Embrithosaurus schwarzi, Watson, 1914

Brandt, Marc Johan Van Den, Abdala, Fernando & Rubidge, Bruce Sidney, 2020, Cranial morphology and phylogenetic relationships of the Middle Permian pareiasaur Embrithosaurus schwarzi from the Karoo Basin of South Africa, Zoological Journal of the Linnean Society 188, pp. 202-241 : 236-238

DENTITION OF EMBRITHOSAURUS SCHWARZI

Embrithosaurus has been distinguished from all other Tapinocephalus AZ pareiasaurs, since Broom’s (1903) description, by the presence of up to nine cusps on the maxillary and mandibular teeth. This differs from the much more common arrangement of seven cusps per tooth for all other pareiasaur species from the biozone, and has, therefore, been considered taxonomically important by most authors. For example, Haughton & Boonstra’s (1929) pareiasaurian classification, focusing on teeth, provided a basic separation of the then 12 Tapincocephalus AZ taxa recognized at that time into two groups: Embrithosaurus with nine-cusped teeth, and all other 11 taxa with less than nine marginal cusps on their teeth. Similarly, Kuhn (1969) considered all forms from the Tapinocephalus AZ in only two genera: Embrithosaurus (four species) with nine cusps per tooth and Bradysaurus (nine species) having fewer than nine cusps.

However, some authors have used either a range of seven to nine cusps ( Broom, 1903; Lee, 1997a) or exactly nine-cusped teeth ( Haughton & Boonstra, 1929; Boonstra, 1934; Kuhn, 1969) as the distinguishing feature to separate Embrithosaurus from the other basal Tapinocephalus AZ forms. Broom (1935) and Boonstra (1969) did not specify the number of marginal cusps in their diagnoses of Embrithosaurus . Broom (1903) remarked that the upper and lower teeth of Embrithosaurus have nine cusps per tooth, and added that the damaged upper front teeth ‘apparently’ have seven cusps ( Broom, 1903: 127), noting his assessment as tentative. All the upper front teeth on the premaxilla are damaged and it is not possible to determine the number of cusps. Besides for being larger (as in all pareiasaurs), the damaged premaxillary teeth appear similar in shape to the adjacent maxillary teeth, which are well preserved and bear nine marginal cusps. This suggests the same number of cusps for the premaxillary teeth. The teeth of the right dentary teeth are well exposed and all reveal nine cusps per tooth along the entire tooth row. Based on the holotype, Embrithosaurus thus has nine cusps on all teeth.

However, Lee (1997a) settled on a range of seven to nine cusps for Embrithosaurus , even though the holotype shows nine cusps on all teeth. It appears that this is the result of accommodating two specimens in Embrithosaurus : he provisionally referred NHMUK R7782, based on a process of elimination ( Lee, 1997a: 253, 254) (even though the dentary appears to have seven-cusped teeth); and he assigned GSP-CBT-112 ( Lee, 1997a) based on the presence of three pelvic autapomorphies (even though the maxilla of GSP- CBT-112 definitely has seven-cusped teeth). Based on dental features alone, we do not recognize the referred specimens NHMUK R7782 and GSP-CBT-112 as Embrithosaurus , as they have only seven marginal cusps and, therefore, do not match the morphology of the holotype of Embrithosaurus . The next step of our study will include postcrania and will be presented elsewhere.

Moreover, it is important to note that the wider maxillary tooth shape of Embrithosaurus , unique amongst the Tapinocephalus AZ forms, is created by: (1) the higher number of cusps; (2) cusps arranged more regularly around the crown, with smaller gaps or closer spacing between the central cusps and the first lateral cusps; and (3) the lack of a central, elongated, isolated trident. In all other co-occurring species there are fewer cusps, and the three central cusps on the teeth occur on the distal tip of a long, narrow, isolated trident at the apex of the crown, following which there are large gaps or spaces to the first lateral marginal cusps on the mesial and distal margins of the crowns, creating an elongated and narrower tooth shape [as in B. seeleyi holotype NHMUK 49426, B. baini holotype NHMUK R1971, B.baini NHMUK R1970 ( Fig.17D) and Nochelesaurus alexanderi holotype SAM-PK-6239]. In contrast, in Embrithosaurus , the first lateral marginal cusps on the mesial and distal margins are placed much closer to the three central cusp complex, and the central three cusps, therefore, are not found on a long, narrow, trident. This creates a more evenly distributed or regularly spaced cusp arrangement and the wider tooth shape of the holotype of Embrithosaurus ( Fig. 17A) and referred specimen SAM-PK-9098 ( Fig. 17C). This unique tooth morphology can be used to distinguish Embrithosaurus from co-occurring taxa. It is also important to note that, although the marginal cusp arrangement of Embrithosaurus is more regular than the co-occurring taxa, due to more cusps and the smaller spaces between the central three cusps and the first lateral cusps, the cusp arrangement of Embrithosaurus does not match the almost perfectly regular (evenly spaced) cusp arrangement defined in phylogenetic character 66, which is only found in the more derived pareiasaurs.

Based on dental features alone (excluding postcrania), our interpretation, therefore, agrees with the diagnoses of Haughton & Boonstra (1929), Boonstra (1934) and Kuhn (1969), who state exactly nine cusps for all teeth as an important diagnostic feature, and differs from Broom (1903) and Lee (1997a), who accepted a range of seven to nine cusps for Embrithosaurus .

VALIDITY OF HISTORIC DIAGNOSTIC FEATURES USED FOR EMBRITHOSAURUS SCHWARZI

Broom (1903) described the holotype of Embrithosaurus schwarzi (SAM-PK-8034) as a member of Pareiasaurus serridens and later ( Broom, 1935) admitted this was an error after Watson (1914a) noted that the specimen was not at all similar to P. serridens . It appears that Broom was misled in his referral of the new specimen to P. serridens , by a combination of distortion and poor preparation, which confused the anatomy. According to Lee (1997a), the derived, largebodied P. serridens is characterized by, amongst other features: a smooth skull with cranial ornamentation consisting of a network of fine grooves and pits, and an absence of central bosses of each cranial element, upper jaw teeth with 9–11 cusps and large marginal bosses on the ventral surface of the cheek. The cranial ornamentation of SAM-PK-8034 has been variously over-prepared and eroded into, creating the impression of a smoother skull, with lower central bosses, than if undamaged. The condition of nine cusps on the teeth probably also influenced Broom’s (1903) incorrect placement of SAM-PK- 8034 in P. serridens , as at that time all Tapinocephalus AZ pareiasaurian specimens were known to only possess seven cusps on their teeth ( Seeley, 1888, 1892). Pareiasaurs from the Tapinocephalus AZ , including Embrithosaurus , do not carry bosses on the ventral margins of the cheeks as P.serridens does. Broom (1903) may have misinterpreted the distorted arrangement of the bosses of the posteroventral corner of the right cheek of SAM-PK-8034 as carrying ventral bosses, as there are two very close, almost unabated corner bosses, where the anterior boss extends anteriorly onto the ventral margin of the cheek.

Broom (1903) correctly noted the general bilateral symmetry of the ornamentation (bosses) of the skull, but he failed to note the significant non-symmetrical placement and size of the bosses on the posterior marginsoftheleftandrightquadratojugal.Thisamount of variability of both the arrangement and the size in the cranial ornamentation, on the same specimen, should caution workers against pareiasaurian species diagnoses based on minor morphological differences.

Watson (1914b), in his brief generic diagnosis for Embrithosaurus , included only two cranial characteristics: the skull being deep (i.e. large or extensive cheek overhangs) and narrow (i.e. pointed snout/skull). The skull of the holotype of Embrithosaurus does have a large cheek, but this is similar in size to that of the holotype of B. baini and Nochelesaurus and, therefore, does not distinguish the species. Regarding the narrow snout, the skull of the holotype of Embrithosaurus is mediolaterally compressed and twisted to the right, creating the appearance of a narrow, pointed snout, which Haughton & Boonstra (1929) and Boonstra (1934) also considered as diagnostic. Lee (1997a), finally, correctly pointed out that the narrow snout as a character was a taphonomic artefact.

Haughton & Boonstra’s (1929) use of the nine-cusped teeth for Embrithosaurus as the most important distinguishing feature for this species is correct. They also noted the pronounced, bulbous bosses on the posterior margin of the cheek flange, common to all species of basal South African pareiasaurs, except for Nochelesaurus , which has more distinct, pointed bosses. This is an important and valid diagnostic condition for Embrithosaurus , as currently the morphology of the bosses on the posterior edge of the cheeks of pareiasaurs is used for diagnosis in several species. For example, Arganaceras (Jalil & Janvier, 1995) , Obirkovia ( Bulanov & Yashina, 2005) and Elginia ( Newton, 1893) have similar small, long, pointed hornshaped bosses and Scutosaurus ( Lee, 1994) has larger, pointed, rugose bosses.

Boonstra (1934) provided a more detailed cranial diagnosis for Embrithosaurus , based on the holotype and specimen NHMUK R7782. However, many of the features he listed are common to most pareiasaurs from the Tapinocephalus AZ. The only unique diagnostic feature noted for Embrithosaurus is the presence of nine-cusped teeth, which Boonstra (1934: 29) specifically mentions as a ‘feature of taxonomic importance’. The large, bulbous bosses on the posterior edge of the cheek, shared with B. baini , is also correctly noted. We disagree with Boonstra (1934) that Embrithosaurus has a transversely wide postparietal, as we discovered unambiguous external cranial sutures confirming a narrow, square postparietal on the holotype skull. Boonstra (1934: 29) noted a ‘very slightly bulging’ maxilla and lacrimal, which we agree with as there is a moderately swollen maxillarylacrimal ridge formed by the bosses of the maxilla and lacrimal.

Lee (1997a) did not identify unique cranial autopomorphies for Embrithosaurus . He presented three pelvic characters as autapomorphies ( Watson, 1914b; Haughton & Boonstra, 1930b). Lee pointed out the lack of a large ‘lump’ or boss on the maxilla distinguishing Embrithosaurus from B. baini , but did not note the presence of the small, distinct, partially sheered-off left maxillary boss on the holotype of Embrithosaurus . Both these taxa have maxillary bosses and swollen maxillary–lacrimal ridges, but are much larger in B. baini . The holotype of Nochelesaurus has no maxillaryboss. Lee(1997a) alsopointedoutthelarge, low, rounded, posterior cheek bosses that Embrithosaurus has in common with B. baini , distinguishing these taxa from Nochelesaurus , which has more discrete, pointed bosses on the posterior cheeks.

Lee (1997a: 247) proposed an autapomorphic condition in the distal portion of the paroccipital process of the opisthotic of B. baini , noting it as ‘hugely swollen’. This brief description we interpret as a combination of two features: the anteroposteriorly expanded distal portion of the paroccipital process and an oblique, dorsomedially oriented, swollen, knob-like posterior projection. The distal portions of the paroccipital processes of Embrithosaurus and Nochelesaurus are expanded to 55 mm, whereas it is longer in Bradysaurus baini at 80 mm. However, the dorsomedially elongated, posteriorly projecting, knob-like process, on the medial edge of the posterodorsal corner of the paroccipital process (absent in Nochelesaurus ) is present in both Embrithosaurus and Bradysaurus baini , and although it is larger in Bradysaurus baini , it is not an autapomorphy of that taxon.

Relatively small cheek flanges compared to all other pareiasaur species (measured by the angle between the ventral margin of the cheek and the horizontal maxillary tooth row) are present in the holotypes of Embrithosaurus schwarzi (SAM-PK-8034: left cheek 24°, right cheek 34°, average 29°, Nochelesaurus alexanderi (SAM-PK-6239: right cheek 33°) and Bradysaurus baini (NHMUK R1971: left cheek 35°, right cheek 20°, average 27.5°).

Nochelesaurus and Embrithosaurus have downwardpointing maxillary teeth that emanate from a narrow, vertical premaxilla and maxilla. This character distinguishes these two taxa from B. baini (e.g. NHMUK R1971, SAM-PK-5002), which has medially inflicted, almost horizontal, backwards pointing teeth, emanating from a medially curved, partially horizontal maxilla.

The occurrence of the four bosses on the posterior edge of the left quadratojugal of Embrithosaurus is present in all cranial specimens of basal South African pareiasaurs, including specimens of B. baini (NHMUK R1971 and SAM-PK-5002), B. seeleyi (NHMUK 49426) and Nochelesaurus alexanderi (SAM-PK-6239). With an asymmetrical posterior boss arrangement, we apply the arrangement pattern of the left cheek as the condition in Embrithosaurus , comprising a large corner boss, followed dorsally by a small boss with large gaps above and below it, followed dorsally by two almost continuous large bosses. This arrangement pattern also occurs on the holotypes of B. baini (NHMUK R1971) and Nochelesaurus alexanderi (SAM-PK-6239).

Unreported by previous workers, Embrithosaurus (SAM-PK-8034) and Nochelesaurus (SAM-PK-6239) show a very flat, lateral surface of the quadratojugal, interrupted by distinct, high, pointed bosses. In contrast, in B. baini (NHMUK R1971) , there are no distinct, high bosses on the lower part of the lateral surface of the quadratojugal, which is instead thickened and swollen by several fused posterodorsally oriented rugose ridges.

As in most pareiasaurs, no septomaxilla could be identified on the holotype of Embrithosaurus . Tsuji et al. (2013) suggest that the single report of a septomaxilla in Anthodon by Boonstra (1934) is probably a misidentification, and argues that the septomaxilla has probably been lost in pareiasaurs.