Ectenosaurus tlemonectes, Kiernan & Ebersole, 2023

Kiernan, Caitlin R. & Ebersole, Jun A., 2023, Two new plioplatecarpine mosasaurs (Mosasauridae; Plioplatecarpinae) of the genus Ectenosaurus from the Upper Cretaceous of North America, PaleoBios 40 (1894), pp. 1-28 : 5-16

ECTENOSAURUS TLEMONECTES SP. NOV.

FIGS. 2–10

ZooBank LSID — urn:lsid:zoobank.org:act:E2E61DA4-

E21D-4643-A0B6A8C04418062

Ectenosaurus clidastoides ( Merriam, 1894) in Russell (1967: 158).

“ Ectenosaurus composite” in Bell (1997).

Ectenosaurus sp. in Konishi (2008).

Ectenosaurus sp. in Willman et al. (2021).

Diagnosis— A medium-sized plioplatecarpine mosasaur referable to the genus Ectenosaurus and exhibits the following unique combination of characteristics: a) very prominent median dorsal ride on frontal, extending back at least to interorbital embayment, wider posteriorly than anteriorly; b) olfactory canal long and tightly bracketed by broad, hour-glass shaped descending processes; c) frontal subtriangular with weakly developed, incipient interorbital embayments; d) suprastapedial process broadly fused with infrastapedial process, the latter partially overlapping the former; e) mandibular condyle deeply concave, saddle-shaped; f) medial face of squamosal rugose and highly vascularized; g) angular surface for contact with surangular posteriorly recurved; h) premaxilla dorsally planar, except for prominent median keel; i) step-wise constriction present anteriorly on internarial bar; j) short edentulous rostrum on premaxilla; k) premaxillary-maxillary contact terminates between the third and fourth maxillary teeth, at the anteriorly deepest portion of the maxilla; l) teeth with smooth enamel at bases, overprint of striations and flut- ing higher on crown, carinate and fluted toward cusps. Characteristics a, c, e, g, h, i, j, k, and l distinguish the taxon from Ectenosaurus clidastoides and E. everhartorum , as well as from ALMNH:Paleo:5452, are considered apomorphic, and justify the erection of a new species. Feature b) distinguishes the taxon from ALMNH:Paleo:5452, but it is unclear whether this characteristic is present in other members of the genus.

Holotype — YPM VP4673, a mostly complete, disarticulated skull and mandibles, several cervical and dorsal vertebrae representing a medium-sized mosasaur (skull ~ 70 cm; total length ~4.0 m, based of dimensions of Platecarpus "ictericus" provided in Russell (1967) after Williston (1898). Cranial material includes premaxilla, both maxillae, both dentaries, the posterior ends of both splenials, the anterior ends of both angulars, almost complete left posterior mandibular unit (i.e., surangular, articular, prearticular) and less complete left posterior mandibular unit (i.e., fragmentary surangular, articular, prearticular, plus coronoid), frontal, portions of both jugals, almost complete left quadrate, proximal fragment of right quadrate, the posterior end of the right squamosal, fragmentary right pterygoid, much of the neurocranium (including the basiocciptal, supraoccipital, the left parocciptal process of the opisthotic, and portions of the prootic), several associated teeth or tooth fragments, along with numerous other miscellaneous bone fragments believed to belong to the skull. The only preserved postcranial material consists of several cervical and dorsal vertebrae or fragments of vertebrae, all badly distorted and weathered.

Occurrence— Collected June 1875 by S. W. Williston from erosional gullies in the Smoky Hill Chalk Member of the Niobrara Chalk Formation (upper Coniacian-lower Campanian; Fig. 1B), Gove County, Kansas ( Fig. 1A). The precise locality and stratigraphic level within the Smoky Hill Chalk are unknown .

Etymology— Species name from Τλήμων [tlḗmōn],

Classical Greek for "patient" + νήΚΤης [nḗktēs], Classi- cal Greek for "swimmer," referring to the 147 years that elapsed between the holotype's discovery in 1875 and its eventual recognition as a distinct taxon in 2022.

Description

Frontal— The frontal ( Fig. 2A, B) is an elongate, triangular bone with very slight antoribital bulges and short, roughly triangular posterolateral alae. The posterior quarter of the frontal is weathered and broken into three fragments. As preserved, the midline length is ~ 200 mm, with a maximum width of 87.9 mm and an interorbital width of 68 mm. A very well-developed median dorsal ridge ( Fig. 2A) extends at least three-quarters of the frontal's length (~ 132.8 mm), widening markedly posteriorly (from 3.9 mm to 8.6 mm). Ventrally ( Fig. 2B, C), a pair of broad, rounded, and roughly violin-shaped descending processes (the crista cranii frontalis of some authors) bracket the olfactory tract, almost completely enclosing it for nearly half its length (59.3 mm) before opening anteromedially into a flat amygdaliform sulcus to accommodate the olfactory lobes. Though neither the prefrontals nor postorbitofrontals are preserved, articulation between these elements and the frontal is evident from well-defined, striated depressions on the ventral surface of the frontal. These articular surfaces are completely separated from one another by a wedge or ridge (the ventral separation ridge of Konishi and Caldwell 2007), oriented approximately 90˚ to the sagittal plane, permitting the frontal to form a small portion of the supraorbital border. Among other plioplatecarpines, this trait has only been reported in Russellosaurus and Plesioplatecarpus , but it is also present in Yaguarasaurus Páramo (2000) , the halisaurines, and in the mosasaurine Clidastes (sensu C. propython Cope, 1869 ); it is presumed to be plesiomorphic for the Mosasauridae ( DeBraga and Carroll 1993) . The contact with the parietal is roughly transverse and possesses no prominent posteromedial projections onto the parietal. The parietal is not preserved, and it is unclear whether, and to what degree, it may have invaded/overlapped the frontal anteromedially.

Quadrates and Squamosals— The left quadrate ( Fig. 3A‒C) is reasonably complete but lacks most of the tym- panic ala and portions of the anterior quadratic shaft. The suprastapedial process is medially deflected and is large and elongate, as is typical of plioplatecarpines, measuring ~60% of the quadrate height. The borders of the suprastapedial process are straight medially, slightly concave laterally, and expand distally to form a bulbous, irregularly rounded extremity that is fused with the infrastapedial process. In the case of Ectenosaurus , the "infrastapedial" process is formed by a greatly developed posteroventral ascending quadratic rim (see Bell 1997 and Palci et al. 2020 for reviews of issues of homoplasy and inconsistent terminology related to this structure) and includes a flange-like process that overlaps the su- prastapedial posterodistally. Though the two processes contact to varying degrees in multiple halisaurine, plioplatecarpine, and mosasaurine taxa, the particular mode of contact seen in Ectenosaurus is distinctive, in that the distal end of the suprastapedial is not only solidly fused with but partially embraced by a projection from the infrastapedial process. We revisit this character state in the description of ALMNH:Paleo:5452.

Only a small portion of the thin tympanic ala is preserved. The mandibular condyle is deeply concave and saddle-shaped, possibly the most pronounced example of this condition in any mosasaur. Though the quadrate of FHSM VP 401 is crushed, Ectenosaurus clidastoides does not appear to share this feature. It is possibly incipient in Selmasaurus johnsoni Polcyn and Everhart (2008) but absent in S. russelli . Herein it is considered an apomorphy for Ectenosaurus tlemonectes . The suprastapedial pit is severely distorted anteroposteriorly, obscuring its original dimensions, but it appears to have been long and subrectangular.

The posterior end of the right squamosal ( Fig. 3D‒F) is preserved and in good condition. The medial face ( Fig. 3E) is marked by an elaborate series of ridges and troughs that would have received corresponding surfaces on the supratemporal. Numerous large foramina are also present. The ventroposterior fossa for articulation with the quadrate ( Fig. 3D) forms a rough isosceles triangle and is shallowly concave. Anterodorsally there is a small but distinct knob marking the posterior terminus of the groove for insertion of the postorbitofrontal processes.

Dentigerous elements— The premaxilla and the majority of both maxillae and dentaries are present in YPM VP4673. The premaxilla ( Fig. 4A‒C) is well preserved and complete except for the posteriormost portion of the internarial bar. There is a very short, spatulate edentulous rostrum (5.0 mm). Dorsally, the premaxilla is planar and bears a small median keel or crest (19 mm) near the base of the internarial bar that is subtriangular in lateral profile ( Fig. 4C). This structure has not been observed in any other specimen of Ectenosaurus . The posterior terminus of the premaxillary-maxillary suture occurs at the deepest anterior portion of the maxilla ( Fig. 4D, E), at a position even with the fourth maxillary tooth. The internarial bar narrows abruptly behind the premaxillomaxillary suture (the "step-wise constriction" described by Willman et al. 2021), is triangular in cross section, and expands posteriorly to receive the anterior process from the frontal.

Neither maxilla is complete, so the tooth count can only be estimated. There are ten preserved tooth positions present in the right maxilla and 11 in the left, though there may have been as many as 12 teeth present in each, considering both are missing the posteriormost portion of the tooth row. As noted by Willman et al. (2021:744), the laterally expanded anterior portion of the external naris extends from the fourth to the midpoint of the seventh tooth in YPM VP4673, a distance of 3.5 tooth bases, which is slightly longer than that of E. everhartorum (2.5 tooth bases long), but shorter than E. clidastoides (four tooth bases long).

The dentaries ( Fig. 5A‒E) are gracile, delicately constructed bones, more so than in either Ectenosaurus clidastoides or E. everhartorum . Both the dorsal and ventral margins are straight. There are 13 teeth in each dentary. The anterior portions of the dentaries in FHSM VP- 401 are restored with plaster, so a precise tooth count for the generic neotype is not possible. According to Willman et al. (2021), the restored dentaries are too short and do not match the maxillae, and the dentary tooth count of "15 and 14 teeth in the right and left dentaries, respectively" is too low. Based on the unrestored premaxilla and maxillae, they estimate the original dental formula for E. clidastoides to have been 2=17=17 (premaxilla=maxillae=dentary). This trend toward hyperdonty in Ectenosaurus is incipient in E. tlemonectes , where the dental formula is 2=~12=13, nearer to the formula seen in most plioplatecarpines (2=12=12).

As noted by Willman et al. (2021:746, 752; fig. 8f), in YPM VP4673 the strap-like medial parapet is raised above the level of the lateral wall in the dentaries, as in E. clidastoides and E. everhartorum , as well as the highly derived Plioplatecarpus marshi Dollo (1882) , a surprising similarity given the latter's phylogenetic distance from Ectenosaurus . Willman et al. (2021:752) explained this as a case convergent evolution, a conclusion with which we concur.

A small edentulous prow is present on the dentary, which on the right is severely deformed, probably from a healed injury ( Fig. 6); the deformation involves the first dentary tooth. A very large and elongate, longitudinally oriented foramen is present below the tenth tooth position, a character also present in E. everhartorum , where the foramen occurs between the tenth and eleventh dentary teeth ( Willman et al. 2021:746).

A fragment of the left pterygoid is present, preserving the bases of seven teeth and the base of the ectopterygoid process. The preserved portion of the tooth row is straight as in E. everhartorum ( Willman et al. 2021) , as opposed to the sigmoidal shape common to plioplatecarpines.

Jugals— Portions of both jugals are preserved ( Fig. 7C‒F). Posteroventral processes are present on both, though a considerable degree of asymmetry exists between the two elements. The process is small on the right jugal ( Fig. 7C, D), but far more developed on the left ( Fig. 7E, F), where it forms a scalloped posterior border for the vertical ramus. While we suspect this asymmetry is the result of pathology, additional examples of the species will be needed to determine which jugal exhibits the abnormality. A deep triangular depression on the lateral surface of the vertical ramus marks the point of articulation with the postorbitofrontal. The vertical ramus is short, less than 50% of the length of the horizontal ramus.

Coronoids— In most russellosaurines, the coronoid is a short, saddle-shaped wedge of bone straddling the anterodorsal margin of the surangular, which it encloses to varying degrees between simple medial and lateral descending processes or wings that reach the prearticular. In the plioplatecarpines, these wings are usually short, compared with those of mosasaurines. In YPM VP4673, only the right coronoid ( Fig. 7A, B) is preserved and is slightly crushed lateromedially.As with most russellosaurines, the coronoid is a slender, short, selliform element divided into lateral and medial wings which straddle the anterodorsal surface of the surangular. The coronoid dorsal margin is concave, and the ascending posterior process was well-developed, though the apex of the crest is missing. The stout medial wing would have overlapped the dorsomedial margin of the surangular, and the interior surface of this wing is highly vascularized, marked by numerous foramina. In contrast, the anterolateral wing is very short, and hardly more than a low, diagonal ridge. The shallow anteromedial bifurcation ("cleft" of Russell 1967:53) is long and well-defined, extending more than a third of the bone's length. We will discuss the coronoid of Ectenosaurus further in our description of ALMNH:Paleo:5452.

Articular and Surangular— The majority of both the left and right surangular-articular units are preserved ( Fig. 8), though the left is the most complete. The dorsal margin of the surangular is slightly concave and broadens anteriorly as it approaches the contact with the coronoid, expanding from a height of about 3–~ 9 cm. The surangular-articular suture is obvious on the lateral side of the left mandible ( Fig. 8A, B, D), curving beneath the glenoid fossa before running more or less straight beneath the coronoid to the intramadibular joint. The right and left articulars are both missing their posteroventral corners, but the ventral margins appear to have been generally straight. Most of the glenoid fossa is located on the articular. The posterodorsal tongue of the surangular terminates abruptly in a rugose, highly vascularized vertical face that forms the anterior wall of the glenoid fossa ( Fig. 8C). Willman et al. (2021:747; fig. 12) stressed the presence of a "distinct notch on the dorsal surface...just prior to the glenoid fossae" in E. clidastoides , E. everhartorum , and YPM VP4673 ( Fig. 8C) as a character "absent in all other mosasauroids" and therefore considered diagnostic of the genus by those authors. In YPM VP4673, portions of the glenoid fossa are bordered dorsally and medially on both the angular and surangular by a thick, raised rim with rounded edges ( Fig. 8B, C).

Splenial-Angular— Portions of both splenials and angulars are present ( Fig. 9A–L), including the articulating faces for all four elements. Willman et al. (2021:747) described the posterior surface of the splenials in E. everhartorum and E. clidastoides as possessing a "robust, dorsomedially expanding flange" giving an "overall square shape, in which they are almost equidimensional in height and width." Neither splenial of YPM VP4673 has a complete articulating surface, and the left is slightly crushed lateromedially, but both have curving margins and appear to have been more rounded than square ( Fig. 9J). Therefore, we do not consider a splenial with a squared articulating surface as diagnostic of the genus Ectenosaurus . Only the anterior termini of the angulars are preserved ( Fig. 9A‒H). While the left is less crushed, the right includes enough of the lateral wing that the surface for its contact with the surangular is preserved ( Fig. 9C). This surface is marked by pronounced horizontal striations and, unlike the condition present in E. clidastoides , its ascending margin is distinctly recurved posteriorly, a trait here considered diagnostic of E. tlemonectes .

The splenio-articular joint in mosasaurs is usually described as a simple ball-and-socket joint ( Camp 1942, Russell 1967, Bell 1997, Rieppel and Zaher 2000), a cotyle (splenial)-condyle (angular) articulation that often fails to reflect the wide range of variation between taxa; Rieppel and Zaher (2000:8), for example, described it simply as "the splenial is the receiving part, the angular the received part." In YPM VP4673, the splenio-angular joint is more complex ( Fig. 9L), with each element contributing to the cotyle/condyle arrangement in a contact consisting of a series of more or less vertically oriented ridges and sulci. A deep median sulcus is present on the angular, for example, receiving a matching ridge on the splenial, while strong medial and lateral ridges have matching splenial sulci. While it is possible to deduce the general nature of the splenio-angular joint in YPM VP4673, deformation has altered the details of the articular faces of the angulars ( Fig. 9B, G). The elements do not match, and it is unclear which most closely approximates the original morphology.

An anteroventral rugosity on the right angular appears to be pathologic ( Fig. 9E), as it is absent on the left angular. Considering the severe damage to the anterior end of the right dentary, it is possible this pathology resulted from the same trauma, produced by torsion inflicted on the intramandibular joint.

Neurocranium— As with much of the posterior portion of the skull of YPM VP4673, the occipital unit has suffered considerable weathering prior to collection. However, much of the neurocranium is preserved. Elements identified include the basiocciptal,, supraoccipital, the left parocciptal process of the opisthotic, and portions of the prootic.It is likely that much of the basicranial circulatory pattern can be reconstructed, as the periotic labyrinth is often well preserved. This aspect of the morphology of YPM VP4673 will be the focus of a future collaborative study between the senior author and M.J. Polcyn ( SMU).

Teeth— Almost all the marginal teeth of YP VP4673 have broken free of the bone and been lost during post-exposure weathering and/or during collection of the specimen. The few teeth that remain in situ are incomplete and missing their apices, though the apices can be observed for several replacement teeth. Possible pterygoid teeth are also present, but none are attached to the preserved fragments of the pterygoids. Preserved marginal teeth include the third tooth in the left dentary ( Fig. 5A, B), the third, fourth, fifth, and seventh teeth in the right dentary ( Fig. 5D, E), the fifth right maxillary tooth (with replacement tooth) ( Fig. 4D, E), and the second left maxillary tooth (not figured). None of the four premaxillary teeth are preserved, but the bases indicate they may have been slightly prognathous ( Fig. 4A, B). The bony tooth base is significantly taller in YPM VP4673 than in E. clidastoides , more closely resembling the condition present in E. everhartorum , Selmasaurus johnsoni , and Platecarpus tympaniticus Cope (1869) . For example, the bony base of the fifth right dentary tooth (the most complete, preserving all but the cusp) is 5.8 mm tall labially, while the preserved portion of the crown measures 14.9 mm.

Russell (1967:156) did not describe the dentition of FHSM VP- 401 in any detail, beyond noting the dental formula and that the teeth are "bicarinate, vertically striated, and medially recurved." However, additional observations are made here based on photographs provided to us by M.J. Polcyn ( SMU). Willman et al. (2021) say nothing about the tooth morphology of E. everhartorum . Other than the second left maxillary tooth and the apices of replacement teeth, only broken bases are preserved in FHSM VP- 5515. Fortunately, while most of the teeth in YPM VP4673 are missing, the ten fragmentary teeth, plus the replacement teeth, do allow for a description of the general dental morphology of YPM VP4673, including the apicobasal ornamentation of the enamel. All the teeth appear to be weakly faceted, though faceting has previously been considered absent in russellosaurines ( Street et al. 2021). While the marginal dentition of E. clidastoides appears to be uniformly strongly bicarinate throughout, but absent on the basal region of the crowns. Apically, in YPM VP4673 carinae are only well developed near the enamel bears an overprint of striations and fluting the cusp (most evident in replacement teeth) and never that is usually more developed on the lingual surfaces. extend the full length of a tooth. Mesial carinae are gener- In general, ornamentation is less prominent than in E. ally better developed, and some teeth lack distal carinae. clidastoides , where striations extend the full height of the The carinae lack serrations (termed false denticulation teeth. Anteriorly, teeth in YPM VP4673 are subcircular by Prasad and de Lapparent de Broin (2002) and Street in cross section, becoming slightly elliptical posteriorly. et al. 2021). Except for faceting, ornamentation is all Teeth in YPM VP4673 are posteromedially recurved, as in E. clidastoides . A number of loose teeth believed to have broken off the pterygoids are more elliptical in cross section, lack almost all ornamentation, but have strongly developed carinae.

Vertebrae— Vertebrae are the sole preserved elements of the postcranial skeleton, though only a few cervical (including the axis) and dorsal vertebrae were collected and so the vertebral formula is unknown. The vertebrae that are present have all suffered varying degrees of plastic deformation. Where the cotylar and condylar surfaces are preserved, the interarcticular surfaces appear to have been broadly elliptical. The two cervicals have strongly developed, posteriorly-inclined hypapophyses. The hypapophyses are joined to the ventral rim of the condyle by a narrow, gently concave ridge ( Fig. 10). Willman et al. (2021) stated that this ridge is unknown in all other plioplatecarpines except Selmasaurus johnsoni and Ectenosaurus everhartorum . The hypapophyseal peduncles are slightly crushed but appear to have been elliptical in life.

Taxonomic note

Future work may show YPM VP4673 to be distinct from Ectenosaurus at the generic level as it does differ markedly in some respects (form of the frontal, concave mandibular condyle on quadrate, etc.), while possessing other characters considered diagnostic of the genus (post-glenoid notch, mode of contact between supra- and infrastapedial processes,etc.). In choosing here to include it in Ectenosaurus we have elected the more conservative approach. Curiously, YPM VP4673 was described on three separate occasions ( Konishi 2008, Willman and Konishi 2019, and Willman et al. 2021), whole or in part, and never recognized as a unique species. Konishi (2008) described it as Ectenosaurus sp. , Willman and Konishi (2019) suggested it was conspecific with the then-undescribed E. everhartorum ( FHSM VP- 5515), and Willman et al. (2021), in the description of E. everhartorum , returned to calling it Ectenosaurus sp. Notably, these authors consistently failed to locate both the frontal and splenials, key elements separating E. tlemonectes from E. everhartorum and E. clidastoides .