Asagaolacerta tricuspidens, Evans & Matsumoto, 2015

Evans, Susan E. & Matsumoto, Ryoko, 2015, An assemblage of lizards from the Early Cretaceous of Japan, Palaeontologia Electronica (London, England) 52 (4), pp. 1-36 : 11-18

Asagaolacerta tricuspidens sp. nov.

Figures 7–10

zoobank.org/ B80126CA-645A-44A8-A5D7-556512A4EF4B

Etymology. For the tricuspid teeth

Holotype. SBEI 1566 An association of skull, jaws and postcranial bones of a small lizard with facetted tricuspid teeth ( Figure 7 ).

Type locality and horizon. The Kaseki-Kabe, Shiramine, Hakusan city, Ishikawa Prefecture, Japan. Early Cretaceous Kuwajima Formation, Tetori Group.

Referred material. Possibly SBEI 1621, a right maxilla (not figured).

Differential diagnosis. A small lizard characterised by the presence of sharply pointed tricuspid teeth with a central cusp and flanking anterior and posterior cuspules; tooth shaft cylindrical rather than bulbous, and tooth crown and shaft of similar width (not apically flared nor strongly labiolingually flattened). These dental features distinguish Asagaolacerta from other Kuwajima Formation lizards and from described Late Jurassic and Early Cretaceous taxa from Europe (e.g., Hoffstetter, 1967; Evans and Searle, 2002), North America (Nydam, 2002; Nydam and Cifelli, 2002a, 2002b), and China (Evans and Wang, 2005, 2010, 2012; Evans et al., 2005). Asagaolacerta tricuspidens resembles North American Late Cretaceous borioteiioids like Obamadon gracilis , Tripennaculus sp. , Socognathus brachyodon and Chamops segnis ( Longrich et al., 2012) in having tricuspid teeth, but differs in lacking any swelling of the tooth bases, and in having tooth crowns that are proportionally smaller in relation to the tooth bases. Asagaolacerta tricuspidens also resembles Mongolian Late Cretaceous borioteiioid lizards (sensu Nydam et al., 2007) Altanteius facilis , Mongolochamops reshetovi , Pyramicephalosaurus cherminicus , and Tchingisaurus multivagus ( Alifanov, 2000b; Gao and Norell, 2000) in having tricuspid teeth, and in having a slender jugal with long dorsal ramus forming most of posterior orbital border, but differs in having lower tooth count and step-like jugo-maxillary suture; further differs from Pyramicephalosaurus and Tchingisaurus in lacking any flaring or labiolingual flattening of tooth crowns ( Alifanov, 2000b; Gao and Norell, 2000); resembles Late Cretaceous "mongolochamopine" Cyclurasia multidentata ( Alifanov, 2000b) in jugo-maxillary suture shape but differs in having fewer, larger teeth. Asagaolacerta tricuspidens also differs from many Asian borioteiioids in lacking a hypertrophied splenial and build-up of cementum around tooth bases.

Material. SBEI 1566 ( Figure 7 ) is a small block in which cranial and postcranial elements are superimposed, possibly as an oral pellet, but the specimen appears to represent a single individual (on the basis of non-repetition of parts and consistent size). Of the skull, the elements exposed include the left maxilla, the right and left dentaries and some postdentary bones, the left and right jugal, a partial left frontal, a palatine, and a quadrate, as well as some undetermined elements. The postcranial bones include several vertebrae, parts of the left pelvis, both femora, and scattered phalanges. The pectoral girdle and forelimbs seem to be missing. X-ray microtomography of the block revealed no other significant elements and no associated osteoderms. In comparison with modern skeletal material, the size of the jaws, pelvis, and limb elements are matched by a specimen of the extant teiid lizard Aspidoscelis tigris of 90 mm SVL (total length 290 mm). The body proportions of Asagaolacerta tricuspidens , at least for those elements preserved, seem to have been broadly similar to the proportions of the living species.

Description. The anterior end of the left maxilla is exposed in labial view ( Figure 8.1 ). The narial margin is oblique (unlike the deeper concavity of Kuroyuriella mikikoi ). As preserved, the first eight teeth are slender and gradually increase in length, but the crowns are damaged, and it is not possible to determine whether there was a change from unicuspid to tricuspid within this series.

Of the other clearly identifiable skull elements, the right and left jugals lie together near the lower jaws ( Figure 9.1–3 ). They are roughly L-shaped, but with an oblique angle between the ventral and postorbital rami. The left bone preserves a nearly complete ventral ramus and part of the postorbital ramus, whereas the right preserves most of the postorbital ramus and the ventral ramus, but the latter is overlain by a left palatine. The left ventral ramus bears an elongate lateral facet for the maxilla and was clearly partially overlapped by that bone in its anterior half, leaving a thin strip of the jugal along the ventral margin of the orbit. This arrangement would have given the jugal-maxillary suture a step-like structure in lateral view. It also suggests that the suborbital ramus of the maxilla extended only to mid-orbit. The jugal lacks a posterior spur or tubercle, but the thin bone is drawn into a slight angle. The postorbital ramus is slender and longer than the ventral ramus. The facet for the postorbital (or postorbitofrontal) is confined to the dorsal tip, suggesting that the jugal formed most of the posterior orbital margin and that the postorbital itself was not large. This interpretation is supported by a slight recess on the posterodorsal edge of the jugal that may have received the anterior tip of the squamosal, excluding the postorbital from the margin of the lower temporal fenestra.

Part of a left frontal is preserved in dorsal view at the edge of the association ( Figure 7.2 ), but shows only that the frontals were paired and weakly sculptured with fine lines. A partial left palatine overlaps the ventral ramus of the right jugal. It is preserved in dorsal view and shows only that a foramen perforated the base of the maxillary process, as it does in most lizards. A robust curved bone close to the anterior end of the maxilla is identified as the right quadrate ( Figure 9.4–5 ). It has a narrow conch and a posteriorly extended head, but no obvious notch or pit for the squamosal.

Both dentaries are preserved in association, the right in lingual view, the left in dorsolingual view ( Figures 7 , 8.2–3 ). The right shows 12 tooth positions with space for at least three further small teeth in the symphysial region. The tooth row is ~ 7.4 mm long, and the dentary was originally 8.5– 9.5 mm long. The posterior part of each dentary is obscured, and the tooth row may have extended further. The subdental ridge is deeper anteriorly than posteriorly, and there is little development of a subdental gutter. The anterior tip (left bone) is narrow, with a smooth internal surface and no trace of facets, pits, or grooves in the symphysial region. On the right, the splenial is in association, filling the posterior part of the Meckelian fossa but not extending to the symphysis ( Figure 8.3 ). The fossa remains mainly medial, rather than becoming ventral anteriorly, but does appear to be restricted in height. Parts of the post-dentary series are in association but are poorly exposed and difficult to prepare further due to the superimposition of other elements. A bifurcate bone overlies the posterior part of the right dentary, in roughly the position for a coronoid although this identification remains tentative. Other robust bone fragments further posteriorly may be parts of the postdentary series.

The dentition is well exposed on the right dentary ( Figure 8.3–4 ). The teeth are pleurodont and parallel-sided. The more anterior teeth are narrower and seem to have unicuspid crowns. The six exposed posterior teeth are larger and have a distinctive morphology. They are cylindrical with the crown of similar width to the tooth shaft. The crown is facetted with a strong sharp triangular central cusp and smaller lateral flanking cusps in a clearly tricuspid arrangement ( Figure 8.4 ). Replacement pits are present in the base of several teeth showing replacement was active, lingual, and vertical.

Several vertebrae are scattered across the block ( Figure 10.1 ). They are procoelous with slightly ovoid cotyles (long axis transverse), have low neural spines, and show a weak zygosphenezygantral system that resembles the condition in some modern lacertids, where the anterior articular surface is V-shaped in anterior view, with the zygapophysial and zygosphenoidal facets facing one another at an angle.

There is no trace of the forelimbs or pectoral girdle, but the central part of the block is a mass of overlapping elements that defies further preparation (or micro-CT visualisation) and it is possible that parts of the forelimb skeleton lie within this bone mass. The hind limbs are represented by part of the left pelvis, both femora, and some phalanges (some of which could also be from the manus). The left ilium is exposed in medial view ( Figure 10.2 ), with the iliac blade broken across and superimposed on the underlying limb bones. It is difficult to reconstruct iliac shape with any accuracy but as preserved the blade has a shape similar to the iliac blade of unspecialised tetrapodal lizards. There is no evidence of a strong anterior tubercle. Depressions on the medial surface suggest the sacroiliac joint was close to the ilio-femoral one, rather than displaced distally as in some reduced limbed taxa. Adjacent to the ilium is a broader plate-like bone (not figured) which may be part of the ischium. Deep to these bones, and overlain by other fragmentary (and mostly unidentifiable) elements, there is an almost complete left femur ( Figure 7 ), 11.3 mm as preserved but probably 11.5-12 mm in life as the proximal end is damaged and the distal condyle slightly distorted. The internal trochanter is present deep to the proximal femoral head. Only the distal part of the right femur is visible, but a micro-CT scan shows that the remainder of the bone lies deeper in the rock. By comparison with modern lizards, and taking into account the size of vertebrae and pelvic bones, these femora are similar in size and shape to the femora of living limbed lizards.

Additional specimens that may pertain to Asagaolacerta tricuspidens . SBEI 190 (Li1) and SBEI 193 (Li4) are blocks bearing postcranial remains of similar size and proportions to those on the holotype of Asagaolacerta tricuspidens .

SBEI 190 (Li1) ( Figure 11.1–2 ) is a partial postcranial association bearing sacral vertebrae, posterior dorsal vertebrae, the left pelvis, and a left femur (12 mm long), tibia (7 mm), and fibula on one block and more anterior dorsal vertebrae and ribs on a second block (not figured). The dorsal vertebrae are procoelous with rounded condyles, short square neural spines, weak zygosphenes, and

PALAEO- ELECTRONICA.ORG large oblique rib synapophyses. The sacrum is damaged. The pelvis is preserved in medial view and its components are co-ossified. The ilium is of similar shape to the ilium on the holotype of Asagaolacerta tricuspidens ; the sacral rib facet is positioned just above the level of the acetabulum. The pubis is only partially preserved, but was clearly tapering rather than broad, and the ischium was flask-shaped. The femur is robust and has a thick circular shaft that has only a slight sigmoid curvature. The proximal and distal ends are fully ossified, with a rounded proximal femoral head, a well-developed tuber-like greater trochanter, and a distinct intertrochanteric fossa. The bone is relatively stouter than the femur of the holotype of Asagaolacerta tricuspidens (overall length/distal width ~ 4.7 compared to 6.2 in Asagaolacerta tricuspidens ), although only slightly longer. The tibia is shorter than the femur (~ 58%) but again robust with a wider proximal end and a flange-like cnemial crest that is separated from the proximal end. The fibula was probably slight longer than the tibia but its distal end is broken. There are no associated osteoderms.

SBEI 193 (Li4) ( Figure 11.3–4 ) is a second partial postcranial association, preserved in ventral view. The sacral vertebrae are fused and the conjoined distal ends of the sacral ribs enclose a foramen sacrale on each side. Rather unusually, the first sacral, though incomplete, appears to be less robust than the second. Behind them, the first caudal has a slight keel and robust transverse processes, each of which bears a deep linear groove proximally. The left pelvis is complete with the components conjoined. The right ilium is preserved in medial view. Its elongate blade lies at roughly 45° degrees to the long axis of the acetabular region, with the sacral rib facet lying just above the level of the acetabulum. There is a slight anterodorsal expansion rather than a tubercle. Both femora (~ 8.7 mm) are preserved to one side of the block, although they are partially covered by disarticulated dorsal vertebrae. They are smaller than femora on the holotype of Asagaolacerta tricuspidens but they are otherwise similar in proportions (overall length/distal width ~6.0 in both SBEI 193 and SBEI 1566 ). Again, there are no associated osteoderms .

On the basis of the femoral and, to a lesser degree, iliac shape, one or both specimens may be attributable to Asagaolacerta tricuspidens but without more complete material they cannot be referred with confidence and characters from these specimens were not included in the diagnosis or phylogenetic analyses.

Affinities. Among extant lizards, tricuspid teeth are found mainly in iguanians, lacertids, and teiioids, although they may also occur in other groups. In the Mesozoic, tricuspidy is rarer (Nydam, 2002), but again occurs most commonly in taxa referred to Iguania or Borioteiioidea (sensu Nydam et al., 2007, =Polyglyphanodontia of Gauthier et al., 2012 and Longrich et al., 2012). One exception is the Late Cretaceous Mongolian Parmeosaurus scutatus ( Gao and Norell, 2000) , which our analysis grouped with scincoids. Of described Jurassic and Early Cretaceous lizard taxa, only Ptilotodon wilsoni from the Aptian-Albian Antlers Formation of Texas ( Nydam and Cifelli, 2002b) approaches tricuspidy in having small anterior and posterior expansions that form shoulders on either side of a much larger central cusp. The teeth of Asagaolacerta tricuspidens differ in having distinct cuspules that are almost the same height as the central cusp. In the Late Cretaceous, tricuspidy became more common. It has been recorded in several Campanian-Maastrichtian genera from North America (Nydam, 2002; Longrich et al., 2012) that were once considered teiioid but have more recently (Nydam et al., 2007; Longrich et al., 2012) been classified as borioteiioid - notably Chamops segnis and Leptochamops denticulatus ( Estes, 1964) , Meniscognathus altmani ( Nydam and Voci, 2007) , Socognathus brachyodon ( Gao and Fox, 1996; Longrich et al., 2012), Tripennaculus eatoni ( Nydam and Voci, 2007) , and Obamadon gracilis ( Longrich et al., 2012) , The first four of these have been grouped (with others) into the family Chamopsiidae (e.g., Nydam et al., 2010; Longrich et al., 2012), but they did not always emerge as a monophyletic unit in our analyses ( Figure 12 ). Tricuspidy is also found in a range of borioteiioid taxa from Late Cretaceous deposits in Mongolia ( Alifanov, 2000b), including Altanteius facilis , Cyclurasia multidentata , Mongolochamops reshetovi , Piramicephalosaurus cherminicus , and Tchingisaurus multivagus . A borioteiioid attribution for Asagaolacerta would be consistent with the heterodonty and tricuspidy (e.g., Nydam and Cifelli, 2002a, 2002b), the paired unsculptured frontals, and the long postorbital ramus of the jugal. However, Asagaolacerta tricuspidens differs from many more derived borioteiioids in lacking a hypertrophied splenial and a heavy deposition of cementum around the tooth bases, characters that have been cited as diagnostic of both teiioids and borioteiioids (e.g., Denton and O'Neill, 1995; Gao and Fox, 1996; Nydam et al., 2007).

As for Kuroyuriella mikikoi , we coded Asagaolacerta tricuspidens into the matrix of Gauthier et al. (2012) and Longrich et al. (2012) (94/622 characters, 85% missing data). In all analyses, weighted and unweighted, constrained or unconstrained, ordered or unordered, Asagoalacerta was consistently placed on the stem of Borioteiioidea ( Figure 12 ), although the position of the clade within the squamate tree varied with different protocols (e.g., Figure 5 ), as did the resolution and positions of in-group taxa ( Figure 12 ). Although the strict consensus of the exhaustive analysis (unconstrained, unordered, equally weighted) placed Asagaolacerta in an unresolved position in relation to the fragmentary North American taxa ( Figure 12.1 ), it is important to note that Asagaolacerta did not group with any one of these taxa in the more resolved individual trees from analyses using different protocols (e.g., Figures 12.2–12.4 ), nor when the borioteiioid taxa were run on their own ( Figure 12.5 ).

SQUAMATA Oppel, 1811

Family indet.

Genus HAKUSEPS gen. nov.

zoobank.org/ 6AD065FC-1D2E-488D-ABBB-5C27E306057D

Type species. Hakuseps imberis

Etymology. From Hakusan, white mountain, the name of the city of which Shiramine village forms a part and of the regional volcanic peak, and seps (L), variably used for lizard.

Diagnosis. As for type and only species

Figures

FIGURE 5. Phylogenetic position of Kuroyuriella mikikoi gen. et sp. nov. and Asagaolacerta tricuspidens gen. et sp. nov. in squamate trees, using the morphological data matrix of Gauthier et al. (2012), extended by Longrich et al. (2012). 1, Strict consensus of 1000 trees using unordered equally weighted characters, and run using TNT with 'minisearch.run'. Major clades are condensed. 1Note that Iguania was not monophyletic and comprised four smaller clades whose position was unresolved in relation to Borioteiioidea. Support values at nodes are Bremer/Jacknife/Symmetric resampling; 2, one of three trees (identical at this level) with clades condensed, run using the same matrix as in Figure 5.1 and analysed with TNT (with sectorial search, ratchet [20 iterations], and tree fusion all activated), but with the molecular tree of Wiens et al. (2010) and Pyron et al. (2013) providing the backbone constraint.

FIGURE 7. Asagaolacerta tricuspidens gen. et sp. nov., holotype, SBEI 1566. 1, the main association with the offset second bone group in the smaller image; 2, explanatory drawing of the same. For abbreviations, see Material and Methods.

FIGURE 8. Asagaolacerta tricuspidens gen. et sp. nov., holotype, SBEI 1566. 1, Left partial maxilla in labial view; 2, left dentary in lingual view; 3, right dentary in lingual view; 4, details of teeth on the right dentary (enlarged from 3). For abbreviations, see Material and Methods.

FIGURE 9. Asagaolacerta tricuspidens gen. et sp. nov., holotype, SBEI 1566, details of cranial bones. 1, left and right jugals; 2–3, explanatory drawings of right and left jugal respectively; 4–5, right quadrate in 4, lateral, and 5, anterior views; 6, biradiate bone, possibly coronoid. For abbreviations, see Material and Methods.

FIGURE 10. Asagaolacerta tricuspidens gen. et sp. nov., holotype, SBEI 1566, postcranial elements. 1, presacral vertebrae; 2, left ilium and distal head of right femur. For abbreviations, see Material and Methods.

FIGURE 11. Associated postcranial specimens that may be referable to Asagaolacerta gen. nov. 1–2, SBEI 190; 3–4, SBEI 193. For abbreviations, see Material and Methods.

FIGURE 12. The phylogenetic position of Asagaolacerta tricuspidens gen. et sp. nov. tested using different analytical protocols within TNT. 1, detail of Strict Consensus of 1000 trees using the protocol that yielded the tree in Figure 5.1, node support values Bremer/Jacknife/Symmetric sampling; 2, one of three trees resulting from an analysis using with the molecular backbone constraint tree, but no character ordering or weighting; 3, one of 58 trees from an analysis run as in (2), but with character ordering as per Gauthier et al. (2012) and Longrich et al. (2012), and Implied Weighting (k=7); 4, one of 34 trees from an analysis run as in (3), but without the molecular backbone constraint; 5, 70% MRT of 19 trees resulting from an analysis (characters ordered but equally weighted, no constraints) run using only the boreoteiioid taxa, with Gekko gecko as the outgroup taxon. The 70%MRT is presented rather than the unresolved Strict Consensus to show that Asagaolacerta tricuspidens is usually (88% of trees) placed in the basal position (see text for further discussion). Abbreviation: poly, polyglyphanodont (as used in Longrich et al., 2012).