Journal of South American Earth Sciences 79 (2017) 281e296

Contents lists available at ScienceDirect

Journal of South American Earth Sciences

journal homepage: www.elsevier.com/locate/jsames

Biostratigraphic reappraisal of the Lower Triassic Sanga do Cabral Supersequence from South America, with a description of new material attributable to the parareptile genus Procolophon * Sergio Dias-da-Silva a, , Felipe L. Pinheiro b, Atila Augusto Stock Da-Rosa c, Agustín G. Martinelli d, Cesar L. Schultz d, Eduardo Silva-Neves a, e, Sean P. Modesto f a Centro de Apoio a Pesquisa Paleontologica da Quarta Colonia,^ Universidade Federal de Santa Maria, Rua Maximiliano Vizotto, 598, Sao~ Joao~ do Pol^esine, Rio Grande do Sul, CEP: 97-230-000, Brazil b Laboratorio de Paleobiologia, Universidade Federal do Pampa, Av. Antonio Trilha, 1847, Sao~ Gabriel, Rio Grande do Sul, Brazil c Laboratorio de Estratigrafia e Paleobiologia, Departamento de Geoci^encias, Centro de Ci^encias Naturais e Exatas, Universidade Federal de Santa Maria, Santa Maria, RS, CEP: 97.105-900, Brazil d Departamento de Paleontologia e Estratigrafia, Universidade Federal do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, CEP: 91540-000, Brazil e Programa de Pos-Graduaç ao~ em Biodiversidade Animal, Universidade Federal de Santa Maria, RS, CEP: 97.105-900, Brazil f Department of Biology, Cape Breton University, Sydney, Nova Scotia, B1P 6L2, Canada article info abstract

Article history: The Sanga do Cabral Supersequence (SCS), comprises the Brazilian Sanga do Cabral Formation (SCF) and Received 8 June 2017 the Uruguayan Buena Vista Formation (BVF). So far, the SCS has yielded temnospondyls, parareptiles, Received in revised form archosauromorphs, putative synapsids, and a number of indeterminate specimens. In the absence of 20 July 2017 absolute dates for these rocks, a biostratigraphic approach is necessary to establish the ages of the SCF Accepted 26 July 2017 and the BVF. It is well established that the SCF is Early Triassic mainly due to the presence of the Available online 29 July 2017 widespread Gondwanan reptile Procolophon trigoniceps. Conversely, the age of the BVF is subject of great controversy, being regarded alternatively as Permian, Permo-Triassic, and Early Triassic. The BVF has Keywords: fi Gondwana yielded the de nite procolophonid Pintosaurus magnidentis. Procolophonoidea is one of the most diverse Sanga do Cabral Supersequence and conspicuous terrestrial tetrapod groups of the Lower Triassic Lystrosaurus Assemblage Zone in the Triassic Karoo Basin of South Africa, which preserves tetrapods from the aftermath of the end-Permian extinction Temnospondyli event. Based on a previous interpretation that the fauna of the BVF is Permian, and in the reinterpretation Procolophonidae of disarticulated vertebrae from SCF with ‘swollen’ neural arches as belonging to either seymour- Archosauromorpha iamorphs or diadectomorphs, it was recently suggested that at least part of the SCF is Permian in age, which prompted this comprehensive reevaluation of both SCS's faunal content and geology. Moreoever, new, strikingly large procolophonid specimens (skull, vertebra, and a mandibular fragment) from the SCF are described and referred to the genus Procolophon. The large procolophonid vertebra described here contradicts the recent hypothesis that similar specimens from the SCF belong to seymouriamorphs or diadectomorphs, because its morphology is consistent with that found in Procolophon. There is not a single diagnostic specimen that supports the inference of Permian levels in the SCS. Accordingly, because all diagnostic and biostratigraphically informative fossils from the SCF and the BVF are either Early Triassic or restricted to the Triassic, we conclude that the available biostratigraphic data reinforce an Early Triassic age assignment to the SCS. © 2017 Elsevier Ltd. All rights reserved.

1. Introduction Uruguayan Buena Vista Formation (BVF), is crucial to the under- standing of the biotic recovery that followed the greatest biological The Western Gondwanan Sanga do Cabral Supersequence (SCS), crisis in the history of Earth, the end-Permian extinction event. The including the Brazilian Sanga do Cabral Formation (SCF) and the SCF is traditionally considered Early Triassic, an age assignment mainly due to the presence of the widespread Gondwanan reptile * Corresponding author. Procolophon trigoniceps (Dias-da-Silva et al., 2006a; Cisneros, E-mail address: [email protected] (S. Dias-da-Silva). http://dx.doi.org/10.1016/j.jsames.2017.07.012 0895-9811/© 2017 Elsevier Ltd. All rights reserved. 282 S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296

2008a). The age of the BVF, however, has been subject of great Federal do Pampa (UNIPAMPA), Sao~ Gabriel, RS, Brazil; UFSM, controversy because it has been alternatively regarded as Permian, Laboratorio de Estratigrafia e Paleobiologia, Universidade Federal Permo-Triassic, and Early Triassic (Pineiro~ et al., 2003, 2004, 2007a, de Santa Maria (UFSM), Santa Maria, RS, Brazil. b, c; Dias-da-Silva et al., 2006a; Modesto and Botha-Brink, 2010). In line with a previous interpretation that the fauna of the BVF is 2. Geology and biostratigraphic reappraisal Permian, together with a reinterpretation of disarticulated verte- brae from SCF with ‘swollen’ neural arches as attributable to either 2.1. Geology and geochronology seymouriamorphs or diadectomorphs, it was suggested recently that at least part of the SCF is Permian in age (Pineiro~ et al., 2015). If The Sanga do Cabral Supersequence (SCS) comprises the Bra- correct, it implies that the Permo-Triassic boundary lies within the zilian Sanga do Cabral Formation (SCF) and the Uruguayan Buena SCF, which would make the SCF fauna critical to analyses of Vista Formation (BVF) (Zerfass et al., 2003). Based upon its tetrapod tetrapod survivorship of the end-Permian extinction event. The content, the SCF has long been regarded as Early Triassic, and hypothesis that part of the SCF is Permian, which contradicts all correlated with the Lower Triassic Katberg Formation of the Karoo previous work supporting an Early Triassic age for that formation, Basin in South Africa (for a complete up-to-date set of SCS's pub- has prompted our comprehensive reevaluation of both SCS's faunal lished vertebrate specimens, see Table 1). Together with other content and geology, in order to provide a biostratigraphic reas- Lower Triassic units (Fig. 1), this Western Gondwanan super- sessment for this supersequence. sequence crucially documents biotic recovery following the end- Whereas we re-examine all biostratigraphic evidence for the age Permian Extinction Event, which is traditionally regarded as the of the SCS, procolophonoid fossils remain central to any biostrati- most severe of the ‘Big Five’ mass extinctions. Accordingly, the SCS graphic assessment of this supersequence. Procolophonoidea is the helps to understand the decisive worldwide biotic turnover from only parareptile clade that survived the end-Permian extinction Paleozoic to Mesozoic eras (see Benton et al., 2004; Smith and event (Modesto et al., 2001, 2003; Modesto et al., 2010) and became Botha, 2005; Botha and Smith, 2006; Sahney and Benton, 2008). one of the most diverse and conspicuous group of tetrapods of the The SCF was proposed by Andreis et al. (1980) for rocks cropping out Lower Triassic Lystrosaurus Assemblage Zone in the Karoo Basin of around the city of Rio Pardo, which is located centrally in Rio Grande South Africa (Smith and Botha, 2005; Botha and Smith, 2006; Botha do Sul State (RS) (Fig. 2). There, orange-coloured, fine-grained et al., 2007; Macdougall and Modesto, 2011). In South America, sandstones with localized intraformational conglomerates occur procolophonoids are known from both the Sanga do Cabral and over the pinkish to white, fine sandstones of the Rio do Rasto Buena Vista formations (see Zerfass et al., 2003). The Brazilian Formation. The conglomerates represent shallow braided streams, material comprises several specimens, including fairly complete in an approximate 1:100 thickness/width scale, whereas the orange skulls (e.g. Lavina, 1983; Langer and Lavina, 2000; Cisneros and fine sandstones are interpreted as massive, or presenting horizontal Schultz, 2002; Dias-da-Silva et al., 2006a). Originally ascribed to stratification, and denote a broad semiarid plain. The intraforma- two species, Procolophon pricei (Holotype UFRGS-PV231T; Lavina, tional conglomerates are the source of the majority of vertebrate 1983) and P. brasiliensis (Holotype MCN-PV-1905; Cisneros and fossils, which consist mainly of isolated, fragmentary bones. Schultz, 2002), both have been synonymized under P. trigoniceps Vertebrate remains are rarely preserved in the sandstones. The (Cisneros, 2008a). Although those best preserved and most diag- most commonly represented groups are procolophonids and tem- nostic specimens (i.e. skull and lower jaws) are unequivocally nospondyls (see discussion on the fossiliferous content, below). The referred to P. trigoniceps, the real diversity of procolophonians in the SCF occurs in an EeW outcrop belt in RS, with many fossiliferous SCS is unknown because several other specimens are fragmentary outcrops (Lavina, 1983; Lavina and Barberena, 1985; Da-Rosa et al., and poorly preserved (Langer and Lavina, 2000). Procolophonoid 2009; Dias-da-Silva and Da-Rosa, 2011), and have a probable material from the BVF so far comprises a single specimen of the physical continuation in Uruguay (Andreis et al., 1996), where they procolophonid Pintosaurus magnidentis (Pineiro~ et al., 2004; are known as the Buena Vista Formation (BVF). The BVF crops out Cisneros, 2008b; Modesto and Botha-Brink, 2010). In support of south of the city of Rivera (Wildner et al., 2008; Da Rosa et al., 2010) our reassessment of the age of the SCS, we describe a strikingly large and northeastwards of the city of Melo (Colonia Orozco), Uruguay procolophonid skull, UFSM 11409a, that we assign to the genus (Pineiro,~ 2004; Pineiro~ and Ubilla, 2003). Procolophon. Measuring 85 mm in length, the skull is the largest The known outcrops of the SCF are scattered, sometimes several specimen of the genus to be collected from South America. We also kilometers apart, which challenges their proper correlation to each examine a vertebral and a mandibular element of comparable size other. Although postdepositional tectonics have been super- that were recovered from the same fossiliferous site, but not directly imposed over the Early and MiddleeLate Triassic of Brazil (Da-Rosa associated with the skull (locality of Bica Sao~ Tome, Southern Brazil). and Faccini, 2005), recent correlation has managed to filter these effects and permitted lateral correlation (Da-Rosa and Dias-da- 1.1. Institutional abbreviations Silva, 2009). Apart from a repetitive intercalation of intraforma- tional conglomerates and fine sandstones, there is an overall trend BSPM, Bayerische Staatssammlung für Palaontologie€ und upwards in modification of the composition of intraclasts, from Geologie, München, Germany; CAPPA/UFSM, Centro de Apoio a argillaceous to carbonate nodules, and finally to a mixture of Pesquisa Paleontologica da Quarta Colonia-Universidade^ Federal de argillaceous, carbonate, and fragments of previously deposited Santa Maria; FC-DPV, Departamento de Paleontología de Verte- sandstones (Da-Rosa et al., 2010). This compositional change is brados, Facultad de Ciencias, Montevideo, Uruguay; MCN-PV, probably due to decreased space for sedimentation (accommoda- Vertebrate Paleontology Collection (PV), Museu de Ciencias^ tion) towards the top of the sequence (Zerfass et al., 2003). The Naturais, Fundaçao~ Zoobotanica^ do Rio Grande do Sul, Porto Alegre, small accommodation space has resulted in intense reworking of RS, Brazil; MCT/PURS, Museu de Ciencia^ e Tecnologia da Pontifícia SCF deposits. Accordingly, fossil bones are generally broken and Universidade Catolica do Rio Grande do Sul, Brazil; UFRGS-PV-T, weathered, and no articulated skeletons have been found so far. Vertebrate Paleontology Collection (PV), Triassic (T), Departamento However, a few articulated pieces are present, always preserved in de Paleontologia e Estratigrafia, Instituto de Geociencias,^ Uni- association with carbonate concretions. Within the same layer, versidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, RS, different diagenetic patterns can be observed in both bones and Brazil; UNIPAMPA, Laboratorio de Paleobiologia, Universidade concretions (Bertoni-Machado et al., 2008), indicating that SCF S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 283

Table 1 Tetrapods from the Sanga do Cabral Supersequence.a

Unit Number Anatomical assignation Previous taxonomical Current taxonomical Source assignationb assignationc

SCF UFRGS PV0237T Dermal skull fragment Rhytidosteidae Unchanged Dias-da-Silva et al. (2005) “ UFRGS PV0253T Dermal skull fragment Rhytidosteidae Unchanged Dias-da-Silva et al. (2005) “ UFRGS PV0257T Dermal skull fragment Rhytidosteidae Unchanged Dias-da-Silva et al. (2005) “ UFRGS PV0327T Dermal skull fragment Rhytidosteidae Unchanged Dias-da-Silva et al. (2005) “ UFRGS PV0361T Dermal skull fragment Rhytidosteidae Unchanged Dias-da-Silva et al. (2005) “ UFRGS PV0362T Dermal skull fragment Rhytidosteidae Unchanged Dias-da-Silva et al. (2005) “ MCN PV2606 Dermal skull fragment Sangaia lavinai Unchanged Dias-da-Silva et al. (2005, 2006a); Dias-da-Silva and Marsicano (2011) “ UFRGS PV0250T Maxillary fragment Stereospondyli Unchanged Dias-da-Silva et al. (2005) “ UFRGS PV0506T Mandibular fragment Stereospondyli Unchanged Dias-da-Silva et al. (2005) “ UFRGS PV651T Mandibular fragment Stereospondyli Unchanged Dias-da-Silva et al. (2005) “ UFRGS PV0326T Distal fragment of a femur Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0331T A left humerus Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0332T Distal humeral fragment Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0339T Distal humeral fragment Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0353T Distal femoral fragment Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0375T Distal femoral fragment Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0369T Proximal femoral fragment Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0387T Cleithrum Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0499T Right ilium Stereospondyli Unchanged Dias-da-Silva and Schultz (2008) “ UFRGS PV0357T; Radial fragments Stereospondyli Unchanged Dias-da-Silva and Schultz UFRGS PV0358T (2008) “ UFRGS PV0236T to PV0238T; PV0253T; Appendicular fragments Stereospondyli Unchanged Dias-da-Silva and Schultz PV0255T to PV0258T; PV0330T; PV0333T (2008) to PV0338T; PV0362T to PV0365T; PV0386T; PV0402T; PV0351T to PV0352T; PV0354T to PV0355T; PV0359T; PV0370T to PV0374T; PV0376T “ UMVT 4302; UMVT 4303 Half left skull; right palatal Sangaia lavinai Unchanged Dias-da-Silva et al. (2006a); fragment Dias-da-Silva and Marsicano (2011) “ UFSM 11408 Right preorbital region; Tomeia witeckii Unchanged Feltrin et al. (2008); Da-Rosa distal mandibular ramus et al. (2009); Eltink et al. (2016) “ UFSM 11455 undetermined skull Mastodonsauroidea Unchanged Feltrin et al. (2008); Da-Rosa fragment et al. (2009) “ MCN PV1999a, b Dermal skull fragment and Plagiosteninae Unchanged Dias-da-Silva and Ilha (2009); its opposite natural cast Dias-da-Silva and Milner (2010) “ UNIPAMPA PV 00227, 00249, 00255 Dermal skull fragments Stereospondyli. Unchanged Dias-da-Silva and Da-Rosa (2011); Dias-da-Silva and Dias (2013) “ UNIPAMPA PV 00230 Distal humeral fragment Stereospondyli. Unchanged Dias-da-Silva and Da-Rosa (2011) “ UNIPAMPA PV 00235 Tooth bearing fragment Stereospondyli. Unchanged Dias-da-Silva and Da-Rosa (2011) “ UFRGS PV0506T Fragment of an olecranon Pareiasauridae Stereospondyli Schultz and Dias-da-Silva (reinterpreted as a PGA (1999), Dias-da-Silva et al. fragment) (2005) “ UFRGS PV0362T Skull fragment Pareiasauridae Indet. Schultz and Dias-da-Silva (1999), Cisneros et al. (2005) “ UFRGS-PV231T Skull and vertebrae Procolophon pricei Procolophon trigoniceps Lavina (1983); Cisneros (2008a) “ MCN-PV-1905 Skull Procolophon brasiliensis Procolophon trigoniceps Cisneros and Schultz, 2002; Cisneros (2008a) “ UFRGS-PV0494T Skull Procolophon Procolophon trigoniceps Dias-da-Silva et al. (2006a), this work “ MCT-PUCRS 3790, UFRGS-PV0252T, Dorsal vertebrae Procolophon Unchanged Dias-da-Silva et al. (2006a) UFRGS-PV0498T “ MCN-PV 2722, MCN-PV 2723, Dorsal vertebrae Tetrapoda indet, Procolophon Pineiro~ et al. (2015), this work. MCN-PV 20000, MCN-PV 20001 possibly belonging to either seymouriamorphs or diadectomorphs (continued on next page) 284 S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296

Table 1 (continued )

Unit Number Anatomical assignation Previous taxonomical Current taxonomical Source assignationb assignationc

“ UNIPAMPA 055 Dorsal vertebra Procolophon This work “ UNIPAMPA 0680 Mandibular fragment Procolophon This work “ UFSM 11409a Skull with mandible Procolophon This work “ UNIPAMPA 0653 Skull Teyujagua paradoxa Unchanged Pinheiro et al. (2016) “ UFRGS-PV493T Vertebra Protorosaurid with Unchanged Langer et al. (1996); Langer and tanystropheid affinities Schultz (1997); Langer and Lavina (2000) “ UFSM 11394 Vertebra ?Archosauromorpha Unchanged Da-Rosa et al. (2009) “ UFSM 11467, 11475, 11458, Vertebrae ?Archosauriformes Unchanged Da-Rosa et al. (2009) 11444, 11460 “ UNIPAMPA 00239, 00250, 00267 Vertebrae Archosauromorpha Unchanged Dias-da-Silva and Da-Rosa (2011) “ UFRGS PV351T, PV375T, Appendicular elements Non-mammalian Procolophonidae indet Abdala et al. (2002), this work PV354T, PV353T, PV332T cynodonts “ MCN 1072, MCN 1073 Stapes Lystrosaurus Procolophonian sacral Schwanke and Kellner (1999), ribs Langer and Lavina (2000), this work BVF FC-DPV-285 Skull Dvinosauria indet. Unchanged Marsicano et al. (2000) “ FC-DPV 1182, Vertebrae Varanopidae Diapsida Pineiro~ et al. (2003); Dias-da- 11831199, 1200, 1189, 1333, 1333 Silva et al. (2006a) “ FC-DPV 1194, Vertebra Basal synapsid Diapsida Pineiro~ et al. (2003); Dias-da- Silva et al (2006a) “ FC-DPV 1181 Skull and mandibles Pintosaurus magnidentis Unchanged Pineiro~ et al. (2003); Cisneros (2008b) “ FC-DPV 1598 Partial skull Uruyiella liminea Unchanged Pineiro~ et al. (2007b) “ FCDPV 1280, 1305, 1600 Mandibular elements Mastodonsauridae Unchanged Pineiro~ et al. (2007c) “ FC-DPV 1369 Partial skull Arachana nigra Unchanged Pineiro~ et al. (2012) “ FC-DPV 2641, 2640, 2639, 2637 Skull fragment, vertebrae Archosauromorpha Unchanged Ezcurra et al. (2015)

a Undiagnostic, badly preserved, and fragmentary specimens are ommited. b In a less inclusive possible level. c Either changed in the present work or in articles published after the first description. deposits are time-averaged, but the exact extent of this phenom- the SCS extends to Uruguay as the Buena Vista Formation, this enon cannot be determined precisely. Notwithstanding these formation, at least in the Colonia Orozco area, records a similar shift constraints, the faunal content points to an unequivocal Early in the river patterns: from low-energy in the base, right above the Triassic age for the package, enhanced by the predominance of Yaguarí Formation, to high-energy in its upper levels (Pineiro~ et al., Procolophon trigoniceps in almost all fossil bearing strata. 2007a). This similarity in the fluvial patterns was used by Zerfass Faccini (1989) applied the concepts of sequence stratigraphy to et al. (2003) to suggest that the SCF and the BVF (plus the Talam- the continental Permo-Triassic Gondwanic rocks from Rio Grande paya/Tarjados package from Argentina) are homotaxial. The main do Sul State. He postulated that the SCF should correspond to a problem, however, is to establish the precise age of this orogenic distinct allostratigraphic unit, which he named the Eoscythian event. Sequence, that was separated from the underlying Permian Apart of the Talampaya/Tarjados formations, other geologic package by an unconformity. The underlying Upper Permian units units in Argentina may be coeval with those aforementioned units. are the fluvio-lacustrine Rio do Rasto Formation (RRF) and the Rocha-Campos et al. (2011) studied the Choiyoi igneous province aeolian Piramboia Formation (PF), both laterally interbedded. from the San Rafael Block, obtaining a SHRIMP UePb 251.9 ± 2.7 Ma Zerfass et al. (2003) enhanced the allostratigraphic framework age for the Cerro Carrizalito Formation (Upper Choiyoi section). In proposed by Faccini (1989). Faccini's Eoscythian Sequence was this case, deposition of the fossiliferous Puesto Viejo Formation renamed the Sanga do Cabral Supersequence (SCS), a tectonically (PVF) initiates the Huarpic extensional phase during Early Triassic. controlled, second-order allostratigraphic unit, of which the The PVF was, for a long time, interpreted to be Lower Triassic dominant facies association comprises intraformational massive or (Ottone and Garcia, 1991), and subsequently divided into the basal trough cross-bedded conglomerates, and horizontal bedded sand- mainly grayish Quebrada de los Fosiles Formation (QFF) and the stones that crop out from the Rio Grande do Sul State towards overlying Río Seco de la Quebrada Formation (RSQF; Stipanicic et al., Uruguay (Buena Vista Formation). According to Zerfass et al. (2003), 2007). Recently, Ottone et al. (2014) stated that the basal unit (QFF) the SCS strata in Brazil overlie the aeolian sandstones of the PF (or includes both plant remains (pleuromeians and sphenopsids) and even the underlying RRF). The disconformity that delineates the vertebrates (scattered fish scales, dicynodont synapsids, and re- basal boundary of the SCS was correlated to Veevers et al.’s (1994) mains of the archosauriform Koilamasuchus gonzalezdiazi Ezcurra Gondwanides I paroxysm. et al., 2010). In contrast, the RSQF beds have yielded only tetra- According to Zerfass et al. (2003), this tectonic event was also pods, although it represents a more diverse fauna than that found in responsible for the most striking sedimentary feature of the SCS the QFF, and include cynodonts such as Cynognathus, Diademodon, (with regard to the underlying Permian rocks), i.e. a shift in the and Pascualgnathus, and also dicynodonts (Vinceria and Kanne- fluvial pattern that changed from meandering to braided, reflecting meyeria)(Abdala, 1996; Martinelli et al., 2009). Based on the an uplift in the relief. This change in the fluvial style, from high- tetrapod assemblage, the fossil-bearing levels were correlated to (Late Permian) to low-sinuosity (Early Triassic) is also recorded in the Cynognathus AZ of South Africa and thus referred to the Middle the Karoo Basin (Katberg Fm.), which has also been linked to the Triassic (Anisian), but a SHRIMP 238U/206Pb age of 235.8 ± 2.0 Ma Gondwanides I paroxysm (Smith, 1995) or to climatic shift (Ward from a rhyolitic ignimbrite interdigitated between the QFF and RSQF et al., 2000; Retallack et al., 2003). Reinforcing the concept that at the Quebrada de los Fosiles section points to two possibilities: (1) S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 285

Fig. 1. Lower Triassic units worldwide. The Talampaya Formation (yellow) also contains Permian strata. In the Puesto Viejo Formation, the question mark (?) indicates that the age of this unit remains controversial. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) the Puesto Viejo Group is younger than previously thought; or (2) interrupted sedimentation along the southwestern segment of the the tetrapod fauna in the RSQ beds existed, instead, during the Late Gondwana margin (i.e. Frontal Cordillera, San Rafael Basin), Triassic (early Carnian), some 10 Ma later than the currently induced cratonward thrusting (i.e. Ventana and Cape foldbelts), and accepted age for the Cynognathus Zone (Ottone et al., 2014). How- triggered accelerated subsidence in the adjacent basins (Sauce ever, a single date such as this should be interpreted cautiously Grande and Karoo) located inboard of the deformation, in the latest (Sues, 2016; Martinelli and Soares, 2016; Martinelli et al., 2017) Permian. because, if we do not consider, or dismiss, the shared cynodont taxa In the Cuyo Basin, U-Pb SHRIMP ages from tuffs interbedded in between the Argentinean RSQF and Cynognathus AZ of the Karoo the Potrerillos Formation are 239.2 ± 4.5 Ma, 239.7 ± 2.2 Ma, and Basin in South Africa (Martinelli et al., 2009), we should at least 230.3 ± 2.3 Ma (Middle Triassic), and are associated with the rift consider that other, and more geographical distant, early Carnian stage that followed the extensive post-orogenic volcanism of the associations from Argentina (i.e. Ischigualasto-Villa Union Basin), Choiyoi Group, which in turn is ascribed to slab break-off in Brazil (i.e. Santa Maria Supersequence) and Madagascar (basal Isalo neighbouring areas (Spalletti et al., 2008). The rift-related Triassic II) are more similar in faunal composition and do not record the event represents the culmination of the Gondwanan magmatic cynodonts (e.g., Cynognathus and Diademodon) and dicynodonts cycle, and is interpreted as the result of subduction cessation and (Vinceria) known from the RSQF, which are otherwise unknown in anomalous heating of the upper mantle prior to the break-up of the late Triassic (see also discussion in Martinelli et al., 2017). western Gondwana. Therefore, an age of the RSQF comparable to that of the Chanares~ Sato et al. (2015) also associated the post-Choiyoi magmatism Formation of Argentina or the Dinodontosaurus AZ of Brazil makes with extensional tectonics, but a 252.38 (þ0.09/-0.22) Ma U-Pb age no sense within a biostratigraphic scheme that otherwise fits rela- from the Talampaya Formation places its deposition in the latest tively well along the Triassic. In this context, we mention here the Permian or even on the PeTr boundary (Gulbranson et al., 2015), PVF as one of the stratigraphic units that was possibly deposited which reinforces an Early Triassic of age for the SCS, for the Tarjados under the effect of an Early Triassic tectonism. Formation overlies the Talampaya Formation, and both are In this line of reasoning Lopez-Gamundi (2006) stated that the considered coeval to the SCS (Zerfass et al., 2003, 2004), in which, so San Rafael orogenic phase (SROP, between 280 and 260 Ma) far, the only biostratigraphically informative taxon is P. trigoniceps. 286 S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296

Fig. 2. Permo-Triassic geological basins of Gondwana. (A) paleogeographic reconstruction of southwestern Gondwana (black, Middle and Upper Triassic; grey, Early Triassic), based on South American, African, and Antarctic basins; modified from Zerfass et al. (2003); (B) Chaco-Parana Basin in South America, and location of inserts A and C; (C) generalized geological map of the central portion of Rio Grande do Sul State; modified from Dias-da-Silva and Da-Rosa (2011); (D) stratigraphic cross-section of Upper Permian and Mesozoic units, based on log correlation (logs shown in C), showing the Triassic units (Sanga do Cabral, Santa Maria, and Caturrita formations) as second-order supersequences; modified from Zerfass et al. (2003) and unpublished data. Abbreviations: B, Botswana; BV, Buena Vista Formation; CB, Cabora-Bassa; MZ, Zambia; R, Ruhuhu; SC, Sanga do Cabral Sequence; SF, San Felix; SG, Sauce Grande; SM, Santa Maria Supersequence; SR, San Rafael; W, Waterberg; W-O, Waterberg-Omaruru Fault; Z, Zambezi.

Even in the Karoo Basin, where the P-Tr boundary constitutes a maximum depositional ages for the marine Ecca Group range from reference for all other continental basins, the available radiometric 274 to 250 Ma, whereas in the conformably overlying terrestrial data are conflicting (see discussion in Sato et al., 2015; Hansma Beaufort Group the maximum depositional ages ranged from 257 to et al., 2015). Fildani et al. (2007, 2009) obtained ages between 252 Ma. This indicates that there is an uncertainty interval age of 7 275 Ma and 254 Ma (Kungurian-Early Changhsingian) in the million years (from 257 to 250 Ma) in which the P-Tr boundary is Laingsburg Formation, (middle Ecca Group), implying that the included [252.17 (IUGS 2015/01)]. Across the southern Karoo Basin, overlying Beaufort Group (and thus the P/Tr boundary) should be the Ecca Group tuffs produce maximum depositional ages that get much younger than supposed. Conflicting with this data, Lanci et al. younger upwards in the succession, correlating with tuffs in (2013) obtained a range from 268 to 264 Ma (WordianeEarly southern South America. According to McKay et al. (2015), Capitanian) for the package including the Waterford Formation Karoo tuffs correlate with the ones in the Puesto Viejo Group and (uppermost Ecca Group) and the Abrahamskraal Formation volcanism in the lower Choiyoi Group, in which tuffs in the Ecca (lowermost Beaufort Group). Rubidge et al. (2013) found a range Group are coeval with the upper Choiyoi Group volcanic rocks between 261 and 255 Ma (Late CapitanianeWuchiapingian) for a (265e250 Ma), and tuffs in the Beaufort Group correlate to the set of five volcanic ashes interbedded with fossils from the Triassic (younger than 240 Ma) Puesto Viejo Group, as a reflect of a Pristerognathus, Tropidostoma, and Cistecephalus vertebrate bio- shallowing of magmatism that is geochemically compatible with zones of the Beaufort Group. Also, Coney et al. (2007) state that the back-arc extensional volcanism. P-Tr boundary is located at the contact of the Palingkloof Member A shift in the fluvial pattern observed in the Katberg Formation (252.2 ± 0.7 Ma) and the Elandsberg Member, within the Balfour around the P-Tr boundary (Smith, 1995) was triggered by a pulse of Formation, reinforcing a younger, Early Triassic age for the Katberg thrusting (circa 247 ± 2 Ma) which brought about rapid pro- Formation. gradation of a large sandy braided fan system into the central parts McKay et al. (2015) studied air-fall tuffs thought to be distal of the basin. However, this would not be necessarily applicable to deposits derived from the Permian-Triassic Southern Gondwanan the South American packages that exhibit the same pattern, once volcanic arc, and they found a “tuff gap”. For those authors, the that the tectonic effects of this orogeny would be only local. On the S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 287 other hand, Ward et al.’s (2000) hypothesis would have a global paleontological point of view, no significant age differences can be scope and could be contemporary in all basins, but is not datable detected (Holz and Souto-Ribeiro, 2000). Recently, Dias-da-Silva isotopically. and Pinheiro (personal observation) detected a second (and not so Zerfass et al. (2004) stated that most of Middle-eLate Triassic rare) mode of fossil preservation in the Sanga do Cabral Super- continental rocks from Southern Brazil, Argentina, and Namibia sequence: several specimens were recovered in horizontally were deposited in a set of ‘en echelon basins’ in southern West bedded sandstones (Sh lithofacies mentioned by Zerfass et al., Gondwana. The main structures of the WaterbergeOmaruru Fault 2003), especially in the locality of Bica Sao~ Tome (see Da-Rosa can be connected to NW-strike faults in Brazil. Some of these faults et al., 2009). In this type of facies, the state of preservation of affected also the deposition of the underlying SCS, thereby indi- fossils (e.g. much larger, better preserved, and even partially artic- cating the initiation of extension in Early Triassic times. NNEeSSW ulated specimens, but in this last case, within concretions). In this extension reactivated structures of the Damara Belt in Namibia, contribution, two specimens (among several others still under propagated towards Rio Grande do Sul State forming an anasto- study) were encountered in such condition: UFSM 11409a and mosing normal fault system and related-rift basin by Early-Middle UNIPAMPA 0655. According to Pineiro~ et al. (2015), because fossil Triassic. In fact, NW fractures occur in all known outcrops, although vertebrates from the SCS are preserved in intraformational con- Zerfass et al. (2005) also suggest that EeW fractures are exclusive of glomerates, the interpretation of the fossils as having been the SCS and could preclude the Triassic extensional phase. reworked from stratigraphically lower levels cannot be dismissed. The sedimentary environment of the SCS is thought to be a However, these authors fail to mention that taphononomic data on braided plain, in which small and shallow channels spread north- river channel deposits (as mentioned above) suggest a short-ranged wards into a vast semiarid plain. Actually, there is an overall time-averaging (hundreds to thousand years). Hence, even if northward-trend of the sedimentary facies in the southern border Permian and Triassic vertebrates were found reworked and mixed of the Parana Basin, as inferred by many geologists and geoscience in the Sanga do Cabral Supersequence, it necessarily indicates that researchers (e.g. Holz et al., 2010). This assumption may have this unit should be Triassic rather than Permian. biased the work of Pineiro~ et al. (2007a, 2015), in which the BVF could be the lower and/or the proximal units of the braidplain 2.3. Procolophonid data recorded in the SCF. Geological data demonstrates that litho- and chronostrati- Procolophonids are the most common fossil vertebrates found at graphic relations between Sanga do Cabral (in Brazil) and Buena the Sanga do Cabral Formation outcrops, being represented by Vista (in Uruguay) formations are still inconclusive, because several specimens, including fairly complete skulls (e.g. Lavina, detailed geological mapping encompassing the two units has not 1983; Langer and Lavina, 2000; Cisneros and Schultz, 2002; Dias- yet been carried out. Furthermore, absolute dating in both units is da-Silva et al., 2006a). They were the first group that permitted a required, and, at this point, highly desirable. Notwithstanding, the relative age estimation for this unit with the identification of the physical continuity between the SCF (in southern Brazil) and the genus Procolophon in South America and its occurence with the BVF (in Uruguay), as well as the outstanding shift in the fluvial Lystrosaurus Assemblage Zone of South Africa (e.g. Barberena et al., pattern that differentiates both units from the respective underly- 1981; Lavina, 1983). Two procolophonid species were erected based ing units suggests that they actually correspond to a single allos- upon specimens from the Sanga do Cabral Formation: Procolophon tratigraphic sequence and are probably coeval. Even if we consider pricei (Holotype UFRGS-PV231T; Lavina, 1983) and P. brasiliensis that this shift in the fluvial pattern resulted from ecologic causes (Holotype MCN-PV-1905; Cisneros and Schultz, 2002), which sub- (extinction of most continental plants at the end Permian, according sequently were referred to P. trigoniceps because they fit within the to Ward et al., 2000) or tectonic activity (according to McKay et al., range of individual and ontogenetic variation known for the large 2015; Sato et al., 2015; Gulbranson et al., 2015), the age of the SCS sample of P. trigoniceps from South Africa (Cisneros, 2008a). should be considered as Early Triassic. Although the best preserved and readily diagnostic specimens (i.e. skull and lower jaws) are unequivocally referred to P. trigoniceps, the 2.2. Taphonomic data real diversity of procolophonians in the Sanga do Cabral Super- sequence is unknown due because several other specimens are It has been a traditional point of view that, in the Sanga do Cabral fragmentary and poorly preserved (Langer and Lavina, 2000). Supersequence, the mode of occurrence of fossil vertebrate remains With regards to the only procolophonid so far recovered from is mostly disarticulated, fragmented, and preserved in intraforma- the Uruguayan Buena Vista Formation, Pintosaurus magnidentis was tional conglomerates (specifically in gravel lenses consisting of interpreted as a latest Permian procolophonid by Pineiro~ et al. mudstone and carbonate intraclasts of facies Gmm, Gem, Gt, and Gp (2004), following their previous claim (later challenged by Dias- - massive and stratified gravels) (Holz and Souto-Ribeiro, 2000; da-Silva et al., 2006a; Modesto and Botha-Brink, 2010) that this Zerfass et al., 2003). Holz and Souto-Ribeiro (2000) hypothesized unit is Permian or Permo-Triassic in age. However, in the compre- that SCS vertebrates were already fossilized in some degree before hensive procolophonid phylogenetic analysis by Cisneros (2008b), deposition in the conglomerates, a condition justified by the Pintosaurus was nested in a polytomy of a clade that only includes frequent preservation of very fragile portions of bone, such as Triassic taxa: Coletta seca, Sauropareion anoplus and a large clade neural arches. Such accumulation of bones would have occurred that embraces all other procolophonids (Cisneros, 2008b, fig. 6). during periods of huge floods. According to Holz and Souto-Ribeiro This phylogenetic uncertainty did not allow Cisneros (2008b) to (2000), reworking is evident in the SCS, so it is important to comment specifically on the biostratigraphic utility of Pintosaurus consider some degree of time-averaging. According to Kidwell and with regards to the age of the BVF. Flessa (1995), vertebrate accumulations formed in river channels Pineiro~ et al. (2015) recently described several disarticulated the time resolution for sampled intervals will be 103e104 yr in most ‘swollen’ vertebrae from the same locality (Rincao~ dos Weiss) that fossiliferous sequences, but it is noteworthy that these authors produced vertebrae of the same morphology and were described by consider only the ‘natural’ accumulation of bones in river systems, Dias-da-Silva et al. (2006a) and tentatively assigned by those au- and not the reworking of previously fossilized remains. In the fluvial thors to Procolophon trigoniceps. Pineiro~ et al. (2015) dismissed that facies of the Sanga do Cabral Supersequence, the taphocoenosis taxonomic assignment, arguing that the neural-arch dimensions may present a mixture of several generations, but from a (ca. 1.6:1 for transverse breadth:anteroposterior length) and large 288 S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 size (ca. 25 mm wide) were not exhibited in the largest known no evidence of accessory intervertebral articulations in similar specimens of P. trigoniceps from South Africa (from which nearly all material from the same locality that we studied (Dias-da-Silva et al., articulated specimens of this reptile are known). Furthermore, 2006a) and we see no convincing sign of such features in the illus- Pineiro~ et al. (2015) stated that the short neural spines of Rincao~ dos tration of Pineiro~ et al. (2015). Second, in Paleozoic tetrapods Weiss vertebrae are incompatible with procolophonid affinities, exhibiting ‘swollen’ vertebrae and accessory intervertebral articu- even though the authors acknowledged that this structure is lations, hyposphenes and hypantra are present on the anterior and broken in all the described specimens. Finally, Pineiro~ et al. (2015) posterior surfaces, respectively, of the neural arch (e.g. Diadectes: tentatively identified the remains of accessory intervertebral ar- Sumida and Modesto, 2000: fig. 2; Mesosaurus: Modesto, 1999: fig. ticulations (hypantra and hyposphenes) in the vertebrae in their 3), whereas Pineiro~ et al (2015: fig. 4b, f). identify the opposite or- study. Pineiro~ et al. (2015) proposed alternatively that the Rincao~ ganization in their specimen. dos Weiss ‘swollen’ vertebrae represented either seymouriamorph Vertebrae similar to those described by Pineiro~ et al. (2015) are or diadectomorph vertebrae, which would suggest a Permian age common in the Sanga do Cabral Supersequence, and several pub- for that particular fossiliferous level. lished specimens with similar morphologies are deposited in public There are several problems with Pineiro~ et al.’s(2015)argument collections (e.g. UFRGS PV 0252 T, 0498 T, and MCT/PUCRS 3790; that the Rincao~ dos Weiss ‘swollen’ vertebrae are not assignable to see Dias-da-Silva et al., 2006a). A new, undescribed large vertebra Procolophonidae. First, Dias-da-Silva et al. (2006a) described a (UNIPAMPA 055) is described here together with a large mandib- ventral double ridge on the centra, which is an autapomorphy of ular fragment (UNIPAMPA 0680) (see section 3). Both specimens Procolophonidae that is unknown in other Permo-Carboniferous are compatible in size with the large skull (UFSM 11409a) from the tetrapods with ‘swollen’ vertebrae, including seymouriamorphs, same outcrop. diadectomorphs, and pareiasaurs (Laurin and Reisz, 1995). This We should note that, even if the ‘swollen’ vertebrae described by ventral double ridge is visible also on the material that Pineiro~ et al. Dias-da-Silva et al. (2006a) and Pineiro~ et al. (2015) could not be (2015) described and illustrated (their figs. 3, 4 and 6), yet those attributed to Procolophon, the presence of P. trigoniceps in the Sanga authors did not consider the taxonomic implications of this feature. do Cabral Supersequence is indisputable, based on other several Second, Pineiro~ et al.’s(2015)statement that the neural arch di- specimens (e.g. Cisneros, 2008a). It is possible that these vertebrae mensions of P. trigoniceps never exceed 1:1 ratio (transverse bread- could represent a non-Procolophon, yet-to-be-determined proco- th:anteroposterior length) is false, because a 1.9:1 ratio is exhibited lophonid in the Sanga do Cabral outcrops. However, P. trigoniceps is in posterior dorsal vertebra a P. trigoniceps skeleton figured by Broili by far the most abundant recognizable species recovered in this and Schroeder€ (1936, table 5; S.P.M., personal examination of a cast unit and, so far, the only one from which biostratigraphic infor- of BSPM 1934 VIII 39) (Fig. 3). Similarly exaggerated ‘swollen’ neural mation can be drawn at a specific level. In the coeval Karoo Basin, arches are present in the Early Triassic procolophonid Sclerosaurus where high resolution sedimentary and biostratigraphic studies armatus (Sues and Reisz, 2008: fig. 1). Moreover, the statement by have been carried out, P. trigoniceps has its first appearance 116 m Pineiro~ et al. (2015) that vertebrae of P. trigoniceps from Brazil could above the Permo-Triassic boundary, becoming abundant about not reach the size displayed by the Rincao~ dos Weiss vertebrae is 160 m above it (Botha and Smith, 2006). The “Procolophon abun- incorrect, because it assumes that the P. trigoniceps material from dant zone” of the Karoo Basin, possibly contemporaneous with the South African collections fully document the size range of this spe- Sanga do Cabral Supersequence deposition, corresponds to the cies. Pursuant to this, we examined a cast of the Munich specimen upper Katberg Formation (upper levels of the Lystrosaurus Assem- (No. 1934 VIII 39), which preserves a skull ca. 59.5 mm long (as blage Zone), which is Induan to early Olenekian in age. measured from tip of snout to posterior edge of skull table). This A new and remarkably large procolophonid skull, UFSM 11409a, skeleton exhibits a dorsal vertebra (presacral 20) that is 19 mm measuring 85 mm in length, is also worth mentioning here (for broad, and this is 1 mm shorter in breadth than MCN-PV 2722 more information, see section 3). Based on the presence of eight (which has a transverse breadth of 20 mm), one of the vertebrae large, mesiodistally compressed, bicuspidate molariform teeth studied by Pineiro~ et al. (2015: fig. 6(B)). We conclude that Pineiro~ with chisel-like crowns, we refer UFSM 11409a to Procolophon et al. (2015) statement that P. trigoniceps vertebrae could not ach- trigoniceps. Da-Rosa et al. (2009) preliminarily reported on UFSM ieve the sizes exhibited by the Rincao~ dos Weiss specimens spring 11409a, and further preparation has uncovered its palate and part from an unawareness of the available literature on P. trigoniceps and of the braincase in dorsal view, since its skull roof is missing. This other procolophonids. Furthermore, Pineiro~ et al. (2015) identified prompted us to include it in the dataset by MacDougall et al. (2013), accessory intervertebral articulations (hyposphenes and hypantra) and a phylogenetic analysis (see section 4) resulted in a sister group in the vertebrae available to those authors. The presence of these relationship of UFSM 11409a with Procolophon (Fig. 6). Given that features seem dubious to us for the following reasons: first, we find this is the largest procolophonid specimen ever reported world- wide, a further investigation is necessary in order to verify whether this large specimen only represents an exceptionally large indi- vidual of P. trigoniceps in an advanced ontogenetic stage, or if it could represent a new and robust species of this genus. In any case, together with UNIPAMPA 0655 and all large vertebral elements described previously (see Dias-da-Silva et al., 2006a) and ascribed to Procolophonidae on the basis of midline double ridges on the pleurocentra, all these specimens support an Early Triassic (prob- ably Induan/Olenekian) age for the Sanga do Cabral Formation.

2.4. Temnospondyl data

Remains of stereospondyls and indeterminate temnospondyls comprise several hundreds of scattered and fragmentary dermal “ ” Fig. 3. Procolophon trigoniceps, BSPM 1934 VIII 39, referred specimen in dorsal view. bones ornamented with the so-called spider-web sculpturing. Photograph by O. Rauhut. Scale bar equals 2 cm. Three described specimens of different groups provide valuable S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 289 information regarding taxonomy, phylogeny, biostratigraphy, and, da-Silva and Marsicano (2011), in which Indobrachyops was consequently, the age estimation of the Sanga do Cabral Formation: considered a derived rhytidosteid. Intriguingly, Pineiro~ et al. (2015) (1) the rhytidosteid Sangaia lavinai (Dias-da-Silva et al., 2006b) did not mention this comprehensive phylogenetic analysis of (specimens UMVT 4302, UMVT 4303, and MCN PV2606); (2) the rhytidosteids and did not comment on the possible presence of the capitosauroid Tomeia witecki Eltink et al. (2016) (specimen UFSM lacrimal bone in this Lower Triassic rhytidosteid from India, only 11408); and (3) a still indeterminate plagiosaurine plagiosaurid mentioning its presence in the Permian Trucheosaurus to support stereospondyl (specimen MCN PV1999) (Dias-da-Silva and Ilha, the alleged presence of Permian levels in the Sanga do Cabral For- 2009; Dias-da-Silva and Milner, 2010). The original description of mation. Pineiro~ et al. (2015) also state that Sangaia “possesses some Sangaia lavinai by Dias-da-Silva et al. (2006b) suggested that characters not previously reported in any other Triassic member of it could be a basal rhytidosteid, because it bears a prominent the group”. Sangaia, however, exhibits a unique suite of characters; lacrimal bone, a condition shared with the Upper Permian none of them are exclusive to this taxon, which means that so far Trucheosaurus major from Australia and the Lower Triassic Indo- this taxondbased on the holotype onlydis a metataxon, because it brachyops panchetensis from India. At that time, it was also sug- lacks autapomorphies. Its diagnosis is based upon a combination of gested that S. lavinai was probably closely related to the Indian non-exclusive features that together render this specimen unique, I. panchetensis,butDias-da-Silva et al. (2006b) did not provide a therefore justifying its status as a valid species. phylogenetic analysis to support this inferred basal phylogenetic Recently, a new capitosauroid taxon, Tomeia witecki, increased position. Later, Dias-da-Silva and Marsicano (2011) presented a the diversity of temnospondyls in the SCS (Eltink et al., 2016). The comprehensive computer-based phylogenetic reassessment of holotype was reported brieflybyDa-Rosa et al. (2009) as a new the Rhytidosteidae. These authors used different methodological temnospondyl to be described elsewhere in detail. Capitosauroidea approaches (parsimony, implied weighting, and two different is a large, cosmopolitan stereospondyl during the Early to Late datasets - lower jaws either included or excluded) with quite Triassic, and its first definite representatives come from Lower different results. In the parsimony analysis without mandibular Triassic deposits. Tomeia witecki, despite its incompleteness, pre- characters in the data matrix, Sangaia was recovered as the sister sents a melange of early and late-diverging capitosauroid charac- taxon of the rhytidosteid subfamily Derwentiinae, which includes ters. Indeed, Eltink et al.’s (2016) phylogenetic analysis places Indobrachyops as the sister taxon of Arcadia, Derwentia,andRewana Tomeia within a clade that comprises only Early Triassic cap- (Dias-da-Silva and Marsicano, 2011: fig. 3a and b). On the other itosauroids, in a sister-group relationship with the Madagascan hand, its position changed with the implied weighting approach Edingerella madagascariensis and close to Watsonisuchus spp., from because Sangaia fell outside Derwentiinae, as the sister taxon of Australia, South Africa, and Madagascar, and thus consonant with Mahavisaurus dentatus from the Lower Triassic of Madagascar an Early Triassic age for the Sanga do Cabral Supersequence. (Dias-da-Silva and Marsicano, 2011: Figs. 4a, b and 5a, ba, b). More Another temnospondyl specimen that is conguent with an Early importantly, none of the optimal trees placed Sangaia either in a Triassic age for the Sanga do Cabral Formation is MCN PV2606, an basal position among rhytidosteids or in a sister group relationship indeterminate plagiosternine plagiosaurid stereospondyl (Dias-da- with Indobrachyops panchetensis, from the Lower Triassic of India, Silva and Ilha, 2009; Dias-da-Silva and Milner, 2010). Plagiosauridae which was actually recovered as a derived taxon nested within is an exclusively Triassic group of trematosaurian stereospondyls Derwentiinae (Dias-da-Silva and Marsicano, 2011: Figs. 2b, 3a, b, known from Germany, Greenland, Russia, Spitsbergen, and possibly 4a, b, and 5a, b). the Upper Triassic of Thailand (Suteethorn et al., 1988; Hellrung, Peculiarly, Pineiro~ et al. (2015) contribution presents some 2003; Jenkins et al., 2008). The oldest record of this group comes misquotings and omissions regarding Sangaia and other temno- from the Lower Triassic of Eastern Europe (Shishkin, 1967). When spondyl taxa. According to these authors, Sangaia “is a basal rhy- defending their attribution of a Permian age to the SCF, Pineiro~ et al. tidosteid that possesses some characters not previously reported in (2015) failed to consider MCN PV2606 and mentioned neither Dias- any other Triassic member of the group, such as the presence of a da-Silva and Ilha (2009) nor Dias-da-Silva and Milner (2010). lacrimal bone, which, instead, occurs in the only known putative Regarding the temnospondyl content of the Sanga do Cabral Permian rhytidosteid Trucheosaurus major”. As stated above, the Formation, the available evidence from diagnostic temnospondyl comprehensive reassessment of Rhytidosteidae (Dias-da-Silva and specimens (UMVT 4302, UMVT 4303: Sangaia lavinai; UFSM 11408: Marsicano, 2011) did not recover Sangaia as a basal rhytidosteid. Tomeia witecki; MCN PV2606; a plagiosternine plagiosaurid) only Also, Pineiro~ et al. (2015) affirm that the Permian Trucheosaurus support an Early Triassic age for this unit. major (see Marsicano and Warren, 1998) is the only rhytidosteid The Uruguayan Buena Vista Formation (BVF) records a dvino- that possesses a lacrimal bone. In the original description of Indo- saurid (Marsicano et al., 2000), the plagiosauroid Uruiella liminea, brachyops panchetensis, von Huene and Sahni (1958) considered the rhinesuchid-like temnospondyl Arachana nigra, and mandib- that a lacrimal bone is present. It is important to note that the ular specimens tentatively ascribed to mastodonsauroids. taxonomic status of Indobrachyops was subject of great controversy The BVF temnospondyl reported by Marsicano et al. (2000) is an (see Welles and Estes, 1969; Cosgriff, 1969; Cosgriff and Zawiskie, incomplete specimen comprising a skull table fragment with a left 1979; Warren and Black, 1985; Marsicano and Warren, 1998; orbital margin. The authors stated that it has affinities with a basal Schoch and Milner, 2000). However, none of these controversial temnospondyl clade (Dvinosaurus þ Tupilakosauridae) which statements cast doubt on the presence of a lacrimal bone in Indo- occurs in Laurasian Upper Permian and Lower Triassic deposits. brachyops. As the only exception, Schoch and Milner (2000) stated Given the temporal range of Dvinosaurus þ Tupilakosauridae, their that the “lachrymal is possibly retained, although evidence is poor” presence in the Lower Triassic of South Africa (Karoo Basin), in Indobrachyops. Dias-da-Silva and Marsicano (2011) pointed out Marsicano et al. (2000) had only hypothesized an invasion of that further preparation of Indobrachyops is necessary in order to members of this clade as early as the Late Permian. However, the gather more information regarding its morphology. Moreover, none fragmentary nature of the specimen (not assigned to any previously of the aforementioned contributions (Welles and Estes, 1969; known or new species) casts doubt if it really represents a member Cosgriff, 1969; Cosgriff and Zawiskie, 1979; Warren and Black, of this group, a question that only can be answered with the 1985; Marsicano and Warren, 1998; Schoch and Milner, 2000) collection of more complete and better preserved material. presented a comprehensive computer-based phylogenetic analysis Pinheiro~ et al. (2007a) presented and described Uruyiella liminea to support its taxonomic statement, which was provided by Dias- a temnospondyl with a broadly triangular skull and laterally 290 S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 projecting posterior corners, with rhytidosteid-like dermal orna- nominal temnospondyl taxa with a known Permian range in both ment and laterally placed orbits. Uruyiella lacks both tabular horns the SCF and the BVF. and otic notches. Pineiro~ et al.'s (2007b) phylogenetic analysis placed Uruyiella as the sister taxon of Early Triassic Laidleria (which 2.5. Synapsid data they described as “enigmatic”) creating the ‘bigeneric’ Laidleriidae, which was recovered in a sister-group relationship with Plagio- The synapsid record from the Sanga do Cabral Supersequence is sauridae, both as non-stereospondyls within the basal clade Dvi- based upon few fragmentary and sparse remains, and a resolved nosauria. The authors commented that Laidleria and Plagiosauridae taxonomy for them is likely problematic. The real reasons of the are nested deeply within Stereospondyli in most phylogenies and lack of well preserved and diagnostic remains of synapsids, a group claimed that the close relationship of Uruyiella and Laidleria outside relatively common in other localities worldwide in the Early Stereospondyli would suggest a ghost lineage for the latter genus Triassic, are still unknown, but taphonomic biases in the known into the earliest Triassic and maybe even into the Late Permian, fossiliferous outcrops could be the reason of such condition. The which in turn would suggest survivorship of the Laidleriidae hitherto described synapsid material includes several partial through the Permo-Triassic extinction event. Later, Dias-da-Silva appendicular bones referred to Cynodontia indet. from the Catu- and Marsicano (2011) included both Laidleria and Uruyiella in çaba, Bica Sao~ Tome, and Granja Palmeiras localities (Abdala et al., their comprehensive phylogenetic analysis of Rhytidosteidae, but 2002; Da-Rosa et al., 2009; Dias-da-Silva and Da-Rosa, 2011) and none of their optimal trees recovered these genera as sister taxa two stapes (MCN 1072 and MCN 1073) collected along the highway (Dias-da-Silva and Marsicano, 2011, figs 2, 3, 4, and 5). Moreover, in BR 158 that were tentatively referred to Lystrosaurus (Langer and the cladogram depicted in Fig. 5B, some of the chosen outgroups Lavina, 2000). form a basalmost clade that includes Uruyiella, Peltobatrachus, Pineiro~ et al. (2015) criticized the cynodont affinities of the Rhineceps, Australerpeton, and Rhytidosteus. The Uruguayan taxon specimen described by Abdala et al. (2002), claiming that the major was found to be distantly related to Laidleria in contrast with the trochanter is not observable in any femora (UFRGS-PV-351T, UFRGS- results of Pineiro~ et al. (2007b). Laidleria was nested within Rhyti- PV-354T, UFRGS-PV-375T; Abdala et al., 2002: Figs. 2 and 3 and page dosteidae in a basal position (Node K, fig. 5B) as the sister taxon of 96), arguing instead that they display an overall morphology similar the Permian Trucheosaurus, both more derived than Nonolania to sphenacodontid and basal diapsids. The claim of Pineiro~ et al. (a definite Early Triassic taxon). Pineiro~ et al. (2015) also did not (2015) is likely prompted because a posterolateral projection is at considered all this conflicting information regarding both Uruyiella least clearly observed in UFRGS-PV-351 (labeled as major trochanter and Laidleria. Of course, all this contrasting information does not in Abdala et al., 2002: Figs. 2 and 3). However, this structure is not prevent the possibility that the BVF presents Permian strata but separated from the femoral head by a distinctive notch. As noted by also do not corroborate such claim (further discussion is provided Abdala et al. (2002), the lack of this notch is also common in non- in the end of this section). eucynodont cynodonts, such as procynosuchids, galesaurids, and Pineiro~ et al. (2007c) reported fragmentary temnospondyl thrinaxodontids (e.g., Jenkins, 1971; Kemp, 1980). In derived eucy- mandibles from the Buena Vista Formation of Uruguay, referring nodonts, the major trochanter is a structure usually more distin- them to Mastodonsauridae, claiming that these specimens were guished from the head (e.g. Jenkins, 1971). Nonetheless, in our view, phenetically similar to Gondwanan and Laurasian Early Triassic the isolated proximal portion of femora cannot be referred to Cyn- taxa, such as Watsonisuchus, Wetlugasarus, and Parotosuchus. They odontia with certainty. The intertrochanteric fossa with several also stated that these specimens present some characters not small foramina and delimited by a distal edge, the well-developed previously described for Mesozoic temnospondyls. For instance, internal trochanter (¼ lesser trochanter), and the small, unnotched presence of two parallel rows of teeth on both posterior and middle posterolateral projection (¼ major trochanter) are features observed coronoids (seen in FCDPV 1280) and not previously been reported in several basal archosauromorphs and procolophonids, which are in mastodonsaurids, and argued that a double row of coronoid the most common tetrapod groups (see the other sections) at the teeth is also present in some Permian rhinesuchids from Brazil. It is localities from which these femora have been collected. Therefore, important to point out that at least one Rhinesuchid genus used to the isolated, partial femora of Sanga do Cabral described by Abdala be considered as an Early Triassic survivor, Uranocentrodon from et al (2002: Figs. 2, 3). and those mentioned by Da-Rosa et al South Africa (Broom, 1912; see also Schoch and Milner, 2000; (2009: page 71 and fig. 5). and Dias-da-Silva and Da-Rosa (2011: Schoch, 2013). However, according to Latimer et al (2002, see also Fig. 8 and page 166) are likely not synapsid. Marsicano et al., 2017). the level that produced Uranocentrodon is With regard to the distal portion of femur (UFRGS-PV-353T; uppermost Permian. Even considering the presence of this ‘rhine- Abdala et al., 2002: fig. 4 and page 96) Pineiro~ et al. (2015) stated suchid’ feature, Pinheiro et al. (2007c) tentatively ascribed FC-DPV that it is also similar to that of Lystrosaurus and sphenacodontids. 1280 to a mastodonsauroid. Similarly, Abdala et al. (2002) stated that this specimen could not In their description of Arachana nigra from the BVF, Pineiro~ et al. be attributed with certainty to a cynodont. As stated before, this (2012) placed the unit on the Permo-Triassic boundary. They stated specimen possibly does not belong to Synapsida and provides no that this new Uruguayan taxon displays a transitional morphology, data whatsoever for age estimation. which precludes a taxonomic assignment to any known temno- A distal end of humerus (UFRGS-PV-332T) was described as spondyl group. Notwithstanding, Pacheco et al. (2016) included it in belonging to Cynodontia (Abdala et al., 2002: fig. 5 and page 96). a broad phylogenetic analysis, where Arachana nigra was nested Contrarily, Pineiro~ et al. (2015) stated that it shares a similar size within a clade that only includes Triassic taxa (Pacheco et al., 2016: and poorly developed trochlear and capitular articulations with pages 9e10, figs 5e6), revealing that its morphology is consistent varanopid eupelycosaurs and, in particular according to those au- with that present in taxa that evolved and lived after the Permo- thors, that of Varanops brevirostris from the Lower Permian of Texas Triassic event. (Campione and Reisz, 2010). In our view, UFRGS-PV-332T has a It is important to point out that all relevant, published infor- combination of features that suggest exclusion from Synapsida: the mation about temnospondyls from both SCF and BVF should be humeral distal end lacks the flaring supinator process, the ectepi- reviewed and discussed by Pineiro~ et al. (2015) as they were pro- condylar groove, and the ectepicondylar foramen. In contrast, these posing a paradigmatic change for the age of the SCF. At this point, structures are commonly present in most non-mammaliaform we affirm that there is no compelling evidence for the presence of synapsids, including most Permian non-therapsid families such as S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 291 varanopids (Varanops and Watongia; Reisz and Laurin, 2004; Langer and Lavina, 2000); (2) dorsal vertebrae, exhibiting a Campione and Reisz, 2010), ophiacodontids (e.g. Ophiacodon; morphology similar to that of Proterosuchus, Euparkeria, and/or Romer and Price, 1940), and sphenacodontids (Romer, 1922; Romer Erythrosuchus, and a fragment of ilium referred to Arch- and Price, 1940), as well as Late Permian and Triassic therapsids, osauriformes indet. (Da-Rosa et al., 2009); (3) several remains such as many dicynodonts (e.g., Cistecephalus, Dicynodontoides, classified only as Archosauromorpha indet. due to their incom- Ischigualastia; Cox, 1965; Cluver, 1978; Angielczyk et al., 2009), pleteness (Da-Rosa et al., 2009; Dias-da-Silva and Da-Rosa, 2011); basal therocephalians (Cynariognathus; Cys, 1967), and most cyn- and (4) the recently described taxon Teyujagua paradoxa, based on a odonts (e.g., Thrinaxodon, Exaeretodon, Chiniquodon; Bonaparte, well preserved skull with associated lower jaw and cervical verte- 1963; Jenkins, 1971; Abdala, 1999). Therefore, the suggestion of brae, and recovered phylogenetically as the sister taxon to Arch- Pineiro~ et al. (2015) is entirely superficial because the Lower osauriformes (Pinheiro et al., 2016). Below the Permo-Triassic Permian Varanops brevirostris (Campione and Reisz, 2010) has a boundary, the global record of archosauromorphs is extremely typical supinator process and associated groove that is not present scarce with only four nominal species from the Upper Permian in UFRGS-PV-332T. The area of the entepicondylar foramen of of Europe and Africa: Protorosaurus speneri from Germany and UFRGS-PV-332T is partially broken but the mediodistal edge of it is England (Meyer, 1832; Gottmann-Quesada and Sander, 2009); preserved. The foramen seems to be large and, as suggested by Archosaurus rossicus and Eorasaurus olsoni from Russia (Tatarinov, Pineiro~ et al. (2015), the way it passes through the entepicondyle is 1960; Sennikov, 1988, 1997; Gower and Sennikov, 2000), and not as characteristic as in therapsids (e.g., Jenkins, 1971; Angielczyk Aenigmastropheus parringtoni from Tanzania (Ezcurra et al., 2014). et al., 2009). The articular condyles are only partially preserved and Particularly in South America, the Permian record consists of a little can be said about them. However, they seem to be not as single incomplete humerus of a basal Archosauromorpha indet. developed as in therapsids. from the Rio do Rastro Formation (Martinelli et al., 2016). Other The lack of supinator process and ectepicondylar foramen, the remains of Archosauromorpha (portion of a basicranium and three wide distal end, the well-developed entepicondyle and entepi- vertebrae) were described from the Buena Vista Formation and condylar foramen, and the poorly developed condyles of UFRGS- considered being Permo-Triassic in age (Pineiro~ and Ubilla, 2003; PV-332-T (Abdala et al., 2002: fig. 5 and page 96) are features Ezcurra et al., 2015). The sparse material does not add much in- shared with procolophonids (e.g. deBraga, 2003). Due to the formation on age and, as discussed above, the tetrapod fauna of this abundance of procolophonids in the Sanga do Cabral Super- unit does not support a Permian age (Dias-da-Silva et al., 2006a; sequence and the combination of features aforementioned, we Modesto and Botha-Brink, 2010). consider UFRGS-PV-332-T (Abdala et al., 2002: fig. 5 and page 96) Thus, the archosauromorph record from the Sanga do Cabral as a likely Procolophonidae indet. Notably, the maximum distal Formation is likely more indicative of a Triassic rather than Permian width of this humerus is 25 mm, representing a relative large age. The almost complete vertebra UFRGS-PV-493-T approaches procolophonid in this fauna association, as also indicated by other closely the morphology of Triassic tanystropheids rather than pro- procolophonid specimens (see also Procolophonid Data section). torosaurids (Wild, 1973; Nosotti, 2007; Peyer, 1937; Ezcurra et al., The other remains referred to Cynodontia are so fragmentary that 2014). The mid-sized vertebrae and fragment of ilium of arch- their affinities are impossible to access. osauriforms reported by Da-Rosa et al. (2009) from Bica Sao~ Tome The isolated stapes referred as to Lystrosaurus were never site, from the same layer yielding Procolophon, are also concomitant formally ilustrated and a personal observation of the material in- with the fossil record of basal archosauriforms (e.g. Proter- dicates that these specimens can be reidentified as the sacral ribs of osuchidae, Erythrosuchidae, Euparkeriidae) of the Lower Triassic of procolophonians. They have the similar morphology to sacral ribs I Africa, Russia, and China (e.g. Nesbitt, 2011; Smith and Botha, 2005; and II of Procolophon (see deBraga, 2003:Fig. 8B), and they have an Sookias and Butler, 2013). Although being by far the most complete oval, obliquely oriented rugose surface (located at the supposed archosauromorph so far recovered from the Sanga do Cabral part of the stapedial plate) that is the contact for the ilium. On the Supersequence, Teyujagua paradoxa has only limited biostrati- other hand, the surface to contact the transverse process is smaller. graphic significance. T. paradoxa was recovered by Pinheiro et al. Considering also the taphonomic environment in which these (2016) as the sister taxon to the Archosauriformes, the origin of specimens were discovered, the presence of two isolated stapes which dates back to the late Guadalupian/early Lopingian (Ezcurra and no other unequivocal dicynodont remains seems highly un- et al., 2014). Primitive archosauriforms are recovered both in likely to us. Lopingian and Lower Triassic strata, and T. paradoxa is interpreted as In conclusion, the data on Cynodontia is extremely limited and a relict taxon that survived the Permo-Triassic boundary (Pinheiro we state that there is no positive evidence to justify at present the et al., 2016). In fact, Prolacerta broomi, the sister taxon to the clade occurrence of this group in the Sanga do Cabral Supersequence. The of T. paradoxa þ Archosauriformes is known from the Lower Triassic stapes referred to Lystrosaurus are reinterpreted here as sacral ribs Lystrosaurus Assemblage Zone of South African Karoo and coeval of procolophonids, and further analyses are needed. In our view, deposits from Antarctica (see Modesto and Sues, 2004). only new discoveries of more complete specimens will permit to Although the record of archosauromorphs from the Sanga do elucidate the real diversity of synapsids in the Sanga do Cabral Cabral Supersequence still needs to be investigated in detail, a Formation. We conclude that the synapsid data is not bio- preliminary assessment suggests a taxonomic diversity that is stratigraphically informative and certainly cannot be used to sup- common to other Early Triassic tetrapod faunas worldwide. port a Permian age for the SCF. 2.7. Data on two problematic specimens 2.6. Archosauromorph data Based upon two specimens from the Sanga do Cabral Formation, At present, the record of archosauromorphs in the Sanga do Schultz and Dias-da-Silva (1999) suggested the possibility that Cabral Formation is fairly limited, but provides support for a Lower pareaisaurs would have crossed the Permo-Triassic Boundary sur- Triassic age and is suggestive of a high taxonomic diversity. Arch- viving during Early Triassic in the region that now comprises osauromorph specimens include: (1) a few isolated vertebrae Southern Brazil. Schultz and Dias-da-Silva (1999) identified these referred to Protorosauria indet., with tanystropheid affinities elements (UFRGS PV0506T and UFRGS PV0362T) as an olecranon (Langer et al., 1996; Langer and Schultz, 1997; Dias-da-Silva, 1998; process and skull fragment, respectively. Later, Cisneros et al. (2005) 292 S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 stated that the skull fragment does not preserve any feature diag- nostic of Pareiasauria, and regarded it to be an indeterminate specimen. Dias-da-Silva et al. (2005) reinterpreted UFRGS PV0506T as a glenoid/postglenoid area of an indeterminate temnospondyl. Pineiro~ et al. (2015) again disregarded the conclusions of both Cisneros et al. (2005) and Dias-da-Silva et al. (2005) in preference of the earlier interpretations (Schultz and Dias-da-Silva, 1999). We herein state that the putative pareiasaur specimens do not preserve any diagnostic character whatsoever that would be indicative of pareiasaurian affinities, in agreement with Cisneros et al. (2005) and Dias-da-Silva et al. (2005). Accordingly, the indeterminate na- ture of these specimens renders them biostratigraphically uninformative.

3. Description of new procolophonid specimens

3.1. Skull

UFSM 11409a is a partial skull with occluded mandible (Fig. 4), which we assign to Procolophon trigoniceps based on the presence of eight bicuspid maxillary teeth and a broad, rounded subtemporal emargination. Although much of the skull roof has been weathered away, the basal skull lengthdfrom the anterior margin of the premaxilla to the posterior margin of the occipital condyledis 85 mm, making UFSM 11409a the largest known skull of Procolo- phon trigoniceps. The left premaxilla houses four teeth. The ante- riormost bone forming the dorsal process has been weathered down to nearly its narial (posterior) margin and, judging from the procumbent dorsal process in specimens with more complete premaxillae (e.g. Carroll and Lindsay, 1985: fig. 5), ca. 3 mm of the anterior end of the snout is missing. The maxillae are indistin- guishable from previously described maxillae of the species (e.g. Carroll and Lindsay, 1985; Cisneros, 2008a). The interdigitating pattern between the upper and lower dentition described for the species by Carroll and Lindsay (1985) is present and well preserved on the left side of UFSM 11409a, but it is disrupted by slight Fig. 4. Partial skull of a large procolophonid, UFSM 11409a, in (A, B) ventral, (C, D) dorsal, (E, F) right lateral and (G, H) left lateral. crushing on the right side. Consonant with its large size, each Abbreviations: a, angular; art, articular; co, coronoid; d, dentary; ect, ectopterygoid; j, maxilla accommodates nine tooth positions. The first maxillary jugal; l, lacrimal; m, maxilla; n, nasal; p, parietal; pal, palatine; pm, premaxilla; po, tooth position is occupied by a unicuspid tooth, whereas all of the postorbital; pt, pterygoid; q, quadrate; qj, quadratojugal; sa, surangular; sof, suborbital more distal (posterior) positions appear to house bicuspid teeth. foramen; sp, splenial; q, squamosal; st, supratemporal. Scale bar equals 2 cm. The mandible is preserved in full occlusion with the skull, and the posterior ends of both mandibular rami have been weathered, and by Carroll and Lindsay (1985:fig. 1a, b). A hard carbonate layer both factors limit description of this structure. What can be encrusts the palate, and the occluded mandible makes mechani- observed, is indistinguishable from those specimens described by cal preparation of its ventral surface very difficult. It is clear, Carroll and Lindsay (1985). however, that there are at least two groups of small teeth. Based The nasals, the lacrimals, the frontals, the prefrontals, and the on other specimens (Cisneros, 2008a:Fig. 3), the anterior group of approximate two-thirds of the parietals have been lost to erosion. small teeth presumably represents the vomerine dentition, and The ventral half of the orbit is preserved from just anterior to the the posterior group represents the pterygoidal and probably jugal to the parietal's contribution to the posterior orbital margin. palatine dentition. Four teeth positioned posterolaterally on the The posterior end of the skull table preserves the weakly bicon- right side form a short row and presumably contribute to the cave margin formed by the paired parietals and the supra- dental row that parallels the medial margin of the pterygoid and temporals. The subtemporal emargination arises immediately borders the interpterygoid vacuity (Cisneros, 2008a:fig. 3a). As posterior to the distalmost maxillary tooth, peaks about 1 cm (or pointed out by Cisneros (2008a), neither position nor number of ca. four tooth positions) further posteriorly, and then descends to palatal teeth are constant in Procolophon due to individual the condylar region. Apart from some minor weathering to the variation. quadratojugal horn, the left posterolateral corner of the skull roof is well preserved, clearly showing a relatively large quadrate fo- ramen formed by the squamosal, the quadratojugal, and the 3.2. Vertebra and mandibular fragment quadrate, and that the squamosal extends slightly ventral to the quadratojugal in occipital view. The loss of most of the skull roof UNIPAMPA 0655 (Fig. 5a, b, c, d, and e) is a well-preserved has allowed preparation of the dorsal surface of the palate and specimen with high neural spines, laterally flanked by distinct the braincase. The morphology and the organization of the palatal buttresses that extend until the dorsal margins of the post- and braincase elements conforms to that described and illustrated zygapophysis, similar to P. trigoniceps (deBraga, 2003). Its trans- by Carroll and Lindsay (1985), except that the cultriform process verse processes lack the characteristic ‘wing-like’ morphology of the parabasisphenoid is relatively broader than reconstructed described by Pineiro~ et al. (2015), and do not protrude from the S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 293

Fig. 5. Procolophonoid material from the Sanga do Cabral Formation. UNIPAMPA 0655, procolophonoid posterior dorsal vertebra in (A) anterior, (B) posterior, (C) right lateral, (D) dorsal and (E) ventral views. UNIPAMPA 0680 Procolophon right mandibular ramus in (F) lateral, (G) medial, and (H) occlusal views.

Fig. 6. Strict consensus of 18 most parsimoniuous trees of 130 steps each, with a consistency index of 0.6308. In all trees, UFSM 11409a was recovered in a sister-group relationship with Procolophon.

prezygapophyseal buttresses, and this is compatible with being an Tome, a site that has so far produced at least three skulls (UFSM element from the posterior dorsal series. Although considerably 11409a, UFSM 11448, and UFSM/CAPPA 0189) as well as skull abraded, the ventral margin of the pleurocentrum has paired fragments clearly attributable to P. trigoniceps (e.g. Da Rosa et al., ventrolateral concavities separated by a double median ridge (a 2009). procolophonid synapomorphy, as discussed in 2.3. Procolophonid Data). UNIPAMPA 0655 (as well as other already mentioned vertebrae with similar morphology) were found in the same layer 4. Phylogenetic analysis as several maxillary fragments (one of them, UNIPAMPA 0680, depicted in Fig. 5f, g, and h) bearing a set of anteroposteriorly The data matrix comprises 24 taxa plus UFSM 11409a and 59 compressed bicuspidate molariform teeth, a morphology consid- characters. It is a modified version of the dataset by McDougall et al. ered as diagnostic of Procolophon (Carroll and Lindsay, 1985; (2013). Character coding for Procolophon and UFSM 11409a are Cisneros, 2008a). The specimens were found at the Bica Sao~ provided below: 294 S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296

Procolophon Acknowledgments

211111011020211112100100021(1/2)00(1/2)3101(0/1)10012101 We thank Rodrigo Temp Müller, Cristian Pereira Pacheco, and 010111 101101101. Jossano de Rosso Morais for technical support, and Oliver Rauhut for generously providing images of BSPM 1934 VIII 39. Conselho Nacional de Desenvolvimento Científico e Tecnologico financially UFSM 11409a supported the present work (CNPq, research grant to SDS, process number 306352/2016-8; research grant to CLS, process number ?11?11?1?1202111?21??1?00?1100 1310???0012???????????? 458187/2014-3; project funding to FLP, process number 407969/ ?????? 2016-0). New Opportunities Fund award from the Canadian Foun- The revised data matrix was run with PAUP 4.0b10 (Swofford, dation for Innovation and a grant from the Nova Scotia Research 2004) and subjected to a branch-and-bound search. Following the and Innovation Trust financially supported S.P.M. We are greatly instructions of MacDougall et al. (2013), characters 1, 8, 18, 26, 31, indebted to Dr. Michael Maisch and an anonymous referee for the and 32 were ordered. The analysis found 18 optimal trees, each of valuable comments and suggestions that greatly improved the final which is 130 steps in length, with a consistency index of 0.6308. In version of this article. all trees, UFSM 11409a was recovered in a sister-group relationship with Procolophon. Strict and majority rule consensus trees both present the same topology. The strict consensus tree is depicted in References Fig. 6. The sister-group relationship of UFSM 11409a with Procolo- phon reinforces our taxonomic assignment, preliminarily based on Abdala, F., 1996. Redescripcion del craneo y reconsideracion de la validez de Cynognathus minor (Eucynodontia-Cynodonthidae) del Triasico Inferior de tooth morphology. Mendoza. Ameghiniana 33, 115e126. Abdala, F., 1999. Elementos postcraneanos de Cynognathus (Synapsida-Cynodontia) del Triasico Inferior de la Provincia de Mendoza, Argentina. Consideraciones 5. Conclusions sobre la morfología del húmero en cinodontes. Rev. Espanola~ Paleontol. 14, 13e24. The available data do not support the presence of unquestion- Abdala, F., Dias-da-Silva, S., Cisneros, J.C., 2002. First record of nonmammalian cynodonts (Therapsida) in the Sanga do Cabral Formation (early Triassic) of able Permian fauna within the Sanga do Cabral Supersequence as southern Brazil. Palaeontol. Afr. 38, 93e98. claimed by Pineiro~ et al. (2015: page 12) work, neither in the SCF Andreis, R.R., Bossi, G.E., Montardo, D.K., 1980. O Grupo Rosario do Sul (Triassico) no Rio Grande do Sul-Brasil. XXXI Congresso Brasileiro de Geologia, Camboriú, nor in the BVF. The complete absence of shared vertebrate taxa (on e fi Sociedade Brasileira de Geologia. Anais 2, 659 673. speci c level) between the BVF and the SCF cannot yet be fully Andreis, R.R., Ferrando, L., Herbst, R., 1996. Terrenos Carboniferos y Permicos de explained, however, and as suggested by Modesto and Botha-Brink Republica Oriental del Uruguay. In: El Sistema Permico en la Republica (2010), both units may have slightly different ages. Probably, the Argentina y en la Republica Oriental del Uruguay. Academia Nacional del Uruguay, Cordoba, Argentina, pp. 309e343. BVF is slightly older than the SCF, but both were deposited during Angielczyk, K.D., Sidor, C.A., Nesbitt, S.J., Smith, R.M.H., Tsuji, L.A., 2009. Taxonomic the Early Triassic. revision and new observations on the postcranial skeleton, biogeography, and The statement by Pineiro~ et al. (2015) that ‘(the attribution of biostratigraphy of the dicynodont genus Dicynodontoides, the senior subjective synonym of Kingoria (Therapsida, Anomodontia). J. Vert. Paleontol. 29, the described vertebrae to seymouramorphs) is the hypothesis 1174e1187. that fits best the currently prevailing biostratigraphic scheme’ is Barberena, M.C., Lavina, E.L., Becker, M.R., 1981. Sobre a presença de tetrapodos na clearly a circular argument, as the proposed Permian age for the Formaçao~ Sanga do Cabral (Grupo Rosario do Sul), Triassico do Rio Grande do Sul, Brasil. In: Congresso Latino-Americano de Palaeontologia. Porto Alegre, SCS is mainly based upon these same vertebrae. The lack of e ~ Anais, vol. 2, pp. 295 306. biostratigraphic accuracy is evident in Pineiro et al. (2015) prop- Benton, M.J., Tverdokhlebov, V.P., Surkov, M.V., 2004. Ecosystem remodeling among ositions, because they ignore the confirmed presence of a bio- vertebrates at the Permian-Triassic boundary in Russia. Nature 432, 97e100. stratigraphically important genus (Procolophon) that occurs well Bertoni-Machado, C., Dias-da-Silva, S., Holz, M., Schultz, C.L., 2008. Assinaturas tafonomicas^ da tafocenose de vertebrados da Superseqüencia^ Sanga do Cabral above the P-Tr boundary in correlative deposits, in favor of (Triassico Inferior, sul do Brasil), evidencias^ de time averaging e suas tentative attributions of fragmentary remains. implicaçoes~ em analises bioestratigraficas. In: VI Simposio Brasileiro de Pale- ~ ~ With regards to the temnospondyl content of the Sanga do ontologia de Vertebrados, Ribeirao Preto. Paleontologia em Destaque. Ediçao Especial, vol. 1, pp. 47e48. Cabral Formation, it is consistent with taxa of overall and world- Bonaparte, J.F., 1963. Descripcion del esqueleto postcraneano de Exaeretodon wide Early Triassic distribution, whereas the BVF specimens do not (Cynodontia-Traversodontidae). Acta Geol. Lilloana 4, 5e54. represent definite Permian temnospondyls. Botha, J., Smith, R.M.H., 2006. Rapid vertebrate recuperation in the Karoo Basin of South Africa following the end-Permian extinction. J. Afr. Earth Sci. 45, The available synapsid material likely belongs to other groups, 502e514. possibly procolophonids and/or basal archosauromorphs, both of Botha, J., Modesto, S.P., Smith, R.M.H., 2007. Extended procolophonoid reptile which are abundant in Sanga do Cabral localities. Therefore, we survivorship after the end-Permian extinction. South Afr. J. Sci. 103, 54e56. Broili, F., Schroder,€ J., 1936. Über Procolophon Owen. In: Beobachtungen an regard synapsids to be extremely rare at the worked localities, and Wirbeltieren der Karrooformation. Verlag der Bayerischen Akademie der Wis- the described specimens have no biostratigraphic utility regarding senschaften, vol. 1936, pp. 239e255. the age of the Sanga do Cabral Supersequence. Broom, R., 1912. Note on the temnospondylous stegocephalian, Rhinesuchus. Trans. Geol. Soc. S. Afr. 14, 79e81. In conclusion, on the basis of a comprehensive and parsimo- Campione, N.E., Reisz, R.R., 2010. Varanops brevirostris (Eupelycosauria, Varanopi- nious reevaluation of the available geological and paleontological dae) from the Lower Permian of Texas with discussion of varanopid data regarding the Sanga do Cabral Formation, an Early Triassic age morphology and interrelationships. J. Vert. Paleontol. 30, 724e746. is well supported for this South Brazilian unit, and there is Carroll, R.L., Lindsay, W., 1985. Cranial anatomy of the primitive reptile Procolophon. Can. J. Earth Sci. 22, 1571e1587. no compelling evidence whatsoever to assume that any known Cisneros, J.C., 2008a. Taxonomic status of the reptile genus Procolophon from the outcrop/level could be Permian in age. Furthermore, there is not a Gondwanan Triassic. Palaeontol. Afr. 43, 7e17. single fossil in BVF that unequivocally points to a Permian age for Cisneros, J.C., 2008b. Phylogenetic relationships of procolophonid parareptiles with remarks on their geological record. J. Syst. Paleontol. 6, 345e366. that unit also, so that all the SCS should be more properly consid- Cisneros, J.C., Schultz, C.L., 2002. Procolophon brasiliensis n. sp., a new procolophonid ered Early Triassic. Beyond fossil data, it is important to point out reptile from the Lower Triassic of southern Brazil. Neues Jahr. Geol. Palaontol.€ that a chronostratigraphy based on absolute dating, both in SCF and Mon. 2002, 641e648. fi Cisneros, J.C., Abdala, F., Malabarba, M.C., 2005. Pareiasaurids from the Rio do Rastro in BVF, is necessary in order to provide a re ned estimation of the Formation, Southern Brasil: biostratigraphic implications for permian faunas of age of the Sanga do Cabral Supersequence. the Parana Basin. Rev. Bras. Paleontogia 8, 13e24. S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296 295

Cluver, M.A., 1978. The skeleton of the mammal-like reptile Cistecephalus with Fildani, A., Weislogel, A., Drinkwater, N.J., McHargue, T., Tankard, A., Wooden, J., evidence for a fossorial mode of life. Ann. South Afr. Mus. 76, 213e246. Hodgson, D., Flint, S., 2009. U-Pb zircon ages from the southwestern Karoo Coney, L., Reimold, W.U., Hancox, P.J., Mader, D., Koeberl, C., McDonald, I., Struck, U., Basin, South Africa e implications for the Permian-Triassic boundary. Geology Vajda, V., Kamo, S.L., 2007. Geochemical and mineralogical investigation of the 37, 719e722. PermianeTriassic boundary in the continental realm of the southern Karoo Gottmann-Quesada, A., Sander, P.M., 2009. A redescription of the early arch- Basin, South Africa. Palaeoworld 16, 67e104. osauromorph Protorosaurus speneri Meyer, 1832, and its phylogenetic re- Cosgriff, J.W., 1969. Blinasaurus, a brachyopid genus from Western Australia and lationships. Palaeont. Abt. A 287, 123e220. new south wales. J. Roy. Soc. West. Aust. 52, 65e88. Gower, D.J., Sennikov, A.G., 2000. Early archosaurs from Russia. In: Benton, M.J., Cosgriff, J.W., Zawiskie, J.M., 1979. A new species of the Rhytidosteidae from from Kurochkin, E.N., Shishkin, M.A., Unwin, D.M. (Eds.), The Age of Dinosaurs in the Lystrosaurus zone and a review of the Rhytidosteoidea. Palaeontol. Afr. 22, Russia and Mongolia. Cambridge University Press, Cambridge, pp. 140e159. 1e27. Gulbranson, E.L., Ciccioli, P.L., Montanez,~ I.P., Marenssi, S.A., Limarino, C.O., Cox, C.B., 1965. New Triassic dicynodonts from South America, their origin and Schmitz, M.D., Davydov, V., 2015. Paleoenvironments and age of the Talampaya relationships. Phil. Trans. R. Soc. Lond. B 248, 57e514. Formation, the Permo-Triassic boundary in northwestern Argentina. J. South Cys, J.M., 1967. Osteology of the pristerognathid Cynariognathus platyrhinus (Rep- Amer. Earth Sci. 63, 310e322. tilia, Theriodontia). J. Paleontol. 41, 776e790. Hansma, J., Tohver, E., Schrank, C., Jourdan, F., Adams, D., 2015. The timing of the Da-Rosa, A.A.S., Dias-da-Silva, S., 2009. Vertebrate-bearing outcrops of the Sanga do Cape orogeny: new 40Ar/39Ar age constraints on deformation and cooling of the Cabral Formation, lower Triassic of southern Brazil, and correlation possibilities. Cape fold belt. South Afr. Gond. Res. 32, 122e137. In: Reunion Anual de Comunicaciones de la Asociacion Paleontologica Hellrung, H., 2003. Gerrothorax pustuloglomeratus, ein Temnospondyle (Amphibia) Argentina, Buenos Aires. Ameghiniana, vol. 2009, p. 34. mit knocherner€ Branchialkammer aus dem Unteren Keuper von Kupferzell Da-Rosa, A.A.S., Faccini, U.F., 2005. Delimitaçao~ de blocos estruturais de diferentes (Süddeutschland). Stutt Beitr. Natur. Ser. B, Geol. Palaontol.€ 330, 1e130. escalas em seqüencias^ mesozoicas do Estado do Rio Grande do Sul, implicaçoes~ Holz, M., Souto-Ribeiro, A., 2000. Taphonomy of the south-Brazilian Triassic ver- bioestratigraficas. Gaea 1, 16e23. tebrates. Rev. Bras. Geocien.^ 30, 487e490. Da-Rosa, A.A.S., Pineiro,~ G., Dias-da-Silva, S., Cisneros, J.C., Feltrin, F.F., Neto, L.W., Holz, M., França, A.B., Souza, P.A., Iannuzzi, R., Rohn, R., 2010. A stratigraphic chart of 2009. Bica Sao~ Tome, um novo sítio fossilífero para o Triassico Inferior do sul do the Late Carboniferous/Permian succession of the eastern border of the Parana Brasil. Rev. Bras. Paleontol. 12, 67e76. Basin, Brazil, South America. J. S. Am. Earth Sci. 29, 381e399. Da-Rosa, A.A.S., Dias-da-Silva, S., Pinero,~ G., Marsicano, C., Schultz, C.L., Horn, B.L., Jenkins, F.A., 1971. The postcranial skeleton of African cynodonts. Bull. Peabody Mus. 2010. Comparaçao~ litofaciologica entre as formaçoes~ Sanga do Cabral (Eotrias- Nat. Hist. 36, 1e216. sico, Bacia do Parana, sul do Brasil) e Buena Vista (Permo-Triassico, Cuenca Jenkins Jr., F.A., Schubin, N.H., Gatesy, S.M., Warren, A., 2008. Gerrothorax pulcher- Norte, Uruguai). In: VII Simposio Brasileiro de Paleontologia de Vertebrados. Rio rimus from the upper Triassic fleming fjord Formation of East Greenland and a de Janeiro: Boletim de Resumos, vol. 17. reassessment of head lifting in temnospondyl feeding. J. Vert. Paleontol. 28, deBraga, M., 2003. The postcranial skeleton, phylogenetic position, and probable 935e950. lifestyle of the Early Triassic reptile Procolophon trigoniceps. Can. J. Earth Sci. 40, Kemp, T.S., 1980. The primitive cynodont Procynosuchus, structure, function, and 527e556. evolution of the postcranial skeleton. Phil. Trans. Roy. Soc. Lond., Ser. B 288, Dias-da-Silva, S., 1998. Novos achados de vertebrados fosseis na Formaçao~ Sanga do 217e258. Cabral (Eotriassico da Bacia do Parana). Acta Geol. Leopoldensia 21, 101e108. Kidwell, S.M., Flessa, K.W., 1995. The quality of the fossil record, populations, spe- Dias-da-Silva, S., Da-Rosa, A.A.S., 2011. Granja Palmeiras, a new fossiliferous site for cies, and communities. Ann. Rev. Ecol. Syst. 26, 269e299. the Lower Triassic of southern Brazil. Rev. Bras. Paleontol. 14,157e14,168. Lanci, L., Tohver, E., Wilson, A., Flint, S., 2013. Upper permian magnetic stratigraphy Dias-da-Silva, S., Dias, E.V., 2013. A comprehensive survey of Triassic stereospondyls of the lower Beaufort group, Karoo Basin. Earth Planet Sci. Let. 375, 123e134. from Southern Brazil with comments on their overall significance. Bull. - New Langer, M.C., Lavina, E.L., 2000. Os amniotas do Neopermiano e Eotriassico da Bacia Mex. Mus, Nat. Histor. Sci. 61, 93e103. do Paranadrepteis e “repteis mamaliformes”. In: Holz, M., de Ros, L.F. (Eds.), Dias-da-Silva, S., Ilha, A.L.R., 2009. On the presence of a pustulated temnospondyl in Paleontologia Do Rio Grande Do Sul. CIGO/UFRGS, Porto Alegre, pp. 210e235. the Lower Triassic of southern Brazil. Acta Palaeontol. Pol. 54, 609e614. Langer, M.C., Schultz, C.L., 1997. Further comments on the Brazilian protorosaurs. Dias-da-Silva, S., Marsicano, C., 2011. Phylogenetic reappraisal of Rhytidosteidae Ameghiniana 34, 537. (Stereospondyli, Trematosauria), temnospondyl amphibians from the permian Langer, M., Schultz, C.L., Lavina, E.L., 1996. A possible Protorosauria Lydekker, 1888 and Triassic. J. Syst. Paleontol. 9, 305e325. (Reptilia, Diapsida) from the southern Brazilian lower Triassic. An. Acad. Bras. Dias-da-Silva, S., Milner, A.R., 2010. The pustulated temnospondyl revisited - a Cien.^ 68, 289e290. plagiosternine plagiosaurids from the Lower Triassic of Brazil. Acta Paleontol. Latimer, E.M., Hancox, P.J., Rubidge, B.S., Shishkin, M.A., Kitching, J.W., 2002. The Pol. 55, 561e563. temnospondyl amphibian Uranocentrodon, another victim of the end-Permian Dias-da-Silva, S., Marsicano, C., Schultz, C.L., 2005. Early Triassic temnospondyl skull extinction event. South Afr. J. Sci. 98, 191e193. fragments from southern South America (Parana Basin, Brazil). Rev. Bras. Laurin, M., Reisz, R.R., 1995. A reevaluation of early amniote phylogeny. Zool. J. Linn. Paleontol. 8, 165e172. Soc. 113, 165e223. Dias-da-Silva, S., Modesto, S.P., Schultz, C.L., 2006a. New material of Procolophon Lavina, E.L., 1983. Procolophon pricei sp. n. um novo reptil procolofonídeo do (Parareptilia, Procolophonoidea) from the lower Triassic of Brazil, with remarks Triassico do Rio Grande do Sul. Iheringia. SerZool. 9, 51e78. on the ages of the Sanga do Cabral and Buena Vista formations of South Lavina, E.L., Barberena, M., 1985. Anfíbios ritidosteídeos e lidekkerinídeos da For- America. Can. J. Earth Sci. 43, 1695e1693. maçao~ Sanga do Cabral (Triassico Inferior do Rio Grande do Sul) - implicaçoes~ Dias-da-Silva, S., Marsicano, C., Schultz, C.L., 2006b. Rhytidosteid temnospondyls in bioestratigraficas e geocronologicas. Iheringia, Ser. Zool. 10, 19e27. Gondwana, a new taxon from the lower Triassic of Brazil. Palaeontology 49, Lopez-Gamundi, O., 2006. Permian plate margin volcanism and tuffs in adjacent 381e390. basins of west Gondwana: age constraints and common characteristics. J. South Dias-da-Silva, S., Schultz, C.L., 2008. Early Triassic postcranial temnospondyl re- Amer. Earth Sci. 22, 227e238. mains from southern Brazil (Sanga do Cabral Formation, Parana Basin). Rev. Macdougall, M.J., Modesto, S.P., 2011. New Information on the skull of the Early Bras. Paleontol. 11, 51e58. Triassic parareptile Sauropareion anoplus, with a discussion of tooth attachment Eltink, E., Da-Rosa, A.A.S., Dias-da-Silva, S., 2016. A Capitosauroid from the Lower and replacement in procolophonids. J. Vert. Paleontol. 31, 270e278. Triassic of South America (Sanga do Cabral Supersequence: Parana Basin), its MacDougall, M.J., Modesto, S.P., Botha-Brink, J., 2013. The postcranial skeleton of the Phylogenetic Relationships and Biostratigraphic Implications. Historical Biology. Early Triassic parareptile Sauropareion anoplus, with a discussion of possible life http://dx.doi.org/10.1080/08912963.2016.1255736. history. Acta Palaeontol. Pol. 58, 737e749. Ezcurra, M.D., Lecuona, A., Martinelli, A., 2010. A new basal archosauriform diapsid Marsicano, C.A., Warren, A.A., 1998. The first Palaeozoic rhytidosteid: Trucheosaurus from the Lower Triassic of Argentina. J. Vert. Paleontol. 30, 1433e1450. major (Woodward, 1909) from the late permian of Australia, and a reassessment Ezcurra, M.D., Scheyer, T., Butler, R.J., 2014. The origin and early evolution of Sauria, of the Rhytidosteidae (Amphibia, Temnospondyli). Bull. Nat. Hist. Mus. Lond., reassessing the Permian saurian fossil record and the timing of the crocodile- Geol. 54, 147e154. lizard divergence. PLoS One 9, e89165. Marsicano, C., Perea, D., Ubilla, M., 2000. A new temnospondyl amphibian from the Ezcurra, M.D., Velozo, P., Meneghel, M., Pineiro,~ G., 2015. Early archosauromorph Lower Triassic of South America. Alcheringa 24, 119e123. remains from the Permo-Triassic Buena Vista formation of north-eastern Marsicano, C.A., Latimer, E., Rubidge, B., Smith, R.M.H., 2017. The Rhinesuchidae and Uruguay. PeerJ 3, e776. early history of the Stereospondyli (Amphibia: Temnospondyli) at the end of Faccini, U.F., 1989. O Permo-Triassico do Rio Grande do Sul. Uma analise sob o ponto the Palaeozoic. Zool. J. Linn. Soc. 2017, zlw032. http://dx.doi.org/10.1093/zoo- de vista das sequencias^ deposicionais [Master Dissertation]. Universidade linnean/zlw032. Federal do Rio Grande do Sul, Porto Alegre, 133 pp. Martinelli, A.G., Soares, M.B., 2016. Evolution of South American cynodonts. Rev Feltrin, F.F., Dias-da-Silva, S., Pineiro,~ G., Da-Rosa, A.A.S., 2008. Analise filogenetica Mus Argentino cien Nat “Bernardino Rivadavia”. Nueva Ser. 6, 183e197. preliminar de um mastodonsauroide (Temnospondyli: Stereospondyli) da For- Martinelli, A.G., Kammerer, C.F., Melo, T.P., Paes Neto, V.D., Ribeiro, A.M., Da- maçao~ Sanga do Cabral, Triassico Inferior do sul do Brasil. In: Encontro Regional Rosa, A.A.S., Schultz, C.L., Soares, M.B., 2017. The African cynodont Aleodon da Sociedade Brasileira de Paleontologia - Paleo RS 2008. Paleontologia em (Cynodontia, Probainognathia) in the Triassic of southern Brazil and its Destaque, Porto Alegre, 62, 23. biostratigraphic significance. Plos One 12 (6), e0177948. http://dx.doi.org/ Fildani, A., Drinkwater, N.J., Weislogel, A., McHargue, T., Hodgson, D.M., Flint, S.S., 10.1371/journal.pone.0177948. 2007. Age controls on the Tanqua and Laingsburg deep-water systems, new Martinelli, A.G., de la Fuente, M., Abdala, F., 2009. Diademodon tetragonus Seeley, insights on the evolution and sedimentary fill of the Karoo Basin. S. Afr. J. 1894 (Therapsida, Cynodontia) in the Triassic of South America and its Sediment. Res. 77, 901e908. biostratigraphic implications. J. Vert. Paleontol. 29, 852e862. 296 S. Dias-da-Silva et al. / Journal of South American Earth Sciences 79 (2017) 281e296

Martinelli, A.G., Francischini, H., Dentzien-Dias, P.C., Soares, M.B., Schultz, C.L., 2016. Romer, A.S., Price, L.I., 1940. Review of the Pelycosauria. Geol. Soc. Am. Spec. Pap. 28, The oldest archosauromorph from South America, postcranial remains from the 1e538. Guadalupian (mid-Permian) Rio do Rasto Formation (Parana Basin), southern Rubidge, B.S., Erwin, D.H., Ramezani, J., Bowring, S.A., de Klerk, W.J., 2013. High- Brazil. Hist. Biol. 29, 76e84. precision temporal calibration of Late Permian vertebrate biostratigraphy, U-Pb McKay, M.P., Weislogel, A.L., Fildani, A., Brunt, R.L., Hodgson, D.M., Flint, S.S., 2015. zircon constraints from the Karoo Supergroup, South Africa. Geology 41, U-PB zircon tuff geochronology from the Karoo Basin, South Africa, implications 363e366. of zircon recycling on stratigraphic age controls. Intern. Geol. Rev. 57, 93e410. Sahney, S., Benton, M.J., 2008. Recovery from the most profound mass extinction of Meyer, H. von, 1832. Palaeologica zur Geschichte der Erde und ihrer Geschopfe.€ all time. Proc. Roy. Soc. Ser. B 275, 759e765. Frankfurt am Main: Verlag von Siegmund Schmerber, p. 560. Sato, A.M., Llambías, E.J., Basei, M.A.S., Castro, C.E., 2015. Three stages in the Late Modesto, S.P., 1999. Observations of the structure of the early permian reptile Paleozoic to Triassic magmatism of southwestern Gondwana, and the re- Stereosternum tumidum Cope. Palaeontol. Afr. 35, 7e19. lationships with the volcanogenic events in coeval basins. J. South Amer. Earth Modesto, S.P., Botha-Brink, J., 2010. Problems of correlation of South African and Sci. 63, 48e69. South American tetrapod faunas across the PermianeTriassic boundary. J. Afr. Schoch, R.R., 2013. The evolution of major temnospondyl clades: an inclusive Earth Sci. 57, 242e248. phylogenetic analysis. J. Syst. Palaeontol. 1, 1e33. Modesto, Sean P., Scott, Diane M., Botha-Brink, Jennifer, Reisz, Robert R., 2010. A Schoch, R.R., Milner, A.R., 2000. Stereospondyli. Handbuch der Palaoherpetologie,€ new and unusual procolophonid parareptile from the Lower Triassic Katberg Part 3B. München: Verlag Dr. Friedrich Pfeil, p. 203. Formation of South Africa. J. Vertebr. Paleontol. 30 (3), 715e723. http:// Schultz, C.L., Dias-da-Silva, S., 1999. A possible new pareiasaurid in the Sanga do dx.doi.org/10.1080/02724631003758003. Cabral Formation, lower Triassic (?) of Southern Brazil. Paleontol. em Destaque Modesto, S.P., Sues, H.-D., 2004. The skull of the Early Triassic archosauromorph 26, 49. reptile Prolacerta broomi and its phylogenetic significance. Zool. J. Linn. Soc. 140, Schwanke, C., Kellner, A.W.A., 1999. Sobre o primeiro registro de Synapsida no 335e351. Triassico basal do Brasil. In: Congr. Bras. Paleont., 16, Crato, 1999. B. Res, p. 101. Modesto, S.P., Sues, H.D., Damiani, R., 2001. A new Triassic procolophonoid reptile Sennikov, A.G., 1988. Novyye rauizukhdy iz triasa Yevropeyskoy chasti SSSR. Pale- and its implications for procolophonoid survivorship during the Permo-Triassic ontol. Zhur. 124e128. extinction event. Proc. R. Soc. Lond. B 268, 2047e2052. Sennikov, A.G., 1997. An enigmatic reptile from the upper permian of the Volga river Modesto, S.P., Damiani, R., Neveling, J., Yates, A.M., 2003. A new Triassic owenettid basin. Paleontol. J. 31, 94e101. parareptile and the mother of mass extinctions. J. Vert. Paleontol. 23, 715e719. Shishkin, M.A., 1967. Plagiosaurs from the Triassic of the USSR. Paleontol. Zhur. Nesbitt, S.J., 2011. The early evolution of archosaurs: relationships and the origin of 86e92 [In Russian]. major clades. Bull. Amer. Mus. NatHist. 352, 1e292. Smith, R.M.H., 1995. Changing fluvial environments across the Permian-Triassic Nosotti, S., 2007. Tanystropheus longobardicus (Reptilia, Protorosauria), re- boundary in the Karoo Basin, South Africa and possible causes of tetrapod ex- interpretations of the anatomy based on new specimens from the Middle tinctions. Palaeogeo. Palaeoclim. Palaeoecol. 117, 81e104. Triassic of Besano (Lombardy, northern Italy). Mem. Soc. Ital. Sci. Nat. Mus. Civ. Smith, R., Botha, J., 2005. The recovery of terrestrial vertebrate diversity in the South Storia Nat. 35, 1e88. Milano. African Karoo Basin after the end-Permian extinction. Compt. Ren. Paleovol. 4, Ottone, E.G., Garcia, G.B., 1991. A lower Triassic miospore assemblage from the 555e568. Puesto Viejo Formation. Argent. Rev. Palaeobot. Palynol. 68, 217e232. Sookias, R.B., Butler, R.J., 2013. Euparkeriidae. In: Nesbitt, S.J., Desojo, J.B., Irmis, R.B. Ottone, E.G., Monti, M., Marsicano, C.A., de la Fuente, M.S., Naipauer, M., (Eds.), Anatomy, Phylogeny and Palaeobiology of Early Archosaurs and Their Armstrong, R., Mancuso, A.C., 2014. A new Late Triassic age for the Puesto Viejo Kin, vol. 379. Geol. Soc. Spec. Publ, London, pp. 35e48. Group (San Rafael depocenter, Argentina), SHRIMP UePb zircon dating and Spalletti, L.A., Fanning, C.M., Rapela, C.W., 2008. Dating the Triassic continental rift biostratigraphic correlations across southern Gondwana. J. South Amer. Earth. in the southern Andes, the Potrerillos Formation, Cuyo Basin, Argentina. Geol. Sci. 56, 186e199. Acta 6. http://dx.doi.org/10.1344/105.000000256. Pacheco, C.P., Eltink, E., Müller, R.T., Dias-da-Silva, S., 2016. A new Permian tem- Stipanicic, P.N., Gonzalez Díaz, L.I., Zavattieri, A.M., 2007. Grupo Puesto Viejo nom. nospondyl with Russian affinities from South America, the new family Konz- transl. por Formacion Puesto Viejo ‘Gonzalez Díaz, 1964, 1967’, nuevas inter- hukoviidae, and the phylogenetic status of Archegosauroidea. J. Syst. Palaeontol. pretaciones paleontologicas, estratigraficas y cronologicas. Ameghiniana 44, 15, 241e256. 759e761. Peyer, B., 1937. Die Triasfauna der Tessiner Kalkalpen XII. Macrocnemus bassanii Sues, H.-D., 2016. Dating the origin of dinosaurs. Proc. Nat. Acad. Sci. U.S.A. 113, Nopcsa. Schweiz. Paleaontol. Abhand. 59, 1e140. 480e481. Pineiro,~ G., 2004. Faunas del Permico y Permo-Triasico de Uruguay, Bioestratigrafía, Sues, H.-D., Reisz, R.R., 2008. Anatomy and phylogenetic relationships of Scle- Paleobiogeografía y Sistematica [Master Dissertation]. Universidad de la rosaurus armatus (Amniota, Parareptilia) from the Bundsandsteind (Triassic) of República, Montevideo, p. 215. Europe. J. Vert. Paleontol. 28, 1031e1042. Pineiro,~ G., Ubilla, M., 2003. Unidades Permo-Triasicas en la Cuenca Norte, pale- Sumida, S.S., Modesto, S.P., 2000. A phylogenetic perspective on locomotory stra- ontología y ambientes. In: Veroslavsky, G., Ubilla, M., Martínez, S. (Eds.), tegies in early amniotes. Am. Zool. 41, 586e597. Cuencas sedimentarias de Uruguay, geología, paleontología y recursos naturales Suteethorn, V., Janvier, P., Morales, M., 1988. Evidence for a plagiosauroid amphibian e Mesozoico. DI.R.A.C./FCien, Uruguay, pp. 33e49. in the upper Triassic Huai Hin Lat Formation of Thailand. J. Southeast Asian Pineiro,~ G., Verde, M., Ubilla, M., Ferigolo, J., 2003. First basal synapsids (“pelyco- Earth Sci. 2, 185e187. saurs”) from South America, late Permiane?Early Triassic of Uruguay. Swofford, D., 2004. PAUP, Phylogenetic Analysis Using Parsimony (And Other J. Paleontol. 77, 389e392. Methods). Sinauer Associates, Sunderland, MA, version 4. . Pineiro,~ G., Rojas, A., Ubilla, M., 2004. A new procolophonoid (Reptilia, Parareptilia) Tatarinov, L.P., 1960. Discovery of pseudosuchians in the upper permian of SSSR. from the upper permian of Uruguay. J. Vert. Paleontol. 24, 814e821. Paleontol. J. 74e80 [In Russian]. Pineiro,~ G., Marsicano, C., Goso, C., Morosi, E., 2007a. Temnospondyl diversity of the Veevers, J.J., Cole, D.I., Cowan, E.J., 1994. South. Afr. Karoo Basin Cope Fold Belt. Geol. Permian-Triassic Colonia^ Orozco local fauna (Buena Vista formation) of Soc. Am. 184, 223e279. Uruguay. Rev. Bras. Paleontol. 10, 169e180. von Huene, F., Sahni, M.R., 1958. On Indobrachyops panchetensis gen. et sp. nov. from Pineiro,~ G., Marsicano, C., Lorenzo, N., 2007b. A new temnospondyl from the Permo- the upper Panchets (Lower Trias) of the Raniganj coalfield. Mono. Palaeontol. Triassic Buena Vista formation of Uruguay. Palaeontology 50, 627e640. Soc. India 2, 1e14. Pineiro,~ G., Marsicano, C., Damiani, R., 2007c. Mastodonsaurid temnospondyls from Ward, P.D., Montgomery, D.R., Smith, R., 2000. Altered river morphology in South the upper Permianelower Triassic of Uruguay, the earliest record from South Africa related to the Permian-Triassic Extinction. Science 289, 1740e1743. America. Acta Palaeontol. Pol. 52, 695e703. Warren, A.A., Black, T., 1985. A new rhytidosteid (Amphibia, Labyrinthodontia) from Pineiro,~ G., Ramos, A., Marsicano, C., 2012. A rhinesuchid-like temnospondyl from the early Triassic Arcadia Formation of Queensland, Australia, and the re- the Permo-Triassic of Uruguay. Compt. Rend. Palevol. 11, 65e78. lationships of Triassic temnospondyls. J. Vert. Paleontol. 5, 303e327. Pineiro,~ G., Ferigolo, J., Ribeiro, A.M., Velozo, P., 2015. Reassessing the affinities of Welles, S.P., Estes, R., 1969. Hadrokkosaurus Bradyi from the Upper Moenkopi For- vertebral remains from Permo-Triassic beds of Gondwana. Compt. Rend. Pale- mation of Arizona with a Review of the Brachyopid Labyrinthodonts, vol. 84. vol. 14, 387e401. Univ. Calif. Publ. Geol. Sci., pp. 1e61 Pinheiro, F.L., França, M.A.G., Lacerda, M.B., Butler, R.J., Schultz, C.L., 2016. An Wild, R., 1973. Die Triasfauna der Tessiner Kalkalpen. XXIII. Tanystropheus lon- exceptional fossil skull from South America and the origins of the archosauri- gobardicus Bassani (Neue Ergebnisse). Schweiz. Palaontol.€ Abhand. 95, 1e162. form radiation. Sci. Rep. 6, 22817. Wildner, W., Ramgrab, G.E., Lopes, R.C., Iglesias, C.M.F., 2008. Mapa Geologico do Reisz, R.R., Laurin, M., 2004. A reevaluation of the enigmatic Permian synapsid Estado do Rio Grande do Sul, Escala 1,750.000. CPRM, Porto Alegre. Watongia and of its stratigraphic significance. Can. J. Earth Sci. 41, 377e386. Zerfass, H., Lavina, E.L., Schultz, C.L., Garcia, A.J.V., Faccini, U.F., Chemale Jr., F., 2003. Retallack, G.J., Smith, R.M.H., Ward, P.D., 2003. Vertebrate extinction across Sequence stratigraphy of continental Triassic strata of Southernmost Brazil, A Permian-Triassic boundary in Karoo Basin, South Africa. GSA Bull. 115, contribution to southwestern Gondwana palaeogeography and palaeoclimate. 1133e1152. Sediment. Geol. 161, 85e105. Rocha-Campos, A.C., Basei, M.A., Nutman, A.P., Kleiman, L.E., Varela, R., Llambias, E., Zerfass, H., Chemale Jr., F., Schultz, C.L., Lavina, E.L., 2004. Tectonics and sedimen- Canile, F.M., da Rosa, O., 2011. 30 million years of Permian volcanism recorded in tation in Southern South America during Triassic. Sediment. Geol. 166, the Choiyoi igneous province (W Argentina) and their source for younger ash 265e292. fall deposits in the Parana Basin: SHRIMP UePb zircon geochronology evidence. Zerfass, H., Chemale Jr., F., Lavina, E.L., 2005. Tectonic controls of the Triassic Santa Gond. Res. 19, 509e523. Maria supersequence of the Parana Basin, southernmost Brazil, and its corre- Romer, A.S., 1922. The locomotor apparatus of certain primitive and mammal-like lation to the Waterberg basin, Namibia. Gond. Res. 8, 163e176. reptiles. Bull. Amer. Mus. Nat. Hist. 46, 517e606.