Andrade & Bertini, 2008

Historical Biology Vol. 20, No. 2, June 2008, 101–136

A new Sphagesaurus (Mesoeucrocodylia: Notosuchia) from the Upper Cretaceous of Monte Alto City (Bauru Group, Brazil), and a revision of the Sphagesauridae Marco Brandalise de Andradeab* and Reinaldo J. Bertinia† a

Nu´cleo de Evoluc¸a˜o e Paleobiologia de Vertebrados, Departamento de Geologia Aplicada, Instituto de Geocieˆncias e Cieˆncias Exatas, Universidade Estadual Paulista, Rio Claro, SP, Brazil; bDepartment of Earth Sciences, Faculty of Sciences, University of Bristol, Bristol, England, United Kingdom Since the description of Sphagesaurus, mostly dental material has been reported, apart from two incomplete skulls. Here we describe a new species of Sphagesaurus, from Monte Alto City, Southeastern Brazil, which includes the skull and most of the mandible. Distinctive characters (e.g. antorbital fenestra; robust quadrate; anterior mandibular teeth incisiform; ornamented sulcate palate) allow differentiation from S. huenei. Several characters allow assignment to the genus Sphagesaurus (e.g. teardrop-like oblique molariform teeth), while new information is provided (e.g. premaxilla, pterygoid and mandible morphology; jugal foramen; occipital surface; battery of mandibular teeth). A revision of the Family Sphagesauridae Kuhn 1968 is given. A preliminary phylogenetic analysis supports a sister-taxon relationship for S. huenei and the new species. The phylogenetic relationship of notosuchians is explored. Sphagesaurids were terrestrial notosuchians that evolved during the Upper Cretaceous of South America, known only from the Adamantina Formation, Campanian – Maastrichtian (Upper Cretaceous) from Brazil.

Keywords: Sphagesaurus; Mesoeucrocodylia; Notosuchia; Cretaceous; Bauru Group; Brazil Institutional abbreviations: AMNH, American Museum of Natural History, New York, USA; DGM, DNPM, Departamento Nacional da Produc¸a˜o Mineral, Rio de Janeiro, Brazil; MACN, Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina; MLP, Museo de La Plata, La Plata, Argentina; MPMA, MMA-R, Museu de Paleontologia de Monte Alto, Monte Alto, Brazil; RCL, PUC, Museu de Cieˆncias Naturais da Pontifı´cia Universidade Cato´lica de Minas Gerais, Belo Horizonte, Brazil; UCD, University College Dublin, Dublin, Ireland; UFRJ, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil; URC, Museu de Paleontologia e Estratigrafia ‘Prof. Dr. Paulo Milton Barbosa Landim’, Universidade Estadual Paulista, Rio Claro, Brazil

Introduction Historical background In 1950 Sphagesaurus huenei was described as a notosuchid mesosuchian, based on two teeth from different localities (Presidente Prudente and Catanduva), Adamantina Formation of the Western Sa˜o Paulo State, Brazil (Price 1950). The material, although scarce, had characteristic features allowing its recognition as a crocodilian, as the teeth had long single roots and a constriction between root and crown. Apart from other known species, the crowns had a wide base, basal-apical ridges, either a posterior or an anterior serrated keel and a triangular cross-section. In fact, from its morphology, Price (1950) was able to predict the peculiar pattern in which these teeth should be implanted, in an oblique position, as in Notosuchus. In the diagnosis and the description there are also references to the particular way

the teeth should occlude, with mandibular teeth inversely oriented with respect to the maxillary teeth. This type of occlusion resulted in the bevelling of the posterior carina into a sharp cutting edge, at least in maxillary teeth (Price 1950, pp. 79– 80).

Prior to its formal description, Sphagesaurus was first noticed in 1931 by von Huene, who recorded its existence and identified the fossil teeth as crocodilian, but regretfully these specimens were considered lost.

The specimens described by Price (1950) were collected even earlier, in 1917, and are housed by DNPM (Brazil).

Subsequent discoveries were rare, resulting in a few minor notes on more general issues, until near the end of the century (Price 1955; Arid and Vizotto 1965; Steel 1973; Buffetaut 1982; Bertini 1993; Bertini et al. 1993).

Price (1950, pp. 81 – 82) originally placed Sphagesaurus in the family Notosuchidae, due to the inferred oblique disposition of teeth and the morphology of the crown,

*Corresponding author. Email: marco.b.andrade@bris.ac.uk; marcobranda@yahoo.com.br † Email: rbertini@rc.unesp.br ISSN 0891-2963 print/ISSN 1029-2381 online q 2008 Taylor & Francis DOI: 10.1080/08912960701642949 http://www.informaworld.com

102 M.B. de Andrade and R.J. Bertini though in a provisional manner. Kuhn (1968) assigned the genus to its own family, Sphagesauridae, on the basis of tooth morphology only. Gasparini (1971) preferred not to recognise Sphagesauridae based solely on teeth characters, but rather considered Sphagesaurus a mesosuchian of uncertain relationships. The adoption of phylogenetic systematics (Hennig 1950, 1966) led to the use of non-eusuchian Mesoeucrocodylia as a reasonable synonym of the previous Mesosuchia (Whetstone and Whybrow 1983; Benton and Clark 1988; Clark 1994). Sphagesaurus was then defined as a basal mesoeucrocodilian, or Mesoeucrocodylia indet.

Specimens including the skull are rare and poorly preserved. Kellner et al. (1995) reported the existence of DGM 1411-R, a badly crushed Sphagesaurus rostrum and symphysis (Kellner et al. 1995; Kellner and Campos 1999; Pol 2003; Bertini and Andrade 2004; Andrade 2005; Candeiro and Martinelli 2006), which remains undescribed. Nonetheless, the specimen corroborates the oblique implantation inferred by Price (1950).

The description of RCL-100 (Pol 2003) represented a major advance to the knowledge of Sphagesaurus.

The specimen included a skull and symphysis from S. huenei, which allowed a better understanding of its functional morphology and evolutionary relationships.

Important information was added to the particular disposition of teeth (‘opposing triangles’) with an alternate occlusion pattern. Furthermore, the wear facets present in a few teeth supported the occurrence of mandibular movements.

In 1990, new material was discovered (Figures 1– 3) in Central-Northern Sa˜o Paulo State, Southeastern Brazil (Adamantina Formation), at the vicinities of Monte Alto City. The specimen is composed by a set of well preserved skull and mandible, in occlusion. The first reports (Bertini 1993; Bertini and Arruda-Campos 1995; Bertini and Carvalho 1999; Andrade and Bertini 2003) included observations based mostly on general features, suggesting the material might be assigned to Uruguaysuchidae. Unpublished works (Andrade 2005; Andrade and Bertini 2007a) collected further morphological information on the specimen and included the material in a phylogenetic study of notosuchians, obtaining a close relationship with S. huenei. Andrade et al. (2006) described some of its morphological features, related to the choanae. Here we extensively describe this new species and provide new morphological information that allows a revision of the Genus Sphagesaurus and the Family Sphagesauridae.

Geological setting: a review of time and space of Sphagesauridae Sphagesaurids are found in deposits of the Adamantina Formation, an important sedimentary unit from the Bauru

Group, Parana´ Basin. More accurately, all specimens of Sphagesaurus, including MPMA 15-001/90, were collected in the Southeastern area of the Adamantina unit, which occurs within the State of Sa˜o Paulo, Southeastern Brazil (Price 1950; Bertini 1993; Bertini et al. 1993; Kellner et al. 1995; Carvalho and Bertini 2000; Pol 2003; Bertini and Andrade 2004; Candeiro and Martinelli 2006). Nonetheless, RCL-100 was reported originally as from Bueno´polis City, Sa˜o Paulo State (Pol 2003), but was in fact collected in the Fazenda Bueno´polis Rail Station, located in the vicinities of Cravinhos City, Northern Sa˜o Paulo State, Southeastern Brazil (C. G. Cartelle, personal communication). All Sphagesaurus specimens ever collected came from the ‘Sa˜o Paulo State’ portion of the Adamantina Formation.

The Bauru Group sensu Batezelli et al. (2003) is an extensive depositional unit, comprising about 330,000 km2 of Upper Cretaceous deposits. It lies over rocks of either the mid Cretaceous Caiua´ Group sensu Fernandes (1992), or the older basaltic Serra Geral Formation, depending on the area of occurrence. The Bauru Group is also referred as Bauru Basin, as it may represent an independent depositional cycle itself (e.g.

Fernandes and Coimbra 1996, 2000; Kellner and Campos 1999; Santucci and Bertini 2001; Batezelli et al. 2003).

Fernandes and Coimbra (1996, 2000) consider the Bauru Basin to consist of both the Bauru and the Caiua´ groups, as these are seen as intergrading and isochronous units.

Batezelli et al. (2003) include in the Bauru Basin only the Bauru Group, due to the existence of extensive unconformity surfaces between Bauru and Caiua´ sedimentary units, each one thus comprising an independent depositional cycle.

The holotype of this new species was collected in the vicinity of Monte Alto City, Central-Northern Sa˜o Paulo State (Figure 4). The stratigraphic column for the State includes only the Arac¸atuba, Adamantina and Marilı´a formations. Cretacic outcrops in the neighbourhood of Monte Alto City are composed only of Adamantina and Marilı´a sediments. The Adamantina deposits extend over a large area of Central-Northern to Western Sa˜o Paulo State, usually above the Arac¸atuba Formation, upon which it broadly interbeds. Except for Adamantina exposed areas, the unit is covered by sediments of the Marilı´a Formation, with discrete interbedding (Carvalho and Bertini 2000; Bertini and Andrade 2004). The Adamantina sediments are usually composed of fine to medium-grained sandstone, with brown to reddish colour. The Adamantina Formation is interpreted as a set of fluvial deposits (Santucci and Bertini 2001; Batezelli et al. 2003).

Although there is some debate about the age of the Upper Cretaceous deposits of the Bauru Group, the Adamantina Formation is probably Campanian to early Maastrichtian (Gobbo-Rodrigues 2001; Santucci and Bertini 2001). The opposing hypothesis (Dias-Brito et al.

Historical Biology 103

Figure 1. MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype. (a) Lateral view with mandible. (b) Parietal view. (c) Palatal view. Note the characteristic disposition of the teeth, with oblique implantation and the arched disposition of molariform teeth.

Scale bar Âź 20 mm. Anatomical abbreviations according to Appendix A.

104 M.B. de Andrade and R.J. Bertini

Figure 2. Graphic representation of MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype, showing main suture lines recognisable in the specimen. (a) Lateral view with mandible. (b) Parietal view. (c) Palatal view. Sutures represented by stronger lines; dashed lines representing inferred sutures. Scale bar Âź 20 mm.

Historical Biology 105

Figure 3. MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype. Frontal view (a) with mandible, and occipital view (b) without the mandible. Mandible in dorsal (c) and ventral (d) views. Sutures in the occipital view (b) shown by superimposed lines.

Note the extreme robustness of the mandible and the characteristic disposition of the teeth, with oblique implantation. Scale bar Âź 20 mm.

106 M.B. de Andrade and R.J. Bertini

Figure 4. Geographical and stratigraphical range of Sphagesaurus montealtensis MPMA 15-001/90, at the Bauru Group, Southeastern Brazil (based in Fernandes and Coimbra 1996; Batezelli 1998; Batezelli et al. 2003). Type-locality at the Bairro Cachoeira, in the vicinity of the Monte Alto City, indicated. The lithological column (right) shows the distribution of the Cretaceous geological unities in the State of Sa˜o Paulo, and the relative position of the fossil material in the Adamantina Formation (modified from Batezelli 2003; Batezelli et al. 2003).

2001; Candeiro et al. 2006) argues for a Campanian depositional hiatus, coincident to the Adamantina– Marilı´a contact, and thus for an earlier age for the Adamantina Formation (Turonian– Santonian). Nevertheless, considering the existence of several gradational contacts between these units (Batezelli 1998, 2003; Batezelli et al. 2003), the Campanian hiatus hypothesis seems to lack a strong support. Subsurface geophysical data, provided by Paula e Silva et al. (2003), also supports both the existence of a Caiua´ / Bauru unconformity and the intergraded contact between the Adamantina and Marilı´a formations. This means that (i) the Bauru Group depositional cycle immediately follows the Caiua´ cycle and that (ii) no Campanian hiatus is possible, as respectively proposed by Batezelli (1998) and Gobbo-Rodrigues (2001).

Besides Sphagesaurus, the Adamantina Formation has yielded a representative series of fossil vertebrates, including fishes, lizards, snakes, turtles, maniraptoriforms, sauropods and at least one mammaliform.

Invertebrates include molluscs, ostracods and conchostraceans (Mezzalira 1989; Bertini 1993; Bertini et al.

1993; Kellner and Campos 1999; Santucci and Bertini 2001; Candeiro et al. 2006). It is widely agreed that

crocodilomorphs are especially abundant in the Bauru Group (Bertini 1993; Bertini et al. 1993; Kellner and Campos 1999; Candeiro et al. 2006).

Systematic palaeontology Crocodylomorpha Walker 1970 Crocodyliformes Hay 1930 (sensu Benton and Clark 1988) Mesoeucrocodylia Whetstone and Whybrow 1983 Notosuchia Gasparini 1971 (sensu Sereno et al. 2001) Family: Sphagesauridae Kuhn 1968 Diagnosis of the family As for the genus Sphagesaurus.

Sphagesaurus Price 1950 Type-species: Sphagesaurus huenei Price 1950 1931 Crocodylia indet. Von Huene 1950 Sphagesaurus huenei Price 1993 Sphagesaurus huenei (Price) Bertini, pp. 216– 217; Figure 35 (p. 453)

Historical Biology 107 1993 Uruguaysuchidae indet. Bertini, p. 185 – 188, Figures 29– 30 (pp. 440 –441) 1993 Sphagesaurus sp. Bertini et al., p. 79, Figures 6– 10 1995 Sphagesaurus sp. Kellner et al.

1995 Uruguaysuchidae Bertini & Arruda-Campos 2003 Sphagesaurus huenei (Price) Pol 2003 Uruguaysuchidae Andrade & Bertini 2006 Sphagesaurus sp. Andrade et al.

2007a Sphagesaurus sp. Andrade & Bertini Etymology After Spha´geus, meaning ‘slayer’ or ‘butcher’; and Saurus, meaning ‘reptile’.

surface); symphysis including pairs 1– 5, composing a battery of teeth; mandibular pairs 3 and anterior with caniniform –incisiform morphology; at least mandibular pairs 4 and 5 molariform; at least mandibular pair 5 with a strong oblique implantation (45 – 90 degrees rotation, in respect to the maxillary alveolar surface); mandibular pairs 1 –5 increase in size; maxillary molariforms stout, anteriorly rounded and posteriorly compressed, with a teardrop-like cross-section; mandibular molariforms stout, anteriorly compressed and posteriorly more rounded (reverse teardrop-like cross-section); maxillary molariforms with one single posterior carina forming a cutting edge, opposing mandibulary molariforms with an anterior one.

Sphagesaurus huenei Price 1950

Geographical Range Southwestern to Northern Sa˜o Paulo State, Southeastern Brazil.

Stratigraphical Range Adamantina Formation, Bauru Group; Campanian to Maastrichtian (Gobbo-Rodrigues 2001; Santucci and Bertini 2001; Batezelli et al. 2003), Upper Cretaceous.

Diagnosis for the genus Skull as high as wide; rostrum slightly elongated; naris anteriorly directed; internarial bar absent; orbits lateral; jugal anterior and posterior rami horizontal and level to each other; jugal foramen single, anteriorly directed, positioned on the lateral surface of the anterior jugal ramus; ascending jugal ramus and postorbital bar inclined posteromedially; triangular antorbital fossa delimited anteriorly by jugal –maxillary and lachrymal crests; elongated supratemporal fenestrae, with parallel main axis coincident with the sagittal axis of the skull; occipital surface exposed posterodorsally (exoccipitals) above, and posteroventrally (basioccipital, basisphenoid) below the foramen magnum; triangular choanae; posterolateral palatine processes present, delimiting anterolateral borders of the choanal opening; palatine and ectopterygoid contact posterior to the suborbital fenestra posterior end; pterygoid and jugal excluded from suborbital fenestra; lachrymal contacting the nasal; edentulous region on the front of the premaxilla; well developed alveolus, posterior to the hypertrophied caniniform; basioccipital wider than basisphenoid; basioccipital – quadrate contact present, well developed; teeth all single cusped; eight upper and nine lower (mandibulary) teeth; reduced premaxillae, bearing at least one pair of hypertrophied caniniform teeth; six to seven maxillary teeth, all molariform, with a strong oblique implantation (45 – 90 degrees of paramesial rotation, in respect to the maxillary alveolar

1931 1950 1993 1993 1995 2003

Crocodylia indet. Von Huene Sphagesaurus huenei Price Sphagesaurus huenei (Price) Bertini Sphagesaurus huenei (Price) Bertini et al.

Sphagesaurus sp. Kellner et al.

Sphagesaurus huenei (Price) Pol

Etymology Specific name after Friedrich von Huene, palaeontologist who first recorded the existence of the fossil material from this species, in 1931.

Holotype DGM 332-R, isolated molariform teeth.

Referred materials DGM 333-R (Price 1950) and URC R†015 (Bertini 1993; Bertini et al. 1993), isolated molariform teeth.

DGM 1411-R (Kellner et al. 1995; Kellner and Campos 1999), badly crushed rostrum and mandibular symphysis, with several well preserved teeth. RCL-100 (Pol 2003), a partially preserved skull and symphysis; skull with dorsoventral compression of the temporal region, a transversal break that separates the skull in two halves (at the orbit) and partially eroded temporal region (Pol 2003; Andrade 2005).

Type-locality Guajussara, an abandoned railway cut from Estrada de Ferro Sorocabana Co., between Presidente Prudente and Santo Anasta´cio cities, Southwestern Sa˜o Paulo State (Price 1950).

Geographical range Southwestern to Northern Sa˜o Paulo State, Southeastern Brazil. Referred materials from Northern Sa˜o Paulo State: DGM 333-R, from the vicinities of Catanduva

108 M.B. de Andrade and R.J. Bertini City; RCL-100, from Fazenda Bueno´polis Rail Station, vicinities of Cravinhos City, as corrected from Pol (2003) upon detailed information provided by C. G. Cartelle (personal communication). Referred materials from Southwestern Sa˜o Paulo State: URC R†015, from an abandoned stone mining facility, 1 km Southwest of Santo Anasta´cio City; DGM 1411-R, from vicinity of the Presidente Prudente City.

Stratigraphical range Adamantina Formation, Bauru Group; Campanian to Maastrichtian (Gobbo-Rodrigues 2001, Santucci and Bertini 2001, Batezelli et al. 2003), Upper Cretaceous.

Diagnosis for the species Antorbital fenestra absent; premaxilla –maxilla suture extending far posteriorly from the hypertrophied caniniform, on the lateral surface of the skull; maxillae forming a posterior wall above the anterior margin of the suborbital fenestra, contacting prefrontal and lachrymal; jugal anterior ramus robust, enlarging progressively towards its contact with the maxilla; palate surface plain; maxillae palatine rami straight and level to each other, not inclined; naso-oral fenestra in premaxilla – maxilla suture; quadrate light, with a trirradiate cross-section; quadrate medial ventral crest thin; quadrate crests reaching from the outer tip of each quadrate, contacting over basisphenoid; pterygoid – basisphenoid surface concave in contact; maxillary dental series forming a straight line in ventral view; two premaxillary and six maxillary teeth; five mandibulary teeth on the symphysis, pairs 1– 3 caniniform, at least pairs 4– 5 molariform; surface of molariform teeth presenting a thick coarsely pebbled enamel coat.

Sphagesaurus montealtensis sp. nov 1993 Uruguaysuchidae Bertini 1995 Uruguaysuchidae Bertini & Arruda-Campos 2003 Uruguaysuchidae Andrade & Bertini 2006 Sphagesaurus sp. Andrade et al.

2007a Sphagesaurus sp. Andrade & Bertini Etymology Specific name after the City of Monte Alto, with a high number of fossil sites from the Adamantina and Marı´lia formations, one of them the type-locality for this species.

Holotype MPMA 15-001/90, skull and mandible preserved in occlusion, the only specimen known. The skull lacks the anteriormost tip of the premaxillae (under the nasal opening) and the central and posterior left portion of the skull; the mandible lacks the anteriormost section of the

symphysis, most of the left ramus and part of the right ramus. The remaining skull and mandible are exceptionally well preserved, without any kind of deformation or abrasion to the surfaces. Previously referred with the provisional coding MMA-R-14, in the unpublished PhD thesis by RJB (Bertini 1993).

Type-locality Bairro Cachoeira, locality at the base of the Serra da ´ gua Limpa, about 8 km Northwest of the City of Monte A Alto, Northern Sa˜o Paulo State, Southeastern Brazil.

Stratigraphical range and associated sediment Adamantina Formation, Bauru Group; Campanian to Maastrichtian (Gobbo-Rodrigues 2001; Santucci and Bertini 2001; Batezelli et al. 2003), Upper Cretaceous.

Found among a fine-grained rose to reddish sandstone, with contribution of a siltic matrix. The sediment directly associated with the specimen was extremely hard and resistant.

Diagnosis for the species One small antorbital fenestra, elliptical, with main axis dorsoventrally oriented; jugal anterior ramus slender, enlarged only on the contact with the maxilla; sickle-like medial process present on the ventral surface of the anterior jugal ramus; prefrontal descending process narrow, elliptical in cross-section; longitudinal sagittal sulcation over the palate surface; palatine rami of the maxillae straight, inclined; palate surface concave, ‘V-shaped’ in cross-section; quadrate robust, with triangular crosssection; stout ventral quadrate ridge directed mesially; quadrate ridge reaching from the distal end of the quadrate to laterosphenoid, then anteroventrally to the pterygoid; pterygoid–basisphenoid surface plain and straight at their contact; pterygoid depression lateral to the choanae; pterygoid descending process well developed, robust and straight, directed posteroventrally 40–50 degrees; premaxillae with one pair of teeth; premaxillary teeth caniniform, without lateral compression; premaxilla ventral ramus extends slightly to the first maxillary teeth; first pair of maxillary teeth small, molariform; maxillary pairs 1–7 all molariform with strong oblique implantation (45–90 degrees of paramesial rotation, in respect to the maxillary alveolar surface); maxillary dental series forming a concave line in ventral view, concavity laterally and anteriorly oriented; symphysis high throughout the entire length, anterior section narrow in cross-section, posterior section extremely wide; mandibular rami higher than symphysis, strongly diverging backwards to the mandibular fenestra, but parallel at the mandibular fenestra; well developed coronoid process; nine mandibulary teeth, pairs 1–5 on the symphysis, 6–9 on the diverging mandibular rami; at least mandibulary teeth pairs 2–3 incisiform, pairs 4–9 molariform; mandibulary pairs 5

Historical Biology 109 and posterior with a strong oblique implantation (45–90 degrees of paramesial rotation, in respect to the alveolar surface of the mandible); mandibular pairs 1–5 increase in size, pairs 5–9 decrease in size, with opposing teeth in the maxilla following the same pattern.

Description of Sphagesaurus montealtensis Rostrum The premaxillae are high, directing onto the nasals.

The area is not completely preserved, but premaxillae should not contact each other above the naris. The two premaxillae meet under the nasal opening, where the area is partially damaged. The surface formed by the premaxillae below the naris is shallow and could not support teeth under the naris. The external naris is single, terminal, with no evidence of an internarial bar or internasal septum. The maxillae are also high and well preserved on both sides (Figures 1 and 2). There is prominent maxillary longitudinal ridge that separates the dorsal region of the maxilla from the alveolar margin (Figures 1 and 3). The maxilla does not take part of the orbit, due to a lachrymal – jugal contact.

A slender posteroventral process of the maxilla

is present, as seen in lateral and ventral views (Figures 5 and 6), extending between the anterior end of the jugal and the ectopterygoid, partially separating these elements. The nasals are elongated, widening near the contact with the lachrymals and narrowing posteriorly, near the contact with the frontal. The anterior medial section and the tips of the nasals are missing.

The rostrum ornamentation consists of a striated irregular pattern developing over the nasals, premaxillae and maxillae, but not over the smooth alveolar region. Ornamentation also occurs on the jugals, lachrymals, prefrontals, frontal and the skull table (Figure 1), although less developed.

The lachrymals contact the nasals, isolating the prefrontal from the maxillae. A dorsolateral lachrymal crest develops lateral to the contact with the prefrontal, directed anteroposteriorly. A jugal–maxillary crest is present below and anterior to the orbit. The jugal– maxillary and lachrymal crests delimit the antorbital fossa ( ¼ preorbital depression), which is triangular and shallow. In the antorbital fossa, between maxilla and lachrymal, there is a small but well developed antorbital fenestra, its main axis steep, almost vertical. The structure can be seen on both sides of the specimen and, although the left fenestra is not perfectly preserved, it has the same

Figure 5. MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype: detail of the antorbital fenestra and the lachrymal region, present both on the right (a) and left (b) sides of the specimen. Note that the morphology is the same in both sides and the main axis of the fenestra is mostly vertical.

110 M.B. de Andrade and R.J. Bertini morphology as the right fenestra (Figure 5). The lachrymal takes part in the posterior to the dorsal borders of the antorbital fenestra. The lachrymal ventral end exceeds the antorbital fenestra, reaching the jugal anterior ramus at the anteroventral border of the orbit. The right prefrontal can be identified, although sutures are poorly preserved. It is small, rhomboid and mostly flat. The prefrontal contacts the lachrymal (laterally), nasals (anteriorlly) and the frontal (medially). At the triple suture between the right prefrontal lachrymal and nasal there is a small damaged area, where the bony surface was not preserved. The right frontal is oriented parallel to the sagittal plane and its anteriormost border is locate anterior to the nasal-frontal medial contact. The left frontal is preserved, but sutures are not identifiable. The descending process of the prefrontal (prefrontal pillar) is damaged on both sides of MPMA 15-001/90, and does not hint whether it maintained contact with an ascending palatine process.

The remaining descending process, near to its origin inside the orbit, only reveals that the process was narrow, with a transversely elliptical cross-section (Figure 6).

Skull roof The frontal is wide and short. The anterior border contacts the nasals in an irregular suture located slightly

anterior to the orbits, at about the same level as the antorbital fenestrae. The posterior margin barely contacts the external border of the supratemporal fenestrae, in an arched line. Although the interfrontal suture is absent in adult fossil mesoeucrocodilians, it may be present in younger specimens, such as in Comahuesuchus (Bonaparte 1991; Martinelli 2003) or in Mariliasuchus (Carvalho and Bertini 1999; Andrade 2005; Vasconcellos and Carvalho 2005, 2006; Zaher et al. 2006). There is no interfrontal suture in MPMA 15-001/90, and the structure is solid, suggesting the specimen was either a mature specimen, or at least a subadult. Along the midline there is a discrete, low and robust sagittal ridge, along the entire frontal length (Figure 1).

The parietal is a single structure, long and not wide, building the internal border of the supratemporal fenestrae. It bears a sagittal crest, but its morphology cannot be accurately determined due to preservation.

Nevertheless, the sagittal crest is low between the supratemporal fenestrae and does not differ significantly from the simple narrow surface found in several notosuchians. In the distal end of the parietal surface, the crest widens and becomes higher, forming a crownlike structure. Only Notosuchus terrestris seems to have

Figure 6. Lateral detail of the orbit to evidence the remains of the descending prefrontal process. (a) In MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype, the structure is indicated by a white pointer. (b) The prefrontal pillars of Melanosuchus niger (Eusuchia, Alligatoridae), for comparison. Scale bar Âź 20 mm.

Historical Biology 111 a similar crown-like structure, though in this species the posterior parietal surface is levelled with the skull table and usually is not described as sagittal crest. On the distal end, the right postorbital is well preserved, while the left is only partially retained. In the anterolateral border of the right postorbital there is a posterior palpebral, projecting over the posterodorsal margin of the orbit. The palpebral is robust, thick and curved anteriorly. The palpebral seems to be fused to the postorbital, which may be an artificial preservational feature. The postorbital only contacts the parietal inside the supratemporal fenestra, not on the skull table. Only the right squamosal is preserved. It contacts the postorbital in the lateral border of the supratemporal fenestra and the parietal on the posterior border. The ventral and dorsal squamosal rami are subequal in size, so the ventral ramus and the suture with the quadrate cannot be seen in dorsal view.

The right supratemporal fenestra is complete, while only part of the left one remains. It has a wide external border, elliptic in shape. The internal border is more rounded and much smaller than the external border, indicating that the physiological cross-sectional area

of the muscle associated with the supratemporal fenestra was reasonably small, although the area for muscle attachment was extensive. The main axis of the external border is mostly anteroposteriorly oriented and the fenestrae are parallel, as in Notosuchus. The posterior section of the fenestrae is expanded medially (Figure 1).

Temporal region Only the right jugal is preserved (Figure 1). It is mostly gracile, contacting the maxilla (anterior ramus), postorbital (ascending ramus), quadratojugal (posterior ramus) and ectopterygoid (central body). Both the anterior and posterior rami are slender and subequal in size, with a subcircular cross-section. The anterior ramus only widens at the contact with the maxilla, developing more dorsally (reaching the lachrymal) than ventrally, due to the presence of the maxillary posteroventral process. On the anterior part of its contact with the ectopterygoid, the jugal forms a small but evident sicklelike medial process (Figures 5 and 6). On the anterior ramus of the right jugal there is a clearly preserved

Figure 7. Jugal foramen in Sphagesaurus. (a) MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype. (b) RCL-100 cast for S. huenei. Note that the RCL-100 specimen is transversely broken on the region of the jugal foramen, therefore hindering its identification. Scale bar Âź 20 mm.

112 M.B. de Andrade and R.J. Bertini neurovascular foramen, close to the contact with the maxilla. The foramen is positioned on the ventrolateral surface of the ramus and opens anteriorly (Figure 7).

The posterior jugal ramus delimits the ventral border of the laterotemporal fenestra. Its posteriormost end only slightly exceeds the fenestra. The ascending jugal ramus comprises part of the postorbital bar, along with the ventral postorbital ramus. As the sutures in this region are not well preserved, it is difficult to determine if there is a contact between the postorbital ventral process and the

ectopterygoid. There is a gradual change from the postorbital body to the postorbital bar and then to the jugal body, making poor limits for this structure. The proximal end of the postorbital bar (postorbital ventral ramus) is located on the anteroventral section of the postorbital, well apart from the dorsal border of the postorbital.

The postorbital bar itself is inclined posteromedially, robust, short and straight, with a subcircular cross-section.

Like most of the structures on the posterior section of the skull, only the right quadratojugal remains intact.

Figure 8. MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype. (a) Palate surface. (b) Detail of the palate, shown as to evidence the ‘V’ cross-section. (c) Detail of the rostrum, showing the premaxilla –maxilla suture. Note the existence of sediment over the surface of the bone and teeth, the molariform morphology of the first maxillary teeth and its slightly oblique implantation.

Historical Biology 113 The quadratojugal contacts the postorbital posterior to the postorbital bar, but it is difficult to determine the actual sutures, as well as the contact with the quadrate.

On the other hand, the suture with the jugal is present on the posteroventral border of the laterotemporal fenestra.

It is not certain if a projection of the quadratojugal over the jugal is present, as the laterotemporal fenestra corners are more round than acute, but if this is the case, such a projection is incipient, at best. The laterotemporal fenestra is mostly triangular, anteroposteriorly elongated and exposed in both lateral and dorsal views (Figure 1).

The laterotemporal and the supratemporal fenestrae are subequal in size, considering the external border of the supratemporal fenestra. The laterotemporal fenestra is much more ample than the internal border of the supratemporal fenestra though, and must have supported a bigger muscle volume.

Palate and choanae The secondary bony palate is formed by the ventral rami of the premaxillae and the maxillae, along with the palatines. The anteriormost region of the premaxillae palatine rami is obliterated by rock matrix and cannot be accessed. The premaxillae – maxillae suture on the palate is completely closed, without the development of a nasooral fenestra at this position (Figure 8). There are no traces of maxillo-palatine fenestrae. Palatines constitute at least the posterior section of the floor of the nasopharyngeal channel. Palatines develop mostly between the suborbital fenestrae. The anterior borders of the palatines do not extend anterior to the suborbital fenestrae, as in neosuchians. The suture rather follows a pattern common in notosuchians, where the maxilla-palatine contact is located in a more posterior position, mesially to the anterior borders of the suborbital fenestrae. Posterolateral divergent projections originate from the main body of the palatines, reaching the pterygoid and forming the entire palatine bar. The distal end of the left palatine process is missing, as is the left ectopterygoid. The posterolateral process from the right palatine clearly broadens at the contact with the pterygoid surface. This creates a long spatulated end. The palatine also contacts the ectopterygoid posteromedial end at the posterior border of the suborbital fenestra. This prevents the pterygoid from taking part of the suborbital fenestra and also excludes the ectopterygoid from the palatine bar. The posterolateral divergent processes of the palatines also form the anterolateral border of the choanae, while the posterior border is formed by the pterygoids.

The right ectopterygoid is preserved. It is robust and forms the lateral border of the suborbital fenestra.

The anterior ramus extends at least to the anteriormost border of the suborbital fenestra, contacting the maxilla.

The anterior ectopterygoid ramus only contacts the jugal laterally, due to the presence of the posteroventral maxillary process. The ectopterygoid posterior ramus contacts at least the pterygoid and the palatine, not reaching the anterolateral border of the choanae.

The choanae form a large opening posterior to the nasopharyngeal duct, with an overall triangular shape.

It includes a robust interchoanal septum, which is wider at the middle, but tapers anteriorly. It posterior end is only slightly narrower than the middle section. The ventral border of the septum is subcircular in cross-section (Figures 1 and 9) (Andrade et al. 2006). The pterygoids seem to be fused, as no sagittal suture is visible. The right lateroventral wing is preserved, showing that it was well developed, robust and thick (Figure 9). In lateral view, the wing is straight and inclined posteroventrally at 40– 50 degrees. Over the ventral surface of the pterygoids, close to the posterior border of the internal naris, there is a parachoanal fossa ( Ÿ pterygoid depression; Andrade et al. 2006). This must have been paired, but only the right one is preserved. It is circular and exposed ventromedially, almost entirely enclosed by the pterygoid, only with a small contribution from the posterolateral process of the palatine (Figure 9). The parachoanal fossae are probably fenestrae, but it was not possible to determine if they truly perforate the pterygoid, as the structure is filled with sediment (Andrade et al. 2006).

Basicranium and quadrate Both quadrates are incomplete, although the right one is almost completely preserved. Both mandibular condyles are missing. Proximal and distal areas of the quadrate ventral surface and proximal dorsal surface remain covered by matrix, as well as the laterosphenoids, and some features could not be observed (e.g. nerve passages). The distal end of the right quadrate is missing, revealing the internal structure of the bone. Although it is strong and robust, the quadrate is highly pneumatic, with the development of internal cavities (Figure 10).

The broken end of the quadrate reveals a triangular cross-section on the distal section, with rhomboid margins. The anteroventral corner constitutes a ridge, which runs from the lateral, towards the pterygoids.

The quadrate is strongly inclined, as seen in lateral view (Figure 1). The position of the quadrate condyle can be estimated, and should be lower and posterior relative to the basioccipital (Figure 3) and the occipital condyle, though it is not possible to state if it is positioned below or in the same level of the dental series.

Not much of the basisphenoid is preserved in MPMA 15-001/90 (Figures 1, 2 and 10). It is broadly exposed, forming a flat surface facing more ventrally than posteriorly. The basisphenoid has an indented suture with the pterygoid anteriorly, and a small irregular suture

114 M.B. de Andrade and R.J. Bertini

Figure 9. Pterygoid and perichoanal morphology from MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype. (a) Internal lateral detail of the pterygoid wing, showing the parachoanal fossa. (b) External lateral view of the pterygoid wing.

with the quadrate laterally. The laterosphenoids could not be identified, as the braincase is mostly covered by sandstone matrix.

Occipital region The occipital region is only partially preserved, and many details are missing (Figure 3). The entire surface probably supported strong muscles, due to its general concave construction. Nevertheless, the squamosals dorsal surface does not overhang the occipital surface.

The supraoccipital is poorly preserved, including the area where a supraoccipital crest would develop. Lateral to the supraoccipital, the occipital surface of squamosal is proportionally well developed and contributes significantely to the occipital surface. The surface is smooth and concave, facing posterodorsally. The right exoccipital ( Âź otoccipital) is mostly preserved, although the posterolateral end is missing. It has a broad and wide surface, forming two planes set apart by a horizontal ridge (occipital crest). The upper plane faces posterodorsally, whereas the lower one faces posteroventrally. The exoccipital distal end is curved posteromedially, with a strongly concave surface in both planes. The supraoccipital, exoccipital and the squamosal occipital surfaces are curved mostly in the

horizontal plane, giving them a concave shape. Most of the basioccipital is missing, including the occipital condyle and the actual limits of the foramen magnum.

The remaining surface shows a smooth suture with the quadrate, and is wider than the basisphenoid.

The presence of a few openings over the exoccipital surface may represent passage for nerves, although preservation does not allow either a specific interpretation of each structure, or even their confidant characterisation as foramina (Figure 10).

Mandible The mandible of MPMA 15-001/90 is robust and ‘Y-shaped’ in general outline (Figure 3), with a long symphysis and high divergent mandibular rami.

The symphysis is well preserved, extending at least to the fifth mandibular tooth. The symphysis lacks only the anteriormost tip, which supported the first pair of teeth.

Pairs 2 to 5 are preserved on the symphysis, as posterior teeth on the mandibular rami. They are disposed in the symphysis as a sagittal arrangement of closely paired teeth. On its anteriormost portion, the symphysis has a subcircular cross-section and is slightly higher than wider. The middle and posterior symphyseal sections become progressively wider, but not deeper, so the cross

Historical Biology 115 section of these regions would be elliptical and proportionally flat. The symphysis is mostly horizontal, but the anterior section shows a small dorsal inclination anteriorly. The right mandibular ramus is mostly preserved, showing a well developed coronoid process.

The tooth-bearing portions are highly divergent (over 45 degreees) and contain four teeth. At the middle section, where the mandibular fenestra is partially preserved, the right ramus is only slightly divergent from the sagittal plane, and both rami would be almost parallel.

The mandibular fenestra is only partially preserved on the right ramus, and likely was elliptical or elongated.

The dentary supports most teeth, but the splenial apparently forms the medial edges of the posteriormost alveoli for teeth pairs 6– 9. The splenials meet on the posterior section of the symphysis and this contact is slightly exposed in ventral view, with no projection over the suture of the dentaries. Only the anterior section of the right angular was preserved. It barely projects over the dentary beyond the maxillary fenestra and is mostly exposed on the lateral part of the mandible. The right surangular is also preserved only on its anterior portion (Figure 11), and is divided into internal and external rami. Both rami are subequal in size and well developed.

The external ramus seems to be bilobate, projecting over the dentary and beyond the anterior border of the mandibular fenestra. The internal ramus is single, projecting over the splenial, but not beyond the anterior border of the mandibulary fenestra. No part of the articular is preserved.

Figure 10. Quadrate and basicranium from MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype, seen in lateroventral view. Note the pneumatic structure of the quadrate and its triangular cross-section.

Alveolar margin and dentition The alveolar margin (premaxillae and maxillae) is evidently differentiated, with lateroventral inclination, due to the prominent maxillary longitudinal ridge.

The mandibular alveolar margin is inclined laterodorsally. Only a few neurovascular foramina are present over the alveolar margin (premaxillae, maxillae and mandible). A small foramen is preserved at the premaxillary – maxillary suture on the right side of the specimen, while the corresponding area on the left side is not preserved. Five maxillary foramina are easily identifiable on the right maxilla, only four of them preserved on the left element. The first three pairs correspond to the first two pairs of maxillary teeth.

The fourth foramen (only preserved on the right maxilla) is located above the fourth maxillary tooth and the fifth pair, over the fifth tooth. Between the third and fourth foramina there is a considerable distance where there is no evidence of other neurovascular structure. This gap could be interpreted as an artefact of preservation, but the same pattern can be seen also on the left side (Figure 12).

There seems to be at least five pairs of neurovascular foramina on the mandible, though preservation hinders a definite identification. On the left side, a small foramen is located below the mandibular teeth pairs 2 – 3, but the structure is absent from the right side. Below the third mandibulary teeth, there is an elongated damaged area on each side of the specimen, which probably correspond to two foramina on each side. Below the fourth mandibular tooth there are other two small foramina, reasonably preserved on both sides of the specimen. A single additional foramen may be present below the fifth tooth, and also below the sixth. The presence of the last two pairs is uncertain though, due to preservation problems.

Furthermore, foramina may have occurred anteriorly, below the first mandibulary tooth, but the area is not preserved. In all cases, neurovascular foramina are positioned distant to the alveoli.

The dentition bears several characters previously used to diagnose Sphagesaurus, and some distinctive characters for S. montealtensis. Only one pair of hypertrophied caniniform teeth is present on the premaxilla. Just the right premaxillary caniniform is present, and its crown was mostly eroded. Seven pairs of teeth were present on the maxilla (Figure 13), and the series shows an arched disposition in ventral view. The maxillary teeth increase in size from pairs 1–3, and then decrease to the end of the series. Most of them are well preserved, apart from the seventh from the right series. The first pair of maxillary teeth in MPMA 15-001/90 is very small, but instead of creating a functional diastema, the corresponding pair of dentary teeth is well developed and occludes with this first pair of maxillary ones (Figure 13).

116 M.B. de Andrade and R.J. Bertini There are eight preserved pairs of mandibular teeth.

An anterior pair probably preceded them, but the tip of the mandible is broken. Nevertheless, the anterior left tip of the mandible shows evidence of rooting. The second pair is procumbent, proportionally small and lacks any surface ornamentation. The third right tooth, a robust incisiform, is partially worn out on its labial surface. The same occur with the left fourth tooth, a molariform. The remaining teeth are mostly preserved, but still partially covered by extremely resistant rock matrix, which prevents detailed examination. The presence of rock matrix also prevented examination of the alveoli, although they were probably well defined and separated from each other.

Ornamentation can barely be identified in most elements. There is no thick enamel layer over the teeth, where the teeth are exposed. It is not clear if this is a distinctive character or related to preservation. A distal carina can be identified in most maxillary teeth, while the mandibular pairs 5 to 8 also show a similar carina located on the anterolabial surface of the teeth. In both cases, carinae are poorly preserved. It is not possible to identify the actual number and morphology of denticles. Additional basiapical keels or striae are also poorly preserved, and identifiable only in a few posterior teeth (Figure 13).

Morphological comparison between MPMA 15001/90 and RCL-100 Differences between the two most complete specimens of Sphagesaurus indicate the characters distinguishing S. huenei (RCL-100) and S. montealtensis (MPMA

15-001/90). The most strikingly difference between these specimens is the presence of a small antorbital fenestra on both sides of MPMA 15-001/90 (Figure 5). This fenestra is somewhat different from what can be found in many mesoeucrocodilians, as its main axis is mostly dorsoventrally oriented. RCL-100 has no antorbital fenestra. Both have the antorbital fossa, although in RCL-100 the lachrymal crest that composes the upper limits for this structure is not as evident as in MPMA 15-001/90, probably due to problems of preservation. A maxillary longitudinal ridge separating the dorsal region of the maxilla from its ventral ( ¼ alveolar) margin is also present in both species.

In S. montealtensis the ridge is much more prominent and the alveolar margin faces ventrolaterally, while in S. huenei this surface is mostly vertical, facing laterally. One additional distinction between both species is the presence of an extended posteroventral process of the maxilla of MPMA 15-001/90. Ultimately, this process is related to the poor ventral expansion of the anterior end of the jugal and partially prevents the jugal–ectopterygoid contact. In S. huenei, the process exists but is discrete, and the jugal contacts the ectopterygoid closer to the last maxillary alveolus.

The jugal anterior ramus of MPMA 15-001/90 is not as dorsoventrally high as that of RCL-100, and enlarges only at the very contact with the maxilla. The jugal–maxillary ridge is more prominent and well defined in S. huenei, especially in the anteriormost section of the jugal. At this point, the ridge overhangs the ventral portion of the jugal in RCL-100 (Pol 2003), which is not evident in MPMA 15001/90. Nevertheless, this difference in the morphology of the ridge may be attributed to ontogenetic development, as RCL-100 is larger than MPMA 15-001/90. A sickle-like

Figure 11. Coronoid process and surangular from MPMA 15-001/90 Sphagesaurus montealtensis sp. nov., holotype. (a) Dorsolateral view of the external furcated ramus of the surangular, progressing over the dentary. (b) Internal lateral view, showing the surangular internal ramus, progressing over the splenial. Note the internal outline of the mandibular fenestra, and the participation of the splenial on the back of the mandibular posterior alveoli. Scale bar ¼ 20 mm.

Historical Biology 117 medial process is present on the right jugal of MPMA 15001/90, right below the jugal foramen. This process is completely absent in S. huenei. The jugal foramen found in MPMA 15-001/90 is not exclusive to S. montealtensis or even Sphagesaurus. It was not previously registered in RCL-100 because the transverse fracture of the specimen broke the jugal on the exact point where the foramen is located. Still the specimen shows the structure, which has part of its surface eroded and enlarged. Comparison with MPMA 15-001/90 allows the unequivocal identification of this foramen for Sphagesaurus (Figure 7). The observation of this feature is of special importance, as it was previously considered to be an exclusive synapomorphy of Comahuesuchus and Mariliasuchus (Zaher et al. 2006).

In RCL-100 the quadrate is gracile, with a thin ventral crest ( ¼ quadrate medial ridge; Pol 2003) that extends medially from the quadrate condyle, along its entire posteroventral margin. This crest clearly separates the posterior and the ventral surfaces of the quadrate. On the other hand, the quadrate of MPMA 15-001/90 is robust, and the medial ridge is stout and enlarged, hardly resembling a crest. A cross-section of RCL-100 quadrate would reveal a star-like outline, while the quadrate cross-section for MPMA 15-001/90 is triangular, with rhomboid tips (Figure 10). Furthermore, in RCL-100 the crests of both quadrates converge, meeting over the basioccipital surface. On the other hand, in MPMA 15-001/90 the ridges turn anteriorly on the proximal end of the quadrate, projecting anteriorly, towards the pterygoid (Figures 1 and 10). Thus, the

proximal ends of the quadrate medial ridges are parallel and do not meet. Also, the surface from the pterygoid to the basioccipital is mostly flat in MPMA 15-001/90, whereas in RCL-100 this surface is mostly concave. Apart from these differences, in both species the exoccipital is excluded from the ventral surface of the quadrate by the quadrate medial ridge and the extensive basioccipital–quadrate contact.

The evidence for a parietal crest between the supratemporal fenestrae may be misleading in both specimens. MPMA 15-001/90 has rock matrix over the crest and the exact morphology cannot be properly defined. RCL-100, on the other hand, seems not to show a parietal crest or its remains, but the specimen had suffered considerable damage on the posterior region of the skull. A comparison with MPMA 15-001/90 suggests that RCL-100 was subject to significant dorsoventral compression in the parietal region, and it is possible that a parietal crest could be present in Sphagesaurus huenei, although probably eroded from RCL-100.

Both specimens differ in the naso-oral fenestra, as the structure is not present on the premaxilla –maxilla suture of the MPMA 15-001/90 palate, meaning that it must be either absent or positioned anteriorly to this suture (Figure 8). As previously described by Pol (2003), RCL100 has a naso-oral fenestra where the premaxilla and maxilla meet. This fenestra has a general rhomboid or lozenge shape, rather than a more common subcircular to elliptical shape found in most Mesoeucrocodylia.

In RCL-100 it is also elongated, extending posteriorly

Figure 12. Neurovascular foramina on the alveolar margins of Sphagesaurus montealtensis sp. novum. In the lateral view of the right side of the specimen, where structures are mostly preserved (right of the figure). In the lateral view of the left side, where preservation is poor (left of the figure). Note the foramen on the premaxilla – maxilla suture preserved on the right side, the disposition and number of maxillary foramina and the occurrence of a gap in the distribution of foramina. Scale bar ¼ 20 mm.

118 M.B. de Andrade and R.J. Bertini

Figure 13. Occlusion on Sphagesaurus montealtensis sp. novum. (a) Anterolateral view, showing main aspects of the dentition. (b) Ventrolateral view, showing the second mandibulary incisiform and the first maxillary teeth. (c) The broken tip of the mandible, exposed as a cross-section. Note that the third mandibulary tooth is badly worn. Note also the rostrum ornamentation and the presence of a jugal foramen. Scale bar ¼ 10 mm.

between the second pair of upper teeth. It is shallow and seemed to be filled with sediment. Although in MPMA 15-001/90 the premaxilla – maxilla suture is covered by matrix, it seems reasonably safe to state that does not have a similar rhomboid fenestra on the same position, extending posteriorly between the second pair of upper teeth. The palate of MPMA 15-001/90 has a strongly ornamented surface, with a main sagittal sulcation (Figure 8), creating an ‘arched’ structure. The rostrum

would show that the palate has a ‘V-shape’, in crosssection. These are all absent in RCL-100, and this shows a flat unornamented palate.

A comparison between choanal and perichoanal structures in both specimens is difficult, as RCL-100 has this region poorly preserved. The parachoanal fossae of MPMA 15-001/90 are likely to be homologous to what was previously identified a pterygoid fossa in RCL-100 (Pol 2003, p. 820). This structure was erroneously

Historical Biology 119 labelled as pterygoid foramen (Pol 2003, p. 819, Figure 1B), being the terminology used in the description the correct one (D. Pol, personal communication). Nevertheless, the presence of a fenestra inside the fossa cannot be corroborated either in RCL-100 or in MPMA 15001/90 due to the incompleteness of the first and the presence of sandstone matrix in the second.

In the alveolar margin of both specimens, there is a reasonably low number of neurovascular foramina, which roughly corresponds to the number of teeth for the maxillary and mandibulary series. This is present in other notosuchians, though. Furthermore, there seems to be a gap on the occurrence of foramina at the medial maxillary alveolar surface. This gap is located above the third – fourth maxillary teeth and separates an anterior series of foramina from a posterior series. In MPMA 15001/90 these series include respectively three and two foramina (Figure 12). This gap seems to be present at least on the right side of the RCL-100 specimen, although with four anterior and three posterior foramina (Pol 2003, p. 819; Figure 1A). Despite differences in number of foramina, the gap seems to be consistent.

Nevertheless, preservation may hinder the identification of such structures and the description of further specimens is important to corroborate the information and support the evidence. Such a gap was never previously reported, and may constitute an additional distinctive character for Sphagesaurus. Nevertheless, it should be noticed that while a wider range of Sphagesaurus and related forms is described, the distribution and the absolute number of foramina should not be used to support the distinction or even the erection of species within the group, as: (i) the number and distribution of foramina can be easily biased by preservation; (ii) the knowledge on the intraspecific variability of this characteristic in fossil crocodilomorphs is poor, at best; (iii) the influence of ontogenesis over the number of foraminae is unknown.

The premaxillary and maxillary dentitions are similar in most aspects. Sphagesaurus huenei seems to have tooth elements more developed than those of MPMA 15-001/90, including the proportional size of teeth, the development of ridges and the lack of an enamel layer. Still, we prefer not to assume a more gracile dentition as a diagnostic character for Sphagesaurus montealtensis, as a number of factors could be involved (e.g. ontogeny, preservation).

Regarding its distinctive morphology, MPMA 15-001/90 has a maxillary dental series positioned in a concave line (the concavity is labial). This, combined with the prominent maxillary longitudinal ridge that separates the dorsal from the ventral regions of the maxilla (that latter of which faces ventrolaterally), allows the lateral surface of the maxilla to be seen in ventral view (Figures 1 and 8).

In RCL-100, the dental series forms a straight line and the alveolar margin is mostly vertical, so the lateral surfaces

of the maxillae cannot be seen in ventral view (as in Pol 2003; Figures 1–2, pp. 819–820).

The preserved section of the symphysis of RCL-100 seems to be similar when compared to the preserved mandible tip of MPMA 15-001/90. RCL-100 preserves caniniform teeth with acute apices (pairs 1–3), while MPMA 15-001/90 have the corresponding teeth with blunt apices and can be defined as incisiform teeth. The first oblique element is the fifth pair (fourth preserved, in MPMA 15-001/90), preserved in RCL-100 as two fragments of crown tips attached to the palate (Pol 2003).

Furthermore, the second pair in MPMA 15-001/90 is smaller, more gracile and lacks any surface ornamentation.

Nevertheless, both Sphagesaurus species had the first three pairs procumbent, the fourth pair molariform and the fifth pair with oblique implantation (Figure 8).

The disposition of symphyseal teeth is also characteristic of the genus, as each tooth is positioned closer to its pair on the other dentary than to the next teeth on the same dentary. This creates a battery of five pairs of teeth, arranged in tandem. Furthermore, the mandibular pairs 2 – 3 show eroded labial surfaces both in RCL-100 and MPMA 15-001/90, indicating they were probably used in at least a similar way.

Morphological comparisons reveal great similarities between Sphagesaurus huenei and the new species, as well as distinctive characters. Even though, two important characters could not be compared.

The maxillary posterior wall, clearly present in S. huenei, could not be identified in MPMA 15-001/90 due to the extensive presence of sandstone matrix. Also, it was not clear if the differences present on the teeth of both species were related to species differentiation or preservation. In a conservative approach, we prefer to maintain both characteristics as diagnostic for S. huenei.

Although most of the comparative overview presented here was based on MPMA 15-001/90 and RCL-100, a few morphological aspects of DGM 1411-R are known (Kelner et al. 1995; Pol 2003) and can be included. Though only a part of the rostrum and symphysis is preserved, the information provided by the specimen indicates that DGM 1411-R fit into the definition hereby proposed for Sphagesaurus huenei.

The general characteristics, the development of the premaxilla, the morphology of anterior symphyseal teeth and the lack of an antorbital fenestra are common with RCL-100, rather than with MPMA 15-001/90. Nevertheless, it should be noticed that the identification of DGM 1411-R as S. huenei is still provisory, as there are morphologic features that await proper description.

The information on the specimen is important and shall contribute to the knowledge on the morphology of teeth, maxillary foramina and ontogenetic development in Sphagesaurus.

120 M.B. de Andrade and R.J. Bertini Phylogenetic analysis Methodology applied A preliminary phylogenetic analysis (Hennig 1950, 1966) was done to verify the inferred sister-taxon relationship between Sphagesaurus montealtensis and S. huenei. The analysis used PAUP for Windows 4.0b10 (Swofford 2002), with the heuristic algorithm (1,00,000 replicates; TBR swapping; random addition sequence of 100 replicates). The data was based on Andrade (2005), composed of both new and previously used characters (Benton and Clark 1988; Clark 1994; Gomani 1997; Buckley et al. 2000; Ortega et al. 2000; Martinelli 2003; Pol 2003; Sereno et al. 2003) (Appendix C). The matrix (183 characters; 21 terminals) (Appendix D) used Sphenosuchia and Protosuchia as consecutive outgroups.

Terminals included 13 species of Notosuchia (sensu Gasparini 1971) and a single terminal for Baurusuchidae (including Baurusuchus pachecoi Price 1945 and Stratiotosuchus maxhechti Campos et al. 2001). Eusuchians were included (Alligatoridae, Crocodylidae, Bernissartia), as well as Sebecus. Characters were treated as unordered (Fitch 1971) and equally weighted, to avoid a priori assumptions. The automatic collapse of zero-length branches was applied, to avoid the possible grouping of unsupported clades. The analysis used the shortest possible transition for each character, avoiding a priori assumption of either ACCTRAN or DELTRAN optimisation. Bremer support (Bremer 1994) and bootstrap (Felsenstein 1985) were used to evaluate results.

Bremer support ( ¼ branch decay) was calculated with the use of TreeRot.v2 (Sorenson 1999). Bootstrap was calculated by heuristic search for 200 replicates, using random addition sequence (50 replicates).

Results Only a single most parsimonious tree was obtained (Figure 14). Bootstrap provides 50% or higher support for 77.7% of the clades, while branch decay provided support for all clades (no polytomies). Most clades (72.2%) had a bootstrap of 57% or higher and a branch decay of 3 or higher. The majority of the clades (55.5%) had a bootstrap index of 80% or higher and a branch decay of 5 or higher. Nevertheless, four branches (28.6%) showed a weaker support, with bootstrap index lower than 50% and a branch decay no greater than 2.

The analysis strongly indicates a sister-taxon relationship between Sphagesaurus montealtensis and S. huenei (bootstrap ¼ 100%; branch decay ¼ 9), confirming the morpho-anatomical observations. A closest relationship between Sphagesaurus and Notosuchus, proposed originally by Price (1950), did not find support only by the inclusion of Mariliasuchus in this analysis. Nevertheless, a sister–clade relationship between Sphagesauridae and

Notosuchidae (Notosuchus þ Mariliasuchus) found a good support (bootstrap ¼ 80%; branch decay ¼ 5), confirming the general idea originally hypothesised by Price (1950). This analysis corroborates Mariliasuchus amarali as sister-group to Notosuchus (bootstrap ¼ 80%; branch decay ¼ 4), supporting its original classification as a Notosuchidae (as in Carvalho and Bertini 1999; contra Zaher et al. 2006). Comahuesuchus does not figure as the sister group of Mariliasuchus, as proposed by Zaher et al.

(2006), but instead it is presented as the sister clade of Sphagesauridae þ Notosuchidae. Chimaerasuchus does not figure as sister-group to Sphagesauridae, as usually presented by several other works (Pol 2003; Pol and Norell 2004a, 2004b; Pol and Apesteguia 2005; Zaher et al.

2006), but instead it occupies a more basal position in the topology, as Baurusuchidae.

Other notosuchians sharing multicusped teeth crowns constitute a poorly supported clade (bootstrap , 50%; branch decay ¼ 2) related to Uruguaysuchus. Chimaerasuchus was the only mesoeucrocodile, among the species with multicusped teeth, nesting outside this clade. Notosuchia (sensu Sereno et al. 2001) found a strong support (bootstrap ¼ 81%; branch decay ¼ 5), but with the exclusion of Araripesuchus and Anatosuchus, which are often referred as notosuchians (e.g.

Gasparini 1971; Sereno et al. 2001, 2003; Pol and Apesteguia 2005; Zaher et al. 2006).

Araripesuchus shows a strong bond with Neosuchia (sensu Benton and Clark 1988) (bootstrap ¼ 78%; branch decay ¼ 5), as previously presented in the bibliography (Buckley et al. 2000; Ortega et al. 2000; Fiorelli 2005; Fiorelli and Calvo 2005; Turner and Calvo 2005; Turner 2006; contra Sereno et al. 2003; Pol and Norell 2004a, 2004b; Pol and Apesteguia 2005; Zaher et al. 2006).

Anatosuchus finds a reasonable support as sister-group to Araripesuchus þ Neosuchia (bootstrap ¼ 73%; branch decay ¼ 3), as in Martinelli (2003). The sister-taxon relationship between Anatosuchus and Comahuesuchus predicted by Sereno et al. (2003) and followed by Turner and Calvo (2005) and Turner (2006) is not supported, as previously suggested by Martinelli (2003); Andrade et al.

(2006) and Zaher et al. (2006).

Discussion The new species, Sphagesaurus montealtensis, shows morphological features that allow a better understanding of the evolution and diversity of the Mesoeucrocodylia throughout the Cretaceous of South America.

The existence of variation in the presence of an antorbital fenestra within the same genus is unexpected, although this presumably may have happened within Baurusuchus (see Carvalho et al. 2005). Nevertheless, the identification of the same morphology in both sides of the skull of S. montealtensis gives reliability to the present

Historical Biology 121

Figure 14. Preliminary phylogenetic analysis of the notosuchian Mesoeucrocodylia, showing the single most parsimonious tree obtained, where MPMA 15-001/90 appears as the sister-taxon of Sphagesaurus huenei (9; 100%). Main taxonomic groups indicated.

Support indexes shown above (bootstrap) and below (Bremmer decay) each branch. Length ¼ 477 steps (min ¼ 227; max ¼ 948); consistency index ¼ 0.476; retention index ¼ 0.653; rescalonated consistency index ¼ 0.311; homoplasy index ¼ 0.524. Matrix with 21 terminals and 183 characters, unordered and equal weighted; automatic collapse of branches with zero-length applied; shortest transition possible assumed for each node; bootstrap replicates ¼ 200. Taxa examined in Appendix B.

interpretation. Another uncommon characteristic is the presence of a concave and ornamented palate, which may be related to specific feeding habits and food intake.

The battery of symphyseal teeth is unique to Sphagesaurus, and was certainly of functional relevance and directly related to food intake.

The lack of a naso-oral fenestra on the premaxilla – maxilla suture is also important. The rhomboid fenestra should be easily identifiable in S. montealtensis, even with rock matrix, if it had a morphology similar to Sphagesaurus huenei. Among fossil Mesoeucrocodylia, a similar type of naso-oral fenestra seems to be present

122 M.B. de Andrade and R.J. Bertini in Mariliasuchus, as observed in specimens MN –6756 – V, MN – 6298 and MZSP– PV – 51 (Zaher et al. 2006). It is not clear though, if this fenestra is homologous to the typical naso-oral fenestra found in eusuchians and some fossil forms (e.g. Sebecus, Notosuchus). Furthermore, a similar structure was identified in Hamadasuchus as a possible palatine exposure of the vomer (Larsson and Sues 2007), and the problem deserves further investigation.

The internal structure of the quadrate is not usually described in mesoeucrocodiles. Bonaparte (1991) considers the pneumatic structure of the quadrate as a common condition found in Notosuchus and protosuchians. Though the observations refer mostly to the presence of multiple fenestrae closer to the dorsal proximal end of the quadrate, Notosuchus actually possess a highly pneumatic quadrate (as seen in MACN – Pv –N – 23, MACN – Pv –RN –1037, MACN – Pv –RN – 1045 and MLP –64 – IV– 16 –30). Eusuchians, on the other hand, do not show the same pneumatic structure in the distal end of the quadrate. Still, the structure of the quadrate is unknown for most notosuchians and the character should be observed in other taxa before its use in phylogenetic reconstruction.

The jugal neurovascular foramen found in MPMA 15-001/90 allows a brief review of the character, its use in previous works and occurrence in mesoeucrocodiles.

The presence of this structure was first noticed and represented by Martinelli (2003, p. 562, Figure 2C), in the jugal of Comahuesuchus. The structure was introduced as a character by Andrade (2005; character 43). In this unpublished work, the presence of the single enlarged and anteriorly directed jugal foramen is recognised for Comahuesuchus and also described for Mariliasuchus and Sphagesaurus. Furthermore, Andrade (2005) includes a third state in the character, for the existence of ventrally directed small foraminae (usually four, but often three or two) in eusuchians. This information was also used in posterior works (Andrade and Bertini 2005, 2007a; Andrade et al. 2005).

The occurrence of the jugal single foramen was first published as a character by Zaher et al. (2006), but identified only in Mariliasuchus and Comahuesuchus.

Nobre and Carvalho (2006) also report the existence of a jugal foramen in Adamantinasuchus. The occurrence of this character in Sphagesaurus and other mesoeucrocodiles shows that this structure is not restricted to Comahuesuchus and Mariliasuchus, and thus cannot be considered an exclusive synapomorphy of these notosuchians. The jugal foramen may rather provide phylogenetic information supporting a broader, more inclusive group.

MPMA 15-001/90 certainly cannot be understood as an uruguaysuchid, as previously proposed (Bertini 1993; Bertini and Arruda-Campos 1995; Andrade and Bertini

2003). Instead, both morphological and phylogenetic analysis strongly supports an intimate relationship between MPMA 15-001/90 and Sphagesaurus huenei.

They both share several diagnostic and exclusive characteristics: (i) the extreme rotation of teeth crowns in the maxilla and at least the fifth mandibular teeth, as possibly also in posterior elements; (ii) an hypertrophied caniniform in the upper dentition, followed by seven single-cusped molariform teeth; (iii) the symphyseal teeth battery composed of three anterior procumbent teeth, followed by molariform teeth; (iv) parachoanal fossae developing next to the posterior border of the choanae; (v) basioccipital wider than the basisphenoid, contacting the quadrate extensively and isolating the exoccipitals from the quadrate ventral surface. Apart from these, the general morphology of S. huenei and S. montealtensis is consistent throughout (e.g. terminal undivided naris; jugal foramen present, anteriorly oriented; jugal anterior and posterior rami horizontal, aligned; edentulous region in the premaxilla, below the naris; palatines do not extend anterior to the suborbital fenestra, separating the ventral rami of the maxilla).

Differences between MPMA 15-001/90 and Sphagesaurus huenei cannot be attributed to ontogeny, sexual dimorphism or even taphonomic bias, given their nature, number and preservation of the specimen. They are also not expected to be intraspecific (e.g. antorbital fenestra cannot be present or absent). The results indicate that MPMA 15-001/90 represents a different species, closely related to Sphagesaurus huenei and thus assignable to the same family and genus. The new species, Sphagesaurus montealtensis (Figure 15), can be distinguished from S. huenei by the presence of: (i) antorbital fenestra; (ii) ‘V-shaped’ ornamented palate with a sagittal sulcation; (iii) robust quadrate; (iv) procumbent incisiforms on the symphysis, progressing posteriorly to molariforms; (v) further reduction of the premaxilla, which supports only the first tooth of the upper series, the hypertrophied caniniform; (vi) presence of a sickle-like medial process, in the ventral surface of the jugal anterior ramus.

Sphagesaurus montealtensis also reveals characteristics previously unknown for the genus, such as the composition of the suborbital fenestra and the morphology of the supratemporal fenestrae, pterygoids, choanae, mandible and occipital elements. Furthermore, provides evidence for the presence of the parachoanal fossae, the jugal foramen and also a strong suggestion for a consistent gap in the distribution of foramina in the alveolar margin of the maxilla. The new information contributes to the phylogenetic analysis, allowing a more comprehensive understanding of the genus Sphagesaurus and its evolutionary relationships.

Previous works related Sphagesaurus only to Notosuchus (Woodward 1896; Martinelli 2003) and Chimaerasuchus (Pol 1999; Pol 2003; Pol et al. 2004; Pol

Historical Biology 123

Figure 15. Sphagesaurus montealtensis reconstructed. Inferred general aspect of the skull in dorsal (upper left) and lateral (upper right) views, including the mandible, and restoration of complete specimen, as in life (below). Dashed lines represent missing parts and unidentified sutures. Inferred structures, including post-cranial elements, were reconstructed by comparison with taxa with close relationships, as given by the phylogenetic analysis (S. huenei, Mariliasuchus, Notosuchus, Comahuesuchus). Inclusion of vegetal material in the diet based on the morphology of teeth, mandible and supratemporal fenestrae. Scale bar Âź 20 mm.

and Norell 2004a, 2004b; Pol and Apesteguia 2005; Fiorelli and Calvo 2005; Jouve et al. 2006; Zaher et al.

2006). The results presented here partially corroborate the phylogenetic study of Martinelli (2003) for a closer relationship to Notosuchus, as originally proposed by Price (1950), since Mariliasuchus amarali Carvalho and Bertini 1999 was not considered in these studies.

The closest relationship obtained in this analysis occurs with Notosuchidae (Notosuchus Ăž Mariliasuchus; Carvalho and Bertini 1999; Andrade 2005; Andrade and Bertini 2005; contra Zaher et al. 2006). The resemblances with Sphagesauridae are evident, mostly related to dental morphology. This includes the oblique positioning of the crowns (less rotated in Notosuchidae), the teardrop-like cross-section of the maxillary dental series and the way the teeth occlude to each other. The reduction of the premaxilla is also present in Notosuchidae, although not to the same extent. The results of the phylogenetic analysis show that the characteristic maxillo-palatine fenestrae, found in both Notosuchus and Mariliasuchus, are unlikely to be convergent features of these genera (as in Carvalho et al. 2004; Zaher et al.

2006). Maxillo-palatine fenestrae may in fact constitute one of the best diagnostic character for Notosuchidae, due to its very specific morphology (size, distance between fenestrae, position). These fenestrae are absent

in both species of Sphagesaurus, as in all other Mesoeucrocodylia.

There are important similarities with Comahuesuchus and notosuchids. The jugal foramen is present in Sphagesaurus, Comahuesuchus and Mariliasuchus. In all those taxa the nasals extend laterally, showing an almost triangular profile. The internal naris is triangular, with gracile palatine bars clearly defined and the suborbital fenestrae reduced in length (when compared with other mesoeucrocodilians). Baurusuchids are also closely related, sharing the general triangular structure of the choanae, although the palatine bar is much more robust (Andrade et al. 2006). Only sphagesaurids and baurusuchids share the presence of parachoanal fossae, although the presence of these structures is better optimised as a convergence of Sphagesaurus and baurusuchids, in the phylogenetic hypothesis obtained.

In fact, in Stratiotosuchus these fossae are located next to the lateral border of the choanae, while in Sphagesaurus the structure is positioned posterolaterally to the internal naris. The inclination and morphology of the Sphagesaurus pterygoid wing is intermediate between that found among other derived notosuchians (Notosuchus, Mariliasuchus, Comahuesuchus, baurusuchids). The inclination is the same as shown by most mesoeucrocodilians, including eusuchians.

124 M.B. de Andrade and R.J. Bertini In opposition, baurusuchids show an almost vertical, well developed wing, while notosuchids seems to have a smaller, more horizontal structure. Sphagesaurus also resemble notosuchids, as the wing is straight, while in baurusuchids it shows a curved surface. Comahuesuchus seems to follow the same pattern shown by notosuchids, but preservation of specimens may interfere with the morphologic interpretation. All derived notosuchians have the pterygoid excluded from the suborbital fenestrae by a palatine – ectopterygoid contact (Martinelli 2003; Andrade 2005; Andrade et al. 2006). This characteristic clearly separates derived notosuchians from more basal groups (Anatosuchus, Araripesuchus, Uruguaysuchus, Simosuchus). They also show palatines that do not extend anterior to the suborbital fenestrae.

In opposition, palatines that surpass the anterior border or the suborbital fenestrae can be seen in Anatosuchus, Araripesuchus, Sebecus, most eusuchians (Price 1959; Sereno et al. 2003) and seems to be present in Uruguaysuchus (Rusconi 1933). Unfortunately this characteristic is poorly known for basal groups of notosuchians and needs clarification at least for Candidodon and Malawisuchus. Furthermore, derived notosuchians have the occipital surface above the occipital crest facing posterodorsally, while Anatosuchus, Araripesuchus, Sebecus and most eusuchians have this surface vertical, facing posteriorly.

The phylogenetic relationships of Chimaerasuchus paradoxus Wu et al. 1995 pose as a special problem and deserves a closer attention. Its phylogenetic position is still problematic within the derived notosuchians, as the skull of this species is limited to a partial rostrum and hemimandible (Wu et al. 1995; Wu and Sues 1996).

The bizarre dental morphology found in Chimaerasuchus is hardly comparable to any known Mesoeucrocodylia, though Pol (2003) and subsequent works relate Chimaerasuchus and Sphagesaurus intimately. As in Sphagesaurus, Chimaerasuchus shows strong reduction of the premaxilla and its dentition, but the crowns are not rotated. Furthermore, the molariform teeth are single cusped in Sphagesaurus and multicusped in Chimaerasuchus. The dental morphology of the molariform teeth of Chimaerasuchus may be considered similar to the molariform teeth of Malawisuchus, as both have well developed series of cusps arranged in tandem (Wu et al.

1995; Wu and Sues 1996; Gomani 1997). Still, in Chimaerasuchus the crown is wide and flat, while in Malawisuchus the crowns are laterally compressed, with cusps showing a different morphology. The overall similarity between these taxa can be easily the result of convergence. Besides, neither this phylogenetic hypothesis nor other published analysis supports this hypothesis. On the other hand, the morphology of the choanae, pterygoids, suborbital fenestrae and the occipital surface are unknown in Chimaerasuchus, and nasals are only

partially preserved. The discovery of both Sphagesaurus postcranial remains and Chimaerasuchus skull material is important, and shall allow a more comprehensive comparison among these taxa.

The morphologic study of the specimens and the phylogenetic analysis, despite its preliminary character, may also contribute with other problems on the evolution of Mesoeucrocodylia. As previously addressed (Martinelli 2003; Andrade 2005; Andrade et al. 2006), the structure of the occipital, palatal and choanal regions clearly shows that Anatosuchus and Comahuesuchus are not closely related and certainly are not sister– clades (contra Sereno et al. 2003; Turner and Calvo 2005; Turner 2006). Instead, they represent different patterns of skull construction and eventual rostral similarities are more likely to be convergences, which can be explained by common palaeoecological aspects (e.g. composition of diet, foraging mode). This study also does not corroborate Araripesuchus as a Notosuchia (sensu Sereno et al. 2001), as originally proposed by Gasparini (1971). The closer relationship between Araripesuchus and neosuchians has been supported by several previous studies (e.g. Buckley et al. 2000; Ortega et al. 2000; Pol 2003; Pol and Apesteguia 2005; Turner and Calvo 2005; Turner 2006; contra Fiorelli 2005; Fiorelli and Calvo 2005; Zaher et al. 2006). The future recognition of undisputable characters uniting Araripesuchus and notosuchians may contribute to the debate, but a better approach could be the recognition of the four species included in the genus as a lineage of its own, which may either be related to Neosuchia or to Notosuchia.

The grouping of the basal notosuchians (Candidodon, Malawisuchus, Simosuchus and Uruguaysuchus) in a single clade is not common, and Uruguaysuchus and Simosuchus usually appear as more basal lineages (Pol 2003; Pol and Norell 2004a, 2004b; Pol and Apesteguia 2005; Zaher et al. 2006). The results obtained support from Buckley et al. (2000), Turner and Calvo (2005) and Turner (2006), who propose a group with the same general structure (only without Candidodon, which is not present in these analysis). Characters common to these notosuchians are often present in other mesoeucrocodilians (e.g. antorbital fenestra) and the presence of multicusped teeth might be a devious characteristic.

In each case, teeth may represent a special adaptation.

In the case of Candidodon, molariform teeth do not show a strong lateral compression, and seems to be crushing elements (Carvalho 1994). In Malawisuchus the molariform teeth resemble the carnassial teeth of mammals (Clark et al. 1989; Gomani 1997). The teeth of Uruguaysuchus show a greater resemblance with the dental elements of Simosuchus, which in both cases are compressed, multicusped and ‘shovel-like’ (Rusconi 1933; Buckley et al. 2000). Nevertheless, Simosuchus have several important differences from all these species

Historical Biology 125 (e.g. short and wide skull, wide internarial bar; palatines that do not meet, choanal morphology), and is the only crocodilomorph to show a dentition solely composed by multicusped teeth (Buckley et al. 2000).

Currently, there is no evidence supporting any of those species as part of the group of derived notosuchians.

Both Malawisuchus and Simosuchus have been originally referred as members of the Notosuchidae (Gomani 1997; Buckley et al. 2000), but these works do not include most notosuchian species (e.g. Sphagesaurus, Mariliasuchus).

Buckley et al. (2000) seems to ignore the family Uruguaysuchidae, and in the cladogram presented, ‘Node A’ could roughly be interpreted as Notosuchidae.

Nevertheless, the topology shows that Uruguaysuchus and Malawisuchus share a closer relationship with Simosuchus than to Notosuchus (Buckley et al. 2000) and a more sensible approach would be to consider Notosuchidae monospecific and redefine Uruguaysuchidae, with the exclusion of Araripesuchus (according to the given phylogeny). Apart from these problems, definition of the family Notosuchidae is much more precise and these species certainly do not possess most of the diagnostic characters of this clade. Candidodon have already been referred to a clade of its own, the phylotaxon Candidodontidae, but with the inclusion of Mariliasuchus (Carvalho et al. 2004). In this analysis Candidodon figures as sister-taxon to Uruguaysuchus and Candidodontidae would be a junior synonym to Notosuchia (sensu Sereno et al. 2001), due to the position of Mariliasuchus.

Despite divergent results present in the bibliography, it seems reasonable to consider Candidodon Malawisuchus and Simosuchus as related to Uruguaysuchus and provisionally part of the family Uruguaysuchidae, but certainly not to Notosuchidae. Eventually, new data may allow the development of more reliable statements on the evolution and taxonomy of these species.

The taxa here included in the Uruguaysuchidae constitute a South American/ African clade, mostly restricted to pre-Senonian times. The basalmost species, Malawisuchus, comes from Africa. Uruguaysuchus and Candidodon compose a more exclusive clade within the Uruguaysuchidae, both from South America. Simosuchus is the only exception, as it comes from the late Upper Cretaceous (Maastrichtian) of Madagascar. The derived notosuchians are almost exclusively related to Upper Cretaceous of Gondwana. Chimaerasuchus, from the Albian of China, is again an exception. Nevertheless, its biochronological range may be biased and a more accurate definition on the age of the Wulong Formation is still necessary (Andrade 2005; Andrade and Bertini 2007b). The presence of Chimaerasuchus in a Laurasian sedimentary unit can be explained by dispersion.

Furthermore, the group composed by sphagesaurids, notosuchids and Comahuesuchus shows a specific distribution, limited to the Upper Cretaceous of South

America and may have evolved after the complete break up of South America from other Gondwanic landmasses.

This suggest a faunal substitution hypothesis for the evolution of South American notosuchians, where uruguaysuchids and Araripesuchus dominated the landscape during the Lower Cretaceous and the derived notosuchians prevailed in post-Senonian times. While may be absent from South America, at least Araripesuchus and uruguaysuchids survived in other gondwanic landmasses until the Maastrichtian, as in the case of Araripesuchus tsangatsangana and Simosuchus clarki (Buckley et al. 2000; Turner 2006). Nevertheless, because of the preliminary character of this analysis and the lack of key taxa, it is not possible to support this or any other particular palaeobiogeographic model, and the hypothesis presented here is merely speculative. Even so, the specific distribution of sphagesaurids, notosuchids and Comahuesuchus is consistent with the phylogenetic hypothesis obtained.

Conclusions The phylogenetic analysis shows that the new species is the sister-taxon of Sphagesaurus huenei. A few important differences with previously published analysis are also present, regarding the evolution of notosuchians: (i) Chimaerasuchus does not figure as the sister-taxon of Sphagesaurus, occupying a less exclusive relationship with this taxon; (ii) Sphagesaurus, Notosuchus and Mariliasuchus form a group with a similar morphology (e.g. single-cusped teeth; molariforms with rotated crowns); (iii) Comahuesuchus does not figure as the sister-taxon of Mariliasuchus or Anatosuchus; (iv) other basal notosuchians form a group of species with multicusped molariform teeth, consistent with the family Uruguaysuchidae, though this group is poorly supported.

Considering the diversity of results obtained by several studies, including this, the phylogenetic position of Sphagesaurus certainly lies within Notosuchia, more specifically with the derived notosuchians. Despite divergent results, Sphagesaurus is consensually related to Baurusuchidae, Notosuchidae and Comahuesuchidae (Martinelli 2003; Pol 2003; Pol and Norell 2004a, 2004b; Pol and Apesteguia 2005; Jouve et al. 2006; Zaher et al.

2006). As a result of its phylogenetic relationships and its particular morphology, Sphagesaurus must be included in a family of its own. Furthermore, Chimaerasuchus is not to be considered a sphagesaurid, due to the lack of agreement on its phylogenetic position.

More general inferences on crocodilomorph evolution find support with the current bibliography.

The occurrence of a notosuchian group with multicusped molariform teeth have already been introduced by Buckley et al. (2000), Turner and Calvo (2005) and Turner (2006). The group is consistent with the family

126 M.B. de Andrade and R.J. Bertini Uruguaysuchidae. Anatosuchus and Araripesuchus may be either a basal notosuchian or, as suggested by this analysis, basal groups of Mesoeucrocodylia related to Neosuchia (e.g. Buckley et al. 2000; Ortega et al. 2000; Pol 2003; Pol and Apesteguia 2005; Turner 2006; Zaher et al. 2006). Anatosuchus does not share an important group of characters present in the derived notosuchians and cannot be considered as closely related to Comahuesuchus (Martinelli 2003; Andrade et al. 2006).

The morphological characteristics present in MPMA 15-001/90 allow the differentiation from Sphagesaurus huenei and the recognition of a new species, Sphagesaurus montealtensis (Figure 15). Additional information provided by this study, as well as other previous contributions (Pol 2003; Andrade et al. 2006) offer enough evidence for a redefinition of the Family Sphagesauridae Kuhn 1968. Sphagesaurids were terrestrial mesoeucrocodilians from the notosuchian (sensu Sereno et al. 2001) branch, which evolved during the Upper Cretaceous of South America. They had reduced dentition with middle and posterior teeth showing strongly rotated crowns, occluding as opposing triangles (Price 1950; Kuhn 1968; Pol 2003), but also skulls as high as wide with lateral orbits, terminal naris, evident antorbital fossa and a long symphysis (Pol 2003).

Sphagesaurids also probably had palatine posterolateral processes composing the anterolateral border of a triangular choanae, pterygoids excluded from the border of the suborbital fenestrae (Andrade et al. 2006), nine mandibulary teeth, elongated and parallel supratemporal fenestrae, jaw and pterygoid flanges robust, reduced premaxilla, parachoanal fenestrae and strong occipital musculature, as in Sphagesaurus montealtensis. Both species of Sphagesaurus: (i) had eight upper teeth, being the first pair hypertrophied caniniforms, though the first postcaniniform diverge in its placement between the two species; (ii) a symphyseal battery of teeth is present, including pairs 1 –5. Currently, sphagesaurids are only known from the Adamantina Formation (Bauru Group sensu Batezelli et al. 2003), Campanian –Maastrichtian (Upper Cretaceous) of Brazil. The limited palaeobiogeographic and biochronologic distribution of sphagesaurids may reflect the isolation of South America during the Upper Cretaceous and the evolution of the family after the break up of Gondwana. It can also be related to the adaptation of these notosuchians to a specific mode of life in a restricted palaeoenviroment, as suggested by the specialised dentition and robust skull that characterises the group.

Acknowledgements The authors are grateful to A. C. Arruda-Campos (MPMA), for the access to the MPMA 15-001/90 specimen. We would also like to thank A. Kramarz

(MACN), A. W. A. Kellner (MN –UFRJ), D. A. Campos (DNPM), F. Novas (MACN), J. F. Bonaparte (MACN), J. O. Calvo (Proyecto Dino), L. Carvalho (MN –UFRJ), M. Reguero (MLP), S. A. K. de Azevedo (MN –UFRJ), S. Bargo (MLP) and Z. N. B. Gasparini (MLP), for access to materials either under their care or study.

Corrections by Diego Pol (MEF), Alan Turner (AMNH) and Gareth Dyke (UCD) greatly contributed to the development of the manuscript. The authors thanks to Ca´stor Cartelle Guerra (PUC), for important information on the outcrop of the RCL-100 specimen; Jose´ F. Bonaparte (MACN), for valuable suggestions regarding the study of Sphagesaurus and Notosuchus specimens; Diego Pol (MEF), for his suggestions, criticism and generous comments on the morphology of Sphagesaurus; A. E. P. Pinheiro (NEPV-IGCEUNESP), for useful comments on the morphology of baurusuchids. Credit also is due to: Michael J. Benton (DES-University of Bristol), for his helpful revision of the manuscript; Lilia M. D. Bertini (NEPV-IGCEUNESP), for preparation of the specimen; Felipe A. Elias (NEPV-IGCE-UNESP), for the pictorial reconstruction of S. montealtensis; Simon Powell (DES-University of Bristol), for valuable directions on image treatment.

Financial support for this study was provided by the Coordenac¸a˜o de Aperfeic¸oamento de Pessoal de Nı´vel Superior (CAPES), Brazil. MBA is currently supported by the Conselho Nacional de Desenvolvimento Cientifico e Tecnolo´gico (CNPq – Proc. no. 200391/2006 –8), Brazil.

Note added in proof stage After this MS was accepted and already in proof stage, the work

by Marinho and Carvalho (2007; reference below) was published as a book chapter, regarding Sphagesauridae and also addressing Adamantinasuchus (see Nobre and Carvalho 2006). Readers should refer to this work for further information on Sphagesauridae.

Marinho, TS, Carvalho, IS. 2007. Revision of the Sphagesauridae Kuhn, 1968 (Crocodyliformes, Mesoeucrocodylia). In: Carvalho, IS, Cassab, RCT, Schwanke, C, Carvalho, MA, Fernandes, ACS, Rodrigues, MA, Carvalho, MSS, Arai, M, Oliveira, MEQ (Org.). Paleontologia: Cena´rios de Vida. 1st ed., Rio de Janeiro (Brazil): Intercieˆncias, vol. 1, p. 589–599.

References Andrade, MB. 2005. Revisa˜o sistema´tica e taxonoˆmica dos Notosuchia (Metasuchia, Crocodylomorpha). Dissertac¸a˜o de Mestrado, Universidade Estadual Paulista. 239 p.

Andrade MB, Bertini RJ. 2003. Morfologias craniana e denta´ria de um novo crocodilomorfo do Creta´ceo Superior brasileiro (Notosuchia, Uruguaysuchidae) e comenta´rios sobre seus possı´veis ha´bitos alimentares. In: Resumos do XVIII Congresso Brasileiro de Paleontologia. Brası´lia: Sociedade Brasileira de Paleontologia/ Universidade de Brası´lia. p. 43 –44.

Andrade MB, Bertini RJ. 2005. Morphological and anatomical observations about Mariliasuchus amarali and Notosuchus terrestris (Mesoeucrocodylia), and their phylogenetical relationships with

Historical Biology 127 other ‘notosuchians’ from South America. In: Boletim do II Congresso Latino-Americano de Paleontologia de Vertebrados. Rio de Janeiro: Museu Nacional/UFRJ. p. 23–24.

Andrade MB, Bertini RJ. 2007a. A new Sphagesaurus (Crurotarsi, Mesoeucrocodylia) from Brazil and the evolution of the notosuchian crocodylomorphs. In: Progressive Palaeontology. 4.

Abstracts. Bristol: Department of Earth Sciences/ University of Bristol. 11 p.

Andrade MB, Bertini RJ. 2007b. The Chimaerasuchus paradox: critical review of a poorly known fossil crocodile. In: Progressive Palaeontology. 4. Abstracts. Bristol: Department of Earth Sciences/University of Bristol. 26 p.

Andrade MB, Bertini RJ, Pinheiro AEP. 2005. Phylogenetical relationships of cf. Araripesuchus wegeneri (Archosauria: Crocodylomorpha). In: Gondwana. 12. Abstracts. Mendoza: Academia Nacional de Ciencias. 44 p.

Andrade MB, Bertini RJ, Pinheiro AEP. 2006. Observations on the palate and choanae structures in Mesoeucrocodylia (Archosauria, Crocodylomorpha): phylogenetic implications. Rev Bras Paleontol 9(3):323–332.

Arid FM, Vizotto LD. 1965. Crocodilı´deos fosseis nas proximidades de Santa Ade´lia (SP). Cieˆnc Cult. 17(2):138–139.

Batezelli, A. 1998. Redefinica˜o litoestratigra´fica da Unidade Arac¸atuba e da sua extensa˜o regional na Bacia Bauru no Estado de Sa˜o Paulo.

Dissertac¸a˜o de Mestrado, Universidade Estadual Paulista. 110 p.

Batezelli, A. 2003. Ana´lise da sedimentac¸a˜o creta´cica no Triaˆngulo Mineiro e sua correlac¸a˜o com a´reas adjacentes. Tese de Doutoramento, Universidade Estadual Paulista. 183 p.

Batezelli A, Saad AR, Etchebehere MLC, Perinotto JAJ, Fulfaro VJ. 2003. Ana´lise estratigra´fica aplicada a Formac¸a˜o Arac¸atuba (Grupo Bauru – KS) no Centro-Oeste do Estado de Sa˜o Paulo.

Geocieˆncias 22:5–19.

Benton MJ, Clark JM. 1988. Archosaur phylogeny and the relationships of Crocodylia. In: Benton MJ, editor. The phylogeny and classification of the tetrapods Volume 1: Amphibians, Reptiles.

Birds. Oxford: Clarendon Press. p. 295 –338.

Bertini, RJ. 1993. Paleobiologia do Grupo Bauru, Creta´ceo Superior continental da Bacia do Parana´, com eˆnfase em sua fauna de amniotas. Tese de Doutoramento, Universidade Federal do Rio de Janeiro. 493 p.

Bertini RJ, Andrade MB. 2004. Ocorreˆncias paleogeogra´ficas e litoestratigra´ficas de Sphagesaurus Price (1950) (Notosuchia: Crocodylomorpha). Desdobramentos biocronolo´gicos. In: Boletim de resumos do IV Simpo´sio Brasileiro de Paleontologia de Vertebrados. Rio Claro: UNESP. p. 5–6.

Bertini RJ, Arruda-Campos AC. 1995. Ocorreˆncia de Notosuchia (Crocodylomorpha) no Creta´ceo Superior continental da Bacia do Parana´. In: Atas do XIV Congresso Brasileiro de Paleontologia.

Uberaba: Sociedade Brasileira de Paleontologia. p. 20 –21.

Bertini RJ, Carvalho IS. 1999. Distribuic¸a˜o cronolo´gica dos crocodilomorfos notossu´quios e ocorreˆncias nas bacias creta´cicas brasileiras. In: Boletim do V Simpo´sio sobre o Creta´ceo no Brasil / I Simpo´sio sobre el Creta´cico de Ame´rica Del Sur. Serra Negra: Universidade Estadual Paulista. p. 517 –523.

Bertini RJ, Marshall LG, Gayet M, Brito P. 1993. Vertebrate faunas from the Adamantina and Marı´lia formations (Upper Bauru Group, Late Cretaceous, Brazil), in their stratigraphic and paleobiogeographic context. Neues Jahrb Geol Palaontol Abh. 188(1):71–101.

Bonaparte JF. 1991. Los vertebrados fosiles de la Formacion Rio Colorado, de la Ciudad de Neuquen y cercanias, Creta´cico Superior, Argentina. Rev Museo Argent Cienc Nat. 4(3):31–63.

Bremer K. 1994. Branch support and tree stability. Cladistics. 10: 295– 304.

Buckley GA, Brochu CA, Krause DW, Pol D. 2000. A pug-nosed crocodyliform from the Late Cretaceous of Madagascar. Nature.

405:91–94.

Buffetaut E. 1982. Radiation evolutive, pale´oe´cologie et bioge´ographie des crocodiliens me´sosuchiens. Mem Soc Ge´ologique Fr. 142:1–88.

Campos DA, Suarez JM, Riff D, Kellner AWA. 2001. Short note on a new Baurusuchidae (Crocodyliformes, Metasuchia) from the Upper Cretaceous of Brazil. Bol Museu Nacl Se´r Geologia. 57:1–7.

Candeiro CRA, Martinelli AG. 2006. A review of the paleogeographical and chronostratigraphical distribution of mesoeucrocodylians

from the upper Cretaceous beds from the Bauru (Brazil) and Neuquen (Argentina) groups, Southern South America. J S Am Earth Sciences. 22:116–129.

Candeiro CRA, Martinelli AG, Avilla LS, Rich TH. 2006. Tetrapods from the Upper Cretaceous (Turonian–Maastrichtian) Bauru Group of Brazil: a reappraisal. Cretaceous Res. 27:923–946.

Carvalho IS. 1994. Candidodon: um crocodilo com heterodontia (Notosuchia, Creta´ceo inferior–Brasil). An Acad Bras Cienc.

66(3):331–446.

Carvalho IS, Bertini RJ. 1999. Mariliasuchus: um novo Crocodylomorpha (Notosuchia) do Creta´ceo da Bacia Bauru, Brasil.

Geologia Colomb. 24:83–105.

Carvalho IS, Bertini RJ. 2000. Contexto geolo´gico dos notossu´quios (Crocodylomorpha) creta´cicos do Brasil. Geologia Colomb.

25:163– 183.

Carvalho IS, Ribeiro LCB, Avilla LS. 2004. Uberabasuchus terrificus sp. nov., a new Crocodylomorpha from the Bauru Basin (Upper Cretaceous), Brazil. Gondwana Res. 7(4):975–1002.

Carvalho IS, Campos ACA, Nobre PH. 2005. Baurusuchus salgadoensis, a new Crocodylomorpha from the Bauru Basin (Cretaceous), Brazil. Gondwana Research. 8(1):11–30.

Clark JM. 1994. Patterns of evolution in Mesozoic Crocodyliformes.

In: Fraser NC, Sues H-D, editors. In the shadows of dinosaurs: early Mesozoic tetrapods. New York: Cambridge University Press. p. 84–97.

Clark JM, Jacobs LL, Downs WR. 1989. Mammal-like dentition in a Mesozoic crocodilian. Science. 240(6):1064–1066.

Dias-Brito D, Musacchio EA, Castro JC, Maranha˜o MSAS, Sua´rez JM, Rodrigues R. 2001. Grupo Bauru: uma unidade continental do Creta´ceo do Brasil – concepc¸o˜es baseadas em dados micropaleontolo´gicos, isoto´picos e estratigra´ficos. Rev Pale´obiologie.

20(1):245–304.

Felsenstein J. 1985. Confidence limits on phylogenies: an approach using bootstrap. Evolution. 39:783– 791.

Fernandes, LA. 1992. A cobertura creta´cica suprabasa´ltica no Estado do Parana´ e Pontal do Paranapanema (SP): os grupos Bauru e Caiua´.

Dissertac¸a˜o de Mestrado, Universidade de Sa˜o Paulo. 171 p.

Fernandes LA, Coimbra AM. 1996. A Bacia Bauru (Creta´ceo Superior, Brasil). An Acad Bras Cieˆnc. 68(2):195– 205.

Fernandes LA, Coimbra AM. 2000. Revisa˜o estratigra´fica da parte Oriental da Bacia Bauru (Neocreta´ceo). Rev Bras Geocieˆncias.

30(4):717–728.

Fiorelli, LE. 2005. Nuevos restos de Notosuchus terrestris Woodward, 1896 (Crocodyliformes: Mesoeucrocodylia) del Creta´cico Superior (Santoniano) de la Provincia de Neuque´n, Patagonia, Argentina. Tesis de Licenciatura, Universidad Nacional de Co´rdoba. 79 p.

Fiorelli LE, Calvo JO. 1896. Nuevos restos de Notosuchus terrestris Woodward, 1896 (Crocodyliformes: Mesoeucrocodylia) del Creta´cico Superior (Santoniano) de la Provı´ncia de Neuque´n, Patagoˆnia, Argentina. In: Boletim do II Congresso LatinoAmericano de Paleontologia de Vertebrados. Rio de Janeiro: Museu Nacional/UFRJ. p. 110–111.

Fitch WM. 1971. Toward defining the course of evolution: minimum change for a specific tree topology. Syst Zool. 20:406–416.

Gasparini ZB. 1971. Los Notosuchia del Cretacico de America del Sur como un nuevo infraorden de los Mesosuchia (Crocodylia).

Ameghiniana. 8(1):83–103.

Gobbo-Rodrigues, SR. 2001. Caro´fitas e ostra´codos do Grupo Bauru, Creta´ceo Superior Continental do Sudeste do Brasil. Dissertac¸a˜o de Mestrado, Universidade Estadual Paulista. 137 p.

Gomani EM. 1997. A crocodyiliform from the Early Cretaceous Dinosaur Beds, Northern Malawi. J Vert Paleontol. 17(2): 280–294.

Hay OP. 1930. Second Bibliography and Catalogue of the fossil vertebrate of North America. Washington: Carnegie Institute of Washington. 1074 p.

Hennig W. 1950. Grundzu¨ge einer theorie phylogenestichen systematik. Berlin: Deutscher Zentralverlag. 263 p.

Hennig W. 1966. Phylogenetic Systematics. Lawrence: University of Illinois Press. 263 p.

128 M.B. de Andrade and R.J. Bertini Huene F. von. 1931. Verschiedene mesozoische Wirbeltierreste aus Sudamerika. Neues Jahrbuch fur Mineralogie und Palaeontologie Abhandlungen. 66:181.

Jouve S, Iarochene M, Bouya B, Amaghzaz M. 2006. A new species of Dyrosaurus (Crocodylomorpha, Dyrosauridae) from the early Eocene of Morocco: phylogenetic implications. Zoolog J Linnean Soc. 148:603 –656.

Kellner AWA, Campos DA. 1999. Vertebrate Paleontology in Brazil – a review. Episodes 22(3):238–251.

Kellner AWA, Campos DA, Price LI. 1995. Material of Sphagesaurus (Sphagesauridae, Crocodilia) from the Late Cretaceous of Brazil.

In: Atas do XIV Congresso Brasileiro de Paleontologia. Uberaba: Sociedade Brasileira de Paleontologia. p. 70 –71.

Kuhn O. 1968. Die vorzeitlichten krokodile. Krailing Bei Mu¨nchen: Verlag Oeben. p. 1–124.

Larsson HCE, Sues H-D. 2007. Cranial osteology and phylogenetic relationships of Hamadasuchus rebouli (Crocodyliformes: Mesoeucrocodylia) from the Cretaceous of Morocco. Zoolog J Linnean Soc. 149:533–567.

Martinelli A. 2003. New cranial remains of the bizarre notosuchid Comahuesuchus brachybuccalis (Archosauria, Crocodyliformes) from the Late Cretaceous of Rio Negro Province (Argentina).

Ameghiniana. 40(4):559–572.

Mezzalira S. 1989. Os fo´sseis do Estado de Sa˜o Paulo. Sa˜o Paulo: Instituto Geolo´gico. 141 p.

Nobre PH, Carvalho IS. 2006. Adamantinasuchus navae, a new Gondwanan Crocodylomorpha (Mesoeucrocodylia) from the Late Cretaceous of Brazil. Gondwana Res. 10:370–378.

Ortega F, Gasparini Z, Buscalioni AD, Calvo JO. 2000. A new species of Araripesuchus (Crocodylomorpha, Mesoeucrocodylia) from the Lower Cretaceous of Patagonia (Argentina). J Vert Paleontol.

20(1):57– 76.

Paula e Silva F, Chang HK, Caetano–Chang MR. 2003. Perfis de refereˆncia do Grupo Bauru (K) no Estado de Sa˜o Paulo.

Geocieˆncias. 22:21–32.

Pol, D. 1999. El esqueleto postcraneano de Notosuchus terrestris (Archosauria: Crocodyliformes) del Creta´cico Superior de la Cuenca Neuquina y su informacion filogene´tica. Tesis de Licenciatura, Universidad de Buenos Aires. 158 p.

Pol D. 2003. New remains of Sphagesaurus (Crocodylomorpha: Mesoeucrocodylia) from the Upper Cretaceous of Brazil. J Vert Paleontol. 23(4):817–831.

Pol D, Apesteguia S. 2005. New Araripesuchus remains from the early Late Cretaceous (Cenomanian-Turonian) of Patagonia. Am Museum Novitates. 3490:1– 38.

Pol D, Norell MA. 2004a. A new crocodyliform from Zos Canyon, Mongolia. Am Museum Novitates. 3445:1–36.

Pol D, Norell MA. 2004b. A new gobiosuchid crocodyliform taxon from the Cretaceous of Mongolia. Am Museum Novitates. 3458: 1–31.

Pol D, Ji S, Clark JM, Chiappe LM. 2004. Basal crocodyliforms from the Lower Cretaceous Tugulu Group (Xinjiang, China), and the phylogenetic position of Edentosuchus. Cretaceous Res. 25: 603– 622.

Price LI. 1945. A new reptile from the Cretaceous of Brazil. Notas Preliminares e Estudos – DGM 25:1– 8.

Price LI. 1950. On a new crocodilian, Sphagesaurus, from the Cretaceous of the State of Sa˜o Paulo, Brazil. Anais da Academia Brasileira de Cieˆncias. 22(1):77–85.

Price LI. 1955. Novos crocodilı´deos dos arenitos da Se´rie Bauru.

Creta´ceo do Estado de Minas Gerais. An Acad Bras Cieˆnc. 27: 487– 498.

Price LI. 1959. Sobre um crocodilı´deo notossu´quio do Creta´ceo brasileiro. Bol Departamento Nacl Produc¸a˜o Miner. 188:1–55.

Rusconi C. 1933. Sobre reptiles creta´ceos del Uruguay (Uruguaysuchus aznarezi, n.g.n. sp.) y sus relaciones com los notosu´chidos de Patagonia. Bol Inst Geologia y Perforaciones. 19(8):1–64.

Santucci RM, Bertini RJ. 2001. Distribuic¸a˜o paleogeogra´fica e biocronolo´gica dos titanossauros (Saurischia, Sauropoda) do Grupo Bauru, Creta´ceo Superior do Sudeste Brasileiro. Rev Bras Geocieˆncias. 31(3):307–314.

Sereno PC, Larsson HCE, Sidor CA, Gado B. 2001. The giant crocodyliform Sarcosuchus from the Cretaceous of Africa.

Science. 294:1516–1519.

Sereno PC, Sidor CA, Larsson HCE, Gado B. 2003. A new notosuchian from the Early Cretaceous of Niger. J Vert Paleontol. 23(2): 477 –482.

Sorenson MD. 1999. TreeRot, version 2. Boston: Boston University.

Steel R. 1973. Crocodylia. In: Steel R, editor. Handbuch der Palaoherpetologie. Stuttgart: G. Fischer. 116 p.

Swofford DL. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), Version 4.0b10. Sunderland, MA: Sinauer Associations, Inc. Turner AH. 2006. Osteology and phylogeny of a new species of Araripesuchus (Crocodyliformes: Mesoeucrocodylia) from the Late Cretaceous of Madagascar. Hist Biol. 18(3):255–369.

Turner AH, Calvo JO. 2005. A new sebecosuchian crocodyliform from the Late Cretaceous of Patagonia. J Vert Paleontol. 25:87–98.

Vasconcellos FM, Carvalho IS. 2005. Esta´gios de desenvolvimento de Mariliasuchus amarali, Crocodyliformes Mesoeucrocodylia da Formac¸a˜o Adamantina, Creta´ceo Superior da Bacia Bauru, Brasil.

Anu Inst Geocieˆncias. 28(1):49– 69.

Walker AD. 1970. A revision of the Jurassic reptile Hallopus victor (Marsh), with remarks on the classification of crocodiles. Philos Trans R Soc Lond B. 257:323– 372.

Whetstone KN, Whybrow PJ. 1983. A ‘cursorial’ crocodylian from the Triassic of Lesotho (Basutoland), Southern Africa. Occasional Papers of the Museum of Natural History of the University of Kansas. 106:1–37.

Woodward AS. 1896. On two Mesozoic crocodylians, Notosuchus (genus novum) and Cynodontosuchus (genus novum) from the red sandstones of the Territory of Neuquen (Argentine Republic). An Museo Plata – Se´rie Paleontologia Argent. 4(1–2):1–20.

Wu X-C, Sues H-D. 1996. Anatomy and phylogenetic relationships of Chimaerasuchus paradoxus, an unusual crocodyliform reptile from the Lower Cretaceous of Hubei, China. J Vert Paleontol.

16(4):688–702.

Wu X-C, Sues H-D, Sun A. 1995. A plant-eating crocodiliform reptile from the Cretaceous of China. Nature. 376:678–680.

Zaher H, Pol D, Carvalho AB, Ricomini C, Campos D, Nava W. 2006.

Redescription of the cranial morphology of Mariliasuchus amarali, and its phylogenetic affinities (Crocodyliformes, Notosuchia). Am Museum Novitates. 3512:1–40.

Appendix A: Anatomical abbreviation An angular Bo basioccipital BPo postorbital bar Bs basisphenoid c caniniform tooth Ch choanae (internal naris) De dentary Elf lateral eustachian foramen Eo exoccipital Ept ectopterygoid FAO antorbital fenestra FLT laterotemporal fenestra FM mandibular fenestra fg foraminal gap fj jugal neurovascular foramen fm maxillary neurovascular foramen Fr frontal fs neurovascular foramen at the Pmx-Mx suture FSO suborbital fenestra FST supratemporal fenestra IcS interchoanal septum JP jugal sickle-like ventral process

Historical Biology 129 Ju jugal La lachrymal m maxillary molariform tooth ma mandibular tooth MP posteroventral maxillary process Mx maxilla N external naris n exoccipital neurovascular passage Na nasal NPD nasopharyngeal duct Orb orbit Pal palatine Par parietal PF parachoanal fossa Pfr prefrontal Pi prefrontal pillar Pmx premaxilla Po postorbital Pt pterygoid Qj quadratojugal Qu quadrate San surangular SER surangular external ramus SIR surangular internal ramus So supraoccipital Sp splenial Sq squamosal Vf vagi foramen

MLP – 64 – IV – 16 – 8(209), MLP – 64 – IV – 16 – 10(221), MLP–64– IV–16– 11, MLP–64– IV–16– 12, MLP–64– IV–16–13, MLP–64–IV–16–14, MLP–64–IV–16–15, MLP–64– IV–16– 16, MLP–64– IV–16– 17, MLP–64– IV–16–18, MLP–64–IV–16–20, MLP–64–IV–16–21, MLP – 64 – IV – 16 – 22, MLP – 64 – IV – 16 – 23, MLP – 64–IV–16–24, MLP–64–IV–16–25, MLP–64–IV–16– 28, MLP – 64 – IV – 16 – 30, MLP – 64 – IV – 16 – 31(206), MPCA–Pv–528;MPCA–Pv–789/1;MPCA–Pv–791.

Sphagesaurus huenei – DGM – 332 – R, DGM – 333 – R, DGM–1411–R,RCL-100,URCR†015.

Sphagesaurusmontealtensis – MPMA15-001/90.

Stratiotosuchusmaxhechti – DGM1477-R,URCR†73.

Appendix C. Character list used in the analysis Total number of characters 183, distributed as follows: 120 cranial (65.57%); 28 dental (15.3%); 32 postcranial (17.5%).

Main anatomical element indicated in bolt, in the description of the character. Characters either new or obtained from previous works, designated by code and original number, presented in chronological order inside brackets. ‘m’ indicates characters modified from the original. Codes for bibliographic origin of characters as follows: A ¼ Buckley et al. 2000; Andrade 2005; C ¼ Clark 1994; G ¼ Gomani 1997; M ¼ Martinelli 2003; O ¼ Ortega et al. 2000; Pa ¼ Pol 1999; Pb ¼ Pol 2003; S ¼ Sereno et al. 2003; W ¼ Wu and Sues 1996.

General (3 characters; 1.64% of total)

(1) Appendix B. Specimens examined Type specimens in bold. The number in parentheses is the original designation for MLP specimens of Notosuchus, given

by A. Smith Woodward. Notosuchus lectotype elected by Gasparini (1971), from the assemblage of specimens used in the original description (Woodward 1896).

Caiman crocodilus – URC R†76.

Araripesuchus patagonicus – MUCPv– 267, MUCPv– 268, MUCPv– 268b, MUCPv – 269 (holotype), MUCPv – 270.

Comahuesuchus brachybuccalis – MACN – Pv – N – 31, MOZ– P– 6131, MUCPv –202.

Cynodontosuchus rothi – MLP – 64 –IV – 16 – 25.

Geosaurus araucanensis – MLP –72 – IV – 7– 1, MLP – 72 –IV –7 – 2.

Geosaurus sp. – MACN – Pv– N – 64, MACN – Pv– N – 95.

Mariliasuchus amarali – MN – 6298– V, MN– 6756– V, UFRJ– DG – 50 – R, URC R†67, URC R†68, URC R†69, URC R†74 and URC R†75.

Notosuchusterrestris – MACN–Pv–N–22,MACN–Pv–N– 23, MACN–Pv–N–24, MACN–Pv–N–43, MACN–Pv– N– 107, MACN – Pv–RN – 1015, MACN – Pv– RN– 1037, MACN–Pv–RN–1038, MACN–Pv–RN–1039, MACN– Pv–RN–1040, MACN–Pv–RN–1041, MACN–Pv–RN– 1043, MACN – Pv – RN – 1044, MACN – Pv – RN – 1045, MACN–Pv–RN–1046, MACN–Pv–RN–1047, MACN– Pv–RN–1048, MACN–Pv–RN–1118, MACN–Pv–RN– 1119, MLP – 64 – IV – 16 – 1, MLP – 64 – IV – 16 – 5(253), MLP – 64 – IV – 16 – 6(203), MLP – 64 – IV – 16 – 7(219),

(2)

(3)

Skull surface [O01]: (0) smooth or ornamented with an irregular pattern of ridges rugosities and anastomosing grooves; (1) ornamented with circular to polygonal pits, with eventual sulcation (not anastomosed).

Skull height, in posterior view [C03m; S06m; P03m; A02]: (0) skull higher than wider or subequal; (1) skull larger than higher.

Orientation of the orbits [A3]: (0) lateral; (1) laterodorsal.

Rostrum (28 characters; 15.3% of total) (4)

(5)

(6)

(7)

(8) (9)

Proportional length of the rostrum in lateral view [W4m]: (0) short, orbits at the skull midlength; (1) long, orbits at the posterior half of the skull.

Rostrum height, anterior view [C03]: (0) rostrum higher than wider; (1) rostrum subquadrate; (2) rostrum wider than higher.

Rostrum, in dorsal view [C02]: (0) is narrow, abruptly widening to adjust the skull outline; (1) gradually fits the skull outline, with a general triangular shape.

External nares [C06; S02m þ 07m; Pb06m]: (0) terminal, opening anteriorly; (1) lateral or semi-lateral; (2) anterodorsal or dorsal.

Internarial bar [S7]: (0) absent; (1) gracile, narrow; (2) present as a wide bar.

Anterorbital region [C68m; W16m; S03m]: (0) smooth and flat; (1) vestigial or small antorbital fenestra; (2) well developed antorbital fenestra.

130 M.B. de Andrade and R.J. Bertini (10)

(11) (12)

(13)

(14)

(15) (16)

(17)

(18)

(19)

(20)

(21)

(22)

(23) (24)

Anteriormost portion of the premaxilla, at the region at the front and below the narial opening [C5; S9]: (0) narrower than the lateral part of the premaxilla; (1) high, subequal to the height of the lateral part of the premaxilla.

Premaxilla–maxilla foramen [Pa149; O13; Pb135]: (0) absent; (1) present.

Premaxilla–maxilla suture, lateral view [A13m]: (0) straight, vertical; (1) curved or composed by two planes (lower vertical, upper diagonal), with a posterodorsal process, flanking the nasals; (2) straight, diagonal.

Inferior border of the premaxilla–maxilla suture, in lateral view [C09m]: (0) without constriction or notch, region between premaxilla and maxilla flat, composing a straight border; (1) evident constriction, forming a notch.

Premaxilla–maxilla suture, in dorsal view [C09m; O14m; S10m; Pb09m]: (0) without constriction, with flat surface; (1) evident constriction.

Maxilla [A16; Z192m]: (0) does not contribute to the orbit; (1) contributes to the orbit.

Maxilla, proportional number of neurovascular foramina relative to the number of teeth [A17m]: (0) small number of foramina, usually 1–2 for each tooth; (1) greater number of foramina, widely surpassing the number of teeth.

Maxilla, distance between neurovascular foramina and teeth [A17m]: (0) small distance, foramina positioned close to teeth; (1) foramina clearly apart from teeth.

Inferior margin of the maxilla [W29; M24]: (0) not different than the remaining surface from maxilla; (1) smooth surface, mesially inclined.

Inferior margin of the maxilla, in lateral view [C79m]: (0) concave at the anteriormost region and convex at the posteriormost region; (1) straight; (2) convex at the anteriormost region and convex and straight at the posteriormost region; (3) concave at the anteriormost region, convex at the midlength and concave at the posteriormost region, ‘festooned’.

Anterior border of the nasals [C13m]: (0) short, at best with a small stiliform projection over the naris (eventually not in contact to the external naris due to isolation by the premaxillae); (1) moderately developed, projecting over the naris as a narrow lamina, without covering the naris; (2) well developed, completely covering the naris.

Nasals, general shape [A21]: (0) triangular, posterior region wider than anterior region; (1) rectangular; (2) triangular, with the anteriormost region wider than the posterior region.

Lachrymal [C11; M05]: (0) does not contact nasal, prevented by large contact among prefrontal–maxilla; (1) with small contact to nasal; (2) with large contact with nasal, preventing any contact among prefrontal–maxilla.

Lachrymal size [A23]: (0) small; (1) well developed.

Relative position of the anteriormost margin of the prefrontal [A25]: (0) with anteriormost margin at the same relative position as the anteriormost margin

(25)

(26)

(27) (28)

(29) (30) (31) (32)

of frontal; (1) anteriormost margin surpasses the anteriormost margin of frontal.

Posterior margin of prefrontal [A26]: (0) short, limited to anterodorsal border of the orbit; (1) elongated, composing the dorsal border of the orbit.

Prefrontal pillars, construction [C15m; S34m; Pb15m]: (0) incomplete, without contact between the descending ramus of the prefrontal and palatine; (1) complete, with contact between the descending ramus of the prefrontal and palatine.

Prefrontal pillars, structure [C15m]: (0) small contact area; (1) wide contact area.

Frontal, shape of anterior border [A28]: (0) straight; (1) triangular to stiliform, projecting forward between nasals.

Frontal anterior border, suture [A29]: (0) straight or with minor interdigitation; (1) strongly interdigitated.

Interfrontal longitudinal ridge [C22; A30]: (0) absent; (1) present.

Rostral transversal crest [A31]: (0) absent; (1) present.

Frontal, position of anterior border [A32]: (0) between orbits; (1) at the same position that the anteriormost orbital border, in dorsal aspect; (2) positioned ahead of the orbits, in dorsal aspect.

Skull table (8 characters; 4.37% of total) (33)

(34)

(35) (36) (37)

(38)

(39)

(40)

(41)

Frontal posterior border [C23; S20; M08]: (0) short, with limited contact with the postorbital; (1) well developed, fairly contacting the postorbital and contributing to the supratemporal fenestra.

Parietal surface [C22m; M09m; S26m]: (0) flat and wide; (1) flat, but narrow due to the development of the supratemporal fenestra; (2) sagittal crest.

Anterolateral process of postorbital [C28; S24]: (0) absent; (1) present.

Supratemporal fenestra, shape [A36m]: (0) circular to subcircular; (1) elliptical, main axis clearly identifiable.

External border of the supratemporal fenestra, orientation of the main axis in dorsal view [A37m]: (0) diagonal, projection of the main axes converge posteriorly; (1) parallel, projection of the axes do not meet; (2) diagonal, projection of the main axes converge anteriorly.

Proportional size of the supratemporal fenestra (main axis of the internal border) [C68m; S04m]: (0) smaller than the diameter of the orbit; (1) subequal or bigger than the diameter of the orbit.

Relation among the internal and external borders of the supratemporal fenestra [A39]: (0) without significant difference; (1) external border slightly larger; (2) external border much larger.

Area posterior to the supratemporal fenestra, where lies the parietal–squamosal suture [A40]: (0) ample and flat horizontal surface; (1) surface extremely narrow and high, forming a crest transversal to the skull.

Ventrolateral ramus of squamosal, in dorsal view [A41]: (0) only slightly developed, suture with the quadrate covered by the superior lateral (temporal)

Historical Biology 131 ramus of the squamosal in dorsal view; (1) well developed, suture with the quadrate exposed in dorsal view.

Temporal region (26 characters; 14.2% of total) (42)

(43)

(44)

(45)

(46)

(47)

(48) (49)

(50) (51)

(52) (53) (54)

(55)

(56)

(57) (58)

Jugal anterior ramus, shape in lateral view [S16m; M07m]: (0) narrow throughout, widening directly at the contact with the maxilla; (1) gradually widening anteriorly; (2) sudden widening from the base of the ramus, ‘leaf-shaped’.

Jugal anterior ramus, external surface [A43; Z193]: (0) well developed single neurovascular foramen, directed anteriorly to anterolateraly; (1) even surface, either flat or ornamented, without any kind or number of foramina; (2) two or more foramina, all small, facing ventrally.

Jugal anterior ramus, length [Pa134; Pb122; M29]: (0) short, anteriormost margin does not reach the anterior margin of the orbit in lateral view; (1) long, either reaching or surpassing the anterior margin of the orbit in lateral view.

Jugal anterior ramus, relative position in lateral view [A45]: (0) horizontal; (1) inclined diagonally, anterior border ventral to the base of the ramus.

Jugal anterior ramus, occurrence of an external lateral crest [Pa133m; O145m; Pb121m]: (0) absent; (1) present.

Jugal anterior ramus, cross-section [C18m]: (0) circular to subcircular; (1) elliptical, with evident lateral compression.

Jugal posterior ramus [new]: (0) straight; (1) dorsally arched.

Postorbital bar, relation to dermis [C25m]: (0) subdermic, distinct, originating mesially from the jugal ramus; (1) dermic, gradually narrowing.

Postorbital bar [new]: (0) straight; (1) posteroventrally bended at midlength.

Postorbital bar, constitution from ectopterygoid [C26m; S22]: (0) does not receive contribution from ectopterygoid; (1) receive contribution from ectopterygoid.

Postorbital bar, ectopterygoid–postorbital contact [C26m; Pa158; O36; Pb144]: (0) absent; (1) present.

Postorbital bar, general structure [C25m]: (0) gracile; (1) robust.

Postorbital bar, inclination in lateral view [A54]: (0) vertical; (1) diagonally inclined, distal end fairly anterior to the proximal end.

Postorbital bar, dorsal end next to the postorbital body [C30; S25]: (0) bar gradually expanding towards the main body of the postorbital, without a well defined limit; (1) constriction delimiting the distinction between the postorbital bar and the postorbital body.

Postorbital bar, occurrence of vascular foramen on the lateral edge of the postorbital margin [C27; S23; T27]: (0) absent; (1) present.

Postorbital bar, cross-section [C26; S21]: (0) subcircular; (1) elliptical, with lateral compression.

Postorbital bar, implantation of the proximal end to the postorbital body [A58]: (0) postorbital bar next to the

(59) (60)

(61)

(62)

(63)

(64)

(65)

(66) (67)

(68)

laterodorsal border of the postorbital body; (1) postorbital bar next to the ventral portion of the postorbital body.

Laterotemporal fenestrae [O46]: (0) facing laterally; (1) facing laterodorsally.

Quadratojugal mesial border, ornamentation of the posterodorsal margin of the laterotemporal fenestra) [S18m]: (0) absent; (1) present, either ornamented with a discrete crest or a well defined spine (Spina quadratojugalis).

Quadratojugal anterodorsal ramus, development [C19m; S19m]: (0) narrow and gracile; (1) wide and robust.

Quadratojugal anterodorsal ramus, contribution to postorbital bar [C19m; S19m]: (0) does not contribute to postorbital bar; (1) contribute to postorbital bar.

Quadratojugal anterodorsal ramus, contact with postorbital bone [C19m; S19m]: (0) contact posterior region of postorbital body; (1) contact anterior region of postorbital body.

Quadratojugal, contact with jugal [O39]: (0) suture between jugal and quadratojugal lies next to the posterior vertex of the laterotemporal fenestra; (1) suture between jugal and quadratojugal lies below the laterotemporal fenestra, due to the development of a small process from the quadratojugal.

Quadrate fenestrae [C45m; S35m]: (0) with no more than one fenestra; (1) with at least two well defined fenestrae.

Quadrate, surface [O154]: (0) surface flat and even; (1) presenting one depression with triangular shape.

Relative position of the quadrate condyle, in lateral/posterior views [W24m; S46m; Pb104]: (0) at the same height than the occipital condyle, clearly above the teeth row; (1) below the occipital condyle, approximately at the same height of the teeth row; (2) clearly below the teeth row and the occipital condyle.

Quadrate, medial articulation facet of the condyle [O53]: (0) small, with the same dimensions than the lateral articular facet; (1) large, bigger than the lateral articular facet, projecting ventrally.

Basicranium (10 characters; 5.47% of total) (69)

(70)

(71)

(72)

Basisphenoid, at the ventral portion of the skull [C56; S36; T56m]: (0) exposed ventrally; (1) almost completely covered by the pterygoids and basioccipital.

Basioccipital–quadrate contact [new]: (0) small or absent; (1) well developed, excluding the basisphenoid from the exoccipital and this last element from the ventral surface of the quadrate.

Basioccipital and occipital condyle [G32m; O176m; Pb112m]: (0) facing posteriorly; (1) facing posteroventrally.

Contact between quadrate, squamosal and exoccipitals [C49m; T49m; M14m]: (0) without significant contact; (1) with well developed contact, lateral to cranioquadrate passage.

132 M.B. de Andrade and R.J. Bertini (73)

(74) (75)

(76)

(77)

(78) (79)

Occipital surface, in dorsal view [new]: (0) overall flat; (1) ‘U-shaped’, concave but flat in the area posterior to the skull table; (2) ‘V-shaped’, exoccipitals posteromedially oriented from near the medial line.

Exoccipital surface, above the occipital crest [new]: (0) faces posteriorly; (1) faces posterodorsally.

Insertion area for the m. depressor mandibulae, at the surface of occipital [A72]: (0) slightly developed surface, narrow and low, smaller than the muscle attachment area of the exoccipitals; (1) well developed surface, with muscle attachment area similar to the surface of the exoccipitals.

Lateral occipital surface of the squamosal [A73]: (0) flat or slightly bended posteriorly; (1) strongly bended posteriorly, so the border is positioned posteriorly to the occipital condyle.

External surface of the occipital portion of the squamosal, inclination [A74]: (0) faces posteriorly; (1) faces posterodorsally.

Occipital surface of supraoccipital, inclination [A75]: (0) faces posteriorly; (1) faces posterodorsally.

Occipital surface of supraoccipital, in dorsal view [A76]: (0) surface either flat and even or concave; (1) in ‘V’, projecting posteriorly forming a vertically oriented medial ridge, or even a crest.

Palate and perichoanal elements (23 characters; 12.57% of total) (80)

(81)

(82)

(83)

(84) (85) (86)

(87)

Naso-oral fenestra ( ¼ incisive foramen) [C07; O11– 12m; S29; Pb07]: (0) absent; (1) present, limited by the premaxilla; (2) present, limited by both the maxilla and premaxilla.

Development of the palatine rami of the maxilla [C10m; S33m; Pb10m]: (0) rami slightly developed; (1) rami well developed.

Contact between palatine rami of the maxilla [C10m; S33m; Pb10m]: (0) rami do not contact each other at the palate surface; (1) rami contact each other at the palate surface, eventually separated by vomer-palatine or vomer-pterygoids, but always forming a bony palate, separating nasal and oral cavities.

Palatines [C37; T37]: (0) do not contact each other and do not contribute to a secondary palate; (1) contact each other, as part of the secondary palate.

Bony surface of the secondary palate [O175; M47]: (0) flat and even, or slightly convex; (1) concave.

Maxillo-palatine fenestrae [A82]: (0) absent; (1) present.

Suborbital fenestrae, shape of anterior border [new]: (0) rounded, smooth; (1) in sharp angle, forming a notch, fenestrae with the shape of a wide fissure.

Suborbital fenestrae, composition of lateral border [O61m]: (0) jugal takes part of the lateral border; (1) both ectopterygoid and maxilla compose the lateral border, excluding the jugal.

Suborbital fenestrae, composition of anteromedial border [new]: (0) composed exclusively by the palatines; (1) palatine ramus of the maxilla contributes to the anteromedial border, by means of a narrow and elongated process, directed posteriorly.

(89) Suborbital fenestrae, composition of posterior border [M35]: (0) pterygoid takes part of the posterior border; (1) posterior border composed exclusively by the palatine and ectopterygoid, with pterygoid excluded by palatine – ectopterygoid contact.

(90) Palatine anterior border [Z129]: (0) do not exceed the anterior borders of the suborbital fenestrae; (1) clearly exceed the anterior border of the suborbital fenestra, directed anteriorly.

(91) Internal nares, shape, in palatal view [A87; Z195m]: (0) anterior border usually straight or slightly arched, posterior border bended, with the overall shape of a reversed triangle; (1) slightly elongated, from rectangular or elliptical/ subcircular; (2) anterior border ‘V-shaped’ due to the presence of posterolateral palatine processes ( ¼ palatine bar), posterior border straight or slightly bended, with the overall shape of a triangle.

(92) Internal nares, perichoanal crest delimiting at least the posterior border of the choanae [A88]: (0) absent, borders smooth; (1) present.

(93) Internal naris in adult specimens, orientation [A89m]: (0) facing ventrally; (1) facing posteroventrally.

(94) Pterygoid ventral rami, size [A90]: (0) small; (1) well developed.

(95) Pterygoid ventral rami, inclination [A91]: (0) slightly inclined, posteroventrally oriented; (1) well inclined, ventrally oriented.

(96) Pterygoid ventral rami, structure [A91]: (0) gracile, with a laminar profile; (1) robust, thick.

(97) Fusion of the caudal portion of pterygoids [C41; O58; M12]: (0) absent; (1) present.

(98) Ectopterygoid medial process of the posterior ramus [A93; Z196m]: (0) absent or incipient, ectopterygoid excluded from the internal naris by pterygoid-palatine contact; (1) present and well developed, contributing to the anterolateral border of the internal naris.

(99) Internal naris anterior border [C44m; W59; S30m]: (0) formed by either maxilla or palatines, in an anteriormost position, anterior to the suborbital fenestrae; (1) formed by palatines, text to the posterior margins of the suborbital fenestrae; (2) formed by pterygoids, positioned far posteriorly to the suborbital fenestra.

(100) Internal naris posterior border [C44m; W59m; S30m]: (0) maxillae or palatines; (1) pterygoids.

(101) Internal naris, length (compared to the length of suborbital fenestrae) [C42m; S31m]: (0) clearly smaller than the suborbital fenestrae; (1) subequal in length.

(102) Interchoanal septum [C69m; S32m; Pb69m]: (0) absent; (1) present, laminar; (2) present, robust.

(88)

Historical Biology 133 (103) Parachoanal fossae [new]: (0) absent; (1) present.

Mandible (16 characters; 8.75% of total) (104) Mandibular symphysis, length [O151m]: (0) short, limited to the anteriormost portion of the rostrum, do not extend posteriorly further than the maxilla– premaxilla suture; (1) long, extending posteriorly beyond the maxilla–premaxilla suture, to a position below the 2nd–4th maxillary teeth.

(105) Mandibular symphysis, structure [W17; S44]: (0) shallow, spatulated anteriorly; (1) deep.

(106) Mandibulary symphysis, contribution of splenials [C77; Pb77]: (0) do not take part of the symphysis or at least do not take part with ventral exposure; (1) clearly take part of the symphysis, with ventral exposure.

(107) Disposition of mandibulary rami, ant the anterior and middle sections [Pb155m]: (0) mandibulary rami very close to each other, parallel; (1) mandibulary rami confluent, with a ‘V’ or ‘Y’shape; (2) mandibulary rami parallel, but distant to each other, with the shape of a ‘U’, forming an arch.

(108) Dentary, lateral aspect [A107]: (0) anterior potion as deep as the posterior one; (1) dentary ramus gradually expand posteriorly, the posterior region been deeper than the anterior region; (2) dentary ramus suddenly expand posteriorly, the posterior region been deeper than the anterior region.

(109) Lateral surface of dentary, at midsection [O81; M42]: (0) flat, with lateral compression, with high lateral margin; (1) without lateral compression, lateroventral surface convex.

(110) Dentary alveolar margin [W29m; M24m]: (0) undifferentiated from the remaining dentary surface; (1) region flat and smooth, inclined mesially.

(111) Shape of the dentary alveolar margin, in lateral view [O84; S38]: (0) straight or with a single elevation (if bearing an hypertrophied caniniform); (1) sinusoidal, undulated, with at least two concave regions and two tooth bearing elevations alternating to each other.

(112) Shape of anteromedial margin of surangular [C74; S41; Pb74]: (0) straight, coronoid process absent; (1) dorsally arched (coronoid process).

(113) Surangular anterior border [new]: (0) single or lightly furcated, directed to the lateral surface of the mandible; (1) clearly furcated and divergent, the medial ramus directed toward the splenial and the lateral ramus directed toward the dentary.

(114) Angular, height of anterior ramus, in lateral view [C70; W18]: (0) narrow; (1) high, excluding completely the posteroventral ramus of the dentary from the internal border of the maxillary fenestra.

(115) Angular, length of the anterior ramus [A114]: (0) short, not surpassing the anterior border of the mandibular fenestra; (1) moderately elongated, slightly surpassing the mandibular fenestra; (2) very long, reaching far beyond the fenestra.

(116) Prearticular [C72; S39]: (0) absent; (1) present.

(117) General proportions of glenoid fossa [W23m; M22m; S45; Pb103]: (0) wider than longer or subequal; (1) longer than wider.

(118) Posterior border of the glenoid fossa [W23m]: (0) posterior border even with the glenoid surface or, at best, incipient; (1) with a well developed posterior border, limiting anteroposterior movements from the mandible.

(119) Retroarticular process [S47m]: (0) posterodorsally oriented; (1) slightly developed or directed posteriorly; (2) posteroventrally oriented.

(120) Angular, extension of the insertion area for m. pterigoideus posterior at the medial surface [C76; S42; P76]: (0) absent; (1) present.

Dentition (28 characters; 15.3% of total) (121) Teeth apex, shape [A142]: (0) apex usually rhomboid; (1) apex usually acute.

(122) Teeth apex, inclination [A143]: (0) without inclination or lingually inclined; (1) inclined posteriorly or posterolingually.

(123) Maxillary/dentary posterior teeth, surface [A126m]: (0) smooth; (1) well striated by a great number of almost microscopic anastomosed ridges, with a general pattern from base to apex (but not exclusively); (2) macroscopic striation (base-apex), composed by gracile narrow ridges; (3) small number of robust ridges (baseapex), large and wide, similar to carinae, usually over the entire surface of the each crown.

(124) Total number of premaxillary teeth [W27m; O133m; Pb105m]: (0) one; (1) two; (2) three; (3) at least four.

(125) Hypertrophied caniniform at the premaxilla [A119]: (0) present, without anterior teeth; (1) present, preceded by one tooth; (2) present, preceded by two teeth; (3) present, preceded by 3 teeth; (4) absent.

(126) Premaxillary tooth posterior to the premaxillary hypertrophied caniniform [A120]: (0) absent; (1) present.

(127) Distribution of premaxillary teeth [S74m]: (0) over the whole alveolar surface of premaxilla; (1) edentulous region between premaxillary teeth, composing a medial diastema at the anteriormost region of the jaws.

(128) Premaxillary posteroventral extensions embracing partially or completely the base of the crown of the first maxillary tooth [new]: (0) absent; (1) present.

(129) Total number of maxillary teeth [W30m; O164m; S51m; Pb107m]: (0) no more than seven teeth; (1) no less than 10 teeth.

(130) Anterior maxillary dentition [A124]: (0) all maxillary teeth caniniform (subisometric and isomorphic); (1) hypertrophied caniniform preceded by 3–4 smaller teeth and followed by smaller caniniform teeth; (2) hypertrophied caniniform preceded by 1–2 smaller caniniform teeth and followed by smaller caniniform teeth; (3) hypertrophied caniniform, preceded by 1–2 smaller caniniform teeth and followed by smaller

134 M.B. de Andrade and R.J. Bertini

(131)

(132)

(133) (134)

(135)

(136)

(137)

(138)

(139) (140)

(141)

molariform teeth; (4) slightly enlarged molariform, preceded by 1–2 molariform-caniniform teeth and followed by molariforms; (5) all maxillary teeth molariform (subisometric and isomorphic).

Area occupied by the maxillary teeth, in palatal view [A125]: (0) proportionally small teeth, occupying only a marginal portion of the ventral surface of the maxilla; (1) proportionally well developed teeth, occupying large area the maxillary ventral surface (at least one third of the surface available).

Mesial and distal surfaces of maxillary teeth [B104m; S53m]: (0) heterogeneous carina, composed by anisomorphic tubercle-like denticles, developed preferentially at the posterior border; (1) either a smooth surface or a homogenous carina (crenulations may appear as a result of superficial ornamentation), extending over most of the anterior and posterior tooth surfaces; (2) homogenous carina, serrated with true denticulation (ziphodont dentition), extending over most of the anterior and posterior tooth surfaces.

Implantation of maxillary teeth [P137]: (0) not oblique; (1) oblique.

Dental implantation at the maxilla (anterior and middle teeth) [O19m; M38m]: (0) teeth set disposed in a groove, the roots originally isolated from each other only by soft tissue; (1) teeth set in isolated alveoli.

Transverse section of posterior maxillary teeth [B116m; O104m; S52m; Pb140m]: (0) strong lateral compression; (1) transverse section circular to subcircular, without significant lateral compression; (2) transverse section ‘teardrop-like’ ( ¼ triangular), with asymmetric lateral compression occurring on the distal margin.

Relative position of the last maxillary tooth [O18m; M37]: (0) last tooth in anterior to the anteriormost border of the suborbital fenestra; (1) last tooth positioned posteriorly to the anteriormost border of the suborbital fenestra.

Implantation of posterior teeth at maxilla and dentary [O18m; M38–39m]: (0) teeth set disposed in a groove, the roots originally isolated from each other only by soft tissue; (1) teeth set in isolated alveoli.

Occurrence of abrasion surfaces in dentary and maxillary teeth [A132]: (0) teeth without abrasion surfaces due to masticatory action; (1) anteroposterior wear surface, indicating capacity for proal/propalinal movements of the mandible; (2) diagonal wear surface, indicating capacity for lateral movements of the mandible.

Number of dentary teeth [A133]: (0) no more than 10; (1) at least 11.

Orientation of the anterior dentary teeth [A134]: (0) vertical or subvertical; (1) mildly procumbent, anteriorly inclined; (2) strongly procumbent, anteriorly inclined, the first pair of teeth almost horizontal.

Dentary symphyseal teeth battery [new]: (0) absent; (1) present, teeth from each pair closer to each other than to other teeth in the same hemimandible.

(142) Length of the dentary teeth occluding at the maxillary/premaxillary contact [C80; S54]: (0) small to medium sized, subequal to other surrounding teeth; (1) hypertrophied, at least twice longer than surrounding teeth.

(143) Middle and posterior dentary teeth (posterior to the maxillary/premaxillary suture) [C81m; O20m; S55m; Pb81m]: (0) gradually bigger and than smaller, the same trait occurring with the occluding teeth at maxilla; (1) diminishing posteriorly; (2) gradually bigger and than smaller, the opposite occurring with the occluding teeth at maxilla.

(144) Implantation of the middle and posterior dentary teeth [new]: (0) not oblique or slightly altered; (1) oblique (more than 30 degrees).

(145) Implantation of middle dentary teeth [O18m; M39m]: (0) teeth set disposed in a groove, the roots originally isolated from each other only by soft tissue; (1) teeth set in isolated alveoli.

(146) Transverse section of middle and posterior dentary teeth [B116m; O104m; S52m; Pb140m]: (0) strong lateral compression; (1) transverse section circular to subcircular, without significant lateral compression; (2) transverse section ‘teardrop-like’ ( ¼ triangular) to lozenge shaped, with asymmetric lateral compression occurring mostly on the anterior margin.

(147) Constriction at the crown-root transition, in posterior maxillary and middle/posterior dentary teeth [B117; S50; Pb157m]: (0) absent; (1) present.

(148) Number of cusps of posterior teeth [G46m; B113m; Pb162m]: (0) single apical cusp; (1) multicusped teeth, with two or more cusps.

(149) Lingual cingulus at the base of the crown of the middle and posterior teeth [A145]: (0) absent; (1) present, with accessory cusps and styli.

Axial elements (16 characters; 8.75% of total) (150) Cervical vertebrae [C92; S57; P92]: (0) amphiplatyc or amphicoelic; (1) procoelic.

(151) Axis centrum, length [A147]: (0) centrum short, as long as high; (1) centrum clearly longer than higher.

(152) Axis, neural spine laminae [Pb152m]: (0) slightly developed, limited to the posterior half of the neural arch; (1) well developed over the whole extension of the neural arch due to the presence of prespinal and postspinal laminae, extending both anteriorly and posteriorly to the neural channel.

(153) Anterior cervical vertebrae, neural spine laminae [C90m; Pb90m]: (0) prespinal and postspinal laminae present; (1) laminae absent, neural spine rod-shaped.

(154) Anterior cervical vertebrae, structure of neural spine [A151]: (0) base narrow, gracile; (1) base short, wide, robust.

(155) Third cervical vertebra (CIII), development of prezygapophysis [A152]: (0) slightly developed, projecting slightly anterior to the centrum; (1) well developed, clearly projecting anterior to the centrum.

Historical Biology 135 (156) Posterior cervical vertebrae, neural spine laminae [C90m; Pb90m]: (0) prespinal and postspinal laminae present; (1) laminae absent, neural spine rod-shaped.

(157) Posterior cervical vertebrae, structure of the neural spine [A153]: (0) base narrow, gracile; (1) base short, wide, robust.

(158) Posterior cervical vertebrae, development of the hypapophysis [Pb91m; W37m]: (0) absent or slightly developed, no more than a sagittal ridge in the anterior portion of the centrum ventral surface; (1) present, well developed, laminar shaft projecting ventrally from the centrum anteroventral surface.

(159) Dorsal vertebrae, development of the hypapophysis [A155]: (0) absent or slightly developed, no more than a sagittal ridge in the anterior portion of the centrum ventral surface; (1) present, well developed, laminar shaft projecting ventrally from the centrum anteroventral surface.

(160) Dorsal vertebra [C93; S58; Pb93]: (0) amphiplatyc or amphicoelic; (1) procoelic.

(161) Caudal vertebrae [C94m; Pb94m]: (0) amphiplatyc or amphicoelic; (1) procoelic, the first vertebra eventually biconvex.

(162) Paramedian dorsal osteoderms (trunk) [C95; S61; Pb95]: (0) wider than longer, rectangular; (1) as long as wide, with variable shape (square to subcircular); (2) longer than wider, elliptical.

(163) Anterolateral process developed at the anterior border of dorsal osteoderms [C96; S62; Pb96]: (0) absent; (1) present.

(164) Number of osteoderm dorsal rows (trunk) [C97; S63; Pb97m]: (0) two paramedial rows; (1) four rows, two paramedial and two accessory.

(165) Accessory osteoderms (trunk) [Pb97m]: (0) absent; (1) present.

(166) Ventral osteoderms (trunk) [C100; S66; Pb100]: (0) absent; (1) present.

Appendicular elements (16 characters; 8.75% of total) (167) Anterior surface of scapula [C82; O120m; Pb82]: (0) curved; (1) straight.

(168) Coracoid length, proportional to the scapula [C83m; S59m; Pb83]: (0) much smaller, no more than half the length of the scapula; (1) smaller, approximately 60–75% of the length of the scapula; (2) subequal.

(169) Glenoid surface of coracoid extended on an oblique plane and the glenoid tip facing outwards and posteroventrally [O122m]: (0) absent; (1) present.

(170) Styliform process of coracoid [O118]: (0) absent; (1) present.

(171) Proximal head of humerus [O123]: (0) facing backwards, posterodorsally; (1) facing dorsally, with a lateromedial major axis.

(172) Internal tuberosity at the proximal articulation of the humerus [O124]: (0) slightly developed, with the articular surface dorsally oriented; (1) well developed, with articular facet ventral or oblique.

(173) Ligamentary depression at the surface of humerus

(174) (175) (176) (177)

(178)

(179)

(180) (181)

(182) (183)

[O125]: (0) lateral to the internal tuberosity and below the proximal articulation of the humerus; (1) located laterally to the articulation of the proximal end of humerus.

Humerus, lateral aspect of the deltapectoral crest [O126]: (0) convex; (1) concave.

Ulna, lateral compression [O168]: (0) absent; (1) present at least at the distal end.

Radial [O127]: (0) longer than wider; (1) length subequal to width.

Radiale, proximal end [Pb117]: (0) symmetric, similar to distal articulation; (1) asymmetric, mesial exposure more representative than lateral.

Ilium, proportional length between the preacetabular and postacetabular processes [C84; S60; Pb84]: (0) subequal; (1) postacetabular process clearly longer (approximately four times longer).

Ilium, orientation of the postacetabular process [W41; Pb110]: (0) posteriorly or posteroventrally directed; (1) posterodorsally directed, positioned well above the preacetabular process.

Ilium, presence of the supracetabular crest [Pb116]: (0) absent; (1) present.

Femur torsion [O149]: (0) femur with light torsion, the difference in the orientation between the proximal and distal articulation facets approximately equals to 30 degrees; (1) femur with evident torsion, the difference in the orientation between the proximal and distal articulation facets approximately equals to 60 degrees.

Femur, position of the 4th trochanter [A178]: (0) anteromedial; (1) posteromedial.

Tibia, proximal end [O87]: (0) single concavity; (1) medial crest separating two concavities.

Appendix D. Matrix used in the analysis 21 terminals included, presented in alphabetical order, after the outgroups. 183 characters for each taxon. Characters grouped, with periods (‘.’) indicating clusters of 10; periods not originally included in the matrix. Coding varies between 0 – 5; ‘?’ ¼ missing entries; ‘-’ ¼ inapplicable characters; ‘{}’ indicating variable condition of a character within the terminal.

Outgroup 1 – Sphenosuchia 0001101220.0101010020.021110?100.02120101010011000010.

0000001000.0000111000.1021111100000??0000-.?0000?0000.?

00?100100.00?001011?11?24?00121101011010.0010100000.1100 100000.?210000001.0100001000.001

Outgroup 2 – Protosuchia 0000101220.1?100??001.{01}{12}1110?000.01{01}201{01}001.

0010000010.0010001000.1?{01}0111000.1021110002.000100000.?000010000.?00?100100.00?0010110.????4??010.11010{01}1010 .0110100000.??00?00000.?010010001.0100001000.001 Alligatoridae 1111212{01}01.0210010030.1211111{01}{01}0.12000??010.1221 101101.1101110011.0011000010.0100000001.1110001101.01011

136 M.B. de Andrade and R.J. Bertini 01021.0100101110.1000200101.10{12}3310011.0101110011.01?0 11{01}001.1100111111.11011{01}1210.1011110110.110 Anatosuchus 0101210020.01000??002.0111110000.0000010010.0111101?0?.

0001????1?.001?001000.1?00?1001?.111000?001.10000?1011.110 0012100.00?110011?.10?34?1012.0101?110?0.00?01??00?.?????? ????.??????????.??????????.???

Araripesuchus gomesii 1100111120.1101010031.1211111010.0210110010.0111101000.1?

01100110.0011001000.1100?1000?.1110001001.1000001011.1101 111110.00?1201011.11024?0012.010111?010.00?0110000.?1?????

100.0010000200.01?01??1??.10?

Araripesuchus patagonicus 1100211120.1?01010021.12101??110.0210110010.0111101000.??

01100110.0011001000.1100100001.1110001001.1000001011.010 1?11110.00?1101011.11034?0012.010111?010.00?0110000.????? ????0.?0100?0??0.01001????1.1??

Baurusuchidae 0001002101.1?1101002?.?1111?1101.0101112111.0111{01}1100?.

0011100110.1101002100.1121111102.1110001?10.2011111111.11 11011001.01?1101121.1103210002.0201101000.0120110000.???? ?????0.01?00??2?0.????11????.???

Bernissartia 1111212001.0211?10030.1??1111??0.0?0000?010.12?1101?0?.110 11100?1.0011????10.0100?????1.1110001??1.??01101?21.01001 11110.10???00100.0013310011.0101?1?01?.0?201?0000.??00?11 110.1001110?10.1011???1??.11?

Candidodon 0000110010.1010001112.?00??????0.0010010010.011?00100?.??

0110?110.??0?00??0?.1?11?1??10.111?0110??.10000?1011.1101? ?1???.0?????????.000321?003.0101011???.??????101?.?????????

0.?1????????.????1?????.???

Chimaerasuchus 0??0100010.0111001101.0011??????.??????????.?????1????.?? ????????.??????????.?????????2.01?1??????.??????????.???100??0 ?.???1?010{12}1.1101100005.11011?110?.??1011?100.0011?1110 ?.??0????10?.??????1101.???

Comahuesuchus 0010210000.0110?01122.01010??100.00?1?1?021.0100111?1?.000 1??????.??????2?00.112111110?.1110001110.201??1?011.1??111 2101.01?1??????.010?4??002.01?100?011.001?0??00?.??????????. ??????????.??????????.???

Crocodylidae 1111212{01}01.0211010030.?211111100.020000?010.1221101{01}0 1.1101110011.0010000010.0100000001.1110001101.0101101021.010 0101110.1000200101.10{12}3310011.0101110011.0120110001.11001 11111.1101101210.1011110110.110

Malawisuchus 0000110110.0110101112.?210110000.02?0110010.00101000?0.??0 110?110.00010?2?00.101111?0??.111101100?.10000?1011.0001 011101.00?1001021.1{01}03210003.01010011?0.??10101110.??1 1?11100.020??10???.????1?????.00?

Mariliasuchus 0100110000.0200101122.0?01010101.0111100021.0100101110.

0001101110.0001102100.111111101?.1110101010.2010011011.12 01011201.01110?1021.01122{01}0104.0010210102.001002100?.0 0110110??.??????????.????1?????.???

Notosuchus 0100100020.1100?01122.0211010101.0111111011.0111100100.

0001101110.1?01102100.1111111011.1110101010.2010011111.12 01011201.01?11?1021.01?3300104.0?10210101.0010020000.0011 011000.00000?0100.01001111?1.001 Sebecus 0001012101.0111010021.1210111100.0101002010.0211101010.

1110101010.0011001000.11101?00?1.1111001101.10011?1011.11 010111?0.0000200101.11034?0011.0201001011.002010000?.???? ??????.??????????.??????????.???

Simosuchus 0000211221.1200001112.2211110000.111110?021.021100?000.

0001101110.1110002100.1011111101.1101001001.?0?0?110??.??

00012001.01?1200121.00034?0015.0101001010.0010101100.??11 ?11100.?2?101????.??????????.???

Sphagesaurus huenei 0000110000.1?00001112.?????10???.0??????0??.?001010000.00 01101110.0001????01.112??????2.11100010?0.201???1?11.12?10 ?????.0?????????.0131011104.101121?2?1.1001?2100?.?0??????? ?.??????????.??????????.???

Sphagesaurus montealtensis 0000110010.1100001112.0211110101.0111111011.0001010110.

0001101110.1001?0??01.112111110?.1111001010.2010011011.12 11011201.01111?????.0130001104.1?1?21?201.1001?2?00?.???? ??????.??????????.??????????.???

Thalattosuchia 0001112121.0200010010.021110?100.0112012100.0011000010.

0010001000.000000{12}000.0010101101.11100000??.10000?101 1.1001110110.00???10111.11?14?0010.0101111011.0010110000. ?100?00000.001001?2??.??????????.???

Uruguaysuchus 1000110020.1210110?22.00011??100.0110010010.?01000?000.??0 1101110.??1?002000.1?11111101.111?00?001.1000011011.11010 111?1.00????0101.10032100?3.0?0?00?010.?0?0?0110?.????????0 0.010?????00.101?1?0101.001