Unusual lizard fossil from the Miocene of Nebraska and a minimum age for cnemidophorine teiids

Teiid lizards are well represented in the fossil record and are common components of modern ecosystems in North and South America. Many fossils were referred to the cnemidophorine teiid group (whiptails, racerunners and relatives), particularly from North America. However, systematic interpretations of morphological features in cnemidophorines were hampered by the historically problematic taxonomy of the clade, and the biogeography and chronology of cnemidophorine evolution in North America is poorly understood from the fossil record. Few fossil cnemidophorines were identified with an apomorphy-based diagnosis, and there are almost no fossil cnemidophorines that could be used to anchor node calibrations. Here, I describe a cnemidophorine from the Miocene Ogallala Group of Nebraska and diagnose the fossil using apomorphies. In that process, I clarify the systematic utility of several morphological features of cnemidophorine lizards. I refer the fossil to the least inclusive clade containing Aspidoscelis, Holcosus and Pholidoscelis. The most conservative minimum age of the locality of the fossil is a fission-track date of 6.3 Ma, but mammal biochronology provides a more refined age of 9.4 Ma, which can be used as a minimum age for the crown cnemidophorine clade in divergence time analyses. The fossil indicates that a cnemidophorine lineage that does not live in Nebraska today inhabited the area during the Miocene. I refrain from naming a new taxon pending discovery of additional fossil material of the lineage to which the fossil belonged.

Teiid lizards are well represented in the fossil record and are common components of modern ecosystems in North and South America. Many fossils were referred to the cnemidophorine teiid group (whiptails, racerunners and relatives), particularly from North America. However, systematic interpretations of morphological features in cnemidophorines were hampered by the historically problematic taxonomy of the clade, and the biogeography and chronology of cnemidophorine evolution in North America is poorly understood from the fossil record. Few fossil cnemidophorines were identified with an apomorphybased diagnosis, and there are almost no fossil cnemidophorines that could be used to anchor node calibrations. Here, I describe a cnemidophorine from the Miocene Ogallala Group of Nebraska and diagnose the fossil using apomorphies. In that process, I clarify the systematic utility of several morphological features of cnemidophorine lizards. I refer the fossil to the least inclusive clade containing Aspidoscelis, Holcosus and Pholidoscelis. The most conservative minimum age of the locality of the fossil is a fission-track date of 6.3 Ma, but mammal biochronology provides a more refined age of 9.4 Ma, which can be used as a minimum age for the crown cnemidophorine clade in divergence time analyses. The fossil indicates that a cnemidophorine lineage that does not live in Nebraska today inhabited the area during the Miocene. I refrain from naming a new taxon pending discovery of additional fossil material of the lineage to which the fossil belonged.
from the Cretaceous of North America and Eurasia. Barbatteiidae, another potential sister clade of extant Teiidae, is known from the Cretaceous of Europe [11,12].
Most fossil cnemidophorines from the Neogene of North America are isolated and fragmented dentaries and maxillae. The identifications of those fossils were hindered by problematic taxonomy of cnemidophorines [13,22,26], and few of those fossils were diagnosed with apomorphies. Divergence time analyses indicate that most crown cnemidophorines originated during the Oligocene and Miocene [33,34], and many teiid skeletal elements are distinct from those of other squamates [13,35] and so should be diagnosable through an apomorphy-based approach [36]. Thus, the paucity of unambiguously identified fossil cnemidophorines from the Neogene of North America is striking, as is the absence of cnemidophorines from the late Palaeogene. The evolutionary history of cnemidophorines is not well understood from the Neogene fossil record of North America.
I describe a nearly complete and unusual fossil dentary of a cnemidophorine lizard from the Miocene Ogallala Group of Cherry County, Nebraska. The fossil is a marginal tooth-bearing element like most other known teiid fossils from the Neogene, but it is largely complete and preserves morphologies that allow for an apomorphy-based referral to the crown cnemidophorine clade. The fossil provides new information on the temporal and biogeographic history of cnemidophorines in North America. A single fossil referred to cf. Cnemidophorus was previously reported from Cherry County in the Valentine Formation (figure 1), but the fossil was lost before publication [ formations [37]. Thus, the maximum age of the fossil is 13.5 ± 0.01 Ma based on interpolation of the age of the Hurlbut Ash between dated horizons [38] or 13.55 ± 0.09 Ma based on 40 Ar/ 39 Ar dates for the Hurlbut Ash [39], and the minimum age is 6.6 ± 0.3 Ma based on fission-track of glass from near the top of the Ash Hollow Formation [37,40].
The age of the fossil is further constrained by biochronology of North American mammals in the Valentine and Ash Hollow formations. Although there are no mammals that can be attributed to the exact site where YPM VP 4707 was collected, fossil mammals from other known localities in the Valentine and Ash Hollow formations in Cherry County are characteristic of the Barstovian and Clarendonian North American land mammal ages (NALMAs) [41]. The Barstovian precedes the Clarendonian. There are no documented localities from the Ogallala Group anywhere in Cherry County that contain taxa characteristic of the younger Hemphillian NALMA, which follows the Clarendonian and is considered to extend to the early Pliocene [40,41]. Key taxa from Clarendonian localities in Cherry County include Eubelodon (a gomphothere) and Barbourofelis (a feliform carnivoran) [41]. The Clarendonian NALMA is currently considered to extend to 9.4 Ma [42]. That age could change if the dates bracketing the extinction of taxa at the end of the Clarendonian (e.g. Eucastor, Aelurodon) are revised, or if specimens of those taxa are found that post-date 9.4 Ma [40,42]. Conservatively, the minimum age of the fossil is 6.3 Ma, but based on mammal biochronology, a minimum age of 9.4 Ma is assigned.

Terminology, taxonomy and specimens examined
Osteological terminology follows Evans [43] unless otherwise noted. Taxonomy of cnemidophorines follows Tucker et al. [33,44]. I accept the phylogenetic hypotheses of Tucker et al. [33,44] in my diagnoses, most importantly, that Aspidoscelis, Holcosus and Pholidoscelis form a clade to the exclusion of other extant teiid lizards, and that Aspidoscelis and Holcosus are sister taxa. A complete list of specimens examined is in electronic supplementary material, file S1. I examined at least one specimen of all cnemidophorines besides Contomastix. I also examined Callopistes and all tupinambines besides Crocodilurus.

Description
YPM VP 4707 is a mostly complete right dentary that preserves the tooth row, four teeth, the symphysis and some of the coronoid and angular processes ( figure 2a,b). The posterior portion of the dentary is royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 200317 significantly taller than the anterior portion of the dentary, and the dentary dramatically tapers in height anteriorly. The Meckelian groove is long and open along the entire dentary. Anteriorly, the groove faces ventrally and is restricted by the suprameckelian lip (sensu [50]). The posterior portion of the groove has a deep mediolateral dimension, at least in part due to a lateral expansion of the dentary at the coronoid facet (figure 2c). The Meckelian groove leaves space for the splenial to extend far anteriorly, almost to the symphysial facet (figure 2a,d). The suprameckelian lip has a relatively short dorsoventral dimension posteriorly, but markedly increases in height anteriorly. There is an inset articulation facet for the splenial along the ventral margin of the posterior half of the suprameckelian lip. Lateral to the suprameckelian lip, there is a posterior exposure of the intramandibular septum (the intramandibular lamella; [51,52]), which would separate the anteromedial process of the coronoid from the surangular. The intramandibular septum is dorsoventrally short, especially anteriorly (figure 2d). The posterior portion of the septum reaches anteriorly to the fourth tooth (numbered on figure 2a,b), and the anterior portion, a slight ridge in ventral view, reaches the second tooth. The subdental gutter is relatively deep anteriorly and shallower posteriorly. Some of the tooth replacement pits invade the subdental gutter (figure 2c). The subdental gutter extends across the entire tooth row and has a moderately broad lingual dimension, although it narrows anteriorly near the symphysis.
The tooth row has a length of 11.37 mm. There are 19 apparent tooth positions, but it is difficult to distinguish the mesialmost tooth positions, so that count should be viewed as tentative. Three distal teeth and one mesial tooth are preserved. The bases of the three distal teeth are adjacent but do not contact. Dentition is pleurodont and is heterodont in terms of both tooth size and cusp morphology. The distalmost tooth has a mesiodistally expanded base, and the two teeth mesial to that tooth taper less distinctly from the base to the crown. The crowns of the three distal teeth are asymmetrically and mesiodistally bicuspid, and the mesial crown is smaller than the main crown. The main (distal) crown of the third tooth slopes ventrodistally instead of ventrally to create a slight distal shoulder. The mesial crowns of all three biscuspid teeth are apically worn. The mesial tooth is unicuspid and is substantially smaller than the distal teeth. A replacement pit is present on the second tooth, and pits are present within dental tissues at other tooth positions that lack teeth. The replacement pits are deep and almost circular. There are substantial but not excessive deposits of basal cementum at the bases of the teeth and at tooth positions lacking teeth, and the cementum does not fill the subdental gutter. The labial face of the dentary is moderately convex and has rugose sculpturing across much of its surface, but rugosities are absent immediately ventral to the three distal teeth and on the anteriormost portion of the bone (figure 2b). The coronoid facet is subtriangular and extends anteriorly to the fourth tooth, and ventrally to the ventrolateral extent of the posterolateral expansion of the dentary. The facet is deep, moderately textured, and has a marked ventral boundary. There are nine distinct nutrient (mental) foramina that extend posteriorly to the second tooth, although most of the foramina are concentrated anteriorly. The posterior portion of the dentary is less complete than the rest of the fossil, and so the posterior extent of the coronoid and angular processes is uncertain, as is the presence of a separate surangular process.

Diagnosis
YPM VP 4707 is referred to the least inclusive clade containing Aspidoscelis, Holcosus and Pholidoscelis based on the presence or absence of the following morphological features; the clade for which the feature is hypothesized to be an apomorphy is listed in parentheses: pleurodont tooth implantation (Lepidosauria), substantial deposits of basal cementum (Teiidae), an elongate Meckelian groove providing space for a hypertrophied splenial (Teiidae), large subcircular tooth replacement pits (Teiidae), an open Meckelian groove (Teiidae), Meckelian groove restricted anteriorly (Teiinae), the absence of a large incision between the coronoid and surangular processes (Teiinae), sculpturing on the labial surface of the dentary both anteriorly and posteriorly (least inclusive clade containing Aspidoscelis, Holcosus and Pholidoscelis), dentary tapers in height posteriorly to anteriorly ( present in several cnemidophorines, see below), and bicuspid distal teeth ( present in several cnemidophorines, see below).
Asymmetrically bicuspid distal teeth and an open Meckelian groove are present in Teiidae, but are also present in Lacertidae, which is closely related to Teiidae [34]. Gymnophthalmidae+Alopoglossidae, the sister clade of Teiidae among extant squamates [34], can also have bicuspid teeth, but the Meckelian groove is generally closed and fused [56,57]. Amphisbaenia, the sister taxon of Lacertidae, possesses several unique derived features, including unicuspid teeth, a short dentary, a low tooth count (less than 10 teeth), and dentition described as acrodont, subacrodont or subpleurodont [58,59]; the dentition of teiids has also been described as subpleurodont [60]. YPM VP 4707 has substantial deposits of cementum at its tooth bases, large subcircular tooth replacement pits, and an elongate Meckelian groove providing space for a hypertrophied splenial, all three of which are apomorphies of Teiidae to the exclusion of other squamates ( [4,35]; see below).
Borioteiioids can have all three dentary apomorphies commonly used to diagnose teiid fossils [4] that are present in YPM VP 4707. The fossil is excluded from Polyglyphanodontia because it lacks distal teeth with transversely oriented cusps, from Chamopsiidae because it lacks a marked symphysial boss and possesses an intramandibular septum [61], from Prototeius because the distal teeth are bicuspid and are not massive or blunt-cusped and the Meckelian groove is anteriorly restricted [51], and from members of Macrocephalosauridae, which have flared, multicuspid teeth, sometimes with obliquely oriented cusps [62].
Members of Barbatteiidae were also reported to possess the three dentary apomorphies of Teiidae, and the clade was hypothesized to be the sister group of Teiidae sensu stricto [11,12]. YPM VP 4707 is differentiated from documented dentaries of barbatteiids because it tapers in height anteriorly, it lacks an incision between the coronoid and angular processes, and it possesses mesiodistally expanded distal tooth bases, labial sculpturing, and a better-developed subdental gutter in terms of both dorsoventral depth and lingual width [12]. YPM VP 4707 is part of total clade Teiidae sensu stricto (excluding Borioteiioidea and Barbatteiidae).
The anterior restriction of the Meckelian groove excludes the fossil from Tupinambinae and Callopistes, whose members have a completely unrestricted and medially facing Meckelian groove royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 200317 [51,63], but is a feature shared with Teiinae ( figure 3). Tupinambines and Callopistes also differ from the fossil by possessing a large incision in between the coronoid and angular processes of the dentary. Tupinambines and Callopistes are further distinguished from YPM VP 4707 because they have more extensive deposits of basal cementum, which fills the subdental gutter and creates mesial and distal walls between teeth (subthecodont implanation) [53,63]. The teiines Dicrodon and Teius also have considerably more basal cementum relative to cnemidophorines and YPM VP 4707 (Fig 3G; [63]). The fossil lacks the transverse tooth cusps of Dicrodon (figure 5g) and Teius [60,64,65].
Labial sculpturing was previously reported to occur in some adult individuals of Ameiva and Cnemidophorus without reference to any particular species [26] and before those genera were split up. Among examined extant teiids (see electronic supplementary material, file S1), only Aspidoscelis and Holcosus have sculpturing on the labial surface of the dentary (figure 4a-c)  royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 200317 the anterolabial surface of the bone was previously documented in Pholidoscelis [17]. I interpret labial sculpturing as an apomorphy of the least inclusive clade containing Aspidoscelis, Holcosus and Pholidoscelis.
Most specimens of Aspidoscelis lack labial sculpturing. In both Aspidoscelis and Holcosus, sculpturing occurs often, but not universally, on larger and more robust specimens that were probably skeletally mature adults, as was observed by Norell [26]. Holcosus festivus MVZ 79608 is a larger specimen that lacks sculpturing. Sculpturing is present on all examined specimens of Aspidoscelis deppii, some specimens of Aspidoscelis tigris   ; figure 4a). The sculpturing of YPM VP 4707 is most similar in its rugose texture and its broad extent to examined specimens of A. tigris. The unusually complex phylogeny of Aspidoscelis [66,67] and intraspecific and intra-clade variability in dentary sculpturing makes the evolution of that morphology difficult to interpret within Aspidoscelis. Some specimens of A. tigris have highly reduced sculpturing, including larger specimens (e.g. TxVP M-8630). Aspidoscelis hyperythrus, a smaller species which may be closely related to A. deppii [67], lacks sculpturing altogether.
Tooth cusp morphology is known to vary ontogenetically and interspecifically in teiids [68]. All examined specimens of Holcosus, Ameiva, Cnemidophorus, Kentropyx and Medopheos have exclusively tricuspid distal dentary teeth, and most have tricuspid median teeth as well. Aurivela, Contomastix and Ameivula were also reported to have tricuspid distal teeth [22,60,69], although the one specimen of Aurivela examined here has bicuspid distal teeth. At least one large and presumably adult specimen of Ameiva is known to have bicuspid distal teeth [17]. In Holcosus, Ameiva, Cnemidophorus, Kentropyx and Medopheos, the transition from unicuspid teeth to tricuspid teeth occurs in the anterior portion of the tooth row, often abruptly. Many Aspidoscelis have bicuspid teeth throughout the tooth row, but some species, such as A. uniparens and A. sonorae, have mostly tricuspid teeth. Intraspecific and ontogenetic variation in tooth morphology is particularly well documented in Pholidoscelis [17]. Small specimens have unicuspid mesial teeth, bicuspid median teeth and bicuspid or tricuspid distal teeth, while in larger specimens the teeth generally lose their cusps and become bicuspid or unicuspid, including the distal teeth [17][18][19]. Some specimens progress to have molarized crushing teeth [17,19]. My observations agree with those of previous studies. Some smaller Pholidoscelis that I examined had mostly tricuspid distal teeth (e.g. Pholidoscelis chrysolaemus UF 99352), but in larger specimens, distal teeth were all bicuspid (figures 3d and 4d) or are mostly bicuspid except for the teeth at the last two or three tooth positions, which were tricuspid (e.g. Pholidoscelis chrysolaemus UF 99646).
The third tooth of YPM VP 4707 is sloped ventrodistally to create a slight dorsal shoulder on the distal face of the tooth. That morphology also occurs in some specimens of Aspidoscelis that have bicuspid distal teeth (e.g. A. deppii, figures 3a and 4a). I interpret the third tooth of YPM VP 4707 as representing a shoulder instead of a third cusp, but recognize that there is some ambiguity. Additionally, the three distalmost tooth positions of YPM VP 4707 are not preserved and could have been tricuspid.
The presence of bicuspid teeth in the distal portion of the tooth row of YPM VP 4707 is most consistent with either Aspidoscelis or Pholidoscelis. The propensity to have bicuspid distal teeth in adult specimens is probably a derived feature of the clade containing Aspidoscelis, Holcosus and Pholidoscelis, with an apparent reversal in Holcosus. However, that feature cannot be used by itself to identify fossil Aspidoscelis and Pholidoscelis, because bicuspid distal teeth occur in at least a few specimens of other cnemidophorines.
In all examined teiids, the posterior portion of the dentary is taller than the anterior portion, but YPM VP 4707 is unusual in that the anterior portion of the dentary strongly and abruptly tapers in height. I observed more moderate tapering of the dentary in C. lemniscatus (figures 3f and 4f ), A. tigris (e.g. TxVP M-15034, M-15035) and H. undulatus (figures 3c and 4c). The anteriormost portion of the dentary tapers in D. guttulatum (figures 3g and 4g). Tapering was also previously reported in Pholidoscelis [17,18], and strong tapering more comparable with the fossil was illustrated for one large specimen of P. griswoldi [19]. A tapering dentary appears to at least be derived in Teiinae; the anterior height of the dentary decreases uniformly in Tupinambinae and Callopistes.
Posterolateral expansion of the dentary is most pronounced in Holcosus among extant teiids (figure 5c). Correspondingly, the ventral border of the coronoid facet is most distinct in Holcosus. The facet has a less defined ventral border in Aspidoscelis, Pholidoscelis and other cnemidophorines, and expansion of the dentary is less exaggerated in those taxa as well. In all cnemidophorines, lateral expansion is most exaggerated in larger specimens. Expansion of the dentary is comparable between YPM VP 4707 and larger specimens of Holcosus.
The presence of a posterior extension of the intramandibular septum (the intramandibular lamella) was reported to diagnose polyglyphanodontids to the exclusion of extant teiids [51]. I observed that morphology in several extant teiids, including Aspidoscelis and Holcosus (figure 3a-c).

Discussion
To my knowledge, YPM VP 4707 is the oldest published record of a crown cnemidophorine from North America. The fossil was identified using apomorphies grounded in a phylogenetic hypothesis based on analyses of targeted-sequence capture data [33,44]. The minimum age of the total clade containing Aspidoscelis, Holcosus and Pholidoscelis is at least 6.3 Ma given a fission-track age estimate of glass from the Ash Hollow Formation [37], but is 9.4 Ma based on the currently accepted temporal extent of the Clarendonian NALMA. The fossil can be used in divergence time analyses as a minimum age of crown cnemidophorines and could be further used to calibrate the minimum age of the split between the clade containing Aspidoscelis, Holcosus and Pholidoscelis, and the clade containing Ameiva, Cnemidophorus, Kentropyx and Medopheos. Few fossils were previously considered reliable for bracketing minimum clade ages in divergence time analyses of Teiidae, and only one cnemidophorine of indeterminate phylogenetic affinity was used recently [33,63]. YPM VP 4707 is useful for future studies seeking to temporally calibrate the evolutionary history of teiids and cnemidophorines in particular.
The resurrection of the genera Aspidoscelis, Holcosus and Pholidoscelis for clades previously accommodated in Cnemidophorus and Ameiva [67,69,70] and the clarification of the phylogeny of Teiinae [33,44] helped elucidate the systematic significance of two morphologies in cnemidophorines. Labial sculpturing of the dentary was reported in adult specimens of Ameiva and Cnemidophorus by Norell [26]. Sculpturing is restricted to three clades formerly in Ameiva and Cnemidophorus, the genera royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 200317 Holcosus, Pholidoscelis and Aspidoscelis. Labial dentary sculpturing is hypothesized to be an apomorphy of the clade containing those three genera. Similarly, bicuspid distal tooth crowns were previously reported in certain groups of Ameiva and Cnemidophorus [22,60], but Pholidoscelis and Aspidoscelis are the only genera previously assigned to Ameiva and Cnemidophorus that consistently have bicuspid distal teeth. The presence of bicuspid tooth crowns cannot be used by itself to identify Aspidoscelis or Pholidoscelis, however, because bicuspid distal teeth are present in rare specimens of Ameiva [17] and Aurivela.
Fossil cnemidophorines reported from the early Miocene of Florida [21,22] are consistent with divergence time hypotheses indicating a Miocene or Oligocene origin of crown cnemidophorine clades [33]. However, those fossils were not [22] or could not [21] be referred to the crown clade with an apomorphy-based approach, and so the early evolution of crown cnemidophorines has been poorly understood from the fossil record. YPM VP 4707 confirms the presence of crown cnemidophorines in the middle-late Miocene and indicates that the entire clade occupied at least part of its modern range in North America at that time. It is noteworthy that the fossil is a part of the clade containing Aspidoscelis, Holcosus and Pholidoscelis. Those three genera either inhabit Nebraska in the modern biota (Aspidoscelis) or are geographically more proximate to the area than are other cnemidophorines (e.g. Kentropyx, Ameiva, Ameivula) that are now mostly found in South America.
Aspidoscelis sexlineatus is the only teiid that occurs in Nebraska or immediately adjacent to Nebraska in the modern biota [71]. Because YPM VP 4707 is not clearly part of crown Aspidoscelis, much less the clade containing A. sexlineatus, it appears that a different lineage of crown cnemidophorine was found in the area at least as recently as the middle-late Miocene. Aspidoscelis tigris, the A. tesselatus complex, the A. neotesselatus complex and Aspidoscelis velox are currently found in southern and western Colorado [71]. Pholidoscelis is currently restricted to islands in the Caribbean, and Holcosus is found in Central and South America and in Mexico as far north as central Tamaulipas [70,72]. More fossils from across North America are needed to explore the historical biogeography of cnemidophorine lizards and to determine whether modern tropical or island clades like Holcosus and Pholidoscelis were once found farther north on the mainland or on the mainland at all, respectively.
The present survey of extant teiids indicates that the overall morphotype of YPM VP 4707 is distinctive. However, the fossil cannot be definitively excluded from total or crown Aspidoscelis, Pholidoscelis or Holcosus, so for now I refrain from establishing a new taxon; new fossil material may reveal that YPM VP 4707 is referable to one of those clades. Additional fossils of the lineage to which YPM VP 4707 belonged are needed to refine its systematic position, to determine whether a new taxon is warranted, and to further investigate the biogeographic history of cnemidophorine lizards during the Neogene in North America.
Ethics. YPM VP 4707 is reposited at the Yale Peabody Museum. Data accessibility. Electronic supplementary material is available in S1. Competing interests. The author declares no competing interests. Funding. Funding was provided by a Lundelius award in vertebrate paleontology from the Jackson School of Geosciences at the University of Texas at Austin.