Ectenosaurus shannoni, 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 : 16-21

ECTENOSAURUS SHANNONI SP. NOV.

FIGS. 11–13

Zoobank LSID — urn:lsid:zoobank.org:act:4D791F56-

9245-41CF-9290-A0AA0F7C6DA.

Ectenosauru s sp. in Kiernan (2002)

Diagnosis— A medium-sized plioplatecarpine mosasaur

referable to a new species of Ectenosaurus based on the structure of the frontal, which possesses the following characteristics: a) a dorsal surface entirely lacking a median ridge, but with transverse doming anteriorly; b) a parietal dorsal table with rounded lateral margins and lacking parasagittal crest; c) a splenial with an elliptical posterior articulating surface. The species can be further distinguished from Ectenosaurus clidastoides and SMU 7650 by the simpler contact between the frontal and parietal and by a relatively shorter frontal. It can be distinguished from E. everhartorum and E. clidastoides by the form of the posterior terminus of the splenial, which is neither square nor rounded, but elliptical. Characteristics separating Ectenosaurus shannoni sp. nov. from E. tlemonectes include the absence of a median ridge on the frontal in E shannoni , the general ventral morphology of the frontal, a more elaborate sutural mode of contact between the coronoid and surangular in E shannoni , the differences in the splenio-angular joint between the two taxa, and a greater degree of fusion between the supra- and infrastapedial processes of the quadrate in E. tlemonectes .

Holotype — ALMNH: Paleo:5452 consists of several bones from a disarticulated skull and axial skeleton, including the frontal, parietal, right splenial, right coronoid, a fragment of the left quadrate, numerous indeterminate skull fragments, an incomplete scapula, and several badly weathered vertebrae, including five cervicals, thirteen dorsals, five pygals, and four caudals. Much of the speci- men is weathered, especially the scapula and vertebrae, probably as a result of prolonged exposure to the elements before collection. Other than subaerial weathering, the specimen is well preserved, lacking crushing and plastic deformation. The specimen was originally deposited in the collections of the Geological Survey of Alabama under the catalog number GSA V1048 , but in 2005 was transferred to the ALMNH ( Ebersole and Dean 2013) and subsequently assigned a new catalog number, ALMNH:Paleo:5452.

Occurrence— The specimen was collected in November 1976 by S.W. Shannon and J. Kidd from erosional gullies within the lower 21 meters of the unnamed member of the Mooreville Chalk (Selma Group, lower Campanian , Upper Cretaceous , Fig. 2B), located in the NE 1/4, SE 1/4 of Sec. 30, T22 N, R1 E., Greene County, Alabama, USA ( Fig. 1A). The locality has since been reclaimed for farmland and is no longer available for study. Until recently, the Mooreville Chalk was subdivided into a thick lower unnamed member and the thin upper Arcola Limestone Member. Gentry et al. (2022) recognized a basal member of the Mooreville Chalk , informally designated the Erie Bend member. In this paper "unnamed member" refers specifically to the portion of the Mooreville Chalk that overlies the contact with the Erie Bend member and underlies the Arcola Limestone Member .

Etymology— The species epithet shannoni honors the late Samuel Wayne Shannon (1951‒2020) for his contri- butions to Alabama Cretaceous vertebrate paleontology throughout the 1970s (see Shannon 1974, 1975, 1977); also for his co-discovery of the holotype specimen during graduate work at the University of Alabama, Tuscaloosa, USA, and for his early encouragement of the senior author's interest in the Mosasauridae .

Description

Frontal— The frontal of ALMNH:Paleo:5452 ( Fig. 11A, B) is essentially complete, but suffered some minor subaerial weathering prior to collection, particularly on its dorsal side. In overall appearance, the frontal is strikingly similar to that of Ectenosaurus clidastoides , though it is relatively shorter and stouter. Anteriorly, the frontal forms a narrow bifurcate process to enclose the rear of the internarial bar within a V-shaped notch. This process is essentially complete and would have been intermediate in length between that of E. clidastoides and E. tlemonectes (unknown for E. everhartorum ). A median dorsal ridge is completely absent in ALMNH:Paleo:5452, in contrast with E. clidastoides , E. tlemonectes , and SMU 76350. Instead, the anteriormost portion of the dorsal surface ( Fig. 11A) of the frontal is transversely arched or domed anteriorly, a trait absent in other specimens referred to the genus. This arching gives way to a generally planar dorsal surface, which thickens parasagittaly, then thins abruptly at the antorbital margins, and forms a broad, shallow sulcus posteriorly. Tiny foramina are present across the entire dorsal surface. The posterior sulcus deepens and expands to form a shallow depression that would have met the dorsal parietal table, forming the mesokinetic axis of the skull, movement along which may have been inhibited by a series of short vertical keels and overlapping flanges arising from the posterior margin of the frontal and the anterior margin of the parietal. This mode of contact appears to have been similar to, but simpler and less restrictive than, that seen in E. clidastoides , where each side of the dorsal table the parietal is overlapped by a posterior projection from the frontal. These projections are absent in E. shannoni .

The frontal alae are more prominent than in either E. clidastoides or E. everhartorum , linguiform in outline, and diverge posterolaterally at an angle of about 50˚ from the sagittal plane. There are well-developed supraorbital embayments, as with E. clidastoides and E. everhartorum , and within this embayment the frontal is emgarinate on the orbit. Some asymmetry is present in the frontal of ALMNH:Paleo:5452, and this is especially evident in the relative positions of the antorbital bulges. It is unclear whether this asymmetry is pathological or an artifact of preservation, but given the otherwise well-preserved, undistorted nature of the element, the former seems most likely.

The preserved portion of the frontal measures ~ 20 cm in length and is widest at its posterior margins (13.4 cm) and between the antorbital bulges (~ 11.1 cm; a precise measurement is not possible due to the aforementioned asymmetry). The frontal narrows substantially within the supraorbital embayments (8.8 cm), producing a waisted or hour-glass shaped dorsoventral profile.

As mentioned previously, any discussion of the frontal of Ectenosaurus is severely hampered by the ventral surface of the neotype frontal being almost completely obscured beneath the left mandible. In ALMNH: Paleo:5452 ( Fig. 11B), the frontal is bisected ventrally by the olfactory canal, which is bordered on either side by prominent descending processes that run roughly parallel before converging abruptly, then immediately diverging where the canal opens to accommodate the olfactory bulb. This forms a roughly X-shaped configuration, with the descending processes thickest where they almost meet at the crux of the X. Posteroventrally, a pair of deep grooves bracket an elongated stilliform boss that contacts the frontoparietal suture .

As in E. tlemonectes , a narrow transverse ventral separating ridge situated within the supraorbital embayment would have prevented the prefrontal from contacting the postorbitofrontal. While the margins of this ridge are fairly straight in E. tlemonectes and reach the supraorbital rim, in E. shannoni they are concave and laterally recurved, allowing the frontal much broader access to the orbit, so that the entire embayment, as well as the antorbital bulge, are emarginate. In ALMNH:Paleo:5452, the wedge is notched anteriorly, providing a shallow fold into which the tongue-shaped supraorbital process of the prefrontal articulates. Unlike the prefrontal, the postorbitofrontal would have been largely excluded from the supraorbital border, its anterior process is enclosed dorsally on three sides within a shallow, boomerang-shaped sulcus, with the separating ridge forming the anterior rim of this sulcus. This approaches the condition present in E. clidastoides more closely than in E. everhartorum , where the frontal appears to comprise most of the supraorbital border. However, in E. everhartorum the prefrontal and postorbitofrontal are much more widely separated than in either E. shannoni or E. tlemonectes and the ventral separating ridge appears to be absent.

Parietal— Only the incomplete anterior half of the parietal is preserved on ALMNH:Paleo:5452 ( Fig. 11C, D). Prior to collection, the parietal split horizontally into dorsal and ventral halves that were subsequently glued together during preparation. The dorsal parietal table ( Fig. 11C) can be divided into two regions: a broad anterior subrhomboidal shelf that merges smoothly into an elongate posterior shaft with rounded, parallel sides, which would have branched into the missing suspensorial rami. There is no evidence of a parasagittal ridge. A large subcircular parietal foramen, very slightly longer than wide, is centered on the dorsal table, and set back from the frontoparietal suture by a distance roughly equal to twice its diameter. Dorsally, the position and shape of the foramen compares well with that of both FHSM VP- 401 and SMU 76350. The surface of the table is planar. Where it would have contacted the frontal, the margin of the dorsal table forms a distinct W-shaped anterior margin, the central peak bounded by short flanges that must have been overlapped dorsally by posteriorly directed projections from the frontal. This is similar but not identical to the condition present in SMU 76350, where the frontoparietal contact appears to have been more complex ( Bell et al. 2013; fig. m).

Most of the lateral descending process (descensus processus parietalis of Konishi and Caldwell 2011) is not preserved and what is present has been severely crushed. Postorbital processes at the anterolateral corners of the dorsal table appear to have been extremely short, so that the parietal would have contributed little to the anterior supratemporal fenestrae.

The ventral surface of the parietal ( Fig. 11D) is dominated by a large elliptical foramen, at least twice as long as wide, in contrast with its circular dorsal opening. It is unclear whether this dorsoventral variance is an artifact of plastic deformation or accurately reflects the anatomy of E. shannoni . In ALMNH:Paleo:5452 the foramen is surrounded by a steep-sided, narrow-walled parapet of bone whose borders converge just behind the foramen. The lateral walls of this parapet rise to meet its dorsal rim at a 90˚ angle, and the narrow dorsal rim is planar. This structure corresponds to what was referred to by Konishi et al. (2015) as the ventral triangular eminence, present in halisaurines and many russellosaurines (though it displays a wide range of variation within both groups), and its existence contradicts Holmes and Sues' (2000: 310) statement that in mosasaurs "the ventral margin of the foramen is not conspicuously raised."

Quadrate— Only a small portion of the left quadrate was collected ( Fig. 12A‒D), but it preserves the contact and fusion of the suprastapedial process with the infrastapedial and the presence of a dorsomedially directed flange from the infrastapedial that overlaps the expanded dorsodistal end of the suprastapedial, a trait already mentioned in the description of YPM VP4673 as unknown in any mosasaur other than Ectenosaurus . Indeed, in ALMNH:Paleo:5452 the contact is better preserved than in any other specimen of the genus previously described and can be viewed in its entirety, allowing for a better understanding of the precise nature of this unique contact, while also suggesting the trait may have been less developed in E. shannoni than in either E. clidastoides or E. tlemonectes . In ALMNH:Paleo:5452, the infrastapedial process forms a delicate cheliform structure that not only overlaps the suprastapedial posterodorsally, but ventrally as well, embracing it in a sort of pincer grip while also creating a cradle that abuts the suprastapedial terminus. Complete fusion between the two processes can only be observed ventromedially ( Fig. 12B). While the full extent of fusion is unclear,, the processes are less extensively fused than in E. clidastoides . The infrastapedial process of YPM VP4673 ( Fig. 3B) may have enfolded the suprastapedial in much the same way, but preservation makes this difficult to determine with any certainty; fusion appears to have been more extensive in the Niobrara Chalk specimen. In neither ALMNH:Paleo:5452 nor YPM VP4673 does the dorsodistal flange enter the deep, polygonal fossa for insertion of M. depressor mandibulae, but contacts and by G.L. Bell, Jr. (M.J. Polcyn personal communication parallels the fossa's distal rim. In ALMNH:Paleo:5452, the 2022). Fortunately, the restored portions are unpainted posterior (external) flange is highly vascularized, though and easily discernible from the fossil itself. As in most the anterior (ventral) flange is not, and there is a distinct russellosaurines, the coronoid is a short, saddle-shaped bulge along the ventrolateral rim of the suprastapedial wedge of bone straddling the anterior rim of the suranprocess that does not appear in either E. clidastoides or gular and forming a posteromedial sutural contact with E. tlemonectes . a buttress on the surangular ( Fig. 13E, F). This sutural

Splenial— Much of the right splenial is preserved (Fig. contact is more complex in ALMNH:Paleo:5452 than in 13A‒C), including the base of the lateral ala. As the me- YPM VP4673 and even more so than in FHSM VP- 401, dial ala has been lost, it is only possible to approximate with the interdigitating grooves and ridges being more the true shape of the articular face at the intramandibu- numerous and more closely spaced. Based on the right lar joint, which appears to have formed an elongated coronoid figured in Willman et al. (2021: fig. 11b), it also rhombus. The "robust, dorsomedially expanding flange" appears more complex than in FHSM VP- 5515. Howreported by Willman et al. (2021) for both Ectenosaurus ever, until more careful examinations of the latter two everhartorum and E. clidastoides , giving a squarish shape specimens can be conducted, as well as comparison with to the articulating surface in those species, is absent in the Big Bend specimen ( SMU 76350), we consider this ALMNH:Paleo:5452. While the splenio-angular joint in feature in ALMNH:Paleo:5452 a provisional apomorphy ALMNH:Paleo:5452 is simpler than that of YPM VP4673, for E. shannoni . it would still have been more complex than the ball- In ALMNH:Paleo:5452, the posterior crest of the coroand-socket condyle/cotyle arrangement often used as a noid is taller than in E. tlemonectes , but less developed blanket characterization of this joint in mosasaurs (see than in E. clidastoides ; and incomplete preservation discussion above). Instead, the articular face is divided makes it difficult to compare with FHSM VP- 5515. The an- into a broad vertical keel positioned just medial to its terior bifurcation is less developed than in E. tlemonectes . center and a sulcus located between the lower terminus and is more dorsally located. In ALMNH:Paleo:5452, of the keel and the medioventral rim of the splenial; these the lateral descending wing (largely restored) is a thin, would have received matching structures from the an- relatively delicate, ventrally-directed process, while the gular. A well-defined longitudinal ridge marks the base medial descending wing is much shorter and ventromeof the lateral wing. dially directed, forming a stout, triangular wedge in an-

Coronoid— The majority of the right coronoid (Fig. teroposterior cross-section. Though it is clear the lateral 13D‒F) is preserved, though it should be noted that wing was shorter than the medial, the precise difference some portions have been heavily restored with plaster cannot be determined due to the incompleteness of the latter. Numerous foramina are present within the ventral fold and, as in YPM VP4673, are especially prominent on the interior surface of the medial descending wing.

Vertebrae— A total of twenty-seven vertebrae are preserved with ALMNH:Paleo:5452, including five cervicals, thirteen dorsals, five pygals, and four caudals. Most of the vertebrae consist of badly weathered centra lacking most of the cortical surface, neural arches, synapophyses. zygopophyses, and transverse processes. The best cervical vertebra ( Fig. 13F) preserves a distinct anteroventral ridge extending from the anterior rim of an elliptical hypapophyseal peduncle to the ventral rim of the cotyle. In lateral profile, this ridge forms a shallowly concave arc, as seen in YPM VP4673 and other species of Ectenosaurus . The articular surfaces of the pygals are sub-square in anteroposterior view. Only one of the caudles is well preserved ( Fig. 13G); it is subcircular in anteroposterior profile and preserves two rounded, posteroventrally oriented peduncles for articulation with the haemal arch.