A Carboniferous synapsid with caniniform teeth and a reappraisal of mandibular size-shape heterodonty in the origin of mammals

1.1. Present study

Here, we describe the sphenacodontian Shashajaia bermani gen. et sp. nov. from the C–P Halgaito Formation (Cutler Group) of southern Utah, USA and investigate its implications for the early evolution of mandibular heterodonty in synapsids. The material was collected from a conglomerate at the base of a multitaxic bonebed preserving faunal elements shared with Carboniferous assemblages in New Mexico [34,35], including Sagenodus copeanus, Edaphosaurus, Ophiacodon navajovicus, Eryops, Sphenacodon and limnoscelid diadectomorphs, among others. The vertebrate assemblages of the Cutler Group were some of the first terrestrial assemblages that included large-bodied vertebrate herbivores and specialist predators [36–38]. The dental morphology of the new taxon supports an expansion of tooth morphospace and pronounced size-shape heterodonty in the common ancestor of sphenacodontids and therapsids during late Carboniferous times. This ecomorphological diversification coincided with the late Palaeozoic remodelling of land-based food webs, new vertebrate dietary guilds, and fills a crucial gap in the synapsid fossil record.

2.1. Fossil specimens

The fossils herein referred to Shashajaia bermani were collected from the basal conglomerate at the Birthday Bonebed locality in Valley of the Gods, SE Utah, Carnegie Museum (CM) locality no. 3345 [35]. The specimens consist of a well-preserved left dentary with dentition collected in 2019 and an unassociated dentary fragment collected from the same lens by the authors in 2015 [39,40]. The fossils were initially prepared mechanically using pneumatic tools and pin vice and were further inspected using computed tomography at the University of Southern California Molecular Imaging Center. The fossils were µCT-scanned on a GE Phoenix Nanotom at 22 µm (CM 96529) and 36 µm (CM 91209) resolution at 120 kV and 100 µA.

2.2. Ecomorphological analyses

3.1. Systematic palaeontology

Synapsida [46]

Eupelycosauria [47]

Sphenacodontia [20]

Shashajaia bermani gen. et sp. nov. (figure 2; electronic supplementary material, figures S1–S3).

• Download figure
• Open in new tab
• Download PowerPoint

Etymology—’Berman's Bear heart.’ The genus name derives from the Navajo ‘shash’ (=bear) and ‘ajai’ (= heart). The species name honours David S Berman for his decades of research on fossils of sphenacodontians and others from the Bears Ears region of southern Utah, and which laid the foundation for the present study.

Holotype—CM 96529 (Carnegie Museum of Natural History, Pittsburgh), an isolated left dentary preserving the dentition (figure 2; electronic supplementary material, figures S1 and S2).

Referred specimen—CM 91209, partial dentary preserving portion of the postcanine toothrow (electronic supplementary material, figure S3).

Diagnosis—Small non-mammalian sphenacodontian that can be distinguished from others by the unique combination of characters: slender, gently bowed dentary that deepens slightly anteriorly near symphysis; shallow, lateral groove positioned posterodorsally on dentary just below the postcanine toothrow; at least 24 lower tooth positions; anterior incisor- and canine-like dentary teeth consecutively increase in size posteriorly, with the fourth lower tooth separated by a short, concave diastema and positioned on a raised buttress; the remaining postcanine teeth are greatly reduced in height relative to the depth of the dentary (as in basal therapsids). Shares with Haptodus, Ianthodon and Palaeohatteria triangular, anteroposteriorly wide posterior cheek teeth that lack mesiodistal cutting edges. Shares with the sphenacodontids Sphenacodon and Dimetrodon festooned infolding of plicidentine.

3.2. Detailed description and discussion

The more complete of the two dentaries (CM 96529) is 12 cm long as preserved, with an estimated total jaw length of about 15 cm assuming similar proportions to Palaeohatteria and Pantelosaurus [28]. The main body is gently bowed and shallowest at mid-length so that the mid-level of the alveolar margin is substantially lower than either the incisiform teeth or the coronoid eminence. As in sphenacodontids and some therapsids, the dentary deepens dorsoventrally toward the symphysis and the anterior teeth are elevated on a thickened platform above the level of the postcanines [48]. The lateral surface of the dentary is textured with numerous large neurovascular pits that connect to an internal canal—the inferior alveolar nerve canal—which spans most of the length of the dentary to about the level of the caniniform tooth. The posterior portion of the ramus bears a shallow but distinctive lateral groove just ventrolateral to the alveolar margin in both specimens. The medial surface bears a smooth, elongated fossa below the alveolar shelf that would have contributed to the Meckelian canal, and an anteroventral facet that would have accepted the splenial (figure 2b; electronic supplementary material, figure S2). As in Cutleria and therapsids, but unlike sphenacodontines, the splenial exposure probably would have been limited near the symphysis (based on its facet on the medial surface of the dentary).

Among the more striking features of Shashajaia are the prominent heterodont dentition, with well-developed anterior canine-like teeth that are up to 2.5 times taller than the postcanines. There are at least 24 preserved tooth positions in total. All of the teeth appear to exhibit a subthecodont implantation, seated slightly deeper on the medial side than on the lateral side, and with festooned infolding of plicidentine at the attachment site forming a ‘four-leaf clover’ cross-sectional shape of the pulp cavity as in Sphenacodon and Dimetrodon limbatus [21] (figure 2d). Similar festooned plicidentine has been reported in indeterminate materials from the late Carboniferous Ada Formation of Oklahoma (Ghzelian stage) ([39]: fig. 11). The first three dentary teeth in Shashajaia consecutively increase in size so that the third is the tallest, reminiscent of the pattern in Sphenacodon in which the third tooth position usually accommodates the largest tooth (electronic supplementary material, figure S9) [49,50]. There is a short diastema (less than 1 cm) between the third and fourth tooth positions (caniniforms). A diastema has been observed previously between the third and fourth tooth positions in some specimens of Haptodus garnettensis ([3]: fig. 10) and an indeterminate sphenacodontian from the Sangre de Cristo Formation ([51]: fig. 4). The fourth (caniniform) tooth is situated on a raised buttress of alveolar bone and is at least twice the height of the succeeding postcanine teeth. As in sphenacodontids and therapsids, the postcanines are greatly reduced in height relative to the dorsoventral depth of the dentary (less than 40%). The more posterior cheek teeth lack the typical ‘teardrop’ shape of sphenacodontids and are instead more similar in outline to those of Haptodus, Ianthodon and Palaeohatteria. The teeth are triangular with a slightly crooked tip and anteroposteriorly wide bases that nearly come into contact with preserved neighbouring teeth (see [28]: fig. 4 and [27]: fig. 16). However, in Shashajaia, the crowns are comparatively lower in aspect ratio and have a wider base, hyperbolic in outline. Distally, the teeth are medially compressed but are only weakly carinated and thus without mesiodistal cutting edges or serrations (figure 2c), suggesting they were probably not important for shearing or tearing.

3.3. Morphometric results

The phylogenetic position of Shashajaia among Sphenacodontoidea presents an opportunity to assess the diversification of tooth morphospace in synapsids prior to the origin of therapsid-like heterodonty. Results of the morphometric analysis indicate that the first two PCs represent the overwhelming majority of tooth shape variation (71%) and were therefore used in subsequent plots to distinguish both (i) morphological variation along the toothrow (figure 3a; electronic supplementary material, figure S6) and (ii) overall variation within different taxonomic groups (figure 3c; electronic supplementary material, figures S7 and S8). PC1 encompassed the majority of tooth shape variation (58.2%) and chiefly reflects variation in tooth aspect ratio, which forms a continuum ranging from tall, slender (more negative) to short, stout (more positive) tooth crowns (figure 3b). This continuum along PC1 generally marks the sequential change in tooth shape from the front to the back of the toothrow (electronic supplementary material, figure S7) and shows a weak but negative correlation with tooth size in sphenacodontians (electronic supplementary material, figure S8). PC2 (12.8%) reflects overall tooth curvature, illustrating whether the curve manifests across the entire crown (more negative) or is concentrated towards the apex (more positive) (figure 3b). Overall, therapsids occupy a greater area of morphospace than more basal (pelycosaur-grade) synapsids (figure 3a,b). Dinocephalians exhibit the greatest morphospace occupation of all subclades, although therocephalians are more widely distributed across PC2 (electronic supplementary material, figure S7). Although sphenacodontians display the greatest dental disparity of all pelycosaur-grade synapsids, Shashajaia's morphospace corresponds even more closely with therapsids as it shows a similarly broad distribution across PC1 (electronic supplementary material, figures S6 and S7).

• Download figure
• Open in new tab
• Download PowerPoint

In addition to filling a greater portion of the tooth morphospace, therapsids generally exhibited more positive PC1 values throughout the toothrow and with a significant increase (lower aspect ratio) toward the back of the row than in the more basal pelycosaur-grade synapsids where the teeth are comparatively more homodont (figure 3a; electronic supplementary material, figure S6). Variance in PC1 across an individual toothrow—a quantitative index for shape heterodonty—was therefore also greater in most therapsid clades than in basal synapsid groups (figure 3c); the highest therapsid PC1 variance was found among biarmosuchians, dinocephalians, gorgonopsians and cynodonts, and the lowest therapsid PC1 variance among non-dicynodont anomodonts and small therocephalian insectivores (e.g. Tetracynodon). Remarkably, PC1 variance in Shashajaia substantially overlapped the greatest therapsid values, probably reflecting the strong functional heterodonty of the tall, slender incisor and canine teeth versus the short, triangular postcanines. Size variation was loosely associated with variations in tooth aspect ratio, with the anterior teeth on average having a larger relative size than the more posterior teeth in both basal synapsids and therapsids. Primitively, the lower postcanine field of therapsids is known to show a marked decrease in tooth height relative to the overall jaw depth [48] beginning at about the level of the fifth or sixth tooth position (electronic supplementary material, figure S6), although sphenacodontids also exhibit comparatively short postcanine crowns relative to the dentary depth. Nevertheless, in the majority of basal synapsids, a functional tooth in any given position is not significantly smaller than its preceding tooth, but Shashajaia with its diminutive postcanines is more similar to the therapsids in this respect (electronic supplementary material, figures S6 and S7).

Time-series data (figure 3d) show a discordance between patterns of dental morphology and body size evolution in Palaeozoic synapsids. Mean body size gradually increased in both synapsid herbivores and carnivores from the late Carboniferous to Middle Permian with a short-term decrease at the Early–Middle Permian boundary (figure 3d ‘Olson's extinction’), corroborating similar findings by Brocklehurst & Brink [22] and Brocklehurst et al. [54]. Though initially lower on average than sympatric carnivores, mean herbivore size surpassed that of carnivores during the Early Permian, a possible ecological consequence of predator pressure on C–P herbivores and overall greater numbers of large herbivores during the late Early Permian. Nevertheless, expansion of body size disparity was not accompanied by similar increases in PC1 variance. In fact, mean PC1 variance was relatively stable or slightly decreased for each geologic stage from the Kasimovian to the Kungurian. Among late Carboniferous taxa, Shashajaia showed the highest PC1 variance (0.026). Among all of the taxa sampled, comparable PC1 variance was only re-encountered later in the Middle Permian therapsids (Capitanian stage), which included the largest known Palaeozoic synapsid herbivores and carnivores: the dinocephalians (100–1000 kg). Given this temporal lag and the inversion of large herbivores relative to carnivores, it is unlikely that regional specialization of tooth function in Shashajaia and the first therapsids were influenced by the same ecological factors. Moreover, the Halgaito assemblage from which Shashajaia originated includes abundant aquatic and semi-terrestrial taxa that proportionately outnumbered the fully terrestrial taxa (in contrast with the assemblage of the Permian Organ Rock Formation which was progressively more terrestrial) [35,55].

The comparisons outlined above bring into focus the ecological backdrop under which therapsid-like heterodonty evolved. During the C–P transition, herbivore richness and body size disparity dramatically increased [56] which, in addition to climate change, may have impacted plant species richness during Permian times [54]. Simultaneously, synapsid predators like the sphenacodontians showed increasing body size disparity [22,23], and functional innovations in their marginal dentition suggest underappreciated dietary diversity among the group ranging from generalist faunivores to more specialized apex predators [21]. While it is often assumed that sphenacodontids preyed on large synapsid herbivores, like Edaphosaurus [20], there is ample evidence that sphenacodontids originated as small-bodied faunivores—as in a documented predator–prey association between Dimetrodon milleri and the amphibian Zatrachys [20]. Shashajaia reflects this early sphenacodontian ecology, with its relatively gracile, upwardly curved dentary (figure 2) suggesting low stress, low power jaw functionality focused on speed and unsuited to extended struggles with prey during feeding [57]. The dental modification in Shashajaia probably compensates for the low-biting efficiency at the tip of the dentary, using instead the high velocity at this part of the jaw [6,7] alongside the enlarged anterior teeth to maximize the impact on the prey and penetrate deep into its flesh [40,58]. Given pelycosaur-grade synapsids were not particularly agile [20], this likely jaw function suggests Shashajaia was an ambush predator that fed on smaller animals, which could be quickly caught and swallowed without struggle. Thus, prior to the establishment of large herbivore-dominated tetrapod communities, as consumers shifted from aquatic to terrestrial feeding, C–P sphenacodontoids like Shashajaia most likely used their enlarged anterior dentition primarily for gripping and puncturing small to mid-sized prey, including perhaps small reptiles, amphibians and fishes [4,59,60]. Though the dentitions of Carboniferous taxa appear to have relatively smooth mesiodistal edges, the canine and antecanine dentition became modified at a later stage with serrations and denticles, forming a ‘ziphodont’ dentition. Ziphodonty in some large-bodied Dimetrodon species [21] and numerous therapsids [61] might indicate more efficient processing of protein in active terrestrial animals that had relatively greater metabolic requirements than the earliest sphenacodontians. Prior to this, the emphasis on prey capture may reflect increasing environmental heterogeneity as the Carboniferous drew to a close [62,63], with seasonal fluvial palaeoenvironments like that of the Birthday Bonebed [35] providing an ideal mix of terrestrial and semi-aquatic prey that allowed basal synapsid faunivores to experiment with varying their exposure to aquatic resources, enabling their trophic ecologies to become more firmly planted in the terrestrial realm.