Eichstaettisaurus schroederi, BSPG, 1937

OSTEOLOGY OF EICHSTAETTISAURUS SCHROEDERI

The holotype of E. schroederi ( BSPG 1937 I 1 a & b) consists of two slabs representing part and counterpart. One slab includes all the preserved bone elements in dorsal view, and has been partially prepared on the opposite side, revealing portions of the ventral side of the skull. The counterpart preserves only the impression of those elements in the soft calcareous matrix, but provides some good morphological detail .

The condition seen in E. schroederi is marked as ‘X’, and the opposite condition is marked as ‘Y’. For feature number 8, a third condition is added, ‘Z’, which is exclusive of E. gouldi . Missing data denoted as ‘?’. Shared features across Eichstaettisaurus and Ardeosaurus = 3; between E. schroederi and A. brevipes only = 1; between E. schroederi and A. digitatellus only = 4; and between A. brevipes and A. digitatellus only = 7.

Cranium

The skull is depressed ( Fig. 2A, B View Figure 2 ) and the bones are unsculptured. The snout is short, broad, and rounded anteriorly. The premaxillae are paired and possess a short nasal process that contacts the anterior end of the nasals forming the anteromedial border of the external nares. The maxillary process in each premaxilla is short, and has a smoothly rounded posterior end that contacts the maxilla ( Fig. 3 View Figure 3 ). The anterior end of the premaxillary process of each maxilla has a similar shape, and it is likely that a soft tissue connection existed between the two elements. No anterior ethmoidal foramina are visible on the premaxillae, indicating that the medial ethmoidal nerves probably exited through the large external nares.

The nasals are paired and their anterior halves form the medial – posteromedial borders of the external nares. No nasal foramina are observable dorsally, and there are no ventrolateral projections. On the ventral side, the straight suture between the nasals is visible and there is no midline crest. The septomaxillae can be seen within the nasal capsule on both sides. They seem to be flattened anteriorly, although this could be because of the deformation of the specimen.

Each maxilla bears a short premaxillary process, which forms a steep angle with the maxillary nasal process. The nasal process on each side contacts the nasal dorsomedially and the prefrontal posteriorly. Small alveolar foramina for the cutaneous branches of the superior alveolar nerve and artery are visible close to the margin of the tooth row, but the anterior superior alveolar foramen is not visible externally. It is probably present close to the base of the medial margin of the nasal process and thus not visible in lateral aspect. The suborbital process of each maxilla extends to the midpoint of the orbit with its posterior tip lying ventral to the anterior end of the jugal.

A lacrimal bone is absent, and the anterior border of the orbit is formed mostly by the prefrontals, which lack any surface ornamentation and possess a faint ridge projecting posterodorsally. This ridge does not form a prefrontal boss, as seen amongst many iguanians. The lacrimal foramen is not visible laterally, and it was probably present posteriorly, with the lacrimal duct opening into the orbital cavity, as observed in most lizards.

Both jugals are preserved. Whereas the left element is apparently complete, the right one is broken, preserving only its postorbital process. Each jugal is a semilunate element extending far anteriorly, approaching the prefrontal. Each one lacks a posteroventral process, and ascends posteriorly in a smooth curve contacting the postorbital on its posterior margin. The posterior end of the postorbital process of the jugal approaches the anterior end of the squamosal, but does not contact it.

The postorbitals and postfrontals are evident on both sides of the skull, but the right elements are better preserved, and in articulation. The postfrontals are forked medially, lying lateral to the frontoparietal articulation, and each possesses a single distal process that clasps the postorbital posteriorly.

Each postorbital is a triradiate element that is separate from the postfrontal and forms most of the posterior margin of the orbit. This differs from previous interpretations, in which the postorbitals are considered to be fused to the postfrontals ( Cocude-Michel, 1963; Hoffstetter, 1966; Estes, 1983). The ventral process of the postorbital contacts the orbital margin of the jugal and its dorsal process contacts the postfrontal also on the orbital margin. The latter feature is uncommon amongst lizards because the dorsal process of the postorbital usually contacts the posterior margin of the distal process of the postfrontal. The posterior process of the postorbital is relatively elongate, extending more than half the length of the upper temporal fenestra, and lying dorsal to the squamosal.

The squamosals are similar to those of most other lizards in being relatively slender and each having a posteroventral process contacting the cephalic condyle of the quadrate. Both squamosals lack a dorsal process, a feature that is usually present in iguanians and teiids. A relatively short supratemporal bone is present between the posterior end of the squamosal and the parietal, being short in length and abutting against the lateral margin of the supratemporal process of the parietal.

The frontals are fused to each other, and form most of the dorsomedial margin of the orbits. They are widely expanded posteriorly, but are constricted at their midpoint and are slightly expanded anteriorly. The anterolateral processes of the frontal are elongate, but do not reach the maxillae, allowing for a short contact between the nasal and prefrontal. A much shorter anteromedial process intrudes between the nasals. In ventral view ( Fig. 2C View Figure 2 ), the frontals bear subolfactory processes (best seen posteriorly), which seem to never touch medially at any point, but matrix partially obscures their central portion. There is seemingly no midventral crest of the frontals. Posteriorly, the frontals contact the parietals along a weakly anteriorly convex suture. There is no evidence of parietal tabs of the frontal either dorsally or ventrally.

The parietals are paired elements, which is uncommon to a number of squamates, but seen in some extant geckoes [e.g. Nephrurus ( Evans, 2003) , Sphaerodactylus ( Daza et al., 2008) , and Pygopus (AMNH R 140843)] and Xantusia (AMNH R 150172 and AMNH R 150174). The pineal foramen is present close to the centre of the parietal table. The lateral margins of the parietals are emarginated and form most of the medial border of the upper temporal fenestrae. The supratemporal processes are relatively short and bear a medial excavation. A posteromedial process is absent, as also seems to be the case for the nuchal fossa. No frontal tabs are present either dorsally or ventrally.

The palatines do not contact in the midline, being widely separated from each other, and are relatively short anteroposteriorly ( Fig. 2A, B View Figure 2 ). Their anterior margins are almost entirely visible in dorsal aspect, contacting the posteroventral margin of the prefrontals, and thus the palatines make little or no contact with the frontals anterodorsomedially. Most of their posterior ends form a wide and nearly straight contact with the anteromedial processes of the pterygoids.

The pterygoids are widely separated from each other. Each pterygoid has a relatively wide, undivided anteromedial process and, based on the right side of the skull (which is better preserved than the left side), seems to be inclined medially. The transverse process of each pterygoid is short and directed anterolaterally, contacting the ectopterygoid on the left side (but see below). The quadrate process of the pterygoid is slender and extends far posteriorly, although disarticulated from the quadrates as preserved in BSPG 1937 I a, b. As the slab with the bone elements in the holotype is preserved in dorsal aspect, and the palate is not clearly visible in ventral view, it cannot be determined whether a ventral flange was present on the transverse process of each pterygoid, or whether or not teeth are absent or present on the palate .

The left ectopterygoid is visible and a small broken element lateral to the transverse process of the right pterygoid might represent its right counterpart, although it may also represent part of the anterior extension of the right jugal. The left ectopterygoid is semilunate and has no lateral process. Its posterior end lies dorsal to the pterygoids, as is the case in some geckoes (e.g. Coleonyx – AMNH R 89271) and anguids (e.g. Diploglossus – AMNH R 154690).

The supraoccipital is visible in dorsal view, bearing an ossified processus ascendens tecti synocti that contacts the parietal posteriorly. It cannot be determined whether anterolateral processes of the supraoccipital are present. A pair of posterolaterally orientated crests are present on the supraoccipital. These are similar to the ‘crests’ formed on the supraoccipital by the posterior semicircular canal in some geckoes (e.g. Hoplodactylus pacificus – MCZ R-141790 and Coleonyx variegatus – AMNH R 89271) and some other lizards of reduced size (e.g. Feylinia corrori – MCZ R-42886 and MCZ R-106990). These apparent ‘crests’ are thus considered herein as a consequence of a reduced degree of ossification due to diminutive size.

On the ventral side of the slab containing skeletal elements ( Fig. 2C View Figure 2 ), parts of the basioccipital and basisphenoid are visible. These elements are not fused to each other, and the basisphenoid is flat on its ventral surface. The basipterygoid processes are stout and have slightly expanded distal ends, but are no longer in articulation with the pterygoids.

The quadrates have been tilted posteriorly and their anterior surfaces face dorsally ( Fig. 1A, B View Figure 1 ). They each have a shallow anterior concavity, and the two articular condyles are of similar size.

Mandibles

The dentaries are visible on both sides of the slab containing the skeletal elements, but they are more readily observed in medial aspect on the ventral side of the specimen ( Fig. 2D View Figure 2 ). The Meckel’s canal is closed medially by the fusion of the subdental ridge to the ventral crest of the dentary in its anterior half, as is also observed in extant geckoes and xantusiids ( Evans, 2008). The splenial is not visible, and it cannot be determined whether it is overlain by the matrix, not preserved, or perhaps fused with the dentary as in xantusiids ( Estes, de Queiroz & Gauthier, 1988). In dorsal aspect, most of each mandible is obscured beneath the skull. However, part of the posterior end of each dentary can be observed, and each part opens medially (contrasting with the anterior and middle portions of the dentaries), where the splenial usually articulates.

The coronoids are partially visible on both sides. The left one bears a relatively elongate posteroventral process, but no posterior process. The right counterpart is mostly overlain by the pterygoid transverse process, and a broken element that might represent either the jugal or the ectopterygoid. The coronoid bears a very short anteromedial process that is directed ventrally.

Both surangulars are visible dorsally and each has a thickened dorsal border, forming the labial margin of the adductor fossa. The latter is relatively deep, but narrow, and faces dorsomedially. A short, posteriorly directed retroarticular process is present posterior to the glenoid, and lacks any lateral notching.

Dentition

The dentary teeth are not visible in the holotype, but the maxillary teeth are visible on both sides ( Fig. 3 View Figure 3 ), along with some of the premaxillary teeth. The teeth are small, narrow, conical, and appear to be unicuspid, with no differentiation between the premaxillary and maxillary regions. The tooth count cannot be established with precision for the premaxillae, but there are c. 22 teeth preserved in situ on the right maxilla, and the total tooth count in each jaw half is estimated to be at least 30. Variation in tooth height throughout the dental arcade indicates that tooth replacement was present and was still active at the time of death. The teeth are positioned medially on the jaw labial margin (or parapet), and are unfused to the jaw margins, suggesting pleurodonty.

Postcranium

The preserved vertebral column includes the axis, the remaining cervical vertebrae, the dorsals, sacrals, and anterior caudals. The number of cervicals is estimated to be seven, because the eighth presacral vertebra seems to be the first one bearing ribs long enough to reach the sternum ( Fig. 4A). Only the neural arches and neural spine of the axis are visible, with its intercentrum, and the atlas, being hidden from view. The postaxial cervicals bear very low neural spines, and most bear no ribs. The first ribs appear at the level of the sixth or seventh cervical.

Based on the identification of the eighth presacral as the first dorsal, the total number of dorsals is estimated to be 24, giving a total of 31 presacrals. This is a higher presacral vertebral count in comparison to that of most iguanians (with 24 presacrals), and a number of scincomorphans (e.g. cordylids, gerrhosaurids, lacertids, and teiioids) and gekkotans, many of which bear 26 presacrals. Yet, a slight increase in the presacral count as occurs in E. schroederi (around 30 presacrals) is observed in a variety of taxa, including xantusiids and some gekkotans, as well as scincids, xenosaurids, anguids, and varanids ( Hoffstetter & Gasc, 1969). The neural spines in the dorsal region are also small, and the last presacrals seem to lack any ribs (thus characterizing lumbar vertebrae). One striking feature of the vertebral morphology is the presence of accessory intervertebral articulations (zygosphenes and zygantra; Fig. 4B). The zygosphenes face dorsolaterally as in many iguanians, cordylids, gerrhosaurids, lacertids, and gekkotans (although they are absent in pygopodids), and are notched anteriorly, as it is also the case in all aforementioned groups, as well as in teiids. Estes (1983), following Broili (1938), reported a procoelic condition for the intervertebral articulations. However, as previously noted by Hoffstetter (1964) and Evans et al. (2004), it is not possible to determine the structure of centra with any confidence in this specimen as it is preserved in dorsal view.

Two sacral vertebrae can be identified just medial to the anterior half of the ilia and anterior to the first caudal that bears short transverse processes ( Fig. 4C). Five anterior caudals are preserved, the remaining part of the tail having been autotomized ( Figs 1 View Figure 1 , 4C). The soft tissue impression of the cartilaginous rod that developed subsequently to the loss of the tail can be seen on both slabs, and was previously illustrated under UV-light by Tischlinger & Wild (2009).

The ribs are holocephalous with circular articular surfaces. There are 21 or 22 pairs of presacral ribs. The posterior-most presacral vertebrae do not bear any ribs, therefore constituting a ‘lumbar’ region. There are no accessory tuberculi anteriorly or posteriorly to the rib heads, and the ribs lack any degree of pachyostosis. Gastralia are absent, as is the case for all other squamates, but inscriptional ribs are visible attached to the distal ends of most of the dorsal ribs, but absent from the posterior-most four.

Of the pectoral girdle, the clavicles, left scapula, and part of both coracoids are preserved ( Fig. 4A and D). The clavicles lack the dorsally located cranial curvature observed in most autarchoglossans ( Estes et al., 1988), as well as the posterior process that is seen in some scincids and cordyloids (e.g. Mabuya mabouia – AMNH 141128 and Gerrhosaurus major – AMNH 173621). It is not possible to determine whether the clavicles have a proximoventral fenestra, but they are expanded in this region.

The left scapulocoracoid exhibits no suture between the scapula and coracoid. The scapula is short, has an expanded acromion process dorsally, and lacks a scapular ray defining a separate scapular emargination dorsal to the scapulocoracoid fenestra ( Fig. 4D). The supracoracoid foramen is visible medially on the scapulocoracoid ( Fig. 4D), located at the level of, and just posterior to, the anterior coracoid ray and anterior (or primary) coracoid emargination of most lizards. The anterior (or primary) coracoid ray, delimiting the anterior coracoid emargination ventrally, is partially overlain by one of the left dorsal ribs. The left coracoid is mostly obscured beneath the vertebral column, and the right coracoid is preserved mostly as an impression in the calcareous matrix. The coracoids are relatively small when compared with most other lizards studied by us, being similar in dimension to the scapula.

Both forelimbs are preserved in articulation ( Figs 1 View Figure 1 , 4A). Details of the proximal surface of the humerus are better seen on the left element ( Fig. 4A), in which the bicipital fossa faces dorsally. The distal end of the humerus is twisted posteriorly relative to the longitudinal axis of the humeral shaft, and the state of preservation hampers identification of either an ectepicondylar or an entepicondylar foramen ( Fig. 4E, F). The ulna has a well-developed olecranon process as well as an expanded distal epiphysis. The radius is present, but its distal epiphysis has suffered some degree of weathering, so the styloid process is not preserved ( Fig. 4G, H).

The right manual carpus ( Fig. 4G) preserves the fourth distal carpal proximal to the fourth metacarpal, but no other distal carpals are distinguishable owing to poor preservation of this region. The metacarpals increase in length from the first to the fourth metacarpal, with the fifth being approximately half the length of the fourth. The phalangeal formula is 2-3-4-5-3, and the penultimate phalanges are elongated in comparison to the intermediate ones between the first and the penultimate ones ( Tables 2 and 4).

For the pelvic girdle, the pubes, ischia, and the impressions of both ilia are preserved. The pubes have relatively narrow pubic aprons, which are directed anteromedially, and lack pubic tubercles. The obturator foramen is not visible, probably because of the crushing present on the acetabular margin of the pubes. The ischia possess enlarged plates, which seemingly contributed to a long symphysial contact between both elements. Each ilium contains a long posteriorly directed blade, but it cannot be determined whether a preacetabular process ( Russell & Bauer, 2008) is present.

M1 – 5, manual digits 1 to 5; I – IV, phalanges I to IV on each corresponding digit. Missing data denoted as ‘?’. The phalangeal measurements for E. schroederi represent the mean values for the right and left manus, whereas on A. digitatellus measurements were taken from the left manus only, as the limits between phalanges are mostly nondiscernible on the right counterpart.

P1 – 5, pedal digits 1 to 5; I – IV, phalanges I to IV on each corresponding digit. Missing data denoted as ‘?’. The phalangeal measurements for E. schroederi represent the mean values for the right and left pedes (apart from digit II), whereas on A. digitatellus measurements were taken from the left pes only, as the limits between phalanges are mostly nondiscernible on the right counterpart.

M1 – 5, manual digits 1 to 5; P1 – 5, pedal digits 1 to 5; I – IV, phalanges I to IV on each corresponding digit. Missing data denoted as ‘?’.

Both hindlimbs are preserved in articulation and in posterior aspect. The shaft of the femur is almost straight, diverging from the more common sigmoidal shape observed in lizards. The femoral heads are badly crushed along with the other elements in the acetabular region, and the distal ends of the femora have only their posterior margin exposed. The tibiae and fibulae lay in close apposition to each other, and their distal epiphyses are not totally ossified.

The astragalus and calcaneum are not totally fused to each other and the tibial facet on the astragalus is almost flat, lacking a crest for articulation with the tibia ( Fig. 8 A, B, D). The distal tarsals are crushed against the metatarsals, and their morphology cannot be described in detail. However, it seems that both distal tarsals three and four are present. The first metatarsal is slightly shorter than metatarsals II – IV, the latter being subequal in length. The fifth metatarsal has its proximal head hooked medially, contacting the fourth distal tarsal, and is significantly shorter than the other metatarsals. The phalangeal formula is 2-3-4-5-4, and the penultimate phalanges are elongated compared with the intermediate ones, being between 30 and 56% longer than the antepenultimate phalanx ( Fig. 8 A, B, D; Tables 3 and 4). Phalanges II (digit III) and I and III (digit IV) of the right pes, as well as phalanges I and II (digit III), and I (digit V) of the left pes have better preserved articulatory surfaces. Distally, these phalanges have a hemispherical (convex) shape, and at least some phalanges (e.g. phalanx II, digit III, right pes) articulate with a concave proximal surface of the phalanx distally to it. This condition is different from the typical ginglymoid bicondylar articulation of the intermediate phalanges of most lizards, and would have allowed a greater range of horizontal movement. Horizontal movement also occurs in the crescentic/cup-shaped articulation in the intermediate phalanges of Gekko gecko ( Russell, 1975) , but in the latter it is probably greater than in E. schroederi owing to its expanded distal phalangeal articulations.

The unguals form claws that are slightly curved and dorsoventrally deepened ( Russell, 1975), or tall, at their bases – claw height defined as the distance measured from dorsal to ventral at the base of the claw ( Zani, 2000; Tulli et al., 2009; Crandell et al., 2014). Higher (or deeper) claws in E. schroederi , as well as in G. gecko ( Fig. 9 View Figure 9 ), bear a ventral expansion in lateral view, below the level of the contact with the penultimate phalanx, the functional consequences of which are further discussed below (see Functional morphology). In E. schroederi , average pedal claw height (measured from those claws visible in lateral aspect) equals 0.70 mm. The ratio between claw height/area of contact between the claw and the penultimate phalanx equals 1.707, indicating that the claw is about 70% higher than the area of contact with the penultimate phalanx.

LEPIDOSAURIA HAECKEL, 1866 SQUAMATA OPPEL, 1811 GEKKONOMORPHA FURBRINGER €, 1900 ARDEOSAURUS MEYER, 1860