Taxonomic identity

Various apomorphies or differential features identified by previous authors, and most recently by Gee et al. (Reference Gee, Parker and Marsh2020), for Anaschisma browni (e.g., alary process of the premaxilla, anteroposteriorly short lacrimal with broad contribution to the orbit) permit confident referral of SMP VP-44 to this taxon. The latter feature is particularly important in separating this material from the only other diagnostic metoposaurid material on the eastern seaboard—the small, referred specimen of “Metoposaurus” bakeri from Nova Scotia (first referred by Baird, Reference Baird1986, and accepted by all other authors; see Sues and Olsen, Reference Sues and Olsen2015, fig. 12, for photograph), which lacks this contact, like its conspecifics from Texas. While identification of the Zions View material is in agreement with previous authors’ interpretations dating back to at least the 1990s, it bears noting that the specimen has never been properly described, with justification for taxonomic referral, either implicitly by way of interpretive figures or explicitly by documentation of apomorphic features or unique combinations of characters. Such justification is essential for ensuring rigorous propagation of data that can be independently assessed by other workers in some fashion. Large temnospondyl material from the east coast of North America should be scrutinized more carefully for several reasons: (1) these localities were spatially closer to other metoposaurid-bearing regions in Europe and Morocco at that time (not coeval at present, but possible spatiotemporal range extensions); (2) these localities are less well-constrained stratigraphically and temporally; and (3) the east coast is the only region in North America where a large-bodied, non-metoposaurid taxon (Calamops paludosus) occurs in the Late Triassic (Sues and Schoch, Reference Sues and Schoch2013).

The remainder of the material can only be referred based on circumstantial evidence and by the absence of any features that clearly differ from either Metoposauridae (for which most diagnostic features are cranial; e.g., Buffa et al., Reference Buffa, Jalil and Steyer2019) or Anaschisma browni specifically. For example, there are no diagnostic features of the clavicle per the diagnosis of Buffa et al. (Reference Buffa, Jalil and Steyer2019); stereospondyls generally have similar interclavicular morphology and ornamentation (e.g., Warren and Snell, Reference Warren and Snell1991). SMP VP-515 is nearly complete, so the subtler nuances in contour allow for differentiation from even the closely related Middle Triassic metoposauroid Callistomordax kugleri Schoch, Reference Schoch2008, which has a narrower, more rhomboidal interclavicle with prominent equant anterior and posterior styluses. SMP VP-514 is much more fragmentary, but it at least differs in no overlapping regions from SMP VP-515. Beyond this, metoposaurid interclavicles are essentially differentiated only by the size of the circular pitted region on the ventral surface by a qualitative binary metric: small or large. Anaschisma browni and “Metoposaurus” bakeri typically have been characterized by relatively large circular pitted regions (e.g., Case, Reference Case1922, Reference Case1931, Reference Case1932; Sulej, Reference Sulej2002, Reference Sulej2007; though see Lucas et al., Reference Lucas, Rinehart, Heckert, Hunt and Spielmann2016, for a contrary example from Rotten Hill), like in SMP VP-515, although the intraspecific variability remains to be better assessed (e.g., Spielmann and Lucas, Reference Spielmann and Lucas2012). From a differential sense, the interclavicles can thus only be differentiated from some taxa and compared favorably with others. Stereospondyl clavicles are arguably even more highly conserved than interclavicles, and the only diagnostic features of Metoposauridae (following Buffa et al., Reference Buffa, Jalil and Steyer2019) relate to the clavicular blade and the articulation with the interclavicle. The latter cannot be assessed here, and the former refers to an intraspecifically variable indentation in the medial edge (Sulej, Reference Sulej2007) that also occurs in brachyopoids (Warren and Marsicano, Reference Warren and Marsicano2000), which is not completely preserved in either SMP VP-360 or SMP VP-516. Even if a metoposaurid identity is accepted, isolated clavicles usually cannot be referred to a particular taxon. Size of circular ornamentation has sometimes been used to differentiate taxa (e.g., Hunt, Reference Hunt1993; Sulej, Reference Sulej2002). One feature noted by Gee et al. (Reference Gee, Parker and Marsh2020) is the presence or absence of a sensory groove on the clavicle (absent in A. browni, but present in Panthasaurus maleriensis Chowdhury, Reference Chowdhury1965). The absence of this feature on SMP VP-360 and SMP VP-516, therefore, at least partially excludes some taxa from consideration.

The hemimandible (SMP VP-37) warrants particular attention because this is the only region of the skeleton that is known from the trematosauroid Calamops paludosus from the lower Stockton Formation in Pennsylvania (Sues and Schoch, Reference Sues and Schoch2013). Of the four diagnostic features listed by Sues and Schoch (Reference Sues and Schoch2013), the only one that can be assessed in SMP VP-37 is the presence of rows of deep pits on the symphysis. Although there are pits on the symphysis of SMP VP-37, this is a common feature of metoposaurids as a continuation of the ventrolabial ornamentation pattern, and these pits are neither deep nor arranged in rows. Other differential features of C. paludosus include the absence of adsymphyseal teeth (found in most metoposaurids; Konietzko-Meier and Wawro, Reference Konietzko-Meier and Wawro2007; Sulej, Reference Sulej2007) and the absence of a well-ossified surface medial to the symphyseal “fang.” SMP VP-37 differs in no regards from other metoposaurids, which have a highly conserved mandibular morphology (see Brusatte et al., Reference Brusatte, Butler, Mateus and Steyer2015, for some possible differentiating features).

Based on the conventional resemblance-based method of identification, there is little reason to suspect that all temnospondyl material from Zions View belongs to anything other than a metoposaurid. More specifically, based on the taxonomic affinities of the skull, probably all of the material belongs to Anaschisma browni. No attributes of any given specimen are inconsistent with this interpretation. Our referrals are thus based primarily on four lines of circumstantial evidence: (1) the near total absence of large-bodied non-metoposaurids in the Late Triassic of North America, (2) the tendency for metoposaurid-bearing deposits to be monotaxic with respect to temnospondyls, (3) the apomorphy-based identification of SMP VP-44 (skull), and (4) our ability to differentiate SMP VP-37 from Calamops paludosus. These lines of evidence have historically been regarded as relatively robust, or at least to be the most viable given the nature of the fossil record.

Lastly, we note that while our detailed study agrees with previously published interpretations regarding the taxonomy of the Zions View metoposaurid material, this study is the first to provide any justification or corresponding evidence for this interpretation. There are inherent limitations to studies that report occurrences without providing the relevant data to support the conclusions. Furthermore, the recognition of C. paludosus as a non-metoposaurid trematosauroid (Sues and Schoch, Reference Sues and Schoch2013) indicates the presence of at least one other large-bodied stereospondyl at a local geographic scale and underscores the need for explicit rationale for taxonomic assignment of material that does not rely largely on absolute size alone. The relatively poor record of temnospondyls on the eastern seaboard only further underscores the point that metoposaurid affinities cannot be assumed for all large specimens. Detailed re-study and re-description of historic specimens is important both for the sake of reproducibility (because taxonomic identifications are hypotheses rather than permanent, static facts) and for facilitating future studies by other workers who may be unable to personally examine the material but who may be interested in incorporating either qualitative (e.g., comparative anatomy) or quantitative (e.g., morphometrics) data from these specimens.

Ontogenetic maturity

Ontogenetic assessment of metoposaurid specimens remains of interest given the debate over Apachesaurus gregorii, which is represented only by relatively small individuals. Based strictly on size alone, SMP VP-44 was at least a sub-adult individual (midline length of 42 cm); metoposaurid skulls rarely exceed 55 cm in length (e.g., see measurements provided by Colbert and Imbrie, Reference Colbert and Imbrie1956; Dutuit, Reference Dutuit1976; Sulej, Reference Sulej2007; Lucas et al., Reference Lucas, Rinehart, Heckert, Hunt and Spielmann2016). Most patterns of metoposaurid ontogeny hinge on the interpretation of A. gregorii because it comprises most of the very small (<20 cm skull length) metoposaurid specimens (Gee and Parker, Reference Gee and Parker2020, table 1). Furthermore, because metoposaurids were paedomorphic, with underdeveloped neurocrania and postcrania, qualitative attributes used for non-paedomorphic temnospondyls (e.g., degree of ossification) are not useful here. These issues have been previously addressed and are not repeated here. The observation that the sutures of the posterior skull have separated slightly does suggest some immaturity of SMP VP-44. Beyond this, no other feature confers robust information one way or the other.

Intraspecific variation within Anaschisma browni.—Metoposaurids are among the most abundantly known Mesozoic temnospondyls, with dozens of known complete skulls among thousands of other fragmentary specimens. Such a sample size is relatively rare for extinct tetrapods and makes metoposaurids an excellent case study for studying intraspecific variation. There are several reasons why the documentation of intraspecific variation is critical for paleontological research at large. Firstly, increased sample size and hypothesized intraspecific variation permits continuous assessment of qualitative taxonomic diagnoses and differentiation—there are numerous examples across the tree of life of taxa named for specimens whose morphological differentiators are now recognized as the result of intraspecific variation (whether due to ontogeny or to other factors). This consideration is particularly salient for metoposaurids because the standard of morphological differentiation that warranted taxonomic distinction within the clade was historically quite low (e.g., Branson and Mehl, Reference Branson and Mehl1929). As such, these narrow typological standards led to an extensive number of taxa erected on minor or single differences that were often questionably valid or that ignored non-taxonomic explanators (e.g., taphonomic distortion, ontogenetic transformation). For example, features such as “eyes not as far forward” (Branson, Reference Branson1905, p. 588–589) for “Anaschisma brachygnatha” (=Anaschisma browni), “clavicle of striking characteristics” (Branson and Mehl, Reference Branson and Mehl1929, p. 233) for “Buettneria major” (=A. browni), and “small size” (Schoch and Milner, Reference Schoch and Milner2000, p. 121–122) for both Apachesaurus gregorii and Arganasaurus lyazidi Dutuit, Reference Dutuit1976 have been previously utilized to diagnose taxa. In the same vein, modern taxonomy, informed by phylogenetics, must also account for variation in scoring of phylogenetic matrices. In the context of the Metoposauridae, this consideration is important given the markedly different topological results of three recent studies (Buffa et al., Reference Buffa, Jalil and Steyer2019; Chakravorti and Sengupta, Reference Chakravorti and Sengupta2019; Gee et al., Reference Gee, Parker and Marsh2020). Some of the disparity results from differing characters (especially the use of predominantly quantitative characters by Chakravorti and Sengupta, Reference Chakravorti and Sengupta2019), but differing scoring (i.e., whether or not polymorphisms were scored when noted in the literature) should also be considered. Not scoring polymorphisms when they are known may contribute to artificially well-resolved topologies because characters appear more discretely informative than they truly are.

With respect to intraspecific variation, the most studied metoposaurid is Metoposaurus krasiejowensis Sulej, Reference Sulej2007. Sulej (Reference Sulej2007) identified several examples of intraspecifically variable cranial sutures, such as the relationship of the lacrimal to the orbit (a longstanding phylogenetic character), contact between the postorbital and the parietal, and contact between the prefrontal and the maxilla. In all three cases, the condition is documented to vary at the level of the individual. Furthermore, Teschner et al. (Reference Teschner, Sander and Konietzko-Meier2018) examined a broad range of humeri and identified distinct histotypes, while Antczak and Bodzioch (Reference Antczak and Bodzioch2018) identified variation in ornamentation patterns on the skull and pectoral girdle. Such variation would be further expanded if Sulej's (Reference Sulej2002) original subspecies designation were restored (i.e., Metoposaurus diagnosticus krasiejowensis vs. M. d. diagnosticus von Meyer, Reference von Meyer1842), or if his species-level framework was utilized without his subspecies addition (i.e., Metoposaurus krasiejowensis = M. diagnosticus).

In contrast, the published osteology of most metoposaurids is based exclusively or primarily from single localities, which in some instances might represent a single event that records a single population. Spatial restriction is true for Metoposaurus krasiejowensis and M. algarvensis Brusatte et al., Reference Brusatte, Butler, Mateus and Steyer2015 in Europe (Sulej, Reference Sulej2007; Brusatte et al., Reference Brusatte, Butler, Mateus and Steyer2015), all three of the Moroccan taxa (Arganasaurus azerouali Dutuit, Reference Dutuit1976, A. lyazidi, Dutuitosaurus ouazzoui; Khaldoune et al., Reference Khaldoune, Jalil, Germain and Steyer2016; Buffa et al., Reference Buffa, Jalil and Steyer2019), and the Indian taxon (Panthasaurus maleriensis; Chakravorti and Sengupta, Reference Chakravorti and Sengupta2019). The North American Apachesaurus gregorii has been described from many localities (Long and Murry, Reference Long and Murry1995), but most of this material can only be said to belong to small-bodied individuals and is not referable under an apomorphy-based system, notwithstanding doubts about the validity of the original interpretation of the taxon as a metoposaurid with a distinctly smaller adult body size (Gee and Parker, Reference Gee and Parker2020). Elimination of the single diagnostic postcranial feature, elongate intercentra (following Gee et al., Reference Gee, Parker and Marsh2017; Gee and Parker, Reference Gee and Parker2020), substantially reduces the number of properly referred specimens of Ap. gregorii. Thus, only Metoposaurus diagnosticus, “Metoposaurus” bakeri, and Anaschisma browni are confidently known from several localities that span an appreciable geographic range. It is worth noting that indeterminate material from localities in Algeria, France, Italy, and Madagascar (in addition to already known metoposaurid-bearing localities or regions) could hypothetically belong to known taxa and thus extend geographic ranges if restudied (Koken, Reference Koken1913; Corroy, Reference Corroy1928; Leonardi, Reference Leonardi1952; Lehman, Reference Lehman1971; Dutuit, Reference Dutuit1978; Cuny and Ramboer, Reference Cuny and Ramboer1991; Fortuny et al., Reference Fortuny, Arbez, Mujal and Steyer2019). Metoposaurus diagnosticus is confined to Germany (Schoch and Milner, Reference Schoch and Milner2000; Sulej, Reference Sulej2002, Reference Sulej2007; Milner and Schoch, Reference Milner and Schoch2004), while “M.” bakeri is known almost exclusively from Texas, with only the additional single specimen in Nova Scotia (Baird, Reference Baird1986; Sues and Olsen, Reference Sues and Olsen2015). Anaschisma browni is definitively recorded from at least five states (Arizona, New Mexico, Pennsylvania, Texas, Wyoming) and numerous localities within each state that collectively form a sizeable range within North America. Documented intraspecific variation in this taxon is thus more likely to capture a fuller range of variation given that even coeval localities in close geographic proximity probably represent distinct populations, let alone those from geographically distant localities.

SMP VP-44 exhibits essentially no deviations from the typical morphology of Anaschisma browni (Case, Reference Case1922; Lucas et al., Reference Lucas, Rinehart, Heckert, Hunt and Spielmann2016). The skull is relatively symmetrical, although slight transverse distortion of the skull has resulted in a slight offset along the longitudinal axis of major features (e.g., anterior orbital margin). The course of the lateral line grooves is also consistent with previous descriptions, although it may be noted that the post-orbital groove is angled more sharply posteromedially as it passes from the jugal onto the postorbital and then onto the supratemporal; in some other individuals, the groove passes medially onto the postorbital and then posteriorly onto the supratemporal. This variation falls within the range previously reported for A. browni and is not considered taxonomically meaningful in this instance. It may be noted (only in a qualitative sense) that specimens of A. browni from the Popo Agie Formation have proportionately larger nares that are nearly the same size as the orbits (e.g., Branson and Mehl, Reference Branson and Mehl1929; Gee et al., Reference Gee, Parker and Marsh2020) when compared to conspecifics from more near-equatorial latitudes in which the nares are distinctly smaller than the orbits (e.g., Case, Reference Case1922; Lucas et al., Reference Lucas, Rinehart, Heckert, Hunt and Spielmann2016). The importance of this, if it is biologically important, is unclear. The relative size of the nares and the orbits in SMP VP-44 is somewhat intermediate to the end members of A. browni, but this should not be interpreted as a suggestion of gradational evolution associated with biogeographic trajectories. The absence of any substantial morphological variation across a broad geographic range in A. browni is not necessarily surprising given the morphological conservatism across metoposaurids in general, but this description provides clear, positive evidence for a hypothesis of intraspecific conservatism.

Temnospondyl turnover in North America

The Late Triassic temnospondyl record of North America is notable for two attributes. Firstly, North America preserves the most complete and diverse record of small-bodied temnospondyls from the Late Triassic (Wilson, Reference Wilson1948; Bolt and Chatterjee, Reference Bolt and Chatterjee2000), including putative stem caecilians (Pardo et al., Reference Pardo, Small and Huttenlocker2017; So et al., Reference So, Kufner, Huttenlocker, Pardo and Lovelace2018) and unequivocal crown batrachians (Stocker et al., Reference Stocker, Nesbitt, Kligman, Paluh, Marsh, Blackburn and Parker2019). Small-bodied stereospondyls are rare in any setting in the Mesozoic, and their ecology and phylogenetic relatedness to Lissamphibia remain to be more comprehensively explored. Secondly, the record of large-bodied taxa is conspicuously dominated by metoposaurids, a pattern that otherwise occurs only in regions with a substantially smaller history and body of collection of both metoposaurids and of terrestrial tetrapods at large (e.g., Morocco, Madagascar). As noted above, the only record of large-bodied, non-metoposaurid temnospondyls from the Late Triassic of North America is Calamops paludosus from the lower Stockton Formation (Solesbury Member; Carnian). Assuming that the correlation of the middle-upper New Oxford Formation with the Lockatong Formation is correct, C. paludosus likely pre-dates all metoposaurid occurrences in eastern North America. However, we note that the specimens collected near Emigsville (including AMNH FARB 1850 and 1863) are potentially from near the base of the New Oxford Formation, which may make them close to temporally equivalent with C. paludosus.

The longstanding perception of metoposaurid exclusivity remains true in the Late Triassic of western North America, which is also the source of the diminutive temnospondyl taxa. In contrast, the Middle Triassic record of temnospondyls in this region is substantially more diverse, similar to other geographic regions, and preserves brachyopoids, non-metoposaurid trematosauroids, and capitosauroids (e.g., Welles and Cosgriff, Reference Welles and Cosgriff1965; Schoch, Reference Schoch2000; Rinehart et al., Reference Rinehart, Lucas and Schoch2015). All three of these clades persist into the Late Triassic in other geographic regions where they sometimes co-occur with metoposaurids, as in western Europe (e.g., Milner and Schoch, Reference Milner and Schoch2004; Sulej and Majer, Reference Sulej and Majer2005), or succeed metoposaurids, as in India (e.g., Sengupta, Reference Sengupta2003). The extensive sampling in the southwestern region of the United States for more than a century suggests that this stark faunal disparity is not the result of poor sampling, but rather that metoposaurids were genuinely the only large-bodied temnospondyls in these basins by the Late Triassic. However, the fossil record of this transition is not captured in this part of North America, with the regional Tr-3 unconformity representing a several-million-year gap between deposition of the Holbrook Member of the Moenkopi Formation (Anisian) and the next occurrences of temnospondyls in the late Carnian of Wyoming and Texas. There are no definitive records of temnospondyls from the Ladinian of North America, although some of the other poorly constrained horizons on the east coast cannot be totally excluded from being Ladinian (e.g., the Economy Member of the Wolfville Formation; Sues and Olsen, Reference Sues and Olsen2015). Consequently, the mechanisms that contributed to the development of a metoposaurid zone of exclusivity by the latest Carnian (Otischalkian LVF) in North America remain unclear. Climatic variation along paleolatitudinal gradients (e.g., Sues and Olsen, Reference Sues and Olsen2015) and climatic instability (e.g., Whiteside et al., Reference Whiteside, Lindström, Irmis, Glasspool, Schaller, Dunlavey, Nesbitt, Smith and Turner2015) are popular hypotheses invoked for explaining the distribution of amniotes (e.g., Dunne et al., Reference Dunne, Farnsworth, Greene, Lunt and Butler2020), such as the absence of the herbivorous sauropodomorph dinosaurs, but hypotheses specifically regarding the distribution of large-bodied stereospondyls have not been formulated. A large body of recent work has emerged examining the role of the Carnian Pluvial Episode (CPE) in shaping terrestrial tetrapod evolution, but few have explicitly addressed temnospondyls. Buffa et al. (Reference Buffa, Jalil and Steyer2019) and Fortuny et al. (Reference Fortuny, Arbez, Mujal and Steyer2019) suggested a correlation between metoposaurid diversification and the CPE, which has been increasingly linked to major events of Triassic tetrapod evolution (e.g., Simms and Ruffell, Reference Simms and Ruffell1989, Reference Simms and Ruffell1990; Ruffell et al., Reference Ruffell, Simms and Wignall2016; Dal Corso et al., Reference Dal Corso, Bernardi, Sun, Song, Seyfullah, Preto and Gianolla2020), particularly to the radiation of dinosaurs (e.g., Benton et al., Reference Benton, Bernardi and Kinsella2018; Bernardi et al., Reference Bernardi, Gianolla, Petti, Mietto and Benton2018; Mancuso et al., Reference Mancuso, Benavente, Irmis and Mundil2020). However, as those workers also noted, it remains to be tested how this climate event would have favored some clades of large-bodied stereospondyls over others; most studies remain focused on amniotes. The occurrence of the non-metoposaurid metoposauroid Callistomordax kugleri in the upper Ladinian (Schoch, Reference Schoch2008) confers another line of evidence that the timing of metoposaurid origination and diversification may be tied to global events in the middle Carnian.

A likely possibility that would reconcile this uncertainty is that presently un- or underrecognized physiological disparity between taxa is a major explanator of Mesozoic temnospondyl distribution (e.g., Muhkerjee et al., Reference Mukherjee, Ray and Sengupta2010). It has long been known that the majority of Mesozoic taxa are medium- to large-bodied aquatic taxa, but it has not been shown that they were all ecologically similar; the frequent colloquial parallel to extant crocodilians may obfuscate the true ecological diversity of large stereospondyls. Indeed, finite element analysis (FEA) of various stereospondyls to infer feeding behavior has identified the potential for ecological differentiation among the distinct morphotypes (e.g., Fortuny et al., Reference Fortuny, Marcé-Nogué, de Esteban-Trivigno, Gil and Galobart2011, Reference Fortuny, Marcé-Nogué, Gil and Galobart2012, Reference Fortuny, Marcé-Nogué, Steyer, de Esteban-Trivigno, Mujal and Gil2016, Reference Fortuny, Marcé-Nogué and Konietzko-Meier2017; Lautenschlager et al., Reference Lautenschlager, Witzmann and Werneburg2016; Konietzko-Meier et al., Reference Konietzko-Meier, Gruntmejer, Marcé-Nogué, Bodzioch and Fortuny2018). Additionally, an increasing body of histological literature has identified disparate physiological capabilities and growth patterns among closely related taxa (e.g., Konietzko-Meier and Klein, Reference Konietzko-Meier and Klein2013; Muhkerjee et al., Reference Mukherjee, Sengupta and Rakshit2020), indicating differing responses to local environmental conditions. Metoposaurids are actually the best sampled clade in this regard. Most of the comparative work has focused on disparities between the Moroccan Dutuitosaurus ouazzoui (Steyer et al., Reference Steyer, Laurin, Castanet and de Ricqlès2004) and the Polish Metoposaurus krasiejowensis (Konietzko-Meier and Klein, Reference Konietzko-Meier and Klein2013; Konietzko-Meier and Sander, Reference Konietzko-Meier and Sander2013). The former appears to have been under more pronounced environmental stress that forced it to aestivate for a prolonged time (evidenced by discrete lines of arrested growth) compared to much shorter, sporadic cessations in growth in the latter. Teschner et al. (Reference Teschner, Sander and Konietzko-Meier2018) further documented two distinct histotypes within M. krasiejowensis that implied distinct environmental conditions for perhaps two distinct local populations. Most recently, histological data from the relatively high paleolatitude Indian taxon, Panthasaurus maleriensis, indicates a similarly temperate climate to that of Poland, with a similar environmental signal to that found in M. krasiejowensis (Teschner et al., Reference Teschner, Chakravorti, Sengupta and Konietzko-Meier2020).

Other clades of large-bodied stereospondyls are generally not as well-sampled histologically, thereby precluding robust comparisons with Metoposauridae. However, a number of studies do hint at cryptic ecological and physiological differentiation that could in turn explain the different geographic ranges of various clades. Muhkerjee et al. (Reference Mukherjee, Ray and Sengupta2010, Reference Mukherjee, Sengupta and Rakshit2020) suggested that some paracylotosaurid capitosaurs were more terrestrially capable than other stereospondyls, perhaps indicating a more shallow-water spatial occupation when in water. Some differences could also relate to the tolerance of various taxa. For example, the ability to tolerate differing environments seems to have been particularly high for the long-ranging plagiosaurid Gerrothorax pulcherrimus Nilsson, Reference Nilsson1934, and it has been suggested that this taxon was a “pioneer species” that inhabited more unstable areas in which other taxa could not survive (Sanchez and Schoch, Reference Sanchez and Schoch2013). Similarly, greater tolerance for cold water temperatures has been suggested to explain the persistence of post-Triassic brachyopoids at high paleolatitudes by means of exclusion of crocodilians (Warren et al., Reference Warren, Rich and Vickers-Rich1997). Most other clades remain very poorly sampled—there is essentially no modern histology on non-metoposaurid trematosaurs, for example—but the available data already demonstrate some apparent differences between and within clades. Integrating these data with the observable external anatomy would likely elucidate a clearer narrative regarding stereospondyl distribution.

In lieu of histological data for most metoposaurids, the paleolatitudinal range of these clades can provide alternative insight. All metoposaurids ranged from ~35°N to 35°S (see Brusatte et al., Reference Brusatte, Butler, Mateus and Steyer2015, fig. 13), as demonstrated by the central European and Indian populations, which bracket this range. Other stereospondyls are known from higher paleolatitudes in both hemispheres, particularly brachyopoids, although Late Triassic temnospondyls as a whole were more abundant at the lower latitudes where presumably seasonality was less pronounced (Dunne et al., Reference Dunne, Farnsworth, Greene, Lunt and Butler2020), much like lissamphibians today (Pyron and Wiens, Reference Pyron and Wiens2013). The North American metoposaurid-bearing basins, specifically those in the southwestern United States, would have been the closest to the equator; the Argana Basin in Morocco lay intermediate between these basins and those of central Europe. In both North America and Morocco (the western part of metoposaurids’ range), there is no evidence for other large-bodied stereospondyl taxa co-existing with or succeeding metoposaurids. The general paleoclimate of the southwestern United States has been well studied from a variety of perspectives, especially in the Chinle Formation (e.g., Dubiel et al., Reference Dubiel, Parrish, Parrish and Good1991, Reference Dubiel, Hasiotis, Davidson, Leleu and North2011; Atchley et al., Reference Atchley, Nordt, Dworkin, Ramezani, Parker, Ash and Bowring2013; Baranyi et al., Reference Baranyi, Reichgelt, Olsen, Parker and Kürschner2018; Jin et al., Reference Jin, Dworkin, Atchley and Nordt2018; Rasmussen et al., Reference Rasmussen, Mundil, Irmis, Geisler, Gehrels, Olsen and Kent2020) and indicates a humid climate during deposition of the lower units, with a progressively arid and seasonally pronounced shift throughout the Norian. Similar seasonality has been inferred from histological data of the Moroccan metoposaurids (Steyer et al., Reference Steyer, Laurin, Castanet and de Ricqlès2004), while the Polish and Indian taxa seem to have been subjected to a comparatively mild climate that did not lead to growth cessations (Konietzko-Meier and Klein, Reference Konietzko-Meier and Klein2013; Teschner et al., Reference Teschner, Chakravorti, Sengupta and Konietzko-Meier2020). This physiological variation is most likely tracking environmental variation since metoposaurid anatomy is so highly conserved, suggesting that the paleolatitudinal restriction of metoposaurids at large also owes to environmental barriers, whether by availability of corridors for dispersal or through various aspects of the climate (e.g., seasonality, ephemerality of freshwater bodies).

Considering the sampling in Late Triassic terrestrial ecosystems, including those that produce other temnospondyl remains, the apparent latitudinal restriction of metoposaurids seems fairly robust as a crude approximation of the true range. In this vein, the recent report of metoposaurid “osteoderms” from Zimbabwe (~40°S) by Barrett et al. (Reference Barrett, Sciscio, Viglietti, Broderick, Suarez, Sharman and Jones2020) warrants some discussion here. This is an intriguing novel occurrence, not yet addressed by temnospondyl workers, that has already been propagated by several studies (Dunne et al., Reference Dunne, Farnsworth, Greene, Lunt and Butler2020; Sciscio et al., Reference Sciscio, Viglietti, Barrett, Broderick, Munyikwa, Chapelle, Dollman, Edwards, Zondo and Choiniere2020) and has been utilized as a line of evidence for broader claims, such as biome analogy with geographically disparate deposits (Barrett et al., Reference Barrett, Sciscio, Viglietti, Broderick, Suarez, Sharman and Jones2020). However, there are several reasons to exercise caution with this purported occurrence. Firstly, metoposaurids do not have osteoderms (Sciscio et al., Reference Sciscio, Viglietti, Barrett, Broderick, Munyikwa, Chapelle, Dollman, Edwards, Zondo and Choiniere2020 more generically termed these “dermal bones”), and the photographs indicate only flat elements with circular ornamentation. These appear to be pectoral in nature, which, as previously noted here, are very hard to properly refer in isolation and when fragmentary. The authors neither described the material (it is only figured in the supplementary data) nor justified their taxonomic identification and, most crucially, the specimens were not collected. Reported occurrences without voucher specimens are problematic for the sake of reproducibility (Bell et al., Reference Bell, Gauthier and Bever2010). Circumstantial evidence (e.g., regional monotaxicity, temporal occurrence) does not lend particular credence to a specific temnospondyl clade in the same way that it does for North America (e.g., chigutisaurids occur in both the lower and upper horizons of the Elliot Formation of South Africa; Warren and Damiani, Reference Warren and Damiani1999). Finally, phytosaurs occur at high paleolatitudes of both hemispheres without metoposaurids (e.g., Bonaparte, Reference Bonaparte1966; Brusatte et al., Reference Brusatte, Butler, Niedźwiedzki, Sulej, Bronowicz and Satkūnas2013), while other stereospondyl clades have been reported to occur alongside phytosaurs in the absence of metoposaurids. Examples include chigutisaurids in India (Sengupta, Reference Sengupta1988), plagiosaurids in Thailand (Suteethorn et al., Reference Suteethorn, Janvier and Morales1988), and plagiosaurids and capitosauroids in Luxembourg (Milner et al., Reference Milner, Duffin and Delsate1996). In Europe, metoposaurids co-occur with phytosaurs and at least one other large-bodied temnospondyl clade (Milner and Schoch, Reference Milner and Schoch2004; Sulej and Majer, Reference Sulej and Majer2005). Lastly, metoposaurids were replaced by other clades in Europe (by capitosauroids; Sulej and Majer, Reference Sulej and Majer2005) and India (by chigutisaurids; Sengupta, Reference Sengupta1988, Reference Sengupta2003) during the late Norian and Rhaetian. Radioisotopic dating of the phytosaur-bearing site in Zimbabwe constrained a maximum age of 209.2 ± 4.5 Ma, near the Norian-Rhaetian boundary (Barrett et al., Reference Barrett, Sciscio, Viglietti, Broderick, Suarez, Sharman and Jones2020). Most metoposaurid-bearing localities are considered to occur in the Carnian, including those from India and Morocco (Buffa et al., Reference Buffa, Jalil and Steyer2019). The Malagasy metoposaurids, the geographically closest to the Zimbabwe occurrence, are of uncertain stratigraphic placement (Fortuny et al., Reference Fortuny, Arbez, Mujal and Steyer2019), but the upper Isalo II interval (Norian) does not contain any metoposaurid material (Burmeister et al., Reference Burmeister, Flynn, Parrish and Wyss2006), and the material is probably from the lower Isalo II and thus Carnian in age. The only metoposaurid records from the Rhaetian (208.5–201.3 Ma) are from North America (Hunt, Reference Hunt1993; Spielmann et al., Reference Spielmann, Lucas and Heckert2007; Spielmann and Lucas, Reference Spielmann and Lucas2012), and there are practically no other post-Norian occurrences of temnospondyls other than brachyopoids (see Konietzko-Meier et al., Reference Konietzko-Meier, Werner, Wintrich and Sander2019, for summary, and Maisch et al., Reference Maisch, Matzke and Sun2004, for a possible Jurassic trematosauroid). Therefore, the occurrence of a metoposaurid from Zimbabwe should be considered unreliable until more diagnostic material is both collected and described. We emphasize that this discussion should not be interpreted to mean that we are confident that the material is not metoposaurid, but rather that the only data that support metoposaurid affinities are weakly supported circumstantial lines of evidence (i.e., large-bodied stereospondyl co-occurring with a phytosaur). This discussion also underscores one of the key motivations of our own study: to provide a rigorous, detailed justification for taxonomic identification of material that may be considered to be a geographic outlier.