Cryptovaranoides microlanius, Whiteside & Chambi-Trowell & Benton, 2022
|
publication ID |
https://doi.org/10.1126/sciadv.abq8274 |
|
DOI |
https://doi.org/10.5281/zenodo.7401234 |
|
persistent identifier |
https://treatment.plazi.org/id/33429E46-FFB4-3E11-1142-FCFA8A87CACE |
|
treatment provided by |
Felipe |
|
scientific name |
Cryptovaranoides microlanius |
| status |
sp. nov. |
Type species— C. microlanius sp. nov. (by monotypy)
Etymology—The genus name Cryptovaranoides is derived from the Greek word crypto, “hidden,” new Latin varan (from Arabic Waran meaning “dragon-lizard”), and the Greek suffix oides meaning “like.” The generic name is thus “hidden, lizard-like” (animal), referring to the fact that this fossil has remained unrecognized in a collection for nearly 70 years and that it was a small lizard-like animal (skull length maximum about 30 mm), living on karst limestone, perhaps hiding in the grykes (crevices). The species name microlanius derives from the Greek “micro” = small and Latin “lanius” meaning butcher. The recurved blade-like teeth demonstrate that the animal was a small efficient predator that could capture and immobilize prey quickly.
Holotype — NHMUK PV R36822 , a partly articulated anterior skeleton and associated partial skull and lower jaws preserved on and in the red matrix of a small block of conglomeratic limestone. The left mandible and a partial skull ( Fig. 1 View Fig , A and B), including both maxillae and associated bones, are close to the articulated partial skeleton. The left side of the skull comprising the maxilla, prefrontal, jugal, and lacrimal in near-life positions ( Fig. 1 View Fig , B and E) has twisted away from the right side. The partial skeleton ( Fig. 1C View Fig ) includes the right mandible with associated quadrate, jugal, postorbital, otoccipital, basicranial bones, articulated cervical and anterior dorsal vertebrae, as well as pectoral girdle and forelimb elements. Nearby are further articulated dorsal vertebrae, followed by a basioccipital ( Figs. 1 View Fig , A and D, and 7B) from a larger individual of Cryptovaranoides . Bone preservation ( Fig. 1 View Fig , A and E) is superb and detail in the CT scans ( Fig. 1 View Fig , B to D and F) is often excellent, allowing us to identify numerous characters of modern squamates.
Referred fossils—Referred isolated bones (all NHMUK PV) are as follows: left maxilla R 36999; right maxilla R 37279; fragment of right maxilla R 37280; left dentary R 37001; fragment of left dentary R 37282; right dentary R 37281; left coronoid R 37273; left frontal R 37274; left quadrate R 37604; braincase R 37377; premaxilla R 37378; sacral vertebrae R 37275; cervical vertebra R 37276; dorsal vertebra R 37277.
Locality and age—Triassic fissure deposit in Carboniferous Limestone, Cromhall Quarry, Tortworth, Gloucestershire, United Kingdom. Age is Rhaetian, probably late Rhaetian (equivalent of Cotham Member, Lilstock Formation), 201.6 to 201.8 Ma ( 15).
Diagnosis—Characterized by the following apomorphies: (i) maxillary and dentary with frequently replaced conical, pleurodont recurved teeth ( Figs. 1 View Fig , A to C, E, and F, 2, C and D, 3, A to G, and 4A) displaying labio-lingually compressed apices (upper part of the crown) and a prominent sharp carina on mesial and distal edges ( Fig. 3 View Fig , F and G). The mesial compressed region is somewhat more pronounced than the distal, and overall cross section of compressed region is oculiform-lanceolate. The mesial region of mid and posterior
maxillary teeth develops a pronounced blade-like flange. The lingual surface of the dentary and maxillary teeth has 2 to 10 vertical striae (higher number on younger, smaller unworn teeth), including a few branching striae (observed on dentary teeth), on compressed upper part of crown; (ii) tripartite maxilla ( Figs. 1B View Fig , 2A View Fig , and 3 View Fig , A to C, and fig. S1, A and F) with high facial process that curves medially encompassing a prominent nasal fossa bordered dorsally with a distinct ridge; also, there are long anterior and posterior processes; (iii) lateral faces of dentary and maxilla are perforated by many foramina ( Fig. 3 View Fig , A and D; c 20+ for the large maxillae) for branches of the
mandibular and maxillary nerve (CN5), respectively, with a range of sizes; about 20% of maxillary foramina have maximum length greater than observable lateral tooth width; mandibular foramina usually form double row anteriorly; (iv) premaxillae fused in larger individuals ( Fig. 4 View Fig , B and C) with four teeth on one premaxilla and three on the other; (v) the fused premaxillae dentition includes a median tooth derived from the premaxilla with the greater number of teeth; (vi) the sphenoid contains a vidian canal ( Fig. 5 View Fig , E and F) that is enclosed as it passes over the basipterygoid processes and an internal carotid foramen that opens into a deep hypophysial fossa
( Fig. 5B View Fig ) bounded posteriorly by the dorsum sellae; (vii) pronounced dorsum sella roofed by a prominent well-developed anteriorly directed crista sellaris that partially covers the fossa ( Fig. 5 View Fig , A and B); (viii) abducens canal present ( Fig. 5 View Fig , E and I); (ix) metotic fissure subdivided with recessus scala tympani developed anteriorly (floored by a prominent occipital recess) bounded posteriorly by the crista tuberalis ( Fig. 4 View Fig , D to H); (x) long thin squamosal with dorsal and posterior processes ( Figs. 1B View Fig , 2A View Fig , and 6 View Fig , B and C, and fig. S1, A, D, and F), the latter articulated as a “peg-in-socket” with the notch on quadrate; (xi) quadrate with notch for squamosal on lateral side of cephalic condyle ( Figs. 1 View Fig , C and F, and 6, D to F); (xii) jugal: posterior aspect very short and truncated with straight edge; short dorsal process and longer anterior process ( Fig. 1F View Fig ); (xiii) postorbital forms a slender postorbital bar with the dorsal process of the jugal ( Figs. 1F View Fig , 2A View Fig , and 7C View Fig , and fig. S1, A and F); (xiv) coronoid is relatively large ( Figs. 1E View Fig , 2C View Fig , and 7 View Fig , I and J, and fig. S1, B and C); about 20% of lower jaw length and more than 80% of jaw height (measured dorsoventrally from top of coronoid process to ventral base of mandible) in medial view and with a prominent posterior recess, presumably for the bodenaponeurosis; (xv) vomer ( Fig. 7A View Fig ) is in anterior sutural contact with maxilla ( Fig. 3C View Fig shows vomer facet on the maxilla), separating the vomeronasal opening from the choana; (xvi) ventrally the vomer has two prominent postero-anterior longitudinal toothed ridges that somewhat converge anteriorly ( Fig. 7A View Fig ).
Skull
The reconstructed skull ( Fig. 2 View Fig , A to E, and fig. S1) is based largely on the holotype, augmented by isolated bones of larger individuals. The holotype is a juvenile, based on its small size (skull length, 14 mm) and unfused cervical neural arches and centra. Isolated bones suggest
that an adult skull could be 30 mm long and up to 16 mm wide. The reconstruction started with the palate ( Fig. 2D View Fig ) because we have all elements, with small, conical teeth, on the palatine, on longitudinal ridges of the vomer, and on the pterygoid anterior ramus; teeth are absent from the sphenoid, ectopterygoid, and most of the pterygoid. The proportions of the skull viewed laterally ( Fig. 2A View Fig ) and dorsally ( Fig. 2E View Fig ) follow the ventral view ( Fig. 2D View Fig ), and elements in color, except the frontals and braincase, are based on the three-dimensional (3D) CT scans. These indicate broad skull roof bones and evident streptostyly ( Fig. 2A View Fig ), with a narrow, conch-shaped quadrate, a thin rod-shaped squamosal,andthelowertemporalopeningiswidewithasmall,truncated jugal.
The palatal reconstruction ( Fig. 2D View Fig and fig. S1E) is the most reliable as we have all the bones (but sometimes only the left or right element) from the palate represented in the CT scan images. The palatine, vomer, and anterior ramus of the pterygoid all bear small conical teeth in the anterior and mid-palate. Teeth are absent from the sphenoid and ectopterygoid, and there are no signs of them on the mid- or posterior parts of the pterygoid and its flange. There is a fossa for the base of the epipterygoid on the dorsal surface of the mid-region of the pterygoid (the fossa columellae; fig. S1H), but the CT scan is not distinct enough to ascertain its shape. Therefore, neither do we know the shape of the base of the epipterygoid. There is a possible epipterygoid (labeled “?” in Fig. 1C View Fig ). However, this is a speculative identification as that element is in two parts, with one part about the right size for an epipterygoid in this skull. If the element was originally a complete unit (rather than from two different bones), then this “epipterygoid” must be some other unidentified bone.
The sphenoid forms the larger element of the basicranial region that constitutes a relatively large feature of the skull in ventral view. The other bone is the basioccipital with a maximum width, across the basioccipital tubera, of more than 42% the width between the lateral margins of the quadrate condyles. These proportions are similar to many lizards [calculated from skulls in ( 16, 17)] such as Varanus salvator (overall between 33 and 80% for lizards and usually about 37 to 50% but can reach 80% in, e.g., the skink Typhlosaurus ) compared to the smaller proportions found in rhynchocephalians such as Sphenodon (c. 30%), Planocephalosaurus [calculated at c. 27% from ( 18)], Clevosaurus [calculated at c. 25% from ( 19)], Diphydontosaurus [calculated at c. 20% from ( 20)], and Gephyrosaurus [calculated at c. 22% from ( 21)]. The subtemporal fenestra is relatively large and indicates that the adductor musculature was substantial.
The sphenoid is a thin bone, formed from the fused endochondral basisphenoid and the dermal parasphenoid ( 17); this was difficult to process from the CT scan because it is highly fractured, crushed in many places, with some missing bone, and sediment has filled the interstices. This is almost certainly a consequence, at least in large part, of the juvenility of the animal in life and therefore incomplete ossification. However, with additional close-up photography, the major features are shown in Fig. 5 View Fig (A to I). The base of the parasphenoid process lies between the trabeculae cranii ( Fig. 5D View Fig ), which diverge posteriorly forming the floor of the deep hypophysial fossa ( Fig. 5 View Fig , A and B), either side of which is a large foramen for the internal carotid artery ( Fig. 5B View Fig ). The hypophysial fossa is bounded posteriorly by a pronounced dorsum sella, and this is roofed by a prominent well-developed anteriorly directed crista sellaris that partially covers the fossa ( Fig. 5 View Fig , A and B). The sphenoid contains the closed vidian canal with the anterior opening at the base of the basipterygoid process ( Fig. 5 View Fig , E and F). The vidian canal, incorporating the internal carotid artery and the base of the palatine artery, enclosed by the sphenoid as it passes over the basipterygoid process is a squamate apomorphy ( 22). The position of the posterior opening is more difficult to discern. There is a foramen positioned on the lateral “wings” of the sphenoid ( Fig. 5G View Fig ) somewhat behind the posterior base of the basipterygoid process that may be toward the posterior entry. However, because of repeated fractures and some crushing and missing bone, it is not possible to follow this foramen continuously toward the front of the sphenoid, although there are enclosed foramina in the anterior region of the bone ( Fig. 5 View Fig , H and I). It is possible that the enclosed vidian canal is quite short and the posterior entry may be as suggested in Fig. 5C View Fig . The anterior exit of the abducens canal ( Fig. 5E View Fig ) is confirmed from the CT slice across the anterior of the sphenoid ( Fig. 5I View Fig ). The posterior opening of the abducens canal is more problematic to identify but is suggested to be as shown in Fig. 5B View Fig .
The vomers form a large portion of the palate and are bounded laterally by extensive choanae. We have searched carefully for facets on the posterior side of the premaxillae or the anterior margin of the vomer ( Fig. 7A View Fig ), but they appear to be lacking in the juvenile and an isolated vomer is absent in the larger bones of the collection. The anterior region of the vomer imaged from the CT scan has a small, truncated process positioned medially, which probably formed the connection to the premaxilla. The large, fused premaxillae ( Fig. 4 View Fig , B and C) is important as it is composed of a right bone with four teeth and a left bone with three teeth. The fusion of the two bones results in the most mesial tooth of the right premaxilla forming a median tooth; this is a synapomorphy of modern Squamata . The asymmetry in number of tooth positions of the two premaxillae is also found in the juvenile holotype.
Although not discernible on the juvenile holotype, Cryptovaranoides has incisive processes on a larger specimen of the premaxillae ( Fig. 4B View Fig ), and the connection with the vomer was probably ligamentous. An important aspect of the palate is the connection between the maxilla and the vomer in the most anterior region; the vomerine facet, which is ridged, is distinct above the subdental shelf of the maxilla ( Fig. 3C View Fig ), and the CT scan image of the vomer ( Fig. 7A View Fig ) has a clear laterally positioned and matching maxillary facet. The vomer facet on the maxilla demonstrate that the maxilla-vomer connection was strong and not simply a vomeromaxillary overlap. These positions result in a small gap on both lateral sides of vomer where it would have contacted the premaxillae. As there appears to be no palatal posterior process of the premaxillae, which would (in squamates) often border the vomeronasal fenestra on the medial side, we suggest that these small gaps are the external openings of the vomeronasal openings. There is a small emargination and a groove on the lateral side of the vomer at the anterior end, which we suggest is the vomeronasal notch ( Figs. 2D View Fig and 7A View Fig ). A vomeronasal opening bordered by the vomer posteromedially and the maxilla laterally is common in extant squamates, but an anterior premaxilla margin is rare. However, just such an arrangement of bones occurs in the amphisbaenians Amphisbaena alba , which has a similarly small external vomeronasal opening, and Diplometopon ( 23). We can also compare Cryptovaranoides with, for example, Varanus ( 24, 25) where the vomeronasal opening lies anterior to the vomer/maxillary suture and is therefore separate from the choana. No other constriction (which might suggest an alternative position for the vomeronasal fenestra) in the choana is observed in Cryptovaranoides . We have therefore concluded that, like Varanus , Cryptovaranoides was neochoanate.
The hypothesized position of the vomeronasal opening is strengthened by the recognition of a right septomaxilla ( Fig. 1B View Fig ) that lies near the separated premaxillae and anterior of the disarticulated vomer and maxilla of the holotype. In life, the septomaxillae lay above and posterior to the external vomeronasal openings. The anterior chamber formed by the septum dividing the ventral space would have housed the Jacobson’s organ for chemoreception. This can be observed in the dorsal reconstruction ( Fig. 2E View Fig and fig. S1D).
We have reconstructed the extent of the nasals ( Fig. 2F View Fig ) based on the facets of the maxillae, indistinct features on the postero-dorsal processes of the premaxillae, and (isolated) frontals. However, the dorsal reconstruction is less certain than the ventral view as CT scan images of the holotype do not show the nasal, frontal, postfrontal, or parietal. The parietal and nasal are shown as single elements analogously to Varanus but may well have been paired to match the frontals. On the basis of the lack of observable anterior nasal facets on the maxilla, we speculate that the anterior part of the olfactory chambers may have been unprotected by the nasal(s) and, if so, presumably would have been covered by thick toughened soft tissue instead.
There are isolated frontals (e.g., NHMUK PV R37274; Fig. 7 View Fig , F to H), and we are reasonably sure that these are referrable to Cryptovaranoides because the lateral and ventral facets in the anterior and mid-region are complementary to prefrontal facets including those of the long curved posterior process. Also, like the CT scans of the prefrontal, the frontal bone is highly ornamented. The frontal bone ( Fig. 7 View Fig , F to H)
has facets for the nasal anteriorly and prefrontal anterolaterally; these extend posteriorly along the lateral side. There are also posterior facets that lie laterally and ventrally for the (presumed) postfrontal [similar to those in Elgaria ; figure 1.79 B of ( 17)] and a small underlap for the parietal posteriorly ( Fig. 7H View Fig ). The olfactory canal running along the ventral anterior half of the frontal is pronounced, but the crista cranii is relatively low as it is in teiids and iguanids ( 17) but less well developed in comparison to many squamates. The cartilaginous planum supraseptale would have attached to the crista cranii ( 17). In similar manner to iguanids, there is a median raised area at the end of the olfactory canal but, unlike that group, it is distinctively ridged in Cryptovaranoides . Palpebrals may have been absent or present, but we have no clear evidence either way.
The discovery of a good braincase NHMUK PV R37377 ( Fig.4 View Fig , D to I), albeit missing the mid-region of the fused exoccipitals and some lateral bone, has allowed a reasonable understanding of this region of the skull with the paroccipital processes and the recognition of an otoccipital (fusion of exoccipital and opisthotic). On the right-hand side, the vagus foramen is inferred to lie above two foramina ( Fig. 4 View Fig , F and G) that we identify as hypoglossal foramina, which are positioned on the lower medial region of the broken exoccipital near its contact with the basioccipital. The hypoglossal foramina lie posteriorly to the medial opening of the recessus scala tympani (mrst). The crista tuberalis, which separates the anterior cavity of the metotic fissure ( 10, 17), is present and protects the lateral opening of the recessus scala tympani (lrst; Fig. 4H View Fig ). There was presumably a lateral vagus opening posterior to the crista tuberalis and anterior to the missing mid-region of the exoccipital, but the fossil provides no evidence. The occipital recess that floors the rst is observed clearly and on the separate basioccipital ( Fig. 7B View Fig ) with the same morphology, confirming that the braincase is from Cryptovaranoides . We confirm that the metotic fissure is divided as in squamates, based on the distinct and separate rst separated by the crista tuberalis ( 10), the (probable) vagus foramen likely positioned between the mrst and the hypoglossal foramina. Furthermore, where the metotic fissure is undivided as in Sphenodon ( 17), there is no occipital recess in the basioccipital, nor in basal sphenodontians and rhynchocephalians such as Diphydontosaurus and Gephyrosaurus , but it is pronounced in Cryptovaranoides ( Fig. 4 View Fig , F and I) as in many squamates. This braincase specimen has a partial suture line between the exoccipital part of the otoccipital and the basioccipital ( Fig. 4 View Fig , D and E), but the isolated basioccipital ( Fig. 7B View Fig ) and others in the collection indicate that fusion of the elements was uncommon. Fractures in the fossil (fig. S1D) probably indicate the suture with the supraoccipital. The supraoccipital has an ascending process ( Fig. 4D View Fig ), not extensive but clearly present.
There are difficulties in reconstructing the mid- to posterior region of the skull as it is uncertain whether the postorbital connected directly to the squamosal, frontal, and/or the parietal. If the facet for the jugal on the postorbital has been correctly recognized ( Fig. 7C View Fig ), then the bone size (relative to others) would suggest not. Moreover, posterolateral facets, partly positioned ventrally, on the frontal ( Fig. 7F View Fig ) indicate that a postfrontal was probably present. Also, there is a dorsal facet on the postorbital that indicates a suture to a bone, which is probably a postfrontal, missing from the fossil. A postfrontal is therefore likely to have connected the postorbital to the squamosal, but the evidence is not conclusive. The reconstruction of the position of the squamosal is based on its relative proportion to other preserved bones. It is probable that the long thin squamosal formed the lateral boundary of the upper temporal fenestra, and the supratemporal (also based on proportional size to other bones) would have formed part of the medial wall as in varanids. The dorsal process of the squamosal ( Fig. 6 View Fig , B and C) would have formed a buttress for the posterolateral region of the skull ( 11).
The same degree of uncertainty about the upper temporal region pertains to the lateral view. However, the two identical jugals and one postorbital ( Fig. 1 View Fig , C, E, and F) make it clear that the postorbital bar was very slender but probably complete (fig. S1, A and F). We have constructed the lower part of the quadrate based on isolated quadrates and the proportionate position of the cotyles within the lower jaw. The mid-region and snout reconstruction are more certain. We have been able to produce the same reconstruction with the CT scan images and another with the best-preserved isolated maxilla, NHMUK PV R36999 ( Fig. 3A View Fig ). The overall effect is of a skull with a long low profile ( Fig. 2A View Fig and fig. S1, A and F). The proportions of skull bones in the holotype and from isolated bones suggest that the antorbital region is slightly longer than the postorbital, and we have reconstructed the skull based on this likelihood. However, without a parietal, there is some uncertainty, and the lengths of the antorbital and postorbital regions may have been more equal. The lower temporal fenestra is open and the jugal has only a small truncated posterior process. The lacrimal is similar in proportion (8% of the antorbital region in lateral view) to some living squamates such as V. salvator [( 17), figure 1.89]. The maxilla has a profusion of foramina (for the maxillary nerve CN5) positioned all over the lateral surface of the bone ( Fig. 3A View Fig and fig. S1A). This may indicate a highly sensitive integument over the region, which would have enabled the animal to sense prey in dark areas such as leaf litter or possibly indicates a nocturnal habit. However, the profuse foramina may perhaps rather relate to an amphibious mode of life as in the extant crocodile lizard Shinisaurus crocodilurus .
Without an identified nasal, we cannot be sure whether the reconstruction of the external naris is accurate, although the lack of an anterior nasal facet suggests that it would have been similar in life. The septomaxilla is exposed in our reconstruction and resembles that of varanids.
Mandible
The mandible ( Fig. 2 View Fig , B and C) is well preserved on both sides ( Fig. 1 View Fig , C and F), showing all elements and numerous teeth. The lower jaw reconstruction ( Fig. 2 View Fig , B and C, and fig. S1, B and C) is based on the left-side element ( Fig. 1 View Fig , A, B, and E) and the 80% complete right-side specimen ( Figs. 1 View Fig , A, C, and F, and 6, D, G, and H) that can be viewed in situ medially and from the lateral ( Fig. 1F View Fig ), dorsal, and ventral sides in a CT scan image. One isolated dentary, NHMUK PV R37281, constitutes the other element in the reconstructions.
The holotype clearly shows where the surangular, angular, and prearticular meet the coronoid, splenial, and dentary. It also shows the relative positions of the splenial, angular, and dentary on the mid-anterior part of the mandible ( Figs. 1E View Fig and 3E View Fig ), confirmed by facets on the isolated bone NHMUK PV R37001. In addition, the isolated coronoid ( Fig. 7 View Fig , I and J) shows excellently preserved facets for the dentary, surangular, and prearticular. The presence of a splenial facet on another isolated coronoid guided us in our reconstruction of that part of the medial view. Our positioning of the posterior parts of the surangular, angular, and the prearticular is guided by features on the scans ( Fig. 1 View Fig , C and F) and is therefore an
informed interpretation, but margins of the bones are not certain. The glenoid is also somewhat unclear, so the shape and structure of the articular surfaces is not as well defined as we would have liked. The condyles on the holotype quadrate are also relatively poorly preserved, but these are intact on some isolated elements to provide good evidence for the reconstructions.
The retroarticular is long and narrow, and there is a rudimentary medial (angular) process ( Figs. 1C View Fig and 6 View Fig , G and H) that lies medial and slightly posterior to the articular cotyles. The dorsal edge of the anterior lateral part of the retroarticular is presumably the ventral site for the tympanic membrane.
Viewed laterally and medially, the coronoid eminence is formed from the coronoid ( Fig. 2B View Fig and fig. S1, B and C) only, rather than the dentary andsurangularasinrhynchocephalians.Inlateralview,the Cryptovaranoides coronoid forms about 41% of the dorso-ventral width of the mandible (measured from the upper apex of the coronoid to the ventral margin of the mandible); corresponding figures [calculated from ( 16, 17)] in squamates are Lialis (31%), Diploglossus (39%), Gonatodes (39%), and juvenile Gallotia (41%), rising to the near 70% of the gecko Pachydactylus . In Huehuecuetzpalli ( 26), the sister taxon to the crown-Squamata, the calculated figure is about 27%, or in basal rhynchocephalians, e.g., Gephyrosaurus [calculated from ( 21)], it is 10%. The Cryptovaranoides coronoid bone, in medial view (fig. S1C), occupies about 84% of the dorso-ventral width of the mandible, as in the anguid Diploglossus (79%) or the gecko Gonatodes (80%), and it is higher than Huehuecuetzpalli (74%) and much greater than Gephyrosaurus (53%) and Diphydontosaurus [64%; calculated from ( 20)]. The ventral region of the coronoid lacks the arched shape found in many extant lizards, e.g., Diploglossus or the lacertoid Gallotia , but is similar to the base of the coronoid of the pygopodid Lialis or indeed in most snakes. The morphology of the Cryptovaranoides coronoid and its aspect on the lower jaw is therefore within the range of modern Squamates.
Laterally, the isolated dentary bones show many mental foramina ( Fig. 3D View Fig ), forming a lengthwise double row in places, particularly in the mid-dentary, and some of the foramina are large relative to the width of the jaw. This complements the multiple foramina on the maxilla and suggests that the animal had a highly sensitive integument covering the snout and the anterior of the lower jaw. It is plausible that this sensitivity was an adaptation for a semi-aquatic life to sense prey such as small arthropods and fish in the water. The “scincomorph” Saurillodon marmorensis ( 5) has a similar profusion of foramina on the anterior of the dentary.
Dentition
Tooth implantation is pleurodont, within a pronounced subdental gutter and lingual wall ( Figs. 1E View Fig , 3 View Fig , B, C, and E to G, and 4A) with teeth firmly attached to the higher labial wall. Teeth are conical, slightly recurved, with a sharp unserrated carina.
The teeth show an iguanid-type regular replacement, with examples of a new smaller tooth emplaced lingually in the resorption pits of the older teeth ( Fig. 4A View Fig ). The marginal tooth-bearing bones of Cryptovaranoides show evidence of frequent tooth replacement, especially in the anterior half of the jaws. In about 25% of the isolated specimens of maxillae and dentaries, there are examples of infrequent replacement in posterior teeth, but the substantial bone of attachment (or cementum) present in basal rhynchocephalians such as Gephyrosaurus ( 21) or, particularly, Deltadectes ( 27) is not observed.
The middle and posterior teeth on both the dentary and maxilla are slightly larger than the anterior, and anterior teeth can be less recurved ( Figs. 1E View Fig and 3 View Fig , D and E). However, the dentition is essentially homodont along the dentary, contrasting with the heterodonty found, for example, in Pseudopus ( 28). The teeth are cone-shaped with a circular cross section toward the tooth base but prominently recurved and labio-lingually compressed in the upper apical half of the crown ( Fig. 3G View Fig ). There is greater variation in the maxillary dentition, with larger and more flanged teeth in mid-region ( Fig. 3F View Fig ). The teeth, unlike those of some varanids, are not serrated but have a sharp carina (resulting in a formidably sharp blade) running mesial-distally across the labio-lingually compressed crown apex of each tooth ( Fig. 3 View Fig , F and G). However, the apices of the mid and posterior maxillary teeth differ from the dentary teeth by the presence of a distinctive mesial flange ( Fig. 3F View Fig ), which produces a razor-blade edge. In profile, the dentary and anterior maxillary teeth closely resemble the recurved teeth of the anguimorphs Anniella , Pseudopus , Heloderma , Shinisaurus , or scincids such as Typhlosaurus and have some similarity (although the curvature is not as extreme) to those of boid snakes. The mid and posterior mesially flanged maxillary teeth of Cryptovaranoides resemble dentary teeth of Ophisaurus ventralis , displaying a mesial flange of the apex [( 28), figure 12E], and may indicate a similar diet of arthropods, small vertebrates, and their eggs.
Postcranial skeleton
The cervical vertebrae are well preserved in articulated series ( Figs. 1C View Fig and 7L View Fig ). The neural arches and pleurocentra are not fused, suggesting that the animal was a juvenile, in contrast with isolated adult cervical vertebrae in the collection where these elements are fused. Ribs are visible on some of the vertebrae. The scan shows probable bicapitate (= bicipital or dichocephalous) ( 29) ribs on CV6 and CV7 ( Fig. 7N View Fig ), and an additional image indicates another on CV5. Although modern Squamata are frequently considered to have the synapomorphy only unicapitate (= single headed) ribs [( 30), Ch. 86], bicapitate cervical ribs are present in Varanus [( 29, 31), figure 1A]; Hoffstetter and Gasc [( 29), p. 252] state that “a double articulation can often be seen in Varanus between the first cervical ribs and the corresponding vertebrae.” There are gaps between the vertebrae indicating that intercentra were present (but displaced in the specimen) on CV3 and posteriorly. Some images of bones on the scans are identified as intercentra ( Fig. 1C View Fig ).
There is a further series of dorsal vertebrae ( Fig. 1D View Fig ), but although they may well be from the same individual, they are not connected to the articulated anterior skeletal elements that reach the 11th vertebra in the holotype ( Fig. 1C View Fig ). There are individual vertebrae in the collection, some certainly and others probably belonging to Cryptovaranoides , which we can recognize by similarities in morphology. These bones have not only wider lateral processes and prominent zygapophyses but also rudimentary zygantrum and zygosphene accessory articulations.
There are a few isolated sacral vertebrae, of which the most complete is shown ( Fig. 7M View Fig ). The bones are characterized by broad pleurapophyses fused to the vertebrae. The sacrum consists of two vertebrae fused at the centra, the lateral ends of the pleurapophyses and probably at the neural arch. There is a large foramen sacrale between the pleurapophyses.
Both scapulae and coracoids ( Figs. 1C View Fig and 7N View Fig ) are present in the holotype. On the right coracoid, there are two large openings, one
possibly the primary coracoid fenestra ( Fig. 7 View Fig , K and N) but that may be an artifact of the CT scan or a damaged area. The left scapula is shown in Fig. 1C View Fig . We have reconstructed the scapulocoracoid (fig. S1G) based on the holotype specimens, and this displays the anterior emargination between the scapula and coracoid. If the possible primary coracoid fenestra is excluded, our reconstruction bears most similarity to Heloderma among extant squamates. Left and right clavicles are imaged from the CT scan ( Figs. 1C View Fig and 7N View Fig ). The clavicle is thicker ventrally where it lies in a facet of the interclavicle. That distal part was orientated latero-dorsally, but the upper part was free of the interclavicle, thinner and twisted so that it curved dorsally and slightly posteriorly. In the holotype of Cryptovaranoides , the gracile interclavicle is essentially T-shaped ( Fig. 1C View Fig ) rather than anchor-shaped. There is no indication of an anterior process. The posterior process has been broken postmortem, so only a small part is present in the fossil.
Left and right humeri are represented in the holotype ( Fig. 1 View Fig , C and D). The distal end is notable for its large entepicondylar and ectepicondylar foramina with openings on the anterior and posterior sides. The left humerus of the holotype shows a proximal end bearing the deltopectoral crest. From CT scan images, the radius is slightly longer than the ulna ( Figs. 1D View Fig and 7L View Fig ), and there is a small protuberance distally, possibly a styloid process, on the left bone. There is no evidence of an epiphysis or a patella ulnaris in the ulna, perhaps because the holotype is a juvenile.
The carpals are somewhat spread out on the right wrist and include the ulnare, a probable radiale, and possibly the pisiform. There are small bones that might be distal carpals, some displaced to the distal ends of the metacarpals, but they cannot be identified further. All five metacarpals are seen beyond the extremities of the right forelimb epipodials ( Fig. 7L View Fig ), but metacarpal 3 is damaged and lacks the middle section. Metacarpals 2 and 4 and probably 3 are similar in length, but they are about 20% longer than metacarpal 5, which is slightly longer than metacarpal 1. Metacarpal 1 has an expanded proximal head, which suggests, as for squamates, that distal carpal 1 has fused with the proximal head of the metacarpal. No clear elements beyond the distal ends of the metacarpals have been recovered in the CT scan, so we are unable to make any substantial comment on the phalanges.
Phylogenetic analysis
A serious problem at present in any phylogenetic analysis of fossil Lepidosauria or Squamata is that the molecular and morphological analyses differ in fundamental topology ( 7, 10, 11, 22, 32, 33). For example, in morphological analyses, the Iguania are frequently placed in a basal position and Gekkonomorpha nested higher in the tree, whereas phylogenomic trees all agree that Gekkonomorpha is the basal clade, and Iguania are derived. This affects clade definitions and makes placement of fossil taxa difficult. For example, the Anguimorpha from morphological analyses generally includes skinks and amphisbaenians, whereas Scincomorpha and Amphisbaenia are more basally placed in the molecular tree. The snakes are placed basally within the Toxicofera in the phylogenomic tree ( 34) but are more crownward in the combined molecular and morphological tree ( 32).
Therefore, we follow current practice ( 10, 11, 32) in performing phylogenetic analyses with a combined molecular-morphological data matrix. As in previous analyses [e.g., ( 34, 35)], the tree ( Fig. 8 View Fig ) shows the maximum clade credibility (MCC) tree from molecular data, with Gekkonomorpha basal and the Toxicofera (Anguiformes, Iguania, and Serpentes) a major clade. Our analysis identifies Rhynchocephalia as a clade and positions some basal taxa as either Pan-squamata ( Sophineta , Megachirella , and Marmoretta ) and others ( Taytalura and Gephyrosauru s) as Lepidosauria. However, our phylogeny of Squamata ( Fig. 8 View Fig ) conforms with recent analyses ( 10, 11, 32, 34, 35), in discovering Gekkonomorpha basal, and the major clades Unidentata, Toxicofera, Skinkomorpha, Laterata, Anguiformes, Iguania, and Serpentes. Unexpectedly, we failed to retrieve complete resolution of clades between Unidentata and Toxicofera, and this relates to uncertainty in the placement of several Jurassic and Cretaceous fossil lizards (e.g., Ardeosaurus , Globaura , and Tepexisaurus ).
The new Triassic lizard, C. microlanius , is nested deep in the phylogeny of crown Squamata as successively Squamata, Unidentata, Toxicofera, Anguiformes, Neoanguimorpha, and Anguioidea ( 34), which suggests that it is related to modern xenosaurids, anguids, and helodermatids ( Fig. 8 View Fig ). Because this is a fossil taxon, we lack molecular data; however, morphology-only phylogenetic analyses confirm that C. microlanius is a crown squamate and an anguimorph (further details in the Supplementary Materials).
Character analysis
We step through numerous characters that confirm that Cryptovaranoides belongs to each of these clades (further details in the Supplementary Materials). Cryptovaranoides shares 10 synapomorphies with the clade Pan-squamata: Angular does not extend posteriorly to reach articular condyle; coronoid eminence formed by coronoid bone only, not in the combined process of the dentary and prominent dorsal expansion of the surangular; quadratojugal not present as a separate element; short overlap in quadrate-pterygoid contact; scapulocoracoid emargination/fenestration present; jugal posterior process absent; coronoid postero-medial process present; jugal closely approaches level of prefrontal below orbit; jugal entirely exposed above labial margin of maxilla; coronoid reaches lateral side of surangular.
Unequivocal squamate apomorphies in Cryptovaranoides are fused premaxillae and premaxillary unfused median tooth. A further 13 squamate synapomorphies are as follows: cephalic head of mobile quadrate with laterally positioned notch, peg-in-notch articulation with rod-shaped squamosal, vomer and maxilla meet at anterior margin of fenestra exochoanalis, prominent choanal fossa on anterior margin of ventral surface of palatine (this is particularly noticeable in larger unregistered bones in the collection where the posterior margin of the fossa extends posteriorly to the anterior edge of the suborbital fenestra/foramen), subdivision of embryonic metotic fissure by the crista tuberalis into vagus (jugular) foramen and recessus scala tympani, enclosed vidian canal exiting anteriorly at base of each basipterygoid process, no quadrate/quadratojugal foramen, medially positioned posterior myohyoidal foramen on mandible, fusion of exoccipitals and opisthotics forming an otoccipital, coronoid anteromedial process fits into sulcus beneath tooth-bearing border of dentary, frontal underlaps (or barely overlaps) parietal laterally on frontoparietal suture, palatine extends posteriorly so pterygoid enters sub-orbital fenestra, and trunk vertebrae lack intercentra.
Cryptovaranoides shows five synapomorphies of Unidentata: frontoparietal suture moderately interdigitated, rugose ornamentation over dorsum, jugal lies ventral to lacrimal, posterodorsal trending ridge delineates anterior limits of naso-lacrimal fossa, and septomaxilla probably contacts dorsal surface of palatal shelf of
85
maxilla (septomaxillary facet on maxilla). Cryptovaranoides shares four synapomorphies with Anguimorpha: frontal underlaps parietal laterally on frontoparietal suture, short overlap in quadrate-pterygoid contact, long ventral longitudinal ridges converging toward midline of vomer, and lacrimal probably arches dorsally over lacrimal duct and floors lacrimal duct with medial process posteriorly.
No known copyright restrictions apply. See Agosti, D., Egloff, W., 2009. Taxonomic information exchange and copyright: the Plazi approach. BMC Research Notes 2009, 2:53 for further explanation.
|
Kingdom |
|
|
Phylum |
|
|
InfraClass |
Diapsida |
|
Order |
|
|
Genus |