Redescription of the holotype of the Miocene

Download PDF
11 downloads 0 Views 4MB Size Report
Comment
Oct 1, 2018 - Colección de Paleontología, Santa Cruz de la Sierra, Bolivia. Anatomical abbreviations ...... património natural de la humanidad. Caracas: ...
Historical Biology An International Journal of Paleobiology

ISSN: 0891-2963 (Print) 1029-2381 (Online) Journal homepage: http://www.tandfonline.com/loi/ghbi20

Redescription of the holotype of the Miocene crocodylian Mourasuchus arendsi (Alligatoroidea, Caimaninae) and perspectives on the taxonomy of the species Giovanne M. Cidade, Andrés Solórzano, Ascánio Daniel Rincón, Douglas Riff & Annie Schmaltz Hsiou To cite this article: Giovanne M. Cidade, Andrés Solórzano, Ascánio Daniel Rincón, Douglas Riff & Annie Schmaltz Hsiou (2018): Redescription of the holotype of the Miocene crocodylian Mourasuchus�arendsi (Alligatoroidea, Caimaninae) and perspectives on the taxonomy of the species, Historical Biology, DOI: 10.1080/08912963.2018.1528246 To link to this article: https://doi.org/10.1080/08912963.2018.1528246

Published online: 01 Oct 2018.

Submit your article to this journal

View Crossmark data

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ghbi20

HISTORICAL BIOLOGY https://doi.org/10.1080/08912963.2018.1528246

Redescription of the holotype of the Miocene crocodylian Mourasuchus arendsi (Alligatoroidea, Caimaninae) and perspectives on the taxonomy of the species Giovanne M. Cidade

a

, Andrés Solórzano

b,c

, Ascánio Daniel Rincón

b

, Douglas Riffd and Annie Schmaltz Hsioua

a Departamento de Biologia, FFCLRP, Universidade de São Paulo, Ribeirão Preto, São Paulo, Brazil; bLaboratorio de Paleontología, Centro de Ecología, Instituto Venezolano de Investigaciones Científicas, Caracas, Venezuela; cPrograma de Doctorado en Ciencias Geológicas, Facultad de Ciencias Químicas, Universidad de Concepción; Barrio Universitario s/n, Concepción, Chile; dInstituto de Biologia, Universidade Federal de Uberlândia, Uberlândia, Minas Gerais, Brazil

ABSTRACT

ARTICLE HISTORY

The Miocene crocodyliform fauna of South America is one of the most diverse of the world, and the late Miocene Urumaco Formation of Venezuela has one of its most important assemblages. Mourasuchus (Caimaninae) is one of the most peculiar crocodyliforms of the South American Miocene due to its unusual morphology, which prompted peculiar feeding habits to be proposed for this taxon. In this paper we present a redescription of the holotype of the species Mourasuchus arendsi (CIAAP-1297) from the Urumaco Formation of Venezuela. The redescription offered a thorough reassessment of the skull, mandibles and postcranium that comprise the holotype of M. arendsi, providing a comprehensive morphological description of this specimen for the first time. The data provided by this description prompted a review of the taxonomic status of M. arendsi, which has enabled the possibility of M. arendsi being a junior synonym of M. atopus to be considered and thoroughly discussed in this paper. An eventual confirmation of the synonymy does not change the phylogeny of the Caimaninae clade. This contribution also offers assessments on the ontogenetic status of the holotype of M. arendsi and on the differences on the closure of the scapulocoracoid synchondroses between Mourasuchus specimens.

Received 6 August 2018 Accepted 21 September 2018

Introduction The crocodyliform record of the Cenozoic is dominated by eusuchians, and that of the South American Cenozoic in particular by alligatoroid caimanines (Brochu 1999, 2011; Riff et al. 2010; Bona et al. 2013a; Souza et al. 2016; Cidade et al. 2017). The Miocene is the epoch in which the crocodylian and caimanine fossil record reaches its apex in South America (see Riff et al. 2010). The diversity described until now for the Miocene includes 27 fossil crocodylian species, of which 16 are caimanines. One of the most prolific geological units for crocodylian species diversity is the late Miocene Urumaco Formation of Venezuela, which also yields a rich vertebrate fossil record of fishes, turtles and mammals, among other groups, including other fossil crocodylians such as the caimanines Caiman, Melanosuchus and Purussaurus and the gavialoids Gryposuchus, Hesperogavialis and Ikanogavialis (see Sill 1970; Bocquentin-Villanueva and Buffetaut 1981; Linares 2004; Aguilera et al. 2006; Riff and Aguilera 2008; Scheyer and MorenoBernal 2010; Scheyer et al. 2013; Scheyer and Delfino 2016; Bona et al. 2017; Foth et al. 2017). The presence of the tomistomine crocodyloids Thecachampsa has also been proposed (Aguilera, 2004), but other authors consider the materials assigned to it as belonging to gavialoids (Scheyer and Moreno-Bernal 2010). Mourasuchus (Price 1964) is one of the most characteristic forms of the South American crocodyliform fauna, mainly due to its peculiar morphology of a long, wide, dorsoventrally flattened skull (Price 1964; Langston 1965; BocquentinVillanueva 1984; Gasparini 1985; Cidade et al. 2017). CONTACT Giovanne M. Cidade [email protected] 14040-901, Ribeirão Preto, São Paulo, Brazil © 2018 Informa UK Limited, trading as Taylor & Francis Group

Published online 01 Oct 2018

KEYWORDS

Mourasuchus; Crocodylia; Alligatoroidea; Caimaninae; Miocene; South America

Uncommon feeding habits have been proposed for the group due to that morphology, such as filter-feeding (see Riff et al. 2010), or gulp-feeding (Cidade et al. 2017). Four species have been assigned to Mourasuchus: M. amazonensis (Price 1964); M. atopus (Langston 1965); M. arendsi (Bocquentin-Villanueva 1984) and M. pattersoni Cidade et al. (2017). The holotypes of the last two species are from the Urumaco Formation, from which there are also several specimens assigned as ‘Mourasuchus’ sp. (Scheyer et al. 2013; Scheyer and Delfino 2016). This fully characterizes Mourasuchus as an important component of the crocodylian fauna of the unit. The holotype of M. arendsi is from the Upper Member of the Urumaco Formation (Bocquentin-Villanueva 1984; Figure 1). The holotype of M. pattersoni was also probably collected in the Upper Member, although there are uncertainties about its provenance (see Langston 2008; Cidade et al. 2017). The holotype of Mourasuchus arendsi (CIAAP-1297) was only briefly described by Bocquentin-Villanueva (1984). Additionally, the original description was illustrated only with drawings, and not with photographs. Scheyer and Delfino (2016) also studied the specimen but did not described it in detail. In this work, the holotype of M. arendsi is redescribed, and the taxonomic implications of the specimen reanalysis are discussed. Additionally, the ontogenetic status of the specimen and the closure of the scapula-coracoid synchondrosis in fossil caimanines are discussed.

Departamento de Biologia, FFCLRP, Universidade de São Paulo; Bandeirantes Avenue, 3900,

2

G. M. CIDADE ET AL.

Figure 1. Map of the Venezuelan state of Falcón showing the ‘Corralito’ locality where the holotype was found according to Bocquentin-Villanueva (1984). Adapted from Cáceres et al. (2016) and Cidade et al. (2017).

Institutional abbreviations AMNH, American Museum of Natural History, New York, United States; CIAAP, Centro de Investigaciones Antropológicas, Arqueológicas y Paleontológicas, Universidad Nacional Experimental Francisco de Miranda, Coro, Venezuela; DGM, Divisão de Geologia e Mineralogia, Museu de Ciências da Terra, Rio de Janeiro, Brazil; MACN, Museo Argentino de Ciéncias Naturales, Buenos Aires, Argentina; MCNC-PAL, Museo de Ciéncias Naturales de Caracas, Caracas, Venezuela; MCZ, Museum of Comparative Zoology, Cambridge, United States; TMM, Texas Memorial Museum, Austin, United States; UCMP, University of California Museum of Paleontology, Berkeley, United States; UFAC, Universidade Fedeeral do Acre, Rio Branco, Brazil; YPFB-LIT-PAL, Yacimientos Petrolíferos Fiscales Bolivia, Centro de Tecnología Petrolera – Litoteca – Colección de Paleontología, Santa Cruz de la Sierra, Bolivia. Anatomical abbreviations amk = anteromedial knob of the orbit; ang = angular; ar = articular; at = atlas; ati = atlas intercentrum; atr = atlantal rib; ax = axis; axr = axial rib; bo = basioccipital; cr = coronoid; crc = coracoid; cv = cervical vertebra; d = dentary; dpf = diapophysis; emf = external mandibular fenestra; en = external naris; ex = exoccipital; f = frontal; fic = foramen intermandibularis caudalis; hpf = hypapophysis; ift = infratemporal fenestra; j = jugal; l = lacrimal; m = maxilla; n = nasal; ns = neural spine; o = orbit; oc = occipital condyle; of = occlusal pits; p = parietal; pf = prefrontal; pm = premaxilla; po = postorbital; pt = pterygoid; q = quadrate; qj = quadratojugal; san = surangular; sc = scapula; so = supraoccipital; sp = splenial; sq = squamosal; stf = supratemporal fenestra; str = subtemporal ramus of the jugal.

Material and methods The holotype of Mourasuchus arendsi (CIAAP-1297) includes an almost complete skull (Figures 2, 3 and 4), an almost complete right mandibular ramus (Figure. 5 and 6), a posterior fragment of the left mandibular ramus (Figure 7) and several postcranial remains. The original description of Mourasuchus arendsi (Bocquentin-Villanueva 1984) mentioned that the

holotype has vertebrae, namely the articulated six first cervical (including atlas and axis), two isolated cervical vertebrae and six ‘very damaged’ dorsal vertebrae. However, only the articulated six first cervical vertebrae (Figure 8), two putative isolated cervical vertebrae (Figure 9) and three putative isolated dorsal vertebrae (Figure 10) could be found. The articulated left scapula and coracoid (Figure 11) were not described or mentioned in the original description (Bocquentin-Villanueva 1984), but these elements are registered with the same catalogue number of the holotype (CIAAP-1297). Additionally, Scheyer et al. (2015, fig. 34) picture an isolated scapula as belonging to the holotype. The coloration and preservation status of the scapula are similar to the holotype and thus may belong to it, but this cannot be assured as we did not analyse it directly. Thus, this scapula is not described as belonging to the holotype in this paper, but future assessments are required to settle this issue. The skull of CIAAP-1297 is currently only visible in dorsal view (Figure 2), and is firmly attached to a wooden support facing the ventral portion. Any attempts to remove the support in order to expose the ventral portion of the skull could result in serious damage of the specimen, which is very fragile. Because of this, no attempts were made to expose the skull in ventral view. The morphological data matrix used in the phylogenetic analysis was scored in the software Mesquite, version 2.75 (Maddison and Maddison 2011). The analysis was performed based on the matrix of Souza-Filho et al. (in press), in which most taxa, characters and scorings are based in Brochu (2011; see Supplemental Online Material for details), with 94 operational taxonomic units (OTUs), with 93 eusuchian taxa in the ingroup and the non-eusuchian crocodyliform Bernissartia fagesii as outgroup and 187 characters. The taxa involved in the analysis include a new caimanine species (the specimen UFAC-2507) that is being described by Souza-Filho et al. (in press), while the other taxa had already been included in previous analyses (Bona 2007; Brochu 2011; Bona et al. 2013a; Hastings et al. 2013; Pinheiro et al. 2013; Scheyer et al. 2013; Fortier et al. 2014; Salas-Gismondi et al. 2015; Cidade et al. 2017). The only change performed was the combination of the scorings of Mourasuchus atopus and M. arendsi for the exploratory phylogenetic analyses, which reduced the total of 94 taxa used by Souza-Filho et al. (in press) to the 93 used in the exploratory analysis. As in Souza-Filho

HISTORICAL BIOLOGY

3

Figure 2. Skull of the holotype of Mourasuchus arendsi (CIAAP-1297) in dorsal view with schematic drawing. en = external naris; f = frontal; ift = infratemporal fenestra; j = jugal; l = lacrimal; m = maxilla; n = nasal; o = orbit; p = parietal; pf = prefrontal; pm = premaxilla; po = postorbital; q = quadrate; qj = quadratojugal; so = supraoccipital; sq = squamosal; stf = supratemporal fenestra. Scale = 10 cm.

Figure 3. Right premaxilla of the holotype of Mourasuchus atopus (UCMP-38,012, A) and premaxillae of the holotype of M. arendsi (CIAAP-1297, B) in dorsal view. Adapted from Cidade et al. (2017; Figure 5). en = external naris; of = occlusal pits; pm = premaxilla. Scale = 5 cm.

Figure 4. Occipital area of the skull of the holotype of Mourasuchus arendsi (CIAAP-1297) with schematic drawing. Adapted from Scheyer & Delfino (2016, fig. 16). bo = basioccipital; ex = exoccipital; oc = occipital condyle; pt = pterygoid; q = quadrate; so = supraoccipital; sq = squamosal. Scale = 10 cm.

et al. (in press), we did not include Orthogenysuchus olseni Mook, 1924, from the Eocene of the United States, in the analysis. This species was recovered as sister-taxon to Mourasuchus by many previous analyses (e.g. Brochu 1999; Aguilera et al. 2006; Bona 2007; Scheyer et al. 2013; Bona et al. 2013a; Fortier et al. 2014), but according to Salas-Gismondi et al. (2015) ongoing preparation of

the holotype and only known species of O. olseni prompted many changes in the scoring of the characters for this taxon that have not been published yet. The complete list of taxa and characters used in the analyses, as well as the complete matrix of scored characters by taxon, is available at the Supplemental Online Material.

4

G. M. CIDADE ET AL.

Figure 5. Right mandibular ramus of the holotype of Mourasuchus arendsi (CIAAP-1297) in medial (A) and lateral (B) views. Scales = 5 cm.

Figure 6. Posterior portion of the right mandibular ramus of the holotype of Mourasuchus arendsi (CIAAP-1297) in medial (A and B), lateral (C) and dorsal (D) views. ang = angular; ar = articular; cr = coronoid; d = dentary; emf = external mandibular fenestra; fic = foramen intermandibularis caudalis; san = surangular; sp = splenial. Scales = 10 cm.

The analysis was performed in the software Tree Analysis using New Technology (TNT – Goloboff et al. 2008), with 10,000 replications, a random seed value of ‘0’ and 20 cladograms saved per replication. The branch swapping algorithm selected was ‘tree-bisection-reconnection’. The characters were unordered and non-additive.

Geological background The Urumaco Formation was deposited from the latest part of the middle Miocene to the end of the late Miocene (Linares 2004). The approximate thickness of this unit is 2000 m. It shows a diverse lithology dominated by sandstone, claystone, siltstone and limestone. Differences in the proportion of these lithologies allows to the recognition of Lower, Middle and Upper Members (Linares 2004). The Lower Member is a sequence of lutites and limonites interleaved with limestone levels. Its sediments are interpreted as coming from large coastal

lagoons of variable salinity, swampy areas and storm-carried sediments. The Middle Member is comprised of a non-cyclical sequence of lutites, sandstones and sandy limestones. Its sediments are interpreted as near-coastal deposits shaped by coastal bars, inter-tributary channels and meander bars that progradate. The Upper Member, in which the holotype of Mourasuchus arendsi was collected, is a non-cyclical sequence of claystones and sandstones without calcareous elements. Its sediments are interpreted as river channel, foodplain and high fronts deposits with occasional marine influence. (see Linares 2004). The dominant paleoenvironment during the sedimentation of the Urumaco Formation is still not clear. One hypothesis is that the sedimentation of the Formation occurring in a complex of marginal and near coastal environments (Díaz de Gamero and Linares 1989; Hambalek et al. 1994). Another hypothesis suggests that the Formation was probably deposited in a prograding strandplain-deltaic complex (Quiroz and Jaramillo 2010).

HISTORICAL BIOLOGY

5

Figure 7. The fragment of left mandibular ramus of the holotype of Mourasuchus arendsi (CIAAP-1297) in dorsal (A), medial (B) and lateral (C) views. ang = angular; d = dentary; emf = external mandibular fenestra; fic = foramen intermandibularis caudalis; san = surangular; sp = splenial. Scales = 5 cm.

Results Systematic palaeontology

CROCODYLIA (Gmelin 1789) (sensu Benton and Clark 1988) ALLIGATOROIDEA (Gray 1844) (sensu Norell et al. 1994) CAIMANINAE (Brochu 1999) (sensu Brochu 2003a; Following Norell 1988) Mourasuchus (Price 1964) Mourasuchus arendsi (Bocquentin-Villanueva 1984) Figures: 2–11. Holotype CIAAP-1297 (almost complete skull, almost complete right mandibular ramus, posterior fragment of the left mandibular ramus, six first cervical vertebrae, two putative isolated cervical vertebrae, three putative isolated dorsal vertebrae, left scapula and coracoid).

Locality and horizon of the holotype. Upper Member of the Urumaco Formation, late Miocene, 70º 18ʹ W and 11º 18ʹ S. Approximately 1 km to the southeast of the hamlet Corralito, Urumaco municipality, Democracía Department, Falcón State, Venezuela (from Bocquentin-Villanueva 1984); the same locality known as ‘Corralito’ in other works (e.g. Scheyer et al. 2013; Scheyer and Delfino 2016) Referred specimens. CIAAP-1333, a fragment of a right maxilla, from the type locality (Bocquentin-Villanueva 1984); UFAC-5716, an almost complete rostrum lacking the orbital area and the skull table; UFAC-5883, a posterior portion of the skull lacking most of the rostrum; UFAC1818, an incomplete right premaxilla; UFAC-1799, two almost complete, articulated premaxillae, all from the late Miocene Solimões Formation of Brazil (Souza-Filho and Guilherme 2011).

6

G. M. CIDADE ET AL.

Figure 8. The six first cervical vertebrae of the holotype of Mourasuchus arendsi (CIAAP-1297) in right (A) and left (B) lateral views. at = atlas; ati = atlas intercentrum; atr = atlantal rib; ax = axis; axr = axial rib; cv = cervical vertebra; dpf = diapophysis; hpf = hypapophysis; ns = neural spine. Scales = 10 cm.

Figure 9. Two putative cervical vertebrae of the holotype of Mourasuchus arendsi (CIAAP-1297), both in right lateral view. Scales = 5 cm.

HISTORICAL BIOLOGY

7

Figure 10. Three putative dorsal vertebrae (A, B and C) of the holotype of Mourasuchus arendsi (CIAAP-1297): (A) is in posterior view, (B) in right lateral view and (C) in anterior view. Scales = 5 cm.

Figure 11. Scapulocoracoids of the holotypes of Mourasuchus atopus (UCMP-38,012, A), M. arendsi (CIAAP-1297, B) and M. pattersoni (MCNC-PAL-110-72V, C). The black squares highlight the beginning of the closure of the scapulocoracoid synchondrosis in (B) and a possible beginning of the closure of the same structure in (C). crc = coracoid; sc = scapula. Scales = 5 cm (A and B), 1 cm (C).

Emended diagnosis incisive foramen enlarged anteriorly and narrow posteriorly (autapomorphy); differs from M. amazonensis and M. pattersoni from having a circular external naris and a lateromedially slender and dorsoventrally low jugal; differs from M. atopus and M. pattersoni from having lateromedially expanded palatines. Emended from Bocquentin-Villanueva (1984). Redescription of CIAAP-1297 Premaxillae The right premaxilla is complete in CIAAP-1297. The left premaxilla is mostly reconstructed with epoxy resin and only the posteriormost portion of it is preserved (Figures 2 and 3). The right premaxilla exhibits four occlusal pits for the first four

mandibular teeth (Figures 2 and 3). This number differs from the amount observed in the other holotypes of Mourasuchus: M. amazonensis (DGM 526-R, one), M. atopus (UCMP-38,012, two) and M. pattersoni (MACN-PAL-110-72V, three) (Cidade et al. 2017). Such difference is not considered taxonomically or systematically relevant as the number of occlusal pits varies individually and ontogenetically in living crocodylians (Kälin 1933; Langston 2008; Souza-Filho and Guilherme 2011). The external naris projects dorsally as in most eusuchians (Brochu 1997, 2011) and is small and roughly circular (Figure 3), differing from the lateromedially enlarged external naris present in M. pattersoni (Cidade et al. 2017) and drawn for M. amazonensis by Price (1964). Even though the external naris of the holotype of M. atopus cannot be fully assessed as only the right premaxilla is preserved, the shape of this element indicates an external naris

8

G. M. CIDADE ET AL.

with a morphology more like that observed in M. arendsi than that of M. pattersoni and M. amazonensis (Cidade et al. 2017). The incisive foramen of the holotype of Mourasuchus arendsi was drawn by Bocquentin-Villanueva (1984) as being enlarged anteriorly and narrow posteriorly, in the shape of a ‘reversed teardrop’. This morphology differs from the incisive foramen of the other Mourasuchus species that preserve the structure, such as the small, circular foramen seen in M. pattersoni and the large, tri-lobed foramen of M. amazonensis (see Cidade et al. 2017; Figure 7). However, the morphology of the incisive foramen of the holotype of Mourasuchus arendsi could not be reassessed due to the non-exposure of the ventral side of the skull, and also because of the fact that the incisive foramen cannot be fully seen in dorsal view (Figures 2 and 3). The right premaxilla of Mourasuchus arendsi displays a very developed narial rim lateral to the external naris (Figure 3), a structure probably also present in M. atopus (although the incompleteness of the premaxilla precludes a definitive statement) but absent in M. pattersoni and appearently in M. amazonensis (see Price 1964; Cidade et al. 2017). Whether this developed narial rim of M. arendsi is homologous to the lateral border of the external naris in M. pattersoni or represents a different structure altogether is discussed by Cidade et al. (2017). Aside from the developed narial rim, there are no signs of a thin crest around the external naris as observed in the caimanine Tsoabichi (Brochu 2010) or of a deep notch lateral to the external naris, as in many aligatorines (see Brochu 2011) and the caimanine Gnatusuchus (SalasGismondi et al. 2015). Each suture between the premaxillae and the maxillae are visible, but the sutures between the premaxillae and the nasals are not visible for the most part (Figure 2). The dorsal premaxillary process is long (sensu Brochu 2011, Character 90–1) as in all Mourasuchus and in Caiman brevirostris (see Fortier et al. 2014; Cidade et al. 2017). There are no signs of a notch for the occlusion of the fourth dentary tooth, which makes it more likely that it occluded in a pit, as in all alligatoroids except Leidyosuchus (see Brochu 2011). Maxillae The maxillae exhibit the same morphology as in other Mourasuchus specimens (see Price 1964; Langston 1965; Cidade et al. 2017), as it is anteroposteriorly and lateromedially expanded (Figure 2) in comparison with other crocodylians. The surface of the maxillae are flat, without developed structures present in other crocodylians such as pronounced bosses, preorbital ridges or canthi rostralii. Part of the posterior portion of the left maxilla is incomplete, and missing fragments have been reconstructed with a resin (Figure 2). The sutures of the maxillae with the nasals medially are not fully visible except for the posteriormost portion of the suture between the right maxilla and the right nasal (Figure 2). The sutures between both maxillae with the prefrontals, lacrimals and jugals, posteriorly, are not fully visible but can be distinguished as in Figure 2. Nasals, prefrontals and lacrimails The nasals also exhibit the same anteroposteriorly elongated morphology typical of Mourasuchus, even though only some parts of the suture between the two bones are visible (Figure 2). The nasals contact the premaxillae anteriorly, but

not the external naris (Figure 2), a morphology that is also illustrated for M. amazonensis (see Price 1964). The right prefrontal and lacrimal are completely preserved. In the left side the area between these two bones is significantly eroded, with the anterior portion of the erosion being replaced by a resin (Figure 2). The posterior portion of the left lacrimal is also eroded. Neither sutures between the prefrontals and the lacrimals, and between the prefrontals and the frontal are fully visible (Figure 2). However, the prefrontals meet medially, in a visible suture, preventing the nasals from contacting the frontal posteriorly, as in all Mourasuchus species and in many caimanine taxa (see Brochu 2011; Cidade et al. 2017; Fernández-Blanco et al. 2018). The sutures of the lacrimals with the jugals are also not visible. Jugals, quadratojugals and quadrates The right jugal is completely preserved, while the left jugal is severely eroded, especially in its limit with the lacrimal, in the area adjacent to the anterolateral border of the orbit and in the medial portion of the subtemporal ramus (Figure 2). The subtemporal rami of the jugals in M. arendsi are lateromedially slender and dorsoventrally low, similar to those of M. atopus. In M. amazonensis, this structure differs from that of M. arendsi in being lateromedially wide and dorsoventrally flattened, whereas in M. pattersoni the subtemporal rami are lateromedially wide and dorsoventrally high (see Cidade et al. 2017). The suture between the right jugal and the right quadratojugal is visible, but that between the respective left bones is not (Figure 2). Both quadratojugals are almost completely preserved, showing significant processes along the lower temporal bar and along the posterior margin of the infratemporal fenestrae. However, it is not possible to observe whether the quadratojugal reached the dorsal extremity of the infratemporal fenestrae in the posterior margins of these fenestrae, as the respective areas in both sides of the skull are eroded (Figure 2). Both quadrates are present, but most of their dorsal surfaces are eroded (Figure 4). Frontal, postorbitals and parietal The frontal is almost completely preserved. The only absent parts are the lateral parts of the posterior portion of the bone, which are adjacent to the medial borders of the orbits (Figure 2). The suture of the frontal with the parietal is not fully visible (Figure 2). The anteromedial border of the orbits exhibits a prominent knob, as in all Mourasuchus species (see Cidade et al. 2017), although the posteriormost portions of the knobs are not preserved due to the non-preservation of the frontal in the respective areas. M. arendsi does not display a preorbital crest, as in most crocodylians except Purussaurus mirandai, P. neivensis, Caiman and Melanosuchus (see Langston 1965; Aguilera et al. 2006; Barrios 2011). The right postorbital is completely preserved, while the dorsal suface of the left postorbital is not (Figure 2). The right postorbital forms the anterolateral margin of the right supratemporal fenestra, which is the only one of the supratemporal fenestrae that preserves its dorsal surface in M. arendsi (Figure 2). The fenestra is overhung by the dermal bones, as in most caimanines except Culebrasuchus, and reduced, as in most caimanines except Culebrasuchus and Purussaurus (see Hastings

HISTORICAL BIOLOGY

et al. 2013; Cidade et al. 2017), and is not obliterated as in most individuals of Paleosuchus (see Brochu 1997). The parietal is completely preserved, exhibiting an elevated median crest as seen in all Mourasuchus specimens that preserve this bone (see Gasparini 1985; Cidade et al. 2013; Bona et al. 2013b; Scheyer and Delfino 2016). The parietal comprises the medial border of the supratemporal fenestrae. Squamosals and supraoccipital The right squamosal is almost completely preserved, while the left squamosal is severely eroded, especially in dorsal view (Figure 2). The squamosal forms the posterolateral border of the supratemporal fenestra, and in M. arendsi exhibits the same accentuated hypertrophy (‘horn’) throughout the entire dorsal surface of the bone as seen in all Mourasuchus specimens that preserve the squamosal (see Gasparini 1985; Cidade et al. 2013; Bona et al. 2013b; Scheyer and Delfino 2016). In dorsal view, the sutures of the right squamosal with the parietal and the postorbital are visible, while the suture with the supraoccipital is not. The sutures between the left squamosal with the parietal and the supraoccipital are also not distinguishable (Figure 2). In occipital view, the squamosals contact the supraoccipital medially and the exoccipitals ventrolaterally, but the limits are not discernible due to the poor preservation of the surface of the bones in the occipital portion of the skull, including cracks between the squamosals and the supraoccipital and between the right squamosal and the right exoccipital (Figure 4). The supraoccipital is only partially preserved. It is completely preserved in dorsal view, but in occipital view most of its surface is covered by a resin (Figure 4). The dorsal exposure of the supraoccipital is large to the point of excluding the parietal from the posterior margin of the skull table (Brochu 2011; Character 160–3; Figure 2), as in all caimanines except Tsoabichi, Paleosuchus, Purussaurus mirandai and P. neivensis (see Aguilera et al. 2006; Brochu 2011; Cidade et al. 2017). In occipital view, the supraoccipital contacts the exoccipitals ventrally, but the suture between the supraoccipital and the exoccipitals are not preserved in M. arendsi as the area is significantly eroded (Figure 4). Exoccipitals and basioccipital Due to the same erosions, both exoccipitals exhibit severe fragmentation in their most medial parts (Figure 4). The exoccipitals contact the quadrate laterally and the basioccipital ventrally. The limit between the right exoccipital and the right quadrate is visible, but that between the respective left bones is not. The limits between the exoccipitals and the basioccipital are not fully visible, but it can be distinguished that the exoccipitals send slender processes lateral to the basioccipital plate (Brochu 2011; Character 176–2; Figure 4), as in all caimanines except Culebrasuchus (see Hastings et al. 2013; Cidade et al. 2017; it is worthwhile to notice, however, that Culebrasuchus is recovered as an alligatorine by SalasGismondi et al. 2015; Bona et al. 2018; and at least outside caimanine by the strict consensus of Hastings et al. 2016). The basioccipital is completely preserved, including the occipital condyle and the basioccipital plate (Figure 4), which bears a median crest that serves as attachment point to the muscles M. basioccipitovertebralis and M. occipitotransversalis profundus in living Crocodylia (Iordansky 1973).

9

Braincase Despite the impossibility of observing the skull in ventral view, some elements from the anterolateral and ventral areas of the braincase can be observed. However, the bones are eroded in several portions and very little of these areas can be described, aside from statements that at least part of the pterygoids, the laterosphenoids and possibly the basisphenoids are preserved. In each side of the anterolateral area of the braincase, there is a large foramen that probably represents the trigeminal foramen, whereby the mandibular and maxillary rami of the trigeminal nerve (= V cranial nerve) exit in living crocodylians (Holliday and Witmer 2009; George and Holliday 2013). Dentaries An almost complete right mandibular ramus and a medial fragment of the left mandibular ramus (Figures. 5, 6 and 7) belong to the holotype of Mourasuchus arendsi. As with the skull, most of their surfaces are covered by gypsum, hindering the observation of the limits between the bones. While the right mandibular ramus preserves all mandibular bones (Figures 5 and 6), the fragment of the left mandibular ramus preserves only the most distal portions of the dentary and the splenial, and the most proximal portions of the surangular and the angular (Figure 7). The right dentary does not preserve the mandibular symphysis. There are 33 alveoli preserved in the right dentary: eight in the anteriormost preserved portion and 25 in the posteriormost preserved portion. Between those, there is an area where the gypsum precludes the observation of alveoli. In the fragmented left dentary, 15 alveoli can be observed, although a space covered with gypsum in the posteriormost portion of the fragment could represent one last alveolus. Some of the alveoli of the right dentary and most of the left dentary preserve the most basal portion of the teeth, but no complete tooth is preserved. Despite the incompleteness of the mandibular rami, it is probable that the number of alveoli in Mourasuchus arendsi is not different from the more than 40 alveoli recorded for M. atopus (Langston 1965). The contact between the right dentary with the splenial is not fully visible for the most part, except for part of its posteriormost region, but the contact as a whole can be interpreted as in Figure 6. The contact between the left dentary and the left splenial is only visible at its posteriormost portion. The suture of the right dentary with the surangular is visible (Figure 6), but that of the left dentary with the surangular is not (Figure 7). The suture of the right dentary with the angular is fully visible only in part of its anteriormost portion, while the sutures between the corresponding left bones are not fully distinguishable, being only interpreted as in Figure 7. The dentary of M. arendsi is linear in the area between the fourth and the tenth alveoli, a character present only in Mourasuchus and Culebrasuchus in the Caimaninae clade (see Hastings et al. 2013; Bona et al. 2013a; Cidade et al. 2017). Splenial, coronoid and surangulars The right splenial is not completely preserved, and none of its posterior sutures with the surangular, the coronoid and the angular can be observed (Figure 6). The suture between the

10

G. M. CIDADE ET AL.

left splenial and the angular is not fully visible either (Figure 6). The right surangular is only partially preserved, while the left surangular preserves only its anteriormost portion. The sutures of the right surangular with the articular and with the angular in lateral view are visible for the most part (Figure 6-C), while the sutures with the coronoid and with the angular in dorsal view are not visible (Figure 6-D). Only the right coronoid is preserved, but is incomplete. The most dorsal area of the right coronoid is not preserved (Figure 6). The suture of the right coronoid with the angular is not visible. Angulars and articular The right angular is almost completely preserved, lacking only part of the portion ventral to the adductor fossa and to the area close to the posterior border of the external mandibular fenestra, both in medial view (Figure 6). The angular-articular suture is visible throughout most of its extension (Figure 6). The left angular preserves only its anteriormost portion. It forms the ventral border of the external mandibular fenestra and of the foramen intermandibularis caudalis (Figure 7). The right articular is almost completely preserved. The only absent parts are the anterior portion of the ventral process (Figure 6), which contacts the surangular and the angular, in dorsal view, in living Crocodylia (see Iordansky 1973), and the most lateral portion of the anterior concavity of the glenoid fossa (Figure 6). The remaining of the glenoid fossa, including the posterior concavity, and the retroarticular process are completely preserved. Vertebrae The atlas is not completely preserved. The atlas intercentrum is almost completely preserved, except mainly for its posterior extremities, where the parapophyseal processes, which articulate with the axial ribs in living crocodylians, would probably be located (see Brochu 1997). The left element of the neural arch of the atlas is almost complete, while the right element preserves only its posteriormost area, which articulates with the axis. The atlas of Mourasuchus arendsi is plate-shaped in lateral view, a common morphology among Alligatoridae (Brochu 1997). The right atlantal rib is almost completely preserved, except for most of its posterior region. The dorsal margin of the rib bears a prominent process (Brochu 2011; Character 6–1), as in most Alligatoroidea (see Brochu 1997; Cidade et al. 2017). The left atlantal rib is mostly absent. Only a fragment of bone located lateral to the atlas may represent the proximal extremity of the left atlantal rib (Figure 8), but this interpretation is tentative due to the poor preservation of the specimen. The axis is almost complete, except for erosions in the neural spine, in the ventral area of the centrum and in the hypapophysis. Anterior to the axis vertebral centrum, the odontoid process is preserved. The right axial rib is preserved only in its anteriormost area (Figure 8; see also Tineo et al. 2014, Figure 5). The right axial rib tuberculum contacts the diapophysis of the axis (Brochu 2011; Character 10–1), a common feature among Alligatoridae (Brochu 1997), and is narrow sensu Brochu (2011, Character 9–1) as in most Brevirostres crocodylians (see Brochu 2011; Cidade et al. 2017).

The only part of the proximal extremity of the left axial rib that may be preserved is a fragment lateral to the odontoid process and to the axial centrum (Figure 8). Dorsal to this fragment, there is an eroded area between the odontoid process and the axial centrum that may correspond to the area in which the tuberculum of the left axial rib would articulate with the axis (Figure 8). However, both these interpretations are only tentative. The axial hypapophysis is not completely preserved. It is located toward the anterior end of the centrum (Brochu 2011; Character 15–1), as in most eusuchians (see Brochu 2011; Cidade et al. 2017). The third to the sixth cervical vertebrae are almost complete except for minor erosions. All of them preserve their centra, pedicels, pre and post-zygapophyses and neural arches, including the neural spines. The neural spine of the third cervical vertebra, however, is significantly broken (Figure 8). All the other neural spines exhibit the shape of a wide vertical blade, as described for isolated cervical vertebrae of the Mourasuchus specimen YPFBLIT-PAL-01, from the late Miocene Yecua Formation of Bolivia, by Tineo et al. (2014). The diapophyses and parapophyses are preserved only on the left side of the vertebrae (Figure 8). On the right side, these structures are either absent or significantly eroded, while the area in which the right parapophysis of the third cervical vertebra would be situated is hidden behind the atlantal rib. The diapophyses are oval in transversal section and ventrolaterally oriented, similar to the isolated cervical vertebrae described by Tineo et al. (2014). All hypapophyses are partially preserved, but with various degrees of erosion: the hypapophysis of the third cervical vertebra is significantly eroded (Figure 8). Regardless of erosions, the cervical hypapophyses of the holotype of M. arendsi may be considered small in comparison with extant crocodylians (e.g. Paleosuchus palpebrosus, AMNH-R-97,326), which is similar to the condition described for M. pattersoni (Langston 2008). The condyle of the sixth cervical vertebra is completely preserved, while the condyles of the other vertebrae are firmly articulated with the cotyle of the immediate posterior vertebra and cannot be fully observed. The neurocentral sutures of the cervical vertebrae can be fully distinguished only on the right side of the fourth vertebra and on the left side of the third and fourth vertebrae. In all of them, the sutures are almost closed (sensu Brochu 1996), with some areas still seeming to be open. This indicates that the specimen was close to the osteological maturity (sensu Brochu 1996), even though it had not reached a full maturity (CA Brochu personal communication to GMC). This perspective about the age of the specimen, however, still needs to be more thoroughly assessed. The cervical vertebrae of the holotype, and particularly the centra, of Mourasuchus arendsi are anteroposteriorly short, as noted in previous works (Bocquentin-Villanueva 1984; Langston 2008; Tineo et al. 2014). Other specimens of Mourasuchus exhibit a similar morphology, such as the holotype of M. pattersoni (Langston 2008) and YPFB-LIT-PAL-01, from the late Miocene Yecua Formation of Bolivia (Tineo et al. 2014). This feature has been associated by previous authors with a reduced neck movement ability of Mourasuchus compared to extant crocodylians (Langston 2008; Tineo et al. 2014; Cidade et al. 2017), but this hypothesis still needs to be thoroughly tested through biomechanical and/or morphofunctional analyses.

HISTORICAL BIOLOGY

One of the putative cervical vertebrae is almost completely preserved (Figure 9, left). The centrum, including the cotyle and the condyle, the pedicel and the right diapophysis and post-zygapophysis are preserved. The left diapophysis, both pre-zygapophyses, the hypapophysis and the neural spine are partially eroded, while the left post-zygapohysis and both parapophyses are almost completely eroded. The other putative cervical vertebra is less complete (Figure 9, right). It preserves only the centrum, including the cotyle and the condyle, the pedicel and the hypapophysis which has a small size. Both parapophyses and diapophyses are eroded, while the neural spine is absent. The three putative dorsal vertebrae exhibit variable degrees of preservation. One of them is almost complete (Figure 10A), preserving centrum, pedicel, neural spine, right diapophysis and left pre-zygapophysis, while only the most proximal portion of the left diapophysis is preserved. The second vertebra is also almost complete (Figure 10-B), exhibiting centrum, neural spine and pedicels completely preserved, while both diapophyses and the right parapophysis preserve only their most proximal portions. The third vertebra is also almost complete, but significantly deformed towards its medial axis (Figure 10-C). The centrum and the pedicel are preserved, as well as the neural spine, which is slightly eroded and also deformed in a medial direction. The right diapophysis is completely preserved but also deformed, while the left diapophysis only preserves its most proximal portion. The right post-zygapophysis is the only one preserved, while all other zygapophyses are absent. Scapula and coracoid The scapula and the coracoid are articulated and exhibit a distorted shape, especially in their extremities (Figure 11). This shape coincides with that of an articulated scapula and coracoid of Caiman crocodilus illustrated by Brühl (1862) as being in a ‘natural context’ and with that of other specimens of living crocodylians (GMC, personal observation). In this way, no taphonomical processes would be necessary to explain the distorted shape observed in these two articulated bones. The scapula is almost completely preserved, except for the dorsal part of the bone, which preserves only its central portion. The coracoid is also almost complete, with some erosions in the anterior and posterior surfaces of its ventral portion. Additionally, the coracoid foramen is not visible, being probably covered by grypsum. Aside from these features, the overall morphology of both the scapula and the coracoid correspond to that of living crocodylians described by Mook (1921). The deltoid crest of the scapula, which is an origin point of the muscle M. deltoids scapularis inferior (Fürbringer 1876; Brochu 1997), is thin (Brochu 2011; Character 24–0), as in most Alligatoroidea and all Caimaninae taxa except Purussaurus mirandai (see Brochu 2011). The synchondrosis between the scapula and the coracoid is in the beginning of the closure (Figure 11). Given the fact that the specimen had probably not reached its full osteological maturity, this means that the closure of the synchondrosis starts relatively early in the ontogeny (Brochu 2011; Character 25–1), a character that is present only in the Caimaninae clade (Brochu 1999). This feature has been described for

11

caimanine extant taxa (Brochu 1995) and also suggested to be present in the fossil species Necrosuchus ionensis (Brochu 2011). Other two specimens of Mourasuchus preserve scapula and coracoid. The holotype of M. atopus (UCMP-38,012) has an almost complete scapula and a right coracoid formed only by its most proximal portion, which are not articulated (Langston 1965). The holotype of M. pattersoni (MCNCPAL-110-72V) preserves both scapulae and coracoids, with the left elements being articulated (Figure 11-C) while the right ones are not (see Langston 2008; Figure 6). The scapula and coracoid of M. atopus do not show any sign of a closure of the synchondrosis, while the left scapular and coracoid M. pattersoni exhibit what may correspond to the beginning of the closure of the synchondrosis. The possible reasons for these differences are discussed below. Aside from this, no significant difference is noted between the scapulae and the coracoids of these three species. Mourasuchus arendsi as a possible junior synonym of M. atopus Upon describing Mourasuchus arendsi a new species, Bocquentin-Villanueva (1984) observed that it was ‘very similar’ to M. atopus. In fact, the following characters have been used in the literature as taxonomic relevant differences between the two species: i) ‘less parallel’ maxillae in M. arendsi than in M. atopus (Bocquentin-Villanueva 1984; Scheyer and Delfino 2016); ii) smaller external naris in M. arendsi than in M. atopus (Bocquentin-Villanueva 1984); iii) anterior margin of the supratemporal fenestrae comprised by both the postorbital and the parietal bones in M. arendsi, while comprised only by the postorbital in M. atopus (Bona et al. 2013a; Scheyer and Delfino 2016); iv) presence of a dorsal median crest in the parietal in M. arendsi, and its absence in M. atopus (Scheyer and Delfino 2016); v) a more developed anteromedial knob in the margins of the orbits in M. arendsi than in M. atopus (BocquentinVillanueva 1984); vi) the presence of four occlusal pits of the mandibular teeth in the premaxillae of M. arendsi, compared to two in M. atopus (Bocquentin-Villanueva 1984); vii) presence of a pronounced notch at the lateral edge of the jugals in M. arendsi and its absence in M. atopus (Scheyer and Delfino 2016); viii) lateromedially more expanded palatine bones in M. arendsi than in M. atopus (Bocquentin-Villanueva 1984). However, all of these proposed diagnostic characters of Mourasuchus arendsi relative to M. atopus have issues. The first four characters (‘i, ii, iii and iv’) cannot be observed in the holotype of M. atopus. The incomplete preservation of the maxillae and the premaxillae in the holotype of this species (see Langston 1965, fig. 28), respectively, precludes any conclusion about whether the characters ‘i’ and ‘ii’ were present in M. atopus, while the absence of the parietal in the holotype of the last species (see Langston 1965, fig. 28) makes any consideration on whether the characters ‘iii’ and ‘iv’ were present in this taxon impossible. The differences observed in the characters ‘v’ and ‘vi’ can be rather explained by ontogenetic and/or individual variations. The distinct sizes of the knobs located in the anteromedial margins of the orbits (Figure 12) may be the result of ontogenetic variation, since the larger specimens (holotypes of M. arendsi, CIAAP-1297 and M. amazonensis, DGM-526-R) exhibit larger knobs than the

12

G. M. CIDADE ET AL.

1933; Langston 2008; Souza-Filho and Guilherme 2011; Scheyer and Delfino 2016), which undermines the consideration of such difference to be of taxonomic importance for Mourasuchus. Regarding character ‘vii’, the jugals of M. atopus and M. arendsi actually exhibit the same morphology (sensu Cidade et al. 2017; Character 187–0; Figure 13) and, in any case, a pronounced notch at the lateral edge of the jugals is absent in both taxa. Character ‘viii’ is the only one which may be a taxonomically important difference. Studies on the ontogenetic development of the palatine in extant Caimaninae have not found similar ontogenetic variations in the morphology of the palatines (e.g. Augusta 2013; Fernández-Blanco et al. 2018) and even though individual variations regarding the lateromedial expansion of the palatines have been recorded in the extant Caiman latirostris (Freiberg and Carvalho 1965), these are not as accentuated as the difference between the morphologies of the palatines of Mourasuchus arendsi (see Bocquentin-Villanueva 1984; Figure 1) and M. atopus (see Langston 1965, figs. 19 and 28). Scheyer and Delfino (2016) considered that the difference in this character could be due to ontogenetic variation due to the possible status of the holotype of M. atopus as a juvenile. However, the presence of lateromedially constricted palatines similar to M. atopus in the holotype of M. pattersoni, a specimen with a skull length of 106 cm and ontogenetically mature at the time of death (Langston 2008; Cidade et al. 2017; Figure 3) argues against this perspective. However, further studies on extant caimanines about ontogenetic development, individual variation and sexual dimorphism are needed for a better understanding of any possible systematic relevance of differences in the morphology of the palatines. These perspectives notwithstanding, the most significant hindrance to solve the taxonomic situation of Mourasuchus arendsi is the fact that the only available depiction of the palatines (and of the ventral portion of the skull of the holotype as a whole) is the drawing published in the original description (BocquentinVillanueva 1984). As noted previously, the ventral portion of the skull of the holotype cannot be currently examined due to risk of damaging the specimen. As such, an eventual reanalysis of the ventral portion of the skull of M. arendsi is essential to solve the taxonomic issues involving this species. Until such study is performed, the possibility that M. arendsi is a junior synonym of M. atopus due to the possible lack of taxonomically relevant differences between the two species is a hypothesis to be considered, but currently each of these two species may be considered as distinct from each other and as valid species.

Discussion Figure 12. Knob of the anteromedial margin of the orbits of the holotype of Mourasuchus atopus (UCMP-38,012, A) and M. arendsi (CIAAP-1297, B). amk = anteromedial knob of the orbit; o = orbit. Scales =5 cm.

holotype of M. atopus (UCMP-38,012), a smaller specimen which, with a skull length estimated in 73.9 cm (Langston and Gasparini 1997), may represent a juvenile specimen (see Langston 1966; Scheyer and Delfino 2016). Regarding character ‘vi’, differences in the number of occlusal pits such as that between M. atopus and M. arendsi (see Cidade et al. 2017; Figure 5) are also present in extant crocodylians and are long known to vary among individuals and ontogenetic stages in the extant taxa (see Kälin

Exploratory phylogenetic analysis In order to evaluate which consequences a synonymy between Mourasuchus atopus and M. arensi could have for the phylogeny of the genus and of Caimaninae, an exploratory phylogenetic analysis was performed based on the matrix of SouzaFilho et al. (in press) in which the scoring of the two species was combined into one. The combined scoring is available in the Supplemental Online Material. As the two species did not have any character scored differently, no scoring changed were made aside from filling missing data of M. atopus with characters scored for M. arendsi.

HISTORICAL BIOLOGY

13

Figure 13. Morphology of the subtemporal ramus of the jugal in the holotype of Mourasuchus atopus (UCMP-38,012, A), M. arendsi (CIAAP-1297, B), Mourasuchus sp. (UFAC-1424, C) and M. pattersoni (MCNC-PAL-110-72V, D). Adapted from Cidade et al. (2017; Figure 8). str = subtemporal ramus of the jugal. Scales = 5 cm (A) and 10 cm (B, C and D).

The scoring of both species as a singles operational unit resulted in a topology (Figures 14 and 15) that does not change from the topology obtained either for Mourasuchus, for Caimaninae or for Eusuchia as a whole by Souza-Filho et al. (in press). The unit ‘M. atopus+ M. arendsi’ was recovered at the same placement that the two species occupy as a clade: as a sister-taxon to the other species of the genus, M. amazonensis and M. pattersoni. Ontogenetic assessment and the differences on the closure of the scapulocoracoid synchondrosis in Mourasuchus The neurocentral sutures of the cervical vertebrae of the holotype of Mourasuchus arendsi (CIAAP-1297) are partially closed, which indicates that the specimen was close to osteological maturity upon death, although not having reached it. The significant length of the skull (91cm) is congruent with an individual in adulthood. Histological techniques of estimation of age through analysis of osteoderms or long bones (see Buffrénil 1980; Ikejiri 2012) are not possible as CIAAP-1297 preserves none of these bones. Aside from CIAAP-1297, two other Mourasuchus specimens preserve the scapula and the coracoid: the holotypes of

M. atopus (UCMP-38,012) and M. pattersoni (MCNC-PAL110-72V). The only significant difference between these three specimens regarding these bones is the presence of a visible beginning of a process of closure of the synchondrosis between the scapula and the coracoid in M. arendsi, the absence of any sign of the closure in M. atopus and what appears to be a very incipient beginning of closure in the synchondrosis of the left elements of M. pattersoni (Figure 11). This last feature cannot be assured due to the poor preservation of the specimen. Nevertheless, if the beginning of the closure is present in M. pattersoni, it is significantly less developed than that present in M. arendsi. Due to the doubt about whether the synchondrosis of Necrosuchus ionensis was beginning to close or not (Brochu 2011), the holotype of M. arendsi is the only fossil caimanine specimen for which an unequivocal beginning of the closure of the scapulocoracoid synchondrosis can be observed. Such differences observed among the species of Mourasuchus for the closure of the synchondrosis cannot be explained solely by different ontogenetic stages, as the holotype of M. arendsi exhibits an intermediate size (with a skull length of 91 cm) between the holotype of M. atopus (skull length estimated in 73.9 cm by Langston and Gasparini 1997), and the holotype of M. pattersoni (with a skull length of 106 cm – GMC personal observation).

14

G. M. CIDADE ET AL.

Figure 14. Exploratory phylogenetic analysis of the Caimaninae clade performed in this work considering Mourasuchus atopus and M. arendsi as different species.

Figure 15. Exploratory phylogenetic analysis of the Caimaninae clade performed in this work considering Mourasuchus atopus and M. arendsi as the same species (under the name Mourasuchus atopus), showing how the topology does not change if compared to when the two species are considered distinct.

HISTORICAL BIOLOGY

As such, this difference may be explained either by intraspecific or interspecific variation or by sexual dimorphism. Intraspecific variation for the closure of this synchondrosis was reported for Caiman crocodilus by Brochu (1995). Sexual dimorphism can be considered if the holotype of Mourasuchus arendsi is assumed to be a female, as extant female crocodylians exhibit a slower growth rate than males (see Webb et al. 1978; Chabreck and Joanen 1979; Rootes et al. 1991; Brochu 1996, 2003b; Wilkinson and Rhodes 1997; Tucker et al. 2007). Therefore, female individuals may exhibit late ontogenetic changes while having a smaller size than that of a male individual while undergoing the same change. Accordingly, the presence of the beginning of the closure of the synchondrosis in CIAAP-1297 could be the result of such earlier reaching of morphological maturity, while its absence or lesser development in the holotype of M. pattersoni, a larger specimen, would be explained by the fact that the holotype of M. pattersoni would be a male individual. The absence of closure in the holotype of M. atopus, however, can also be understood by the smaller size of the specimen, which could represent a juvenile (see Langston 1966; Scheyer and Delfino 2016), regardless of its sex. However, all these interpretations must still be thoroughly tested in future studies.

Conclusions The redescription of the holotype (CIAAP-1297) of Mourasuchus arendsi (Bocquentin-Villanueva 1984) presented in this paper offers the first thorough reassessment of the morphology of the dorsal surface of the skull. Additionally, it contains an almost completely new, detailed assessment of the morphology of the mandibles and of the postcranium of the holotype. This includes the description of a left scapula articulated with a left coracoid that were not mentioned as belonging to the holotype in the original description. The ventral portion of the skull could not be reassessed as the skull is firmly attached to a plaster jacket that could not be removed. Nevertheless, the redescription provided data that evidenced that M. arendsi is a species distinct from M. amazonensis and M. pattersoni, but the known differences with M. atopus have been revealed to be few. The only possibly significant difference between M. arendsi and M. atopus is that between the lateromedial width of the palatine of the two species. This difference may be a full distinctive character that makes M. arendsi a different species from M. atopus, but intraspecific variation in this character has been observed in Caiman latirostris and future studies on extant caimanines may confirm this character to be intraespecifically variable or reveal it to be ontogenetically variable. If any of these possibilities turns out to be true, then the possibility that M. arendsi is a junior synonym of M. atopus must be properly considered. The exploratory phylogenetic analysis performed showed that this possible synonymy does not change the phylogenetic relationships of the Caimaninae clade. The ontogenetic assessment of the holotype based on the closure of the neurocentral sutures indicates that it was close to osteological maturity upon death but had not yet reached it. The significant length of the skull (91 cm) is also congruent with an individual in adulthood.

15

The scapulocoracoid of the holotype of M. arendsi exhibits a beginning of closure. The same elements in M. atopus do not show any sign of closure, while the left scapulocoracoid of M. pattersoni exhibit what may be the beginning of the closure, although this is significantly less developed than that present in M. arendsi. Such differences cannot be explained by ontogenetic differences, but rather by intraspecific or interspecific variation or sexual dimorphism. In this last hypothesis, the holotype of M. arendsi would be assumed to be a female individual, while the holotype of M. pattersoni could be a male individual.

Acknowledgments We are indebted to Gina Oneda (CIAAP) for access to CIAAP-1297 for study. We thank Carl Mehling (AMNH), Patricia Holroyd (UCMP), Stella Alvarez (MACN), Hyram Moreno (MCNC), Rodrigo Machado (MCT), Jonas P. de Souza-Filho and Andréa Maciente (UFAC), Chris Sagebiel (TMM) and Jessica Cundiff (MCZ) for access to fossil specimens under their care. We thank the editor Gareth Dyke and three anonymous reviewers for comments and suggestions that greatly improved the manuscript. The first author thanks Jorge CarrilloBriceño for support on his visit to the CIAAP collection. We thank Paula Bona (Museo de La Plata, Argentina), for discussions, providing relevant literature and for a critical reading of the first version of the manuscript and Christopher Brochu (University of Iowa, United States) for discussions and for providing relevant literature. We are grateful to Torsten Scheyer (Paläontologisches Institut und Museum, Universität Zürich, Swtizerland) for pictures of the holotype of Mourasuchus arendsi and other specimens. The Willi Hennig Society is thanked for making TNT freely available for download.

Disclosure statement No potential conflict of interest was reported by the authors.

Funding This work was supported by the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) [under grants 2013/04516-1 to GMC and 2011/14080-0 to ASH]; by the Conselho de Desenvolvimento Científico e Tecnológico (CNPq) [grant 40808/2016-7 to GMC and 309434/2015-7 to ASH]; by the Instituto Venezolano de Investigaciones Científicas (IVIC) [grant 1096 to AR] and by the Fundação de Amparo à Pesquisa de Minas Gerais (FAPEMIG) [under grant APQ-02490-12 to DR];Conselho Nacional de Desenvolvimento Científico e Tecnológico [309434/20157,40808/2016-7];Fundação de Amparo à Pesquisa do Estado de Minas Gerais [APQ-02490-12];Fundação de Amparo à Pesquisa do Estado de São Paulo [2011/14080-0,2013/04516-1];Instituto Venezolano de Investigaciones Científicas [1096];

ORCID Giovanne M. Cidade http://orcid.org/0000-0001-8621-5122 http://orcid.org/0000-0003-1862-2724 Andrés Solórzano http://orcid.org/0000-0002-7102-5358 Ascánio Daniel Rincón

References Aguilera OA. 2004. Tesoros Paleontológicos de Venezuela: Urumaco, património natural de la humanidad. Caracas: Universidad Nacional Experimental Francisco de Miranda, Editora Arte.. Aguilera OA, Riff D, Bocquentin-Villanueva J. 2006. A new giant Purussaurus (Crocodyliformes, Alligatoridae) from the upper Miocene Urumaco Formation, Venezuela. J Sys Palaeontol. 4(3):221–232.

16

G. M. CIDADE ET AL.

Augusta BG 2013. Revisão sistemática e ontogenética dos materiais cranianos atribuídos ao gênero Mariliasuchus (Crocodyliformes, Notosuchia) e suas implicações taxonômicas e paleobiológicas [Master’s Degree Dissertation]. São Paulo (Brazil): University of São Paulo. Barrios F 2011. Nuevos restos de Alligatoridae Cuvier, 1807 (Eusuchia, Crocodylia) del Neógeno del Sur de la província de Salta, Implicancias sistemáticas [Thesis of Licenciature]. Salta (Argentina): Universidad Nacional de Salta. Benton MJ, Clark JM. 1988. Archosaur phylogeny and the relationships of the Crocodylia. In: Benton MJ, editor. The phylogeny and classification of the tetrapods. Oxford: Clarendon Press. Vol.1; p. 295–338. Bocquentin-Villanueva J. 1984. Um nuevo Nettosuchidae (Crocodylia, Eusuchia) proveniente de la Formación Urumaco (Mioceno Superior), Venezuela. Ameghiniana. 21:3–8. Bocquentin-Villanueva J, Buffetaut E. 1981. Hesperogavialis cruxenti n. gen., n. sp., nouveau gavialide (Crocodylia, Eusuchia) du Miocène Supériur (Huayquerien) d’Urumaco (Venezuela). Geobios. 14:415–419. Bona P. 2007. Una nueva especie de Eocaiman Simpson (Crocodylia, Alligatoridae) del Paleoceno Inferior de Patagonia. Ameghiniana. 44 (2):435–445. Bona P, Degrange FJ, Fernández MS. 2013b. Skull anatomy of the bizarre crocodylian Mourasuchus nativus (Alligatoridae, Caimaninae). Anat Rec. 296(2):227–239. Bona P, Ezcurra MD, Barrios F, Fernandez-Blanco MV. 2018. A new Palaeocene crocodylian from southern Argentina sheds light on the early history of caimanines. Proc Roy Soc B. 285:20180843. Bona P, Fernandez-Blanco MV, Scheyer TM, Foth C. 2017. Shedding light on the taxonomic diversity of the South American Miocene caimans: the status of Melanosuchus fisheri (Crocodylia, Alligatoroidea). Ameghiniana. 54:681–687. Bona P, Riff D, Gasparini Z. 2013a. Late miocene crocodylians from Northeast Argentina: new approaches about the austral components of the Neogene South American crocodylian fauna. Ear Envir Sci Tran R Soc Ed. 103:1–20. Brochu CA. 1995. Heterochrony in the crocodylian scapulocoracoid. J Herpetol. 29(3):464–468. Brochu CA. 1996. Closure of neurocentral sutures during crocodilian ontogeny: implications for maturity assessment in fossil archosaurs. J Vert Paleontol. 16(1):49–62. Brochu CA 1997. Phylogenetic systematics and taxonomy of Crocodylia [Phd thesis]. Austin (TX): University of Texas. Brochu CA. 1999. Phylogenetics, taxonomy, and historical biogeography of Alligatoroidea. Mem Soc Vert Paleo. 6:9–100. Brochu CA. 2003a. Phylogenetic approaches toward crocodilian history. Ann Rev Ear Plan Sci. 31:357–397. Brochu CA. 2003b. Osteology of Tyrannosaurus rex: insights from a nearly complete skeleton and high-resolution computed tomographic analysis of the skull. Mem Soc Vert Paleo. 7:1–138. Brochu CA. 2010. A new alligatoroid from the lower eocene green river formation of Wyoming and the origin of caimans. J Vert Paleontol. 30:1109–1126. Brochu CA. 2011. Phylogenetic relationships of Necrosuchus ionensis Simpson, 1937 and the early history of caimanines. Zool J Linn Soc. 163:S228–S256. Brühl CB. 1862. Das Skelett der Krokodilinen. Vienna: Wilhelm Braumüller. Buffrénil V. 1980. Données preliminaires sur la structure des marques de croissance squeletiques chez les crocodilians actuels et fossils. Bull Soc Zool Fr. 105:355–361. Cáceres C, Rincón AD, Solórzano A, Núñez-Flores M, Ruiz-Ramoni D, Sánchez L. 2016. The first record of Necronectes (Decapoda: portunidae) in Venezuela. Rev Bra de Paleontol. 19(1):145–148. Chabreck RH, Joanen T. 1979. Growth rates of American alligators in Louisiana. Herpetologica. 35:51–57. Cidade GM, Riff D, Souza-Filho JP, Hsiou AS, Montefeltro F. 2013. Description of nine new specimens of Mourasuchus nativus (Alligatoroidea, Caimaninae), and comments on ontogenetic development and intraspecific variation of the skull table. J Vert Paleontol Prog Abstr. 2013:107.

Cidade GM, Solórzano A, Rincón AD, Riff D, Hsiou AS. 2017. A new Mourasuchus (Alligatoroidea, Caimaninae) from the late Miocene of Venezuela, the phylogeny of Caimaninae and considerations on the feeding habits of Mourasuchus. PeerJ. 5:e3056. Díaz de Gamero ML, Linares OJ. 1989. Estratigrafía y paleontología de la Formación Urumaco, del Mioceno tardío de Falcón noroccidental. VII Con Geol Venez Mem. 1:419–439. Fernández-Blanco MV, Cassini GH, Bona P. 2018. Skull ontogeny of extant caimans: a three-dimensional geometric morphometric approach. Zool. 129(2018):69–81. Fortier DC, Souza-Filho JP, Guilherme E, Maciente A, Schultz CL. 2014. A new specimen of Caiman brevirostris (Crocodylia, Alligatoridae) from the late miocene of Brazil. J Vert Paleontol. 34(4):820–834. Foth C, Fernandez-Blanco MV, Bona P, Scheyer TM. 2017. Cranial shape variation in jacarean caimanines (Crocodylia, Alligatoroidea) and its implications in the taxonomic status of extinct species the case of Melanosuchus fisheri. J Morphol. 2017:1–15. Freiberg MA, Carvalho AL. 1965. El yacare sudamericano Caiman latirostris (Daudin). Physis XXV. 70:351–360. Fürbringer M. 1876. Zur vergleichenden Anatomie der Schultermuskeln. Part III. Gegenb Morphol Jahrb. 1:636–816. Gasparini Z. 1985. Un Nuevo cocodrilo (Eusuchia) Cenozoico de América del Sur. Col Trab Paleont MME–DNPM. sér Geol. 27:51–53. George ID, Holliday CM. 2013. Trigeminal nerve morphology in alligator mississipiensis and its significance for Crocodyliform facial sensation and evolution. Anat Rec. 296:670–680. Gmelin J. 1789. Linnei Systema Naturae. Leipzig: G.E. Beer. Goloboff PA, Farris JS, Nixon KC. 2008. TNT: a free program for tree analysis using new technology. Cladistics. 24(5):774–786. Gray JE. 1844. Catalogue of tortoises, crocodilians, and amphisbaenians in the collection of the British Museum. London: British Museum (Natural History). Hambalek N, Rull V, De Digiacomo E, Díaz de Gamero ML. 1994. Evolución paleoecológica y paleoambiental de la secuencia del Neogeno en el surco de Urumaco, estudio palinológico y litológico. Bol Soc Venez Geol. 191–192:7–19. Hastings AK, Bloch JI, Jaramillo CA, Rincon AF, MacFadden BJ. 2013. Systematics and biogeography of crocodylians from the Miocene of Panama. J Vert Paleontol. 33(2):239–263. Hastings AK, Reisser M, Scheyer TM. 2016. Character evolution and the origin of Caimaninae (Crocodylia) in the new world tropics: new evidence from the Miocene of Panama and Venezuela. J Paleontol. 90(2):317–332. Holliday CM, Witmer LM. 2009. The epipterygoid of Crocodyliforms and its significance for the evolution of the orbitotemporal region of Eusuchians. J Vert Paleontol. 29(3):715–733. Ikejiri T. 2012. Histology-based morphology of the neurocentral synchondrosis in Alligator mississippiensis (Archosauria, Crocodylia). Anat Rec. 295:18–31. Iordansky NN. 1973. The Skull of the Crocodylia. In: Gans C, Parsons TS, editors. Biology of the Reptilia volume 4. New York (NY): Academy Press; p. 201–262. Kälin JA. 1933. Beiträge zur vergleichenden Osteologie des Crocodilidenschädels. Zool Jahrb Abt Anat Ontog Tie. 57:535–714. Langston W. 1965. Fossil crocodilians from Colombia and the Cenozoic history of the Crocodilia in South America. Univ Calif Publ Geol Sci. 52:1–168. Langston W. 1966. Mourasuchus Price, Nettosuchus Langston, and the family Nettosuchidae (Reptilia: crocodilia). Copeia. 4:882–885. Langston W. 2008. Notes on a partial skeleton of Mourasuchus (Crocodylia, Nettosuchidae) from the Upper Miocene of Venezuela. Arq Mus Nac. 66:125–143. Langston W, Gasparini Z. 1997. Crocodilians, Gryposuchus, and the South Americans gavials. In: Kay RF, Madden RH, Ciffelli RL, Flynn JJ, Colombia LV, editors. Vertebrate paleontology in the neotropics: the Miocene fauna of La Venta, Colombia. Washington (DC): Smithsonian Institution. p. 113–154. Linares OJ. 2004. Bioestratigrafía de la fauna de mamíferos de las formaciones Socorro, Urumaco y Codoré (Mioceno medio-Plioceno temprano) de la región de Urumaco, Falcón, Venezuela. Paleobio Tropic. 1:1–26.

HISTORICAL BIOLOGY

Maddison WP, Maddison DR 2011. Mesquite: a modular system for evolutionary analysis. Version 2.75. [accessed 2018 Sep 23]. http://mesquite project.org/. Mook CC. 1921. Notes on the postcranial skeleton in the Crocodilia. Bull Am Mus Nat Hist. 44:67–100. Norell MA 1988. Cladistic approaches to paleobiology as applied to the phylogeny of alligatorids [Phd thesis]. New Haven (CT): Yale University. Mook CC. 1924. A new crocodilian from the Wasatch Beds. Am Mus Nov. 137:1–4. Norell MA, Clark JM, Hutchison JH. 1994. The Late Cretaceous alligatoroid Brachychampsa montana (Crocodylia): new material and putative relationships. Am Mus Nov. 3116:1–26. Pinheiro AEP, Foriter DC, Pol D, Campos DA, Bergqvist LP. 2013. A new Eocaiman (Alligatoridae, Crocodylia) from the Itaboraí Basin, Paleogene of Rio de Janeiro, Brazil. Hist Biol. 25(3):327–337. doi:10.1080/08912963.2012.705838. Price L. 1964. Sobre o crânio de um grande crocodilídeo extinto do Alto de Rio Juruá, Estado do Acre. Acad Brasil Ciên. 56:59–66. Quiroz L, Jaramillo CA. 2010. Stratigraphy and sedimentary environments of Miocene shallow to marginal marine deposits in the Urumaco trough, Falcón basin, Western Venezuela. In: SánchezVillagra MR, Aguilera OA, Carlini AA, editors. Urumaco and Venezuelan palaeontology – the fossil record of the Northern Neotropics. Bloomington (IN): Indiana University Press; p. 153–172. Riff D, Aguilera O. 2008. The world’s largest gharials Gryposuchus: description of G. croizati n. sp. (Crocodylia, Gavialidae) from the upper miocene urumaco formation, Venezuela. Palaeo Zeit. 82:178–195. Riff D, Romano PSR, Oliveira GR, Aguilera OA. 2010. Neogene crocodile and turtle fauna in Northern South America. In: Hoorn C, Wesselingh F, editors. Amazonia landscape and species evolution: a look into the past. London: Wiley-Blackwell; p. 259–280. Rootes WL, Chabreck RH, Wright VL, Brown BW, Hess TJ. 1991. Growth rates of American alligators in estuarine and palustrine wetlands in Louisiana. Estuaries. 14:489–494. Salas-Gismondi R, Flynn JJ, Baby P, Tejada-Lara JV, Wesselingh FP, Antoine P-O. 2015. A Miocene hyperdiverse crocodylian community reveals peculiar trophic dynamics in proto-Amazonian mega-wetlands. Proc Roy Soc B. 282:20142490. Scheyer TM, Aguilera OA, Delfino M, Fortier DC, Carlini AA, Sánchez R, Carrillo-Briceño JD, Quiroz L, Sánchez-Villagra MR. 2013.

17

Crocodylian diversity peak and extinction in the late Cenozoic of the northern Neotropics. Nat Commun. 4:1907. Scheyer TM, Delfino M. 2016. The late Miocene caimanine fauna (Crocodylia: alligatoroidea) of the urumaco formation, Venezuela. Palaeontol Electron. 19.3.48A:1–57. Scheyer TM, Moreno-Bernal JW. 2010. Fossil crocodylians from Venezuela in the context of South American faunas. In: SánchezVillagra MR, Aguilera OA, Carlini AA, editors. Urumaco and Venezuelan palaeontology – the fossil record of the Northern Neotropics. Bloomington (IN): Indiana University Press; p. 192–213. Scheyer TM, Straehl FR, Sánchez-Villagra MR. 2015. Crocodile in a Tree. Zurich: Scidinge Hall Verlag. Sill W. 1970. Nota preliminar sobre un nuevo gavial del Plioceno de Venezuela y una discusión de los gaviales Sudamericanos. Ameghiniana. 7:151–159. Souza RG, Cidade GM, Campos DA, Riff D. 2016. New crocodylian remains from the solimões formation (Lower Eocene–pliocene), state of Acre, southwestern Brazilian Amazonia. Rev Bras Paleontol. 19(2):217–232. Souza-Filho JP, Guilherme E. 2011. Ampliação da diagnose e primeiro registro de Mourasuchus arendsi (Crocodylia-Nettosuchidae) no Neógeno da Formação Solimões, Amazônia Sul-Ocidental. In: Carvalho IS, editor. Paleontologia: Cenários de Vida, Volume 4 (1st edition). Rio de Janeiro (Brazil): Interciencia; p. 399–408 Souza-Filho JP, Souza RG, Hsiou AS, Riff D, Guilherme E, Negri FR, Cidade GM. in press. A new caimanine (Crocodylia, Alligatoroidea) species from the Solimões Formation of Brazil and the phylogeny of Caimaninae. J Vert Paleontol.doi: 10.1080/02724634.2018.1528450 Tineo DE, Bona P, Pérez LM, Vergani GD, González G, Poiré DG, Gasparini ZN, Legarreta P. 2014. Palaeoenvironmental implications of the giant crocodylian Mourasuchus (Alligatoridae, Caimaninae) in the Yecua Formation (late Miocene) of Bolivia. Alcheringa. 39 (2):224–235. Tucker AD, Limpus CJ, McDonald KR, Hamish I, McCallum H. 2007. Growth dynamics of freshwater crocodiles (Crocodylus johnstoni) in the Lynd River, Queensland. Austral J Zool. 54(6):409–415. Webb GKW, Messel H, Crawford J, Yerbury MJ. 1978. Growth rates of Crocodylus porosus (Reptilia: crocodilia) from Arnhem Land, Northern Australia. Austral Wild Res. 5(3):385–399. Wilkinson PM, Rhodes WE. 1997. Growth rates of American alligators in coastal South Carolina. J Wild Manag. 61(2):397–402.