is partly eroded (Figs 5, 12). The left ectopterygoid remains partly in articulation with the left maxilla and jugal. The ectopterygoid is short and anterolaterally oriented. Each ectopterygoid preserves a posteromedially oriented suture with the pterygoid. Only a small part of the ectopterygoid is preserved in Ba. macrodactylus. All that can be said about it is that it is somewhat posteromedially oriented. The well-preserved ectopterygoid of E. schroederi (Fig. 9) is short and robust. It possesses a significant dorsal process, but not a strong posteromedial dorsal overlap of the pterygoid. It has a strong anteroposterior orientation. Ardeosaurus PMU R.58 preserves a robust and elongate ectopterygoid (see Mateer, 1982) with a strong dorsal ridge and an elongate anterolateral process contributing to the lateral margin of the suborbital fenestra. The dorsomedial overlap of the pterygoid extends to the main body of the pterygoid. Presence or absence of an anterior palatine contact is uncertain. Specimen NHM PR 38006 retains the posterior part of the ectopterygoid (Fig. 10A; see also Estes, 1983). Impressions of both ectopterygoids are preserved in SNSB-BSPG 1923 I 501 (Fig. 10B). Mandible The left mandible is preserved as a combination of bone (anteriorly and posteriorly; Figs 2, 3, 5, 13) and impressions of bones from the level of the coronoid to an area just anterior to the glenoid fossa. Only the dentary is exposed for the right mandible (Figs 5, 7). Figure 13. View largeDownload slide The left mandible of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) in lateral view. The dentary is incomplete posteriorly and most of the mandible has been lost from the posterior end of the dentary to the level of the mandibular glenoid region. The latter region and the retroarticular process is somewhat damaged, but note the presence of a clear mandibular outline on the underlying matrix. Figure 13. View largeDownload slide The left mandible of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) in lateral view. The dentary is incomplete posteriorly and most of the mandible has been lost from the posterior end of the dentary to the level of the mandibular glenoid region. The latter region and the retroarticular process is somewhat damaged, but note the presence of a clear mandibular outline on the underlying matrix. Because of the state of preservation and/or orientation of the skeletons, little of the mandible is visible in Ba. macrodactylus (Fig. 8A) and E. schroederi (Figs 9A, 11). The available Ardeosaurus material preserves part of the middle and posterior parts of the mandibles (Fig. 10; see also Hoffstetter, 1964; Mateer, 1982; Estes, 1983). Even E. gouldi, which is preserved in ventral view, offers few details beyond the general shapes of the dentary, surangular and parts of the prearticular (Evans et al., 2004). These elements are of a generalized squamate form in E. gouldi. The dentaries converge anteriorly, probably (although not certainly) being connected via a symphysis. The dentary possesses a posterior emargination such that the surangular is visible laterally between the surangular and angular processes. A small part of the angular is probably preserved on the left mandible. The prearticular may have possessed a short, tapering, retroarticular process. Dentary The dentary is straight along its longitudinal axis and the dorsal and ventral margins are parallel (Figs 2, 3, 13). Nine mental foramina are preserved on the preserved part of the dentary. Of these, fourth is relatively small compared to the rest. The mandible possesses only very limited exposure in Ba. macrodactylus (Fig. 8A). The anterior tip of the left dentary is preserved in dorsomedial review. The exposed part of the dentary seems to be straight. Four teeth are visible in that view. Only the lateral surface of the dentary is not available in E. schroederi (Fig. 11), and there is no informative morphology there. Coronoid An impression of the left coronoid preserves a view of part of its medial surface (Fig. 13). The coronoid is dorsally arched or chevron-shaped and bears at least a short dorsal coronoid process. The full of extent of the coronoid process is uncertain (Figs 6, 13) because it remains unclear whether the impression captures its full extent. The coronoid is not visible in Ba. macrodactylus. Only the dorsal part of the coronoid is visible in a meaningful way in E. schroederi. Its dorsal process is moderately elongate and dorsally rounded. It appears to be chevron-shaped as it is in S. dyspepsia and many other squamates, but not rhynchocephalians and most snakes. Angular Part of the left angular is preserved in articulation with the other known parts of the mandible. Only the posterior end is preserved (Figs 6, 13). It appears to show a posteriorly tapering element with a small dorsal expansion just anterior to the level of the mandibular glenoid. It lays lateral to the surangular and the prearticular (see below). Surangular The posterior part of the left surangular is preserved as a combination of bone and impressions in articulation with the rest of the left mandible (Figs 6, 13). Sutures with the coronoid are not clearly preserved as bone or as impressions. A small dorsal buttress is present just anterior to the mandibular glenoid. Prearticular Because the prearticular and articular commonly fuse in squamates, we follow the terminology of Conrad & Norell (2007) in referring to this element as the prearticular. The preserved part of the prearticular lies ventral to and posterior to the mandibular glenoid (Figs 6, 13). No clear suture with the surangular is preserved. The retroarticular process is posteriorly directed and shows moderate posterior broadening. A dorsal pit is preserved on the dorsal surface of the retroarticular process and some torsion is present. Dentition A total of 18 marginal teeth are visible in addition to the three pterygoid teeth mentioned above. Two premaxillary are preserved with the left premaxilla (Fig. 5). Impressions of three teeth are preserved with the impression of the right premaxilla (Fig. 5; see above). Five maxillary tooth impressions are preserved with the left maxilla (Figs 5, 7). Eight dentary tooth positions are preserved on the left mandible (Fig. 13). The teeth are closely spaced and curved with sharp tips. Tooth implantation cannot be determined because all of the marginal tooth-bearing bones are preserved in lateral view. Vertebral and costal morphology Presacral vertebrae and ribs Eight presacral vertebrae with elongate ribs are preserved in dorsolateral (anteriorly; Figs 3, 14, 15) and dorsal (posteriorly; Figs 3, 15) view and in association with the two sacral vertebrae. Thus, these vertebrae are the last eight presacrals and indicate the absence of lumbar vertebrae. Because the anterior presacrals are not preserved (Figs 2, 3), the length of the presacral column is unknown. The preserved presacral vertebrae are exposed in dorsal view, but their neural spines are all damaged such that their morphology is unknown (Fig. 15). The antepenultimate vertebra is heavily damaged such that the neural arch is missing. Figure 14. View largeDownload slide The preserved dorsal, appendicular and part of the caudal region of the skeleton of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in situ. Note the overlying ribs of Compsognathus longipes. A, photograph. B, reconstructed. White areas indicate parts preserved in bone. Figure 14. View largeDownload slide The preserved dorsal, appendicular and part of the caudal region of the skeleton of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in situ. Note the overlying ribs of Compsognathus longipes. A, photograph. B, reconstructed. White areas indicate parts preserved in bone. Figure 15. View largeDownload slide Posterior dorsal vertebrae and associated ribs of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in dorsal view. Figure 15. View largeDownload slide Posterior dorsal vertebrae and associated ribs of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in dorsal view. The preserved presacral vertebrae neural arches and associated pre- and postzygapophyses are short and robust. Each postzygapophysis possesses a transverse posterior margin. The presence or absence of intercentra or a notochordal canal can be neither confirmed nor denied. A clear synapophysis is preserved in the second preserved presacral vertebra. The presacral ribs associated with the more anterior preserved dorsal vertebrae are more than four times as long as the prezygapophyseal–postzygapophyseal length. By contrast, the last rib is about half as long as the penultimate rib and only about twice the length of the vertebra to which it attaches. Bavarisaurus macrodactylus clearly preserves intercentra in its presacral vertebrae (Fig. 16). It possesses between 23 and 25 presacral vertebrae (Fig. 4; Cocude-Michel, 1963; Evans, 1994b). I estimate the count at 25 based on the preserved bones and the bone impressions. The last dorsal vertebra has an elongate and apparently mobile rib (Fig. 17); there were no lumbar vertebrae. The last rib is approximately 1.3 times as long as the last vertebra. As preserved, the ribs appear to shorten more gradually as they approach the sacrum in Ba. macrodactylus than in S. dyspepsia. The antepenultimate rib is significantly longer than the penultimate, which is approximately 1.5 times as long as the last rib. Figure 16. View largeDownload slide Anterior dorsal vertebrae and left forelimb of Bavarisaurus macrodactylus (SNSB-BSPG 1873 III 501) as preserved in (generally) ventral view. Figure 16. View largeDownload slide Anterior dorsal vertebrae and left forelimb of Bavarisaurus macrodactylus (SNSB-BSPG 1873 III 501) as preserved in (generally) ventral view. Figure 17. View largeDownload slide Posterior dorsal vertebrae, partial sacrum and pelvis and anterior caudal regions of of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501) as preserved in (generally) ventral view. Figure 17. View largeDownload slide Posterior dorsal vertebrae, partial sacrum and pelvis and anterior caudal regions of of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501) as preserved in (generally) ventral view. Eichstaettisaurus schroederi possesses 29 presacral vertebrae (Fig. 18). The last five of these vertebrae possess shorter ribs than those immediately anterior to them. The ribs on vertebrae 25–28 are approximately 1.5 times as long as the vertebrae to which they attach; the last rib is slightly shorter than the associated vertebra. Figure 18. View largeDownload slide Eichstaettisaurus schroederi (holotype part – counter-part not shown; SNSB-BSPG 1937 I 1a). Figure 18. View largeDownload slide Eichstaettisaurus schroederi (holotype part – counter-part not shown; SNSB-BSPG 1937 I 1a). The skeletons that have been referred to Ardeosaurus all lack complete enough axial columns to confidently reconstruct a specific presacral vertebral count. Even so, each preserves approximately 28 presacral vertebrae (e.g. Fig. 19; see also Hoffstetter, 1964; Mateer, 1982; Estes, 1983). Figure 19. View largeDownload slide A, cast of the holotype of Ardeosaurus brevipes (NHM PR 38006) in dorsal view. B, skeleton of SNSB-BSPG 1923 I 501 (labelled A. cf. brevipes) in dorsal view. Figure 19. View largeDownload slide A, cast of the holotype of Ardeosaurus brevipes (NHM PR 38006) in dorsal view. B, skeleton of SNSB-BSPG 1923 I 501 (labelled A. cf. brevipes) in dorsal view. Sacral vertebrae The two sacral vertebrae are not well preserved in S. dyspepsia. Each is partly obscured by the impression of an overlying dorsal C. longipes rib (Figs 2, 3, 14, 15). The anterior sacral vertebra is almost completely obscured and no meaningful details available beyond the fact that it exists. The second sacral vertebra preserves a broken part of its right sacral rib. The preserved part of the latter suggests a subovate cross-section. Bavarisaurus macrodactylus does not preserve any clear details for the centrum or for the neural arch for either sacral vertebra (Fig. 17). Both sacrals preserve incomplete impressions of their neural arches. The sacral vertebra preserves as a distal part of the sacral rib on the right side, but this fragment only indicates that the expansion is slightly less expansive than that of the second sacral rib. The left side preserves an incomplete impression of the sacral rib that contacts the impression for the left ilium. A clear impression is present for the distal part of the left rib of the second sacral. The right side preserves an impression for the proximal part of the rib. Distally, the rib itself is preserved. The right side of the sacrum is better preserved than the left in E. schroederi and preserves parts of both sacral ribs (Fig. 20). The left side preserves only the first sacral rib. The exposed part of the first sacral rib is anteroposteriorly narrow. The second sacral rib is much broader than the first and possesses a posterior lamina that gives it an anteriorly arched posterior margin. It is also slightly expanded anterolaterally. The anterior part of this lamina slightly dorsally overlaps the posterolateral part of the first sacral rib. Figure 20. View largeDownload slide Pelvis, left hind limb and preserved caudal region of Eichstaettisaurus schroederi (holotype part; SNSB-BSPG 1937 I 1a). Figure 20. View largeDownload slide Pelvis, left hind limb and preserved caudal region of Eichstaettisaurus schroederi (holotype part; SNSB-BSPG 1937 I 1a). The sacrum and sacral ribs of P. bavarica are damaged and I cannot competently describe them at this time. The cast of the A. brevipes holotype has a damaged sacrum lacking clear sacral ribs and neural arches. The specimen described in Mateer (1982) is also damaged, but a damaged sacral rib is preserved. Almost nothing is known of the sacrum in SNSB-BSPG 1923 I 501. The sacrum and its sacral ribs are incompletely preserved as impressions on the matrix. The sacral ribs are broad and do not appear to contact distally. Caudal vertebrae Approximately 47 caudal vertebrae are preserved in a nearly consecutive series (Figs 2, 3, 14). These are looped into the anterior part of the body of C. longipes, associated with the second through fourth elongate ribs of the dinosaur. A study by Ostrom (1978) convincingly argued that all of these caudal vertebrae belong to the same animal and this is confirmed by direct observation. The first caudal is damaged such that the transverse processes are not preserved. The second caudal preserves broad transversely oriented transverse processes that do not taper distally. The third, fourth and fifth caudals are not well preserved, but they do demonstrate that the transverse processes were becoming posterolaterally oriented by the level of the third caudal vertebra. The fifth caudal transverse processes demonstrate the presence of a faint autotomy septum. These autotomy septa are present throughout the rest of the caudal series; transverse processes are clearly preserved through the ninth caudal vertebra. No autotomy foramina are evident on any of the caudals preserving transverse processes. Posterior to the ninth caudal vertebra, there is a discontinuity within the caudal vertebral series. There is no way to know how many caudals are missing, but the next vertebra is similar in length to the ninth caudal vertebra, yet possesses no clear transverse processes. No subsequent vertebra possesses a clear transverse process. Also, the prezyapophyses and postzygapophyses are very weakly developed throughout the rest of the vertebrae. The autotomy planes are located near the midlength of the vertebrae. Bavarisaurus macrodactylus possesses an incomplete caudal series, but the first 23 caudals are preserved in an unbroken string (Fig. 4). In contrast to S. dyspepsia, the first transverse process is posterolaterally oriented (Fig. 17). The second transverse process is more transversely oriented than the first caudal vertebra. Subsequent transverse processes are similarly oriented; that is, they are mostly transversely oriented with a slight posterolateral inclination (Figs 4, 17, 21). Bavarisaurus macrodactylus possesses autotomy planes on every preserved caudal vertebra. Figure 21. View largeDownload slide Left hind limb zeugopodium and pes, and proximal and middle tail of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501) with impressions of the dorsal view and bones preserved in ventral view. Figure 21. View largeDownload slide Left hind limb zeugopodium and pes, and proximal and middle tail of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501) with impressions of the dorsal view and bones preserved in ventral view. Although most of the skeleton of Ba. macrodactylus is preserved in ventral view, the tail is twisted such that the neural spine is partly visible on the specimen’s left side beginning on caudal vertebra fourth caudal vertebra. The fifth caudal vertebra clearly demonstrates a broad and dorsally oriented neural spine. Its dorsal margin is squared. Chevrons are also preserved and visible beginning with the fifth caudal vertebra. The chevron on the fifth caudal vertebra is slightly longer than the centrum. The neural spines become increasingly diminished in dorsoventral and anteroposterior lengths. By the level of the eighth caudal vertebra, the neural spine is very reduced. Its dorsal height is approximately equal to one-third of the centrum length. The eighth caudal vertebra also preserves a very clear autotomy plane and demonstrates the absence of a clear pseudospine. By the level of the 17th caudal vertebra, the neural spine appears to be absent. Eichstaettisaurus schroederi preserves the first six caudal vertebrae (Fig. 20). These vertebrae possess well-developed transverse processes. These transverse processes are not as laterally extensive as in S. dyspepsia or in Ba. macrodactylus. These transversely oriented transverse processes appear to be broadest at their midlength and taper distally. Beyond the preserved vertebrae, the E. schroederi preserves a stain that seems to be a continuation of the tail. This trace is equal to approximately half of the precaudal length of the animal. The caudal series in NHM PR 38006 is incomplete and damaged (Fig. 19A). Only the faintest impressions of the caudal series is preserved in the Ardeosaurus specimen SNSB-BSPG 1923 I 501 (Fig. 19B). The caudals of CM 4026 are represented only by impressions on the matrix. The first five caudal vertebrae are well preserved in the Ardeosaurus described in Mateer (1982), the sixth retains part of a transverse process and a centrum, and the following five are known only from partial centra. The preserved transverse processes extend directly mediolaterally. These are the best preserved Ardeosaurus caudal vertebrae considered here. Much of the caudal series is preserved in P. bavarica (Hoffstetter, 1964), but the vertebrae are damaged. Autotomy septa are preserved in the caudal vertebrae posterior to the transverse processes. Appendicular skeleton Pectoral girdle Nothing is known of the pectoral girdle. Only the left humerus and associated part of the proximal forelimb zeugopodium is preserved (Figs 3, 22). Figure 22. View largeDownload slide The preserved part of the left humerus of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) in ventral view along with some of the dorsal ribs in dorsal view. Figure 22. View largeDownload slide The preserved part of the left humerus of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) in ventral view along with some of the dorsal ribs in dorsal view. Most of the Solnhofen lizards preserve very little or no pectoral girdle elements. Bavarisaurus macrodactylus preserves only an incomplete right clavicle. It appears to be slightly dorsally expanded. The subtriangular scapulae and the medially expanded clavicles are preserved in E. schroederi (Fig. 18). A scapular emargination (defined by a scapular epicoracoid bar) is absent and a posterior (secondary) coracoid emargination cannot be confirmed, but a primary emargination is present. An incomplete pectoral girdle is preserved as a partial impression in A. brevipes (NHM PR 38006). Although this impression includes a small amount of the scapulocoracoid unit, virtually no details are available for it. Two little of the clavicle is preserved to offer meaningful details in this specimen and in SNSB-BSPG 1923 I 501, for which the scapulocoracoid is similarly poorly known. More of the scapulocoracoid is preserved in CM 4026 and demonstrates the absence of a scapular epicoracoid bar, but the presence of at least a primary coracoid emargination (see Hoffstetter, 1964). Palaeolacerta bavarica preserves only a small part of the clavicle, but also a complete scapulocoracoid (Hoffstetter, 1964; Estes, 1983). This specimen preserves part of the cartilaginous skeleton, including costal cartilages and, perhaps, a part of the epicoracoid. Although this scapulocoracoid has been suggested as possessing a single coracoid emargination and no scapular emargination (Hoffstetter, 1964; Estes, de Queiroz & Gauthier, 1988), a secondary coracoid emargination may be present and obscured by the partial epicoracoid. Further study of the specimen will be necessary to determine whether this is correct. Humerus The left humerus is incompletely preserved in ventral view (Figs 14, 22). The proximal head is not preserved. An impression remains of approximately two-thirds of the diaphysis. The distal part of the humerus is preserved in total. The preserved part of the humerus is approximately 11.5 mm long, but the complete humerus was likely approximately 13.5 mm long. The ectepicondylar–entepicondylar breadth is approximately 4.1 mm. The narrowest preserved part of the diaphysis is approximately 0.9 mm across. Thus, the humerus is quite gracile. An impression of the deltopectoral crest is preserved. The entepicondyle is robust with a distally concave proximal surface that extends laterally to form an acute anterior angle. The ectepicondyle is not as robust as the entepicondyle. It extends only slightly beyond the level of the radial condyle. A small ectepicondylar foramen is present at about the same distal level as the proximal margin of the radial condyle. A broad entepicondylar notch (sensuRieppel, 1980) is proximally rounded and distally expands to span the ulnar condyle. The subovate radial condyle is approximately 1.8 mm long and 1.1 mm across. The ulnar condyle is 1.3 mm long and 1 mm across. Similar parts of the humerus are known in Ba. macrodactylus (Fig. 16), although both humeri are known for the latter. Bavarisaurus macrodactylus possesses a similarly robust entepicondyle to that of S. dyspepsia. Whereas the entepicondyle is attenuated in S. dyspepsia (Fig. 22), it is distally squared in Ba. macrodactylus (Fig. 16). The morphology of the entepicondylar notch is similarly deeply developed between the two species, but does not expand distally as much in Ba. macrodactylus. The greatest distal width of the humerus is approximately 4.5 mm; its narrowest observable width is 1.4 mm. The length of the radial condyle may only be estimated because its distal terminus is hidden by the articulated radius; it is approximately 1.9 mm long and 1.3 mm across. The ulnar condyles are small, partly obscured and damaged. Both humeri are preserved in posterior view in E. schroederi (Fig. 18). Neither deltopectoral crest is visible, nor are the entepicondyles clearly exposed. The posterodistal part of each humerus is slightly eroded. Even so, presence of an ectepicondylar foramen is confirmed. The specimen of Ardeosaurus described by Mateer (1982; PMU.R58) possesses a humerus that is approximately 8.2 mm long, or approximately 51% the length of the skull roof at the midline. This is a proportion that is more or less conserved across the compared Ardeosaurus specimens here (Table 1). The broadest part of the distal epicondyles of the humerus in SNSB-BSPG 1923 I 501 is approximately 4.2 times the minimum measureable shaft diameter. Table 1. Osteometrics of selected Jurassic and Cretaceous squamates   PCl  SRl  Ml  Hl  dwH  mwH  Fl  mwF  Tl  Pl  Schoenesmahl dyspepsia    15.1  18.3  13.5  4.1  0.9  18.2  2.1  16.9  30.6  Bavarisaurus macrodactylus  91  20      4.7  1.3  22  2.1  18.1  33.5  Eichstaettisaurus schroederi  85.5  17  18.2  10    0.9  15.3  1.3  10.8  15.4  ‘Eichstaettisaurus’ gouldi  47    13.5  7      7    5.4    ‘Ardeosaurus’ (Mateer, 1982)  78.2  16.3  19.4  8.2      11.8    6.4  15.8  NHM PR 38006  58.1  11  13.1  6.3  2.2    8.3  0.7  5.3  10.7  BSP 1923 I 501  84.4  17.3  20.5  9.8  2.9  0.9  12.5  1  8.1  17.9  Chometokadmon fitzingeri  128.8  27.2    16.7  4.6  1.4  19.6  2  17    Scandensia ciervensis  36.9    6.4  3.7      4.8    3.3  6.7  Hoyalacerta sanzi  39.8    8.5  1.8      4.4    2.9    Huehuecuetzpalli mixtecus  112.4  28.6  30.2  15.7  5.3  1.8  24.5  3.1  21.6  37.1  ‘Ardeosaurus’ digitatellus  89  19          13.1  1  8.3  16.2  Palaeolacerta bavarica  41  8.1  9.9  4.6      5    4.2      PCl  SRl  Ml  Hl  dwH  mwH  Fl  mwF  Tl  Pl  Schoenesmahl dyspepsia    15.1  18.3  13.5  4.1  0.9  18.2  2.1  16.9  30.6  Bavarisaurus macrodactylus  91  20      4.7  1.3  22  2.1  18.1  33.5  Eichstaettisaurus schroederi  85.5  17  18.2  10    0.9  15.3  1.3  10.8  15.4  ‘Eichstaettisaurus’ gouldi  47    13.5  7      7    5.4    ‘Ardeosaurus’ (Mateer, 1982)  78.2  16.3  19.4  8.2      11.8    6.4  15.8  NHM PR 38006  58.1  11  13.1  6.3  2.2    8.3  0.7  5.3  10.7  BSP 1923 I 501  84.4  17.3  20.5  9.8  2.9  0.9  12.5  1  8.1  17.9  Chometokadmon fitzingeri  128.8  27.2    16.7  4.6  1.4  19.6  2  17    Scandensia ciervensis  36.9    6.4  3.7      4.8    3.3  6.7  Hoyalacerta sanzi  39.8    8.5  1.8      4.4    2.9    Huehuecuetzpalli mixtecus  112.4  28.6  30.2  15.7  5.3  1.8  24.5  3.1  21.6  37.1  ‘Ardeosaurus’ digitatellus  89  19          13.1  1  8.3  16.2  Palaeolacerta bavarica  41  8.1  9.9  4.6      5    4.2    All measurements are in millimetre. dwH, maximum width of the distal end of the humerus; Fl, femur length; Hl, humerus length; Ml, mandible length; mwF, minimum (midshaft) width of the humerus; mwH, minimum (midshaft) width of the humerus; PCl, precaudal length; Pl, pes length; SRl, skull roof length; Tl, tibia length. View Large Table 1. Osteometrics of selected Jurassic and Cretaceous squamates   PCl  SRl  Ml  Hl  dwH  mwH  Fl  mwF  Tl  Pl  Schoenesmahl dyspepsia    15.1  18.3  13.5  4.1  0.9  18.2  2.1  16.9  30.6  Bavarisaurus macrodactylus  91  20      4.7  1.3  22  2.1  18.1  33.5  Eichstaettisaurus schroederi  85.5  17  18.2  10    0.9  15.3  1.3  10.8  15.4  ‘Eichstaettisaurus’ gouldi  47    13.5  7      7    5.4    ‘Ardeosaurus’ (Mateer, 1982)  78.2  16.3  19.4  8.2      11.8    6.4  15.8  NHM PR 38006  58.1  11  13.1  6.3  2.2    8.3  0.7  5.3  10.7  BSP 1923 I 501  84.4  17.3  20.5  9.8  2.9  0.9  12.5  1  8.1  17.9  Chometokadmon fitzingeri  128.8  27.2    16.7  4.6  1.4  19.6  2  17    Scandensia ciervensis  36.9    6.4  3.7      4.8    3.3  6.7  Hoyalacerta sanzi  39.8    8.5  1.8      4.4    2.9    Huehuecuetzpalli mixtecus  112.4  28.6  30.2  15.7  5.3  1.8  24.5  3.1  21.6  37.1  ‘Ardeosaurus’ digitatellus  89  19          13.1  1  8.3  16.2  Palaeolacerta bavarica  41  8.1  9.9  4.6      5    4.2      PCl  SRl  Ml  Hl  dwH  mwH  Fl  mwF  Tl  Pl  Schoenesmahl dyspepsia    15.1  18.3  13.5  4.1  0.9  18.2  2.1  16.9  30.6  Bavarisaurus macrodactylus  91  20      4.7  1.3  22  2.1  18.1  33.5  Eichstaettisaurus schroederi  85.5  17  18.2  10    0.9  15.3  1.3  10.8  15.4  ‘Eichstaettisaurus’ gouldi  47    13.5  7      7    5.4    ‘Ardeosaurus’ (Mateer, 1982)  78.2  16.3  19.4  8.2      11.8    6.4  15.8  NHM PR 38006  58.1  11  13.1  6.3  2.2    8.3  0.7  5.3  10.7  BSP 1923 I 501  84.4  17.3  20.5  9.8  2.9  0.9  12.5  1  8.1  17.9  Chometokadmon fitzingeri  128.8  27.2    16.7  4.6  1.4  19.6  2  17    Scandensia ciervensis  36.9    6.4  3.7      4.8    3.3  6.7  Hoyalacerta sanzi  39.8    8.5  1.8      4.4    2.9    Huehuecuetzpalli mixtecus  112.4  28.6  30.2  15.7  5.3  1.8  24.5  3.1  21.6  37.1  ‘Ardeosaurus’ digitatellus  89  19          13.1  1  8.3  16.2  Palaeolacerta bavarica  41  8.1  9.9  4.6      5    4.2    All measurements are in millimetre. dwH, maximum width of the distal end of the humerus; Fl, femur length; Hl, humerus length; Ml, mandible length; mwF, minimum (midshaft) width of the humerus; mwH, minimum (midshaft) width of the humerus; PCl, precaudal length; Pl, pes length; SRl, skull roof length; Tl, tibia length. View Large Radius and ulna Only the proximal parts of the radius and ulna are preserved and visible (Figs 14, 22, 23). Both are partly obscured by overlying ribs and vertebrae. The proximal part of the radius is expanded such that it is slightly broader than the radial condyle of the humerus. The preserved proximal part of the ulna is similar in breadth to that of the radial head. Figure 23. View largeDownload slide Posterior dorsal and anterior caudal vertebrae, sacrum, pubis, ischium and proximal femora of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) in dorsal view. Figure 23. View largeDownload slide Posterior dorsal and anterior caudal vertebrae, sacrum, pubis, ischium and proximal femora of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) in dorsal view. The radii and ulnae are more completely preserved in the other Solnhofen lizards, but the limited visible parts of those bones in S. dyspepsia preclude detailed comparisons. Pelvic girdle and hind limb The right side of the pelvis and parts of both hind limbs are preserved in association with the axial skeleton in S. dyspepsia (Figs 2, 3, 14, 23, 24). The anterior part of the pubis is partly overlaid by a dorsal rib from C. longipes. Part of what is probably an ischium is narrowly visible near the proximal end of pubis and projecting under the first caudal vertebra. The right femur is preserved as part of a cast of that was made before the specimen was further prepared (Fig. 24). Almost the entire left hind limb is preserved as a combination of bones and impressions of bones. Figure 24. View largeDownload slide Preserved dorsal, sacral and right hind limb region of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) as preserved in a cast of Compsognathus longipes (SNSB-BSPG AS I 563b). The right femur of S. dyspepsia gen. et. sp. nov. shown here was largely lost some time after creation of the mould for this cast. Figure 24. View largeDownload slide Preserved dorsal, sacral and right hind limb region of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) as preserved in a cast of Compsognathus longipes (SNSB-BSPG AS I 563b). The right femur of S. dyspepsia gen. et. sp. nov. shown here was largely lost some time after creation of the mould for this cast. No ilium is preserved with S. dyspepsia. Parts of the ilia are preserved in Ardeosaurus, E. schroederi and Ba. macrodactylus. However, none of those preserved ilia show details of the iliac anatomy, nor do they offer any indication of any unusual characteristics. Pubis The right pubis is preserved in dorsal view lying near the distal end of the left humerus and the proximal end of the right femur (Fig. 23). The pubis is robust with a broad symphyseal part that is more elongate than the tubercular part. The symphyseal part is approximately 4.3 mm whereas the tubercular part is 3.9 mm in length. The broadest visible part of the symphyseal ramus is approximately 2.8 mm in anteroposterior width. The symphyseal part of the pubis forms a slightly obtuse angle with the tubercular part. The pubic tubercle is robust and appears to be distally squared. The obturator foramen is anteroventrally oriented. A fossa is present extending posterodorsal to the obturator foramen and contiguous with it. This fossa is somewhat broader than the obturator foramen. The posterior margin of the symphyseal part of the pubis is nearly transverse, suggesting an anteriorly broad thyroid fenestra. The right pubis is well preserved in ventral view in Ba. macrodactylus (Figs 4, 17). It is similar to the pubis of S. dyspepsia in that the symphyseal part (6.7 mm long) is longer than the tubercular part (3.4 mm long), but the Ba. macrodactylus pubis is less robust than S. dyspepsia. The broadest preserved and visible part of the symphyseal part of the pubis (including the impression of its anterior margin) in Ba. macrodactylus is approximately 2.1 mm wide. Bavarisaurus macrodactylus has a pubis in which the symphyseal part forms a very obtuse angle with the tubercular part; the symphyseal part is anteromedially oriented in the Ba. macrodactylus and the anterior margin of the thyroid fenestra extends far anterior to the level of the pubic tubercle. The pubic tubercle is robust, but it is anterolaterally attenuated. The obturator foramen is anteroventrally oriented. Eichstaettisaurus schroederi preserves largely complete pubes, but these are partly overlaid by vertebrae, ribs, the proximal parts of the femora and part of the ilium (Figs 18, 20). Although few details of these pubes are present, the state of their preservation indicates that they were more anteromedially oriented than those of S. dyspepsia and have an orientation similar to that present in Ba. macrodactylus. The Ardeosaurus specimens NHM PR 38006 (Fig. 19A), SNSB-BSPG 1923 I 501 (Fig. 19B) and CM 4026 possess very incompletely known pubes. The specimen described in Mateer (1982) has a clearly preserved right pubis with an elongate symphyseal part and a short tubercular part. Ischium A flat piece of bone is seen extending ventral to the first caudal vertebra. Originally tentatively identified as the left pubis (Ostrom, 1978), this element probably represents the right ischium (Figs 2, 3, 23). The preserved part of the ischium demonstrates a relatively straight dorsal margin. Although partly hidden, the ventral margin suggests that ischium is gently constricted just distal to its acetabular part before expanding posteroventrally. The extent of this expansion and the distal morphology are obscured by the second caudal vertebra. No details of the ischium are available for Ardeosaurus. None of the ischium is visible in Ba. macrodactylus or in E. schroederi. Femur A complete right femur is preserved on a cast of the specimen (Fig. 24); this femur was largely lost subsequent to the making of the mould for this cast. This femur is approximately 16.9 mm long. The preserved part of the left femur (Fig. 14) matches this length. The narrowest part of the femur occurs at approximately midshaft and measures approximately 2.1 mm across (Table 1). The left femur is preserved in dorsal view; the right femur is preserved in posterodorsal view. The femur cast demonstrates a mild dorsal arching to the femur distally paired with a ventral proximal arching. Both femora are represented in Ba. macrodactylus (Fig. 4). The left femur is preserved only as an impression on the rock and lacks its proximal end. The right femur is nearly complete and lacks only the internal trochanter, which has been eroded away. The right femur is 21.9 mm long and approximately 2.1 mm wide at its narrowest point. The femur shows the same curvature present in the femur of S. dyspepsia. Eichstaettisaurus schroederi preserves both gracile femora as bone (Figs 18, 20). The femora are approximately 15.3 mm long and the minimum measurable diameter is approximately 1.3 mm. Although E. schroederi possesses the curvature of the femur, it appears less pronounced than in S. dyspepsia and Ba. macrodactylus. Eichstaettisaurus gouldi preserves incomplete femora (Evans et al., 2004). The right femur is slightly more complete proximally and appears to preserve the proximal head, or most of the proximal head. It is approximately 7 mm long, about the same length as the humerus (Table 1). The femur is more than 1.5 times the tibia length in all of the specimens of Ardeosaurus considered here. The femur is approximately 9.6 mm long in the holotype of A. brevipes, 11.8 mm long in the specimen described in Mateer (1982), approximately 12 mm long in SNSB-BSPG 1923 I 501 and 13.1 mm long in CM 4026. Palaeolacerta bavarica preserves both femora, but neither is complete (Hoffstetter, 1964; Estes, 1983). When complete, the femur would have been approximately 5 mm long, approximately 1.2 times the length of the reconstructed tibia. Tibia and fibula Distal parts of the right tibia and fibula are preserved near the anterior caudal vertebrae (Figs 2, 3, 25). Based on their position and the position of the right femur, they were probably in natural articulation during fossilization, but were subsequently damaged and/or eroded. The ends of the left tibia and fibula are preserved as bone, but the middle parts of these bones are preserved only as impressions on the matrix. Although the middle parts of the impressions are partly obscured by overlying ribs from C. longipes, the state of preservation and those of the bones around them suggest that the preserved parts are undisturbed. The tibia is 17.3 mm long and the fibula is 17.8 mm long. Figure 25. View largeDownload slide Left hind limb zeugopodium and pes, and some caudal vertebrae of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) as preserved in the anterior rib cage of Compsognathus longipes (SNSB-BSPG AS I 563b). Figure 25. View largeDownload slide Left hind limb zeugopodium and pes, and some caudal vertebrae of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) as preserved in the anterior rib cage of Compsognathus longipes (SNSB-BSPG AS I 563b). The tibiae and fibulae of Ba. macrodactylus are each preserved as a combination of bones and bone impressions (Figs 4, 21). The tibia is approximately 18.7 mm long and the fibula is almost exactly the same length. Thus, the tibia is longer than the femur in S. dyspepsia, but the femur is longer than the tibia in Ba. macrodactylus. Eichstaettisaurus schroederi possesses proportionately shorter limbs compared to the skull length (Fig. 18; Table 1) than does S. dyspepsia and Ba. macrodactylus. The tibia is approximately 10.8 mm long in E. schroederi. Thus, it is much shorter than the femur (Figs 18, 20). The femur is mostly straight in E. schroederi, but there is a distinctive anterodistal curve at the level of the fourth trochanter (Fig. 20). Well-preserved hind limb zeugopodia are preserved as bone in the A. breviceps holotype and the Ardeosaurus specimen described in Mateer (1982), as clear impressions in SNSB-BSPG 1923 I 501, and as a combination of bone and impressions in CM 4026. Palaeolacerta bavarica preserves partial hind limb zeugopodia. The left tibia is nearly completely preserved, but only the distal part of the fibula is preserved. The right zeugopodium is poorly preserved. Tarsals Only the left calcaneum is clearly preserved in S. dyspepsia (Fig. 25), and it is slightly damaged medially. The pronounced fibular facet has a distinct lip. Damage to the medial surface of the calcaneum precludes determining whether the non-preservation of the astragalus is the result of damage or an unfused condition. Even so, the preserved part of the calcaneum demonstrates that the tibial and fibular condyles are separated by a narrow space similar to less than one-half the width of the distal end of the tibia. No distal tarsals are preserved. The fused astragalocalcanea are bilaterally preserved in Ba. macrodactylus (Figs 4, 21). The tibial and fibular facets are separated by a distance of about one-quarter the distal tibial breadth. The calcaneum has a modest lateral tuberosity that extends only slightly beyond the level of the distal fibular head. The lateral tuberosity has a well-defined proximodistal ridge. Eichstaettisaurus schroederi preserves damaged tarsals (Figs 18, 20) that offer few morphological details. Ardeosaurus specimens lack well-preserved tarsals. The specimen described in Mateer (1982; PMU.R58) preserves a robust and fused astragalocalcaneum with a lateral tuberosity. Both NHM PR 38006 and SNSB-BSPG 1923 I 501 preserve an unfused astragalus and calcaneum (Fig. 19). A lateral tuberosity is preserved in CM 4026, but it is not as robust as in PMU.R58, and the potential fusion of the astragalus and calcaneum remains uncertain. Pedes Parts of all five left metatarsals and parts of four digits are preserved (Fig. 25). Only the distal part of the metatarsal and the proximal phalanx are preserved for the first digit. Only the proximal parts of the second and third metatarsals are missing. None of the phalanges are preserved in digit II, but all of the phalanges are preserved on digit III where there are four phalanges, including the ungual. The penultimate and antepenultimate phalanges are of subequal length. No phalangeal sesamoids are preserved, but their original absence cannot be confidently inferred due to the state of preservation. The fourth metatarsal and digit are almost as complete, but the ungual is not preserved. Four phalanges are preserved in the fourth metatarsal. This digit almost certainly possessed five phalanges. Although the fifth metatarsal is nearly complete, the proximomedial surface is hidden. Regardless, this metatarsal was clearly very short. Only one phalanx is preserved on the fifth digit. The proportions of the pedes are similar in Ba. macrodactylus (Figs 4, 21) as in S. dyspepsia. All five digits are preserved either as bone or as clear impressions. The left metatarsals I–V in Ba. macrodactylus measure 5.6, 8.5, 10.8, 11.3 and 3.2 mm, respectively. The pedal phalangeal formula in Ba. macrodactylus is 2-3-4-5-4. Eichstaettisaurus schroederi has shorter pedes with respect to its skull and its metatarsals I–IV are more constant in length, although the fifth metatarsal is still relatively short (Figs 18, 20). The right metatarsal lengths, from I to V are 3.2, 3.7, 4.2, 4.2 and 1.6 mm. The pedal phalangeal formula for E. schroederi (based on comparisons of the two sides of the specimen) is 2-3-4-5-4. The pedes are too incompletely preserved in E. gouldi to determine specifics about the proportions or the pedal phalangeal formula (see Evans et al., 2004). The pedes are preserved in the holotype of A. brevipes, but few details are available from the cast of the specimen. The metatarsals II–IV are of similar length to one another, but becoming progressively longer. Only the distal part of metatarsal I is visible and only on the right pes. Metatarsal I appears to be slightly more than one-half the length of metatarsal II. Although the phalangeal count in digit V is uncertain, it probably possessed four phalanges; the phalangeal formula appears to have been 2-3-4-5-4. Other known Ardeosaurus material includes pedes preserved to a greater or lesser degree. The specimen SNSB-BSPG 1923 I 501 possesses metatarsals and phalanges for the left pes that are preserved mostly as impressions (Fig. 26). The first digit is represented only by the proximal part of the metatarsal and the second digit is also incomplete. The metatarsals II–IV are subequal in length and robustness. The fifth metatarsal is only approximately 38% the length of the fourth metatarsal. The known digital formula is ?-?-4-5-4. The specimen described in Mateer (1982) has a pedal phalangeal formula of ?-3-4-5-4. The pedes in CM 4026 are also known mostly from impressions and the right is more complete than the left. The proximal parts of the metatarsals are damaged, but the metatarsals appear to get progressively longer from I to IV. The digital formula is 2-3-4-5-4. Figure 26. View largeDownload slide Right hind limb of SNSB-BSPG 1923 I 501 (labelled Ardeosaurus cf. brevipes) in dorsal view. Figure 26. View largeDownload slide Right hind limb of SNSB-BSPG 1923 I 501 (labelled Ardeosaurus cf. brevipes) in dorsal view. Palaeolacerta bavarica possesses incompletely preserved pedes (Hoffstetter, 1964). Metatarsals II–IV of the right pes appear to be subequal in length. The pedes are relatively short in P. bavarica compared to the Ardeosaurus, Ba. macrodactylus and S. dyspepsia material. A pedal phalangeal formula cannot be confidently reconstructed. PHYLOGENY Historical treatments of the Solnhofen squamates The Solnhofen squamates Ardeosaurus spp., Ba. macrodactylus and E. schroederi have all been known for more than three-quarters of a century. Prior to the implementation of cladistic analyses, they were typically associated with geckos. Ardeosaurus is known from multiple specimens representing at least two species (Estes, 1983). Ardeosaurus spp. have been considered ‘scincomorphs’ (Romer, 1956) and basal geckos (Cocude-Michel, 1961; Hoffstetter, 1964;,Kluge, 1967; Estes, 1983; Carroll, 1988; Alifanov, 2000). The holophyly of Ardeosaurus has never been questioned or tested. Originally, Ba. macrodactylus was described as Homoesaurus macrodactylus (Wagner, 1852) and believed to be a rhynchocephalian (Wagner, 1852; Grier, 1914). Subsequent to its identification as a squamate (Huene, 1952), the holotype was given the name Ba. macrodactylus (Hoffstetter, 1953). Two studies by Cocude-Michel suggested possible gekkotan (Cocude-Michel, 1961) or iguanian affinities (Cocude-Michel, 1963). Many other studies have suggested that Bavarisaurus (traditionally often including the S. dyspepsia) may be close to geckos (Hoffstetter, 1964; Estes, 1983; Carroll, 1988) and possibly closely related to Ardeosaurus, E. schroederi and the Asian form Yabeinosaurus tenuis Endo & Shikama, 1942 (Hoffstetter, 1964; Estes, 1983). One more recent non-cladistic study placed Bavarisaurus within Iguania as a part of Mosasauromorpha along with Mosasauridae, Aigialosauridae, Dolichosauridae and Paravaranidae (Alifanov, 2000). Surprisingly few cladistic phylogenetic analyses have addressed the relationships of the Solnhofen squamates, especially given their relative completeness, the excellent studies addressing their morphology (e.g. Cocude-Michel, 1961; Ostrom, 1978; Mateer, 1982; Estes, 1983; Evans, 1994b, among others), and the fact that they are some of the earliest- and longest-known squamates. Many of the relevant studies have overlapping authorship (e.g. Evans & Barbadillo, 1998, 1999; Evans & Wang, 2005; Evans, Wang & Li, 2005), meaning that relatively few researchers have considered the placement of these animals. Bavarisaurus macrodactylus and Ardeosaurus have been conspicuously absent from some recent global phylogenetic analyses of Squamata (e.g. Wiens et al., 2010; Gauthier et al., 2012; Reeder et al., 2015). Most analyses addressing the Solnhofen squamates have found Ba. macrodactylus to be outside of crown-group squamates (Evans & Barbadillo, 1998, 1999; Evans & Wang, 2005; Evans et al., 2005; Evans, Raia & Barbera, 2006). Many of those studies also find Ardeosaurus to be just outside of the crown group (Evans & Barbadillo, 1998; Evans & Wang, 2005; Evans et al., 2005). Ardeosaurus has also been found to form a polytomy with Eichstaettisaurus, Hoyalacerta sanzi Evans & Barbadillo, 1999, Scandensia ciervensis Evans & Barbadillo, 1998, Iguania and scleroglossans (Evans & Barbadillo, 1999). Eichstaettisaurus (sometimes including E. gouldi) has variably been found to be outside crown-group squamates (Evans & Barbadillo, 1998), as forming a polytomy with the crown group (Evans & Barbadillo, 1999), as a basal scincogekkonomorph (Evans et al., 2006), or as part of a basal anguimorph radiation (Evans & Wang, 2005; Evans et al., 2005). Conrad (2008) found Bavarisaurus (including the S. dyspepsia material), Ardeosaurus and E. schroederi to be basal members of Scincogekkonomorpha. A composite Eichstaettisaurus (including E. schroederi and E. gouldi) was found to be a proximal outgroup to Gekkota in a recent global study of squamate relationships (Gauthier et al., 2012). A subsequent study (Reeder et al., 2015), whose morphological data and fossil taxon sampling relied heavily upon the Gauthier et al.’s (2012) study, found the Eichstaettisaurus chimaera to be outside of the crown-group Squamata. The latter result is reminiscent of pre-2006 morphological studies (Evans & Barbadillo, 1998, 1999; Evans & Wang, 2005; Evans et al., 2005). The present analysis Data matrix I modified the morphological character list presented by Conrad, Balcarcel & Mehling (2012), mainly by adding data from soft tissue (Schwenk, 1988), salivary compounds (Fry et al., 2006) and osteology (Smith, 2009; Gauthier et al., 2012), and also with additions of a few characters from other studies (see character descriptions below). Taxon selection was performed following a modified version of the protocol set forth in a recent study wherein terminal units were ‘selected carefully in an effort to bracket ancestral states for well-supported squamate clades’ (Gauthier et al., 2012: 58). Importantly, that study omitted several taxa important for testing the relationships and holophyly of many major squamate clades. The Gauthier et al. (2012) study omitted most of the Jurassic and Early Cretaceous squamates that have been considered as falling near the trunk of the squamate tree, including A. brevipes, Ba. macrodactylus, Chometokadmon fitzingeri Costa, 1866, H. sanzi, Sca. ciervensis, and Y. tenuis. Therefore, I have used a different set of terminal units than was used in Gauthier et al. (2012). I have included all of those taxa mentioned above. My analysis, designed to identify the relationships among the major squamate clades, includes 220 fossil and extant terminals. None of these terminals are coded as composites above the species level (e.g. Eichstaettisaurus and Palaeoxantusia of Gauthier et al., 2012). Some are coded at the specimen level, including specimens previously referred to Ardeosaurus, Eoxanta lacertifrons Borsuk-Białynicka, 1988, Globaura venusta Borsuk-Białynicka, 1988 and Slavoia darevskii Sulimski, 1984. The present morphological data matrix includes 839 morphological characters. Following recent applications of this matrix (Conrad et al., 2011a, b, 2012), three characters (236, 242 and 364) were deactivated because they are supplanted by others or, in the case of character 364, because it was originally included only for mapping biogeography. I used NEXUS Data Editor (NDE; Page, 2001) to assemble and manage the data matrix. I performed an analysis using the New Technology Search in (200 replicates) the computer program T.N.T. (Goloboff, Farris & Nixon, 2003, 2008, 2010) with the ‘ratchet’ and ‘drift’ options employed. For the T.N.T. analysis, I rooted the tree on Pamelina polonica Evans, 2009, but rooted the tree such that kuehneosaurids are holophyletic for generating apomorphy lists. The apomorphy list for lepidosaurs remains unchanged when all three included kuehneosaurids are used as the outgroup. The results of the analysis were exported in Nexus format and opened in PAUP* 4.0b10 (Swofford, 2001) and a list of unequivocal apomorphies was generated using the ‘describe’ function for the strict consensus. Phylogenetic hypotheses with kuehneosaurids as outgroups The shortest tree length recovered by the analysis is 5497 steps and 192 trees of that length were recovered. The consistency index, excluding uninformative characters, for these trees is 0.1694 and the retention index is 0.6046. I report the strict consensus of the recovered trees here (Fig. 27). Figure 27. View largeDownload slide Temporally calibrated lepidosaur interrelationships based on the morphological analysis presented in the paper. Figure 27. View largeDownload slide Temporally calibrated lepidosaur interrelationships based on the morphological analysis presented in the paper. The strict consensus of shortest trees is unique among recently proposed sets of relationships and demonstrates the distinctiveness of S. dyspepsia with respect to Ba. macrodactylus, the polyphyly of the traditional usage of Ardeosaurus, and the relationships of E. schroederi and E. gouldi. For the character states addressing clades to which S. dyspepsia belongs, the nature of the character states for S. dyspepsia are noted. Bharatagama rebbanensis The Early Jurassic lepidosaur B. rebbanensis was originally described as a possible acrodontan iguanian (Evans et al., 2002), but this position has not been cladistically tested until now. The present analysis recovers B. rebbanensis as a rhynchocephalian close to Pleurosaurus goldfussi Meyer, 1845 with which it shares only a single unambiguous synapomorphy: the anterior marginal teeth are procumbent (character 216, state 1). It shares three unambiguous synapomorphies with Pl. goldfussi, Palaeopleurosaurus posidinae Carroll, 1985 and Pamizinsaurus tlayuaensis Reynoso, 1997 (character 219, state 1). This group is part of larger clade also containing Priosphenodon avelasi Apestaguía & Novas, 2003, Kallimodon cerinensis Cocude-Michel, 1963, Cynosphenodon huizachalensis Reynoso, 1996 and Sp. punctatus with which B. rebbanensis shares the unambiguous synapomorphy: presence of caniniform teeth (character 215, state 1). This large clade is the ‘clevosaur’ radiation and B. rebbanensis shares one of that clade’s unambiguous synapomorphies: addition of teeth to the posterior end of the marginal tooth row (character 221, state 3). Although some of these character states are plesiomorphic and/or convergent between rhynchocephalians and squamates, the analysis recovers no unambiguous support for B. rebbanensis being a squamate. Squamata In the present analysis, Squamata (all taxa sharing a more recent common ancestor with Lacerta viridis than with Sp. punctatus) is supported by 19 unambiguous synapomorphies. These are: 1. The nasals, as a unit, are narrower than the breadth of the external naris (character 373, state 1). This is reversed in S. dyspepsia. 2. The frontal unit (the bilateral width) is wider than it is long (character 387, state 1). This character cannot be confidently coded in S. dyspepsia. 3. The frontal possesses concave lateral margins and with a minimum width less than 60% of the posterior border width (character 57, state 2). 4. The frontoparietal suture is transverse (character 70, state 1). The frontoparietal suture is anteriorly arched in S. dyspepsia. 5. The parietal lacks frontal tabs (character 74, state 0). This character cannot be confidently coded in S. dyspepsia. 6. The postfrontal is reduced to an irregular piece of bone (character 92, state 1). The postfrontal is anteroposteriorly elongate in many squamates, including S. dyspepsia. 7. Palatine teeth are absent (character 115, state 1). This character cannot be coded in S. dyspepsia. 8. The pterygoid lacks teeth (character 118, state 2). This is reversed within Squamata. 9. There is no pterygoid–vomer contact (character 119, state 1). This character cannot be confidently coded in S. dyspepsia. 10. Absence of a midline pterygoid contact anterior to the pyriform recess (character 122, state 0). This character cannot be coded in S. dyspepsia. 11. Presence of a vidian canal formed by the parabasisphenoid enclosing the internal carotid artery and the basal of the palatine artery as they pass the basipterygoid process (character 693, state 1). 12. The quadrate lacks a distinct pterygoid lappet (character 160, state 1). This character cannot be confidently coded in S. dyspepsia. 13. Presence of three teeth in each premaxilla (character 403, state 2). This character cannot be confidently coded in S. dyspepsia. 14. Presence of a midline premaxillary tooth (character 404, state 1). This character cannot be confidently coded in S. dyspepsia. 15. Vertebrae procoelous (character 231, state 2). Vertebral centrum morphology cannot be confidently coded for S. dyspepsia. 16. Transverse processes absent from anterior presacral vertebrae (character 236, state 2). 17. Anterior dorsal ribs lacking distal broadening (character 255, state 0). This character cannot be confidently coded in S. dyspepsia. 18. A primary coracoid emargination is present (character 260, state 1). This character cannot be confidently coded in S. dyspepsia. 19. The pubis is anteromedially narrow (character 284, state 0) as opposed to the expanded pubis found in many rhynchocephalians. Scincogekkonomorpha Huehuecuetzpalli mixtecus Reynoso, 1998, H. sanzi and Iguanomorpha are the basalmost squamate clades in the current analysis and form an unresolved tetrachotomy with higher squamates. Bavarisaurus macrodactylus, Ch. fitzingeri and Eichstaettisauridae are the basalmost scincogekkonomorphs. Scincogekkonomorpha is, here, united by six unambiguous synapomorphies. 1. The postfrontal is medially forked and spanning the frontoparietal contact laterally (character 89, state 1). 2. The postfrontal is anteroposteriorly elongate – it is longer anteroposteriorly than its mediolateral width (character 92, state 0). 3. The teeth are pointed and curved (character 212, state 2) as opposed to peg-like and vertical. 4. The symphyseal part of the pubis is more than 1.5 times the length of the length of the proximal (tubercular) portion (character 283, state 2). 5. Twenty-five presacral vertebrae are present (character 410, state 7). This character cannot be scored in S. dyspepsia. 6. The penultimate manual phalanges are subequal to or shorter than the antepenultimate phalanges (character 767, state 0). This character cannot be scored in S. dyspepsia. Scincogekkonomorphs exclusive of Ba. macrodactylus Two unambiguous synapomorphies unite this clade. 1. Presence of paired, rather than fused, frontals (character 55, state 0). This is reversed in S. dyspepsia. 2. The presence of moderately long anterolateral processes of the frontal, extending anteriorly to about the same level as the midline anterior projection of the frontal (character 378, state 1). This character cannot be scored in S. dyspepsia. Higher scincogekkonomorphs Scincogekkonomorphs exclusive of Ba. macrodactylus and Ch. fitzingeri are united by four unambiguous synapomorphies. 1. The frontals possess an interorbital constriction such that the frontal is hourglass shaped (character 58, state 1). 2. No nuchal fossa is present on the parietal (character 83, state 0). This character cannot be scored in S. dyspepsia. 3. The jaw adductor musculature attaches to the ventral surface of the parietal (character 86, state 1). 4. Enlargement of the skull roof scales into broad, thin, plates (character 296, state 2). This character cannot be scored in S. dyspepsia. The traditionally rooted morphological analysis recovers a holophyletic Scleroglossa within Scincogekkonomorpha. Ardeosauridae, ‘Ardeosaurus’ digitatellus, Eichstaettisauridae and Polyglyphanodontia are successively more proximal scleroglossan outgroups. Polyglyphanodontia and Jucaraseps Polyglyphanodontia here includes Jucaraseps grandipes as the sister taxon to a clade containing Sineoamphisbaena hexatabularis and Cryptolacerta hassiaca. The Jucaraseps–Sineoamphisbaena–Cryptolacerta clade is united by five unambiguous synapomorphies. 1. Frontals with anterior and posterior borders of subequal width (character 57, state 0). 2. The arietal supratemporal process length is equal to less than one-half the length of the skull table part of the parietal (character 80, state 1). 3. The coronoid process is short and broad (character 192, state 0), as opposed to tall and narrow. 4. The clavicles are rodlike and lack proximal expansion and notching (character 258, state 1). 5. The prefrontal extends medially for more than one-half the breadth of the frontal (or one-quarter of the width of an azygous frontal; character 574, state 1). Polyglyphanodontia is united by six unambiguous synapomorphies in the current analysis. 1. The posterior process of the postorbital extends for more than three-quarters of the length of the supratemporal fenestra (character 96, state 2). 2. An ectopterygoid–palatine contact is present anteriorly within the suborbital fenestra and the contact is broader than the smallest diameter of the suborbital fenestra (character 124, state 2). 3. Zygosphenes and zygantra are present in the dorsal vertebrae and the zygosphene articular surface faces ventrolaterally (character 235, state 2). 4. The posterior margin of the palatine is squared (character 390, state 1). 5. The parietal has a midline contact with the supraoccipital (character 565, state 1). Polyglyphanodontia and Scleroglossa A clade containing Polyglyphanodontia and Scleroglossa to the exclusion of the Eichstaettisauridae (see below) is united by seven unambiguous synapomorphies. 1. The premaxilla is azygous (character 11, state 1). 2. Parietal tabs are present on the frontal (character 69, state 1). 3. Frontal tabs are present on the ventral surface of the parietal (character 74, state 2). 4. Descending processes of the parietal are present and anteroposteriorly narrow and elongate projections (character 76, state 2). 5. The supratemporal is anteroposteriorly short and does not extend much beyond the anterior margin of the suspensorium (character 88, state 0). 6. No medial tubercle is present on the medial margin of the articular (character 209, state 0). 7. Hypapophyseal keels are present in the cervical vertebrae (character 242, state 1). Eichstaettisauridae, Polyglyphanodontia and Scleroglossa This clade is united to the exclusion of other squamates by two unambiguous synapomorphies. 1. The clavicles are rodlike and lack proximal expansion and notching (character 258, state 1). 2. The penultimate manual phalanges are longer than the antepenultimate manual phalanges (character 767, state 1). Eichstaettisauridae: Eichstaettisauridae is here defined as: all taxa sharing a more recent common ancestor with E. schroederi than with Gekko gecko, L. viridis, Scincus scincus, Anguis fragilis or Boa constrictor. Eichstaettisauridae includes, in the current analysis, E. schroederi, Sca. ciervensis and ‘Eichstaettisaurus’ gouldi. Although E. schroederi and ‘Eichstaettisaurus’ gouldi may form a clade to the exclusion of Sca. ciervensis, no evidence of this is recovered here. The two unambiguous eichstaettisaurid synapomorphies recovered here are listed below. 1. Fused (not paired) frontals (character 55, state 1). This is here reconstructed as convergent in S. dyspepsia. 2. The vertebrae are amphicoelous (character 231, state 1). ‘Ardeosaurus’ digitatellus ‘Ardeosaurus’ digitatellus, Eichstaettisauridae, Polyglyphanodontia and Scleroglossa are united here by only one unambiguous synapomorphy: the squamosal possesses no distinct dorsal process (character 100, state 1). Ardeosauridae I offer a revised systematic definition for Ardeosauridae: all taxa sharing a more recent common ancestor with A. brevipes than with Ba. macrodactylus, G. gecko, L. viridis, Sci. scincus, An. fragilis or Bo. constrictor. As used here, Ardeosauridae includes three specimens usually referred to as Ardeosaurus, including the cast of the holotype of A. brevipes (NHM PR 38006), the specimen referred to it described in Mateer (1982; PMU.R58) and the specimen considered in Estes (1983) to be A. cf. digitatellus (SNSB-BSPG 1923 I 501). Note that the latter specimen is nearer to the A. brevipes than it is to the holotype of ‘Ardeosaurus’ digitatellus, which is more closely related to scleroglossans and eichstaettisaurids. Ardeosauridae, here, also includes P. bavarica and S. dyspepsia. The present analysis recovers three unambiguous ardeosaurid synapomorphies. 1. A well-developed posteroventral process is present on the jugal (character 48, state 0). 2. There is no lateral flange of the parietal at the frontoparietal suture (character 72, state 0). 3. The humerus possesses a hook-like and postglenoid process (character 407, state 1). Palaeolacerta, Schoenesmahl and A. brevipes Within Ardeosauridae, PMU.R58 is the basalmost taxon. All other ardeosaurids are united to the exclusion of PMU.R58 by the shared presence of a weakly inclined anterodorsal (narial) border of the external naris wherein the anterior margin of the nasal process is not distinctly offset from the premaxillary process (character 29, state 1). Schoenesmahl and A. brevipes A clade containing S. dyspepsia, SNSB-BSPG 1923 I 501 and NHM PR 38006 is here recovered to the exclusion of other specimens and is united by a single unambiguous synapomorphy: the retroarticular process is posteriorly broadened (character 207, state 1). SNSB-BSPG 1923 I 501 and NHM PR 38006 The specimens SNSB-BSPG 1923 I 501 and NHM PR 38006 are found to be sister terminals. The analysis recovers one unambiguous synapomorphy for them: the astragalus and calcaneum remain unfused (character 290, state 0; see below). Phylogenetic topology rerooted on Dibamidae The present phylogenetic analysis yields a novel phylogenetic hypothesis network that shows similarities with morphological phylogenetic hypotheses, but also has some relationships reminiscent of phylogenetic networks based on genetic data (compare topologies in Fig. 28A–G). Figure 28. View largeDownload slide Comparative phylogenetic relationships for Squamata. A, a simplified phylogenetic network based on the present analysis with two possible rooting points (Iguanomorpha and Dibamidae). Note that Lacertoidea is paraphyletic with respect to Anguimorpha in this network. B, the present morphological network rooted to Iguanomorpha; note that Gekkonomorpha, Scincoidea and Dibamidae form a clade to the exclusion of lacertoids and anguimorphs. C, the present morphological network rooted to Dibamidae; note that iguanomorpha, Lacertoidea and Anguimorpha form a clade exclusive of Scincoidea and that D, Polyglyphanodontia, Eichstaettisauridae, Ardeosauridae and Bavarisaurus macrodactylus are part of Iguanomorpha with this topology and rooting. E, a simplified phylogenetic network based on some recent molecular analyses (e.g. Vidal & Hedges, 2009; Pyron et al., 2013) with possible rooting points on Dibamidae and Iguanomorpha indicated. F, the molecular network rooted on Iguanomorpha; note the highlighted similarities with B. G, the molecular network rooted on Dibamidae; note the highlighted similarities with C. H, the morphological network presented by Gautheir et al. (2012); note the broad scale differences of this phylogenetic network as compared to A and E. Figure 28. View largeDownload slide Comparative phylogenetic relationships for Squamata. A, a simplified phylogenetic network based on the present analysis with two possible rooting points (Iguanomorpha and Dibamidae). Note that Lacertoidea is paraphyletic with respect to Anguimorpha in this network. B, the present morphological network rooted to Iguanomorpha; note that Gekkonomorpha, Scincoidea and Dibamidae form a clade to the exclusion of lacertoids and anguimorphs. C, the present morphological network rooted to Dibamidae; note that iguanomorpha, Lacertoidea and Anguimorpha form a clade exclusive of Scincoidea and that D, Polyglyphanodontia, Eichstaettisauridae, Ardeosauridae and Bavarisaurus macrodactylus are part of Iguanomorpha with this topology and rooting. E, a simplified phylogenetic network based on some recent molecular analyses (e.g. Vidal & Hedges, 2009; Pyron et al., 2013) with possible rooting points on Dibamidae and Iguanomorpha indicated. F, the molecular network rooted on Iguanomorpha; note the highlighted similarities with B. G, the molecular network rooted on Dibamidae; note the highlighted similarities with C. H, the morphological network presented by Gautheir et al. (2012); note the broad scale differences of this phylogenetic network as compared to A and E. I rerooted the current cladistic network so that Dibamidae was the basalmost squamate clade following recent phylogenetic hypotheses based on genetic data (Fig. 28; Vidal & Hedges, 2004, 2005, 2009; Pyron, Burbrink & Wiens, 2013). The resulting tree was, of course, problematic because the non-squamate taxa (e.g. Kuehneosauridae, Rhynchocephalia) were forced inside of Squamata. Those problematic results aside, the rerooted analysis offers some noteworthy results. Among them is the recovery of a group including iguanomorphs and anguimorphs to the exclusion of Scincoidea (Fig. 28C). Lacertoidea is still the sister-group to Anguimorpha in that clade, but the topology is similar in that iguanomorphs and anguimorphs are united to the exclusion of scincoids. This clade recovers no unambiguous synapomorphies (although 21 ambiguous synapomorphies are recovered). The rerooted cladogram recovers an unconventional ‘Iguanomorpha’ (Fig. 28C, D). Polyglyphanodontia is here found to be the basalmost iguanomorph clade and the outgroup to a polychotomy composed of Ardeosauridae, E. schroederi, E. gouldi and the rest of the included species. This Iguanomorpha is supported by a single unambiguous synapomorphy: the prefrontal extends more than half the width of the frontal (unilaterally; character 580, state 1). Ardeosauridae is supported by the same synapomorphies in the rerooted tree as in the cladogram rooted to the kuehneosaurids. TAPHONOMY It was recently demonstrated that some supposed gut contents of Coelophysis bauri are not gut contents, but a product of a taphonomic accident (Nesbitt et al., 2006) based on the ribs from both sides of the specimen lying on top of the supposed prey item. Although this was the case for a juvenile Coe. bauri associated with a preserved adult, another Coe. bauri was shown to have eaten a crocodylomorph archosaur (Nesbitt et al., 2006). Coelurosaurian theropods diverged by the Middle Jurassic (e.g. Rauhut, Milner & Moore-Fay, 2010; Novas et al., 2012; Choiniere et al., 2014). Given that at least three compsognathid coelurosaurs are known to have consumed squamates (Ostrom, 1978; Evans, 1994b; Chen, Dong & Zhen, 1998; Dal Sasso & Maganuco, 2011), coelurosaurs have been important predators of squamates at least since the Late Jurassic and continue to be in modern times (see below). Compsognathus longipes clearly ate S. dyspepsia. As preserved, the S. dyspepsia clearly overlies the right ribs of C. longipes and is overlaid by the left ribs. A majority of the preserved elements of S. dyspepsia occur along the dorsal part of what was originally the body cavity of C. longipes, just ventral to the dorsal vertebral centra. It is clear that S. dyspepsia was ingested by the coelurosaur not long before the death of the latter. The position of S. dyspepsia within the C. longipes and its preservation offer some insights about the way S. dyspepsia was ingested. Schoenesmahl dyspepsia is oriented generally posteriorly–anteriorly within the body of C. longipes. Nonavian theropods apparently shared the two-part, proventriculus-gizzard stomach (Varricchio, 2001) found in modern birds (e.g. Beebe, 1906; Romer, 1949; Webster & Webster, 1974). Given such a stomach in C. longipes, the head of the squamate was in or near the gizzard of the dinosaur. The abdomen and right hind limb were in the area of the proventriculus. The left hind limb and pes, and the tail were in the proximal part of the proventriculus or, possibly, the distal part of the esophagus. Parts of the skeleton of S. dyspepsia remain articulated. The articulation suggests that the specimen was bitten or torn into two pieces. This is consistent with the ways in which some modern birds (e.g. Loggerhead Shrikes, Lanius ludovicianus) kill and partly dismember squamate prey (Andrews, 1990; Yosef, 1996; Polcyn et al., 2002; Yosef, 2004). Lanius spp. disarticulate the skull from the vertebral column to kill their squamate prey, then secure it by impalement or by wedging it in the fork of a tree branch so that they can pull it apart (Yosef, 1996, 2004). Many predatory birds, including raptors, may hold down relatively large prey items with their pedes and pull them apart with their beaks (Snively & Russell, 2007; Snively et al., 2013). Given the presence of potentially very dexterous forelimbs and manus in nonavian theropods (Sereno, 1993; Gishlick, 2001; Carpenter, 2002), C. longipes might have been capable of stabilizing a prey item (such as a lizard) with its manus and using the manus and teeth to dismember the prey. The orientation of the preserved part of the left humerus of S. dyspepsia with respect to the preserved skull elements offers some weak suggestion that the humerus might have remained articulated with the anterior part of the skeleton when the squamate was ingested (Figs 2, 3). This is bolstered somewhat by the preservation of two mid-dorsal vertebrae lying perpendicular to the C. longipes vertebral column and also to the remaining S. dyspepsia vertebrae. If the association between the skull, the humerus and these two vertebrae is not an artefact of preservation, then the anterior half of the precaudal skeleton may have been ingested as a unit. This piece would have been approximately equivalent to three-quarters the length of the skull of its predator. The posterior presacral vertebrae, hind limbs and the tail remain in partial articulation. Exclusive of the tail, this potential second piece would be similar in length to the anterior portion. Ostrom’s (1978) study identified the folded nature of the lizard’s caudal vertebrae, giving some indication that the tail may have been folded over itself multiple times in the gut of C. longipes and that it may have been consumed whole. The skull of S. dyspepsia is somewhat disarticulated. None of the braincase is visible, the preserved skull roofing bones remain mostly associated and/or slightly out of articulation. The left mandible lies well outside of what might otherwise be expected as the body cavity of C. longipes, except for the associated posterior gastralia. It appears that the abdomen was blown out by postmortem buildup of gases associated with decomposition, probably in the gizzard part of the stomach. DISCUSSION Distinctiveness of S. dyspepsia The traditional association between S. dyspepsia with Ba. macrodactylus is plausible given their temporal and geographic co-occurrence, their similar size and the shared possession of very elongate hind limbs compared to their bodies (Figs 29A, B). Even so, S. dyspepsia is demonstrably distinct from Ba. macrodactylus (see Diagnosis, above). Each of these taxa is generally similar in size to the Solnhofen squamates Ardeosaurus (SNSB-BSPG 1923 I 501; Fig. 30A), the PMU.R58 specimen (Fig. 30B), A. digitatellus (Fig. 30C) and E. schroederi (Fig. 30D), but those taxa have less elongate hind limbs than either Ba. macrodactylus or S. dyspepsia. All of these taxa are much larger than P. bavarica (Fig. 31A) and the cast of the holotype of A. brevipes (Fig. 31B). The distinctiveness of the S. dyspepsia is demonstrated by details of the skull roofing bones, mandible and limb proportions. Although S. dyspepsia and Ba. macrodactylus share the possession of similarly elongate hind limbs as compared to their respective skulls (Fig. 29), the proportions of the femora, zeugopodia and pedes are different between the two taxa (Table 1). The NHM PR 38006 specimen can be coded for 71 characters and SNSB-BSPG 1923 I 501 can be coded for 79 (Supporting Information, Appendices SI and SII). These two specimens have 42 characters for which they both may be coded and they are identical in all of those character states. Figure 29. View largeDownload slide Skeletal reconstructions of A, Schoenesmahl dyspepsia gen. et. sp. nov. and B, Bavarisaurus macrodactylus in dorsal view with speculative body silhouettes. Figure 29. View largeDownload slide Skeletal reconstructions of A, Schoenesmahl dyspepsia gen. et. sp. nov. and B, Bavarisaurus macrodactylus in dorsal view with speculative body silhouettes. Figure 30. View largeDownload slide Skeletal reconstructions of some Solnhofen lizards in dorsal view with speculative body silhouettes. A, SNSB-BSPG 1923 I 501 (referred to Ardeosaurus brevipes here). B, PMU.R58, a specimen previously referred to A. brevipes (Mateer, 1982), but considered to be distinct genus and species here. C, CM 4026, holotype of Ardeosaurus digitatellus. D, SNSB-BSPG 1937 I 1a,b, holotype of Eichstaettisaurus schroederi. Figure 30. View largeDownload slide Skeletal reconstructions of some Solnhofen lizards in dorsal view with speculative body silhouettes. A, SNSB-BSPG 1923 I 501 (referred to Ardeosaurus brevipes here). B, PMU.R58, a specimen previously referred to A. brevipes (Mateer, 1982), but considered to be distinct genus and species here. C, CM 4026, holotype of Ardeosaurus digitatellus. D, SNSB-BSPG 1937 I 1a,b, holotype of Eichstaettisaurus schroederi. Figure 31. View largeDownload slide Skeletal reconstructions of some Solnhofen lizards in dorsal view with speculative body silhouettes. A, Palaeolacerta bavarica reconstructed after. B, NHM PR 38006, Ardeosaurus brevipes. Figure 31. View largeDownload slide Skeletal reconstructions of some Solnhofen lizards in dorsal view with speculative body silhouettes. A, Palaeolacerta bavarica reconstructed after. B, NHM PR 38006, Ardeosaurus brevipes. Phylogeny Lepidosaur phylogeny Divergence between Rhynchocephalia and Squamata occurred by around 245 Mya (Jones et al., 2013), antedating the first appearance of squamates in the fossil record by at least 65 Myr. The oldest putative squamate record is B. rebbanensis (Evans, Prasad & Manhas, 2001). The present analysis suggests that B. rebbanensis is a derived rhynchocephalian (see above). However, this taxon is known from very fragmentary remains (Evans et al., 2001) and the discovery of more complete fossils might show otherwise, especially given the homoplastic nature of some of the characters holding B. rebbanensis in its present phylogenetic position. Tikiguana estesi was originally described as a Late Triassic squamate (Datta & Ray, 2006) have been convincingly refuted (Hutchinson et al., 2012). Consequently, the earliest unequivocal squamates are some basal snakes (Caldwell et al., 2015; see below), a heavily armoured squamate from the earliest Late Jurassic of China (IVPP V 17816; Conrad et al., 2013), and the Solnhofen squamates. Because of their ages, these taxa are important for inclusion in any phylogenetic analysis attempting to identify the basal relationships of Squamata. Parviraptor was originally described as an anguimorph (Evans, 1994a) and has sometimes been recovered as such in subsequent cladistic analyses (Fig. 32A; Evans & Wang, 2005; Evans et al., 2005, 2006). Parviraptor has also been suggested as a close relative of gekkotans (Fig. 32B; Conrad, 2008), near the base of crown-group squamates (Fig. 32C; Bolet & Evans, 2012), or a basal snake (Caldwell et al., 2015). The present analysis confirms the results of a recent study (Caldwell et al., 2015) demonstrating that Parviraptor estesi and Portugalophis lignites represent basal snakes. Figure 32. View largeDownload slide Published phylogenetic hypotheses for Squamata and/or Lepidosauria focusing on the parts of the trees addressing the Solnhofen taxa. A, Evans et al. (2006); B, Conrad (2008); C, Bolet & Evans (2012); D, Gauthier et al. (2012). For the purposes of brevity and clarity, some of the clades that were more densely sampled in the illustrated cladograms have been collapsed and some of the taxon names have been modified to conform to the current usage (e.g. Anguiformes, Norellius nyctisaurops, etc.). Figure 32. View largeDownload slide Published phylogenetic hypotheses for Squamata and/or Lepidosauria focusing on the parts of the trees addressing the Solnhofen taxa. A, Evans et al. (2006); B, Conrad (2008); C, Bolet & Evans (2012); D, Gauthier et al. (2012). For the purposes of brevity and clarity, some of the clades that were more densely sampled in the illustrated cladograms have been collapsed and some of the taxon names have been modified to conform to the current usage (e.g. Anguiformes, Norellius nyctisaurops, etc.). The Solnhofen squamates have appeared in only a limited fashion in phylogenetic analyses (e.g. Fig. 32). Bavarisaurus macrodactylus is usually recovered as falling outside of the squamate crown (Fig. 32A; Evans & Barbadillo, 1998, 1999; Evans & Wang, 2005; Evans et al., 2005) or just within the basal squamate dichotomy (Fig. 32B, C; Conrad, 2008; Bolet & Evans, 2012). Ardeosaurus is also usually recovered as a basal member of the squamate crown (Fig. 32B, C; Evans & Barbadillo, 1998, 1999; Evans & Wang, 2005; Evans et al., 2005). Eichstaettisaurus is variably found as a basal squamate (Fig. 32A–C; Evans & Barbadillo, 1998, 1999), as a basal anguimorph (Evans & Wang, 2005; Evans et al., 2005), or as a proximal outgroup to Gekkota (Fig. 32D; Gauthier et al., 2012). Other basal and/or squamates have an equally spotty inclusion in cladistic phylogenetic analyses. Jucaraseps grandipes, an Early Cretaceous squamate from Spain, was only recently described and has been suggested to be a basal squamate (Fig. 32C; Bolet & Evans, 2012). Chometokadmon fitzingeri (Fig. 33) was described as a basal anguimorph close to the anguimorph crown-group (Anguiformes sensuConrad, 2008; Conrad et al., 2011a) at a similar phylogenetic level as Dorsetisaurus (Fig. 32A; Evans et al., 2006). The most extensive recent morphological analysis of lepidosaur interrelationships (Gauthier et al., 2012) unfortunately did not include the Solnhofen squamates. The lone Solnhofen-occurring taxon included in Gauthier et al. (2012) was E. schroederi, but that taxon was compositely coded with the Cretaceous E. gouldi as ‘Eichstaettisaurus’ (Gauthier et al., 2012: 29). Despite being some of the most completely known pre-Cretaceous squamates (Figs 3, 4, 19, 29–31, 33), the other Solnhofen squamates were not even mentioned in that study. Figure 33. View largeDownload slide The skull of Chometokadmon fitzingeri reconstructed after Evans et al. (2006). Semi-opaque grey layers indicate missing parts restored here. Figure 33. View largeDownload slide The skull of Chometokadmon fitzingeri reconstructed after Evans et al. (2006). Semi-opaque grey layers indicate missing parts restored here. There is very little consensus regarding the phylogeny of the major squamate groups based on morphological data (e.g. Estes et al., 1988; Evans, 1998; Lee, 1998; Caldwell, 2000; Evans & Wang, 2005; Evans et al., 2005, 2006; Conrad, 2008; Gauthier et al., 2012; reviewed by Conrad, 2008). By contrast, recent analyses using exclusively or primarily molecular data for tree reconstruction are very consistent with one another (e.g. Townsend et al., 2004; Vidal & Hedges, 2005; Wiens et al., 2010; Wiens et al., 2012b; Pyron et al., 2013; Reeder et al., 2015), but are totally at odds with morphological phylogenetic hypotheses. Most analyses of molecular data find dibamids as basal within Squamata (Fig. 28; e.g. Vidal & Hedges, 2009; Pyron et al., 2013, but see, e.g., Wiens et al., 2012a; Reeder et al., 2015 wherein gekkotans are the sister taxon to dibamids). The present analysis suggests the polyphyly of taxa usually attributed to Ardeosaurus (Fig. 27). A specimen referred to A. brevipes (PMU.R58; Mateer, 1982) possesses proportionately larger eyes and a smaller supratemporal fenestra (Fig. 34D) than in the holotype of that species. The current systematic hypothesis might allow for PMU.R58 to be referred to the genus Ardeosaurus, but only if P. bavarica and S. dyspepsia were referred to Ardeosaurus, too. Given the morphological diversity present between these five ardeosaurid specimens (compare Figs 6, 29A, 30A–C, 31), generic separation is recommended. By contrast to the situation with PMU.R58, it is impossible to maintain the ‘Ardeosaurus’ digitatellus holotype (CM4026) in a holophyletic Ardeosaurus given the present cladistic networks because doing so would require Ardeosaurus to include one or more major squamate clades (Figs 27, 28D). Figure 34. View largeDownload slide Reconstructed skulls of some Solnhofen lizards in dorsal view with restored parts shown as semi-opaque grey layers. A, NHM PR 38006, Ardeosaurus brevipes. B, SNSB-BSPG 1923 I 501, referred to A. brevipes here. C, CM 4026, Ardeosaurus digitatellus. D, PMU.R58, a specimen previously referred to A. brevipes (Mateer, 1982), but considered to be distinct genus and species here, drawn after Mateer (1982). Figure 34. View largeDownload slide Reconstructed skulls of some Solnhofen lizards in dorsal view with restored parts shown as semi-opaque grey layers. A, NHM PR 38006, Ardeosaurus brevipes. B, SNSB-BSPG 1923 I 501, referred to A. brevipes here. C, CM 4026, Ardeosaurus digitatellus. D, PMU.R58, a specimen previously referred to A. brevipes (Mateer, 1982), but considered to be distinct genus and species here, drawn after Mateer (1982). The holotype for the type species (A. brevipes; Figs 10A, 19A, 34A) is lost and is currently represented only by NHM PR 38006, a cast. The specimen SNSB-BSPG 1923 I 501 (Fig. 10B, 19B, 34B) traditionally has been referred to A. cf. digitatellus with the caveat that it and A. digitatellus (CM 4026; Figs 30C, 34C) might simply represent adults of A. brevipes. The specimen SNSB-BSPG 1923 I 501 is even labelled ‘Ardeosaurus brevipes’ in the collection. The two specimens NHM PR 38006 and SNSB-BSPG 1923 I 501 are very similar in form and in the comparable details of their morphology (Figs 10, 19) as well as their proportions (Figs 19, 30A, 31B; Table 1). Based on the present analysis and observations of the two skeletons, I see no reason to separate the two at the species level. My only reservations regarding such a referral is that I see no clear juvenile features (e.g. open sutures, enlarged pineal foramen, incompletely ossified distal limb elements) in the holotype (which is a noticeably smaller individual than SNSB-BSPG 1923 I 501) and the specimens are incompletely preserved. The two specimens might represent two very closely related species. It is even possible that the two specimens preserve sexual dimorphism within A. brevipes, but this would be difficult to substantiate and there is no apparent evidence of that. It seems prudent to refer them both to A. brevipes. Comparison with the molecular network Rerooting a standard genetics-based phylogenetic network so that iguanians and their fossil relatives are basal reveals a cladogram with similarities to the present morphological one; similarly rerooting the present morphology-based phylogenetic network so that dibamids are basal presents a cladogram with unexpected similarities to molecular-based phylogenetic hypotheses (Fig. 28E–G; see above). Recent phylogenetic analyses based on molecular data consistently find dibamids and gekkotans as basal squamates, lacertids as being closely associated with amphisbaenians, and a close relationship between snakes, iguanians and anguimorphs (Fig. 28E, G). When the molecular network is rerooted to make Iguania basal, the resulting cladogram shares with phylogenetic hypothesis produced for the present study a close relationship between scincoids, dibamids and gekkonomorphs (Fig. 28F). Granted, the present study also finds amphisbaenians and snakes to be a part of the scincoid-dibamid-gekknomorph clade (Fig. 27, 28B) – a result not supported in the rerooted molecular cladogram. Even so, the suggestion of a relationship between those three groups to the exclusion of anguimorphs and lacertoids must possess some cryptic molecular support based on that topology. By contrast, no rerooting offers any support for the Anniella-Krypteia clade suggested in a recent morphological study (Gauthier et al., 2012); that result may reflect the limited taxon sampling of that study. Besides the Anniella-Krypteia clade, the Gauthier et al. (2012) study recovers the same basic squamate interrelationships as the first major cladistic analysis of Squamata (Estes et al., 1988) – including some traditional groups (e.g. a Scincomorpha including scincoids and lacertoids) not recovered in the present morphological analysis (Fig. 28H). A combination of expanded taxonomic sampling and the addition of morphological data of a type that is not usually incorporated into phylogenetic analyses may be partly responsible for these unanticipated results (Table 2). Table 2. Taxonomy of selected Jurassic European squamates from recent studies and this study Specimen #  Estes (1983), ‘family’  Estes (1983), species  This study, ‘family’  This study, species  BSP 1873 III 501  Bavarisauridae  Bavarisaurus macrodactylus  Bavarisauridae  Bavarisaurus macrodactylus  NHM PR 38006  Ardeosauridae  Ardeosaurus brevipes  Ardeosauridae  Ardeosaurus brevipes  SNSB-BSPG AS I 563b  Bavarisauridae  Bavarisaurus cf. macrodactylus  Ardeosauridae  Schoenesmahl dyspepsia  BSP 1937 I 1a, b  Ardeosauridae  Eichstaettisaurus schroederi  Eichstaettisauridae  Eichstaettisaurus schroederi  BSP specimen  Bavarisauridae  Palaeolacerta bavarica  Ardeosauridae  Palaeolacerta bavarica  CM 4026  Ardeosauridae  Ardeosaurus digitatellus  Indeterminate  Unnamed genus digitatellus  BSP 1923 I 501  Ardeosauridae  Ardeosaurus cf. digitatellus  Ardeosauridae  Ardeosaurus brevipes  PMU.R58  Ardeosauridae  Ardeosaurus brevipes  Ardeosauridae  Unnamed genus and species  Specimen #  Estes (1983), ‘family’  Estes (1983), species  This study, ‘family’  This study, species  BSP 1873 III 501  Bavarisauridae  Bavarisaurus macrodactylus  Bavarisauridae  Bavarisaurus macrodactylus  NHM PR 38006  Ardeosauridae  Ardeosaurus brevipes  Ardeosauridae  Ardeosaurus brevipes  SNSB-BSPG AS I 563b  Bavarisauridae  Bavarisaurus cf. macrodactylus  Ardeosauridae  Schoenesmahl dyspepsia  BSP 1937 I 1a, b  Ardeosauridae  Eichstaettisaurus schroederi  Eichstaettisauridae  Eichstaettisaurus schroederi  BSP specimen  Bavarisauridae  Palaeolacerta bavarica  Ardeosauridae  Palaeolacerta bavarica  CM 4026  Ardeosauridae  Ardeosaurus digitatellus  Indeterminate  Unnamed genus digitatellus  BSP 1923 I 501  Ardeosauridae  Ardeosaurus cf. digitatellus  Ardeosauridae  Ardeosaurus brevipes  PMU.R58  Ardeosauridae  Ardeosaurus brevipes  Ardeosauridae  Unnamed genus and species  Changes are in bold. View Large Table 2. Taxonomy of selected Jurassic European squamates from recent studies and this study Specimen #  Estes (1983), ‘family’  Estes (1983), species  This study, ‘family’  This study, species  BSP 1873 III 501  Bavarisauridae  Bavarisaurus macrodactylus  Bavarisauridae  Bavarisaurus macrodactylus  NHM PR 38006  Ardeosauridae  Ardeosaurus brevipes  Ardeosauridae  Ardeosaurus brevipes  SNSB-BSPG AS I 563b  Bavarisauridae  Bavarisaurus cf. macrodactylus  Ardeosauridae  Schoenesmahl dyspepsia  BSP 1937 I 1a, b  Ardeosauridae  Eichstaettisaurus schroederi  Eichstaettisauridae  Eichstaettisaurus schroederi  BSP specimen  Bavarisauridae  Palaeolacerta bavarica  Ardeosauridae  Palaeolacerta bavarica  CM 4026  Ardeosauridae  Ardeosaurus digitatellus  Indeterminate  Unnamed genus digitatellus  BSP 1923 I 501  Ardeosauridae  Ardeosaurus cf. digitatellus  Ardeosauridae  Ardeosaurus brevipes  PMU.R58  Ardeosauridae  Ardeosaurus brevipes  Ardeosauridae  Unnamed genus and species  Specimen #  Estes (1983), ‘family’  Estes (1983), species  This study, ‘family’  This study, species  BSP 1873 III 501  Bavarisauridae  Bavarisaurus macrodactylus  Bavarisauridae  Bavarisaurus macrodactylus  NHM PR 38006  Ardeosauridae  Ardeosaurus brevipes  Ardeosauridae  Ardeosaurus brevipes  SNSB-BSPG AS I 563b  Bavarisauridae  Bavarisaurus cf. macrodactylus  Ardeosauridae  Schoenesmahl dyspepsia  BSP 1937 I 1a, b  Ardeosauridae  Eichstaettisaurus schroederi  Eichstaettisauridae  Eichstaettisaurus schroederi  BSP specimen  Bavarisauridae  Palaeolacerta bavarica  Ardeosauridae  Palaeolacerta bavarica  CM 4026  Ardeosauridae  Ardeosaurus digitatellus  Indeterminate  Unnamed genus digitatellus  BSP 1923 I 501  Ardeosauridae  Ardeosaurus cf. digitatellus  Ardeosauridae  Ardeosaurus brevipes  PMU.R58  Ardeosauridae  Ardeosaurus brevipes  Ardeosauridae  Unnamed genus and species  Changes are in bold. View Large Ardeosauridae, Bavarisauridae and Eichstaettisauridae Although only Ardeosaurus was mentioned in the original description of Ardeosauridae (Camp, 1923), Y. tenuis has been (sometimes tenuously) considered a possible ardeosaurid (Romer, 1956; Hoffstetter, 1962, 1964; Estes, 1983; Kluge, 1987), as has E. schroederi (Hoffstetter, 1962, 1964; Estes, 1983; Kluge, 1987). Eichstaettisauridae was originally erected for E. schroederi (Kuhn, 1958), but the latter species was subsequently referred to Ardeosauridae and Eichstaettisauridae was considered a junior synonym of Ardeosauridae (Estes, 1983). Here, Ardeosauridae is defined as a node-based clade: all taxa sharing a more recent common ancestor with A. brevipes than with Ba. macrodactylus, An. fragilis, Coluber constrictor, G. gecko, Iguana iguana, L. viridis or Sci. scincus. Eichstaettisaurus schroederi is not found to fall within Ardeosauridae in the present analysis (Fig. 27); this result is consistent with recent studies addressing the Solnhofen squamates (Evans & Barbadillo, 1998, 1999; Evans & Wang, 2005; Evans et al., 2005; Conrad, 2008). As with other cladistic analyses addressing Y. tenuis, that species is here found to be a nested scincogekkonomorph, not associated with Ardeosauridae (Evans et al., 2005; Conrad, 2008). The present analysis demonstrates the distinctiveness of Eichstaettisauridae. Eichstaettisauridae is defined as all taxa sharing a more recent common ancestor with E. schroederi than with Ba. macrodactylus, An. fragilis, Col. constrictor, G. gecko, I. iguana, L. viridis or Sci. scincus. Note that should subsequent studies demonstrate that E. schroederi is an ardeosaurid, then Eichstaettisauridae will then be considered a junior synonym of Ardeosauridae based on this definition. Bavarisauridae here differs from prior understanding of the group, which typically suggests the inclusion of Ba. macrodactylus and P. bavarica (Fig. 31A; Estes, 1983; Kluge, 1987). Palaeolacerta bavarica was described more than 50 years ago (Cocude-Michel, 1961), but a complete description of the type material has not yet appeared. Published data addressing the morphology of P. bavarica (Cocude-Michel, 1961; Estes, 1983) demonstrate its distinctiveness from known contemporary squamates (e.g. paired frontals, fused parietal, pineal foramen apparently located at the frontoparietal suture, shape of the supratemporal processes of the parietal, the apparent absence of hook-like postglenoid humeral process; Fig. 35). Palaeolacerta bavarica and S. dyspepsia, both usually considered close to ‘bavarisaurids’, are here found to be ardeosaurids. The current phylogenetic context would render any Bavarisauridae that could be defined here as monospecific (Fig. 27). Figure 35. View largeDownload slide The skull of Palaeolacerta bavarica reconstructed after Hoffstetter (1964). Semi-opaque grey layers indicate missing parts restored here. The outlined grey areas associated with the supratemporal process indicate that those processes are partly preserved, but the bone surfaces are incomplete. The dotted lines indicate tentative margins for restored bone margins and/or sutures. Figure 35. View largeDownload slide The skull of Palaeolacerta bavarica reconstructed after Hoffstetter (1964). Semi-opaque grey layers indicate missing parts restored here. The outlined grey areas associated with the supratemporal process indicate that those processes are partly preserved, but the bone surfaces are incomplete. The dotted lines indicate tentative margins for restored bone margins and/or sutures. Limb proportions of basal squamates The holotype of S. dyspepsia was traditionally allied with Bavarisaurus partly because of its long hind limbs. The hind limb in S. dyspepsia is more than 4.3 times the length of the skull roof at the midline (Fig. 29A). Bavarisaurus macrodactylus also has an elongate hind limb as compared to its skull (as restored) – the hind limb is more than 3.6 times the length of the skull roof (Fig. 29B). Such elongate limbs are unusual among squamates. Chometokadmon fitzingeri, an Early Cretaceous squamate of a similar grade to Ba. macrodactylus in this analysis, possesses a hind limb only approximately 1.3 times the length of the dorsal skull roof. Scandensia ciervensis and E. schroederi form an unresolved trichotomy with ‘Eichstaettisaurus’ gouldi in the present analysis. The skull is too incompletely preserved in Sca. ciervensis for a confident reconstruction, but the hind limb in that taxon is approximately 2.3 times the length of the mandible. A. brevipes specimens NHM PR 38006 (Fig. 31B) and SNSB-BSPG 1932 I 501 (Fig. 30A) possess hind limbs that are 2.2 times the skull roof length. This contrasts the proportions of PMU R.58 (Fig. 30B) and, to a greater degree, CM 4026 (Fig. 30C) which possess hind limbs that are 2.1 and 2.0 times the skull lengths, respectively. The hind limb is approximately 2.4 times the length of the skull in E. schroederi (Fig. 30D) and approximately 2.9 times the length of the skull in Hu. mixtecus. Thus, it appears that the very elongate hind limb appeared independently in S. dyspepsia and Ba. macrodactylus. This may also be reflected in the differing proportions of the hind limb between the two taxa. Size Because of the missing anterior precaudal vertebrae and the partly disarticulated nature of the skull, the exact precaudal and total body lengths of S. dyspepsia remain uncertain. As restored (Fig. 29A), I have reconstructed the skull roof (SRL) as being approximately 15 mm long. The precaudal (PCL) and total body lengths (TL) are reconstructed as approximately 75 and 250 mm, respectively. The reconstructed caudal length is a somewhat shorter caudal length (approximately 175 mm) than the 190 mm suggested previously (Ostrom, 1978). This, even though I used the Ostrom (1978) study along with observations of the specimen to reconstruct the animal. Even with this somewhat shorter tail, the animal still possesses a very long tail compared to the reconstructed precaudal length. If the precaudal length is reasonable at 75 mm, then the tail constitutes approximately 70% of the total body length and more than twice the PCL length. Bavarisaurus macrodactylus has a PCL of approximately 91 mm (Fig. 29B) after the SRL is reconstructed (Fig. 8B). Eichstaettisaurus schroederi (SNSB-BSPG 1937 I 1a, b) is very complete and may be confidently reconstructed with a PCL of 85.5 mm, somewhat smaller than that of Ba. macrodactylus, but larger than the reconstructed size of S. dyspepsia. The cast of the A. brevipes holotype (NHM PR 38006) has a reconstructed SRL of approximately 11 mm and a PCL of approximately 58 mm. This is significantly smaller than the specimen referred to that species here (SNSB-BSPG 1923 I 501), which has a PCL of approximately 84 mm. Schoenesmahl dyspepsia and C. longipes Clear predator–prey interactions are rarely preserved in the fossil record. In the case of S. dyspepsia and C. longipes, there can be little doubt about the relationship between the former as a prey item of the latter. I suggest that the total length of S. dyspepsia was approximately 250 mm. This is noteworthy because the total length of C. longipes is approximately 890 mm (Ostrom, 1978). Thus, S. dyspepsia is more than one-quarter the length of the predator which ate it. As described above (see Taphonomy), the lizard may have been bitten into two pieces. The front half of the lizard seems to have suffered more trauma and is less completely known than the back part. The hind limbs seem mostly intact and the tail is folded into the anterior part of the gastrointestinal tract. The folded condition of the tail is significant. Specifically, it raises the question: how did the tail come to be folded into the gullet of this coelurosaur? Certainly, modern birds lack the morphological equipment to fold a lizard tail into their mouths. Predatory birds tear their prey into pieces they can swallow (Yosef, 1996, 2004; Fargallo et al., 2003; Young, Brodie & Brodier, 2004) or swallow it whole and usually head-first (e.g. Newton & Newton, 1859; Brooker & Ridpath, 1980; Sherbrooke, 1990; Fargallo et al., 2003). I posit that the coelurosaur used its manus to place into its mouth the lizard, and subsequently used its manual digits to fold the tail into its oropharynx. Despite the presence of numerous predator–prey interactions captured in the fossil record (e.g. Chen et al., 1998; Evans et al., 2004; Dal Sasso & Maganuco, 2011), the S. dyspepsia–C. longipes interaction is unusual in producing a situation wherein both taxa can be diagnosed and each is a holotype. Evolution of predator avoidance in Solnhofen squamates The recent revelation that snakes appeared by the Early Jurassic significantly extends the unknown lineages for many squamate. The Solnhofen squamates form a paraphyletic assemblage with respect to other known squamates, but most are similar in possessing relatively short limbs (Fig. 30, 31). Most of these species are of similar size to one another, but P. bavarica (Fig. 31A) is much smaller than the others. Elongate limbs independently evolved in Ba. macrodactylus and S. dyspepsia (Fig. 29). Exactly what this means with regards to the evolution of predator avoidance and ecology of squamates within the Solnhofen system cannot be confidently reconstructed. Apparently, S. dyspepsia and Ba. macrodactylus were somehow different from other Solnhofen forms. Ardeosaurus brevipes (Fig. 30A, 31B) and ‘Ardeosaurus’ digitatellus (Fig. 30C) were relatively large-headed as compared to PMU R.58 (Fig. 30B), E. schroederi (Fig. 30D) and P. bavarica (Fig. 31A), but this may be related to prey choice (Pianka & Vitt, 2003) rather than predator avoidance. The limb proportions may be related to predator avoidance and, perhaps, with diurnality versus nocturnality. Modern lizards with relatively elongate limbs are generally diurnal and cursorial and/or scansorial whereas species with limbs that are shorter compared to their PCL are typically crevice-dwelling and/or nocturnal (see data in Müller et al., 2011 and Pianka & Vitt, 2003) and possess relatively short limbs. Note that arboreal or semi-arboreal forms may have essentially any limb proportions (see Pianka & Vitt, 2003; Müller et al., 2011). Perhaps the short limbed Solnhofen forms were mostly nocturnal or crepuscular. Because the emerging coelurosaurs may also have functioned well in low-light conditions (Schmitz & Motani, 2011), the elongate limbs of Ba. macrodactylus and S. dyspepsia may represent a shift toward relying less upon nocturnal habits for defence and more on cursoriality. CONCLUSIONS Schoenesmahl dyspepsia is an ardeosaurid closely related to A. brevipes, but distinct from it in various ways including the possession of proportionately much longer hind limbs. The very elongate hind limbs in S. dyspepsia are shared with Ba. macrodactylus with which it has traditionally been allied. Even so, the relative proportions of the femur and tibia are different between these two long-limbed forms. This, and differences in the skull definitively differentiate S. dyspepsia from Ba. macrodactylus. Bavarisaurus macrodactylus, ‘Ardeosaurus’ digitatellus, Ardeosauridae and Eichstaettisauridae represent four separate Jurassic squamate radiations. The traditional understanding of Ardeosaurus with all the specimens usually referred to that taxon must be rejected as polyphyletic (Figs 27, 28D). Bharatagama rebbanensis, previously suggested as an early iguanomorph, is hypothesized to be a rhynchocephalian here. The present morphological analysis recovers a traditional topology wherein Iguanomorpha is basal within Squamata and a gekkonomorph-scincoid-lacertoid-anguimorph clade constitutes are united in a Scleroglossa. Even so, rerooting the phylogenetic network presented here produces a cladogram showing important similarities with recent genetics-based phylogeneis (Fig. 28A–C, E–G). This contrasts with another large-scale recent morphological phylogeny (Fig. 28H). The evolution of elongate limbs in the Solnhofen system may represent an advancement within the arms race between the emerging coelurosaurian theropod radiation and squamates. Regardless of what strategy S. dyspepsia relied upon for predator avoidance, it failed for SNSB-BSPG AS I 563b. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher's web-site: Appendix SI. Morphological characters used in the phylogenetic analysis. Appendix SII. Morphological phylogenetic data matrix. [Version of Record, published online 18 December 2017; http://zoobank.org/urn: lsid: zoobank.org:pub:B59389B9-28C8-43C2-9FA6-EC07E494C4A4] ACKNOWLEDGEMENTS This material is based, in part, upon work supported by the United States National Science Foundation under grant no. NSF EAR 1325457 and startup funds provided by New York Institute of Technology College of Osteopathic Medicine. First and foremost, I thank O. Rauhut for inviting me to work on SNSB-BSPG AS I 563b. He and A. Lopez-Arbarello showed me tremendous kindness and excellent hospitality on my visits to München. Additionally, M. Moser was always helpful on those trips. I am grateful for the hospitality and specimen access for the numerous institutions I visited to collect phylogenetic and specimen data, including M. Arnold, D. Frost, D. Kizirian, C. Mehling, M. Norell, C. Osmone, R. Pascocello (AMNH), A. Henrici, S. Trauth (CM), A Resetar, W. Simpson (FMNH), Xu X. (IVPP), F. Manthi (KNM), S. McLeod, V. Rhue (LACM), J. Müller, D. Schwarz-Wings (MB), J. Rosado (MCZ), M. Forir (MINS), R. Allain, S. Bailon (MNHN), C. Spencer (MVZ), P. Barrett (NHM), M. Carrano, A Wynn (NMNH), D. Pagnac (SDSMT), S. Evans (UCL), P. Holroyd (UCMP), D. Brinkman, J. Gauthier, G. Watkins-Colwell, K. Kyskowski (YPM) and M. Borsuk-Białynicka (ZPAL). I am grateful to A. Herrel for help with literature regarding bird and squamate predator–prey interactions. For assistance with figures and for helpful conversations, I am immensely grateful to A. M. Balcarcel, B. Beatty, C. M. Mehling, M. Mihlbachler and C. Osmone. Additionally, I thank C. Balcarcel, M. Balcarcel, C. Chisam, S. S. Conrad, K. T. Darby, R. Jennings and F. Lewis for help and support over the course of this work. REFERENCES Adobe Systems . 2010. Adobe Photoshop CS6 Extended . Alifanov VR . 2000. Macrocephalosaurs and the early evolution of lizards of Central Asia. Transactions of the Paleontological Institute of the Russian Academy of Sciences  272: 1– 126. Andrews P . 1990. Owls, caves, and fossils . Chicago: University of Chicago Press. Apestaguía S , Novas FE . 2003. Large Cretaceous sphenodontian from Patagonia provides insight into lepidosaur evolution in Gondwana. Nature  425: 609– 612. Google Scholar CrossRef Search ADS PubMed  Beebe W . 1906. The bird: its form and function . New York: Holt and Company. Google Scholar CrossRef Search ADS   Blakey R . 2011. Global palaeogeography . Flagstaff: Northern Arizona University. Bogert CM , Porter AP . 1967. A new species of Abronia (Sauria, Anguidae) from the Sierra Madre del Sur of Oaxaca, Mexico. American Museum Novitates  2279: 1– 21. Bolet A , Evans SE . 2012. A tiny lizard (Lepidosauria, Squamata) from the Lower Cretaceous of Spain. Palaeontology  55: 491– 500. Google Scholar CrossRef Search ADS   Borsuk-Białynicka M . 1988. Globaura venusta gen. et sp. n. and Eoxanta lacertifrons gen. et sp. n. – non-teiid lacertoids from the Late Cretaceous of Mongolia. Acta Palaeontologica Polonica  33: 211– 248. Brooker MG , Ridpath MG . 1980. The diet of the Wedge-Tailed Eagle, Aquila audax, in western Australia. Australian Wildlife Research  7: 433– 452. Google Scholar CrossRef Search ADS   Caldwell MW . 2000. On the aquatic squamate Dolichosaurus longicollis Owen, 1850 (Cenomanian, Upper Cretaceous), and the evolution of elongate necks in squamates. Journal of Vertebrate Paleontology  20: 720– 735. Google Scholar CrossRef Search ADS   Caldwell MW , Nydam RL , Palci A , Apestaguía S . 2015. The oldest known snakes from the Middle Jurassic-Lower Cretaceous provide insights on snake evolution. Nature Communications  6: 5996. Camp CL . 1923. Classification of the lizards. Bulletin of the American Museum of Natural History  48: 289– 481. Carpenter K . 2002. Forelimb biomechanics of nonavian theropod dinosaurs in predation. Senckenbergiana Lethaea  82: 59– 76. Google Scholar CrossRef Search ADS   Carroll RL . 1982. A short limbed lizard from the Lystrosaurus zone (Lower Triassic) of South Africa. Journal of Paleontology  1982: 183– 190. Carroll RL . 1985. A pleurosaur from the Lower Jurassic and the taxonomic position of the Sphenodontida. Palaeontographica Abteilung A  189: 1– 28. Carroll RL . 1988. Vertebrate paleontology and evolution . New York: W. H. Freeman and Company. Chen P-J , Dong Z-M , Zhen S-N . 1998. An exceptionally well-preserved theropod dinosaur from the Yixian Formation of China. Nature  391: 147– 152. Google Scholar CrossRef Search ADS   Choiniere JN , Clark JM , Forster CA , Norell MA , Eberth DA , Erickson GM , Chu H , Xu X . 2014. A juvenile specimen of a new coelurosaur (Dinosauria: Theropoda) from the Middle-Late Jurassic Shishugou Formation of Xinjiang, People’s Republic of China. Journal of Systematic Palaeontology  12: 177– 215. Google Scholar CrossRef Search ADS   Cocude-Michel M . 1961. Les Sauriens des calcaires lithographiques de Bavière d’âge portlandien infèrieur. Bulletin de la Société Géologique de France  2: 707– 710. Cocude-Michel M . 1963. Les rhynchocéphales et les sauriens des calcaires lithographiques (Jurassique-Superieur) d’Europe occidentale. Nouvelles archives du Muséum d’histoire naturelle de Lyon  7: 1– 187. Conrad JL . 2004. Skull, mandible, and hyoid of Shinisaurus crocodilurus Ahl (Squamata, Anguimorpha). Zoological Journal of the Linnean Society  141: 399– 434. Google Scholar CrossRef Search ADS   Conrad JL . 2008. Phylogeny and systematics of Squamata (Reptilia) based on morphology. Bulletin of the American Museum of Natural History  310: 1– 182. Google Scholar CrossRef Search ADS   Conrad JL , Ast JC , Montanari S , Norell MA . 2011a. A combined evidence phylogenetic analysis of Anguimorpha (Reptilia: Squamata). Cladistics  27: 230– 277. Google Scholar CrossRef Search ADS   Conrad JL , Balcarcel AM , Mehling CM . 2012. Earliest example of a giant monitor lizard (Varanus, Varanidae, Squamata). PLoS ONE  7: e41767. Google Scholar CrossRef Search ADS PubMed  Conrad JL , Daza JD . 2015. Naming and re-diagnosing the Cretaceous gekkonomorph (Reptilia, Squamata) from Öösh (Övörkhangai, Mongolia). Journal of Vertebrate Paleontology  35: e980891. Google Scholar CrossRef Search ADS   Conrad JL , Norell MA . 2007. A well-preserved iguanian (Squamata: Reptilia) from the Cretaceous of the Gobi and identification of a new clade of Iguania. American Museum Novitates  3584: 1– 47. Google Scholar CrossRef Search ADS   Conrad JL , Rieppel O , Gauthier JA , Norell MA . 2011b. Osteology of Gobiderma pulchrum (Monstersauria, Lepidosauria, Reptilia). Bulletin of the American Museum of Natural History  362: 1– 89. Google Scholar CrossRef Search ADS   Conrad JL , Wang Y , Xu X , Pyron A , Clark JM . 2013. Skeleton of a heavily armored and long legged Middle Jurassic lizard (Squamata, Reptilia). Society of Integrative and Comparative Biology  33( electronic supplement): 108. Costa OG . 1866. Nuove osservazioni e scoperte intorno ai fossili della calcarea ad illioliti di Pietraroja. Atti dell’Accademia delle scienze, fisiche e matematiche di Napoli  2: 1– 12. Dal Sasso C , Maganuco S . 2011. Scipionyx samniticus (Theropoda: Compsognathidae) from the Lower Cretaceous of Italy. Memorie della Società di Scienze Naturali e del Museo Civico di Storia Naturale di Milano  37: 1– 281. Datta PM , Ray S . 2006. Earliest lizard from the Late Triassic (Carnian) of India. Journal of Vertebrate Paleontology  26: 795– 800. Google Scholar CrossRef Search ADS   Endo R , Shikama T . 1942. Mesozoic reptilian fauna in the Jehol mountainland, Manchoukuo. Bulletin of the Central National Museum of Manchoukuo  3: 1– 19. Estes R . 1983. Sauria terrestria, Amphisbaenia . New York: Gustav Fischer Verlag. Estes R , de Queiroz K , Gauthier J . 1988. Phylogenetic relationships within Squamata. In: Estes R , Pregill G , eds. Phylogenetic relationships of the lizard families . Stanford: Stanford University Press, 119– 281. Evans SE . 1980. The skull of a new eosuchian reptile from the Lower Jurassic of South Wales. Zoological Journal of the Linnean Society  70: 203– 264. Google Scholar CrossRef Search ADS   Evans SE . 1994a. A new anguimorph lizard from the Jurassic and Lower Cretaceous of England. Palaeontology  37: 33– 49. Evans SE . 1994b. The Solnhofen (Jurassic: Tithonian) lizard genus Bavarisaurus: new skull material and a reinterpretation. Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen  192: 37– 52. Evans SE . 1998. Crown group lizards (Reptilia, Squamata) from the Middle Jurassic of the British Isles. Palaeontographica Abteilung A  250: 123– 154. Evans SE . 2001. The early Triassic ‘lizard’ Colubrifer campi: a reassessment. Palaeontology  44: 1034– 1041. Evans SE . 2003. At the feet of the dinosaurs: the early history and radiation of lizards. Biological Reviews  78: 513– 551. Google Scholar CrossRef Search ADS PubMed  Evans SE . 2009. An early kuehneosaurid reptile from the Early Triassic of Poland. Palaeontologica Polonica  65: 145– 178. Evans SE , Barbadillo LJ . 1998. An unusual lizard (Reptilia: Squamata) from the Early Cretaceous of Las Hoyas, Spain. Zoological Journal of the Linnean Society  124: 235– 265. Google Scholar CrossRef Search ADS   Evans SE , Barbadillo LJ . 1999. A short-limbed lizard from the Lower Cretaceous of Spain. Special Papers in Palaeontology Series  60: 73– 85. Evans SE , Prasad GVR , Manhas BK . 2001. Rhynchocephalians (Diapsida: Lepidosauria) from the Jurassic Kota Formation of India. Zoological Journal of the Linnean Society  133: 309– 334. Google Scholar CrossRef Search ADS   Evans SE , Prasad GVR , Manhas BK . 2002. Fossil lizards from the Jurassic Kota Formation of India. Journal of Vertebrate Paleontology  22: 299– 312. Google Scholar CrossRef Search ADS   Evans SE , Raia P , Barbera C . 2004. New lizards and rhynchocephalians from the Lower Cretaceous of southern Italy. Acta Palaeontologica Polonica  49: 393– 408. Evans SE , Raia P , Barbera C . 2006. The Lower Cretaceous lizard genus Chometokadmon from Italy. Cretaceous Research  27: 673– 683. Google Scholar CrossRef Search ADS   Evans SE , Wang Y . 2005. Early Cretaceous lizard Dalinghosaurus from China. Acta Palaeontologica Polonica  50: 725– 742. Evans SE , Wang Y . 2007. A juvenile lizard specimen with well-preserved skin impressions from teh Upper Jurassic/Lower Cretaceous of Daohugou, Inner Mongolia, China. Naturwissenschaften  9: 431– 439. Google Scholar CrossRef Search ADS   Evans SE , Wang Y . 2009. A long-limbed lizard from the Upper Jurassic/Lower Cretaceous of Daohugou, Ningcheng, nei Mongol, China. Vertebrata Pal Asiatica  47: 21– 34. Evans SE , Wang Y , Li C . 2005. The Early Cretaceous lizard genus Yabeinosaurus from China: resolving an enigma. Journal of Systematic Palaeontology  3: 319– 335. Google Scholar CrossRef Search ADS   Fargallo JA , Laaksonen T , Korpimäki E , Pöyri V , Griffith SC , Valkama J . 2003. Size-mediated dominance and begging behaviour in Eurasian kestral broods. Evolutionary Ecology Research  5: 549– 558. Fraser NC . 1982. A new rhynchocephalian from the British Upper Trias. Palaeontology  25: 709– 725. Fry BG , Vidal N , Norman JA , Vonk FJ , Scheib H , Ramjan SFR , Kuruppu S , Fung K , Hedges SB , Richardson MK , Hodgson WC , Ignjatovic V , Summerhayes R , Kochva E . 2006. Early evolution of the venom system in lizards and snakes. Nature  439: 584– 588. Google Scholar CrossRef Search ADS PubMed  Gauthier JA , Kearney M , Maisano JA , Rieppel O , Behlke DB . 2012. Assembling the squamate tree of life: perspectives from the phenotype and the fossil record. Bulletin of the Peabody Museum of Natural History  53: 3– 308. Google Scholar CrossRef Search ADS   Gishlick AD . 2001. The function of the manus and forelimb of Deinonychus antirrhopus and its importance for the origin of avian flight. In: Gauthier JA , Gall LF , eds. New perspectives on the origin and early evolution of birds . New Haven: Yale Peabody Museum, 301– 318. Goloboff PA , Farris JS , Nixon KC . 2003. T.N.T.: tree analysis using new technology, Version 1.0 . Available at: www.zmuc.dk/public/phylogeny: Goloboff, Farris, and Nixon Goloboff PA , Farris JS , Nixon KC . 2008. TNT, a free program for phylogenetic analysis. Cladistics  24: 774– 786. Google Scholar CrossRef Search ADS   Goloboff PA , Farris JS , Nixon KC . 2010. T.N.T.: tree analysis using new technology, Version 1.1 . Willi Hennig Society Edition. Available at: www.zmuc.dk/public/phylogeny: Goloboff, Farris, and Nixon Gray JE . 1842. Description of two hitherto unrecorded species of reptiles from New Zealand; presented to the British Museum by Dr. Dieffenbach. In: Gray JE , ed. The zoological miscellany, Part 2 . London: Treuttel, Würtz & Co., 57– 72. Gray JE . 1856. Notice of a new species of nocturnal lizard from Mexico. Annals and Magazine of Natural History, Series 3  2: 270. Grier NM . 1914. A new rhynchocephalian from the Jura of Solnhofen. Annals of Carnegie Museum  9: 86– 91. Hoffstetter R . 1953. Les sauriens anté-cretacées. Bulletin de la Museum Nationale d’Histoire Naturelle  25: 345– 352. Hoffstetter R . 1962. Revue des récentes acquisitions concernant l’histoire et la systématique des squamates. Colloques Internationaux du Centre National de la Recherche Scientifique  45: 243– 278. Hoffstetter R . 1964. Les Sauria du Jurassique supérieur et specialement les Gekkota Bavière et de Mandchourie. Senckenbergiana Biologie  45: 281– 324. Hoffstetter R . 1967. À-propos des genres Ardeosaurus et Eichstäettisaurus (Reptilia, Sauria, Gekkonoidea) du Jurassique Supérieur de Franconie. Comptes rendus du Societie Geologique, France  8: 592– 595. Huene FFV . 1952. Revision der Gattung Pleurosaurus auf Grund neuer und alter Funde. Palaeontographica Abteilung A  101: 167– 200. Huene FFV . 1956. Paläontologie und Phylogenie der Niederen Tetrapoden . Jena: G. Fischer. Hutchinson MN , Skinner A , Lee MSY . 2012. Tikiguana and the antiquity of squamate reptiles (lizards and snakes). Biology Letters  2012: 1– 5. Jones MEH , Anderson CL , Hipsley CA , Müller J , Evans SE , Schoch RR . 2013. Integration of molecules and new fossils supports a Triassic origin for Lepidosauria (lizards, snakes, and tuatara). BMC Evolutionary Biology  13: 1– 21. Google Scholar CrossRef Search ADS PubMed  Kluge AG . 1967. Higher taxonomic categories of gekkonid lizards and their evolution. Bulletin of the American Museum of Natural History  135: 1– 60. Kluge AG . 1987. Cladistic relationships in the Gekkonoidea (Squamata, Sauria). Miscellaneous Publications: Museum of Zoology, University of Michigan  173: 1– 54. Kuhn O . 1958. Ein neuer lacertilier aus dem fränkischen Lithographieschiefer. Neues Jahrbuch für Geologie und Paleontologie Monatschefte  1958: 437– 440. Lee MSY . 1998. Convergent evolution and character correlation in burrowing reptiles: towards a resolution of squamate relationships. Biological Journal of the Linnean Society  65: 369– 453. Google Scholar CrossRef Search ADS   Mateer NJ . 1982. Osteology of the Jurassic lizard Ardeosaurus brevipes (Meyer). Palaeontology  25: 461– 469. Meyer H . 1845. System der fossilen Saurier. Neues Jahrbuch für Mineralogie, Geognosie, Geologie und Petrefakten-Kunde  1845: 278– 285. Meyer H . 1855. Briefliche Mitteilung an Prof. Bonn. Neues Jahrbuch für Geologie und Paleontologie  1855: 326– 327. Montero R , Gans C . 1999. The head skeleton of Amphisbaena alba Linneaus. Annals of Carnegie Museum  68: 15– 79. Müller J , Hipsley CA , Head JJ , Kardjilov N , Hilger A , Wuttke M , Reisz RR . 2011. Eocene lizard from Germany reveals amphisbaenian origins. Nature  473: 364– 367. Google Scholar CrossRef Search ADS PubMed  Nesbitt SJ , Turner AH , Erickson GM , Norell MA . 2006. Prey choice and cannibalistic behaviour in the theropod Coelophysis . Biology Letters  2006: 611– 614. Google Scholar CrossRef Search ADS   Newton A , Newton E . 1859. Observations on the birds of St. Croix, West Indies, made, between February 20th and August 6th, 1857 and between March 4th and September 28th, 1858. Ibis  1: 59– 69. Google Scholar CrossRef Search ADS   Nopcsa FB . 1908. Zur kenntnis der fossilen Eidechsen. Beiträge zur Paläontologie und Geologie Österreich-Ungarns und des Orients  21: 33– 62. Novas FE , Ezcurra MD , Agnolin FL , Pol D , Ortíz R . 2012. New Patagonian Cretaceous theropod sheds light about the early radiation of Coelurosauria. Revista del Museo Argentino de Ciencias Naturales, New Series  14: 57– 81. Google Scholar CrossRef Search ADS   Oelrich TM . 1956. The anatomy of the head of Ctenosaura pectinata (Iguanidae). Miscellaneous Publications: Museum of Zoology, University of Michigan  94: 1– 122. Oppel M . 1811. Die Ordnungen, Familien, und Gattungen der Reptilien als Prodrom einer Naturgeschichte derselben . München: Joseph Lindauer. Google Scholar CrossRef Search ADS   Ostrom J . 1978. The osteology of Compsognathus longipes Wagner. Zitteliana  4: 78– 118. Page RDM . 2001. NDE: Nexus Data Editor for Windows . Glasgow: Page, R. D. M. Pianka ER , Vitt LJ . 2003. Lizards: windows to the evolution of diversity . Berkeley: University of California Press. Polcyn MJ , Rogers JV II , Kobayashi Y , Jacobs LL . 2002. Computed tomography of an Anolis lizard in Dominican amber: systematic, taphonomic, biogeographic, and evolutionary implications. Palaeontologia Electronica  5: 1– 13. Pyron RA , Burbrink FT , Wiens JJ . 2013. A phylogeny and revised classification of Squamata, including 4161 species of lizards and snakes. BMC Evolutionary Biology  13: 1– 53. Google Scholar CrossRef Search ADS PubMed  Rauhut OWM , Foth C , Tischlinger H , Norell MA . 2012. Exceptionally preserved juvenile megalosauroid theropod dinosaur with filamentous integument from the Late Jurassic of Germany. Proceedings of the National Academy of Sciences  109: 11746– 11751. Google Scholar CrossRef Search ADS   Rauhut OWM , Milner AC , Moore-Fay S . 2010. Cranial osteology and phylogenetic position of the theropod dinosaur Proceratosaurus bradleyi (Woodward, 1910) from the Middle Jurassic of England. Zoological Journal of the Linnean Society  158: 155– 195. Google Scholar CrossRef Search ADS   Reeder TW , Townsend TM , Mulcahy DG , Noonan BP , Wood PL Jr , Sites JW Jr , Wiens JJ . 2015. Integrated analyses resolve conflicts over squamate reptile phylogeny and reveal unexpected placements of fossil taxa. PLoS ONE  2015: 1– 22. Reisdorf AG , Wuttke M . 2012. Re-evaluating Moodie’s opisthotonic-posture hypothesis in fossil vertebrates part I: reptiles—the taphonomy of the bipedal dinosaurs Compsognathus longipes and Juravenator starki from the Solnhofen Archipelago (Jurassic, Germany). Palaeobiodiversity and Palaeoenvironments  91: 119– 168. Google Scholar CrossRef Search ADS   Reynoso V-H . 1996. A Middle Jurassic Sphenodon-like sphenodontian (Diapsida: Lepidosauria) from Huizachal Canyon, Tamaulipas, Mexico. Journal of Vertebrate Paleontology  16: 210– 221. Google Scholar CrossRef Search ADS   Reynoso V-H . 1997. A “beaded” sphenodontian (Diapsida: Lepidosauria) from the Early Cretaceous of Central Mexico. Journal of Vertebrate Paleontology  17: 52– 59. Google Scholar CrossRef Search ADS   Reynoso V-H . 1998. Huehuecuetzpalli mixtecus gen. et sp. nov.: a basal squamate (Reptilia) from the Early Cretaceous of Tepexi de Rodríguez, Central México. Philosophical Transactions of the Royal Society of London, Series B: Biological Sciences  353: 477– 500. Google Scholar CrossRef Search ADS   Rieppel O . 1980. The postcranial skeleton of Lanthanotus borneensis (Reptilia, Lacertilia). Amphibia-Reptilia  1: 95– 112. Google Scholar CrossRef Search ADS   Romer AS . 1949. The vertebrate body . Philadelphia: W. B. Saunders Company. Romer AS . 1956. Osteology of the reptiles . Chicago: The University of Chicago Press. Schmitz L , Motani R . 2011. Nocturnality in dinosaurs inferred from scleral ring and orbit morphology. Science  332: 705– 708. Google Scholar CrossRef Search ADS PubMed  Schwenk K . 1988. Comparative morphology of the lepidosaur tongue and its relevance to squamate phylogeny. In: Estes R , Pregill G , eds. Phylogenetic relationships of the lizard families . Stanford: Stanford University Press, 569– 597. Sereno PC . 1993. The pectoral girdle and forelimb of the basal theropod Herrerasaurus ischigualestensis . Journal of Vertebrate Paleontology  13: 425– 450. Google Scholar CrossRef Search ADS   Sherbrooke WC . 1990. Predatory behavior of captive Greater Roadrunners feeding on horned lizards. Wilson Bulletin  102: 171– 174. Smith KT . 2009. Eocene lizards of the clade Geiseltaliellus from Messel and Geiseltal, Germany, and the early radiation of Iguanidae (Reptilia: Squamata). Bulletin of the Peabody Museum of Natural History  50: 219– 306. Google Scholar CrossRef Search ADS   Snively E , Cotton JR , Ridgely R , Witmer LM . 2013. Multibody dynamics model of head and neck function in Allosaurus (Dinosauria, Theropoda). Palaeontologia Electronica  16: 1– 29. Snively E , Russell AP . 2007. Functional morphology of neck musculature in the Tyrannosauridae (Dinosauria, Theropoda) as determined via a hierarchical inferential approach. Zoological Journal of the Linnean Society  151: 759– 808. Google Scholar CrossRef Search ADS   Sulimski A . 1984. A new Cretaceous scincomorph lizard from Mongolia. Palaeontologia Polonica  46: 143– 155. Swofford DL . 2001. PAUP*: phylogenetic analysis using parsimony, Version 4.0b10 . Washington: Smithsonian Institution. Townsend TM , Larson A , Louis E , Macey JR . 2004. Molecular phylogenetics of Squamata: the position of snakes, amphisbaenians, and dibamids, and the root of the squamate tree. Systematic Biology  53: 735– 757. Google Scholar CrossRef Search ADS PubMed  Uetz P . 2016. The reptile database . Hamberg: Reptarium Association. Varricchio DJ . 2001. Gut contents from a Cretaceous tyrannosaurid: implications for theropod dinosaur digestive tracts. Journal of Paleontology  75: 401– 406. Google Scholar CrossRef Search ADS   Vidal N , Hedges SB . 2004. Molecular evidence for a terrestrial origin of snakes. Proceedings of the Royal Society of London, Series B: Biological Sciences  271( suppl.): S226– S229. Google Scholar CrossRef Search ADS   Vidal N , Hedges SB . 2005. The phylogeny of squamate reptiles (lizards, snakes, and amphisbaenians) inferred from nine nuclear protein-coding genes. Comptes Rendus Biologies  328: 1000– 1008. Google Scholar CrossRef Search ADS PubMed  Vidal N , Hedges SB . 2009. The molecular evolutionary tree of lizards, snakes, and amphisbaenians. Comptes Rendus Biologies  332: 129– 139. Google Scholar CrossRef Search ADS PubMed  Viohl G . 2000. Die Solnhofener Plattenkalke, oberer Jura. In: Meischner D , Pinna G , eds. Europäische Fossillagerstätten . Berlin: Springer, 143– 150. Google Scholar CrossRef Search ADS   Wagner A . 1852. Neu aufgefundene Saurier, Überreste aus den lithographischen Schiefern und dem obern Jurakalke . München: Abhandlungen der Bayerischen Akademie der Wissenschaften, Mathematisch-Physikalische Klasse. Wagner JA . 1861. Neue Beiträge zur Kenntniss der urweltlichen Fauna des lithographischen Schiefers. Abhandlungen der königlichen bayerische Akademie der Wissenschaft, II Classe  9: 65– 124. Webster D , Webster M . 1974. Comparative vertebrate morphology . New York: Academic Press. Wiens JJ , Hutter CR , Mulcahy DG , Noonan BP , Townsend T , Sites JW Jr , Reeder TW . 2012a. Resolving the phylogeny of lizards and snakes (Squamata) with extensive sampling of genes and species. Biology Letters  8: 1043– 1046. Google Scholar CrossRef Search ADS   Wiens JJ , Hutter CR , Mulcahy DG , Noonan BP , Townsend TM , Sites JW Jr , Reeder DM . 2012b. Resolving the phylogeny of lizards and snakes (Squamata) with extensive sampling of genes and species. Biology Letters  8: 1043– 1046. Google Scholar CrossRef Search ADS   Wiens JJ , Kuczynski CA , Townsend T , Reeder TW , Mulcahy DG , Sites JW Jr . 2010. Combining phylogenomics and fossils in higher-level squamate reptile phylogeny: molecular data change the placement of fossil taxa. Systematic Biology  59: 675– 688. Google Scholar CrossRef Search ADS   Yosef R . 1996. Lanius ludovicianus – Loggerhead Shrike. In: Rodewald P G , ed. The Birds of North America, No. 231 . Ithaca: Cornell Lab of Ornithology . Available at: https://birdsna.org/Species-Account/bna/species/logshr. Google Scholar CrossRef Search ADS   Yosef R . 2004. Response to: “How the horned lizard got its horns”. Science  305: 1909. Google Scholar PubMed  Young KV , Brodie ED Jr , Brodier ED III . 2004. How the horned lizard got its horns. Science  304: 65. Google Scholar CrossRef Search ADS PubMed  © 2017 The Linnean Society of London, Zoological Journal of the Linnean Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Zoological Journal of the Linnean Society Oxford University Press

A new lizard (Squamata) was the last meal of Compsognathus (Theropoda: Dinosauria) and is a holotype in a holotype

Loading next page...
 
/lp/ou_press/a-new-lizard-squamata-was-the-last-meal-of-compsognathus-theropoda-KrRtBfYRma
Publisher
Oxford University Press
Copyright
© 2017 The Linnean Society of London, Zoological Journal of the Linnean Society
ISSN
0024-4082
eISSN
1096-3642
D.O.I.
10.1093/zoolinnean/zlx055
Publisher site
See Article on Publisher Site

Abstract

Abstract A lizard preserved in the abdominal cavity of the Compsognathus longipes holotype traditionally has been referred to Bavarisaurus macrodactylus Hoffstetter. New observations on that specimen reveal that it differs from B. macrodactylus in the shape of the frontoparietal suture, the presence of paired parietals and the orientation of the parietal supratemporal processes. Phylogenetic analysis of 220 saurians and 839 morphological characters demonstrate the paraphyly of Ardeosaurus as traditionally used and the distinctiveness of three major squamate radiations in the Solnhofen Formation, and suggests rhynchocephalian affinities for the early lepidosaur Bharatagama rebbanensis. The gut content specimen is a diagnostic new species – a squamate holotype in a dinosaur holotype. It is an ardeosaurid based on the presence of a posteroventral process of the jugal, absence of a lateral parietal flange at the frontoparietal suture and a posteriorly broadened mandibular retroarticular process. Ardeosaurus, Bavarisaurus, Eichstaettisaurus, Palaeolacerta, Squamata INTRODUCTION Squamata is known from nearly 10 000 extant named species (Uetz, 2016) inhabiting essentially every non-polar environment (Pianka & Vitt, 2003). Squamata diverged from Rhynchocephalia by the Middle Triassic (Jones et al., 2013), but has a known unequivocal fossil record extending only to the Middle Jurassic (Evans, 1994a; Caldwell et al., 2015). Although earlier fossils have been suggested as representing squamates (Carroll, 1982; Evans, Prasad & Manhas, 2002; Datta & Ray, 2006), these suggestions have either been refuted (Evans, 2001; Hutchinson, Skinner & Lee, 2012) or are untested (Bharatagama rebbanensis Evans et al., 2002; see below). Currently, the earliest known, definitive, squamate is from the Middle Jurassic of Britain (Evans, 1998; Caldwell et al., 2015). The earliest known complete squamates are from the Late Jurassic of China (Evans & Wang, 2007, 2009; Conrad et al., 2013) and Europe (Meyer, 1855; Nopcsa, 1908; Grier, 1914; Huene, 1956; Kuhn, 1958; Cocude-Michel, 1961; Hoffstetter, 1964, 1967; Ostrom, 1978; Mateer, 1982; Estes, 1983; Evans, 1994a, b; Evans & Barbadillo, 1998; Evans, 2003; Conrad, 2008). Germany hosts the richest record of complete Jurassic squamates so far known, and these occur within the Solnhofen Formation from the Late Jurassic (Fig. 1). Five squamate species have been named from these limestones. Ardeosaurus brevipes Meyer, 1855, Ardeosaurus digitatellus Grier, 1914, Bavarisaurus macrodactylus Hoffstetter, 1953, Eichstaettisaurus schroederi Kuhn, 1958 and Palaeolacerta bavarica Cocude-Michel, 1961 are each known from complete or partial remains of articulated specimens from Solnhofen (see Estes, 1983; Evans, 1994b). Eichstaettisaurus schroederi and P. bavarica are each represented only by one specimen (Cocude-Michel, 1961; Cocude-Michel, 1963; Estes, 1983; Evans & Barbadillo, 1998). Figure 1. View largeDownload slide World palaeomap with modern pullout of Deutschland, highlighting the area from which SNSB-BSPG AS I 563b is recovered. The palaeomap is a reconstruction of the Late Jurassic and is drawn after Blakey (2011). Figure 1. View largeDownload slide World palaeomap with modern pullout of Deutschland, highlighting the area from which SNSB-BSPG AS I 563b is recovered. The palaeomap is a reconstruction of the Late Jurassic and is drawn after Blakey (2011). Ardeosaurus is a problematic taxon with two named species. Ardeosaurus brevipes is known by a cast of the holotype (NHM PR 38006) and a referred specimen (PMU.R58; Meyer, 1855; Camp, 1923; Estes, 1983). Ardeosaurus digitatellus is known only from the holotype (CM 4026; Cocude-Michel, 1963; Estes, 1983), but a second specimen (SNSB-BSPG 1923 I 501) has been referred to A. cf. digitatellus by Estes (1983). The relationships of these four specimens have not been cladistically tested. Bavarisaurus macrodactylus is known from the holotype, but a second specimen has been referred to that species (Huene, 1956; Evans, 1994b) or to Ba. cf. macrodactylus (Ostrom, 1978; Estes, 1983). The referred specimen of Bavarisaurus is preserved inside the abdomen of the holotype of the theropod dinosaur Compsognathus longipes Wagner, 1861 (Figs 2, 3; see Ostrom, 1978; Estes, 1983; Evans, 1994b). This specimen is of uncertain provenance and may come from the Solnhofen Formation, but might be from the Painten, Torleite or Rögling Formations with a likely age of either latest Kimmeridgian or earliest Tithonian (Viohl, 2000; Rauhut et al., 2012; Reisdorf & Wuttke, 2012). Figure 2. View largeDownload slide Body of Compsognathus longipes (holotype; SNSB-BSPG AS I 563) with the skeleton of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b) inside it. See also Figure 3. Figure 2. View largeDownload slide Body of Compsognathus longipes (holotype; SNSB-BSPG AS I 563) with the skeleton of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b) inside it. See also Figure 3. When the morphology of Ba. macrodactylus was reviewed in 1994, the gut content specimen was considered to be representative of Ba. macrodactylus (Evans, 1994b). Even so, in observing the holotype (Fig. 4) and the referred specimen (Figs 2, 3), several morphological differences became apparent. As a result, I revisited the morphology and/or the published data regarding the morphology of all of the available Solnhofen Jurassic squamates and have undertaken a comparative study of these specimens. The current paper has four major goals. Figure 3. View largeDownload slide Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b) as it lies in the body of Compsognathus longipes (SNSB-BSPG AS I 563). A, restored body silhouette of the dinosaur constructed and restored body silhouette showing its hypothesized orientation inside the gut based on the preserved skeletal associations. B, drawing of the lizard skeleton and the associated bones of the dinosaur. Note that the right femur is not currently preserved, but is represented by a cast of the specimen made before further preparation led to the femur’s removal. Figure 3. View largeDownload slide Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b) as it lies in the body of Compsognathus longipes (SNSB-BSPG AS I 563). A, restored body silhouette of the dinosaur constructed and restored body silhouette showing its hypothesized orientation inside the gut based on the preserved skeletal associations. B, drawing of the lizard skeleton and the associated bones of the dinosaur. Note that the right femur is not currently preserved, but is represented by a cast of the specimen made before further preparation led to the femur’s removal. Figure 4. View largeDownload slide Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501). Figure 4. View largeDownload slide Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501). Identify the differences between Ba. macrodactylus and the squamate preserved within the abdomen of the C. longipes holotype (SNSB-BSPG AS I 563b). Establish the identity of SNSB-BSPG AS I 563b. Present a phylogenetic analysis of Jurassic lizards with a focus on the parts of the tree pertaining to the Solnhofen specimens, including all of the putative Ardeosaurus species. MATERIAL AND METHODS Institutional abbreviations AMNH, American Museum of Natural History, New York City, NY; SNSB-BSPG, Staatliche naturwissenschaftliche Sammlungen Bayerns-Bayerische Staatssamlung für Paläontologie und Geologie, Munich, Germany; CM, Carnegie Museum of Natural History, Pittsburgh, PA; FMNH, Field Museum, Chicago, IL; GM, Geiseltal Museum, Martin-Luther University, Halle/Saale, Germany; IGM, Institute of Geology, Mongolian Academy of Sciences – American Museum of Natural History Expeditions, field numbers; IVPP, Institute of Vertebrate Paleontology and Paleoanthropology, Beijing, People’s Republic of China; KNM, Kenyan National Museums, Nairobi, Kenya; LACM, Los Angeles County Museum, Los Angeles, CA; MB, Museum für Natkurkunde, Berlin, Germany; MCZ, Museum of Comparative Zoology, Cambridge, MA; MINS, Missouri Institute of Natural Science, Springfield, MO; MNHN, Muséum national d’Histoire naturelle, Paris, France; MPN, Museo di Paleontologia, Napoli, Italy; NHM, The Natural History Museum, London, UK; NMNH, National Museum of Natural History, Washington, D.C.; PMU, Paleontological Museum of Uppsala University, Uppsala, Sweden; SDSMT, South Dakota School of Mines and Technology, Rapid City, SD; UCL, University College London, London, UK; UCMP, University of California Museum of Paleontology, Berkeley, CA; UF, Florida State Museum (University of Florida), Gainesville, FL; USNM, National Museum of Natural History, Smithsonian Institution, Washington, D.C.; YPM, Yale Peabody Museum, New Haven, CT; ZPAL, Zakład Paleobiologii, Polska Akademia Nauk (Paleobiological Institute, Polish Academy of Sciences), Warsaw, Poland. Abbreviations a, angular; as, astragalus; asca, astragalocalcaneum; bc, basioccipital c, coronoid; CA, caudal vertebra; colf, columellar fossa; d, dentary; Dr, dorsal rib; e, epipterygoid; ec, ectopterygoid; ecf, ectepicondylar foramen; ecp, ectepicondyle; ent, entepicondyle; f, frontal; fe, femur; fi, fibula; h, humerus; inc, intercentrum; is, ischium; j, jugal; L, left; l, lachrymal; n, nasal; m, maxilla; ma, manus; mn, mandible; mt. metatarsal; nea, neural arch; of, obturator foramen; oo, opisthotic; or, orbit; p, parietal; pa, palatine; pbs, parabasisphenoid; pe, pes; pf, postfrontal; pif, pineal foramen; plr, palatine ramus; pm, premaxilla; po, postorbital; pra, prearticular; prf, prefrontal; pt, pterygoid; pu, pubis; q, quadrate; qpr, quadrate process; R, right; r, radius; rac, radial condyle; rap, retroarticular process; S1, first sacral vertebra; S2, second sacral vertebra; Sr, sacral rib; sa, surangular; sk, skull; so, supraoccipital; sq, squamosal; st, supratemporal; stf, supratemporal fenestra; ti, tibia; tvp, transverse process; ul, ulna; ulc, ulnar condyle. Specimen reconstructions Skull and whole-body reconstructions were made from direct observations, new photographs and/or published photos and interpretive drawings. For those reconstructed taxa that were directly observed (BMNH PR 38006, SNSB-BSPG 1923 I 501, CM 4026, SNSB-BSPG 1873 III 501 and SNSB-BSPG 1937 Ia, b), digital photographs and notes were taken, and physical and digital line drawings were made with the specimen present. Where useful, I used published illustrations and photographs to help interpret these specimens. When specimens were not directly observed (i.e. PMU.R58 and MPN 539), published photos and interpretive drawings were used and the reconstructions are less detailed and more tentative. Digital line drawings were made in Adobe Photoshop CS6 (Adobe Systems, 2010). The midline was identified for each specimen and the alignments were modified accordingly for reconstruction. When one side of the specimen was more complete or better preserved, it was copied and reversed as a separate Photoshop layer to help reconstruct both sides of the skull. Individual bones were treated similarly and aligned, often as semi-transparent layers, to help reconstruct the specimen. Each illustration is based on a single specimen. SYSTEMATICS Squamata Oppel, 1811 Eichstaettisauridae Kuhn, 1958 Schoenesmahl dyspepsia gen. et sp. nov. Type species:Schoenesmahl dyspepsia sp. nov. Diagnosis: As for the type and only known species. Etymology: Schoenesmahl (Deutsch: schöne Mahl ‘beautiful meal’ where ‘beautiful’ is used here as a synonym of elegance), referring to the fact that it was the last meal of the holotype of C. longipes Wagner, 1861 whose generic name means ‘elegant jaw’ or ‘comely jaw’ (Greek). Dyspepsia (Greek: ‘difficult digestion’), referring to the undigested nature of this last meal of C. longipes. Holotype: SNSB-BSPG AS I 563b, an incomplete skeleton lacking nasals, vomers, palatines, postorbitals, quadrates, anterior presacral vertebrae, pectoral girdles, most of the radii and ulnae, manus, ilium, ischium, tarsals and the distal pedal phalanges from digits I, II, IV and V. Type locality and horizon: Kelheim (Land Bayern; Fig. 1), Solnhofen Formation (Tithonian) of Germany (Wagner, 1861). Known distribution: Known only from the type locality and horizon. Diagnosis: Schoenesmahl dyspepsia differs from Ba. macrodactylus in the possession of a U-shaped (rather than a W-shaped) frontoparietal suture, paired parietals and anteroposteriorly (rather than mediolaterally oriented) parietal supratemporal processes. Schoenesmahl dyspepsia differs from E. schroederi in possessing a weakly inclined nasal process and an anteroposteriorly elongate prefrontal, frontals with parallel-sided interorbital margins (rather than an hourglass-shaped frontal). Schoenesmahl dyspepsia differs from Eichstaettisaurus gouldi Evans, Raia & Barbera, 2004 in possessing posteriorly broadened retroarticular process of the articular. Schoenesmahl dyspepsia differs from E. schroederi and E. gouldi in lacking well-developed subolfactory processes of the frontal. DESCRIPTION Skull form Cranial form Although the skull is mostly disarticulated and slightly spread apart (Fig. 5), many details may be reconstructed (Fig. 6). The snout is elongate relative to the orbital and temporal areas (sensuMontero & Gans, 1999). This differs from the recent rhynchocephalian Sphenodon punctatus Gray, 1842, which possesses a relatively short snout compared to the post-snout part of the skull; however, basal rhynchocephalians have relatively elongate snouts (Evans, 1980; Fraser, 1982), suggesting that this may be the more plesiomorphic lepidosaur condition. Figure 5. View largeDownload slide Skull elements of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b). A, photograph. B, interpretive drawing. Dark grey areas in (B) indicate areas preserved by faint impressions and/or outlines. Lighter grey areas indicate areas where the bone surfaces have been lost, but some bone is preserved. White areas indicate bones or clear bone impressions. Figure 5. View largeDownload slide Skull elements of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b). A, photograph. B, interpretive drawing. Dark grey areas in (B) indicate areas preserved by faint impressions and/or outlines. Lighter grey areas indicate areas where the bone surfaces have been lost, but some bone is preserved. White areas indicate bones or clear bone impressions. Figure 6. View largeDownload slide Reconstruction of the skull of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b). A, left lateral view. B, dorsal view. Semi-opaque shadows indicate elements or bone surfaces not preserved. White areas indicate bones or clear bone impressions. C, only the preserved bones in left lateral view and D, dorsal view. Figure 6. View largeDownload slide Reconstruction of the skull of Schoenesmahl dyspepsia gen. et sp. nov. (holotype; SNSB-BSPG AS I 563b). A, left lateral view. B, dorsal view. Semi-opaque shadows indicate elements or bone surfaces not preserved. White areas indicate bones or clear bone impressions. C, only the preserved bones in left lateral view and D, dorsal view. The exact shape of the external naris remains uncertain due to the disarticulated nature of the skull bones and lack of preserved nasal bones. The external naris appears to have been relatively elongate as compared to the many lepidosaurs (e.g. Sp. punctatus, Abronia mixteca Bogert & Porter, 1967, Xenosaurus grandis Gray, 1856 and many others). This is suggested by the gently sloping anterior margin of the maxillary nasal process and the elongate nasal process of the premaxilla (Figs 5–7). The orbits are not particularly large and their greatest reconstructed diameter is slightly less than the probable length of the antorbital snout (Fig. 6). Based on the length of the supratemporal processes of the parietal and the parietal itself (Fig. 5), the supratemporal fenestrae may have been significantly smaller than the orbits (Fig. 6). The elements of the palate and their contacts are not sufficiently known to confidently reconstruct the size and shape of the suborbital fenestrae or the interpterygoid vacuity. Figure 7. View largeDownload slide Detail of the area of the skull of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) showing the impression of the left maxilla and some of the surrounding bones. Figure 7. View largeDownload slide Detail of the area of the skull of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) showing the impression of the left maxilla and some of the surrounding bones. Mandibular form The robust lower jaw is preserved as a combination of bones and impressions of bone (Figs 3, 5). The long axis of the mandible is straight (Fig. 6). The dentary constitutes more than one-half the length of the mandible. No data are available for the state of Meckel’s canal. Dermal skull roof Premaxilla Parts of each premaxilla are preserved (Fig. 5). The right is preserved as an impression, but the left is partially preserved in lateral view. The preserved parts of the left premaxilla include a robust body, two teeth, and part of the elongate, posterodorsally oriented, nasal process. The posterior part of the nasal process is preserved only as an impression. The paired premaxillae form the anterior margin of the external naris. The premaxillary body is narrow. The right premaxillary impression includes three tooth positions, but the original number of tooth positions remains uncertain. The preserved parts indicate that three or four were probably present on each premaxilla. The breadth of the complete left dental margin appears to be too small for any more than four tooth positions. The nasal process tapers distally. Its posterior surface has a distinct and deep ventral ridge. The posterior margin of this is vertical to a level just dorsal to the level of the main premaxillary body. Distal to that level, the nasal process consistently tapers toward its tip. The body of the premaxilla is joined with the posterior nasal process ridge by a dorsomedially oriented flange (Fig. 5). Anterior to this flange is a minor depression that would have lain in the anteroventral edge of the bony external naris. Based on the shape of the anterolateral edge of the body of the premaxilla, the premaxilla narrowly overlaid the premaxillary process of the maxilla anterolaterally. The presence and nature of the anterior ethmoidal foramina is unknown. Bavarisaurus macrodactylus possesses paired premaxilla bearing very narrow nasal processes (Fig. 8). The premaxilla is preserved as a combination of bone and impression. The body of the premaxilla is, apparently, similar in width relative to the snout tip as it is in S. dyspepsia. Two tooth positions are preserved as infilled impressions from the eroded premaxillary body, but there probably three or four tooth positions per premaxilla. Figure 8. View largeDownload slide The skull of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501). The skull is preserved in ventral view, meaning that the bone impressions preserved on the matrix represent dorsal views. A, photograph of the skull and partial mandible as preserved. B, reconstructed skull. White areas indicate parts preserved in bone. Semi-opaque grey areas overlapping white (e.g. premaxilla) indicate areas represented by bone, but on surfaces that are not visible in the reconstructed dorsal view. Black lines indicate visible suture patterns and/or bone margins. Grey areas without outline indicate unknown elements tentatively restored here. Figure 8. View largeDownload slide The skull of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501). The skull is preserved in ventral view, meaning that the bone impressions preserved on the matrix represent dorsal views. A, photograph of the skull and partial mandible as preserved. B, reconstructed skull. White areas indicate parts preserved in bone. Semi-opaque grey areas overlapping white (e.g. premaxilla) indicate areas represented by bone, but on surfaces that are not visible in the reconstructed dorsal view. Black lines indicate visible suture patterns and/or bone margins. Grey areas without outline indicate unknown elements tentatively restored here. Eichstaettisaurus schroederi possesses a short and broad premaxillary nasal processes on its paired premaxillae (Fig. 9). A proximomedial emargination that appears to be bilaterally preserved in SNSB-BSPG 1937 I 1 may be an emargination associated with an ethmoidal nerve. Gauthier et al. (2012) did not code this structure as if it were an ethmoidal branch and the nature of this condition remains uncertain. Figure 9. View largeDownload slide The skull of Eichstaettisaurus schroederi (holotype; SNSB-BSPG 1937 I 1a) in dorsal view. A, photograph of skull. B, reconstructed skull. Semi-opaque grey layer indicates missing parts restored here. Figure 9. View largeDownload slide The skull of Eichstaettisaurus schroederi (holotype; SNSB-BSPG 1937 I 1a) in dorsal view. A, photograph of skull. B, reconstructed skull. Semi-opaque grey layer indicates missing parts restored here. The premaxillae are paired in all Ardeosaurus specimens for which they are preserved (CM 4026, SNSB-BSPG 1923 I 501 and NHM PR 38006). As a unit, the premaxillae have a broad body and a moderately broad nasal process as indicated by the preserved part of the broken base and the distal part overlying the nasals. Premaxillary teeth are visible in none of the Ardeosaurus specimens observed here, and they were hidden in the specimen described by Mateer (1982). Palaeolacerta bavarica appears to possess an unpaired premaxilla with an elongate nasal process (Hoffstetter, 1964; Estes, 1983). Because of damage to the specimen and the nature of the published figures, the breadth of the premaxillary body is uncertain. Maxilla Only the left maxilla is known and it is preserved only as an impression on the rock (Figs 5, 7). This impression retains its contacts with the left lacrimal and jugal and has been moved slightly out of position with the ectopterygoid contact. The lithographic limestone preserves exquisite details of the lateral (external) surfaces of the maxilla, including five of its teeth. The overlying postfrontal only slightly obscures details of the posterodorsal margin of the maxillary nasal process. The maxilla is subtriangular with a tall nasal process, an elongate premaxillary ramus and somewhat shorter suborbital ramus (Figs 5–7). The gently and continuously tapering premaxillary ramus is attenuated at its anterior end and was overlapped by the body of the premaxilla at the level of the dental margin. Its dorsal margin describes a gentle posterodorsal curve without a clear distinction between the subnarial surface and the anterior margin of the nasal process. There is, however, a small and weakly developed spur at a level that probably indicates the anterior limit of the articulation with the nasal bone (Fig. 7). The suborbital ramus apparently did not reach the level of the middle of the orbit. No dermal sculpturing is apparent on the lateral surface of the maxilla. Seven superior labial foramina are present near the dental margin of the maxilla. The posteriormost of these is posteroventrally oriented, whereas the other six are more anteromedially oriented. The most anterior of these foramina is located dorsal to the level of the other six. Additionally, two other small foramina are present more dorsally, on the nasal process of the maxilla. The more anterior of the two is located just dorsal to the third superior labial foramen; the more posterior of them lies slightly more dorsal than the anterior one and at a level between the third and fourth superior labial foramina. The maxillary suborbital ramus is more obtusely attenuated than the premaxillary ramus. As preserved, a small sliver of the ectopterygoid is visible ventral to the level of the jugal. However, it is unclear whether the ectopterygoid abuts the posterior tip of the maxillary suborbital ramus, or clasps it. Impressions of five tooth positions are associated with the left maxilla. The most anterior preserved tooth impression is present ventral to and slightly anterior to the anteriormost maxillary ethmoidal foramen. The second tooth impression is present immediately posterior to the first. The other three tooth impressions are located below the level of the last four maxillary ethmoidal foramina. Few details are available regarding the maxilla of Ba. macrodactylus. Eichstaettisaurus schroederi has well-preserved maxillae, but details of the dental margin are hidden as preserved. Eichstaettisaurus schroederi differs from S. dyspepsia in that the nasal process is steeply inclined and the apex of the nasal process lies anterior to the midpoint of the maxilla (Fig. 9). The posterior margin of the nasal process also has a stronger overlap of the anterior part of the lacrimal in E. schroederi. Few details are available for the maxilla in Ardeosaurus. The holotype of A. brevipes offers the clearest view of the morphology. The nasal process is dorsally tall and possessed what appears to be a steep posterior margin to the external naris. Between eight and ten superior labial foramina are probably present, but the exact number is uncertain. The maxilla is mostly ventral to (rather than lateral to) the suborbital process of the jugal. The relationships between the maxilla and the prefrontal and frontal remain unclear. Prefrontal Although preserved with only its ventral aspect visible (Figs 5, 7), the prefrontal is well preserved. It is a robust bone with a mediolaterally broad palatine process. The nasal cavity area is concave and generally similar to that of other non-snake squamates (e.g. Oelrich, 1956; Conrad, 2004). The orbital process of the prefrontal is elongate and extends to approximately the level of the middle of the orbit, fitting into a shallow groove on the ventrolateral aspect of the frontal. The orbital process of the prefrontal is similarly long in Ba. macrodactylus (Fig. 8), but is much shorter in E. schroederi (Fig. 9). As illustrated by Evans et al. (2004), the Catalonian Eichstaettisaurus sp. has prefrontals bearing orbital processes that are longer than those of E. schroederi, but that do not reach the midpoint of the orbit. All of the observed Ardeosaurus specimens and the specimen described by Mateer (1982) are similar in that the orbital processes of their prefrontals extending nearly to the midpoint of the orbit. The specimens SNSB-BSPG 1923 I 501 and NHM PR 38006 are similar in having subtriangular orbital processes and, apparently, ovate anterior processes. By contrast, CM 4026 has a laterally concave margin such that the orbital processes become very narrow posteriorly. The orbital process of PMU.R58 is subtriangular (see Mateer, 1982), but seems less continuously anterolaterally expanded than in SNSB-BSPG 1923 I 501 and NHM PR 38006. Although the prefrontals are incompletely preserved in P. bavarica, the orbital processes seem to be mostly preserved and complete (Hoffstetter, 1964). The orbital process is more similar to that known for PMU.R58 than other known Solnhofen lizards. Lacrimal The left lacrimal is preserved, and only as a clear impression of a bone articulated with the jugal and maxilla (Figs 5–7). The jugal contact with the lacrimal is gently bowed posteriorly. The maxillary contact with the lacrimal demonstrates that the maxilla did not reach the orbital margin. A small lateral ridge is present at the orbital rim and is continuous with the same on the anterior part of the jugal. Although part of the right lacrimal is preserved in Ba. macrodactylus (Fig. 8), no details are apparent. The bilaterally preserved lacrimals of E. schroederi are small and simple. They are only slightly medially emarginated for contribution to the lacrimal foramen. Although the maxilla probably did not reach the orbital margin in E. schroederi, it did more extensively overlie the lacrimal. The E. schroederi lacrimal apparently lacks the small orbital ridge (Fig. 9) seen in the lacrimal of S. dyspepsia. It is impossible to determine definitive presence or absence of a lacrimal in any of the known Ardeosaurus based on specimens or Mateer (1982). Jugal An incomplete left jugal is preserved in articulation with the lacrimal and maxilla in S. dyspepsia (Figs 5, 7). Most of the preserved part is visible in medial view, but the anterior bit is preserved as an impression of the lateral surface. The part preserved as bone is eroded and lacks the natural bone surfaces. Its posterior part is narrowly obscured by the overlying left pterygoid and ectopterygoid. Because the postorbital process is preserved in medial view, there is no way to tell whether it possesses dermal sculpturing. The posterior and ventral margins of the jugal describe an obtuse angle. The suborbital part is curved somewhat anterodorsally and its dorsal and ventral margins are continuous with those of the lacrimal. Based on the impression of the lateral aspect of the jugal, the suborbital process of the jugal lies mostly dorsal to the maxilla (Figs 5, 6), as opposed to medial to it as in many squamates (e.g. most iguanians, most scincids and most gekkotans). At its anterior end, the jugal shows a continuation of the orbital ridge from the lacrimal, but this quickly disappears posteriorly. A small posteroventral process is preserved. It is directed posteriorly. The postorbital bar is similarly broad to the suborbital process. Eichstaettisaurus schroederi and A. brevipes share with S. dyspepsia the condition the jugal lying mostly dorsal to the maxilla. This differs from the condition present in the putative basal gekkonomorph Norellius nyctisaurops Conrad & Daza, 2015, and all of the modern gekkotans observed here. Presence or absence of a posteroventral process of the jugal cannot be confirmed in E. schroederi, but one is apparently absent in E. gouldi (Evans et al., 2004). This process is present in Ardeosaurus specimens for which it may be scored (Mateer, 1982; NHM PR 38006, SNSB-BSPG 1923 I 501). Only a small piece of the postorbital ramus of the jugal appears to be preserved in P. bavarica (Hoffstetter, 1964), but this may be a fragment of the mandible. Postfrontal and postorbital The left postfrontal is preserved slightly out of normal articulation with the frontal and parietal. It is triradiate and each side is nearly equal in length (Figs 5, 6). The frontal process tapers anteriorly and forms a slightly obtuse angle with the parietal process. The postfrontal spans the frontoparietal contact. The parietal process ends more bluntly than the frontal process. The postorbital process is attenuated. Presumably, the space between the postorbital and parietal processes originally bore a facet receiving the postorbital bone, but the state of preservation does not allow that to be confirmed. No postorbital is preserved. Small bits of the postorbital and postfrontal are preserved in Ba. macrodactylus and many additional details are visible as impressions on the matrix (Fig. 8). The postorbitofrontal bar is a robust and anteroposteriorly elongate structure. As in S. dyspepsia, the postfrontal of Ba. macrodactylus clasped the frontoparietal suture. However, the frontal process is much more robust than the parietal process in Ba. macrodactylus and the postorbital process is much broader and distally rounded. The postorbital of Ba. macrodactylus also is robust and its main body is anteroposteriorly and mediolaterally broad. Exclusive of its posterior (squamosal) process, it is nearly subequal in anteroposterior and mediolateral lengths. The squamosal process is attenuated and shorter than the main body of the postorbital. The postfrontal and postorbital are well preserved in E. schroederi (Fig. 9). The postfrontal clasps the frontoparietal contact with a relatively small frontal process and an elongate parietal process. The lateral margin of the postfrontal curved in dorsal view receives the concave medial surface of the postorbital. The posterior part of the postfrontal and the posteromedial surface of the postorbital body are slightly depressed, forming a shallow supratemporal fossa at the anterior end of the supratemporal fenestra. The robust postfrontal is mediolaterally and anteroposteriorly broad and possesses a robust squamosal process that tapers posteromedially. The specimen of Ardeosaurus described by Mateer (1982) possesses a postfrontal with a weakly forked medial surface that spans the frontoparietal suture. The frontal narrowly overlaps the anterior part of the postfrontal. The frontal process is approximately one-third the length of the parietal process. The postorbital is robust and elongate. The postorbital–postfrontal contact is elongate and is equivalent to more than one-half the length of the supratemporal fenestra. An apparent suture remains between the postfrontal and postorbital in NHM PR 38006 (Fig. 10A), but it is poorly defined. A similar suture is faintly visible in SNSB-BSPG 1923 I 501 (Fig. 10B). The area is damaged in CM 4026. Two small pieces of bone visible in the photograph in Hoffstetter (1964) appears to represent the left postfrontal and a fragment of the postorbital. The postfrontal is anteroposteriorly elongate and spans the frontoparietal suture. If these elements are correctly identified, they demonstrate that the two bones are unfused. Figure 10. View largeDownload slide A, the skull of NHM PR 38006 (cast of the holotype of Ardeosaurus brevipes) in dorsal view. B, the skull of SNSB-BSPG 1923 I 501 (labelled A. cf. brevipes) in dorsal view. Figure 10. View largeDownload slide A, the skull of NHM PR 38006 (cast of the holotype of Ardeosaurus brevipes) in dorsal view. B, the skull of SNSB-BSPG 1923 I 501 (labelled A. cf. brevipes) in dorsal view. Squamosal The posterior part of a squamosal, probably the right, is reserved near the parietal (Figs 5, 6). It shows the characteristic posteroventral downcurve that is common to most squamates. There is no indication of a dorsal process. This fragment is similar to the more complete squamosals present in Ba. macrodactylus (Fig. 8) and E. schroederi (Fig. 9). The squamosals of the Ardeosaurus specimens NHM PR 38006 and SNSB-BSPG 1923 I 501 are robust, but poorly preserved (Fig. 10). The posterior parts of the left squamosal in NHM PR 38006 and the right in SNSB-BSPG 1923 I 501 are the best preserved and each seems to demonstrate the absence of a dorsal process. The squamosals are well preserved in CM 4026 and clearly lack dorsal processes. By contrast, the specimen described and illustrated in Mateer (1982) demonstrates the presence of a robust dorsal process. Frontal The unpaired frontal is preserved in ventral view (Figs 5, 7). The anterior part is preserved and exposed, as is the left posterolateral part. The middle of the prefrontal is obscured by overlying pieces of the right mandible and a piece of bone that is of unknown identity, but which may represent the nasal process of the right maxilla. Because the frontal is preserved in ventral view, it is impossible to tell if there was dermal sculpturing on its dorsal surface. The frontal is hourglass-shaped with expanded anterior and posterior ends (Figs 5–7). The posterior end is broadest, but the antorbital part is also expanded and more than 1.8 times the width of the narrowest part of the bone in the posterior part of the orbit. The nasofrontal suture is W-shaped as indicated by the impression of the dorsal surface of the anterior part of the frontal (Figs 6, 7). A midline anterior projection anterior projection of the frontal would have laid deep to the nasal(s) in the articulated specimen. Short anterolateral processes are also present and extend to about the same anterior level as the midline projection. These processes are parallel-sided and anterolaterally oriented. They may partly underlie or invade the contact between the prefrontal and nasal. Their lateral surfaces have prefrontal facets. Dorsal to the level of the prefrontal facet, the frontal has a thin but well-developed superficial lamina that would have partly overlain the frontal process of the prefrontal in natural articulation. There is no development of the crista cranii of the frontal (Fig. 7). Posterior to the midpoint of the orbit, the frontal is expanded laterally. The frontoparietal suture is broad and somewhat anteriorly curved. Similar to S. dyspepsia, E. schroederi possesses an anteriorly arched frontoparietal suture (Figs 9, 11), albeit a less arched one. By contrast, Ba. macrodactylus possesses a more sinuous frontoparietal suture wherein the parasagittal anterior extensions of the parietal extend far forward of the posteromedial extension of the frontal (Fig. 8). Eichstaettisaurus gouldi seems to possess a less well-developed version of this type of suture (see Evans et al., 2004), such that it is somewhat W-shaped. The frontoparietal suture is transverse in A. brevipes (Mateer, 1982). Figure 11. View largeDownload slide Part of the mandible, braincase, palate and skull roof of Eichstaettisaurus schroederi (holotype; SNSB-BSPG 1937 I 1a) visible in ventral view. The specimen is extensively visible in dorsal view (see Fig. 9), and has been partly prepared from ventral view to show some relevant elements. Figure 11. View largeDownload slide Part of the mandible, braincase, palate and skull roof of Eichstaettisaurus schroederi (holotype; SNSB-BSPG 1937 I 1a) visible in ventral view. The specimen is extensively visible in dorsal view (see Fig. 9), and has been partly prepared from ventral view to show some relevant elements. The frontals are paired and possess a transverse frontoparietal suture in all of the Ardeosaurus specimens considered here. Parietals and supratemporals The parietals are preserved as a combination of small bits of bone and bone impressions with the left parietal being much more completely represented than the right (Fig. 5). They remain in articulation with the frontal. A small piece of what may be a left palatine partly obscures the lateral margin of the left parietal. The right parietal apparently passes out of the exposed plane on the matrix and is largely lost beyond the level of the rim of the pineal foramen. The paired parietals are joined at the midline by a straight suture. Together, the two parietal bodies form a skull table that is subequal in length and width (Figs 5, 6). A large pineal foramen is present within the parietal. Its anterior margin lays approximately one-third of the parietal body length from the middle of the frontoparietal suture. The frontoparietal suture is a gentle anterior curve, lacking frontal tabs. The visible parts of the lateral parietal margin indicate that it was anteroposteriorly straight and unconstricted. Only the left supratemporal process is preserved, but it demonstrates the anteroposterior (rather than mediolateral or posterolateral) orientation. The supratemporal process was relatively short and was approximately one-third the length of the body of the parietal. A transverse posterior margin of the parietal is present between the anteromedial bases of the supratemporal processes. Apparently, the jaw adductor musculature attached ventrally on the parietal. No discernible supratemporal is preserved in S. dyspepsia. Bavarisaurus macrodactylus possesses a broad parietal table similar to that of S. dyspepsia. It has been suggested that the parietal was fused as a single element (Estes, 1983). A later study disagreed, identifying a faint midline suture (Evans, 1994b). A faint feature that may represent a suture is present, but only anteriorly and just to the right of the midline. Because it is not as well defined in its appearance as the other sutures visible in the skull and because it is visible only at one point, I posit that this short furrow represents something other than a suture and that the parietals form a single, fused, element as suggested by Estes (1983). Eichstaettisaurus schroederi has been suggested as possessing incompletely fused parietal bones (Estes, 1983; Evans & Barbadillo, 1998, 1999). Inspection of the holotype suggests that this is not the case. The parietals are fused (Figs 9A, 11). The apparent incompletely fused suture is actually a small break or crack along the midline of the bone caused by diagenesis as the middle part of the parietal was pushed ventrally and the lateral sides pushed dorsomedially. This is confirmed by the ventral view of the specimen where the anterior part of the parietal is visible (Fig. 11). The parietal supratemporal processes are directed almost completely mediolaterally in Ba. macrodactylus (Fig. 8), contrasting the condition present in S. dyspepsia (Figs 5, 6), E. schroederi (Fig. 9) and Ardeosaurus (e.g. Fig. 10; Hoffstetter, 1964; Mateer, 1982). The left supratemporal process is the only part of the parietal preserved as bone in Ba. macrodactylus. The supratemporal process remains in contact with the supratemporal. The supratemporal bone is elongate and retains short contacts with the posterior part of the squamosal. Bavarisaurus macrodactylus is generally considered to lack a pineal foramen (Cocude-Michel, 1961; Hoffstetter, 1964; Estes, 1983). One study suggested that the pineal foramen is present (Evans, 1994b), but this was based on the presence of pineal foramen in the S. dyspepsia specimen which was, at the time, considered to be a specimen of Ba. macrodactylus. Re-examination of the holotype of Ba. macrodactylus reveals no indication of a pineal foramen on the parietal impression (Fig. 8). All of the Ardeosaurus considered here possess elongate, fused, parietals with posteriorly oriented supratemporal processes that are more than one-half the length of the parietal table. These parietals lack anterolateral expansions at the frontoparietal suture and a small pineal foramen located at about the midpoint of the parietal table. They also preserve vermiculate dermal sculpturing on their dorsal surfaces. The specimen described by Mateer (1982) preserves a small and weakly developed posteromedial process of the parietal. This may also be present in NHM PR 38006, but it appears to be absent in PMU.R58. The badly crushed parietal of P. bavarica is clearly azygous (see Hoffstetter, 1964; Estes, 1983). Unlike other known Solnhofen lizards, a large parietal foramen is present at the frontoparietal suture (Hoffstetter, 1964; Estes, 1983). Although the posterior part of the parietal is more damaged than the anterior part, elongate supratemporal processes are visible mostly by their bony outlines, and extend posterolaterally (see Hoffstetter, 1964; Estes, 1983). Palate Pterygoid Both pterygoids are preserved in ventral view in S. dyspepsia (Figs 5, 12). The right pterygoid (Fig. 12) is exposed a short distance away from (anteroventral to) the maxilla and (posterodorsal to) the left premaxilla. The left pterygoid is preserved and lies between the parietal and the jugal. Both pterygoids retain their contacts with the ectopterygoids. The right pterygoid is incomplete anteriorly with its palatine process. The left pterygoid is more complete, but its quadrate process has been eroded. Figure 12. View largeDownload slide Right pterygoid of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in ventral view. Anterior is toward the bottom and slightly offset to the left. The presumed columellar fossa has partly collapsed, creating the small circle of broken bone, and the quadrate process is somewhat damaged. Note the presence of a few small pterygoid teeth anteromedial of the underside of the columellar fossa, at the base of the vomerine process. Figure 12. View largeDownload slide Right pterygoid of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in ventral view. Anterior is toward the bottom and slightly offset to the left. The presumed columellar fossa has partly collapsed, creating the small circle of broken bone, and the quadrate process is somewhat damaged. Note the presence of a few small pterygoid teeth anteromedial of the underside of the columellar fossa, at the base of the vomerine process. The triradiate pterygoid possesses an elongate quadrate process, a robust transverse process that ends in an ectopterygoid articulation, and a narrow palatine process. A short ventral ridge is preserved on each pterygoid near the confluence of the transverse and palatine processes. This ridge is best preserved on the right pterygoid wherein the bases of three small teeth and a pit associated with a fourth tooth position are preserved (Figs 5, 12). The robust transverse process extends anterolaterally and is medially joined with the palatine process by a broad anterior flange – a ‘web’ of bone that unites those two processes. As a result of this anterior flange, the posterior margin of the suborbital fenestra would have been broadly rounded in the articulated skull. The lateral margin of the transverse process extends posteromedially toward the more anteroposteriorly oriented medial border of the pterygoid. Thus, the lateral, medial and suborbital fossa margins of the pterygoid together form an isosceles triangle wherein the medial and lateral margins of the bone describe the most acute angle. The columellar fossa is not directly observable in either of the ventrally exposed pterygoids. Even so, each pterygoid shows a circular area of collapsed bone within the main body of the pterygoid, near the confluence of the three major processes. This area of collapsed bone may represent an area of weakness created by the overlying columellar fossa. The quadrate processes are eroded on their exposed surfaces in each pterygoid, but they clearly were elongate and mediolaterally compressed. The quadrate process is approximately 135% the length of the transverse process as measured from the presumed columellar fossa. Bavarisaurus macrodactylus preserves no part of the pterygoid. The pterygoids are visible through the orbits in E. schroederi (Fig. 9) and demonstrate a relatively broad and transversely oriented parietal–palatine contact. Additionally, the suborbital fenestra is relatively narrower at its posterior terminus in E. schroederi than in S. dyspepsia. Pterygoids are in the Ardeosaurus specimens PMU.R58 (Mateer, 1982). The transverse processes are relatively narrow and the anterior contact with the palatine is, apparently, transversely broad (Mateer, 1982). The quadrate process is not preserved or visible. Ectopterygoid The short and robust ectopterygoids are exposed in ventral view and each is partly eroded (Figs 5, 12). The left ectopterygoid remains partly in articulation with the left maxilla and jugal. The ectopterygoid is short and anterolaterally oriented. Each ectopterygoid preserves a posteromedially oriented suture with the pterygoid. Only a small part of the ectopterygoid is preserved in Ba. macrodactylus. All that can be said about it is that it is somewhat posteromedially oriented. The well-preserved ectopterygoid of E. schroederi (Fig. 9) is short and robust. It possesses a significant dorsal process, but not a strong posteromedial dorsal overlap of the pterygoid. It has a strong anteroposterior orientation. Ardeosaurus PMU R.58 preserves a robust and elongate ectopterygoid (see Mateer, 1982) with a strong dorsal ridge and an elongate anterolateral process contributing to the lateral margin of the suborbital fenestra. The dorsomedial overlap of the pterygoid extends to the main body of the pterygoid. Presence or absence of an anterior palatine contact is uncertain. Specimen NHM PR 38006 retains the posterior part of the ectopterygoid (Fig. 10A; see also Estes, 1983). Impressions of both ectopterygoids are preserved in SNSB-BSPG 1923 I 501 (Fig. 10B). Mandible The left mandible is preserved as a combination of bone (anteriorly and posteriorly; Figs 2, 3, 5, 13) and impressions of bones from the level of the coronoid to an area just anterior to the glenoid fossa. Only the dentary is exposed for the right mandible (Figs 5, 7). Figure 13. View largeDownload slide The left mandible of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) in lateral view. The dentary is incomplete posteriorly and most of the mandible has been lost from the posterior end of the dentary to the level of the mandibular glenoid region. The latter region and the retroarticular process is somewhat damaged, but note the presence of a clear mandibular outline on the underlying matrix. Figure 13. View largeDownload slide The left mandible of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) in lateral view. The dentary is incomplete posteriorly and most of the mandible has been lost from the posterior end of the dentary to the level of the mandibular glenoid region. The latter region and the retroarticular process is somewhat damaged, but note the presence of a clear mandibular outline on the underlying matrix. Because of the state of preservation and/or orientation of the skeletons, little of the mandible is visible in Ba. macrodactylus (Fig. 8A) and E. schroederi (Figs 9A, 11). The available Ardeosaurus material preserves part of the middle and posterior parts of the mandibles (Fig. 10; see also Hoffstetter, 1964; Mateer, 1982; Estes, 1983). Even E. gouldi, which is preserved in ventral view, offers few details beyond the general shapes of the dentary, surangular and parts of the prearticular (Evans et al., 2004). These elements are of a generalized squamate form in E. gouldi. The dentaries converge anteriorly, probably (although not certainly) being connected via a symphysis. The dentary possesses a posterior emargination such that the surangular is visible laterally between the surangular and angular processes. A small part of the angular is probably preserved on the left mandible. The prearticular may have possessed a short, tapering, retroarticular process. Dentary The dentary is straight along its longitudinal axis and the dorsal and ventral margins are parallel (Figs 2, 3, 13). Nine mental foramina are preserved on the preserved part of the dentary. Of these, fourth is relatively small compared to the rest. The mandible possesses only very limited exposure in Ba. macrodactylus (Fig. 8A). The anterior tip of the left dentary is preserved in dorsomedial review. The exposed part of the dentary seems to be straight. Four teeth are visible in that view. Only the lateral surface of the dentary is not available in E. schroederi (Fig. 11), and there is no informative morphology there. Coronoid An impression of the left coronoid preserves a view of part of its medial surface (Fig. 13). The coronoid is dorsally arched or chevron-shaped and bears at least a short dorsal coronoid process. The full of extent of the coronoid process is uncertain (Figs 6, 13) because it remains unclear whether the impression captures its full extent. The coronoid is not visible in Ba. macrodactylus. Only the dorsal part of the coronoid is visible in a meaningful way in E. schroederi. Its dorsal process is moderately elongate and dorsally rounded. It appears to be chevron-shaped as it is in S. dyspepsia and many other squamates, but not rhynchocephalians and most snakes. Angular Part of the left angular is preserved in articulation with the other known parts of the mandible. Only the posterior end is preserved (Figs 6, 13). It appears to show a posteriorly tapering element with a small dorsal expansion just anterior to the level of the mandibular glenoid. It lays lateral to the surangular and the prearticular (see below). Surangular The posterior part of the left surangular is preserved as a combination of bone and impressions in articulation with the rest of the left mandible (Figs 6, 13). Sutures with the coronoid are not clearly preserved as bone or as impressions. A small dorsal buttress is present just anterior to the mandibular glenoid. Prearticular Because the prearticular and articular commonly fuse in squamates, we follow the terminology of Conrad & Norell (2007) in referring to this element as the prearticular. The preserved part of the prearticular lies ventral to and posterior to the mandibular glenoid (Figs 6, 13). No clear suture with the surangular is preserved. The retroarticular process is posteriorly directed and shows moderate posterior broadening. A dorsal pit is preserved on the dorsal surface of the retroarticular process and some torsion is present. Dentition A total of 18 marginal teeth are visible in addition to the three pterygoid teeth mentioned above. Two premaxillary are preserved with the left premaxilla (Fig. 5). Impressions of three teeth are preserved with the impression of the right premaxilla (Fig. 5; see above). Five maxillary tooth impressions are preserved with the left maxilla (Figs 5, 7). Eight dentary tooth positions are preserved on the left mandible (Fig. 13). The teeth are closely spaced and curved with sharp tips. Tooth implantation cannot be determined because all of the marginal tooth-bearing bones are preserved in lateral view. Vertebral and costal morphology Presacral vertebrae and ribs Eight presacral vertebrae with elongate ribs are preserved in dorsolateral (anteriorly; Figs 3, 14, 15) and dorsal (posteriorly; Figs 3, 15) view and in association with the two sacral vertebrae. Thus, these vertebrae are the last eight presacrals and indicate the absence of lumbar vertebrae. Because the anterior presacrals are not preserved (Figs 2, 3), the length of the presacral column is unknown. The preserved presacral vertebrae are exposed in dorsal view, but their neural spines are all damaged such that their morphology is unknown (Fig. 15). The antepenultimate vertebra is heavily damaged such that the neural arch is missing. Figure 14. View largeDownload slide The preserved dorsal, appendicular and part of the caudal region of the skeleton of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in situ. Note the overlying ribs of Compsognathus longipes. A, photograph. B, reconstructed. White areas indicate parts preserved in bone. Figure 14. View largeDownload slide The preserved dorsal, appendicular and part of the caudal region of the skeleton of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in situ. Note the overlying ribs of Compsognathus longipes. A, photograph. B, reconstructed. White areas indicate parts preserved in bone. Figure 15. View largeDownload slide Posterior dorsal vertebrae and associated ribs of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in dorsal view. Figure 15. View largeDownload slide Posterior dorsal vertebrae and associated ribs of Schoenesmahl dyspepsia gen. et sp. nov. (SNSB-BSPG AS I 563b) as preserved in dorsal view. The preserved presacral vertebrae neural arches and associated pre- and postzygapophyses are short and robust. Each postzygapophysis possesses a transverse posterior margin. The presence or absence of intercentra or a notochordal canal can be neither confirmed nor denied. A clear synapophysis is preserved in the second preserved presacral vertebra. The presacral ribs associated with the more anterior preserved dorsal vertebrae are more than four times as long as the prezygapophyseal–postzygapophyseal length. By contrast, the last rib is about half as long as the penultimate rib and only about twice the length of the vertebra to which it attaches. Bavarisaurus macrodactylus clearly preserves intercentra in its presacral vertebrae (Fig. 16). It possesses between 23 and 25 presacral vertebrae (Fig. 4; Cocude-Michel, 1963; Evans, 1994b). I estimate the count at 25 based on the preserved bones and the bone impressions. The last dorsal vertebra has an elongate and apparently mobile rib (Fig. 17); there were no lumbar vertebrae. The last rib is approximately 1.3 times as long as the last vertebra. As preserved, the ribs appear to shorten more gradually as they approach the sacrum in Ba. macrodactylus than in S. dyspepsia. The antepenultimate rib is significantly longer than the penultimate, which is approximately 1.5 times as long as the last rib. Figure 16. View largeDownload slide Anterior dorsal vertebrae and left forelimb of Bavarisaurus macrodactylus (SNSB-BSPG 1873 III 501) as preserved in (generally) ventral view. Figure 16. View largeDownload slide Anterior dorsal vertebrae and left forelimb of Bavarisaurus macrodactylus (SNSB-BSPG 1873 III 501) as preserved in (generally) ventral view. Figure 17. View largeDownload slide Posterior dorsal vertebrae, partial sacrum and pelvis and anterior caudal regions of of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501) as preserved in (generally) ventral view. Figure 17. View largeDownload slide Posterior dorsal vertebrae, partial sacrum and pelvis and anterior caudal regions of of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501) as preserved in (generally) ventral view. Eichstaettisaurus schroederi possesses 29 presacral vertebrae (Fig. 18). The last five of these vertebrae possess shorter ribs than those immediately anterior to them. The ribs on vertebrae 25–28 are approximately 1.5 times as long as the vertebrae to which they attach; the last rib is slightly shorter than the associated vertebra. Figure 18. View largeDownload slide Eichstaettisaurus schroederi (holotype part – counter-part not shown; SNSB-BSPG 1937 I 1a). Figure 18. View largeDownload slide Eichstaettisaurus schroederi (holotype part – counter-part not shown; SNSB-BSPG 1937 I 1a). The skeletons that have been referred to Ardeosaurus all lack complete enough axial columns to confidently reconstruct a specific presacral vertebral count. Even so, each preserves approximately 28 presacral vertebrae (e.g. Fig. 19; see also Hoffstetter, 1964; Mateer, 1982; Estes, 1983). Figure 19. View largeDownload slide A, cast of the holotype of Ardeosaurus brevipes (NHM PR 38006) in dorsal view. B, skeleton of SNSB-BSPG 1923 I 501 (labelled A. cf. brevipes) in dorsal view. Figure 19. View largeDownload slide A, cast of the holotype of Ardeosaurus brevipes (NHM PR 38006) in dorsal view. B, skeleton of SNSB-BSPG 1923 I 501 (labelled A. cf. brevipes) in dorsal view. Sacral vertebrae The two sacral vertebrae are not well preserved in S. dyspepsia. Each is partly obscured by the impression of an overlying dorsal C. longipes rib (Figs 2, 3, 14, 15). The anterior sacral vertebra is almost completely obscured and no meaningful details available beyond the fact that it exists. The second sacral vertebra preserves a broken part of its right sacral rib. The preserved part of the latter suggests a subovate cross-section. Bavarisaurus macrodactylus does not preserve any clear details for the centrum or for the neural arch for either sacral vertebra (Fig. 17). Both sacrals preserve incomplete impressions of their neural arches. The sacral vertebra preserves as a distal part of the sacral rib on the right side, but this fragment only indicates that the expansion is slightly less expansive than that of the second sacral rib. The left side preserves an incomplete impression of the sacral rib that contacts the impression for the left ilium. A clear impression is present for the distal part of the left rib of the second sacral. The right side preserves an impression for the proximal part of the rib. Distally, the rib itself is preserved. The right side of the sacrum is better preserved than the left in E. schroederi and preserves parts of both sacral ribs (Fig. 20). The left side preserves only the first sacral rib. The exposed part of the first sacral rib is anteroposteriorly narrow. The second sacral rib is much broader than the first and possesses a posterior lamina that gives it an anteriorly arched posterior margin. It is also slightly expanded anterolaterally. The anterior part of this lamina slightly dorsally overlaps the posterolateral part of the first sacral rib. Figure 20. View largeDownload slide Pelvis, left hind limb and preserved caudal region of Eichstaettisaurus schroederi (holotype part; SNSB-BSPG 1937 I 1a). Figure 20. View largeDownload slide Pelvis, left hind limb and preserved caudal region of Eichstaettisaurus schroederi (holotype part; SNSB-BSPG 1937 I 1a). The sacrum and sacral ribs of P. bavarica are damaged and I cannot competently describe them at this time. The cast of the A. brevipes holotype has a damaged sacrum lacking clear sacral ribs and neural arches. The specimen described in Mateer (1982) is also damaged, but a damaged sacral rib is preserved. Almost nothing is known of the sacrum in SNSB-BSPG 1923 I 501. The sacrum and its sacral ribs are incompletely preserved as impressions on the matrix. The sacral ribs are broad and do not appear to contact distally. Caudal vertebrae Approximately 47 caudal vertebrae are preserved in a nearly consecutive series (Figs 2, 3, 14). These are looped into the anterior part of the body of C. longipes, associated with the second through fourth elongate ribs of the dinosaur. A study by Ostrom (1978) convincingly argued that all of these caudal vertebrae belong to the same animal and this is confirmed by direct observation. The first caudal is damaged such that the transverse processes are not preserved. The second caudal preserves broad transversely oriented transverse processes that do not taper distally. The third, fourth and fifth caudals are not well preserved, but they do demonstrate that the transverse processes were becoming posterolaterally oriented by the level of the third caudal vertebra. The fifth caudal transverse processes demonstrate the presence of a faint autotomy septum. These autotomy septa are present throughout the rest of the caudal series; transverse processes are clearly preserved through the ninth caudal vertebra. No autotomy foramina are evident on any of the caudals preserving transverse processes. Posterior to the ninth caudal vertebra, there is a discontinuity within the caudal vertebral series. There is no way to know how many caudals are missing, but the next vertebra is similar in length to the ninth caudal vertebra, yet possesses no clear transverse processes. No subsequent vertebra possesses a clear transverse process. Also, the prezyapophyses and postzygapophyses are very weakly developed throughout the rest of the vertebrae. The autotomy planes are located near the midlength of the vertebrae. Bavarisaurus macrodactylus possesses an incomplete caudal series, but the first 23 caudals are preserved in an unbroken string (Fig. 4). In contrast to S. dyspepsia, the first transverse process is posterolaterally oriented (Fig. 17). The second transverse process is more transversely oriented than the first caudal vertebra. Subsequent transverse processes are similarly oriented; that is, they are mostly transversely oriented with a slight posterolateral inclination (Figs 4, 17, 21). Bavarisaurus macrodactylus possesses autotomy planes on every preserved caudal vertebra. Figure 21. View largeDownload slide Left hind limb zeugopodium and pes, and proximal and middle tail of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501) with impressions of the dorsal view and bones preserved in ventral view. Figure 21. View largeDownload slide Left hind limb zeugopodium and pes, and proximal and middle tail of Bavarisaurus macrodactylus (holotype; SNSB-BSPG 1873 III 501) with impressions of the dorsal view and bones preserved in ventral view. Although most of the skeleton of Ba. macrodactylus is preserved in ventral view, the tail is twisted such that the neural spine is partly visible on the specimen’s left side beginning on caudal vertebra fourth caudal vertebra. The fifth caudal vertebra clearly demonstrates a broad and dorsally oriented neural spine. Its dorsal margin is squared. Chevrons are also preserved and visible beginning with the fifth caudal vertebra. The chevron on the fifth caudal vertebra is slightly longer than the centrum. The neural spines become increasingly diminished in dorsoventral and anteroposterior lengths. By the level of the eighth caudal vertebra, the neural spine is very reduced. Its dorsal height is approximately equal to one-third of the centrum length. The eighth caudal vertebra also preserves a very clear autotomy plane and demonstrates the absence of a clear pseudospine. By the level of the 17th caudal vertebra, the neural spine appears to be absent. Eichstaettisaurus schroederi preserves the first six caudal vertebrae (Fig. 20). These vertebrae possess well-developed transverse processes. These transverse processes are not as laterally extensive as in S. dyspepsia or in Ba. macrodactylus. These transversely oriented transverse processes appear to be broadest at their midlength and taper distally. Beyond the preserved vertebrae, the E. schroederi preserves a stain that seems to be a continuation of the tail. This trace is equal to approximately half of the precaudal length of the animal. The caudal series in NHM PR 38006 is incomplete and damaged (Fig. 19A). Only the faintest impressions of the caudal series is preserved in the Ardeosaurus specimen SNSB-BSPG 1923 I 501 (Fig. 19B). The caudals of CM 4026 are represented only by impressions on the matrix. The first five caudal vertebrae are well preserved in the Ardeosaurus described in Mateer (1982), the sixth retains part of a transverse process and a centrum, and the following five are known only from partial centra. The preserved transverse processes extend directly mediolaterally. These are the best preserved Ardeosaurus caudal vertebrae considered here. Much of the caudal series is preserved in P. bavarica (Hoffstetter, 1964), but the vertebrae are damaged. Autotomy septa are preserved in the caudal vertebrae posterior to the transverse processes. Appendicular skeleton Pectoral girdle Nothing is known of the pectoral girdle. Only the left humerus and associated part of the proximal forelimb zeugopodium is preserved (Figs 3, 22). Figure 22. View largeDownload slide The preserved part of the left humerus of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) in ventral view along with some of the dorsal ribs in dorsal view. Figure 22. View largeDownload slide The preserved part of the left humerus of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) in ventral view along with some of the dorsal ribs in dorsal view. Most of the Solnhofen lizards preserve very little or no pectoral girdle elements. Bavarisaurus macrodactylus preserves only an incomplete right clavicle. It appears to be slightly dorsally expanded. The subtriangular scapulae and the medially expanded clavicles are preserved in E. schroederi (Fig. 18). A scapular emargination (defined by a scapular epicoracoid bar) is absent and a posterior (secondary) coracoid emargination cannot be confirmed, but a primary emargination is present. An incomplete pectoral girdle is preserved as a partial impression in A. brevipes (NHM PR 38006). Although this impression includes a small amount of the scapulocoracoid unit, virtually no details are available for it. Two little of the clavicle is preserved to offer meaningful details in this specimen and in SNSB-BSPG 1923 I 501, for which the scapulocoracoid is similarly poorly known. More of the scapulocoracoid is preserved in CM 4026 and demonstrates the absence of a scapular epicoracoid bar, but the presence of at least a primary coracoid emargination (see Hoffstetter, 1964). Palaeolacerta bavarica preserves only a small part of the clavicle, but also a complete scapulocoracoid (Hoffstetter, 1964; Estes, 1983). This specimen preserves part of the cartilaginous skeleton, including costal cartilages and, perhaps, a part of the epicoracoid. Although this scapulocoracoid has been suggested as possessing a single coracoid emargination and no scapular emargination (Hoffstetter, 1964; Estes, de Queiroz & Gauthier, 1988), a secondary coracoid emargination may be present and obscured by the partial epicoracoid. Further study of the specimen will be necessary to determine whether this is correct. Humerus The left humerus is incompletely preserved in ventral view (Figs 14, 22). The proximal head is not preserved. An impression remains of approximately two-thirds of the diaphysis. The distal part of the humerus is preserved in total. The preserved part of the humerus is approximately 11.5 mm long, but the complete humerus was likely approximately 13.5 mm long. The ectepicondylar–entepicondylar breadth is approximately 4.1 mm. The narrowest preserved part of the diaphysis is approximately 0.9 mm across. Thus, the humerus is quite gracile. An impression of the deltopectoral crest is preserved. The entepicondyle is robust with a distally concave proximal surface that extends laterally to form an acute anterior angle. The ectepicondyle is not as robust as the entepicondyle. It extends only slightly beyond the level of the radial condyle. A small ectepicondylar foramen is present at about the same distal level as the proximal margin of the radial condyle. A broad entepicondylar notch (sensuRieppel, 1980) is proximally rounded and distally expands to span the ulnar condyle. The subovate radial condyle is approximately 1.8 mm long and 1.1 mm across. The ulnar condyle is 1.3 mm long and 1 mm across. Similar parts of the humerus are known in Ba. macrodactylus (Fig. 16), although both humeri are known for the latter. Bavarisaurus macrodactylus possesses a similarly robust entepicondyle to that of S. dyspepsia. Whereas the entepicondyle is attenuated in S. dyspepsia (Fig. 22), it is distally squared in Ba. macrodactylus (Fig. 16). The morphology of the entepicondylar notch is similarly deeply developed between the two species, but does not expand distally as much in Ba. macrodactylus. The greatest distal width of the humerus is approximately 4.5 mm; its narrowest observable width is 1.4 mm. The length of the radial condyle may only be estimated because its distal terminus is hidden by the articulated radius; it is approximately 1.9 mm long and 1.3 mm across. The ulnar condyles are small, partly obscured and damaged. Both humeri are preserved in posterior view in E. schroederi (Fig. 18). Neither deltopectoral crest is visible, nor are the entepicondyles clearly exposed. The posterodistal part of each humerus is slightly eroded. Even so, presence of an ectepicondylar foramen is confirmed. The specimen of Ardeosaurus described by Mateer (1982; PMU.R58) possesses a humerus that is approximately 8.2 mm long, or approximately 51% the length of the skull roof at the midline. This is a proportion that is more or less conserved across the compared Ardeosaurus specimens here (Table 1). The broadest part of the distal epicondyles of the humerus in SNSB-BSPG 1923 I 501 is approximately 4.2 times the minimum measureable shaft diameter. Table 1. Osteometrics of selected Jurassic and Cretaceous squamates   PCl  SRl  Ml  Hl  dwH  mwH  Fl  mwF  Tl  Pl  Schoenesmahl dyspepsia    15.1  18.3  13.5  4.1  0.9  18.2  2.1  16.9  30.6  Bavarisaurus macrodactylus  91  20      4.7  1.3  22  2.1  18.1  33.5  Eichstaettisaurus schroederi  85.5  17  18.2  10    0.9  15.3  1.3  10.8  15.4  ‘Eichstaettisaurus’ gouldi  47    13.5  7      7    5.4    ‘Ardeosaurus’ (Mateer, 1982)  78.2  16.3  19.4  8.2      11.8    6.4  15.8  NHM PR 38006  58.1  11  13.1  6.3  2.2    8.3  0.7  5.3  10.7  BSP 1923 I 501  84.4  17.3  20.5  9.8  2.9  0.9  12.5  1  8.1  17.9  Chometokadmon fitzingeri  128.8  27.2    16.7  4.6  1.4  19.6  2  17    Scandensia ciervensis  36.9    6.4  3.7      4.8    3.3  6.7  Hoyalacerta sanzi  39.8    8.5  1.8      4.4    2.9    Huehuecuetzpalli mixtecus  112.4  28.6  30.2  15.7  5.3  1.8  24.5  3.1  21.6  37.1  ‘Ardeosaurus’ digitatellus  89  19          13.1  1  8.3  16.2  Palaeolacerta bavarica  41  8.1  9.9  4.6      5    4.2      PCl  SRl  Ml  Hl  dwH  mwH  Fl  mwF  Tl  Pl  Schoenesmahl dyspepsia    15.1  18.3  13.5  4.1  0.9  18.2  2.1  16.9  30.6  Bavarisaurus macrodactylus  91  20      4.7  1.3  22  2.1  18.1  33.5  Eichstaettisaurus schroederi  85.5  17  18.2  10    0.9  15.3  1.3  10.8  15.4  ‘Eichstaettisaurus’ gouldi  47    13.5  7      7    5.4    ‘Ardeosaurus’ (Mateer, 1982)  78.2  16.3  19.4  8.2      11.8    6.4  15.8  NHM PR 38006  58.1  11  13.1  6.3  2.2    8.3  0.7  5.3  10.7  BSP 1923 I 501  84.4  17.3  20.5  9.8  2.9  0.9  12.5  1  8.1  17.9  Chometokadmon fitzingeri  128.8  27.2    16.7  4.6  1.4  19.6  2  17    Scandensia ciervensis  36.9    6.4  3.7      4.8    3.3  6.7  Hoyalacerta sanzi  39.8    8.5  1.8      4.4    2.9    Huehuecuetzpalli mixtecus  112.4  28.6  30.2  15.7  5.3  1.8  24.5  3.1  21.6  37.1  ‘Ardeosaurus’ digitatellus  89  19          13.1  1  8.3  16.2  Palaeolacerta bavarica  41  8.1  9.9  4.6      5    4.2    All measurements are in millimetre. dwH, maximum width of the distal end of the humerus; Fl, femur length; Hl, humerus length; Ml, mandible length; mwF, minimum (midshaft) width of the humerus; mwH, minimum (midshaft) width of the humerus; PCl, precaudal length; Pl, pes length; SRl, skull roof length; Tl, tibia length. View Large Table 1. Osteometrics of selected Jurassic and Cretaceous squamates   PCl  SRl  Ml  Hl  dwH  mwH  Fl  mwF  Tl  Pl  Schoenesmahl dyspepsia    15.1  18.3  13.5  4.1  0.9  18.2  2.1  16.9  30.6  Bavarisaurus macrodactylus  91  20      4.7  1.3  22  2.1  18.1  33.5  Eichstaettisaurus schroederi  85.5  17  18.2  10    0.9  15.3  1.3  10.8  15.4  ‘Eichstaettisaurus’ gouldi  47    13.5  7      7    5.4    ‘Ardeosaurus’ (Mateer, 1982)  78.2  16.3  19.4  8.2      11.8    6.4  15.8  NHM PR 38006  58.1  11  13.1  6.3  2.2    8.3  0.7  5.3  10.7  BSP 1923 I 501  84.4  17.3  20.5  9.8  2.9  0.9  12.5  1  8.1  17.9  Chometokadmon fitzingeri  128.8  27.2    16.7  4.6  1.4  19.6  2  17    Scandensia ciervensis  36.9    6.4  3.7      4.8    3.3  6.7  Hoyalacerta sanzi  39.8    8.5  1.8      4.4    2.9    Huehuecuetzpalli mixtecus  112.4  28.6  30.2  15.7  5.3  1.8  24.5  3.1  21.6  37.1  ‘Ardeosaurus’ digitatellus  89  19          13.1  1  8.3  16.2  Palaeolacerta bavarica  41  8.1  9.9  4.6      5    4.2      PCl  SRl  Ml  Hl  dwH  mwH  Fl  mwF  Tl  Pl  Schoenesmahl dyspepsia    15.1  18.3  13.5  4.1  0.9  18.2  2.1  16.9  30.6  Bavarisaurus macrodactylus  91  20      4.7  1.3  22  2.1  18.1  33.5  Eichstaettisaurus schroederi  85.5  17  18.2  10    0.9  15.3  1.3  10.8  15.4  ‘Eichstaettisaurus’ gouldi  47    13.5  7      7    5.4    ‘Ardeosaurus’ (Mateer, 1982)  78.2  16.3  19.4  8.2      11.8    6.4  15.8  NHM PR 38006  58.1  11  13.1  6.3  2.2    8.3  0.7  5.3  10.7  BSP 1923 I 501  84.4  17.3  20.5  9.8  2.9  0.9  12.5  1  8.1  17.9  Chometokadmon fitzingeri  128.8  27.2    16.7  4.6  1.4  19.6  2  17    Scandensia ciervensis  36.9    6.4  3.7      4.8    3.3  6.7  Hoyalacerta sanzi  39.8    8.5  1.8      4.4    2.9    Huehuecuetzpalli mixtecus  112.4  28.6  30.2  15.7  5.3  1.8  24.5  3.1  21.6  37.1  ‘Ardeosaurus’ digitatellus  89  19          13.1  1  8.3  16.2  Palaeolacerta bavarica  41  8.1  9.9  4.6      5    4.2    All measurements are in millimetre. dwH, maximum width of the distal end of the humerus; Fl, femur length; Hl, humerus length; Ml, mandible length; mwF, minimum (midshaft) width of the humerus; mwH, minimum (midshaft) width of the humerus; PCl, precaudal length; Pl, pes length; SRl, skull roof length; Tl, tibia length. View Large Radius and ulna Only the proximal parts of the radius and ulna are preserved and visible (Figs 14, 22, 23). Both are partly obscured by overlying ribs and vertebrae. The proximal part of the radius is expanded such that it is slightly broader than the radial condyle of the humerus. The preserved proximal part of the ulna is similar in breadth to that of the radial head. Figure 23. View largeDownload slide Posterior dorsal and anterior caudal vertebrae, sacrum, pubis, ischium and proximal femora of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) in dorsal view. Figure 23. View largeDownload slide Posterior dorsal and anterior caudal vertebrae, sacrum, pubis, ischium and proximal femora of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) in dorsal view. The radii and ulnae are more completely preserved in the other Solnhofen lizards, but the limited visible parts of those bones in S. dyspepsia preclude detailed comparisons. Pelvic girdle and hind limb The right side of the pelvis and parts of both hind limbs are preserved in association with the axial skeleton in S. dyspepsia (Figs 2, 3, 14, 23, 24). The anterior part of the pubis is partly overlaid by a dorsal rib from C. longipes. Part of what is probably an ischium is narrowly visible near the proximal end of pubis and projecting under the first caudal vertebra. The right femur is preserved as part of a cast of that was made before the specimen was further prepared (Fig. 24). Almost the entire left hind limb is preserved as a combination of bones and impressions of bones. Figure 24. View largeDownload slide Preserved dorsal, sacral and right hind limb region of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) as preserved in a cast of Compsognathus longipes (SNSB-BSPG AS I 563b). The right femur of S. dyspepsia gen. et. sp. nov. shown here was largely lost some time after creation of the mould for this cast. Figure 24. View largeDownload slide Preserved dorsal, sacral and right hind limb region of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) as preserved in a cast of Compsognathus longipes (SNSB-BSPG AS I 563b). The right femur of S. dyspepsia gen. et. sp. nov. shown here was largely lost some time after creation of the mould for this cast. No ilium is preserved with S. dyspepsia. Parts of the ilia are preserved in Ardeosaurus, E. schroederi and Ba. macrodactylus. However, none of those preserved ilia show details of the iliac anatomy, nor do they offer any indication of any unusual characteristics. Pubis The right pubis is preserved in dorsal view lying near the distal end of the left humerus and the proximal end of the right femur (Fig. 23). The pubis is robust with a broad symphyseal part that is more elongate than the tubercular part. The symphyseal part is approximately 4.3 mm whereas the tubercular part is 3.9 mm in length. The broadest visible part of the symphyseal ramus is approximately 2.8 mm in anteroposterior width. The symphyseal part of the pubis forms a slightly obtuse angle with the tubercular part. The pubic tubercle is robust and appears to be distally squared. The obturator foramen is anteroventrally oriented. A fossa is present extending posterodorsal to the obturator foramen and contiguous with it. This fossa is somewhat broader than the obturator foramen. The posterior margin of the symphyseal part of the pubis is nearly transverse, suggesting an anteriorly broad thyroid fenestra. The right pubis is well preserved in ventral view in Ba. macrodactylus (Figs 4, 17). It is similar to the pubis of S. dyspepsia in that the symphyseal part (6.7 mm long) is longer than the tubercular part (3.4 mm long), but the Ba. macrodactylus pubis is less robust than S. dyspepsia. The broadest preserved and visible part of the symphyseal part of the pubis (including the impression of its anterior margin) in Ba. macrodactylus is approximately 2.1 mm wide. Bavarisaurus macrodactylus has a pubis in which the symphyseal part forms a very obtuse angle with the tubercular part; the symphyseal part is anteromedially oriented in the Ba. macrodactylus and the anterior margin of the thyroid fenestra extends far anterior to the level of the pubic tubercle. The pubic tubercle is robust, but it is anterolaterally attenuated. The obturator foramen is anteroventrally oriented. Eichstaettisaurus schroederi preserves largely complete pubes, but these are partly overlaid by vertebrae, ribs, the proximal parts of the femora and part of the ilium (Figs 18, 20). Although few details of these pubes are present, the state of their preservation indicates that they were more anteromedially oriented than those of S. dyspepsia and have an orientation similar to that present in Ba. macrodactylus. The Ardeosaurus specimens NHM PR 38006 (Fig. 19A), SNSB-BSPG 1923 I 501 (Fig. 19B) and CM 4026 possess very incompletely known pubes. The specimen described in Mateer (1982) has a clearly preserved right pubis with an elongate symphyseal part and a short tubercular part. Ischium A flat piece of bone is seen extending ventral to the first caudal vertebra. Originally tentatively identified as the left pubis (Ostrom, 1978), this element probably represents the right ischium (Figs 2, 3, 23). The preserved part of the ischium demonstrates a relatively straight dorsal margin. Although partly hidden, the ventral margin suggests that ischium is gently constricted just distal to its acetabular part before expanding posteroventrally. The extent of this expansion and the distal morphology are obscured by the second caudal vertebra. No details of the ischium are available for Ardeosaurus. None of the ischium is visible in Ba. macrodactylus or in E. schroederi. Femur A complete right femur is preserved on a cast of the specimen (Fig. 24); this femur was largely lost subsequent to the making of the mould for this cast. This femur is approximately 16.9 mm long. The preserved part of the left femur (Fig. 14) matches this length. The narrowest part of the femur occurs at approximately midshaft and measures approximately 2.1 mm across (Table 1). The left femur is preserved in dorsal view; the right femur is preserved in posterodorsal view. The femur cast demonstrates a mild dorsal arching to the femur distally paired with a ventral proximal arching. Both femora are represented in Ba. macrodactylus (Fig. 4). The left femur is preserved only as an impression on the rock and lacks its proximal end. The right femur is nearly complete and lacks only the internal trochanter, which has been eroded away. The right femur is 21.9 mm long and approximately 2.1 mm wide at its narrowest point. The femur shows the same curvature present in the femur of S. dyspepsia. Eichstaettisaurus schroederi preserves both gracile femora as bone (Figs 18, 20). The femora are approximately 15.3 mm long and the minimum measurable diameter is approximately 1.3 mm. Although E. schroederi possesses the curvature of the femur, it appears less pronounced than in S. dyspepsia and Ba. macrodactylus. Eichstaettisaurus gouldi preserves incomplete femora (Evans et al., 2004). The right femur is slightly more complete proximally and appears to preserve the proximal head, or most of the proximal head. It is approximately 7 mm long, about the same length as the humerus (Table 1). The femur is more than 1.5 times the tibia length in all of the specimens of Ardeosaurus considered here. The femur is approximately 9.6 mm long in the holotype of A. brevipes, 11.8 mm long in the specimen described in Mateer (1982), approximately 12 mm long in SNSB-BSPG 1923 I 501 and 13.1 mm long in CM 4026. Palaeolacerta bavarica preserves both femora, but neither is complete (Hoffstetter, 1964; Estes, 1983). When complete, the femur would have been approximately 5 mm long, approximately 1.2 times the length of the reconstructed tibia. Tibia and fibula Distal parts of the right tibia and fibula are preserved near the anterior caudal vertebrae (Figs 2, 3, 25). Based on their position and the position of the right femur, they were probably in natural articulation during fossilization, but were subsequently damaged and/or eroded. The ends of the left tibia and fibula are preserved as bone, but the middle parts of these bones are preserved only as impressions on the matrix. Although the middle parts of the impressions are partly obscured by overlying ribs from C. longipes, the state of preservation and those of the bones around them suggest that the preserved parts are undisturbed. The tibia is 17.3 mm long and the fibula is 17.8 mm long. Figure 25. View largeDownload slide Left hind limb zeugopodium and pes, and some caudal vertebrae of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) as preserved in the anterior rib cage of Compsognathus longipes (SNSB-BSPG AS I 563b). Figure 25. View largeDownload slide Left hind limb zeugopodium and pes, and some caudal vertebrae of Schoenesmahl dyspepsia gen. et. sp. nov. (SNSB-BSPG AS I 563b) as preserved in the anterior rib cage of Compsognathus longipes (SNSB-BSPG AS I 563b). The tibiae and fibulae of Ba. macrodactylus are each preserved as a combination of bones and bone impressions (Figs 4, 21). The tibia is approximately 18.7 mm long and the fibula is almost exactly the same length. Thus, the tibia is longer than the femur in S. dyspepsia, but the femur is longer than the tibia in Ba. macrodactylus. Eichstaettisaurus schroederi possesses proportionately shorter limbs compared to the skull length (Fig. 18; Table 1) than does S. dyspepsia and Ba. macrodactylus. The tibia is approximately 10.8 mm long in E. schroederi. Thus, it is much shorter than the femur (Figs 18, 20). The femur is mostly straight in E. schroederi, but there is a distinctive anterodistal curve at the level of the fourth trochanter (Fig. 20). Well-preserved hind limb zeugopodia are preserved as bone in the A. breviceps holotype and the Ardeosaurus specimen described in Mateer (1982), as clear impressions in SNSB-BSPG 1923 I 501, and as a combination of bone and impressions in CM 4026. Palaeolacerta bavarica preserves partial hind limb zeugopodia. The left tibia is nearly completely preserved, but only the distal part of the fibula is preserved. The right zeugopodium is poorly preserved. Tarsals Only the left calcaneum is clearly preserved in S. dyspepsia (Fig. 25), and it is slightly damaged medially. The pronounced fibular facet has a distinct lip. Damage to the medial surface of the calcaneum precludes determining whether the non-preservation of the astragalus is the result of damage or an unfused condition. Even so, the preserved part of the calcaneum demonstrates that the tibial and fibular condyles are separated by a narrow space similar to less than one-half the width of the distal end of the tibia. No distal tarsals are preserved. The fused astragalocalcanea are bilaterally preserved in Ba. macrodactylus (Figs 4, 21). The tibial and fibular facets are separated by a distance of about one-quarter the distal tibial breadth. The calcaneum has a modest lateral tuberosity that extends only slightly beyond the level of the distal fibular head. The lateral tuberosity has a well-defined proximodistal ridge. Eichstaettisaurus schroederi preserves damaged tarsals (Figs 18, 20) that offer few morphological details. Ardeosaurus specimens lack well-preserved tarsals. The specimen described in Mateer (1982; PMU.R58) preserves a robust and fused astragalocalcaneum with a lateral tuberosity. Both NHM PR 38006 and SNSB-BSPG 1923 I 501 preserve an unfused astragalus and calcaneum (Fig. 19). A lateral tuberosity is preserved in CM 4026, but it is not as robust as in PMU.R58, and the potential fusion of the astragalus and calcaneum remains uncertain. Pedes Parts of all five left metatarsals and parts of four digits are preserved (Fig. 25). Only the distal part of the metatarsal and the proximal phalanx are preserved for the first digit. Only the proximal parts of the second and third metatarsals are missing. None of the phalanges are preserved in digit II, but all of the phalanges are preserved on digit III where there are four phalanges, including the ungual. The penultimate and antepenultimate phalanges are of subequal length. No phalangeal sesamoids are preserved, but their original absence cannot be confidently inferred due to the state of preservation. The fourth metatarsal and digit are almost as complete, but the ungual is not preserved. Four phalanges are preserved in the fourth metatarsal. This digit almost certainly possessed five phalanges. Although the fifth metatarsal is nearly complete, the proximomedial surface is hidden. Regardless, this metatarsal was clearly very short. Only one phalanx is preserved on the fifth digit. The proportions of the pedes are similar in